Defining recurrent urinary tract infections and quantifying bladder cancer risk in primary care in England: a...

Around 30-40% of people with bladder cancer in England present with symptoms or diagnosis of urinary tract infection (UTI) in the year before a diagnosis of bladder cancer. National Institute for Health and Care Excellence guidelines advise referral of people aged 60 or older with unexplained recurrent UTIs to specialists, but recurrence is ambiguously defined, risking delayed diagnosis and worsened outcomes. We aimed to better define and quantify the effect of recurrent UTIs on the likelihood of bladder cancer.

In this nationwide retrospective case-control study, we collected data from people aged 18 years or older with bladder cancer diagnosed between Jan 5, 1998, and Dec 31, 2018, and who had at least one UTI before diagnosis from the National Cancer Registration and Analysis Service (NCRAS) and the Clinical Practice Research Datalink (CPRD) in England. NCRAS diagnosis and dates were preferentially used. We matched each patient with bladder cancer (ie, case) to five patients without bladder cancer (ie, controls) using age at cancer diagnosis, sex, and general practice. Primary analyses examined the associations between UTI frequency (one, two, three, four, and five or more episodes) and bladder cancer using conditional logistic regression, within predefined look-back intervals up to 5 years before the index UTI (eg, the most recent UTI before their cancer diagnosis).

We identified 92 277 patients registered with English primary care practices contributing to CPRD and 13 227 contributing to NCRAS, including 17 584 patients with bladder cancer (cases) and 87 920 without bladder cancer (controls). After excluding 7513 duplicates and 42 526 controls who had no UTIs before a diagnosis, then further excluding 1529 patients (427 cases and 1102 controls) who had a UTI 5-10 years before the index UTI, we included 53 936 patients for analysis (17 157 cases and 36 779 controls). There were more male (9992 [58·2%]) than female (7165 [41·8%]) patients with bladder cancer; with a median age of 75·0 years (IQR 67·0-81·0); and patients were White (16 028 [93·4%]), Asian (189 [1·1%]), Black (139 [0·8%]), or mixed race (31 [0·2%]) and of other (91 [0·5%]) and unknown ethnic groups (679 [4·0%]). There was a higher proportion of female than male patients without bladder cancer (19 858 [54·0%] vs 16 921 [46·0%]). A dose-response relationship between recurrent UTI and bladder cancer risk was seen in the first 6 months only, with the largest effect for five or more UTIs at 0-6 months (adjusted odds ratio 13·05 [95% CI 11·60-14·68]); patients with five or more UTIs had around 2·5-times higher odds than those with three UTIs in the same interval (4·95 [4·51-5·43]). Associations attenuated with increasing time from the index UTI.

Our data suggest that recurrent UTIs within 6 months were a strong signal of bladder cancer risk. Our findings can help to refine existing guideline recommendations for patients with recurrent UTIs who might benefit from further investigations to rule out possible bladder cancer.

Barts Charity and the National Institute for Health Research Policy Research Unit in Cancer Awareness, Screening, and Early Diagnosis.