Declining Artesunate-Mefloquine Efficacy against Falciparum Malaria on the Cambodia–Thailand Border

Emerging resistance in Southeast Asia raises concern over possible spread or similar evolution of resistance to other artemisinin-based combination therapies in Africa.


Historical Perspective
The western Cambodia-southeastern Thailand border, which comprises the areas around the town of Pailin in Cambodia and the provinces of Trat and Chanthaburi in Thailand, has been an epicenter of drug-resistant malaria (3) (Figure). Resistance to chloroquine and sulfadoxinepyrimethamine occurred there in the late 1950s and 1960s, respectively. Mefl oquine was introduced there in 1983, fi rst on the Thai side and in the form of mefl oquine in combination with sulfadoxine-pyrimethamine. Mefl oquine resistance led to replacement by artesunate-mefl oquine in Thailand in 1995. In Chanthaburi and Trat Provinces, considered by the Thai national malaria control program to be the areas with the highest level of mefl oquine resistance, the dosages were 12 mg/kg artesunate and 25 mg/kg mefl oquine or a maximum adult dose of 600 mg artesunate and 1,250 mg mefl oquine, given for 2 days. This regimen was 99% effi cacious when fi eld tested in the same border areas in 1993 (4). It also ensured better compliance than an extended (3-day) regimen. With the exception of the southeastern border of Thailand with Cambodia, the regimen remained effective throughout Thailand even when the control program switched to a 3-day treatment course in 2007 in accordance with the World Health Organization (WHO) recommendation.
In Cambodia in 2000, artesunate-mefl oquine became the fi rst-line drug for the treatment of falciparum malaria. The dosages were 12 mg/kg artesunate and 20 mg/kg mefl oquine, or a maximum adult dose of 600 mg artesunate and 1,000 mg mefl oquine, given for 3 days, in accordance with the WHO recommendation. The lower mefl oquine dose was based on the perception that Cambodians were slightly smaller than Thais in body build. Malaria control programs of each country regularly monitor antimalarial therapeutic effi cacy at selected sentinel sites. of response to combination artesunate (≈12 mg/kg) and mefl oquine (≈20 mg/kg) co-blister packs (artesunate for 3 days and mefl oquine for 1 day) showed 85.7% effi cacy at day 28 follow-up (5) (Table). A repeat study in Pailin in 2004, which used the same drug combination but more precise dosing and follow-up at 42 days, found effi cacy to be 79.3% (5). To exclude cases of reinfection from analysis, parasite variants were identifi ed by using nested PCR amplifi cation of 3 polymorphic genes for merozoite surface protein 1 (msp1), msp2, and glutamate-rich protein (7). In Thailand's Trat Province in 2003, effi cacy of 78.6% (95% confi dence interval 66.4%-91.1%) was reported from a 28day follow-up study of 44 patients who received the same total dosage of this combination in a 2-day regimen (6). Each of the 3 studies used directly observed therapy, followed the WHO standard in vivo study protocol, and obtained from their respective national control programs reliable drugs that worked effectively at other sentinel sites. Although each of the 3 studies was small, in aggregate they attest to a worrisome level of in vivo artesunate-mefl oquine resistance.

Resistance to Artesunate-Mefl oquine
The above clinical observations are supported by molecular evidence. High copy numbers of the Plasmodium falciparum multidrug resistance 1 (pfmdr1) gene, a marker of multidrug resistance (8), predicted recrudescence in the 2004 Pailin study, even after PCR correction and adjustment for age and parasite density (9). Thus, clinical and molecular evidence indicate that artesunate-mefl oquine failures are occurring on the Cambodia-Thailand border. In contrast, artesunate-mefl oquine remains effective in eastern Cambodia and elsewhere in Thailand.
In fi eld situations, artesunate-mefl oquine failure is likely to be worse than in controlled studies because of poorer compliance and variations in drug quality. Some of the results presented above might have overestimated artesunate-mefl oquine therapeutic effi cacy because 1) 28day follow-up is inadequate time to evaluate drugs with a long elimination half-life, such as mefl oquine for which up to 60% of recrudescence may occur between day 28 and day 42 (9), and 2) the current genotyping method is likely to misclassify some recrudescence as reinfection because of the method's limited ability to detect minor subpopulations of parasites that carry drug resistance mutations (J. J. Juliano, pers. comm.). In our opinion, PCR correction to classify recurrent infections is not necessary for a lowtransmission area such as the Cambodia-Thailand border, where infection frequently results from occupational exposure in the jungles and patients usually remain at low risk for reinfection while in the village during study follow-up.
These treatment failures most likely result from mefl oquine resistance rather than artemisinin resistance because increased pfmdr1 copy number has been linked most closely with mefl oquine resistance in vivo (10,11). Monitoring of in vitro antimalaria drug susceptibility by the Pasteur Institute of Cambodia since 2001 also suggests progressive loss of mefl oquine sensitivity in western Cambodia (2). Most likely, mefl oquine resistance in this area had already reached a level too extreme for the drug to be further protected by artesunate. To some extent, the development of resistance to artesunate-mefl oquine in areas of preexist-  *ACT, artemisinin-based combination therapy; ATS, artesunate; MFQ, mefloquine; day 0, first 24 h of enrollment and start of therapy. †Also in this study, increased copy numbers of Plasmodium falciparum multidrug resistance 1 gene were found to be associated with parasite recrudescence, and as many as 44% of patients did not clear parasites until after 48 hours.
ing mefl oquine resistance mirrors the experience with the mefl oquine-sulfadoxine-pyrimethamine combination that was introduced to deal with widespread preexisting sulfadoxine-pyrimethamine resistance (3). Fortunately, no clear evidence points to artesunate failure, although it is possible. No mutations thought to be associated with artemisinin resistance have been detected in Cambodia, and no isolates resistant to artesunate in vitro have been found (12). However, the increased parasite clearance times recorded in Pailin are worrisome. In the Pailin 2004 study, as many as 34% of patients cleared parasites between 48 and 72 hours, and 10% after 72 hours (9). When artesunate was fi rst introduced in the early 1990s, failure to clear parasites by 48 hours was rare among patients with uncomplicated falciparum malaria in southeastern Thailand. Additionally, increased copy numbers of pfmdr1 gene have recently been shown to be associated with in vitro resistance to artemisinin (and to lumefantrine and quinine) (13).

