Journal of Crohn's and Colitis (2014) 8, 16751683

Available online at www.sciencedirect.com

ScienceDirect

Hospitalisation, surgical and medical

recurrence rates in inflammatory bowel
disease 20032011A Danish
population-based cohort study
Marianne K. Vester-Andersen a,, Ida Vind a , Michelle V. Prosberg a ,
Bo G. Bengtsson b , Thomas Blixt c , Pia Munkholm d , Mikael Andersson e ,
Tine Jess e,f , Flemming Bendtsen a
a

Gastrounit, Medical Section, Hvidovre Hospital, Hvidovre, University of Copenhagen, Denmark

Department of Medical Gastroenterology, Rigshospitalet, Copenhagen, University of Copenhagen, Denmark
c
Department of Internal Medicine, Frederiksberg Hospital, Frederiksberg, University of Copenhagen, Denmark
d
Digestive Disease Centre, Medical Section, Herlev Hospital, Herlev, University of Copenhagen, Denmark
e
Department of Epidemiological Research, National Health Surveillance and Research, Copenhagen Denmark
f
Department of Clinical Epidemiology, University of Aalborg, Aalborg, Denmark
b

Received 26 April 2014; received in revised form 17 July 2014; accepted 26 July 2014
KEYWORDS
Inflammatory bowel disease;
Epidemiology;
Recurrence rates;
Hospitalisation;
Biological treatment

Abstract
Objective: The aim of this study is to evaluate the cumulative probability of recurrence and
admission rates in an inflammatory bowel disease (IBD) inception cohort diagnosed in 20032004.
Methods: Data on medications, phenotypes and surgery for 513 individuals with ulcerative
colitis (UC, n = 300) and Crohn's disease (CD, n = 213) were obtained from medical records and
linked to population-based health administrative database information. The admission rates and
cumulative probability of recurrences were estimated, and the association with the baseline
factors and medication was tested.

Abbreviations: 6-MP, 6-Mercaptopurine; Anti-TNF, Anti-tumour necrosis factor alpha; AZA, Azathioprine; CD, Crohn's disease; CE, Capsule
endoscopy; CPR, Central person registration; CT, Computed tomography; ECCO, European Crohn's and Colitis Organisation; EC-IBD, European
Collaborative Study Group on Inflammatory Bowel Disease; NPR, The National Patient Registry; MRI, Magnetic resonance imaging; IBD,
Inflammatory bowel disease; IBDU, Inflammatory bowel disease unclassified; IMM, Immunomodulators; RCT, Randomised clinical trial; SMI,
Small bowel imaging; UC, Ulcerative colitis; US, Ultra sonography.
Corresponding author at: Gastrounit, Medical Division, Hvidovre University Hospital, Kettegaard All 30, DK-2650 Hvidovre, Denmark.
Tel.: + 45 41429908.
E-mail addresses: marianne@kajbaek.dk (M.K. Vester-Andersen), ida.vind@regionh.dk (I. Vind), michelle_prosberg84@hotmail.com
(M.V. Prosberg), bobengtsson@dadlnet.dk2 (B.G. Bengtsson), tn@jtnautomatik.dk (T. Blixt), pia.munkholm@regionh.dk (P. Munkholm),
aso@ssi.dk (M. Andersson), tjs@ssi.dk (T. Jess), flemming.bendtsen@regionh.dk (F. Bendtsen).

http://dx.doi.org/10.1016/j.crohns.2014.07.010
1873-9946/ 2014 European Crohn's and Colitis Organisation. Published by Elsevier B.V. All rights reserved.

1676

M.K. Vester-Andersen et al.

Results: The cumulative risk of first recurrence after 1, 5 and 7 years was 40%, 63%, and 66% in CD
patients and 51%, 75%, and 79% in UC patients, respectively. The cumulative risk of first surgical
relapse was 6%, 18%, and 23% at 1, 5 and 7 years in CD respectively. One hundred and CD patients
(66%) and 142 UC patients (47%) had at least one IBD-related hospitalisation. The hospitalisation rate
decreased from 7.0 days/person-year in year one to 0.9 day at year 5 in CD, and from 4.7 days to
0.4 days for UC patients. Age above 40, current smoking, stricturing behaviour, and disease
localisation (colonic, ileocolonic, and upper-GI) at diagnosis were predictors of recurrence in CD. In
UC, age above 40 and former smoker status were predictors of recurrence and left-sided and
extensive colitis were predictors of first-time hospitalisation.
Conclusion: In an era of improved treatment options, the recurrence rates, but not the surgery
or hospitalisation rates, have decreased for CD but not for UC. The phenotypic characteristics at
diagnosis predict the risk of recurrence and hospitalisation.
2014 European Crohn's and Colitis Organisation. Published by Elsevier B.V. All rights reserved.

