Optimal bedside urinalysis for the detection of proteinuria in hypertensive pregnancy: a study of diagnostic accuracy

Objective To compare semi-quantitative visual and automated methods of urine testing with fully quantitative point of care urinalysis for the detection of significant proteinuria (0.3 g/24 hours) in pregnancy complicated by hypertension. Design A prospective comparative study. Setting A large teaching maternity hospital. Sample One hundred and seventy-one pregnant women referred to the obstetric day-care unit for assessment of newly arisen hypertension. Methods Early morning urine specimens were tested with four dipstick techniques (Multistix 8SG visual and automated and microalbumin/creatinine ratio visual and automated; Bayer, Elkhart, USA) as well as a fully quantitative measure of the microalbumin/creatinine ratio with the DCA 2000 (a point of care assay for albumin; Bayer). These results were compared to a 24- hour urine protein measurement and measures of diagnostic accuracy/prediction are reported. Main outcome measures Significant proteinuria (0.3 g/24 hours) measured by laboratory assay. Results Automated dipstick urinalysis using the Clinitek 50 has significantly better predictive values for significant proteinuria (LR+ 4.27, 95% CI 2.78 to 6.56; LR- 0.225, 95% CI 0.14 to 0.37) than conventional visual dipstick urinalysis (LR+ 2.27, 95% CI 1.47 to 3.51; LR- 0.635, 95% CI 0.49 to 0.82). Dipstick microalbumin/creatinine ratio testing did not improve overall detection rates with automated or visual testing. Fully quantitative point of care measurement of albumin/creatinine ratio (ACR) was significantly better than any dipstick technique (LR+ 14.6, 95% CI 6.74 to 31.8; LR- 0.069, 95% CI 0.030 to 0.16). Conclusions This study confirms that in pregnancy automated dipstick urinalysis is a more accurate screening test for the detection of proteinuria than visual testing. ACR testing can offer a significant improvement over conventional urinalysis if a fully quantitative method of detection is employed that uses pregnancy-specific thresholds. Dipstick assessment of ACR does not improve the detection rate of significant proteinuria.