Reproducibility of the ambulatory arterial stiffness index in hypertensive patients

Background We studied the repeatability of the ambulatory arterial stiffness index (AASI), which can be computed from 24-h blood pressure (BP) recordings as unity minus the regression slope of diastolic on systolic BP. Methods One hundred and fifty-two hypertensive outpatients recruited in Nijmegen (mean age = 46.2 years; 76.3% with systolic and diastolic hypertension) and 145 patients enrolled in the Systolic Hypertension in Europe (Syst-Eur) trial (71.0 years) underwent 24-h BP monitoring at a median interval of 8 and 31 days, respectively. We used the repeatability coefficient, which is twice the SD of the within-participant differences between repeat recordings, and expressed it as a percentage of four times the SD of the mean of the paired measurements. Results Mean AASI (crude or derived by time-weighted or robust regression) and 24-h pulse pressure (PP) were similar on repeat recordings in both cohorts. In Nijmegen patients, repeatability coefficients of AASI and PP were similar to 50%. In Syst-Eur trial patients, repeatability coefficient was similar to 60% for AASI and similar to 40% for PP. For comparison, repeatability coefficients for 24-h systolic and diastolic BP were similar to 30%. Differences in AASI between paired recordings were correlated with differences in the goodness of fit (r(2)) of the AASI regression line as well as with differences in the night-to-day BP ratio. However, in sensitivity analyses stratified for type of hypertension, r(2), or dipping status, repeatability coefficients for AASI did not widely depart from 50 to 60% range. Conclusion Estimates of mean AASI were not different between repeat recordings, and repeatability coefficients were within the 50-60% range.