Abstract 3582: A Risk Score to Predict In-hospital Mortality Following Primary Angioplasty from New York State Reporting System
Background: Although several risk scores exist that risk-stratify patients during elective percutaneous coronary intervention (PCI), few have been developed in primary angioplasty setting. Therefore, we developed a primary angioplasty risk score using the 2000 –2002 New York State PCI reporting system.
Methods: 7321 patients who had primary angioplasty were divided into model development (N=3706) & validation cohorts (N=2366). Stepwise logistic regression was performed to identify independent predictors of in-hospital mortality(pr= 0.2; pe=0.1). In the final model, independent predictors with p<<26>0.05 were retained. The point risk score was developed as the ratio of regression coefficient distance between variable and base category divided by the constant (risk for each 5-yr higher age=0.31).
Results: Risk factors identified with corresponding point scores were Age (80yrs =5); Prior stroke(3); COPD(2); Current heart failure(2); Prior heart failure(4); Ejection fraction(>45%=0;35–45%=2;25–35%=3);Cr> 2.5mg/dl(7), Hemodynamic instability(4), CPR(6), cardiogenic shock(8) & multivessel disease(1). The overall mortality in the development & validation datasets was 3.5% & 2.5% respectively. The risk score was categorized into following classes: low (≤ 5), intermediate (6 –10), high (11–15) and very high risk (>15). Significant correspondence between the observed and predicted mortality in the validation and development cohorts was seen (Figure⇓)
Conclusion: This novel risk score identifies patients at variable risk following primary angioplasty. If validated, it could be a useful adjunct during decision making in the setting of primary angioplasty.