The Urinary Protein-to-Creatinine Ratio in Canadian Women at Risk of Preeclampsia: Does the Time of Day of Testing Matter?

OBJECTIVES: To determine the performance of a protein-to-creatinine ratio threshold of 30 mg/mmol in pregnant women investigated for hypertension according to the time of day of the sample. METHODS: This prospective study included ambulatory pregnant women investigated for hypertensive disorders. A single voided random urine specimen was obtained to determine the protein-to-creatinine ratio, followed immediately by a 24-hour urine collection. Statistical analyses included Spearman correlation, sensitivity, specificity, predictive values, likelihood ratios, and receiver-operator characteristic curves with 95% confidence intervals. A P value < 0.05 was considered statistically significant. RESULTS: Among the 91 specimens analyzed, 47.3% showed significant proteinuria in the 24-hour collection and 33% were first morning samples. The protein-to-creatinine ratio and 24-hour urinary protein excretion were highly correlated (r = 0.92, P < 0.001). The diagnostic accuracy of the protein-to-creatinine ratio threshold of 30 mg/mmol was lower in first morning samples than in samples obtained during the rest of the day, with sensitivity 58% and 90%, specificity 93% and 100%, positive predictive value 88% and 100%, negative predictive value 72% and 92%, positive likelihood ratio 8 and not calculable, and negative likelihood ratio 0.45 and 0.1, respectively. The receiver-operator characteristic area under the curve was 0.94 (95% CI 0.86 to 1) for first morning samples and 1.0 (95% CI 0.99 to 1) for other samples. CONCLUSION: A protein-to-creatinine ratio threshold of 30 mg/mmol reliably identifies significant proteinuria, but its reliability is reduced in first morning samples. Consequently, such samples should not be used for this purpose.