Minimum Detectable Change in Arterial Pulse Wave Conduction Time From Heart to Toe as a Function of the Number of Cardiac Pulses Included: A Prospective Noninterventional Study

Background and objectiveNoninvasive assessments of arterial stiffness are used to evaluate a patient's vascular age and predict the risk of future cardiovascular events across various clinical conditions.These predictions are based on pulse wave velocity (PWV) measurements because PWV increases with increased arterial vessel stiffness.PWV is calculated from the measured pulse wave conduction time (PWCT) between two sites a known distance apart along an arterial pathway.Another application of PWV determinations is to assess an intervention's effect on vascular stiffness by measuring changes in PWV.For this purpose, an estimate of the minimum detectable change (MDC) in measured PWCT is needed, which itself depends on the number of pulses used (n) to estimate PWCT.The objective of this study was to provide initial quantitative estimates of the minimum detectable percentage change (MDC%) and its dependence on n. MethodsEight young adults with an age of 26.14 ± 0.83 years and a BMI of 23.20 ± 1.78 kg/m² participated after signing an institutional review board-approved consent form.With the subject supine, a photoplethysmographic (PPG) pulse was recorded from the great toe of one foot while an EKG was recorded continuously at 1000 samples/s for 40 minutes.For each sequential pulse pair, the PWCT was calculated as the measured time difference between the EKG R-wave peak and the PPG pulse arrival at the great toe, with pulse arrival determined by the maximum of the second derivative of the toe's PPG signal.This process produced N sequential PWCT values, where N depended on each subject's EKG RR interval over the 40minute recording interval.For each subject's PWCT time series, an MDC% was determined for p-values of 0.05, 0.01, and 0.001.The MDC dependence on n was determined by calculating the MDCs for n values of 5, 10, 20, 40, 50, and 100.The MDC% was calculated using the equation (t-value × √2 × SD), where t-value is the two-sided Student's t-value for the n used, and SD is the standard deviation. ResultsThe principal finding for the estimated MDC% was that MDC% for a given p-value decreased nonlinearly as the number of pulses (n) used to estimate MDC% increased.A key feature of this relationship was that, for any p-value, MDC% improves only slightly for sample sizes greater than 40-50 pulses.At a sample size of n = 50, the MDC% ranges from 5.8% to 10.0% for p-values of 0.05 and 0.001, respectively. ConclusionsThese findings offer initial guidance on the number of pulses to use when investigating a specific intervention's short-term effect on PWCT and, by extension, PWV estimates.Because the absolute values of the estimated MDC% come from a small group of healthy young adults, the determined MDC% are likely lower bounds for what might be observed in elderly patients with medical comorbidities or if the time between assessments is extensive.However, because the number of pulses needed to accurately estimate the MDC% was based on a large number of sequential PWCTs for each subject, the recommended 40-50 pulses should not be significantly affected, but further study of this relationship in elderly populations is warranted.