Serum homocysteine concentrations, gemfibrozil treatment, and progression of coronary atherosclerosis

Serum homocysteine concentrations, gemfibrozil treatment, and progression of coronary atherosclerosis

Atherosclerosis 172 (2004) 267–272 Serum homocysteine concentrations, gemfibrozil treatment, and progression of coronary atherosclerosis Mikko Syvänn...

97KB Sizes 0 Downloads 45 Views

Atherosclerosis 172 (2004) 267–272

Serum homocysteine concentrations, gemfibrozil treatment, and progression of coronary atherosclerosis Mikko Syvänne a,∗ , Roslyn A. Whittall b , Ursula Turpeinen c , Markku S. Nieminen a , M. Heikki Frick a , Y. Antero Kesäniemi d , Amos Pasternack e , Steve E. Humphries b , Marja-Riitta Taskinen a a b

Department of Medicine, Division of Cardiology, Helsinki University Central Hospital, P.O. Box 340, Haartmaninkatu 4, Helsinki, 00029 HUS, Finland Department of Medicine, Centre for Cardiovascular Genetics, The Rayne Institute, Royal Free and University College London Medical School, London, UK c Department of Clinical Chemistry, Helsinki University Central Hospital, Helsinki, Finland d Department of Medicine and Biocenter Oulu, Oulu University Hospital and University of Oulu, Oulu, Finland e Department of Medicine, Tampere University Hospital, Tampere, Finland Received 15 April 2003; received in revised form 17 September 2003; accepted 23 October 2003

Abstract The present study aimed to assess the effect of gemfibrozil on serum total homocysteine (tHcy) concentration and to evaluate the influence of tHcy on the angiographically determined progression of coronary atherosclerosis in a randomised, placebo-controlled trial of 395 post-coronary bypass men with low HDL cholesterol levels. The baseline levels of tHcy and those after 16 months of randomised therapy were measured by high-pressure liquid chromatography. Patients were genotyped for the thermolabile variant of N5 ,N10 -methylenetetrahydrofolate reductase (MTHFR) (677C > T substitution). Gemfibrozil therapy was associated with a median 18% increase in tHcy levels (P < 0.001). In the gemfibrozil group increases in tHcy and HDL cholesterol were related (r = 0.217, P = 0.004), but changes in tHcy and triglycerides were not. Levels of tHCy were not associated with baseline extent or progression of coronary-artery disease. Subjects homozygous for the rare MTHFR T allele had 34% higher median tHcy concentrations than CC homozygotes or CT heterozygotes, but responses to gemfibrozil did not differ significantly among genotypes. The MTHFR genotype was not associated with extent or progression of coronary atherosclerosis. We conclude that gemfibrozil causes a significant elevation in tHcy levels, but the clinical relevance of this is unknown at present. © 2003 Elsevier Ireland Ltd. All rights reserved. Keywords: Homocysteine; Gemfibrozil; Coronary-artery disease

1. Introduction Plasma total homocysteine (tHcy) concentration has emerged as a risk marker for atherosclerosis and ischaemic clinical events [1–5]. Recently, three small short-term studies suggested that the fibric acid derivatives fenofibrate and bezafibrate elevate plasma tHcy concentrations [6–8]. These results seem paradoxical, as in many studies fibrates have retarded progression of atherosclerosis [9–11] and diminished clinical cardiovascular events [12,13]. A recent small study suggested that another fibrate, gemfibrozil, might not raise tHcy levels [14].

The present study was undertaken to examine the effect of gemfibrozil on tHcy in a larger cohort and over longer term than the previous studies. A second objective was to investigate the relation between homocysteine levels and angiographically determined progression of coronary atherosclerosis. Furthermore, we assessed whether the thermolabile variant of N5 ,N10 -methylenetetrahydrofolate reductase (MTHFR), caused by a C-to-T substitution at nucleotide 677 (677C > T) [15,16], modulated these associations. 2. Materials and methods 2.1. Patients and study outline

Corresponding author. Tel.: +358-50-4271938; fax: +358-9-47174574. E-mail address: [email protected] (M. Syvänne).

The design [17] and principal results [10,18] of the Lopid Coronary Angiography Trial (LOCAT) have been reported.

