Objective: To investigate the effect of riboflavin supplementation on plasma homocysteine (tHcy) concentrations in healthy elderly people with sub-optimal riboflavin status.
Design: A double-blind, randomized, placebo-controlled riboflavin supplementation trial.
Setting: Community based study in Northern Ireland.
Subjects: From a screening sample of 101 healthy elderly people, 52 had sub-optimal riboflavin status (erythrocyte glutathione reductase activation coefficient, EGRAC≥1.20) and were invited to participate in the study.
Intervention: The intervention had two parts. Part 1 was a 12 week randomized double blind, placebo-controlled intervention with riboflavin (1.6 mg/day). Following completion of part 1, the placebo group went on to part 2 of the study which involved supplementation with folic acid (400 µg/day) for 6 weeks followed by folic acid and riboflavin (1.6 mg/day) for a further 12 weeks, with a 16 week washout period post-supplementation. The purpose of part 2 was: (a) to address the possibility that homocysteine-lowering in response to riboflavin may be obscured by a much greater effect of folate, and that, once folate status was optimized, a dependence of homocysteine on riboflavin might emerge; and (b) to demonstrate that these subjects had homocysteine concentrations which could be lowered by nutritional intervention.
Results: Although riboflavin supplementation significantly improved riboflavin status in both parts 1 and 2 of the study (P<0.001 for each), tHcy concentrations were unaffected (P=0.719). In contrast, folic acid supplementation (study part 2) resulted in a homocysteine lowering of 19.6% (P=0.001).
Conclusion: Despite the metabolic dependency of tHcy on riboflavin, it did not prove to be an effective homocysteine-lowering agent, even in the face of sub-optimal riboflavin status.
Mild elevations in total plasma homocysteine (tHcy) are associated with an increased risk of cerebro-, peripheral- and cardio-vascular disease (Eikelboom et al, 1999). Homocysteine is formed by the demethylation of methionine. Once formed, homocysteine is either remethylated to methionine, or undergoes a trans-sulphuration reaction to form cysteine. The remethylation of homocysteine is catalysed by the enzyme methionine synthase which requires vitamin B12, in the form of methylcobalamin (Finkelstein, 1990), and folate in the form of 5-methyltetrahydrofolate (5-MTHF; Finkelstein, 1990), as co-factor and co-substrate respectively. In the trans-sulphuration pathway, the pyridoxal 5′-phosphate (vitamin B6)-dependent enzyme, cystathionine β-synthase (CBS), catalyses the conversion of homocysteine to cysteine (Mudd et al, 1995). In addition to the roles played by vitamins B6, B12 and folate in either the trans-sulphuration or remethylation pathways, a fourth B-vitamin, riboflavin, is required for the function of both pathways. The co-substrate for methionine synthase, 5-MTHF, is generated by the action of the enzyme methylenetetrahydrofolate reductase (MTHFR) which in turn requires riboflavin, in the form of flavin adenine dinucleotide (FAD) as a prosthetic group (Bates & Fuller, 1986). In the trans-sulphuration pathway, riboflavin, in the form of flavin mononucleotide (FMN), is required for the generation of the active co-enzyme form of vitamin B6, pyridoxal 5′-phosphate (PLP; McCormick, 1989), which serves as a co-factor for CBS.
Low riboflavin status is frequently found in older populations, with 49–78% (Madigan et al, 1998; Bailey et al, 1997) of elderly subjects classified as having sub-optimal riboflavin status (erythrocyte glutathione reductase activation coefficient, EGRAC≥1.20). It is also widely recognized that plasma homocysteine (tHcy) concentrations increase with age (Selhub et al, 1993; Andersson et al, 1992). In support of the possibility that riboflavin may be an important determinant of tHcy, Lakshmi et al (1990) noted a 2–4-fold increase in the concentration of tHcy in the skin of rats in response to an experimentally induced riboflavin deficiency. Sparse and conflicting data exist, however, on the effect of riboflavin intake and status on plasma tHcy in humans. A large (n=2435) cross-sectional study from The Netherlands (De Bree et al, 2001) recently observed that dietary folate intake was the only B-vitamin independently inversely associated with plasma tHcy. Jacques et al (2001) observed only a modest (but significant) association between dietary riboflavin and tHcy in the Framingham Offspring cohort (n=1960). In contrast, Shimakawa et al (1997) reported a strong inverse relationship between dietary intake of riboflavin and plasma tHcy concentrations after multivariate adjustment (r=−0.763, P<0.01). In addition, Hustad et al (2000) found plasma riboflavin to be an independent determinant of plasma tHcy in a cross-sectional study of 423 healthy blood donors.
