Predictive utility of prior negative urine cultures in women with suspected recurrent uncomplicated urinary tract infections

Purpose: Guidelines recommend treating women who have symptoms of an uncomplicated urinary tract infection with antimicrobials without performing a urine culture. However, 10% to 50% of women with urinary tract infection symptoms are found to have a negative culture. Urinalysis data are useful to predict a negative culture. We evaluated how a previous negative culture predicts the likelihood of a subsequent negative culture. Materials and Methods: We gathered retrospective data on women 18 years old or older with symptoms of an uncomplicated urinary tract infection who submitted urine cultures as outpatients from 2011 to 2017. Univariate analysis and multivariable regression models were used to determine the likelihood ratios and risk ratios of predicting a negative culture. Results: Of the 20,759 patients 9,271 (44.7%) had a negative culture, defined as less than 103 CFU/ml, and 6,958 (33.5%) had at least 1 prior culture, including 4,510 (64.8%) with at least 1 prior negative culture and 2,634 (58.4%) with a subsequent negative culture. Variables associated with an increased likelihood of another negative culture were a prior negative culture (LR 1.43, 95% CI 1.387-1.475), prior negative culture and negative urinalysis (LR 1.839, 95% CI 1.768-1.913), and vaginal irritation and/or discharge (LR 1.335, 95% CI 1.249-1.427, each p <0.001). Urinalysis had 83% specificity and 78% positive predictive value. These values were significantly enhanced if the patient had a prior negative culture without a prior positive culture (95% and 87%, respectively). Conclusions: In women with recurrent urinary tract infection symptoms a previous negative culture and negative urinalysis are highly predictive of another negative culture. Women with recurrent urinary tract infection symptoms, and negative urinalysis and urine cultures may benefit from further evaluation.