Objective: To determine in healthy men: (1) whether an extended genotype of the fibrinogen gene cluster using the G/A-455 and the BclI polymorphism of the beta fibrinogen gene and TaqI of the alpha fibrinogen gene explains a significantly larger proportion of variance in plasma fibrinogen levels in either smokers or non-smokers than a single polymorphism (G/A-455); (2) whether there is any evidence for genotype-smoking interaction in the determination of fibrinogen levels.
Design: A cross sectional study of healthy, white men recruited at the screening for entry into the Thrombosis Prevention Trial.
Setting: The subjects were drawn from four general practices in the United Kingdom.
Results: The frequency of the rare alleles in the sample was 0.19 for the G/A-455 polymorphism (A-455), 0.15 for BclI (B+), and 0.27 for TaqI (T+) alleles. BMI and age made significant contributions to the variance in plasma fibrinogen levels only in non-smokers of 5.4% and 2.3% respectively and, in the group as a whole, smoking accounted for 6.6% of the variance. In the non-smokers, of the individual polymorphisms only the G/A-455 showed a significant association with plasma fibrinogen levels (p = 0.03). The mean fibrinogen in non-smokers homozygous for the G-455 allele was 2.54 g/l v 2.85 g/l in those homozygous for the A-455 allele, with the polymorphism explaining 3.6% of the variance in plasma fibrinogen levels in this group. On investigation of the association of fibrinogen levels with combined genotypes, the most significant effect was seen with the combination of the G/A-455 and TaqI polymorphisms, with those with no "fibrinogen raising alleles" having a mean fibrinogen of 2.57 g/l v 3.10 g/l for those with four "fibrinogen raising alleles" (p = 0.0036), and this combination explained 8.9% of the variance in plasma fibrinogen levels (p < 0.005). Although the contribution to variance was greater with the G/A-455/TaqI combination than the G/A-455 polymorphism alone (8.9% v 3.6%), this did not reach significance (p = 0.09). By contrast, in the smoking group, the only significant contribution to the difference in plasma fibrinogen levels was the G/A-455 genotype alone which, after adjustment for BMI and age, contributed 3.8% to the variance (p < 0.05). No interaction was shown between smoking and genotype.
Conclusion: These data suggest that in non-smokers an extended genotype using the G/A-455 beta fibrinogen gene polymorphism and the TaqI alpha fibrinogen gene polymorphism explains a larger proportion of the variance in plasma fibrinogen levels than any one polymorphism alone, but that smoking has an overriding effect so that other variables such as age and BMI make little additional contribution.