Solomon1  claims that many patients with clinical features suggestive of cobalamin (Cbl) deficiency who have normal serum Cbl, methylmalonic acid (MMA), and homocysteine (HCys) levels may respond to Cbl therapy, while many patients with low serum Cbl, high serum MMA, or high serum HCys levels may fail to respond to Cbl therapy. He concludes that these tests are unreliable for the diagnosis of Cbl deficiency. I do not think that there is much support for his conclusion.

First of all, Solomon writes that resolution of signs and symptoms consistent with Cbl deficiency in 12 patients occurred prior to any Cbl therapy. How can he then attribute to Cbl therapy the resolution of signs and symptoms of those who received Cbl?

Second, the response to pharmacologic doses of Cbl does not support the diagnosis of Cbl deficiency. The dose of Cbl for a therapeutic trial of Cbl deficiency is 1 μg Cbl daily for 10 days. This dose would produce optimal reticulocyte response within 7 to 10 days and, if continued, would produce a complete hematologic response.2  Patients who do not respond to small doses of Cbl but respond to pharmacologic doses of Cbl are not Cbl deficient.

Third, the criteria that Solomon has used for demonstrating a response to Cbl therapy (5 fL reduction of mean corpuscular volume [MCV] or an increase in hematocrit of 0.05 point within 3 months of Cbl therapy) are nonspecific. Such a response can be seen in a patient who has folate-deficiency anemia treated with high doses of Cbl.2  Similar hematologic changes may be seen in any patient with alcoholic liver disease or ethanol abuse who stops drinking ethanol; in patients recovering from acute hemolytic anemia, anemia associated with acute infection; or in patients with hypothyroidism treated with thyroid extract, irrespective of whether they are given Cbl.

Fourth, of 8 patients who met Solomon's criteria for hematologic response, only 2 demonstrated an increase in hematocrit within 3 months of Cbl therapy. The other 6 patients showed only reduction of MCV. If the reduction of MCV were due to Cbl therapy, why did the hematocrit not increase?

Finally, Solomon reports that low or intermediate Cbl levels were present in 46% of the responders and 56% of the nonresponders, and increased MMA values were present in 73% of responders and 88% of nonresponders. Consequently, he claims that these tests are unreliable. In the presence of low serum Cbl or high serum MMA, failure of response to Cbl does not exclude Cbl deficiency, but indicates that the anemia or neurologic abnormalities were not due to Cbl deficiency.

Serum Cbl, MMA, and HCys, like any other laboratory tests, may give false-positive or false-negative results in certain situations. One has to be familiar with these situations in order to interpret the laboratory results properly rather than claiming that these tests are unreliable.

Solomon LR. Cobalamin-responsive disorders in the ambulatory care setting: unreliability of cobalamin, methylmalonic acid, and homocysteine testing.
Herbert V. Megaloblastic anemia.
N Engl J Med