Estimation of the cut-off value in cardiovascular autonomic nervous function tests: not-age-related criteria or the age-related 5th percentile

Abstract

The aim of the study was to estimate the cut-off value of cardiovascular tests using not-age-related criteria or the age-related 5th percentile in 165 diabetic patients. The cANP, pANP, and smNP were assessed using previously recommended standardized test procedures. Prevalence of overall definite (borderline) cANP was 35.2% (23.0%) when using the 5th percentile or 16.4% (27.9%) when using not age-related criteria (p for the difference < 0.0001) (p = 0.3205). Prevalence of pANP was 54.0% and of smNP 37.0%. Concerning cANP, the number of test results below the 5th percentile (N5) correlated significantly with the number of abnormal test results using not age-related criteria (Nnar) (p < 0.000001) but the slope of the regression line differed substantially from 1. Using the 5th percentile as the cut-off value for cANP testing, sensitivity and specificity were calculated for not age-related criteria for respiratory sinus arrhythmia (90.5%, 80.2%), Valsalva test (30.4%, 98.0%), lying-to-standing test (46.3%, 98.7%), orthostatic systolic blood pressure fall (69.0%, 78.3%), and overall definite cANP (44.8%, 92.8%). The statistical analysis revealed that the age-related 5th percentile is superior to the not-age-related cut-off values in diabetic patients. We therefore suggest that age-related normal values (percentiles) have to be applied when cANP is estimated.