TRACKING CENTRAL HYPOVOLEMIA WITH ECG IN HUMANS

Abstract

Heart period variability (HPV) metrics have been suggested for use in medical monitoring of trauma patients. This study sought to ascertain the use of various HPV metrics in tracking central blood volume during simulated hemorrhage in individual humans. One hundred one healthy nonsmoking volunteers (58 men, 43 women) were instrumented for continuous measurement of electrocardiogram and beat-by-beat finger arterial blood pressure. Stroke volume (SV) was estimated from the arterial pulse wave and used to reflect central blood volume. Progressive lower body negative pressure (LBNP) was applied in 5-min stages until the onset of impending hemodynamic decompensation (systolic blood pressure <70 mmHg and/or presyncopal symptoms). HPV was assessed with analysis of R-to-R intervals using both linear (time and frequency domains) and nonlinear (e.g., complexity, fractality) methods. Application of increasing LBNP caused progressive reductions of SV, whereas arterial pressures changed only minimally and late. Group LBNP stage means for each HPV metric changed progressively and were strongly correlated with the mean decrease in SV (|r| > or = 0.87). To ascertain the utility of the HPV metrics to track individual responses to central hypovolemia, the difference scores for each HPV metric were correlated at each successive LBNP level, with percentage change in SV at the subject level. This cross-correlation of difference scores revealed that none of the HPV metrics showed strong and consistent correlations (|r| < or = 0.49) with percentage change in SV across successive LBNP levels. Although aggregate group mean values for HPV metrics are well correlated with SV changes during central hypovolemia, these metrics are less reliable when tracking individual reductions in central volume during LBNP. HPV metrics, therefore, may not be useful in monitoring hemorrhagic injuries in individual patients.