Guidelines for the treatment of uncomplicated malaria in Africa have recently undergone a paradigm shift away from inexpensive monotherapies, such as chloroquine and sulphadoxine-pyrimethamine, to artemisinin-based combination therapies (ACT). Two of the most important artemisinin-based combinations for use in Africa are artemether-lumefantrine (AL) and dihydroartemisinin-piperaquine (DP). AL has been shown to be highly efficacious and well tolerated, and has become the most widely recommended first-line regimen in Africa [1, 2]. DP, a newer ACT, has shown excellent efficacy in multiple trials from Asia and Africa and is considered a highly promising drug for global deployment due to its simple once-a-day dosing and extended post-treatment prophylactic effect [3–7]. Although ACT has been shown to be highly efficacious, concerns remain about its safety and tolerability, especially in certain patient populations, such as young children and those with co-morbidities such as HIV.

There are a number of methodological challenges in evaluating the safety of anti-malarial drugs, such as establishing causality between drugs and adverse events in the setting of malaria. The World Health Organization (WHO) has established guidelines for evaluating the efficacy of anti-malarial drugs, but no similar guidelines exist for safety monitoring [8–10]. A lack of uniformly applied methods for adverse events reporting limits the ability to compare data from different studies. Additionally, most anti-malarial safety data come from clinical trials evaluating single episodes of malaria, while in practice African children are often treated repeatedly for recurring episodes of malaria over a relatively short period of time. Despite these challenges, comparative data on the safety and tolerability of different forms of ACT are critical for making informed policy decisions.

Results of a study comparing the efficacy AL and DP for the treatment of uncomplicated malaria using a longitudinal study design in a cohort of young children living in an area of very high malaria transmission intensity in Uganda were recently published [11]. This cohort included HIV-unexposed (HIV-uninfected children born to HIV-uninfected mothers), HIV-exposed (HIV-uninfected children born to HIV-infected mothers), and HIV-infected infants enrolled from six weeks to 12 months of age and followed up to one year. Results presented here include detailed data on the safety and tolerability of these therapies with follow-up extended for two additional months, in the context of repeated treatments, HIV co-infection, and the concomitant use of trimethoprim-sulphamethoxazole (TS) prophylaxis and anti-retroviral (ARV) therapy.

Results from this longitudinal randomized clinical trial suggest that both AL and DP are safe for treating uncomplicated malaria in young HIV-infected and uninfected infants and children. Adverse events were uncommon and generally of mild severity, with only cough, diarrhoea, vomiting, and anaemia occurring in more than 1% of treatments with study drugs. There were no significant differences in adverse events between the two treatment arms, although recent treatment with DP was associated with an increased risk of vomiting. Concomitant use of TS and ARVs were not associated with an increased risk of common adverse events.

This study contributes to a body of literature suggesting that AL and DP are both safe and well tolerated across a wide range of epidemiological settings. AL has been extensively studied in clinical trials primarily from Asia and Africa, and was added to the WHO Essential Medicines List in 2002. A review of AL safety and tolerability in 1,869 patients (33% children) from Asia reported gastro-intestinal complaints (nausea, vomiting, and diarrhoea), headache, and dizziness as the most commonly reported adverse events [16]. Several randomized clinical trials from Africa have reported adverse events associated with AL similar to those from Asia, with safety and tolerability profile found to be equivalent or superior to comparator regimens such as amodiaquine + sulphadoxine-pyrimethamine or amodiaquine + artesunate [17–22]. DP has also been shown to be very safe and well tolerated in studies from Asia and Africa. In a review of 2,600 patients from studies primarily done in Asia, gastrointestinal complaints were the most common adverse events associated with DP, with no evidence of severe drug toxicity [5]. In three clinical trials comparing AL (n = 597) versus DP (n = 613) in Africa using similar protocols, the risks of adverse events between these two treatment groups were similar [3, 6, 7]. In these studies a total of 13 serious adverse events (1.1%) were reported (AL = 4, DP = 9), and all were felt to be unrelated to the study drugs. The most severe adverse event in this trial, death due to presumed haemolytic anaemia, occurred in a child found to have G6PD deficiency. Haemolysis following treatment with DP in a child with G6PD deficiency has been described in a prior study from Laos [23]. However, it is unclear whether haemolytic anaemia is causally associated with DP or just a consequence of G6PD deficiency alone [23].

