The association of homocysteine, folate, vitamin B12, and vitamin B6 with fracture incidence in older adults: a systematic review and meta-analysis.

BACKGROUND: Diverse conclusions have been drawn regarding the association of homocysteine (HCY) deficiency and supplements of B vitamins with fracture incidence in older adults. The aim of this meta-analysis was to investigate the association of HCY and B vitamins (folate, vitamin B12, and B6) with fracture incidence in older adults and whether supplements of B vitamins reduce the risk of fracture.
METHODS: The PubMed, Embase, and Cochrane library databases were systematically searched from their inception dates to 1 July 2019 to identify relevant published articles. Meta-analysis was performed to pool hazard ratios (HRs) or risk ratios (RRs) and 95% confidence intervals (CIs) using a random effects model.
RESULTS: A total of 28 studies fulfilled the inclusion criteria. High serum HCY was an independent risk factor for fractures in older persons (HR =1.25, 95% CI: 1.12 to 1.40), but only at the highest quartile level (>15 µmol/L) (HR =1.71, 95% CI: 1.37 to 2.12), rather than the second and third quartile. Multiple sensitivity and subgroup analyses supported the consistency and stability of this result. A severe deficiency of folate, instead of vitamin B12 and B6, was found to increase the risk of fracture in older adults (HR =1.46, 95% CI: 1.06 to 2.02; 1.24, 95% CI: 0.79 to 1.95; 1.36, 95% CI: 0.90 to 2.06, respectively). For the interventional effect, there was no significant association of combined folate and vitamin B12, combined folate, vitamin B12 and B6, or single vitamin B6 supplementation with the decrease of fracture risk. DISCUSSION: This meta-analysis revealed that significantly elevated serum level of HCY is positively associated with fracture incidence in older adults, yet the necessity and threshold for intervention by B vitamins require further large-scale high-quality clinical trials to validate. PROSPERO IDENTIFIER: CRD42019122586. 2021 Annals of Translational Medicine. All rights reserved.

Introduction

The increased prevalence of aging-related fractures among people worldwide leads to a higher morbidity and mortality accompanied by a heavy economic burden on public health (1-3). Multiple factors, including deterioration of bone mineral density (BMD), low body mass index, malnutrition, neurological diseases, and metabolic disorders such as diabetes and homocystinuria contribute to the elevation of fracture incidence in older adults (4-8). Some risk factors such as low bone density, calcium deficiency and vitamin D deficiency have been well investigated. However, ageing-related metabolic products, as potential risk factors and promising interventional targets for fractures in older adults, were poor investigated but has drawn ever-increasing attention in the last two decades. Among them, the role of homocysteine (HCY), one of the most investigated metabolites in geriatric cardiovascular diseases, in aging-related fractures remains unclear.

The sulfur-containing amino acid, HCY, is produced during the metabolism of methionine. Numerous studies have shown that HCY is a risk factor of aging-related diseases, such as cardio-/cerebrovascular diseases, diabetes, and neural disease (9-13). In addition, accumulated evidence indicates that high serum homocysteine (HHCY) is associated with higher incidence of skeletal deformities, including osteoporosis, which is a key risk factor for osteoporotic fracture (5,14,15). Furthermore, the intervention of HCY using B vitamins has been reported to reduce the incidence of targeted aging-related diseases (10,11,16,17). Hence, we hypothesized that the risk of osteoporotic fractures, as a major aging-related problem, can also be reduced by supplementation of B vitamins.

Folate, vitamin B12, and vitamin B6, as members of the B vitamin group, are the 3 key enzymes in the metabolism of HCY that can be conveniently supplemented in the older community via nutrient intake (18). Evidence shows that supplementation of folate, vitamin B12, and/or vitamin B6 lower the level of serum HCY which in theory can subsequently reduce osteoporotic fracture incidence (15). However, mixed conclusions have been drawn regarding the association between the supplements of B vitamins with fracture incidence in older adults (15,19-21).

