INTRODUCTION. 1 From the Division of Human Nutrition, Wageningen University, Wageningen,

Transcription

1 Effect of daily vitamin B-12 and folic acid supplementation on fracture incidence in elderly individuals with an elevated plasma homocysteine concentration: B-PROOF, a randomized controlled trial 1 5 Janneke P van Wijngaarden, Karin MA Swart, Anke W Enneman, Rosalie AM Dhonukshe-Rutten, Suzanne C van Dijk, Annelies C Ham, Elske M Brouwer-Brolsma, Nikita L van der Zwaluw, Evelien Sohl, Joyce BJ van Meurs, M Carola Zillikens, Natasja M van Schoor, Nathalie van der Velde, Johannes Brug, Andre G Uitterlinden, Paul Lips, and Lisette CPGM de Groot ABSTRACT Background: Elevated plasma homocysteine concentrations are a risk factor for osteoporotic fractures. Lowering homocysteine with combined vitamin B-12 and folic acid supplementation may reduce fracture risk. Objective: This study [B-vitamins for the PRevention Of Osteoporotic Fractures (B-PROOF)] aimed to determine whether vitamin B-12 and folic acid supplementation reduces osteoporotic fracture incidence in hyperhomocysteinemic elderly individuals. Design: This was a double-blind, randomized, placebo-controlled trial in 2919 participants aged $65 y with elevated homocysteine concentrations (12 50 ). Participants were assigned to receive daily 500 mg vitamin B-12plus400mg folic acid or placebo supplementation for 2 y. Both intervention and placebo tablets also contained 600 IU vitamin D 3. The primary endpoint was time to first osteoporotic fracture. Exploratory prespecified subgroup analyses were performed in men and women and in individuals younger than and older than age 80 y. Data were analyzed according to intention-to-treat and per-protocol principles. Results: Osteoporotic fractures occurred in 61 persons (4.2%) in the intervention group and 75 persons (5.1%) in the placebo group. Osteoporotic fracture risk was not significantly different between groups in the intention-to-treat analyses (HR: 0.84; 95% CI: 0.58, 1.21) or per-protocol analyses (HR: 0.81; 95% CI: 0.54, 1.21). For persons aged.80 y, in per-protocol analyses, osteoporotic fracture risk was lower in the intervention group than in the placebo group (HR: 0.27; 95% CI: 0.10, 0.74). The total number of adverse events (including mortality) did not differ between groups. However, 63 and 42 participants in the intervention and placebo groups, respectively, reported incident cancer (HR: 1.56; 95% CI: 1.04, 2.31). Conclusions: These data show that combined vitamin B-12 and folic acid supplementation had no effect on osteoporotic fracture incidence in this elderly population. Exploratory subgroup analyses suggest a beneficial effect on osteoporotic fracture prevention in compliant persons aged.80 y. However, treatment was also associated with increased incidence of cancer, although the study was not designed for assessing cancer outcomes. Therefore, vitamin B-12 plus folic acid supplementation cannot be recommended at present for fracture prevention in elderly people. The B-PROOF study was registered with the Netherlands Trial Register (trialregister.nl) as NTR1333 and at clinicaltrials.gov as NCT Am J Clin Nutr 2014;100: Keywords bone, folic acid, fractures, homocysteine, vitamin B-12 INTRODUCTION Osteoporosis is a chronic, multifactorial disorder, characterized by low bone mass and microarchitectural deterioration of bone tissue with fractures as a major consequence (1). Fractures lead to pain, impairment in physical and social functioning, loss of quality of life, and an increased risk of mortality (2). Because of further aging of the population, the number of fractures and their socioeconomic burden is expected to increase substantially in the coming decades (3). An elevated circulating plasma homocysteine concentration has been identified as an independent risk factor for osteoporotic fractures in observational studies (4 10), a finding that is consistent with meta-analyses (11, 12) and mechanistic studies (13, 14). Elevated homocysteine concentrations ($15) are 1 From the Division of Human Nutrition, Wageningen University, Wageningen, Netherlands (JPvW, RAMD-R, EMB-B, NLvdZ, and LCPGMdG); the Department of Epidemiology and Biostatistics and the EMGO Institute for Health and Care Research (KMAS, ES, NMvS, JB, and PL) and the Department of Internal Medicine, Endocrine Section (PL), VU University Medical Center, Amsterdam, The Netherlands; the Department of Internal Medicine, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands (AWE, SCvD, ACH, JBJvM, MCZ, NvdV, and AGU); and the Department of Internal Medicine, Section of Geriatric Medicine, Academic Medical Center, Amsterdam, The Netherlands (NvdV). 2 B-PROOF (B-vitamins for the PRevention Of Osteoporotic Fractures) is supported and funded by The Netherlands Organization for Health Research and Development (ZonMw; grant ), The Hague; an unrestricted grant from NZO (Dutch Dairy Association), Zoetermeer; The Netherlands Consortium on Healthy Aging (NCHA), Leiden/ Rotterdam; the Ministry of Economic Affairs, Agriculture, and Innovation (project KB ), The Hague; Wageningen University, Wageningen; the VU University Medical Center, Amsterdam; and Erasmus MC, Rotterdam. All organizations are based in The Netherlands. 3 Supplemental Table 1 is available from the Supplemental data link in the online posting of the article and from the same link in the online table of contents at 4 JPvW, KMAS, and AWE contributed equally to the contents of the article. 5 Address correspondence to L de Groot, Division of Human Nutrition, Wageningen University, PO Box 8129, 6700 EV Wageningen, The Netherlands. lisette.degroot@wur.nl. Received April 25, Accepted for publication September 2, First published online October 1, 2014; doi: /ajcn Am J Clin Nutr 2014;100: Printed in USA. Ó 2014 American Society for Nutrition

