Factor V Leiden, Prothrombin 20210 Gene Variant, and Risk of Myocardial Infarction
To The Editor:
Doggen et al1 reported a case-control study assessing the effect of factor V Leiden and the prothrombin 20210 gene variant on the risk of myocardial infarction in men. The authors stated in their conclusions that “the 20210 G→A variant of prothrombin is associated with an increased risk of myocardial infarction” and that “the combined presence of major cardiovascular risk factors and carriership of a coagulation defect increases the risk considerably.” A review of their data in Tables 1 through 4 leads us to conclude that neither claim is supported by statistical significance. In Table 2, the prevalence of the prothrombin variant in cases (1.8%) is not significantly different from the prevalence in controls (1.2%) by the χ2 test, with an OR of 1.5 and a CI (0.6–3.8) that includes unity. A larger study would be required to prove that an OR of 1.5 is truly increased. The likelihood that this OR of 1.5 is not significantly different from 1.0 is also suggested by the fact that it is based on the unexpected finding of a lower prevalence of the variant in the controls (1.2%) than in the general population(2%) rather than a higher prevalence in the cases.
The second conclusion, that a coagulation defect added to the effect of a metabolic defect, also does not stand up to scrutiny of the data. In Table 3, the ORs of 6.1 and 3.3, with their overlapping CIs of 3.0–12.5 and 2.5–4.2, respectively, indicate that smoking increased the risk of myocardial infarction, with or without a coagulation defect, but do not prove that factor V Leiden or the prothrombin variant significantly added to the risk. Without access to the original data, we cannot calculate the P value for this comparison of age-adjusted ORs, but the wide overlap of the CIs suggests that a significant difference between the 2 ORs was not found. Similar observations apply to the subset analyses of the effect of obesity and hypertension in Table 4. Only the pooled data for any “metabolic defect” and any coagulation defect, shown in Table 3, appear possibly to have achieved statistical significance. The authors should have explicitly stated the presence or absence of statistical significance for each of the comparisons of risk that they published in Tables 1 through 4, a minimum of 23 comparisons, for which a P value of <0.05 may be found at least once by chance. We conclude that the authors have not proven that the prothrombin 20210 gene variant, in the absence of factor V Leiden, increases the risk of myocardial infarction.
- Copyright © 1999 by American Heart Association
Doggen CJM, Cats VM, Bertina RM, Rosendaal FR. Interaction of coagulation defects and cardiovascular risk factors. Circulation. 1998;97:1037–1041.
The results of the “Study of Myocardial Infarctions Leiden” showed that the variant in the prothrombin gene (20210 G→A) increased the risk of myocardial infarction by 50% (OR, 1.5; 95% CI, 0.6–3.8).R1 A similar effect was found for another prothrombotic defect, factor V Leiden, which increased the risk by 40% (OR, 1.4; 95% CI, 0.8–2.2). The effect was higher in those individuals with additional risk factors (ie, smoking, obesity, diabetes, hypertension, or hypercholesterolemia).
Hultin and Grimson question our conclusions because of concerns about statistical significance. We agree with their observations about the absence of statistical significance. We also believe this is largely irrelevant. Contrary to what Hultin and Grimson state, we do not think it possible to prove anything, nor do we attempt to do so. Single studies should not swing opinions, whether their results are significant or not. Our first aim is to estimate effects, ie, relative or absolute risks. In this case, our best estimate for the risk of myocardial infarction for carriers of the 20210 variant of the prothrombin gene is a 50% risk increase. Indeed, the 95% CI, which might be seen as a range of plausible values for this OR of 1.5, is 0.6 to 3.8, ie, does not exclude even a small protective effect, or a nearly fourfold increased risk. The most plausible effect, however, is the observed risk of 1.5 (maximum likelihood estimator). The consistently increased risk of myocardial infarction or coronary heart disease in the presence of the prothrombin variant in other studies lends further support to this estimate.R2 R3 R4
Hultin and Grimson may have missed the point about interaction: for each of the major cardiovascular risk factors, a coagulation defect had a higher relative risk in the presence than in the absence of that risk factor. For instance, for nonhypertensives, carriership of a coagulation defect only increased the risk 1.2-fold; for hypertensives, this was 3.3-fold (similarly so for smokers, obese individuals, diabetics, and hypercholesterolemics). It is striking that this was the case for all these major risk factors, and this may be of public health importance because in these individuals, the risk of the genetic clotting defect is superimposed on a risk already elevated by classic cardiovascular risk factors. Again, our findings concerning interaction are consistent with the results of a previous study among young women.R2 Consistency between studies is, in our view, more important than statistical significance: we try to elucidate a biological problem by clinical, epidemiological, and biochemical means. For this aim, the results of additional studies will be helpful.
Doggen CJM, Manger Cats V, Bertina RM, Rosendaal FR. Interaction of coagulation defects and cardiovascular risk factors: increased risk of myocardial infarction associated with factor V Leiden or prothrombin 20210A. Circulation. 1998;97:1037–1041.
Rosendaal FR, Siscovick DS, Schwartz SM, Psaty BM, Raghunathan TE, Vos HL. A common prothrombin variant (20210 G to A) increases the risk of myocardial infarction in young women. Blood. 1997;90:1747–1750.
Watzke HH, Schüttrumpf J, Graf S, Huber K, Panzer S. Increased prevalence of a polymorphism in the gene coding for human prothrombin in patients with coronary heart disease. Thromb Res. 1997;87:521–526.
Arruda VR, Siquiera LH, Chiaparini LC, Coelho OR, Mansur AP, Ramires A, Annichino-Bizzacchi JM. Prevalence of the prothrombin gene variant 20210 G→A among patients with myocardial infarction. Cardiovasc Res. 1998;37:42–45.