3. Results
A study design flowchart is presented in Figure 1. Of the 460-singleton pregnant woman recruited into this study, 423 (92%) completed the follow-up until delivery. In 11 cases (2.6%), GCF-PLAP samples were unsatisfactory for analysis and were excluded from the study; therefore, 412 cases (89.6%) were available for analysis.
The baseline characteristics of the study population are summarized in Table 1. Of the 423 pregnant women recruited and followed throughout pregnancy, 18 of them subsequently developed PE (4.3%), and five (1.2%) required delivery before 37 weeks of gestation. Maternal age, maternal weight and height, body mass index, and smoking status did not differ significantly between patients with PE and controls. Systolic blood pressure (p<0.001), diastolic blood pressure (p=0.007), and median arterial blood pressure (p=0.006), measured during early pregnancy, were significantly higher in women who developed PE when compared to controls. No statistically significant differences in periodontal parameters were identified between patients who developed PE and controls, and no statistically significant association was identified between periodontal clinical diagnosis and the subsequent development of PE (p=0.617) (Table 2). The median maternal GCF-PLAP concentration was 63.7 (interquartile range [IQR]: 88.9) pg/ml in healthy patients, 46.6 (IQR: 47.2) pg/ml in patients with gingivitis, and 42.4 (IQR: 46.7), 41.1 (IQR: 51.8), 34.7 (IQR: 46.88) pg/ml at periodontitis stage I, stage II – III and stage IV, respectively, without statistically significant differences among them (p=0.407).
GCF-PLAP concentrations at 11-14 weeks were compared between patients who subsequently developed PE and controls. The median maternal GCF-PLAP concentration was significantly higher in the PE group than that of the controls (77.5 pg/ml (IQR: 41.5) vs. 41.3 pg/ml (IQR: 50.1), p=0.015) (Figure 2A). In addition, PLAP concentrations were also measured in paired plasma and GCF samples from 80 women from the same cohort. The median plasma PLAP concentrations were 24.2 pg/ml (IQR: 2.5) and 24.6 pg/ml (IQR: 7.6) in the control and PE groups, respectively. In the paired GCF samples, median PLAP concentrations were 66.1 pg/ml (IQR: 4.3) and 99 pg/ml (IQR: 17.8) in the control and PE groups, respectively (p=0.011) (Figure 2B). There was no correlation between plasma and GCF-PLAP concentrations. The observed amount of PLAP in GCF was 3- to 6-fold higher than in plasma samples.
Multiple logistic regression analysis identified an association between first-trimester systolic blood pressure (OR: 1.07; 95% CI 1.00-1.015; p=0.004) and GCF-PLAP concentrations (OR: 1.008, 95% CI 1.000-1.015; p=0.034) (Table 3) in women who subsequently developed PE. The results of the bootstrap analysis were similar to those observed in the logistic regression model (Table 3). The GCF-PLAP concentration combined with systolic blood pressure at 11-14 weeks of gestation was found to be a good predictor of PE, with a specificity of 72%, a sensitivity of 83%, a PPV of 12%, and an NPV of 99%. The positive likelihood ratio was 2.9, and the negative likelihood ratio was 0.3. The model correctly classified 72% of the women who developed PE. The AUC for GCF-PLAP concentrations alone at 11-14 weeks of gestation was 0.67; for systolic blood pressure, 0.74, and for GCF-PLAP concentrations and systolic blood pressure, 0.77 (95% CI: 0.70 – 0.85) (Figure 2C). In the sub-analysis, dividing the PE pregnancies into preterm PE (≤ 37 weeks) and term PE (> 37 weeks), the observed AUC was 0.85 (95% CI: 0.81 – 0.93) for preterm PE and 0.72 (95% CI: 0.58 - 0.82) for term PE (Figure 2D and 2D, respectively). All five cases (100%) of preterm PE observed in the current study were correctly classified by the model.