Vitamin B12 deficiency was linked to higher fracture hazard after 50

By Dr. Liji Thomas, MDReviewed by Lauren HardakerOct 2 2026

A study found that abnormal vitamin B12 levels at both ends of the spectrum were associated with fractures, suggesting the relationship is more complex than deficiency alone.

Study: Vitamin B12 status and long-term fracture risk in a multicenter propensity score–matched cohort. Image credit: Tatjana Baibakova/Shutterstock.com

A recent retrospective study in the journal Scientific Reports found an association between vitamin B12 deficiency and long-term fracture hazard in a propensity-matched cohort. However, the authors cautioned that this should not be interpreted as meaning that vitamin B12 deficiency causes or predicts fracture.

B12 deficiency may affect bones and falls

Fractures can have lasting consequences for middle-aged and older adults, contributing to ill-health, reduced function, and loss of independence. Their causes are often complex, with fracture risk shaped by nutrition, neuromuscular function, underlying health conditions, medications, and factors that make falls more likely.

Vitamin B12 could play a role in several of these pathways. The nutrient is essential for cell division and neurological function, and deficiency may contribute to skeletal fragility and increase susceptibility to falls.

Vitamin B12 and folate are key cofactors in homocysteine metabolism, and deficiency can contribute to hyperhomocysteinemia, which has been associated with altered collagen cross-linking and impaired bone quality. B12 deficiency may also impair osteoblast function, while neurological effects such as peripheral neuropathy, muscle weakness, and unstable gait could further increase fall risk.

Despite these possible links, previous studies examining vitamin B12 and fractures have produced inconsistent results. Many involved relatively small or selected populations and focused on hip fractures or fracture incidence alone. Interpreting the relationship is further complicated by the fact that abnormal B12 levels can reflect broader nutritional problems, supplementation, underlying illness, or patterns of healthcare use rather than a specific effect of the vitamin itself. Even undergoing B12 testing may identify a group with different health characteristics from the wider population.

Comparing B12 status with subsequent fractures

The researchers used electronic health records from the TriNetX Global Collaborative Network, drawing on data from multiple healthcare organizations. After applying the initial eligibility criteria, the study included 115,735 patients with vitamin B12 deficiency and more than 1.76 million with concentrations in the normal range. All had undergone B12 testing between 2016 and 2023 and had at least one healthcare visit during the previous five years, helping the researchers capture relevant medical history.

Patients were grouped according to their B12 concentrations: deficiency was defined as 199 pg/mL or less, borderline levels as 200–299 pg/mL, the normal reference range as 300–900 pg/mL, and high levels as above 900 pg/mL. Those with conflicting B12 measurements during the previous five years were excluded to keep the categories clearly separated.

To reduce the possibility that an existing or imminent fracture influenced B12 testing, the researchers used a one-year landmark design. Participants had to be alive and free of fracture, osteoporosis, and repeated falls when outcome follow-up began. These conditions were also exclusion criteria if recorded during the five years before the index measurement through the landmark period, although a single prior fall was permitted and accounted for in the matching process. Patients with several conditions that could affect bone health, as well as those with recent critical illness, were also excluded.

The eligible patients were then matched according to a broad range of demographic characteristics, health conditions, medications and available laboratory measures, leaving 115,735 people in each of the B12-deficient and normal-B12 groups. From one year after the initial B12 measurement, the researchers tracked incident fractures across multiple parts of the body as the primary outcome.

The analysis went beyond fractures alone. Secondary outcomes included lower-limb and osteoporotic fractures, osteoporosis without fracture, all-cause falls, and repeated falls. Hyperhomocysteinemia served as a positive control, while lipoma was used as a negative control to help assess nonspecific associations and residual confounding.

Open wounds provided an additional comparison for whether greater exposure to injury, rather than bone fragility, might partly explain the fracture findings. The researchers also carried out sensitivity and subgroup analyses to test how consistently the main association appeared under different analytical conditions.

Participant characteristics

Before matching, the vitamin B12-deficient cohort was older, with a higher prevalence of diabetes and metformin use. After propensity-score matching, measured baseline characteristics were well balanced between the groups. However, the follow-up period was shorter in the vitamin B12-deficient group, and the authors suggest that this could be due to different mortality rates in the groups.

