The PROMISE-ACHD Score

Figures

• 

  Central Illustration. PROMISE-ACHD Risk Score for Predicting Perioperative Outcomes in Complex Congenital Heart Disease The PROMISE-ACHD score was developed to predict the risk of in-hospital mortality or Society of Thoracic Surgeons–defined major perioperative complications in adults with congenital heart disease (ACHD) undergoing complex cardiac surgery. The score incorporates nine routinely available clinical and imaging variables. Event rates increased stepwise across low-, intermediate-, and high-risk categories. Restricted cubic spline analysis demonstrated a monotonic increase in predicted risk with increasing PROMISE-ACHD score. The PROMISE-ACHD score outperformed the PEACH score in discrimination (AUC: 0.724 vs 0.658) and demonstrated greater net benefit in decision curve analysis, supporting its clinical utility for individualized preoperative risk assessment in complex ACHD surgery. Abbreviations as in Figures 1, 2, and 5.

• 

  Figure 1. Receiver-Operating Characteristic Curves Comparing the Discriminatory Performance of the PROMISE-ACHD and PErioperative ACHd Scores The apparent area under the curve (AUC) was significantly higher for the PROMISE-ACHD score than for the PEACH score (0.724 vs 0.658; DeLong test P < 0.001). Risk reclassification analyses demonstrated superior discrimination with the PROMISE-ACHD score, including a category-free net reclassification improvement (NRI) of 0.49 (95% CI: 0.33-0.64; P < 0.001), event NRI of 0.09 (95% CI: −0.05 to 0.23), nonevent NRI of 0.40 (95% CI: 0.34-0.46), and an integrated discrimination improvement (IDI) of 0.058 (95% CI: 0.015-0.102; P = 0.01). PEACH = PErioperative ACHd; PROMISE-ACHD = Perioperative Risk Of Morbidity and mortalIty for Surgical Evaluation in complex ACHD.

• 

  Figure 2. Incidence of the Composite Endpoint According to PROMISE-ACHD Score Categories Patients were stratified into low-risk (<3 points), intermediate-risk (3-5 points), and high-risk (>5 points) groups based on the total PROMISE-ACHD score. The incidence of the composite endpoint increased stepwise across categories (8.1%, 18.7%, and 36.7%, respectively; P for trend <0.001). Findings were consistent across prespecified subgroups and when continuous predictors were modeled using restricted cubic splines. PROMISE-ACHD = Perioperative Risk Of Morbidity and mortalIty for Surgical Evaluation in complex ACHD; STS = Society of Thoracic Surgeons.

• 

  Figure 3. Calibration of the PROMISE-ACHD Model for Predicting Perioperative Outcomes The PROMISE-ACHD model demonstrated good calibration and acceptable overall accuracy, with a Hosmer–Lemeshow test P value of 0.711, a Brier score of 0.143, a calibration slope of 1.0, and mean and maximal absolute differences of ±0.02 and ± 0.056, respectively. Internal validation using 1,000 bootstrap resamples demonstrated stable model performance with minimal optimism, supporting the robustness of the model. PROMISE-ACHD = Perioperative Risk Of Morbidity and mortalIty for Surgical Evaluation in complex ACHD.

• 

  Figure 4. Decision Curve Analysis Comparing the Net Clinical Benefit of the PROMISE-ACHD and PErioperative ACHd Models Both models demonstrated greater net benefit than treat-all and treat-none strategies across a wide range of threshold probabilities. The PROMISE-ACHD model provided consistently higher net benefit than the PEACH score between threshold probabilities of 0.10 and 0.50, indicating superior clinical utility across the surgical risk spectrum. Abbreviation as in Figure 1.

• 

  Figure 5. Subgroup Analyses of the Association Between the PROMISE-ACHD Score and the Primary Endpoint Forest plots display the association between the PROMISE-ACHD score and the composite endpoint across major clinical subgroups. Odds ratios and P values for interaction are shown. Associations were generally consistent across subgroups, with significant interactions observed for sex, atrial fibrillation/flutter, and elevated N-terminal pro–brain natriuretic peptide. Associations in patients with lung disease or active infective endocarditis were not statistically significant, likely reflecting true heterogeneity or limited statistical power. ACHD = adult congenital heart disease; ACHS = adult congenital heart surgery; AVVR = atrioventricular valve regurgitation; eGFR = estimated glomerular filtration rate; NT-proBNP = N-terminal pro–brain natriuretic peptide; PROMISE-ACHD = Perioperative Risk Of Morbidity and mortalIty for Surgical Evaluation in complex ACHD.