Effect of folic acid on fenofibrate-induced elevation of homocysteine and cysteine

Effect of folic acid on fenofibrate-induced elevation of homocysteine and cysteine Vojtech Melenovsky, MD, PhD,a,c Tomas Stulc, MD,a Viktor Kozich, MD, PhD,b Barbora Grauova, MD, PhD,a Jakub Krijt, PhD,b Dan Wichterle, MD,a Tomas Haas, PhD,a Jan Malik, MD, PhD,a Jaromir Hradec, MD, PhD, FESC,a and Richard Ceska, MD, PhDa Prague, Czech Republic, and Baltimore, Md

Background

An elevated total plasma homocysteine (tHcy) level is considered to be an independent risk factor for atherosclerosis. It has been reported that lipid-lowering therapy with fibric acid derivatives (fibrates) increases tHcy and total plasma cysteine (tCys) levels. The aim of this study was to determine whether therapy with folic acid, a potent tHcylowering agent, could modify the fenofibrate-induced elevation of plasma aminothiols.

Methods Patients with combined hyperlipidemia (n ⫽ 37) were randomized to receive 9 weeks of treatment with micronized fenofibrate 200 mg/day (F group) or fenofibrate 200 mg/day plus folic acid 10 mg/every other day (F⫹F group). tCys and tHcy levels were determined before and after the therapy with high performance liquid chromatography. Results

The tHcy level increased significantly in the F group by 51.3% and in the F⫹F group by 14.6% (betweengroup difference P ⫽ .001). Total plasma cysteine (tCys) increased similarly after both treatments (P ⫽ .72). The serum creatinine level increased in the F group by 20.7% and in F⫹F group only by 9.8% (P ⫽ .04). The increase of tHcy level in F group correlated with an increase of tCys and creatinine levels (r ⫽ 0.74 and 0.64, respectively). The effects on the lipid profile did not differ by treatment group.

Conclusions Folic acid effectively reduces the fenofibrate-induced elevation of tHcy and creatinine, but it does not affect the elevation of the tCys. Folic acid has neutral effect on the lipid-lowering action of fenofibrate. Clinical efficacy of fenofibrate might be improved by folic acid coadministration. (Am Heart J 2003;146:1.) Epidemiological studies demonstrated that an elevated total plasma homocysteine (tHcy) level is associated with an increased risk of all forms of atherosclerotic vascular disease.1 A higher risk of atherosclerosis has also been recently found in subjects with increased levels of plasma total cysteine (tCys).2,3 The vascular toxicity of aminothiols may be produced by oxidative injury of endothelium by reactive oxygen species generated during tHcy and tCys auto-oxidation, leading to endothelial dysfunction.4 – 6 However, the causal relationship between atherosclerosis and plasma aminothiols still needs to be established.

From the aThird Department of Internal Medicine, and bInstitute of Inherited Metabolic Diseases, First Faculty of Medicine, Charles University, Prague, Czech Republic, and c Division of Cardiology, Department of Medicine, The Johns Hopkins Medical Institutions, Baltimore, Md. Supported by the research project (J13/98 1111 0000 2-1) and grants (IGA 5986-3/ 2000, 6134-3/2000, and 6548-3) from the Ministry of Education and Ministry of Health, Czech Republic. Submitted June 12, 2002; accepted November 13, 2002.

Reprint requests: Vojtech Melenovsky, MD, PhD, Division of Cardiology, Department of Medicine, The Johns Hopkins Hospital, 600 N Wolfe St, Carnegie 534, Baltimore, MD 21287. E-mail: [email protected]. © 2003, Mosby, Inc. All rights reserved. 1097-6744/2003/$30.00 ⫹ 0 doi:10.1016/S0002-8703(03)00122-4

Fibrates (fibric acid derivatives) are lipid-lowering drugs that are used in patients with elevated triglyceride levels, low high-density lipoprotein levels, or both. The effects of fibrates are mediated through the activation of nuclear PPAR␣ (peroxisome proliferator-activated receptor ␣) receptors, leading to increased cellular free fatty acid catabolism, decreased hepatic very low-density lipoprotein secretion, and increased lipolysis by lipoprotein lipase.7 It has been reported that therapy with fibrates leads to an elevation of tHcy level8 –14 by 20% to 50% and also to an elevation of tCys level10,15,16 by an unknown mechanism. It is known that administration of folic acid, B vitamins (B6 and B12), or both reduces tHcy level by 15% to 25%, even in the absence of vitamin deficiency.17 The principal aim of this study was to determine whether the administration of folic acid could modify the fibrate-induced elevation of plasma tHcy and tCys levels.