0021-9150/$ – see front matter © 2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.atherosclerosis.2003.10.003


M. Syvänne et al. / Atherosclerosis 172 (2004) 267–272

Briefly, we enrolled 395 men with low HDL cholesterol levels as their main lipid abnormality to receive gemfibrozil or placebo in a randomised, double-blind study for on average 2.5 years. All had previously undergone a successful coronary bypass operation at least 6 months before baseline and were clinically stable. Subjects with significant renal dysfunction, defined as serum creatinine concentrations ≥175 ␮mol/l (2.0 mg/dl), were excluded. Other exclusion criteria were manifest diabetes mellitus, hypertension not adequately controlled with beta blockers, and smoking. The patients received dietary counselling at baseline and regularly during the trial. Coronary and bypass graft angiograms were obtained at baseline and at the end of the trial. A validated computer-assisted quantitative method was used to analyse the angiographic change over time. We found that gemfibrozil significantly retarded progression of coronary and vein-graft atherosclerosis [10]. Lipoprotein levels predicted angiographic change in that LDL and IDL were directly and HDL3 inversely related to worsening atheroma in native coronary vessels. In vein grafts, VLDL and IDL were the main lipoproteins associated with accelerated progression of obstructions [18]. 2.2. Laboratory procedures Blood samples were obtained by venipuncture after an overnight fast. Serum was promptly separated and frozen at −80 ◦ C and kept frozen until the present analyses. Serum tHcy concentrations (homocysteine-cysteine, mixed disulfides, and protein-bound homocysteine) were measured by high-performance liquid chromatography as follows. To 100 ␮l of serum we added 10 ␮l of internal standard (10 mg/l, 2-mercaptoethylamine) and 100 ␮l of reducing agent (10 g/l TCEP-HCl) and incubated the mixture at room temperature for 5 min. Proteins were precipitated by addition of 100 ␮l of perchloric acid (0.6 M containing 1 mM EDTA). To derivatize the thiols in 50 ␮l of clear supernatant, we added 50 ␮l of thiol-specific reagent, ammonium-7-fluoro-2,1,3-benzoxadiazole-4-sulphonate (SBD-F), 1 g/l in borate buffer, and heated this for 1 h at 60 ◦ C. We used Lichrospher 100 RP-18, 5 ␮m, 250 mm × 4.6 mm column with a mobile phase of 5% acetonitrile in 0.1 M phosphate buffer, pH 2.3 at a flow rate of 1 ml/min. The detection was by fluorometry. The excitation was at 385 nm and the emission at 515 nm. The detection limit of the assay is 1 ␮mol/l. Interassay CV is 5–8% at 9–35 ␮mol/l. Samples taken at baseline and after 16 months of randomised therapy were used. Other biochemical measurements were performed as reported previously [17]; LDL cholesterol was calculated by the Friedewald equation. 2.3. DNA extraction and genotyping Genomic DNA was extracted from peripheral whole blood by the salting-out method [19]. For the MTFHR 677C > T polymorphism the primers used were as described in [20]

and PCR conditions, digestion with HinfI and MADGE gel electrophoresis as described in [21]. 2.4. Statistical analysis Data are reported as median (25th, 75th percentile). Logarithmic transformations were used to normalise skewed distributions. Group means were compared by unpaired t-tests or ANOVA. Within-group changes from baseline to on-trial were evaluated with paired t-tests. Correlations were assessed by Pearson’s or Spearman’s coefficients.

