OBJECTIVE:To identify predictors of symptomatic urinary tract infection (UTI) after 20 weeks' gestation.
STUDY DESIGN:A retrospective cohort analysis was conducted of all deliveries at three North Carolina hospitals between 1990 and 1993. A total of 7403 deliveries remained after exclusions (pre-pregnancy diabetes, HIV-positive, structural urologic abnormalities, no prenatal care) and restrictions (black or white race, county of residence). Cystitis and pyelonephritis were identified by clinician diagnosis. Multiple logistic regression was conducted.
RESULTS:Prior UTIs (both before and earlier in pregnancy), nonprivate clinics, and a history of chlamydia (white women only) doubled the risk of symptomatic UTIs after 20 weeks' gestation. The strongest predictor of pyelonephritis was prior antenatal UTIs (adjusted incidence odds ratio ϭ 5.3, 95% confidence interval of 2.6 -11.0), followed by less education (Ͻ12 years), a history of chlamydia, nonprivate clinics, illicit drug use, sickle cell hemoglobinopathy, and being unmarried.
CONCLUSION:Medical history and demographic factors predict cystitis and pyelonephritis after 20 weeks' gestation. Prospective studies of pyelonephritis predictors and screening strategies are warranted.Urinary tract infections (UTIs) are common in pregnancy and have been associated with preterm delivery and low birth weight.1-3 Lower UTIs in the mother are more likely to progress to pyelonephritis during pregnancy, and pyelonephritis can result in sepsis and acute adult respiratory distress syndrome. 4,5 In 1960, Kass 6 demonstrated that detection and treatment of asymptomatic bacteriuria (ASB) in pregnancy could diminish the likelihood of progression to symptomatic UTI. This was followed by various cohort studies describing the epidemiology of UTIs during pregnancy.3,6 -16 These studies were limited by the breadth of risk factors explored and their statistical methodology. Thus, we conducted this study to identify demographic, behavioral, and medical history factors during and before pregnancy that are predictive of a symptomatic UTI in pregnancy after 20 weeks' gestation. Of particular interest was the identification of risk factors for pyelonephritis and separation of the correlated effects of black race and sickle cell hemoglobinopathies.
MATERIALS AND METHODSA retrospective cohort was assembled of deliveries at three central North Carolina hospitals from 1990 to 1993: University of North Carolina Hospitals (UNCH), Duke University Medical Center (DUMC), and Durham Regional Hospital (DRH). Study details are published elsewhere. 17 Briefly, computerized obstetrics databases were analyzed and supplemented with medical chart abstracts. The original sources for all variables were routine patient questionnaires administered at the first or second prenatal visit by each clinic and medical charts maintained by the clinicians. A subset of charts was compared with the computerized data for validation.17 Approval from human subjects was received from all participating hospitals.Of the 9215 pregna...