Prediction of one-year survival in high-risk patients with acute coronary syndromes: results from the SYNERGY trial

Abstract

Background: Despite advances in pharmacologic therapy and invasive management strategies for patients with non-ST-segment elevation acute coronary syndromes (NSTE ACS), these patients still suffer substantial morbidity and mortality.

Objective: The objective of this study was to analyze independent predictors of 1-year mortality in patients with high-risk NSTE ACS.

Design and participants: A total of 9,978 patients were assigned to receive enoxaparin or unfractionated heparin (UFH) in this prospective, randomized, open-label, international trial.

Measurements: Vital status at 1 year was collected. Univariable and multivariable predictors of 1-year mortality were identified. Three different multivariable regression models were constructed to identify: (1) predictors of 30-day mortality; (2) predictors of 1-year mortality; (3) predictors of 1-year mortality in 30-day survivors. The last model is the focus of this paper.

Results: Overall, 9,922 (99.4%) of patients had 1-year follow-up. Of the 56 patients (37 UFH-assigned and 19 enoxaparin-assigned) without 1-year data, 11 patients were excluded because of withdrawal of consent, and 45 could not be located. One-year mortality was 7.5% (7.7% enoxaparin-assigned patients; 7.3% UFH-assigned patients; P = 0.4). In patients surviving 30 days after enrollment, independent predictors of 1-year mortality included factors known at baseline such as increased age, male sex, decreased weight, having ever smoked, decreased creatinine clearance, ST-segment depression, history of diabetes, history of angina, congestive heart failure, coronary artery bypass grafting, increased heart rate, rales, increased hematocrit, lowered hemoglobin, and higher platelet count. Factors predictive of mortality during the hospitalization and 30-day follow-up period were decreased weight at 30 days from baseline, atrial fibrillation, decreased nadir platelet, no use of beta-blockers and statins up to 30 days, and not receiving an intervention (c-index = 0.82).

Conclusions: Easily determined baseline clinical characteristics can be used to predict 1-year mortality with reasonable discriminative power. These models corroborate prior work in a contemporary aggressively managed population. A model to predict 1-year mortality in patients surviving at least 30 days may be quite helpful to healthcare providers in setting expectations and goals with patients after ACS.

Figures

Figure 1

Kaplan–Meier curves for survival from death through 1-year follow-up by treatment.

Figure 2

Nomogram for 30-day to 1-year mortality (reduced model). CABG Coronary artery bypass grafting, Hgb hemoglobin. In Figure 2, find the value most closely matching the patient’s risk factors and circle the corresponding point assignment. Sum the points for all predictive factors. Then use Figure 3 to determine probability of death from 30 days to 1 year after randomization. Example: A 70-year-old male with baseline creatinine clearance of 80, weight of 80 kg, a baseline hemoglobin of 12, a baseline platelet count of 300, and a nadir platelet count of 200, who had a PCI on day 2 of his hospitalization and was discharged from the hospital without a statin would have a total score of [13(age) + 20(male sex) + 15(CrCl) + 5(wt) + 22(Hgb) + 13(baseline platelet) + 0(nadir platelet) + 0(afib) + 17(statin) + 20(CABG)] = 125. This score corresponds to a predicted probability of death at 1-year follow-up of ∼10% in patients who survived to 30 days.

Figure 3

Plot of mortality rates associated with scores calculated from the nomogram in Figure 2. Table shows the predicted values for common point tallies.