In 2010, the World Health Organization (WHO) estimated that 216 million cases of malaria occurred and were responsible of 650,000 deaths 1 . More and more people travel from northern countries to malaria-endemic areas. As a result, imported malaria is not exceptional in Europe, where approximately 7,000 cases were notified in 2010 to the WHO Regional Office for Europe 2 . The actual case numbers might be higher, due to under-notification. With its historical links with West and Central Africa, France is the leading European country in terms of cases of malaria diagnosed with 3,560 cases reported in 2011 by the National Malaria Reference Centre 3 . In Europe and other northern countries, travellers are treated by highly effective anti-malarial therapy, by contrast to malaria-endemic countries where appropriate treatments are not always available. A recent publication has shown that 18 different regimens were used in Europe, with the most prescribed being the fixed combination, atovaquone-proguanil (AP), commercially known as Malarone® 4 .

AP has been licensed in France since 1997 and marketed since late 2001 and it remains the principal treatment of acute malaria. Since the National Experts Committee recommended in 2007 that AP, together with artemether-lumefantrine, be used as first-line regimen, ‘old drugs’ such as quinine or mefloquine were downgraded to second-line treatment 5 . Atovaquone belongs to the family of hydroxy-naphthoquinones, which anti-malarial activity was first described 40 years ago 6 7 . Its mode of action is original, blocking the electron transport chain of the parasite’s mitochondria 8 . Used on its own, atovaquone has limited value, as shown by a significant relapse rate 9 . Its association with proguanil has shown excellent efficacy on acute malaria in numerous clinical trials, due to a synergistic effect 10 11 12 13 14 15 16 17 . AP is also widely used as an efficient and well-tolerated chemoprophylaxis for travellers.

Efficacy and tolerance of AP were extensively reported from clinical trials performed in malaria-endemic areas 10 11 12 13 14 . However, these results might not be valid in the case of imported malaria, due to epidemiological and biological differences (i e, study conditions, immune status, parasitaemia, heterogeneity of parasite strains, perception of side effects). Most of the studies comparing AP to other drugs were performed in endemic countries, and few observational or retrospective studies from non endemic countries have been published, amongst which only was a comparative trial, yet with a limited number of subjects has been published 18 19 20 21 22 23 24 . A recent international, prospective, observational study analysed a large cohort (504 cases) of imported Plasmodium falciparum malaria cases. It provided interesting data on the use of AP with a relatively large series (n = 253), but heterogeneity of practices, due to ‘centre effects’ between participating European centres may limit the interpretation of the study 4 .

The aim of this study was to bridge this gap in knowledge by providing data on the use of AP in ‘real life’ conditions in France, using a large series of uncomplicated imported cases of P. falciparum.

This ‘real life condition’ prospective, observational study of 553 patients treated by AP for uncomplicated P. falciparum malaria seems to be the largest series assessing the use of AP in the field of imported malaria. Patient profile (mostly young male adults of African origin living in Europe and infected in West Africa) are similar to those observed in the majority of studies on imported malaria 18 19 20 21 22 23 24 26 27 28 . The high percentage of HIV-infected patients (4.4%) observed in the study may be explained by the fact that the majority of patients are of African origin and because the nine recruiting centres are travel clinics linked to infectious diseases departments where a majority of the HIV patients living in Paris area are followed.

Not surprisingly, the rate of chemoprophylaxis and exposure prophylaxis was low. Clinical and biological presentation had no specificity but it is of interest to note that nearly one third of patients had no fever at admission, which might be misleading for non-experienced practitioners. Compared to some other studies describing non-comparative cohorts of malaria patients, follow-up, even too low, might be considered satisfactory given such a ‘real life’ design for this cohort, since outcome data are available for the majority at day 3 and since nearly half of patients were seen one month after treatment 16 19 20 21 26 27 28 29 . By contrast the loss to follow-up rate at one month in the recent European study was much lower, at 25% 4 .

With regard to efficacy, if a majority of patients (95%) were fever-free at day 3, nearly a third of them were still parasitaemic confirming that AP is slow-acting 15 16 19 20 30 31 . The analysis of efficacy in a per-protocol approach gives a cure rate of 99% (three relapses of 265 patients with a follow-up at one month) which is comparable to other treatments 15 16 22 32 33 . Details on relapses were available for only two cases. In both cases the reason was probably suboptimal plasmatic AP level: consecutive to obesity (115 kg) in one case and to a poor absorption in the second case, since the patient had not taken food with the drug 34 . As a consequence, physicians should re-assess AP dosage in obese patients and should insist on food intake with AP to optimize its absorption.

