OBJECTIVE The aim of this study was to evaluate the diagnostic accuracy of random urine protein-creatinine ratio for prediction of significant proteinuria (> or = 300 mg/24 h) in patients with new-onset mild hypertension in late pregnancy. METHODS Medical records of 185 consecutive pregnant patients with new onset of mild hypertension in late pregnancy were reviewed. Random urine samples were taken before 24-hour urine collection. The predictive values of the random urine protein-creatinine ratio for diagnosis of significant proteinuria were estimated by using at least a 300-mg protein level within the collected 24-hour urine as the gold standard. RESULTS Thirty-nine patients (21%) had significant proteinuria. There was a significant association between 24-hour protein excretion and the random urine protein-creatinine ratio (rs = 0.56, P <.01). With a cutoff protein-creatinine ratio greater than 0.19 as a predictor of significant proteinuria, sensitivity and specificity were 85% and 73%, respectively. Positive and negative predictive values of the test were 46% and 95%, respectively. CONCLUSIONS The random urine protein-creatinine ratio was a poor predictor for significant proteinuria in patients with new-onset mild hypertension in late pregnancy.