[Simple cardiovascular reflex tests in prediction of sudden death after myocardial infarction: which method to...

Some evidence exists that heart rate response to simple provocative maneuvers may predict sudden cardiac death (SD) after myocardial infarction (MI). However optimal test has not been not established yet. Aim of this study was to compare prognostic value of different noninvasive reflex tests after MI.

Four reflex tests were consecutively performed in 188 patients on days 4-11 of MI (68% men, age 34-75 years, 93.6% on beta-blockers, without heart failure NYHA IV on the day of tests). Time- and frequency domain heart rate variability measures were obtained during 5 min at active standing and at bed rest with controlled breathing 6 and 15 per minute. In addition difference between average maximal and minimal heart rate at first minute of breathing 6 per minute (HRD) and Valsalva ratio (VR) were calculated. ROC analysis was used to determine cut-off values of studied measures for dichotomization of patients into those with low- and high-risk of SD and these values were used in logistic regression analysis.

During follow up for 2.1+/-0.8 years there were 9 SD. Univariate predictors of SD were follows: HRD <3.36; VR <1.13; pNN 50 <2.5, total spectral power <1021 ms(2), LF power <229 ms(2) and HF power <65 ms(2) at active standing; pNN 50 <2.3, LF power <129 ms(2) and HF power <111 ms(2) during controlled breathing 15 per minute. HF power <65 ms(2) during active standing (OR 28.8, 95% CI 4.1-104.2; p=0.0001, positive predictive value 29.4%) and VR <1.13 (OR 6.0, 95% CI 1.02-34.3; p=0.04, positive predictive value 11.5%) were independent predictors of SD. For combination of these parameters OR increased to 34.9 (95% CI 6.7-181.6; p<0.001), positive predictive value to 50%.

Among simple noninvasive reflex tests in this small group of patients with routine beta-blockers use and without severe heart failure active standing with calculation of HF power seems preferable method for prediction of SD after MI. Its predictive value may be enhanced by combination with Valsalva ratio.