Clin Rheumatol (2013) 32:355360

DOI 10.1007/s10067-012-2129-7

ORIGINAL ARTICLE

Urinary tract infections in patients with rheumatoid arthritis

D. Puntis & S. Malik & V. Saravanan & M. Rynne &
C. Heycock & J. Hamilton & C. A. Kelly

Received: 27 April 2012 / Revised: 11 October 2012 / Accepted: 19 November 2012 / Published online: 14 December 2012
# Clinical Rheumatology 2012

Abstract Co-morbidity from rheumatoid arthritis (RA) has

recently focussed on outcomes of cardiovascular and pulmonary disease, but serious infections are an increasingly
well-recognised complication of RA. Recent work has demonstrated how the incidence of pneumonia can be reduced in
RA, but little attention has been paid to the incidence of
urinary tract infection (UTI) in RA or to the associated comorbidity. The aim of this study was to describe the incidence of UTI leading to hospitalisation in a large cohort of
patients with RA and investigate which factors contributed
to this. This study assessed all patients with RA hospitalised
over a 12-month period with a discharge diagnosis including
UTI. Patients were identified through a PAS records search
in a single large centre. Historical case controls without RA
matched for age and gender were identified from the literature. Clinical notes were manually examined by two observers. We recorded: age, gender, duration of RA, number of
UTI, all RA therapy, co-morbidity, results of urine and
blood cultures with antimicrobial sensitivities, readmission
rates, treatment and outcome. We calculated the relative risk
(RR) of developing UTI in patients with RA and the factors
influencing this. From a population of 2,200 RA patients,
the overall annual incidence of hospitalisation with UTI
amongst RA patients was 2.09 %, as against 0.97 and
0.91 % for two control groups (RR02.16 and 2.29). Most
patients (90 %) were female, and the group mean age was
76 years. The use of long-term oral steroids as sole therapy
D. Puntis : S. Malik : V. Saravanan : M. Rynne : C. Heycock :
J. Hamilton : C. A. Kelly (*)
Department of Rheumatology, Queen Elizabeth Hospital,
Sheriff Hill,
Gateshead, UK NE96SX
e-mail: clive.kelly@ghnt.nhs.uk
D. Puntis : S. Malik : V. Saravanan : M. Rynne : C. Heycock :
J. Hamilton : C. A. Kelly
Department of Medicine, Queen Elizabeth Hospital, Sheriff Hill,
Gateshead, UK NE96SX

was associated with a RR of 6.8 for UTI (p00.002) while

failure to take disease-modifying anti-rheumatic drugs
(DMARDs) was associated with a similar RR of 6.7
(p00.001). Positive cultures for Escherichia coli were found
in 51 % of RA patients. Relevant co-morbidities included
permanent catheters, vaginal prolapse, cancer and diabetes.
Recurrence of UTI within a year was common. RA was
associated with a higher-than-expected incidence of UTI,
particularly among older females. This was associated with
the use of long-term oral steroids and the absence of
DMARDs. Other factors included female gender, greater
age and long disease duration. We recommend avoidance
of long-term oral steroids but consideration of low-dose
prophylactic antibiotics in those patients with recurrent UTI.
Keywords Co-morbidity . Infection . Rheumatoid arthritis

Introduction
Rheumatoid arthritis (RA) shortens life expectancy [1] for a
variety of reasons. Excess deaths occur in part due to infection [2, 3], much of which has been reported as respiratory
in origin [4, 5]. Morbidity from sepsis is also common, and a
prevalence as high as 45 % over 10 years has been recorded
[6, 7]. An increase in hospitalisation for serious infections
among patients with RA has recently been confirmed in a
longitudinal cohort [8] where risk factors included the presence of extra-articular manifestations and prior use of oral
steroids [9].
Recent published work has demonstrated the ability to
significantly reduce the mortality associated with pneumonia
in RA by using a combination of evidence-based measures
[10]. However, there is an impression that RA patients now
appear to be hospitalised as a result of urinary tract infections
(UTI) as commonly as from pneumonia. This complication of
RA has attracted very little attention recently, although an

356

association has been previously reported with secondary Sjgren's syndrome [11]. However, the frequency, associations and
outcome of UTI have not been investigated in unselected
patients with RA previously.

