U. S. Food and Drug Administration
Center for Food Safety and Applied Nutrition
Office of Nutritional Products, Labeling, and Dietary Supplements
November 28, 2000

Letter Regarding Dietary Supplement
Health Claim for Folic Acid,
Vitamin B₆, and Vitamin B₁₂
and Vascular Disease

(Docket No. 99P-3029)

Jonathan W. Emord, Esq.
Emord & Associates, P.C.
1050 Seventeenth Street, NW
Suite 600
Washington, DC 20036

Re: Petition for Health Claim: Folic Acid, Vitamin B₆, and Vitamin B₁₂ Dietary Supplements and Vascular Disease [Docket No. 99P-3029]

Dear Mr. Emord:

This letter is in further response to your petition of May 25, 1999, requesting the Food and Drug Administration (FDA) to authorize the health claim "As part of a well-balanced diet, rich in fresh fruits and vegetables, daily intake of at least 400 µg folic acid, 3 mg vitamin B₆ and 5 µg vitamin B₁₂ may reduce the risk of vascular disease." FDA has reevaluated the petition in response to the court decision directing the FDA to consider qualified health claims for dietary supplement labeling (Pearson v. Shalala, 164 F.3d 650 (D.C. Cir. 1999)) when the evidence in support of the claim does not meet the significant scientific agreement standard.

I. Procedure and Standard for Evaluating the Claim

In reconsidering this claim, FDA proceeded as described in the October 6, 2000, Federal Register notice entitled "Food Labeling; Health Claims and Label Statements for Dietary Supplements; Update to Strategy for Implementation of Pearson Court Decision." 65 Fed. Reg. 59,855 (2000). FDA previously had evaluated your petitioned claim under the "significant scientific agreement" standard by which the health claim regulations require the agency to evaluate the scientific validity of claims. Under this standard, FDA may issue a regulation authorizing a health claim only "when it determines, based on the totality of publicly available scientific evidence (including evidence from well-designed studies conducted in a manner which is consistent with generally recognized scientific procedures and principles), that there is significant scientific agreement, among experts qualified by scientific training and experience to evaluate such claims, that the claim is supported by such evidence." 21 C.F.R. § 101.14. FDA informed you by letter (copy enclosed) on November 30, 1999 (the November 1999 letter) that the evidence in the petition did not meet the significant scientific agreement standard, and the agency denied the petition.

For claims that do not meet the significant scientific agreement standard, FDA must consider, under Pearson, whether to allow a qualified claim about the substance-disease relationship. Because current FDA regulations do not provide for qualified claims, FDA must also consider the circumstances under which it will exercise enforcement discretion for a qualified claim. Consistent with the Pearson decision, the agency considers whether consumer health and safety would be threatened by the qualified claim, and, if not, whether the evidence in support of the claim is outweighed by evidence against the claim, either quantitatively or qualitatively. See 164 F.3d at 650, 659 & n.10. If the evidence for the claim outweighs the evidence against the claim, and there is no health or safety threat, the agency considers whether a qualified claim can meet the general health claim requirements of 21 C.F.R. § 101.14, other than the requirement to meet the significant scientific agreement standard and the requirement that the claim be made in accordance with an authorizing regulation. These requirements were not challenged in Pearson and therefore still apply.

In the October 6 notice, FDA explained that it would consider exercising enforcement discretion for a dietary supplement health claim that did not meet the significant scientific agreement standard if the scientific evidence for the claim outweighed the scientific evidence against the claim, if the claim included appropriate qualifying language, and if the other criteria listed in the notice were met. In that event, the agency explained, FDA would send a letter to the petitioner outlining the agency's rationale for its determination that the evidence did not meet the significant scientific agreement standard and stating the conditions under which the agency would ordinarily expect to exercise enforcement discretion for the claim. See 65 Fed. Reg. at 59,856. The agency also stated that, conversely, if the scientific evidence for the claim did not outweigh the scientific evidence against the claim, or the substance posed a threat to health, or the other criteria for the exercise of enforcement discretion were not met, FDA would issue a letter denying the claim and explaining its reasons for doing so. See 65 Fed. Reg. at 59,856.

The deadlines for FDA action in 21 C.F.R. § 101.70(j) apply to petitions for health claims. FDA initially responded to your petition within the prescribed 190 day time frame for issuing either a denial or a proposed regulation to authorize the health claim. In its November 1999 letter, the agency stated that after FDA completed a rulemaking to reconsider the general health claim regulations for dietary supplements, it would reconsider the petition. The agency subsequently agreed to decide by November 24, 2000, whether it would allow dietary supplement manufacturers to make the health claim with a qualifying statement. November 24 is the 540th day after the filing of the Vitamin B petition pursuant to 21 C.F.R. § 101.70(j). Following this agreement, a joint motion for a stay was filed in a lawsuit brought on your clients' behalf against the agency involving this petition as well as a petition for a health claim involving Vitamin E Dietary Supplements and Heart Disease (Docket No. 99-4375). The court granted the motion, and stayed the litigation until November 24, 2000. FDA is issuing this decision, consistent with its updated implementation strategy, on November 28, 2000.

II. Summary of Review

FDA reviewed the evidence in your petition for the claim, "As part of a well-balanced diet, rich in fresh fruits and vegetables, daily intake of at least 400 µg folic acid, 3 mg vitamin B₆ and 5 µg vitamin B₁₂ may reduce the risk of vascular disease," and provided you with its decision in the November 1999 letter. FDA determined at that time that there was not significant scientific agreement that the totality of publicly available scientific evidence supported the proposed claim.

In its November 1999 letter, FDA observed that most of the information presented in your petition addressed the relationship between folic acid, vitamin B₆ and vitamin B₁₂ and vascular disease risk via (a) relationships between vitamin intake or vitamin status and circulating levels of homocysteine, an amino acid formed during the metabolism of methionine, an amino acid derived from food, and (b) relationships between circulating levels of homocysteine and specific vascular disease endpoints.

The November 1999 letter explained FDA's evaluation under the significant scientific agreement standard of the scientific data submitted in support of the proposed claim. In reviewing the evidence for an association between the B vitamins and homocysteine, FDA found that there was a sound basis for associations between homocysteine levels and folic acid and, to a lesser extent, vitamin B₆ and vitamin B₁₂. The data reviewed, however, did not establish an association between the B vitamins and vascular disease risk, because lowering of homocysteine levels has not been demonstrated to affect vascular disease risk in the general population. Lacking such evidence, homocysteine level cannot be considered a validated surrogate marker for vascular disease risk, and the studies of changes in homocysteine levels with intake of collectively, the B vitamins, cannot be inferred as supporting changes in cardiovascular risk.

For all these reasons, the agency concluded that the totality of publicly available scientific evidence provided an inadequate basis to support a relationship between supplements containing the B vitamins and reduced risk of vascular disease.

Under its updated plan for implementation of Pearson (65 Fed. Reg. 59,855 (2000)), FDA has considered the scientific evidence on the putative relationship between the B vitamins and the risk of vascular disease, focusing on whether evidence in the petition supports a qualified claim. FDA also considered any new evidence from human studies that have become available since its 1999 decision, including information submitted in your letter of October 19, 2000. We first evaluated new evidence derived from human studies for its potential impact on our original determination that there was no significant scientific agreement for the claim that the B vitamins reduce the risk of vascular disease. We then considered the totality of evidence presented in the petition and in the new studies to determine if the evidence for the claim outweighed the evidence against the claim. We conducted both the original and current scientific evaluations consistent with the principles articulated in FDA's Guidance for Industry: Significant Scientific Agreement in the Review of Health Claims for Conventional Foods and Dietary Supplements (December, 1999).

Based on its review of the scientific evidence published subsequent to November 30, 1999, FDA finds that: (1) although the totality of the publicly available scientific evidence demonstrates a lack of significant scientific agreement as to the validity of a relationship between the B vitamins and risk of vascular disease in the general population, the scientific evidence in support of a qualified claim ⁽¹⁾ outweighs the scientific evidence against the claim; and (2) it may appropriately exercise enforcement discretion with respect to use of the qualified claim about the strength of the scientific evidence in the general population, provided that the general conditions stated in the Pearson implementation notice and the specific conditions set forth in this letter are met.

III. Review of the Scientific Evidence

A. 1999 Scientific Review

A full discussion of FDA's November 1999 decision on the relationship between the B vitamins and vascular disease is provided in the agency's November 30, 1999 letter (copy attached). FDA determined at that time that, based on the totality of publicly available scientific evidence, there was not significant scientific agreement, among experts qualified by scientific training and experience to evaluate such claims, for a claim about the B vitamins and risk of vascular disease.

Specifically, with respect to the association of homocysteine and vascular disease, we found that the observational data provide evidence for an association of homocysteine levels and vascular disease risk. The data did not provide evidence, however, that allowed us to determine whether co-varying changes in homocysteine levels and vascular disease risk represent a causal relationship between homocysteine and vascular disease, or whether these two factors co-vary because of a common relationship to some other factor, perhaps the disease process itself. Without such evidence, it was not possible to predict whether interventions designed to change homocysteine levels would also change vascular disease risk.

With respect to the association of the B vitamins and homocysteine, we found that the available data provided a sound basis for an association between folic acid and, to a lesser extent, vitamins B₆ and B₁₂ on homocysteine levels. However, these data did not establish an association between the B vitamins and vascular disease risk in the general population, because lowering of homocysteine levels has not been demonstrated to affect vascular disease risk in the general population. Lacking such evidence, homocysteine level cannot be considered a validated surrogate marker for vascular disease and the studies of changes in homocysteine levels with B vitamin intake cannot be inferred as supporting changes in cardiovascular disease risk.

Lastly, with respect to an association of B vitamins and vascular disease, we found that the limitations in the designs and conflicting results obtained in the available studies provided an inadequate basis to support a direct effect of the B vitamins on vascular disease outcomes. The findings strongly suggested that well-designed and controlled clinical trials are necessary to establish whether the B vitamins may reduce the risk of vascular disease.

B. Current Scientific Review

FDA first considered whether any significant new human study data had become available in the months since its November 30, 1999 decision. The agency also conducted a literature search to ensure that all relevant scientific literature was included in the current review.

Specifically, in the current review, we looked for scientific evidence regarding an association between (a) homocysteine and vascular disease, (b) the B vitamins and homocysteine, and (c) the B vitamins and vascular disease. We did not include preclinical studies (studies not performed in humans) because there are numerous human studies available and the usefulness of data from preclinical studies is limited in that such studies cannot fully simulate human physiology and disease. Additionally, preclinical studies cannot accurately estimate appropriate intake levels or magnitude of effects in humans.

In evaluating the scientific evidence that has become available since the November 1999 letter, FDA reviewed results of randomized placebo-controlled intervention trials, non-randomized intervention trials, prospective studies, case-control studies, and cross-sectional studies that have become available since issuance of the November 1999 letter, including publicly available information cited in your letter of October 19, 2000. ⁽²⁾

1. Intervention Trials

Intervention studies provide the strongest evidence for testing a hypothesis. Trials designed to prevent disease ("prophylactic" trials) or trials designed to treat established disease ("therapeutic" trials) are two types of intervention studies. In an intervention study, the investigator controls whether the subjects receive quantified intakes of the substance of interest (i.e., the intervention). In contrast, in an observational study, the investigator does not have control over exposure.

In an intervention study, the procedure by which participants are assigned to specific groups (e.g., different treatment groups) is one means by which bias can be minimized. With random allocation of subjects to specific treatment or intervention groups, the groups formed can be expected to be generally alike at the beginning of the trial (i.e., the groups tend to be comparable in all factors that influence the outcome, whether these are known or unknown at the time of randomization). This is not the case when participants are assigned non-randomly to treatment groups. Randomization also eliminates selection bias on the part of participants or investigators. Randomized controlled clinical trials are considered the most persuasive studies, and when results of such studies are available, they will be given the most weight in the evaluation of the totality of the evidence. See Guidance for Industry: Significant Scientific Agreement in the Review of Health Claims for Conventional Foods and Dietary Supplements, at 5.

Another important consideration in an intervention trial is the possible introduction of bias in assessment of outcomes resulting from the expectation of investigators or participants. Protection against this source of bias is provided when neither the investigator nor the subject knows the group to which the subject is assigned. A trial conducted in this manner is said to be "double blind." A study is identified as "single blind" when the investigator knows the assignments but the subjects do not. In a "non-blinded" intervention trial, both the investigator and the subject know the group to which the subject is assigned. Double-blinded intervention studies are generally given greater weight than single-blinded studies in evaluation of the evidence, and non-blinded studies are given the least weight.

a. Randomized controlled intervention trials with homocysteine-lowering as the endpoint

FDA reviewed a number of randomized intervention studies that have appeared since its November 1999 letter. Homocysteine-lowering with increased intakes of B vitamins is of interest, given the hypothesis that high levels of homocysteine are a cause of increased risk of vascular disease. The randomized controlled intervention trials of Chait et al. (1999), Appel et al. (2000) and Riddell et al. (2000) considered the inverse relationship between increased intakes of the B vitamins from foods and dietary supplements and circulating levels of homocysteine. These studies are summarized below.

As part of studies to compare the efficacy of a prepared meal plan with that of a self-selected diet in producing clinically relevant changes in major end points associated with hypertension, dyslipidemia, and type 2 diabetes, Chait et al. (1999) fed participants with these conditions singly or in combination a prepared meal plan (244 subjects) or self-selected diets (247 subjects) for 10 weeks. The prepared meal plan met nutritional guidelines for sodium, total and saturated fat, cholesterol, refined sugars, fiber and complex carbohydrate and was fortified to provide at least 100 percent of recommended daily allowances for 23 micronutrients (Chait et al., 1999). The participants on the self-selected diet were instructed to consume a fixed number of servings of foods consisting primarily of breads and starches, fruit, low-fat milk, vegetables and lean meats. The results showed that intakes of nutrients known to influence homocysteine concentrations (e.g., folate), as found in the meal plan diets at levels consistent with 100 percent of their recommended daily intakes, can effectively decrease homocysteine in persons with risk factors for cardiovascular disease.

Appel et al. (2000) treated a total of 118 normal volunteers with three diets differing in amount and source of factors that affect coronary heart disease (e.g., fruits, vegetables, fat, saturated fat, saturated fat). The "control" diet was low in fruits and vegetables and dairy products and had a fat content typical of United States consumption. A second diet was rich in fruits and vegetables but was otherwise similar to the control diet. Another diet was a combination diet rich in fruits and vegetables and low-fat dairy products and reduced in saturated and total fat content. Appel et al. (2000) reported that after 12 weeks on the diets, there was a significant decrease in homocysteine in those fed the combination diet. The change in homocysteine was significantly and inversely associated with change in serum folate but not with changes in serum vitamin B₁₂ or vitamin B₆.

