Cross-Sectional and Longitudinal Associations Between Anemia and Frailty in Older Australian Men: The Concord Health and Aging in Men Project

• Published in: Journal of the American Medical Directors Association Vol. 16; no. 7; p. 614
• Main Authors: Hirani, Vasant, Naganathan, Vasi, Blyth, Fiona, Le Couteur, David G, Kelly, Patrick, Handelsman, David J, Waite, Louise M, Cumming, Robert G
• Format: Journal Article
• Language: English
• Published: United States 01.07.2015
• Subjects: Aged; Aged, 80 and over; Anemia - epidemiology; Cross-Sectional Studies; Epidemiologic Studies; Frail Elderly - statistics & numerical data; Humans; Longitudinal Studies; Male; older men; population study; frailty; Anemia
• ISSN: 1538-9375
• DOI: 10.1016/j.jamda.2015.02.014

Anemia and frailty are both common in older people and are associated with adverse health outcomes. There have been some cross-sectional studies of anemia and frailty but no longitudinal studies. The objectives of this study were to examine cross-sectional and longitudinal associations between anemia and frailty in older Australian men. A total of 1666 men aged 70 years and older from the Concord Health and Aging in Men Project were assessed at baseline (2005-2007), 1314 men came for the 2-year follow-up between 2007 and 2009, and of those, 917 men returned for the 5-year follow-up between 2012 and 2013. The main outcome measurement was frailty, assessed using the Cardiovascular Health Study method. Anemia was defined as a hemoglobin levels <13.0 g/dL. Covariates included age, income, body mass index, measures of health, estimated glomerular filtration rate, and inflammatory markers (white cell count and albumin). The prevalence of anemia was 14.6% at baseline, 16.2% at 2-year follow-up, and 19.4% at 5-year follow-up. Prevalence of frailty was 9.1% at baseline and 9.7 % at both 2- and 5-year follow-up. Among men aged 70-74 at baseline, prevalence of frailty was 4.5%, but at 5-year follow-up the prevalence was 9.0%. There were significant cross-sectional associations between anemia and frailty in unadjusted [odds ratio, [OR 5.03 (95% confidence interval, CI 3.50, 7.25, P < .0001)] and in fully adjusted analysis [OR 2.90 (95% CI 1.87, 4.51, P < .0001)]. Generalized estimating equations time-lag models were used to examine the longitudinal associations between repeated measurements of hemoglobin and frailty. There were significant associations between measurements of anemia and frailty in unadjusted [OR 2.51 (95% CI 1.58, 4.00, P < .0001] and in fully adjusted analysis (OR 1.80, 95% CI 1.14, 2.85, P = .01). Anemia was associated with frailty in both cross-sectional and longitudinal analyses, and anemia precedes frailty in men who were nonfrail at baseline. Low hemoglobin levels among patients may alert clinicians to the increased risk of frailty.