1. Introduction
Ulcerative colitis (UC) and Crohn's disease (CD) are inflammatory bowel diseases (IBDs) of unknown aetiology. The clinical
course is often unpredictable, although continuous effort has
been made to identify the factors that predict outcome.1,2 The
disease course ranges from indolent with prolonged periods of
remission to a chronic continuous course with incapacitating
symptoms. Severe cases may lead to medical refractory disease
and surgery. In the late 1990s biological therapy (i.e., anti-TNF
agents) was introduced in the treatment of IBD and treatment
with immunomodulators (IMM) is now more frequently used.3,4
A recent meta-analysis showed that the surgical rates over the
last decades have been declining5; however, a causal relationship with the altered treatment regimens could not be demonstrated. Several population-based studies from Europe and
North America have addressed the clinical course of CD over
decades and have focused on trends in medication use and
surgery rates,37 whilst fewer studies describe the clinical
course of UC in this way.8,9 However, all of these studies were
developed before the introduction of biological therapy and in a
period where immunomodulators were offered to fewer
patients. Furthermore, the relapse rates prior to this new
treatment era were reported to be rather high (69%86% in CD
and 78% in UC) after 5 years of follow-up in population-based
studies.1012 Patients with IBD are mainly treated in outpatient
clinics and only hospitalised during serious and incapacitating
flare-ups.13 More recent studies indicate that hospitalisation
rates have declined, whilst outpatient visits have increased.14
This change is most likely a consequence of institutional
changes although a lowered need of hospitalisation could
potentially be associated with improved remission rates.
The aim of the present study was to assess the recurrence
and admission rates in a Danish population-based cohort in the
new treatment era after 7 years of follow-up. Additionally, we
assessed the association of baseline factors and IBD treatment
with the recurrence rates, admission rates and surgery.

2. Methods
2.1. Study population
From January 1, 2003 to December 31, 2004, all patients
diagnosed with UC, CD and IBD-unclassified (IBDU) in a

well-defined Copenhagen area from a reference population

of 1,211,634 residents (23% of the total population of
Denmark) were included in an inception cohort.15 Registration
was carried out at 28 specialist practitioners (gastroenterologists and surgeons) and in 10 departments of gastroenterology
and internal medicine, including 2 paediatric units, covering
all 8 hospitals in the region. Medical records were reviewed
from November 1, 2011 to November 30, 2012. In all cases the
diagnosis was re-evaluated according to the Copenhagen
criteria16,17. The follow-up period was from the date of
diagnosis to December 31, 2011. Data on medical treatment,
disease localisation and behaviour, diagnostic procedures
(endoscopy, MRI-/CT-scans, ultrasound, capsule endoscopy)
and surgical procedures were registered prospectively.
Phenotypic classification was recorded by the use of the
Vienna classification in CD.18 In UC, the extent of disease was
defined as proctitis, left-sided colitis (inflammation with
maximum extent to the splenic flexure) and extensive colitis
(inflammation beyond the splenic flexure). Details on the
inclusion, exclusion and phenotypic definitions have been
described previously and can be assessed in the supplementary
material.15,19
The unique 10-digit personal identification number
(CPR-number) allocated to all Danish citizens at birth
enabled us to perform complete follow-up in the Danish
National Patient Registry20 with respect to hospital admissions and surgical procedures.

2.2. Definitions
In Denmark an admission is registered with a primary code
(diagnosis of action) and a secondary code(s), referring to
conditions or diseases that are relevant to the primary
diagnosis. An admission was defined as IBD-related when the
primary codes were CD (DK50.x), UC (DK51.x) or pre-defined
conditions related to IBD (see supplementary material).
Recurrences were recorded as all type, non-surgical,
or surgical as previously described by Romberg-Camps et
al.7 and in earlier publications of the European Collaborative
Study Group on Inflammatory Bowel Disease (EC-IBD)10,21
with modifications. Non-surgical (medical) recurrence was
defined as an episode of increased disease activity, leading
to an increased dose of current medication or the addition of
a new medication. Surgical recurrence was defined as an

Hospitalisation and recurrence rates in IBD - a 7-year population-based follow-up study

episode of increased disease activity requiring a surgical
intervention, i.e., 1) an intestinal resection or 2) perianal
fistula operation or abscess drainage with or without a
change in medication.
The Regional Ethics Committee approved the study
(H-1-2011-088) and permission was obtained from the Danish
Data Registry (01769 HVH-2012-027).

2.3. Statistical analysis

The data are given as numbers and percentages or medians
and ranges when appropriate. Survival curves displaying
cumulative recurrence rates were derived with the Aalen
Johansen method.22 The hospitalisation rates were calculated as the total number of hospitalisation days divided by
the total number of days under observation. We performed
unadjusted Cox regression analyses to evaluate the association between covariates, including; age (b 40/ 40 years),
sex (female/male), smoking status (never/current/former
smoker), disease localisation (CD) (L1/L2/L3/L4), disease
behaviour (CD) (B1/B2/B3), disease extent (UC) (E1/E2/E3),
use of 5-ASA (y/n), steroids (y/n), steroids within 3 months
of diagnosis (y/n), high-dose steroids above 50 mg (y/n),
azathioprine/6-mercaptopurine (y/n) or anti-TNF agents
(y/n) and the four endpoints (time to first all-type
recurrence, medical recurrence, first admission and first
resection [CD]/colectomy [UC]). The reference levels are
specified in Table 2. Subsequently, we performed an all
entry adjusted Cox regression analysis with all covariates
except steroids (y/n) and steroids within 3 months of
diagnosis (y/n). Only events occurring after the date of
diagnosis and amongst patients being followed-up were
included in the Cox regression analysis. A sensitivity analysis
of the Cox regression was performed only on patients aged
17 years of age or above, in order to test if the risk of
recurrence and surgery was due to an increased risk
in children aged below 17. A P value of b 0.05 was considered statistically significant. The statistical analysis was
performed with SAS (version 9.3) statistical software
(SAS Institute, Cary, NC, USA).

