Effect of antibiotic consumption on self-reported recovery from uncomplicated urinary tract infection

Background Randomised trials provide high-quality data on the effects of antibiotics for urinary tract infections (UTI) but are prone to selection bias and may not accurately reflect effects in those who actually consume antibiotics. Aim To estimate the effect of antibiotic consumption, rather than prescription, on self-reported time to recovery in women with uncomplicated UTI. Design and setting Secondary analysis of data from a randomised control trial on point of care diagnostic to guide antibiotic prescribing in primary care. Antibiotic prescribing decisions were not mandated by the trial protocol. Methods 413 adult women presenting with UTI symptoms and who either consumed (n=333) or did not consume (n=80) one or more doses of antibiotics during a 14-day follow-up were included. Antibiotic consumption was standardised using defined daily dose (DDD) and described as days-worth of antibiotics. Cox-proportional hazard model was used to estimate the effect of antibiotic consumption upon time to self-reported recovery, controlling for potential confounders and propensity of consuming antibiotics based on baseline symptom severity and history of UTI treatment. Results The crude median time to recovery was 3 days longer among patients who did not consume antibiotics (10 days; 95% CI= 8.5-11.5) compared to those who consumed antibiotics (7 days; 95% CI= 6.5-7.5). Patients who consumed >0 to <3, [≥]3 to [≤]5, and >5 days-worth of antibiotics recovered at approximately 2 to 3 times faster than the rate of those who consumed no antibiotics, adjusted HR 3.2 (95% CI 2.0, 5.1), 2.8 (1.8, 4.3), 2.3 (1.5, 3.7), respectively. Conclusion Consuming antibiotics was associated with a reduction in self-reported time to recovery in this study which is consistent with results from previous trials.