Prediction Model for Pre-Eclampsia Using Gestational-Age-Specific Serum Creatinine Distribution

Abstract

Pre-eclampsia (PE) is a pregnancy-related disease, causing significant threats to both mothers and babies. Numerous studies have identified the association between PE and renal dysfunction. However, in clinical practice, kidney problems in pregnant women are often overlooked due to physiologic adaptations during pregnancy, including renal hyperfiltration. Recent studies have reported serum creatinine (SCr) level distribution based on gestational age (GA) and demonstrated that deviations from the expected patterns can predict adverse pregnancy outcomes, including PE. This study aimed to establish a PE prediction model using expert knowledge and by considering renal physiologic adaptation during pregnancy. This retrospective study included pregnant women who delivered at the Wonju Severance Christian Hospital. Input variables, such as age, gestational weeks, chronic diseases, and SCr levels, were used to establish the PE prediction model. By integrating SCr, GA, GA-specific SCr distribution, and quartile groups of GA-specific SCr (GAQ) were made. To provide generalized performance, a random sampling method was used. As a result, GAQ improved the predictive performance for any cases of PE and triple cases, including PE, preterm birth, and fetal growth restriction. We propose a prediction model for PE consolidating readily available clinical blood test information and pregnancy-related renal physiologic adaptations.

Keywords: creatinine; gestational age; pre-eclampsia; pregnancy; renal hyperfiltration.

Figure 1

Prediction performances of all PE cases using the four feature sets. Any cases of PE (PE only + PE with other adverse pregnancy outcomes) were determined as the disease group. Prediction model was established using logistic regression. Boxplots include minimum, first quartile (25%), median, third quartile, and maximum values. Points in the boxplot referred to as outliers indicate cases showing more than 1.5 times the IQR, biased from the matched median value. Y-axis indicates AUC for predicting PE-any, and the box plot summarizes 100 levels of AUCs. Green-, purple-, orange-, and red-colored boxplots were obtained from the testing dataset, while grey-colored boxplots were curated from the training dataset. Abbreviations: PE, pre-eclampsia; SCr, serum creatinine; Mt, measured time; GAQ, quartile groups of gestational-age-specific SCr; IQR, interquartile range; AUC, area under receiver operating characteristic curve.

Figure 2

Prediction performances of PE-only case using the four feature sets. An exclusively diagnosed PE type among the three adverse pregnancy outcomes was defined as the disease group. Prediction model was established using logistic regression. Boxplots include minimum, first quartile (25%), median, third quartile, and maximum values. Points in the boxplot referred to as outliers indicate cases showing more than 1.5 times the IQR, biased from the matched median value. Y-axis indicates AUC for predicting PE-only, and the box plot summarizes 100 levels of AUCs. Green-, purple-, orange-, and red-colored boxplots were obtained from the testing dataset, while grey-colored boxplots were curated from the training dataset. Abbreviations: PE, pre-eclampsia; SCr, serum creatinine; Mt, measured time; GAQ, quartile of gestational-age-specific SCr; IQR, interquartile range; AUC, area under receiver operating characteristic curve.

Figure 3

Prediction performances of PE with early PTB using the four feature sets. Pregnant women with PE and early PTB were categorized as the disease group. Prediction model was established using logistic regression. Boxplots include minimum, first quartile (25%), median, third quartile, and maximum values. Points in the boxplot referred to as outliers indicate cases showing more than 1.5 times the IQR, biased from the matched median value. Y-axis indicates AUC for predicting PE + PTB_early, and the box plot summarizes 100 levels of AUCs. Green-, purple-, orange-, and red-colored boxplots were obtained from the testing dataset, while grey-colored boxplots were curated from the training dataset. Abbreviations: PE, pre-eclampsia; PTB, preterm birth; SCr, serum creatinine; Mt, measured time; GAQ, gestational-age-specific SCr quartile; IQR, interquartile range; AUC, area under receiver operating characteristic curve.

Figure 4

Prediction performances of the triple (PE + PTB + FGR) outcomes using four feature sets. Cases with the triple adverse pregnancy outcomes of PE, PTB, and FGR was determined as disease group. Prediction model was established using logistic regression. Boxplots include minimum, first quartile (25%), median, third quartile, and maximum values. Points in the boxplot referred to as outliers indicate cases showing more than 1.5 times the IQR, biased from the matched median value. Y-axis indicates AUC for predicting PE + PTB + FGR, and the box plot summarizes 100 levels of AUCs. Green-, purple-, orange-, and red-colored boxplots were obtained from the testing dataset, while grey-colored boxplots were curated from the training dataset. Abbreviations: PE, pre-eclampsia; PTB, preterm birth; FGR, fetal growth restriction; SCr, serum creatinine; Mt, measured time; GAQ, quartiles of gestational-age-specific SCr; IQR, interquartile range; AUC, area under receiver operating characteristic curve.

Figure 5

Prediction performances of PE according to GW. The prediction performances of PE-any (A) and PE-triple (B) cases are summarized according to GWs. Prediction model was established using logistic regression. Features for the model are SCr, Mt, and GAQ. Boxplots include minimum, first quartile (25%), median, third quartile, and maximum values. Points in the boxplot referred to as outliers indicate cases showing more than 1.5 times the IQR, biased from the matched median value. Y-axis indicates AUC for predicting PE, and the box plot summarizes 100 levels of AUCs in each GW category. Green-, purple-, orange-, and red-colored boxplots were obtained from the testing dataset, while grey-colored boxplots were curated from the training dataset. The PE-triple case denotes patients with PE, PTB, and FGR. Abbreviations: PE, pre-eclampsia; PTB, preterm birth; FGR, fetal growth restriction; GW, gestational week; AUC, area under receiver operating characteristic curve.

Figure 6

Final PE prediction model. Abbreviations: PE, pre-eclampsia; PTB, preterm birth; SCr, serum creatinine; Mt, measured time; GAQ, quartiles of gestational-age-specific SCr; HTN, hypertension; DM, diabetes; LP, linear predictor.