In this nested case–control study performed in high risk women we found lower PlGF concentration from 18 + 0 to 20 + 0 weeks of gestation in sera of women who later developed early-onset preeclampsia. Serum sFlt-1/PlGF ratio over 40 at 26 + 0 to 28 + 0 weeks of gestation has high specificity and sensitivity in identifying women who developed early-onset disease. Women who were later diagnosed with late-onset, severe form of preeclampsia had significantly increased levels of sFlt-1 already at 12 + 0 to 14 + 0 weeks of gestation compared to women with late-onset, non-severe form of the disease. This is a new finding in a study with a small sample size and needs to be confirmed in a larger study population.
We would like to emphasize that these findings are likely to be specific to high risk women.
Earlier studies have reported that the ratio of sFlt-1/PlGF is a better predictor of early preeclampsia than the two markers alone . We found that serum sFlt-1/PlGF ratio was significantly higher from 26 + 0 to 28 + 0 weeks of gestation in women who later developed early-onset preeclampsia compared to the three other study groups. However, in our study PlGF alone was a better predictor of early pre-eclampsia at 18 + 0 to 20 + 0 weeks of gestation. In the study of Levine and coworkers  sFlt-1/PlGF ratio was significantly higher already from 17 to 20 weeks of gestation in women who developed preterm preeclampsia. Moore Simas and coworkers  found that sFlt-1/PlGF ratio measured at 22 to 26 weeks of gestation was highly predictive of early-onset preeclampsia. In these two studies significantly increased sFlt-1/PlGF ratio was also found at 25 to 30 weeks of gestation in women destined to develop late-onset preeclampsia. In contrast, we did not observe significant difference in sFlt-1/PlGF ratio between the late onset preeclampsia group and the two control groups at any of the measured time points. In our study all women who were destined to develop early-onset preeclampsia could be identified at 26 + 0 to 28 + 0 weeks of gestation, 4.0 to 6.3 weeks before the diagnosis of preeclampsia, by sFlt1/PlGF ratio, cut-off point set so, that no false positives existed. In other words, we were able to identify the developing early-onset disease among the high-risk population at least one month before the clinical diagnosis was made. From the clinical perspective, these findings are significant by helping obstetricians to make decisions on the management of the high-risk women; how intensive follow-up is required, whether hospitalization is essential, and whether it is necessary to be prepared to early delivery in a tertiary care center.
Interestingly, recent studies have proposed that sFlt-1/PlGF ratio may be useful not only as a predictor of preeclampsia but also in the differential diagnosis of hypertensive diseases of pregnancy, and, as well, serve as a prognostic parameter in patients with established preeclampsia [16, 17]. Verlohren and coworkers proposed that sFlt-1/PlGF ratio may be used in individualized risk stratification in patients with clinical preeclampsia, and clinical management can be adapted accordingly . Rana and coworkers , studied angiogenic factors in 616 women with a suspected preeclampsia, and found that in early-onset (<34 weeks) disease sFlt-1/PlGF ratio predicts adverse outcome occurring within two weeks. Even in association with the atypical presentation of preeclampsia, with relatively normal blood pressure or with no proteinuria, sFlt-1/PlGF ratio performed well. Rana and coworkers showed an inverse correlation between sFlt-1/PlGF ratio and the remaining duration of pregnancy.
Many studies have demonstrated low levels of PlGF already in the first trimester in women who later developed preterm or term preeclampsia [10, 19–22]. Cowans and coworkers  and Noori and coworkers  found significantly lower PlGF levels from the first trimester of pregnancy in women who later developed early-onset or preterm preeclampsia, but not in women with term preeclampsia. We found significant differences in the serum PlGF concentrations between the early-onset preeclampsia group and the three other study groups, however, not earlier than 18 + 0 to 20 + 0 weeks of gestation. We showed that at 18 + 0 to 20 + 0 weeks of gestation, 7.3 to 13.4 weeks before the diagnosis, PlGF could predict early-onset preeclampsia in our high-risk population with an AUC 91.4%. The performance of PlGF as a predictive marker improves over the course of pregnancy so that, 26 + 0 to 28 + 0 weeks of gestation, PlGF has AUC 99.8% (PPV 80%, NPV 100%) in predicting early-onset preeclampsia in this study population. However, during the second half of pregnancy, sFlt-1/PlGF ratio performed even better (PPV 100%, NPV 100%).
Our study is in line with most earlier studies in finding significant differences of sFlt-1 concentration between women who are destined to develop preeclampsia and controls not earlier than during the second half of pregnancy [15, 20, 25]. However, some studies have reported elevated sFlt-1 concentration already from thirteen to twenty weeks of gestation [18, 26–28]. Analyses performed using first trimester serum samples have resulted in negative findings [19, 21, 22]. Vatten and coworkers  studied a cohort consisting of 154 women who later developed preterm preeclampsia (diagnosis before 37 weeks of gestation), 190 women with term preeclampsia (diagnosis after 37 weeks of gestation) and 392 control women. They found that sequential change of sFlt-1 concentration between first and second trimester strongly predicts preeclampsia. Moreover, in the preterm preeclampsia group the sequential change in sFlt-1 concentration was steeper than in the term preeclampsia group. One major difference between our study and the studies of others was that we did not find any differences between the late onset preeclampsia and the two control groups at any of the measured time points.
Vasoactive agents may behave differently not only in early-onset and late-onset preeclampsia, but also in severe and non-severe cases. This might reflect the differences in pathogenesis of subtypes of preeclampsia. Early-onset preeclampsia is considered more as a placental disease whereas late-onset more as a maternal disease. Early-onset preeclampsia is considered a consequence of abnormal placentation and it has often a familial predisposition suggesting a genetic component and a high recurrence risk . In the early-onset form of the disease placental insufficiency often results in fetal growth restriction. Late-onset disease emerges from maternal predisposing risk factors, eg. metabolic factors, associated with obesity, chronic hypertension, diabetes, and interacting with a normal placenta . Our finding of significantly higher serum sFlt-1 concentration in maternal serum already from 12 + 0 to 14 + 0 weeks of gestation in women who developed late-onset, severe, preeclampsia compared to women who developed late-onset, non-severe form of the disease could reflect this phenomenon. Women with late-onset, severe preeclampsia gave birth to significantly lighter newborns, whereas women with late-onset, non-severe form of the disease had significantly higher BMI in early pregnancy arguing for a metabolic etiology.
The strength of our study is a well-characterized prospective cohort. However, we acknowledge that the number of women who developed early-onset preeclampsia was small even if we studied women with clinical risk factors for preeclampsia. One limitation of our study is that we do not have serum samples available after 30 weeks of gestation. Therefore we are not able to study how these biomarkers behave close to the diagnosis of late-onset preeclampsia. Our prospective study reflects the true composition of early-onset and late-onset preeclampsia in women at high-risk, which may explain differences between the present and some earlier studies. If the proportion of women with early-onset preeclampsia compared to women with late-onset preeclampsia is higher than usually observed in the clinical practise, and the early- and late-onset preeclampsia groups are analysed together, the results may be biased. Moreover, many preanalytical and methodological differences may exist between studies. These include the different sample material (plasma or serum) and the storage and handling of the samples. The assays may have a different specificity and sensitivity and the study population may differ between studies.