1677

Table 1 Demographic and clinical characteristics of the

513 IBD patients included in the study.19
At diagnosis
Age (years), median (IQR)
UC
CD
Male/female ratio
UC (%)
CD (%)
Vienna classification CD
A1 b 40 years
A2 40 years
Localisation (CD)
L1 terminal ileum
L2 colon
L3 ileocolon
L4 upper gastrointestinal
B1 inflammatory
B2 stricturing
B3 penetrating
Extension of disease (UC)
E1 proctitis
E2 left-sided
E3 extensive
Smoking habits at diagnosis
Current smoker
UC
CD
Former smoker
UC
CD
Never smoked
UC
CD

37.3 (1.593.5)
31.8 (7.185.2)
151 (50)/149 (50)
99 (47)/114 (54)
N (%)
133 (62)
80 (38)
N (%)
57
88
50
18
166
20
27

(27)
(41)
(24)
(9)
(78)
(9)
(13)

At diagnosis
93 (31)
127 (42)
80 (27)
At diagnosis
45 (15)
86 (40)
72 (24)
25 (12)
137 (46)
97 (46)

n: number; (%): percentage of total; IQR: interquartile range.

3. Results
The baseline characteristics of the 513 patients included in
the present study, as previously reported,19 are shown in
Table 1. After diagnostic reassessment, 213 patients fulfilled
the diagnostic criteria for CD and 300 fulfilled the diagnostic
criteria for UC, encompassing the follow-up cohort. In total,
28 patients did not fulfil the diagnostic criteria and were
excluded from further analyses. During re-evaluation, new
information regarding the residence site at the time of
diagnosis was achieved, and 16 were not incident cases, and
were, thus excluded from further analyses. Finally, 5 of the
original 27 IBDU patients (6 non-IBD, 3 non-incident, 5
changed to UC and 8 changed to CD after diagnostic
re-evaluation) could be classified as neither CD nor UC
(remained IBDU) and, due to the small sample size, were not
included in this study for further analyses. Flowcharts of the
patient selection are illustrated in a previous study.19 The
median follow-up time was 7.5 years (IQR 5.68.1) in UC
patients and 7.7 years (IQR 7.18.4) in CD patients.19

Overall 45 patients were younger than 17 years of age at

diagnosis (24 CD and 21 UC). Of these, 29 patients were
14 years or older at diagnosis. In the UC group, 35 patients
(12%) underwent colectomy. Sixty-two patients in the CD
group (29%) underwent a surgical resection during
follow-up.19

3.1. Recurrences in CD
In CD, the cumulative risks of first all-type recurrence were
40% (95% CI 3149), 63% (95% CI 5174), and 66% (95% CI
5477) after 1, 5 and 7 years. Fig. 1AB shows the
cumulative risk of the first (1A) and second (1B) recurrence
in CD for all types of recurrence. Of patients experiencing a
relapse, 54% had their second relapse within 2 years after
the first relapse (Fig. 1B) and 56% had their third relapse
within the second year after the second relapse. The
medical recurrences dominate the first all-type recurrences

1678

M.K. Vester-Andersen et al.

Figure 1 A. Plot of first all type relapse in Crohn's disease patients years after diagnosis. B. Plot of the second all type relapse in
Crohn's disease patients years after first relapse.

with cumulative risks of 36% (95% CI 2845), 58% (95% CI 4769),

and 61% (95% CI 4972) at 1, 5 and 7 years, respectively.
Overall, the cumulative risks of a first resection surgical relapse
were 6% (95% CI 411), 18% (95% CI 1325), and 23% (95% CI
1730) at 1, 5 and 7 years, respectively (Fig. 2). The risk of
having a second resection surgical relapse was 32% (95% CI

Figure 2

1751) 5 years after the first relapse. Two patients had a third
resection surgical relapse within 1 year of the second relapse.
As shown in Table 2, current smoking at diagnosis was
associated with an increased risk of first all-type recurrence
compared to non-smoking behaviour (P = .04), whereas
patients older than 40 years of age at diagnosis had a

Plot of the first resection surgical relapse in Crohn's disease patients years after diagnosis.

Hospitalisation and recurrence rates in IBD - a 7-year population-based follow-up study

reduced risk of all-type recurrence (P b .05) and first
medical recurrence (P b .05). Patients with stricturing
disease at diagnosis (B2) compared to patients with
inflammatory behaviour had an increased risk of a first
resection surgical recurrence (P = .02), whereas patients
with disease localisation of the colon, ileocolon and upper GI
compared to patients with localisation of the terminal ileum
had a reduced risk of first surgical recurrence (P = .01)
(Table 2).
The use of anti-TNF was associated with an increased risk
of first resection surgical recurrence and first resection in CD
(HR 2.82 95% CI 1.226.50 and HR 3.34 95% CI 1.507.44,
respectively). Furthermore, the use of AZA/6-MP was
associated with an increased risk of colectomy in UC
(HR 10.99 95% CI 3.4834.7). The use of high dose steroids
increased the risk of hospitalisation (HR 2.74 95% CI 1.75

1679

4.29) and the use of 5-ASA reduced the risk of colectomy

(HR 0.14 95% CI 0.040.58) in UC. AZA/6-MP did not
influence the risk of the first medical relapse in CD or UC;
however, AZA/6-MP was associated with a reduced risk of
first medical recurrence in CD in the unadjusted analysis
(HR 0.69 95% CI 0.441.06).