Riddell et al. (2000) recruited 62 free-living subjects (i.e., non-institutionalized individuals living independently) to participate in a 12-week randomized trial to compare three approaches for increasing dietary folate to about 600 µg/day. Blood samples were collected at recruitment. Participants were then asked to follow a fat-modified diet (30 percent of energy as fat (10 percent as saturated fat), 50 percent as carbohydrate, and 20 percent as protein) for a 2-week run-in period. Participants were randomly assigned to a control group (15 subjects), or to 1 of 3 intervention groups (a dietary folate group (15 subjects), a cereal group (16 subjects), and a supplement group (16 subjects)). During the intervention phase, the control group continued to consume the fat-modified diet that was followed in the run-in period. Subjects in the dietary folate group were instructed to increase their consumption of folate-rich foods. Cereal-group participants increased their intakes through consumption of fortified breakfast cereals and the supplement group was instructed to take a folic acid supplement daily. The results showed that increased intake of folate from foods, breakfast cereals, or supplements led to significant increases in serum folate. Significant decreases in homocysteine occurred in participants in the "cereal" and "supplement" groups. A smaller but not statistically significant decline in homocysteine occurred in the "dietary folate" group. Reductions in homocysteine were significantly negatively correlated with increases in serum folate.

b. Randomized placebo-controlled intervention trials with homocysteine-lowering and other physiologic endpoints

FDA also reviewed randomized, placebo-controlled intervention trials that have become available since issuance of the November 1999 letter. Several studies examined the effects of supplemental intakes of folic acid and vitamin B₆, along with other vitamins, on circulating levels of homocysteine and on parameters of endothelial function (e.g., flow-mediated dilation of specific arteries) (Bellamy et al., 1999; Woo et al., 1999; Title et al., 2000; and Wilmink et al., 2000), or on hemorheological properties of blood of patients with end-stage renal disease (McGregor et al., 2000), or on rate of abnormal exercise electrocardiology tests (Vermeulen, Stehouwer, et al., 2000), or on factors that may affect the susceptibility of low-density-lipoprotein cholesterol to oxidation (Woodside et al., 1998 and Woodside et al., 1999). These studies are summarized briefly below.

Bellamy et al. (1999) investigated the effect of oral folic acid on endothelial function in healthy adults with mild hyperhomocysteinemia. Eighteen healthy subjects entered a randomized double-blind placebo controlled crossover study of oral folic acid (5 mg daily for 6 weeks) with a 6-week interval between treatments. Flow-mediated (endothelium-dependent) and endothelial-independent glyceryl trinitrate-mediated brachial artery dilatations were measured by high resolution vessel wall tracking. Folate supplementation reduced homocysteine significantly and enhanced endothelium-dependent responses.

Woo et al. (1999) treated 17 asymptomatic adults with homocysteine values above the 75th percentile with 10 mg folic acid/day or placebo for 8 weeks to evaluate whether oral folic acid might improve endothelial function in the arteries. Folic acid supplementation resulted in higher serum folate levels and significantly lower homocysteine levels. There was also a significant improvement in endothelium-dependent dilation (increase in diameter) of the brachial artery as assessed by high resolution ultrasound after folic acid treatment.

Title et al. (2000) studied 75 patients with angiographically proven coronary artery disease treated with folic acid (5 mg) or a combination of 5 mg folic acid and antioxidant vitamins daily for 4 months. Title et al. (2000) reported that plasma folate increased and plasma homocysteine decreased in response to consumption of folic acid and folic acid plus antioxidant vitamins treatments. Endothelium-dependent flow-mediated dilation of the brachial artery (measured by Doppler flow techniques) was statistically significantly improved by folic acid treatment but was not further improved by the addition of antioxidant vitamins.

Wilmink et al. (2000) evaluated the effects of an acute oral fat load on endothelial function and oxygen radical production and evaluated whether two weeks of consumption of folic acid (10 mg/day) could prevent fat-induced endothelial changes. Twenty volunteers free of medication, without renal or endocrine disorders, and without a history of hypertension or cardiovascular disease and said to have normal serum folate levels, participated. The results showed that consumption of folic acid for 2 weeks prior to the test prevented the lipid-induced decrease in flow-mediated dilation of the brachial artery and prevented the lipid-induced increase in urinary radical-damage end products.

Elevated plasma homocysteine, dyslipidemia and hemorheological abnormalities occur commonly in patients with end-stage renal disease. McGregor et al. (2000) treated twenty-one end-stage renal disease patients (i.e., those on hemodialysis or ambulatory peritoneal dialysis) with 5 mg folic acid daily for three months and reported that while homocysteine decreased on average 33 percent, there were no statistically significant changes in hemorheological properties of blood in the folic acid-treated group.

Vermeulen, Stehouwer, et al. (2000) investigated the effects of homocysteine-lowering treatment on parameters of subclinical atherosclerosis in healthy siblings of patients with premature atherothrombotic disease. Seventy-eight siblings of vascular disease patients were treated with 5 mg folic acid and 250 mg vitamin B₆ for 2 years. Plasma folate increased and homocysteine decreased significantly (p<0.001). There was a non-significant decreased rate of abnormal exercise electrocardiology tests (p<0.35) in the treated group but no apparent effect of the treatment on ankle-brachial pressure indices or on carotid and peripheral arterial outcome variables.

The mechanisms by which homocysteine may affect the vasculature are not understood. Oxidative modification of low-density lipoprotein (LDL) cholesterol is considered to play a role in atherogenesis and investigations have been carried out to determine whether homocysteine plays a role in this process. Woodside et al. (1998) and Woodside et al. (1999) carried out an intervention trial to assess the effects of antioxidants and B-group vitamins on the susceptibility of LDL-cholesterol to oxidation. One hundred thirty-two participants (clinically healthy men, aged 30-49 years, with homocysteine levels > 8.34 µmol/l) were randomly assigned to take B vitamins alone, a mixture of antioxidant vitamins alone, or a combination of both for 8 weeks. Results showed that homocysteine was significantly decreased (p<0.001) in groups taking the B vitamins and that the antioxidant combination itself had no effect on serum homocysteine. The results of the study indicated that the B vitamins did not have an antioxidant effect.

c. Non-randomized placebo-controlled single-blind trial with homocysteine-lowering as the endpoint

We identified one report of a non-randomized single-blind trial of treatment of coronary artery disease patients with folic acid, vitamin B₆ and vitamin B₁₂ to lower serum homocysteine. Lobo et al. (1999) carried out a placebo-controlled, single-blind dose-ranging study of effects of B vitamins in patients with coronary artery disease. Patients with confirmed coronary artery disease were non-randomly assigned to groups given placebo or one of three daily supplements of folic acid (400 µg, 1 mg, or 5 mg) for three months. Groups contained 22-27 patients each. Patients in the treatment arms of the study also received 500 µg vitamin B₁₂ and 12.5 mg vitamin B₆ daily. Serum homocysteine was measured after 30 and 90 days. The decrease in homocysteine was similar in all treatment groups, and the data suggest that daily administration of 400 µg/day of folic acid combined with vitamin B₁₂ and vitamin B₆ may be equivalent to higher doses in reducing homocysteine levels in patients with coronary artery disease.

d. Non-randomized, non-blinded intervention trials with homocysteine-lowering and progression of vascular disease as endpoints

Circulating levels of homocysteine can be determined after fasting or after a methionine load. Post-methionine load homocysteinemia is the increase in homocysteine in the blood that is observed in some individuals following an intake of a measured amount of the amino acid methionine. The ingestion of the specified amount of methionine is referred to as the "methionine load." The concentration of homocysteine measured in the blood after the methionine load is referred to as the "post-methionine load homocysteine." Individuals can be identified as "hyperhomocysteinemic" on the basis of measurements of their fasting homocysteine level or on the basis of their post-methionine load homocysteine levels.

We identified two reports of non-randomized studies of the treatment of patients with established vascular disease with high levels of folic acid and vitamin B₆ to reduce the recurrence of vascular complications of the existing disease (DeJong et al., 1999; Vermeulen, Rauwerda, et al., 2000). Evaluation of the possible effects of homocysteine-lowering on subsequent vascular disease risk in this population are of interest because hyperhomocysteinemic vascular patients, compared to those with normo-homocysteinemia, have a two- to four-fold higher incidence of cardiovascular morbidity and mortality. Mild hyperhomocysteinemia, whether determined after fasting or after methionine loading, is thus associated with increased risk and severity of atherosclerotic vascular disease in this patient population. Both fasting and post-methionine load hyperhomocysteinemia are generally responsive to treatment with vitamin B₆ and folic acid. This population thus presents a sensitive model for studying the relationship between homocysteine-lowering and vascular disease risk, although caution in extrapolating the results to the general population, or even to other patient populations, must be exercised.

DeJong et al. (1999) studied 273 consecutive patients with peripheral arterial occlusive vascular disease. Two-hundred thirty two patients participated in the intervention study. Seventy patients with post-methionine load hyperhomocysteinemia were treated with a combination of 5 mg folic acid and 250 mg pyridoxine (vitamin B₆). The 162 patients who were normo-homocysteinemic following methionine loading served as the reference group. Data obtained during follow-up for 1-63 months (median, 20 months) indicated that treatment with folic acid and pyridoxine reduced the elevated vascular disease risk for this hyperhomocysteinemic patient population, resulting in an adjusted odds ratio (RR, relative risk) for new cardiovascular events that was not statistically significantly different between the vitamin-treated hyperhomocysteinemic patients and the normo-homocysteinemic reference group.

Vermeulen, Rauwerda, et al. (2000) studied 224 consecutive patients with atherothrombotic cerebrovascular disease. The 52 patients who were hyperhomocysteinemic following methionine loading started treatment with 5 mg folic acid and 250 mg pyridoxine daily, and the post-methionine load normo-homocysteinemic patients (n=172) served as the reference group. Data obtained during follow-up (median, 57 months) indicated that the vitamin treatment resulted in a decreased incidence of new cardiovascular events in the hyperhomocysteinemic group, and an adjusted odds ratio (relative risk, RR) for new cardiovascular events that was not significantly different between the vitamin-treated hyperhomocysteinemic patients and the normo-homocysteinemic reference group.

These two new reports (DeJong et al., 1999; Vermeulen, Rauwerda, et al., 2000) are significant because they suggest that treatment using folic acid and vitamin B₆ in hyperhomocysteinemic patients with established vascular disease may reduce the risk of new cardiovascular events.

However, the results of these studies must be used cautiously. The most important limitations of these two studies are that the vitamin treatment was non-randomized and non-blinded. Thus, alternative explanations for the findings must be considered (e.g., differences in other, unmeasured cardiovascular risk factors between hyperhomocysteinemic and normo-homocysteinemic patients, more aggressive treatment of other recognized risk factors in the hyper-homocysteinemic group, lifestyle changes in hyperhomocysteinemic subjects that may have changed their cardiovascular risk). Moreover, the studies were performed in patients with premature vascular disease and utilized very high levels of the vitamins (5 mg folic acid and 250 mg vitamin B₆) and it is not clear whether the results can be generalized to other patient groups or to the general population. Nevertheless, the data are consistent with a protective effect of folic acid and vitamin B₆ against vascular disease.

e. Non-randomized, non-blinded intervention trials with homocysteine-lowering and other physiologic endpoints

FDA also reviewed non-randomized, non-blinded intervention trials with other than clinical disease endpoints that have become available since issuance of the November 1999 letter. Several studies examined the effects of treatments with folic acid and vitamin B₆, along with other vitamins, on circulating levels of homocysteine alone (van der Griend et al., 1999) or in combination with effects on measures of vascular function (Constans et al., 1999). Other studies examined the effects of treatments with the B vitamins on homocysteine levels and such endpoints as oxidation of LDL-cholesterol (Bunout, Garrido et al., 2000; Weiss et al., 1999), while others examined the effects of vitamin treatments on areas of plaque in blood vessels of patients with confirmed atherosclerosis (Peterson and Spence, 1998; Hackam, et al., 2000). These studies are summarized below.

Van der Griend et al. (1999) studied homocysteine-lowering effects after 8 weeks of treatment with a combination of folic acid (0.5 mg) and pyridoxine (100 mg) in 49 hyperhomocysteinemic persons (i.e., persons with high levels of homocysteine; fasting homocysteine greater than 12 µmol/L and/or post-methionine load homocysteine greater than 38 µmol/L in this study). One group consisted of 33 patients with documented premature arterial disease and the second group consisted of 16 of their first-degree relatives. Low dose-vitamin therapy statistically significantly reduced fasting homocysteine concentrations (reduction, 32 percent) and post-methionine load homocysteine concentrations (reduction, 30 percent) in patients with premature arterial disease as well as in their healthy first-degree relatives. Reductions in post-methionine load hyperhomocysteinemia were significantly less in men than in women and in relatives compared to patients. This study provides the first data on homocysteine-lowering with dietary levels of vitamin therapy in patients with arterial disease.

Constans et al. (1999) studied the effect on plasma homocysteine and two markers of endothelial injury (i.e., soluble thrombomodulin, von Willebrand factor) of three months of treatment with folic acid and vitamin B₆. Eighteen patients with various forms of vascular arterial or venous vascular disease whose hyperhomocysteinemia was detected after presentation with the vascular disease were treated for three months with 5 mg folic acid and 250 mg vitamin B₆ daily. After three months, there was a statistically significant decrease in post-methionine load homocysteinemia, a significant decrease in soluble thrombomodulin, and no change in von Willebrand factor.

Bunout, Garrido et al. (2000) assessed the effect of a supplement of folic acid and antioxidant vitamins on homocysteine levels and in vitro LDL-cholesterol oxidation in patients with coronary artery disease. Twenty-three patients with angiographically proven coronary artery disease were supplemented for 15 days with 1.3 mg folic acid, 0.8 µg vitamin B₁₂, and a mixture of vitamins A, E and -carotene. During supplementation, serum folate increased and serum homocysteine decreased significantly, as did in vitro LDL-cholesterol oxidation (measured via content of thiobarbituric acid-reactive substances).

Oxidation of homocysteine in the blood leads to the formation of oxygen radicals. These may lead to oxidative modification of LDL-cholesterol and promote atherosclerotic vascular changes. Weiss et al. (1999) determined fasting and post-methionine load homocysteine and susceptibility of LDL-cholesterol particles to ex vivo oxidation in 69 male patients with symptomatic peripheral arterial occlusive disease. A sub-set of these patients (n = 15) with hyperhomocysteinemia were treated with 5 mg folic acid and 300 mg pyridoxal phosphate (vitamin B₆) for 4 weeks and received an intramuscular injection of 1 mg vitamin B₁₂ at the start of the study. The authors reported that vitamin-treated patients had statistically significant reductions in fasting and post-methionine load homocysteine and there was a decrease in the in vitro susceptibility of their LDL-cholesterol particles to oxidation.

Peterson and Spence (1998) and Hackam et al. (2000) reported that when thirty-eight patients with unexplained progression of atherosclerosis and homocysteine values greater than 14 µmol/L were treated with a combination of 2.5 µg folic acid, 250 µg vitamin B₁₂ and 25 mg vitamin B₆ daily for a mean of 4.4 years, there was a statistically significant decrease in plaque areas determined by ultrasonography.

