The impact of methodologic differences on low frequency blood pressure variability assessments: Implications for individuals with spinal cord injuries

PURPOSE To compare differences in low frequency (LF) blood pressure variability (BPV) power with spontaneous vs paced breathing and with autoregressive analysis vs Welch periodograms. Applying these methods to individuals with spinal cord injury (SCI), we aimed to assess the stability of "autonomically complete" LF BPV threshold classifications over repeated measures and compare these thresholds to another measure of dynamic sympathetic activation, pressure recovery time following Valsalva maneuvers. METHODS Five-minute supine continual blood pressures were recorded with spontaneous and paced (0.25 Hz) breathing. Three 15-second Valsalva maneuvers were completed to calculate pressure recovery time. RESULTS Data from 83 individuals with SCI and 59 uninjured controls demonstrated a moderately strong (R2 = 0.29) proportional bias on Bland-Altman analysis, with larger LF BPV powers being increasingly inaccurate with spontaneous vs paced breathing. Welch periodograms generated significantly lower LF BPV powers compared to 12th order autoregressive models (p = 0.001), though this had little bearing on threshold classification. Across 42 individuals with SCI, threshold measurements of LF BPV power demonstrated poor repeatability with a 55% coefficient of variation and individuals vacillating from autonomically complete to incomplete (or the inverse) in 15% of tests. In 24 individuals with SCI, abnormal pressure recovery time poorly correlated to LF BPV power (R2 = 0.04-0.06) and correctly corresponded to LF BPV defined autonomically complete SCI in only 42-58% of tests. CONCLUSION Measurements of LF BPV power are significantly impacted by testing conditions and analytic techniques. While refinement is possible, more comprehensive measures of autonomic completeness after SCI are likely needed.