There is a range of factors influencing fracture incidence, among which the compromise of BMD plays one of the most key roles in increasing fracture risk in the older population. Accumulating evidence has stated a strong association of HHCY with a decrease of BMD (22-24). Recent studies have shown that HCY-lowering intervention can also improve BMD, indicating the causation of HCY and BMD (6,23,25-27). Furthermore, 1 study (25) analyzed cross-sectional data involving 6,100 participants from 3 large Dutch studies including the cross-sectional data of the B-PROOF (B-Vitamins for the Prevention of Osteoporotic Fractures) study and 2 cohorts of the Rotterdam Study (RS-I and RS-II), and confirmed this association.

The underlying mechanism may be its effect on osteoclasts and osteoblasts, but the natural mechanism remains unclear (28). Both in vitro and in vivo studies have indicated that HHCY can disrupt the process of collagen cross-linking in bone tissue and then cause compromised bone strength (29-32). Moreover, owing to HCY’s role in cerebrovascular and neural diseases, it has emerged from increasing evidence that HCY can increase the risk of stroke and Parkinson’s disease, harming the coordination function and subsequently raising the risk of falling and fracture incidence (11,33).

The aim of this meta-analysis was to investigate the association of HCY and B vitamins (folate, vitamin B12, and B6) with fracture incidence in older men and women and whether supplementation with B vitamins can reduce the risk of fracture.

We present the following article in accordance with the PRISMA reporting checklist (available at https://dx.doi.org/10.21037/atm-21-2514).

Methods

This meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Review and Meta-analyses statement (PROSPERO identifier: CRD42019122586).

Search trials

Published articles were retrieved utilizing two methods. First, 3 scientific databases (PubMed, Embase, and the Cochrane Library) were searched to identify published articles evaluating the association of HCY, deficiency and supplements of B vitamins (folate, vitamin B12, and B6) with the incidence of fracture using a combination of keywords and MeSH terms “homocysteine”, “homocysteic acid”, “HCY”, “vitamin”, “cobalamin”, “cyanocobalamin”, “B12”, “pyridoxal“, “pyridoxine”, “pyridoxamine”, “B6”, “folate”, “folic acid”, “folacin”, “B9” AND “fracture” OR “bone”. Second, all reference lists of relevant articles (reviews, systematic reviews, meta-analyses, and included studies) were further screened manually to retrieve additional studies that were not listed in the databases. The last search was updated on 1 August 2020 with no restrictions on language, date, or journal of publication.

Inclusion and exclusion criteria

Studies were selected based on the following inclusion criteria: (I) participants, mean age of people enrolled in the trials was older than 50 years; (II) interventions, high or low level of HCY, folate, vitamin B6, or B12; (III) control group, high or low level of HCY, folate, vitamin B6, or B12; (IV) outcome, trials evaluated the relationship of HCY, folate, vitamin B6, or B12 with the risk of fracture; and (V) study design, randomized or quasi-randomized controlled clinical trials, retrospective/prospective matched cohort studies.

The exclusion criteria were as follows: (I) comments, reviews, meta-analysis, editorials, and other non-original trials; (II) congress proceedings and abstracts; (III) animal experiments; and (IV) studies providing no data to calculate evaluation indexes [hazard ratio (HR) or risk ratio (RR)]. For articles containing overlapping data, those presenting the most comprehensive data or that were published the most recently were selected. We first removed redundant and unrelated records by reading titles and abstracts. Then the full texts of remainders were downloaded to confirm their eligibility based on above criteria.

Quality assessment

The methodological qualities of eligible studies were estimated utilizing two methods based on data type. First, for the data of dichotomous variables, the Newcastle-Ottawa Scale (NOS) with 9 factors considered was used. A score of 1 was given to a study for each item. The quality scale ranged from a score of 0 to 9 and studies with high scores were considered good reports. Studies with scores ≥7 were regarded as high-quality reports.