Taken together, the intervention studies show an association between intakes of the B vitamins at varying levels (folic acid, 400 µg [0.4 mg] to 10 mg; vitamin B₆, 12.5 to 300 mg; vitamin B₁₂, 0.5 to 1 mg) and reductions in homocysteine and changes in other physiologic endpoints in both patient and general populations. However, the utility of these other endpoints in resolving uncertainties regarding a causal relationship between homocysteine and vascular disease is not known because these other endpoints themselves are not recognized as validated surrogate markers of vascular disease. The two new intervention trials (DeJong et al., 1999; Vermeulen, Rauwerda, et al., 2000), however, do suggest a protective effect of vitamin treatment in hyperhomocysteinemic, vascular disease patients, although the non-random, non-blinded nature of the studies warrants caution in interpretation. None of the available intervention trials evaluated a relationship between vitamin intakes and decreased risk of vascular disease in the general population.

2. Observational Studies

Observational studies (sometimes called "epidemiological" studies) include several types: population or correlational, retrospective case control, and prospective cohort. These types of studies can provide information on the association between the B vitamins and vascular diseases; however, such studies often do not provide a sufficient basis for determining whether a substance-disease association reflects a causal, rather than a coincidental, relationship. Population or correlational studies use grouped data to examine the relationship between dietary exposure and health outcome among populations. Such studies do not examine relationships for individuals and have traditionally been regarded as useful for generating rather than testing hypotheses regarding diet-disease or substance-disease relationships. Therefore, FDA did not give population studies much weight in the current evaluation.

In case-control studies, subjects with existing diagnosed disease (the cases) are enrolled in a study. The subjects are matched by identifiable characteristics (e.g., age, race, gender) to disease-free subjects (the controls). Specific characteristics (e.g., diet) of the two groups are then compared to discern an association between dietary habits and risk of disease. In prospective or cohort studies, subjects are recruited within a specific group of people (the cohort) and the characteristics of the subjects are determined. The study tracks the subjects over an extended period of time to determine whether they develop, for example, a disease under investigation. At the end of the follow-up period, the characteristics of subjects who developed the disease during the follow-up period are compared to those of subjects who did not develop the disease to discern characteristics that are associated with risk of the disease. Prospective studies are generally considered to be the most persuasive type of observational study. Therefore, FDA weighted these more heavily than other types of observational studies.

a. Prospective Studies

In a prospective study, cohorts (i.e., groups of persons who share a common experience within a defined time period) are assembled and followed forward in time for the development [or progression] of disease. Particular groups may be chosen because they are accessible (e.g., volunteers) or because medical records or history of exposure are readily available (nurses, physicians). Other cohorts can be assembled from participants sharing unique medical conditions (e.g., renal transplant recipients). The major advantage of prospective studies is that the cohort is classified with respect to exposure to a factor or factors before the disease develops. Prospective studies permit the calculation of incidence rates among those exposed or not exposed. One of the most important features of prospective studies is that they permit the observation of many outcomes. Further, blood samples obtained at enrollment into the study can subsequently be assayed for a variety of components, some of which may not have been of interest at the initiation of the study. The need to follow a cohort over time presents special obstacles, however. Attrition (e.g., loss of participants from follow-up due to loss of interest, migration, or death) and change of status of participants (e.g., change in smoking habits, change in dietary or drinking habits) are two such obstacles that may affect interpretation of findings.

We reviewed a number of prospective studies that have become available since our November 1999 letter. Three of these studies (Aronow and Ahn, 2000; Ducloux et al., 2000; Omland et al., 2000) looked at relationships between circulating levels of homocysteine and risk of vascular disease. The study of Voutilainen et al. (2000), which did not provide data on homocysteine levels, looked for an association between serum folate levels and risk of coronary events. These studies are summarized below.

Aronow and Ahn (1997) reported that high plasma homocysteine and lower plasma folate and vitamin B₁₂ levels were associated with a higher prevalence of coronary artery disease in 153 older men and 347 older women. These authors recently reported the results of a new 31-month follow-up that showed that plasma homocysteine is an independent risk factor for new coronary events in older subjects with established coronary artery disease (Aronow and Ahn, 2000).

Ducloux et al. (2000) evaluated the relevance of a single determination of fasting total homocysteine in 207 stable renal transplant patients who were followed for a mean duration of 21.2 months (range, 14-26 months). They reported that total homocysteine is an independent cardiovascular risk factor in stable renal transplant recipients. Total homocysteine was negatively correlated with folate concentration but not with cobalamin (vitamin B₁₂) concentration. Total homocysteine was also closely related to serum creatinine concentration, a clinical measure of renal function.

Omland et al. (2000) studied homocysteine levels obtained during the first 24 hours following hospital admission in 579 patients with acute coronary syndromes to determine whether there was a relationship between post-event homocysteine levels and long-term mortality. Follow-up was for a median interval of 628 days. The authors reported that serum homocysteine on hospital admission was an independent predictor of long-term survival in patients with acute coronary syndromes. Patient age, history of myocardial infarction and serum creatinine were also identified as independent predictors of serum homocysteine concentrations.

Voutilainen et al. (2000) tested the hypothesis that low serum folate concentrations were associated with an increased risk of acute coronary events in 734 middle-aged men without prior coronary disease and found that during a 63-month follow-up, moderate-to-high levels of serum folate were associated with a significantly reduced incidence of acute coronary events. Levels of homocysteine were not determined.

b. Case-control studies

Case-control studies are retrospective studies. As is the case with cross-sectional or survey studies, the markers (e.g., homocysteine) are measured at the same time or soon after the identification of cases of the disease under investigation. In case-control studies, however, unlike cross-sectional studies, a distinct control group is usually identified as the standard for comparison. Controls may be drawn from a variety of sources. An important assumption underlying the validity of case-control studies is that the controls are representative of the general population. In case-control studies, it is also important to remove the influence of confounding variables (i.e., factors known to be associated with the risk factor of interest and causally related to the disease under investigation). Matching of cases and controls is one method used to control confounding.

Retrospective case-control studies have several advantages, at least compared to prospective studies. Such studies are relatively inexpensive to carry out and their results can be obtained relatively quickly. The number of subjects in a retrospective study can be small, since the study is initiated by the identification of cases, which are often compared to a similar number of controls. Even when two or three controls are selected for each case, the total numbers are small compared with the numbers needed for a prospective study. These advantages are offset, however, by a number of significant disadvantages. The most serious problems associated with the use of retrospective studies include those related to the selection of an appropriate control group and failure to adjust for confounding factors (e.g., renal impairment). Perhaps most importantly, measurement of the factors of interest at the time of diagnosis of the disease eliminates the possibility of determining whether an elevated risk factor (e.g., homocysteine) is causally or coincidentally related to the disease process itself. For these reasons, case-control studies are generally given less weight than prospective studies.

FDA reviewed case-control studies that have become available since its November 1999 letter. All of the studies summarized below (Kristensen et al., 1999; Turgan et al., 1999; Hoogeveen et al., 2000(a); van der Molen et al., 2000; Bunout, Petermann et al., 2000; Chambers et al., 2000; Martyn, 2000; Langman et al., 2000; Mansoor et al., 2000; Leowattana et al., 2000) looked for associations between circulating levels of homocysteine and risk of vascular disease.

Kristensen et al. (1999), in a study of 89 young persons (18-44 years of age) who had experienced a first stroke, reported that a moderately increased post-methionine load homocysteine (see above) was associated with a 4.8-fold increased risk for stroke. Fasting homocysteine levels (i.e., levels of homocysteine measured after a specified interval without food intake) did not differ between these patients and controls (n = 41).

Turgan et al. (1999) studied 40 patients with a confirmed diagnosis of an acute coronary syndrome and normal renal function and 30 apparently healthy controls with no evidence of ischemic heart disease. With one exception, there were no significant differences between patients and controls in a range of serum or plasma parameters measured, including vitamin B₁₂, folate, creatinine and HDL- and LDL-cholesterol. Homocysteine was significantly elevated in patients versus controls and positive correlations between homocysteine and age and between homocysteine and creatinine were observed in the patient group but not in the control group.

Hoogeveen et al. (2000(a)) measured homocysteine in 171 subjects who died (76 of cardiovascular disease) and in a stratified random sample of 640 survivors in the Hoorn Study. The Hoorn Study is cross-sectional survey of glucose tolerance and other cardiovascular risk factors in a 50 to 75-year old general Caucasian population numbering 2,484 in the Netherlands. Hyperhomocysteinemia (defined as homocysteine greater than 14 µmol/L) was found to be related to 5-year mortality independent of other major risk factors (e.g., gender, systolic blood pressure, current smoking habit, serum cholesterol, diabetes mellitus) and appeared to be a stronger risk factor for mortality in type 2 diabetic patients than in non-diabetic subjects.

Placental vasculopathies, including placental infarctions, are serious complications of pregnancy and can lead to fetal growth restriction. Signs of placental vasculopathy include atherosclerosis, narrowing necrosis, and thrombosis, and these signs resemble changes seen in patients with vascular disease. Van der Molen et al. (2000) studied 165 women with placental vasculopathy and 139 matched controls with normal pregnancy outcomes. Mild hyperhomocysteinemia was confirmed among women with placental vasculopathy, for which homozygosity for a mutated 5,10-methylene tetrahydrofolate reductase gene was also found to be a new risk factor.

Bunout, Petermann et al. (2000) looked for an association between homocysteine, folate and vitamin B₁₂ in 84 Chilean subjects with angiographically proven stable coronary artery disease or peripheral vascular disease. The control group consisted of 66 asymptomatic subjects. No differences in homocysteine were observed between coronary or vascular patients and their control counterparts. However, serum folate levels were significantly lower in coronary and vascular patients than in their respective asymptomatic control groups.

Chambers et al. (2000) performed two parallel case-control studies among European and Indian Asian coronary heart disease patients. The number of cases was 551 and the number of controls was 1,025. The results showed that moderately elevated homocysteine was associated with increased risk of coronary heart disease in both European and Indian Asian patients. Levels of homocysteine tended to be higher in Indian Asian patients. The role of renal impairment was not considered as a possible confounding variable (Martyn, 2000).

Langman et al. (2000) carried out a case-control study in individuals with confirmed thromboembolism and no history of atherosclerosis. Baseline homocysteine levels for 232 cases of coronary heart disease identified during follow-up averaging 3.3 years were similar to levels for a reference cohort of 537 participants. The RR (relative risk) for coronary heart disease comparing the highest and lowest quintiles of homocysteine was 1.3 (95% CI (confidence interval), 0.5-3.2). After accounting for other risk factors, however, only plasma pyridoxal-5'-phosphate, the form in which vitamin B₆ is found in the blood, was associated with incidence of coronary heart disease (women only). The RR for the highest versus the lowest quintile of pyridoxal-5'-phosphate was 0.28 (95% CI, 0.1-0.7).

Sixty-five patients with peripheral vascular disease and 65 age- and sex-matched healthy control subjects were included in the case-control study of Mansoor et al. (2000). Concentrations of plasma homocysteine were significantly higher and concentrations of pyridoxal-5'-phosphate were significantly lower in patients than in controls, while serum folate and serum vitamin B₁₂ concentrations did not differ between the two groups.

Leowattana et al. (2000) investigated associations between serum homocysteine, vitamin B₁₂ and folate levels and vascular disease in 178 patients with coronary artery disease and 178 normal age- and sex-matched controls. Paired t-tests showed that serum homocysteine concentrations were significantly higher in patients than in controls and that serum folate and serum vitamin B₁₂ levels did not differ significantly between the groups.

c. Cross-sectional studies

A cross-sectional study or survey is another type of observational study. In this type of study, both specific factor(s) (e.g., dietary intakes, elevated levels of homocysteine in the blood) and health outcome or endpoint (e.g., arterial wall thickening, vascular disease) are ascertained at the same time. Cross-sectional studies are those in which members of a defined population are examined for the presence or absence of a specific disease endpoint or for a presumed surrogate marker of the disease. Because cross-sectional studies or surveys measure risk factors and disease endpoints at the same time, they cannot establish the temporal sequence of events necessary for drawing causal inferences from the results obtained. Such studies have traditionally been regarded as useful for generating, rather than testing, hypotheses regarding, for example, diet-disease relationships.

Several cross-sectional studies that have recently become available are summarized below. Some of these new studies looked at associations between circulating levels of homocysteine and risk of vascular disease (Hughes and Ong, 2000; Abdelmouttaleb et al., 2000; Chao et al., 1999). Other studies looked at associations between homocysteine and carotid artery wall thickening (Willinek et al., 2000; Tsai et al., 2000; Mazza et al., 2000) or on associations between homocysteine and specific localizations of vascular disease (Hoogeveen et al., 2000(b)). Another study (Hak et al., 2000) looked at other physiologic factors that may affect circulating levels of homocysteine.

Hughes and Ong (2000) attempted to determine whether the higher rates of coronary heart disease in South Asian Indians (versus Malays and Chinese) in Singapore were associated with differences in blood levels of homocysteine, folate, and vitamin B₁₂. Hughes and Ong (2000) reported that, although there were ethnic differences for plasma folate and vitamin B₁₂ among a random sample of 726 subjects, aged 30 to 69 years, there was no evidence that homocysteine plays a role in the increased susceptibility of Indians to coronary heart disease.

Abdelmouttaleb et al. (2000) reported that in 215 subjects undergoing diagnostic coronary angiography, the level of homocysteine and the B vitamins did not contribute to the ability to discriminate for the presence of coronary artery disease in patients undergoing coronary angiography. The comparison group for this study consisted of 171 subjects without a history of myocardial infarction, coronary heart disease, other cardiovascular events, or cancer. Age, sex, total cholesterol, triglycerides and apolipoprotein B concentrations were found to be independently related to the presence of angiographically visible coronary artery disease.

Chao et al. (1999) studied homocysteine levels in 116 coronary artery disease (CAD) and 76 non-CAD patients, all of whom underwent coronary angiography. They reported that risk of CAD in the two highest quartiles of homocysteine (> 14 µmol/L) was 3.2 to 4-times higher than in the lowest quartile (homocysteine less than < 7.9 µmol/L).

Structural alterations (e.g., thickening) in the walls of the carotid artery have been investigated as potentially useful predictors of atherosclerotic changes occurring in the coronary and cerebral arteries. Willinek et al. (2000) studied 75 male and female asymptomatic subjects with normal serum folate concentrations, and reported that high-normal homocysteine levels (mean 10.5 µmol/L) were associated with an increased prevalence of carotid artery wall thickening. Other predictors of carotid artery wall thickening included age, body mass index and LDL-cholesterol.

Tsai et al. (2000) performed ultrasound measurements of carotid artery intimal-media thickness in 1,467 individuals in the National Heart, Lung and Blood Institute's (NHLBI) Family Heart Study and reported that increased total homocysteine was significantly associated with an increase in carotid intimal-media wall thickness in participants 55 years or older. A similar but not statistically significant trend was found in participants less than 55 years of age. No association between total homocysteine and coronary heart disease was observed.

