To the Editor: Anemia is associated with poor outcomes in older people,1 and low hemoglobin (Hb) levels are independently associated with greater risk of frailty.2 The etiology of anemia in people aged 65 and older divides approximately into thirds: iron, vitamin B12, or folate deficiency; chronic disease; and unexplained.3 Studies examining the relationship between Hb and erythropoietin (EPO) in older populations have reported conflicting results. Some have found that older people have an impaired EPO response to iron deficiency anemia,4 anemias of malignancy or infection,5 and unexplained anemias.6 Other studies have reported no difference in EPO response to anemia with aging.7 The aim of this study was to investigate the relationship between Hb and EPO across age and across varying degrees of frailty. One hundred ten subjects aged 75 and older were recruited from long-stay inpatient wards (n=30), a day hospital (n=40), and the community (n=40). Thirty healthy controls younger than 35 were also studied. Forty percent of all participants were male, and all were Caucasian. A single observer (REH) measured frailty indicators in all subjects. A Frailty Index (FI) was constructed from 30 variables including self-reported data, symptoms, comorbidities, and performance-based tests using a method developed previously.8 Participants were categorized according to Fried criteria: exhaustion (Energy and Vitality Score on the Medical Outcomes Study 36-item Short-Form Survey <40%), weight loss (≥10 pounds in preceding year), slow walking speed (6-minute walking distance of <210 m), low handgrip strength, and low physical activity. Individuals with three or more components were defined as frail. The local research ethics committee approved the study. Standard complete blood count was performed using a Coulter counter and EPO using enzyme immunoassay. Differences between study groups were analyzed using independent sample t-tests for normally distributed variables, Mann-Whitney U for variables not normally distributed and chi-square analyses for prevalence of anemia. Correlations between log-transformed EPO and Hb were explored using Pearson correlation coefficients. The FI is not meant to be dichotomized into frail or robust, but an empirical cutoff point of 0.25 has been proposed.8 Of the 110 older subjects, 69 (63%) had a FI of 0.25 or greater, and 63 (57%) had three or more Fried frailty indicators (kappa=0.77). The difference in Hb between young controls and fit older people was not statistically significant, but Hb was lower with greater patient frailty (Table 1). In older people, there was a significant negative correlation between Hb and FI (correlation coefficient (r)=−0.399, P<.01). A significant inverse relationship between Hb and log-transformed EPO was preserved in older people defined as fitter using FI (<0.25: r=−0.408, P<.01) or Fried phenotype (≤2 frailty indicators: r=−0.329, P<.05). This correlation was lost in those who were frail (FI≥0.25: r=−0.115, P=.38; Fried frail: r=−0.191, P=.17). This study has important limitations. Numbers of subjects were small, and the etiology of anemia was not determined, but anemia was subclinical, normocytic, and with normal white blood cell and platelet counts, which make iron, vitamin B12, or folate deficiency and myelosuppression less likely.9 In addition, this was a cross-sectional design, and no inferences can be drawn regarding anemia, low EPO and causality of frailty development. Despite these limitations, this study contributes to the discussion about anemia and aging. It confirms that anemia may not be an inevitable part of the aging process,10 because there was no significant difference in Hb between young controls and fit older people. The significant negative correlation between Hb and the FI, which depicts frailty as a continuum rather than a binary outcome, provides additional insights into the association between frail status and anemia.2 Loss of correlation between Hb and EPO supports the hypothesis that anemia of aging is associated with inappropriately low EPO levels.6 That the correlation seems to be lost only in patients who are frail, a consistent finding across different approaches to the concept, should prompt further research into the temporal relationships between falling Hb, EPO production, and development of frailty. Conflict of Interest: Dr. Ruth E. Hubbard: currently funded by Peel Medical Research Trust, London. The Trust played no role in the design, analysis and interpretation of data, or writing of the study. Author Contributions: Dr. Ruth E. Hubbard contributed to study design, recruited patients, conducted the functional tests, analyzed data, and cowrote the letter. Dr. M. Sinead O'Mahony designed the study and cowrote the letter. Professor Ken W. Woodhouse contributed to study design and cowrote the letter. Sponsor's Role: No sponsor.