In a randomized, placebo-controlled clinical trial, treatment with hydroxyurea (HU) reduced crisis rates in adult patients with severe sickle cell anemia. No serious acute toxicity was seen, but the safety of long-term therapy could not be evaluated. The rationale for the use of HU was based on its ability to increase fetal hemoglobin (HbF) synthesis and the inhibitory effect of HbF on polymerization of sickle cell hemoglobin. Surprisingly, final HbF levels in patients assigned to HU did not differ markedly from their pretreatment levels (5% v 9%). As Steinberg et al showed, when HU patients were divided into quartiles based on final HbF levels, those in the highest quartile had an 18% mean HbF, while those in the lowest quartile had a mean of only 4%. Higher HbF levels were associated with lower crisis rates, but the association was not statistically significant. Further analyses suggested noncompliance of patients in the lower HbF quartiles in the later months of the study as a significant factor responsible for the difference in HbF levels. When HU patients were divided into quartiles based on their 2-year crisis rates, those with the fewest crises had lower neutrophil counts and higher mean corpuscular volumes (MCVs) and F-cell counts; similar but less marked changes were seen in patients assigned to placebo. Multivariable analyses showed a significant relationship between crisis rate and, in the early months of the study, F-cell count and, in later months, MCV. Lower neutrophil counts were associated with lower crisis rates in all months of the study. The HbF in F cells inhibits their sickling and decreases the likelihood of vaso-occlusion and infarction. If infarction does occur, lower neutrophil counts may limit the extent of tissue destruction and the severity of pain. Further study is needed to clarify the interplay of these two factors as mediators of the effect of HU in lowering crisis rates in sickle cell anemia.