Riboflavin has previously been included in only two interventions in which tHcy was the primary end-point (Olszewski et al, 1989; Lakshmi & Ramalakshmi, 1998). In the first study (Olszewski et al, 1989), 12 myocardial infarction patients were treated with a combination of troxerutin, choline, vitamin B6, vitamin B12, folate and riboflavin for 21 days. A study design such as this makes it impossible to determine the effect of riboflavin alone on fasting tHcy levels. In the second study (Lakshmi & Ramalakshmi, 1998), 20 women with clinical and marked biochemical riboflavin deficiency (EGRAC=1.80) were given a pharmacological dose of riboflavin (10 mg) for 15 days. This resulted in a significant improvement in riboflavin status but there was no significant change in plasma tHcy concentration. This study, however, was not blinded and had no placebo group or washout period.
Although the effect of supplementation with other relevant B-vitamins on plasma tHcy concentration is well documented, there is clearly a lack of information regarding the effect of riboflavin. Therefore, the aim of this study was to investigate the effect of riboflavin supplementation on fasting tHcy concentrations in a group of healthy elderly people who were pre-screened in order to select individuals with sub-optimal riboflavin status.
Ethical approval was granted by the University of Ulster Research Ethical Committee and subjects gave written, informed consent. Subjects aged 60 y and over were recruited between January 1998 and April 1998 through senior citizen groups and local ‘Folds’ (which provide sheltered accommodation but no medical support for healthy elderly people who live independently within a housing complex and take care of themselves). All potential subjects were interviewed, using a short medical questionnaire, regarding general health, drug and supplement use. The exclusion criteria were: B-vitamin supplementation; gastrointestinal disease; haematological disorders; drugs known to affect vitamin B6 or riboflavin metabolism (antacids containing magnesium or ioniazid); vascular, hepatic or renal disease; impaired cognitive function (score <7 on Hodkinson 10-Point Mental State Questionnaire; Quereshi & Hodkinson, 1974); serum creatinine concentration ≥130 µmol/l; or a serum vitamin B12 concentration less than 111 pmol/l (150 ng/l).
The adequacy of sample size was established from power calculations using typical variances from other similar investigations (Madigan et al, 1998). All potential subjects were initially screened for riboflavin status. Only individuals with sub-optimal riboflavin status (erythrocyte glutathione reductase activation coefficient, EGRAC≥1.20) were invited to participate in the intervention.
Figure 1 shows the study design. A laboratory technician, who was not involved with the study in any other way, randomly assigned subjects to two groups (1 or 2). Both group 1 and group 2 participated in the first part of this intervention (weeks 0–12); only group 1 went on to participate in the second part of the intervention (weeks 12–46). Part 1 of the intervention was a randomized, double-blind, placebo-controlled trial during which group 1 took a placebo supplement (weeks 0–12), while group 2 received riboflavin (1.6 mg) supplementation daily for 12 weeks. Following completion of part 1 of the study (the placebo controlled trial), the involvement of group 2 was complete. Group 1 (placebo group from part 1 of the study) went on to part 2 of the study and received low dose folic acid (400 µg/day) for the next 6 weeks (weeks 12–18), followed by folic acid and riboflavin supplementation for the last 12 weeks (weeks 18–30). All group 1 subjects gave a washout blood sample that was taken at least 16 weeks after the intervention had ended. Thus, for part 2 of the study, each individual acted as his/her own control.
