Guidelines recommend treating women with symptoms of an uncomplicated urinary tract infection (UTI) with antimicrobials without performing a urine culture. However, 10-50% of women with symptoms of a UTI are found to have a negative culture (NC). Urinalysis (UA) data is useful in predicting a NC. We evaluated how a previous NC predicts the likelihood of a subsequent NC.
Retrospective data on women (age > 18) with symptoms of uncomplicated UTIs, submitting urine cultures as outpatients from 2011 to 2017 were gathered. Univariate analyses and multivariable regression models were used to determined likelihood ratios (LR) and risk ratios for predicting a NC.
Of 20,759 patients, 9,271 (44.7%), had a NC (<103 CFU/mL), and 6,958 (33.5%) had at least one prior culture. Of these, 4,510 (64.8%) had at least one prior NC and 2,634 (58.4%) had a subsequent NC. Variables associated with increased likelihood of another NC: prior NC (LR 1.43 (1.387-1.475), p<0.001), prior NC and negative UA (LR 1.839 (1.768-1.913), p<0.001) and vaginal irritation/discharge (LR 1.335 (1.249-1.427) p<0.001). UA had a specificity and positive predictive value (PPV) of 83% and 78%, respectively. These values are enhanced significantly if the patient had a prior NC without a prior PC (95% and 87%, respectively).
In women with recurrent UTI symptoms, a previous negative culture and negative UA is highly predictive of another negative culture. Women with recurrent UTI symptoms, negative UA and urine cultures may benefit from further evaluation.
The Journal of urology. 2019 May 07 [Epub ahead of print]
Jason E Cohen, Emily M Yura, Liqi Chen, Anthony J Schaeffer
Department of Urology, Northwestern University Feinberg School of Medicine., Department of Preventative Medicine Biostatistics Collaboration Center, Northwestern University Feinberg School of Medicine.