3. Results Baseline and on-trial tHcy data were available for 184 patients in the placebo group and 178 patients in the gemfibrozil group. Fig. 1 shows that baseline concentrations were similar in both groups. The median value (25th, 75th percentile) was 11.9 (9.5, 15.2) ␮mol/l in the placebo group and 12.6 (9.9, 15.1) ␮mol/l in the gemfibrozil group, P = 0.50. tHcy levels after 16 months of therapy were unchanged in the placebo group (12.0 [9.9, 14.3] ␮mol/l, P = 0.61), whereas a highly significant elevation was observed in the gemfibrozil group (14.1 [11.9, 17.5] ␮mol/l, P < 0.001), between-group on-trial P < 0.001. In the placebo group, median change from baseline to on-trial was only 0.6% (25th, 75th percentile −14, 23). In those allocated to gemfibrozil therapy, median increase in tHcy concentration was 17.7% (−2, 42). In the combined study population, baseline tHcy concentration was positively correlated with age (r = 0.191, P < 0.001) but not with body mass index, total, HDL or LDL cholesterols, serum triglyceride, glucose, or insulin levels (data not shown). In both the placebo and gemfibrozil groups, the change in tHcy from baseline to 16 months was not related to age, adiposity, or baseline lipids, glucose or insulin. tHcy change was also unrelated to changes in total cholesterol, LDL cholesterol, and triglycerides. In both groups there was an inverse relation with baseline tHcy (placebo, r = −0.441, P < 0.001; gemfibrozil, r = −0.416, P < 0.001). By contrast, in the treated group only there was a significant relation between the change in tHcy and the change in HDL cholesterol (r = 0.217, P = 0.004) (Fig. 2). MTHFR 677C > T genotype was obtained for 296 patients, 148 in each randomised group, 138 of whom in the placebo group and 136 in the gemfibrozil group also had complete tHcy concentration data. Overall the distribution of genotypes was as expected from Hardy-Weinberg proportions; the frequency of the rare allele was 0.28 (95% CI 0.24–0.31) and this was not different in the placebo and treated group (P > 0.2). As shown in Table 1, TT homozygotes had significantly higher tHcy concentrations at baseline and during randomised therapy than heterozygotes or CC homozygotes. In those receiving fibrate treatment

M. Syvänne et al. / Atherosclerosis 172 (2004) 267–272


Fig. 1. Serum total homocysteine (tHcy) concentrations at baseline and during randomised therapy with gemfibrozil or placebo. Lines within boxes represent median values; lower and upper ends of boxes represent 25th and 75th percentiles, respectively.

Fig. 2. Association between changes in serum total homocysteine (tHcy) and HDL cholesterol concentrations.

median (25th, 75th percentile) change in tHcy concentration was 15 (−2, 43)% in the CC group, 18 (−5, 40)% in the CT group and 22 (−13, 33)% in the TT group although this difference did not reach statistical significance (ANOVA, P = 0.776, CC + CT versus TT, P = 0.476). The number of distal anastomoses of grafts placed at the bypass operation (median 4, range 1–8) was used as a measure of the extent of coronary atherosclerosis at baseline. The number of anastomoses was not related to baseline tHcy (Spearman’s rho = 0.083). The MTHFR genotypes also had similar numbers of anastomoses (median 4 in each genotype). We defined global angiographic progression as any deterioration in coronary dimensions exceeding random variation

[18,22]. Neither baseline nor on-trial tHcy differed in those showing or not showing progression of coronary atherosclerosis. This was true for crude analysis in the whole cohort and in an analysis adjusting for the group allocation. The presence of new vein-graft lesions in the follow-up angiogram, a characteristic strongly influenced by gemfibrozil therapy [10], was also unrelated to tHcy. Finally, the quantitative changes in native coronary dimensions (change in average coronary segment diameters, reflecting primarily diffuse progression and change in minimum luminal diameter of coronary stenoses, a measure of focal progression) were unrelated to tHcy. None of these measures of disease progression differed among the MTHFR 677C > T genotypes (not shown).

Table 1 Serum total homocysteine concentrations according to the methylenetetrahydrofolate reductase gene C667T polymorphism

Baseline placebo [n] On-trial placebo [n] Baseline gemfibrozil [n] On-trial gemfibrozil [n]





11.65 (9.5, 14.3) [76] 11.75 (9.85, 13.95) [76] 11.55 (9.5, 14.9) [62] 13.2 (11.6, 16.7) [62]

11.85 (9.3, 14.9) [50] 11.35 (9.7, 13.7) [50] 13.1 (10.0, 14.85) [67] 13.9 (12.15, 17.45) [67]

14.35 (10.2, 21.55) [12] 17.6 (13.45, 21.5) [12] 18.2 [15.05, 29.3] [7] 20.6 (17.8, 25.35) [7]

0.069 0.001 0.001 0.007

Data are ␮mol/l, median (25th, 75th percentile).


