Epidemiology and microbiological features associated with greater risk of recurrent urinary tract infection

Problem

The epidemiology and microbiology of recurrent urinary tract infection (rUTI) is not fully understood. Presently, rUTI is defined as 2 UTIs within 6 months or 3 within a year period. The proportion of women with ≥1 UTI who go on to experience rUTI is unknown, and there is no underpinning evidence for the definition of rUTI in use. We aimed to estimate the percentage of women with ≥1 UTI who had rUTI at least once, describe their characteristics, and identify the features that increase the risk of reinfection within 6 months from a rUTI episode.

Approach

The Infections in Oxfordshire Research Database was used to extract all urine culture results from women 16 and older reported between April 2008 and March 2019 on urines sent from healthcare settings in Oxfordshire. We included urine cultures from community settings or within 48 hours of admission to hospital, and excluded other cultures within 28 days of hospital admission, those explicitly for antenatal screening, and samples sent within 28 days from an index sample (as likely related to the index infection). We restricted the analysis of rUTI to positive cultures. To consider multiple UTIs within individuals, we used frailty models to assess the likelihood of reinfection with same species, adjusted for age.

Findings

We retrieved 697626 urine culture results performed on samples from 201927 women, 27% were positive cultures and 28%, mixed growth. In the 84809 women with ≥1 UTI, the risk of having two UTIs within 6 months was 11 times higher than of having three UTIs in that period. In the study period, 18.5% women had ≥1 rUTI episode. The age distribution of index UTI and first rUTI was markedly different, favouring women over 60. Of all positive urine cultures, 65% were of Escherichia coli (E.coli). A similar proportion was observed in rUTI episodes. Overall, the risk of reinfection with E.coli was 1.54(95%CI=1.49-1.60) times higher in those previously infected with this bacterial species than with other species. After an index rUTI, women more likely to have ≥2 UTIs within 6 months than no UTIs in the same period were older (median age=71, IQR=55-81 versus 65, IQR=43-78 years).

Consequences

This is the first study using a UK dataset to describe the population of women with rUTI, and to highlight the age distribution. This is important because even using the most restrictive definition of UTI, based on positive urine cultures, shows that this problem is experienced by almost a fifth of women who have a single UTI, and overwhelmingly affecting older women. We found that the current rUTI definition is met within 6 months of an index UTI and not 12. Exploring alternative definitions would allow us to determine an optimal definition that permits early intervention.