Hyperhomocysteinemia, MTHFR 677C→T Polymorphism, and Stroke
To the Editor:
We welcome the report from Madonna and co-workers examining the role of prothrombotic and homocysteine (Hcy) pathway polymorphisms in risk of ischemic stroke in young adults.1 We wish to comment on several issues raised by their article relating to the design of studies of genetic risk factors for complex phenotypes such as ischemic stroke, which we believe to be important when interpreting their findings. As the report points out, genetic predisposition to a complex human phenotype such as stroke is unlikely to be mediated by a large influence of 1 or 2 genes. Many observers agree that it is likely to result from a combination of relatively small individual effects of several genes, each predisposing to stroke via their influence on intermediate phenotypic traits, such as hypertension or hyperlipidemia.2–4⇓⇓ This assumption has several implications for the design of epidemiological studies examining candidate genetic risk factors for stroke.
First, given the small anticipated effect size associated with any single candidate polymorphism, the group sample sizes required to robustly demonstrate an association will be very large. This is important to avoid a potentially erroneous conclusion that no association exists (type 2 error). This point is particularly relevant in the case of the MTHFR 677C→T polymorphism. Most prospective and retrospective studies to date have indicated that mildly elevated plasma Hcy is independently associated with ischemic stroke, and that the MTHFR TT genotype is associated with elevated Hcy. Paradoxically, most studies have not demonstrated an association between ischemic stroke and TT genotype.5 One potential explanation for this apparent inconsistency is that the majority of studies lacked sufficient power to detect an association. Given a background rate of 17% in the general population, as reported in the article by Madonna and colleagues, and an anticipated relative risk of 20% associated with the TT genotype, we estimate that at least 1018 subjects per group are required to detect an effect with 80% power (α=0.05). The practical difficulties that this creates may be overcome by pooling data from multiple centers, or by employing alternative family-based designs, such as that currently underway by the NIH-funded Siblings With Ischemic Stroke Study (SWISS).
Also, given the heterogeneity of underlying biological mechanisms predisposing to ischemic stroke, careful definition of stroke subtype and intermediate traits (eg, hypertension, atrial fibrillation) is important when studying candidate genetic polymorphisms.2,3⇓ This is important because a susceptibility gene that increases stroke risk by influencing 1 trait (eg, hypertension) is unlikely to be overrepresented in a study sample that contains a large proportion of patients with stroke due to other mechanisms (eg, cardiogenic embolism). Stratification by subtype and intermediate traits may increase the ability to detect the influence of prespecified candidate genes, by decreasing “noise” and reducing variability in the sample. This is relevant when interpreting studies examining the role of Hcy pathway polymorphisms. Although greatly elevated Hcy is associated with arterial thrombosis in vivo and activation of coagulation pathways in vitro, mildly elevated Hcy levels in the range described in the report by Madonna et al have been reported to cause cerebrovascular disease primarily by predisposing to large- and small-vessel atherosclerosis.6,7⇓ In studies of this nature, underrepresentation of subjects with large- and small-vessel stroke may further dilute the ability to detect an association of stroke with Hcy pathway polymorphisms, such as MTHFR TT genotype.
Finally, it is important to control for potential modification of genetic influences by environmental factors, such as vitamin status in studies of Hcy pathway genes.2–4⇓⇓ Unfortunately, this information was not provided in the report. It is particularly relevant to studies of the MTHFR 677C→T substitution, as data consistently indicate that the adverse effect of the substitution on plasma Hcy are abolished by adequate folate intake, suggesting that the TT genotype may be relatively more important in populations with low folate intake. This is relevant when interpreting their results in the US context, as recent data indicate that folate levels have increased in the US population since folate fortification of cereals in 1998, with accompanying reductions in plasma Hcy.8
- ↵Madonna P, de Stefano V, Coppola A, Cirillo F, Cerbone AM, Orefice G, Di Minno G. Hyperhomocysteinemia and other inherited prothrombotic conditions in young adults with a history of ischemic stroke. Stroke. 2002; 33: 51–56.
- ↵Hassan A, Markus HS. Genetics and ischaemic stroke. Brain. 2000; 123: 1784–1812.
- ↵Brass LM, Alberts MJ. Genetic epidemiology and family studies of stroke.In: Alberts MJ, ed. Genetics of Cerebrovascular Disease. New York, NY: Futura; 1998: 159–182.
- ↵Eikelboom JW, Hankey GJ, Anand SS, Lofthouse E, Staples N, Baker RI. Association between high homocyst(e)ine and ischemic stroke due to large- and small-artery disease but not other etiologic subtypes of ischemic stroke. Stroke. 2000; 31: 1069–75.
We thank Dr Kelly et al for their letter that emphasizes major aspects concerning the design of our study as well as the role of some genetic predisposing factors. As acknowledged by Kelly et al, we interpret our data to support the concept that, rather than single genes, the combination of several genes, each causing a small effect per se, is important to lead to complex human phenotypes such as stroke. Nevertheless, our data emphasize that, similar to factors such as cigarette smoking, hypertension, diabetes, and hyperlipidemia, hyperhomocysteinemia is more common in cases than in controls and is important to help identify subjects with a history of early-onset ischemic stroke.
From studies on other conditions associated with the risk of ischemic events, hypercholesterolemia rather than genetic defects leading to hypercholesterolemia are associated with the clinical setting. In keeping with this, candidate genes thought to play a role in arterial thrombosis (including ischemic stroke) correlate with quantitative/qualitative changes in the levels of the protein and poorly with ischemic events.1 This is also true in the present report: rather than with the ischemic event, MTHFR TT mutation correlates with serum levels of homocysteine.2 As Dr Kelly et al are aware, there is a variety of combinations of genetic/environmental factors leading to hyperhomocysteinemia. Despite this, appropriate interventions (folate) correct hyperhomocysteinemia in subjects with as much as in those without homozygosity for the MTHFR TT mutation. This does not imply that it is not worth looking for the association of this polymorphism with the event in larger sample sizes; it only suggests that the impact that these findings may have for the identification of subjects at risk is likely to be rather limited.
The point of stratification of the sample(s) according to the type of event (atherothrombotic, cardioembolic) is well taken. Several groups believe that rather than in atherosclerosis, homocysteine is involved in platelet activation and thrombosis.3 Nor we have found conclusive clinical data that at low concentrations, homocysteine is atherogenic while it is thrombogenic at higher concentrations. Thus we are not entirely sure that stratifying the sample(s) according to the type of event will help clarify the role of the MTHFR TT mutation in early-onset ischemic stroke.
Stroke welcomes Letters to the Editor and will publish them, if suitable, as space permits. They should not exceed 750 words (excluding references) and may be subject to editing or abridgment. Please submit letters in duplicate, typed double-spaced. Include a fax number for the corresponding author and a completed copyright transfer agreement form (published in the January and July issues).
- ↵Brattström L, Wilcken D, Öhrvick J, Brudin L. Common methylenetetrahydrofolate reductase gene leads to hyperhomocysteinemia but not to vascular disease. Circulation. 1998; 98: 2520–2526.
- ↵Davì G, Di Minno G, Coppola A, Andria G, Cerbone AM, Madonna P, Tufano A, Falco A, Marchesani P, Ciabattoni G, Patrono C. Oxidative stress and platelet activation in homozygous homocystinuria. Circulation. 2001; 104: 1124–1128.