A study was conducted in a sample of 140 children with sickle cell anemia to evaluate the relationship between hematological variables (%HbF, %HbA2, %Hb, and mean cell volume) and disease severity. A patient's severity status was determined by whether he/she was hospitalized, had a transfusion, and/or had a pain crisis at 2 evaluation periods; the first was based on a patient's history taken at the initial assessment visit to the Wayne State Comprehensive Sickle Cell Center, and the second was based on a 1-3 year follow-up at the center. Fetal hemoglobin was a strong predictor of a patient's hospitalization and transfusion status. A decrease in %HbF of 4.76% (one SD of %HbF) was associated with a 3.58 fold (95% confidence interval, 1.18-7.28) greater odds of being hospitalized both prior to initial assessment and on follow-up, compared to not being hospitalized at either evaluation. Similarly, a decrease in %HbF of 4.76% was associated with a 5.56 fold (95% confidence interval, 1.67-18.96) greater odds of having a transfusion both prior to initial assessment and on follow-up compared to not having a transfusion at either evaluation. Patients who were both hospitalized and transfused at initial assessment and on follow-up (n = 12) had a mean %HbF of 7.59%, while patients who were not hospitalized or transfused at either evaluation (n = 19) had a mean %HbF of 13.61%.Fecal hemoglobin was not a significant predictor of pain crises in this sample of patients. None of the other hematological variables were significant predictors of disease severity in this study. The strong relationship between %HbF and disease severity identified in this study suggests that a single %HbF and disease severity identified in this study suggests that a single %HbF measurement may be useful in predicting important aspects of the clinical course of children with sickle cell anemia.