Index

03-16 Osteoarthritis Research Society International (2008)

Does the hip powder of Rosa canina (rosehip) reduce pain in
osteoarthritis patients?

17-20 PhytotheraPy research 20 (2006):

The evidence for clinical efficacy of rose hip and seed:
a sysTemaTic review

20-27 scand. J rheumatol (2005):

a powder made from seeds and shells of a rose-hip
subspecies (rosa canina) reduces sympToms of knee
and hip osTeoarThriTis:
a randomized, double-blind, placebo-conTrolled
clinical Trial

27-35 PhytomedIcIne 11 (2004):

a herbal remedy, hyben viTal (sTand. powder of a
subspecies of rosa canina fruiTs), reduces pain and
improves general wellbeing in paTienTs wiTh
osTeoarThriTis—a double-blind, placebo-conTrolled,
randomised Trial

36-41 current theraPeutIc research Vol. 64 (2003):

The effecTs of a sTandardized herbal remedymade
from a subType of rosa canina in paTienTs wiTh
osTeoarThriTis: a double-blind, randomized,
placebo-conTrolled clinical Trial

42-45 InflammoPharmacology Vol. 7 (1999):

The anTi-inflammaTory properTies of rose-hip

45-50 InflammoPharmacology Vol. 7 (1999):

rose hip inhibiTs chemoTaxis and chemiluminescence of
human peripheral blood neuTrophils in viTro and
reduces cerTain inflammaTory parameTers in vivo

51 oVerVIeW on aBstracts
 03

Does the hip powder of Rosa canina

(rosehip) reduce pain in osteoarthritis
patients?
Review
Does the hip powder of Rosa canina (rosehip) reduce pain in osteoarthritis patients? - a meta-
analysis of randomized controlled trials R. Christensen M.Sc.y, E. M. Bartels Ph.D., D.Sc.yz, R. D. Altman
M.D., Professorx, A. Astrup M.D., Ph.D., Professork and H. Bliddal M.D., Ph.D.y*
The Parker Institute, Musculoskeletal Statistics Unit, Frederiksberg Hospital, Denmark, Copenhagen
University Library, Denmark. David Geffen School of Medicine, University of California, Los Angeles,
CA 90024, USA. Department of Human Nutrition, Faculty of Life Sciences, University of Copenhagen,
Denmark

Summary
Objective: Meta-analysis of randomized controlled trials (RCTs) - of a hip powder of Rosa canina
(rosehip) preparation for symptomatic treatment of osteoarthritis (OA), in order to estimate the em-
pirical efficacy as a pain reducing compound.

Method:
RCTs from systematic searches were included if they explicitly stated that OA patients were random-
ized to either rosehip or placebo. The primary outcome was reduction in pain calculated as effect
size (ES), defined as the standardized mean difference (SMD). As secondary analysis the number of
responders to therapy was analyzed as Odds Ratios (OR), and expressed as the Number Needed
to Treat (NNT). Restricted Maximum Likelihood (REML) methods were applied for the meta-analyses
using mixed effects models.

Results:
The three studies (287 patients and a median trial-duration of 3 months) e all supported by the man-
ufacturer (Hyben-Vital International) e showed a reduction in pain scores by rosehip powder (145
patients) compared to placebo (142 patients): ES of 0.37 [95% confidence interval (CI): 0.13e0.60],
P = 0.002. Test for homogeneity seemed to support that the efficacy was consistent across trials (I 2 =
0%). Thus it seems reasonable to assume that the three studies were measuring the same overall ef-
fect. It seemed twice as likely that a patient allocated to rosehip powder would respond to therapy,
compared to placebo (OR = 2.19; P = 0.0009); corresponding to a NNT of six (95% CI: 4-13)
patients.

Conclusions: Although based on a sparse amount of data, the results of the present meta-analysis
indicate that rosehip powder does reduce pain; accordingly it may be of interest as a nutraceuti-
cal, although its efficacy and safety need evaluation and independent replication in a future large-
scale/long-term trial. ® 2008 Osteoarthritis Research Society International. Published by Elsevier Ltd.
All rights reserved.

Key words: Rosa canina, Rosehip, Meta-analysis, Osteoarthritis, Dietary supplements, Knee, Hip, Herbal therapy.

Introduction tive protein (CRP)5. Drug therapy in OA consists

Osteoarthritis (OA) is a common joint disor-
der and may occur in any synovial joint in the
body, although the condition is most common
in hands, knees, hips and spine1.The clinical
problems, along with the pathological and ra-
diographic changes, include joint pain, stiffness,
movement with a restricted range and cracking
of joints (crepitus)2. OA has traditionally been re-
garded as a non-inflammatory condition3, but
improved detection methods show that inflam-
matory pathways are up-regulated in OA4; with,
e.g., a low-level increase by groups in C-reac-
mainly of analgesics and non-steroidal anti-in-
flammatory drugs (NSAIDs). Paracetamol is the
oral analgesic of first choice, and if successful,
the preferred long-term oral analgesic. Howev-
er, NSAIDs must be considered in patients with
no response
*Address correspondence and reprint requests to: Professor H.
Bliddal, The Parker Institute, Frederiksberg Hospital, DK-2000 F,
Denmark. Tel: 45-38164151; Fax: 45-38164159; E-mail: henning.
bliddal@frh.regionh.dk Received 21 November 2007; revision
accepted 2 March 2008. to paracetamol6,7. Disease-modifying to
paracetamol6,7.
04
Disease-modifying OA drug (DMOAD)-therapy
remains to be developed in order to slow down
disease progression as demonstrated by, e.g., a
reduced joint space narrowing on plain X-rays8.
According to the consensus statement follow-
ing the outcome measures in rheumatology
(OMERACT) III conference, a core set of out-
come measures for phase III clinical trials point-
ed towards four variables which should be eval-
uated in trials with patients suffering from either
knee, hip, or hand OA: pain, physical function
(i.e., disability), patient global assessment; and
e for studies of 1 year or longer e joint imaging9.
It is mandatory to perform continuous follow-up
on clinical interventions, which are assessed on
scales typically referred to as subjective10,11.
A standardized hip powder of Rosa canina
made from the seeds and husks of the fruits from
a subtype of R. canina hip powder (i.e., rose-
hip), the common wild-briar hedgerow rose, has
been evaluated in (short-term) randomized
controlled trials (RCTs)12. According to the best-
evidence synthesis, there are contradictory
results with regard to scientific evidence for R.
canina extracts13. Evidence from early in vitro
studies indicates that R. canina hip powder
preparations exert anti-inflammatory properties
via reduced chemotaxis of peripheral blood
neutrophils and monocytes in healthy subjects,
and a reduction in CRP is seen after 4 weeks
supplementation in patients with OA14,15. The
proposed mechanism of action has been fo-
cused on the preparations’ anti-oxidative ca-
pacities, and a specific galactolipid (called
GOPO) has been identified (in vitro) as antiin-
flammatory, and as such possibly the reason for
the preparation’s proposed pain reducing prop-
erty16. These considerations over active ingredi-
ents have recently been confirmed by others,
as extracts of R. canina fruits have shown potent
anti-inflammatory and anti-nociceptive activi-
ties 17 and R. canina hip powder extracts (an or-
ganic solvent) may inhibit both cyclooxygenase
(COX)-1 and -218. In a clinical trial it has been
shown that R. canina hip powder may have
some efficacy in hip and knee OA patients19.
In the present systematic review on clinical ef-
ficacy of giving a R. canina hip powder prepa-
ration for symptomatic treatment of OA, with
explicit meta-analysis of the available RCTs20 -
our primary aim was to obtain up-to-date, ev-
idence-based estimates that could provide a
detailed view of the symptomatic efficacy of
R. canina compounds used in the treatment of
OA. The results of this analysis may be crucial
for the evaluation whether or not these prepa-
rations will be relevant for future large-scale (i.e.,
phase III) clinical trials.
Materials and methods
Study selection, assessment of eligibility criteria, data extraction,
and statistical analysis were performed based on a predefined
protocol according to the Cochrane Collaboration guidelines
(http://www.cochrane.org/resources/ handbook/index.htm).
RETRIEVAL OF PUBLISHED STUDIES
RCTs of R. canina hip powder treatment vs pla-
cebo were identified through a systematic lit-
erature search in the following bibliographic
databases: Medline via PubMed (mid 1950s
to October 2007), EMBASE via WebSpirs (1980
to October 2007), CINAHL via WebSpirs (1982
to October 2007), Biosis Previews via WebSpirs
(1980 to October 2007), Web of Science (1945-
54 to October 2007), Scifinder (1907 to October
2007), Scopus (1966 to October 2007), and the
Cochrane Library from 1966 to October 2007.
Following the searches, reference lists of original
reports and review articles retrieved through the
described searches, were thoroughly checked
for further relevant studies. Finally, we searched
conference abstracts over the past 2 years via
the established international societies of rheu-
matology, i.e., the OsteoArthritis Research So-
ciety International (OARSI), EUropean League
Against Rheumatism (EULAR) and the Ameri-
can College of Rheumatology (ACR). Since the
available studies were expected to be few,
a broad, less specific search strategy was ap-
plied: (Rosehip OR ‘‘Rose hip’’ OR ‘‘Rosa canina’’
OR ‘‘dog rose’’ OR Rosaceae OR Litozin OR hy-
ben OR GOPO OR i-flex) AND (OA OR osteoar-
throsis). Controlled, randomized and clinical trials
were deliberately not part of an explicit search
strategy, since we wished to find any work deal-
ing with R. canina hip powder in the treatment
of OA. With the awareness of a higher propor-
tion of noise in the chosen searches, full refer-
ences were reviewed for possible RCTs, and full
text references were obtained for further scru-
tiny, where relevant.
INCLUSION AND EXCLUSION CRITERIA
We included RCTs comparing a preparation
containing R. canina hip powder with a pla-
cebo intervention. Studies were selected if the
included patients were (explicitly) described
as having clinical or radiographic evidence of
OA. Two reviewers (RDA, HB) crosschecked and
agreed on diagnostic criteria in each trial. We
excluded studies in conditions such as non-OA
joint pain, rheumatoid arthritis (RA), pain due to
surgery or injury, and studies with mixed patient
groups such as those with both OA and RA, un-
less the subgroup data for OA were available.
No language restrictions applied.
QUALITY ASSESSMENT
The quality of studies was assessed based on
randomization, masking and withdrawal. The
complete reports of the RCTs that were

selected for inclusion in the meta-analysis were
scored by two reviewers for quality (RC, EMB)
using a validated instrument21. The score was
given as follows: if the study was described as
randomized (þ1); if the study was described as
double masked (þ1); if there was a (detailed)
description of withdrawals and attrition rates/
detailed outcome data and theanalysis was
performed according to the intention-to-treat
(ITT) principle (þ1). In addition, if the random al-
location and the double blinding were properly
described and appropriately put into practice,
each item received 1 point extra. Conversely, if
the methods (randomization and
masking) were not considered appropriate, 1
point was subtracted from each item.
DATA
EXTRACTION
AND OUTCOME
MEASURES
Two reviewers (RC, EMB) undertook data ex-
traction independently. Disagreements were re-
solved by discussion. A customized form was
used to record the following: authors of the study,
year of publication, trial design, study length,
number of patients randomized (i.e., the ITT
population, Ntotal), the number of patients for
whom detailed outcome data was available
for meta-analysis in each group (E/exposed = R.
canina´hip powder and C/control = placebo)
included in the individual-study statistical tests
(NE and NC, respectively), average patient age,
sex, site of OA. Note that in order to estimate the
relative number of responders to therapy, we in-
cluded the ITT population (based on the NEITT
and NC-ITT, respectively) in the denominator. The
number of responders per se, was assessed as
the number of patients in each trial defined
by the authors as being a responder; the num-
ber of responders in both the R. canina hip pow-
der and placebo group were based on the
same criterion11. As it seemed relevant to con-
sider the available efficacy in cross-over trials
as being subjected to carry-over bias19, we only
report (i.e., include) data from the first period.
The primary outcome measure was the mag-
nitude of pain reduction22. The secondary out-
comes were the reported changes in the aver-
age level of applied painkillers; the extracted
(or estimated) reported number of respond-
ers per group following intervention11. Disabil-
ity and patient’s global assessment following
05
therapy9 were not included as outcomes in the
present meta-analysis, since we expected that
these endpoints would not have been reported
consistently.
STATISTICAL ANALYSIS
As a preliminary review of the available data23
supported the notion that the available cross-
over trials had been reporting carry-over bias24,
we chose to include only data from the first
period, as any pooled efficacy meta-analysis
including data from both periods would imply
a risk of (accumulating) carry-over bias25,26. For
each of the continuous outcomes (i.e., pain and
rescue medications), we calculated the test sta-
tistics based on the available data, using stan-
dard formulae26,27. Based on these statistics and
the number of observations in each group, we
were able to estimate the standardized mean
difference (SMD) for each study28 - which was
applied as effect size (ES)29. The corresponding
variance (SE2) was calculated based on the
individual study SMD and the number of pa-
tients included (SE2 = 1/NE + 1/NC + SMD2/[2{NE
+ NC}])28. As the unadjusted (Cohen’s) SMD in
principle does not treat the variance (SE2) as an
estimate, we applied (i.e., via multiplication) the
Hedges’ bias-correction (J = 1 - 3/[4  x df - 1]; i.e.,
df = NE + NC - 2) by default - adjusting for small
sample bias30. SMDs were signed so that posi-
tive values (>0) indicated a benefit of R. canina
hip powder: clinically, |ES| ≥ 0.2 is considered
small,|ES|≥ 0.5 is moderate (and would prob-
ably be recognized clinically31), and|ES| ≥ 0.8 is
large6,7,32,33. The Odds Ratio (OR) was estimated
for the dichotomous efficacy data (i.e.,
responders to therapy)34. To combine the indi-
vidual study results we did (generic inverse vari-
ance) meta-analyses via mixed effects model
procedures using SAS software (version 9.1.3,
by SAS Institute Inc., Cary, NC, USA)35. We ap-
plied the restricted maximum likelihood (REML)
method36,37 to estimate the between study vari-
ance and the combined efficacy28,38. The het-
erogeneity (between trials) was examined with
a standard Q test (testing the hypothesis of ho-
mogeneity: x2(k-1) )39.
However, as measures of the extent of hetero-
geneity might be considered preferable to test
of its presence, we evaluated possible inconsis-
tency between effect measures via the I2 sta-
tistic40 - which can be interpreted as the per-
centage of variability in effect estimates due to
heterogeneity41. As it is often sensible to use one
statistic for meta-analysis and re-express the re-
sults using a second more easily interpretable
statistic26, we estimated the Number Needed to
Treat (NNT), with 95% confidence intervals
(CI) on the basis of the combined OR value42,
since this method enables direct translation into
06
clinical practice 10,43,44; applying the overall event
rate in the placebo group as a proxy for base-
line risk45,46. The software ‘Visual Rx’’ is designed
to calculate NNT (and NNH) from the pooled re-
sults of a meta-analysis and produce a graphi-
cal graphical display of the result47:
http://www.nntonline.net/ebm/visualrx/try.
asp48.
Results
CHARACTERISTICS OF TRIALS
The Quality of Reporting of Meta-analyses
(QUOROM)-recommended flowchart20 in
Fig. 1 displays the eligibilitydetails of the stud-
ies identified by the combined search
strategy. Studies with clearly irrelevant ob-
jectives/designsas well as abstracts and re-
views/theme articles, were separated from
possible studies for inclusion: initially the
search strategy revealed 37 potential refer-
ences, which were considered at abstract
level. When removing obviously residual lit-
erature and abstracts later reported in full,
we retrieved 15 studies for further scrutiny -
including critical assessment of the reported
references12-19,49-55.
Among these, five papers were excluded as
a consequence of being reviews12,13,50,54,55;
one study only considered intestinal mi-
croflora in patients with irritable bowel syn-
drome49; three studies were categorized
as in vitro16-18. Among the remaining six po-
tentially relevant studies14,15,19,51-53 two were
excluded as a consequence of being con-
trolled trials, reported as case-control trials
with explicit focus on in vitro-inflammatory
properties14, and inhibition of chemotaxis
and chemiluminescence15, respectively. This
left four trials19,51-53 potentially relevant for in-
clusion in the meta-analysis23. However, fol-
lowing personal contact with Dr Winther
and Dr Rein, it appeared that the patent
registration from Rein et al.51 was based
on an unpublished subgroup-analysis of the
Norwegian study52 and was, quote: ‘‘a re-
hash of a another study’’. Accordingly, we
were able to include three (assumed) mutu-
ally independent RCTs19,52,53.
Table I shows the baseline characteristics of
the included studies. All trials were support-
ed by Hyben-Vital International (Tullebølle,
Langeland, Denmark): one study was per-
formed in an outpatient clinic in Norway52,
while the two others included patients (from
outpatient clinics) in Denmark19,53. Overall,
the trials randomized 306 OA patients to ei-
ther R. canina hip powder or placebo, allo-
cating 153 patients to each group. The Dan-
ish trials19,53 applied a cross-over design, and
excluded patients with other rheumatic dis-
eases than OA, and those who received glu-
cosamine or intra-articular glucocorticoids
6 weeks prior to the study. The Norwegian
study52 included OA patients with pain for
at least 6 months, who were on a waiting list
for either hip or knee surgery, or on a list for
final evaluation for surgery. As presented in
Table I the majority of the participating pa-
tients were women (62%) suffering from knee
OA (61%) with a median age of 66 years.
PAIN REDUCTION
As presented in Fig. 2(A): the meta-analysis
of the three studies reporting changes in
pain scores produced a statistically signifi-
cant (P = 0.0019) combined ES of 0.37 (95%
CI: 0.13-0.60) - favoring R. canina hip powder
compared to placebo. Test for homogene-
ity seemed to support that the efficacy was
consistent across trials (Q = 0.18; /2 = 0%).
Thus, it seems reasonable to assume that the
three (mutually independent) studies mea-
sured the same overall effect. Apparently
the pain reducing property of R. canina hip
powder seemed more pronounced in the
population examined in the study by War-
holm et al.52, which included patients who
were on a waiting list for either hip or knee
surgery, or on a list for final evaluation for sur-
gery.
USE OF RESCUE MEDICATION
As presented in Fig. 2(B): the meta-analysis
of the three studies reporting changes in the
use of ‘rescue medication’ produced a sta-
tistically significant (P = 0.018)
combined ES of 0.28 (95% CI: 0.05e0.51) - fa-
voring R. canina hip powder compared to
placebo. Test for homogeneity seemed to
support that the efficacy was consistent
across trials (Q = 1.25; I 2 = 0%). Thus it seems
reasonable to assume that the three
(mutually independent) studies measured
the same overall effect. Apparently R. cani-
na hip powder did not reduce the patients’
consumption of painkillers in the population
reported by Warholm et al. (i.e., ES < 0.2)52,
 and was, quote: ‘‘a rehash of a another study’’. Accordingly,
we were able to include three (assumed) mutually indepen-
dent RCTs19,52,53.
Table I shows the baseline characteristics of the included
studies. All trials were supported by Hyben-Vital Interna-
tional (Tullebølle, Langeland, Denmark): one study was
52
while based
performed in anonoutpatient
the diaries
clinic inof the consump-
Norway , while the
two others included patients (from outpatient clinics) in
tion of ‘rescue
Denmark 19,53 medication’
. Overall, investigated
the trials randomized 306 OAin
the study by Winther et al.19 - the use
patients to either R. canina hip powder or placebo, of R.
allocat-
ing 153 patients to each group. The Danish trials19,53
applied a cross-over design, and excluded patients with
Potentially relevant studies identified and
screened for retrieval (k0 = 15)
Trials retrieved for more detailed evaluation
(k1 = 6)
Potentially appropriate RCTs
to be included in systematic review
(k2 = 4)
RCTs included in pooling
(k = 3)
Fig. 1. Flow chart of the search strategy and selection of trials.
NUMBER OF RESPONDERS TO THERAPY
In order to assess this secondary outcome,
the arbitrary ‘‘responders to therapy’’, the fol-
lowing data was extracted: (1) Warholm et
al. used a simple yes-or-no questionnaire
(about relief of pain) after 4 months thera-
py in both groups52 (R. canina hip powder:
31/50 vs placebo: 21/50); (2) Rein et al. de-
fined a responder as one who showed at
least one category of pain improvement53
(R. canina hip powder: 31/56 vs placebo:
18/56); (3) Winther et al. used any reduction
in western ontario and mcmaster (WOMAC)
score for joint pain after the initial 3 weeks of
treatment as a response criterion e however,
they did not report any explicit numbers
following 3 months treatment19. We assessed
the number of responders (any reduction in
WOMAC pain) in each group following a
Monte-Carlo simulation based on the report-
ed means and standard deviations (SDs)
(i.e., table 319) assuming that a univariate
normal distribution apply56 (R. canina hip
powder: 32/47 vs placebo: 26/47). As present-
ed in Fig. 2(C): the meta-analysis of the
studies reporting the number of patients re-
sponding to therapy as a dichotomized (yes/
no) count, produced a statistically
studies reporting changes in the use of ‘rescue medica-
tion’ produced a statistically significant (P ¼ 0.018)
combined ES of 0.28 (95% CI: 0.05e0.51) e favoring
R. canina hip powder compared to placebo. Test for
homogeneity seemed to support that the efficacy was con- 07
sistent across trials (Q ¼ 1.25; I 2 ¼ 0%). Thus it seems
canina
reasonablehip powder
to assume thatresulted in a significantly
the three (mutually indepen-
dent) studies measured the same overall effect. Appar-
reduced
ently R. caninausehip of analgesics,
powder compared
did not reduce the patients’ to
placebo.
consumption of painkillers in the population reported by
Warholm et al. (i.e., ES < 0.2)52, while based on the dia-
ries of the consumption of ‘rescue medication’ investigated
Studies excluded because of:
• Review article (5)
• In vitro study (3)
• Patients not diagnosed with OA (1)
Studies excluded because of:
• Non-randomized (case-control) study design (2)
Trial excluded as:
• Patent registration/Subgroup analysis (1)
highly significant (P = 0.00089) combined
OR of 2.19 (95% CI: 1.38 - 3.48) - favoring R.
canina hip powder compared to placebo;
i.e., it is more than twice as likely that a pa-
tient allocated to R. canina hip powder will
respond to therapy, compared to placebo.
Test for homogeneity seemed to support
that the observed efficacy was consistent
across trials (Q = 0.52;/2 = 0%), supporting the
assumption that the three (mutually inde-
pendent) studies were measuring the same
overall effect. In absolute terms: the total
number of responders (across the three tri-
als) on R. canina hip powder and placebo
was 94/153 (61.4%) and 65/153 (42.5%), re-
spectively. On the basis of the average
number of responders within the placebo
groups, the combined OR corresponded to
a NNT of six (95% CI: 4 -13) patients.
08 968 R. Christensen et al.: Does the hip po

ADVERSE EVENTS in the

Data are number (%) or mean � SD. QS: Jadad Quality Score (range: 0e5); PG and CO indicate Parallel-Group and Cross-Over Trial design, respectively. BMI: body-mass index; K: knee
OA; H: hip OA; N: neck OA; S: shoulder OA; Ha: hand OA. NE and NC are the number of patients included in the analyses in the exposed and control groups (i.e., rosehip and placebo),
142
48

47

47
Definite sample

NC
powde

AND SAFETY
sics, c

size
CONSIDER-

145
NUMBE

48

50

47
NE
In o

ATIONS
‘‘respo
(1) W

BMI (kg/m2)

27.0 � n.a.
27.3 � 5.0
(abou

19e41]
[range:
(R. ca
Focusing on adverse events, there seemed

n.a.

n.a.
et al.
to be the same amount of mild cases of one c
31/56
gastrointestinal discomfort after intervention

Age (years)

65.2 � 11.1

67.0 � 11.7

65.6 � n.a.
tion in
vs control52. The same number of patients

Median:
joint p

38e92]
[range:
seemed to experience ‘acid regurgitation’ criterio

65.6
in both the study by Rein et al.53 and Winther bers
numb
et al.19: one case in each group (R. canina in eac

Joint affected

(61.4% KOA)
hip powder and placebo) - both leading

K: 59, H: 46,
N: 18, S: 14,

K: 58, H: 21,
K: 56, H: 44
Summary of baseline characteristics of all participants in the eligible trials
on th

(56% KOA)

(53% KOA)

(78% KOA)
to discontinuation. In the study by Winther (i.e., t

K&H: 15
Ha: 40

K: 188
et al.19, mild unwanted effects (reported as tion a
47). A
being non-significant) that did not cause studie
withdrawal, were explicitly reported; based apy a

65 (65.0%)

71 (63.4%)

54 (57.4%)

190 (62.1%)
Women (%)
on these data we re-calculated empirical cally h
OR-values with 95% (‘‘exact’’57) confidence (95%
pared
limits for these rare incident cases58: (1) ‘Fre- a pati
quent voiding’ [OR = 3.07 (0.24 - 162.65)]; therap
(2) ‘Diarrhea’ [OR = 1.00 (0.07 - 14.07)]; (3) seem

94
100

112

306
ITT
Table I

‘Constipation’ [OR = 2.02 (0.10e120.54)]; (4) tent a

sump
‘Short episode of mild urticaria’ [OR = 2.02 were

Mean: 3.3,
Median: 3,
(0.10 -120.51)].
(months)
Duration

the to

SD: 0.5
4

3
R. ca
Discussion and 6
age n
The main result of our analysis was a small to comb
moderate short-term efficacy of prepara-
32.7%

CI: 4e
Design

CO

CO
PG

PG

tions with R. canina hip powder with a small

but clinically relevant reduction of pain in
ADVER
OA patients. However, the available data
2 � 5 � 0.5 g/day ¼ 5 g/day

2 � 5 � 0.5 g/day ¼ 5 g/day

2 � 5 � 0.5 g/day ¼ 5 g/day

are sparse, since we had only three clinical Foc

same
trials evaluating the efficacy in 145 patients
Intervention

after i
after use of R. canina hip powder for 3-4 seem
Hyben-Vital ,

Hyben-Vital�,

months. One assumption that is prudent in

�

by Re
5 g Rosehip
powder/day

order to make statistical inference following (R. ca

LitoZin�,

contin
meta-analysis, is that the eligible studies in- effect
cluded can be assumed to be mutually withdr
respectively. n.a.: Data not available.

independent, which might be an issue within we re

Mean: 4.3,
Median: 4,

the context of clinical efficacy of R. canina confid

‘Frequ
SD: 0.5
QS

hip powder. Dr Rein had access to the origi-

4

[OR ¼
nal data from Warholm et al.52 - enabling (0.10e
a patent registration51 prior to the publica- 2.02 (
tion by Rein et al.53; recalling that Dr Winther
2003

2004

2005
Year

was the co-author on this paper53 before

Discu
the Winther et al. paper was published
Warholm

in 200519. However, meta-analyses are de-

Winther

The
Overall
Study

short-
Rein

pending on the international peer-review

der w
system, which has been applied
 9

in all of the included papers19,52,53. A com- available60. Our meta-analysis supports the
bined analysis (i.e., meta-analysis) of homo- conclusion previously stated by Chrubasik:
geneous results, quantifies the magnitude of ‘‘Moderate evidence exists for the use of a
clinical efficacy per se59. Thus, it seems pos- powder of the seeds and husks of a Rosa
sible that the empirical magnitude of clinical canina subspecies in patients suffering from
efficacy following use of R. canina hip pow- osteoarthritis’’12.
Osteoarthritis and CartilagetoVol.
der is comparable 16, No.
other 9
nutraceuticals 969

A Favors Placebo Favors Rosehip

Warholm (2003)
Rein (2004)
Winther (2005)
ES = 0.37
Combined (Pain)
2
= 0.18; I2 = 0%; Z = 3.10 (P = 0.0019) (0.13 to 0.60)

B
Warholm (2003)
Rein (2004)
Winther (2005)
ES = 0.28
Combined (Rescue medication)
2 2
= 1.25; I = 0%; Z = 2.37 (P = 0.018) (0.05 to 0.51)

-0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8

Effect Size (SMD)
C

Warholm (2003)

Rein (2004)

Winther (2005)

Combined (Responders)
χ 2 = 0.52; I2 = 0%; Z = 3.33 (P = 0.00089) OR = 2.19
(1.38 to 3.48)

0.1 1.0 10.0

Odds Ratio

Fig. 2. Efficacy of R. canina hip powder (i.e., Rosehip) compared to placebo in OA patients presented as SMDs and OR. (A) Change (i.e.,
reduction) in self-reported pain; (B) change (i.e., reduction) in self-reported use of analgesics; (C) the number of patients defined/estimated
as being a responder. Every square represents the individual study’s effect measure with 95% CI indicated by horizontal lines. Square sizes
are proportional to the precision of the estimate. The overall estimate from the meta-analysis and its CI are shown at the bottom of each
subplot (AeC), represented as a diamond. The center of the diamond represents the pooled point estimate, and its horizontal lines represent
the CI.

