First trimester screening of maternal placental protein 13 for predicting preeclampsia and small for gestational age: in-house study and systematic review

Abstract

Objective: To describe normative levels of PP13 in first trimester of pregnancy and determine the accuracy of PP13 in predicting preeclampsia and small for gestational age (SGA) infants.

Methods: We measured PP13 in archived first trimester serum samples from an unselected maternal cohort of 2989 women. Associations of PP13 levels and diagnostic accuracy in predicting adverse pregnancy outcomes were assessed using multivariate logistic regression models. Due to inadequate number of cases we then conducted a systematic review and subsequent meta-analysis of predictive accuracy. Structured searches including all languages were completed in electronic databases and supplemented by cross-checking reference lists of relevant publications. Characteristics, data extraction and quality assessment of studies was conducted by independent assessors.

Results: Overall, 2678 women were included in the in-house study with 71 (2.7%) preeclampsia cases, 5 (0.2%) early-onset preeclampsia (≤34 weeks) cases; and 191 (7.1%) and 41 (1.5%) infants SGA<10th and <3rd centile. Median (IQR) normative level of PP13 in unaffected pregnancies was 53.5 (37.7-71.8) pg/ml. The area under the receiver operating characteristic curve (AUC) for multivariate models was 0.72 (95%CI 0.66-0.78) for preeclampsia; 0.82 (95%CI 0.63-0.99) for early-onset preeclampsia; 0.73 (95%CI 0.69-0.77) for SGA<10th centile; and 0.83 (95%CI 0.78-0.88) for SGA<3rd centile. Eight studies were included in the systematic review, normative levels of PP13 were assessed in four studies but these were variable; and meta-analysis was performed on seven studies. Sensitivity rates of PP13 based on 5% fixed false positive rates were 24%, 45% and 26% for preeclampsia, for early-onset preeclampsia and SGA, respectively. There was no evidence of between-study heterogeneity.

Conclusions: First trimester PP13, in combination with maternal characteristics and other serum biomarkers was inadequate for screening purposes and predicting women at risk.