A common variant of the enzyme 5,10-methylenetetrahydrofolate reductase (MTHFR) has been correlated with mild elevations on plasma homocysteine in the general population. Mazza et al. (2000) assessed the contribution of serum homocysteine and MTHFR polymorphism in 130 middle-aged patients with non-insulin-dependent diabetes mellitus without nephropathy, and reported that basal levels of homocysteine as well as MTHFR polymorphism did not predict significant changes in intimal medial thickness of the common carotid artery.

Hoogeveen et al. (2000(b)) investigated whether the strength of the association between hyperhomocysteinemia and peripheral arterial disease differs according to the location of the obstruction (i.e., aortoiliac, femoropopliteal or distal; measured by means of Doppler flow velocity) in a random sample of 50-75 year old men and women (n=631) in the Netherlands. These authors reported that there was a statistically significant relationship between increments of 5 µmol/L homocysteine and increased risk for aortoiliac but not femoropopliteal or distal obstructions. The authors also reported that systolic blood pressure was significantly associated with aortoiliac obstructions.

The post-menopausal state has been found to be associated with higher plasma homocysteine levels, but data are inconsistent and studies have not adjusted for age, an important confounder. Hak et al. (2000) reported that levels of homocysteine were significantly higher in 93 post-menopausal Dutch women than in 93 age-matched pre-menopausal women. The authors noted that their results strengthen the hypothesis that estrogens influence homocysteine levels and that increased cardiovascular disease risk in post-menopausal women may be mediated via effects of estrogen on homocysteine.

As noted above, these studies examined associations between homocysteine and a number of endpoints but do not provide direct evidence for a relationship between the B vitamins and risk of vascular disease. In some studies (e.g., Hughes and Ong, 2000; Abdelmouttaleb et al., 2000; Mazza et al., 2000), expected associations between homocysteine and specific vascular disease endpoints were not found.

IV. Agency's Consideration of Significant Scientific Agreement

As discussed in section II, a major factor in FDA's November 30, 1999 decision not to authorize a health claim for the B vitamins and vascular disease was that collectively the results from the available data did not provide a sufficient basis to establish a relationship between the vitamins and reduced risk of vascular disease. In its current review, FDA focused on whether any newer studies resolve the previous uncertainties regarding the nature of the relationship between homocysteine and vascular disease (i.e., whether the relationship is causal or coincidental).

A. Conclusions from the new data

FDA reviewed more than 35 new studies, including randomized placebo-controlled intervention trials, non-randomized intervention studies, prospective studies, case-control studies, and cross-sectional studies that became available since the November 1999 letter. These studies addressed associations of homocysteine and vascular disease, associations of the B vitamins and homocysteine and other physiological endpoints, and associations of the B vitamins and vascular disease.

1. Association between homocysteine and vascular disease

Some of the new evidence addresses the relationship between circulating levels of homocysteine and vascular disease risk. This linkage is based on the presumption that homocysteine level is a validated surrogate marker for vascular disease risk. However, evidence is needed to demonstrate that homocysteine is a valid surrogate marker for vascular disease risk. This evidence must establish that interventions that alter homocysteine levels also affect disease risk. To date, such evidence is not available.

A number of new studies looked at effects of vitamin intervention on both homocysteine levels and on a variety of endpoints associated with vascular disease, including (a) parameters of endothelial function (Bellamy et al., 1999; Woo et al., 1999; Title et al., 2000; and Wilmink et al., 2000); (b) hemorheological properties of blood (McGregor et al., 2000); (c) rate of abnormal exercise electrocardiology tests (Vermeulen, Stehouwer et al., 2000); (d) susceptibility of low-density-lipoprotein cholesterol to oxidation (Woodside et al., 1998, Woodside et al., 1999); (e) carotid wall thickening (Willinek et al., 2000; Tsai et al., 2000; Mazza et al., 2000); (f) specific localizations of vascular disease (Hoogeveen et al., 2000(b)); or (g) soluble thrombomodulin (Constans et al., 1999; Bunout, Garrido et al., 2000; Hackam et al., 2000). Although in most of these studies circulating levels of homocysteine and the other physiologic endpoints measured responded to the vitamin interventions, the endpoints assessed also are not validated surrogate markers of vascular disease because changes in these endpoints have not been shown to reduce the risk of vascular disease. Therefore, the relevance of the effect of reducing homocysteine on these other measured endpoint relative to risk of vascular disease is unknown.

None of the available intervention studies provided evidence that lowering of homocysteine levels would also lower the risk of vascular disease in the general population. Therefore, these newer studies do not provide a basis to establish homocysteine as a validated surrogate marker of vascular disease risk.

Several of the observational studies that have become available since the November, 1999, review evaluated the relationship between homocysteine levels and disease risk. Observational studies are generally limited in their ability to ascertain the actual food or nutrient intake for the population studies. They also often do not provide a sufficient basis for determining whether a substance/disease association reflects a causal rather than a coincidental relationship.

Three prospective studies (Ducloux et al., 2000; Aronow and Ahn, 2000; Omland et al., 2000) looked at relationships between circulating levels of homocysteine and risk of vascular disease and reported positive associations in stable renal transplant patients and in those with established vascular disease.

New case-control studies (Kristensen et al., 1999; Turgan et al., 1999; Hoogeveen et al., 2000(a); van der Molen et al., 2000; Bunout, Petermann et al., 2000; Chambers et al., 2000; Langman et al., 2000; Mansoor et al., 2000; Leowattana et al., 2000) looked for associations between circulating levels of homocysteine and risk of vascular disease. The new studies generally report associations between elevated homocysteine and increased risk of vascular disease (Kristensen et al., 1999; Turgan et al., 1999; Hoogeveen et al., 2000(a); Chambers et al., 2000, Mansoor et al., 2000; Leowattana et al., 2000). By contrast, the report of Bunout, Petermann et al. (2000) did not find an association between homocysteine levels and risk of vascular disease.

In case-control studies, differences in the recall or ascertainment of exposure history according to disease status are an important potential source of bias. In blood-based case-control studies (e.g., Langman et al., 2000; Bunout, Petermann et al., 2000; Hoogeveen et al., 2000(a); Chambers et al., 2000; van der Molen et al., 2000; Kristensen et al., 1999; Leowattana et al., 2000; Turgan et al., 1999), there is an additional concern that the disease process itself may alter blood levels of possible determinants of the disease (e.g., homocysteine).

In cross-sectional studies (Hoogeveen et al., 2000(b); Hak et al., 2000; Hughes and Ong, 2000; Willinek et al., 2000; Tsai et al., 2000; Abdelmouttaleb et al., 2000; Mazza et al., 2000; and Chao et al., 1999), homocysteine levels and other putative markers for vascular disease (e.g., carotid intimal-medial wall thickness, blood concentrations of folate and vitamin B₁₂, localization of vascular obstructions, methylene tetrahydrofolate reductase gene polymorphism) are assessed concurrently and cases are typically persons with early or preclinical disease.

The results of the new cross-sectional studies that looked at associations between homocysteine and vascular disease have been mixed. Hughes and Ong (2000), Tsai et al. (2000), and Abdelmouttaleb et al. (2000) did not find evidence that homocysteine plays a part in risk of coronary heart disease, while Chao et al. (1999) found that elevated homocysteine was associated with increased risk of coronary artery disease. Mixed results regarding associations between homocysteine levels and other putative markers of vascular disease have also been reported (Tsai et al., 2000; Mazza et al., 2000; Hoogeveen et al., 2000(b)).

It is not possible to determine from the new observational data whether elevated homocysteine levels preceded clinical changes (e.g., atherosclerosis). In the first instance (i.e., if homocysteine levels are causally related to risk of vascular disease), then interventions designed to affect homocysteine levels would likely alter risk of vascular disease. To date, such data are lacking.

The intervention and observational studies (i.e., prospective, case-control, and cross-sectional) identified since November 1999 provide suggestive but not conclusive evidence for an association of homocysteine levels and risk of vascular disease. The studies do not provide information regarding whether the changes in vascular disease endpoints measured are caused by homocysteine or whether homocysteine levels and the disease endpoint are co-varying, perhaps because of a common relationship to some other factor (e.g., the disease process itself).

2. Association between the B vitamins and homocysteine

The randomized studies of Chait et al. (1999), Appel et al. (2000), Riddell et al. (2000), Bronstrup et al. (1998), Bellamy et al. (1999), Woo et al. (1999), Title et al. (2000), Wilmink et al. (2000), McGregor et al. (2000), Vermeulen, Stehouwer, et al. (2000), Woodside et al. (1998), Woodside et al. (1999), and den Heijer et al. (1998) confirm the previously-reported association between intakes of the B vitamins and lowering of levels of homocysteine. The well-designed studies of Chait et al. (1999), Appel et al. (2000), Riddell et al. (2000) and Bronstrup et al. (1998) show that intakes of the vitamins at levels attainable through careful dietary choices or through use of fortified cereals or dietary supplements can significantly lower circulating levels of homocysteine in individuals with risk factors for vascular disease and in individuals in the general population.

However, these data, when considered with that reviewed in the November 1999 letter, do not establish an association between the B vitamins and vascular disease risk, because lowering of homocysteine levels has not been demonstrated in the general population to affect vascular disease risk.

The results of a non-randomized single-blind trial of treatment of coronary artery disease patients with B vitamins to lower serum homocysteine (Lobo et al., 1999) are consistent with the results from the randomized trials. Data from other non-randomized intervention studies also confirm earlier observations that, in general, increased intakes of folic acid, in combination with vitamins B₆ and B₁₂, will reduce fasting levels of homocysteine and post-methionine load homocysteine concentrations (Constans et al., 1999; Bunout, Garrido et al., 2000; van der Griend et al., 1999; Weiss et al., 1999; Peterson and Spence, 1998, and Hackam et al., 2000).

Although supplementation with folic acid and vitamin B₆ has been shown to decrease plasma homocysteine, there are no data that assess an effect of such treatment on the risk of vascular disease in the general population. Lacking such evidence, homocysteine levels cannot be considered a validated surrogate marker for vascular disease risk in the general population, and the studies of changes in homocysteine levels with intakes of the B vitamins cannot be inferred as supporting changes in cardiovascular disease risk in the general population.

3. Association between the B vitamins and vascular disease

New evidence for a direct association of the B vitamins and vascular disease risk is provided by two non-randomized intervention trials and a prospective study. The non-randomized, non-blinded intervention studies of DeJong et al. (1999) and Vermeulen, Rauwerda, et al. (2000) are of interest because they prospectively investigate the effects of homocysteine-lowering treatment with folic acid (5 mg) plus vitamin B₆ on the course of peripheral arterial occlusive disease and atherothrombotic vascular disease in patients with post-methionine hyperhomocysteinemia. The data show that the relatively high cardiovascular risk of patients with premature vascular disease who have hyperhomocysteinemia after methionine loading and who are treated with high levels of folic acid and vitamin B₆ drops to levels similar to that of comparable patients who are normo-homocysteinemic. These findings are consistent with a protective effect of treatment with folic acid and vitamin B₆.

An important limitation of these two studies is that the vitamin treatment was non-randomized and non-blinded. Thus, alternative explanations for the findings must be considered (e.g., differences in other, unmeasured cardiovascular risk factors between hyperhomocysteinemic and normo-homocysteinemic patients, more aggressive treatment of other recognized risk factors in the hyper-homocysteinemic group, lifestyle changes in hyperhomocysteinemic subjects that may have changed their cardiovascular risk). Additionally, the studies were performed in patients with premature vascular disease and utilized very high levels of the vitamins (5 mg folic acid and 250 mg vitamin B₆). It is not clear whether the results can be generalized to other patient groups or to the general population. Nevertheless, the data are consistent with a protective effect of folic acid and vitamin B₆ against vascular disease.

The prospective cohort study of Voutilainen et al. (2000), which did not provide data on homocysteine levels, also found a protective effect of folate status on risk of coronary events. However, as an observational study, this type of evidence does not provide a basis for determining whether this association is causal or coincidental.

B. Summary of conclusions from the totality of data

Based on our review of all of the publicly available scientific evidence, including the literature reviewed for the November 1999 letter and that has become available since the November 1999 letter, we conclude that:

We find that the evidence provides a sound basis for associations between intake of folic acid, and, to a lesser extent, intake of vitamin B₆ and vitamin B₁₂, and homocysteine levels. However, considered together, the data do not establish an association between the vitamins and risk of vascular disease because lowering of homocysteine levels has not been demonstrated to affect vascular disease risk in the general population. Lacking such evidence, levels of homocysteine cannot be considered a validated surrogate marker for vascular disease risk and studies of changes in homocysteine levels and intake of folic acid, vitamin B₆ and vitamin B₁₂ cannot be inferred as supporting changes in vascular disease risk. The data do not provide evidence that allows us to determine whether observed co-varying changes in homocysteine levels and vascular disease risk represent a causal relationship between homocysteine and vascular disease, or whether these two factors were co-varying because of a common relationship to some other factor, perhaps the disease process itself. We also find that the limitations in the types of studies and conflicting results in the studies provided an inadequate basis to demonstrate that increasing the intake of B-vitamins will decrease the risk of vascular disease in the general population.

C. Reviews

You cited a number of review articles (Andreotti et al., 2000; Brattstrom and Wilcken, 2000; Kullo, et al., 2000; Scott and Sutton, 1999; Temple et al., 2000; and Ueland et al., 2000) in your October 19, 2000, letter. FDA reviewed those articles and one other such article (Christen et al., 2000) and finds that they do not corroborate the causal role of homocysteine in the etiology and progression of vascular disease.

Specifically, Andreotti et al. (2000) state that "at present, it cannot be excluded that hyperhomocysteinemia represents a mere marker of manifest or subclinical disease without contributing to its development." Brattstrom and Wilcken (2000) note that "the results of the many ongoing homocysteine-lowering trials with folic acid in vascular patients may clarify whether folate therapy is relevant to cardiovascular risk."

Kullo et al. (2000) note that, although screening for hyperhomocysteinemia should be considered in certain situations, "prospective trial data that demonstrate a reduction in vascular events as a result of treatment of hyperhomocysteinemia are not yet available." Kullo et al. (2000) state further that "prospective randomized trials are needed to prove a beneficial effect of increased folic acid intake in the setting of secondary prevention." Temple et al. (2000) noted that studies have demonstrated a positive correlation between hyperhomocysteinemia and atherosclerosis but that "to date, no studies have assessed the effects on morbidity and mortality when treating high homocysteine concentrations in atherosclerotic patients."

Ueland et al. (2000) note that only placebo-controlled intervention studies with total homocysteine (tHcy) lowering vitamins and clinical endpoints can provide additional valid arguments for the debate over whether tHcy is a causal CVD risk factor. Christen et al. (2000) state that "randomized trials are necessary to test reliably whether lowering homocysteine levels will decrease risks of cardiovascular disease."