Aims and objectives

We aimed to record the details of all patients with a dual
diagnosis of RA and UTI and determine the range of
responsible pathogens, the use of imaging, and to review
therapeutic intervention. In addition, we wished to investigate
which factors might contribute to the risk of UTI.

Methods
Our Foundation Trust serves a population of 220,000 people
in and around Gateshead and receives referrals from another
80,000 outside of our traditional catchment area. This population generates 2,500 new referrals and 9,000 follow-up
attendances to the department annually, with 2,200 RA
patients included among the latter. The methodology closely
followed that of our previous studies [4, 10]. All patients
admitted to the Queen Elizabeth Hospital in Gateshead
during the 12 months from November 20092010 with
UTI and a prior diagnosis of RA [12] were identified from
the hospital database. The clinical records of these patients
were then examined manually and demographic and clinical
data extracted. We excluded those patients whose notes
failed to confirm the diagnosis of either RA or UTI. We
specifically recorded age, gender, duration of RA, number
of previous UTI, all comorbid conditions, treatment, and
outcome of the index admission. We also noted the results
of midstream urine and blood cultures with antibiotic sensitivities, subsequent prescribing of antibiotics and the results
of any imaging of the urinary tract undertaken.
We defined UTI as a symptomatic positive urine culture
or an episode where clinical symptoms and signs had been
clearly documented with abnormal urinalysis (haematuria
and proteinuria), but culture results were unavailable. Where
a patient had more than one episode of UTI as an inpatient,
the earliest episode was used as the index case. Any other
episodes in the study year, including those diagnosed in the
community and identified by laboratory results, were also
noted. This allowed us to calculate recurrence and readmission rates as a result of UTI in our population.
Completeness of ascertainment was ensured by comparing database records with those from the Department of
Rheumatology and Department of General Medicine. A
cross-check was made at subsequent clinic review in case
patients on our RA database had been admitted to a neighbouring hospital with UTI over the study period. The cause

Clin Rheumatol (2013) 32:355360

of death in fatal cases was ascertained from the medical

notes and cross-checked with death certificate data. In cases
of failure to attend clinic, these data were obtained from
the patient's general practitioner, as was the cause of all outof-hospital deaths in our RA population during the 3 years
from 2009 to date.
A large-scale American study on UTI prevalence in the
general population was used as one source of control data as
the population demographics were a close match to those of
our RA group [13]. We also calculated the incidence of UTI
from within our own Gateshead population by examining
hospital records for all admissions with UTI for the same
12-month period (November 2009November 2010) for
people aged over 64 years using the same methods as for
the RA patients. These data allowed us to calculate the
relative risk of developing UTI in patients with RA [14].
Our computerised database (RheMos) identified the total
number of RA patients in Gateshead and the relative proportions in each of the treatment subgroups: (1) index group
who were taking DMARDs and/or biologic therapy, (2) all
patients on DMARDs and/or biologic therapy, (3) patients
on no treatment for RA and (4) patients on oral steroids
alone. We examined the treatment profile of our total RA
population and compared that to the profile of those RA
patients admitted with UTI. This allowed us to identify
associations between treatment and the frequency of UTI
by calculating the relative risks of each treatment group
compared to the index group, having adjusted the data for
confounders including age, gender and disease duration. We
used Fisher's exact test to calculate the probabilities of an
effect of treatment on the chances of being admitted with
UTI in our RA population. The MannWhitney U test was
used to compare demographics between the RA populations
with and without UTI. This retrospective observational
study was registered with Gateshead Health Foundation
Trust, and prior ethical approval was obtained.