2 B-VITAMIN SUPPLEMENTATION AND INCIDENT FRACTURES 1579 prevalent in 30 50% of persons aged.65 y (15, 16). Treatment with vitamin B-12 and folic acid, which both play a central role in homocysteine metabolism (17), is effective in normalizing homocysteine concentrations (18, 19). Three randomized controlled trials investigated the effect of B-vitamin supplementation on fracture risk (20 22). Among stroke survivors (mean age: 71 y), a large protective effect of 2-y supplementation of 1.5 mg vitamin B-12 and 5 mg folic acid was observed on hip fracture risk in the trial by Sato et al. (21). However, in the Heart Outcomes Prevention Evaluation-2 (HOPE-2) 6 trial, no effect of 5-y supplementation of 1 mg vitamin B-12, 2.5 mg folic acid, and 50 mg vitamin B-6 was observed on fracture incidence in persons with high cardiovascular disease risk (mean age: 69 y) (22). In the VITAmins TO Prevent Stroke (VITATOPS) study, there was also no effect of treatment with 2 mg folic acid, 25 mg vitamin B-6, and 500 mg vitamin B-12 during a mean of 2.8 y on osteoporotic fracture incidence observed in patients with cerebrovascular disease (mean age: 63 y) (20). Given the conflicting results and low generalizability to the general older population, further investigation is needed. We conducted the B-vitamins for the PRevention Of Osteoporotic Fractures (B-PROOF) study to assess the efficacy [through intention-to-treat (ITT) analysis] and effectiveness [through perprotocol (PP) analysis] of 2-y oral supplementation with 500 mg vitamin B-12 and 400 mg folic acid in the prevention of osteoporotic fractures in Dutch elderly people with elevated plasma homocysteine concentrations. SUBJECTS AND METHODS Study design B-PROOF is a randomized, placebo-controlled, double-blind multicenter trial for which the design and methods were described in detail previously (23). The B-PROOF study is registered with the Netherlands Trial Register as NTR1333 and with clinicaltrials.gov as NCT Setting and participants The included participants had to be living independently or with assistance but not residing in a nursing home, aged $65 y with elevated homocysteine concentrations (12 50 ). Exclusion criteria were a serum creatinine concentration $150, cancer diagnosis in the past 5 y, and severe immobility (being bedridden or using a wheelchair permanently). In total, 2919 participants were included. Randomization and intervention Participants were randomly assigned in a 1:1 ratio in blocks of 4 to receive daily either an oral tablet containing 500 mg vitamin B-12 plus 400 mg folic acid or a placebo tablet. Permutation within blocks did not occur. Tablets in both treatment arms contained 600 IU vitamin D 3 to ensure normal vitamin D status (24). The intervention and placebo tablets were indistinguishable 6 Abbreviations used: B-PROOF, B-vitamins for the PRevention Of Osteoporotic Fractures; HOPE-2, Heart Outcomes Prevention Evaluation- 2; MMA, methylmalonic acid; ITT, intention to treat; PP, per protocol; VITATOPS, VITAmins TO Prevent Stroke. in taste, smell, and appearance. Randomization was stratified for study center, sex, age (65 80 y,.80 y), and homocysteine concentration (12 18,.18 ). The individual participation period comprised 2 y. As planned at the start of the B-PROOF study (23), to increase power, individuals who finished their participation.1 y before the close of the trial (n = 678) were invited to extend their participation and take the intervention or placebo tablets for 1 more year. In total, 393 participants agreed and extended their participation. Participants with extended follow-up (n = 393) were significantly older (mean 6 SD: compared with y; P, 0.001) and had higher serum methylmalonic acid [MMA; median (IQR): 0.23 ( ) compared with 0.22 ( ) ; P = 0.021] than did participants without extended follow-up (n = 2526). There were no differences with regard to sex and concentrations of homocysteine, vitamin B-12, folate, and holotranscobalamin. The Medical Ethics Committee of Wageningen University approved the study protocol, and the Medical Ethics Committees of Erasmus MC and VU University Medical Center gave approval for local feasibility. All participants provided written informed consent. Outcomes and follow-up At baseline and the 2-y follow-up, a broad set of measurements was performed. In the current article, the primary outcome of the study is reported (i.e., time to first osteoporotic fracture) as well as the secondary outcome (time to first occurrence of any fracture) and adverse events. Fracture assessment Fractures were reported by the participants on a study calendar that was returned every 3 mo during the intervention period. In addition, participants were asked about the occurrence of fractures at the follow-up measurement with the use of a