Scott and Sutton (1999) concluded following their review that "our present understanding is that tHcy is associated with atherosclerotic disease, probably in a causal role." They continued that "we hope that careful large-scale, prospective, randomized clinical trials can establish the degree of protection afforded the atherosclerotic disease population and the optimal folate dose."

Taken together, these review articles do not indicate that there is significant scientific agreement for a (causal) relationship between homocysteine levels and risk of vascular disease. All of the reviews emphasized the need for controlled clinical trials to establish the validity of a risk reduction relationship between the intakes of the B vitamins and vascular disease. A number of such trials are in progress (Clark and Collins, 1998; Clark, 1998; Hankey and Eikelboom, 1999).

D. Summary regarding significant scientific agreement

In sum, there are no studies that demonstrate a causal relationship between the B vitamins and risk of vascular disease in the general population, or that demonstrate the validity of homocysteine-lowering as a validated surrogate marker of disease risk. Therefore, the agency finds that the more recent data do not alter our November 30, 1999 conclusion that the scientific evidence is not sufficient for a relationship between the B vitamins and reduced risk of vascular disease in the general population. Based on its evaluation of the totality of the publicly available scientific evidence, the agency concludes that there is not significant scientific agreement among qualified experts that a relationship exists between the B vitamins and reduced risk of vascular disease.

This conclusion is consistent with the results of the Institute of Medicine of the National Academy of Sciences (IOM/NAS) (1998) review of the available data. IOM/NAS (1998) found that, although there is an inverse relationship between folate intake and homocysteine concentration, there are conflicting data on the association among indicators of folate status or metabolism, homocysteine concentration, and risk of vascular disease. Specifically, IOM/NAS noted that "whether increasing intake of folate could reduce the risk of vascular disease remains to be demonstrated."

V. Agency's Consideration of A Qualified Claim

In the Pearson implementation notice, the agency stated that it would consider exercising enforcement discretion for a dietary supplement health claim when the following conditions are met: (1) the claim is the subject of a health claim petition that meets the requirements of §101.70; (2) the scientific evidence in support of the claim outweighs the scientific evidence against the claim, the claim is appropriately qualified, and all statements in the claim are consistent with the weight of the scientific evidence; (3) consumer health and safety are not threatened; and (4) the claim meets the general requirements for health claims in § 101.14, except for the requirement that the evidence supporting the claim meets the significant scientific agreement standard and the requirement that the claim be made in accordance with an authorizing regulation. Thus, in the absence of significant scientific agreement, FDA has considered, under Pearson, whether the weight of the scientific evidence in support of the claim outweighs the scientific evidence against the claim, and, if so, whether the use of a qualified claim would be safe.

A. Weight of the Scientific Evidence

As noted in the previous section, there is scientific evidence that intakes of the B vitamins are effective in lowering circulating levels of homocysteine. However, the usefulness of homocysteine as a surrogate marker for cardiovascular disease remains to be demonstrated. Additionally, while there are two non-randomized, non-blinded trials in vascular disease patients showing a positive relationship between intake of folic acid and vitamin B₆ and reduced risk of recurrence of vascular events, there are no studies to date that demonstrate a causal relationship between the B vitamins and risk of vascular disease in the general population. A number of on-going randomized trials are examining the effects of increased intakes of the three vitamins in individuals at risk of a recurrence of a myocardial infarction or stroke.

FDA looked for scientific evidence that would support a qualified claim in the following areas: (1) associations of B vitamins and risk of vascular disease; and (2) associations of homocysteine and risk of vascular disease. FDA focused on the most persuasive types of evidence, i.e., intervention studies and prospective studies from the previous (November 1999) and current reviews. Among observational studies, a finding of beneficial effects in prospective studies would provide the strongest evidence of a relationship between B vitamins and risk of vascular disease (Table 1).

As noted above, associations of B vitamins and homocysteine are well-established in both the general and diseased populations. Therefore, FDA will not further discuss studies to assess this relationship in this section, but rather is relying on the discussion of this topic in Section IV.

1. B vitamins and risk of vascular disease

The non-randomized, non-blinded intervention trials of DeJong et al. (1999) and Vermeulen, Rauwerda, et al. (2000), in patients with established vascular disease, provide evidence for a protective effect of the B vitamins against progression of vascular disease. These studies directly measured intakes of folic acid and vitamin B₆ and also directly measured disease endpoints (i.e., progression of established vascular disease; occurrence of additional vascular events). Although a randomized trial would provide stronger evidence of a beneficial effect, because this type of study controls for potential biases that may arise in the absence of randomization and blinding procedures, the non-randomized, non-blinded trials of DeJong et al. (1999) and Vermeulen, Rauwerda, et al. (2000) suggest a relationship between intake of folic acid and vitamin B₆ and reduced risk of adverse outcomes associated with progression of vascular disease in patient populations.

No corresponding studies in a general population support a direct (causal) relationship between intake of the B vitamins and reduced risk of vascular disease. Thus, uncertainty remains regarding whether the relationship between the B vitamins and reduced risk of vascular disease found in the population with vascular disease would be seen in the general population.

A number of prospective studies have examined the relationship between status for the B vitamins (i.e., vitamin intakes or serum vitamin levels) and risk of vascular disease in populations that were disease-free at the time of enrollment in the studies. Among the reports utilizing data from a follow-up study of the National Health and Nutrition Examination Survey I (NHANES I), that of Ford et al. (1998) found a slight but non-significant inverse association between serum folate and cardiovascular disease mortality. Another prospective study (Giles et al., 1998) reported an inverse association of serum folate with coronary heart disease in persons younger than 55 years and a positive association in persons older than 55 years. The third study utilizing the NHANES I follow-up data (Giles et al., 1995) found a slightly increased risk for stroke associated with low serum folate levels. Follow-up studies of participants in the Nutrition Canada survey showed a significant association between decreased serum folate levels and increased risk of fatal coronary heart disease (Morrison et al., 1998; Morrison et al., 1996). Plasma pyridoxal phosphate but not folate or vitamin B₁₂ was found to be associated with incidence of coronary heart disease among participants in the Atherosclerosis Risk in Communities (ARIC) study (Folsom et al., 1998). A report from the Physicians Health Study (Chasen-Taber et al., 1996) showed a non-statistically significant association of folate or vitamin B₆ with myocardial infarction or coronary heart disease mortality, and a report from the large Nurse's Health Study (Rimm et al., 1998) indicated that high folate and vitamin B₆ intakes were associated with lower incidence of myocardial infarction and fatal coronary heart disease. The cross-sectional study of Selhub et al. (1993), using data from the Framingham population, showed that the lowest levels of folate intake were associated with higher circulating homocysteine levels. The most recent prospective study, that of Voutilainen et al. (2000), showed, in a study of 734 coronary disease free persons, that increased folate was strongly associated with decreased acute coronary events.

In summary, the two non-randomized, non-blinded intervention studies in populations with established vascular disease, most of the prospective studies that started with disease-free populations at the time of enrollment, and the cross-sectional study showed a positive relationship between B vitamin intakes and reduced risk of vascular disease. In the studies noted above that did not show a statistically significant association, the odds ratios for risk of vascular disease associated with reduced vitamin status were generally suggestive of increased risk with poorer vitamin status. Small size may have prevented the studies from showing a statistically significant effect.

Thus, the most persuasive evidence (i.e., the two non-randomized, non-blinded intervention studies) show an effect in a population with established vascular disease, but the ability to generalize their results to the general population is uncertain. The prospective studies are strongly suggestive of a relationship, but because of study design limitations do not allow a determination as to whether the relationship can be specifically attributed to the B vitamin intakes or whether other dietary or lifestyle factors are responsible for the observed relationships.

2. Homocysteine and risk of vascular disease

As noted above, the validity of homocysteine as a surrogate marker for vascular disease risk remains uncertain. The non-randomized, non-blinded intervention studies of DeJong et al. (1999) and Vermeulen, Rauwerda, et al. (2000) that showed a relationship between B vitamin intakes and reduced progression of disease in hyperhomocysteinemic patients with existing vascular disease appeared to be mediated via a reduction in homocysteine levels.

We reviewed a number of other trials that related changes in circulating homocysteine to other physiologic effects, such as carotid wall thickening, endothelial parameters, LDL-cholesterol oxidation, and plaque area. These studies are of limited usefulness, however, because of the validity of these other endpoints as surrogate markers for vascular disease is uncertain (Bostom and Garber, 2000; Probstfield et al., 1993).

A number of prospective studies have examined the relationship between elevated homocysteine and risk of vascular disease in individuals in the general population and in populations with vascular disease. While the prospective study data of Alfthan et al. (1994) indicated that serum homocysteine was not a risk factor for atherosclerotic disease in persons with a confirmed myocardial infarction, the data of Perry (1995) suggested that homocysteine level was an independent risk factor for stroke, and the new data of Aronow and Ahn (2000) suggested that homocysteine was an independent predictor for new coronary event in patients with prior coronary artery disease. The data of Wald (1998) suggested that elevated homocysteine conferred an increased risk of ischemic heart disease. Omland et al. (2000) suggested that homocysteine level was an independent predictor of long-term survival in patients with acute coronary syndromes. Ducloux et al. (2000) studied stable renal transplant recipients and found that total homocysteine was negatively correlated with folate but not vitamin B₁₂, and that homocysteine, age, and serum creatinine were independent risk factors for cardiovascular events.

A review of other available prospective studies shows that some have reported positive associations between intake or status for the B vitamins and risk of vascular disease (Arnesen et al., 1995; Nygard et al., 1997), while others have not (Alfthan et al., 1994; Verhoef, Hennekens, et al., 1997; Verhoef, Kok, et al., 1997; Evans et al., 1997).

The most serious limitations of the studies of DeJong et al. (1999) and Vermeulen, Rauwerda, et al. (2000) include their non-randomized, non-blinded design, their small numbers of patients treated, and the high doses of folic acid and vitamin B₆ used. Although the designs of the prospective studies do not allow a determination as to whether the observed association of B vitamin status with vascular disease is causal, all but two of the studies suggest a relationship between the B vitamins and risk of vascular disease in the general population.

3. Conclusions

Based on its review of the scientific evidence, FDA concludes that the weight of the scientific evidence for a claim relating intake of the B vitamins and reduced risk of vascular disease outweighs the scientific evidence against the claim because:

Based on its review of the available prospective studies, the agency has concluded that the weight of the evidence for the relationship is greater than the evidence against it and that a qualified claim is appropriate.

B. Safety Considerations

Many of the intervention trials mentioned above used levels of the B vitamins that are greatly in excess of recommended daily intakes (e.g., 5- to 50-or more times higher). For this reason, we have considered below the most recent information on the safety of folic acid, vitamin B₆ and vitamin B₁₂ in supplements that might carry a qualified claim.

1. Folic Acid.

FDA addressed issues of safety of high intakes of folate in its health claims and fortification proposed (58 Fed. Reg. 53,254 (1993), and 58 Fed. Reg. 53,305 (1993), respectively) and final (61 Fed. Reg. 8752 (1996), and 61 Fed. Reg. 8781 (1996), respectively) rules. In its proposed rule for the health claim on folate and neural tube birth defects, the agency noted (58 Fed. Reg. at 53,256) that the Public Health Service, in its recommendation for use of folate by women of child-bearing age for reduction in risk of neural tube birth defects (DHHS, PHS, 1992), included a caution statement in its recommendation that "because the effects of higher intakes are not well known but include complicating the diagnosis of vitamin B₁₂ deficiency, care should be taken to keep total folate consumption at less than 1 mg per day, except under the supervision of a physician." FDA stated further in the folate/neural tube defects final rule (61 Fed. Reg. at 8766) that the potential adverse effect that has been most extensively documented is a masking of the anemia of vitamin B₁₂ deficiency, while irreversible neurologic damage progresses. FDA also noted that other groups at risk from excessive intakes of folate include pregnant women, persons on antiseizure (i.e., antiepileptic) medications, and those on antifolate medications, and that there were no data to identify the magnitude of other possible risks of increased folate intake or to establish safe use at daily intakes above 1,000 µg (mcg).

With respect to the safety of foods and supplements bearing the folate/neural tube defects health claim, §101.79(c)(2)(i)(F) states, in relevant part, that "[c]laims on foods that contain more than 100 percent of the Daily Value (DV) (400 mcg) when labeled for use by adults and children 4 or more years of age...shall identify the safe upper limit of daily intake with respect to the DV. The safe upper limit of daily intake value of 1,000 mcg (1 mg) may be included in parentheses." Although these regulations apply to conventional foods and not to dietary supplements, the safety evaluation is relevant to intakes from all sources, including dietary supplements. Once ingested, all forms of folate are converted into biologically active coenzyme forms and participate in folate-dependent enzymatic reactions. Thus, all forms of folate can contribute to potentially excessive intakes.

Safety concerns were also extensively discussed in FDA's final rule on fortification of enriched cereal-grain products (61 Fed. Reg. 8781) and its final rule that amended the food additive regulation for folic acid (61 Fed. Reg. 8797 (1996)) with a final decision to limit the addition of folic acid to specified amounts in a limited range of products (i.e., enriched cereal-products subject to a standard of identity, breakfast cereals, infant formula, medical foods, food for special dietary use, and foods represented as meal-replacement products) to minimize the risk of consumption in excess of 1 mg total folate per day by persons in the general population.

Subsequent to FDA's regulations, the IOM/NAS set the Tolerable Upper Intake Level (UL) for folic acid as 1,000 µg/day from fortified food or supplements. IOM/NAS (1998) recognized that one of the potential hazards of high intakes of folate is the masking of the anemia of vitamin B₁₂ deficiency, in that high intakes of folate may partially and temporarily correct the anemia while the neurological damage of vitamin B₁₂ deficiency progresses. IOM/NAS (1998) set the UL for all adults of 1 mg per day or 1,000 µg/day because of the devastating and irreversible neurological consequences of vitamin B₁₂ deficiency, the data suggesting that pernicious anemia may develop at a younger age in some racial or ethnic groups, and the uncertainty about the extent of occurrence of vitamin B₁₂ deficiency in younger age groups (IOM/NAS, 1998).

Significant increases in folate status indicators among persons in the general population (Lawrence et al., 1999 and 2000) and among women of childbearing age (PHS/DHHS, 2000) have been reported recently. In 1999, the median serum folate levels in one study were so high that they could not be assessed, since they exceeded the maximal value (20.0 ng/ml) of the method used to measure serum folate in the clinical laboratory (Lawrence et al., 2000). The potential for excessive intakes of folate from all sources is of concern, and FDA is monitoring changes in the indicators of folate status through the NHANES surveys. Therefore, consistent with § 101.79(c)(2)(i)(F), FDA intends to exercise enforcement discretion with respect to supplements bearing a qualified claim that contain more than 400 µg folate/dose only if they carry a statement indicating that the upper limit of safe daily intake is 1000 µg folic acid from all sources.