Second, a quality assessment tool based on Cochrane risk-of-bias criteria was used to evaluate the methodological quality of the included studies involving the data of randomized controlled trials (RCTs). This tool contains 7 items used to assess bias in each trial that included the randomization sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other bias, and each paper was described as low risk, high risk, or unclear risk.

Data extraction

To reduce potential bias, 2 investigators (A & B) conducted the process of data extraction blindly and independently. Any discrepancies were resolved by consultation with a third investigator. Each included study deemed appropriate for inclusion listed the first author’s surname, publication year, journal, country of origin, participant characteristics (number, age, and gender), doses of folate, vitamin B6, or vitamin B12; baseline serum, follow-up, fracture type, statistical index, and adjustments factors of Cox-regression. If the studies had more than two groups or factorial designs and permitted multiple comparisons, the information and data that ruled out more factors was extracted. When those data were our outcomes of interest, we pooled them with the data from primary trials.

Statistical analysis

The correlation of folate, vitamin B6, vitamin B12, or HCY with fracture incidence was assessed by calculating the pooled HRs or RRs and their 95% confidence interval (CI). Based on the practice recommendation of the Cochrane Handbook, heterogeneity was assessed using the I-square (I2) statistic. Meanwhile, an I2<50% was considered not significant. When heterogeneity was significant, the potential sources of heterogeneity were identified by analyzing the methodological variability of the included studies. To reduce potential bias, a random-effects model was used for all subgroup meta-analyses. Additionally, sensitivity analysis was performed by omitting studies 1 by 1 to evaluate the impact of a single trial on the overall pooled estimate. Egger’s test was used to evaluate the possibility of publication bias. The software STATA, version 12.0 (Stata Corp, College Station, TX, USA), was applied for all statistical analyses, and P<0.05 was considered statistically significant.

Results

Selection process

The detailed article search and study selection process are listed in . A total of 23,376 articles were retrieved after the initial search of the chosen electronic databases. Of the 23,376 articles scanned, 23,260 failed the selection criteria, and the 116 remaining articles were subjected to full-text check. Among them, 89 studies were excluded because 11 articles presented no fracture data relating to serum level of HCY; 31 studies were involved in the level of HCY and osteoporosis; 10 studies were involved in the relationship of gene polymorphism and homocysteinemia; 6 were involved in the relationship of gene polymorphism and fracture; 14 were reviews; 5 were meta-analyses; 9 were irrelevant articles; 1 was a study rationale and design; 1 was a survival analysis; and 1 article was retracted (Table S1). Of the included studies, 1 was identified as originating from the reference list of the relevant studies scanned for in the databases. After rigorous selection, eventually, 28 studies involving 60,318 participants (26,508 non-interventions and 33,810 therapeutic interventions) were used for the meta-analysis.

Figure 1

Study selection flow diagram. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

Study characteristic and quality assessment

The detailed characteristics of the studies utilized in the meta-analysis are presented in . Articles were published between 2004 and 2019, with sample sizes ranging of 189–12,064 participants. All included studies were rated as high-quality via the risk of bias assessment (Figure S1).

