et al recently presented in this journal a novel parameter of heart rate variability called prevalent low-frequency oscillation of heart rate (PLF). 1 Using 2 groups of postinfarction patients derived from the EMIAT (European Myocardial Infarction Amiodarone Trial) and ATRAMI (Autonomic Tone and Reflexes After Myocardial Infarction) studies, they specifically compared and showed a predictive value for PLF that was stronger than that for heart rate turbulence (HRT) parameters using univariate analysis. They further showed that with Cox's multivariate regression model, the relative risk for PLF remained high (4.6 and 3.6 in EMIAT and ATRAMI, respectively), whereas HRT parameters turbulence slope and turbulence onset lost their independence. Their study is noteworthy because in previous comparisons of measures of HRT and heart rate variability in large studies, HRT has consistently outperformed heart rate variability. [2][3][4] We would like to point out that the prognostic ability of risk factors is changing significantly with declining mortality from acute ischemic events as a result of improved treatment protocols, and that the superiority of PLF may not hold true in a contemporary population. For example, we found that late potentials lost their predictive ability entirely in a modern population of patients, 90% of whom received percutaneous coronary intervention and 84% received statins. 5 In contrast, the hazard ratio of HRT in multivariate analysis improved from 3.2 in the EMIAT placebo arm population, in whom only 60% of the patients had received thrombolysis, 2 to 5.9 in the modern population. 4 There were also some methodological differences in the analysis of HRT between the PLF article and previous HRT publications. For example, combining turbulence slope and turbulence onset usually provides the best predictive ability. 2,3 Wichterle et al did not show how the combined HRT parameters performed against PLF. The end point for the ATRAMI population was cardiac mortality or resuscitated cardiac arrest, in contrast to all-cause mortality, which is commonly used in HRT studies. Finally, it is unclear why Wichterle et al used different criteria for admissible Holter recordings for their 2 patient populations based on the number of ventricular premature complexes-a minimum of 1 in the ATRAMI versus 5 for the EMIAT. It is generally accepted that a larger number of ventricular premature complexes for HRT calculation produces more stable values.
Mari