Cardiac biomarkers predict 1-year mortality in elderly patients undergoing hip fracture surgery

This prospective study included 152 elderly patients (mean age, 80 years; range, 72-88 years) with a hip fracture treated surgically. Comorbidities were evaluated, and B-type natriuretic peptide was measured at baseline and at postoperative days 4 and 5 in addition to troponin I. Major cardiac events were recorded, and 1-year mortality was assessed. Comorbidity models with the important multivariate predictors of 1-year mortality were analyzed. Overall, 9 patients (6%) experienced major cardiac events postoperatively during their hospitalization. Three patients (2%) died postoperatively, at days 5, 7, and 10, from autopsy-confirmed myocardial infarction. Three patients (2%) experienced a nonfatal myocardial infarction, and 3 patients (2%) experienced acute heart failure. At 1-year follow-up, 37 patients (24%) had died. Age older than 80 years (P=.000), renal failure (P=.016), cardiovascular disease (P=.003), respiratory disease (P=.010), Parkinson disease (P=.024), and dementia (P=.000) were univariate predictors of 1-year mortality. However, in the multivariate model, only age older than 80 years (P=.000) and dementia (P=.024) were important predictors of 1-year mortality. In all comorbidity models, age older than 80 years and dementia were important predictors of 1-year mortality. Postoperative increase in B-type natriuretic peptide was the most important predictor of 1-year mortality. Receiver operating characteristic curve analysis showed a threshold of 90 ng/mL of preoperative B-type natriuretic peptide (area under the curve=0.773, 95% confidence interval, 0.691-0.855, P<.001) had 82% sensitivity and 62% specificity to predict 1-year mortality. Similarly, a threshold of 190 ng/mL of postoperative B-type natriuretic peptide (area under the curve=0.753, 95% confidence interval, 0.662-0.844, P<.001) had 70% sensitivity and 77% specificity to predict the study endpoint.