Abstract 3792: Baseline Physical Function Is An Independent Predictor Of Long Term Mortality After Myocardial Infarction: A Ten-year Study
Background: Physical function, as measured by the Short Form-36 (SF-36) Physical Composite Score (PCS), is a predictor of mortality in heart failure, but its utility in myocardial infarction (MI) is unknown.
Methods: MI patients admitted from July 1995 to December 1996 completed the SF-36 during hospitalization. A cardiologist blind to the PCS evaluated clinical variables using standard criteria. All-cause mortality was assessed in May 2007 using the Social Security Death Index.
Results: Among the 273 patients (57% men, mean age 64.8 ± 12.1 years at the time of enrollment), lower PCS (worse physical function) was seen in women, older patients and those with hypertension, diabetes, previous MI, higher Killip class, non Q-wave MI, renal insufficiency, and longer hospital stay. There were 32, 96, and 152 deaths at 1, 5, and 10 years post-MI. Higher baseline PCS (Quartile 4, Figure⇓) was associated with lower mortality. In Cox proportional hazards models, higher PCS was associated with lower mortality at 1 year [HR=0.95 (0.92– 0.97), p<0.01], 5 years [HR=0.96 (0.95– 0.98), p<0.01], and 10 years [HR=0.96 (0.95– 0.97), p<0.01]. This association was independent of demographics, co-morbidities, ejection fraction, MI severity and MI treatment. PCS alone was a good predictor of mortality at 1 year, 5 and 10 years (area under receiver-operator curve 0.75, 0.70, 0.67 respectively), although much of the predictive value of PCS was due to differences in mortality in the first 3 years after MI.
Conclusions: The SF-36 PCS obtained during hospitalization for MI is independently associated with long-term mortality. Including the SF-36 PCS in clinical assessments may improve risk stratification in MI.