3.2. Recurrences in UC
In UC patients, the cumulative risks of first all-type
recurrence were 51% (95% CI 4061), 75% (95% CI 6185)
and 79% (95% CI 6489) at 1, 5 and 7 years, respectively
(Fig. 3) The cumulative risks of a first medical recurrence
were 49% (95% CI 3959), 73% (95% CI 5983) and 77% (95% CI
6387) at 1, 5 and 7 years, respectively.

Table 2 Association between the baseline factors and first all-type and first medical recurrences, first IBD-related
hospitalisation and resection/colectomy in CD and UC patients as determined by the adjusted Cox proportional hazard model.
Crohn's disease

Smoking status at diagnosis

Never smoker
Current Smoker
Former smoker
Age at diagnosis
A1 b 40 years
A2 40 years
Disease localisation at diagnosis
L1 terminal ileum
L2 colon
L3 ileocolon
L4 upper gastrointestinal
Disease behaviour at diagnosis
B1 inflammatory
B2 stricturing
B3 penetrating
Ulcerative colitis

Smoking status at diagnosis

Never smoker
Current smoker
Former smoker
Age at diagnosis
A1 b 40 years
A2 40 years
Disease extent at diagnosis
E1 proctitis
E2 left-sided
E3 extensive

First all type

recurrence

First medical
recurrence

First resection

First IBD-related
hospitalisation

First resection
surgical
recurrence

HR

HR

HR

HR

HR

95% CI

95% CI

Ref.
Ref.
1.65 1.112.44 1.4 NS
1.11 NS
1.05 NS

95% CI

95% CI

95% CI

Ref.
0.77 NS
0.62 NS

Ref.
1.08 NS
0.93 NS

Ref.
1.03 NS
0.92 NS

Ref.
Ref.
Ref.
0.58 0.390.85 0.62 0.420.93 1.45 NS

Ref.
0.97 NS

Ref.
1.22 NS

Ref.
1.07 NS
1.26 NS
0.75 NS

Ref.
0.93 NS
1.18 NS
0.65 NS

Ref.
Ref.
0.34 0.170.70 0.93 NS
0.32 0.150.69 0.74 NS
0.34 NS
0.88 NS

Ref.
0.31 0.140.68
0.32 0.140.77
0.27 0.080.96

Ref.
1.42 NS
1.69 NS

Ref.
0.9 NS
1.29 NS

Ref.
Ref.
8.63 4.2217.7 1.66 NS
0.97 NS
1.50 NS

Ref.
3.45 1.478.09
1.07 NS

First all type

Recurrence

First medical
Recurrence

Colectomy

First IBD-related
Hospitalisation

.First resection
surgical
recurrence

HR

HR

HR

HR

HR

95% CI

95% CI

95% CI

95% CI

Ref.
Ref.
Ref.
0.97 NS
0.93 NS
0.85 NS
1.60 1.102.34 1.70 1.172.47 1.05 NS

Ref.
0.89 NS
1.34 NS

N
NA
NA

Ref.
Ref.
Ref.
0.54 0.390.76 0.56 0.400.79 0.80 NS

Ref.
0.74 NS

NA
NA

Ref.
0.79 NS
0.70 NS

Ref.
0.77 NS
0.64 NS

95% CI

Ref.
Ref.
NA
0.22 0.070.7 1.84 1.113.03 NA
0.29 0.100.90 2.30 1.353.94 NA

HR: hazard ratio; 95% CI: 95% confidence interval; Ref.: reference level; NS: non-significant; NA: not applicable. Bold text indicates
significant results of the analyses.

1680

M.K. Vester-Andersen et al.

Figure 3

Plot of the first all type relapse in ulcerative colitis patients years after diagnosis.

Table 2 shows the adjusted analysis of UC patients.

Patients older than 40 years of age at diagnosis had a
reduced risk of both first all-type recurrence (P = b .05) and
medical recurrence (P = b .05). Patients who were formerly
smoking compared to non-smoking patients at diagnosis had
an increased risk of both all-type recurrence (P = .03) and
medical recurrence (P = .01) (Table 2).

3.3. Hospitalisation
A total of 140 CD (66%) and 142 UC (47%) patients were
hospitalised with IBD or IBD related conditions during
follow-up (13 cases (9 CD, 4 UC) had a primary code of
hospitalisation different than DK50.x and DK51.x; see
supplementary material for details). In the first year after
diagnosis, 104 (49%) CD patients and 99 (33%) UC patients
had been admitted to the hospital. The median time to first
hospitalisation was 44 days (IQR 6419) in CD patients and
65 days (IQR 10585) in UC patients. The median number
of days hospitalised was 13 days (IQR 729) in CD and
10 days (IQR 420) in UC. The hospitalisation rate of
all-type hospitalisations decreased from 7.0 days per
person-year (95% CI 6.77.4) in CD in the first year to 0.9
(95% CI 0.81.0) days per person-year in the fifth year. For
UC patients, the rate of IBD hospitalisations decreased from
4.7 (95% CI 4.55.0) days per person-year during the first
year to 0.4 (95% CI 0.30.4) days per person-year by the fifth
year.
We did not find any significant predictive factors of the
first hospitalisation in CD patients (Table 2).
UC patients with left-sided colitis (E2) and extensive
colitis (E3) were at an increased risk of first time
hospitalisation (P b .05) (Table 2).