Table 1

Characteristics of the included trials and participants

Study, Year	Country	Journal	Participants	F/M (% women)	Age mean (SD)	HCY	Folate	Vitamin	Follow-up (years)	Fracture type	HR/RR/OR
Lopez et al. 2017	Norway	JBMR	6,837	1,610/5,227 (76.5%)	62.3 (10.9)	−	+	VB12+6	Mean 10.0 years; median 3.3 [interquartile range (IQR), 2.6–3.5]; 11.1 (IQR, 9.1–12.2) for the extended follow-up	Hip	HR
Torbergsen et al. 2015	Norway	Clinical Nutrition	189	135/54 (71.4)	82.6 (8.6)	−	+	B6/B12	NR	Hip	OR
Lewerin et al. 2014	Sweden	Osteoporos Int	760	0/760 (0)	Median 75.3 [70–81]	+	−	−	Mean 5.9 (4.7–7.4)	All; vertebral F	HR
van Wijngaarden et al. 2014	Netherlands	Am J Clin Nutr	2,919	1,461/1,458 (50.1)	74.1 (6.5)	+	+	B12	> 2	All	HR
Li et al. 2014	China	J Diabetes Invest	292	191/101 (65.4)	54.3 [41–65]	+	+	B12	NR	All	OR
Urano et al. 2014	Japan	Geriatr Gerontol Int	663	663/0 (100.0)	NR	-	+	−	Mean 5.1, SD 3.4, max 13	All	HR
Gommans et al. 2013	New Zealand	BMC Geriatrics	8,164	2,944/5,218 (36.1)	62.6 (12.5)	+	−	−	Mean 3.4 (0.5–10.5)	All; hip; wrist; thoracic spine	RR
Kuroda et al. 2013	Japan	Calcif Tissue Int	1,475	1,475/0 (100.0)	66.6 (9.0)	+	−	−	≤19	Vertebral	OR
Enneman et al. 2012	Netherlands	Bone	503	503/0 (100.0)	68.5	+	−	−	7.0 (2.3)	Osteoporotic fracture	HR
Maghraoui et al. 2012	Morocco	J Clin Densitom	188	188/0 (100.0)	57.9±8.5	+	−	−	NR	Osteoporotic vertebral fracture	RR
Shiraki et al. 2011	Japan	J Bone Miner Metab	251	251/0 (100.0)	70.5±8.9	+	−	−	3.2±2.0	Osteoporotic vertebral fracture	HR
Armitage et al. 2010	UK	JAMA	12,064			+	+	Vit B12			RR
Zhu et al. 2009	Australia	Osteoporos Int	1,213	1,213/0 (100.0)	75.2±2.7	+	−	−	5	Osteoporotic fracture	HR
LeBoff et al. 2009	USA	J Clin Endocrinol Metab	800	800/0 (100.0)	70.8±6.2	+	−	−	5	Hip	OR
McLean et al. 2008	USA	J Clin Endocrinol Metab	1,002	603/399 (60.2)	75.3 (4.9)	+	+	VB12+6	14	Hip	HR
Yazdanpanah et al. 2007	Netherlands	Bone	5,304	3,140/2,164 (59.2)	67.6 (7.75)	+	+	VB12+6	7.4±3.3	Osteoporotic fracture (hip, pelvis and proximal humerus)	HR
Sawka et al. 2007	Canada	Arch Intern Med	5,522	1,559/3,963 (28.2)	68.8 (7.1)	+	−	−	5	All	HR
Périer et al. 2007	France	Osteoporos Int	671	671/0 (100.0)	62.2 (9.0)	+	−	−	9.8±1.2	All	HR
Gjesdal et al. 2007	Norway	J Bone Miner Res	4,766	2,639/2,127 (55.4)	65–67	+	+	Vit B12	13	Hip	HR
Gerdhem et al. 2007	Sweden	J Bone Miner Res	996	996/0 (100.0)	75 (0)	+	+	Vit B12	7	Osteoporotic fracture	HR
Sato et al. 2005	Japan	Bone	433	230/203(53.1)	75.4 (5.4)	+	+	Vit B12	9	Osteoporotic fracture (hip)	HR
Ravaglia et al. 2005	Italy	Journal of Gerontology	702	374/328 (53.3)	73.0 (6.0)	+	+	−	4	Osteoporotic fracture	RR
Dhonukshe-Rutten et al. 2005	Netherlands	J Bone Miner Res	1,253 (partially overlap with van Meurs’ Study)	651/602 (52.0)	76 (6.6)	+	-	Vit B12	3	Osteoporotic fracture	RR
Sato et al. 2005*	Japan	The American Journal of Medicine	199	199/0 (100.0)	71.0 (5.9)	+	−	−	4.9	Osteoporotic fracture (hip)	HR
van Meurs et al. 2004	Netherlands	N Engl J Med	2,406	1,292/1,114 (53.7)	73.9 (7.8)	+	−	−	8.1±3.7; 5.7±1.9; 2.7±0.7	Osteoporotic fracture	HR
McLean et al. 2004	USA	N Engl J Med	1,999	1,174/825 (58.7)	70.0 (7.0)	+	−	−	m 12.3, f 15	Osteoporotic fracture (hip)	HR

*, article retracted. NR, not reported; HR, hazard ratio; RR, risk ratio; OR, odds ratio; HCY, homocysteine; SD, standard deviation.

