A randomized trial on effectiveness of artemether-lumefantrine versus artesunate plus amodiaquine for unsupervised treatment of uncomplicated Plasmodium falciparum malaria in Ghanaian children

Artemisinin-based combination therapy (ACT) is advocated as the way forward in antimalarial treatment to overcome the global spread of Plasmodium falciparum drug resistance [1]. ACT rapidly diminishes parasite biomass leading to clinical and parasitological cure while at the same time gametocytocidal activity might be able to reduce overall malaria transmission [2]. Initially, efficacy and safety of ACT had been evaluated in Southeast Asia [3], and only recently, data became available from randomized clinical trials in sub-Saharan Africa, where the highest malaria burden occurs in children under five years of age [4]. Based on recommendations of the World Health Organization (WHO), several sub-Saharan countries introduced ACT as first-line treatment for uncomplicated malaria, although the comparatively high costs of a treatment dose are a burden to their health systems [5]. Currently, artemether-lumefantrine (AL) and artesunate plus amodiaquine (ASAQ) are the only widely available drugs, which are produced on a large scale complying with good manufacturing practices (GMP).

In 2004 Ghana implemented ASAQ as national first-line treatment for uncomplicated P. falciparum malaria, after experts had reviewed data on efficacy of chloroquine (CQ) indicating widespread CQ resistance throughout the country [6]. At that time, limited data on the efficacy and safety of ASAQ and AL were available from sub-Saharan Africa [7, 8]. Subsequently, overdosing of locally produced prepacked ASAQ for use in adults resulted in severe toxicity in several cases and fuelled the call for alternative drugs with better tolerability, e.g. assumed for AL. Clinical trials in Uganda showed high efficacy and good tolerability of AL when used as a six-dose therapy for uncomplicated malaria in children, with a risk of treatment failure due to recrudescence of less than 2% [9, 10]. The high cure rates of AL were confirmed in studies from Tanzania [11‐13], suggesting that this combination is still effective in areas with considerable resistance to other antimalarial drugs.

However, despite the promising results of AL, there are substantial limitations of this regimen, including the twice-daily dosing schedule, the preferable administration with fatty food, and selection for drug-resistant parasites. Although adequate clinical and parasitological responses in Ugandan children following unsupervised treatment were similar to supervised treatment, other studies showed that unsupervised therapy with AL resulted in lower plasma levels of lumefantrine with an increased risk for early reinfection [9, 10].

The aim of this study was to compare the effectiveness, tolerability and acceptance of unsupervised AL to that of unsupervised ASAQ in an area of intense malaria transmission in Ghana.

In this randomized effectiveness trial from Ghana cure rates of ASAQ and AL for the treatment of uncomplicated falciparum malaria were around 90%, qualifying both ACTs as highly effective in children. LCFs until day 28 uncorrected and, as a trend, corrected for reinfections were more frequently observed in AL-treated children than in ASAQ-treated individuals. This observation of similar ACPRs but higher LCFs after treatment with AL, could reflect a difference in the prevention of recurrent clinical malaria episodes between the two ACTs.

One limitation of the study is its premature termination due to the high frequency of anemia cases (>50%). The discontinuation was necessary through the stopping rules of the study protocol despite the fact that the occurrence of anemia was not different in both study arms and was not related to treatment. Accordingly, the analysis was performed with a decreased number of study participants providing a power of 65% to detect differences between the study arms. Therefore, the lack of a significantly different effectiveness in one of the study arms has to be interpreted with caution. Nevertheless, cure rates could be provided within the given confidence intervals and the findings of different LCFs in both arms are valid independent of the estimated power.

Another limitation of the study is the restriction of the follow-up period to 28 days bearing the risk of underestimation of treatment failures. Although logistically difficult, only longitudinal clinical trials with follow-up time over several months are appropriate for comparing the benefits and risks of different antimalarial drugs, especially in highly endemic areas [20]. Recently, an efficacy trial comparing ASAQ and AL was performed in Ghana with a follow-up time of one whole year showing high day-28 ACPRs of 95.3% and 94.2%, respectively, with no difference in the occurrence of reinfections or adverse events [21]. However, the differentiation between reinfection and resistance is even more difficult with increasing follow-up periods.

With respect to the interpretation of PCR-controlled cure rates, it has to be considered, that the definition of reinfections is critical for the assessment of drug efficacy and effectiveness. In the study presented here, a recrudescence was defined as a recurrent P. falciparum parasitaemia with detection of at least one msp1/msp2 amplicon of identical length within the 28-day follow-up period compared to amplicons found on day 0. This definition was chosen very stringent to identify all possible treatment failures. Remarkably, although most infections in our study area are multiclonal and the genetic diversity is high [14], some amplicons occur at high frequency. In principle, this increases the likelihood that amplicons with identical lengths were found in reinfections, leading to an overestimation of recrudescences and, consequently, treatment failures. Recently, the pitfalls of different genotyping approaches in antimalarial drug trials were analysed in relation to the transmission intensity [22]. The authors concluded that, especially regarding high endemicity settings, methods to overcome the inaccuracy of genotyping to distinguish recrudescenses from new infections were limited and complex, and comparative results should be interpreted, rather to rely on the absolute efficacy of a respective drug. But not only treatment efficacy, also post-treatment prophylaxis is of major concern in areas of intense malaria transmission, where timely recurrent symptomatic malaria episodes require re-treatment, and subsequently influence the acquisition of immunity and increase the risk of development of drug resistance [23].

