A head-to-head comparison of four artemisinin-based combinations for treating uncomplicated malaria in African children: a randomized trial

Abstract

Background: Artemisinin-based combination therapies (ACTs) are the mainstay for the management of uncomplicated malaria cases. However, up-to-date data able to assist sub-Saharan African countries formulating appropriate antimalarial drug policies are scarce.

Methods and findings: Between 9 July 2007 and 19 June 2009, a randomized, non-inferiority (10% difference threshold in efficacy at day 28) clinical trial was carried out at 12 sites in seven sub-Saharan African countries. Each site compared three of four ACTs, namely amodiaquine-artesunate (ASAQ), dihydroartemisinin-piperaquine (DHAPQ), artemether-lumefantrine (AL), or chlorproguanil-dapsone-artesunate (CD+A). Overall, 4,116 children 6-59 mo old with uncomplicated Plasmodium falciparum malaria were treated (1,226 with AL, 1,002 with ASAQ, 413 with CD+A, and 1,475 with DHAPQ), actively followed up until day 28, and then passively followed up for the next 6 mo. At day 28, for the PCR-adjusted efficacy, non-inferiority was established for three pair-wise comparisons: DHAPQ (97.3%) versus AL (95.5%) (odds ratio [OR]: 0.59, 95% CI: 0.37-0.94); DHAPQ (97.6%) versus ASAQ (96.8%) (OR: 0.74, 95% CI: 0.41-1.34), and ASAQ (97.1%) versus AL (94.4%) (OR: 0.50, 95% CI: 0.28-0.92). For the PCR-unadjusted efficacy, AL was significantly less efficacious than DHAPQ (72.7% versus 89.5%) (OR: 0.27, 95% CI: 0.21-0.34) and ASAQ (66.2% versus 80.4%) (OR: 0.40, 95% CI: 0.30-0.53), while DHAPQ (92.2%) had higher efficacy than ASAQ (80.8%) but non-inferiority could not be excluded (OR: 0.35, 95% CI: 0.26-0.48). CD+A was significantly less efficacious than the other three treatments. Day 63 results were similar to those observed at day 28.

Conclusions: This large head-to-head comparison of most currently available ACTs in sub-Saharan Africa showed that AL, ASAQ, and DHAPQ had excellent efficacy, up to day 63 post-treatment. The risk of recurrent infections was significantly lower for DHAPQ, followed by ASAQ and then AL, supporting the recent recommendation of considering DHAPQ as a valid option for the treatment of uncomplicated P. falciparum malaria.

Trial registration: ClinicalTrials.gov NCT00393679; Pan African Clinical Trials Registry PACTR2009010000911750

Conflict of interest statement

Prof. Umberto d‚Alessandro has received research funding from Sigma Tau and Sanofi-Aventis. Dr. Quique Bassat has received in the past three years speaker fees and travel grants from Sigma Tau SP, Pomezia, Rome, Italy. All other authors declare no competing interests.

Figures

Figure 1. Trial profile up to day 28.

Adj TF, adjusted treatment failure; Unadj TF, unadjusted treatment failure.

Figure 2. Proportion of patients whose treatment was failure-free by day of follow-up.

ITT population; data pooled over all sites.

Figure 3. Six pair-wise comparisons at day 28 and day 63.

OR (circles), 95% CI (horizontal bars), and non-inferiority limit (vertical bars) for each pair-wise analysis for PCR-unadjusted (left panel) and -adjusted (right panel) ACPR at days 28 (filled circles) and 63 (open circles) (ITT population). An OR and 95% CI

Figure 4. Treatment efficacy by pair-wise comparison and by site.

Risk difference and 95% CI for three pair-wise analyses by site for PCR-unadjusted (left panel) and -adjusted (right panel) ACPR at days 28 (filled circles) and 63 (open circles) (ITT population) (A) ASAQ versus DHAPQ; (B) AL versus ASAQ; (C) AL versus DHAPQ. BF, Burkina Faso; GA, Gabon; MZ, Mozambique; NG, Nigeria; RW, Rwanda; UG, Uganda; ZM, Zimbabwe.

Figure 5. Gametocyte prevalence by treatment and day of follow-up.

(A) All patients regardless of gametocytemia at enrollment; (B) patients with gametocytes at enrollment excluded.

Figure 6. Hematological recovery by treatment and day of follow-up.

Hematological recovery determined by Hb changes compared to day 0. The boxplots contain the following information: median (bold line), first and third quartile (box), whiskers extending to 1.5× interquartile range, and all more extreme values.