Table 2

Characteristics of the included studies involving the effectiveness of folate, vitamin B12, and/or vitamin B6 on fractures

Study, year	Country	Treatment/Placebo	Women No. (%)	Age (y)	Treatment (intake mg/d)			Follow up (years)	Plasma HCY level (μmol/L)
					Folate	Vit B12	Vit B6		Baseline	Follow-up	Change
Araghi et al. 2019	Netherlands	1,298	598 (46.1)	71.0 (68.0–76.0)	0.4	0.5	−	5–7	14.0 (12.8–15.9)	–	-
Garcia Lopez et al. 2018	Norway	1,021	370 (36.2)	57.4±9.5	1	−	−	3	9.80±2.90	8.85±2.18	−0.88±2.19
Garcia Lopez et al. 2017	Norway	1,708	398 (23.3)	62.7±11.2	0.8	0.4	40	3.2–3.4	12.17±2.78	8.84±3.22	−2.97±3.76
		1,703	390 (22.9)	62.3±10.9	0.8	0.4	−	3.2–3.4	12.06±4.69	9.13±3.71	−2.82±4.57
		1,705	401 (23.5)	62.0±10.9	−	−	40	3.2–3.4	12.30±5.60	12.25±5.04	0.20±4.65
		1,721	421 (24.5)	62.3±10.7	−	−	−	3.2–3.4	12.29±5.06	12.42±5.26	0.46±4.61
van Wijngaarden et al. 2014	Netherlands	1,461	736 (50.4)	74.0±6.6	0.4	0.5	−	2–3	14.3 (13.0–16.5)	10.3 (8.9–12.0)	−4.4±3.3
		1,458	725 (49.7)	74.2±6.4	−	−	−	2–3	14.5 (13.0–16.7)	14.3 (12.3–17.0)	−0.2±4.1
Gommans et al. 2013	New Zealand	580	–	–	2	0.5	25	0.5–10.5	14.3±8.5	10.5 4.8	–
		584	–	–	−	−	−	0.5–10.5		14.3±6.1	–
Yazdanpanah et al. 2007	Netherlands	5,304	–	67.66±7.75	+	+	+	–	–	–	–
Sawka et al. 2007	Canada	2,758	796 (28.9)	68.8±7.1	2.5	1	50	–	11.5±?	9.3±?	–
		2,764	763 (27.6)	68.9±6.8	−	−	−	–		12.3±?	–
Armitage et al. 2010	UK	6,033	2,052 (17.0)	64.2±8.9	2	1	−	84 mo	13.5±4.8	–	3.3±0.2
		6,031			−	−	−	84 mo			–

+, plus; −, minus; ? standard deviation indicates that standard deviation is unknown. SD, standard deviation; HCY, homocysteine.

Table 3

Characteristics of the included studies evaluating the association of the level of plasma folate, vitamin B12, vitamin B6 with fracture risk

Study, year	Country	Subjects	Women, n (%)	Age mean (SD)	HCY, µmol/L	Folate, nmol/L	Vit B12, pmol/L	Vit B6, nmol/L	Follow-up (years)	Fracture type
Ravaglia et al. 2005	Italy	702	374 (53.3)	73.0±6.0	>15	≤9.30	≤190	≤14	4	All
Dhonukshe-Rutten et al. 2005	Netherlands	1267	652 (51.5)	76±6.6	>16	–	≤200	–	3	All
Gerdhem et al. 2007	Sweden	996	996 (100.0)	75	14.1	–	–	–	7	All
		946	946 (100.0)	75	–	–	308	–	7	All
		978	978 (100.0)	75	–	18	–	–	7	All
Gjesdal et al. 2007	Norway	4,482	2,445 (54.6)	65–67	>15	–	–	–	12.6	Hip
		4,490	2,453 (54.6)	65–67	–	<2.9	–	–	12.6	Hip
		4,487	2,450 (54.6)	65–67	–	–	–	–	12.6	Hip
McLean et al. 2008	USA	1,002	603 (60.2)	75.3±4.9	>14	<7	<148	<20	4	Hip
Li et al. 2014	China	292	190 (65.1)	53.7–55.3 (43.9, 63.5)	Per 5 increase	Per 10 increase	Per 100 increase	–	–	Hip
Torbergsen et al. 2015	Norway	189	135 (71.4)	82.6±8.6	–	–	–	–	–	Hip