For this study, we used low doses of both riboflavin and folic acid in order to make the results more relevant for any future public health policies aimed at the prevention of hyperhomocysteinemia through modification of dietary intakes, probably via the fortification of foods. A supplementation level of 1.6 mg/day riboflavin for 12 weeks was chosen based on data from a previous riboflavin intervention (Madigan et al, 1998) carried out at this centre (but not concerned with tHcy), which demonstrated that this amount significantly improved riboflavin status in elderly people. We therefore chose the same regimen in order to observe what effect, if any, a change in riboflavin status would have on plasma tHcy. In the case of folic acid, a previous study (Ward et al, 1997) from our laboratory demonstrated that 200 µg/day for 6 weeks produced significant tHcy lowering in young adults. In the current study in elderly people we doubled this dose (400 µg/day) in order to ensure that tHcy concentrations in this group were, in the event that riboflavin supplementation was unsuccessful, indeed lowerable. In order to maximize compliance throughout both parts of the study, subjects were visited in their own homes every 3 weeks and supplied with the exact number of supplements; any unused tablets from the previous 3 weeks were returned at this stage and counted. Subjects were instructed to follow their usual diet and refrain from commencing any form of vitamin supplementation throughout the intervention and until after the washout blood sample.
Group 1 subjects provided six 20 ml fasting blood samples in total: a screening sample, then samples at baseline (week 0), week 12; week 18; week 30; and after a 16 week washout (ie week 46). Group 2 subjects gave three 20 ml fasting blood samples in total: a screening sample, then samples at baseline (week 0) and week 12. All blood samples were collected after an overnight fast, in the subjects' own home with subjects in the sitting position at the time of collection. After collection and processing, samples were stored at −70°C for batch analysis.
Total plasma homocysteine was measured by immunoassay (Nexo et al, 2000); red cell folate (Molloy & Scott, 1997), serum folate (Molloy & Scott, 1997) and serum vitamin B12 (Kelleher & O'Brien, 1991) were measured by microbiological assay. Measurement of EASTAC (Mount et al, 1987) and EGRAC (Powers et al, 1983) were by enzyme assay on the Cobas Fara centrifugal analyzer (Roche Diagnostics, Welwyn Garden City, UK). For all assays, samples were analysed blind, in duplicate (except for EGRAC where triplicate samples were measured), and within 6 months after sampling. Average values of duplicate or triplicate measurements are reported. Samples were re-analysed if the CV between duplicate or triplicate measurements was greater than 10%. Quality control was provided by repeated analysis of stored (−70°C) batches of pooled erythrocytes (for EASTAC and EGRAC), plasma (tHcy), serum (vitamin B12, folate) or red cell folate lysates (red cell folate) covering a wide range of values. MTHFR genotype was identified by polymerase chain reaction (PCR) amplification followed by HinF1 restriction digestion (Goyette et al, 1994).
All statistical analysis was performed using the Statistics Package for the Social Sciences (SPSS) computer software package (Chertsey, UK). Groups 1 and 2 were compared at baseline using independent samples t-test. The response of subjects in groups 1 and 2 to the placebo-controlled stage of the intervention was examined using two-way ANOVA. Response of group 1 to riboflavin supplementation when given after repletion with folic acid was examined using repeated measures ANOVA with LSD; P-values <0.05 were considered significant.
A total of 101 healthy elderly people were initially recruited and screened for riboflavin status. Some 52 (51.5%) subjects were found to have sub-optimal riboflavin status (EGRAC≥1.20) and were invited to participate in the intervention, of which 48 agreed to participate and 45 volunteers (mean age 68.4 y) completed the study. By week 12, one individual (group 1) had died and one individual (group 2) had dropped out because of illness; by week 30 a further subject (group 1) withdrew because of the long duration of the study. Throughout the study, the compliance of all subjects was excellent (ie only one subject missed more than two tablets during the intervention period, that subject was included in the analysis). The general characteristics of groups 1 and 2 are shown in Table 1. There were no significant differences in any of the variables examined between groups 1 and 2 at baseline.
Baseline plasma tHcy concentrations ranged from 5.36–33.07 µmol/l with 42 of 46 participants classified as normohomocysteinaemic at baseline (ie plasma tHcy≤15 µmol/l; Kang et al, 1992). Serum folate was significantly correlated with plasma tHcy at baseline (r=−0.376, P=0.010), but no significant correlation with tHcy was found for red cell folate, serum vitamin B12, EASTAC or EGRAC (data not shown). Of the 45 subjects who completed all stages of the intervention study, n=5 (11.1%), n=21 (46.7%) and n=19 (42.2%) were found to be homozygous, heterozygous and wild-type, respectively, for the C→T677 (thermolabile) variant of MTHFR.