We are confident
OA patients. However, that the efficacy
the available estimate
data are sparse, sinceis tailed
We are interpretation,
confident that thealthough it seems
efficacy estimate is robustlikely
per
robust per se, as it is based on very consistent that a reduction in analgesics could have an
2
we had only three clinical trials evaluating the efficacy in se, as it is based on very consistent findings (I ¼ 0%) e
145 patients after use of R. canina hip powder for 3e4 thus, a new trial (of the same duration) would be63expected
findings (I 2 = 0%) - thus, a new trial (of the
months. One assumption that is prudent in order to make
impact on a major public health scale . When
to result in a similar magnitude of small to moderate clinical
same
statistical inference would
duration) followingbe expectedistothat
meta-analysis, result
the focusing
efficacy (ES on the In
� 0.4). explicit, although
an (assumed) average arbitrary
knee OA pop-
in a similar magnitude of small to moderate
eligible studies included can be assumed to be mutually outcome ‘responders to therapy’,
ulation 61,62
, this ES would correspond to a mean it reduction
seems
independent,
clinical which might
efficacy be anInissue
(ES 0.4). within the context
an (assumed) av- that an OR of 2.19 corresponding to anmm)
in the visual analog scale (VAS) for pain (0e100 ES ofof
of clinical efficacy of R. canina hip 61,62, powder. Dr Rein had 6 mm; i.e., approximately 10% pain reduction. Apparently
erage
access toknee OA data
the original population
from Warholm etthis al.52ES would
e enabling 0.43
the use- of
64
indicates
R. canina hip a powder
small toleadsmoderate
to a significant clinical
reduc-
correspond to a mean reduction in the visual
a patent registration 51
prior to the publication by Rein efficacy
tion in 6,7,33
the . When translated into the num-
use of rescue medication, corresponding to
53
analog scale
et al. ; recalling (VAS) for pain (0 -100 mm) of on
that Dr Winther was the co-author ber of patients who would need R. canina hip
a small clinical efficacy. This does not allow more detailed
this paper53 before the Winther et al. paper was published interpretation, although it seems likely that a reduction in
6inmm;
200519 i.e., approximately
. However, meta-analyses 10% arepain reduction.
depending on the powder
analgesicstherapycould have(compared
an impact on ato majorplacebo)
public healthin
Apparently the use system,
international peer-review of R. caninawhich haship beenpowder
applied order
scale63to ‘‘treat’
. When ’ one patient,
focusing the combined
on the explicit, although arbitraryes-
19,52,53
leads
in all ofto theaincluded
significant
papersreduction in the analysis
. A combined use of timated NNT was to
outcome ‘responders sixtherapy’,
patients. The magnitude
it seems that an OR of
(i.e., meta-analysis) of homogeneous results, quantifies 2.19 corresponding to an ES of 0.4364 e indicates a small
rescue
the magnitude medication, corresponding
of clinical efficacy per se59. Thus, to a small
it seems ofto R. canina
moderate hip powder
clinical efficacy6,7,33as. When
a pain reducing
translated into the
clinical
possible that efficacy. This does
the empirical not allow
magnitude more
of clinical de-
efficacy agent
number is of more
patients pronounced
who would need than R. canina thehipprima-
powder
following use of R. canina hip powder is comparable to therapy (compared to placebo) in order to ‘‘treat’’ one pa-
other nutraceuticals available60. Our meta-analysis tient, the combined estimated NNT was six patients. The
supports the conclusion previously stated by Chrubasik: magnitude of R. canina hip powder as a pain reducing
‘‘Moderate evidence exists for the use of a powder of the agent is more pronounced than the primary analgesic of
10
ry analgesic of choice in clinical practice,
paracetamol/acetaminophen, which com-
pared to placebo has an ES - 0.13 (95% CI:
analgesic of choice in clinical practice,
paracetamol/acetaminophen, which com-
pared to placebo has an ES of 0.13 (95% CI:
0.04-0.22), and thus of questionable clinical
significance65. Hence R. canina hip powder
might have an impact as an over-the-coun-
ter (OTC) preparation in the future. The pa-
tients studied in the present meta-analysis
(see Table I)19,52,53 represent a fairly homog-
enous OA population with a clinically rele-
vant age distribution. The exact degree of
OA (i.e., radiographic data) was not given
in the studies, in one study52, however, pa-
tients were presumably endstage, which
might be the reason for a continuous use
of pain medication in this study, in spite of
a significant effect of R. canina hip powder
on self-reported pain. An increasing inter-
est has been noted over the last years for
dietary supplements for OA54 with a special
emphasis on glucosamine and chondroi-
tine66,67. Glucosamine only shows significant
efficacy in Rottapharm-supported pivotal
trials, of which three well conducted trials
had a pooled efficacy of ES ¼ 0.27 (95% CI:
0.12-0.43)68. Never the less the efficacy of
glucosamine has been heavily debated,
among many things because of the great
heterogeneity between efficacy out-
comes69-71.
In the present analysis of R. canina hip pow-
der, the lack of heterogeneity between
studies gives credit to an efficacy. The draw-
back of this observation is - as with the Rot-
tapharm product - that the same company
sponsored all three studies on R. canina hip
powder. Ideally, other similar products from
other manufacturers should be tested to
substantiate the outcome or even better,
the presumed analgesic of choice in clinical
practice, paracetamol/acetaminophen,
which compared to placebo has an ES of
0.13 (95% CI: analgesic of choice in clinical
practice, paracetamol/acetaminophen,
which compared to placebo has an ES of
0.13 (95% CI: 0.04e0.22), and thus of ques-
tionable clinical significance65. Hence R.
canina hip powder might have an impact
as an over-the-counter (OTC) preparation in
the future. The patients studied in the present
meta-analysis (see Table I)19,52,53 represent a
fairly homogenous OA population with a
clinically relevant age distribution. The exact
degree of OA (i.e., radiographic data) was
not given in the studies, in one study52, how-
ever, patients were presumably endstage,
which might be the reason for a continuous
use of pain medication in this study, in spite
of a significant effect of R. canina hip pow-
der on self-reported pain. An increasing in-
terest has been noted over the last years for
dietary supplements for OA54 with a special
emphasis on glucosamine and chondroi-
tine66,67. Glucosamine only shows significant
efficacy in Rottapharm-supported pivotal
trials, of which three well conducted trials
had a pooled efficacy of ES = 0.27 (95% CI:
0.12-0.43)68. Never the less the efficacy of
glucosamine has been heavily debated,
among many things because of the great
heterogeneity between efficacy out-
comes69-71.
In the present analysis of R. canina hip pow-
der, the lack of heterogeneity between
studies gives credit to an efficacy. The draw-
back of this observation is - as with the Rot-
tapharm product - that the same compa-
ny sponsored all three studies on R. canina
hip powder. Ideally, other similar products
from other manufacturers should be tested
to substantiate the outcome or even bet-
ter, the presumed active ingredient (e.g.,
GOPO) should be isolated, patented, and
tested in a strictly controlled clinical trial,
following guidelines for Good Clinical Prac-
tice (GCP) and consolidated standards of
reporting trials (CONSORT)72. Such initiatives
would increase the external validity of any
proposed herbal therapy73. Patients with
chronic painful diseases seek complemen-
tary-alternative therapy for various reasons.
Ramsey et al. has previously reported from
a US cohort, that alternative medicine use is
highly prevalent among those with OA (47%)
and that levels of expenditure for those who
do consume these services ($1,127 per
year) approximate expenditures on more
traditional medical care ($1,148 per year)74.
The traditional medical approach has only
been able to offer slight improvements with
regard to pain65,75 with a definite problem of
a rather frightening list of adverse events75,76.
In contrast, alternative medications are re-
peatedly found (i.e., report) to have almost
no adverse effects; this has been shown for

both glucosamine77, chondroitin78, avoca-
do/soybean unsaponifiables (ASU)79 among
many e as well as R. canina hip powder.
We turn to the question: whether treatment
of OA via prescription of anti-oxidants is
dream or reality?80 In a shortterm cross-over
trial 1 g of calcium ascorbate for either knee
or hip OA was given for 14 days, resulting in
a small to moderate (statistically significant)
pain reduction compared to placebo81,
which is equivalent to our results for R. cani-
na hip powder. It is, however, noteworthy
that data have been presented that the
anti-inflammatory properties of R. canina
hip powder is unrelated to its vitamin C con-
tent14,15. In regard to anti-oxidants, however,
a recent large-scale meta-analysis found
that treatment with b-carotene, vitamin A,
and vitamin Emay increasemortality, while
a potential role for vitaminCremains to be
clarified82. By consequence, large-scale tri-
als on anti-oxidants are still relevant, and no
final conclusion may be drawn regarding
safety.
Alternative therapy should be subjected to
a similar scrutiny of effect vs adverse effects
as ordinary medications42,83. The alternative
OTC market is huge74 whether it is efficacious
or not54, and with an inevitable influence
on both direct and indirect costs84. With re-
gard to R. canina hip powder a large-scale
trial is justified by the magnitude of clinical
efficacy demonstrated in this meta-analysis
of shortterm trials e an efficacy in the area
of 0.4 SMD-points. In a parallel group design
this would correspond to 133 OA patients in
each group in order to assess a statistically
significant effect (P < 0.05, two-tailed) with a
proper statistical power (90%)85. In order to
monitor the clinical efficacy applying these
R. canina hip powder products, the next
RCT should be of at least half a year dura-
tion, although a 1-year trial with sufficient
imaging would be even better9.
We emphasize the need for future studies
applying empirically validated outcomes
(e.g., WOMAC86, knee and osteoarthritis
outcome core (KOOS)87 or the Lequesne
index88)11, and that these studies explicitly
report the number of so-called responders
according to the OMERACT-OARSI response
11
criterion11,89. Also, a study should strictly
adhere to the CONSORT statement72,73,90,
and besubjected to central registration
(e.g., http://www. clinicaltrials.gov).
In conclusion, the dry powder of R. canina
L. fruit (i.e., R. canina hip powder) seems to
have a consistent, small to moderate effi-
cacy on pain in OA patients; however, an
efficacy only observed in short-term clinical
trials (3-4 months). The adverse events were
similar to placebo in the available literature,
and it seems safe to apply this herbal rem-
edy, though long-term safety remains to be
tested. The results of the present meta-anal-
ysis - that R. canina hip powder does reduce
pain e should be further substantiated in a
large-scale (i.e., phase III) trial.
Conflicts of
interest
RC is statistical editor in the Cochrane Mus-
culoskeletal Group (CMSG, Australian edito-
rial base); the present meta-analysis is not
a Cochrane review. The funding agencies
(The Danish Rheumatism Association and
The Oak Foundation) had no role in study
design, data collection, data synthesis, data
interpretation, writing the report, or the de-
cision to submit the manuscript for publica-
tion.
None of the authors is affiliated with or fund-
ed by any manufacturer of a R. canina hip
powder agent.
Acknowledgements
This study was supported by grants from the
Oak Foundation, The Danish Rheumatism
Association, and Frederiksberg Hospital. We
acknowledge the personal and scientific
supports of Professor Bente Danneskiold-
Samsøe, M.D., Head of the Parker Institute.
We thank Mette Gad, M.A., for linguistic sup-
port.
12
References
1. Dieppe PA, Lohmander LS. Pathogenesis
and manage ment of pain in osteoar
thritis. Lancet 2005;365(9463): 965-73.
2. Altman RD. Criteria for classification
of clinical osteoarthritis. J Rheumatol
Suppl 1991;27:10-2.
3. Altman R, Asch E, Bloch D, Bole G,
Borenstein D, Brandt K, et al.
Development of criteria
for the classification and reporting of
osteoarthritis. Classification of osteoarthri
tis of the knee. Diagnostic and Therapeu
tic Criteria Committee of the American
Rheumatism Association. Arthritis Rheum
1986;29(8):1039-49.
4. Abramson SB. Inflammation in osteoarthri
tis. J Rheumatol Suppl 2004; 70:70-6.
5. Spector TD, Hart DJ, Nandra D, Doyle DV,
Mackillop N, Gallimore JR, et al. Low-level
increases in serum C-reactive protein
are present in early osteoar
thritis of the knee and predict progres
sive disease. Arthritis Rheum 1997;40(4)
:723-7.
6. Jordan KM, Arden NK, Doherty M,
Bannwarth B, Bijlsma JW, Dieppe P, et al.
EULAR Recommendations 2003: an
evidence based approach to the man
agement of knee osteoarthritis: report of
a Task Force of the Standing Committee
for International Clinical Studies Includ
ing Therapeutic Trials (ESCISIT). Ann
Rheum Dis 2003;62(12):1145-55.
7. Zhang W, Doherty M, Arden N,
Bannwarth B, Bijlsma J, Gunther KP, et al.
EULAR evidence based recommenda
tions for the management of hip osteoar
thritis: report of a task force of the EULAR
Standing Committee for International
Clinical Studies Including Therapeutics
(ESCISIT). Ann Rheum Dis 2005;64(5)
:669-81.
8. Abadie E, Ethgen D, Avouac B, Bouvenot
G, Branco J, Bruyere O, et al. Recommen
dations for the use of new methods to
assess the efficacy of disease-modifying
drugs in the treatment of osteoarthritis.
Osteoarthritis Cartilage 2004;12(4):263-8.
9. Bellamy N, Kirwan J, Boers M, Brooks
P, Strand V, Tugwell P, et al. Recommenda
tions for a core set of outcome
mea sures for future phase III clinical
trials in knee, hip, and hand osteoarthritis.
Consensus development at
OMERACT III. J Rheumatol
1997;24(4):799-802.
10.Tugwell P, Shea B, Boers M, Brooks P, Simon
LS, Strand V, et al. Evidence-Based Rheu
matology. 1st edn. London: BMJ Books;
2004.
11.Dougados M. Monitoring osteoarthritis
progression and therapy. Osteoarthritis
Cartilage 2004;12(SupplA):S55-60.
12.Chrubasik C, Duke RK, Chrubasik S. The
evidence for clinical efficacy of rose hip
and seed: a systematic review. Phytother
Res 2006;20(1):1-3.
13.Ameye LG, Chee WS. Osteoarthritis and
nutrition. From nutraceuticals to function
al foods: a systematic review of the
scientific evidence. Arthritis Res Ther
2006;8(4):R127.
14.Winther K, Rein E, Kharazmi A. The anti-
inflammatory properties of rose-hip. In
flammopharmacology 1999;7(1):63-8.
15. Kharazmi A, Winther K. Rose hip inhibits
chemotaxis and chemiluminescence of
human peripheral blood neutrophils in
vitro and reduces certain inflammatory
parameters in vivo. Inflammopharma-
cology 1999;7(4):377-86.
16. Larsen E, Kharazmi A, Christensen LP,
Christensen SB. An antiinflammatory
galactolipid from rose hip (Rosa cani
na) that inhibits chemotaxis of human
peripheral blood neutrophils in vitro.
J Nat Prod 2003;66(7):994-5.
17. Orhan DD, Hartevioglu A, Ku¨ peli E,
Yesilada E. In vivo anti-inflammatory and
antinociceptive activity of the
crude extract and fractions from
Rosa canina L. fruits. J
Ethnopharmacol 2007;112:394-400.
18. Jager AK, Eldeen IM, van SJ. COX-1 and
-2 activity of rose hip. Phytother Res 2007
Dec;21(12):1251-2.
19. Winther K, Apel K, Thamsborg G. A

powder made from seeds and shells of a
rose-hip subspecies (Rosa canina)
reduces symptoms of knee and hip os
teoarthritis: a randomized, double-blind,
placebo-controlled clinical trial. Scand J
Rheumatol 2005;34(4):302-8.
20. Moher D, Cook DJ, Eastwood S, Olkin I,
Rennie D, Stroup DF. Improving the
quality of reports of meta-analyses
of randomised controlled trials: the
QUOROM statement. Quality of Report
ing of Metaanalyses. Lancet
1999;354(9193):1896-900.
21. Jadad AR, Moore RA, Carroll D,
Jenkinson C, Reynolds
DJ, Gavaghan DJ, et al. Assessing the
quality of reports of randomized clinical
trials: is blinding necessary? Control Clin
Trials 1996;17(1):1-12.
22. Huskisson EC. Measurement of pain.
J Rheumatol 1982;9(5):768-9.
23. Christensen R, Sørensen LB, Bartels EM,
Astrup A, Bliddal H. Rose-hip in osteoar
thritis (OA): a meta-analysis. Ann Rheum
Dis 2007;66(SupplII):495 (Abstract).
24. Elbourne DR, Altman DG, Higgins JP,
Curtin F, Worthington HV, Vail A. Meta-
analyses involving cross-over trials:
methodologi cal issues. Int J Epidemiol
2002;31(1):140-9.
25. Curtin F, Elbourne D, Altman DG.
Meta-analysis combining parallel and
cross-over clinical trials. III: the issue of
carry-over. Stat Med 2002;
21(15):2161-73.
26. Deeks JJ, Higgins JPT, Altman DG,
Higgins JPT, Green S, Eds. Analysing and
Presenting Results. 4.2.5. edn. Co chrane
Handbook for Systematic Reviews of
Interventions 4.2.5 Chichester, UK: John
Wiley & Sons, Ltd; 2005:97-166 (Updated
May 2005).
27. Altman DG, Bland JM. Interaction
revisited: the difference between two
estimates. BMJ 2003;326(7382):219.
28. Normand SL. Meta-analysis: formulating,
evaluating, combining, and reporting.
Stat Med 1999;18(3):321-59.
29. Cohen J. Statistical Power Analysis for
13
the Behavioral Sciences. 2nd edn.
Lawrence Earlbaum Associates; 1988.
30. Hedges LV. Distribution theory for Glass’s
estimator of effect size and related esti
mators. J Educational Statistics
1981;6(2):107-28.
31. Norman GR, Sloan JA, Wyrwich KW.
Interpretation of changes in heal-
threlated quality of life: the remark
able universality of half a standard
deviation. Med Care 2003;41(5):582-92.
32. Kazis LE, Anderson JJ, Meenan RF.
Effect sizes for interpreting changes in
health status. Med Care 1989;27
(Suppl3):S178-89.
33. Zhang W, Doherty M, Leeb BF,
Alekseeva L, Arden NK, Bijlsma JW, et al.
EULAR evidence based recommenda
tions for the management of hand
osteoarthritis: report of a Task Force
of the EULAR Standing Committee for
International Clinical Studies Including
Therapeutics (ESCISIT). Ann Rheum
Dis 2007;66(3):377-88.
34. Bland JM, Altman DG. Statistics notes:
the odds ratio. BMJ 2000; 320(7247):1468.
35. Littell RC, Milliken GA, Stroup WW,
Wolfinger RD, Schaben berger O. SAS for
Mixed Models. 2nd edn. Cary, NC: SAS
Institute Inc; 2006.
36. Patterson HD, Thompson R. Recovery
of inter-block information when block
sizes are unequal. Biometrika
1971;58(3):545-54.
37. Wang MC, Bushman BJ. Integrating
Results through Meta-Analytic Review Us
ing SAS Software. 1st edn. Cary, NC: SAS
Institute Inc; 1999.
38. van Houwelingen HC, Arends LR, Stijnen T.
Advanced methods in meta-analysis:
multivariate approach and meta-regres
sion. Stat Med 2002;21(4):589-624.
39. Cochran WG. The combination of esti
mates from different experiments.
Biometrics 1954;10(1):101-29.
40. Higgins JP, Thompson SG. Quantifying het
erogeneity in a metaanalysis. Stat
Med 2002;21(11):1539-58.
41. Higgins JP, Thompson SG, Deeks JJ,
14
Altman DG. Measuring inconsistency
in meta-analyses. BMJ 2003;327(7414):
557-60.
42. Christensen R, Kristensen PK, Bartels
EM, Bliddal H, Astrup A. Efficacy and
safety of the weight-loss drug rimona
bant: a meta-analysis of andomised
trials. Lancet 2007;370:1706-13.
43. Osiri M, Suarez-Almazor ME, Wells GA,
Robinson V, Tugwell P. Number needed to
treat (NNT): implication in rheuma-
tology clinical practice. Ann Rheum Dis
2003;62(4):316-21.
44. Kristensen LE, Christensen R, Bliddal H,
Geborek P, Danneskiold- Samsoe B,
Saxne T. The number needed to
treat for adalimumab, etanercept
and infliximab based on ACR50 response
in three randomized controlled trials on
established rheumatoid arthritis: a
systematic literature review. Scand J
Rheumatol 2007;36(6):411-7.
45. Cates CJ. Simpson’s paradox and
calculation of number needed to treat
from meta-analysis. BMC Med Res
Methodol 2002;2:1.
46. Altman DG, Deeks JJ. Meta-analysis,
Simpson’s paradox, and the number
needed to treat. BMC Med Res
Methodol 2002;2:3.
47. Edwards A, Elwyn G, Mulley A. Explaining
risks: turning numerical data into
meaningful pictures. BMJ
2002;324(7341):827-30.
48. Visual Rx. Version 1.0-1.7. Copyright®
2000-2003. Christopher Cates,
<http://www.nntonline.net>; 2003.
49. Nobaek S, Johansson ML, Molin G,
Ahrne S, Jeppsson B. Alteration of
intestinal microflora is associated with
reduction in abdominal bloating
and pain in patients with irritable bowel
syndrome. Am J Gastroenterol
2000;95(5):1231-8.
50. Rossnagel K, Willich SN. Value of
complementary medicine exemplified
by rose-hips. Gesundheitswesen
2001;63(6):412-6.
51. Rein E. Hansen OT, Hansen M, Assignee.
Composition and method for alleviating
joint pain and stiffness. United States
Patent. Patent US 6,485,752 B1.
2002 Nov 26.
52. Warholm O, Skaar S, Hedman E, Mølmen
HM, Eik L. The effects of a standardized
herbal remedy made from a subtype of
Rosa canina in patients with osteoart
hritis: a double-blind, randomized,
placebo-controlled clinical trial.
Curr Ther Res Clin Exp 2003; 64(1):21-31.
53. Rein E, Kharazmi A, Winther K. A herbal
remedy, Hyben Vital (stand. powder of a
subspecies of Rosa canina fruits),
reduces pain and improves general well
being in patients with osteoarthritis - a
doubleblind, placebo-controlled,
randomised trial. Phytomedicine 2004;
11(5):383-91.
54. Ernst E. Complementary or alternative
therapies for osteoarthritis. Nat Clin Pract
Rheumatol 2006;2(2):74-80.
55. Chrubasik JE, Roufogalis BD, Chrubasik S.
Evidence of effectiveness of herbal
antiinflammatory drugs in the treatment
of painful osteoarthritis and chronic low
back pain. Phytother Res 2007;21(7):
675-83.
56. Fan X, Felso¨valyiA`, Sivo SA, Keenan S
C. SAS for Monte Carlo Studies: a Guide
for Quantitative Researchers. 3rd edn.
Cary, NC, USA: SAS Institute Inc; 2001.
57. Agresti A. A survey of exact inference for
contingency tables. Stat Sci 1992;7(1):
131-77.
58. Thomas DG. Algorithm AS 36: exact
confidence limits for the odds ratio in a
2  2 table. Appl Stat 1971;20(1):105-10.
59. Thompson SG, Pocock SJ. Can meta-
analyses be trusted? Lancet
1991;338(8775):1127-30. 60. Zhang W,
Moskowitz RW, Nuki G, Abramson S,
Altman RD, Arden N, et al. OARSI recom
mendations for the management of hip
and knee steoarthritis, part I: criti
cal appraisal of existing treatment guide
lines and systematic review of current re

search evidence. Osteoarthritis Carti
lage 2007;15(9):981-1000.
61. Tubach F, Ravaud P, Baron G, Falissard B,
Logeart I, Bellamy N, et al. Evaluation of
clinically relevant changes in patient re
ported outcomes in knee and hip
osteoarthritis: the minimal clinically
important improvement. Ann Rheum Dis
2005;64(1):29-33.
62. Tubach F, Ravaud P, Baron G, Falissard B,
Logeart I, Bellamy N, et al. Evaluation of
clinically relevant states in
patient reported outcomes in knee and
hip osteoarthritis: the patient accept
able symptom state. Ann Rheum Dis
2005;64(1):34-7.
63. Woolf AD. Healthcare services for those
with musculosk eletal conditions: a rheu
matology service. Recommen
dations of the European Union of Medi
cal Specialists Section of Rheumatology/
European Board of Rheuma
tology 2006. Ann Rheum Dis 2007;66(3):
293-301.
64. Chinn S. A simple method for converting
an odds ratio to effect size for use in me
ta-analysis. Stat Med 2000;19(22):
3127-31.
65. Towheed T, Maxwell L, Judd M, Catton
M, Hochberg M, Wells G.
Acetaminophen for osteoarthritis.
Cochrane Database Syst Rev
2006;1:CD004257.
66. McAlindon TE, LaValley MP, Gulin JP,
Felson DT. Glu
cosamine and chondroitin for treatment
of osteoarthrtis: a systematic quality as
sessment and meta-analysis. JAMA
2000;1469-75.
67. Richy F, Bruyere O, Ethgen O,
Cucherat M, Henrotin Y, Reginster
JY. Structural and symptomatic efficacy
of glucosamine and chondroitin in knee
osteoarthritis: a com prehensive meta-
analysis. Arch Intern Med
2003;163(13):1514-22.
68. Reginster JY. The efficacy of glucosamine
15
sulfate in osteoarthritis: financial and nonfi
nancial conflict of interest. Arthritis Rheum
2007; 56(7):2105-10.
69. Towheed TE, Anastassiades TP.
Glucosamine and chon droitin for treating
symptoms of osteoarthritis: evidence
is widely touted but incomplete. JAMA
2000;283(11):1483-4.
70. Towheed TE. Current status of glucosamine
therapy in osteoarthritis. Arthritis Rheum
2003;49(4):601-4.
71. Vlad SC, LaValley MP, McAlindon TE, Felson
DT. Glucosamine for pain in osteoarthritis:
why do trial results differ? Arthritis Rheum
2007; 56(7):2267-77.
72. Altman DG, Schulz KF, Moher D, Egger M,
Davidoff F, Elbourne D, et al. The revised
CONSORT statement for reporting random
ized trials: explanation and elabora
tion. Ann Intern Med 2001;134(8):663-94.
73. Gagnier JJ, Boon H, Rochon P, Moher
D, Barnes J, Bombardier C. Recommenda
tions for reporting randomized controlled
trials of herbal interventions: explanation
and elaboration. J Clin Epidemiol 2006;
59(11):1134-49.
74. Ramsey SD, Spencer AC, Topolski TD, Belza
B, Patrick DL. Use of alternative therapies
by older adults with osteoarthritis. Arthritis
Rheum 2001;45(3):222-7.
75. Bjordal JM, Ljunggren AE, Klovning A, Slord
al L. Non-steroidal antiinflammatory drugs,
including cyclo-oxygenase-2 inhibitors, in
osteoarthritic knee pain: meta-analysis of
randomised placebo controlled trials. BMJ
2004;329(7478):1317.
76. Lee C, Straus WL, Balshaw R, Barlas S, Vogel
S, Schnitzer TJ. A comparison of the effica
cy and safety of nonsteroidal antiinflam
matory agents versus acetaminophen in
the treatment of osteoarthritis: a meta-
analysis. Arthritis Rheum 2004;51(5):746-54.
77. Towheed TE, Maxwell L, Anastassiades TP,
Shea B, Houpt J, Robinson V, et al. Glu
cosamine therapy for treating os
teoarthritis. Cochrane Database Syst Rev
2005;2:CD002946.
78. Reichenbach S, Sterchi R, Scherer M, Trelle
16

S, Burgi E, Burgi U, et al. Meta-analysis: Osteoarthritis Outcome Score (KOOS) e

chondroitin for osteoarthritis of the knee or validation and comparison
hip. Ann Intern Med 2007;146(8):580-90. to the WOMAC in total knee replacement.
79. Christensen R, Bartels EM, Astrup A, Bliddal Health Qual Life Outcomes 2003;1(1):17.
H. Symptomatic efficacy of avocado/soy 88. Lequesne MG, Mery C, Samson M,
bean unsaponifiables Gerard P. Indexes of severity for osteoar
(ASU) in osteoarthritis (OA) patients: a me thritis of the hip and knee. Validation
ta-analysis of randomized controlled trials. - value in comparison with other assess
Osteoarthritis Cartilage 2007 Nov 23 [Epub ment tests. Scand J Rheumatol Suppl
ahead of print]. 1987;65: 85-9.
80. Henrotin Y, Kurz B. Antioxidant to treat os 89. Pham T, van der HD, Altman RD, Anderson
teoarthritis: dream or reality? Curr Drug Tar JJ, Bellamy N, Hochberg M, et al. OMERAC
gets 2007;8(2):347-57. TeOARSI initiative: Osteoarthritis Research
81. Jensen NH. Reduced pain from osteoar Society International set of
thritis in hip joint or knee joint during treat responder criteria for osteoarthritis clinical
ment with calcium ascorbate. A random trials revisited. Osteoarthritis Cartilage
ized, placebocontrolled cross-over trial in 2004;12(5):389-99.
general practice. Ugeskr Laeger 2003; 90. Moher D, Schulz KF, Altman D. The
165(25):2563-6. CONSORT statement: revised recommen
82. Bjelakovic G, Nikolova D, Gluud LL, dations for improving the quality of reports
Simonetti RG, Gluud C. Mortality in ran of parallel-group randomized trials. JAMA
domized trials of antioxidant supple 2001;285(15):1987-91.
ments for primary and secondary preven
tion: systematic review and meta-analysis.
JAMA 2007; 297(8):842-57.
83. Berlin JA, Colditz GA. The role of
meta-analysis in the regulatory process for
foods, drugs, and devices. JAMA
1999;281(9):830-4.
84. Rabenda V, Manette C, Lemmens R, Mari
ani AM, Struvay N, Reginster JY. Direct and
indirect costs attribut able to osteoarthritis
in active subjects. J Rheumatol
2006;33(6):1152-8.
85. Campbell MJ, Julious SA, Altman DG. Esti
mating sample sizes for binary, ordered
categorical, and continuous outcomes in
two group comparisons. BMJ
1995;311(7013):1145-8.
86. Bellamy N, Buchanan WW, Goldsmith CH,
Campbell J, Stitt LW. Validation study of
WOMAC: a health status in strument for
measuring clinically important patient rel
evant outcomes to antirheumatic drug
therapy in patients with osteoarthritis of
the hip or knee. J Rheumatol
1988;15(12):1833-40.
87. Roos EM, Toksvig-Larsen S. Knee injury and
 17

The eVIdence for clInIcal effIcacy of

rose hIP and seed: a systematIc reVIeW
C. chrubasik1, R K. Duke2,3 and S. Chrubasik1,3
1
Institute of Forensic Medicine, University of Freiburg i.Br., Albertstr. 9, 79104 Freiburg i.Br., Germany
2
Pharmaceutical Chemistry, Faculty of Pharmacy, University of Sydney, NSW 2006, Australia
3
Herbal Medicines Research and Education centre, faculty of pharmacy, university of sydney,
NSW 2006 australia

Background: The objective of this review is to evaluate whether clinical research has gained any
evidence of effectiveness of Rosa canina preparations.

Methods: several databases and other sources were searched to identify randomized controlled
trials of rosa canina preparations.

Results: Trials were described in a narrative way, taking into consideration methodological quality
scores. four trials were included in this review and two were identified as subgroup analyses.

Conclusion: moderate evidence exists for the use of a powder of the seeds and husks of a rosa
canina subspecies in patients suffering from osteoarthritis. copyright © 2006 John wiley & sons, ltd.

keywords: osteoarthritis; pain; rose hip and seed; nutraceutical.