SD, standard deviation; HCY, homocysteine.

HCY and fracture risk

A total of 13 trials (5,14,34-44) evaluating the association of HCY level with fracture risk demonstrated that an increased homocysteine level was an independent risk factor for osteoporotic fractures in older persons (HR =1.25, 95% CI: 1.12 to 1.40) (). Subgroup analysis by excluding two trials without HR data and sensitivity analysis performed by omitting studies 1 by 1 demonstrated consistent results. No significant publication bias was identified using Egger’s test (P>0.05, Figure S2). Furthermore, no significant difference was observed in fracture risk between women (5,14,34,37-43) and men (5,14,34,39,44) (HR =1.20, 95% CI: 1.06 to 1.37; 1.41, 95% CI: 1.02 to 1.94) ().

Figure 2

Forest plot of the pooled results involving the association of HCY level with fracture risk. HCY, homocysteine.

Figure 3

Forest plot of the pooled results involving the association of HCY level with fracture risk of men and women. HCY, homocysteine.

HCY stratification and fracture risk

For different levels of HCY, the results of meta-analysis generated from 10 studies (5,14,34,36-40,42,43) showed that a HCY level in the highest quartile (or >15 µmol/L) was associated with an increased risk of fracture (HR =1.71, 95% CI: 1.37 to 2.12); however, the second and third quartile generated from 8 studies (5,36,37,39-43) showed no significant association (HR =0.96, 95% CI: 0.80 to 1.16; 1.14, 95% CI: 0.95 to 1.37) (). Sensitivity analyses performed by excluding 2 trials that provided insufficient data to calculate the HR and its 95% CI or 3 studies involving only participants with stroke, diabetes, and cardiovascular diseases, respectively, demonstrated consistent results.

Figure 4

Forest plot of the pooled results involving the association of HCY stratification and fracture risk. HCY, homocysteine.

HCY and fracture site

Meta-analysis results generated from 2 studies (38,39) showed that high HCY level was associated with an increased risk of total fractures (HR =1.13, 95% CI: 1.01 to 1.26) as well as nonvertebral and hip fractures (HR =1.36, 95% CI: 1.15 to 1.60; 1.58, 95% CI: 1.24 to 2.00), rather than vertebral fractures (HR =1.13, 95% CI, 0.54 to 2.38) (). Based on HCY levels, a subgroup analysis for hip fractures demonstrated that the results did not change either with overall data (HR =1.19, 95% CI: 0.96 to 1.48), or in the second, third, and highest quartile (HR =0.77, 95% CI: 0.56 to 1.06; HR =1.18, 95% CI: 0.87 to 1.61; HR =1.53, 95% CI: 1.24 to 1.89, respectively) (Figure S3).

Figure 5

Forest plot of the pooled results involving the association of HCY level and fracture risk at different sites. HCY, homocysteine.

Folate, vitamin B12, vitamin B6 and fracture risk

Our analysis revealed that 4 trials (35,39,45,46), 3 trials (34,35,39), and 2 trials (35,45) evaluating serum folate level, vitamin B12, and vitamin B6 with fracture risk had found that a severe deficiency of folate, instead of vitamin B12 and B6 increased the risk of fracture in older adults (HR =1.46, 95% CI: 1.06 to 2.02; 1.24, 95% CI: 0.79 to 1.95; 1.36, 95% CI: 0.90 to 2.06). Meanwhile, low serum folate, vitamin B12, and vitamin B6 level was not significantly associated with an increased risk of fracture (HR =0.79, 95% CI: 0.56 to 1.12; 1.01, 95% CI: 0.43 to 2.38; 1.16, 95% CI: 0.64 to 2.10) ().

