Reproducibility of noninvasive vascular laboratory measurements of the peripheral circulation.

This study reports the reproducibility of noninvasive vascular laboratory measurements for the diagnosis of peripheral arterial occlusive disease and compares their variability to other clinical measurements, including pulse rate, hemoglobin, white blood cell count, blood urea nitrogen, and creatinine. To study the reproducibility of these measurements, we considered three components that affect the repeatability; the variation associated with the measurement process, the variation associated with transient changing characteristics of the patients with time, and the variation among the patients. It is useful to consider the variation of each component relative to the variation among the patients, measured in units of standard deviation. The variation associated with the measurement process (expressed as the relative precision) was determined by repeating the measurements in pairs at random times throughout the study period. The relative precision was 0.3% for the measurements of ankle-arm systolic blood pressure ratio and treadmill walking time, 2.7% for pulsatility index, and 1.2% to 4.2% for the other clinical, hematologic, and biochemical measurements. The temporal variations were determined by repeating the measurements over a 6-week period in 15 patients. The normalized long-term fluctuation was 1.7% for ankle-arm systolic pressure ratio, 3.2% for treadmill walking time, 7.5% for pulsatility index, and 1.3% to 5.5% for the other measurements. For each test, the 95% confidence interval for an individual patient's measurements and the values above which the difference between two successive measurements can be considered significantly different (95%) have been calculated. It is concluded that the reproducibility of standard vascular laboratory measurements compares favorably to the other clinical, hematologic, and biochemical measurements evaluated in this study.