Background The gold standard for diagnosing proteinuria in pregnancy is ≥300 mg of protein in a 24-h urine collection.1 Practical considerations have reduced the utility of this investigation and National Institute for Health and Clinical Excellence have recommended that a diagnosis of significant proteinuria can be made using a spot urinary protein-to-creatinine ratio (PCR).2 Given the reported variation in protein assay performance, we assessed the diagnostic accuracy of PCR at St Mary's Hospital, a tertiary referral centre in Manchester.

Method Creatinine Jaffé and Roche Diagnostics Urinary/cerebrospinal fluid methods were used to measure PCR and 24-h urine, respectively. Data was collected from January 2009-October 2010 from antenatal patients who had a spot PCR and 24-h urine protein assessment. Maternal outcome data was also analysed.

Results There were 141 patients and 154 occasions where patients had both tests performed. There was a significant correlation between PCR and 24-h protein measurements (Spearman r=0.863; p<0.0001) with an AUROC of 0.92 (95% CI 0.85 to 0.99). The sensitivity and specificity of PCR at ≥20, ≥30 and ≥60 mg/mmol were 99% & 47%, 93% and 65% and 72% & 99%, respectively. Using the standard threshold of 30 mg/mmol in this cohort would have given a false negative result in 8/136 women, but a false positive result in 6/17 women.

Conclusion In the presence of an abnormal PCR (≥30 mg/mmol), National Institute for Health and Clinical Excellence recommends no further urine protein assessments. Our data suggest that in women with PCR <60 mg/mmol the false positive rate is such that a second assessment could improve diagnostic accuracy and avoid unnecessary obstetric intervention.