3.4. First-time surgical intervention in CD

Stricturing disease (B2) was associated with resective
surgery (P b .05). Disease localisation of the colon (L2) or
ileocolon (L3) compared to localisation of the terminal ileum
was associated with a reduced risk of first resective surgery
(P b .05) (Table 2).

3.5. First-time surgical intervention in UC

In the unadjusted analysis, patients with extensive disease
had an increased risk of colectomy (HR 2.66 95% CI 1.08
6.52). In the adjusted analysis, as shown in Table 2,
left-sided and extensive colitis reduced the risk of colectomy
(HRE2 0.22 95% CI 0.070.70 and HRE3 0.29 95% CI 0.100.90,
respectively). Only patients with extensive (n = 28) and
left-sided colitis (n = 7) (at the time of surgery) had a
colectomy (3 E1 had progressed to E2 and 4 E1 had progressed
to E3 at the time of colectomy, whilst 9 E2 had progressed to
E3 at time of colectomy).
Repeating the Cox regression analyses without the
children below 17 years of age did not affect the point
estimates of the predicting variables of the outcomes for all
types of recurrence, medical recurrence, and resection/
colectomy in neither CD nor UC apart from the risk of first
resective recurrence in patients with stricturing disease
(B2), which became insignificant (HR 2.1; 95% CI 0.85.3)
whilst the risk of first resection in B2 was still significantly
increased (HR 8.6; 95% CI 4.217.7).

3.6. Use of azathioprine/6-mercaptopurine during

follow-up
In total 137 CD patients (64%) and 77 UC patients (26%)
received AZA/6-MP, respectively during follow-up. The
median time to receive AZA/6-MP from the time of diagnosis
was 188 days (IQR 62690) in CD and 433 days (IQR 148874)
in UC.
Amongst the 62 CD patients undergoing resective surgery
and the 35 UC patients undergoing colectomy, 29 CD
patients (47%) and 18 UC patients (51%) had received AZA/
6-MP before the first surgical intervention. The cumulative
probabilities of AZA/6-MP use prior to the first intestinal
resection (in the entire cohort, when censuring at surgery,
LFU or death) were 41%, 59% and 62% at 1, 5 and 7 years,
respectively. The cumulative risks of receiving AZA/6-MP in
UC before colectomy were 12%, 24% and 27% at 1, 5 and
7 years, respectively. The median durations of treatment
amongst operated patients receiving AZA/6-MP before
surgery were 885 days (IQR 1662128) in CD and 428 days
(IQR 245939) in UC.

Hospitalisation and recurrence rates in IBD - a 7-year population-based follow-up study

Table 3 Overall use of anti-TNF agents amongst 213 CD
patients as well as in relation to surgery.
Use of anti-TNF agents amongst
213 CD patients during follow-up
Type of anti-TNF agent, n (%)
Infliximab
Adalimumab
Time to first infusion (years)
Median years (IQR)
Number of all-type relapses after
anti-TNF treatment, n (%)
Number of medical relapses after
anti-TNF treatment, n (%)
Number of resective surgical
relapses after anti-TNF
treatment, n (%)

First treatment course

49 (98.0)
1 (2.0)
2.2
34 (68)
29 (58)
14 (28)

Use of anti-TNF according to Vienna First treatment course

classification at diagnosis, n (%)
A1 b 40 years
A2 40 years
L1 terminal ileum
L2 colon
L3 ileocolon
L4 upper GI
B1 inflammatory
B2 stricturing
B3 penetrating

39
11
6
22
17
5
36
2
12

Use of anti-TNF agents amongst 62

operated patients

First treatment course

Anti- TNF before first intestinal

resection, n (%)
Anti- TNF after first intestinal
resection, n (%)

12 (19)

(78)
(22)
(12)
(44)
(34)
(10)
(72)
(4)
(24)

8 (13)

n: number; (%): percentage of total; IQR: interquartile range.

3.7. Use of anti-TNF agents amongst CD patients

The use of anti-TNF agents is shown in Table 3. The median
time from diagnosis to the first infusion was 2.2 years
(IQR 0.54.7). The treatment regimen was based on past
recommendations with either on demand treatment, i.e., a
single dose of anti-TNF in the case of symptoms (5 patients
(10%)) or short treatment rounds (33 patients of the 49
infliximab users (67%) had 8 consecutive infusions).

4. Discussion
Treatment regimens have changed during the last decades
and RCTs and studies from tertiary centres report a change
in the clinical course of CD and UC associated with the
widespread use of IMM and the introduction of anti-TNF.2327
However, the present study is, to the best of our knowledge,
the first based on a population-based cohort with long-term
follow-up developed in this new treatment era that addresses
the risk and predictors of recurrence and hospitalisation. Our