In accordance with these assumptions and in addition to the study mentioned above a recent comparative efficacy study in Ghanaian children showed lower parasitological and clinical failure rates in the ASAQ compared to the AL arm [24]. Another trial from Ghana showed a high PCR-corrected efficacy of ASAQ and AL of 100% and 97.5%, respectively [25]. Together with recently published efficacy data the presented study shows that ASAQ is highly efficacious, if not superior to AL, in treating uncomplicated malaria in Ghana.

These results are in contrast to most comparative efficacy trials from Central and East Africa. In the Republic of Congo, in Uganda and Tanzania including Zanzibar, AL showed higher efficacy than ASAQ, especially a higher potency in preventing reinfections [11‐13, 26‐28]. In one study from an area with holoendemic malaria transmission in Uganda, both regimens, ASAQ and AL, were highly efficacious for clearance of infections but the risk of reinfection by day 42 was found to be high in both study arms. In that study, although initial cure rates were above 90%, 66% of ASAQ-treated and only 50% of AL-treated patients developed recurrent parasitaemia within 28 days [28]. Only in Nigeria, one of a few West-African countries from where data is available, the efficacies of ASAQ and AL after supervised treatment were similarly high, with PCR-corrected cure rates of 98.4% and 100%, respectively [29], while uncorrected ACPRs were 87.0% and 82.5%, respectively [30].

To explain the observed differences, one has to consider that not only the prevalences of molecular markers for ASAQ and AL drug resistance vary in different geographical regions [31, 32], but also other parasite characteristics, the genetic background of the host and multiple environmental factors.

The lower effectiveness of AL in our trial compared to previous studies might be explained by unreliable or incorrect drug intake when unobserved. Furthermore, intestinal absorption of lumefantrine, the long-acting partner drug in AL, can be critically influenced by concomitant intake of fat-containing food [33]. Nevertheless, the only other comparative effectiveness study performed in Tanzania showed superiority of AL over ASAQ [11]. In the present study, a fat-containing diet was emphatically advised throughout, yet it forms a possible source of treatment failure due to sub-therapeutic plasma drug concentrations. AL tablets need to be taken twice daily, with equal number of tablets per dose. Therefore, skipping single doses, not taking them entirely, or not coadministering fat-containing food could also have resulted in a higher frequency of subtherapeutic drug levels in the AL arm. Nevertheless, good compliance with AL treatment seems to be possible, as shown in a trial on home management of malaria in Ghana [34]. In Gabon, the outcome after once-daily ASAQ treatment was significantly improved by direct observation [35]. In this context, it seems reasonable to hypothesize that the fixed-dose combination of artesunate plus amodiaquine, which is currently replacing the co-blister packed ASAQ, will further increase therapeutic adherence and treatment effectiveness. Indeed, the acceptance of both ACTs was high and only few caregivers reported general problems with the study medications.

The trial was discontinued due to a high frequency of anemia cases in our study population. Nevertheless, anemia occurred equally often in both study arms and haematologic recovery was similar and adequate in both arms. The safety profiles of the study drugs were good and our findings are in agreement with previous drug safety monitorings in Senegal [36, 37]. It is noteworthy that six ASAQ-treated children and only three children in the AL arm had to be excluded after repeated vomiting of the study drug. Though not significant, this could be an indicator of a poorer tolerability of ASAQ, which has also been shown in a recently published trial from Ghana [24]. Pharmacokinetics after combined administration of artesunate and amodiaquine, unlike the extensive investigations of AL, were just recently systematically tested in fifteen healthy volunteers [38]. Drug interactions, frequency and nature of adverse events, namely severe transaminitis (n = 1), neutropenia (n = 2), and hypersensitivity (n = 1), were called a matter of concern. The risk of ASAQ-induced neutropenia was reported to be increased in HIV-positive individuals, especially when concomitantly treated with antiretrovirals [39]. Therefore, despite its high efficacy, ASAQ for combination therapy has to be carefully monitored in areas with high HIV prevalence.

ASAQ and AL were both well tolerated and highly effective for treatment of Ghanaian children with uncomplicated falciparum malaria. Together with recent efficacy data, the more effective prevention of LCFs after unobserved ASAQ compared to AL treatment, affirms its appropriate role as national first-line therapy in Ghana. Monitoring of efficacy, safety, and drug resistance markers are essential to guide national treatment policies, however, further comparative effectiveness studies of ACTs are needed in the West African region.