Results
Risk of admission
From 2,200 RA patients in the rheumatology database at the
time of this study, there were 54 confirmed admissions with
UTI involving 46 different RA patients during the study
period. This was after exclusion of patients whose notes
contained inadequate documentation to confirm the diagnosis of RA (two patients) or UTI (three patients). These
figures represented an annual incidence of UTI of 2.09 %
in our RA population. By comparison, frequency of inpatient admission with UTI for 7484-year-old women among
the general American population was 968/100,000 over a 1year period (13), equating to an annual incidence of 0.97 %.

Clin Rheumatol (2013) 32:355360

357
Comparing Treatment Between the RA Study
Group and the Total QEH RA Population
Study Group
Total QE Hospital RA population

100
90
80
70

Percent

This demonstrated a relative risk of admission with UTI

among our RA patients of 2.16. Within our local population,
we had 498 documented admissions with UTI in the population aged 64 years and over during the same 12-month
period, of whom 89 % were female. The total number of
people in this age group within our catchment zone was
54,890. This equates to an annual incidence of 0.91 % and
equates to a relative risk of admission among our agematched RA population of 2.29.
The median age of the RA patients with UTI was 74 years
(range, 6493 years), and all but two were female. In addition, the average number of total UTI per patient in the RA
group was 3.1, as compared to an average of 1.7 in the
control group, for the year. This indicated a higher recurrence rate overall, although the readmission rate was not
significantly increased at 18 % among RA patients.

60
50
40
30
20
10
0
DMARD or
Biologic

Steroids +
DMARD

No DMARD

Steroids + No
DMARD

Fig. 2 Comparison of treatment between the patients with RA admitted with UTI and the overall RA population

RA therapy
Pathogens
The percentages of patients in the various treatment subgroups for our total RA population are compared to those in
the UTI study group in Fig. 1. There is a significant difference between the two populations for all treatment subgroups, with those patients developing UTI being less
likely to be taking DMARDs and biologics and more likely
to be on oral steroids or no treatment. A total of 18 (40 %)
patients in the study group were on oral steroids, either with
or without DMARDs, whereas in our overall RA population, only 10 % of patients were taking oral prednisone
[p00.002]. Also, 18 (40 %) patients with UTI were not
receiving DMARDs, by comparison with 6 % of the RA
population overall [p00.001]. The relative risks of requiring
an inpatient stay for UTI associated with the different treatment subgroups are shown in Fig. 2. These data remain
significantly different between the groups after correction
for gender, age and disease duration.
UTI Pathogens
2% 2%

E.coli
Klebsiella
Proteus

52%

Enterobacter cloacae
No Growth
Strep agalactiae

11%

Imaging
Among 15 patients (33 %) referred for imaging of the renal
collecting system, abnormalities were found in five. This
comprised renal cysts (three), bilateral renal atrophy (one)
and unilateral hydronephrosis (one).
Co-morbidity and outcome

22%

2%

Urine cultures were sent in all patients, and the pathogens cultured are shown in Fig. 3. All patients were
treated with antibiotics with co-amoxiclav used in 40 %
and trimethoprim in 25 %. Other antibiotic regimes
included tazobactam/piperacillin and meropenem for
patients who had more severe infection or who were
allergic to penicillin. Of concern was the observation
that two patients were treated with trimethoprim in spite
of taking methotrexate. Furthermore, only 46 % of
patients had their DMARDs suspended while hospitalised for UTI.

Morganella Marganii

9%

Fig. 1 Frequency with which different pathogens were identified

during the 54 admissions of RA patients with UTI

Twenty patients had significant co-morbidity which may

have contributed to the risk of developing UTI. These
included underlying malignancy (nine), indwelling catheter (seven) and type 2 diabetes (four). Two patients
died during their admission, one as a result of the
infection and another from infarction of the intestine.
The median age of the patients admitted with UTI was
13 years greater than that of the RA population as a
whole, with a median RA duration of 14 years which
was 3 years longer than the median duration of RA in
the overall patient population. These data are summarised in
Table 1.

