A nomogram model for predicting maternal cardiovascular complications and neonatal adverse outcomes in pregnant patients with pulmonary arterial hypertension

Abstract

Background: Pulmonary arterial hypertension (PAH) during pregnancy significantly increases maternal and fetal mortality risk. We developed nomogram prediction models from retrospective data to assess maternal cardiovascular risks and neonatal adverse outcomes.

Methods: Our study included 170 pregnant women, divided into training (70%) and validation (30%) sets. Predictors of outcomes were identified using logistic regression in the training set, and nomograms were constructed to predict maternal cardiovascular complications and neonatal adverse outcomes. Model performance was evaluated through internal validation.

Results: Predictors of cardiovascular complications included severe PAH (OR = 4.80), New York Heart Association (NYHA) classification ≥ III (OR = 25.94), ST-T changes (OR = 25.18), total bilirubin (OR = 1.49), albumin (OR = 0.87) and lactate dehydrogenase level (OR = 1.01). The nomogram showed high predictive accuracy with concordance indices of 0.96 and 0.91, areas under the ROC curve of 0.96 and 0.93. Neonatal outcome predictors included gestational age at termination (OR: 0.93), maternal platelet count level (OR: 0.99), and B-type natriuretic peptide level (OR: 1.00). The corresponding nomogram showed concordance indices in the training set and validation set were 0.92 and 0.73, respectively, with area under the ROC curve values of 0.92 and 0.73.

Conclusions: Nomogram models based on the above factors useful tools for predicting cardiovascular complications and neonatal adverse outcomes in pregnant women with PAH, potentially aiding in early detection and timely intervention. Further validation is needed to confirm their accuracy in broader clinical settings.

Keywords: Pregnancy; maternal and neonatal outcomes; prediction model; pulmonary arterial hypertension.

Figure 1.

The flow chart of eligibility, inclusion, and exclusion criteria for the study population.

Figure 2.

Nomogram for preoperative estimation of cardiovascular complications risk in pregnant patients with pulmonary arterial hypertension (PAH) and its ROC curves in the training set and validation set. (A) Nomogram to estimate the risk of cardiovascular complications. To utilize the nomogram, locate each variable on its respective axis, then draw a upwrd line to the points” axis to determine the individual score. Sum all variables scores and project the total tothe“total points” axis to estimate the probabilities of cardiovascular complications, shown at the bottom of the nomogram. For PAH classification, 0 = mild severity, 1= moderate severity, 2 = severe severity. For ST-T changes, 1 = no changes, 2 = presence of changes. For NYHA classification ≥ III, 0 = absence, 1 = presence. (B) ROC curve of the nomogram prediction model for cardiovascular complications in the training set (n = 118). (C) ROC curve of the nomogram prediction model for cardiovascular complications in the validation set (n = 52). PAH, pulmonary arterial hypertension. TBIL, total bilirubin; ALB, albumin; LDH, lactate dehydrogenase; NYHA, New York Heart Association; ROC, receiver operating characteristic.

Figure 3.

Calibration curves and decision curve analysis (DCA) curves of the nomogram model for preoperative estimation of cardiovascular complications in the training set and validation set (A), calibration curve of the nomogram model in the training set. (B) Calibration curve of the nomogram model in the validation set. (C) DCA curves of the nomogram model in the training set. (D) DCA curves of the nomogram model in the validation set. The calibration plots show close agreement between predicted and observed event rates. The Hosmer–Lemeshow goodness-of-fit test yielded p = 0.86 (training) and p = 0.99 (validation), indicating good model calibration. DCA demonstrates the clinical utility of the model across a range of threshold probabilities.

Figure 4.

Nomogram for prediction of neonatal adverse outcomes risk in pregnant patients with pulmonary arterial hypertension (PAH) and its ROC curves in the training set and validation set (A), nomogram to estimate the risk of neonatal adverse outcomes. (B) ROC curve of the nomogram prediction model for neonatal adverse outcome in the training set (n = 93). (C) ROC curve of the nomogram prediction model for neonatal adverse outcome in the validation set (n = 41). BNP, B-type natriuretic peptide; PLT platelet count; ROC, receiver operating characteristic.

Figure 5.

Calibration curves and decision curve analysis (DCA) of the nomogram model for preoperative estimation of neonatal adverse outcomes in the training set and validation set (A), calibration curve of the nomogram model in the training set. (B) Calibration curve of the nomogram model in the validation set. (C) DCA curves of the nomogram model in the training set. (D) DCA curves of the nomogram model in the validation set. The calibration plots show close agreement between predicted and observed event rates. The Hosmer–Lemeshow goodness-of-fit test yielded p = 0.38 (training) and p = 0.81 (validation), indicating adequate model calibration. DCA confirms the clinical utility of the model across a wide range of threshold probabilities.