This study addresses some of the limitations from previously published studies on the safety and tolerability of AL and DP. First there is limited data on the safety of these drugs in children under 12 months of age, an important population in Africa. Study participants initiated ACT therapy at four months of age and over 50% of the treatments given for uncomplicated malaria were in children younger than 12 months. Both medications were well tolerated by infants, with diarrhoea the only adverse event associated with this younger age group, independent of the treatment given. Second, most anti-malarial clinical trials are limited to single episodes of uncomplicated malaria in the same patient, which precludes the ability to analyse the effect of repeated treatments. The longitudinal design used in this study allowed for the follow-up of children for up to one year, and observe the effects of repeated treatments of ACT drugs. Repeated therapy was found to be generally safe and well tolerated with the exception of a significantly higher risk of vomiting following repeat treatment with DP within 2-4 weeks of a previous dose. Although the extended half-life of piperaquine [24] provides a prolonged post-treatment prophylactic effect [3, 6, 7], repeat therapy with DP over a short period of time may increase the risk of adverse events and should be investigated further. This study also provides data on the safety and tolerability of ACT in unique patient populations including those taking TS prophylaxis and HIV-infected patients taking ARVs. Almost 40% of the patients were concomitantly taking TS when treated for malaria, however, this was not associated with an increased risk of adverse events. The concomitant use of ARVs and ACT is of greater concern because drug toxicities and the potential for drug interactions [25]. Artemether is metabolized via CYP3A4 to the more active compound dihydroartemisinin. Lumefantrine is also metabolized by CYP3A4. Less is known about the metabolism of piperaquine. There are three main classes of ARVs: protease inhibitors (PIs), non-nucleoside reverse transcriptase inhibitors (NNRTIs), and nucleoside reverse transcriptase inhibitors (NRTIs). PIs can be either potent inducers or inhibitors of CYP3A4. In a study of healthy subjects the PI lopinavir/ritonavir increased the lumefantrine area under the curve (AUC) by 193%, though no significant toxicities were reported [26]. NNRTIs can also be inducers or inhibitors of CYP3A4 with efavirenz primarily inhibiting CYP3A4 and nevirapine primarily inducing CYP3A4. Nucleoside reverse transcriptase inhibitors (NRTIs) are not thought to have clinical significant interactions with commonly used ACTs [25]. Prior studies have found an association between amodiaquine-based anti-malarial therapy and neutropaenia in HIV-infected children [27]. In this study, no association between the risk of adverse events and the concomitant use of ARVs was found. However, this study had a limited number of patients taking ARVs (n = 66), and only included NNRTI + NRTI containing regimens.

There were several limitations to this study. Firstly, although every effort was made to apply standardized definitions for adverse events, treatment assignment was not blinded and results were limited by largely subjective reports of symptoms from parents or guardians. Secondly, only haemoglobin levels were regularly measured during the period of malaria follow-up, preventing the detection other laboratory associated adverse events. Finally, this study was not powered to specifically test for hypothesis of differences in the risk of adverse events between the various subgroups. Therefore, the possibility of type II errors cannot be ruled out, especially in those subgroups with small samples sizes such as HIV-infected children taking ARVs.

In summary, both AL and DP were safe and well tolerated for the treatment of uncomplicated malaria in a cohort of children uniquely characterized by their young age, repeated therapy, and the inclusion of HIV-exposed and HIV-infected patients. However, the occurrence of haemolytic anaemia leading to death in a G6PD deficient patient highlights the importance of monitoring for rare, but serious adverse events. As the use of ACT is scaled up in Africa, continued evaluation of the safety and tolerability of these drugs in diverse patient populations is essential.