Subtle cobalamin deficiency states, where low serum cobalamin levels are not accompanied by megaloblastic anemia or malabsorption of free cobalamin, often display metabolic evidence of cellular depletion as shown by the deoxyuridine suppression test. However, the suppression test abnormalities are usually mild and are sometimes atypical; moreover, their response to cobalamin therapy has never been documented. Four patients with this subtle defect, at least three of whom had food-cobalamin malabsorption, were therefore tested before and after cobalamin treatment. Each patient had low serum cobalamin levels but did not have megaloblastic anemia, and all but one had normal serum levels of methylmalonic acid and total homocysteine. Two patients had mildly but typically cobalamin-deficient deoxyuridine suppression test results (baseline values 15.7% and 12.8%; normal less than 8.5%). The other two patients had normal or borderline baseline values (5.4% and 8.9%) that became abnormal on incubation with methyl tetrahydrofolate (16.1% and 12.3%), a pattern previously noted in subtle acquired and hereditary cobalamin deficiencies. After 6 months of cobalamin therapy, the deoxyuridine suppression test abnormalities reversed in all four patients. These findings show that the mild deoxyuridine suppression test stigmata of subtle cobalamin deficiency respond to therapy and thus represent true metabolic deficiency; the unusual abnormality induced in vitro by added methyl tetrahydrofolate responds as well, indicating that it, too, represents metabolic cobalamin deficiency. The findings provide further proof that subtle cobalamin deficiency often exists even when megaloblastic anemia and malabsorption of free cobalamin are lacking, and that the deoxyuridine suppression test can be a reliable tool for its identification.