Heart rate dynamics predict poststroke mortality

Background: Impaired cardiovascular autonomic regulation assessed by heart rate (HR) variability provides prognostic information in patients with heart disease as well as among elderly. Reduced HR variability has been described in stroke patients, but the prognostic significance of HR variability measures after stroke has not been studied. Methods: A series of 84 patients (mean age 59 +/- 12years) with an acute first-ever ischemic stroke underwent a comprehensive clinical investigation, laboratory tests, and 24-hour EKG recordings and were followed up for 7 years (mean 83 +/- 54 months). Various conventional and newer qualitative measures of HR variability were analyzed from the baseline 24-hour EKG. Risk factors for all-cause mortality were assessed. Results: During the follow-up, 33 (39%) patients died and 51 survived. Among all the variables analyzed, abnormal long-term HR variability measure power-law slope beta (beta < -1.5), reflecting an altered distribution of spectral characteristics over ultra and very low frequency bands, was the best univariate predictor of death ( hazard ratio 4.5, 95% CI 2.2 to 9.5, p < 0.001). High age, poor Scandinavian Stroke Scale score, and abnormal short-term HR variability scaling measure (alpha) also predicted mortality in univariate analysis. In multivariate analysis, after adjustment for age, the only independent predictor of the risk of death was abnormal long-term power-law slope beta ( hazard ratio 3.8, CI 1.8 to 8.2, p < 0.001). Conventional HR variability measures showed no prognostic power. Conclusion: Abnormal long-term HR dynamics predict poststroke mortality. This measure may have value in the risk stratification of stroke patients.