Study population methodes

The German Commission e monograph (blu-
menthal, 1998) summarizes the indications for
rose hip and seed in traditional medicine which
include the prevention and treatment of colds
and influenza-like infections, infectious diseases,
prophylaxis and therapy of vitamin c deficien-
cies, fever, for increase in the immune mecha-
nism during general exhaustion, gastric spasms,
gastric acid deficiency, prevention of inflamma-
tion of the gastric mucosa and gastric ulcers, as
‘stomach tonic’, for intestinal diseases, for diar-
rhea, as prophylaxis of intestinal catarrhs, as a
laxative, for gallstones, gall- and discomforts and
ailments, diseases and discomforts of the lower
urinary tract, dropsy, as a ‘tonic for kidneys’, as
a diuretic, for gout, disorders of uric acid metab-
olism, arthritis, sciatica, diabetes, inadequate
peripheral circulation, as an astringent, for lung
ailments, and as an eye rinse.
The monograph stated that the effectiveness
of the herb for most of its claimed applications
was not documented. investigations in rats and
rabbits failed to demonstrate an increased di-
uresis and a hypoglycaemic effect, respectively
(anon., 1998). however, a potent antioxidative
effect was seen in vitro (anon., 1998).
The aim of this study was to evaluate whether in
the meantime clinical research has gained any
evidence of efficacy for rose hip and seed.
computerized literature searches were carried
out by the authors (medline, pubmed, co-
chrane collaboraTion library, embase
(ovid technologies) back to 1985 and also
manually to identify randomized controlled
studies (rcT) investigating preparations of rosa
canina (‘or’ rosehip ‘or’ rose hip ‘or’ rose hip and
seed, hagebutte (medline rosa ‘or’ fruit; drug
effects)). The following data were extracted
from each study: authors’ names; date of pub-
lication; country of origin; type of study, includ-
ing number of study centres; participants (num-
bers, disease(s), characteristics of the study
population (age, size, weight, gender) ); dura-
tion of acute exacerbation or chronic disease;
baseline values with details on pain and previ-
ous treatments; additional treatments; types of
outcome measures; summary statistics; timing
of outcome assessment; withdrawals and drop-
outs; and adverse events. methodological
quality and level of evidence were assessed
as described in a previous review (gagnier et
al., 2004): Quality items: (a) eligibility criteria
specified, (b) randomization appropriate, (c)
treatment allocation concealed, (e) similarity
at baseline, (f) outcome measures and control
interventions explicitly described, (g) co-inter-
ventions comparable, (h) outcome measures
relevant, (i) adverse events and (J) drop-outs
fully described, (k) sample size based on a pri-
ori power calculation, (l) intention-to-treat ana-
18

sis, (n) point estimates and measures of variabi- among one confirmatory study with a clinical
lity presented for the primary outcome measu- relevant effect and/ or multiple exploratory
re, (o) appropriate timing giving a Total score rcTs, insufficient – one low quality rcT, conflic-
(Ts) of 13; levels of evidence strong – pooling of ting – inconsistent findings among multiple trials,
data or at least 2 confirmatory studies (consi- no evidence from trials - no RCTs
dering items k and n) demonstrating a clinical
relevant effect, moderate – consistent findings
table 1. (a) eligibility criteria specified, (B) randomization appropriate, (c) treatment allocation
conce-aled, (e) similarity at baseline, (f) outcome measures and control interventions explicitly
described, (g) co-interventions comparable, (h) outcome measures relevant, (I) adverse events
and (J) drop-outs fully described, (K) sample size based on a priori power calculation, (l) intention-
to-treat analysis, (n) point estimates and measures of variability presented for the primary outcome
measure, (o) appropriate timing giving a total score (ts) of 13

Osteoarthritis Cartilage 2004; 12 Suppl 2

Phytomedicine Curr Ther Res Clin Exp
2004; 11: 383–391 2003; 64: 21–31 80 (subgroup) 145 (subgroup)
n = 112 n = 100 n = 94 (112) n = 32 (112)
5 g/day vs placebo 5 g/day vs placebo 5 g/day vs placebo 5 g/day vs placebo
Cross-over Parallel Cross-over Cross-over
Over 3 months Over 4 months Over 3 months Over 3 months
A OA multiple sites Hip, knee Hip, knee Hand
B Yes Yes Yes Yes
C Yes Yes Yes Yes
E Yes Don‘t know Yes Yes
F Yes Yes Yes Yes
G Yes Yes Yes Yes
H Yes Yes Yes Yes
I Yes Yes Yes Yes
J Yes Yes Yes Yes
K Yes Yes No No
L Yes Yes Yes Yes
N No No No No
O Yes Yes Yes Yes
TS 11 10 11 11

results a total of 88 (30 pubmed, 24 medline),
citations were screened and 4 rcTs identified
(warholm et al., 2003; rein et al., 2004a, b; win-
ther and kharazmi, 2004), however, two were
identified as subgroup analyses (rein et al.,
2004a; winther and kharazmi, 2004). all trials
were carried out with a powder of the seeds
and husks of a rosa canina subspecies in pa-
tients suffering from osteoarthritis. a full descrip-
tion of the studies is placed on the webpage
http://remed-chrubasik.uniklinikfreiburg. de. The
two main studies were of high quality (Ts 10, 11,
Table 1), but not confirmatory. relief of joint pain
was greater after 3 and 4 months of treatment
with 5 g powder/day compared with placebo,
respectively (n = 112, p < 0.01; n = 100, p < 0.05).
likewise, activities of daily living were more im-
proved and consumption of rescue medication
was significantly less.
discussion
our systematic review shows that clinical evi-
dence of effectiveness has only been gained in
the field of osteoarthritis. There is evidence that
nutritional supplementation with a dry powder
of a rosa canina subspecies may decrease
both osteoarthritic pain and the consump-
tion of additional synthetic pain medications.
The proprietary powder has a potent antioxi-
dativeeffect (daels-rakotoarison et al., 2002),
inhi-bited chemotaxis and chemiluminescence
of human peripheral blood neutrophils in vitro
and reduced certain inflammatory parameters
in vivo (kharazmi and winther, 1999; winther et
al., 1999). a galactolipid contributes to the an-
tiinflammatory principle (larsen et al., 2003).
painful arthritis is usually treated with nonsteroi-
dal antiinflammatory drugs (nsaids) (pincus et
al., 2000), although their use is often associated
with adverse gastrointestinal events that may
be life threatening in some patients (smalley
et al., 1995). The cost of health care resources
spent on preventing and managing these side-
effects was calculated to be around one Ca-
nadian dollar for every day of nsaid treatment
(rahme et al., 2001). safer therapies are therefore
required and have led to the introduction

of selective cox-2 inhibitors for the treatment
of chronic pain (Grainger and Cicuttini, 2004).
However, recently, rofecoxib (vioxx®) although
associated with a statistically significantly lower
incidence of upper gastrointestinal bleedings
(watson et al., 2004) was voluntarily withdrawn
from the market due to increased risk of car-
diovascular events (davies and Jamali, 2004).
some nutraceuticals may be promising alterna-
tives to synthetic medications in the treatment
of musculoskeletal pain, although conclusive
studies are required for all of them: proprietary
preparations from devil’s claw (Gagnier et al.,
2004), willow bark (Chrubasik et al., 2000a, b),
ginger (chrubasik et al., 2005), avocado-soybe-
an (ernst, 2003) and glucosamine supplemented
with and without shark chondroitin (mcalindon
et al., 2000). confirmatory studies (considering
items k and n) are also required for the propri-
etary rose hip and seed preparation in order to
prove the effectiveness beyond any doubt in
the treatment of osteoarthritis. more research is
needed to clarify the other supposed rose hip
and seed effects.
References
Anon. 1998. rosa canina l. in hager hand- dT. 2000. glucosamine and chondroitin
buch der pharmazeutischen praxis,
5th edn. band 4, hänsel r, keller k, rim
pler h, schneider g (eds). springer-
verlag: berlin, heidelberg, new york, 3:
447–452.
blumenthal m. 1998. The complete geman
commission e monographs. The
american botanical council: austin,
Texas, 368–369.
chrubasik s, eisenberg e, balan e, wein
berger T, luzzati r, conradt c. 2000a.
Treatment of low back pain exacer
bations with willow bark extract: a
randomized double-blind study. am J
med 109: 19–14.
chrubasik s, kunzel o, model a, conradt c,
black a. 2000b. Treatment of low back
pain with a herbal or synthetic anti-
rheumatic: a randomized controlled
study. willow bark extract for low
back pain. rheumatology 40: 1388–1393.
chrubasik s, pittler mh, roufogalis bd.
2005. zingiberis rhizome. a comprehen
sive review on the ginger effect and ef
ficacy profiles. phytomedicine in press
daels-rakotoarison da, gressier b, Trotin f
et al. 2002. effects of rosa canina fruit
extract on neutrophil respiratory burst.
phytother res 16: 157–161.
davies nm, Jamali f. 2004. cox-2 selective
inhibitors cardiac toxicity: getting to the
heart of the matter. J pharm pharm sci
19
7: 332–336.
ernst e. 2003. avocado-soybean unsapon
ifiab-les (asu) for osteoarthritis – a sys
tematic review. clin rheumatol 22: 285–
288.
gagnier JJ, chrubasik s, manheimer e. 2004.
harpagophytum procumbens for osteo-
arthritis and low back pain: a system
atic review. bmc complement altern
med 4: 13.
grainger r, cicuttini fm. 2004. medical man-
agement of osteoarthritis of the knee
and hip joints. med J aust 180: 232–236.
kharazmi a, winther k. 1999. rose hip inhibits
chemotaxis and chemiluminescence of
human peripheral blood neurophils in
vitro and reduces certain inflamma
tory parameters in vivo. inflammophar-
macology 7: 377–386.
larsen e, kharazmi a, christensen lp, chris-
tensen sb. 2003. an antiinflammatory
galactolipid from rose hip (rosa canina)
that inhibits chemotaxis of human
peripheral blood neutrophils in vitro.
J nat prod 66: 994–995.
mcalindon Te, lavalley mp, gulin Jp, felson
for treatment of osteoarthritis: a system
atic quality assessment and meta-
analysis. J am med assoc 283: 1469–
1475.
pincus T, swearingen c, cummins p, callahan
lf. 2000. preference for nonsteroidal anti-in
flammatory drugs versus ac etaminophen
and concomitant use of both types of drugs
in patients with osteoarthritis. J rheumatol
27: 1020–1027.
rahme e, Joseph l, kong sx, watson dJ, lelorier
J. 2001. cost of prescribed nsaid-related
gastrointestinal adverse events in elderly pa
tients. Br J clin pharmacol 52: 185–192.
rein e, kharazmi a, winther k. 2004a. a herbal
remedy, hyben vital (stand. powder of a
subspecies of rosa canina fruits), reduces
pain and improves general wellbeing in pa-
tients with osteoarthritis--a double-blind,
placebo-controlled, randomised trial.
phytomedicine 11: 383–391.
rein e, kharazmi a, Thamsborg g, winther k.
2004b. A herbal remedy made from a
subspecies of rose-hip rosa canina, reduces
symptoms of knee and hip osteoarthritis.
osteoarthr cartil 12 (suppl 2): 80.
smalley we, ray wa, daugherty Jr, griffin
mr.1995. nonsteroidal anti-inflammatory
drugs and the incidence of hospitalizations
for peptic ulcer disease in elderly patients.
am J epidemiol 141: 539–545.
20

Warholm O,Skaar s, Hedman E, Molmen HM,
Eik l. 2003. The effects of a standardized
herbal remedy made from a subtype of
rosa canina in patients with osteoarthri-
tis: a double-blind, randomized, placebo-
controlled clinical trial. Curr Ther
Res 64: 21–31. Watson dJ, yu Q, bolognese Ja,
reicin as, simon TJ. 2004. The upper gastrointes-
tinal safety of rofecoxib vs. nsaids: an updated
combined analysis. curr med res opin 20: 1539–
1548.
winther k, kharazmi a. 2004. a powder prepared
from seeds and shells of subtype of rose-hip
rosa canina reduces pain in patients with
osteoarthritis of the hand – a double blind,
placebo-controlled study. osteoarthr cartil
12 (suppl 2): 145. Winther K, Rein E, Kharazmi
A. 1999. The anti- inflammatory properties of
rose-hip. inflam- mopharmacology 7: 63–68.

a PoWder made from seeds and shells

of a rose-hIP suBsPecIes (rosa canina)
reduces symPtoms of Knee and hIP oste-
oarthrItIs: a randomIzed, douBle-BlInd,
PlaceBo-controlled clInIcal trIalK

K Winther1, K apel2, g thamsborg2

1
department of clinical biochemistry, copenhagen county hospital gentofte, and 2department
of rheumatology, copenhagen county hospital glostrup, university of copenhagen, denmark

objective: The aim of this study was to determine whether a herbal remedy made from a subspe-
cies of rose-hip (Rosa canina) might reduce symptoms of osteoarthritis and consumption of rescue
medication in patients suffering from osteoarthritis.

methods: ninety-four patients with osteoarthritis of the hip or knee were enrolled in a randomized,
placebocontrolled, double-blind crossover trial. forty-seven patients were given 5 g of the herbal
remedy daily for a period of 3 months and the remaining patients were given a similar amount of
placebo. The group initially treated with placebo was then changed to rose-hip and vice versa for
another 3-month period. upon inclusion and after 3 weeks and 3 months of each treatment peri-
od, pain, stiffness, disability, and global severity of the disease were scored on a western ontario
and mcmaster universities (womac) questionnaire. after 3 weeks of treatment, patients, if possib-le,
were allowed to reduce their consumption of ‘rescue medication’. data were analysed on the
basis of intention to treat.

results: rose-hip resulted in a significant reduction in womac pain (p<0.014) as compared to place-
bo, when testing after 3 weeks of treatment. The consumption of ‘rescue medication’ signi-ficantly
declined as a result of active treatment (p<0.027). womac disability, stiffness, and global assess-
ment of severity of the disease were not altered by 3 weeks but decreased significantly (p<0.018,
p<0.038, and pv0.035, respectively) after 3 months of treatment.

conclusion: The data suggest that the present herbal remedy can alleviate symptoms of oste-
oarthritis and reduce the consumption of ‘rescue medication’.
osteoarthritis is a disease that reaches younger sportspersons of both sexes, many middle-aged
people, and the majority of the older population.
it has recently been claimed that long-term treatment with glucosamine sulfate can repair the
destroyed cartilage, which is normally thought to be the main element of the disease (1). however,

most treatment is still directed against symptoms
of the disease, such as pain and stiffness, which
are responsible for the main reduction in daily
activities often reported in osteoarthritis.
non-steroidal anti-inflammatory drugs (nsaids),
acetylsalicylic acid, and glucocorticoids are
often used for treatment of such symptoms, al-
though treatments can result in serious side ef-
fects such as bleeding, gastric erosions, and liver
and kidney damage (2, 3). cyclooxygenase- 2
inhibitors, which selectively inhibit the enzyme
cyclooxygenase, have also exerted unfavour-
able effects (4) and the daily cost of the treat-
ment is still very high. paracetamol, which for a
decade was regarded as a safe drug, was re-
cently reported to enhance the risk of upper
gastrointestinal problems (5). for these reasons
there has been a search for new compounds
that could minimize pain and stiffness without
the serious side effects mentioned above. vari-
ous herbal remedies, especially extracts of gin-
ger and avo cado/soybean unsaponifiables,
have shown promising results in patients with
osteoarthritis (6, 7). more focus on remedies of
a herbal origin might therefore, in the future,
change the treatment of patients with osteoar-
thritis by a consumption pattern with fewer side
effects.
inflammatory cells such as polymorphonucleat-
ed leucocytes participate in inflammation and
tissue damage by liberating proteolytic and hy-
drophilic enzymes as well as oxygen radicals.
we have found that a standardized dry powder
made from seeds and shells of a subtype of rose-
hip (rosa canina) reduces the migration rate of
polymorphonucleated leucocytes in vitro and
the serum concentration of c-reactive protein
in humans (8), an effect unrelated to the high
vitamin c content of rose-hip (9).
moreover, some of the osteoarthritic volunteers
who participated in these preliminary studies
claimed that their pain symptoms were dramat-
ically reduced after a period of treatment (8).
This encouraged us to investigate whether a
standardized powder made from the same wild
type of rose-hip (rosa canina) would alleviate
symptoms such as pain and stiffness and im-
prove daily functions in osteoarthritic patients.
we also wanted to evaluate whether an effect,
if present, was of sufficient magnitude to influ-
ence the daily consumption of pain relieving
medicine.
Patients and methods
study population
Patients were recruited from the outpatient clin-
21
ics of the department of rheumatology of co-
penhagen university hospital in glostrup and of
the institute for clinical research. The study was
approved by the ethics committee of vejle and
copenhagen counties (no. 9980042 pmc). pa-
tients were recruited after announcements in
local newspapers. The primary inclusion criteria
were age over 35 years and symptomatic knee
or hip osteoarthritis. osteoarthritis of the knee
or hip was diagnosed according to the clinical
and radiological criteria of the american col-
lege of rheumatology (10, 11). major exclusion
criteria were inflammatory arthritis, fibromyal-
gia, depression, and substantial abnormalities
in haematological, hepatic, renal, or metabolic
functions. furthermore, we excluded patients
who received glucosamine sulfate, chondroitin
sulfate, intra-articular hyaluronate, or systemic
or intra-articular glucocorticoids in the 6 weeks
preceding enrolment.
design and treatment
The study was a randomized, double-blind, pla-
cebocontrolled, crossover trial with three suc-
cessive periods: a 14-day run-in period and two
subsequent treatment periods of 3 months. af-
ter the run-in period, patients were allocated to
receive active medication and placebo in ran-
dom order in the two treatment periods (figure
1). allocation was carried out in blocks of four by
a computer program. active medication com-
prised biologically standardized rosehip pow-
der (litozin). all capsules were produced from
the same batch. identical capsules containing
an inactive powder of similar taste, smell, and
colour were produced for placebo. The dos-
age was a total of 5 g of rose-hip powder ad-
ministered daily as five capsules each of 0.5 g
of the rose-hip or placebo, to be taken in the
morning and again in the evening along with
a meal. compliance with study treatment was
established by asking the patient about missed
doses and by counting the number of returned
capsules.
The rose-hip powder used has been on the
market as a herbal remedy in the scandinavian
countries for almost a decade. it is produced
from fruits of a selected subtype of rosa canina.
The plants are always grown in standardized
fields according to good agricultural practice
and harvesting takes place only when the fruits
are mature. immediately after harvesting, the
fruits are frozen. when the fruits are thawed lat-
er on, a special laser technique is used to en-
sure optimal fruits for the production of powder.
a computerized technique ensures that the
drying process never exceeds 40°c and the dry
powder, which contains elements of the seeds
22
08

as
na.well
Theas the shells
plants are alwaysof thegrownrose-hip, is controlled
in standardized figure 1. flow diagram.
regarding vitaminto
fields according andgoodmineral content.practice
agricultural patients
using nsaids regularly wereonly advised randomIsatIon
and harvesting takes place when to theconti-
fruits (n=94)
nue using the same dosage
are mature. immediately after harvesting, the during the entire
study. however,
fruits are frozen.patients
when the werefruitsadvised
are thawed to redu-la- actIVe treatment PlaceBo treatment
ce
terintake of otherlaser
on, a special analgesics
technique if possible,
is usedsuch to en- as (n=47) (n=47)
paracetamol
sure optimal fruits or synthetic opioids after
for the production the first
of powder.
1 Personal reasons 1 Personal reasons
3aweeks of each of the two treatment periods.

3 Weeks
3 Weeks
computerized technique ensures that the 1 started Prednisolon 1 did not feel that
during the study period, the
drying process never exceeds 40°c and the dry patients were ins- 1 difficulty to swallov treatment worked
tructed
powder,not to change
which containstoelements
another genericof the seeds type capsules
of
asthe
wellsame
as theanalgesic
shells of the or rose-hip,
to use similar tablets
is controlled
containing
regarding vitamin a differentand mineralquantity of the
content. same
patients
painkiller. neither was patients
using nsaids regularly were advised to conti- allowed to start 1 Personal reasons 2 Personal reasons

3 month
3 month
up 1 acid regurgitation 1 acid regurgitation
nueany new
using the type same of pain
dosage relieving
duringmedication.
the entire
1 Insisted on knowing
The consumption
study. however, patients of analgesics
were advised was recorded
to redu- sequense of
daily by the
ce intake patients
of other in a diary.
analgesics The change
if possible, such asin treatment
consumption
paracetamol or synthetic opioids after the first
of analgesics, in each of two
treatment
3 weeks ofperiods,each of was the two estimated
treatment by periods.
subtrac-
ting thethe
during consumption
study period, of medication
the patientsinwere the past
ins-
2tructed
weeks not fromtothat changeof thetoinitial
another 2 weeks.
generic no other
type
cointerventions
of the same analgesic for osteoarthritis
or to use similarwere allowedtablets
during the entire
containing study period.
a different quantity of the same
1 Intercurrent surgery
3 Weeks

3 Weeks
painkiller. neither was patients allowed to start
outcome
up any new measures
type of pain relieving medication.
symptoms
The consumption of osteoarthritis
of analgesics werewas assessed
recorded by
the
dailywestern
by the ontario
patientsand in a mcmaster
diary. The change universities in
(womac)
consumption osteoarthritis
of analgesics, index,in eacha validated,
of the two di-
1 Personal reason 1 With drew as
3 month

sease-specific
treatment periods, questionnaire
was estimated addressing
by subtrac-seve-

3 month
treatment did not
rity
tingofthejoint pain (fiveofquestions),
consumption medication stiffness (two
in the past work any longer
questions),
2 weeks from limitation
that of the of initial
physical2 weeks. function
no other (17
questions),
cointerventions and patients’ global were
for osteoarthritis assessmentallowed of
disease
during the severity
entirereferring
study period.to the 48 h before as- all withdrawals are given in the boxes
sessment (11). The visual analogue scale version
outcome
of the index measures
was used, that is with the patient
symptomseach
assessing of osteoarthritis
question bywere a 100 assessed
mm visual by statistical
pairs wasanalysis
used throughout. subana-lysis com-
the western
analogue ontario
scale. and mcmaster
a higher womac universities
score rep- based
paringon a within-patients
parallel groups was sd of 10%, we
performed cal-
using the
(womac)
resents worse osteoarthritis
symptom index, severity, a with
validated,
2500 mm di- culated that a sample
mann–whitney size are
test. data of 90given
patients in a
as mean
sease-specific
being the worst questionnaire
possible total addressingscoreseve- (12). crossover
values ±sd. design would give a power of 90% in
rity of joint scores
womac pain (five werequestions),
assessed stiffness at (two the detecting more than a 15% difference in the
questions), limitation
begin-ning, after 3 ofweeks, physicaland function at (17 the womac
Results score of joint pain at the 5% level of
questions),
end of each and ofpatients’
the two global assessment
treatment peri- significance.
Patients statistical analysis was based on
of disease severity referring
ods. womac score of joint pain was the to the 48 h before the intention-to-treat
a total of 94 patients, principle with last
comprising 54 obser-
women
assessment
primary outcome(11). The measurevisual together
analogue withscalethe vation
(mean age 66 years; range 38–92) andtest
carried forward. The wilcoxon 40 for
men
version of theof
consumption index was used,
analgesics that during
ta-ken is with the the matched pairs was used throughout.
(mean age 65 years; range 48–85) were enrol- subana-
patient
two assessing
different treatment each periods.
questionwomac by a 100scores mm lysis
led comparing
in the study parallel groups was performed
and randomized to either re-
visual analogue scale. a higher
of stiffness, limitation of physi-cal function, and womac score using
ceive placebo first and then active given
the mann–whitney test. data are as
treatment
represents
patients’ worseassessment
global symptom severity,of disease with 2500
severity mean values ±sd.
(group a, n=547) or active treatment first and
mm occurrence
and being the worst possibleevents
of adverse total score were(12).sec- then placebo (group b, n=547).There were no
womac
ondary scores were
outcome assessed at the begin-
measures. results
significant differences in gender or age on com-
ning, after 3 weeks, and at the end of each of Patients
paring the a and b groups (data not given).
the two treatment
statistical analysis periods. womac score of ain total of 94 group
the entire patients,the comprising
mean body54mass women index
joint pain
based on a was the primary outcome
within-patients sd of 10%,measure we cal- (mean
(bmi) wasage 2766 years;
kg/m2range (range38–92) and
19–41). in40 men a
group
togetherthat
culated with a the consumption
sample size of of 90analgesics
patients in ta-a (mean
the bmi agewas6527.3
years; range(range
kg/m2 48–85)19–39)
were enrol-
and in
ken during the two different
crossover design would give a power of 90% in treatment periods. led in the study and 2 randomized to either re-
group b, 26.6 kg/m (range 22–41), a non-signi-
womac scores
detecting more of thanstiffness,
a 15%limitation
difference of physi-
in the ceive
ficantplacebo firstinand
difference. groupthena active treatment
13 of the patients
cal function,
womac score and of joint patients’
pain atglobal the 5%assessment
level of sig- (group a, n=547)
were taking or 18
nsaids, active treatment 10
paracetamol, firstsynthe-
and
of diseasestatistical
nificance. severity and occurrence
analysis was based of adverse
on the then placebo
tic opioids such(group b, n=547).
as tramadol There
and were no
codeine, and
events were secondary outcome
intention-to-treat principle with last obser-vation measures. significant differences in gender or
19 did not use any rescue medication at all.age on com-
carried forward. The wilcoxon test for matched in group b the corresponding numbers of pa-
 23
09
ti-ents were: nsaid 15, paracetamol 21, syn- was 92.5% with hyben-vi-tal and 90.5% with pla-
paringopioids
thetic the a and b groups
6, and (data not given).
no medication at all 17.in opioids 6, and no medication at all 17. These
cebo.
the entire group the mean body
These values were not significantly differentmass index values
Primarywere
outcomenot significantly
measure different from the
(bmi) was
from kg/m2 (range
the 27values reported 19–41).
in ingroup
group a. a values
womac scores forgroup
reported in a. There
joint pain, wasentire
for the no signifi-
study
the bmi
There waswas 27.3 kg/m difference
no significant
2
(range 19–39)
in theand num-in cant difference
population, are in the number
given in Tableof2.patients
after 3 drop-
weeks
group b, 26.6 kg/m 2
(range 22–41),
ber of patients drop-ping out of the study a non-signi- ping
of out of
active the studywomac
treatment, when comparing the two
scores for joint pain
ficant comparing
when difference. inthegroup a 13 of the
two different patients
treatments different treatments or the a and b
declined from 33.7±19.4 to 29.4±18.3, a delta groups (for
were taking nsaids, 18 paracetamol,
or the a and b groups (for details see figure 10 synthe- 1). details see figure 1). There were no
reduction of 7.4±14.9 mm (pv0.001), compared significant
tic opioids
There were such as tramadol
no significant and codeine,
differences and
in osteoar- differences
to a change in from
osteoarthritic
33.7±19.4characterization on
to 35.3±21.5, a del-
19 didcharacterization
thritic not use any rescue medicationthe
on comparing at aall.
andin comparing the a and b groups, as
ta change of 2.1±16.8 (pv0.299), when placebo detailed in
group b the corresponding numbers
b groups, as detailed in Table 1. compliance of pati- Table 1. compliance was 92.5% with
treatment was given (Table 2). The change com- hyben-vi-
ents were: nsaid 15, paracetamol 21, synthetic tal and 90.5% with placebo.

table 1. characterization of osteoarthritis.

all patients Placebo–active active–Placebo p-value Pa vs. aP
(n=594) (n=547) (n=547)
knee osteoarthritis 58 29 29 29
hip osteoarthritis 21 11 10 10
hip and knee osteoarthritis 15 7 8 8

Initial Womac scores

pain 33.7 (19.4) 30.4 (18.1) 37.0 (20.4) 37.0 (20.4)
stiffness 39.2 (19.4) 35.6 (22.0) 42.5 (26.2) 42.5 (26.2)
adl 35.3 (21.6) 34.0 (21.1) 36.7 (22.2) 36.7 (22.2)
pgad 43.9 (24.4) 43.6 (22.6) 44.3 (26.8) 44.3 (26.8)

table 2. Womac scores for pain, stiffness, daily activities (adl), and patients’ evaluation of disease
severity (Pgad) in all the included patients (n594). data given are mean values with sd in parentheses.
p-value
placebo vs. active
start 3 weeks delta value 3 months delta value 3 weeks 3 months
pain placebo 33.7 (19.4) 35.3 (21.5) 2.1 (16.8) 2.1 (16.8) 5.1 (18.3
)1
0.014 0.125
active 33.7 (19.4) 29.4 (18.3) 7.4 (14.9)2 7.4 (14.9)2 7.0 (19.7)3
stiffness placebo 39.2 (24.4) 40.0 (24.2) 3.3 (19.0) 3.3 (19.0) 5.0 (23.2) 0.198 0.038
active 39.2 (24.4) 34.0 (20.5) 7.5 (16.7)4 7.5 (16.7)4 8.0 (21.6)5
adl placebo 35.3 (21.6) 39.7 (25.3) - 20.7 (221.4) - 20.7 (221.4) - 20.2 (25.5) 0.165 0.018
active 35.3 (21.6) 35.9 (27.7) 2.2 (22.7)6 2.2 (22.7)6 6.4 (17.5)7
pgad placebo 43.9 (24.4) 42.3 (21.2) 8.2 (25.1)8 8.2 (25.1)8 7.8 (28.8)9 0.682 0.035
active 43.9 (24.4) 39.2 (22.4) 8.2 (22.6)10 8.2 (22.6)10 14.1 (28.1)11

1pv0.005, 2pv0.001, 3pv0.003, 4pv0.001, 5pv0.006, 6pv0.002, 7pv0.003, 8pv0.004, 9pv0.031, 10pv0.002, 11pv0.001. the p-values
givenare relative to pretreatment values (initial values).

Primary outcome measure ment was significantly higher when active treat-
paring the two different groups was statistically design, that the intake of nsaids was unchan-
ment was given (82%) than when placebo was
significantscores
womac at theforp<0.014
joint pain, for the
level. entire
after study
3 months ged during the two different treatment periods
population, given (49%) (p<0.004) (figure 2). after 3 months
of treatment,are thegiven
same in Table
pattern 2. after
was 3observed,
weeks of (p<0.803) (data not given). a decline of 40% in
of treatment, the percentages of responders in
although the changes were not statisticallypain
active treatment, womac scores for joint sig- the consumption of paracetamol (data availa-
declined from 33.7±19.4 to 29.4±18.3, a delta the two groups, although still in favour of active
nificant (p<0.125). The percentage of patients ble in 21 patients) was observed as a result of
reduction of 7.4±14.9 mm (pv0.001), compared treatment, did not differ significantly.
experiencing a reduction in the womac score active treatment (p<0.052).
to a
for change
joint pain from
after 33.7±19.4
the initialto335.3±21.5,
weeks of atreat-
del-
ta change of 2.1±16.8 (pv0.299), when
ment was significantly higher when active treat- placebo diaries of the consumption of ‘rescue medica-
when a mann–whitney subanalysis was applied
treatment
ment was given
was given (82%) (Table
than2). The change
when placebo com-
was on theindicated,
tion’ consumptionin accordance with the
of paracetamol studyof
in each
paring the two different groups was
given (49%) (p<0.004) (figure 2). after 3 months statistically design, that the intake of nsaids was unchan-
the two groups, during the first 3 month of treat-
significant
of treatment, at the
the percentages
p<0.014 level.ofafter 3 months
responders in ged during
ment, thepowder
rosehip two different treatment
resulted periods
in a significant
of treatment, the same pattern was
the two groups, although still in favour of active observed, reduction in the number of tablets taken40%
(p<0.803) (data not given). a decline of in
during
although the
treatment, didchanges
not differwere not statistically sig-
significantly. the consumption of paracetamol (data
a 2-week period (14.0±24.0; p<0.031) compared availa-
nificant (p<0.125). The percentage of patients blean
to in 21 patients) was
insignificant observed
increase as a result
of 7.9±15.5 of
tablets
experiencing
diaries of the a reduction in of
consumption the‘rescue
womac score
medica- active treatment
observed (p<0.052).
as a result of placebo treatment. The
for joint pain after the initial 3
tion’ indicated, in accordance with the study weeks of treat- between-group difference was 51% (p<0.027).
10
10
24
when a mann–whitney subanalysis was applied (p<0.035) when active treatment was given,
on when a mann–whitney
the consumption subanalysis was
of paracetamol applied
in each of (p<0.035)
as compared when active treatment
to placebo. The werewas given,
no signifi-
The
on consumption
the consumption of weak
of opioids
paracetamol
the two groups, during the first 3 month of treat- (data in availa-
each ofsed,
as separately,
compared the
to group
placebo.
cant difference in the alleviation of symptoms initially
The treated
were no with
signifi-
ble
ment, in
the two only seven
groups,
rosehip patients)
during resulted
powder showed
the first 3inmonth a similar re-
of treat-
a significant placebo
cant difference
comparing and then
patients actively
in the with treated
alleviation
osteoarthritis (group
of symptoms
of the a)
duction
reduction ininthe
ment, rosehip theconsumption
powder
number resulted during
of tablets active
intaken treat-
a significant
during and the
comparing
hip group
to patients initially
patients given active
with osteoarthritis
with osteoarthritis treatment
of the of knee. the
ament
2-week (p<0.0313)
reduction period (data not
in the(14.0±24.0;
number given).taken
ofp<0.031)
tablets as relatively
compared during and
hip then
to placebo
patients with(group
as a carry-over effect can blunt the impact of b).
osteoarthritis group ofa showed
the knee.
tofewer
a 2-weekpatients
an insignificant were
period (14.0±24.0;taking weak
increase p<0.031) opioids,
of 7.9±15.5 a sua-
compared
tablets atreatment
significant
as a carry-over inimprovement effect can
a crossover in activities
design, bluntwe the of
also daily
impact
analy- li-of
nalysis
observed was not
to an insignificant performed
as a resultincrease
of placebo on weak
of 7.9±15.5 opioids.
treatment. tablets
The ving (adl)
treatment
sed, function
separately, in a crossover and a
the groupdesign,reduction
initiallywe in patients’
also analy-
treated with
secondary
observed asoutcome
a result measures
of placebo
between-group difference was 51% (p<0.027). treatment. The overall feeling of discomfort
placebo and then actively treated (groupwith
sed, separately, the group from
initially their
treateddisease a)
Thebetween-group
consumption difference of weak opioids was 51% (data (p<0.027).
availa- patients
placebo
and theglobal and assessment
group then
initially actively
given oftreated
disease
active severitya)
(group
treatment
womac
ble onlyscores
Theinconsumptionseven for of stiffness,
weak showed
patients) limitation
opioids (data ofavaila-
a similar physi-
re- (pgad)
andthen
and theafter group
placebo3 weeks initially and
(group given 3 months
b). activea
group of active
treatment
showed
cal
duction function,
ble in onlyin the sevenand patients’
patients) during
consumption global
showed assessment
a similar
active re-
treat- treatment.
aand The
then placebo
significant impact
improvement on pain
(groupinb). and
group of
activities stiffness,
a showed
daily alt-li-
of disease
duction in severity
the for
consumption the
ment (p<0.0313) (data not given). as relatively entire
during study
active popu-
treat- hough present, did not attain
ving (adl) function and a reduction in patients’li-
a significant improvement in statistical
activities of signifi-
daily
lation are given (data in Table 2.given).
after 3asmonths ofcance (Table 3).patients and in group
mentpatients
fewer (p<0.0313) were taking notweak opioids, relatively
a sua- ving (adl)
overall feeling functionof discomfort from b
a reduction showed
theirin patients’
disease a
treatment,
fewer waspatientsthere was a
were takingon significant
weak reduction
opioids, a sua-instatistically
overall feeling sig-nificant reduction in pain, stiff-
nalysis not performed weak opioids. patients global of discomfortoffrom
assessment disease theirseverity
disease
womac symptom scores
nalysis was not performed on weak opioids.for stiffness (p<0.037), ness, and pgad asassessment
a result of active treatment.
(pgad) after 3 weeks and 3 months of severity
patients global of disease active
womac scores for limitation of physical func- These
(pgad) changes, after however, did 3not returnof toactive
pre-
figure 2. Percentage of patients experiencing a treatment. The 3impact weekson and pain months
and stiffness, alt-
treatment
treatment. levels
The impact during thepain following placebo
figure 2. in
reduction Percentage
the Womac of score
patients for experiencing
joint pain aftera hough present, did noton and stiffness,
attain statistical alt-
signifi-
treatment
hough period,did
present, suggesting
not attain carryover
statistical (Table
signifi-
3reduction
weeks ofintreatment
the Womac in thescore
groupfor joint
initiallypain after
given cance (Table 3).
3).cance
a comparison (Table 3).of the a and b groups regard-
3 weeksand
placebo of treatment
in the group in the group
initially giveninitially
activegiven ing pain and stiffness yielded mann–whitney
placebo and in the group initially given active
treatment. patients in group b showed a statistically sig-
p-values of 0.001 and 0.016, respectively, when
treatment. patientsreduction
nificant in group inb pain, showed a statistically
stiffness, and pgad sig-
evaluating after the initial 3 weeks of treatment.
nificant reduction in
as a result of active treatment. These changes,pain, stiffness, and pgad
100 – although this com-parison between groups
as a result
however, did of notactive returntreatment.
to pretreatment These changes, levels
100 – was still also in favour of active treatment after
however,
during the did not return
following placebo to pretreatment
treatment levels
period,
the first 3 months of treat-ment, statistical sig-
during
suggesting the following
carryover placebo
(Table 3). treatment
a further
comparison period, of
nificance was not obtained and statisti-
suggesting
the a and b carryover
groups (Table 3).
regarding paina comparison
and stiffness of
50 – cally significant changes in womac parameters
the a and
yielded b groups regarding
mann–whitney p-values pain of and
0.001 stiffness
and
50 – were not observed whencomparing the initial
yielded
0.016, mann–whitney
respectively, when p-values of 0.001 and
3-month periods of the two evaluating
different treatments. after the
0.016,3 weeks
initial respectively, of treatment. when evaluating
although after the
forthis com-
49 % 82 %
an identical pattern as described womac
49 % 82 % initial
parison 3 weeks
between of treatment.
groups
data was also ob-served for rescue medication was although
still also in this
favour com- of
0
0 (data not given). There was no significant differ-of
parison
active between
treatment groups
after the was
first still
3 also
months in favour
of treat-
active
ment,
ence treatment
instatistical
dropout after
significance
rate the first
or milder was 3 months
unwanted not obtainedof treat-
sideef-
secondary outcome measures ment,
and
fects statistical
further
reported significance
statistically
during significant
treatment was (Tablenot obtained
changes
4). in
secondary
womac outcome
scores for measures
stiffness,
tion improved (p<0.018), whereas patients’ limitation of physi- and
womac furtherparametersstatistically were significant
not observed changes when in
womac
cal function,scores
and for stiffness,
patients’ limitation
global
global assessment of disease severity declined of
assessment physi- womac parameters
comparing the initial 3-month periods of the were not observed when
ofcal function,
disease and
severity patients’
for the
0(p<0.035) when active treatment was given, global
entire assessment
study popu- comparing
two different the initial 3-month
treatments. an identical periodspattern of the
of
lation disease
are severity
given in for
Table the2.
as compared to placebo. The were no signifi- entire
after 3study
months popu-
of two
as different
described treatments.
for womac an
data identical
was pattern
also ob-
lation
treatment, are given
there in
was Table
a 2. after
significant
cant difference in the alleviation of symptoms 3 months
reduction of
in as described for womac
served for rescue medication (data not given). data was also ob-
treatment,
womac
comparing there
symptom
patientswas
scores a significant
withfor reduction
stiffness (p<0.037),
osteoarthritis of the in served
There was for no rescue medication
significant difference (datainnot given).
dropout
womac
womac symptom
scores for scores
limitation for
hip to patients with osteoarthritis of the knee.stiffness
of (p<0.037),
physical func- There
rate or was
milder no significant
unwanted difference
side effects in dropout
reported
womac
tion
as a improvedscores(p<0.018),
carry-over for limitation
effect of physical
whereas
can blunt func-
patients’
the impact of rate ortreatment
during milder unwanted (Table 4). side effects reported
tion
global
treatment improved
assessment (p<0.018),
of disease
in a crossover whereas
design,severity patients’
declined
we also analy- during treatment (Table 4).
global assessment of disease severity declined
table 3. Womac scores for pain, stiffness, daily activity (adl), and patients’ evaluation of disease
table 3.(Pgad)
severity Womac in scores
group a for(placebo
pain, stiffness,
first, thendaily activity
active (adl), and
treatment) andpatients’
in groupevaluation
B (active treatment of disease
severity
first, then (Pgad)
placebo). in group
data given a (placebo
are mean first,values
then active
with sdtreatment)
in parentheses. and in group B (active treatment
first, then placebo). data given are mean values with sd in parentheses.
Initial value 3 weeks delta value 3 months delta value 3 weeks delta value 3 months delta value
Initial value 3 weeks delta value 3 months delta value 3 weeks delta value 3 months delta value

