The relationship of iron status with cognition and dementia risk in older people is contentious. We have examined the longitudinal relationship between serum ferritin and cognition in 800 community-dwelling Australians 60 years or older. Iron studies (serum iron, transferrin saturation, serum ferritin) were performed in 1994/5 and 2003/4 and clinical and cognitive assessments were conducted in 2003/4 for 800 participants of the Busselton Health Study. All participants completed the Cambridge Cognitive test (CAMCOG). Those with CAMCOG scores <84 underwent expert clinical review for cognitive disorders, including the Clinical Dementia Rating scale. Mean serum iron (18.3 μmol/l) and transferrin saturation (28.5%) in 2003/4 did not differ significantly from 1994/5 whereas mean serum ferritin decreased from 162 μg/l in 1994/5 to 123 μg/l in 2003/4, possibly reflecting aging or dietary changes. No relationships were observed between serum iron or transferrin saturation and presence or absence of dementia (p> 0.05). In participants without dementia (n=749), neither serum ferritin in 1994/5 or 2003/4 nor change in serum ferritin between these times was related to total CAMCOG or executive function scores, with or without adjustment for gender, age, National Adult reading test, or stroke history (all p> 0.05). No relationships were observed between ferritin and cognition for participants with possible or probable dementia (n=51). All participants identified as HFE C282Y homozygous or with serum ferritin >1,000 ng/ml had normal CAMCOG scores. We conclude abnormal body iron stores (low or high) are unlikely to have clinically significant effects on cognition or dementia risk in community-dwelling older people.
Journal of Alzheimers Disease Vol. 20, Issue 2, p. 617-623