Methods Patients Patients with combined hyperlipidemia (total cholesterol level ⱖ240 mg/dL and triglyceride level ⱖ200 mg/dL) indicated for fibrate therapy were recruited

American Heart Journal July 2003

2 Melenovsky et al

from a university-based lipid clinic. There were no age or sex restrictions. Patients with diabetes mellitus, uncontrolled thyroid disease (thyroid-stimulating hormone level ⬍0.5 or ⬎7.0 mIU/L), renal insufficiency (serum creatinine level ⬎1.7 mg/dL), macrocytosis (mean cell volume ⬎100 fl), B12 deficiency (B12 level ⬍200 ng/L), or malignancies were excluded. Drugs known to influence tHcy metabolism (including multivitamin supplements) and any hypolipidemic medications were discontinued at least 4 weeks before the start of the study. Patients were asked to maintain their regular dietary, smoking, and drinking habits throughout the course of the study. Wheat and common cereal products are not fortified with folic acid in the Czech Republic. The institutional ethics committee approved the study protocol, and all patients signed an informed consent form.

Study design This trial was a single-center, randomized, prospective study, consisting of 2 parallel arms of treatment. Patients were assigned to treatment group by adaptive random allocation procedure for age, sex, cigarette smoking status, and initial levels of folic acid and triglycerides.18 Patients in the fenofibrate group (F group) received micronized fenofibrate only (200 mg/day, Lipanthyl 200M, Laboratoires Fournier), patients in the fenofibrate plus folate group (F⫹F group) received fenofibrate (200 mg/day) and folic acid (10 mg every other day, Acidum folicum, Leciva). The length of treatment was 9 weeks. Because th highest dose of folic acid recommended for treatment of hyperhomocysteinemia is 5 mg/day17 and the only available formulation in Czech Republic contains 10 mg of folic acid per tablet, we decided to administer 10 mg every other day. This administration scheme completely saturates body folate stores.19

Total serum cholesterol, high-density lipoprotein cholesterol, creatinine, and triglyceride levels were determined enzymatically with a multianalyzer HITACHI 717 and commercially available kits. Full blood counts were analyzed by an automatic analyzer (Coulter Counter STKF). Plasma folate levels were determined with an enzyme-linked immunosorbent assay (Abbott Laboratories, Abbott Park, Ill). Serum levels of B12 were determined by a chemiluminiscence method with kits from Bayer.

Statistical methods The primary end point was the tHcy level at the end of study. The sample size needed for the detection of 5 ␮mol/L difference in post-treatment tHcy level was estimated to be 17 persons per study arm (power ⬎80% and significance level ⬍5%). The estimates of the difference and variability were made on the basis of reported effects of fenofibrate8 and folic acid17 on tHcy. The between-group differences before and after treatment were tested with the Wilcoxon rank sum test or Fisher exact test (for categorical variables). The difference in tCys level at the end of study was tested with non-parametric analysis of covariance, with the baseline level as a covariate. Within-group changes were tested with the Wilcoxon matched pairs test. Pearson correlation coefficients were calculated between the changes of individual variables for both treatment arms. The values are expressed as means plus or minus SDs or as medians plus or minus interquartile ranges (IQRs) for variables with significantly skewed distribution (Shapiro-Wilk test). The changes of variables are expressed as mean percentage. A 2-sided hypothesis was used for all tests, and a P value ⬍.05 was considered to be significant.

Results Assessments Laboratory examinations were performed at the beginning and at the end of the treatment period. Blood samples were taken after overnight fasting and were directly processed. The plasma for analysis of aminothiol compounds was separated from blood cells by centrifugation at 2000 g for 15 minutes at 4°C, and samples were immediately stored at ⫺75°C until analysis. Plasma tHcy and tCys levels were determined with high performance liquid chromatography (Shimadzu LC-10A system) after reduction of disulfides and protein-bound aminothiols with tris(2-carboxyethyl)phosphine and derivatization with fluorogenic reagent SBD-F (ammonium-7-fluorobenzo-2-OXA-1,3-diazole-4sulfonate).20 The coefficients of variation (CV) for the assessment of aminothiols were ⬍3% (intraassay CV) and ⬍4% (interassay CV).

