An elevated serum concentration of the metabolite, homocysteine (Hcys): 1) can indicate folate or vitamin B12 deficiency, 2) is an independent risk factor for vascular disease. The metabolite, methylmalonic acid (MMA), is elevated in deficiency of vitamin B12, but not folate. The purpose of this study was to determine the effect of self-selected vitamin supplementation and other variables on serum Hcys and MMA concentrations in elderly men and women. Serum concentrations of Hcys, MMA, folate and vitamin B12 were measured for elderly volunteers, age 68-96 years, and compared for those consuming (26 men, 25 women) and not consuming (24 men, 25 women) self-selected vitamin supplements. Compared with the nonsupplemented group, the supplemented group had lower mean serum MMA (208 +/− 162 vs. 241 +/− 98 nmol/L [+/− SD]) and Hcys (9.5 +/− 2.6 vs. 11.2 +/− 2.7 mumol/L); and higher serum vitamin B12 (391 +/− 174 vs 292 +/− 107 pmol/L), and serum folate (46 +/− 15 vs. 24 +/− 10 nmol/L) p < 0.05. Among all 100 subjects, the prevalence of serum vitamin B12 < 221 pmol/L (300 pg/mL) was 18; MMA > 271 nmol/L, 16; Hcys > 16.2 mumol/L, 3; folate < 5.0 nmol/L, none. Based on serum vitamin B12 < 221 nmol/L with elevated serum MMA, vitamin B12 deficiency was probable in seven subjects, of whom two were supplemented. All three subjects with elevated serum Hcys had elevated serum MMA as well, suggesting vitamin B12 deficiency or renal insufficiency. A stepwise linear regression model for serum Hcys explained 61.7% of the variance, and included (in order) serum creatinine, folate, vitamin B12, albumin, age and body mass index (BMI). A model with serum MMA replacing serum vitamin B12 explained 64.1% of the variance in serum Hcys. Folate did not enter the model for supplemented subjects, supporting a “threshold effect”: serum Hcys was inversely related to serum folate at lower serum folate (nonsupplemented subjects), but at higher serum folate (supplemented subjects), the relationship was flat. In supplemented subjects, serum Hcys was still related to vitamin B12 status, confirming that tissue deficiency of the vitamin was present. Results showed potential usefulness of serum MMA and Hcys in identifying subclinical or tissue deficiency of vitamin B12. Clinicians should be aware of the risk of vitamin B12 deficiency in older people and of current screening algorithms using serum metabolites. These elderly volunteers had generally good folate status; nevertheless, some subjects seemed likely to benefit from an improvement in folate status that would reduce their serum Hcys within the normal range. The role of serum creatinine in the normal range in predicting serum Hcys, a vascular disease risk factor, remains unexplained.