Table 2 shows the response of plasma tHcy and B-vitamin status to 12 weeks placebo (group 1) or riboflavin (1.6 mg/day) supplementation (group 2; ie part 1 of the study). As expected, neither the riboflavin status nor plasma tHcy concentrations of the placebo group (group 1) changed during the 12 week period. In response to 1.6 mg riboflavin daily for 12 weeks riboflavin status improved significantly in group 2, but no significant change in fasting tHcy was observed. When the homocysteine response to riboflavin supplementation was examined in the upper half of baseline tHcy concentrations (data not shown), no significant (P=0.318) homocysteine-lowering effect was found (placebo group (group 1), n=12: pre-tHcy (µmol/l)=13.44±1.59, post-tHcy (µmol/l)=13.33±3.51; riboflavin supplementation group (group 2), n=11: pre-tHcy (µmol/l)=15.40±6.31, post-tHcy (µmol/l)=16.22±3.51).
In part 2 of the intervention, group 1 was followed (weeks 12–46) to investigate the effect of riboflavin supplementation on plasma tHcy when given after optimization of folate status. After taking a placebo supplement daily during the first part of the intervention (weeks 0–2), group 1 took folic acid (400 µg/day) for 6 weeks (weeks 12–18) which resulted in a significant homocysteine-lowering of 19.6%, from 11.42±3.46 to 9.18±1.96 µmol/l, as shown in Table 3. In response to riboflavin (1.6 mg/day) in combination with folic acid (400 µg/day) for the next 12 weeks of the study (weeks 18–30), the results showed that, although riboflavin supplementation successfully improved riboflavin status, it failed to lower plasma tHcy (Table 3). Although the folate status of group 1 continued to increase significantly during the riboflavin intervention (owing to continued supplementation with folic acid during weeks 18–30), this did not affect plasma tHcy concentrations which had reached a plateau by week 18, before riboflavin supplementation commenced. Following a 16 week washout, tHcy, EGRAC and red cell folate concentrations had all returned to baseline.
For inclusion in this intervention we used an EGRAC≥1.20 as a cut-off to define sub-optimal riboflavin status in our screening sample. Even after applying stricter cut-off values for riboflavin status of either EGRAC≥1.25 or EGRAC≥1.30, the results showed no change in plasma tHcy in response to riboflavin supplementation (EGRAC≥1.25, n=24; tHcy before=11.00±5.20 µmol/l, tHcy after=11.12±5.94 µmol/l; EGRAC≥1.30, n=13: tHcy before=9.87±2.31 µmol/l, tHcy after=9.70±2.32 µmol/l; for the purpose of this additional analysis individuals who received riboflavin either before (group 1) or after (group 2) folic acid were grouped together).
This study investigated the potential role of riboflavin as a homocysteine-lowering agent, which, to date, has been largely overlooked in the literature. The results of the first part of this study (ie the 12 week randomized, placebo-controlled trial) show that, despite the metabolic dependency of homocysteine on both FAD and FMN, riboflavin supplementation in highly compliant subjects did not result in any homocysteine-lowering effect. This was also found to be true when the response of tHcy to riboflavin supplementation was examined in individuals who had the highest tHcy concentrations at baseline. Also, there was no significant correlation between riboflavin intake and tHcy concentrations in the baseline sample of 101 elderly people.
In the second part of this riboflavin intervention study, the placebo group (group 1) was followed for a further 6 weeks with folic acid (400 µg/day), then a combination of folic acid (400 µg/day) and riboflavin (1.6 mg/day) for a final 12 weeks. This part of the study examined the possibility that homocysteine-lowering in response to riboflavin may be obscured by a much greater effect of folate status. The hypothesis addressed in part 2 was therefore that once folate status had been optimized, a dependence of tHcy on riboflavin might emerge, similar to the dependence of tHcy on vitamin B12 that emerges following folic acid supplementation, as recently reported by our group (Quinlivan et al, 2002). The results show that, in response to folic acid supplementation, tHcy concentrations decreased significantly (by 19.6%), demonstrating that baseline tHcy concentrations were indeed lowerable by nutritional intervention in these subjects. Riboflavin supplementation, however, even when given after optimization of folate status, again proved to have no significant homocysteine-lowering effect. By 16-weeks post-supplementation, riboflavin status, folate status and tHcy concentrations had all returned to baseline values. It could be argued that as tHcy had already shown such a marked response to folic acid supplementation, a further lowering, in response to riboflavin supplementation, would have been unlikely. We have recently demonstrated (McKinley et al, 2001), however, that low-dose vitamin B6 supplementation significantly lowered plasma tHcy concentrations in healthy elderly subjects when given after optimization of folate status.
