大口 昭英
日本産科婦人科学会雑誌 65(11) 2604-2615 2013年11月
Although many prediction methods for pregnancy-induced hypertension (PIH), especially preeclampsia (PE), have been developed, there are no clinically useful single markers, whose positive likelihood ratios (LR+s) are ≥ 10. LR+s of maternal factors, such as primipara, obesity, past history of PIH, family history of hypertension, blood pressure levels, and uterine artery indices are all <10; and LR+s of angiogenesis-related factors, such as placental growth factor (PlGF), soluble fms-like tyrosine kinase 1 (sFlt-1), soluble endoglin (sEng) and sFlt-1/PlGF ratio, are also<10. In the first clinical study, we developed a prediction method of PE with onset at <36 weeks of gestation using the onset threshold of the sFlt-1/PlGF ratio at 26-31 weeks of gestation and using multivariate models containing maternal factors and angiogenesis-related factors at 19-31 weeks of gestation, which showed LR+ of >10. In the second clinical study, while changing the laboratory methods from enzyme-linked immunosorbent assay (ELISA) to electrochemiluminescence immunoassay (ECLIA), we developed a prediction method of the imminent onset of PE with onset at <4 weeks after blood sampling during 19-31 weeks of gestation, resulting in improvements of LR+, sensitivity (SE), and the positive predictive value (PPV). In addition, we developed a three-step approach: the first step is a positive selection using either blood pressure levels of 120/80mmHg at 20-23 weeks of gestation or a past history of PIH, the second step is a positive selection using plasma levels of PlGF at 19-31 weeks of gestation, and the final third step is a positive selection using plasma levels of sFlt1 at 19-31 weeks of gestation. This approach resulted in improvement of cost performance while showing LR+ of >10, SE of >0.80 and PPV of >0.20. In Guideline 2009 for the Care and Treatment of Hypertension in Pregnancy (PIH) (edited by the Japan Society for the Study of HYPERTENSION IN PREGNANCY), the editors clearly concluded that there are no clinically useful ways to predict PE. We tried to solve this difficult clinical problem by the following 5 approaches: first, the outcome was changed from PE to PE with onset at <36 weeks of gestation, and imminent onset of PE with onset at <4 weeks from blood sampling at 19-31 weeks of gestation; second, cutoff level of the sFlt-1/PlGF ratio was changed from normal reference ranges to the onset threshold of PE; third, multivariate logistic regression models using maternal factors and angiogenesis-related factors were introduced; fourth, the measurement method of angiogenesis-related factors was changed from ELISA to ECLIA; finally, concomitant measurements of sFlt-1 and PlGF were changed to serial measurements. These approaches successfully resulted in not only high LR+, high SE and high PPV, but also good cost performance. Measurements of sFlt-1 and PlGF by ECLIA are very easy, fast and automated. In addition, the measurements of sFlt-1 and PlGF at 19-31 weeks of gestation informed us about the imminent risk of PE. When should we introduce the measurements of sFlt-1 and PlGF into clinical practice for predicting and managing PE? Now.