Screening models combining maternal characteristics and multiple markers for the early prediction of preeclampsia in pregnancy: a nested case-control study.

To identify maternal laboratory markers to predict the risk of preeclampsia (PE) in different stages of pregnancy, we analysed 67, 25, and 73, pregnancies developing PE at 11-13, 16-20, and 24-28 wks, respectively. Routine laboratory markers were measured in whole blood or serum and binary logistic regression analysis was used to identify predictive models. At 11-13 wks of gestation, patients who went on to develop PE showed significantly higher concentrations of alanine aminotransferase, aspartate aminotransferase, α-L-fucosidase, 5'-nucleotidase, glutamyl transpeptidase, cholinesterase, and uric acid; plateletcrit was also higher. At 16-20 wks, inhibin A concentration and plateletcrit were significantly elevated. At 24-28 wks, platelets, plateletcrit, and glucose concentration were significantly elevated. Logistic regression analysis showed that an elevation in 5'-nucleotidase was independently associated with PE at 11-13 wks. The combination of inhibin A, diastolic blood pressure, and body mass index was a significant predictor for PE at 16-20 wks, while the combination of glucose and systolic blood pressure was a significant predictor for PE at 24-28 wks. In conclusion, when combined with maternal characteristics, the measurement of 5'-nucleotidase, inhibin A, and glucose levels, represents a potentially valuable risk assessment for PE.Impact statement Preeclampsia (PE) may be viewed as a spectrum of disorders with a severity that is reflected in the levels of specific biomarkers. Consequently, there is a clear need for additional biomarkers that can be used to stratify pregnancies as high or low risk soon after conception. At 11-13 wks of gestation, maternal assays for platelets, plateletcrit, alanine aminotransferase, aspartate aminotransferase, α-L-fucosidase, 5'-nucleotidase, glutamyl transpeptidase, cholinesterase, and uric acid, demonstrated significantly higher values in patients with PE when compared with normal controls. Furthermore, assay results for inhibin A and platelets showed increased values at 16-20 wks of gestation. Assays performed at 24-28 wks of gestation revealed elevated levels of platelets, plateletcrit, and glucose. Our analysis indicated that increases in the levels of 5'-nucleotidase, inhibin A, and glucose, are effective and significant biomarkers that could be used in combination with maternal characteristics to screen for PE at 11-13, 16-20, and 24-28 wks of gestation, respectively. These findings provide a new basis for our understanding of the aetiology underlying PE. Further studies that consider the entire population are now needed and should include the investigation of laboratory markers across different stages of pregnancy. Long-term follow up would also be necessary if we are to explore the full role of laboratory markers in the pathophysiology of PE.