In this study, low-dose folic acid was a very effective homocysteine-lowering agent, even though all subjects were generally healthy, had normal red cell and serum folate status, and 42 out of 46 (91.3%) subjects were normo-homocysteinaemic. Previous studies have demonstrated that doses of folic acid as low as 200 µg/day appear to effectively lower homocysteine concentrations in both hyperhomocysteinaemic (Guttormsen et al, 1996) and normohomocysteinaemic subjects (Ward et al, 1997). If homocysteine is viewed as a functional indicator of folate status, then it seems probable that current definitions of ‘normal’ folate status may need to be revised as folic acid-responsive homocysteine concentrations are found in individuals who have what is currently defined as ‘normal’ folate status.
It is possible that riboflavin supplementation had no effect on tHcy because riboflavin status at baseline was not low enough to adversely affect the activity of MTHFR, the FAD-dependent enzyme involved in homocysteine remethylation. As demonstrated by Bates and Fuller (1986), MTHFR activity in rats was not adversely affected at EGRAC values of 1.27 and 1.31, but was affected when much higher EGRAC values of 1.81 and above (indicative of more severe deficiency) were attained. Therefore, we looked at the homocysteine response to riboflavin supplementation in a sub-sample of individuals after applying a stricter cut-off for low riboflavin status (EGRAC≥1.25 and EGRAC≥1.30), but still observed no homocysteine-lowering effect. The effect of riboflavin supplementation on tHcy in individuals with severe riboflavin deficiency could not be examined in the current study since no subject had an EGRAC value greater than 1.40.
Several years ago a C→T677 polymorphism in the MTHFR gene was discovered (Frosst et al, 1995). This polymorphism, which is commonly referred to as thermolabile MTHFR, is autosomal recessive in nature, exhibits a lower specific activity in lymphocytes (∼30% of wild-type) and decreased activity after in vitro heating to 46°C in individuals possessing the TT (homozygous) genotype (Frosst et al, 1995). The in vitro work of Guenther et al (1999) found that the mutant enzyme was approximately 10 times more likely than the wild-type enzyme to dissociate from its FAD prosthetic group and become inactivated. This in vitro work, therefore, suggests that MTHFR activity in individuals with the TT genotype may be particularly sensitive to riboflavin status. Indeed, data from 2 recent cross-sectional studies (Hustad et al, 2000; McNulty et al, 2001) indicate that riboflavin is a determinant of tHcy, but the effect is driven by subjects who are homozygous for the thermolabile variant of MTHFR. This raises the possibility that riboflavin supplementation may only effectively lower tHcy when given to individuals with the TT genotype. Only five such subjects were found in the current sample. These TT subjects were randomized to different treatment groups at the start of the study, making it impossible, therefore, to examine the effect of riboflavin supplementation on tHcy in different genotype groups for thermolabile MTHFR. Placebo-controlled studies are, therefore, necessary to investigate the effect of riboflavin supplementation on plasma tHcy in individuals who are homozygous for the thermolabile (C→T677) variant of MTHFR.
In conclusion, despite the metabolic basis for this study, riboflavin did not prove to be an effective homocysteine-lowering agent in this group of elderly people who had sub-optimal riboflavin status at baseline. Further investigation is required, however, to determine its effect in different population sub-groups, in individuals with severe riboflavin deficiency, and, importantly, in individuals who are homozygous for thermolabile MTHFR (5–18% of healthy populations; Schneider et al, 1998).
We wish to acknowledge Clonmel Healthcare, Tipperary, Ireland for providing us with the folic acid supplements, and the volunteers who kindly participated in the study. This study was supported by EU Project BMH 4983549 and Abbott Germany.
About this article
Effects of short-term folic acid and/or riboflavin supplementation on serum folate and plasma total homocysteine concentrations in young Japanese male subjects
European Journal of Clinical Nutrition (2006)