Homocysteine, glycine betaine, and N,N-dimethylglycine in patients attending a lipid clinic

We recruited nondiabetic subjects (n = 158) attending a lipid disorders clinic, a subset of whom (n 46) had established cardiovascular disease. Glycine betaine, N,N-dimethylglycine, and carnitine were measured in fasting plasma and urine samples. The concentrations and excretions were related to known cardiovascular risk factors in multivariate regression models. The relationships between homocysteine and plasma and urinary glycine betaine were highly significant (P < .002), comparable with the known relationships with folate and plasma creatinine. The regression coefficient for plasma glycine betaine was consistently approximately -0.1 in 5 different regression models (3 best-subsets and forward and backward stepwise regression models) for predicting homocysteine using 23 variables. Plasma glycine betaine was higher in males than in females, and the difference was associated with a difference in percentage of body fat. Its concentration included a constant factor of approximately 20 mu mol/L that was independent of any of the variables investigated here. In the total group, body fat, homocysteine, and carnitine were significant predictors of plasma glycine betaine. Carnitine, an important betaine that is involved in lipid metabolism positively correlated with both homocysteine and glycine betaine. In our sample, the urinary excretion of glycine betaine was outside the reference range in 14 of the 158 subjects and the betaine fractional clearances were above the reference range in 23 subjects. Fractional clearance correlated strongly with plasma homocysteine (r = 0.50), and this relationship may be stronger in patients with known vascular disease. Urinary loss of glycine betaine may contribute to hyperhomocysteinemia and the development of cardiovascular disease. (c) 2004 Elsevier Inc. All rights reserved.