Using a commercial kit method we obtained a vitamin D metabolite level within the normal range in a patient with biopsy-proven osteomalacia. This suggested that the ethanol extraction method employed had not removed serum factors known to falsely elevate the measurement of vitamin D metabolites. We therefore compared the levels measured after ethanol extraction and after purification by chromatography on Sep-pak C 18 or Sephadex LH-20. Vitamin D metabolite levels after ethanol extraction of sera correlated with, but were higher than, those after Chromatography on Sep-pak C 18 cartridges ( r = 0.84; 134 ± 76 vs 76 ± 46 nmol/l: mean ± Sd; p < 0.01). Results were similar after Chromatography on Sep-pak C 18 and Sephadex LH-20 ( r = 0.95; 79 ± 46 vs 68 ± 41 nmol/l). Sera from 5 patients (4 with osteomalacia, 1 with chronic pancreatitis/malabsorption) had vitamin D metabolite levels in the normal range after ethanol extraction but had low levels after Sep-pak C 18 Chromatography; four of these sera also had low levels after Chromatography on Sephadex LH-20. These findings indicate that Chromatography of serum on Sep-pak C 18 cartridges corrects falsely elevated vitamin D metabolite levels measured by protein binding assay.