Performance of dipstick test for diagnosing urinary tract infection among adult patients in a hospital setting in Sri Lank

Introduction: This prospective study compared the diagnostic performance of dipsticks with microscopic urinalysis and with urine culture for rapid diagnosis of urinary tract infection (UTI) in adults.
Methods: A total of 1292 urine samples received at a Teaching Hospital in Sri Lanka were analysed. Sensitivity, specificity, predictive values, likelihood ratios, and discriminative power of urinalysis were evaluated against positive urine culture results.
Results: A pure growth of >105 colonies/ml was obtained from 406 (31.4%) samples received by the laboratory, with Escherichia coli (51.9%) being the most common isolate. There was concordance between leukocyte esterase (LE) and microscopy results (83.05%; kappa = 0.650). Combined dipstick and microscopy analysis reported the highest sensitivity (81.5%) and specificity (99.2%), considering culture as the gold standard. Urine dipstick and microscopy and NIT or LE combination showed positive and negative predictive values of >98% and >86%, respectively. They also reported a positive likelihood ratio of >90 and were the best predictors of urine positive or negative cultures (AUC >0.8). Moreover, the highest degree of concordance with culture results was demonstrated for the same combined test parameters (>90%), with a fair quality of agreement.
Conclusions: Combined performance of either LE or NIT result may reasonably be used to rule in UTI in specific cohorts of patients in resource-limited settings. Confirmatory culture can then be done only for dipstick positive patients, reducing the testing turnaround time and financial burden for hospitals. The concurrent use of urine microscopy can increase the diagnostic accuracy of the dipstick test.