RESUMO
Endothelial function at the arterial level has been extensively assessed by a non-invasive method using flow-mediated dilatation (FMD) of the brachial artery. Early disturbances have been found in patient groups prone to later development of manifest macrovascular atherosclerosis. A possible non-invasive means of studying blood-flow regulation and function at the microcirculatory level is through videophotometric capillaroscopy. The most stable variable in such an investigation is the time-to-peak (TtP) flow after a brief arterial occlusion. The short-term reproducibility of such assessments is excellent but the coefficient of variation (CV) in long-term studies is reported to be in the order of slightly less than 20%. The aim of the present methodological study was to evaluate different sources of variations in such microcirculatory assessments in order to be able to propose design recommendations that minimize the number of patients and recordings needed to achieve sufficient statistical power in longitudinal studies. We used a symmetric design with 144 recordings of TtP after a one-minute arterial occlusion in healthy volunteers. We did six occlusions each time in the capillaries of two fingers on each occasion, and repeated the procedure three times with an interval of at least one week between each investigation. All recordings were analyzed off-line using a cross-correlation technique with the Capiflow system. Each analysis was performed at least three times, giving a total of slightly less than 500 assessments. In our material (n = 10) TtP had a mean of 6.3 s (95% confidence interval 5.2-7.4). The correlation between repeated measurements in a single capillary during a single session was r > 0.91 (CV 6%). The between-finger CV was 8% (r = 0.84). The CV of measurements between different days was about 20% when single measurements were compared. However, the CV decreased to less than 13% when the mean of at least two time-to-peak assessments on each occasion was used. In conclusion, the methodological error including day-to-day variation could be minimized using the mean of at least two repeated assessments of post-occlusive hyperemia at each time point in a longitudinal study. This finding should be taken into consideration in the design of future longitudinal studies.