group a (n=547)
pain
group a 30.4(n=547)
(18.1) 34.5 (23.1) - 22.5 (13.6) 36.3 (20.4) 2.3 (14.9) 29.9 (17.7) 5.1 (15.6)1 31.9 (23.4) 5.9 (21.9)
stiffness
pain 35.6
30.4(22.0)
(18.1) 37.1
34.5(25.9)
(23.1) - 21.4
- 22.5(19.3)
(13.6) 38.0
36.3(23.6)
(20.4) 3.2
2.3(22.8)
(14.9) 31.4
29.9(19.0)
(17.7) 5.95.1
(19.2) (2)
(15.6)1 33.8
31.9(25.5)
(23.4) 6.85.9
(21.9) (3)
(21.9)
adl
stiffness 34.0
35.6(21.1)
(22.0) 36.1
37.1(22.3)
(25.9) - 20.3
- 21.4(10.5)
(19.3) 38.3
38.0(20.3)
(23.6) 2.5
3.2(14.7)
(22.8) 33.5
31.4(17.6)
(19.0) 5.3 (13.8) 4
5.9 (19.2)(2) 33.0 (23.0)
33.8 (25.5) 7.6 (19.9) 5
6.8 (21.9)(3)
pgad
adl 43.6
34.0(22.6)
(21.1) 40.2
36.1(22.3)
(22.3) - 8.1
20.3(23.5)
(10.5) 48.9 (25.5)
38.3
6
(20.3) 5.2
2.5(29.6)
(14.7) 41.1
33.5(21.5)
(17.6) 9.6
5.3(25.5)
7
(13.8)4 38.1
33.0(22.9)
(23.0) 15.3
7.6(28.6)
8
(19.9)5
pgad 43.6 (22.6) 40.2 (22.3) 8.1 (23.5) 48.9 (25.5)6 5.2 (29.6) 41.1 (21.5) 9.6 (25.5)7 38.1 (22.9) 15.3 (28.6)8
group B (n=547)
pain
group B 37.0(n=547)
(20.4) 28.9 (19.0) 9.6 (13.9)9 33.8 (17.6) 8.1 (17.4)10 36.0 (20.0) 7.0 (18.5)11 34.9 (20.6) 7.8 (20.9)12
stiffness
pain 42.5
37.0(26.2)
(20.4) 36.2
28.9(21.7)
(19.0) 8.99.6
(14.4) 13
(13.9)9 39.8
33.8 (17.6) 9.2
(21.6) 8.1(21.4)
14
(17.4)10 42.8
36.0 (20.0) 7.8
(22.4) 7.0(18.0)
15
(18.5)11 44.0
34.9(24.5)
(20.6) 7.8
7.8 (18.0)
15
(20.9)12
adl
stiffness 36.7
42.5(22.2)
(26.2) 38.0
36.2(34.2)
(21.7) - 20.4 (28.2)13
8.9 (14.4) 37.0
39.8(18.1)
(21.6) 5.39.2
(15.0) (17)
(21.4)14 43.6
42.8(27.9)
(22.4) - 21.2 (29.1)15
7.8 (18.0) 45.3
44.0(32.7)
(24.5) - 21.2 (29.1)15
7.8 (18.0)
pgad
adl 44.3
36.7(26.8)
(22.2) 37.6
38.0(23.3)
(34.2) 6.8 (19.4)
- 20.4
18
(28.2) 44.4
37.0(39.3)
(18.1) 12.6
5.3 (28.0)
19
(15.0)(17) 44.5
43.6(20.2)
(27.9) - 9.3
21.2(27.4)
(29.1) 41.5
45.3(19.3)
(32.7) - 9.3
21.2(27.4)
(29.1)
pgad 44.3 (26.8) 37.6 (23.3) 6.8 (19.4)18 44.4 (39.3) 12.6 (28.0)19 44.5 (20.2) 9.3 (27.4) 41.5 (19.3) 9.3 (27.4)

p<0.042, 2p<0.076, 3p<0.095, 4p<0.002, 5p<0.025, 6p<0.018, 7pv0.022, 8p<0.001, 9p<0.001, 10p<0.011, 11p<0.031, 12p<0.012,
1

p<0.001, 14p<0.022, 15p<0.037, 16p<0.084, 17p<0.068, 18p<0.044, 19p<0.010, 20p<0.049. P-values given are relative to pretreat-
13

ment values. P-values given in parantheses indicate borderline significance.

p<0.042, 2p<0.076, 3p<0.095, 4p<0.002, 5p<0.025, 6p<0.018, 7pv0.022, 8p<0.001, 9p<0.001, 10p<0.011, 11p<0.031, 12p<0.012,
1

p<0.001, 14p<0.022, 15p<0.037, 16p<0.084, 17p<0.068, 18p<0.044, 19p<0.010, 20p<0.049. P-values given are relative to pretreat-
13

ment values. P-values given in parantheses indicate borderline significance. 25

table 4. dropout rate and unwanted effects in 3ceiving

weeksplacebo.
of treatment, may explain
we suggest that thisthechange
lack of
patients after 3 months while on placebo or ac- significance
in consumption when pain was evaluated
of additional painkillers, after
which3
tive treatment. months of treatment. furthermore, the powder
patients were allowed to reduce after the first
was well tolerated and did not give rise to any
3 weeks of treatment, may explain the lack of
serious adverse effects; in fact, stiffness and glo-
Pla- ac- p- significance when pain was evaluated after 3
cebo tive value bal assessment of disease severity significantly
months of treatment. furthermore, the powder
declined and daily activities significantly impro-
was well tolerated and did not give rise to any
Dropped out during treatment 7 7 ns ved after 3 months of active treatment. our re-
serious adverse effects; in fact, stiffness and glo-
sults are supported by the findings in a recent
Reasons for dropout 2 0 ns bal assessment of disease severity significantly
felt that treatment did not work 3 3 ns norwegian study in which treatment with pow-
declined and daily activities significantly impro-
for personal reasons 1 1 ns der from the same subtype of rosehip resulted
acid regurgitation 0 1 ns ved after 3 months of active treatment. our re-
difficulty to swallow capsules 0 1 ns
in improved joint mobility and less joint pain in
sults are supported by the findings in a recent
started prednisolone treatment 0 1 ns patients on a waiting list for either hip or knee
intercurrent surgery 1 0 ns norwegian
surgery duestudy in which treatment
to osteoarthritis (13). with pow-
insisted on knowing kind of treatment der from the same subtype of rosehip resulted
Milder unwanted effects reported during
in improved
There joint mobility
are, however, and less
reservations to joint pain in
our conclu-
treatment that did not cause withdrawal patients on a waiting list for either
sion. The dose was possibly not optimal and hip or kneea
frequent voiding 1 3 ns surgery due to osteoarthritis (13).
longterm study is needed to confirm that the
diarrhoea 2 2 ns
constipation 1 2 ns reduction in symptoms is persistent, and that
short episode of mild urticaria 0 1 ns There are, treatment
long-term however, reservations
does not result to our in conclu-
side ef-
sion. The dose was possibly
fects different from what was observed not optimal andwith a
longterm study is needed to confirm that the
placebo.
discussion
Discussion reduction in symptoms is persistent, and that
This study
This study shows
shows thatthat a a standardized
standardized rose-hip
rose-hip The presenttreatment
long-term data, however,does not seem result in well
to fit side intoef-
powder, made from a subtype of rosa canina,
powder, made from a subtype of rosa canina, fects different
earlier, more basic, fromreports
what was from observed
our laboratory with
has a
has a beneficial
beneficial symptomatic
symptomatic effect
effect in in patients
patients placebo.
indicating that the present version of rose-hip
with knee
with knee and and hiphip osteoarthritis.
osteoarthritis. The
The percen-
percen- powder, when used in higher doses, reduces
tage of patients who reported at least
tage of patients who reported at least some
some The present
pain data, however,
in osteoarthritis and affects seemmechanisms
to fit well into of
reduction in womac pain after 3 weeks of ac-
reduction in womac pain after 3 weeks of ac- earlier, more to
importance basic,
joint reports
diseasefrom ourit laboratory
(8, 9). is also en-
tive treatment
tive treatment was was82% 82%compared
compared to to
a 49% re-
a 49% indicating to
couraging that
notethethatpresent
in anotherversion of rose-hip
study ai-ming
duction in the group treated with
re-duction in the group treated with placebo. placebo. a powder, when used in higher
to test patients with osteoarthritis, on the doses, reduces
wait-
aplacebo
placeboeffecteffectofofthe thesame
same magnitude
magnitude as as re-
re- painlistinforosteoarthritis
ing and affects mechanisms
hip or knee replacement, the pre-sent
ported here was also reported in a
ported here was also reported in a recent study recent study of importance
powder, given in to ajoint disease
similar dose,(8, 9). it is pain
reduced also
evaluating the
evaluating the impact
impact of of aa ginger
ginger extract
extract onon encouraging to note that in
and improved mobility, suggesting that the pow- another study ai-
pain from osteoarthritis
pain from osteoarthritis of the knee (6).
(6). in that ming
der may to testworkpatients
in both with osteoarthritis,
the early and ate on the
stages
study, which
study, which reported
reported a a 50%
50% reduction in pain waiting
of list for hip
osteoarthritis or knee replacement, the pre-
(13).
during placebo
during placebo compared to a 66% reduc- sent powder, given in a similar dose, reduced
tion during
tion during active treatment, early testing was pain and
When improvedtomobility,
responders treatment suggesting
were asked that
also performed.
also performed. The The placebo
placebo impact, in both the powder
about the time maybefore
work insome bothalleviation
the early and ate
of pain
studies, might
studies, might havehave declined
declined if the studies had stages of osteoarthritis
occurred, the earliest (13). response reported was
been running
been running for for aa longer
longer period
period of time. within 2 weeks. moreover, a certain carry-over
effect
when was demonstrated
responders in the present
to treatment were asked study,
we used
We used a validated, disease-specific, and and
about carry-over
the time beforewas also somedemonstrated
alleviation of in pain
ano-
very sensitive
very sensitive questionnaire and were able to ther study using
occurred, the same
the earliest rosehipreported
response powder was and
demonstrate
demonstrate a a reduction
reduction in joint pain and stiff-
stiff- an identical study design (14).
within 2 weeks. moreover, a certain carry-over This may indicate
ness as
ness as well
well as an improved physical function that
effect the was present powder does
demonstrated in the not work like
present study,the
in
in these
these patients
patients after treatment with the pre- traditional
and carry-over painkillers
was also normally used in the
demonstrated trea
in ano-
sent
sent rose-hip
rose-hip powder. pain pain that was significant- ment
ther studyof osteoarthritis.
using the same as reported
rosehip powder earlier, and one
ly
ly reduced
reduced afterafter 3 weeks of treatment did not mode of action might be an
an identical study design (14). This may indicateanti-inflammatory
attain significancewhen
attain statistical significance whentested
testedafter
af- action
that the mediated
present powder by leucocyte
does not work neutrophils
like the
ter 3 months.
3 months. duringthe
during thecourse
courseofof the
the 3-month (8, 9). indeed,
traditional we were
painkillers able to used
normally showin thattheC-rea
trea
treatment
treatment periodperiod in which the patients recei- tive protein and also the chemotaxis
ment of osteoarthritis. as reported earlier, one of neutr
ved
ved active treatment, there was, was, however,
however, a phil
mode leucocytes
of actionwere mightdecreased in vitro as well
be an anti-inflammatory
significant
significant reduction
reduction in the consumption of tra- as in vivo,
action using concentrations
mediated of the present
by leucocyte neutrophils (8,
ditional
ditional painkillers
painkillers such as paracetamol and subtype of rose-hip, comparable
9). indeed, we were able to show that c-rea to the dose
synthetic
synthetic opioids
opioids as compared to the group re- used in this and
tive protein studyalso (8, 9).
thechemotaxis
chemotaxisof ofleuco-
neutr
ceiving placebo. we suggest that this change cyte neutrophils also significantly declined
in consumption of additional painkillers, which when measured in a subfraction of the present
patients were allowed to reduce after the first patients (15). it seems likely therefore that one
26
mechanism of the present powder might be of
an anti-inflammatory origin. indeed, the anti-
inflammatory hypothesis seems to be receiving
increased attention. we have shown that the
anti-inflammatory impact of the present sub-
type of rose-hip was not related to vitamin c
and suggested that another possibly unknown
active ingredient might be found in the rosehip
powder (8). an active ingredient that can inhibit
the chemotaxis of human neutrophil leucocytes
has been isolated recently from the present sub-
type of rose-hip, making the anti-inflammatory
hypothesis more likely (16). a framingham stu-
dy and, more recently, a danish study indicate
that patients with osteoarthritis might benefit
from vitamin c (17, 18). as the present powder
is rich in natural vitamin c, an additional me-
chanism might be of vitamin c origin.
The powder does not seem to be involved in
the arachidonic acid pathway as platelet ag-
gregation was not affected when the powder
was tested in healthy volunteers and patients
on warfarin treatment (19).
This is different from the anti-inflammatory agents
referred to in the introduction of this paper and
we suggest that this might help to explain why
side effects in this study were comparable to
that of the placebo.
In summary, we suggest that the present stan-
dardized powder, made from a subtype of rosa
canina, can alleviate pain to an extent that can
influence the consumption of rescue medica-
tion. it should be emphasized that the present
data may not apply to any type of rose-hip, as
species can be different regarding biological
activity (20). further research should aim to find
the optimal dose, test the impact of long-term
treatment and compare that with the impact
of nsaids, and evaluate the biological activity
of different subtypes of rosehip.
References
1. Reginster Jy, deroisy l, rovati lc, lee RL,
Lejeune e, bruyere o, et al. long-term
effect of glucosamine sulphate on oste-
oarthritis progression: a randomised,
placebo-controlled clinical trial. lancet
2001;357:251–6.
2. Hochberg mc, altman rd, brandt kd,
Clark bm, dieppe pa, griffin mr, et al.
Guidelines for the management of
osteoarthrosis (parts 1 and 2). 2. arthritis
Rheum 1985;38:1535–46.
3. Vane Jr, botting rm. anti-inflammatory
drugs and their mechanism of action.
Inflamm res 1998;47:578–87.
4. Mukherjee d, nissen se, Topol eJ. risk of
cardiovascular events associated with
Selective cox-2 inhibitors. J am med assoc
2001;286:954–9.
5. Rodriques lac, hernandes-diaz s. relative
risk of upper gastrointestinal complications
among users of acetaminophen and non-
steroidal anti-inflammatory drugs. epide-
miology 2001;12:570–6.
6. Altman rd, markussen kc. effect of ginger
extract on knee pain in patients with oste-
oarthritis. arthritis rheum 2001;44:2531–8.
7. Appelboom T, schuermans J, verbruggen
G, henrotin y, reginster Jy. symptoms
modifying effect of avocado/soybean
unsaponifiables (asu) in knee osteoarth-
ritis. scand J rheumatol 2001;30:242–7.
8. Winther k, rein e, kharazmi a. The anti-in
flammatory properties of rose-hip. inflammo
pharmacology 1999;7:63–8.
9. Kharazmi a, winther k. rose-hip inhibits che-
motaxis and chemiluminescence of human
blood neutrophils in vitro and reduces
certain inflammatory parameters in vivo.
Inflammopharmacology 1999;7:377–86.
10. Altman R, asch e, bloch d, borenstein d,
brandt k, et al. development of criteria for
the classification and reporting of
osteoarthritis. classification of osteoarthritis
of the knee. arthritis rheum 1986;29:1039–49.
11 Altman R alarcon g, appelrouth d. The
American college of rheumatology criteria
for the classification and reporting of
osteoarthritis of the hip. arthritis rheum
1991;34:505–14.
12. Bellamy n, buchanan wb, goldschmidt ch,
Campbell J, stitt lw. validation study of
womac: a health status instrument for
measuring clinically important patient
relevant outcomes to antirheumatic drug
therapy in patients with osteoarthritis of the
hip and knee. J rheumatol
1988;15:1833–40.
13. Warholm o, skaar s, hedman e, mølmen
hm, eik l. The effects of a standardised
herbal remedy made from a subtype of
rosa canina in patients with osteoarthritis:
a double-blind, randomised, placebo-con-
trolled clinical trial. curr Therap res
2003;64:21–31.
14 Rein e, kharazmi a, winther k. a herbal
remedy, hyben vital, reduces pain and
improves general wellbeing in patients
with osteoarthritis – a double-blind,
placebo-controlled, randomised trial.
phytomedicine 2004;11:383–91.
15. Thamsborg gm, apel k, rein e, winther k. a
 27

randomized, doubleblind, placebo- 1996;39:648–56.

controlled, crossover trial of the herbal
remedy hybenvital in patients with
osteoarthritis. Thu0214 eular, stockholm,
Sweden, 10–15 June 2002.
16. Larsen e, kharazmi a, christensen lp,
Christensen sb. an antiinflammatory
galactolipid from rose-hip (rosa canina)
that inhibits chemotaxis of human blood
neutrophils in vitro.
J nat prod 2003;66:994–5.
17. Mcalindon Te, Jacques p, zhang y, hannan
MT, aliabadi p, weissmann b, et al. do
anti-oxidant micronutrients protect against
the development and progression of knee
osteoarthrosis? arthritis rheum
18. Jensen nh. reduced pain from
osteoarthritis in hip joint or knee joint during
treatment with calcium ascorbate. a ran-
domised, placebo-controlled, crossover
trial. ugeskr laeger 2003;165:2563–6.
19. Rein e, kharazmi a, winther k. rose-hip
given as a standardised dry powder exerts
anti-inflammatory properties, without influ-
encing platelet aggregation and the coa-
gulation cascade. 1st international
Congress on heart disease, washington,
May 1999.
20. Brandt K, Aakeson b. health promoting
compounds in vegetables and fruit. plant
prod 2002;29:43–44.

a herBal remedy, hyBen VItal (stand. PoW-

der of a suBsPecIes of rosa
canIna fruIts), reduces PaIn and
ImProVes general WellBeIng In PatIents
wiTh osteoarthrItIs—a douBle-BlInd,
PlaceBo-controlled, randomIsed trIal
E. Reina, A. Kharazmib, K. Wintherc,*
a
Institute for clinical research, kolding, denmark
b
Department of clinical microbiology, university hospital, copenhagen, denmark
c
Department of Clinical Biochemistry, Copenhagen County hospital in Gentofte, university of
Copenhagen,

Abstract
The treatment of osteoarthritis, a disease that eventually affects the majority of the older population,
involves the alleviation of symptoms such as pain and stiffness, and the reduction of inflammation.
The double-blind, placebocontrolled, crossover study reported here examined the effect of hyben
vital, a herbal remedy made from a subtype of rosa canina and recently reported to have anti-
inflammatory properties, on the symptoms of osteoarthritis. one hundred and twelve patients with
osteoarthritis were randomly allocated to treatment with either hyben vital 5 g daily or an identical
placebo for 3 months, followed immediately by the alternative treatment. The patients assessed
changes in joint pain and stiffness after each treatment period on a 5-point categorical scale.
general wellbeing, including mood, sleep quality and energy were also assessed and recorded in a
personal diary.
The results in the two arms of the crossover differed markedly. group a (placebo first) showed sig-
nificantly more improvement from hyben vital than from placebo, po0:0078 for pain ando0:0025
for stiffness. but group b (hyben vital first) revealed a positive effect of the same order as for hyben
vital in group a, not only from the active drug, but also from placebo (difference not significant). an
identical pattern was observed when we evaluated general wellbeing from the diary records. when
patients, on the basis of reduction in joint pain, were divided into responders and nonresponders,
28

the first 3 months of active treatment (group a) showed a response rate of 31/47 (66%) compared to
that of placebo (group b) 18/50 (36%), po0:0185: no major side effects occurred in either group. The
data indicate that hyben vital reduces the symptoms of osteoarthritis. we interpret the marked dif-
ferences in the responses of the two groups as indicating a strong ‘‘carryover’’ effect of hyben vital.
© 2004 elsevier gmbh. all rights reserved.

Introduction
inflammatory cells such as polymorphonuclear leukocytes are known to be causally involved in in-
flammation, in pain and tissue damage. The damage is caused by release of proteo-lytic and hydro-
philic enzymes as well as toxic, reactive oxygen radicals derived from cells activated in the tissues
and joints (harris, 1988). The non-surgical therapy of osteoarthritis, a di-sease that attacks many of
the middle-aged and the majority of the older population, invol-ves alleviation of the symptoms as-

sociated with the disease, such as pain and stiff-
ness, and the reduction of inflammation. ace-
tylsalicylic acid and a range of non-steroidal
anti-inflammato-ry drugs including ibuprophen,
indomethacin and naproxen, as well as gluco-
corticoids, have been used for the treatment of
arthritis (hoch-berg et al., 1995; vane and bot-
ting, 1998). These drugs have a variety of toxic
and unwanted ef-fects, including interference
with haemostasis, gastric erosion and adverse
effects on the liver and kidneys (hochberg et
al., 1995; vane and botting, 1998). selective in-
hibitors of the cyc-looxygenase- 2 system have
recently shown promising analgesic and anti-
inflammatory properties, without the side ef-
fects mentioned. They are, however, still expen-
sive and a nega-tive effect on the circulatory
system cannot be excluded (mukherjee et al.,
2001). recently, acetaminophen was shown to
worsen the risk of upper gastrointestinal compli-
cations (rodrigues and hernandes- diaz, 2001).
There is, therefore, still a need for a safe, low-cost
remedy for the long-term treatment of symp-
toms in osteoarth-ritis. we have earlier shown
that a standardised dry powder, hyben vital,
made from the hips of a particular subtype of
rosa canina, reduced both chemotaxis and the
generation of oxygen radicals in polymorpho-
nuclear cells (winther et al., 1999; kharazmi and
winther, 1999).
The plants used for the current preparation of
hyben vital powder are grown in standardised
fields according to good agricultural practice.
Harvesting takes place when the fruits are ma-
ture and all fruits are brought to freezing facili-
ties without delay. later, the selection of opti-mal
fruits for production of the powder is made by
a laser technique and the temperature of the
subsequent controlled drying process never ex-
ceeds 40 c. The powder contains seeds as well
as husks (rosae pseudofructus cum fruc-tibus)
from a certain subtype of r. canina, by name liTo,
and is finally standardised to contain at least
500 mg vitamin c per 100 g hyben vital pow-
der. further constituents are: pectins (to-tal) 58.0
mg/g, b-carotene 57.9 mg/kg, b-sitos-terol 0.5
mg/g, folic acid 1.6 mg/kg, vitamin e 4.6 mg/100
g, mg 170 mg/100 g, zn 1.0 mg/100 g, copper
10.9 mg/100 g and non-quantified but reported
flavonoids. The antiinflammato-ry effect of the
powder is not related to the well-known high vi-
tamin c content of rose hip extracts (kharazmi
and winther, 1999). but we have earlier shown
that, hyben vital modifies inflammation by re-
ducing both chemotaxis and the generation of
oxygen radicals in poly-morphonuclear white
cells (winther et al., 1999; kharazmi and winther,
1999). moreover, many volunteers have claimed
that pain from oste-oarthritis was diminished
after a few weeks of treatment with the pow-
der (winther et al., 1999; kharazmi and winther,
1999). all these findings encouraged us to inves-
tigate whether hyben vital, in a larger controlled
trial as now reported, would affect pain, stiffness
and general wellb-eing and the consumption
of pain-reducing medicines, in particular par-
acetamol and the synthetic opioid Tramadol, in
patients with os-teoarthritis.
Methods
Patients
after we had obtained approval for the trial
from the local ethical committee, 125 cauca-
sian out patients were enrolled through adver-
tisements in local newspapers. The study was
performed according to good clinical practice
and designed to accord, as far as possible, with
the guidelines on conduct of clinical trials on
osteoarthritis devised by the osteoarthritis re-
search society international. The only notable
exception was that the study included patients
with arthritis of various joints instead of confin-
ing it to a single joint (altman et al., 1996). The
vo-lunteers all gave their oral and written in-
formed consent. They had all been earlier diag-
nosed by their own general practitioner or local
rheu-matologist as suffering from osteoarthritis,