structured questionnaire. All of the reported fractures were verified with the participants general practitioner or hospital. Fractures were classified as osteoporotic or nonosteoporotic. Osteoporotic fractures were defined as all fractures except for head, hand, finger, foot or toe fractures; fractures caused by traffic accidents; and fractures caused by cancer (25). Subjects who dropped out of the study or who were unable to complete the follow-up measurements were contacted near the end of the follow-up period to obtain information on incident fractures. In case this was not successful, a participant was regarded as lost to follow-up, and the date of last contact was recorded. Date, type, and cause of fracture were recorded and verified. Compliance Every 6 mo, new tablets were sent to the participants, and they were requested to return any remaining tablets. Participants were defined as compliant when at least 80% of the tablets had been taken during the intervention period, as indicated by the returned tablets. Adverse events Adverse events (i.e., ill health related conditions) were recorded by participants on the study calendars. In addition, all events reported to the study team by phone or otherwise were

3 1580 VAN WIJNGAARDEN ET AL. recorded. If during the study participants reported a cancer diagnosis of any type, except for nonmelanoma skin cancer, they were excluded from further tablet use. In addition, at the end of the intervention period, participants were asked whether they had been diagnosed with cancer during the trial. Reported cases of cancer, except for nonmelanoma skin cancer, were verified with the participant s general practitioner or hospital. If a participant died during the study period, this was reported by relatives. Baseline characteristics Height and weight were measured, and information on demographic factors, lifestyle characteristics, medication use, and medical history were obtained by using a questionnaire. Antiosteoporotic medication use included use of bisphosphonates, strontium-ranelate, selective estrogen receptor modulators, estrogens, androgens, denosumab, or teriparatide at baseline. Plasma homocysteine and serum creatinine (23) and serum 25-hydroxyvitamin D (26) were measured. For homocysteine, follow-up concentrations were measured as well. In addition, baseline serum vitamin B-12 and folate were determined by using an immunoelectrochemiluminescence assay (Elecsys 2010; Roche). Serum holotranscobalamin was determined by the AxSYM analyzer (Abbott Diagnostics), and serum MMA was measured by liquid chromatography tandem mass spectrometry (27). DNA was isolated from buffy coats for genotyping. Methylenetetrahydrofolate reductase (MTHFR) polymorphisms were determined by using Illumina Omni-express array (Illumina, Inc). Statistical analyses Fracture rate was estimated to be 5 6% in untreated elderly people (either living independently or institutionalized). With an expected fracture rate reduction of 34% in the intervention group, a power of 80%, and a significance level of 0.05 (onesided), 1500 participants per treatment group were required (23). Statistical analyses were prepared in advance, when the study was still blinded. Analyses were executed after partial unblinding: group allocation (A/B) was revealed without the researchers knowing which group was the treatment group. Treatment allocation was completely unblinded after completion of the analyses. Baseline characteristics of the treatment groups were compared with chi-square tests for categorical data and unpaired Student s t tests for continuous data. Nonparametric tests were applied if the distribution was skewed. Differences between the treatment groups in changes in homocysteine after 2 y were tested with an unpaired Student s t test. To assess the efficacy of the supplementation, the primary analysis was based on the ITT principle, including all subjects who agreed to start the treatment and who completed the baseline measurements. To assess the effectiveness of the supplementation, prespecified PP analyses were performed that included data only from subjects who were compliant to the study protocol. For dropouts, the time until dropout was used in these analyses. Time-to-event data were analyzed by using the Kaplan-Meier approach and the log-rank test. HRs and 95% CIs were calculated with the use of Cox proportional hazards models. Models were unadjusted; adjusted for age, sex, and study center; and additionally adjusted for baseline homocysteine and holotranscobalamin. Individual time of follow-up was calculated as the time until the first fracture (primary outcome, osteoporotic; secondary outcome, any type), end of the intervention period, date of loss to follow-up, or death, whichever came first. Log-minus-log plots were used to check the proportional hazard models assumption, which was not violated. The difference in number of persons who reported at least one adverse event between treatment groups was tested by using chisquare analysis. For time to cancer, post hoc analyses were performed following the same approach as the fracture analyses. In addition, sensitivity analyses were conducted including cancer cases that could not be fully verified, and this was repeated after excluding relapse cancer cases (relapse of a cancer that was diagnosed.5 