Figure 6

Forest plot of the pooled results involving the association of folate, vitamin B12, vitamin B6, and fracture risk.

Folate, vitamin B12, and/or B6 supplementation and fracture risk

In 2 trials (6,21) the results of evaluating participants receiving vitamin B6 supplements vs. placebo or no treatment showed no significant association with fracture risk (HR =0.62, 95% CI: 0.28 to 1.39). As shown in , similar results generated from 2 trials (15,17) involving dual supplementation (folate and vitamin B12) and 3 trials (19-21) with triple supplementation (folate, vitamin B12, and B6) were concluded (HR =1.00, 95% CI: 0.88 to 1.14; HR =0.97, 95% CI: 0.79 to 1.19).

Figure 7

Forest plot of the pooled results involving the association of folate, vitamin B12, and/or B6 supplementation and fracture risk.

Discussion

In 2004, 2 RCTs (5,14) reported by the New England Journal of Medicine indicated that HHCY was associated with higher fracture risk in the older population. However, with the exception of a study conducted in 2005 which was retracted by the Journal of the American Medical Association (JAMA) in 2016 stating a significant effect on reducing fracture risk in the osteoporotic population by fortification with folate and vitamin B12 (47), a number of prospective randomized cohort intervention trials in the following decade failed to show HCY as an intervention target (15,19-21). The underlying reason for this mixed evidence to date awaits unravelling. Therefore, this meta-analysis was performed to investigate this inconsistence by systematically assessing the association of HCY, folate, vitamin B12, and B6 with fracture risk and whether supplements of B vitamins reduce risk of fracture in the older population.

Based on the existing evidence and results in this meta-analysis, intervention of HCY is supposed to have a beneficial effect on osteoporotic fracture in elderly people. The metabolism of HCY has been investigated by multiple studies to uncover possible prevention targets to reduce the serum level of HCY. Among them, folate, vitamin B12, and vitamin B6, the 3 critical enzymes involved in the metabolism of HCY, are considered effective and are subsequently viewed as promising intervention candidates for fracture incidence in older adults thanks to their convenient supplementation by oral intake (18). Evidence has shown that combined folate and vitamin B12 supplementation for 2–4 years could reduce HCY by 2–5 mmol/L (21); while the single application of vitamin B6 has failed to display a significant effect on reducing serum HCY level. Furthermore, the effect on reducing HCY caused by the combined fortification of folate and vitamin B12 is not changed by either the addition or absence of vitamin B6 supplementation (21). In addition, there is limited evidence showing that vitamin B2 may also play a role in the metabolism of HCY and can possibly reduce HCY with intervention (27). In this investigation, there was insufficient data concerning vitamin B2 supplementation to support our analysis of its effectiveness.