and were reported to have an x-ray verified
dia-gnosis and symptoms of primary osteoar-
thritis in the hip, knee, hand, shoulder or neck,
or some combination of these, for at least the
last 12 months. all reported pain of the affected
joints of at least mild to moderate severity. we
exclu-ded patients with liver or kidney disease
and those known to suffer from allergy or a his-
tory of drug or alcohol abuse. we also excluded
patients with cancer, rheumatoid arthritis, fibro-
myalgia, gout, serious cardiovascular disease,
asthma requiring treatment with steroids, and
any other disease which would substantially in-
fluence the patients’ quality of life. likewise, we
excluded those who had received intra-articu-
lar hyaluronate, glucosamine sulphate, immuno-
suppressive drugs such as gold or peni-cillamine
or injections of glucocorticoids within the 6
weeks prior to the study, and patients who were
found to be unable to co-operate after the first
evaluation.
Trial design
The trial was of a double-blind, placebo-con-
trolled, crossover design, and randomisation of
treatment allocation was performed in blocks
of four with the block size unknown to the in-
vestigators. The design had three immediate-
ly successive periods: a 14 days run-in period
followed by randomised allocation of the two
treatment periods of 3 months each. The two pri-
mary efficacy parameters were: change in joint
pain and the alteration of consumption of con-
comitant ‘‘rescue’’ medication for allevi-ating
pain, evaluated after each of the two, blinded,
3-months treatment periods. The three second-
ary efficacy parameters were: joint stiff-ness,
general wellbeing including mood, ener-gy and
sleep quality, and a subjective overall evalua-
tion of preference for one or other of the study
medications. The run-in period was intended
primarily for patients to become ac-customed
to the ideas of the trial, and to be instructed in
and practise the daily subjective assessment/
record-keeping required, rather than as a for-
mal ‘‘baseline’’. however, we took the opportuni-
ty during this period to measure blood pressure
and removed a routine blood sample for mea-
surement of haemoglobin, cre-atinine, sodium
and potassium, blood glucose and cholesterol.
The patients were then ran-domly allocated, in
blocks of four, by a compu-ter-generated allo-
cation schedule, to receive capsules containing
either a biologically stan-dardised rose hip pow-
der (hyben vital) or an identical placebo. The
capsules were kept in numbered containers. The
daily dosage was five 0.5 g capsules a.m. and
p.m. one or other of the responsible investiga-
tors enrolled all pati-ents. The patients as well as
the research team were kept blind throughout
29
the study. after 3 months, the groups switched
immediately to the alternative treatment for a
further 3 months. immediately after each of the
two treatment periods, a further routine blood
sample was ta-ken and blood pressure was
measured. when the trial had been completed,
all data were entered onto the spreadsheet, af-
ter which the treatment code was broken and
patients were separated into two groups ac-
cording to the treatment sequence they had
received. it transpired that group a started out
with 56 ran-domised patients who took placebo
first, follo-wed by hyben vital, while group b com-
prised the same number of randomised patients
who took hyben vital first, followed by placebo
(fig. 1). The data for the two groups separately
were also entered on the spreadsheet, which
was then mailed to the statistician, who was also
kept blind as to the treatment code.
Methods of assessing clinical effect
Primary efficacy parameters
The cardinal item of information obtained was
the end-of-treatment subjective assessments of
any changes in pain that had occurred during
each of the treatments. These were estimated
by the patients on a 5-step categorical scale
ranging from 0 (no change) to 4 (almost total
relief of pain). here, the higher the score the
greater the clinical benefit, a rise of 1 category
representing 25% improvement. This technique
also allowed us to calculate the number of re-
sponders and non-responders in each group.
Each type of ‘‘rescue’’ analgesic consumed
was noted daily by the volunteers in a diary. All
patients taking nsaids regularly on prescription
from their general practitioners were advised
to continue such treatment, without any
change in dosage, throughout the study.
Three weeks into each of the two treatment
periods we recom-mended the patients to re-
duce their consump-tion of concomitant pain-
relieving medicine, if at all possible. consump-
tion of such medication was recorded daily in
a diary, and at the end of each 3-months treat-
ment period we calculated the consumption of
each type of non-trial pain-relieving medicine.
as patients normally use a wide range of rescue
medications, we simpli-fied accounting of them
by transformation into paracetamol equiva-
lents, as devised by the danish health authori-
ties (lægemiddelkata-loget, 2002). Thus 25 mg
of Tramadol and 25mg of codeine would be
considered as equivalent to 1000 mg paraceta-
mol and aspirin would be considered equal to
paracetamol.
Secondary efficacy measures
The patients made a subjective assessment of
16
30
was recorded daily in a diary, and at the end of secondary efficacy measures
joint stiffness at the end of each treatment peri- creasing disability. an average of each kind of
each 3-months treatment period we calculated The patients made a subjective assessment of
od, on a 5-step categorical scale ranging from measurement was taken for statistical compari-
the consumption of each type of non-trial pain- joint stiffness at the end of each treatment peri-
0 (no change) to 4 (almost total relief of the son of treatments.
relieving medicine. as patients normally use a od, on a 5-step categorical scale ranging from
symptom), as devised for pain. in addition, the
wide range of rescue medications, we simpli- 0 (no change) to 4 (almost total relief of the
patients made a daily subjective assessment of Patients’ overall evaluation of the study
fied accounting of them by transformation into symptom), as devised for pain. in addition, the
the severity of joint pain (in the morning and la- Medication
paracetamol equivalents, as devised by the patients made a daily subjective assessment of
ter in the day), stiffness (in the morning and later On the final day of the trial, before the treat-
danish health authorities (lægemiddelkata- the severity of joint pain (in the morning and la-
in the day) and the state of wellbeing, sleep, en- ment code had been broken, the supervising
loget, 2002). Thus 25 mg of Tramadol and 25mg ter in the day), stiffness (in the morning and later
ergy, and mood was recorded by the pati-ent in physician asked this question of the patient: Tak-
of codeine would be considered as equivalent in the
a diary. each aspect was assessed and ing all day) andinto
aspects theconsideration,
state of wellbeing,
did yousleep,
de-
to 1000 mg paracetamol and aspirin would be energy, and mood was recorded
velop a definite preference for one by the treat-
of the pati-
considered
recorded onequal to paracetamol.
a separate 10-point categorical ent in a
ments, ordiary.
not? each aspect was assessed and
scale, where an increasing score denoted in- recorded on a separate 10-point categorical
fig. 1. flow-chart showing the dropout rate of scale, where
Statistical an increasing score denoted in-
techniques
the different time points of the study. creasing
We based disability. an average
the sample size on of kind an
eachfrom
results of
measurement was taken for statistical compari-
earlier clinical trial using the same dry powder.
125 patients eligible son of from
data treatments.
all the randomised patients were
entered on the spreadsheet. statistical evalua-
1 entered an other study
Patients’
tion was overall
based evaluation of the study
on the intention to treat (iTT),
3 chose other therapy medication
with the last value carried forward. we applied
9 declined to take part on the final
wilcoxon’s testdayforofmatched
the trial, pairs
before the evalu-
when treat-
ment code had been broken, the
ating the study as a simple crossover trial and supervising
112 randomised physician
when we asked
compared this question of the patient:
effects occurring within
Taking all aspects into consideration,
the same group of patients. The mann–whitney did you
n=56 n=56
develop
test a definite
was applied preference of
to comparison forgroups
one of the
a and
9 withdrawals treatments, or not?
b after 3-months treatment. The only exceptions
3 personal reasons
group 3 protocol violation group were simple yes/no questions, to which fisher’s
a 1 high dose steroids B statistical
test techniques
was applied. data given are mean ±sd. any
1 diagnosed cancer we
p based
value equalthetosample
or <0.05size
wasonregarded
results from an
as sta-
n=47 1 acid regurgitation n=50 earlier clinical trial
tistically significant. using the same dry powder.
6 withdrawals
data from all the randomised patients were
3 personal reasons entered on the spreadsheet. statistical evalua-
Results
1 protocol violation tion was based
Description on the intention to treat (iTT),
of patients
1 high dose steroids
1 acid regurgitation
Of the 125 eligiblecarried
with the last value patientsforward. we applied
who responded to
wilcoxon’s
our test for matched
advertisement, we eventuallypairs when
enrolledevalu-
112,
ating the 71
including study
women,as a mean
simpleagecrossover
68 years trial and
(range
outcome measurements
when we
33–93) andcompared
41 men, mean effects
ageoccurring within
64 years (range
35–89)
the same (see flow diagram
group of patients.of fig.
The1).mann–whitney
test was applied to comparison of groups a and
Matching of groups
b after 3-months treatment. The only exceptions
Details are given
were simple yes/no in Table 1. The to
questions, two groups
which were
fisher’s
virtually
test wasidentical
applied.indata their given
demographic
are mean data,
±sd.in
7 withdrawals
the
anyseverity
p valueand equaldistribution
to or <0.05of osteoarthritis
was regardedand as
2 personal reasons in their consumption
statistically significant.of rescue medication; in-
group 2 protocol violation group
a 2 intercurrent surg. B deed,
1 diarhoea resultsthere was no significant between-group
n=43 n=42 difference
descriptioninofany of the 16 items of Table 2. The
patients
mean
of the 125 eligible index
body mass for who
patients the included
responded pati-
to
5 withdrawals
1 personal reasons
ents was 26.9, range 18–42 kg/m2. although
our advertisement, we eventually enrolled 112, only
3 protocol violation 85 patients
including 71 completed
women, mean theage
trial,
68the two
years fi-nal
(range
1 high dose steroids groups of per-protocol patients were
33–93) and 41 men, mean age 64 years (range still not
significantly different. we consider the groups
35–89) (see flow diagram of fig. 1).
therefore to have been very well matched (Ta-
outcome measurements ble 1).
matching of groups
group a: group B: details are given in Table 1. The two groups were
active first Placebo Fifteen patients dropped out before the first 3
virtually identical in their demographic data, in
first months period was finished, leaving 97 patients
the severity and distribution of osteoarthritis and
for the second part of the study and 85 com-
in their consumption of rescue medication; in-
pleted both treatment periods (fig. 1). before
 deed, there was no significant between-group ents, 59 had arthritis of the knee, 46 of the hip,
difference in any of the 16 items of Table 2. The 40 had involvement of the hands, 18 of the neck
mean body mass index for the included pati- and 14 of the shoulder or a combination of the-31
ents was 26.9, range 18–42 kg/m2. although se different joints. The dropouts were correspon-
the
onlycode was broken,
85 patients completeda further 5 were
the trial, exclu-
the two fi- if not otherwise
dingly represented stated.
by allof the
the included
different patients,
joints men-
ded because
nal groups of protocolpatients
of per-protocol violationweredetected
still not 40 wereand
tioned taking
therensaids
wereregularly,
no major40 paracetamol,
disagreements
on evaluation
significantly of the patient’s
different. we consider recordtheform
groupsbe- 12 Tramadol,
between the 3iTTcodeine,
and the2 per-protocol
aspirin, 2 morphine,
analy-
fore the data were entered on the
therefore to have been very well matched (Ta-spreadsheet. and 1 dextropropoxyphen. Thirty of the
sis—hence we refer only to the iTT analysis if not patients
This left 80 patients, 46 women and 34 men, for
ble 1). took no rescue
otherwise stated.medication whatever.
of the included when40a
patients,
a per-protocol analysis. of the randomised pati- subanalysis
were taking of the initial
nsaids values40ofparacetamol,
regularly, the placebo-
ents,
fifteen59 had arthritis
patients of the knee,
dropped 46 of the
out before the hip,
first 40
3 first group (n • 56) versus the active
12 Tramadol, 3 codeine, 2 aspirin, 2 morphine, treatment
had involvement of the hands, 18
months period was finished, leaving 97 patientsof the neck first group (n • 56) were made, there
and 1 dextropropoxyphen. Thirty of the patients were no
and 14 of
for the the shoulder
second part ofor a combination
the study and 85of the-
com- significant differences in body
took no rescue medication whatever. when mass index, age,
a
se different
pleted bothjoints. The dropouts
treatment periodswere
(fig. correspon-
1). before sex, joints involved,
subanalysis consumption
of the initial of nsaid
values of the and
placebo-
dingly represented by all the different
the code was broken, a further 5 were exclu- joints rescue
first groupmedication (Table the
(n ¼ 56) versus 1). active treatment
men-tioned and there were no major
ded because of protocol violation detected on disagree- first group (n ¼ 56) were made, there were no
ments
evaluationbetween
of thethe iTT and
patient’s the per-protocol
record form before Compliance
significant differences in body mass index, age,
analy-sis—hence we refer only to
the data were entered on the spreadsheet. the iTT analysis
This compliance, as calculated
sex, joints involved, consumption fromofthe proporti-
nsaid and
left 80 patients, 46 women and 34 men, for a rescue medication (Table 1).
per-protocol analysis. of the randomised pati-

table 1. Baseline demographic and osteoarthritic characteristics of the study population

Intention-to-treat population Per-protocol-population

Placebo treatment active treatment Placebo treatment active treatment

(n = 56) (n = 56) (n = 39) (n = 56)

age (years) 66.8 ± 11.8 67.1 ± 11.6 67.5 ± 10.6 67.0 ± 10.8
sex
women 34 37 21 26
men 21 20 18 15

bmi (kg/m2) 26.8 ± 5.0 27.7 ± 4.9 27.7 ± 4.9 27.5 ± 5.5
no. of patients with oa of the hip 20 26 15 21
no. of patients with oa of the knee 30 29 24 24
no. of patients with oa of the neck 11 7 10 6
no. of patients with oa of the shoulder 7 7 6 6
no. of patients with oa of the hand 17 23 15 19
no. of patients on nsaids 20 20 17 19
no of patients on paracetamol 21 19 18 16
no. of patients on tramadol 7 5 6 4
no. of patients on codein 1 2 0 0
no of patients on aspirin 1 1 1 1
no of patients on morphine 1 1 1 1
no of patients on dext. ppox. phen. 1 0 0 0
no. of patients with no medication 26 24 19 18

compliance mean rise from 1.02 ± 1.45 after placebo (an

compliance, as calculated from the proporti- improvement of 25%), to 1.91 ± 1.43 (an im-
on of study medication (number of capsules) provement close to 50% of the improvement
returned by the patients, was 92.8 ± 11% for scale) observed after 3 months of hyben vital
hyben vital and 90.6 ± 11% for placebo (non- treatment, p<0.0078. but group b (starting with
on of studydifference).
significant medicationcompliance
(number ofincapsules)
the pla- provement
active close to
treatment) 50% ofno
showed thesignificant
improvement dif-
returned
cebo-first group was 92.3 ±was
by the patients, 92.8
10.0% ± 11%
and for
for ac- scale) observed after 3 months
ference between the two treatments: 1.45 ± of hyben Vital
hyben vital and 90.6 ± 11% for placebo (non-
tive treatment first 90.5 ± 8.0% (nonsignificant treatment,
1.28 units forp<0.0078. but group bas(starting
active treatment, compared with
significant
difference).difference). compliance in the pla- active treatment) showed no significant
with 1.72 ± 1.37 for placebo, p = 0.6084. Table dif-
cebo-first group was 92.3 ± 10.0% and for ac- ference
2, upper between
panel, andthe thetwo treatments:
histograms 1.45
of figs. 2a±
tive treatment
Primary efficacyfirst 90.5 ± 8.0%
measures: pain (nonsignificant 1.28 units for active treatment,
and b illustrate the large between-treatment as compared
difference).
details are given in Table 2. The most important with 1.72 ± 1.37
differences, whenfor groups
placebo, a pand
= 0.6084.
b areTable
com-2,
item of clinical information—the patients’ final pared. group a patients showed a marked and
upper panel, and the histograms of figs. 2A dif-
Primary efficacy
evaluations measures:
of change pain
in pain—showed a re- b illustrate the large between-treatment
ference between the two treatments at every differ-
Details are given in Table
markable difference between 2. The most important
the groups. in ences, when
degree groupswhile
of response, a and b are com-pared.
b showed no consis-
item of clinical information—the patients’
group a (placebo first), there was a highly sig-final group a patients showed a marked
tent pattern of difference between treatments. dif-ference
evaluations of change in pain—showed a re- between
The the two
carryover treatments
effect that weatpostulate
every degree of
as re-
nificant difference in favour of hyben vital—a
markable difference between the groups. in response, while b showed no consis-tent pattern
group a (placebo first), there was a highly sig- of difference between treatments. The carryover
nificant difference in favour of hyben vital—a effect that we postulate as re- sponsible for this
mean rise from 1.02 ± 1.45 after placebo (an between-groups discrepancy (see also discus-
improvement of 25%), to 1.91 ± 1.43 (an im- sion) likewise blunted the level of significance
18
18
32
sponsible for
sponsible for this
this between-groups
between-groups discrepancy discrepancy mol and
mol and seven
seven and and four four andand two two patients,
patients, re- re-
when
(see the
also two treatment
discussion) likewise
(see also discussion) likewise blunted the groups were
blunted lumped
the level
level and aspirin.
spectively, A pattern
handed in very
diaries
spectively, handed in diaries adequate for iTT much like
adequate that forpre-iTT
together:
of there
significance was
when again
the
of significance when the two treatment groupstwono sig-nificant
treatment differ-
groups viously
analyses described
of their for
daily pain,
use
analyses of their daily use of Tramadol, codeineof occurred.
Tramadol, group
codeine a,
ence
werebetween
were lumped together:
lumped the effects
together: of the
there
there was
was two treatments
again
again no sig-
no sig- placebo
and aspirin.
and first,
aspirin. a data
a pattern available
pattern very very much from
much like 12
like thatpatients,
that pre- pre-
(p<0.0991),
nificant difference
nificant data not
difference between shown.
between the an evaluation
the effects
effects of of the of
the showed after 3 months
viously described
viously described for pain,
for pain, a occurred.
mean
occurred. consumption
group a,
group a,
between-
two treatments
two group
treatments (p<0.0991),differences
(p<0.0991), data after
data not only 3-months
not shown.
shown. an an of 227.4
placebo ± 249.5
first, data g. However,
available
placebo first, data available from 12 patients, this
from consumption
12 patients,
treatment
evaluation of did not
of between- attain
between- group statistical
group differences significance,
differences after after was
showed reduced afterto 127.9 ± 143.3 g after 3 months
evaluation showed after 33 months
months a mean
a mean consumption
consumption
although
only 3-months
3-monthsan improvement
treatment did of 50%
did notwas attainobserved
statisti- of active
of 227.4 drug
227.4 ±± 249.5 treatment.
249.5 g. g. however, This
however, this decline of 99.4 ±
this consumption
consumption
only treatment not attain statisti- of
incal
favour of active although
significance, treatmentan (p<0.101)
improvement data not of 163.9 g
was reduced (p<0.0024)
reduced to to 127.9 comprised
127.9 ±± 143.3 a 44%
143.3 gg after reduction.
after 33 monthsmonths
cal significance, although an improvement of was
shown.
50% was We also
observed made in an
favour alternative
of active analysis
treatment of The
of b group,
active drug active
treatment. treatment
This first, with
decline of data
99.4
50% was observed in favour of active treatment of active drug treatment. This decline of 99.4 ±±
the data by
(p<0.101) dataidentifying
not shown.
shown. two categories of sub- available from
163.9 gg (p<0.0024) 15 volunteers,
(p<0.0024) comprised comprised a showed
a 44% after
44% reduction.
reduction. the
(p<0.101) data not 163.9
ject—‘‘responders’’
we also
also mademade an who by
an alternative
alternative definition
analysisshowedof the at
the first
The 3bmonths
group, of active
active treatment
treatment a mean
first, with data va-
data
we analysis of The b group, active treatment first, with
least
data byone category
by identifying
identifying two of improvement
two categories
categories of and ‘‘non-
of subject—
subject— lue of 129.50
available from ±
from 15 91.00 g,
15 volunteers, a value
volunteers, showed close
showed after to what
after the the
data available
responders’’,
‘‘responders’’ who who who byshowed
by definition less improvement
definition showed showed at at was observed
first 33 months
months of in the
of active second
active treatment active
treatment a treatment
a mean
mean va- va-
‘‘responders’’ first
than
least this.
one If we compare
category the A andand
of improvement
improvement B groups
‘‘non- phase
lue of of
of 129.50the group
129.50 91.00ag, patients
value(see Table 2). A
least one category of and ‘‘non- lue ±± 91.00 g, aa value close
close to what
to what
after the first 3 who
responders’’, months of treatment,
showed the overall
less improvement
improvement further
was observed3 months
observed placebo
in the
the second treatment,
active treatmentin the b
treatment
responders’’, who showed less was in second active
outcome of the analysis is that 31/47 (66%) of group, resulted in a non-significant decline to
than this.
than this. ifif we
we compare
compare the the a a and
and bb groups
groups phase of
phase of the
the group
group a patients patients (see (see Table
Table 2). 2). a a
subjects responded to hyben vital, while 18/50 77.70 ± 51.1 g (Tablea2). no significant change
after the
after the first
first 33 months
months of of treatment,
treatment, the the overall
overall further 33 months
further months placebo placebo treatment, treatment, in in thethe bb
(36%) responded to placebo and this was signi- was present when the two groups were lumped
outcome
outcome of the analysis
of the analysis is that 31/47
is that 31/47 (66%) (66%) of
of group, resulted
group, resulted in in a a non-significant
non-significant decline decline to to
ficant at p<0.0128). The corresponding per-pro- together (p<0.1420), data not shown. An eva-
subjects
subjects responded
responded to hyben
to hyben vital,
vital,of while
while 18/50
18/50 77.70
77.70 ± of ± 51.1
51.1 g (Table 2). no significant change
tocol evaluation yielded a p value 0.0428. luation theg two(Table 2). no
groups significant
after 3-months changetreat-
(36%) responded
(36%) responded to to placebo
placebo and and thisthis was
was signi-
signi- was present
was present when when the the twotwo groups
groups were were lumpedlumped
ment showed placebo values of 227.4 ± 249.5 g
ficant at p<0.0128). The corresponding
ficant at p<0.0128). The corresponding per-pro- per-pro- together (p<0.1420),
together (p<0.1420), data not not shown.
shown. an an eva-eva-
and active values of data 128.4 ± 94.3 g. The reduc-
tocol evaluation
tocol evaluation yielded yielded a ap p value
value of of 0.0428.
0.0428. luation
luation of the
of the of two
two groups
groups after 3-months
after 3-months treat-
treat-
Primary efficacy measures: pain tion, in favour active treatment, was 44%, but
ment showed
ment showed placebo placebo values values of of 227.4
227.4 ±± 249.5 249.5 gg
was not statistically significant. when, however, a
table 2.
table 2. Pain
Pain given
given on
on aa scale
scale from
from 00 (no
(no reduc-
reduc- and active
and active valuesvalues of of 128.4
128.4 ±± 94.3 94.3 g.g. The
The reduc-reduc-
subanalysis was made on the delta change in
tion) to
to 44 (almost
(almost total
total relief
relief of
of pain),
pain), consumpti-
consumpti- tion, in favour
tion, in favourof of active
ofrescue treatment,
active medication
treatment, was was 44%, but
tion) consumption from44%, the but
be-
on of
on of rescue
rescue medication
medication givengiven as as paracetamol
paracetamol was not
was
ginning notof statistically
statistically
each of the significant.
significant.
two 3-months when,treatment
when, however,
however,
equivalents (g)
equivalents (g) aa subanalysis
subanalysis was
was made
made on
on
periods (the two initial weeks of treatment) to the
the delta
delta change
change in
in
consumption
consumption
the end of each of rescue
of rescue medication
of themedication
respective from from
periods the be-
the (the
be-
active
active ginning
ginning of
of each
each ofof the
the two
two 3-months
3-months treatment
treatment
Placebo
Placebo treatment p-value
p-value
final two weeks of the 3-month treatment pe-
treatment periods
periods (thewas
(the twoainitial
two initial weeksreduction
weeks of treatment)
of treatment) to
to
riod), there significant in con-
Group A:
Group A: Placebo
Placebo first,
first,then
then active
active treatment
treatment the
sumptionend
the end ofof of each
eachrescue of
of thethe respective
respective periods
medication periods
from active (the
(the
final two
final two weeks
treatment, weeks
whenof of the 3-month
the
comparing 3-month treatment
treatment
placebo and pe- pe-
ac-
pain 1.02±±1.45
1.45 1.91±±1.43
1.43 0.0078
pain
rescuemedication
1.02
medication 227.40
1.91 0.0078 riod),
riod),
tive there was
there
treatment was aa significant
significant
(p<0.006), reduction
datareduction
not shown. in con-
in con-
rescue 227.40±±249.50 127.90±±143.30
249.50 127.90 143.30 0.0024
0.0024
stiffness
stiffness 0.91±±1.38
0.91 1.38 1.91±±1.25
1.91 1.25 0.0025
0.0025
sumption of
sumption of rescue
rescue medication
medication from from active active
treatment, when
treatment,
Secondary when comparing
efficacy comparing
measuresplacebo placebo and and ac- ac-
Group B:
B: Active
Active treatment,
treatment,then
then placebo
placebo
tive treatment
tive
Joint treatment
stiffness, tested (p<0.006),
(p<0.006), on adatadata
scale not
notfromshown.
shown. 0 (no im-
Group
active
active
Placebo
Placebo treatment
treatment p-value
p-value
pain
pain 1.45±±1.28
1.45 1.28 1.72±±1.37
1.72 1.37 0.6084
0.6084
rescuemedication
rescue medication 127.50
127.50±±94.00
94.00 77.70±±51.1
77.70 51.1 0.1452
0.1452
stiffness
stiffness 1.28±±1.35
1.28 1.35 1.71±±1.47
1.71 1.47 0.3850
0.3850
Placebo active
active
Placebo
stiffness estimated
stiffness estimated on on aa scale
scale from
from 00 (no
(no reduction)
reduction) to
to 44 treatment
treatment
100 ––
100
(almost total
(almost total relief
relief of
of stiffness)
stiffness) isis given
given for
for groups
groups a
a and
and B. B.
data given
data given are
are mean
mean ±sd.±sd. non-responders respondersgraded
graded
non-responders responders
75 ––
75
Twenty-three patients handed in medical diari- n=50
n=50
esPrimary
adequate
Primary efficacy
for iTT
efficacy measures:
analysis of
measures: pain
their use of nsaids
pain a
a
patients
üf patients

p=0.0078
p=0.0078
inTwenty-three
accordance patients
with
Twenty-three patients handed in thehanded
protocol.in medical
medical diari-
consumption
diari- 50 ––
50 placebovs
placebo vs
es adequate for iTT analysis of their use of nsaids activetreatment
treatment
during
es adequatethe two fortreatment
iTT analysisperiods wasof
of their use found
nsaidsto active
bein accordance
accordance
identical (data with the
the protocol.
not protocol. consumption
shown). consumption
paracetamol
Percentage üf

in with
25 ––
Percentage

during
and the two
two treatment
acetylsalicylictreatment
acid wereperiods was found
found to
administered to
as 25
during the periods was
be
500 identical
mg tablets (data
and not shown).
Tramadol
be identical (data not shown). paracetamol and paracetamol
codeine as
50and
and and acetylsalicylic
25 mg tablets,
acetylsalicylic acid
acid were administered
administered
respectively.
were Twentyfive-as
as
patients
500 mg handed
mg tablets
tablets andandin medical
Tramadoldiaries adequate
and codeine
codeine as 00
500 Tramadol and as 00 11 22 33 44
for
50 iTT
and analysis
25 mg of their
tablets, daily use
respectively.
50 and 25 mg tablets, respectively. Twentyfive of paraceta-
Twentyfive reliefof
relief ofpain
pain
mol
patientsand handed
patients seven
handed and four anddiaries
in medical
in medical two patients,
diaries adequate
adequate re-
spectively,
for iTT
for handed
iTT analysis
analysis of in diaries
of their
their daily use
daily adequate
use for iTT
of paraceta-
of paraceta-
analyses of their daily use of Tramadol, codeine

100 ––
100

75 ––
75
 0
0 1 2 3 4

33
19
group a (n = 26) showed, in all measurements
active asame
distinct
sharp difference
distinctioninbetween favour of Hyben
groups as forVital.
the
Placebo
treatment The change
primary was highly
parameters. Thesignificant, stiffness
placebo-first group anda
100 –
pain in the
(n = 26) morning
showed, in allgiving p values of
measurements 0.0016
a distinct
non-responders responders graded and 0.0127, respectively, and
difference in favour of hyben vital. The change sleep quality,
75 – mood and significant,
was highly general wellbeing, stiffness and 0.0096,
pain0.0124
in the
n=47
p=0.6084 and 0.0164,
morning re-spectively.
giving p values of but0.0016
in theandhyben-first
0.0127,
B group B, the and
respectively, two setssleepofquality,
results mood
appeared and ge- in-
Percentage üf patients

active treatment
50 – vs placebo distinguishable, and there was
neral wellbeing, 0.0096, 0.0124 and 0.0164, re- not a single in-
stance of anything
spectively. but in the approaching
hyben-first agroupstatistically
b, the
significant difference between
two sets of results appeared indistinguishable, the two treat-
25 –
ment groups,
and there wasas notshown
a single by instance
a meanof p anything
level of
more than 0.50 (details not shown).
approaching a statistically significant difference The majority
of the significant
between the twochangestreatment observed
groups, in as favour
shown
0
0 1 2 3 4 of active treatment in the placebo-first
by a mean p level of more than 0.50 (details not group,
relief of pain were
shown).confirmed,
The majority whenofsub-analysis
the significant comparing
changes
the A and B group after
observed in favour of active treatment 3- months treatment
in the
fig. 2. (a) histograms comparing, in group a was made: stiffness in the morning, p<0.054;
placebo-first group, were confirmed, when sub-
subjects (placebo first then hyben Vital), the pain in the morning,the p<0.036;
analysis comparing a and general wellbe-
b group after 3-
degree of improvement in pain relief from pla- ing,
months treatment was made: stiffnessquality,
p<0.012; mood, p<0.017; and sleep in the
cebo (white columns) as compared with hyben p<0.005.
morning, p<0.054; pain in the morning, p<0.036;
Vital (shaded columns). the height of each co-
general wellbeing, p<0.012; mood, p<0.017;
lumn indicates the percentage of patients who Patients’ preference for treatment
and sleep quality, p<0.005.
experienced pain relief of category 0, 1, 2, 3 or The separate groups again showed a large
4, corresponding to 0%, 25%, 50%, 75% or 100% difference, similar in pattern to that described
Patients’ preference for treatment
relief of pain (p value refers to the added scores above. In group a, 24 againpatients reported that
The separate groups showed a large
comparing the two different treatments). (B) his- they felt most improvement from Hyben Vital,
difference, similar in pattern to that described
tograms comparing in group B subjects (hyben while 8 patients preferred placebo and 9 were
above. in group a, 24 patients reported that
Vital first then placebo), the degree of impro- not sure (p<0.0070). In group B, 12 patients pre-
they felt most improvement from hyben vital,
vement in pain relief from hyben Vital (shaded ferred the first treatment (Hyben Vital) whereas
while 8 patients preferred placebo and 9 were
columns) as compared with placebo (white 20 voted for placebo treatment and 8 did not
not sure (p<0.0070). in group b, 12 patients pre-
columns). the height of each column indicates have any preference (p<0.2153). Comparison
ferred the first treatment (hyben vital) whereas
the percentage of patients who experienced of the A and B groups (fisher’s test) gave a p
20 voted for placebo treatment and 8 did not
a given pain relief of category 0, 1, 2, 3 or 4, value of <0.0040 in favour of hyben vital.
have any preference (p<0.2153). comparison
corresponding to 0%, 25%, 50%, 75% or 100%
of the a and b groups (fisher’s test) gave a p
relief of pain (p value refers to the added scores Routine screening tests
value of <0.0040 in favour of hyben vital.
comparing the two different treatments). Haemoglobin, blood glucose, creatinine and
sodium and potassium levels were unaffected
provement at all) to 4 (almost total relief of stiff- routine
by eitherscreening
treatment.tests nor were there any chang-
secondary
ness) revealedefficacy measures
an almost identical pattern to haemoglobin,
es when those patients blood glucose, creatinine
with blood glucose andlev- so-
that found for pain. inongroup
Joint stiffness, tested a scale frominitial
a, the 0 (nopla-
im- dium and potassium levels were
els above 5.5 mmol/l were analysed separately. unaffected by
provement
cebo value wasat all) to 4
0.91 ± (almost
1.38 (antotal relief of stiff-
improvement of either
an treatment.
unexpected nor were
finding was thatthere any changes
Hyben Vital re-
ness) revealed an almost identical
23% on the scale) as compared to 1.91 ± 1.25 pattern to sulted in a small but significant 8.5% fall of levels
when those patients with blood glucose total
that found for pain. in group
(an improvement of 48%) while on hyben vital a, the initial pla- above 5.5 mmol/l were analysed separately.
cholesterol.
cebo value
therapy), was 0.91group
p<0.0025. ± 1.38B, (an improvement
however, showed of an unexpected finding was that hyben vital re-
23%significant
no on the scale) as compared
difference betweentotreatments:
1.91 ± 1.25 sulted in aeffects
Unwanted small but significant 8.5% fall of total
(an improvement of 48%) while
hyben vital 1.28 ± 1.35 versus placebo 1.71 on hyben vital
± cholesterol.
Although 27 of the original 112 subjects recrui-
therapy), p<0.0025. group b, however,
1.47, p<0.3850 (Table 2). nor was there any sig- showed ted dropped out during the 6-months treatment
no significant
nificant difference
difference when the between
two groupstreatments:
were unwanted
period, onlyeffects
3 of these defaulted because of
hybentogether
taken vital 1.28(p<0.1612),
± 1.35 versus data placebo
not shown. 1.71 a± although 27
adverse effects: of the
acidoriginal 112 subjects
regurgitation recrui-
occurred in
1.47, p<0.3850
comparison of (Table
the two 2).groups
nor wasafter there 3 any
monthssig- ted dropped out during the 6-months
one patient during placebo therapy and in one treatment
nificant
of difference
treatment, althoughwhen the two
in favour groupstreat-
of active were period,active
during only 3treatment,
of these anddefaulted because
one other patientof
taken together
ment, did not (p<0.1612),
attain statistical data not shown. a
significance adverse
with effects:
diarrhoea acid regurgitation
dropped out while on occurred
placebo; in
comparison
(p<0.153), of not
data the shown.
two groups The diaryafterrecords
3 months of one patient during placebo therapy and
for details see Fig. 1. In the remaining group the- in one
of treatment, although in favour
joint pain and stiffness in the morning and later of active treat- during
re wereactive
12 who treatment,
reportedand oneunwanted
milder other patient
ef-
ment,
in the day,did wellbeing,
not attain mood,statistical
energysignificance
and sleep, with diarrhoea
fects. These were dropped out frequency
as follows: while on placebo;
of mic-
(p<0.153), in
available data not shown.
diaries from 47The diary records
patients, showed of for details
turition see fig.
4 (three 1. inon
while the remaining
active groupand
treatment the-
jointsame
the pain and
sharpstiffness in the
distinction morninggroups
between and lateras one while on placebo); waterbrash 3 (presentef-
re were 12 who reported milder unwanted in
in the
for theday, wellbeing,
primary mood, energy
parameters. and sleep,
The placebo-first fects.treatments);
both These were diarrhoea
as follows: 2frequency
(present inofboth
mic-
available in diaries from 47 patients, showed the turition 4 (three while on active treatment and
34
treatments); constipation 2 (1 during placebo
and 1 during both treatments); urticaria 1 (while
on placebo). There were no major side effects
of any kind in the whole group.
Discussion
Interpretation of trial results
The chief advantage of a crossover trial, as
used here, is that in comparing the effects of
two successive treatments on the same ‘‘arm’’
of the trial, each patient acts as his/her per-
fect control, so concern about mismatching of
the groups—an important source of error—can
be forgotten. A wholly uncomplicated crosso-
ver trial with a positive result can be expected
to yield three pieces of information: a within-
group significant comparison of the two test
substances—one from each of the two arms of
the trial (and more or less identical with each
other), and a significant between-groups com-
parison at the crossover point, provided that the
groups have been well matched, since in this
case the patients do not act as their own con-
trols.
Looking at the results of the trial described here,
it is obvious that they are far from this idealised
pattern. That arm of the trial given placebo first
does show a significant, clear-cut difference
between the effects of the two test substances.
so far so good, but the other arm— active sub-
stance first, placebo second—shows no sig-nifi-
cant difference between the two. We belie-ve
that by far the most likely explanation of this dis-
crepancy between the two arms of the trial is a
strong ‘‘carryover’’ effect of Hyben Vital. This is a
common, major complication of cross-over trials
and the reason for the inclusion of a ‘‘washout’’
period after crossover.
The usual tactical response is to write off all data
after the crossover point and to supplement the
single withingroup result obtained in the place-
bo-first arm, with a between-groups comparison
at the crossover point. but this, using the primary
efficacy data of Table 2, also gave a non-signi-
ficant result. This raises the possibility that a car-
ryover effect is not the whole explanation—a
slow onset of the active drug effect could be
another factor. The strength and significance
of the difference between placebo and active
drug seen in group a is supported by several
ancillary aspects. if the reduction in pain sensa-
tion was evaluated after 3-month treatment on
a yes/no basis, there was a significant reduction
of pain from active treatment when compared
to placebo. In agreement with this finding, pre-
ference for treatment a or b was also in favour
of active treatment and the diary recordings
on pain, general wellbeing, mood and sleeping
quality were all statistically significant in favour of
active treatment. Taken together these findings
seem to fully justify confirmation of the action of
hyben vital by a large-scale, parallel, placebo-
controlled, blind study, and this is our intention.
R. canina (the ‘‘dog rose’’, the common wild-
briar rose of english hedgerows) is said to have
been so named because the ancient greeks
believed its root to be effective against the bite
of a mad dog (brewer, 1981). in this context, pliny
the elder used the plant’s classical greek name
‘‘cynorrhodos’’, combining the verbal roots of
‘‘dog’’ and ‘‘red’’ (pliny, 1966). although hyben
vital has been marketed in scandinavia for
several years, modern european interest in the
plant has been concentrated on preparations
made from the hips rather than the root, mainly
because of their high content of vitamin c, and
herbal tea infusions of ‘‘cynorrhodon’’ are still
used today.
It is widely known that rose hips contain signifi-
cant amounts of vitamin c, but it seems highly
unlikely that this accounts for much, or indeed
any, of the activity of hyben vital in this trial. A
large-scale study in 1996 on the framingham
population group showed that the middle and
highest tertiles of daily dietary vitamin c intake
did protect against the long-term progression
of knee osteoarthritis (especially against loss of
cartilage). but the lowest intake tertile—a daily
mean of 81 mg for men and 94 for women—
had no such protective effect (mcalindon et al.,
1996). The vitamin c content of a hyben vital
dosage of 5 g daily, as used in this trial, is only
26 mg, i.e. only one-third of the framingham lo-
west tertile and therefore very unlikely to contri-
bute significantly to the action of hyben vital.
We have earlier shown that hyben vital signi-
ficantly reduces the migration of neutrophils,
when estimated after 1 month of treatment (win-
ther et al., 1999; kharazmi and winther, 1999).
one explanation for the lessening of symptoms
during Hyben Vital treatment could therefore
be a reduction of the inflammation that is an in-
tegral part of the pathogenesis of osteoarthritis
(harris, 1988). This hypothesis has gained increa-
sing interest, as an active ingredient that inhibits
neutrophil chemotaxis, has now been isolated
from the present subtype of rose hip (larsen et
al., 2003). if the present suggestion is correct,
it could also explain the pronounced carry-
over effect; once inflammation has subsided,
it requires a certain interval of time before the

