Age-adjustment of HRV measures and its prognostic value for risk assessment in patients late after myocardial infarction

Abstract

The aims of our study were: (1) to establish the normal limit of the heart rate variability (HRV) indices in a healthy population and in its four age-related subgroups, including a new HRV index, HRV fraction; and (2) to analyse the frequency and predictive value of abnormally low HRV in a population of post-infarction patients in respect to the cut-points chosen (raw or age-adjusted). Methods: Normal population of 296 healthy subjects (81 f, 215 m, aged 47±10 years) and post-infarction population of 298 patients (>3 months after acute MI, 65 f, 233 m, aged 56±10) were examined. The normal population was divided into 4 subsets based on age at entry: <35, 35–44, 45–54 and >54 years. Based on a 24 h ECG the standard HRV analysis was performed to obtain the following indices: mean RR interval, SDNN and SDANN. A new index of HRV, HRV fraction (HRVF, %) was calculated based on a numerical processing of the RR intervals return map. All patients were followed for 24 months. The endpoints of the study were death (of any reason) and cardiac death. Results: Means and normal limits for SDNN, SDANN and HRVF were: 147±36 ms [95% CI 89–220], 136±36 ms [79–212] and 53±9% [35–68]. The HRV values below the lower normal limit (LNL) were observed in 20–25% of post-MI patients. During a 2 year follow-up there were 36 deaths (total mortality 12.1%), while cardiac mortality was 9.1% (27 cases). The prognostic value of the analysed indices was similar (sens ∼53–61%, spec ∼79–84%, PPV 22–26%, NPV 93–94%) irrespective of the cut-points chosen (calculated either for the entire population or age-related). Multivariate Cox regression analysis showed that a decrease of any index below the LNL was associated with a ∼2.5 and ∼4–6 times greater risk for death and cardiac death, irrespective of the cut-points chosen. Conclusions: The age-dependence of the HRV indices does not seem to significantly influence their prognostic value. Thus, a single cut-point of a particular HRV index, based on the entire population, is sufficient to be treated as a risk predictor. In the late phase of myocardial infarction the value of any global HRV index lying below the lower normal limit indicates independently an increased risk of death, especially cardiac death. The new index (HRV fraction) seems to be a promising substitute for currently used standard indices.