UCSF

BACKGROUND: The association between obesity and dementia has been inconsistent, possibly due to changes in body composition often seen in old age. Leptin may be associated with better cognitive function. However, neuroprotection may be inhibited among obese subjects possibly due to leptin resistance. We sought to determine (i) if leptin is associated with risk of dementia or mild cognitive impairment (MCI) in a cohort of very old women, (ii) if this association is modified by obesity, and (iii) if leptin is a stronger risk factor compared with traditional anthropometric measures. METHODS: We studied 579 older women (mean age 82.6 years) from the ongoing prospective cohort Study of Osteoporotic Fractures, who were dementia-free at year-16 examination (our study baseline). Leptin (ng/mL) was measured using year-16 frozen serum, and anthropometric measures were collected during the same visit. Diagnosis of dementia/MCI was determined at year-20 examination. RESULTS: There was evidence for a multiplicative interaction between log leptin and categorical body mass index (p = .03). Among women with body mass index <25kg/m(2) (n = 190), 1SD difference in log leptin (0.91ng/mL) was associated with 32% lower odds of dementia/MCI (OR = .68; 95% CI = .46, .99), after adjustment. The association was not significant among women with body mass index ≥25kg/m(2) (n = 377). Traditional anthropometric measures such as weight, height, and body mass index were not associated with dementia/MCI. CONCLUSIONS: In this cohort of very old women, higher serum leptin was prospectively associated with lower odds of dementia/MCI in women with normal body mass index, but not in overweight or obese women. Leptin may be a better predictor of dementia/MCI than traditional anthropometric measures.