process can be reactivated. as rose hips have
been used in daily household use for centu-
ries, it is surprising that their anti-inflammatory
property has not been detected before now.
a possible explanation is that different species
of rose hip vary in their anti-inflammatory prop-
erties (Brandt and (Akesson, 2002). In another
study testing a possible interaction between
Rose hip and warfarin, we could not show any
effect on coagulation and platelet aggregabil-
ity (Winther, 2000). This suggests that rose hip, un-
like nsaid, aspirin and ginger—another natural
remedy also used for symptoms of osteoarthritis
(Altman and Markussen, 2001)—does not affect
the arachidonic acid and cyclo-oxygena-se
system. This could explain why the incidence of
side effects is lower for Hyben Vital than for the
therapies mentioned above.
Conclusion
We have found that the herbal remedy has a
moderate alleviating effect on joint pain and
improves general wellbeing, sleep quality and
mood in patients with osteoarthritis, without pro-
ducing any side effects. We consider that the
results warrant a largescale dou-ble-blind, long-
term, placebo-controlled and parallel study.
References
Altman, r.d., markussen, k.c., 2001. effect
of a ginger extract on knee pain in
patients with osteoarthritis. arthritis rheum.
44 (11), 2531–2538.
Altman, R.D., Brandt, K., Hochenberg, M.,
Moskowitz, r.d., 1996. Conduct of clinical
trials in patients with osteoarthritis.
Osteoarthr. Cartilage 4, 217–243.
Brandt, k., a( kesson, b., 2002. health promo-
ting compounds in vegetables and fruit.
Plant prod. 29, 43–44.
Brewer’s dictionary of phrase and fable
(revised edition, 1981) 346. in: evans, i.h.
(ed.), cassell, London. p. 1213.
Harris Jr., e.d., 1988. Pathogenesis of
rheumatoid arthritis: a disorder associated
with dysfunctional immunoregulation.
In: Gallin, J.H., Goldstein, i.m. (eds.),
Snyderman, R. (Eds), Inflammation. Basic
Principles and Clinical Correlates.
Raven Press, New York, pp. 751–773.
Hochberg, m.c., altman, r.d., brandt, k.d.,
clarck, b.m., dieppe, p.a., griffin, m.r.,
moskowitz, w., schnitzer, T.J., 1995.
guidelines for the medical management of
osteoarthritis (part 1 and 2). Arthritis Rheum.
38 (11), 1535–1546.
Kharazmi, a., winther, k., 1999. rose-hip inhibits
35
chemotaxis and chemiluminescence of
human blood neutrophils in vitro and
reduces certain inflammatory parameters
in vivo. inflammopharmacology 7 (4),
377–386.
Larsen, e., Kharazmi, A., Christensen, l.p.,
Christensen, s.b., 2003. an antiinflammatory
galactolipid from rose-hip (rosa canina)
that inhibits chemotaxis of human
peripheral blood neutrophils in vitro. J. nat.
prod. 66, 994–995.
Lægemiddelkataloget, 2002. Dansk Lægemid-
del information a/s: issn 0105-287x.
Mcalindon, T.e., Jacques, P., Zhang, Y.,
Hannan, m.T., aliabadi, p., weissman, b.,
Rush, D., Levy, D., Felson, d.T., 1996. do
antioxidant micronutrients protect against
the development and progression of knee
osteoarthritis? arthritis rheum. 39 (4),
648–656.
Mukherjee, d., nissen, s.e., Topol, e.J., 2001. risk
of cardiovascular events associated with
selective cox-2 inhibitors. Jama 286 (8),
954–959.
The older pliny, 1966. natural history vll: books
xxlvxxvll, 149. william heinemann, london
(isbn: 0-434-99419-7).
Rodrigues, l.a.c., hernandes-diaz, s., 2001.
relative risk of upper gastrointestinal
complications among users of
acetaminophen and non-steroidal anti-
inflammatory drugs. epidemiology 12,
570–576.
Vane, J.r., botting, r.m., 1998. anti-
inflammatory drugs and their mechanism of
action. inflamm. res. 47 (2) 578–587.
Winther, k., 2000. rose-hip, in the form of
HybenVital, has no impact on coagulation,
platelet function and fibrinolysis. In: Procee-
dings of the Third international exhibition
and conference on nutraceuticals and
food for vitality, palexpo exhibition and
Conference Centre, geneva, Switzerland,
May 2–3.
Winther, k., rein, e., kharazmi, a., 1999. The
antiinflammatory properties of Rose-hip.
immunopharmacology 7, 63–68.
The EffEcts of a
36

standardIzed herBal remedymade from

a suBtyPe of rosa canIna In PatIents
wiTh osteoarthrItIs:
a douBle-BlInd, randomIzed,
PlaceBo-controlled clInIcal trial
Odd Warholm, md, sigrun skaar, md, ewa hedman, rn,
Hanna Maria Mølmen, rn, and Liv Eik, rn
Department of Orthopaedic surgery, vestfold central hospital, Tønsberg, Norway

Background: a standardized rose-hip powder produced from the seeds and husks of fruit from a
subtype of rosa canina has been reported to inhibit leukocyte functions that cause cell injury in
osteoarthritis.

Objective: The aim of this study was to assess the impact of standardized rose-hip powder on mobil-
ity of the hip and knee joints, activities of daily living, quality of life, and pain in patients with osteoar-
thritis.

Methods: patients with a diagnosis of osteoarthritis of either the hip or knee, verified on radiography,
participated in this randomized, placebo-controlled, double-blind study. Half of the patients were
given five 0.5-g capsules of standardized rose-hip powder twice daily for 4 months, and the other
half received identical placebo capsules twice daily for the same period. Mobility of the hipor knee
was measured in both groups after the initial screening and again after 4 months of therapy.

Results: One hundred patients (65 women, 35 men; mean [sd] age, 65.2 [11.1] years) were divided
into 2 treatment groups of 50 patients each. hip joint mobility improved significantly in the treat-
ment group compared with the placebo group (p=0.033). similarly, pain decreased significantly
in the treatment group compared with the placebo group (p=0.035). Two patients (4%) from each
group withdrew during the early stages of the trial for reasons not related to treatment.

Conclusions: in this study population, standardized rose-hip powder reduced symptoms of osteoar-
thritis, as 64.6% of patients reported at least some reduction of pain while receiving treatment. Stan-
dardized rose-hip powder may improve hip flexion and reduce pain in patients with osteoarthritis.
(curr Ther res clin exp. 2003;64:21–31) copyright © 2003 excerpta medica, inc.

key words: osteoarthritis, stiffness, pain, rose-hip powder, rosa canina.

Introduction oarthritis. As a possible alternative, a standar-

During the past decade, the commonly used
drugs for osteoarthritic pain were aspirin, other
nonsteroidal anti-inflammatory drugs (nsaids),
and corticosteroids. 1 however, side effects
have been associated with prolonged use of
these drugs. During the past 5 years, selective
inhibitors of cyclooxygenase-2 (an enzyme in-
volved in the synthesis of proinflammatory cy-
tokines) have shown promising analgesic and
anti-inflammatory actions without serious ad-
verse effects. 2 however, these drugs are ex-
pensive, and the need remains for a lowcost,
safe remedy for long-term treatment of oste-
dized rose-hip powder* made from the seeds
and husks of fruit from a subtype of Rosa canina
is available. This powder inhibits leukocyte func-
tions that cause cell injury in osteoarthritis. The
plants are grown according to good agricultu-
ral practice in standardized fields in Denmark
and Sweden. When the fruits are mature, they
are harvested and frozen immediately. selec-
tion of optimal fruits for later production of pow-
der is made by a laser technique, and the com-
puterized drying process does not exceed 40 c.
The vitamin and mineral content of the powder
is controlled. uncontrolled exploratory trials3,4 of

this standardized dry rose-hip powder showed
analgesic action in patients with osteoarthritis.
This finding was evidenced by a mean (sd) de-
crease in the serum concentration of c-reactive
protein from 8.25 (4.9) mg/l before treatment to
6.67 (2.6) mg/l after treatment, and inhibition of
polymorphonuclear chemotaxis. These findings
were sufficient to encourage the present trial.
The aim of this study was to assess the impact
of the standardized rose-hip powder on mobi-
lity of the hip and knee joints, activities of daily
living (adls), quality of life, and pain in patients
with osteoarthritis.
Patient and methods
This was a single-center, double-blind, rando-miz-
ed, placebo-controlled study. All patients had
a diagnosis of osteoarthritis of the hip or knee,
verified on radiography, within 12 months before
the study. Patients with pain for 6 months and
who were on a waiting list for either hip or knee
surgery, or on a list for final evaluation for surgery,
were included. patients who reported allergy
to plant products or who had severe asthma
or liver disease were excluded. all patients pro-
vided written informed consent to participate,
and approval from the ethics committee of the
study site (an outpatient clinic in norway) was
obtained.
Patients were randomized in groups of 10 us-
ing an independent computerized system. one
group was randomized to treatment with five
0.5-g capsules of standardized rose-hip pow-
der twice daily for 4 months. The other group
received the same quantity of placebo cap-
sules (identical in appearance, taste, and smell
to the rose-hip powder capsules) for the same
time period as the active treatment group.
Primary outcome measures
Mobility of the hip or knee was measured in both
groups after the initial screening and again af-
ter 4 months of therapy. mobility measurements
included the full range of external and internal
rotation of the hip; maximum flexion and exten-
sion of the hip and knee measured using a go-
niometer (gallus plesner, oslo, norway) during
passive movement; and active voluntary rotati-
on, flexion, and extension by the patient. goni-
ometry can result in some variation if the test is
not conducted by the same researcher at each
visit. for that reason, all measurements were ta-
ken by the same investigator and data given
are expressed as the mean of 3 test episodes.
The measurements of joint movement are pre-
sented in 2 ways: as the numeric measurements
taken and also as a degree of restriction, calcu-
37
lated by subtracting these measurements from
a standard value of 125° for hip flexion, 140° for
knee flexion, and 45° for external and internal
hip rotation.5
Secondary outcome measures
At the start of the trial and again after 1, 2, and 4
months of treatment, patients recorded any dif-
ficulties in performing adls, such as walking, get-
ting into and out of a car, shopping, and getting
up and down from the lavatory. The dif-ficulty
was estimated on a visual analog scale ranging
from 0 (no difficulty) to 10 (great diffi-culty). after
4 months of therapy, patients gave their over-
all assessment of the effectiveness of the study
medication on relief of joint pain using a cat-
egoric scale of 0 (no improvement) to 4 (almost
total relief of pain). The patients also were asked
about relief of pain on a simple yes-or-no ques-
tionnaire after 1, 2, and 4 months of treatment.
During the trial, patients were asked to maintain
their daily dosage of nsaids. Any changes that
did occur were to be recorded in a diary. Com-
pliance was estimated by counting the num-
ber of capsules returned by patients. adverse
events were recorded on the case-report forms
completed at each visit.
Statistical analysis
Statistical analysis was performed on an intent-
to-treat basis. Results in the 2 groups were com-
pared using the mann-whitney test for parallel
data. The wilcoxon signed rank test for matched
pairs was used to compare baseline findings
with those after 1, 2, and 4 months of treatment
in each group separately. The chi-square test
was used for the questionnaires. all data are
presented as mean (sd). statistical significance
was set at p<0.05.
Results
One hundred patients (65 women, 35 men;
mean [sd] age, 65.2 [11.1] years) were enrol-
led. The treatment group comprised 34 women
and 16 men (mean [sd] age, 65.1 [12.2] years).
The placebo group comprised 31 women and
19 men (mean [sd] age, 65.3 [9.9] years). The
demographic and osteoarthritic characteris-
tics of the 100 patients entering the study (in-
tent-to-treat population) and of the 96 patients
who completed the study (per-protocol popu-
lation) are shown in Table i. The demographic
characteristics and consumption of medicine
were similar in the intent-to-treat and per-pro
tocol populations. at baseline, active flexion of
the hip, however, was significantly different in
the active-treatment group versus the placebo
 38
24

group in both the intent-to-treat

the active-treatment group versusand the per-
the placebo woman
withdrewand 1 man
during the in the stages
early placebo group
of the be-
trial: 1
protocol
group in populations. active external
both the intent-to-treat and rotation
the per- of cause
womanofand cardiac
1 man problems and a sore
in the placebo groupthroat,
be-
the hip was
protocol significantlyactive
populations. different in the rotation
external ac-tive- respectively,
cause of cardiac and 1problems
woman and and a1 sore
manthroat,
in the
treatment
of the hip group compareddifferent
was significantly with theinpla-cebo
the ac- treatment group due to the possibility
respectively, and 1 woman and 1 man in the of hip sur-
group
tive-treatment group comparedpopula-tion.
only in the intent-to-treat with the pla- all gery earlier than expected and because
treatment group due to the possibility of hip sur- of the
passive
cebo group movements
only inwere comparable between
the intent-to-treat popula- desire not tothan
gery earlier continue,
expectedrespectively.
and becauseTheseof 4 the
pa-
groups.
tion. all passive movements were comparable tients
desirecomprised 3 hip respectively.
not to continue, joints (1 in theThese
treatment
4 pa-
between groups. group and 2 in the
tients comprised placebo
3 hip group)
joints (1 in theand 1 knee
treatment
Among the 100 patients, there were 44 hip joints joint (in the treatment group). The
group and 2 in the placebo group) and 1 kneebaseline de-
(25
amongin the thetreatment group,
100 patients, there 19were
in the44 placebo
hip joints mographic
joint (in the characteristics,
treatment group). medication,
The baselineandde- os-
group) and 56 knee joints (25 in the
(25 in the treatment group, 19 in the placebo treatment teoarthritic characteristics of the 2 groups
mographic characteristics, medication, and os- were
group,
group)31 andin the placebo
56 knee jointsgroup)
(25 in involved in the
the treatment similar, except
teoarthritic for range of of
characteristics motion
the 2 for active
groups hip
were
trial. all patients had experienced
group, 31 in the placebo group) involved in the osteoarth- flexion and active external hip rotation (p=0.041
similar, except for range of motion for active hip
ritic
trial.pain for 2 to had
all patients 12 years. four patients
experienced (4%)
osteoarth- for treatment
flexion group
and active vs placebo
external group).(p=0.041
hip rotation
withdrew
ritic pain for 2 to 12 years. four patients (4%)1
during the early stages of the trial: for treatment group vs placebo group).

table I. Baseline demographic and osteoarthritic characteristics of the study population.

Intent-to-treat Population Per-Protocol Population

Placebo srhP Placebo srhP
(n=50) (n=50) (n=48) (n=48)
age, y* 65.3 (9.9) 65.1 (12.2) 65.8 (14.7) 65.5 (14.2)
sex, no. (%)
woman 31 (62.0) 34 (68.0) 29 (60.4) 33 (68.8)
man 19 (38.0) 16 (32.0) 19 (39.6) 15 (31.3)
no. (%) of patients
with oa of the hip 19 (38.0) 25 (50.0) 17 (35.4) 24 (50.0)

no. (%) of patients

with oa of the knee 31 (62.0) 25 (50.0) 31 (64.6) 24 (50.0)

hip joint movement, deg*

passive flexion 111.05 (12.76) 116.00 (13.92) 111.47 (13.20) 115.62 (14.09)
active flexion 97.63 (15.49) 105.60 (13.10)† 97.94 (16.01) 105.42 (13.34)‡
passive external rotation 19.72 (11.56) 26.40 (9.74) 20.00 (16.01) 25.62 (9.13)
active external rotation 13.06 (10.17) 20.00 (9.79)§ 13.44 (10.76) 19.17 (9.05)
passive internal rotation 28.61 (11.61) 28.80 (13.17) 29.37 (12.09) 28.75 (13.45)
active internal rotation 21.39 (10.68) 21.20 (12.61) 22.19 (11.10) 21.04 (12.85)
Knee joint movement, deg*
passive flexion 128.71 (14.37) 132.40 (9.14) 128.71 (14.37) 132.08 (9.20)
active flexion 123.55 (14.73) 124.80 (11.77) 120.35 (24.95) 124.58 (11.97)

no. (%) of patients

taking concomitant
medication
none 15 (30.0) 11 (22.0) 14 (29.2) 10 (20.8)
nsaids 20 (40.0) 24 (48.0) 19 (39.6) 23 (47.9)
paracetamol 12 (24.0) 14 (28.0) 12 (25.0) 14 (29.2)
opioids 2 (4.0) 0 (0.0) 2 (4.2) 0 (0.0)
asthma medication 2 (4.0) 0 (0.0) 2 (4.2) 0 (0.0)
antihypertensive 2 (4.0) 2 (4.0) 2 (4.2) 2 (4.2)
heart disease medication 5 (10.0) 3 (6.0) 5 (10.4) 3 (6.3)

srhP=standardized rose-hip powder; oa=osteoarthritis; nsaIds=nonsteroidal anti-inflammatory drugs.

*Values are expressed as mean (sd). †P=0.020 versus placebo. ‡
P=0.039 versus placebo. §P=0.041 versus placebo.

effects of 4 months’ treatment on Joint

Movement
(p=0.031), but not in flexion or external rotation.
Effects of 4 months’ treatment on Joint
movement
Patients receiving standardized rose-hip powder
The between-group comparison at 4 months
The between-group comparison at 4 months
patients receiving standardized
showed significant improvements rose-hip powder
at 4months in showed a
showed a significant
significant difference
difference in
in improvement
improvement
showed hip
passive significant
flexion improvements at 4months
(p=0.003), external in
rotation in passive
in passive hip
hip flexion
flexion (p=0.033),
(p=0.033), but
but not
not in
in inter-
inter-
passive hipand
(p=0.006), flexion (p=0.003),
internal external
rotation rotation
(p<0.001) (Ta- nal or external rotation.
nal or external rotation.
(p=0.006), and internal rotation (p<0.001)
ble ii). The placebo group showed a significant (Ta-
ble ii). The placebo
improvement group hip
in passive showed a significant
internal rotation The same
The same patterns
patterns of
of change
change in
in joint
jointmovement
movement
improvement in passive hip internal rotation
(p=0.031), but not in flexion or external rotation. (and in
(and in pp values)
values) were
werefound
foundwhen
whenhip hipflexion
flexion
 25
39

and rotation
and rotation were actively performed by the thethefollowing
followingimprovements
improvementswere were observed
observed in in
patients (Table iii). however,
patients (Table however, itit should be noted the placebo
the placebo group: group: walking down the
down the street (p<street
that the
that the baseline values for active hip flexion (p< 0.05),
0.05), getting
getting intointo and
and out out
ofof a car
a car (p<0.001),
(p<0.001), put-
and active
and active external
external hip rotation were not iden- putting on/taking
ting on/taking off stockings
off stockings (p<0.001), (p<0.001),
and get-
tical in
tical in the 2 groups (Table i), which makes the andtinggetting
up andup downand fromdown thefrom the lavatory
lavatory (p=0.274).
interpretation of
interpretation of these
these results
results difficult.
difficult. (p=0.274). These improvements
These improvements were not were foundnot at 4found
months
atof treatment
4 months in the placebo
of treatment in the placebogroup.group. in con-
Changes
changes in in passive
passive flexion of the knee did not in trast, the group
contrast, the grouptreated with the
treated with stan-dardized
the stan-
differ
differ significantly between
significantly between the 2 groups (data rose-hip powder showed
dardized rose-hip powder showed significant significant changes
not
not shown).
shown). Active
active treatment
treatment resulted in ame- in the majority
changes of adl functions
in the majority after 1 month
of adl functions after of
an
an (sd)
(sd) improvement
improvement of 2.71 (4.42°)(4.42°) (p=0.012); treatment,
1 month as follows:aswalking
of treatment, follows: downwalking the street
down
this
this value
value improvementwas
improvementwas 3.75° 3.75° (5.32°) in the the(p<0.001), getting into
street (p<0.001), and out
getting intoofand a car out(p<0.05),
of a
placebo
placebo group
group (p=0.005). asimilar pattern oc- carshopping
(p<0.05),(p<0.001),
shoppingputting(p<0.001), on/taking
putting off stock-
on/ta-
curredwhen
curredwhen flexionwas performed actively by ingsoff(<0.001),
king stockings and getting
(<0.001), andupgetting
and down up and from
the
the patients
patients at
at the
the request
request of
of the
the researcher.
researcher. the lavatory
down (p<0.05).(p<0.05).
from the lavatory after 2 months
after 2 months of treat-
ofment, improvement
treatment, improvement was found in all ofin these
was found all
Activities
activities of of daily
daily living
living adls (p<0.001 for all), and
of these adls (p<0.001 for all), and this this group continued
group
Changes
changes in in difficulty performing adls
difficulty performing adls did did not
not to show significant
continued improvement
to show significant in the major-
improvement in
differ significantly between the 2 groups.
differ significantly between the 2 groups. sig- sig- ity of adl performances
the majority of adl performances at month 4at month 4 compared
nificant
nificant improvement
improvement was was observed
observed in the fol-
in the fol- with baseline,
compared withasbaseline,
follows: walking
as follows:downwalkingthe street
lowing adls in the placebo group
lowing adls in the placebo group after 1monthafter 1month (=0.038), getting into and
down the street (=0.038), getting into and out of a car (p=0.054,
out
of
of treatment: walking down
treatment:walking down thethe street
street (p<0.05),
(p<0.05), ofborderline
a car (p=0.054, significant), shop-ping
borderline (p=0.024),
significant), shop- put-
getting into and out of a car (p=0.258),
getting into and out of a car (p=0.258), shop- shop- ting on/taking off stockings
ping (p=0.024), putting on/taking off stockings (p=0.019), and get-
ping (p<0.001), putting on/taking off
ping (p<0.001), putting on/taking off stockings stockings ting up and
(p=0.019), anddowngetting from uptheand lavatory
down (p=0.016).
from the
(p=0.251),
(p=0.251), and and getting
getting up
up and down from
and down from the
the lavatory (p=0.016).
lavatory
lavatory (p= 0.154). after
(p= 0.154). after 22 months
months ofof treatment,
treatment,

table II. Passive hip joint movements before therapy and standardized rose-hip powder (srhP)
and placebo.

type of Baseline, deg* restriction at 4 months Improvement, %

movement of movement, of therapy, deg*
deg

flexion
srhp 116.00 (13.92) 9.0 119.37 (14.09)†‡ 40.0
placebo 111.05 (12.76) 13.9 112.38 (14.27) 6.7
external rotation
srhp 26.40 (9.74) 18.6 28.96 (8.84)§ 17.1
placebo 19.72 (11.56) 25.3 22.50 (11.40) 10.0
Internal rotation
srhp 28.80 (13.17) 16.2 34.38 (13.41)|| 35.0
placebo 28.68 (11.61) 16.4 33.13 (12.09)¶ 24.0
*Values are expressed as mean (sd). †
P 0.003 versus pretreatment. ‡
P 0.033 versus placebo.
§
P 0.006 versus pretreatment. ||
P 0.001 versus pretreatment. ¶
P 0.031 versus pretreatment.

Joint Pain cated that they

compared withhadthepain relief atgroup
placebo both 1indicat-
month
significantly
Joint Pain greater relief of joint pain was (p=0.014) and 4 months (p=0.046) of
ed that they had pain relief at both 1 month treatment,
found in the group
Significantly greater receiving
relief of standardized
joint pain rose-
was but not atand
(p=0.014) 2 months of treatment.
4 months (p=0.046) of treatment,
hip powder than in the placebo
found in the group receiving standardized group after
rose- 4 but not at 2 months of treatment.
months of treatment (p= 0.035;
hip powder than in the placebo group after 4 figure). at month compliance, concomitant medication,
4, 31 ofof48treatment
months (64.6%) patients
(p= 0.035; in figure).
the activetreat-
at month and tolerability
Compliance, concomitant medication,
ment group reported some effect,
4, 31 of 48 (64.6%) patients in the activetreat- ranging up compliance
and tolerability was 98% in the treatment group
to almost total relief of pain,
ment group reported some effect, ranging up whereas 17 of 48 and 97% in
Compliance was the placebo
98% in group. althoughgroup
the treatment pati-
(35.4%) patients reported no effect.
to almost total relief of pain, whereas 17 of 48 in the pla- ents were asked to maintain their daily
and 97% in the placebo group. although pati- doses of
cebo group,
(35.4%) patients 27 reported
(56.3%) patients
no effect. reported no
in the pla- analgesic
ents therapy
were asked to throughout
maintain theirthedaily
study, in the
doses of
effect of treatment, whereas
cebo group, 27 (56.3%) patients reported no ef-21 (43.8%) pati- group receiving the standardized rose-hip
analgesic therapy throughout the study, in the pow-
ents of
fect reported
treatment, various degrees
whereas 21 of improvement.
(43.8%) patients der, 7 (14.6%)
group receivingpatients reduced their
the standardized consumpti-
rose-hip pow-
when pain
reported reliefdegrees
various was assessed on a yes-or-no
of improvement. when on of
der, 7 nsaids, and none
(14.6%) patients increased
reduced it. in
their contrast,
consumpti-
basis,relief
pain significantly more patients
was assessed in the treatment
on a yes-or-no basis, sig- 4 (8.3%)
on patients
of nsaids, andinnone
the placebo
increased group decrea-
it. in contrast,
group compared
nificantly more patients with the placebo
in the treatment groupgroup
indi- 4 (8.3%) patients in the placebo group 4decrea-
sed their consumption ofnsaids, and (8.3%)
 40

26sed their consumption ofnsaids, and 4 (8.3%) tive measures of hip and knee flexion is diffi-
patients increased it. The decrease in nsaid use cult to explain. The large-scale, controlled trial8
patients
in increasedgroup
the treatment it. The was
decrease in nsaid
statistically use
signifi- cado/
of soybeansoybean
avo-cado/ unsaponifiables in 101 instances
unsaponifiables in 101
in the treatment group was statistically
cant (p< 0.016); however, the between-group signifi- instances of osteoarthritis of the hip and knee
of osteoarthritis of the hip and 62 of the 62 of
cant (p< 0.016);
difference however,
was not. the between-group
Three (6.3%) patients in the showed
the kneeashowed
similar, sharp difference
a similar, between the
sharp difference be-
treatment group and 2 (4.2%) inpatients
difference was not. Three (6.3%) in the
the placebo therapeutic
tween response ofresponse
the therapeutic the 2 joints.
of theThe fact
2 joints.
treatment group and 2 (4.2%) in
group decreased their consumption of parace the placebo that fact
The the hipthatjoint
the iship
a ball
jointand
is a socket,
ball and whereas
socket,
group in
amol. decreased
the placebotheir group,
consumption
1 of the of parace
2 (50%) the knee difference, and the possibility
whereas the knee difference, and the possibil- that the
amol. in the placebo group, 1
patients taking an opioid drug (tramadol) of the 2 (50%)
re- pain is differently mediated in the 2 joints
ity that the pain is differently mediated in the 2 is ba-
patients
duced theirtaking an opioidofdrug
consumption that (tramadol)
drug. re- sed on
joints unsupported
is ba-sed conjecture.conjecture.
on unsupported
duced their consumption of that drug.
The only adverse event reported was mild pain is
Pain is the
the cardinal
cardinal symptom
symptom of of osteoarthritis.
osteoarthritis.
The only adverse
gastrointestinal event (2
discomfort reported was mild
[4.2%] patients in due to degeneration of the cartilage and
due to degeneration of the cartilage and lack lack
gastrointestinal
each group). discomfort (2 [4.2%] patients in of joint stability, small intra-articular traumas do
of joint stability, small intra-articular traumas do
each group).
Degree of joint pain relief on a scale from 0 (no occur. injuries
occur. injuries of of this
this kind
kind are are reflected
reflected in in bio-
bio-
chemical responses,
chemical responses,some someof of whichwhich involveinvolvecy-
tokines. cytokines have proinflammatory ef-
cytokines. cytokines have proinflammatory
9
srhP (n=48) 9
ef-
fects that are manifested as as episodes
episodes of of pain,
pain,
Placebo (n=48) joint swelling, and redness. our interest in in these
these
60 –
mechanisms lies in the fact that the
fact that the standar- standar-
dized rose-hip powder used here inhibits inhibits the the
chemotaxis that thatisisaastep step
% of Patients

40 – polymorphonuclear chemotaxis in
in the
the proinflammatory
proinflammatory action
action of various
of various cytoki-
cytokines.
nes.could
This This could be the bebasis
the basis
of the ofeffects
the effects
of theofstan- the
20 – standardized
dardized rose-hiprose-hip
powder powder
on joint onpain.
joint further
4 pain.4
further support
support for an for an anti-inflammatory
anti-inflammatory actionof action this
of this compound
compound is that the is thatserum theconcentration
serum concen- of
0 c-reactive
tration of c-reactive protein, a markerprotein, for in-flammation,
a marker for in-
0 1 2 3 4
decreases
flammation,significantly decreasesduring treatment
significantly with
during
degree of Joint Pain relief the compound,
treatment with the as shown
compound, by a mean as shown (sd) de- by
crease from 8.25 (4.9) mg/l
a mean (sd) decrease from 8.25 (4.9) mg/l to to 6.67 (2.6) mg/l. 3,4

