Vitamin B₁₂ and homocysteine

In their analysis of the relationship between carotid plaque area and vitamin B12 status,1 Julie Robertson and colleagues define vitamin B12 deficiency as a serum concentration of less than 258 pmol/L, with a plasma homocysteine concentration of 14 μmol/L or more or a plasma methylmalonic acid (MMA) level of 271 nmol/L or more. We question the use of these thresholds and the conclusions that are based on them.

Their threshold for vitamin B₁₂ deficiency of 258 pmol/L is almost the same as the sample median value of 253 pmol/L . This value is exceedingly high and probably inappropriately labeled many study patients as having a vitamin B12 deficiency. Among 11 000 elderly women in Ontario and British Columbia, the mean and fifth centile serum vitamin B12concentrations were 300 and 118 pmol/L, respectively, after folic acid food fortification.2 Similar values were found among Ontario women of reproductive age.3

We previously showed that plasma total homocysteine levels can be relatively insensitive for detecting vitamin B₁₂ deficiency.4 A homocysteine concentration at or above 15 μmol/L has positive and negative predictive values of 7.4% and 97.2%, respectively, for the detection of a serum vitamin B12 concentration below 120 pmol/L. For vitamin B12 levels between 120 and 150 pmol/L, the corresponding predictive values are just 6.3% and 94.0%.4 In Fig. 1 of the article by Robertson and colleagues the mean homocysteine concentration was just 12 μmol/L even in the quartile of patients with the lowest serum vitamin B₁₂ concentrations (below 203 pmol/L). Thus, it is unlikely that vitamin B12 deficiency could be efficiently detected on the basis of a homocysteine measurement equal to or greater than 15 μmol/L.

Their lack of use of a comprehensive and suitable definition of vitamin B₁₂ leads us to question the authors' conclusions that “vitamin B12 deficiency is surprisingly common among patients with vascular disease” and that “low serum vitamin B₁₂ levels are a major determinant of elevated homocysteine levels and increased carotid plaque area.”1 With nearly 50% of patients omitted from their major analysis, could not age and serum creatinine alone have explained some of their findings?