Encouragingly, a number of studies have confirmed the positive association of folate with BMD, encouraging more large-scale RCTs to validate the beneficial effect of folate fortification on BMD in older adults, although no significant association was observed between BMD and either vitamin B6 or vitamin B12 (6,23,25-27,48). However, researchers have long been perplexed by whether HCY is a culprit or a bystander of fracture in the elderly population. Although the combined application of folate and vitamin B12 can reduce HCY, the present meta-analysis revealed that this intervention was not associated with a lower risk of fracture in older people. A similar phenomenon was noted in a recent large-scale meta-analysis (49) confirming that vitamin D and/or with calcium supplementation was not associated with the decrease of fracture risk in community-dwelling older adults, whereas many trials had previously validated that calcium and vitamin D supplementation could significantly improve BMD (50-52). If folate, vitamin B12, and vitamin B6 share similar mechanisms, the underlying reason that no threshold-effect on fracture risk was observed by intervention of HCY might be that the intervention was not applied to a properly targeted population, subsequently attenuating the effect on reducing HCY level. Based on this hypothesis, we performed a further subgroup analysis to investigate participants with different serum HCY levels. Intention-to-treat participants by serum HCY was separated into 4 groups: the lowest quartile, the second quartile, the third quartile, and the highest quartile. Our subgroup analysis generated encouraging results showing that only the highest quartile of HCY instead of the other 3 quartiles increased the fracture risk in older people. Furthermore, we reviewed all the intervention trials and found that the mean level of serum HCY, except that of a subgroup in the B-PROOF study, was below 15 µmol/L, which implies that the intervention was not applied to the correct population, leading to no significant effects on decreasing fracture risks (15,19-21,53). Moreover, the subgroup analysis in the B-PROOF study (15,53), with participants over 80 years old and extended follow-up evidence of higher HCY level participants with baseline serum HCY of 15.3 µmol/L (13.6–18.2 µmol/L), supported our hypothesis.

Regretfully, the studies included in our meta-analysis provided no data regarding the extent of lowering HCY. There was 1 study that reported a mean reduction of HCY of 4.5 µmol/L after a 2-year intervention of folate and vitamin B12 (15), while others had a lowering effect of less than 3 µmol/L (6,17,19-21), indicating a potential bias caused by various interventions. Concludingly, all the evidence to date supports the hypothesis that only older people with HCY levels of 15 µmol/L (highest quartile) or above need supplementations of folate and vitamin B12. We also found that the evidence for beneficial effects of vitamin B6 application was not consistent. Garcia Lopez et al. stated that intervention with folate and vitamin B12 combined with vitamin B6 did not produce a greater effect on reducing serum HCY. Contrarily, an overdose of vitamin B6 was positively correlated with the fracture incidence possibly owing to its toxic effect (21). Therefore, any supplementation of vitamin B6, based on the current evidence, is not recommended unless a severe deficiency of vitamin B6 has been confirmed.

Based on the dose-dependent effect of HCY reported in a study, 15 µmol/L is before the J point in the standard curve, indicating a significant effect of B vitamins supplementation with a higher threshold of serum HCY level (39). Guidelines regarding the threshold of serum HCY level have been established in cardiovascular disease, cerebrovascular disease, and chronic kidney disease over the past decade (9,10,16). However, consensus can yet be drawn for intervention of HHCY and osteoporotic fractures due to the inconsistent results from studies to date. We therefore strongly encourage more large-scale RCTs focusing on the threshold of HHCY to investigate if a higher threshold would deliver a significant intervention effect on fracture risks in older adults.

The supplementation of both vitamin D and B vitamins has been validated by strong evidence to be ineffective at lowering fracture risk in community dwelling older people. We tend to believe that in the past decades, researchers have, to some extent, overestimated the beneficial effect of various nutrients including but not only limited to calcium/vitamin D and B vitamins on reducing the risk of osteoporotic fractures in relatively low-risk older people. In the future, we suggest placing increased emphasis on more focalized and economically efficient screening of the high-risk population and the corresponding intervention to achieve a safer and more economical goal, by which the public health resource might be appropriately exploited.

In this study, there were several limitations present. First, no analysis based on evidence to date was able to be carried out to assess the association of people with >15 mmol/L HCY with fracture risk. Hence, no direct evidence is available to validate our hypothesis. Second, HR values from a 199-participants trial by Sato el al. in 2005 was extracted from a graph due to the raw data being inaccessible, which caused some deviation. Besides, some included trials did not test HCY, folate, vitamin B12, and vitamin B6 in all participants, which compromised the analytical power.

Conclusions

This meta-analysis revealed that significantly elevated serum level of HCY is positively associated with fracture incidence in older adults, yet the necessity and threshold for intervention by B vitamins (folate, vitamin B12, and B6) require further large-scale high-quality clinical trials to validate.

The article’s supplementary files as