358

Clin Rheumatol (2013) 32:355360

Fig. 3 Relative risks of

different therapeutic strategies
in RA for admission with UTI

Relative Risk of an Inpatient episode of UTI Associated With Different

Therapeutic Options
8
7

6.7

6.8

No DMARD

Steroids + No DMARD

6
5
4

3.3

3
2
1
1
0
DMARD or Biologic

Discussion
This study has demonstrated a higher incidence of hospitalisation for UTI among our population of RA patients than
would be expected for the general population. This finding
supports previous publications showing an increased frequency
of infection amongst RA patients [3, 6, 8, 9, 15, 16]. Most
patients were elderly females with significant co-morbidity and
a high recurrence rate.
Our data suggest that RA itself, rather than DMARD
therapy, is primarily responsible for the observed increase
in infection. This supports the findings of a multicenter
study that showed no increase in hospitalisation for pneumonia in patients with RA as a consequence of taking
DMARDs [15]. However, this large study did demonstrate
an increased risk of hospitalisation for pneumonia in
patients taking oral steroids and found that this risk increased with increasing steroid dose. The present study
suggests that the same is true for UTI, as the risk of infection
in those taking long-term oral steroids as sole therapy was
increased almost sevenfold, and this was the most readily
modifiable risk factor we identified for UTI. Interestingly,
there is some evidence that Proteus species may be implicated in the development of RA [17], although our study
demonstrated this organism was responsible for infection in
only 11 % of patients. E. coli was the dominant pathogen
accounting for over half of all infections.

Table 1 To compare clinical variables between patients with RA

admitted to hospital with UTI and the overall population of RA patients
using MannWhitney U test

Age (years) median [range]

Gender (% female)
Not on DMARD (%)
On oral steroid (%)
Disease duration (years)

RA with
UTI

Overall RA
population

p value

74 [6493]
97 %
40 %
40 %
14 [236]

61 [1894]
74 %
6%
10 %
11 [138]

0.010
0.008
0.001
0.002
NS

Steroids + DMARD

Previous work has shown that specific measures to reduce the

morbidity and mortality from lower respiratory tract infections
(LRTI) in RA patients are identifiable and effective [10]. Risk
factors for the development of LRTI in RA included use of longterm oral steroids, smoking and prior lung disease. Data also
suggested that vaccination against influenza and pneumococcus
may contribute to the observed reduction in LRTI [10, 18, 19],
although it appears that there is room for improvement in vaccination levels within the RA population at present [20, 21].
Concern that methotrexate might reduce the effectiveness of
vaccination in patients with RA has been expressed [22] and
explored [18], with reassuring findings [18].
We also found that failure to use DMARDs increased
the risk of UTI. The literature has been previously
confused on this topic [23, 24], but there is now a
general consensus that drugs such as methotrexate do
not in themselves contribute to the risk of infection in
RA, but that suspension of such agents during infection
is a wise precaution to allow mobilisation of natural
immune defence mechanisms [18, 19]. Despite all our
patients receiving both written and verbal communication to suspend DMARDs during concurrent antibiotic
therapy, this study has identified further educational
needs as we did not manage to achieve this in more
than half of our patients.
Although vaccination is not an option for patients with recurrent UTI, we suggest consideration be given to the use of longterm low-dose prophylactic antibiotics in patients with recurrent
UTI as patients with RA seem to have a high recurrence rate and
significant co-morbidity. However, it is important to recognise
that trimethoprim must be avoided in patients on methotrexate
because of the risk of neutropaenia, and it was worrying that this
combination was observed in two patients in our study in spite of
the fact that the interaction has been highlighted in the British
National Formulary for 15 years [25]! In each case, the antibiotic
was initiated by an emergency room physician, and we have
highlighted this concern to the physicians involved. In patients
not taking methotrexate, trimethoprim 100 mg once daily is
appropriate prophylaxis. In those on methotrexate, cephalexin
250 mg once daily is an alternative, providing there is no history

Clin Rheumatol (2013) 32:355360

of Clostridium difficile infection previously, in which case nitrofurantoin may be considered.

