Group B streptococci cultured in urine during pregnancy associated with preterm delivery: a selection problem?

To investigate an association between Group B streptococci (GBS) in urine culture during pregnancy and preterm delivery. A population-based cohort consisted of all the pregnant women ( = 36,097) from the catchment area of Lillebaelt Hospital, Denmark, during the period January 2002 -December 2012. The cohort of 34,285 singleton pregnancies used in this study was divided into three groups. Group I ( = 249) included women whose urine culture was positive for GBS; group II ( = 5765) included women whose urine culture was negative for GBS; and group III ( = 28 271) included women whose urine had not been cultured during pregnancy. Primary outcome was preterm delivery before 37 weeks' gestation (PTD). We did not find an association between PTD and GBS bacteriuria in the cultured groups (odds ratios (OR) = 0.89; 95% CI: 0.5-1.4) ( Table 1 ). After controlling for potential confounders, the PTD remained not associated with GBS bacteriuria (adjusted OR = 0.99; 95% CI: 0.6-1.6). Combined, the cultured groups (I and II) were associated with a statistically significant higher risk for PTD, when compared with the group with no urine specimens taken for culture (OR = 1.96; 95% CI: 1.8-2.2 and adjusted or 1.80; 95% CI 1.6-2.0). The cultured group of women differed considerably from the group of women with no urine specimens taken for culture on the vast majority of variables examined. No association between asymptomatic GBS bacteriuria and preterm delivery among women with singleton pregnancy and urine specimens cultured during pregnancy was found. Previous suggestions of such association may have been compromised by a selection problem for testing due to a high-risk profile of pregnancy complications in pregnant women selected for urine culture.