2. Vitamin B₆

No adverse effects have been associated with high intake of vitamin B₆ from food sources (IOM/NAS, 1998). Large oral supplemental doses of pyridoxine have been associated with the development of sensory neuropathy (IOM/NAS, 1998). On the basis of considerations of causality, relevance and the quality and completeness of the database, IOM/NAS selected sensory neuropathy as the critical endpoint on which to base the UL for pyridoxine. For adults 19 years and older, IOM/NAS set the UL at 100 mg/day of vitamin B₆ as pyridoxine.

3. Vitamin B₁₂

With respect to this vitamin, IOM/NAS (1998) concluded, based on a review of data involving high dose intakes, that there appear to be essentially no risks of adverse effects to the general population even at the current ninety-fifth percentile of intake (approximately 37 µg/day)(IOM/NAS, 1998). Although there are extensive data showing no adverse effects associated with high intake of supplemental vitamin B₁₂, the studies in which such intakes were reported were not designed to assess adverse effects.

4. Safety summary

Based on the information above, FDA concludes that use of folic acid, vitamin B₆ and vitamin B₁₂ are safe and lawful under 21 C.F.R. 101.14 provided that daily intakes by adults of folate do not exceed 1 mg/day and daily intakes of vitamin B₆ do not exceed 100 mg/day.

C. Qualified Claim Language

1. Sources and amounts of the B vitamins

FDA has considered whether the proposed claim, which includes the phrase "rich in fresh fruits and vegetables" and which identifies specific amounts of folic acid, vitamin B₆ and vitamin B₁₂, (i.e., "at least 400 µg folic acid, 3 mg vitamin B₆ and 5 µg vitamin B₁₂") is consistent with the available scientific evidence and whether this information should be included in a qualified claim.

a. Fruits and vegetables as sources

FDA notes that, although folate is found in many vegetables and fruits, data obtained from the United States Department of Agriculture's 1995 Continuing Survey of Food Intake by Individuals indicate that the greatest contribution to vitamin B₆ intakes of the U.S. population comes from fortified ready-to-eat cereal, foods with meat, fish, or poultry as the main ingredient, white potatoes and other starchy vegetables and non-citrus fruits (IOM/NAS, 1998).

FDA further notes that vitamin B₁₂ is not a normal constituent of plant foods and that the greatest contribution to vitamin B₁₂ intake in the adult population comes from the category of foods (including sandwiches) with meat, fish or poultry as the main ingredient. Milk and milk drinks and fortified ready-to-eat breakfast cereals also contribute significantly to vitamin B₁₂ intake.

Reference in the claim to fruits and vegetables that are not primary sources of two of the three vitamins that are the subject of the claim could render the claim misleading. Accordingly, FDA concludes that inclusion of the phrase "rich in fresh fruits and vegetables" in a qualified claim would be inappropriate if it implies these are good sources of vitamin B₆ and vitamin B₁₂.

b. Amounts

It is well recognized that folic acid, vitamin B₆, and vitamin B₁₂ can lower homocysteine, but neither the "optimal" intakes of these vitamins relative to homocysteine lowering nor the target homocysteine level is currently known (IOM/NAS, 1998). In randomized studies reported above (e.g., Appel et al., 2000; Bronstrup et al., 1998; Chait et al., 1999; den Heijer et al., 1998; Riddell et al., 2000), significant reductions in homocysteine were achieved, both in persons with risk factors for vascular disease and in healthy individuals, with regimens of dietary intakes or supplement intakes that provided 400 µg to 10 mg (i.e., 400 to 10,000 µg) folate, 0 to 300 mg vitamin B₆, and 6.5 µg to 1,000 µg vitamin B₁₂. (See Table 2.) These data do not allow a determination of a recommended or optimal intake to achieve reductions in homocysteine because dose-response curves cannot be developed from these data. Moreover, the validity of homocysteine as a surrogate marker for vascular disease risk remains uncertain, so identifying effective intake levels of the B vitamins with respect to decreased risk of vascular disease is problematic.

More directly, the absence of data documenting a causal relationship between intake of B vitamins and decreased risk of vascular disease, the absence of data linking homocysteine lowering to decreased risk of vascular disease, and the paucity of dose response data precludes establishing effective intake levels. In deriving new Dietary Reference Intakes for a variety of B vitamins, IOM/NAS (1998) specifically stated that with respect to folate, "it is premature to consider vascular disease risk as an indicator for setting the estimated Average Requirement and Recommended Dietary Allowance for folate" (IOM/NAS, 1998). With respect to vitamin B₆, IOM/NAS (1998) noted that, although the increase in plasma homocysteine concentration after a methionine load or a meal is responsive to and primarily affected by vitamin B₆ status, the data are not sufficient to support using this as the criterion on which to base the estimated average requirement. With respect to vitamin B₁₂, IOM/NAS (1998) noted that serum total homocysteine is commonly elevated in elderly persons whose folate status is normal but who have a clinical response to treatment with vitamin B₁₂. Because a lack of folate, vitamin B₆ or both also results in an elevated serum and plasma homocysteine, this indicator has poor specificity and does not provide a useful basis for deriving the estimated average requirement.

In short, the available scientific information does not support identification of specific amounts of folic acid, vitamin B₆ and vitamin B₁₂ that might be useful for reduction in risk of vascular disease. FDA therefore concludes that a claim that implied that specific amounts or combinations of folic acid, vitamin B₆, or vitamin B₁₂, are necessary to reduce the risk of vascular disease would be misleading.

2. Qualified Claim Language

The agency would consider the following claim to be appropriately qualified:

It is known that diets low in saturated fat and cholesterol may reduce the risk of heart disease. The scientific evidence about whether folic acid, vitamin B₆ and vitamin B₁₂ may also reduce the risk of heart disease and other vascular diseases is suggestive, but not conclusive. Studies in the general population have generally found that these vitamins lower homocysteine, an amino acid found in the blood. It is not known whether elevated levels of homocysteine may cause vascular disease or whether high homocysteine levels are caused by other factors. Studies that will directly evaluate whether reducing homocysteine may also reduce the risk of vascular disease are not yet complete.

The relevant elements of the claim include:

D. Relevant 21 C.F.R. § 101.14 requirements and requirements from 21 C.F.R. § 101.79

Consistent with the Pearson implementation notice, the agency intends to exercise its enforcement discretion with respect to the qualified claim when the claim meets the general requirements for health claims in 21 C.F.R. § 101.14 (65 Fed. Reg. at 59,856). FDA finds that the provision in § 101.14 (d)(2)(2)(vii) stating, "If the claim is about the effects of consuming the substance at other than decreased dietary levels, ...the claim must specify the daily dietary intake necessary to achieve the claimed effect..." does not apply to the qualified claim for folic acid, vitamin B₁₂ and vitamin B₆ and reduced risk of vascular disease. The scientific evidence for this relationship is merely suggestive and does not support the identification of a possible level of effect. Therefore, the agency would consider any labeling suggesting that particular levels of folic acid, vitamin B₁₂ and vitamin B₆ are useful in achieving a claimed effect to be false and misleading under section 403(a) of the act.

Moreover, compliance with certain criteria in § 101.14 will have to be evaluated after-the-fact, because they involve information or circumstances that cannot be determined a priori. For example, FDA will not be able to determine whether the entire claim appears in one place without intervening material, as required by § 101.14(d)(2)(iv), until it actually sees the claim on products in the marketplace.

The agency intends to exercise its enforcement discretion with respect to supplements bearing the qualified claim when the supplements comply with §101.79(c)(2)(i)(F), which states, in relevant part, that "[c]laims on foods that contain more than 100 percent of the Daily Value (DV) (400 mcg) when labeled for use by adults and children 4 or more years of age...shall identify the safe upper limit of daily intake with respect to the DV. The safe upper limit of daily intake value of 1,000 mcg (1 mg) may be included in parentheses."

The agency intends to exercise its enforcement discretion with respect to supplements bearing the qualified claim when the supplements comply with § 101.79(c)(ii)(B) which states that "[d]ietary supplements shall meet the United States Pharmacopeia (USP) standards for disintegration and dissolution, except that if there are no applicable USP standards, the folate in the dietary supplement shall be shown to be bioavailable under the conditions of use stated on the product label."

E. Other Considerations

Folic acid, vitamin B₁₂ and vitamin B₆ dietary supplements bearing the qualified claim, which meet the conditions for the exercise of FDA's enforcement discretion in the Pearson implementation notice and the other conditions set forth in this letter, must still meet all applicable statutory and regulatory requirements under the act. For example, such supplements must be labeled consistent with 21 C.F.R. § 101.36(b)(3). Such supplements should be manufactured in a manner that will not adulterate or misbrand the product. Dietary supplements must not pose an unreasonable risk of illness or injury to the consumer or contain substances that may render the product injurious to health.

VI. Conclusion

FDA has set forth conditions, consistent with, and in addition to, those described in the Pearson implementation notice, under which it intends to exercise enforcement discretion with respect to the use of the qualified claim, as described above, on dietary supplements containing folic acid, vitamin B₆, and vitamin B₁₂ and risk of vascular disease.

FDA concludes that there is not significant scientific agreement for an unqualified claim about the relationship between folic acid vitamin B₆ and vitamin B₁₂ and reduced risk of vascular disease. Thus, a health claim stating that "Folic acid, vitamin B₆ and vitamin B₁₂ may reduce the risk of vascular disease" would be misleading. However, the weight of the scientific evidence for a health claim for the three vitamins outweighs the scientific evidence against such a claim, and the qualified claim that FDA has set forth in this letter is consistent with the weight of the scientific evidence.

Scientific information is subject to change, as are consumer consumption patterns. FDA intends to evaluate new evidence that becomes available to determine whether the weight of the evidence shifts, either in favor of an unqualified claim or in favor of no longer exercising enforcement discretion. For example, scientific evidence may later become available that will support significant scientific agreement or that will no longer support the use of a qualified claim, or that may raise concerns about the conditions that FDA has outlined for the use of this qualified claim. If and when such information becomes available, FDA intends to inform you of that by letter.

We hope that this letter clarifies the issues related to labeling of your product.

2. Your October 19, 2000, letter identified information that became publicly available after you had submitted your May 25, 1999, petition, and after the agency's November 1999 letter. After it issued that letter, the agency did not issue a Federal Register notice asking the public to identify additional, publicly available scientific information or to comment on your petition, but the agency continued to follow the literature, as described above. The agency's own recent literature search identified 22 of the 25 articles that you submitted.
The agency considered three of the 25 articles not relevant to the topic of the health claim. Green et al. (1999) report the effects of intravenous infusions of various minerals and several vitamins, including vitamin B6 and vitamin B12, but not folic acid, on forearm blood flow in eight vascular disease patients. We considered this report not relevant to the topic of the health claim, because the claim involves oral intake of the B vitamins and risk of vascular disease, and because forearm blood flow is not considered a validated marker for vascular disease. The study of Jakubowski et al. (2000) addressed the metabolism of homocysteine in cultured umbilical venous endothelial cells, and the study of Xu et al. (2000) examined endothelial senescence, also in cultured umbilical venous endothelial cells. These preclinical studies were not relevant for the reasons stated above.

Abdelmouttaleb, I., N. Danchin, I. Aimone-Gastin, F. Namour, M. Angioi, M.-A. Gelot, N. Bennani, D. Lambert, C. Jeandel, and J.L. Gueant. Homocysteine, vitamins B₆, B₁₂, folate, and risk of coronary artery disease in patients undergoing diagnostic coronary angiography. Amino Acids. 2000;18:139-146.

Alfthan, G., J. Pekkanen, M. Jauhiainen, J. Pitkaniemi, M. Karvonen, J. Tuomilehto, J.T. Salonen, and C. Ehnholm. Relation of serum homocysteine and lipoprotein(a) concentrations to atherosclerotic disease in a prospective Finnish population-based study. Atherosclerosis. 1994;106:9-19.

Andreotti, F., F.Burzotta, A.Manzoli, and K. Robinson. Homocysteine and risk of cardiovascular disease. Journal of Thrombosis and Thrombolysis. 2000;9:13-21.

Appel, L.J., E.R. Miller, S.H. Jee, R. Stolzenberg-Solomon, P.-H. Lin, T. Erlinger, M.R. Nadeau, and J. Selhub. Effect of dietary patterns on serum homocysteine: Results of a randomized, controlled feeding study. Circulation. 2000;102:852-857.

Arnesen, E., H. Refsum, K.H. Bonaa, P.M. Ueland, O.H. Forde, and J.E. Nordrehaug. Serum total homocysteine and coronary heart disease. International Journal of Epidemiology. 1995;24(4):704-709.

Aronow, W.S. and C. Ahn. Increased plasma homocysteine is an independent predictor of new coronary events in older persons. American Journal of Cardiology. 2000;86(3):346-347.

Aronow, W.S. and C. Ahn. Association between plasma homocysteine and coronary artery disease in older persons. American Journal of Cardiology. 1997;80(9):1216-1218.

Bellamy, M.F., I.F.W. McDowell, M.W. Ramsey, M. Brownlee, R.G. Newcombe, and M.J. Lewis. Oral folate enhances endothelial function in hyperhomocysteinemia subjects. European Journal of Clinical Investigation. 1999;29:659-662.

Bostom,A.G.and C.Garber. Endpoints for homocysteine-lowering trials. The Lancet. 2000;355:511- 512.

Bostom,A.G. Homocysteine: "Expensive creatinine" or important, modifiable risk factor for arteriosclerotic outcomes in renal transplant recipients? Journal of the American Society of Nephrology. 2000;11:149-151.

Brattstrom, L. and D.E.L. Wilcken. Homocysteine and cardiovascular disease: cause or effect? American Journal of Clinical Nutrition. 2000;72:315-323.

Bronstrup, A., M. Hages, R. Prinz-Langenohl, and K. Pietrzik. Effects of folic acid and combinations of folic acid and vitamin B-12 on plasma homocysteine concentrations in healthy, young women. American Journal of Clinical Nutrition. 1998;68:1104-1110.

Bunout, D., A. Garrido, M. Suazo, R. Kauffman, P. Venegas, P. de la Maza, M. Petermann, and S. Hirsch. Effects of supplementation with folic acid and antioxidant vitamins on homocysteine levels and LDL oxidation in coronary patients. Nutrition. 2000;16:107-110.

Bunout, D., M. Petermann, S. Hirsch, P. de la Maza, M. Suazo, G. Barrera, and R. Kauffman. Low serum folate but normal homocysteine levels in patients with atherosclerotic vascular disease and matched healthy controls. Nutrition. 2000;16:434-438.

Chait, A., M.R. Malinow, D.N. Nevin, C.D. Morris, R.L. Eastgard, P. Kris-Etherton, F.X. Pi- Sunyer, S. Oparil, L.M. Resnick, J.S. Stern, R.B. Haynes, D.C. Hatton, J.A. Metz, S. Clark, M. McMahon, S. Holcomb, M.E. Reusser, G.W. Snyder, and D.A. McCarron. Increased dietary micronutrients decrease serum homocysteine concentrations in patients at high risk of cardiovascular disease. American Journal of Clinical Nutrition. 1999;70:881-887.

