Repeatability of adaptive traits among ethnic Tibetan highlanders.

Objectives Connecting traits to biological pathways and genes relies on stable observations. Researchers typically determine traits once, expecting careful study protocols to yield measurements free of noise. This report examines that expectation with test-retest repeatability analyses for traits used regularly in research on adaptation to high-altitude hypoxia, often in settings without climate control. Methods Two hundred ninety-one ethnic Tibetan women residing from 3500 to 4200 m in Upper Mustang District, Nepal, provided three observations of hemoglobin concentration, percent of oxygen saturation of hemoglobin, and pulse by noninvasive pulse oximetry under conditions designed to minimize environmental noise. Results High-intraclass correlation coefficients and low within-subject coefficients of variation reflected consistent measurements. Percent of oxygen saturation had the highest intraclass correlation coefficient and the smallest within-subject coefficient of variability; measurement noise occurred mainly in the lower values. Hemoglobin concentration and pulse presented slightly higher within-subject coefficients of variation; measurement noise occurred across the range of values. The women had performed the same measurements 7 years earlier using the same devices and protocol. The sample means and SD observed across 7 years differed little. Hemoglobin concentration increased substantially after menopause. Conclusions Analyzing repeatability features of traits may improve our interpretation of statistical analyses and detection of variation from measurement or biology. The high levels of measurement repeatability and biological stability support the continued use of these robust traits for investigating human adaptation in this altitude range.