1681

cohort was characterised by a high use of IMM and more

patients were treated with biological therapy compared to
previous studies.19
We found a higher rate of hospitalisations amongst CD
patients (66%) than UC patients (47%) overall, which is in
accordance with previous studies,28,29 and the hospitalisation
rate (expressed as days per person year) was highest the first
year after diagnosis and then decreased over time as
previously reported.6,28 However, in Denmark there seems
to be a declining hospitalisation rate within the first year
of diagnosis, which previously has been reported to be
83%13 compared to 49% in our study. The only predictor
of hospitalisation was disease extent in UC; not surprisingly
patients with more extensive disease, reflecting disease
severity, were at a higher risk of hospitalisation. Whether
the seemingly first year decrease in CD hospitalisations is
related to the introduction of immunomodulating and biological therapy or simply a result of the changed management of
care has not been clearly documented. Comparison of the
hospitalisation rates between countries is difficult due to the
disease management and health care facilities according to
the individual country observed.
The finding of a 5-year cumulative risk of all-type
recurrences of 63% in CD is markedly lower than previous
reports from Nordic population-based studies with all-type
recurrence rates at 5 years of 86% to 93%.11,30 However,
lower recurrence rates have been reported from Southern
European countries.10 The high and early use of IMM in our
cohort may explain the lower recurrence rate in CD in our
study. Although not statistically significant, though of clinical
relevance, the use of AZA/6-MP amongst CD patients could be
associated to a reduced risk of first medical recurrence in CD.
In UC patients we did not find a decrease in the cumulative risk
of all types of recurrences compared to previous studies.12,21
even though the use of immunomodulators in UC in our study
was higher compared to previous studies.7,31 The lack of effect
in this study may be due to the rather late introduction of IMM
during the disease course in UC. Fifty CD patients (24%)
received anti-TNF treatment. This is higher than in previous
studies,3,4,31 but the use of anti-TNF in our cohort was in the
early treatment era and characterised by short treatment
periods based on past recommendations, and treatment was
introduced more than two years after diagnosis. Furthermore
only 12 patients received anti-TNF before surgery.
In both UC and CD, patients older than 40 years of age at
diagnosis were at a reduced risk of an all type recurrence,
which is in agreement with previous studies7,31 This could
implicate a more benign phenotype in the elderly. In CD, we
have previously shown that children aged below 18 years at
diagnosis have more extensive disease.15 Furthermore we
could not show an association between upper GI localisation
and an increased risk of surgical recurrences as reported by
Wolters et al. and Romberg Camps et al.7,10 In our cohort,
8.5% of the CD patients were diagnosed with upper GI
involvement compared to 5.4% and 5%, respectively in the
study by Wolters et al. and Romberg-Camps et al. More
patients with associated upper-GI inflammation not requiring
surgery found by improved diagnostics in our cohort together
with optimised treatment could explain this finding. We did
not stratify based on the treatment duration or time of
implementation due to the small sample size, and introducing
the use of AZA/6-MP and anti-TNF into the model makes our

1682
findings confounded by indication because these treatment
regimens are administered to patients who are at a high risk of
recurrence and surgery. However, some population-based
studies have shown that the recently observed decrease in
surgical rates in CD is associated to the increasing and early
use of thiopurines.3,4 We have not been able to show a
decrease in the surgery rates in our cohort compared to
previous studies19 and the lack of efficacy of IMM in reducing
surgery rates may be due to a late introduction during the
disease course when intestinal damage has already occurred.32 A recent French, open-labelled, randomised controlled trial compared early AZA administration (within
6 months of diagnosis) and conventional management for the
corticosteroid-free and anti-TNF-free period in remission
during 3 years of follow-up and failed to show any beneficial
effect of early AZA.33 However, AZA/6-MP was introduced
with a median time of 188 days in our study, though only 47%
of the patients who had surgery received IMM; therefore
treatment optimisation may improve this outcome. Stricturing
disease behaviour, in our cohort, was associated with an
increased risk of surgical relapse and resection and it could be
speculated that a subclass of patients with strictures develop
fibrostenotic lesions at a very early stage that are not
modifiable by IMM therapy. This group of patients might
benefit from early surgery3436 Finally, smoking habits at
diagnosis were found to influence the risk of recurrence in
both UC and CD. This emphasises the need for patient
education in the outpatient clinics and on-going encouragement on smoking cessation. Our results underline that even
though the medical treatment of IBD has undergone significant
changes during the last decades, well-known disease characteristics are associated with the risk of recurrence and
hospitalisation. We believe that our study supports the
hypothesis that immunosuppressants and, possibly, biological
therapy might be effective in keeping patients from a medical
relapse as reflected by the decrease in overall recurrence
rate.
The strength of this study is the observational design. By
following an inception cohort, with the use of strict
diagnostic criteria, enables us to include only true IBD
patients covering the whole spectrum of disease severity.
Patient inclusion took place over a short period of 2 years
and all patients were followed throughout the same time
period. All Danish citizens are assigned a personal registration number that is registered at any contact with the health
care system or other public authorities. The Danish National
Patient Register (NPR-register)20 includes clinical information, such as diagnoses, surgical procedures, and administrative information, such as municipality, date of activity
and identification of hospital ward on a personal level.
Therefore, by combining Health Administrative Database
registrations with data from medical records, we obtained
complete follow-up data regarding the use of medication,
surgical interventions and admission rates.
We have made clear definitions of the description of a
recurrence. An all type and non-surgical recurrence in UC
also included topical treatment, according to our definition.
This could influence our results. The definition of a medical
relapse is made by using an increase in medication or
initiation of new medication as a surrogate marker for
relapse. We did not use clinical/biochemical findings, such
as mucosal damage, faecal calprotectin, elevated C-reactive

M.K. Vester-Andersen et al.

protein, low protein levels or thrombocytosis, as indicators
of disease activity. Integration of these data in the
definition of a medical relapse would possibly strengthen
the study.
Finally, our study was hampered by the population size.
There is a risk of type II errors as well as stratification
according to the treatment initiation and the duration was
not applicable due to the small sample size.
In conclusion, first time recurrence rates in CD but not in
UC have decreased in an era of intensified immunomodulating
therapy. This decrease was not paralleled by a decrease in the
hospitalisation rates or surgical rates. However, some phenotypic predictors showed that CD patients with complicating
characteristics at diagnosis are at risk of severe outcomes.
Close follow-up to uncover imminent symptoms of recurrence
and treatment optimisation is recommended.