y before inclusion in the study). For all of the outcomes, interaction with treatment was tested for the prespecified covariates sex, baseline age #80 and.80 y, plasma homocysteine #18 and.18, and MTHFR genotype. In case of a significant interaction (P, 0.1), subgroup analyses were performed. Significance was set at a = Data were analyzed by using IBM SPSS Statistics 20 (SPSS, Inc). RESULTS Characteristics of the participants Analyses included 2919 participants [49.9% men; mean 6 SD age: y; median (IQR) plasma homocysteine concentration: 14.4 ( ) ] (Figure 1). At baseline, no significant differences between the intervention (n = 1461) and placebo (n = 1458) groups were observed, except for a 3% higher holotranscobalamin concentration in the intervention group (P = 0.028) (Table 1). The mean change in homocysteine concentration was 24.4 in the intervention group compared with 20.2 in the placebo group (P, 0.001) (Table 2). Primary endpoint: osteoporotic fracture In the ITT analyses, 52 persons sustained a total of 61 osteoporotic fractures (fracture rate = 2.0/100 person-years) in the intervention group compared with 61 persons with 75 osteoporotic fractures (fracture rate = 2.5/100 person-years) in the placebo group (incidence rate ratio: 0.80; 95% CI: 0.55, 1.16). Two fractures could not be verified and were considered noncases. Time to first osteoporotic fracture was not significantly different between the intervention and placebo groups (log-rank P = 0.396). Cox proportional hazards models adjusted for age, sex, and study center showed that persons in the intervention group did not have a significantly lower probability of sustaining an osteoporotic fracture than did persons in the placebo group (HR: 0.84; 95% CI: 0.58, 1.21) (Figure 2A, Table 3). Results were similar after additional adjustment for baseline homocysteine and holotranscobalamin (data not shown). PP analysis was conducted in 2661 compliant participants, including 91.4% of participants in the intervention group and 90.9% in the placebo group. Fracture rates are presented in Table 3. Multivariable Cox proportional hazards models did not show a significantly different osteoporotic fracture risk between the

4 B-VITAMIN SUPPLEMENTATION AND INCIDENT FRACTURES 1581 FIGURE 1 Screening, randomization, and follow-up in the B-vitamins for the PRevention Of Osteoporotic Fractures (B-PROOF) study. Note that the number of participants who dropped out because of death does not equal the total number of deceased participants; some participants (n = 37) dropped out for other reasons and died after they dropped out of the study (intention-to-treat). intervention group and placebo group (HR: 0.81; 95% CI: 0.54, 1.21) (Table 3). Secondary endpoint: any type of fracture The rate of fractures of any type in the intervention group compared with the placebo group was 2.6/100 person-years compared with 3.1/100 person-years, respectively. No significant effects of the intervention in both the ITT and the PP analyses were observed [HR (95% CI): 0.83 (0.59, 1.16) and 0.80 (0.56, 1.15), respectively] (Figure 2B, Table 3). Specific types of fractures are presented in Table 4. Exploratory analyses Interaction of treatment group with age, sex, homocysteine concentration, and MTHFR genotype were not significant in the ITT analyses for either osteoporotic or any type of fractures. In the PP analyses, a significant interaction effect with age was observed (P = and for osteoporotic and any type of fractures, respectively). Persons aged.80 y in the intervention group had a lower probability of sustaining an osteoporotic fracture (HR: 0.27; 95% CI: 0.10, 0.74) than did the placebo group (Table 3). The number needed to treat was 25 (for 2 y). Results were similar when any type of fractures was considered. No effect was observed in persons #80 y, and no significant interaction was observed with sex, homocysteine concentration, or MTHFR genotype. Adverse events Mortality did not differ between the intervention and placebo groups (n = 37 compared with n = 42, respectively; P = 0.571, ITT). In the total number of adverse events, no difference was observed between treatment groups (P = 0.862). However, 63 participants in the intervention group and 42 participants in the

5 1582 VAN WIJNGAARDEN ET AL. TABLE 1 Baseline characteristics of the B-PROOF study population 1 Placebo group (n = 1458) Intervention group (n = 1461) Age, y Sex, 3 % women Study center, 3 % WU VUmc EMC History of fracture, 3,4 % yes Height, cm Weight, kg Current smoker, 3 % Alcohol use, 4 % Light Moderate Excessive Physical activity, 3 min/d 131 (86 193) (81 190) Education, 3 % Low Intermediate High Vitamin B-12 and/or folic acid supplement use, 3,4 % yes Vitamin D supplement use, 3,4 % yes Osteoporotic medication use, 3,4 % yes Biochemical analyses Homocysteine, 14.5 ( ) 14.3 ( ) Vitamin B-12, pmol/l 266 ( ) 267 ( ) Folate, nmol/l 18.9 ( ) 18.8 ( ) Methylmalonic acid, 0.23 ( ) 0.22 ( ) Holotranscobalamin, pmol/l 63.0 ( ) 65.0 ( )* 25-Hydroxyvitamin D, nmol/l Creatinine, n = *P, B-PROOF, B-vitamins for the PRevention Of Osteoporotic Fractures; EMC, Erasmus MC; VUmc, VU University Medical Center; WU, Wageningen University. 2 Mean 6 SD (all such values). Differences were tested by using a t test. 3 Differences were tested by using a chi-square test. 4 Data based on self-report. 