Qinghaosu (Artemisinin): The Price of Success

See allHide authors and affiliations

Science  18 Apr 2008:
Vol. 320, Issue 5874, pp. 330-334
DOI: 10.1126/science.1155165


Artemisinin and its derivatives have become essential components of antimalarial treatment. These plant-derived peroxides are unique among antimalarial drugs in killing the young intraerythrocytic malaria parasites, thereby preventing their development to more pathological mature stages. This results in rapid clinical and parasitological responses to treatment and life-saving benefit in severe malaria. Artemisinin combination treatments (ACTs) are now first-line drugs for uncomplicated falciparum malaria, but access to ACTs is still limited in most malaria-endemic countries. Improved agricultural practices, selection of high-yielding hybrids, microbial production, and the development of synthetic peroxides will lower prices. A global subsidy would make these drugs more affordable and available. ACTs are central to current malaria elimination initiatives, but there are concerns that tolerance to artemisinins may be emerging in Cambodia.

In the fourth decade of the 17 century, Jesuits brought the bark of a Peruvian tree (later named Cinchona) to Europe. This provided for the first time a specific remedy for agues, periodic fevers that were prevalent throughout much of the continent, particularly in and around marshy areas (mal-aria or bad air). In 1880, Alphonse Laveran identified the intraerythrocytic protozoan parasite that caused malaria. The Cinchona alkaloids (quinine, quinidine, cinchonidine, and cinchonine) were shown to arrest the development of the malignant tertian malaria parasites (Plasmodium falciparum) at the mature trophozoite stage (after the first third of their 2-day intraerythrocytic life cycle) and thereby prevent their multiplication in the red blood cells (Fig. 1). Today, the Cinchona alkaloids are giving way to the products of a ubiquitous annual shrub (Artemesia annua, or huang hua hao, but often called qinghao) (Fig. 2).

Fig. 1.

The intraerythrocytic life cycle of P. falciparum. Parasitized red cells circulate for the first third of the 48-hour cycle and then sequester in capillaries and venules. Artemisinins inhibit development of a broader age range of the parasites than do quinine and other antimalarial drugs. The effect on the young rings prevents their development to the more pathological mature parasites that sequester.

Fig. 2.

Commercial cultivation of Artemisia annua (qinghao/huang hua hao) in China.

The antimalarial properties of the traditional Chinese medicine qinghaosu (artemisinin) were discovered by Chinese scientists in 1971 who performed low temperature ethyl ether extractions of Artemesia annua. In a research effort, apparently prompted by the requests of Ho Chi Minh to Zhou En Lai for antimalarial drugs to protect his Vietnamese troops, the scientists identified the active antimalarial principle, characterized its physicochemical properties, conducted in vitro, animal, and human studies, and synthesized derivatives of the more potent dihydroartemisinin (DHA). Artemisinin was first announced to the rest of the world in 1979 (1). At first, biological chemists were puzzled by the apparent stability of the hitherto unknown 15-carbon (sesquiterpene) peroxide structure. A full chemical synthesis was reported 4 years later (2), although, as for quinine, this remains too expensive for commercialization. Trials reporting efficacy in both uncomplicated and severe malaria soon followed (3, 4), but progress thereafter was slow. Instead of accepting the compounds the Chinese had developed, the World Health Organization (WHO)–Special Program for Research and Training in Tropical Diseases (TDR), the pharmaceutical industry, and the U.S. Army elected to develop their own compounds. Time and money was wasted developing artemotil (arteether), the ethyl ether of DHA, an oil-based formulation for intramuscular injection, which the Chinese scientists had discarded earlier in favor of the almost identical artemether (the methyl ether of DHA) (Fig. 3). Initially, the orally active compounds and the water-soluble artesunate, which could be given intravenously, were neglected entirely outside China, but with worsening resistance to all available antimalarials in Southeast Asia, researchers there began to investigate the compounds from China, and an increasing evidence base accumulated supporting the rapidity, reliability, and safety of these drugs in both uncomplicated and severe malaria. The parent drug artemisinin was largely replaced by DHA and its derivatives artesunate and artemether, which have greater antimalarial activity (Fig. 2). Initially artemisinin and its derivatives were used as monotherapies, but it became gradually accepted that antimalarials, like antituberculosis and antiretroviral drugs, should be used in combination (5, 6). A semisynthetic artemisinin derivative (artelinate) was developed, but was not taken beyond animal studies, and another (artemisone) is under development. Artiflene, a structurally dissimilar peroxide derivative of another Chinese plant (yinghaosu) was developed in the early 1990s and proved an effective antimalarial in clinical trials, but it was eventually abandoned because of high production costs and lack of evident advantages over the artemisinin derivatives.

Fig. 3.

Chemical structures. (Top) Artemisinin and its derivatives. 1, artemisinin; 2, artemether; 3, artemotil (arteether); 4, artesunate. (Bottom) The synthetic trioxolane OZ 277 (51).

Pharmacological Properties

Artemisinins kill nearly all of the asexual stages of parasite development in the blood (7), and also affect the sexual stages of P. falciparum (gametocytes), which transmit the infection to others (8), but they do not affect pre-erythrocytic development or the latent stages of P. vivax and P. ovale in the liver (the hypnozoites). The mechanism of action of artemisinins remains uncertain. The integrity of the endoperoxide bridge is necessary (but not sufficient) for antimalarial activity. Ion-dependent alkylation (principally by Fe++) is a likely mode of action (9), and the sarcoplasmic endoplasmic reticulum calcium adenosine triphosphatase (PfATPase 6) has been proposed as the primary target (10). The role of reactive decomposition intermediates such as carbon-centered free radicals remains controversial. The report that P. falciparum parasites from French Guiana with point mutations in the gene encoding PfATPase 6 were relatively resistant to artemether seemed to have fulfilled the molecular Koch's postulates for this target (11), but these findings have not yet been reproduced elsewhere. The synthetic peroxide trioxolanes, which are potent antimalarials, are 2 to 3 orders of magnitude less active in inhibiting PfATPase 6 (12).

