Hydroxycarbamide treatment in sickle cell disease: estimates of possible leukaemia risk and of hospitalization survival benefit.

Using health insurance claims databases we compared the frequency/incidence of acute myeloid leukaemia (AML) and inpatient mortality in sickle cell disease (SCD) subjects taking (n = 1051), or not taking (n = 9203) hydroxycarbamide (HC). Patients taking HC were older (median 19 vs. 17 years of age), had a higher proportion of males (53% vs. 38%), and their median hospitalizations per year was five times greater than in SCD patients not on HC (all P < 0·001). No new AML cases occurred in HC-treated paediatric SCD patients. For adults, the new AML incidence with HC exposure was 10·7/10 000 patient years, vs. 4·0/10 000 patient years in subjects not on HC (P = 0·2), a possible AML risk ratio of 3·18. Adjustment for a probable database bias for AML diagnosis/ascertainment lowered the risk ratio to 0·94 (95% confidence interval = 0·16-5·47). Despite their greater disease severity, the inpatient mortality in SCD adults prescribed HC (0·29%) was lower than that of patients not taking the drug (0·42%, P = 0·032). In this SCD population we find no increased risk for AML with HC treatment. If such a risk is eventually proven, it will probably be lower than that for drugs with known AML association. By contrast, HC treatment appears to confer a survival benefit.