Background

Malaria is one of the most significant causes of morbidity and mortality worldwide, causing approximately 881,000 deaths every year [1]. The current WHO guidelines for the treatment of malaria recommend the use of artemisinin-based combination therapy (ACT) owing to the rising threat of Plasmodium falciparum resistance to monotherapy [2]. Artemether/lumefantrine (AL) was the first fixed-dose combination of ACT to be approved by the European regulatory authorities according to the requirements of the International Committee on Harmonization (ICH). AL is also pre-qualified by WHO for efficacy, safety and quality and is marketed as Coartem® for use as a six-dose regimen in infants, children and adults with acute, uncomplicated infection due to P. falciparum or mixed infections including P. falciparum [3, 4].

The efficacy of AL has been confirmed in many different patient populations around the world and is discussed elsewhere in this supplement [5]. This article reviews the comprehensive safety data on AL that has been reported in the scientific literature from Novartis-sponsored and independent clinical trials.

Novartis-sponsored studies

Six Novartis-sponsored clinical studies [6–11] have been conducted to assess efficacy and safety of the six-dose regimen of AL, with safety data from these summarized in Table 1. These studies were conducted in different regions of the world and so included patients who lived in areas with varying levels of drug resistant P. falciparum and malaria endemicity. Some of the studies included other anti-malarial drugs as active comparators [7, 8], one allowed the inclusion of patients with mixed infections that included P. falciparum [10], and one utilized a dispersible paediatric formulation [11].

Table 1 Key clinical studies evaluating the safety of the six-dose regimen of AL.
Full size table

In all of these studies AL was well tolerated and most reported adverse events were of mild or moderate severity. Reported adverse events occurred mainly in the nervous system or gastrointestinal system and were considered typical of the symptomatology of malaria or concomitant infections. No cardiac, neurological or audiological safety issues were identified in the studies that assessed these safety parameters [6–11].

Comparative results from independent studies

Many independent trials on AL have also been published in the scientific literature. Results from the studies conducted in Africa [12–32] are summarized in Table 2 and the results from studies conducted in Asia [33–39] in Table 3. The safety findings from these trials confirm the results of the Novartis-sponsored studies, namely that the majority of adverse events were mild or moderate in severity, tended to affect the gastrointestinal or nervous system, and were consistent with the symptoms of malaria. Serious adverse events were infrequent and were unrelated or unlikely to be related to study medication.

Table 2 Published studies with the six-dose AL regimen in Africa
Full size table
Table 3 Published studies with the six-dose AL regimen in Asia
Full size table

Data have also been pooled to provide a more complete assessment of the safety and tolerability of AL in a large number (>1,900) of adult and paediatric patients. The results of two of these meta-analyses, Mueller et al and Makanga et al are shown in Table 4[40, 41]. These pooled results confirm earlier data and show that AL is generally well tolerated with the majority of reported adverse events again affecting either the gastrointestinal or nervous systems and of mild or moderate severity [40, 41]. Few adverse events were considered by the investigators to be drug-related - most were regarded as being related to the symptomatology of malaria. Serious adverse events occurred infrequently [40, 41]. There were, however, some differences in the adverse event profile in children when different body weights were compared, but this is unlikely to be of clinical relevance. Instead, it is more likely to be related to the subjective nature of these adverse events and the differences in the ability of young infants and children to verbalize these complaints [41].

Table 4 Pooled analyses of data on the six-dose AL regimen
Full size table

Special safety considerations

Neurological and audiological safety

There have been case reports of neurological problems (including ataxia, nystagmus, tremor and slurred speech) occurring after administration of herbal artemisinin [42] or artesunate monotherapy [43, 44]. However, it is questionable whether these neurological effects are related to artemisinin treatment [45–47]. Indeed, this is not supported by audiology case-control [48–50] or other audiology data [14, 51], although additional analyses are often called for to investigate this fully. There were, however, no neurological or audiological safety concerns identified for AL in the Novartis-sponsored studies [6–10, 52] as shown in Table 1. In addition, audiological data are currently being analysed from a Novartis-sponsored study in 265 patients treated with AL, malarone or mefloquine-artesunate. [Personal communication with Novartis, study A2417). A routine review of the data collected from the first 85 patients, however, did not raise any safety concerns. Neurological or audiological safety concerns have not been identified in the African studies by Bukirwa et al [12], Gürkov et al [14], and Adjei et al [20], as shown in Table 2, Asian studies by Krudsood et al [33], and Mayxay et al [35], as shown in Table 3, and in the pooled analyses reported by Mueller et al [40] or Makanga et al [41], which are summarized in Table 4.

