Zinc for infection prevention in children with sickle cell anemia: a randomized double-blind placebo-controlled trial

Abstract

Data from small clinical trials in the United States and India suggest zinc supplementation reduces infection in adolescents and adults with sickle cell anemia (SCA), but no studies of zinc supplementation for infection prevention have been conducted in children with SCA living in Africa. We conducted a randomized double-blind placebo-controlled trial to assess zinc supplementation for prevention of severe or invasive infections in Ugandan children 1.00-4.99 years with SCA. Of 252 enrolled participants, 124 were assigned zinc (10 mg) and 126 assigned placebo once daily for 12 months. The primary outcome was incidence of protocol-defined severe or invasive infections. Infection incidence did not differ between treatment arms (282 vs. 270 severe or invasive infections per 100 person-years, respectively, incidence rate ratio of 1.04 [95% confidence interval (CI), 0.81, 1.32, p=0.78]), adjusting for hydroxyurea treatment. There was also no difference between treatment arms in incidence of serious adverse events or SCA-related events. Children receiving zinc had increased serum levels after 12-months, but at study exit, 41% remained zinc deficient (<65 μg/dL). In post-hoc analysis, occurrence of stroke or death was lower in the zinc treatment arm (adjusted hazard ratio (95% CI), 0.22 (0.05, 1.00); p=0.05). Daily 10 mg zinc supplementation for 12 months did not prevent severe or invasive infections in Ugandan children with SCA, but many supplemented children remained zinc deficient. Optimal zinc dosing and the role of zinc in preventing stroke or death in SCA warrant further investigation. This trial was registered at clinicaltrials.gov as #NCT03528434.

Graphical abstract

Figure 1.

Incidence of severe or invasive infections. Negative binomial regression analysis with and without adjustment for time treated with hydroxyurea. The bars and point estimates for the CIs correspond to crude incidence rate ratios. ∗The adjusted result is from a Poisson mixed-effects regression model, as the negative binomial counterpart did not converge or had a non-positive variance estimated for the random intercept. †No convergence (no event in at least 1 treatment arm).

Figure 2.

Incidence of clinically defined infections. Negative binomial regression analysis with and without adjustment for time treated with hydroxyurea. The bars and point estimates for the CIs correspond to crude incidence rate ratios. ∗The adjusted result is from a Poisson mixed-effects regression model, as the negative binomial counterpart did not converge or had a non-positive variance estimated for the random intercept. †No convergence (no event in the placebo group).

Figure 3.

Incidence of adverse events. Hazard and incidence rate ratios obtained by Cox regression and negative binomial regression, respectively, with and without adjustment for time treated with hydroxyurea. The bars and point estimates for the CIs correspond to the crude incidence rate ratios. ∗The adjusted result is from a Poisson mixed-effects regression model as the negative binomial counterpart did not converge or had a non-positive variance estimated for the random intercept. †Hazard ratio is computed for this binary outcome with hydroxyurea as a time-varying covariate. Incidence rate ratios are computed for other outcomes. ‡ No convergence (no event in the zinc group).