All 37 patients who were enrolled completed the study, 19 in the F group and 18 in the F⫹F group. The length of treatment was 65 ⫾ 18 days in the F group and 63 ⫾ 11 days in the F⫹F group. No significant adverse effects were observed during the study. Baseline clinical characteristics of both groups are shown in Table I. The comparison of biochemical variables at the baseline is presented in Table II. Of all the tested parameters, the groups differed at the baseline only in tCys level (P ⫽ .01).

Plasma aminothiol levels and B vitamins The effects of both treatments on biochemical variables are summarized in Table III. In the F group, the level of plasma tHcy increased by 51.3% (P ⬍.001). In the F⫹F group, the tHcy level increased only by 14.6% (P ⬍.05). The between-group difference in tHcy con-

American Heart Journal Volume 146, Number 1

Melenovsky et al 3

Table I. Baseline clinical characteristics of patients Fenofibrate (n ⴝ 19)

Parameter Sex (m/f) Age (y) Body mass index at baseline (kg/m2) Coronary heart disease (%) Peripheral vascular disease (%) Cerebrovascular disease (%) Hypertension (%) Current smoker (%) Previous lipid lowering therapy (%)

12/7 50.1 ⫾ 9.6 29.6 ⫾ 4.8 1 (5) 0 (0) 0 (0) 12 (63) 3 (16) 8 (42)

Fenofibrate ⴙ folate (n ⴝ 18)

P*

10/8 50.3 ⫾ 10 27.7 ⫾ 2.9 1 (5) 1 (5) 0 (0) 8 (44) 6 (33) 5 (28)

.74 .96 .25 1.00 .49 1.00 .33 .27 .49

Values are means ⫾ SD or absolute numbers and proportions. *Wilcoxon rank sum test or Fisher exact test.

Table II. Comparison of biochemical variables at the baseline Parameter Homocysteine (␮mol/L)† Cysteine (␮mol/L)† Folate (ng/mL)† Vitamin B12 (pg/mL)‡ Mean red cell volume (fL)‡ Creatinine (mg/dL)† Uric acid (mg/dL)† Total cholesterol (mg/dL)‡ HDL cholesterol (mg/dL)‡ Triglycerides (mg/dL)‡

Fenofibrate (n ⴝ 19) 11.5 ⫾ 3.0 321 ⫾ 32 2.5 ⫾ 1.0 287 ⫾ 128 90 ⫾ 4.9 1.0 ⫾ 0.1 7.0 ⫾ 1.7 289 ⫾ 58 46 ⫾ 7.7 380 ⫾ 292

Fenofibrate ⴙ folate (n ⴝ 18)

P

10.4 ⫾ 1.9 292 ⫾ 34 2.5 ⫾ 0.7 298 ⫾ 120 91 ⫾ 4.7 1.0 ⫾ 0.1 6.6 ⫾ 1.8 278 ⫾ 73 46 ⫾ 7.7 389 ⫾ 239

.32 .01 .75 .61 .42 .89 .37 .49 .89 .46

*P value for between-group difference at the baseline (Wilcoxon rank sum test). †Values are mean ⫾ SD. ‡Values are medians ⫾ IQR.

centration was highly significant at the end of study (P ⫽ .001). The plasma tCys level increased in the F group (by 14.2%, P ⬍.001) and in the F⫹F group (by 17.6%, P ⬍.001), without a significant between-group difference at the final visit (P ⫽ .72). Individual case profiles of tHcy and tCys levels are presented in Figures 1 and 2. The increase of the tHcy level correlated positively with the increase of the tCys level (r ⫽ 0.74, P ⬍.001) and creatinine level (r ⫽ 0.64, P ⬍.01) in the F group. The increase of the tHcy level correlated positively with the increase of the tCys level (r ⫽ 0.49, P ⬍ .05) in the F⫹F group. Plasma folate levels did not change in the F group, but expectedly increased by 167% in the F⫹F group (P ⬍.001). A significant increase of vitamin B12 levels was observed in the F group (by 37%, P ⬍.01). The increase of B12 in the F⫹F group was not significant, nor was the difference in B12 levels at the final visit. The mean red cell corpuscular volume did not change in either group.

