Recruitment procedure and flow of participants through the study. MMA indicates methylmalonic acid.
Proportional effects of different doses of cyanocobalamin on mean methylmalonic acid (MMA) (A), total homocysteine (tHcy) (B), holotranscobalamin (holoTC) (C), and vitamin B12 (D) concentrations after 16 weeks of supplementation. Error bars represent SD.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Eussen SJPM, de Groot LCPGM, Clarke R, et al. Oral Cyanocobalamin Supplementation in Older People With Vitamin B12 Deficiency: A Dose-Finding Trial. Arch Intern Med. 2005;165(10):1167–1172. doi:10.1001/archinte.165.10.1167
Supplementation with high doses of oral cobalamin is as effective as cobalamin administered by intramuscular injection to correct plasma markers of vitamin B12 deficiency, but the effects of lower oral doses of cobalamin on such markers are uncertain.
We conducted a randomized, parallel-group, double-blind, dose-finding trial to determine the lowest oral dose of cyanocobalamin required to normalize biochemical markers of vitamin B12 deficiency in older people with mild vitamin B12 deficiency, defined as a serum vitamin B12 level of 100 to 300 pmol/L (135-406 pg/mL) and a methylmalonic acid level of 0.26 μmol/L or greater. We assessed the effects of daily oral doses of 2.5, 100, 250, 500, and 1000 μg of cyanocobalamin administered for 16 weeks on biochemical markers of vitamin B12 deficiency in 120 people. The main outcome measure was the dose of oral cyanocobalamin that produced 80% to 90% of the estimated maximal reduction in the plasma methylmalonic acid concentration.
Supplementation with cyanocobalamin in daily oral doses of 2.5, 100, 250, 500, and 1000 μg was associated with mean reductions in plasma methylmalonic acid concentrations of 16%, 16%, 23%, 33%, and 33%, respectively. Daily doses of 647 to 1032 μg of cyanocobalamin were associated with 80% to 90% of the estimated maximum reduction in the plasma methylmalonic acid concentration.
The lowest dose of oral cyanocobalamin required to normalize mild vitamin B12 deficiency is more than 200 times greater than the recommended dietary allowance, which is approximately 3 μg daily.
Vitamin B12 deficiency, due to intrinsic factor deficiency, hypochlorhydria, or food-bound malabsorption, affects mainly older people.1-3 Symptoms of vitamin B12 deficiency include anemia, neuropathy, and neuropsychiatric disorders, but it more commonly leads to nonspecific tiredness or malaise in older people.3-5 Approximately 20% of the circulating plasma vitamin B12 is transported as holotranscobalamin (holoTC), which can be taken up by all cells, and the remaining 80% is transported as haptocorrin, which is believed not to be metabolically active.6,7 In the cell, vitamin B12 acts as a cofactor for methionine synthase, an enzyme that remethylates homocysteine (Hcy) to methionine, and for methylmalonyl–coenzyme A mutase, an enzyme that converts methylmalonyl–coenzyme A to succinyl–coenzyme A. In the setting of vitamin B12 deficiency, methylmalonyl–coenzyme A is hydrolyzed to methylmalonic acid (MMA). Thus, elevated plasma concentrations of MMA and total Hcy (tHcy) can be used as biochemical markers to aid in the diagnosis of vitamin B12 deficiency and to monitor the response to cobalamin supplementation.8,9
Active absorption of protein-bound vitamin B12 in food is impaired in individuals with vitamin B12 deficiency, but approximately 1% of orally administered crystalline cobalamin is absorbed by passive diffusion.3,10 Consequently, vitamin B12 deficiency is usually treated by monthly intramuscular injections of 1000 μg of hydroxycobalamin or cyanocobalamin. However, daily dietary supplementation with 1000 to 2000 μg of cyanocobalamin administered orally has been shown to be as effective as11 or even more effective than12 cobalamin administered by intramuscular injections to correct biochemical markers of vitamin B12 deficiency. Previoustrials13,14 that examined the effects on biochemical markers of vitamin B12 status of daily dietary oral supplements ranging from 10 to 100 μg of cobalamin did not determine the lowest effective dose required to correct vitamin B12 deficiency. A major knowledge gap concerns the lowest dose of oral cobalamin supplementation that will normalize elevated MMA concentrations. The aim of the present trial is to determine the lowest dose of cyanocobalamin that is required for a maximal reduction in MMA concentrations in a randomized, parallel-group, double-blind, controlled, dose-finding study in older people with mild vitamin B12 deficiency. The doses used cover the total spectrum from the recommended dietary allowance to the commonly used dose in cobalamin injections.