5 Median; IQR in parentheses (all such values). Differences were tested by using the Mann-Whitney U test. placebo group reported a new, subsequently verified, diagnosis of cancer during the intervention period (chi-square P = 0.038; HR: 1.56; 95% CI: 1.04, 2.31; ITT) (Table 3, Figure 2C). PP analyses (Table 3) and sensitivity analyses, including unverified cancer diagnoses (n = 13) and excluding relapse cases (n =6) (data not shown), provided similar results. TABLE 2 Plasma homocysteine concentrations at baseline and after follow-up according to treatment group for the complete study sample and per age category 1 n Baseline homocysteine, Placebo group Follow-up homocysteine, 2-y change, n Baseline homocysteine, Intervention group Follow-up homocysteine, 2-y change, ITT Total sample ( ) ( ) ( ) 10.3 ( ) ,0.001 Age #80 y ( ) 14.0 ( ) ( ) 10.2 ( ) ,0.001 Age.80 y ( ) 16.4 ( ) ( ) 11.2 ( ) ,0.001 PP Total sample ( ) 14.3 ( ) ( ) 10.2 ( ) ,0.001 Age #80 y ( ) 14.0 ( ) ( ) 10.2 ( ) ,0.001 Age.80 y ( ) 16.2 ( ) ( ) 10.9 ( ) , ITT and PP analyses include only participants for whom follow-up homocysteine concentrations were available; numbers therefore slightly differ from the total ITT and PP samples. ITT, intention-to-treat; PP, per protocol. 2 Median; IQR in parentheses (all such values). 3 Mean 6 SD (all such values). Differences were tested by using a t test. P

6 B-VITAMIN SUPPLEMENTATION AND INCIDENT FRACTURES 1583 Interaction effects for cancer with age (P = 0.085) and sex (P = 0.090) were observed (ITT). Corresponding subgroup analyses showed that the effect was more pronounced in participants aged.80 y (HR: 3.76; 95% CI: 1.24, 11.49) and in women (HR: 2.35; 95% CI: 1.23, 4.50). Differences mainly appeared for colorectal cancer (14 in the intervention group compared with 5 in the placebo group) and other gastrointestinal cancers (7 compared with 1 in the intervention and placebo groups, respectively) (Supplemental Table 1). No interaction effects for baseline homocysteine and MTHFR genotype were observed. FIGURE 2 Time to first osteoporotic fracture (A), first fracture of any type (B), and cancer diagnosis (C), adjusted for age, sex, and study center, derived from Cox proportional hazards analysis (intention-to-treat). Note that the participation period of 2 y was extended for 1 y for a subgroup of 393 participants to increase power. DISCUSSION Daily supplementation of 500 mg vitamin B-12 and 400 mg folic acid in addition to 600 IU vitamin D 3 for 2 y did not significantly reduce osteoporotic fracture risk in elderly individuals aged $65 y with an elevated plasma homocysteine concentration. Prespecified subgroup analysis suggested a reduction in fractures among those aged.80 y who were compliant in taking the supplement. Mortality did not differ between treatment groups over the 2-y intervention period. However, supplementation was associated with a higher cancer incidence, especially colorectal and other gastrointestinal cancers. The subgroup analysis in persons aged.80 y who were compliant and the analyses on mortality and cancer should be considered as exploratory analyses because the study was powered only to detect differences in fracture risk. Compared with the trial by Sato et al. (21) (a strong treatment effect on fractures) and the HOPE-2 (22) and VITATOPS (20) trials (no treatment effect), differences in study design and population with B-PROOF should be noted. With regard to baseline health status, the Sato et al. trial included a very specific, high-fracture-risk population consisting of postischemic stroke, hemiplegic patients. HOPE-2 and VITATOPS included participants with a history of vascular or cerebrovascular disease, whereas B-PROOF included participants primarily on the basis of elevated homocysteine concentrations. In addition, median homocysteine concentrations differed substantially between the studies: 19.9 in the Sato et al. trial, 11.5 in the HOPE-2 trial, 14.3 in VITATOPS, and 14.4 in B-PROOF. Mean age slightly differed between the studies: 71, 69, 63, and 74 y, respectively. In addition, dietary patterns, the presence of fortified food, and/or supplement use might contribute to differences between the populations. Although the sex distribution was similar for the Sato et al. trial and B-PROOF (53% compared with 50% women), fewer women participated in HOPE-2 (28%) and VITATOPS (36%). However, we did not find evidence for an interaction between sex and treatment in our study. Comparison of dose and duration across the 3 trials indicates that higher supplementation dose and longer study duration did not result in more favorable outcomes and so do not explain the differences in results. Finally, in the Sato et al. trial, homocysteine concentrations were higher and the general health status of the participants was worse, which resulted in a higher a priori fracture risk than for participants in HOPE-2, VITA- TOPS, and B-PROOF. In the current study, a significant effect of B-vitamins on fracture risk was observed only in compliant persons aged.80 y. It is known that homocysteine concentrations increase with age, and therefore baseline homocysteine concentrations or changes

