Lack of Benefit in Treating High Homocysteine Levels with Vitamins

Dafna Rippel, MD, MAH, Andrew P. Ignaszewski, MD, FRCPC, Jiri J. Frohlich, MD, FRCPC

Lack of benefit in treating high homocysteine levels with vitamins

Recent studies suggest that most patients at risk of cardiovascular disease will not be helped by homocysteine-lowering treatment with folic acid, B6, and B12.

ABSTRACT: Homocysteine-lowering n 1969, McCully suggested that after the reported lack of association vitamin therapy has been widely pre- elevated levels of plasma homo- between CAD and a specific genetic scribed in the belief that it will cysteine (Hcy) are associated polymorphism affecting MTHFR reduce cardiovascular events. While I with atherosclerosis.1 He noted (677C→T).8,9 Normally it would be epidemiological studies have indi- high prevalence of hyperhomocys- expected that if a particular substance cated a strong association between teinemia in patients with coronary in blood causes a disease and its level high plasma homocysteine levels artery disease (CAD) when compared is related to a genetic polymorphism, and risk of cardiovascular disease, with the general population. A num- then that polymorphism would also be several recent prospective trials of ber of other observations linking related to the disease. homocysteine lowering with folic homocysteine with CAD followed.2,3 While enzyme deficiencies caused

acid, vitamin B6, and vitamin B12 A meta-analysis in 1995 by Boushey by mutations in the CBS gene and the have failed to show any benefit in and colleagues suggested that for each MTHFR gene are related to very high patients with established cardio- 5 μmol/L increase in homocysteine levels of homocysteine with vascular vascular disease. Three recent pro- there was a 70% higher risk of CAD, consequences,9 a number of other spective randomized trials (VISP, a 50% increase in cerebral vascular causes of hyperhomocysteinemia (see NORVIT, and HOPE 2) suggest that disease, and a strong association with Table 1 ) are known to be proven risk measurement of homocysteine is peripheral vascular disease.4 Based on factors for vascular disease, and it is not required for the management of these data, the authors suggested that difficult to separate the effect of these most patients and that treating mod- 10% of the population’s CAD risk is factors from the specific conse- erate hyperhomocysteinemia with attributable to homocysteine. Another quences of hyperhomocysteinemia. In vitamins does not reduce cardiovas- meta-analysis done by Eikelboom and addition, blood sample handling, such cular events. colleagues confirmed these data, as inadequate chilling, can also namely that “there is a strong dose- increase plasma homocysteine levels. dependent positive association between plasma homocysteine levels Dr Rippel is a clinical trainee in the Healthy and risk for cardiovascular disease” Heart Program at St Paul’s Hospital, Van- that “is independent of other known couver, BC. Dr Ignaszewski is head of the risk factors.”5 While these and other Division of Cardiology at St. Paul’s Hospital studies revealed a strong epidemio- and medical director of the Healthy Heart logical correlation, they also empha- Program. Dr Frohlich is academic director sized the need for randomized clinical of the Healthy Heart Program, director of trials to establish causation.6,7 the St. Paul’s Lipid Clinic, and a professor of The first doubts about the causal pathology and laboratory medicine at the role of homocysteine were expressed University of British Columbia.

20 BC MEDICAL JOURNAL VOL. 49 NO. 1, JANUARY/FEBRUARY 2007 Lack of benefit in treating high homocysteine levels with vitamins

