TL;DR
This study investigated whether the amount and composition of fat in bone marrow, measured non-invasively with MRI, could predict future fragility fractures in postmenopausal women. The findings indicate that, contrary to some expectations, bone marrow fat content and composition were not associated with an increased risk of incident fractures. This suggests that current MRI-based bone marrow fat measurements may not be a reliable standalone biomarker for fracture risk prediction in this population.
What Problem Does This Paper Address?
Osteoporotic fractures are a significant health concern, particularly in postmenopausal women, leading to substantial morbidity and mortality. While bone mineral density (BMD) measured by DXA is a standard tool for assessing fracture risk, it doesn't fully capture an individual's risk, as many fractures occur in individuals with seemingly normal BMD. Bone marrow adipose tissue (BMAT), the fat within the bone cavity, has emerged as a potential new biomarker. BMAT's composition and quantity change with age and metabolic status, and some cross-sectional studies have suggested a link between higher BMAT and lower bone quality or increased fracture prevalence. However, longitudinal data on whether BMAT can predict future fractures, especially using advanced MRI techniques like proton density fat fraction (PDFF) and water-fat imaging (WFI), has been limited. This paper aims to fill this gap by prospectively assessing the association between BMAT measurements (at both the lumbar spine and proximal femur) and incident clinical fractures in postmenopausal women.
Methods
The study utilized data from the ADIMOS cohort, a case-control investigation. For this specific analysis, they focused on a longitudinal follow-up of postmenopausal women.
- Participants: The study included postmenopausal women aged 45-95 years. Participants were categorized into two groups: those with recent osteoporotic fractures (cases) and those with osteoarthritis but no history of fragility fractures (controls).
- Imaging: Magnetic Resonance Imaging (MRI) was the primary tool for assessing bone marrow fat.
- Water-Fat Imaging (WFI): This technique was used to quantify the proton density fat fraction (PDFF), representing the percentage of fat within the bone marrow. Measurements were taken at the lumbar spine (L1-L4) and the proximal femur (femoral head, neck, and diaphysis).
- Proton Magnetic Resonance Spectroscopy (¹H-MRS): This method was used to assess BMAT composition, specifically measuring the apparent lipid unsaturation level (aLUL). This was performed on a subset of participants at the L3 vertebral body and the femoral neck.
- Fracture Assessment: Incident clinical fractures were recorded during the follow-up period through questionnaires and detailed interviews conducted by rheumatologists.
- Statistical Analysis: The researchers employed the Fine and Gray regression model to evaluate the association between PDFF/aLUL and incident clinical fractures, accounting for death as a competing event. Multivariable regression with a forward stepwise approach was used to identify and adjust for potential confounding factors (e.g., age, BMD, Charlson Comorbidity Index, number of falls, recent osteoporotic fractures).
Datasets
The study utilized data from the ADIMOS cohort, a case-control investigation that collected information on postmenopausal women. This study is an ancillary analysis of that cohort, focusing on longitudinal follow-up data. No new datasets were introduced.
Key Insights and Ideas
- BMAT as a Predictor? Not So Fast: The central finding is that neither the quantity (PDFF) nor the composition (aLUL) of bone marrow fat, as measured by MRI at the lumbar spine or proximal femur, was associated with an increased risk of future fragility fractures in postmenopausal women. This challenges the growing interest in BMAT as a direct predictor of fracture risk.
- Proximal Femur Fat is High: The study confirmed that bone marrow fat content is significantly higher in the proximal femur (especially the femoral head) compared to the lumbar spine, which aligns with previous observations.
- Existing Fracture Risk Factors Still Reign Supreme: While BMAT measures didn't predict fractures, the study did find that having a "recent osteoporotic fracture" was a significant predictor of future fractures. This reinforces the importance of established risk factors.
- Methodological Nuances Matter: The paper highlights the need for standardized reporting and measurement protocols (like those from the BMAS) for BMAT research, as comparability across studies is crucial.
Key Results
- No Association with WFI-PDFF: Hazard ratios (HR) for incident clinical fractures showed no significant association with WFI-based PDFF at the lumbar spine (HR=0.95, P=0.77) or the femoral neck (HR=1.07, P=0.74). Similar non-significant results were observed for the femoral head and diaphysis.
- No Association with ¹H-MRS: Similarly, ¹H-MRS-based PDFF and aLUL at the L3 vertebral level and femoral neck were not associated with incident clinical fractures.
- Recent Fractures are Key: The variable "recent osteoporotic fractures" was found to be significantly associated with incident clinical fractures (e.g., HR=2.45, P=0.027 in one model).
- Follow-up: The study followed 195 participants for an average of 3.1 years, during which 29 incident clinical fractures occurred.
Limitations and Future Work
- Sample Size: The authors acknowledge that the relatively small sample size, particularly given the low number of fracture events, may have limited the statistical power to detect smaller associations.
- Self-Reported Fractures: Reliance on self-reported fractures through telephone interviews, rather than systematic imaging during follow-up, could lead to underestimation or misclassification of fractures. While rigorous interview protocols were used, recall bias remains a possibility.
- AOM Use: The use of anti-osteoporosis medications (AOM) during the follow-up period could have reduced fracture incidence, potentially masking or obscuring relationships between BMAT and fracture risk.
- Short Follow-up: The follow-up period of 3.1 years was considered relatively short, suggesting the need for longer prospective studies.
- Population Specificity: The cohort was limited to postmenopausal women aged 45+, meaning the findings may not generalize to younger women or men.
- Future Directions: The authors emphasize the need for larger cohorts and longer follow-up periods. They also recommend focusing on proximal femur PDFF measurements, aligning with BMAS recommendations, rather than solely vertebral PDFF. Further research could explore other BMAT characteristics or different populations.
Summary updated