7 1584 VAN WIJNGAARDEN ET AL. TABLE 3 Fracture and cancer rates and HRs as measures of the association of treatment with time to first osteoporotic fracture, time to first fracture of any type, and time to cancer diagnosis in the total group and subgroups 1 Placebo group Intervention group Unadjusted model Adjusted model Outcome n Rate/100 person-years n Rate/100 person-years HR (95% CI) P HR (95% CI) P ITT analysis Osteoporotic fracture (n = 2919) (0.59, 1.23) (0.58, 1.21) Any type of fracture (n = 2919) (0.60, 1.17) (0.59, 1.16) Cancer (n = 2906) (1.05, 2.31) (1.04, 2.31) Age #80 y (n = 2416) (0.86, 2.04) (0.86, 2.04) Age.80 y (n = 490) (1.21, 11.11) (1.24, 11.49) Men (n = 1450) (0.70, 1.96) (0.69, 1.94) Women (n = 1456) (1.23, 4.50) (1.23, 4.50) PP analysis Osteoporotic fracture (n = 2661) (0.55, 1.22) (0.54, 1.21) Age #80 y (n = 2263) (0.68, 1.68) (0.68, 1.67) Age.80 y (n = 398) (0.11, 0.82) (0.10, 0.74) Any type of fracture (n = 2661) (0.57, 1.16) (0.56, 1.15) Age #80 y (n = 2263) (0.68, 1.51) (0.67, 1.49) Age.80 y (n = 398) (0.11, 0.77) (0.10, 0.70) Cancer (n = 2651) (1.02, 2.70) (1.02, 2.70) Stratified analyses were performed if the interaction of sex, age, homocysteine, study center, or MTHFR genotype with treatment was significant (P, 0.10). Values were derived from Cox proportional hazards, ITT, and PP analyses. ITT, intention-to-treat; PP, per protocol. 2 Adjusted for age, sex, and study center. 3 Adjusted for sex and study center. 4 Adjusted for age and study center. in homocysteine concentrations might provide a possible explanation for the results. On the one hand, we did not observe a significant interaction of baseline homocysteine concentration (#18 and.18 ) with treatment. However, homocysteine concentrations appeared to decrease more in compliant persons aged.80 y than in persons aged #80 y (Table 2; post hoc analysis), especially when taking into account the changes over time as shown in the placebo group. Because these analyses should be considered exploratory, future studies are needed to examine this further. The observation of a significantly higher cancer incidence in the intervention group than in the control group was unexpected. It is important to note that B-PROOF was not designed to study cancer as a primary outcome. The limited follow-up time of 2 y, for instance, does not allow firm conclusions about cancer development and long-term cancer risk. The results of the present trial differ from the B-Vitamin Treatment Trialists collaboration meta-analysis of 13 trials in 49,621 individuals, which reported that allocation to folic acid had no significant effects on overall cancer incidence (1904 in the folic acid group compared with TABLE 4 Total number of fractures during the intervention period per fracture type according to treatment group and age category Placebo group Intervention group Total sample (n = 1458) Age #80 y (n = 1205) Age.80 y (n = 253) Total sample (n = 1461) Age #80 y (n = 1220) Head Arm Elbow Wrist Hand Fingers Rib Vertebra Pelvis Hip Leg Ankle Foot Toe Total Age.80 y (n = 241)

8 B-VITAMIN SUPPLEMENTATION AND INCIDENT FRACTURES in the control group; RR: 1.06; 95% CI: 0.99, 1.13) or on cancer incidence at any site (28). It should be noted, however, that the results nearly reached significance. This meta-analysis, primarily involving participants at high risk of cardiovascular disease, tested the effect of a mean daily dose of folic acid of 2 mg (range: mg) for an average duration of 5.2 y (28). The findings were consistent with 2 previous meta-analyses (29, 30) [RR (95% CI): 1.05 (0.99, 1.11) and 1.07 (1.00, 1.14), respectively]. The larger effect sizes observed in B-PROOF may reflect the effects of chance because they were based on only 105 incident cancer events compared with 3713 cancer events in the B-Vitamin Treatment Trialists collaboration meta-analysis (28). However, on the basis of both the previous observations and the results of B-PROOF, chance finding cannot be excluded but a true effect may also have been seen. Cancer risk after B-vitamin treatment should therefore be a focus in future studies. The dose of folic acid provided in B-PROOF (400 mg) was relatively low and well below the Tolerable Upper Intake Level for folic acid of 1 mg/d in Europe (31). In addition, no national mandatory folic acid food fortification exists in The Netherlands. Subgroup analysis in 2 of the 3 meta-analyses on this potential side effect showed that increased risk of cancer was mainly seen in low-dose (#1 mg/d) supplementation rather than in high-dose supplementation, whereas a dose-response effect was absent (29, 30). However, it should be noted that the lowdose trials all had doses higher than ours (ranging from 0.5 to 1.0 mg); in addition, dose-related effects were absent in the third meta-analysis (28). With regard to vitamin B-12 and vitamin D 3, to date little is known about the possible relation between vitamin B-12 and cancer risk or the interaction between folic acid, vitamin B-12 and/or vitamin D 3 and cancer risk. Folate is required for DNA synthesis and DNA methylation, processes that are also important in cancer initiation and progression. It has been hypothesized that folic acid is preventive in the initiation of cancer, whereas it enhances growth and progression of established neoplastic cells (32). As shown in Figure 2C, the curves for cancer incidence diverged shortly after the start of the intervention. This