M. Syvänne et al. / Atherosclerosis 172 (2004) 267–272

4. Discussion The results of the present study showed that during 16 months of gemfibrozil therapy in men with coronary-artery disease and low levels of HDL cholesterol, tHcy concentrations increased by 18% while they remained on average unchanged during treatment with placebo. This extends similar findings in smaller and shorter trials with fenofibrate [6–8] and bezafibrate [7] and suggests that increases in tHcy are a class action of fibric acid derivatives. Our results differ from those of Westphal et al. [14] who found an elevation of tHcy levels on bezafibrate but not on gemfibrozil therapy. Possible explanations for this discrepancy include different patient populations, smaller gemfibrozil dose (900 mg in the previous versus 1200 mg per day in the present study), shorter follow-up in the previous study, and, most likely, chance due to the small sample size (n = 22) in the study by Westphal et al. The mechanism by which gemfibrozil and other fibrates elevate tHcy levels is unknown. Fibrates act by binding to transcription factors known as peroxisome proliferator-activated receptors (PPARs), especially PPAR-␣, expressed mainly in liver, skeletal muscle, myocardium, and kidney [23,24]. This binding modulates the transcription of several genes involved in lipid metabolism, such as the genes encoding lipoprotein lipase, apolipoprotein C-III, acetyl coenzyme A synthase, fatty acid transport protein, and the HDL-related apolipoproteins apoA-I and apoA-II [23]. Ultimately, these transcriptional changes promote lipolysis, reduce VLDL production, and increase HDL production. Furthermore, it is now known that PPARs are expressed in endothelial and smooth muscle cells and macrophages [24]; thus, fibrates may have pleiotropic effects beyond lipid metabolism. We found a strong inverse correlation between baseline tHcy concentration and the change observed during the study, i.e., initially high concentrations tended to decrease and vice versa. This relationship was similar in both the placebo and treatment groups and most likely represents regression to the mean. In the gemfibrozil group only there was also a significant positive association between change in tHcy and change in HDL cholesterol, suggesting a possible a link between the mechanisms by which gemfibrozil increases HDL cholesterol and tHcy levels. One possibility for this would be if one or more of the enzymes in the transsulphuration or remethylation pathways that determine the metabolism of homocysteine contained PPAR responsive elements, but currently there is no evidence from the literature that this is the case. By contrast, no such association was found between the change in tHcy and triglyceride lowering, which is the main lipid effect of gemfibrozil. Nutritional intakes of folate and Vitamins B6 and B12 have profound effects on tHcy levels [1]. This is unlikely to be a confounder in the present study because patients in placebo and gemfibrozil groups received similar dietary counselling; they were advised to follow a low-fat diet rich in vegetables,

one that would also provide adequate amounts of folate and Vitamin B. In this respect, the patients’ diets probably did not greatly change from baseline, as tHcy concentrations remained stable in the placebo group. This may be due to earlier dietary advice in conjunction with coronary events and bypass surgery. In future studies, acquisition of more detailed dietary data is desirable. Fibrates do not seem to influence concentrations of Vitamin B or folic acid [7]. Certain other drugs raise tHcy concentrations [25], but they act by antagonizing folate (methotrexate, phenytoin, carbamazepin) or Vitamin B6 (theophylline, azarabine, estrogen-containing oral contraceptives) [1], unlikely mechanisms for fibrates. tHcy has been associated with the presence of atherosclerotic vascular disease in a large number of retrospective and case–control studies, but the results of prospective studies in initially healthy people have been variable (reviewed in references [1–5]). We studied a group of men with severe coronary atherosclerosis, all of whom had received a coronary bypass graft operation. We found no association between tHcy levels and the number of bypasses required at the time of surgery. This is in broad agreement with the dataset by Nygård et al., encompassing a wider range of the extent of coronary atherosclerosis, where only small and statistically borderline differences in tHcy levels were detected according to the number of angiographically diseased vessels [26]. To our knowledge, this is the first detailed analysis of the progression of coronary-artery disease as assessed by serial coronary angiography in relation to tHcy. We found no association between the progression of native coronary or bypass graft disease and tHcy levels, in contrast to our previous findings with lipoproteins [18]. This negative finding requires confirmation in further studies, preferably with a more diverse patient population and a longer follow-up. Our finding is in apparent conflict with the study of Nygård et al. [26], who found a strong, graded association between tHcy and total mortality over a median follow-up of 4.6 years in a group of patients with angiographically confirmed coronary-artery disease. We hypothesize that if the relation between tHcy and mortality in patients with coronary-artery disease is causal, the underlying mechanisms may relate more closely to the effects of tHcy on endothelial function and thrombogenicity [2] and thus clinical events, than with the anatomical progression of coronary atherosclerosis. Unfortunately, this hypothesis could not be tested in the present study because of the very low incidence of ischaemic events [10]. In this sample the frequency of the 677T allele was lower than that reported in the UK (0.32, 95% CI 0.31–0.34) but not significantly so (P = 0.09) and higher than the 0.23 (0.19–0.28) reported in Baltic subjects [27], but again not significantly so (P = 0.27). In agreement with older (reviewed in reference [28]) and more recent [21,29] data, LOCAT subjects homozygous for the MTHFR 677T thermolabile variant had group median tHcy levels that were approximately 37% higher than those with the CC or CT genotypes, taking into account the baseline and on-trial

