Discussion
Principle Findings
In this study, the predictive value of sFlt-1/PlGF ratio for subsequent PE occurrence following hypertensive symptoms was explored for the first time among Chinese women (including gestational hypertension & chronic hypertension). (1) For the different types of hypertensive disorders in pregnancy (HDP), the sFlt-1 /PlGF ratios in PE and chrHBP with PE were significantly higher than GH and chrHBP, in which the sFlt-1 /PlGF ratios in early-onset PE and chrHBP with PE were higher than the late onset. (2) For gestational hypertension, sFlt-1/PlGF ratio> 36.44 could rule in PE occurrence (AUC=76.17% - 78.39%) in 1-4week with PPV 0.91-0.94 & NPV 0.67 before 34 weeks, while sFlt-1 /PlGF ratio< 51.52 could rule out PE occurrence (AUC=75.29% - 78.39%) in 1-4 weeks with NPV 0.91-0.96 & PPV 0.62-0.73 after 34 weeks. (3) For chronic hypertension, sFlt-1 /PlGF ratio> 47.78 could rule in PE occurrence (AUC=82.5% - 86.36%) in 1-4 weeks with PPV 0.9-0.96 & NPV 1.0 before 34 weeks, while sFlt-1 /PlGF ratio< 28.61 could rule out PE occurence(AUC=72.41% - 73.03%) in 1-4 weeks with NPV 0.88 & PPV 0.43-0.79 after 34 weeks.
Clinical research significance
In the studies on predicting PE with sFlt-1/PlGF ratios, the cutoff point of 38 was widely accepted to rule out the presence of PE among general population. Zeisler and colleagues reported the NPV up to 97.9% for the ratios on PE prediction within 2 weeks, 95.7% within 3 weeks and 94.3% within 4 weeks, with sensitivity of 66.2% and specificity of 83.1% in common[6]. For the prediction of early-onset PE, the ratio of sFlt-1/PlGF of ≥85 was established as a useful tool for screening in the general pregnancy, with sensitivity at 82% and specificity at 95%. [11] As a universal screening method from general population, it is rational to select the cutoff point with high specificity (usually 95% or above) so as to minimize the chance of incorrectly labelling healthy women as patients. [12]
In our study, the cut-offs for early-onset PE with comparable predictive values in hypertensive patients were substantially lower than pre-established value of 85 in the general population, which suggested an increased sensitivity of the test among hypertensive women, irrespective of pregnancy-specific(sFlt-1/PlGF ratio 37 with PPV 0.91-0.94 and NPV 0.67) or chronic conditions(sFlt-1/PlGF 48 with PPV 0.90-0.96 and NPV 1.0). This might suggest that hypertensive women with sFlt-1/PIGF ratios just >38 were at increased risk for PE, in which surveillance and interventions should be properly implemented.
Unlike early-onset PE, improved diagnostic accuracy was previously reported in patients with late-onset PE using the cut-off values of <38 and >110 for the exclusion and prediction of PE, respectively[13]. However, ambiguity still existed in between the two values including 30% of late-onset PE and 20% healthy individuals[14]. In this study, for the exclusion of progression from GH/chrHBP to late-onset PE, we reported the cut-offs at 51.52 (NPV 91-95%) and 28.61(NPV 88%), respectively. However, the sensitivity(60-62%) and specificity(75-79%) were still suboptimal in hypertensive patients.
The discrepancy of pathophysiology between early- and late-onset PE might explain our findings. PE, in nature, is a hypertensive disorder affecting multiple systems during pregnancy, and placenta malformation is thought to be the key process in the development of PE, in particular the early-onset PE[15]. The early-onset PE was known as the mian result of inadequate vascular remodeling and placenta implantation. As shown in our study and the others alike, sFlt-1 and PlGF, the angiogenic factors, were extremely high in the early-onset PE [16].
Unlike early-onset PE, the late-onset PE is considered as the consequence of altered hemodynamics (increased blood volume and cardiopulmonary stress) and senescent placenta during late pregnancy but with less angiogenic factors imbalance [17]. This might also explain higher cutoff point in the late-onset than the early-onset PE (≥110 for late- and ≥85 for early-onset) in literature. Also shown in our sudy, the ratio of sFlt-1/PlGF had superior predictive value for early-onset other than late-onset PE.
Studies on the sFlt-1/PlGF ratios in subtypes of HDP have shown its value in differential diagnosis in PE. Verlohren et al. discovered that the sFlt-1/PlGF ratios in PE/HELLP group, particularly the early-onset subsets (<34 weeks), were significantly higher than the normotensive, chrHBP and GH women (P<0.01)[18]. Limited studies on the sFlt-1/PlGF in chrHBP, showed relatively lower sFlt-1/PlGF ratios ratios even with fluctuation of BP[19][20]. Our results also showed the lowest levels of sFlt-1/PlGF ratios in chrHBP among the 4 subtypes of HDP, while the ratio significantly increased in chrHBP with superimposed PE, similar to the levels of PE progressed from GH, particularly <34 weeks.
The findings in our investigation were consistent with Verlohren’s study, demonstrating the value of sFlt-1/PlGF ratios on differentiating the subtypes of HDP , which also indicated that early-onset PE on the basis of GH and chrHBP may have similar pathological mechanism associated with placenta malformation. While the sFlt-1/PlGF ratios in late-onset PE from chrHBP lower than PE developed from GH might reflect the inconsistent contribution of placenta inflammation in their occurrence.
Limitation
This study only included Chinese women, the results may not be applicable to population with different racial backgrounds. Also this was an observational single-center study with limited sample size. No blinding was employed in the study, which may result in potential over-diagnosis and interventions, though antenatal care was not conducted according to sFlt-1/PlGF ratios. Larger scale of randomized cohort studies are needed to validate our preliminary results.