Free-living older people (aged ≥70 years) were recruited in the Wageningen area of the Netherlands through a database of individuals who had previously indicated interest in participation in such a trial. Individuals with self-reported anemia, surgery or diseases of the stomach or small intestine, or any life-threatening diseases were excluded, as were individuals who reported current use of multivitamin supplements containing folic acid, cobalamin, or pyridoxine hydrochloride and those currently receiving cobalamin injections. The concomitant use of medications known to affect vitamin B12 absorption (eg, proton pump inhibitors, H2-antagonists, and metformin) was permitted if the medication had been provided at least 3 months before enrollment and was scheduled to be continued for the duration of the trial. Individuals who fulfilled these criteria were invited to give a blood sample at a screening visit. People were eligible for the trial if their serum vitamin B12 concentration was between 100 and 300 pmol/L (135-406 pg/mL), their plasma MMA concentration was 0.26 μmol/L or greater, and their serum creatinine concentration was 120 μmol/L or less (≤1.4 mg/dL), the latter reflecting normal kidney function.3Figure 1 shows the recruitment procedure and the flow of participants through the phases of the study. The study protocol was approved by the Medical Ethical Committee of Wageningen University, and written informed consent was obtained from all participants before the screening visit.
Eligible people who agreed to be enrolled in a 3- to 4-week placebo run-in period before randomization and who had proved to be compliant (>90% intake of capsules) during the run-in period were randomized to receive 16 weeks of treatment in a parallel-group design with daily oral doses of 2.5, 100, 250, 500, or 1000 μg of cyanocobalamin (Figure 1). The doses selected for this study were based on the recommended dietary allowance of the Netherlands, which was 2.5 μg daily at the start of the trial, and 1000 μg, which served as a positive control and is administered in the form of intramuscular injections to treat vitamin B12 deficiency. The 100-, 250-, and 500-μg doses were chosen to provide an optimum dose-response curve. We did not include a placebo in the study design for ethical reasons. Randomization was based on plasma MMA concentration at the screening visit, age, and sex. We used strata to ensure a balanced distribution of participants with respect to MMA concentration (0.26−0.309, 0.31−0.359, and ≥0.36 μmol/L), age (≤75 and >75 years), and sex. All investigators and participants were masked to study treatment.
Assuming a within-person SD for MMA of 0.25 μmol/L for changes in plasma MMA concentrations induced by hydroxycobalamin supplementation,15 sample size calculations indicated that 17 participants per group provided 80% power to detect an absolute difference of 0.22 μmol/L in MMA concentrations among the treated groups. To control for an estimated dropout rate of 23%,16 at least 20 participants were to be enrolled in each group.
Cyanocobalamin was administered in capsules that were identical in appearance, smell, and taste in all treatment groups. The mean (SD) measured doses of cyanocobalamin for the capsules intended to contain 2.5, 100, 250, 500, and 1000 μg were 3 (single pooled assessment), 112 (4.7), 270 (3.4), 553 (1.7), and 860 (9.7) μg, respectively.
Participants were asked to maintain their regular diet and to avoid the use of supplements containing B vitamins during the trial. All participants were asked to complete a diary to record their daily intake of capsules, their use of nonstudy medications, and the occurrence of any new illnesses during the trial. No adverse events were reported. Compliance was checked by counting the unused capsules remaining in the capsule dispensers and by verifying pill counts in the participants’ diaries. Mean compliance was 98%, and because compliance for each participant was greater than 90%, data for all participants were included in the analyses.
