Midway Evaluation - Gard Frodahl Svingen
Dimethylglycine and Coronary Heart Disease
Background: Dimethylglycine (DMG) is made from betaine during the folate independent remethylation of homocysteine (Hcy). This process is related to lipid metabolism and methylation status, and a small study suggested adverse cardiovascular risk among patients with a recent acute coronary syndrome and higher plasma DMG levels. Moreover, altered betaine and DMG levels in plasma and urine have also been found among patients with diabetes, a major CHD risk factor and the incidence of which is increased by statin therapy. We investigated the relationship between plasma DMG and cardiovascular risk factors and clinical outcomes among patients with suspected or verified coronary heart disease (CHD). The within-subject reproducibility of plasma DMG was also explored. We further assessed the relationship between statin treatment and the status of one-carbon metabolites and markers of B-vitamin status.
Methods: Two independent cohorts of patients with either suspected angina pectoris (SAP) (n=4164) or an acute myocardial infarction (AMI) (n=3749), followed for median 4.6 and 7.0 years, respectively, provided baseline and long-term clinical, laboratory and endpoint data. The majority of these patients were participants in two large randomized clinical trials, investigating B-vitamin therapy in secondary CHD prevention. Associations between plasma DMG and baseline variables were explored using median linear and logistic regression, and generalized additive modeling. Survival was studied by Cox regression analyses. Assessment of temporal changes in plasma DMG and the within-subject reproducibility was made using mixed linear and random effect modeling, respectively.
Results: Among both cohorts plasma DMG was related to an adverse CHD risk profile. We observed an almost doubled risk of incident AMI among patients with SAP and plasma DMG in the upper vs. the lower quartile, when adjusting for potential confounders. This association was even stronger among patients who did not smoke or had low levels of serum triglycerides or apolipoprotein B (P for interaction <0.03).
We also found that among both cohorts there was a linear relationship between plasma DMG an overall mortality, with those having DMG levels in the highest vs. the lowest DMG quartile carrying about 70% increased risk in age and gender adjusted analyses. These estimates were robust also when adjusting for potential confounders, including cardiac troponin T among patients with SAP, and even stronger for death due to cardiovascular causes. Adding plasma DMG to traditional CHD risk factors improved risk prediction of clinical outcomes in both studies, although stronger with endpoints related to cardiovascular disease. Further, we observed excellent within-subject reproducibility for plasma DMG among all patient, and showed that treatment with folic acid hampered the time-dependent increase in plasma DMG levels seen among patients not allocated to such treatment.
Among patients with SAP, we found a dose-response relationship between statin therapy and plasma betaine, although no alterations in urinary betaine levels were seen. Plasma total Hcy (tHcy) levels were lower among statin users in general, but patients taking very high doses of statin actually had increased plasma tHcy. Despite any dose-response relationship, statin users also tended to have higher plasma choline and DMG levels, whereas urinary choline levels were lower.
Conclusion: During our work, we have shown that plasma DMG levels are related to an unfavourable cardiovascular prognosis, independent of the association with other, established CHD risk factors. Moreover, plasma DMG exerts great within-subject reproducibility. We also showed that statin therapy may influence the choline oxidation pathway, suggesting a potential interplay between statin therapy, one-carbon metabolism, and diabetes.