Fig. 6.

The regression models show a significant influence of preeclampsia for partial correlations between CX3CR1 expression in placenta and birth weight (a) and between CX3CR1 expression in placenta and vascularization index (b). In the upper right corner of each graph, p values are shown. Solid line represents a regression curve in Group 1 and intermittent line in Group 2, grey zones represent 95% confidence intervals. In the model (a), the difference between partial correlations due to influence of preeclampsia was statistically significant at p = 0.001. Beta coefficient (slope) in the linear relations between CX3CR1 expression and birth weight in the absence of preeclampsia was 1.04 (p = 0.007), but in case of preeclampsia, the decrease of the coefficient beta was − 0.35 (p = 0.015). The criterion of the accuracy of the approximation r² was 0.64. In the model (b), FGR was also included. Based on this model, the difference between partial correlations between CX3CR1 expression and V/EVTI due to influence of preeclampsia was statistically significant at p = 0.004. Beta coefficient (slope) in the linear relations between CX3CR1 expression and vascularization index in the absence of preeclampsia was 0.00079 (p = 0.017), but in case of preeclampsia, the decrease of the coefficient beta was − 0.00029 (p = 0.038). The criterion of the accuracy of the approximation r² was 0.47