Author + information
- Istvan Hizoh1,
- Dominika Domokos1,
- Zalan Gulyas1,
- Gyongyver Banhegyi1,
- Zsuzsanna Majoros1,
- Laszlo Major1,
- Timea Ratkai1 and
- Robert Gabor Kiss1
The benefit of transradial primary percutaneous coronary intervention (PPCI) on mortality of patients with ST-segment elevation myocardial infarction (STEMI) has been shown by several trials. Previous risk models have not considered access site as a candidate predictor and most of them were developed using relatively low risk populations of randomized trials. We conducted a prospective cohort study to construct and validate a simple admission risk model for predicting 30-day mortality of STEMI patients undergoing PPCI. We also aimed to create an internet-based risk calculator.
The derivation data set consisted of 750 consecutive patients with a 30-day mortality of 7.6%. We analyzed eight candidate predictors readily available at or soon after presentation: age, gender, ECG localization, onset-to-door time, heart rate, systolic blood pressure (SBP), need for life support on or prior to admission (LS), and access site (AS). To avoid overfitting and determine the best predictors, variable selection was based on backward stepwise logistic regression using 10,000 bootstrap samples. Restricted cubic splines were used to explore presence of non-linear relationships of the continuous predictors to log odds of 30-day mortality. Besides internal verification by bootstrapping, the model has also been validated externally in an independent cohort of 505 patients. Furthermore, we compared the prognostic capacity of the new score with that of previous risk models.
Age (per year, OR: 1.08, p<0.0001, selected in 100.0% of the bootstrap samples), heart rate (per 1/min., OR: 1.04, p<0.0001, 100.0%), LS (OR: 6.24, p<0.0001, 98.1%), AS (radial/femoral, OR: 0.48, p=0.04, 68.0%), and SBP (selected in 99.7% of the samples) were predictive of 30-day mortality. Since SBP has been proved to be non-linearly associated with the logit of the outcome, it was represented by two parameters: SBP (per mmHg, OR: 0.95, p<0.0001) and SBP' (per mmHg, OR: 1.03, p=0.03). ROC curve analysis showed high discriminatory power (apparent c-statistic: 0.88, optimism-corrected c-statistic: 0.87), which was preserved in the validation data set (c-statistic: 0.87). Model fit was good in both cohorts (Hosmer-Lemeshow test, calibration intercept, and slope). For the new risk model the acronym ALPHA (Age, Life support, Pressure, Heart rate, Access site) has been coined. Compared to previous models, the new model achieved the greatest area under the ROC curve (0.87) followed by the CADILLAC model (0.85), the other three risk scoring systems performed somewhat less well (0.82, 0.81, and 0.78 for the TIMI, Zwolle, and PAMI scores, respectively). Both the ALPHA and CADILLAC models predicted better than the PAMI score (ALPHA vs. PAMI: p=0.005; CADILLAC vs. PAMI: p=0.02), the remaining pairwise comparisons revealed no statistically significant differences.
Using this simple tool, mortality risk may be precisely assessed at admission and patients who may benefit most from transradial PCI may be identified.
CORONARY: Acute Myocardial Infarction
Risk model, ST-segment elevation myocardial infarction, Transradial