Derivation and validation of a prediction model for long-term mortality in patients with ST-segment elevation myocardial infarction undergoing percutaneous coronary intervention: the PREDICT-ST-Elevation Myocardial Infarction score.
There are limited prediction models of long-term mortality for patients with ST-elevation myocardial infarction (STEMI) undergoing percutaneous coronary intervention (PCI). Against this background, we aimed to derive and validate a predictive model for long-term mortality in patients with STEMI undergoing PCI.
A total of 23 086 patients from a STEMI network were included in the derivation cohort. Using time-to-event regression analysis, predictors of long-term mortality were identified and used to develop a score ranging from 0 to 206 points, with a score directly proportional to the probability of mortality. The predictive performance of this score was then validated in patients from the EXAMINATION-EXTEND study (n=1498) and Coronary Artery diSease Tracking registry (n=1112). An outcome-based cut-point optimisation analysis was performed to determine the best cut-off value in the derivation and validation cohorts.
The prediction model for long-term mortality in STEMI (PREDICT-STEMI) score comprised seven variables: age, diabetes mellitus, previous myocardial infarction, previous ischaemic stroke/transient ischaemic attack, haemodynamic status, three-vessel disease and mechanical circulatory support. The score showed a Concordance index for long-term mortality of 0.81 (95% CI 0.80 to 0.81) in the derivation and 0.81 (95% CI 0.78 to 0.84) and 0.84 (95% CI 0.81 to 0.87) in the validation cohorts, respectively. The optimal prediction model cut-off was 60 points; compared with those with a low score, patients with a high score had a sixfold increased risk of long-term mortality in both the derivation and validation cohorts.
The PREDICT-STEMI score is a simple tool for predicting long-term mortality and facilitating early risk stratification and inform clinical decision-making.
A total of 23 086 patients from a STEMI network were included in the derivation cohort. Using time-to-event regression analysis, predictors of long-term mortality were identified and used to develop a score ranging from 0 to 206 points, with a score directly proportional to the probability of mortality. The predictive performance of this score was then validated in patients from the EXAMINATION-EXTEND study (n=1498) and Coronary Artery diSease Tracking registry (n=1112). An outcome-based cut-point optimisation analysis was performed to determine the best cut-off value in the derivation and validation cohorts.
The prediction model for long-term mortality in STEMI (PREDICT-STEMI) score comprised seven variables: age, diabetes mellitus, previous myocardial infarction, previous ischaemic stroke/transient ischaemic attack, haemodynamic status, three-vessel disease and mechanical circulatory support. The score showed a Concordance index for long-term mortality of 0.81 (95% CI 0.80 to 0.81) in the derivation and 0.81 (95% CI 0.78 to 0.84) and 0.84 (95% CI 0.81 to 0.87) in the validation cohorts, respectively. The optimal prediction model cut-off was 60 points; compared with those with a low score, patients with a high score had a sixfold increased risk of long-term mortality in both the derivation and validation cohorts.
The PREDICT-STEMI score is a simple tool for predicting long-term mortality and facilitating early risk stratification and inform clinical decision-making.
Authors
Ortega-Paz Ortega-Paz, Laudani Laudani, Brugaletta Brugaletta, Vidal-Calés Vidal-Calés, Ariza-Solé Ariza-Solé, Buera Buera, Cárdenas Manilla Cárdenas Manilla, Carrillo Carrillo, Jimenez Jimenez, Guarinos Guarinos, Muñoz Camacho Muñoz Camacho, De Diego De Diego, Tizón-Marcos Tizón-Marcos, Tomás-Querol Tomás-Querol, Gomez-Lara Gomez-Lara, Jimenez-Diaz Jimenez-Diaz, Jiménez-Quevedo Jiménez-Quevedo, Diletti Diletti, Bordes Bordes, Campo Campo, Gomez-Hospital Gomez-Hospital, Greco Greco, Spagnolo Spagnolo, Capodanno Capodanno, Angiolillo Angiolillo, Sabate Sabate
View on Pubmed