Three recent clinical trials Table 1. Causes of elevated homocysteine levels. Three large prospective trials using vitamin therapy in high-risk popula- Diabetes, rheumatoid arthritis, systemic lupus erythematosas, Chronic medical tions were conducted over the last sev- renal failure, malignant neoplasms and leukemia, hyperproliferative disorders and Table 2 disorders, severe psoriasis, hypothyroidism, acute-phase response to eral years (see ). conditions illness, cardiac and renal transplantation Vitamin Intervention for Stroke Pre- 10 vention (VISP). The aim of this trial Sex hormones, l-DOPA, isoniazid, some anticonvulsant agents was to determine whether the best (phenytoin, carbamazepine), cholesterol-lowering agents medical and surgical management, Drugs (cholestyramine, colestipol, nicotinic acid), metformin, thiazide diuretics, risk-factor modification, and high- cyclosporine, folate antagonists (methotrexate), vitamin B12 antagonists dose multivitamin therapy would (nitrous oxide), vitamin B6 antagonists reduce the incidence of recurrent cerebral infarction, coronary artery dis- Tobacco use, physical inactivity, obesity, stress ease, or death in patients with a non- Lifestyle factors disabling cerebral infarction and fasting High alcohol and coffee intake, folate, vitamin B , and vitamin B total homocysteine levels greater than Dietary factors 6 12 the 25th percentile for North Ameri- deficiencies; increased methionine consumption can stroke patients. A total of 3680 Demographic stroke patients were followed for 2 Increasing age, male gender, postmenopausal state years on average. Although the Hcy characteristics

Table 2. Main data and results of VISP, NORVIT, and HOPE 2 trials.

VISP10 NORVIT11 HOPE 212

Participants 3680 3749 5522

Follow-up (mean, in months) 24 36 60

• Nondisabling ischemic Vascular disease / cerebral infarct 120 Entry criteria <7 days from MI diabetes plus one days other risk factor • Hcy 25th percentile* >35 30–85 >55 Age (years) mean 66 mean 63 mean 69

Homocysteine reduction 18% 27% 18%–20%

1.14‡ Primary endpoints 1.00† 1.22§ 0.95 (relative risk reduction compared to placebo) 1.08 || 1.17‡ MI (RRR compared to placebo) 0.90† 1.23§ 0.98 1.06 || 0.83‡ Ischemic stroke (RRR compared to placebo) 1.00† 0.81§ 0.75 1.02 || Total 1.19‡ • CV death 0.96 Death (RRR compared to placebo) 0.90† 1.21§ • Total death 0.98

‡ *For North American stroke patient population B6 vs no B6 † High-dose vs low-dose vitamin therapy § Folic acid and B vitamin in combination vs placebo || Folic acid and B12 vs no folic acid and no B12

VOL. 49 NO. 1, JANUARY/FEBRUARY 2007 BC MEDICAL JOURNAL 21 Lack of benefit in treating high homocysteine levels with vitamins

