Association between dietary total fiber and bone loss in the Framingham Offspring Study

Abstract

ObjectivesThe effect of dietary fiber on bone health is unclear. This may be due to the complex relationship including its impact on reducing body weight and conflicting results in calcium absorption. To best knowledge, longitudinal data on this issue in older populations at risk of osteoporosis is lacking. Therefore, the objective of this study was to examine the relation of dietary fiber to bone loss longitudinally in older adults.MethodsIn the Framingham Offspring Study, bone mineral density (BMD, g/cm2) was measured at the femoral neck, trochanter and lumber spine (L2–L4) at baseline (1996–2001) and the follow‐up exam (2005–2008) in 653 men and 843 women. At baseline, habitual diet including dietary fiber intake (g/d) was assessed using the Willett validated food frequency questionnaire. Dietary fiber was further adjusted for total energy intake using the residual method developed by Willett. Annualized percent change (%) in BMD was calculated as [(follow‐up BMD ‐ baseline BMD) / baseline BMD] × 100% / (years between two BMD measures). Multivariable regression was used to examine the association between total dietary fiber (as a continuous variable) and percent change in BMD after adjustment for confounders and baseline BMD at the corresponding site.ResultsThe mean time difference between the two BMD assessments is 8.1 years (SD: 1.1 years). Baseline characteristics in men and women are described in Table 1. The estimated average intake of fiber, calcium and vitamin D are slightly below the recommended intake. The main results are presented in Table 2. In men, dietary fiber was significantly associated with less bone loss, indicated by a positive relation with annualized percent change in BMD at the femoral neck (standardized β coefficient=0.11, p=0.015) and at the trochanter (standardized β coefficient=0.09, p=0.05), but no association was found at the lumbar spine. However, the relationship was null at all sites in women.ConclusionsData from the Framingham Offspring Study suggested a possible sex specific difference for the association between fiber intake and change in BMD. In men, a higher fiber intake was suggested to protect against bone loss, particularly at the femoral neck and trochanter sites. In women, there was no evidence suggesting that fiber intake was related to change in BMD. Further studies are needed to assess what may be the contributing factors for such sex differences in fiber intake and BMD change.Support or Funding InformationThis study was supported by NIH grants T32 AR 7598 and NHLBI, Framingham Heart Study (NHLBI/NIH contract #N01‐HC‐25195) and the Boston University School of Medicine. Baseline characteristics in men and women Men Women N 653 843 Age (year)* 58.0 (8.9) 57.1 (8.7) BMI (kg/m2) 28.4 (4.1) 27.0 (5.2) Education: =college (%) 34.1 22.0 Never smokers (%) 81.1 80.4 Physical activity (PASE) 163.8 (84.9) 144.4 (71.7) Total energy intake (kcal) 1994.0 (647.9) 1755.5 (556.8) Total fiber intake (g/day) 19.8 (7.8) 19.4 (7.9) Vitamin D intake (IU/day) 367.3 (261.6) 419.6 (288.1) Calcium intake (mg/day) 821.2 (403.5) 1036.3 (535.2) Caffeine intake (units/day) 272.2 (211.2) 235.6 (194.3) Femoral BMD (g/cm2) 0.97 (0.13) 0.87 (0.13) Trochanter BMD (g/cm2) 0.87 (0.14) 0.71 (0.13) Lumbar spine BMD (L2–4) (g/cm2) 1.31 (0.20) 1.15 (0.19) Such values for mean (standard deviation) Association of dietary fiber (g/d) and annualized percent change (%) in BMD (g/cm2) at different site between baseline (1996–2001) and follow‐up exam (2005–2008) in men and women BMD change (%) Femoral neck Trochanter Lumbar spine N β1 β2 SE p‐value N β1 β2 p‐value N β1 β2 SE p‐value Men Model 1 607 0.01 0.11 0.005 0.02 607 0.01 0.09 0.05 597 0.007 0.05 0.006 0.25 Model 2 584 0.012 0.11 0.005 0.015 595 0.01 0.09 0.05 595 0.008 0.06 0.006 0.18 Women Model 1 789 −0.004 −0.03 0.005 0.49 789 0.002 0.01 0.72 779 −0.001 −0.008 0.006 0.83 Model 2 781 −0.005 −0.04 0.005 0.27 768 −0.003 −0.02 0.63 771 −0.001 −0.006 0.006 0.89 β1unstandardized parameter coefficient; β2 standardized parameter coefficient Model 1 adjusted for total energy intake (kcal/day), age (year), BMI (kg/m2) Model 2 further adjusted for cigarette smoking (never smokers, 0‐=15 cig/day, >15 cig/day), physical activity (PASE), baseline BMD at the corresponding site, total calcium intake (mg/day), total vitamin D intake (IU/day), and caffeine intake (mg/day).