Contributorship
Guarantor of the article: Flemming Bendtsen, Professor, MD,
DMSci
Specific author contributions: Marianne K. Vester-Andersen
contributed to the concept and study design, acquisition and
interpretation of data, and drafting and critical revision of the
manuscript.
Ida Vind contributed to the concept and study design, the
interpretation of the data, and critical revision of the
manuscript. Michelle V. Prosberg contributed to acquisition
of the data, and critical revision of the manuscript. Bo G.
Bengtsson contributed to acquisition of data, and critical
revision of the manuscript. Thomas Blixt contributed to
acquisition of data, and critical revision of the manuscript.
Pia Munkholm contributed to the concept and study design,
and critical revision of the manuscript. Mikael Andersson
contributed to statistical analysis, interpretation of the
data, and critical revision of the manuscript. Tine Jess
contributed to interpretation of the data, and critical
revision of the manuscript. Flemming Bendtsen contributed
to the concept and study design, interpretation of the data,
and critical revision of the manuscript.

Conflict of interest
None.

Acknowledgements
Grant support was received from MSD (No. 39197), Ferring
Pharmaceuticals, the Research Council of Hvidovre Hospital
and the Research Council of the Capital Region of Denmark. The
study sponsors did not contribute to the study design or to the
analysis and interpretation of the data or publication.

Appendix A. Supplementary data

Supplementary data to this article can be found online at
http://dx.doi.org/10.1016/j.crohns.2014.07.010.

Hospitalisation and recurrence rates in IBD - a 7-year population-based follow-up study

References
1. Yarur AJ, Strobel SG, Deshpande AR, Abreu MT. Predictors of
aggressive inflammatory bowel disease. Gastroenterol Hepatol
Oct 2011;7(10):6529.
2. Blonski W, Buchner AM, Lichtenstein GR. Clinical predictors of
aggressive/disabling disease: ulcerative colitis and Crohn disease.
Gastroenterol Clin North Am Jun 2012;41(2):44362.
3. Ramadas AV, Gunesh S, Thomas GAO, Williams GT, Hawthorne AB.
Natural history of Crohn's disease in a population-based cohort from
Cardiff (19862003): a study of changes in medical treatment and
surgical resection rates. Gut Sep 1 2010;59(9):12006.
4. Lakatos PL, Golovics PA, David G, Pandur T, Erdelyi Z, Horvath A,
et al. Has there been a change in the natural history of Crohn's
disease? Surgical rates and medical management in a populationbased inception cohort from Western Hungary between 19772009.
Am J Gastroenterol Apr 2012;107(4):57988.
5. Frolkis AD, Dykeman J, Negrn ME, Debruyn J, Jette N, Fiest KM,
et al. Risk of surgery for inflammatory bowel diseases has decreased
over time: a systematic review and meta-analysis of populationbased studies. Gastroenterology Nov 2013;145(5):9961006.
6. Nguyen GC, Nugent Z, Shaw S, Bernstein CN. Outcomes of
patients with Crohn's disease improved from 1988 to 2008 and
were associated with increased specialist care. Gastroenterology
Jul 2011;141(1):907.
7. Romberg-Camps MJL, Dagnelie PC, Kester AD, H de K MA, Cilissen
M, Engels LG, et al. Influence of phenotype at diagnosis and
of other potential prognostic factors on the course of
inflammatory bowel disease. Am J Gastroenterol Feb 2009;104:
37183.
8. Targownik LE, Singh H, Nugent Z, Bernstein CN. The epidemiology
of colectomy in ulcerative colitis: results from a population-based
cohort. Am J Gastroenterol Aug 2012;107(8):122835.
9. Kaplan GG, Seow CH, Ghosh S, Molodecky N, Rezaie A, Moran
GW, et al. Decreasing colectomy rates for ulcerative colitis:
a population-based time trend study. Am J Gastroenterol Dec
2012;107(12):187987.
10. Wolters FL, Russel MG, Sijbrandij J, Ambergen T, Odes S, Riis L,
et al. Phenotype at diagnosis predicts recurrence rates in
Crohn's disease. Gut Aug 2006;55:112430.
11. Henriksen M, Jahnsen J, Lygren I, Aadland E, Schulz T, Vatn MH,
et al. Clinical course in Crohn's disease: results of a five-year
population-based follow-up study (the IBSEN study). Scand J
Gastroenterol May 2007;42:60210.
12. Henriksen M, Jahnsen J, Lygren I, Sauar J, Kjellevold , Schulz
T, et al. Ulcerative colitis and clinical course: results of a 5-year
population-based follow-up study (the IBSEN study). Inflamm
Bowel Dis Jul 2006;12(7):54350.
13. Munkholm P, Langholz E, Davidsen M, Binder V. Disease activity
courses in a regional cohort of Crohn's disease patients. Scand J
Gastroenterol Jul 1995;30:699706.
14. Bernstein CN, Longobardi T, Finlayson G, Blanchard JF. Direct
medical cost of managing IBD patients: a Canadian populationbased study. Inflamm Bowel Dis Aug 2012;18(8):1498508.
15. Vind I, Riis L, Jess T, Knudsen E, Pedersen N, Elkjaer M, et al.
Increasing incidences of inflammatory bowel disease and
decreasing surgery rates in Copenhagen City and County,
20032005: a population-based study from the Danish Crohn
colitis database. Am J Gastroenterol Jun 2006;101:127482.
16. Munkholm P. Crohn's diseaseoccurrence, course and prognosis. An
epidemiologic cohort-study. Dan Med Bull Jun 1997;44:287302.
17. Langholz E, Munkholm P, Nielsen OH, Kreiner S, Binder V. Incidence
and prevalence of ulcerative colitis in Copenhagen county from
1962 to 1987. Scand J Gastroenterol Dec 1991;26:124756.
18. Gasche C, Scholmerich J, Brynskov J, D'Haens G, Hanauer SB,
Irvine EJ, et al. A simple classification of Crohn's disease: report of

