Is there an association between low dose aspirin and anemia (without overt bleeding)? Narrative review

Overt bleeding associated with low dose aspirin (LDA) is well-recognized, little attention is given to the possibility of association between LDA and occult bleeding, although this is known to occur in healthy volunteers. LDA is used increasingly in primary and secondary prevention of a number of medical conditions, many of which are common in older people, as is anemia. Anemia in older people is associated with adverse outcomes including disability, morbidity and mortality. The purpose of this study was to review the evidence that LDA might cause anemia without overt bleeding.

An extensive narrative review was carried out. Electronic searching (including database links) and reference lists of reports were used to identify studies reporting on use of aspirin ≤325 mg/day and anemia or change in hemoglobin (Hb) without overt bleeding. Data were extracted from reports of trials, adverse drug reactions (ADRs) and prevalence studies of adults aged ≥18 years, published since 1980.

There are few relevant data, with considerable heterogeneity among trial designs, duration, and patient characteristics in studies of LDA. In five randomised trials (n = 5879) in (mostly secondary) prevention, the majority of patients were men without peptic ulcer disease aged 50-70 years and no consistent association between LDA and change in Hb was found. In two smaller studies (n = 609) of primary prevention in healthy patients aged ≥70 years, there was a small but statistically significant fall in Hb with LDA. Observational studies, and data from trials in which use of LDA was not a primary focus of the study, were inconclusive.

It is not clear whether there is an association between LDA and anemia in the absence of overt bleeding, but there may be an association between LDA and fall in Hb in (a subset of) older patients. The available evidence has significant limitations, which are discussed; studies including more older patients, and publication of individual patient data, would help clarify this important matter.