Background: Osteoporosis research has focused on vertebral fractures and trabecular bone loss. However, non-vertebral fractures at predominantly cortical sites account for 80% of all fractures and most fracture-related morbidity and mortality in old age. We aimed to re-examine cortical bone as a source of bone loss in the appendicular skeleton.
Methods: In this cross-sectional study, we used high-resolution peripheral CT to quantify and compare cortical and trabecular bone loss from the distal radius of adult women, and measured porosity using scanning electron microscopy. Exclusion criteria were diseases or prescribed drugs affecting bone metabolism. We also measured bone mineral density of post-mortem hip specimens from female cadavers using densitometry. Age-related diff erences in total, cortical, and trabecular bone mass, trabecular bone of cortical origin, and cortical and trabecular densities were calculated.
Findings: We investigated 122 white women with a mean age of 62·8 (range 27–98) years. Between ages 50 and 80 years (n=89), 72·1 mg (95% CI 67·7–76·4) hydroxyapatite (68%) of 106·5 mg hydroxyapatite of bone lost at the distal radius was cortical and 34·3 mg (30·5–37·8) hydroxyapatite (32%) was trabecular; 17·1 mg (11·7–22·5) hydroxyapatite (16%) of total bone loss occurred between ages 50 and 64 years (n=34) and 89·4 mg (83·7–101·1) hydroxyapatite (84%) after age 65 years (n=55). Remodelling within cortex adjacent to the marrow accounted for 49·9 mg (45·4–53·7) hydroxyapatite (47%) of bone loss. Between ages 50–64 years (n=34) and 80 years and older (n=33), cortical density decreased by 127·8 mg (93·1–162·1) hydroxyapatite per cm³ (15%, p<0·0001) before porosity trabecularising the cortex was included, but 374·3 mg (318·2–429·5) hydroxyapatite per cm³ (43%, p<0·0001) after; trabecular density decreased by 18·2 mg (–1·4 to 38·2) hydroxyapatite per cm³ (14%, p=0·06) before cortical remnants were excluded, but 68·7 mg (37·7–90·4) hydroxyapatite per cm³ (52%, p<0·0001) after.
Interpretation: Accurate assessment of bone structure, especially porosity producing cortical remnants, could improve identifi cation of individuals at high and low risk of fracture and therefore assist targeting of treatment.