A blood sample was collected at the screening and randomization visits and after 8 and 16 weeks of active treatment. Height and weight were also measured at the randomization visit. Participants were asked to be fasting at the randomization visit but were allowed to eat a light breakfast (without fruit, fruit juices, meat, or eggs) at least 1 hour before attending the screening and follow-up visits. The study was carried out between February 27, 2002, and February 28, 2003. A sample of blood for the subsequent measurement of MMA (the primary outcome measure) and tHcy and holoTC (secondary outcome measures) concentrations was collected in a 10-mL vacutainer containing EDTA. This blood sample was placed in ice water and was centrifuged at 2600 rpm for 10 minutes at 4°C within 30 minutes of collection. All plasma samples were stored at –80°C before laboratory analyses. Plasma concentrations of MMA and tHcy were determined by gas chromatography–mass spectroscopy after derivatization with methylchloroformate.17 The plasma concentration of holoTC was measured using the AXIS-Shield radioimmunoassay method.18 A blood sample was collected in a 5-mL gel tube for measurement of serum vitamin B12 (secondary outcome measure) and creatinine levels. The serum samples for vitamin B12 determination were stored at room temperature in the dark for measurement later that day using the Immulite 2000 cobalamin method (Diagnostic Products Corp, Los Angeles, Calif).19 In addition, at the randomization visit, a sample of blood was collected in 5-mL evacuated tubes containing EDTA and were stored at room temperature for measurement later that day of hematologic variables (hemoglobin level, hematocrit level, mean cell volume, and hypersegmentation of neutrophils) and plasma folate concentrations.
Baseline concentrations of the biochemical variables were calculated as the average of the measurements recorded at the screening and randomization visits for each individual. The proportional changes in plasma concentrations of MMA, tHcy, and holoTC and in serum concentrations of vitamin B12 were calculated by dividing each participant’s absolute change in concentration after 16 weeks of treatment by their concentration at baseline. The lowest dose of oral cyanocobalamin required to achieve a maximum reduction in MMA concentrations was determined using a closed test procedure.20 The Kruskal-Wallis test was used to investigate whether differences in median proportional changes were present among dose groups, and the Mann-Whitney test was used to investigate between which 2 dose groups differences in the median changes occurred. In addition, curve fitting that plots the proportional reductions in MMA concentrations against the incremental doses of cyanocobalamin administered was used to assess the dose-response relationship. The best-fit dose-response curves showed a 1-phase exponential decay estimated by the following nonlinear regression equation: Change (%) = (Top – Bottom) × Exp (–k × Cyanocobalamin Dose) + Bottom. This regression equation was used to identify the lowest oral dose of cyanocobalamin required to achieve a maximal reduction in MMA concentrations. This dose was defined as the dose that produces 80% to 90% of the maximum estimated reduction in plasma MMA concentrations. Statistical analyses were conducted using SAS statistical software (SAS Institute Inc, Cary, NC), and curve fitting was performed using GraphPad Prism software (GraphPad Software Inc, San Diego, Calif).
Selected characteristics of the study participants are given in Table 1. At baseline, the study population was, on average, not undernourished since the median body mass index (calculated as weight in kilograms divided by the square of height in meters) was 25.3.21 There were no significant differences in the mean concentrations of MMA, tHcy, holoTC, and vitamin B12 between the screening and randomization visits. The median baseline concentrations of serum vitamin B12 and plasma MMA were well matched by treatment groups, indicating that the randomization procedure had been successful. At baseline, serum vitamin B12 concentrations were correlated with plasma holoTC (ρ = 0.53; P < .001), plasma MMA (ρ = −0.34; P < .001), and tHcy (ρ = –0.25; P = .006) concentrations. Plasma holoTC concentrations were correlated with MMA (ρ = –0.41) and plasma tHcy (ρ = –0.38) concentrations (P < .001 for both), whereas plasma tHcy concentrations were correlated with MMA concentrations (ρ = 0.85; P < .001) but not with folate concentrations (ρ = –0.01; P = .91).
