Effect of venesection in high-altitude residents with CMS. (A-I) Open bar represents 500-mL isovolemic venesection on each of 4 consecutive days (days 1-4; total volume, 2000 mL). Hematocrit was estimated by microcentrifugation (mean of 2 measurements). Arterial oxygen saturation was measured by pulse oximetry (Nonin Onyx; Nonin Medical, Plymouth, MN). Serum for analysis of ferritin, iron, and total iron-binding capacity was stored at 4°C and analyzed within 72 hours at sea level (Medlab, Lima, Peru). Transferrin saturation was calculated as 100(serum iron/total iron-binding capacity). Serum was stored at −20°C for EPO enzyme-linked immunosorbent assay (ELISA) (Medlab). Plasma was stored at −20°C for hepcidin (Bachem, St Helens, United Kingdom), GDF-15 and sTfR (R&D Systems, Abingdon, United Kingdom), and FAM132B ELISAs. In the case of FAM132B, samples were assayed using 2 independent ELISA kits (Cusabio Biotech, Wuhan, China, and Aviscera Bioscience, Santa Clara, CA). Results are provided for the Cusabio Biotech assay because the detection range most closely matched the range of FAM132B values in our samples, but neither assay revealed any evidence of an effect of venesection on FAM132B. Individual time points (mean ± standard error of the mean) were compared with baseline (day 0) using paired Student t tests. *Statistically significant difference (P < .05).