Background: Although alcohol abuse is known to create a variety of adverse effects on hematopoiesis, the associations between ethanol consumption and hematological abnormalities have not been fully established.

Methods: We studied 144 consecutive adult patients who underwent clinical and bone marrow examinations due to abnormal findings in peripheral blood cell counts or red blood cell indices without previously established diagnoses of specific hematological diseases, malignancies, or infections. Assessment included the amount of alcohol consumption, complete blood cell counts, morphological review of peripheral blood and bone marrow, markers of liver status, and erythrocyte folate and serum vitamin B12 levels.

Results: There were 57 (40%) patients who showed a history of hazardous drinking and 87 patients who were either nondrinkers or social drinkers. The incidence of anemia was 51% in the alcohol abusers, as compared with 69% of the nonalcoholics (p < 0.05). A diverse pattern of hematological effects was observed in the alcohol abusers. Abnormal platelet and leukocyte levels were common, especially in the anemic alcoholics. Both increased mean cell volume of erythrocytes (macrocytosis; 67 vs. 18%; p < 0.0001) and mean cell hemoglobin (63 vs. 22%; p < 0.0001) were more frequent in the alcoholics than in the nonalcoholics. Reticulocytosis (37%), thrombocytopenia (41%), and combined cytopenias (34-38%) were also common findings in the alcoholic patients. The blood smears from such patients typically showed round macrocytes, stomatocytes, and knizocytes. Bone marrow aspirates revealed vacuolization of pronormoblasts in 24% of the alcoholic patients. Interestingly, megakaryocytes in the cell periphery were also vacuolized in 20% of the alcohol abusers, especially in those with recent intoxication.

Conclusions: Our findings suggest that alcohol abuse results in diverse patterns of hematological effects and affects several cell lines. Therefore, in patients undergoing bone marrow examinations due to cytopenias, the probabilities for likely findings seem to be different between alcoholics and nonalcoholics. Information on ethanol consumption should be systematically included in the clinical assessment of such patients.