We did not routinely collect data on disease activity in
our patients with RA and UTI. However, it would be interesting to see if those patients with greater disease activity
scores (DAS28) are at greater risk of infection, as a recent
large study has suggested [26]. Their findings related to an
association between increased DAS28 and raised infection
rates, with more LRTI than UTI recorded. Our study group
was significantly older than the RA population as a whole
and had a longer mean disease duration. Older age is also
statistically associated with a higher risk of UTI [27]. However, we found that the association with RA and UTI
remained significant after correction for these variables.
Other factors contributing to the association of UTI with
RA may apply. Personal hygiene in the perineum can be an
issue for older patients with poor hand function, and this
may be an important factor, especially in older females. We
did not address this issue specifically in this study. Poor
fluid intake may be a further factor. This is again common in
older females and may contribute towards UTI, especially in
the warmer months of the year. We did find that hyperglycaemia and the presence of an indwelling catheter were
important contributory factors in a number of our patients.
We did not specifically assess the possibility of IgA deficiency which is recognised in association with RA.

Conclusions
This study confirms that there is an increased rate of admission with UTI among patients with RA and that this is mainly
seen in older females. The evidence suggests that this increased propensity to infections is largely a complication of
RA itself. However, the use of oral steroids increases the risk
of infection, as does the absence of DMARDs. We recommend that the use of long-term oral steroids be minimised
especially in elderly females, that DMARDs be suspended
during intercurrent infection and that the use of appropriate
low-dose prophylactic antibiotics be considered in those with
recurrent UTI. In order to reassess the impact of our recommendations, we intend to reassess the frequency and management of UTI in our RA population in a further 5 years.
Disclosures None.

References
1. Young A, Koduri G, Batley M, Kulinskaya E, Gough A, Norton S
et al (2007) Mortality in rheumatoid arthritis is increased early in
the course of disease, as a result of ischaemic heart disease and
pulmonary fibrosis. Rheumatology 46:350357