Fracture hazard higher in the B12-deficiency group

Fractures occurred in 7.2% of the B12-deficient patients (8,362 individuals), versus 6% (6,962) of the reference group. The adjusted analysis found that the fracture hazard was 33% higher among patients who remained alive and under observation in the deficiency group.

Borderline and high vitamin B12 levels were also associated with increased fracture hazard, though the associations were weaker. This non-monotonic pattern argues against a simple deficiency-specific or dose-response relationship. These may support the coexistence of vitamin B12 abnormalities with conditions or factors other than those with a direct biological link. In particular, the high-B12 association may reflect underlying morbidity, supplementation, or greater healthcare contact.

However, the analysis suggested non-proportionality of the hazard rate over the full period. Further testing using only the 1–5 years after the index measurement showed an increase in fracture hazard by 37%, without evidence of non-proportionality, and fracture-free probability was consistently lower in the deficiency group compared to controls throughout the follow-up.

Secondary outcomes were concordant with the primary outcome. Higher hazards of osteoporosis, osteoporotic fracture, lower-limb fracture, all-cause falls, and repeated falls were found. The increase in hazard ranged from 22% for osteoporosis without fracture to 50% for repeated falls.

The authors suggest that these findings may mean that deficiency is linked to both poorer bone health and fall-related pathways that act together to eventually impact fracture risk.

Hyperhomocysteinemia was associated with a 41% higher hazard in the deficiency group, but not lipoma. Open wounds were associated with a 12% higher hazard. Conversely, high B12 levels were negatively linked to hyperhomocysteinemia. None of these outcomes showed non-proportionality of hazards. However, these are exploratory findings as they do not rule out residual confounding.

The pattern persisted across all sensitivity analyses with similar direction and magnitude of association. A complementary multivariable analysis of the unmatched cohort also found an association between vitamin B12 deficiency and fracture after adjustment. Subgroup analyses suggested stronger associations with fracture among patients aged 50–65 years compared to those above 65, but not in other subgroups including sex, obesity, or proton pump/metformin use. This was also considered an exploratory and hypothesis-generating finding.

The findings are biologically plausible; however, they remain observational. The current study adds to previous knowledge by identifying associations between vitamin B12 levels and a wider range of outcomes, reducing the risk of reverse causation and adjusting for many known confounders.

Testing and follow-up complicate the B12 association

Several limitations temper the findings. Most importantly, the observational design means the study cannot establish that vitamin B12 deficiency causes fractures, and unmeasured factors may still have influenced the association despite extensive adjustment. Vitamin B12 status was also based on the index measurement and was not updated during follow-up, so the analysis could not account for patients whose levels later changed, including those who received supplementation.

The study population itself warrants caution. All participants had undergone clinically indicated vitamin B12 testing, meaning they may differ from the wider population in ways related to nutrition, frailty or underlying illness. The broad exclusion criteria further narrow the population to which the findings can be applied.

Health-record data also had limits. Bone mineral density measurements were unavailable, and potentially important factors such as continuous vitamin D, calcium, and parathyroid hormone levels; muscle strength; sarcopenia; physical activity; diet; and cumulative glucocorticoid exposure could not be fully accounted for. The primary fracture outcome also covered multiple fracture types, so the researchers could not reliably distinguish fractures related to bone fragility from those caused by trauma.

Finally, mortality complicated the long-term comparison. Death was treated as a censoring event, and the B12-deficient group had higher mortality and shorter follow-up. As a result, the primary analysis describes fracture hazard among patients who remained alive and under observation, rather than the absolute cumulative probability of experiencing a fracture.

B12 deficiency emerges as potential fracture risk marker

The large multicenter study described an association between vitamin B12 status at baseline and subsequent fracture hazard in adults aged 50 years and above who underwent vitamin B12 testing, but without causality or predictive ability due to the study design. The primary outcome described fractures across multiple anatomical sites and was not restricted to fragility fractures resulting from bone fragility due to vitamin B12 deficiency.

However, it does not establish that deficiency causes or predicts fracture, or that B12 supplementation reduces the risk. This will require further prospective studies.