Even though rare under AP and comparable to other malaria treatment (1% in this study’s series), the risk of relapse has to be considered by physicians given the potential severe outcome at a time when the diagnosis of malaria may be omitted (long delay after travel in endemic area) 4 21 23 . Given that relapses occur usually between day 14 and day 30 (and in a few cases later) after treatment, physicians should organize a ‘recapture’ system for patients lost to follow-up after day 7, even limited to a phone call, in order to identify promptly a possible recrudescence 4 35 .

Perceived efficacy of AP was satisfying. This evaluation is not as valid as parasitological efficacy but it has not been studied in Africa before, both for curative treatment and for prophylaxis in travellers. It is considered as a significant predictor for compliance to treatment and preventative behaviours 36 37 .

With regard to tolerance of AP, data of this study highlighted a high rate of digestive ADR, mainly at day 3, especially nausea and vomiting. Digestive ADR represented 77% of ADR reported at day 3 and 72% of the totality of ADR reported during follow-up. It seems that digestive ADR were more frequent in backpackers compared to migrants, with no particular explanation, and no comparable data were found in the literature. In 18% of cases (n = 15), vomiting was severe enough to justify a change to second-line treatment. This high occurrence of digestive ADR could be partially explained by the population enrolled in the study since black people are known to have a lower clearance rate of atovaquone compared to white people 38 . Yet there were no relationship between immunity, or African origin, with digestive ADR.

Surprisingly, these digestive side effects are one of the most important from studies. A review of ten trials comparing Atovaquone-Proguanil (AP) with other anti-malarial drugs for uncomplicated malaria report a median rate of nausea and/or vomiting (inter quartile range) of 15.6% (5.2 – 25.0) for Atovaquone-Proguanil whereas other studies did not report this ADR 10 11 15 16 18 20 30 31 , 39 . To discriminate digestive ADR from symptoms related to malaria is difficult in a cohort and only clinical trials would be able to make the distinction. This study didn’t compare AP to other drugs, and this misclassification could be a bias. Nausea at diagnosis was not associated with digestive ADR at day 3, which involves AP rather than acute malaria in the etiology of these adverse effects.

Nevertheless, data from this series on digestive ADR are comparable to the results of a recent cohort study showing a higher switch rate to second-line treatment in patients treated with AP, compared to others 4 .

The design of this study does not enable to guide physicians on their choice of anti-malarial drug. However, the significant risk of vomiting associated with AP needs to be taken into account, especially in patients already complaining of nausea and vomiting at the time of diagnosis. The need for fat intake with AP, as is recommended to improve absorption, may be an additional risk factor for vomiting 38 . Prescription of metoclopramide is probably not a solution since it decreases the bioavailability of atovaquone 40 .

As observed in the literature, this study did not reveal any liver toxicity. Moderate variations in transaminase level observed at day 3 and 7 were not significant and were possibly due to malaria parasite itself 12 14 30 41 .

A drop in haemoglobin level, as observed here, was commonly reported after initiation of treatment of acute malaria due to malaria haemolysis 42 . The absence of correlation between haemoglobin at diagnosis and initial parasitaemia is in line with other studies 43 .

In 2007, French experts’ consensus recommended both AP and artemether-lumefantrine as first-line treatment for acute uncomplicated malaria 5 . As a consequence and because artemether-lumefantrine, even if registered throughout Europe in 1999, is only available since 2007 in France, the use of AP has progressively increased from 25% in 2006 to 46% in 2011 3 . On the one hand, AP’s good efficacy was confirmed by this series while its limitations were described (mainly slow-acting drug, poor absorption, and ADR, such as vomiting). On the other hand, use of artemether-lumefantrine in the particular situation of imported malaria still has a limited experience. Hence, regarding imported malaria, the question of which association between AP or artemisinin combination therapy (ACT) is the best option remains. A clinical trial is currently comparing the use of AP and ACT in the indication of uncomplicated malaria in non-endemic areas and will hopefully bring the answer 44 .

With the introduction of ACT in Europe (artemether-lumefantrine and, since mid-2012, dihydroartemisinin-piperaquine are both authorized in France and in a limited number of European countries) for uncomplicated acute malaria onset, the use of AP will probably decrease in favour of ACT, due to its prompt efficacy and good tolerance.

This observational series of 553 cases, the largest to date in number of patients, describes a large experience in using AP for imported uncomplicated malaria in real life conditions. Despite a non-comparative design, it appears that its efficacy is good and comparable to other similar drugs. The study confirms that AP is a valuable treatment option, while ADR, such as vomiting and its limited absorption in some cases, may be a limitation for its use.

The authors have declared that they have no competing interests.

HC, PR and OB participated in the design of the study and performed the statistical analysis. PR and OB conceived of the study, and participated in its coordination. HC, JC, SM and OB contributed to the draft manuscript. OF, PB, PC, HG, PHC, NG and MB participated in the recruitment of included patients. All authors read and approved the final manuscript.