fits with the hypothesis of an effect on cancer progression, rather than cancer induction. This idea is supported by the observation that the effect on cancer incidence in our study was most pronounced in persons aged.80 y, among whom the presence of latent cancer is speculated to be more likely. The fact that our study population was older (mean age: 74 y) than the populations in the 3 meta-analyses [mean ages ranging from 26 to 69 y (28 30)] may therefore also explain the higher overall HR observed in our study. Further research into the effects of folic acid on cancer progression is warranted, especially in the oldest-old individuals. The major strengths of B-PROOF are its double-blind, randomized, placebo-controlled design and the use of clinical endpoints. It is the first trial, to our knowledge, primarily designed to study the effect of B-vitamin supplementation on fracture prevention in an elderly population with mildly elevated homocysteine concentrations. Another strength is the high compliance with the allocated treatment. A limitation of the study is that we included 2919 instead of 3000 participants as indicated by our sample size calculation. As described previously, there was a preplanned prolonged follow-up, which included, in total, 393 persons to increase power. These persons were slightly different (in terms of age and MMA status) than those without extended follow-up, with potentially a slightly higher a priori risk of fractures. However, the subgroup was too small for further subanalyses. Both the occurrence of a fracture and the diagnosis of cancer were based on self-report, which could be regarded as a limitation. However, structured questionnaires were used and the diagnoses were verified with the participant s general practitioner or hospital. Potential underreporting is expected to be nondifferential for treatment groups. In addition, it should be noted that multiple statistical tests were performed. Although they were prespecified, the occurrence of false-positive findings cannot be ruled out. In conclusion, an overall effect of supplementation of vitamin B-12 and folic acid in reducing fracture risk in elderly individuals with elevated homocysteine concentrations was not observed in B-PROOF. Exploratory analyses suggested a reduced fracture risk in elderly individuals aged.80 y who were compliant in taking the supplement. On the other hand, supplementation of vitamin B-12 and folic acid was also associated with higher cancer risk, although these results should be treated with caution because our study was not designed to study effects of the intervention on cancer. Hence, vitamin B-12 and folic acid supplementation cannot be recommended for fracture prevention. We thank the dedicated coworkers who helped this trial to succeed, especially PH in t Veld, M Hillen-Tijdink, A Nicolaas-Merkus, N Pliester, S Oliai Araghi, and S Smits. We also thank HAP Pols for obtaining funding. The members of the advisory committee, R Clarke and H van Laarhoven, are thanked for their valuable contribution in the evaluation of the cancer results. The authors responsibilities were as follows LCPGMdG: had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis; RAMD-R, MCZ, NMvS, PL, and LCPGMdG: designed the research; JPvW, KMAS, AWE, SCvD, ACH, EMB-B, NLvdZ, and ES: conducted the research; JPvW, KMAS, and AWE: performed statistical analysis and drafted the manuscript; RAMD-R, SCvD, ACH, EMB-B, NLvdZ, ES, JBJvM, MCZ, NMvS, NvdV, JB, AGU, PL, and LCPGMdG: critically revised the manuscript for important intellectual content; and AGU, PL, LCPGMdG, RAMD-R, NMvS, and NvdV: supervised the study. All of the authors analyzed and interpreted data and approved the final manuscript. PL and NMvS received an unconditional grant from Merck and Co for vitamin D assessment in the Longitudinal Aging Study Amsterdam; PL received personal fees from Merck and Co and Bristol-Myers Squibb. None of the other authors declared a conflict of interest. The sponsors did not have any role in the design or implementation of the study, data collection, data management, data analysis, or data interpretation or in the preparation, review, or approval of the manuscript. REFERENCES 1. Consensus development conference. Diagnosis, prophylaxis, and treatment of osteoporosis. Am J Med 1993;94: Lips P, van Schoor NM. Quality of life in patients with osteoporosis. Osteoporos Int 2005;16: Kanis JA, Johnell O. Requirements for DXA for the management of osteoporosis in Europe. Osteoporos Int 2005;16: Leboff MS, Narweker R, LaCroix A, Wu L, Jackson R, Lee J, Bauer DC, Cauley J, Kooperberg C, Lewis C, et al. Homocysteine levels and risk of hip fracture in postmenopausal women. J Clin Endocrinol Metab 2009;94: Gerdhem P, Ivaska KK, Isaksson A, Pettersson K, Vaananen HK, Obrant KJ, Akesson K. Associations between homocysteine, bone turnover, BMD, mortality, and fracture risk in elderly women. J Bone Miner Res 2007;22:

9 1586 VAN WIJNGAARDEN ET AL. 6. Dhonukshe-Rutten RAM, Pluijm SMF, de Groot LCPGM, Lips P, Smit JH, van Staveren WA. Homocysteine and vitamin B12 status relate to bone turnover markers, broadband ultrasound attenuation, and fractures in healthy elderly people. J Bone Miner Res 2005;20: Gjesdal CG, Vollset SE, Ueland PM, Refsum H, Meyer HE, Tell GS. Plasma homocysteine, folate, and vitamin B 12 and the risk of hip fracture: the Hordaland Homocysteine Study. J Bone Miner Res 2007; 22: van Meurs JB, Dhonukshe-Rutten RA, Pluijm SM, van der Klift M, de Jonge R, Lindemans J, de Groot LC, Hofman A, Witteman JC, van Leeuwen JP, et al. Homocysteine levels and the risk of osteoporotic fracture. N Engl J Med 2004;350: McLean RR, Jacques PF, Selhub J, Fredman L, Tucker KL, Samelson EJ, Kiel DP, Cupples LA, Hannan MT. Plasma B vitamins, homocysteine, and their relation with bone loss and hip fracture in elderly men and women. J Clin Endocrinol Metab 2008;93: McLean RR, Jacques PF, Selhub J, Tucker KL, Samelson EJ, Broe KE, Hannan MT, Cupples LA, Kiel DP. Homocysteine as a predictive factor for hip fracture in older persons. N Engl J Med 2004;350: Yang J, Hu X, Zhang Q, Cao H, Wang J, Liu B. Homocysteine level and risk of fracture: a meta-analysis and systematic review. Bone 2012; 51: van Wijngaarden JP, Doets EL, Szczecinska A, Souverein OW, Duffy ME, Dullemeijer C, Cavelaars AE, Pietruszka B, Van t Veer P, Brzozowska A, et al. Vitamin B12, folate, homocysteine, and bone health in adults and elderly people: a systematic review with meta-analyses. J Nutr Metab 2013;2013: Vaes BL, Lute C, Blom HJ, Bravenboer N, de Vries TJ, Everts V, Dhonukshe-Rutten RA, Muller M, de Groot LC, Steegenga WT. Vitamin B(12) deficiency stimulates osteoclastogenesis via increased homocysteine and methylmalonic acid. Calcif Tissue Int 2009;84: Saito M, Fujii K, Marumo K. Degree of mineralization-related collagen crosslinking in the femoral neck cancellous bone in cases of hip fracture and controls. Calcif Tissue Int 2006;79: Selhub J, Jacques PF, Rosenberg IH, Rogers G, Bowman BA, Gunter EW, Wright JD, Johnson CL. Serum total homocysteine concentrations in the third National Health and Nutrition Examination Survey ( ): population reference ranges and contribution of vitamin status to high serum concentrations. Ann Intern Med 1999;131: de Bree A, van der Put NM, Mennen LI, Verschuren WM, Blom HJ, Galan P, Bates CJ, Herrmann W, Ullrich M, Dierkes J, et al. Prevalences of hyperhomocysteinemia, unfavorable cholesterol profile and hypertension in European populations. Eur J Clin Nutr 2005;59: Carmel R, Green R, Rosenblatt DS, Watkins D. Update on cobalamin, folate, and homocysteine. Hematology (Am Soc Hematol Educ Program) 2003;1: Homocysteine Lowering Trialists Collaboration. Lowering blood homocysteine with folic acid based supplements: meta-analysis of randomised trials. BMJ 1998;316: Kuzminski AM, Del Giacco EJ, Allen RH, Stabler SP, Lindenbaum J. Effective treatment of cobalamin deficiency with oral cobalamin. Blood 1998;92: Gommans J, Yi Q, Eikelboom JW, Hankey GJ, Chen C, Rodgers H. The effect of homocysteine-lowering with B-vitamins on osteoporotic fractures in patients with cerebrovascular disease: substudy of VITA- TOPS, a randomised placebo-controlled trial. BMC Geriatr 2013;13: Sato Y, Honda Y, Iwamoto J, Kanoko T, Satoh K. Effect of folate and mecobalamin on hip fractures in patients with stroke: a randomized controlled trial. JAMA 2005;293: Sawka AM, Ray JG, Yi Q, Josse RG, Lonn E. Randomized clinical trial of homocysteine level lowering therapy and fractures. Arch Intern Med 2007;167: van Wijngaarden JP, Dhonukshe-Rutten RA, van Schoor NM, van der Velde N, Swart KM, Enneman AW, van Dijk SC, Brouwer-Brolsma EM, Zillikens MC, van Meurs JB, et al. Rationale and design of the B-PROOF study, a randomized controlled trial on the effect of supplemental intake of vitamin B12 and folic acid on fracture incidence. BMC Geriatr 2011;11: Lips P. Vitamin D deficiency and secondary hyperparathyroidism in the elderly: consequences for bone loss and fractures and therapeutic implications. Endocr Rev 2001;22: van Schoor NM, Visser M, Pluijm SM, Kuchuk N, Smit JH, Lips P. Vitamin D deficiency as a risk factor for osteoporotic fractures. Bone 2008;42: Sohl E, de Jongh RT, Heijboer AC, Swart KM, Brouwer-Brolsma EM, Enneman AW, de Groot CP, van der Velde N, Dhonukshe-Rutten RA, Lips P, et al. Vitamin D status is associated with physical performance: the results of three independent cohorts. Osteoporos Int 2013;24: Heil SG, de Jonge R, de Rotte MC, van Wijnen M, Heiner-Fokkema RM, Kobold AC, Pekelharing JM, Adriaansen HJ, Sanders E, Trienekens PH, et al. Screening for metabolic vitamin B12 deficiency by holotranscobalamin in patients suspected of vitamin B12 deficiency: a multicentre study. Ann Clin Biochem 2012;49: Vollset SE, Clarke R, Lewington S, Ebbing M, Halsey J, Lonn E, Armitage J, Manson JE, Hankey GJ, Spence JD, et al. Effects of folic acid supplementation on overall and site-specific cancer incidence during the randomised trials: meta-analyses of data on 50,000 individuals. Lancet 2013;381: Qin X, Cui Y, Shen L, Sun N, Zhang Y, Li J, Xu X, Wang B, Xu X, Huo Y, et al. Folic acid supplementation and cancer risk: a meta-analysis of randomized controlled trials. Int J Cancer 2013;133: Wien TN, Pike E, Wisloff T, Staff A, Smeland S, Klemp M. Cancer risk with folic acid supplements: a systematic review and meta-analysis. BMJ Open 2012;2:e European Food Safety Authority [Internet]. Tolerable upper intake levels for vitamins and minerals. [cited 15 July 2013]. Available from: Smith AD, Kim YI, Refsum H. Is folic acid good for everyone? Am J Clin Nutr 2008;87: