Factors that influence the heterogeneity of the disease expression of sickle-cell anemia are not well understood. This study examines the ability of a profile of six hematological variables (HB, HCT, RBC, %Hb F, MCV, and %HBA2) to predict the severity of disease measured on 225 patients ranging from 0.2 to 18 years of age. Four subgroups of patients were identified separately in each sex using cluster analysis techniques. In each sex, mean hemoglobin concentration and percent Hb F increased across the four clusters from 7 to 10 gm/dl and from 7% to 16%, respectively. Mean cell volumes were approximately 90, 80, 90, and 75 in groups 1, 2, 3, and 4, respectively; thus MCV did not increase in an orderly progression along with HB and %Hb F. We studied the distribution of four anthropometric, five physical examination, and seven clinical measures of disease severity among clusters. In each sex, subgroups differed significantly (P less than .05) for percent ever hospitalized for sickle-cell anemia, percent ever transfused, and percent with bone-age delays greater than 1 year. In addition, male clusters differed significantly for percent ever having had pneumonia, priapism, or dactylitis, and females differed significantly for height and weight. %Hb F and its inverse relationship with %HBA2 was more highly associated with the measures of severity than the degree of anemia or MCV. This study establishes the utility of a vector of hematological variables as a predictor of heterogeneity of measures of clinical manifestations among young patients with sickle-cell anemia. The role of alpha-thalassemia and genetic factors that affect Hb F levels were considered as possible explanations for the observed heterogeneity.