Figure 1

Total plasma homocysteine level (␮mol/L) in subjects (n ⫽ 37) before and after 9 weeks of treatment with micronized fenofibrate or fenofibrate and folic acid in combination.

American Heart Journal July 2003

4 Melenovsky et al

Table III. Comparison of biochemical variables at the end of study Fenofibrate ⴙ folate

Fenofibrate Parameter Homocysteine (␮mol/L)# Cysteine (␮mol/L)# Folate (ng/mL)# Vitamin B12 (pg/mL)** Mean red cell volume (fl)** Creatinine (mg/dL)# Uric acid (mg/dL)# Total cholesterol (mg/dL)** HDL cholesterol (mg/dL)** Triglycerides (mg/dL)**

(n ⴝ 19)

% change

(n ⴝ 18)

% change

P*

17.5 ⫾ 6.5 366 ⫾ 53 2.2 ⫾ 0.8 304 ⫾ 137 88 ⫾ 3.0 1.2 ⫾ 0.2 5.0 ⫾ 1.8 220 ⫾ 50 46 ⫾ 12 124 ⫾ 71

51.3‡ 14.2‡ 6.4 37.0¶ –0.5 20.7‡ –27.8† –21.3‡ 1.6 –48.3¶

11.7 ⫾ 2.5 340 ⫾ 30 6.3 ⫾ 1.7 352 ⫾ 132 91 ⫾ 3.0 1.1 ⫾ 0.1 4.6 ⫾ 1.0 216 ⫾ 69 42 ⫾ 19 178 ⫾ 124

14.6§ 17.6‡ 167‡ 23.1 –0.1 9.8§ –29.6‡ –13.7‡ 1.7 –39.1‡

.001 .72† ⬍.001 .92 .09 .04 .13 .89 .94 .23

*P value for between-group difference at the end of study (Wilcoxon rank sum test). †P value for between-group difference at the end of study (ANCOVA). ‡P value versus baseline ⬍.001. §P value versus baseline ⬍.05. ¶P value versus baseline ⬍.01. #Values are mean ⫾ SD. **Values are median ⫾ IQR.

Figure 2

Total plasma cysteine level (␮mol/L) in subjects (n ⫽ 37) before and after 9 weeks of treatment with micronized fenofibrate or fenofibrate and folic acid in combination.

Figure 3

Serum creatinine level (mg/dL) in subjects (n ⫽ 37) before and after 9 weeks of treatment with micronized fenofibrate or fenofibrate and folic acid in combination.

Discussion Effects on blood lipids, creatinine, and uric acid The levels of total cholesterol and triglycerides decreased similarly after both treatments. The serum creatinine level increased by 20.7% in the F group (P ⬍.001) and by 9.8% in the F⫹F group (P ⬍.05). The between-group difference was significant (P ⫽ .04) at the end of the study. Individual case profiles of creatinine levels are shown in the Figure 3. The concentration of uric acid decreased in the F and the F⫹F groups to a similar extent, by 27.8% and 29.6%, respectively (both P ⬍.001).

The principal finding of this study is that folic acid substantially attenuates hyperhomocysteinemia induced by fenofibrate. These results confirm a preliminary observation obtained with a small number of subjects.21 Our results are also in agreement with a recent study22 in which a combination of vitamins B6 and B12 and folic acid prevented tHcy elevation after fenofibrate. In the F group of our study, the proportion of subjects with hyperhomocysteinemia (defined as tHcy ⱖ15 ␮mol/L) increased from 21% to 63% of subjects at the final visit, so that the increase of tHcy after fenofibrate administration is considerable. The mechanism of

