Main results
The research presented in this paper focused on the development of a novel tool, based on data-mining methods, to support choosing proper mode of delivery in HDP patients without other indications for cesarean section. In this study, the edema (cut-off value: edema is limited to the ankle joint) was the most important predictor for HDP patients to choose mode of delivery. Then β2-microglobulin (cut-off value: 1.71 ug/ml) and proteinuria (cut-off value: 1) was second most important predictor, and HCT (cut-off value: 36.9%) came the third. Participants with clinical features of no edema and β2-microglobulin ≤1.71ug/ml reached VD rate of 86.4%, which was the highest VD rate among all subgroups.
In the study, with the classification tree, when choosing mode of delivery, the first feature should be considered was edema, which reflected loss of albumin. Former studies usually emphasized the blood pressure when evaluating severity of HDP, however, other important features like edema and proteinuria should also be taken into consideration[7]. Edema was easily to be observed and evaluated during clinical process, and obstetricians should pay attention to this syndrome especially when making a trial of VD. In this study, the cut point of edema for selecting VD or CD was Level 1. When edema was limited to the ankle joint or above, CD should be considered with the top priority.
HDP affects the function and morphology of kidney. The second most important feature for choosing mode of delivery were β2-microglobulin and proteinuria. In former studies, proteinuria was believed to be a manifestation of kidney damage caused by HDP and as a criterion to evaluate the severity of pre-eclampsia, which was often used as a sign of pregnancy termination. In the study, the model kept both β2-microglobulin and proteinuria to distinguish mode of delivery. β2-microglobulin reflects glomerular lesions more sensitively than proteinuria[8]. In this study, when β2-microglobulin > 1.71μg/ml, CD was considered to be the first mode of delivery for HDP.
Proteinuria is defined as the excretion of 300 mg or more of protein in a 24-hour urine collection. A determination of 1+ is considered as the cutoff for the diagnosis of proteinuria[9]. Although, proteinuria as a semiquantitative analysis has the problem of stability, it is still the most convenient and low-cost way to reflect renal function in clinical practice. Some studies showed that the severity of proteinuria was significantly associated with maternal and fetal complications in HDP[10-12]. Proteinuria is still an effective indicator to evaluate the status of HDP. In our study, the CD rate in patients with proteinuria≥1+ was higher than that in patients with proteinuria<1+.
HDP patients, are susceptible to anemia and electrolyte disturbance. The third most important feature when deciding mode of delivery was HCT, which is an indicator of erythrocyte status in disease. HCT is often used as a diagnostic marker for anemia and can also be used as an experimental basis for clinical determination of whether patients need rehydration and electrolyte supplementation. Dong’s study showed that compared with normotensive pregnant women, HDP patients were in a relatively anemic state and the levels of HCT were reduced[13] . In our study, when HCT was less than 36.9%, the probability of CD increased significantly. As a supplementary indicator, HCT reflected the features of HDP in the blood system and could assist obstetricians in the decision-making on mode of delivery.
Strengths and limitations of the study
The study employed the classification tree to elucidate factors that are relevant predictors of delivery modes. The classification tree is gaining broader acceptance in the area of biomedical research, yet they are not in widely use. One of the benefits of classification tree is that the output – a simple decision tree – is analogous to a diagnostic process with which medical professionals are familiar. As a result, the output of classification tree is generally easy to understand.
Limitations of this study include a retrospective design and lack of information on confounders, which might have influenced assignment on mode of delivery. A randomized clinical trial of VD vs CD for HDP patients would be the gold standard clinical design, although this would be very difficult to conduct.
Interpretation
For severe HDP patients especially with complication, obstetricians tend to choose CD because of clear indications for cesarean section. The purpose of the treatment is to prevent the occurrence of eclampsia, reduce morbidity and mortality. However, the ideal mode of delivery for HDP patients without other indications for cesarean section still remains to be established. Some studies concerning mode of delivery were documented. Amorim et al. investigated the effect of delivery modes among women with HDP in Brazil[14] and they found an increased risk of adverse outcomes for HDP patients with CD. Levine LD et al. found that CD in labors were associated with the overall composite of adverse outcomes seen in women with HDP[15].
In clinical practice, how to select mode of delivery for HDP mainly relied on clinical signs and symptoms, which are variable and non-specific (main syndromes: hypertension, headache, visual disturbance, epigastric pain, reduced fetal movements and small for gestational age infant)[16]. It is crucial to synthetically consider various related syndrome before treatment, which put forward a high requirement on obstetricians. In this study, the classification tree model had told you what the cluster of HDP’s features in patients with high or low VD rate. To some extent, it simulated the clinical decision-making process, which provided assistance for inexperienced young doctors.