Table 3. Vitamins and doses used in VISP, NORVIT, and HOPE 2 trials. We are thus faced with a paradox. While homocysteine levels predict the B12 B6 Folic acid likelihood of cardiovascular events 4,7,13 VISP10 and, for that matter, cardiovascular 13 High-dose formulation group 0.40 mg 25.0 mg 2.50 mg and noncardiovascular death, reduc- ing the level of homocysteine by up to Low-dose formulation group 0.006 mg 0.2 mg 0.02 mg 27% does not appear to influence NORVIT11 these events. There are several possi- A group 0.40 mg 40.0 mg 0.80 mg ble explanations. 14 B group 0.40 mg 0.80 mg Recently, Spence has argued that the major cause of hyperhomocys- C group — 40.0 mg — teinemia, particularly in the elderly, is D group — — — vitamin B12 deficiency, and that the 12 HOPE 2 dose of vitamin B12 used in all of these Treated group 1.0 mg 50.0 mg 2.50 mg trials (seeTable 3 ) was not sufficient. Another less likely possibility is that Placebo control group — — — the relatively small decreases in homocysteine are insufficient to influence level was reduced by 18% there was Heart Outcomes Prevention Evalua- outcomes. Perhaps the most plausible no effect on the primary endpoints of tion (HOPE) 2.12 A recent Canadian explanation is that while homocys- stroke, CAD, or death. However, the trial examined whether prolonged teine is a good marker of risk of either baseline homocysteine level was an administration of folic acid with vita- cardiovascular or total mortality, it is independent predictor of vascular mins B12 and B6 reduces the risk of not the root cause of the problem. events. major vascular events in persons at Other metabolic pathways that are high risk of cardiovascular events. In involved in homocysteine metabolism The Norwegian Vitamin Trial this study, 5522 patients with vascular should be explored for causal relation (NORVIT).11 This trial examined the disease or diabetes and another risk to the observed pathologies. potential benefits of vitamin B thera- factor were followed for 5 years on In a recent editorial, Loscalzo sug- py in patients with acute myocardial average. While treatment reduced the gested that vitamin therapy “has other, infarction. A total of 3749 patients Hcy level by 18% to 20%, more par- potentially adverse effects that offset who had myocardial infarction within ticipants in the vitamin treatment its homocysteine-lowering benefits.”15 7 days of beginning therapy were fol- group were hospitalized for unstable He proposed three possible mecha- lowed for 5 years on average. The pri- angina. As in the VISP and NORVIT nisms: promotion of cell proliferation mary endpoints were recurrent myo- studies, there was no effect on the in the plaque by folic acid through its cardial infarction, stroke, or sudden study’s vascular endpoints (death from role in the synthesis of thymidine; death. Treatment with folic acid and cardiovascular causes, myocardial increased methylation potential by vitamin B12 reduced the Hcy level by infarction, or stroke). And once again, folic acid and vitamin B12 (including 27% on average (from 13 to 9.6 the baseline homocysteine level was a DNAmethylation), which could result μmol/L), but had no effect on the end- statistically significant predictor of in promoting development of plaques; points. On the contrary, in the group cardiovascular events. and inhibition of nitric oxide synthase treated with folic acid, vitamin B12, as a result of methylation of l-arginine and B6, there was a trend toward an Possible reasons for to asymmetric dimethylarginine. increased risk of vascular events, results although this was only observed in While the epidemiological studies Two important lessons patients with homocysteine levels clearly associate homocysteine level What can we learn from these studies greater than 13 μmol/L at baseline. As with cardiovascular outcomes,5 the and observations? The first lesson is in the VISP study, the baseline homo- results of the three clinical trials des- that association is not the same as cau- cysteine level was a significant pre- cribed here strongly suggest that treat- sation. While homocysteine has been dictor of vascular events. ment with vitamins does not improve proven to act as a predictor of vascular outcomes. events, lowering homocysteine levels

22 BC MEDICAL JOURNAL VOL. 49 NO. 1, JANUARY/FEBRUARY 2007 Lack of benefit in treating high homocysteine levels with vitamins