American Heart Journal Volume 146, Number 1

fibrate-induced tHcy elevation is not known, but it is absent in fibrate treated PPAR␣-deficient mice,23 so that it is specifically dependent of PPAR␣ receptors. It has been shown in previous studies9,14,15,22 that fibrate therapy does not decrease plasma levels of folic acid, vitamin B12, or both. Also, it seems unlikely that fibrates could produce greater intracellular deficiency of active folate metabolites, because no change of mean read blood cell corpuscular volume was observed in our study. The therapy with fenofibrate was accompanied by a significant increase of tCys level, which confirms the findings of previous reports.15,16 An original finding of our study is that a fenofibrate-induced tCys increase —in contrast to tHcy—is not affected by folic acid administration. Correspondingly, it has been found in a cross-sectional population study that some main factors influencing tHcy levels, such as folate and vitamin intake, were not associated with plasma tCys levels.24 The mechanisms of fibrate-induced tCys elevation and reasons for different response to folate therapy are not known. Another new finding of our study is that fibrate-induced elevation of serum creatinine levels is influenced by folic acid therapy. An increase of serum creatinine after fibrate administration has been consistently observed in previous studies.9,14,15,22 It is known that the kidneys play a major role in plasma tHcy clearance. Patients with renal failure have hyperhomocysteinemia, and markers of glomerular filtration rate (like serum creatinine) correlate with tHcy, even in subjects with normal renal function.25,26 Fibrateinduced impairment of renal function was suggested as an explanation of tHcy increase.9,22 However, the increase of serum creatinine after fibrate administration was not accompanied by a reduction of the glomerular filtration rate assessed with an isotope technique27 or by PAH and insulin clearances.28 Moreover, the fibrateinduced increase of creatinine level was associated with an increase of urinary creatinine excretion,11 which suggests that fibrates may stimulate creatinine synthesis. The pathways of creatinine and tHcy are connected, and this link might explain tHcy elevation after fibrate administration. The proximal precursor of creatinine (phosphocreatine) is synthesized by a transfer active methyl group from S-adenosylmethionine (SAM) to guanidinoacetate. S-adenosylhomocysteine (SAH)—a byproduct of this reaction, is then hydrolyzed to Hcy. Hcy is then converted back to SAM in methylation cycle (Hcy 3 methionine 3 SAM 3 SAH 3 Hcy), in which folate serves as a donor of methyl group.29 Increased creatinine synthesis may lead in this way to a higher production of Hcy. Simultaneously, the use of Hcy in the metylation cycle may be also impaired by a relative depletion of active methyl groups consumed by creatinine synthesis. The observa-

Melenovsky et al 5

tion of the parallel effect of folic acid on creatinine and tHcy levels may indirectly support this hypothesis. In conclusion, we demonstrated that folic acid significantly reduced the elevation of tHcy and creatinine after fenofibrate administration, and it had a neutral effect on the lipid-lowering action of this drug. If it is proven by ongoing trials that tHcy lowering prevents cardiovascular events,30 treatment of the fibrate-induced hyperhomocysteinemia might improve the efficacy of fibrates in the reduction of cardiovascular mortality. This hypothesis would be of considerable clinical importance especially in patients with type 2 diabetes mellitus, because diabetes mellitus is often associated with hyperhomocysteinemia31 and fibrates are frequently used for the treatment of diabetic dyslipidemia.

References 1. Gerhard GT, Duell PB. Homocysteine and atherosclerosis. Curr Opin Lipidol 1999;10:417–28. 2. Jacob N, Bruckert E, Giral P, et al. Cysteine is a cardiovascular risk factor in hyperlipidemic patients. Atherosclerosis 1999;146: 53–9. 3. El-Khairy L, Ueland PM, Refsum H, et al. Plasma total cysteine as a risk factor for vascular disease: the European Concerted Action Project. Circulation 2001;29:103:2544 –9. 4. Chambers JC, McGregor A, Jean-Marie J, et al. Demonstration of rapid onset vascular endothelial dysfunction after hyperhomocysteinemia: an effect reversible with vitamin C therapy. Circulation 1999;99:1156 – 60. 5. Jacobsen DW Hyperhomocysteinemia and oxidative stress: time for a reality check? Arterioscler Thromb Vasc Biol 2000;20:1182– 4. 6. Hogg N. The effect of cyst(e)ine on the auto-oxidation of homocysteine. Free Radic Biol Med 1999;27:28 –33. 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. 8. Melenovsky V, Malik J, Wichterle D, et al. Fenofibrate, but not atorvastatin treatment increases homocystein levels in patients with combined hyperlipidaemia [abstract]. Eur Heart J 2000;21(Suppl S):P3502. 9. Dierkes J, Westphal S, Luley C. Serum homocysteine increases after therapy with fenofibrate or bezafibrate. Lancet 1999;354:219 – 20. 10. de Lorgeril M, Salen P, Paillard F, et al. Lipid-lowering drugs and homocysteine. Lancet 1999;353:209 –10. 11. Jonkers IJ, de Man FH, Onkenhout W. Implication of fibrate therapy for homocysteine. Lancet 1999;354:1208. 12. Schaefer EJ, McNamara JR, Selhub J, et al. Gemfibrozil treatment raised Homocysteine concentration in VA-HIT [abstract]. Circulation 2000;102:II-847. 13. Harats D, Yodfat O, Doolman R, et al. Homocysteine elevation with fibrates: is it a class effect? Isr Med Assoc J 2001;3:243– 6. 14. Westphal S, Dierkes J, Luley C. Effects of fenofibrate and gemfibrozil on plasma homocysteine. Lancet 2001;358:39 – 40. 15. Giral P, Bruckert E, Jacob N, et al. Homocysteine and lipid lowering agents: a comparison between atorvastatin and fenofibrate in