M. Syvänne et al. / Atherosclerosis 172 (2004) 267–272

values in the placebo group. After fibrate treatment, levels rose in all genotype groups and although this rise was greatest in the TT group, such that the TT subjects had median about 50% higher than the CC + CT group combined, this difference was not significantly larger, implying no specific contraindication for fibrate treatment in the TT group. As in previous studies [28], no association was found between MTHFR genotype and extent or progression of coronary atherosclerosis. In conclusion, we found that gemfibrozil, compared with placebo, significantly increased tHcy levels in men with coronary-artery disease, but the clinical relevance of this finding is uncertain. If ongoing clinical trials show benefit from treatment of hyperhomocysteinaemia with folate and Vitamins B6 and B12 , the usefulness of routinely combining such therapy with a fibrate should be tested in further clinical trials.




[14] [15]


Acknowledgements This study was supported by grants from Parke–Davis, a division of Warner Lambert; the Finnish Foundation for Cardiovascular Research; the Aarne Koskelo Foundation; and the Finnish Society of Angiology. S.E. Humphries R.A. Whittall are supported by the British Heart Foundation (RG2000015).

References [1] Hankey GJ, Eikelboom JW. Homocysteine and vascular disease. Lancet 1999;354:407–13. [2] Welch GN, Loscalzo J. Homocysteine and atherothrombosis. New Engl J Med 1998;338:1042–50. [3] Eikelboom JW, Lonn E, Genest Jr J, Hankey G, Yusuf S. Homocystein(e) and cardiovascular disease: a critical review of the epidemiologic evidence. Ann Intern Med 1999;131:363–75. [4] Mangoni AA, Jackson SHD. Homocysteine and cardiovascular disease: current evidence and future prospects. Am J Med 2002;112: 556–65. [5] The Homocysteine Studies Collaboration. Homocysteine and risk of ischemic heart disease and stroke. A meta-analysis. J Am Med Assoc 2002;288:2015–22. [6] De Lorgeril M, Salen P, Paillard F, Lacan P, Richard G. Lipid-lowering drugs and homocysteine. Lancet 1999;353:209– 10. [7] Dierkes J, Westphal S, Luley C. Serum homocysteine increases after therapy with fenofibrate or bezafibrate. Lancet 1999;354:219– 20. [8] Giral P, Bruckert E, Jacob N, Chapman MJ, Foglietti MJ, Turpin G. Homocysteine and lipid lowering agents. A comparison between atorvastatin and fenofibrate in patients with mixed hyperlipidemia. Atherosclerosis 2001;154:421–7. [9] Ericsson C-G, Hamsten A, Nilsson J, Grip L, Svane B, de Faire U. Angiographic assessment of effects of bezafibrate on progression of coronary artery disease in young male postinfarction patients. Lancet 1996;347:849–53. [10] Frick MH, Syvänne M, Nieminen MS, Kauma H, Majahalme S, Virtanen V, et al. Prevention of the angiographic progression of