has not been shown to influence out- Although the hypothesis that high 9. Scott JM. Modification of hyperhomo- comes. We should not plunge into homocysteine levels contribute to cystenemia. In: Carmel R, Jacobsen DW treatment of conditions or metabolites atherogenesis cannot be rejected, (eds). Homocysteine in Health and Dis- that are associated with specific out- treating mild to moderate hyperhomo- ease. Cambridge: Cambridge University comes until there is a clear causal rela- cysteinemia with folic acid and B vit- Press; 2001: 467-476. tionship. In retrospect, the fact that the amins is clearly not indicated based 10. Toole JF, Malinow MR, Chambless LE, et MTHRF polymorphism itself was not on the results of recent clinical trials. al. Lowering homocysteine in patients associated with vascular events should with ischemic stroke to prevent recurrent have been a warning to us not to start Competing interests stroke, myocardial infarction, and death: vitamin treatment in patients with mild None declared. The Vitamin Intervention for Stroke Pre- to moderate hyperhomocysteinemia. vention (VISP) randomized controlled A second lesson is that we should References trial. JAMA 2004;291:565-575. not assume absence of harm when pre- 1. McCully KS. Vascular pathology of homo- 11. Bonaa KH, Njolstad I, Ueland PM, et al.; scribing a combination of “harmless” cysteinemia: Implications for the patho- NORVIT Trial Investigators. Homocys- substances in accepted dosages— genesis of arteriosclerosis. Am J Pathol teine lowering and cardiovascular events something already learned with vita- 1969;56:111-128. after acute myocardial infarction. N Engl min E.12,16,17 While we tried to justify 2. Gauthier GM, Keevil JG, McBride PE. The J Med 2006;354:1578-1588. using vitamin therapy for lowering association of homocysteine and coro- 12. The Heart Outcomes Prevention Evalua- Hcy by assuming that vitamins are not nary artery disease. Clin Cardiol 2003; tion (HOPE) 2 Investigators. Homocys- toxic and that any treatment based on 26:563-568. teine lowering with folic acid and B vita- them is not harmful, there are sugges- 3. Wald DS, Law M, Morris JK. Homocys- mins in vascular disease. N Engl J Med tions from both the NORVIT and teine and cardiovascular disease: Evi- 2006;354:1567-1577. HOPE 2 studies that in fact some of dence on causality from a meta-analysis. 13. Lee KWJ, Hill JS, Walley KR. Relative the study conditions might have been BMJ 2002;325:1202-1206. value of multiple plasma biomarkers as precipitated or worsened by the vita- 4. Boushey CJ, Beresford SA, Omenn GS, risk factors for coronary artery disease min treatment. et al. A quantitative assessment of plas- and death in an angiography cohort. ma homocysteine as a risk factor for vas- CMAJ 2006;174:461-466. Clinical implications cular disease. Probable benefits of in- 14. Spence D. Homocysteine: Call off the Should we measure homocysteine and creasing folic acid intakes. JAMA 1995; funeral [editorial]. J Stroke 2006;37:282- treat high levels with vitamins? First, 274:1049-1057. 283. while the homocysteine level is a sen- 5. Eikelboom JW, Lonn E, Genest J Jr, et al. 15. Loscalzo J. Homocysteine trials—clear sitive indictor of vitamin B12 deficien- Homocyst(e)ine and cardiovascular dis- outcomes for complex reasons. N Engl J cy, it should not replace measurement ease: A critical review of the epidemio- Med 2006;354:1629-1632. of vitamin B12. Second, treatment of logic evidence. Ann Intern Med 1999; 16. Lee IM, Cook NR, Gaziano JM, et al. Vit- very high levels, such as those seen in 131:363-375. amin E in the primary prevention of car- homocysteinuria, should be consid- 6. Homocysteine Lowering Trialists’ Collab- diovascular disease and cancer: The ered since these levels are the cause of oration. Lowering blood homocysteine Women’s Health Study: A randomized underlying vascular events in this dis- with folic acid based supplements: Meta- controlled trial. JAMA 2005;294:56-65. order and are known to respond to vit- analysis of randomised trials. BMJ 17. Miller ER 3rd, Pastor-Barriuso R, Dala Dl, amin treatment.18 Third, it may still be 1998;21:894-898. et al. Meta-analysis: High-dosage vitamin of value to measure homocysteine lev- 7. The Homocysteine Studies Collabora- E supplementation may increase all- els in those rare patients with severe tion. Homocysteine and risk of ischemic cause mortality. Ann Intern Med vasculopathy and no obvious tradi- heart disease and stroke: A meta-analy- 2005;142:37-46. tional risk factors. Similarly, measure- sis. JAMA 2002;288:2015-2022. 18. Mudd SH, Levy HL, Krause JP. Disorders ment may be of use in patients with a 8. Lewis SJ, Ebrahim S, Davey Smith G. of transsulfuration. In: Scriver CR, Sly WS, disastrous family history of vascular Meta-analysis of MTHFR 677C_T poly- Beaudet AL, et al. (eds). The Metabolic disease. In these patient categories, if morphism and coronary heart disease: and Molecular Basis of Inherited Disease. homocysteine values are very high Does totality of evidence support causal 8th ed. New York: McGraw Hill; 2001: (i.e., >20–30 μmol/L) treatment may role for homocysteine and preventive po- 2007-2056. be considered. tential of folate? BMJ 2005;5:331:1058.

VOL. 49 NO. 1, JANUARY/FEBRUARY 2007 BC MEDICAL JOURNAL 23