**3. Results**

The baseline demographics, lipids, and adipokine characteristics of those with PE were compared to those without PE (**Table 1**). The mean age of those who acquired PE was significantly greater than that of those who did not (35.1 vs. 28.44 years; *p* < 0.0001), and their BMI was likewise significantly higher (32.63 vs. 24.99 kg/m2 ; *P* < 0.0001). Except for HDL, which was considerably lower in the PE group compared to those without PE (1.39 vs. 1.569, *p* = 0.043), the lipid profile parameters did not demonstrate any significant differences between the PE group and those without PE (**Table 1**). Leptin levels were statistically substantially higher in the PE group (39.26 vs. 18.46 ng/mL, *P* < 0.0001) than in the control group. Similarly, resistin and visfatin were considerably higher in PEs compared to normotensives (*p* < 0.0001), although adiponectin was significantly lower in PEs compared to non PEs (*p* < 0.0001).

The ROC curves were used to assess the performance of the screening. **Table 2** shows the areas under the ROC curve, the sensitivities and specificities, as well as the threshold points for detecting PE. The accuracy with which biochemical markers can differentiate on the condition of PE was tested in this study. As shown in **Table 2**, the adipokines leptin (92.0%), resistin (91.4%), and adiponectin (90.5%) have good accuracy levels, whereas visfastin (77.1%) has fair accuracy levels in diagnosing PE, according to **Table 2** ratings. With a cut-off point of 50.55 ng/mL, adiponectin had a sensitivity and specificity of 87.8 and 86%, respectively, while leptin had a sensitivity and specificity of 92% with a threshold of 27 ng/mL. Furthermore, resistin had a sensitivity and specificity of 94 and 91%, respectively, with a cut-off point of around 9 ng/mL, whereas visfatin had a sensitivity and specificity of 69 and 83%, with a threshold of 6.67 ng/mL. This suggests that adiponectin, leptin, resistin, and visfatin are effective PE predictors (**Table 2**).

Furthermore, a detailed examination of the ROC plots (**Figure 1**) reveals that they are all far from the diagonal line, which represents 50%, indicating that they are not random guesses but rather meaningful. This indicates that they are quite good at predicting pregnancies that are likely to result in PE. After adjusting for BMI, none of those in the normal BMI category had PE (**Table 3**); as a result, no AUC values for all of the adipokines studied were obtained. The overweight group, on the other hand, had greater AUCs, sensitivities, and specificities. Obese people, on the other hand, had lower sensitivities and specificities. These findings point to a possible influence of BMI on adiponectin, leptin, resistin, and visfatin, as well as a possible negative feedback mechanism in the metabolism of these adipocytokines during pregnancy. However, BMI does not appear to have an effect on the predictive ability of these PE signaling molecules.

There were minor variations in the AUCs, sensitivities, specificities, and threshold points for predicting PE after controlling for family history of hypertension, which is a known confounding factor, but these variations were minor, and the overall effect of these adipocytokines' predictive abilities remained intact (**Table 4**).

**Table 5** shows a multivariate analysis of individual adipokines as PE predictors. Adiponectin, leptin, resistin, and visfatin were included as predictors in Models 1, 2, 3, and 4, respectively, while correcting for confounding factors such as age, parity, BMI, and relative with hypertension, and family history of diabetes and preeclampsia. Based on the criteria analyzed, Model 1 including adiponectin as a predictor was the best model. This means having the greatest Nagelkerke R2 and CCR values of 95 and



*low density lipoprotein cholesterol.*