Chambers, J.C., O.A. Obeid, H. Refsum, P. Ueland, D. Hackett, J. Hooper, R.M. Turner, S.G. Thompson, and J.S. Kooner. Plasma homocysteine concentrations and risk of coronary heart disease in UK Indian Asian and European men. The Lancet. 2000;355:523-527.

Chao, C-L., H.-H. Tsai, C.-M. Lee, S.-M. Hsu, J.-T. Kao, K.-L. Chien, F.-C. Sung, and Y.-T. Lee. The graded effect of hyperhomocysteinemia on the severity and extent of coronary atherosclerosis. Atherosclerosis. 1999;147:379-386.

Chasan-Taber, L., J. Selhub, I.H. Rosenberg, R. Malinow, P. Terry, P.V. Tishler, W. Willett, C.H. Hennekens, and M.J. Stampfer. A prospective study of folate and vitamin B₆ and risk of myocardial infarction in US physicians. Journal of the American College of Nutrition. 1996;15(2):136-143.

Christen, W.G., U.A. Ajani, R.J. Glynn, and C.H. Hennekens. Blood levels of homocysteine and increased risks of cardiovascular disease: Causal or casual? Archives of Internal Medicine. 2000; 160:422-434.

Clarke, R. Lowering blood homocysteine with folic acid based supplements: meta-analysis of randomized trials: Homocysteine Lowering Trialists' Collaboration. British Journal of Medicine. 1998;316:894-898.

Clarke, R. and R. Collins. Can dietary supplements with folic acid or vitamin B₆ reduce cardiovascular risk? Design of clinical trials to test the homocysteine hypothesis of vascular disease. Journal of Cardiovascular Risk. 1998;5:249-255.

Constans, J., A.D. Blann, F. Resplandy, F. Parrot, M. Renard, M. Seigneur, V. Guerin, M. Boisseau, and C. Conri. Three months supplementation of hyperhomocysteinaemic patients with folic acid and vitamin B6 improves biological markers of endothelial dysfunction. British Journal of Haematology. 1999;107:776-778.

DeJong, S.C., C.D.A. Stehouwer, M. van den Berg, T.W. Guerts, L.M. Bouter, and J.A. Rauwerda. Normohomocysteinaemia and vitamin-treated hyperhomocysteinaemia are associated with similar risks of cardiovascular events in patients with premature peripheral arterial occlusive disease: A prospective cohort study. Journal of Internal Medicine. 1999;246:87-96.

den Heijer, M., I.A. Brouwer, G.M.J. Bos, H.J. Blom, N.M.J. van der Put, A.P. Spaans, F.R.Rosendaal, C.M.G. Thomas, H.L. Haak, P.W. Wijermans, and W.B.J. Gerrits. Vitamin supplementation reduces blood homocysteine levels: a controlled trial in patients with venous thrombosis and healthy volunteers. Arteriosclerosis, Thrombosis, and Vascular Biology. 1998;18:356-361.

Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention. Folate status in women of childbearing age United States, 1999. Morbidity and Mortality Weekly Report. 2000;49(42):962-965.

Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention. Recommendations for the Use of Folic Acid to Reduce the Number of Cases of Spina Bifida and Other Neural Tube Defects. Mortality and Morbidity Weekly Report. 1992;41/No.RR- 14:1-7.

Ducloux, D., G. Motte, B. Challier, R.Gibey, and J.-M. Chalopin. Serum total homocysteine and cardiovascular disease occurrence in chronic, stable renal transplant recipients: A prospective study. Journal of the American Society of Nephrology. 2000;11:134-137.

Evans, R.W., B.J. Shaten, J.D. Hempel, J.A. Cutler, L.H. Kuller, for the MRFIT Research Group. Homocyst(e)ine and risk of cardiovascular disease in the Multiple Risk Factor Intervention Trial. Arteriosclerosis, Thrombosis, and Vascular Biology. 1997;17:1947-1953.

Folsom, A.R., F.J. Nieto, P.G. McGovern, M.Y. Tsai, M.R. Malinow, J.H. Eckfeldt, D.L. Hess, and C.E. Davis. Prospective study of coronary heart disease incidence in relation to fasting total homocysteine, related genetic polymorphisms, and B vitamins: The Atherosclerosis Risk in Communities (ARIC) study. Circulation. 1998;98:204-210.

Ford, E.S., T.E. Byers, and W.H. Giles. Serum folate and chronic disease risk: Findings from a cohort of United States adults. International Journal of Epidemiology. 1998;27:592-598.

Giles, W.H., S.J. Kittner, R.F. Anda, J.B. Croft, and M.L. Casper. Serum folate and risk for ischemic stroke: First National Health and Nutrition Examination Survey Epidemiologic Follow-up Study. Stroke. 1995;26: 1166-1170.

Giles, W.H., S.J. Kittner, J.B. Croft, R.F. Anda, M.L. Casper, and E.S. Ford. Serum folate and risk for coronary heart disease: Results from a cohort of US adults. Annals of Epidemiology. 1998;8: 490-496.

Green, D.J., J.G. O'Driscoll, A. Maiorana, N.B. Scrimgeour, R. Weerasooriya, and R.R. Taylor. Effects of chelation with EDTA and vitamin B therapy on nitric oxide-related endothelial vasodilator function. Clinical and Experimental Pharmacology and Physiology. 1999;26:853-856.

Hackam, D.G., J.C. Peterson, and J.D. Spence. What level of plasma homocyst(e)ine should be treated? Effects of vitamin therapy on progression of carotid atherosclerosis in patients with homocyst(e)ine levels above and below 14 µmol/L. American Journal of Hypertension. 2000;13:105-110.

Hak,A.E., K.H. Polderman, I.C.D. Westendorp, C. Jakobs, A. Hofman, J.C.M. Witteman, and C.D.A. Stehouwer. Increased plasma homocysteine after menopause. Atherosclerosis. 2000;149:163-168.

Hankey, G.J. and J.W. Eikelboom. Homocysteine and vascular disease. The Lancet. 1999;354:407- 413.

(a) Hoogeveen, E.K., P.J. Kostense, C. Jakobs, J.M. Dekker, G. Nijpels, R.J. Heine, L.M. Bouter, and C.D.A. Stehouwer. Hyperhomocysteinemia increases risk of death, especially in type 2 diabetes: 5-year follow-up of the Hoorn Study. Circulation. 2000;101:1506-1511.

(b) Hoogeveen, E.K., P.J. Kostense, C. Jakobs, J.A. Rauwerda, J.M. Dekker, G. Nijpels, L.M. Bouter, R.J. Heine, and C.D.A. Stehouwer. Hyperhomocysteinemia is not associated with isolated crural arterial occlusive disease: The Hoorn Study. Journal of Internal Medicine. 2000;247:442-448.

Hughes, K. and C.-N. Ong. Homocysteine, folate, vitamin B12, and cardiovascular risk in Indians, Malays, and Chinese in Singapore. Journal of Epidemiology and Community Health. 2000;54:31-34.

Institute of Medicine, National Academy of Sciences. Dietary Reference Intakes for Thiamin, Riboflavin, Niacin, Vitamin B₆, Folate, Vitamin B₁₂, Pantothenic Acid, Biotin and Choline. National Academy Press, Washington, DC., 1998; pages 260-264.

Jakubowski, H., L. Zhang, A. Bardeguez, and A. Aviv. Homocysteine thiolactone and protein homocysteinylation in human endothelial cells: Implications for atherosclerosis. Circulation Research. 2000;87: 45-51.

Kristensen, B., J. Malm, T.K. Nilsson, J. Hultdin, B. Carlberg, G. Dahlen, and T.Olsson. Hyperhomocysteinemia and hypofibrinolysis in young adults with ischemic stroke. Stroke. 1999;30:974-980.

Kullo, I.J., G.T. Gau, and A.J. Tajik. Novel risk factors for atherosclerosis. Mayo Clinic Proceedings. 2000;75:369-380.

Langman, L.J., J.G. Ray, J. Evrovski, E. Yeo, and D.E.C. Cole. Hyperhomocyst(e)inemia and the increased risk of venous thromboembolism: More evidence from a case-control study. Archives of Internal Medicine. 2000;160:961-964.

Lawrence, J.M., V. Chiu, D.B. Petitti. Fortification of foods with folic acid (Letter to the Editor). New England Journal of Medicine. 2000;343(13):970-972.

Lawrence, J.M., D.B. Petitti, M. Watkins, and M.A. Umekubo. Trends in serum folate after food fortification. The Lancet. 1999;354:915-916.

Leowattana,W., N. Mahanonda, K.Bhuripunyo, and S. Pokum. Association between serum homocysteine, vitamin B12, folate and Thai coronary artery disease patients. Journal Of the Medical Association Of Thailand. 2000;83(5): 536-542.

Lobo, A., A. Naso, K. Arheart, W.D. Kruger, T. Abou-Ghazala, F. Alsous, M. Nahlawi, A. Gupta, A. Moustapha, F. van Lente, D.W.Jacobsen, and K.Robinson. Reduction in homocysteine levels in coronary artery disease by low-dose folic acid combined with vitamins B₆ and B₁₂. American Journal of Cardiology. 1999;83:821-825.

Mansoor, M.A., C. Bergmark, S.J. Haswell, I.F. Savage, P.H. Evans, R.K. Berge, A.M. Svardal, and O.Kristensen. Correlation between plasma total homocysteine and copper in patients with peripheral vascular disease. Clinical Chemistry. 2000;46(3):385-391.

Martyn,C.N. Serum homocysteine and risk of coronary heart disease in UK Indian Asians. The Lancet. 2000;355, 512-513. 2000.

Mazza, A., C. Motti, A. Nulli, G. Marra, A. Gnasso, A. Pastore, G. Federici, and C. Cortese. Lack of association between carotid intima-media thickness and methylenetetrahydrofolate reductase gene polymorphism or serum homocysteine in non-insulin-dependent diabetes mellitus. Metabolism. 2000;49(6):718-723.

McGregor, D., B. Shand, and K. Lynn. A controlled trial of the effect of folate supplements on homocysteine, lipids and hemorheology in end-stage renal disease. Nephron. 2000;85:215-220.

Morrison, H.I., L.F. Ellison, D. Schaubel, and D.T.Wigle. Relationship of dietary folate and vitamin B6 with coronary heart disease in women [Letter to the Editor]. Journal of the American Medical Association. 1998;280(5):417-418.

Morrison, H.I., D. Schaubel, M. Desmeules and D.T. Wigle. Serum folate and risk of fatal coronary heart disease. Journal of the American Medical Association. 1996;275(24):1893-1896.

Nygard, O., J.E. Nordrehaug, H. Refsum, P.M. Ueland, M. Farstad, and S.E. Vollset. Plasma homocysteine levels and mortality in patients with coronary artery disease. New England Journal of Medicine. 1997;337:230-236.

Omland, T., A. Samuelsson, M. Hartford, J. Herlitz, T. Karlsson, B. Christensen, and K. Caidahl. Serum homocysteine concentration as an indicator of survival in patients with acute coronary syndromes. Archives of Internal of Medicine. 2000;160:1834-1840.

Perry, I.J., H. Refsum, R.W. Morris, S.B. Ebrahim, P.M. Ueland, and A.G. Shaper. Prospective study of serum total homocysteine and risk of stroke in middle-aged British men. The Lancet. 1995;346:1395-1398.

Peterson, J.C. and J.D. Spence. Vitamins and progression of atherosclerosis in hyper-homocyst(e)inaemia. The Lancet. 1998;351:263.

Probstfield, J.L., R.P. Byington, D.A. Egan, M.A. Espeland, S.E. Margitic, W.A. Riley, and C.D. Furberg. Methodological issues facing studies of atherosclerotic change. Circulation. 1993;87[suppl II]:II-74-II-81.

Riddell, L.J., A. Chisholm, S. Williams, and J.I. Mann. Dietary strategies for lowering homocysteine concentrations. American Journal of Clinical Nutrition. 2000;71:1448-1454.

Rimm, E.B., W.C. Willett, F.B. Hu, L. Sampson, G.A. Colditz, J.E. Manson, C. Hennekens, and M.J. Stampfer. Folate and vitamin B₆ from diet and supplements in relation to risk of coronary heart disease among women. Journal of the American Medical Association. 1998;279:359-364.

Scott, C.H. and M. St. John Sutton. Homocysteine: Evidence for a causal relationship with cardiovascular disease. Cardiology in Review. 1999;7(2):101-107.

Selhub, J., P.F. Jacques, P.W.F. Wilson, D. Rush, and I.H. Rosenberg. Vitamin status and intake as primary determinants of homocysteinemia in an elderly population. Journal of the American Medical Association. 1993;270(22):2693-2698.

Temple, M.E., A.B. Luzier, and D.J. Kazierad. Homocysteine as a risk factor for atherosclerosis. Annals of Pharmacology. 2000;34:57-65.

Title, L.M., P.M. Cummings, K. Giddens, J.J. Genest, and B.A. Nassar. Effect of folic acid and antioxidant vitamins on endothelial dysfunction in patients with coronary artery disease. Journal of the American College of Cardiology. 2000;36(3):758-765.

Tsai, M.Y., D.K. Arnett, J.H. Eckfeldt, R.R. Williams, and R.C. Ellison. Plasma homocysteine and its association with carotid intimal-medial wall thickness and prevalent coronary heart disease: NHLBI Family Heart Study. Atherosclerosis. 2000;151:519-524.

Turgan, N., B. Boydak, S. Habif, S. Apakkan, D. Ozmen, I. Mutaf, and O. Bayindir. Plasma homocysteine levels in acute coronary syndromes. Japanese Heart Journal. 1999;40:729-736.

Ueland, P.M., H. Refsum, S.A.A. Beresford, and S.E. Vollset. The controversy over homocysteine and cardiovascular risk. American Journal of Clinical Nutrition. 2000;72:324-332.

van der Griend, R., F.J.L.M. Haas, D.H. Biesma, M. Duran, J.A.Th. Meuwissen, and J-D. Banga. Combination of low-dose folic acid and pyridoxine for treatment of hyperhomocysteinaemia in patients with premature arterial disease and their relatives. Atherosclerosis. 1999;143:177-183.

van der Molen, E.F., G.E. Arends, W.L.D.M. Nelen, N.J.M. van der Put, S.G. Heil, T.K.A.B. Eskes, and H.J. Blom. A common mutation in the 5,10-methylenetetrahydrofolate reductase gene as a new risk factor for placental vasculopathy. American Journal of Obstetrics and Gynecology. 2000;182:1258-1263.

Verhoef, P., C.H. Hennekens, R.H. Allen, S.P. Stabler, W.C. Willett, and M.J. Stampfer. Plasma total homocysteine and risk of angina pectoris with subsequent coronary artery bypass surgery. American Journal of Cardiology. 1997;79:799-801.