Likely Resistance Factors
Several reasons may explain the emergence of artesunate-mefl oquine resistance on the Cambodia-Thailand border. First, the concept of rational therapy is poorly reinforced. Improper use of antimalaria drugs based on clinical diagnosis alone or on misdiagnosis as a result of poor microscopy technique or interpretation could have accelerated the onset of resistance. In Cambodia, because of poor transportation and public health infrastructure, artesunate and mefl oquine are made available in the private sector to increase patients' access to the drug. This access, in turn, increases the risks that drug quality will be substandard and drug use will be uncontrolled (5). Because adherence and indication are not adequately emphasized, drugs are consumed in incomplete dosages or for prophylaxis such as before a jungle trip. Social marketing helps to control drug quality but cannot ensure adherence.
Although ACT use has been restricted and prescription is based on microscopic confi rmation at government-run malaria clinics in Thailand, ACT is less well-controlled in Cambodia. Unreliable services and poor diagnostic capabilities at peripheral health facilities further discourage patients from seeking malaria treatment from the public sector and encourage self-purchase of drugs. A recent study of malaria treatment-seeking behavior in Cambodia showed that >80% of the patients initially sought treatment from private providers and pharmacies or consumed drugs on their own (14).
Second, the short half-life of artesunate relative to that of mefl oquine means that tolerance to mefl oquine could develop when treated patients are reinfected (15). Third, the malaria parasites in this region could have a unique ability to develop resistance to any antimalaria drugs (16); their genetics need to be further studied.

Possible Development of New Foci of Resistance
As of 2007, reduced effi cacy of artesunate-mefl oquine was noted in Kampot, a province southwest of Phnom Penh (17). Thus, resistant parasites may be spreading. Although no other ACT is immediately ready for fi eld use, the Greater Mekong subregion will soon need alternatives to artesunate-mefl oquine.
In sub-Saharan Africa, presumptive treatment with ACT may soon become the norm; this drug-use practice may similarly promote evolution of resistance. ACT resistance in Africa could be devastating. This concern has been raised repeatedly (18,19). Although artesunate-mefl oquine is not recommended for African countries, 1 of the ACTs now used is the combination of artemether and lumefantrine (Coartem; Novartis, Basel, Switzerland); lumefantrine is chemically related to mefl oquine. One way to delay the emergence of resistance would be to enforce ACT prescription based on accurate biological diagnosis.

Conclusions
Studies are now under way to replicate these initial fi ndings on the Cambodia-Thailand border (20). Nevertheless, existing data strongly suggest that artesunatemefl oquine resistance exists in this area. This fi nding is a warning message for Africa, where ACT has been used in a large scale but not with parallel effort to enhance rational therapy. Continued surveillance for ACT resistance should be an integral part of any malaria control program that uses these drugs. Dr Wongsrichanalai is an epidemiologist with the US Naval Medical Research Unit No. 2 in Jakarta and Phnom Penh. Her research interests are drug-resistant malaria, malaria diagnosis, and emerging infections.
Dr Meshnick is professor of epidemiology at the University of North Carolina School of Public Health. His research interests are malaria in pregnant women and development of molecular tools to understand the epidemiology of microbial drug resistance.