The
6.67 basic
(2.6) mg/l. mechanism
3,4 of themechanism
The basic anti-inflammatory of the
figure. degree of joint pain relief on a scale from
impact) to 4 (almost total relief of pain) after 4 action of standardized
anti-inflammatory actionrose-hip
of standardizedpowder rose- does
0 (no impact) to 4 (almost total relief of pain)
months of treatment with standardized rose-hip not
hip reside
powder in adoes
blockade of theincyclooxygenase
not reside a blockade of
after 4 months of treatment with standardized
powder (srhP) or placebo. P=0.035 for srhP ver- pathway,
the cyclooxygenaseas is known to be the case
pathway, as isfor the anti-
known to
rose-hip powder (srhP) or placebo. P=0.035 for
sus placebo (mann-Whitney test). inflammatory
be the case for drugs (as-pirin and other
the anti-inflammatory nsaids)
drugs (as-
srhP versus placebo (mann-Whitney test).
and
pirin the
andherbal other remedy
nsaids)ginger.
10,11
and the herbal This was shown
remedy
in a study
ginger. 10,11 measuring platelet aggregation
12
This was shown in a study measuring 12 dur-
Discussion
discussion ing treatment with the same
platelet aggregation during treatment with the standardized rose-
The
The aim aim of
of this controlled study was to answer hip
same powder
standardized in dosesrose-hip
far higher than that
powder in doses usedfar in
the
the following
following questions:
questions: does the standardized the present study. In contrast
higher than that used in the present study. in to drugs inhibiting
rose-hip
rose-hip powder improve
powder improve mobility of the hip the cyclooxygena-se
contrast to drugs inhibiting pathway, the platelet
cyclooxygena- aggre-
and
and knee joints? does itit reduce
knee joints? does reduce the the functional
functional gation was not affected
se pathway, platelet aggregation was not af-by these high doses. In
disability
disability inin performing
performing adlsadlsthatthatgoes
goeswith
with ther-
the fact,
fected the bypowderthese high seemsdoses.to in stabilize
fact, the cell mem-
powder
estricted hip and knee joint movements?
restricted hip and knee joint movements? does does branes, as shown by the
seems to stabilize cell membranes, as shown by finding that erythro-
itit relieve
relieve pain?
pain? wewe found
found that,
that, inin the group tre-
the group tre- cytes from individuals
the finding treated with
that erythrocytes fromthe powder,
individuals
ated with standardized rose-hip
ated with standardized rose-hip powder, (1) powder, (1) when routinely stored
treated with the powder, when in a blood bank, leak
routinely stored less
functional
functional capacity
capacity of the hip,
of the hip, as assessed by
as assessed by hemoglobin
in a blood bank, thanleak ex-pected. 13
less hemoglobin than ex-
an objective method, was improved;
an objective method, was improved; (2) the im- (2) the im- pected. 13
pact
pact on on functional
functional capacity
capacity and and adls,
adls, when
when Natural vitamins C and E are present in stan-
measured subjectively, was less
measured subjectively, was less pronounced; pronounced; dardized
natural vitamins rose-hipcpowder. and e However,
are present it does
in stan-not
and
and (3) (3)pain
painwas was reduced
reduced in approximately
in approximately two seem
dardized likely that these
rose-hip powder.vitamins
however, can explain
it does not the
two thirds of these patients.
thirds of these patients. This response This response rate
rate was present findings because vitamin
seem likely that these vitamins can explain the C was not in-
was comparable
comparable to that
to that reported
reported for ginger,6,7
for ginger, 6,7
ano- volved in the
present findings because anti-inflammatory action
vitamin c was not in- reported
ano-ther natural remedy often used
ther natural remedy often used by patients with by patients for rose-hip powder,
volved in the anti-inflammatory
4
and vitamin actionE has been
reported
with osteoarthritis.
osteoarthritis. reported to be ineffective for symptomatic re-lief
for rose-hip powder,4 and vitamin e has been
of osteoarthritis.14 Also, the prevalence of gastro-
reported to be ineffective for symptomatic re-
The difference between
The difference betweenthe the effects
effects onon objec-
objective intestinal adverse events was low in the present
lief of osteoarthritis.14 also, the prevalence of
measures of hip and knee flexion is difficult to gastrointestinal adverse events was low in the
explain. The large-scale, controlled trial8 of avo- present trial and similar to that of placebo. mo-

trial and similar to that of placebo. Mo-
Reover, several years of use of the powder in the
Scandinavian countries has not disclosed signi-
ficant data on any adverse events.
Although a significant increase was found in
mobility of the hip joint and a significant decre-
ase in pain was found in the majority of patients
who received the standardized rose-hip pow-
der, the clinical benefit of 4 months of treatment
should not be overestimated. Future research
should include long-term studies to evaluate
joint mobility, clinical improvements, and con-
sumption of nsaids and other types of conco-
mitant pain-reducing medicine. It is also impor-
tant to find the active ingredient(s) in rose-hip
and clarify whether the content of such active
ingredient(s) (as well as the content of vitamins
and minerals) differ among subtypes, as spe-
cies of rose-hip can be very different from each
other regarding biological activity.15
Conclusions
in this study population, standardized rose-hip
powder reduced symptoms of osteoarthritis, as
64.6% of patients reported at least some reduc-
tion of pain while receiving treatment. Standar-
dized rose-hip powder may improve hip flexion
and reduce pain in patients with osteoarthritis.
References
1. Hochberg mc, Altman rd, brandt kd, et al.
guidelines for the medical management
of osteoarthritis. part 1. american college
of Rheumatology. Arthritis Rheum.
1995;38:1535–1540.
2. Simon ls, Lanza fl, lipsky pe, et al. prelimi-
nary study of the safety and efficacy of
sc-58635, a novel cyclooxygenase 2
inhibitor: efficacy and safety in two place-
bocontrolled trials in osteoarthritis and
rheumatoid arthritis, and studies of gastroin
testinal and platelet effects. Arthritis
Rheum. 1998;41:1591–1602.
3. Winther k, rein e, Kharazmi A. The anti-in-
flammatory properties of rose-hip. Inflam-
mopharmacology. 1999;7:63–68.
4. Kharazmi a, winther k. rose hip inhibits
chemotaxis and chemiluminescence of
human peripheral blood neutrophils in vitro
and reduces certain inflammatory parame-
ters in vivo. inflammopharmacology.
1999;7:377–386.
5. Kendall fp, mcCreary ek. Muscles: Testing
and Function. 3rd ed. Baltimore: Williams &
Wilkins; 1983.
6. Bliddal h, rosetzsky a, schlichting p, et al. A
41
randomized, placebo-controlled, crossover
study of ginger extracts and ibuprofen in
osteoarthritis. osteoarthritis cartilage.
2000;8:9–12.
7. Altman rd, markussen kc. effects of a
ginger extract on knee pain in patients with
osteoarthritis. arthritis rheum. 2001;44:2531–
2538.
8. Blotman f, maheu e, wulwik a, et al.
efficacy and safety of avocado/soybean
unsaponifiables in the treatment of
symptomatic osteoarthritis of the knee
and hip. A progressive, multicenter, three-
month, randomised, double-blind,
placebo-controlled trial. rev rhum engl ed.
1997;64:825–834.
9. Endres s, ghorbani r, kelley ve, et al. The
effect of dietary supplementation with
n-3 polyunsaturated fatty acids on the
synthesis of interleukin-l and tumor necrosis
factor by mononuclear cells. n engl J med.
1989;320:265–271.
10. Mustafa T, srivastava kc, Jensen kb. drug
development report (9): pharmacology
of ginger, zingiber officinale. J drug dev.
1993;6:25–39.
11. Kiuchi f, iwakami s, shibuya m, et al.
inhibition of prostaglandin and leukotriene
biosynthesis by gingerols and diarylhepta-
noids. chem pharm bull (Tokyo). 1992;
40:387–391.
12 Winther k. impact on coagulation, platelet
function and fibrinolysis. in: proceedings
of the 3rd international exhibition and
Conference on Nutraceuticals and Food
for Vitality; May 3–5, 2000; Geneva, Switzer-
land. Abstract.
13 Winther k, kharazmi a, rein e. rose-hip,
given as a standardised dry powder, exerts
anti-inflammatory and cell preserving
properties in humans. presented at the 2nd
International Congress on Coronary Artery
Disease; October 18–21, 1998; Florence,
Italy.
14. Brand c, snaddon J, Bailey M, Cicuttini F.
Vitamin E is ineffective for symptomatic
relief of knee osteoarthritis: A six month
double blind, randomised, placebo
controlled study. Ann Rheum dis.
2001;60:946–949.
15. Brandt k, a° Kesson B. Health promoting
compounds in vegetables and fruit. Plant
Production. 2002;29:43–44.
42

The antI-Inflammatory ProPertIes of

rose-hIP
K. WInther 1,*, e. reIn1 and a. KharazmI2
1
Department of Clinical Chemistry, Kolding Hospital, Kolding, Denmark
2
Department of Clinical Microbiology, University Hospital (Rigshospitalet), Copenhagen, Denmark

Received 2 December 1998; Revised 4 February 1999; Accepted 5 February 1999

Abstract – The anti-inflammatory properties of rose-hip are described in this short report. Rose-hip
extract reduced chemotaxis of peripheral blood neutrophils and monocytes of healthy subjects
in vitro. Daily intake of rose-hip powder for four weeks by healthy volunteers and patients suffering
from osteoarthritis, resulted in reduced serum c-reactive protein (crP) levels and reduced chemo-
taxis of peripheral blood neutrophils. The results indicate that rose-hip possesses anti-inflammatory
properties and might be used as a replacement or supplement for conventional drug therapies in
patients with osteoarthritis.

1.IntroductIon third patient had pain from osteoarthritis of the

There have been undocumented lay claims that
rose-hip, normally known for its high vitamin C
content, may reduce the pain in patients suffe-
ring from osteoarthritis. We have recently shown
that rose-hip extract reduced the chemotaxis
of peripheral blood polymorphonuclear leuko-
cytes (pmns) and monocytes in vitro [1]. This ac-
tivity was independent of the vitamin C content
of rose-hip. furthermore, the level of CRP and
the chemotaxis of neutrophils were reduced in
healthy subjects under rose-hip treatment. The
purpose of this study was to investigate whether
the natural product rose-hip, administered as
dry powder to volunteers of which four were suf-
fering from clinical osteoarthritis, had any effect
on the clinical signs and symptoms and certain
inflammatory parameters.
2. suBJects and methods
2.1. subjects
Eight male volunteers, free from any known all-
ergic, hepatic, cardiovascular or infectious di-
seases, mean age 52 years (range 47-62), were
entered into the study. Four of them had never
experienced any pain of muscular or joint origin.
The other four had all been engaged in hard
physical work in different areas of construction
for most of their adult life. One had suffered from
clinical osteoarthritis for more than 20 years,
with pain especially in the knee and elbow. The
pain had been alleviated by injections of stero-
id directly into the joints and by acetylsalicylic
acid and nonsteroid anti-inflammatory drugs
(nsaids). The second patient had osteoarthritis
and moderate pain in the knee and the ankle,
periodically relieved by acetylsalicylic acid. The
ankle and had been periodically treated with
nsaids and acetylsalicylic acid. The fourth pa-
tient had osteoarthritis of the elbow and shoul-
der of 10 years duration, normally treated with
aspirin or paracetamol. The volunteers were
treated with 45 grams (high dose) of Hyben Vi-
tal rose-hip daily for four weeks. The treatment
was withdrawn for at least one month, then fol-
lowed by another treatment for four weeks at
a daily dose of 10 grams (low dose). Rose-hip
was taken together with a main meal. After four
weeks of the high dose rose-hip intake, at the
end of treatment-free intervals and at the end
of the low dose intake, the volunteers were as-
ked about the possible side-effects, and blood
samples were collected for clinical chemistry
and pmn chemotaxis studies. all blood samp-les
were taken between 8 : 30 and 9 : 00 am by the
same laboratory technician after 30 mi-nutes of
rest, and analyzed immediately. for chemotaxis,
heparinized blood was taken using vacuotain-
ers. The time-lapse between blood sampling
and chemotaxis was the same for the patients
and control subjects.
2.2. rose-hip
Rose-hip powder of rosa canina was kindly pro-
vided by Hyben Vital, Langeland,
Denmark. The rose-hip powder used in
these studies was a well characterized and
standardized batch containing both seeds and
shell. During the drying procedure of the rose-hip
powder, the temperature never exceeded 40°c
for the in vitro studies, a water extract of rose-hip
was prepared. The extraction took place at 4°c
 28
43
2.3. crP determination 3. results and discussion
2.3. crPcrp
serum determination
was estimated by a turbidometric 3. results and
rose-hip discussion
extract at concentrations as low as 100
Serum
method crp was
using a estimated
hitachi 717by a turbidometric
turbidometer. crp Rose-hip extractthe
µg/ml inhibited at concentrations
chemotaxis of pmns as lowin asvitro
100
method using a hitachi 717 turbidometer.
antiserum was from orion diagnostica, helsinki, crp mg/ml inhibited the chemotaxis of
(data not shown). cell viability after incubation pmns in vitro
antiserum
finland. crp was from orion
dilution bufferdiagnostica,
and human crp Helsinki,
ca- (data not shown).
with rose-hip extractCellwas viability
greaterafter
thanincubation
98%. as
Finland. crp dilution buffer and human
librator was purchased from daco a/s, glos- crp ca- with rose-hip extract was greater
shown in fig. 1, serum crp levels, although than 98%.wi- as
librator was purchased
trup, denmark. The normalfrom daco
range a/s,
in our glos-
labora- shown in Fig.range,
thin normal 1, serum crp levels,
declined although
significantly wi-
both
trup, denmark.
tory is ≥ 10 mg/1.The normal range in our labora- thin normal range, declined significantly
in the high-dose (p ≤ 0.02) and in the low-dose both
tory is ≥ 10 mg/1. in the high-dose
groups (p ≤ 0.05) (p 0.02) and to
as ≤compared in the
the no-thera-
low-dose
2.4. chemotaxis groups (p ≤ 0.05) as compared
py group. The crp levels (mean ± sem) in the to the no-thera-
2.4. Chemotaxis
chemotaxis was carried out using a modified py group.
patient The crp
group werelevels
5.75(mean
± 2.95,±6.67
sem)±in theand
2.67 pa-
Chemotaxis
boyden chamber was carried outfor
assay [2]. using
the ain modified
vitro stu- tient group were 5.75 ± 2.95, 6.67
8.25 ± 4.98 with high dose, low dose and no ± 2.67 and 8.25
boyden
dies, pmns isolated from peripheralin blood
chamber assay [2]. for the vitro stu-
of ±therapy,
4.98 with high dose, low
respectively. dose
in the and no
control therapy,
group the
dies, pmns isolated from peripheral blood
the subjects were preincubated with various of the respectively. in the control group
crp levels were 4.75 ± 0.75, 4.00 ± 0.0 and 7.25the crp levels
subjects
dilutions were preincubated
of rose-hip with
extract for 30 various
rain at 37°c dilu- were
± 1.034.75with± high
0.75, dose,
4.00 ±low 0.0 and
dose7.25 and±no 1.03 with
thera-
tions of rose-hip extract for 30 rain at
following preincubation, chemotaxis of the cel- 37°c fol- high dose, low dose and no therapy,
py, respectively. The neutrophil chemotaxis respective-
lowing preincubation, chemotaxis of
ls towards the chemotactic peptide f-met-leu- the cel-ls ly. The are
data neutrophil
shown in chemotaxis data are shown
Table 1. chemotaxis towards in
towards
phe (tmlp) theatchemotactic
a concentrationpeptide f-met-leu-
of 10-5-5 m or zymo- Table 1. chemotaxis towards fmlp
fmlp declined by approximately 60% and 50%, declined by
phe (tmlp) at a concentration of 10 m or zymo- approximately 60% and 50%, in the high dose
san activated serum (zas) at a dilution of 1 : 200 in the high dose and low dose group, respec-
san activated serum (zas) at a dilution of 1 : 200 and low dose group, respectively, with p values
was tested. for the in vivo studies, the chemota- tively, with p values of 0.01 and 0.02. chemota-
was tested. for the in vivo studies, the chemota- of 0.01 and 0.02. chemota-xis towards zas also
xis of peripheral blood neutrophils from healthy xis towards zas also declined in both the high
xis of peripheral blood neutrophils from healthy declined in both the high dose (p ≤ 0.01) and
control subjects and patients towards fmlp and dose (p ≤ 0.01) and in the low dose groups
control subjects and patients towards fmlp and in the low dose groups (p ≤ 0.02). The decline in
zas was determined. The migrated ceils were (p ≤ 0.02). The decline in chemotaxis of cells
zas was determined. The migrated ceils were chemotaxis of cells from the patients and the
counted by a computer-assisted image analy- from the patients and the controls under treat-
counted by a computer-assisted image analy- controls under treatment with rose-hip was simi-
sis system. ment with rose-hip was similar. The decline in
sis system. lar. The decline in chemotaxis was observed in all
chemotaxis was observed in all the 8 subjects.
the 8 subjects. The mean ± sem values for fmlp
2.5. statistical analysis The mean ± sem values for fmlp were 187 ± 60
2.5. Statistical Analysis were 187 ± 60 compared to 370 ± 39 and for
statistical analysis
statistical analysis of
of the
the data
data waswas performed
performed compared to 370 ± 39 and for zas 414 ± 136
zas 414 ± 136 compared to 673 ± 27 in the high
by using the wilcoxon test for
by using the wilcoxon test for matched matched pairs.
pairs. all compared
dose patient togroup
673 ± compared
27 in the high with dose patient
no therapy.
all data are given as mean ± sem. p
data are given as mean ± sem. p values of ≤ values of ≤ in the high dose control group the mean ± high
group compared with no therapy. in the sem
0.05 were
0.05 were considered
considered significant.
significant. dose control
response to fmlpgroupwasthe101mean
± 45 as ± sem response
compared to
to fmlp
308 ± 22wasand101 ± 45272
to zas as ± compared to 308 ± 22
125 as compared to
and4-to49
600 zas
as 272 ± 125 asto
compared compared
no therapy. to 600 4- 49
10 – as compared to no therapy.
The salient finding of the present study is that
The salient
rose-hip, finding
given of the
as dry present
powder, study crp
lowered is thatle-
rose-hip,
vels given asand
significantly dry powder,
inhibitedlowered
chemotaxis crp le- of
8–
vels significantly and inhibited
peripheral blood neutrophils in human male chemotaxis of
peripheral To
volunteers. bloodour neutrophils
knowledge,inthis human
findingmalehas
not been reported before. There are very has
volunteers. To our knowledge, this finding few
6–
not been
reports reported
in the before.
literature on There
other are very few
properties of
reports in the literature on other
rose-hip. rose-hip has been used as source of properties of
rose-hip.crose-hip
vitamin in tea and has other
been products
used as source [3]. cells of
vitamin
such c in tea and other products
as polymorphonuclear leukocytes [3]. cells
(pmns)
4– suchmonocytes
and as polymorphonuclear
are involved leukocytes (pmns)
in the inflammato-
and
ry monocytes
process are involved
and tissue damage in the inflammato-
in inflammatory
ry process
diseases suchandastissue damage
arthritis in inflammatory
and atherosclerosis [4].
diseases
The damage suchisascaused and
arthritisby theatherosclerosis
release of prote- [4].
2– The damage
olytic is caused
and hydrolytic by the release
enzymes as wellof asprote-
toxic
olytic and hydrolytic enzymes as
oxygen radicals [5]. acetylsalicylic acid, non- well as toxic
No oxygenanti-inflammatory
steroid radicals [5]. acetylsalicylic acid, non-
drugs and glucocorti-
High dose treatment Low dose steroidhave
coids anti-inflammatory
been used fordrugs and glucocorti-
the treatment of the-
0 coids
se have [6,
diseases been used for
7]. These the have
drugs treatment of the-
a variety of
figure 1. c-reactive protein levels in the serum of eight se diseases
side effects [6, such7]. as
These drugs
gastric have a
erosion andvariety
kidney of
volunteers given high dose, low dose or no treatment. the side effects such
disturbances. The as gastricstudy
present erosion and kidney
demonstrates
values are given as mean ± sem in rag/1. there were sta-
tistically significant differences between crP levels from
disturbances.
that administrationThe present study
of rose-hip todemonstrates
patients with
the high dose group and no treatment (p ≤ 0.02) and low osteoarthritis, diagnosed on a
that administration of rose-hip to patientsclinical basis with
dose group and no treatment (p ≤ 0.05).treatment. reduced the levels
osteoarthritis, of the acute
diagnosed on a phaseclinicalprotein
basis,
30
44
reduced the levels of the acute phase protein in conclusion, the anti-inflammatory and pain-
crp and
crp and peripheral
peripheral blood blood neutropbil
neutropbil chemota-
chemota- Inrelieving
conclusion, the anti-inflammatory
properties and pain-
of the natural product ro-
xis. similar
xis. similar results
results were
were found
found in in the
the four
four healthy
healthy relieving properties of the natural product
sehip, combined with its safety, low price and ro-
subjects who had never experienced pain of
subjects who had never experienced pain of sehip,
ease combined with its safety,
of administration, providelowanprice and
attractive
osteoarthritis origin.
osteoarthritis origin. Symptoms
symptoms were were assessed
assessed as as ease of administration, provide an attractive
strategy to use rose-hip as part of a supplement
pain severity
pain severity on on a a scale
scale of of 1-10
1-10 andand change
change in in strategy to use rose-hip
to a therapeutic as part of a supplement
regimen
limitation of joint movement. Alleviation
limitation of joint movement. alleviation of phy- of phy- to a therapeutic regimen for osteoarthritis. a
for osteoarthritis. a large scale placebo-con-
sical symptoms
sical symptoms by by rose-hip
rose-hip in in the
the patients
patients corre-
corre- large scale placebo-con-trolled clinical study
trolled clinical study will be required to extend
lated very well with the reduced chemotaxis of will be required to extend confirmation of the
lated very well with the reduced chemotaxis of confirmation of the anti-inflammatory effect of
peripheral blood neutrophils and reduced level anti-inflammatory effect of rose-hip.
peripheral blood neutrophils and reduced level rose-hip.
of cre after the volunteers stopped taking rose-
of cre after the volunteers stopped taking rose-
hip, the chemotaxis of neutrophils and the levels References
hip, the chemotaxis of neutrophils and the levels acknowledgements
of crp rose to the untreated values. it is interes- 1. K. Winther, A. Kharazmi and B. rangaard
of
tingcrp rose to
to note thatthetheuntreated
initial crp values.
valuesitwere is interes-
hig- Technical
(1997). Cell preserving
ting assistance of and antiinflammato-
kirsten mossin, hanne
her in the patient than the control group. Thehig-
to note that the initial crp values were in- Tamstorf
ry property of rose-hip
and anne asanovski and (hyben vital). pos- of
support
her in the
hibition ofpatient
chemotaxis thanobserved
the control group.
in our study The
was in- thesible clinical implication, 1st. int. congress
hibition of chemotaxis observed in our study was danish rheumatism association is acknow-
comparable to that observed with acetylsalicy- ledged.
on coronary artery diseases: from preven-
comparable
lic acid as reported to thatby observed
matznerwith acetylsalicy-
et al. [8]. on the tion to intervention, p. 68. prague, czech
lic acid as reported by
other hand kemp and smith [9] showed matzner et al. [8].that
on thein- republic.
other hand kemp and smith [9] showed that in- references
cubation of neutrophils in vitro with sodium sali- 2. E Jensen and a. kharazmi (1991).
cubation of neutrophils in vitro with 1. k. winther, a. kharazmi and b. rangaard
cylate increased the chemotaxis of sodium
these cells.sali- computer-assisted image analysis assay
cylate increased theresponse
chemotaxis (1997). cell preserving and antiinflammato-
a similar increased wasofobserved
these cells. in of human neutrophil chemotaxis in vitro,
a similar increased response was observed in ry property of rose-hip (hyben vital). pos-
normal individuals after ingestion of sodium sa- J. immunol. method 144, 43-48.
normal
licylate individuals after ingestion
[9]. some non-steroid of sodium sa-
anti-inflammatory sible clinical implication, 1st. int. congress
3. A. leung and s. foster (1996).
licylate
drugs such [9]. some non-steroid
as ibuprofen at inanti-inflammatory
vivo obtainable on coronary artery
Encyclopedia of common diseases: from preven-
natural
drugs such as ibuprofen
concentrations at in vivo locomotion
inhibited neutrophil obtainable ingredients. John wiley, new york. czech
tion to intervention, p. 68. prague,
by 50%, similar to
concentrations our findings
inhibited with rose-hip
neutrophil locomotion [10- 4. Erepublic.
m. Ridker, M. Cushman, M. J. Stampfer,
12].50%,
by The patients
similar towho complained
our findings with of mild pain
rose-hip [10- 2. R.
e EJensen
Traceyand anda. C.kharazmi
H. Hennekens (1991). (1997).
of osteoarthritis origin, reported
12]. The patients who complained of mild pain that their pain computer-assisted
Inflammation, aspirin and the image analysis
risk of assay
cardio-
declined
of after 14
osteoarthritis daysreported
origin, of rose-hip thatintake.
their painThe of human
vascular neutrophil
diseases chemotaxis
in apparently in vitro,
healthy
pain relieving
declined aftereffect
14 days of rose-hip in these
of rose-hip patients
intake. The men,J. immunol.
New Eng. method
J. Med. 144,
336,43-48.
973-979.
was comparable to that of nsaid
pain relieving effect of rose-hip in these patients and acetylsa- 3. a. leung and s. foster
5. E. D. Harris, Jr. (1988). Pathogenesis(1996). of
licylic
was acid. in all cases
comparable to thatthe pain returned
of nsaid and acetylsa- 12-14 encyclopedia
rheumatoid of common
arthritis: A disorder natural
associated
days after stopping intake. no allergic
licylic acid. in all cases the pain returned 12-14 reactions withingredients. John wiley,
dysfunctional new york. in:
immunoregulation,
or gastrointestinal
days after stoppingdisturbances
intake. no allergicwere observed
reactions 4. e m. ridker, m.Basic
Inflammation: cushman,
principlesm. J. stampfer,
and clinical
or gastrointestinal disturbances weredifference
during therapy. There was no major observed r. e TraceyJ.and
correlates, c. h.i.hennekens
h. Gallin, m. Goldstein (1997).
and
between
during the pain
therapy. Therealleviating
was no effect
major of rose-hip
difference R. Snyderman (eds),
inflammation, aspirin pp.and751-773.
the risk raven
of
given at the two different doses. Three patients Press, new york. diseases in apparently
cardiovascular
between the pain alleviating effect of rose-hip
had total pain relief from rose-hip and were un- 6. M. C. Hochberger,
healthy men, newr.eng. d. Altman,
J. med.K.336,D. 973-979.
given at the two different doses. Three patients
able to distinguish the difference between the 5. Brandt, B. M. Jr.
e. d. harris, Clarck,
(1988). R A. Dieppe, m. r.of
pathogenesis
had total pain relief from rose-hip and were un-
high dose and the low dose. however, one pati- griffin, r. w. moskowitz
rheumatoid arthritis: aand T. J. Schnitzer
disorder associated
able to distinguish the difference between the
ent felt that high dose gave him total relief whe- (1995). Guidelines for the medical
with dysfunctional immunoregulation, in:
high dose and the low dose. however, one pati-
reas low dose decreased the pain dramatically management
ent inflammation:of osteoarthritis
basic principles(part one),
and clinical
but felt
not that high dose gave him total relief whe-
completely. Arthritis Rheum. 38, 1535-1540.
correlates, J. h. gallin, i. m. goldstein and
reas low dose decreased the pain dramatically 7. M.
but not completely. r. C. Hochberger,
snyderman r. d.
(eds), pp.altman,
751-773. K. D.raven
Brandt, b. m.
press, new york.clarck, e a. dieppe, m. r.
6. Griffin, r. w. moskowitz
m. c. hochberger, and
r. d. T. J. Schnitzer
altman, k. d.
table 1. chemotaxis of peripheral blood neutro- (1995). Guidelines for the medical manage-
brandt, b. m. clarck, r a. dieppe, m. r.
phils from the eight volunteers at the end of high ment of osteoarthritis (part two), Arthritis
griffin, r. w. moskowitz and T. J. schnitzer
dose intake, 28 days after cessation of intake Rheum. 38, 1541 - 1546.
(1995). guidelines for the medical
and at the end of low dose intake of rose-hip
management of osteoarthritis (part one),
powder. the results are given as mean ± sem 8. Y. Matzner, r. drexler and m. levy (1984).
arthritis rheum. 38, 1535-1540.
Effect of dipyrone, acetylsalicylic acid and
high dose no therapy low dose 7. m. c. hochberger,
acetaminophen on r. d. altman,
human k. d. che-
neutrophil
brandt, b.
motaxis, eur.m.J.clarck, e a.14,
clin. invest. dieppe,
440-443. m. r.
griffin, r. w. moskowitz and
9. A. Kemp and J. smith (1987). The effect of T. J. schnitzer
chemotaxis (fmlp) 144 ± 38.7a 339 ± 24.0 172 ± 18.0b
chemotaxis (zas) 343 ± 89.7a 637 ± 29.3 432 ± 39.9b
(1995). guidelines
salicylate on humanfor the medical
leukocyte manage-
migration,
a
comparison of high dose with no therapy p ≤ 0.01. ment
clin. of immunoi.
exp. osteoarthritis (part two), arthritis
49, 233-238.
b
comparison of low dose with no therapy p ≤ 0.02. rheum. 38, 1541 - 1546.
10. I rivkin, v. foschi and c. h. rosen (1976).
 45

inhibition of in vitro neutrophil chemotaxis atory agents on human neutrophil functions

and spontaneous motility by anti-inflamm-
atory agents (39518), ptvc. Soe. Exp. Biol.
Med. 153, 236- 240.
11. H. B. Kaplan, h. s. edelson, h. m. korchak,
w. r given, s. abramson and g. weismann
(1984). effect of non-steroidal anti-inflamm-
in vitro and in vivo, biochem. pharmacol.
33, 371-378.
12. E. G. Maderazo, S. E Breaux and C. L.
Woronick (1984). inhibition of human
polymorphonuclear leukocyte cell response
by ibuprofen, J. pharm. sci. 73, 1403-1406.

rose hIP InhIBIts chemotaxIs and

chemIlumInescence of human PerIPhe-
ral Blood neutroPhIls in VItro and re-
duces certaIn Inflammatory Parameters
In VIVo
Arsalan KharazmI1* and KaJ WInther 2
1
Department of Clinical Microbiology, Rigshospitalet afsnit 7806, Tagensvej 20, DK-2200
Copenhagen, Denmark
2
Department of Clinical Chemistry, Kolding Hospital, Kolding, Denmark

Received 25 may 1999; revised 12 July 1999; accepted 15 July 1999

Abstract – Objective and design: The objective of this study was to investigate the leucocyte-rela-
ted antiinflammatory properties of rose hip.

Materials and methods: The effect of rose hip on a number of inflammatory parameters was eva-
luated using the following models: (1) The effect of rose hip extract on chemotaxis and chemilu-
minescence of peripheral blood polymorphonuclear leucocytes (pmns) from healthy subjects in
vitro; (2) The effect of rose hip administered to healthy subjects on serum levels of creatinine and
c-reactive protein and on chemotaxis and chemiluminescence of peripheral blood pmns.

Results: rose hip extract at concentrations higher than 500 mg/ml inhibited the chemotaxis and
chemiluminescence of peripheral blood polymorphonuclear leucocytes in vitro. daily intake of rose
hip powder at doses of 45 grams or lower by healthy subjects resulted in reduced chemota-xis of
peripheral blood pmns and reduced the level of serum creatinine and acute phase protein crp.

Conclusions: RThese results indicate that rose hip possesses antiinflammatory properties and might
be used as a replacement or supplement for conventional drug therapies in some inflammatory
diseases such as arthritis.