Verhoef, P., C.H. Hennekens, M.R. Malinow, F.J. Kok, W.C. Willett, W.C. and M.J. Stampfer. A prospective study of plasma homocyst(e)ine and risk of ischemic stroke. Stroke. 1994;25(10):1924-1930.

Verhoef, P., F.J. Kok, D.A.C.M. Kruyssen, E.G. Schouten, J.C.M. Witteman, D.E. Grobbee, P.M. Ueland,and H. Refsum. Plasma total homocysteine, B vitamins, and risk of coronary atherosclerosis. Arteriosclerosis, Thrombosis, and Vascular Biology. 1997;17:989-995.

Vermeulen, E.G.J., J.A. Rauwerda, P. Erix, S.C .de Jong, J.W.R. Twisk, C. Jakobs, R.J.G.M. Witjes, and C.D.A. Stehouwer. Normohomocysteinaemia and vitamin-treated hyperhomocysteinaemia are associated with similar risks of cardiovascular events in patients with premature atherothrombotic cerebrovascular disease: A prospective cohort study. Netherlands Journal of Medicine. 2000;56:138-146.

Vermeulen, E.G.J., C.D.A. Stehouwer, J.W.R. Twisk, M. van den Berg, S.C. de Jong, A.J.C. Mackaay, C.M.C. van Campen, F.C. Visser, C.A.J.M. Jakobs, E.J. Bulterijs, and J.A. Rauwerda. Effect of homocysteine-lowering treatment with folic acid plus vitamin B₆ on progression of subclinical atherosclerosis: a randomized, placebo-controlled trial. The Lancet. 2000;355:517-522.

Voutilainen, S., T.A. Lakka, E. Porkkala-Sarataho, T. Rissanen, G.A. Kaplan, and J.T. Salonen. Low serum folate concentrations are associated with an excess incidence of acute coronary events: The Kuopio Ischaemic Heart Disease Risk Factor Study. European Journal of Clinical Nutrition. 2000;54:424-428.

Wald, N.J., H.C. Watt, M.R. Law, D.G. Weir, J. McPartlin, and J.M. Scott. Homocysteine and ischemic heart disease: Results of a prospective study with implications regarding prevention. Archives of Internal Medicine. 1998;158:862-867.

Weiss, N., A. Feussner, S. Hailer, F.A. Spengel, C. Keller, and G .Wolfram. Influence of folic acid, pyridoxal phosphate and cobalamin on plasma hmocyst(e)ine levels and the susceptibility of low- density lipoprotein to ex-vivo oxidation. European Journal of Medical Research. 1999;4:425-432.

Willinek, W.A., M. Ludwig, M. Lennarz, T. Holler, and K.O. Stumpe. High-normal serum homocysteine concentrations are associated with an increased risk of early atherosclerotic carotid artery wall lesions in healthy subjects. Journal of Hypertension. 2000;18:425-430.

Wilmink, H.W., E.S.G. Stroes, W.D. Erkelens, W.B. Gerritsen, R. Wever, J.-D. Danga, and T.J. Rabelink. Influence of folic acid on postprandial endothelial dysfunction. Arteriosclerosis, Thrombosis, and Vascular Biology. 2000;20:185-188.

Woo, K.S., P. Chook, Y.I. Lolin, J.E. Sanderson, C. Metreweli, and D.S. Celermajer. Folic acid improves arterial endothelial function in adults with hyperhomocystinemia. Journal of the American College of Cardiology. 1999;34(7):2002-2006.

Woodside, J.V., J.W.G. Yarnell, D. McMaster, I.S. Young, D.L. Harmon, E.E. McCrum, C.C. Patterson, K.F. Gey, A.S. Whitehead, and A. Evans. Effect of B-group vitamins and antioxidant vitamins on hyperhomocysteinemia: A double-blind, randomized, factorial-design, controlled trial. American Journal of Clinical Nutrition. 1998;67:858-866.

Woodside, J.V., I.S. Young, J.W.G. Yarnell, H.E. Roxborough, D. McMaster, E.E. McCrum, K.F. Gey, and A. Evans. Antioxidants, but not B-group vitamins increase the resistance of low-density lipoprotein to oxidation: a randomized, factorial design, placebo-controlled trial. Atherosclerosis. 1999;144:419-427.

Xu, D., R. Neville, and T. Finkel. Homocysteine accelerates endothelial cell senescence. FEBS Letters. 2000;470:20-24.

TABLE 1: B VITAMINS AND RISK OF VASCULAR DISEASE - OBSERVATIONAL STUDIES
Type of study	Authors	Participants/Groups/Numbers	Measurements	Duration d/wk/mo/yr	Findings
Cross-sectional study	Selhub et al., 1993	Elderly participants in Framingham study. Adult survivors, aged 67-96 years, 1,160	B vitamin index   Hcy µmol/L Prevalence % 1 Highest intake 9.5 9.4 2 10.4 15.6 3 11.7 25.5 4 13.0 35.5 5 Lowest intake 14.8 53.2 "Index" intakes for the three B vitamins folate, B6 and B12: 1, (highest intakes), all vitamins >70th percentile 2, all vitamins >50th percentile and at least 1 vitamin <70th percentile 3, vitamin intakes both >50th and <50th percentile 4, all vitamins <50th percentile and at least 1 vitamin >30th percentile 5, (lowest intakes), all three vitamins < 30th percentile.		Mean Hcy and prevalence of high Hcy increased significantly across categories of B vitamin intake indices. Prevalence of high Hcy was almost 6-fold greater among subjects in the lowest intake index category compared with subjects in the highest intake index category.
Prospective study - Coronary heart disease	Afthan et al., 1994	MI cases 191 Controls 269	Hcy 9.6 µmol/L 9.6 µmol/L   NS	Follow-up 9 yr	Serum Hcy is not a risk factor for atherosclerotic disease. RR (95%CI): 1.3 (0.6-2.4), NS.
"	Arnesen et al., 1995	CHD cases 122 Controls 478	Hcy 12.7 µmol/L 11.3 µmol/L p <0.001	Follow-up 4 yr	Serum tHcy independent predictor for risk of MI in general population.
"	Chasen-Taber et al., 1996	MI (men) cases 333 Controls 333		Follow-up 7.5 yr	Low fol, low B6 non-sig assoc with risk of MI. RR (95% CI): 1.7 (0.9-3.3), NS.
"	Morrison et al., 1996	Nutrition Canada Survey Men, women (35-79 yr) 5,056 CHD deaths 165		Follow-up 15 years	RR and 95% CI of fatal CHD by quartiles of ser fol: RR=1.69 (CI, 1.10-2.61) for lowest (ser fol <6.8 nmol/L) vs highest category (ser fol > 13.6 nmol/L).
"	Verhoef et al., 1997	Angina w cor art surg 149 Controls 149	Hcy 10.9 µmol/L 10.4 µmol/L   NS	Follow-up 9 yr	No assoc between plas Hcy & risk of angina with subseq bypass surgery. RR (95% CI): 1.1 (0.4-3.2) NS.
"	Evans et al., 1997	MI, CHD deaths 240 Controls 472	Hcy 12.7 µmol/L 12.9 µmol/L   NS	Follow-up 7 yr	No assoc between Hcy and heart disease. RR 95% CI): 0.9 (0.6-1.6), NS.
"	Nygard et al., 1997	Angiographically confirmed coronary artery disease 587	-	Follow-up 4.6 yr	Plas Hcy-strong predictor mortality in patients with conf cor art dis. RR (95% CI): 4.5 (1.2-16.6).
"	Ford et al., 1998	Nationally representative sample, Adults of NHANES Follow-up study 3,059	Examined effect of serum folate concentration on mortality and chronic disease incidence.	Follow-up 19 years	Hazards ratio for disease of the circulatory system, 1.31 (95% CI, 0.82-2.12). Hazards ratio for all-cause mortality, 1.26 (95% CI, 1.01-1.57) for participants with ser fol <9.3 nmol/L vs other participants. Low levels of ser fol may be associated with mortality from all-causes and cardiovascular disease. Study lacked power to adequately examine disease-specific endpoints.
"	Giles et al., 1998	1st NHANES Follow-up Cohort 1,921 CHD 284		Follow-up 20 yr	Association between ser fol and risk for CHD strongest for persons aged 35-55 years: RR 2.4 (95% CI, 1.1-5.2) for lowest quartile (<9.9 nmol/L) vs highest quartile (>21.8 nmol/L).
"	Morrison et al., 1998	Nutrition Canada Survey cohort Men and women 5,056 CHD deaths 165		Follow-up 15 years	RR and 95% CI of CHD by quartiles of serum folate: RR=0.52 (0.24-1.13) for highest quartile (serum folate >12.7 nmol/L) vs lowest quartile (serum folate < 6.8 nmol/L).
"	Rimm et al., 1998 Nurses' Health Study	Cohort, women 80,082 Nonfatal MI 658 Fatal CHD 281		Follow-up 14 years	RR (95%CI) of CHD for highest quintile of folate intake (696 µg/day) = 0.69 (0.55-0.87) and for highest quintile of vitamin B6 intake (4.6 mg/day) = 0.67 (0.53-0.85)
"	Wald et al., 1998	Ischemic heart disease deaths 229 Controls 1,126	Hcy 13.1 µmol/L 11.8 µmol/L   p <0.001	Follow-up 8.7 yr	33% in ischemic heart disease/5 µmol/L in Hcy. RR (95% CI): 2.9 (2.0-4.1).
"	Folsom et al., 1998	Cases of CHD 232 Reference cohort 537	Hcy 8.9 µmol/L 8.5 µmol/L   NS	Follow-up Aver, 3.3 yr	No diff Hcy cases/controls; plasma vit B6 assoc with CHD in women. RR (95% CI): 1.3 (0.5-3.2), NS.
"	Ducloux et al., 2000	Stable renal transplant recipients 207		Follow-up Mean, 21.2 mo	tHcy corr neg with fol, not B12. Hcy, age, ser creat independent risk factors for CVE. RR (95% CI): 1.06 (1.04-1.09).
"	Voutilainen et al., 2000	Coronary disease-free 734		63 mo	No Hcy data; fol assoc with acute cor events. For acute cor events RR (95% CI): 0.31 (0.11-0.90).
"	Aronow and Ahn, 2000	Coronary artery disease Men 153 (age: 81 ± 9 yr) Women 347	Combined men, women: New events Hcy 16.8 µmol/L No new events 12.8 µmol/L	31 mo	Hcy an independent predictor of new coronary events in patients with prior CAD (RR, 1.078).
"	Omland et al., 2000	Coronary artery disease Survivors 579 Deaths 65	Admission Hcy mean Hcy 12.3 µmol/L mean Hcy 14.3 µmol/L	628 days	Hcy on hospital admission an independent predictor of long-term survival in patients with acute coronary syndromes.
Prospective study - Cerebrovascular disease	Alfthan et al., 1994	Stroke cases 74 Controls 269	Hcy 10.3 µmol/L 9.6 µmol/L   NS	Follow-up ~ 9 yr	Serum Hcy not a risk factor for stroke.
"	Verhoef et al., 1994	Stroke cases 109 Controls 427	Hcy 11.1 µmol/L 10.6 µmol/L   NS	Follow-up 5 yr	Hcy not a risk factor for stroke. RR (95% CI): 0.8 (0.3-2.4)
"	Perry et al., 1995	Stroke cases 107 Controls 118	Hcy 13.7 µmol 11.9 µmol/L   p <0.01	Follow-up ~ 11 yr	Hcy is an independent risk factor for stroke. RR (95% CI): 4.7 (1.1-20.0)
Abbreviations: Hcy or tHcy, homocysteine; fol, folate; d, days; wk, weeks; mo, months; yr, years; MI, myocardial infarction; , increased; , decreased; CAD, coronary artery disease; CHD, coronary heart disease; ser, serum; RR, relative risk; CI, confidence interval; cor art surg, coronary artery surgery.

TABLE 2: B VITAMINS AND RISK OF VASCULAR DISEASE - INTERVENTION STUDIES
Type of study	Authors	Participants/Groups/Numbers	Treatments			Duration	Findings
			Folate	B12	B6
Randomized, controlled intervention trial-dietary	Chait et al., 1999	BP, dyslipidemia, diabetes 244	690µg	12.8µg	4.0mg	10 wk	Hcy
		"    "    " Controls 247	358µg	4.9µg	2.2mg	10 wk
Randomized, controlled intervention trial-dietary	Appel et al., 2000	Normal Control 39	168µg	4.7µg	1.8mg	12 wk	Hcy
		Fruits,vegs 41	314µg	4.6µg	3.1mg	12 wk
		Fruits,vegs,fat,sat fat 38	418µg	6.5µg	2.7mg	12 wk
Randomized, controlled intervention trial-dietary	Riddell et al., 2000	Normal Control	250µg			12 wk
		Diet folate	+ ~350-400 µg fol			12 wk	ser fol
		Cereal	+ ~350-400 µg fol			12 wk	ser fol, Hcy
		Supplement	+ ~450 µg fol			12 wk	ser fol, Hcy
Randomized double-blind placebo-controlled trial - supplements	den Heijer et al., 1998	Healthy volunteers Hcy <16 µmol/L
		Placebo 36 Fol,B6,B12 34 Fol 35 Fol 36 B12 36	- 5mg 5mg 0.5mg -	- 0.4 mg- - 0.4 mg	- 50 mg - - -	8 wk 8 wk 8 wk 8 wk 8 wk	Hcy nc Hcy sig (p<0.001) Hcy sig Hcy sig Hcy nc (p=0.14)
		Healthy volunteers Hcy >16 µmol/L
		Placebo 27 Fol,B6,B12 23	- 5	- 0.4mg	- 50mg	8 wk 8 wk	Hcy nc Hcy sig (p<0.001)
Randomized double-blind placebo-controlled trial - supplements	Bronstrup et al., 1998	Female volunteers (total,150)
		Placebo Fol 51 Fol + B12 49 Fol + B12 50	- 400µg 400µg 400µg	- - 6µg 400µg	- - - -	4 wk 4 wk 4 wk 4 wk	Hcy nc ser fol Hcy ser fol Hcy ser fol Hcy
Abbreviations: Fol, folate; wk, weeks; , increased; BP, blood pressure; , decreased; Hcy, homocysteine; vegs, vegetables; sat, saturated; ser, serum; nc, no change.

Letter Regarding Dietary Supplement Health Claim for Folic Acid with Respect to Neural Tube Defects October 10, 2000

Letter Regarding Dietary Supplement Health Claim for Fiber with Respect to Colorectal Cancer October 10, 2000

Letter Regarding Dietary Supplement Health Claim for Omega-3 Fatty Acids and Coronary Heart Disease October 31, 2000

Guidance for Industry: Significant Scientific Agreement in the Review of Health Claims for Conventional Foods and Dietary Supplements (December 1999)

This document was issued on November 28, 2000.
For more recent information on Dietary Supplements
See http://www.cfsan.fda.gov/~dms/supplmnt.html

Hypertext updated by cjm/kwg 2001-APR-02