On average, the absolute decreases in plasma MMA and tHcy concentrations and the absolute increases in plasma vitamin B12 and holoTC concentrations increased with increasing doses of cyanocobalamin (Table 2). The reductions in MMA concentrations in all cyanocobalamin-treated groups were significant during the first 8 weeks of treatment and remained stable during the second 8 weeks of treatment. The absolute reduction in MMA concentrations of at least 0.22 μmol/L observed after 8 and 16 weeks of supplementation with 500 and 1000 μg of cyanocobalamin indicated that the study had sufficient power to detect differences among the randomly allocated doses of cyanocobalamin. In addition, the absolute effects of cyanocobalamin supplementation on MMA concentrations were assessed using the proportion of the trial population that achieved an MMA concentration below the laboratory reference interval for MMA of 0.26 μmol/L (J.S., oral communication, February 15, 2002). Daily supplementation with 2.5, 100, 250, 500, or 1000 μg of cyanocobalamin resulted in reductions in MMA concentrations to below the reference interval of 0.26 μmol/L in 21%, 38%, 52%, 62%, and 76% of the participants, respectively.
The determination of the lowest dose of cyanocobalamin associated with the maximum reductions in MMA levels or maximum increases in holoTC levels using the closed test procedure19 (which defined the optimum dose as that dose that differed significantly from the lower doses but not from the higher doses) concluded that the intended daily dose of 500 μg of cyanocobalamin was the lowest oral dose associated with a maximum reduction in MMA concentrations and a maximum increase in holoTC concentrations. The proportional reductions in MMA concentrations after daily supplementation with 2.5, 100, 250, and 500 μg of cyanocobalamin differed significantly from each other, whereas the proportional reductions in MMA concentrations did not differ significantly from each other after daily supplementation with 500 and 1000 μg of cyanocobalamin (P = .2).
The proportional decreases in MMA and tHcy levels and the proportional increases in vitamin B12 and holoTC concentrations observed with incremental doses of cyanocobalamin after 16 weeks of supplementation are shown in Figure 2. The mean reduction in plasma MMA concentrations after 16 weeks of supplementation compared with baseline varied from 16% to 33% in the groups receiving 2.5 to 1000 μg of cyanocobalamin per day. The proportional reduction in MMA after 16 weeks of supplementation was calculated using the following formula: 25.82 × Exp (–0.0018626 × Cyanocobalamin Dose) – 39.6.
The lowest daily oral dose of cyanocobalamin that resulted in 80% to 90% of the maximum reduction in MMA concentrations varied from 647 to 1032 μg. On average, such doses of cyanocobalamin reduced plasma MMA concentrations by approximately 33%.
The results of this dose-finding trial demonstrate that the lowest oral dose of cyanocobalamin associated with 80% to 90% of the estimated maximum reduction in plasma MMA concentrations in an older population with mild vitamin B12 deficiency varied from 647 to 1032 μg/d, and such doses reduce plasma MMA concentrations by approximately 33%. However, daily doses of 2.5 to 250 μgof cyanocobalamin produce statistically significant reductions in MMA concentrations of 16% to 23% in this population. The conclusions of this trial are based primarily on reductions in plasma MMA concentrations because MMA reflects tissue levels of vitamin B12.3,9,12
Comparable proportional increases in concentrations of serum vitamin B12 and plasma holoTC were observed in response to the different doses of cyanocobalamin. The dose-finding curve for holoTC demonstrated that daily oral doses of 527 to 759 μg of cyanocobalamin resulted in 80% to 90% of the estimated maximum increase in holoTC concentrations.
In contrast to the dose-finding curves for MMA and holoTC, the curve for tHcy does not show a plateau effect. This finding may be related to the selection criteria, which did not include tHcy because tHcy is not a specific marker of vitamin B12 status but is also affected by folate status and a variety of lifestyle factors.22 Most likely, tHcy concentrations in these participants are less responsive to cyanocobalamin supplementation. Therefore, we cannot assume that, based on these data, a full dose-response curve can be fitted for tHcy.