6 Melenovsky et al

16.

17.

18.

19. 20.

21.

22.

23.

patients with mixed hyperlipidemia. Atherosclerosis 2001;154: 421–7. Bissonnette R, Treacy E, Rozen R, et al. Fenofibrate raises plasma homocysteine levels in the fasted and fed states. Atherosclerosis 2001;155:455– 62. Homocysteine Lowering Trialists’ Collaboration. Lowering blood homocysteine with folic acid based supplements: meta-analysis of randomised trials. BMJ 1998;316:894 – 8. Pocock SJ, Simon R. Sequential treatment assignment with balancing for prognostic factors in the controlled clinical trial. Biometrics 1975;31:103–15. Loew D, Eberhardt A, Heseker H, et al. Plasma kinetics and elimination of folic acid. Klin Wochenschr 1987;65:520 – 4. Krijt J, Vackova M, Kozich V. Measurement of homocysteine and other aminothiols in plasma: advantages of using tris(2-carboxyethyl)phosphine as reductant compared with tri-nutylphosphine. Clin Chem 2001;47:1821– 8. Stulc T, Melenovsky V, Grauova B, et al. Folate supplementation prevents plasma homocysteine increase after fenofibrate therapy. Nutrition 2001;17:721–3. Dierkes J, Westphal S, Kunstmann S, et al. Vitamin supplementation can markedly reduce the homocysteine elevation induced by fenofibrate. Atherosclerosis 2001;158:161– 4. Legendre C, Causse E, Chaput E, et al. Fenofibrate induces a selective increase of protein-bound homocysteine in rodents: a PPAR␣-mediated effect. Biochem Biophys Res Commun 2002;295: 1052– 6.

American Heart Journal July 2003

24. El-Khairy L, Ueland PM, Nygard O, et al. Lifestyle and cardiovascular disease risk factors as determinants of total cysteine in plasma: the Hordaland Homocysteine Study. Am J Clin Nutr 1999;70: 1016 –24. 25. Brattstrom L, Lindgren A, Israelsson B, et al. Homocysteine and cysteine: determinants of plasma levels in middle-aged and elderly subjects. J Intern Med 1994;236:633– 41. 26. Bostom AG, Bausserman L, Jacques PF, et al. Cystatin C as a determinant of fasting plasma total homocysteine levels in coronary artery disease patients with normal serum creatinine. Arterioscler Thromb Vasc Biol 1999;19:2241– 4. 27. Goffin E, Jamar F, Desager JP, et al. Implication of fibrate therapy for homocysteine. Lancet 1999;354:1209. 28. Hottelart C, el Esper N, Achard JM, et al. Fenofibrate increases blood creatinine, but does not change the glomerular filtration rate in patients with mild renal insufficiency. Nephrologie 1999;20: 41– 4. 29. Jacobsen DW. Determinants of hyperhomocysteinemia: a matter of nature and nurture. Am J Clin Nutr 1996;64:641–2. 30. Mangoni AA, Jackson SH. Homocysteine and cardiovascular disease: current evidence and future prospects. Am J Med 2002;112: 556 – 65. 31. Hoogeveen EK, Kostense PJ, Beks PJ, et al. Hyperhomocysteinemia is associated with an increased risk of cardiovascular disease, especially in non-insulin-dependent diabetes mellitus: a populationbased study. Arterioscler Thromb Vasc Biol 1998;18:133– 8.