Direct, reverse or reciprocal causation in the relation between homocysteine and ischemic heart disease

Background: Whether mild hyperhomocysteinemia is a risk factor for ischemic heart disease (IHD) or it is a secondary epiphenomenon remains unknown. We tested the alternative hypotheses that the Hcy-IHD relation is due to direct, reversal or reciprocal causality. Methods: Ninety-four families from 32 pedigrees (296 members) including subjects who died for a premature (< 50 years) IHD and with at least one family member with also a premature IHD were selected. Three Structural Equation Models were created, in which causal (Model 1), reversal (Model 2), and reciprocal (Model 3) tHcy-IHD relation were tested. Results were confirmed by testing "Pearl's instrumental inequalities". Results: A significant tHcy-IHD association was found in Model 1 (OR= 1.38, 95% CI: 1.01 to 1.88, for any increase of + 10 mu mol/l in tHcy), as opposed to a non-significant association in the other models (Model 2: MD = + 1.63 mu mol/l, 95% CI: -1.72 to + 4.99 mu mol/l; Model 3: OR=0.69, 95% CI: 0.17 to 2.78 for tHcy predictor of IHD; MD = -0.46 mu mol/l, 95% CI: -2.41 to 1.48 mu mol/l, for IHD predictor of tHcy). "Pearl's instrumental inequalities" qualify MTHFR as an instrument relative to the tHcy-IHD relation. A suppression effect may explain the non-significant total MTHFR-IHD relation (OR=1.275, 95% CI: 1.02 to 1.71 for the indirect MTHFR-tHcy-IHD path; OR=0.52, 95% CI: 0.17 to 1.64 for the direct MTHFR-IHD path). Conclusion: Our findings support the assumption of a triangular genotype-phenotype-disease mediation process in the Hcy-IHD relation, and indirectly, of a causal relationship between moderately elevated plasma tHcy levels and IHD. (c) 2006 Elsevier Ltd. All rights reserved.