Cardiac safety

Lumefantrine is chemically related to halofantrine, an anti-malarial known to be associated with significant prolongation of QTc interval. Indeed, QTc prolongation is a known class effect of many anti-malarial drugs. As such, cardiac safety has been thoroughly investigated during the preclinical and clinical development of AL.

The effects of lumefantrine and its major metabolite desbutyl-lumefantrine on wild-type hERG K+ channels have been investigated in stably transfected human embryonic kidney cells (HEK293) using a whole cell patch-clamp technique [52]. This in vitro hERG channel assay showed that lumefantrine and desbutyl-lumefantrine have higher IC50 values (approximately 200-fold) than halofantrine (see Table 5). In addition, the calculated cardiac safety indices, which are over 30 for lumefantrine, suggest that lumefantrine and its major metabolite pose an unlikely risk of cardiotoxocity compared with halofantrine [52]. A phase I, parallel group study with AL, moxifloxacin and placebo arms (n = 42 per group) was conducted in healthy volunteers to assess cardiac safety [52]. As shown in Figure 1, this study demonstrated that the six-dose regimen of AL was associated with a mean maximum increase in QTc (Fridericia's formula) of 7.45 msec compared with placebo (90% CI: 4.4 to 10.5 msec) in healthy volunteers. By comparison, the effect on QTc (Fridericia's formula) was much greater after dosing with moxifloxacin, the positive control. A formal analysis of the relationship between the plasma concentration of lumefantrine and the change from mean baseline in QTc (Fridericia's formula) was conducted (see Figure 2). This shows that the 95% confidence limit for the mean lumefantrine plasma concentration does not cross the upper confidence band for the threshold of relevance for change in QTc (Fridericia's formula).

Table 5 In vitro IC50 for hERG tail current in HEK 293 cells and cardiac safety index of antimalaria drugs [52]
Full size table
Figure 1
figure 1

Placebo-corrected mean change in QTc (Fridericia's formula) from time-averaged baseline in healthy volunteers. [Data on file (study A2101), Novartis]. AL = artemether/lumefantrine; CI = confidence interval.

Full size image
Figure 2
figure 2

Relationship between QTc (Fridericia's formula) and maximum plasma concentration (C max ) in healthy volunteers. [Data on file (study A2101), Novartis]. Note: the threshold of relevance for QTc change is 10 msec vs. placebo.

Full size image

One possible explanation for the alterations in QTc seen with AL may be due to malaria itself and recovery from this disease, as well as to stress and anaemia. These conditions can affect cardiac electrophysiology and may lengthen the QT interval [53]. As QT correction formulae are based on a normal heart rate of 60 beats/minute, and patients with malaria tend to have elevated heart rates that decrease with successful treatment and defervescence, overcorrection of the QT interval can also occur. This is even more pronounced in small children, and heart rates higher than 60 beats/minute are routinely seen in healthy children. Despite the changes in QTc associated with malaria, no adverse events attributable to QTc prolongation (e.g. syncope or sudden death) have been reported in clinical trials with AL [6] as shown in Table 1[6–11], Table 3[39], and Table 4[40, 41].

Haematological safety

Preclinical data suggested that repeated exposure to AL may affect blood cell counts and as such, particular attention was paid to haemoglobin levels and haematological adverse events reported during clinical trials. A published pooled analysis of 15 trials conducted in China, India, Thailand, The Gambia, Tanzania, France, The Netherlands and the UK confirmed these results [54]. Indeed, in this analysis, anaemia and thrombocytopaenia were frequently present at baseline, but returned to normal or improved considerably with the resolution of disease [54]. No haematological safety concerns were identified during the safety assessment presented in this review.

Safety of repeated administration

Most safety data on antimalarials are collected as part of clinical trials and these tend to evaluate the treatment of a single episode of malaria. In practice, however, children in areas of high malaria endemicity are likely to suffer from repeated episodes of malaria and this raises concerns over potential toxicity due to repeated short-term exposure to the drugs.

Two longitudinal studies have therefore attempted to address this question by assessing the efficacy, safety and tolerability of amodiaquine-artesunate versus AL [55] and amodiaquine plus sulphadoxine-pyrimethamine versus artesunate-amodiaquine versus AL [56] after repeated use.

These studies in Ghanaian and Ugandan children indicated all treatments were efficacious, but the various treatments were associated with slightly different adverse event profiles. For example, treatment with amodiaquine plus sulphadoxine-pyrimethamine had a greater risk compared with AL of anorexia, weakness and subjective fever and a greater risk compared with artesunate-amodiaquine of weakness and subjective fever. Children treated with AL, however, had a higher risk of elevated temperature than those receiving artesunate-amodiaquine [56]. In both studies, repeated administration of AL did not seem to be associated with an increased risk of adverse drug reactions. The study in Ghanaian children also included neurological examinations for children treated with repeated artemisinin-based regimens. This did not identify any abnormal neurological signs that were considered related to artemisinin [55].