Key words: Rose hip; rosa canina; neutrophil; chemotaxis; CRP; antiinflammatory.
46
1. Introduction
Inflammatory diseases such as arthritis involve
a broad spectrum of different clinical manifes-
tations. Inflammatory cells such as polymorpho-
nuclear leucocytes have been shown to be in-
volved in the inflammatory process and tissue
damage. Inflammatory cytokines such as Tnf
appear to be involved in the amplification of
the disease process. The damage is caused by
the release of proteolytic and hydrolytic enzy-
mes as well as toxic reactive oxygen radicals
from these cells activated in the tissue and joints
(harris, 1988). Therapy of inflammatory diseases
involves alleviation of the symptoms associated
with the disease, such as, relief of pain, reduc-
tion of inflammation and increase of motion.
acetylsalicylic acid (aspirin) and other non-ste-
roid anti-inflammatory drugs such as ibuprofen
methotrexate and naproxen, and glucocorti-
coids have been used for the treatment of ar-
thritis (Hochberger et al., 1995a, b; ridker, et al.,
1997). control of the symptoms with these drugs
requires long term daily treatment. These drugs
have a variety of toxic and other side effects,
such as gastric erosion and adverse effects on
kidneys and liver. Some of these drugs, particu-
larly the glucocorticoids, inhibit the imuune re-
sponse to infections. Therefore, there is a great
need for alternative therapies for the manage-
ment of arthritis which can eliminate the need
for conventional drugs and their side-effects,
particularly for prolonged daily use. in a short
communication we have reported on the anti-
inflammatory activity of rose hip in four subjects
suffering from mild osteoarthritis (Winther et al.,
1999). The purpose of this study was to investi-
gate in more detail the anti-inflammatory pro-
perty of the natural product rose hip, utilizing
in vitro methods in a larger number of healthy
subjects. 2.
Materials and methods
2.1. Rose hip
The extract was prepared by incubating 80 mg
of Hyben Vital rose hip (Langeland, Denmark) dry
powder from Rosa canina with 4 ml of mi-nimal
essential medium (mem) containing 50 units/ml
of penicillin and 0.05 mg/ml of strepto-mycin, for
19 h at 4°C. The extracts were pre-pared from
either the whole fruit powder, the shells or the
seeds. The shells and the seeds were separated
from each other by splitting the dried fruit and
separating the shells from the seeds manually.
They were then ground in a mortar. Chemical
analyses of hyben vital rose hip was performed
by Steins Laboratorium a/s, Holstebro, Denmark.
following incubation of the powders in mem,
the mixtures were centrifuged at 4000 rpm for
10 min. The supernatants were collected, ster-
ile filtered and diluted further. The ph of extract
preparations was adjusted to ph 7.2 before use.
2.2. Chemotaxis
The chemotaxis assay was performed using a
modified Boyden chamber technique as previ-
ously described (Jensen and Kharazmi, 1991).
pmns isolated from peripheral blood of heal-
thy subjects were preincubated with different
dilutions of rose hip extract for 30 min at 37°c.
following preincubation, the chemotaxis of the
cells towards the chemotactic peptide f-Met-
Leu-Phe (fmlp) or zymosan activated serum
(zas) were tested. The migrated cells were
counted by a computer-assisted image analy-
sis system.
2.3 chemiluminescence
Chemiluminescence assay was used as a
measure of oxygen radical generation by ac-
tivated pmns. The method was performed as
previously described (kharazmi et al., 1984).
pmns were preincubated with different dilu-
tions of rose hip extract and then stimulated
with either fmlp or opsonized zymosan. The
oxidative burst response of the activated cells
was measured by a luminometer (1250-lkb
Wal-lace).
2.4. subjects
Thirteen healthy volunteers represented by
both sexes with a mean age of 47 years (range
30-59 years) were included in this study.
All the subjects included were without known
cardio-vascular, immunological, kidney, liver, al-
lergic, rheumatological or haematological dis-
orders. The volunteers were treated with 45 g
of Hyben Vital rose hip daily for 28 days (high
dose), followed for another 28 days during which
Hyben Vital rose hip was not taken. At the end
of this period, the volunteers received another
treatment of rose hip at a dose of 10 g daily for
28 days (low dose). Before inclusion, all volun-
teers went through a screening procedure to
assure that none of the above mentioned dis-
eases were present. Moreover, before inclusion
blood samples were taken for C-reactive pro-
tein (crp) measurement to assure that none of
the included volunteers suffered from unknown
in-fectious diseases.
2.5. Blood chemistry
Blood potassium, sodium, serum creatinine, ala-
nine aminotransferase, alkaline phosphatase,
lactate dehydrogenase, bilirubin, hemoglo-
bin and total cholesterol were also measured
before initiation of the treatment. All measure-
 33
47
before initiation of the treatment. all measure- The inhibition of chemotaxis by rose hip shells at
ments were
ments were performed
performed according
according to to the
the con-
con- the 1000mg/ml
both the 1000µg/mland 500
andµg/ml levels levels
500 µg/ml was signifi-
was
ventional laboratory
ventional laboratoryroutine.
routine.All the
all above
the above men- cantly higher
significantly than than
higher that of seeds
that (p ≤ 0.01
of seeds (p ≤and
0.01p
tioned parameters except serum
mentioned parameters except serum creatini- creatinine ≤ 0.04,
and p ≤respectively). when comparing
0.04, respectively). rose hip
when comparing
andand
ne crp crpwere repeated
were repeated after 5, 10,
after 5, 10,21 21andand 28 shells with whole powder, there was significantly
rose hip shells with whole powder, there was sig-
days of treatment, 28 days after
28 days of treatment, 28 days after stoppingstopping rose higher inhibition
nificantly higher by rose hipbyshells
inhibition roseathip
500 µg/ml
shells at
hip therapy
rose hip therapyand and
again at the
again atend
the of endlow of dose
low (p ≤ 0.03) (p
500µg/ml but≤ not
0.03)atbut
1000
notµg/ml.
at 1000 µg/ml.
treatment. Serum creatinine and
dose treatment. serum creatinine and c-reac- C-reactive
protein
tive (crp)
protein werewere
(crp) tested before
tested therapy,
before after
therapy, 3.3. chemiluminescence
3.3. chemiluminescence
10 and 28 days of treatment, 28 days
after 10 and 28 days of treatment, 28 days fol- following As shown
as shown in in Table
Table 2,2, rose
rose hip
hip extract
extract inhibited
inhibited
cessation
lowing of theoftreatment
cessation and and
the treatment finally at the
finally at the chemiluminescence
the chemiluminescence of ofpmns
pmnsactivated
activatedby by
the end of low dose treatment. rose hiptaken
end of low dose treatment. Rose hip was was opsonized zymosan. Adjustment of
opsonized zymosan. adjustment of ph to physi- ph to physi-
together
taken with a meal
together with aatmeal
12.00atnoon.
12.00on the days
noon. on ological values
ological values in in the
theextract
extractdid
didnot
notinfluence
influence
of blood sampling rose hip was taken
the days of blood sampling rose hip was taken together the inhibitory effect markedly.
the inhibitory effect markedly. vitamin cVitamin Cinincrys-
crys-
with a light
together meal
with at 09.00
a light meala.m.,
at two
09.00 hours
a.m., before
two tallineform
talline formused
usedas ascontrol
controlup
uptotoaaconcentrati-
concentrati-
blood sampling. blood sampling
hours before blood sampling. blood sampling was always on of 5000 µg/ml had almost
on of 5000 µg/ml had almost no effect onno effect onpmn
pmn
performed
was alwaysafter 15 min at
performed rest15
after sitting
min in atarest
chair.
sit- chemiluminescence when the ph of vitamin cC
chemiluminescence when the ph of vitamin
ting in a chair. solution was
solution was adjusted
adjustedtotothe thephysiological
physiological phph
7.2.
2.6. statistical analysis vitamin
7.2. E (alpha-tochopherol)
vitamin e (alpha-tochopherol) was alsowas used as
also
Statistical analysis of the data was performed a known antioxidant control. vitamin E at a con-
2.6. statistical analysis used as a known antioxidant control. vitamin
by using wilcoxon test for matched pairs, p va- centration of over 1µg/ml inhibited chemilumi-
statistical analysis of the data was performed e at a concentration of over 1µg/ml inhibited
lues of < 0.05 were considered significant. nescence.
by using wilcoxon test for matched pairs, p va- chemiluminescence.
lues of < 0.05 were considered significant.
3. results table 1. chemical analyses of the commercially
3. results
3.1. analysis of rose hip
3.1. analysis of rose available hyben Vital rose hip powder. the va-
Table 1 shows the hip chemical analyses of the lues are given for 100 g of dry powder
Table 1 shows
commercially available the chemical Hyben analyses
Vital rose of the hip.
commercially available hyben
Rose hip powder contains proteins, carbohyd- vital rose hip. protein 6.2 g
rose carbohydrate 39.0 g
rates,hipa low powder amount contains
of fat and proteins, carbohyd-
several vitamins fat 4.0 g
rates, a low amount of fat and
such as vitamin A, vitamin B, vitamin C, vitamin several vitamins vitamin c 560 mg
such
E andasvitamin
vitamin K.a,
The vitamin
powder vitamin
b, also c, vitamin
contains seve- energy 916 kJ
eral
and vitaminUptake
minerals. k. The powder
of vitamin alsoCcontains
present seve-in Hy-
ral
ben minerals.
Vital powder uptake wasof as
vitamin
goodcas, present
or even in bet-
hy-
ben vital powder was as good
ter than, that of vitamin C when given in tablet as, or even bet-
ter than,
form. Thethat of vitamin cand
concentrations when givenof
kinetics in uptake
tablet
form.
through Thethe concentrations
gastrointestinal and kinetics
tract of theof uptake
equiva- 100 –
through
lent of 250 themg gastrointestinal
vitamin C in rose tracthip of powder
the equiva- was
lent of 250 mg vitamin c in rose
similar to 500 mg vitamin C in tablet form. The hip powder was
80 –
similar
better to 500 mg vitamin
absorption of vitamin c inCtablet
in roseform.
hip pow- The
%inhibition

better
der may absorption
be due toofavitamin c in rose
larger surface hipof
area pow-
rose
der
hip may
powder be due to a largerto
as compared surface
vitamin area of rose
C tablets. 60 –
hip powder as compared to vitamin c tablets.
3.2. chemotaxis 40 –
3.2. chemotaxis
Initial dose-response experiments were perfor-
med and it was found
initial dose-response that the extract
experiments of rose
were perfor-
hip atand
med concentrations
it was found that equivalent
the extractto 500mg/ml
of rose 20 –
andat
hip higher inhibited chemotaxis
concentrations equivalentoftopmns 500µg/ml in vit-
ro. ashigher
and shown in Fig. 1,
inhibited rose hip extract
chemotaxis of pmns atincon-vit- 0
centrations of 500 µg/ml and
ro. as shown in fig. 1, rose hip extract at con- higher inhibited
chemotaxis of
centrations of500
human µg/mlperipheral
and higher blood
inhibitedneu- 1000 500 1000 500 1000 500 µg/ml
trophils; ph-adjusted
chemotaxis of humanrose hip extract
peripheral blood at these
neu-
concentrations was as strongly inhibitory shell seeds whole powder
trophils; ph-adjusted rose hip extract at as thesethe
non-ph-adjusted rose hip extract.
concentrations was as strongly inhibitory as the The two ma- figure 1. effect of rose hip extract on polymorpho-
jor parts of rose hip
non-ph-adjusted rose– hipshells and seeds
extract. The two -- were
ma- nuclear leukocytes (Pmn) chemotaxis in vitro.
tested separately for their activity
jor parts of rose hip – shells and seeds -- were on pmn che- cells were preincubated with various concentra-
motaxis. it was shown that by
tested separately for their activity on pmn che- far most of the in- tions of rose hip powder as given in the x-axis for
hibitory activity resided in
motaxis. it was shown that by far most of thethe shells (Fig. 1). The 30 min. the data are presented as percent inhibi-
inhibition of
inhibitory chemotaxis
activity resided byin rose
thehip shells(fig.
shells at both1). tion of Pmn chemotaxis for each subject tested.
34
48
34
table 2. effect of rose hip extract on human periphe- figure 2. levels of serum c-reactive protein (crP)
table 2. effect
ral blood of rose hip extract on
polymorphonuclear human periphe-
leucocyte (Pmn) figure
as 2. levels
given in mg/lof from
serum c-reactive
subjects protein
on the start(crP)
(day
ral blood polymorphonuclear
chemiluminescence. the data leucocyte (Pmn)
are presented as as
0), given
10 days in and
mg/l28from subjects
days on the of
during intake start
rose(day
hip
chemiluminescence. the data are
percent inhibition as compared withpresented
control as 0),
and1035days
daysand 28cessation
after days during intake of(day
of treatment rose hip
63)
percent inhibition as compared with control and 35 days after cessation of treatment (day
with 45 g daily intake of rose hip. the results are 63)
rose hip extract with 45 g daily intake of rose hip. the
given as mean 4 ± sem values from 13 subjects.results are
concship extract
rose given as mean
1 subject 2 subject 3 subject the mean value4on ± day
sem10values from 13 subjects.
was significantly lower
(µg/ml)
concs
1 subject 2 subject 3 subject the
than that on day 0 and day 63 (p ≤ 0.05). lower
mean value on day 10 was significantly
(µg/ml)
2500 37 27 57 than that on day 0 and day 63 (p ≤ 0.05).
1000
2500 8
37 8
27 12
57
500
1000 0
8 0
8 nd
12
500
nd: not determined. 0 0 nd 4.2. chemotaxis
nd: not determined. 4.2.
pmnchemotaxis
chemotaxis in the period during which the
4.
4. ex
ex vivo
vivostudies
studies 4.2.
pmnchemotaxis
chemotaxis
volunteers had not in the
taken period
any during
rose hip which
powder the
4.1.
4. ex Blood chemistry
vivo studies
4.1. Blood chemistry PMN chemotaxis
volunteers had not in the
taken
was compared with values obtained in the period
any rose during
hip which
powder
No
4.1. significant
Blood chemistry
no significant changes
changes occuredoccured in in potassium,
potassium, the
was volunteers
compared
preceeding 28had with
days not taken
values
(figs 3 andany 4).rose
obtainedThehip inpow-
mean the±
sodium,
no alanine
significant
sodium, alanine changes aminotransferase,
occured in potassium,
aminotransferase, alkaline
alkaline der
semwas
preceeding
valuecompared
of28
pmn days with
(figsvalues
chemotaxis 3 and obtained
4). Thethe
towards meaninche-
the±
phosphatase,
sodium, alanine lactate dehydrogenase,
aminotransferase,
phosphatase, lactate dehydrogenase, bilirubin, bilirubin,
alkaline preceeding
motactic peptide fmlp was 103.6 ± 60.0 when±
sem value of28
pmn days (Figs
chemotaxis 3 and 4).
towards The mean
the che-
haemoglobin
phosphatase,or
haemoglobin total
total cholesterol
lactate
or dehydrogenase,
cholesterol comparing
comparing va-
bilirubin,
va- sem
tested value
motactic of pmn
on peptide
day 28 of chemotaxis
fmlp
treatment towards
was 103.6 with± rose
60.0the che-
when
hip as
lues from
haemoglobin before
lues from before intake
or total
intake to values
cholesterol obtained
to values comparing
obtained after after
va- motactic
tested on day
compared peptide
with28298.9 fmlp was
of treatment 103.6
±-26.2 when ±
withblood60.0
rose hip when
sam-as
5, 10,
lues 21
from and
before 28 days
intake ofto high
values
5, 10, 21 and 28 days of high dose therapy, va- dose therapy,
obtained va-
after tested
compared on day
with 28 of
298.9 treatment
±-26.2 when
ples were taken 28 days after cessation of rose with rose
blood hip
sam- as
lues
5, obtained
10,obtained
lues 21 and 28 28 days
28daysdays ofafter
high
after stopping intake
dose therapy,
stopping intake and and
va- compared
ples were with
taken 298.9
28 days ±-26.2
after
hip intake (p < 0.001). The mean ± sem values when
cessationblood of sam-
rose
those
lues
those obtained
obtained
obtained 28atat the
days end
end of
the after low dose
stopping
of low doseintaketherapy
and
therapy ples
hip werechemotaxis
intake
for7 –pmn taken 28 days
(p < 0.001). The after
towards meancessation of
± semactiva-
zymosan rose
values
(data
those not shown). hip intake
for7 –pmn (p <
serumchemotaxis0.001).
(zas) whichtowards The mean ±
containszymosan sem values
activa- for
(data obtained
not shown). at the end of low dose therapy ted the biologically
(data not shown). pmn
active chemotaxis
ted serum chemotactic towards
(zas) which zymosan
contains
factor c5athe was activated
biologically
218 ± 60.0 se-
Serum creatinine, however, declined significant- rum
active
as (zas)
6 compared
– which
chemotactic contains
to 529.9 factor the biologically
c5awhen
± 39.9 was 218 tested active
± 60.028
serum creatinine, however, declined significant-
ly compared
serum creatinine, withhowever,
initial value given as mean
declined chemotactic
as6 compared factor
– after cessationto 529.9 c5a ±was
39.9218 ± with
when 60.0 as com-
tested 28
ly compared with initial value givensignificant-
as mean days of treatment rose hip
4ly±compared
sem (90.0 ± with2.1 _mol/1) to values obtained pared
days to 529.9
after ±
cessation39.9 when
of tested
treatment 28
withdaysrose after
hip
4 ± sem (90.0 ± 2.1 initial
µmol/1) value given obtained
to values as mean (p < 0.001). The decline in chemotaxis response
after 10 days: (87.4 ± 1.8 µmol/1) and 28 days cessation
(p 4 – 0.001).of treatment with rose hip (presponse
< 0.001).
4 ± sem
after 10 (90.0
days: ±(87.4 2.1 µmol/1) to values
± 1.8 µmol/1) andobtained
28 days to4<fmlp wasThe 65% decline
in 12 out in chemotaxis
of 13 volunteers:
–decline in chemotaxis response to fmlp was
a
(84.9
after 4 ± 1.9 µmol/1) of intake, respectively 3 (p The
to fmlp was 65% in 12ofout of 13 volunteers: a
(84.9 410± days:
1.9 µmol/1)(87.4 ±of1.8 µmol/1)
intake, and 28 days
respectively 3 (p considerable decline chemotaxis response.
<(84.9
0.001). when
4 ± when treatment
1.9 µmol/1) had
of intake, been stopped
respectively for 65% in 12 out decline
considerable of 13 volunteers:
of chemotaxisa considerable
response.
< 0.001). treatment had been stopped3 for (p The2 – decline in chemotactic response to zas
28 days the serum creatinine levels significantly decline
The of chemotaxis
decline response. The decline in
<
280.001).
days the when serumtreatment
creatinine had levels
been significantly
stopped for was2 – 59%, alsoina chemotactic
considerable decline response in to
12 zasout
increased
28 days the(93.2 ± creatinine
1.9_mol/1) (p < significantly
0.001) and chemotactic
was also response to zasdecline
was 59%, also outa
increased serum
(93.2 ± 1.9µmol/1)levels (p < 0.001) and of 1359%,
volunteers. a considerable
it was the same in 12
subject who
were similar to values obtained before intake. considerable
of013 volunteers.decline in 12 out of 13 volunteers.
increased
were similar(93.2 to values ± 1.9µmol/1)
obtained(p < 0.001)
before intake.and did not respond toit therapywas theinsame both subject
assays. who
were similar to values obtained before intake. It
didwasnotthe same
respond
Day 0
subject
to therapy
Day
whoindid
10 both
Day
not
23
respond
assays.Day 63
to
0
The data on c-reactive protein are given in fig. therapy in both Day 0 assays. Day 10 Day 23 Day 63
The
2. data on
similar to c-reactive
the findingsprotein on serum are given in fig.
creatinine,
The data
2. similar on c-reactive protein are given in fig.
crp valuestowere the findings on serum
also decreased creatinine,
during intake
2.
crp similar
values to the findings
wereinitialalso mean on
decreased serum creatinine,
during intake
of rose hip. The 4 ± sem values of 350 –
crp
of values
rose hip. were
The also decreased
initial mean 4 ± during
sem valuesintake of
crp were 5.38 ± 0.4 mg/l and declined to 3.31 350 –
of rose hip. The±initial mean 4 declined
± sem values of
±crp
0.49weremg/ 5.38and 4.31 0.4 ±mg/l
0.47andmg/l, after 10 and to 3.3128
crp were
± 0.49ofmg/ 5.38 ± 0.4
andrespectively mg/l
4.31 ± 0.47 mg/l,and declined
afterAfter to
10 and 3.31
28
days intake (p < 0.05). stop-
± 0.49ofmg/
days intakeandrespectively
4.31 ± 0.47 mg/l, (p < afterafter
0.05). 10 and 28
stop- 300 –
ping therapy for 28 days, the levels increased to 300 –
days of
ping intake for
therapy respectively
28 days, (p
the < levels
0.05). after
increasedstop-
5.75 ± 0.54 mg/l (p < 0.051 as compared with
(fmlp)

ping
to 5.75 therapy
± 0.54 mg/l for 28 (pdays,
< 0.051 theaslevels
comparedincreased with
(fmlp)

that previously.
to 5.75 ± 0.54
that previously. mg/l (p < 0.051 as compared with 250 –
that previously. 250 –
chemotaxis
chemotaxis

7– 200 –
7– 200 –

6–
(mg/l)

150 –
pmn

6–
(mg/l)

150 –
pmn

4– 100 –
crp

4– 100 –
crp

2– 50 –
2– 50 –

0 0
0 Day 0 Day 10 Day 23 Day 63 0 No treatment Low dose High dose
Day 0 Day 10 Day 23 Day 63 No treatment Low dose High dose

350 –
350 –
 49
35
Figure 3. chemotaxis of peripheral blood poly- 5. dIscussIon
figure 3. chemotaxis of peripheral blood poly- 5. dIscussIon
morphonuclearleukocytes (Pmn) from subjects The studies described in this communication
morphonuclear leukocytes (Pmn) from subjects The studies described in this from communication
during and 28 days after cessation of treatment demonstrate that the extract rose hip in-
during and 28 days after cessation of treatment demonstrate that the extract from rose hip in-
with 45 g (high dose) or 10 g (low dose) daily in- hibited, in vitro, the chemotaxis and oxidative
with 45 g (high dose) or 10 g (low dose) daily in- hibited, in vitro,ofthe
takeof rose hip for 28 days or no treatment. the burst response thechemotaxis
human peripheral and oxidative blood
take of rose hip for 28 days or no treatment. the burst response of the human peripheral blood
chemotaxis is determined towards the chemo- polymorphonuclear leucocytes, important and
chemotaxis is determined towards the chemo- polymorphonuclear leucocytes, important
tacticpetide fmlP. the results are given as mean abundant inflammatory cells involved in theand pa-
tactic petide fmlP. the results are given as mean
± sem number of ceils migrated from 13 subjects. abundant inflammatory
thogenesis cells involved
of arthritis. Furthermore, in the pa-
administrati-
± sem
The number
mean of ceils migrated
chemotaxis values forfrom 13 low
both subjects.
dose thogenesis
on of rose hip of arthritis.
to healthy furthermore,
volunteersadministrati-
for a period
the mean chemotaxis values for both low
and high dose were significantly lower than that dose on28
of of days
rose hip to healthy
inhibited volunteers for response
the chemotactic a period
and
for
7 – high dose were significantly lower than that
no treatment group (p ≤ 0.01 and p ≤ 0.001, of 28 days inhibited
neutrophils the chemotactic
by approximately 60% or response
higher.
for no treatment group (p ≤ 0.01 and p ≤ 0.001,
respectively). moreover,
of neutrophils roseby hip lowered the 60%
approximately levelor ofhigher.
serum
respectively). creatinine and the acute
moreover, rose hip lowered the level of serum phase protein C-re-
6–
active
creatinine protein
and inthe volunteers
acute phase with protein
values within c-re-
4.3. Clinical findings normal range, in
active protein which is below
volunteers with 10values
mg/l. Serum within
No4 –allergic reactions or any other side-effects
4.3. clinical findings creatinine
normal range, levelswhich were iswithin below the 10normal
mg/l. serum range
were observed during therapy. Only two volun- in all the volunteers (males
creatinine levels were within the normal range 55-125 and females
no allergic
teers reactions
complained of or anygastrointestinal
mild other side-effects
gas 45-100 µmol/l). However, the55-125
decline was statis-
in all the volunteers (males and females
were
2 – observed during therapy. only two volun-
disturbances at the end of the study, while on tically
45-100significant
µmol/l). however, and might theindicate
decline enhanced
was statis-
teers
the complained
high dose. of mild gastrointestinal gas glomerular filtration.
tically significant andThe might blood
indicatechemistry
enhanced data
disturbances at the end of the study, while on presented in this study showed that intake of
0 high dose.
glomerular filtration. The blood chemistry data
the rose hip hadinno harmful
Day 0 Day 10 Day 23 Day 63 presented this study effect
showed on that
any of the livei
intake of
functions determined in
rose hip had no harmful effect on any of thethis study.
livei functions determined in this study.
Studies on the inhibition of neutrophil oxidative
350 – burst
studies response by rose hip
on the inhibition of extract
neutrophil showed
oxidative that
this effect was not due to
burst response by rose hip extract showed that vitamin C content of
the extract. This is shown by
this effect was not due to vitamin c content ofthe inability of ph-ad-
300 – justed vitamin
the extract. ThisC to inhibit
is shown by thechemiluminescence
inability of ph-ad-
whereas phadjusted rose
justed vitamin c to inhibit chemiluminescence hip extarct was still as
pmn chemotaxis (fmlp)

inhibitory as non-ph-adjusted
whereas phadjusted rose hip extarct was extract. In order to
still as
determine which part of rose
inhibitory as non-ph-adjusted extract. in order hip exhibited the
250 –
inhibitory
to determine effect on chemotaxis
which part of rose thehip extract
exhibited from
shells, seeds and
the inhibitory effect theon whole powderthe
chemotaxis were pre-
extract
pared and tested in pmn
from shells, seeds and the whole powder were chemotaxis assay. as
200 –
shown in Fig. 1 the major
prepared and tested in pmn chemotaxis assay. inhibitory activity was
found
as shown to reside
in fig.in1the theshells.
major it will be interes-ting
inhibitory activity
150 – to
was identify
found the compound(s)
to reside in the shells. responsible for the
it will be interes-
anti-inflammatory activity of rose hip. The inhi-
ting to identify the compound(s) responsible for
bition of chemotaxis observed in our study was
the anti-inflammatory activity of rose hip.
100 – comparable to that observed with acetylsalicyl-
The inhibition of chemotaxis observed in our
ic acid as reported by Matzner et al. (1984). on
study was comparable to that observed with
the other hand Kemp et al. (1982) showed that
50 – acetylsalicylic acid as reported by matzner et
incubation of neutrophils in vitro with sodium
al. (1984). on the other hand kemp et al. (1982)
salicylate increased the chemota-xis of these
showed that incubation of neutrophils in vitro
cells. Similar increased response was observed
0 with sodium salicylate increased the chemota-
No treatment Low dose High dose
in normal individuals after ingestion of sodium
xis of these cells. similar increased response was
salicylate (Kemp et al., 1982). Some non-steroid
observed in normal individuals after ingestion
anti-inflammatory drugs such as ibuprofen at
figure 4. chemotaxis of peripheral blood poly- of sodium salicylate (kemp et al., 1982). some
attainable concentrations during therapy has
morphonuclear leukocytes (Pmn) from subjects non-steroid
been shown anti-inflammatory
to inhibit neutrophildrugs locomotion such as by
during and 28 days after cessation of treatment ibuprofen at attainable concentrations
50%; a finding which is similar to our findings during
with
with 45 g (high dose) or 10 g (low dose) daily in- therapy
rose has been
hip (rivkin et al., shown
1976; to inhibitetneutrophil
kaplan al., 1984;
take of rose hip for 28 days or no treatment. the locomotion by
maderazo et al., 1984). 50%; a finding which is similar
chemotaxis is determined towards zymosan to our findings with rose hip (rivkin et al., 1976;
activated serurm (zas). the results are given as kaplan et al., 1984; maderazo et al., 1984).
mean ± sem number of cells migrated from 13
subjects. the mear chemotaxis value for high
dose was significantly lower than that for no treat-
ment group (p ≤ 0.00l),
50

6. conclusion Matzner, y., drexler, r. and levy, m. (1984).

Rose hip possesses anti-inflammatory and anti- effect of dipyrone, acetylsalicylic acid and
oxidant properties. These properties are impor- aceta minophen on human neutropbal
tant in alleviation of tissue damage in the in- chemotaxis, eur. J. clin. invest. 14, 440-443.
flammatory sites. As a natural product, rose hip Ridker, p. m., cushman, m., stampfer, m. J.,
has no side effects, is safe and can be adminis- Tracey, r. p. and hennekens, c. h. (1997)
tered easily. it can be designed for daily con- inflammation, aspirin and the risk of
sumption as supplemental part of a therapeutic cardiovascular diseases in apparently
regimen for some inflammatory diseases, or as healthy men, n. eng. J. med. 336, 973-979.
a prophylactic regimen for individuals having a Rivkin, i., foschi, v. and rosen, c. h. (1976).
genetic or environmental predisposition to the- inhibition of in vitro neutrophil chemotaxis
se diseases. a large scale placebo-controlled and spontaneous motility by anti-inflamm-
clinical study will be required to extend con- atory agents (39518), proc. soc. exp. biol.
firmation of the antiinflammatory effect of rose med. 153 236-240.
hip described in this report. Winther, k., rein, e. and kharazmi, a. (1999).
The anti-inflammatory properties of rose-hip
References inflammopharmacol. 7, 63-68.36
Harris, Jr., e. d. (1988). pathogenesis of
rheumatoid arthritis: a disorder associa-
ted with dysfunctional immunoregulation.
in: inflammation. basic principles and
clinical correlates, gallin, J. h., goldstein,
i. m. and snyderman, r. (eds), pp. 751-
773. raven press, new york.
Hochberger, m. c., altman, r. d., brandt,
k. d., clarck, b. m., dieppe, e a., griffin,
m. r., moskowitz, r. w. and schnitzer, T. J.
(1995). Guidelines for the medical mana-
gement of osteoarthritis (part 1), arthritis
rheum. 38, 1535-1540.
Hochberger, m. c., altman, r. d., brandt, k. d.,
clarck, b. m., dieppe, e a., griffin, m. r.,
moskowitz, r. w. and schnitzer, T. J. (1995).
guidelines for the medical management of
oseteoarthritis (part 2), Arthritis Rheum. 38,
1541 - 1546.
Jensen, E and Kharazmi, A. (1991). Computer-
assisted image analysis assay of human
neutrophil chemotaxis in vitro, J. lmmunol.
methods 144, 43-48.
Kaplan, h. b., edelson, h. s., korchak, h. m.,
given, w. e, abramson, s. and weismann,
g. (1984) effect of non-steroidal anti-in-
flammatory agents on human neutrophil
functions in vitro and in vivo Biochem.
Pharmacol. 33, 371-378.
Kemp, A. and Smith, J. (1982). The effect of
salicylate on human leukocyte migration,
clin. exp. lmmunol. 49, 233-238.
Kharazmi, a., hØiby, n., döing, g. and valerius,
n. h. (1986). pseudomonas aeruginosa
exopro teases inhibit human neutrophil
chemiluminescence, infect. immun. 44,
587-591.
Maderazo, e. g., breaux, s. p. and woronick,
c. l. (1986). inhibition of human polymor-
phonuclear leukocyte cell response by
ibuprofen, J. pharm. sci. 73, 1403-1406.
 51

OVERVIEW ON ABSTRACTS
Rein E., Kharazmi A., Thamsborg g., Winther K. (2004): A herbal remedy, made from a subspecies of
rose-hip rosa canina, reduces symptoms of knee and hip osteoarthritis. Abstract. 9th World Congress
of the Osteoarthritis Research Society International, Chicago.

Warmholm O., Skaar S., Hedman E., Molmer H., Ek l. (2001): Alternative Therapies in health and medi-
cine. First International Conference on Complementary Alternative & Integrative Medicine Re-
search, San Francisco.

Winther K., Kharazmi A., Rein e. (2005): A powder made from a subspecies of rose hip (rosa cani-na),
reduces womac symptoms score as well as cholesterol level in patients suffering from oste-oarthritis.
Abstract. 10th World Congress on Osteoarthritis, Boston.

Winther K., Kharazmi a. (2004): A powder prepared from seeds and shells of a subtype of rose-hip
rose canina reduces pain in patients with osteoarthritis of the hand – a double-blind, place-con-
trolled, randomized study. Abstract. 9th World Congress of the Osteoarthritis Research Society Inter-
national, Chicago.

Winther K., kharazmi a. (2004): a subtype of rosa canina exerts anti-inflammatory properties. abs-
tract. The xxix nordic congress in clinical chemistry, Malmö.

Winther K. (2006): breakthroughs in reducing inflammation. abstract. nuTracon conference,

Anaheim.

Winther, K. (2000): impact on coagulation, platelet function and fibrinolysis. abstract.

Copenhagen.

Rein E., Winther k. (2001): ldl-cholesterol and c-reactive protein is influenced by rose-hip, a ran-
domized, double blind, placebo controlled trial. abstract. xiv. international symposium on drugs
affecting metabolism, New York.

Kharazmi a., song J., winther k., hoiby n. (2000): natural medicine and the immune system: immu-no-
modulatory properties of ginseng and antiinflammatory properties of rose-hip. Abstract. vita-foods
International Conference 2000, Geneva.

Winther K., kharazmi a., rein e. (1998): Rose-hip, given as a standardised dry powder, exerts anti-
inflammatory and cell preserving properties in humans. Abstract. 2nd International Congress on
Coronary Artery Disease, Florence.

Rein e., Kharazmi a., winther k. (1999): rose-hip given as a standardised dry powder exerts
antiΩinflammatory properties, without influencing platelet aggregation and the coagulation cas-
cade. 1st International Congress on Heart Disease, Washington.

Bütner S., Jacobsen O., Andersen J.R., Winther K. (2004): The impact of a standardized powder made
from a subtype of Rose-hip on ibdQ symptom score in patients with Crohn’s Disease or Ul-cerative
Colitis. Abstract. Scandinavian Journal of Gastroenterology, Oslo.