Identifying gene-environment interactions on the efficacy of folic acid therapy for hyperhomocysteinemia based on prediction model.

Various genetic and environmental factors or their interactions may result in the failure of folic acid therapy for hyperhomocysteinemia (HHcy). We hypothesized that an optimal predictive model of gene-environment interactions could be constructed to predict the efficacy of folic acid therapy in HHcy. A prospective cohort study of 638 HHcy patients was performed. The patients were treated with oral folic acid (5 mg/d) for 90 days. We used conditional inference tree model to stratify the failure risk of folic acid therapy synthesizing information from a weighted genetic risk score (wGRS) and environmental exposures, simultaneously interpreting the gene-environment interaction network in predicting the efficacy of HHcy. We detected high-order interactions between medical history of stroke, coronary heart disease (CHD), wGRS, and baseline total homocysteine (tHcy) on the failure risk of folic acid therapy. The wGRS in fourth quartile had 3.73-fold increased failure risk of folic acid treatment (odds ratio = 3.73, 95% confidence interval: 1.47-9.45). Stroke was identified as the key discriminator among the variables examined. A total of 3.3% of participants in failure group were at the lowest failure risk of folic acid therapy (nonstroke, non-CHD, baseline tHcy ≤ 31.1 μmol/L, wGRS ≤ 1.05). Individuals with stroke but with wGRS > 1.05 were at the highest failure risk of folic acid therapy (91.0% of participants in failure group). Medical history of stroke, CHD, wGRS, and baseline tHcy were consistently identified as significant risk factors for the failure risk of folic acid therapy. The multiple interactions between genetic and environmental factors can be visually presented via the conditional inference tree model.