Abstract
Background
The fight against diseases such as malaria requires the synthesis of evidence from existing studies to inform decision makers. Indeed, at a cross road of antimalarial drug resistance, several artemisinin-based combination therapies (ACT) with multiple doses are available to fight uncomplicated malaria. However, little is known on how these combinations are combined as well as how different formulations are tested.
Methods
A systematic review was performed to identify randomized trials. Articles were sought by hand-searching and scanning references. Additional covariates effect on treatment outcome was assessed, and a modeling approach to reduce heterogeneity among trials was evaluated. We explored one single interaction effect for all treatment with age as the main covariate in a meta-regression. A Bayesian analysis was used to implement the consistency and inconsistency models under the WinBUGS software. Ranking measure was used to obtain a hierarchy of the competing interventions.
Results
In total, 77 articles meet the inclusion criteria with 15 combinations tested in 36,000 patients. Results were compared to that of frequentist approach and presented according to the Prisma NMA checklist. The consistency model showed a good performance than the inconsistency model under the hypothesis of homogeneity. It was found that compared to artemether-lumefantrine, the dihydro-artemisinin-piperaquine was more effective before (B, OR = 1.83; 95% CI = 1.31-2.56) and after (A, OR = 1.70; 95% CI = 1.20-2.43) covariate adjustment, and occupied the top rank.
Conclusions
The application of the methods described here may be helpful to gain better understanding of treatment efficacy and improve future decisions in malaria programs. Based on the available evidence, this study demonstrated the superiority of DHAP among currently recommended ACT in preventing as well as treating uncomplicated malaria.
Key messages
Choosing the best therapy requires data triangulation and data science. Network meta-analysis could be a solution but need more methodological studies.