19.

20.
21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

1683

the Working Party for the World Congresses of Gastroenterology,

Vienna 1998. Inflamm Bowel Dis Feb 2000;6(1):815.
Vester-Andersen MK, Prosberg MV, Jess T, Andersson M, Bengtsson
BG, Blixt T, et al. Disease course and surgery rates in inflammatory
bowel disease: a population-based, 7-year follow-up study in
the era of immunomodulating therapy. Am J Gastroenterol May
2014;109(5):70514.
Lynge E, Sandegaard JL, Rebolj M. The Danish National Patient
Register. Scand J Public Health Jul 2011; 39(7 Suppl.):303.
Hoie O, Wolters F, Riis L, Aamodt G, Solberg C, Bernklev T,
et al. Ulcerative colitis: patient characteristics may predict
10-yr disease recurrence in a European-wide population-based
cohort. Am J Gastroenterol Aug 2007;102:1692701.
Aalen OO, Johansen S. An empirical transition matrix for
non-homogeneous Markov chains based on censored observations.
Scand J Stat 1978;5(3):14150.
Hanauer SB, Feagan BG, Lichtenstein GR, Mayer LF, Schreiber S,
Colombel JF, et al. Maintenance infliximab for Crohn's disease:
the ACCENT I randomised trial. Lancet May 4 2002;359:15419.
Schnitzler F, Fidder H, Ferrante M, Noman M, Arijs I, Van AG,
et al. Long-term outcome of treatment with infliximab in 614
patients with Crohn's disease: results from a single-centre
cohort. Gut Apr 2009;58:492500.
Colombel JF, Sandborn WJ, Reinisch W, Mantzaris GJ, Kornbluth
A, Rachmilewitz D, et al. Infliximab, azathioprine, or combination therapy for Crohn's disease. N Engl J Med Apr 15
2010;362(15):138395.
Rutgeerts P, Sandborn WJ, Feagan BG, Reinisch W, Olson A, Johanns
J, et al. Infliximab for induction and maintenance therapy for
ulcerative colitis. N Engl J Med Dec 8 2005;353(23):246276.
Peyrin-Biroulet L, Oussalah A, Williet N, Pillot C, Bresler L,
Bigard M-A. Impact of azathioprine and tumour necrosis factor
antagonists on the need for surgery in newly diagnosed Crohn's
disease. Gut Jul 2011;60(7):9306.
Odes S, Vardi H, Friger M, Wolters F, Russel MG, Riis L, et al.
Cost analysis and cost determinants in a European inflammatory bowel disease inception cohort with 10 years of follow-up
evaluation. Gastroenterology Sep 2006;131(3):71928.
Longobardi T, Bernstein CN. Health care resource utilization in
inflammatory bowel disease. Clin Gastroenterol Hepatol Jun
2006;4(6):73143.
Binder V, Hendriksen C, Kreiner S. Prognosis in Crohn's disease
based on results from a regional patient group from the county
of Copenhagen. Gut Mar 1985;26(2):14650.
Solberg IC, Vatn MH, Hie O, Stray N, Sauar J, Jahnsen J, et al.
Clinical course in Crohn's disease: results of a Norwegian
population-based ten-year follow-up study. Clin Gastroenterol
Hepatol Dec 2007;5(12):14308.
Govani SM, Stidham RW, Higgins PDR. How early to take arms
against a sea of troubles? The case for aggressive early therapy
in Crohn's disease to prevent fibrotic intestinal strictures. J
Crohns Colitis Dec 1 2013;7(11):9237.
Cosnes J, Bourrier A, Laharie D, Nahon S, Bouhnik Y, Carbonnel F,
et al. Early administration of azathioprine versus conventional
management of Crohn's disease: a randomized controlled trial.
Gastroenterology Apr 2013;145(4):75865.
Aratari A, Papi C, Leandro G, Viscido A, Capurso L, Caprilli R.
Early versus late surgery for ileo-caecal Crohn's disease.
Aliment Pharmacol Ther Nov 15 2007;26(10):130312.
Rink AD, Fischer IR, Vestweber B, Vestweber K-H. Long-term
outcome of laparoscopic ileocecal resection for Crohn's disease
before the era of biologics. Int J Colorectal Dis Jan 2014;29(1):
12732.
Cullen G, O'toole A, Keegan D, Sheahan K, Hyland JM, O'donoghue
DP. Long-term clinical results of ileocecal resection for Crohn's
disease. Inflamm Bowel Dis Nov 2007;13(11):136973.