coronary and vein-graft atherosclerosis by gemfibrozil after coronary bypass surgery in men with low levels of HDL cholesterol. Circulation 1997;96:2137–43. Diabetes Atherosclerosis Intervention Study Investigators. Effect of fenofibrate on progression of coronary-artery disease in type 2 diabetes: the Diabetes Atherosclerosis Intervention Study, a randomised study. Lancet 2001;375:905–10. Frick MH, Elo O, Haapa K, Heinonen OP, Heinsalmi P, Helo P, et al. Helsinki Heart Study: primary-prevention trial with gemfibrozil in middle-aged men with dyslipidemia: safety of treatment, changes in risk factors, and incidence of coronary heart disease. New Engl J Med 1987;317:1237–45. Bloodfield Rubins H, Robins SJ, Collins D, Fye CL, Anderson JW, Elam MB, et al. Gemfibrozil for the secondary prevention of coronary heart disease in men with low levels of high-density lipoprotein cholesterol. New Engl J Med 1999;341:410–8. Westphal S, Dierkes J, Luley C. Effects of fenofibrate and gemfibrozil on plasma homocysteine. Lancet 2001;358:39–40. Kang SS, Zhou J, Wong PWK, Kowalisyn J, Strokosch G. Intermediate homocysteinemia: a thermolabile variant of methylenetetrahydrofolate reductase. Am J Hum Genet 1988;43:414–21. Frosst P, Blom HJ, Milos R, Goyette P, Sheppard CA, Matthews RG, et al. A candidate genetic risk factor for vascular disease: a common mutation in methylenetetrahydrofolate reductase. Nat Genet 1995;10:111–3. Syvänne M, Taskinen M-R, Nieminen MS, Manninen V, Kesäniemi YA, Pasternack A, et al. A study to determine the response of coronary atherosclerosis to raising low HDL cholesterol with a fibric-acid derivative in men after coronary bypass surgery: the rationale, design, and baseline characteristics of the LOCAT study. Control Clin Trials 1997;18:93–119. Syvänne M, Nieminen MS, Frick MH, Kauma H, Majahalme S, Virtanen V, et al. Associations between lipoproteins and the progression of coronary and vein-graft atherosclerosis in a controlled trial with gemfibrozil in men with low baseline levels of HDL cholesterol. Circulation 1998;98:1993–9. Miller SA, Dykes DD, Polesky HF. A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res 1988;16:1215. Day IN, Humphries SE. Electrophoresis for genotyping: microtiter array diagonal gel electrophoresis on horizontal polyacrylamide gels, hydrolink, or agarose. Anal Biochem 1994;222:389–95. Dekou V, Whincup P, Papacosta O, Ebrahim S, Lennon L, Ueland PM, et al. The effect of C677T and A1298C polymorphisms in the methylenetetrahydrofolate reductase gene on homocysteine levels in elderly men and women from the British regional heart study. Atherosclerosis 2001;154:659–66. Syvänne M, Nieminen MS, Frick MH. Accuracy and precision of quantitative arteriography in the evaluation of coronary artery disease after coronary bypass surgery: a validation study. Int J Vasc Imaging 1994;10:243–52. Fruchart J-C, Brewer HB, Leitersdorf E. Consensus for the use of fibrates in the treatment of dyslipoproteinemia and coronary heart disease. Am J Cardiol 1998;81:912–7. Fruchart JC, Duriez P, Staels B. Peroxisome proliferator-activated receptor-alpha activators regulate genes governing lipoprotein metabolism, vascular inflammation and atherosclerosis. Curr Opin Lipidol 1999;10:245–57. Desouza C, Keebler M, McNamara D, Fonseca V. Drugs affecting homocysteine metabolism: impact on cardiovascular risk. Drugs 2002;62:605–16. Nygård O, Nordrehaug JE, Refsum H, Ueland PM, Farstad M, Vollset SE. Plasma homocysteine levels and mortality in patients with coronary artery disease. New Engl J Med 1997;337: 230–6. Gudnason V, Stansbie D, Scott J, Bowron A, Nicaud V, Humphries SE. on behalf of the EARS Group. C677T (thermolabile alanine/


M. Syvänne et al. / Atherosclerosis 172 (2004) 267–272

valine) polymorphism in methylenetetrahydrofolate reductase (MTHFR): its frequency and impact on plasma homocysteine concentration in different European populations. Atherosclerosis 1998;136: 347–54. [28] Brattstrom L, Wilcken DEL, Ohrvik J, Brudin L. Common methylenetetrahydrofolate reductase gene mutation leads to hyper-

homocysteinemia but not to vascular disease. Circulation 1998;98: 2520–6. [29] Dekou V, Gudnason V, Hawe E, Miller GJ, Stansbie D, Humphries SE. Gene–environment and gene–gene interaction in the determination of plasma homocysteine levels in healthy middle-aged men. Thromb Haemost 2001;85:67–74.