359
2. Mitchell DM, Spitz PW, Young DY, Bloch DA, McShane DJ, Fries
JF (1986) Survival, prognosis, and causes of death in rheumatoid
arthritis. Arthritis Rheum 29:706713
3. Kelly CA, Hamilton J (2007) What kills patients with rheumatoid
arthritis. Rheumatology 46:183184
4. Coyne P, Hamilton J, Heycock C, Saravanan V, Coulson E, Kelly
CA (2007) Acute lower respiratory tract infections in patients with
rheumatoid arthritis. J Rheumatol Sept 34(9):18321837
5. Riise T, Jacobsen BK, Gran JT, Haga HJ, Arnesen E (2001) Total
mortality is increased in rheumatoid arthritis. A 17-year prospective study. Clin Rheumatol 20:123127
6. Franklin J, Lunt M, Bunn D, Symmons D, Silman A (2007) Risk
and predictors of infection leading to hospitalisation in a large
primary-care-derived cohort of patients with inflammatory polyarthritis. Ann Rheum Dis 66:308312
7. Capell HA, Murphy EA, Hunter JA (1991) Rheumatoid arthritis:
workload and outcome over 10 years. Q J Med 79:461476
8. Doran M, Crowson CS, Pond GR, OFallon M, Gabriel SE (2002)
Frequency of infection in patients with rheumatoid arthritis compared with controls: a population-based study. Arthritis Rheum
46:22872293
9. Doran M, Crowson CS, Pond GR, OFallon M, Gabriel SE (2002)
Predictors of infection in rheumatoid arthritis. Arthritis Rheum
46:22942300
10. Housden MM, Bell G, Heycock C et al (2010) How to reduce
mortality from chest infections in rheumatoid arthritis. Clin Med
29:326329
11. Tishler M, Caspi D, Almog Y, Seqal R, Yaron M et al (1992)
Increased incidence of urinary tract infection in patients with
rheumatoid arthritis and secondary Sjgren's syndrome. Ann
Rheum Dis 51:604606
12. Aletaha D, Neogi T, Silman A et al (2012) 2010 rheumatoid
arthritis classification criteria: an American College of Rheumatology/European League Against Rheumatism collaborative initiative. Ann Rheum Dis 69:15801588
13. Griebling TL (2007) Urinary tract infections in women. In: Litwin
MS, Saigal CS (eds) Urologic diseases in America. US Department
of Health and Human Services, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases. Washington (D. C.): US Government Printing Office 2007: pp. 587620.
Available from http://kidney.niddk.nih.gov/statistics/uda/Urinary_
Tract_Infection_in_Women-Chapter18.pdf. Accessed 23 Mar 2007
14. Sheskin DJ (2004) Handbook of parametric and nonparametric
statistical procedures, 3rd edn. Chapman & Hall CRC, Boca Raton
15. Wolfe W, Caplan L, Michaud K (2006) Treatment for rheumatoid
arthritis and the risk of hospitalisation for pneumonia: associations with prednisone, disease-modifying antirheumatic
drugs and anti-tumour necrosis factor therapy. Arthritis
Rheum 54:628634
16. Vandenbroucke JP et al (1984) Survival and cause of death in
rheumatoid arthritis: 25 year prospective follow up. J Rheumatol
11:158161
17. Ebringer A, Rashid T (2006) Rheumatoid arthritis is an autoimmune disease triggered by Proteus urinary tract infection. Clin Dev
Immunol 13:4148
18. Coulson E, Saravanan V, Heycock C, So KL, Morgan L, Hamilton
J et al (2011) Pneumococcal antibody levels after Pneumovax in
patients with rheumatoid arthritis on methotrexate. Ann Rheum
Dis 48:12891291
19. McLean-Tooke A, Aldridge C, Waugh S, Spickett G, Kay L (2009)
Methotrexate, rheumatoid arthritis and infection riskwhat is the
evidence? Rheumatology 48:867871
20. Doe S, Pathare S, Kelly CA, Heycock CR, Binding J, Hamilton J
(2007) Uptake of influenza vaccination in immunosupressant
agents for rheumatological diseases: a follow-up audit of the
influence of secondary care. Rheumatology 46:716717

360
21. Lanternier F, Henegar H, Mouthon L, Blanche P, Guillevin L,
Launay O (2008) Low influenza vaccination rate among adults
receiving immunosuppressive therapy for systemic inflammatory
disease. Ann Rheum Dis 67:1007
22. Kapetanovic MC, Saxne T, Sjoholm A, Truedsson L, Jonsson G,
Geborek P (2006) Influence of methotrexate, TNF blockers and
prednisolone on antibody responses to pneumococcal polysaccharide vaccine in patients with rheumatoid arthritis. Rheumatol Oxf
45:106111
23. Van der Veen MJ, van der Heide A, Kruize AA, Biljsma JW
(1994) Infection rate and use of antibiotics in patients with
rheumatoid arthritis treated with methotrexate. Ann Rheum
Dis 53:224228

Clin Rheumatol (2013) 32:355360

24. Luzi G, Lagan B, Salemi S, Di Rosa R et al (2009) Are glucocorticoids a consistent risk factor for infections in rheumatoid
arthritis patients under treatment with methotrexate and etanercept?
Clin Ter 160:121123
25. Joint Formulary Committee (Great Britain), British Medical Association, Royal Pharmaceutical Society of Great Britain (1997)
British National Formulary. Issue 34; 564
26. Au K, Reed G, Curtis J et al (2011) High disease activity is
associated with an increased risk of infection in patients with
rheumatoid arthritis. Ann Rheum Dis 70:785791
27. Eriksson I, Gustafson Y, Fagerstrm L, Olofsson B et al (2010)
Prevalence and factors associated with urinary tract infections in
very old women. Arch Gerontol Geriatr 50:132135

Copyright of Clinical Rheumatology is the property of Springer Science & Business Media B.V. and its content
may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express
written permission. However, users may print, download, or email articles for individual use.