Incidence of infections in children before hydroxyurea treatment compared to during hydroxyurea treatment. (A) Severe or invasive infections; (B) clinically defined infections. Negative binomial or ∗Poisson mixed effects regression analysis adjusted for zinc treatment arm. Person-years before hydroxyurea = 43.0; after hydroxyurea = 71.8. ^†Both models did not converge or had nonpositive definite G matrix. Incidence rates that differ significantly after adjustment by the Benjamini-Hochberg method are in bold type.