Background:
Existing approaches for the screening and treatment of asymptomatic bacteriuria in pregnancy are based on trials that were done more than 30 years ago. In this study, we reassessed the consequences of treated and untreated asymptomatic bacteriuria in pregnancy.
Methods:
In this multicentre prospective cohort study with an embedded randomised controlled trial, we screened women (aged ≥18 years) at eight hospitals and five ultrasound centres in the Netherlands with a singleton pregnancy between 16 and 22 weeks' gestation for asymptomatic bacteriuria. Screening was done with a single dipslide and two culture media. Dipslides were judged positive when the colony concentration was at least 1×10(5) colony-forming units (CFU) per mL of a single microorganism or when two different colony types were present but one had a concentration of at least 1×10(5) CFU per mL. Asymptomatic bacteriuria-positive women were eligible to participate in the randomised controlled trial comparing nitrofurantoin with placebo treatment. In this trial, participants were randomly assigned 1:1 to receive either nitrofurantoin 100 mg or identical placebo tablets, and were instructed to self-administer these tablets twice daily for 5 consecutive days. Randomisation was done by a web-based application with a computer-generated list with random block sizes of two, four, or six participants rendered by an independent data manager. 1 week after the end of treatment, they provided us with a follow-up dipslide. Women, treating physicians, and researchers all remained unaware of the bacteriuria status and treatment allocation. Women who refused to participate in the randomised controlled trial did not receive any antibiotics, but their outcomes were collected for analysis in the cohort study. We compared untreated and placebo-treated asymptomatic bacteriuria-positive women with asymptomatic bacteriuria-negative women and nitrofurantoin-treated asymptomatic bacteriuria-positive women. The primary endpoint was a composite of pyelonephritis with or without preterm birth at less than 34 weeks, analysed by intention to treat at 6 weeks post-partum. This trial is registered with the Dutch Trial Registry, number NTR3068.
Findings:
Between Oct 11, 2011, and June 10, 2013, we enrolled 5621 women into our screening cohort, of whom 5132 were eligible for screening. After exclusions for contaminated dipslides and patients lost to follow-up, in our final cohort of 4283 women, 248 were asymptomatic bacteriuria positive, of whom 40 were randomly assigned to nitrofurantoin and 45 to placebo for the randomised controlled trial, whereas the other 163 asymptomatic bacteriuria-positive women were followed without treatment. The proportion of women with pyelonephritis, preterm birth, or both did not differ between untreated or placebo-treated asymptomatic bacteriuria-positive women and asymptomatic bacteriuria-negative women (6 [2·9%] of 208 vs 77 [1·9%] of 4035; adjusted odds ratio [OR] 1·5, 95% CI 0·6-3·5) nor between asymptomatic bacteriuria-positive women treated with nitrofurantoin versus those who were untreated or received placebo (1 [2·5%] of 40 vs 6 [2·9%] of 208; risk difference -0·4, 95% CI -3·6 to 9·4). Untreated or placebo-treated asymptomatic bacteriuria-positive women developed pyelonephritis in five [2·4%] of 208 cases, compared with 24 [0·6%] of 4035 asymptomatic bacteriuria-negative women (adjusted OR 3·9, 95% CI 1·4-11·4).
Interpretation:
In women with an uncomplicated singleton pregnancy, asymptomatic bacteriuria is not associated with preterm birth. Asymptomatic bacteriuria showed a significant association with pyelonephritis, but the absolute risk of pyelonephritis in untreated asymptomatic bacteriuria is low. These findings question a routine screen-treat-policy for asymptomatic bacteriuria in pregnancy.
Funding:
ZonMw (the Netherlands Organisation for Health Research and Development).