The conclusions of this trial may reflect the definition of vitamin B12 deficiency and the variable absorption of vitamin B12 in older people. The diagnosis of vitamin B12 deficiency is complicated by the limitations of current assay techniques because serum vitamin B12 concentrations alone may misclassify a significant proportion of individuals with vitamin B12 deficiency.3,9,23 Moreover, there is no consensus about the cutoff points for vitamin B12 deficiency or metabolites to define vitamin B12 deficiency. The present trial enrolled healthy older people with mild vitamin B12 deficiency, defined as serum vitamin B12 levels of 100-300 pmol/L (135 to 406 pg/mL) combined with plasma MMA levels of 0.26 μmol/L or greater in individuals without renal dysfunction. Analysis of a subgroup of participants with more severe vitamin B12 deficiency (using MMA concentrations ≥0.32 μmol/L at baseline, present in 67 participants) resulted in more pronounced changes in MMA, tHcy, holoTC, and vitamin B12 concentrations and confirmed the results of the closed test procedure (data not shown). According to the corresponding dose-finding curves for MMA and holoTC, 830 μg/d would provide 80% of the maximal reduction in MMA levels, and 449 μg/d would provide 80% of the maximal increase in holoTC levels.
Vitamin B12 can be absorbed actively, with a limited capacity of approximately 3 μg per meal in the presence of intrinsic factor and normal functioning of the stomach, pancreas, and terminal ileum. However, the bioavailability of crystalline vitamin B12 is unaffected by the underlying causes of vitamin B12 deficiency, and approximately 1% of crystalline cobalamin (typically used in oral cobalamin supplements) is absorbed by passive absorption.3 This study did not distinguish the extent to which differences in individual responses were due to active as opposed to passive absorption of vitamin B12.
The results of this trial differ from those of Seal et al,13 who compared the effects on serum vitamin B12 and tHcy concentrations of oral cyanocobalamin using daily oral doses of 10 to 50 μg or placebo for 4 weeks in 31 older people who had a pretreatment vitamin B12 concentration between 100 and 150 pmol/L (135-203 pg/mL). Seal et al13 showed that supplementation with 50 μg/d increased serum vitamin B12 levels but had no significant effects on tHcy concentrations. Rajan et al14 compared the effects of sequential daily treatment with 25, 100, and 1000 μg of cyanocobalamin for 6 weeks on serum vitamin B12 and plasma MMA concentrations in 23 older people who had a pretreatment vitamin B12 concentration less than 221 pmol/L (<299 pg/mL) in combination with an MMA concentration greater than 0.27 μmol/L. Rajan et al14 reported that daily treatment with 25 or 100 μg of cyanocobalamin lowered, but did not normalize, MMA levels and that a daily dose of 1000 μg of cyanocobalamin was required to normalize MMA concentrations.
The results of this trial indicate that the lowest dose of oral cyanocobalamin required to normalize biochemical markers of mild vitamin B12 deficiency in older people with a mild vitamin B12 deficiency is more than 200 times greater than the recommended dietary allowance for vitamin B12 of approximately 3 μg/d. Clinical trials are currently assessing the effects of high doses of oral cobalamin on markers of cognitive function and depression. If such trials can demonstrate that the reported associations of vitamin B12 deficiency with cognitive impairment or depression are causal and reversible by treatment,24 the relevance of correction of vitamin B12 deficiency in older people could be substantial. However, the present trial demonstrates that much higher doses of cyanocobalamin are required to normalize vitamin B12 deficiency than were previously believed.
Correspondence: Lisette C. P. G. M. de Groot, PhD, Division of Human Nutrition, Wageningen University, PO Box 8129, 6700 EV Wageningen, the Netherlands (Lisette.deGroot@wur.nl).
Accepted for Publication: November 22, 2004.
Financial Disclosure: None.
Funding/Support: This work was supported by grant 2100.0067 from ZonMw, the Hague, the Netherlands; grant 001-2002 from Kellogg’s Benelux, Zaventem, Belgium; and grant QLK3-CT-2002-01775 from the Foundation to Promote Research Into Functional Vitamin B12 Deficiency and the European Union BIOMED Demonstration Project.
Acknowledgment: We thank the volunteers who took part in this study; Roche Vitamins in Basel, Switzerland, for the supply of cyanocobalamin; DBF in Helmond, the Netherlands, for the production of capsules; Kathleen Emmens, Meng Jie Ji, Janet Taylor, and Jane Wintour for carrying out the holoTC assays at the Clinical Trial Service Unit; and Ove Aaeseth for carrying out the MMA and Hcy assays at the LOCUS of Homocysteine and Related Vitamins.