Comparison with other anti-malarial therapies

The safety of the six-dose regimen of AL has been compared with MAS in two Novartis-sponsored, open-label randomized trials [7, 8, 52], as summarized in Table 1. In these two studies, 314 patients received AL and 105 patients received MAS. Overall, AL was at least as well tolerated as MAS, and had a similar safety profile (Table 6) [52]. Only one patient, who developed urticarial rash, experienced a serious adverse event that was thought to be treatment-related [52]. Independent studies have also compared AL with other anti-malarial therapies as shown in Table 2, Table 3 and Table 4[12–40]. These studies showed that AL was at least as well tolerated as the various comparators (AQSP, ASAQ, ASCD, ASSP, AS plus sulphamethoxypyrazine plus pyrimethamine, atovaquone/proguanil, CQ, CQSP, DP, DNP, MAS, quinine, or SP) used in clinical trials.

Table 6 Most frequently reported adverse events that occurred at a frequency of >10% in either group of patients enrolled in comparator studies [7, 8, 52]
Full size table

Safety in pregnancy

Pregnant women with symptomatic malaria, and especially those in their second and third trimesters, are more likely to develop severe malaria than other women [2]. In addition, this is often complicated by pulmonary oedema and hypoglycaemia [2]. Prompt, safe, and effective treatment is, therefore, recommended in this high risk group to reduce maternal mortality, foetal death and premature labour [2]. ACT can be used in uncomplicated malaria in the second and third trimester. The use of ACT in the first trimester should, however, only be considered if they are the only effective treatment available and the benefits outweigh the risks of treatment [2, 4]. In severe malaria, ACT is preferred over quinine due to the hypoglycaemia associated with quinine. Until recently there has been limited evidence available to support the use of ACT in early pregnancy; however, McGready et al published data on the use of AL in pregnancy in 2009 [57]. These data were from a prospective eight-year study in Thailand on pregnant women exposed to artemisinin antimalarials. In this study, birth outcomes did not differ significantly from community rates of congenital malformations and stillbirth [57]. In addition, data from a prospective observational study to compare the safety of AL and SP in pregnant women treated for symptomatic falciparum malaria was presented at ASTMH Annual Meeting in December 2008 [58]. This study analysed data from 1,001 pregnant women (AL: n = 495; SP: n = 506) and their foetuses/newborns (AL: n = 470; SP: n = 477). There were no clinically relevant differences in perinatal mortality, neonatal mortality, still births, preterm deliveries, gestational age-adjusted low birth weight or birth defects between the two groups.

Post-marketing experience

As of July 2009, 250 million AL treatments (70% of which were for children) have been delivered to malaria-endemic countries. Post-marketing experience has not identified any new specific safety concerns apart from hypersensitivity and skin reactions (allergies), a recognized class effect for artemisinin derivatives [2].

Conclusion

This review summarizes some of the safety and tolerability data on AL (Coartem®) from published Novartis-sponsored and independently-sponsored clinical trials. Data are available to support the use of a six-dose regimen of AL as a safe and well-tolerated treatment for P. falciparum malaria or malaria due to mixed infection including P. falciparum in adults, adolescents, children and infants. Indeed, the safety profile of this drug is similar in both adults and children, and many of the reported adverse events are typical of symptoms of malaria or concomitant infections that are commonly seen in these patient populations. Reported adverse events are mainly of mild or moderate severity, with very few serious adverse events reported and even fewer serious adverse events considered related to treatment. In addition, no notable neurological adverse events or cardiac safety concerns were identified in the many studies that examined the safety of AL. As patients (and especially children) often take repeated courses of treatment for malaria, it is reassuring to note that data are available that show repeated administration of AL is not associated with an increased risk of adverse drug reactions, in particular with regard to neurological adverse events. In addition, a thorough review of the clinical data for AL does not identify any potential cardiac safety issues. Data are also available to show that there were no clinically relevant differences in perinatal mortality, neonatal mortality, still birth, pre-term delivery, low birth weight and spontaneous abortion in women exposed to AL compared with SP during pregnancy.

In conclusion, AL is a safe and well-tolerated treatment for P. falciparum infections that should help fight malaria. Indeed, as of July 2009, 250 million AL (Coartem®) treatments have been delivered and post-marketing experience confirms the favourable safety and tolerability profile for this drug. In addition, this wealth of safety evidence has not identified any safety concerns for AL apart from rare type 1 hypersensitivity reactions, a recognized class effect of artemisinin derivatives [2].