Artemisinin's broad stage specificity of antimalarial action (Fig. 1) has two therapeutic consequences. Killing young circulating ring-stage parasites in P. falciparum infections results in a more rapid reduction in parasitaemia compared with other antimalarials (Fig. 4) and reduces considerably the number of parasites that mature to sequester in and block capillaries and venules (13, 14). This explains the rapidity of clinical responses and the life-saving benefit in severe malaria compared with quinine (which does not stop sequestration because it kills parasites only after they have matured and adhered to vascular endothelium). Reducing gametocyte carriage diminishes the transmission potential of the treated infection.

Fig. 4.

The pharmacodynamic properties of the antimalarial drugs in vivo assessed in terms of parasite clearance (14). Shown are infections of 1012 parasites corresponding to ∼2% parasitemia in an adult. The horizontal dashed line is the threshold for microscopic detection. Weak antimalarials, such as many antibiotics with antimalarial properties, produce parasite reduction ratios (PRR) of about a factor of 10 per asexual cycle, and therefore take weeks to eradicate an infection. Most antimalarials have PRR values of 100 to 1000 per cycle. The artemisinin derivatives are the most potent antimalarial drugs, with PRR values of ∼10,000 per cycle, and therefore take only three to four cycles (6 to 8 days) to remove all the parasites from the blood.

Artemisinin is eliminated by metabolic biotransformation, predominantly by CYP 2B6, to inactive metabolites. The artemisinins are weak inducers of some important drug-metabolizing enzymes and augment their own clearance (15, 16). After oral or parenteral administration, artemether, artemotil, and artesunate are all converted back rapidly to DHA in vivo, which is then eliminated by glucuronidation with an elimination half-life of ∼1 hour, both in healthy volunteers and in patients with malaria (Fig. 5). The broad stage specificity of action ensures that a single daily administration is sufficient for maximal killing of sensitive parasites. A 3-day artemisinin combination treatment (ACT) regimen provides antimalarial activity for two asexual parasite cycles and results in a reduction by a factor of about 100 million in parasite numbers within the infected patient—but this still leaves up to 100,000 parasites for the partner drug to remove, variably assisted by the immune response (14). The artemisinin component of the ACT therefore reduces the probability that a mutant resistant to the partner drug would arise from the primary infection, and, if effective, the partner should kill any artemisinin-resistant parasite that arose.

Fig. 5.

Comparative pharmacokinetics of intramuscular (i.m.) artesunate (2.4 mg/kg), oral artesunate (4 mg/kg), and i.m. artemether (3.2 mg/kg) in acute malaria. The plasma concentrations of the parent compound are shown as a red line for i.m. artesunate, a red dashed line for oral artesunate, and a blue line for i.m. artemether. The commonactive metabolite dihydroartemisinin is shown as gray or green lines.

Artemisinins also have clinically important activity against other parasites. They kill the younger stages of trematodes and are effective in the treatment of schistosomiasis and fascioliasis and in animal models of Clonorchis infections (17, 18). Their in vitro activity against other protozoa is considerably less than against malaria parasites, although they might be of value in the treatment of African and South American trypanosomiasis (19). Artemisinins have anti-inflammatory properties and also inhibit angiogenesis and cell growth in several neoplastic cell lines, which suggests a potential role in cancer chemotherapy (20, 21).

Artemisinin Combination Treatment

When artemisinins are given alone, 7-day regimens are required to maximize cure rates. Adherence with 7-day treatment courses is poor, so the combination partner in ACTs is usually a slowly eliminated antimalarial drug. This allows a complete treatment course to be given in 3 days (22). The first ACT to be evaluated systematically was artesunate-mefloquine (3, 23). This was deployed in 1994 on the northwest border of Thailand—an area of mefloquine-resistant falciparum malaria—and has retained efficacy over the subsequent 14 years (24). The first fixed-dose ACT (artemether-lumefantrine) followed soon afterward. Other ACTs have combined artesunate with existing drugs (sulfadoxine-pyrimethamine or amodiaquine) (6). Their evaluation (25) coincided with increasing realization that, whereas mortality from most infectious diseases (with the exception of HIV-AIDS) was declining, malaria mortality was rising (26, 27). This was attributed directly to the continued use of increasingly ineffective antimalarial drugs, mainly chloroquine and sulfadoxine-pyrimethamine (SP). Where the partner drug had not already fallen to resistance, the new ACTs were effective and well tolerated, but they were more expensive than the failing monotherapies.

Malaria affects the poorest people in the poorest countries. Many affected people can barely afford the US $.10 equivalent to buy ineffective chloroquine. Malaria control activities are heavily dependent on aid, and initially this was not forthcoming for the ACTs, so the necessary policy changes were delayed (28). In the last few years, however, there has been a massive increase in donor funding, spearheaded by the Global Fund to Fight AIDS, Tuberculosis, and Malaria (GFATM). Politicians in temperate countries, whose constituents have long been free from the burden of malaria, began to appreciate the humanitarian and economic burden that malaria imposes on the developing world. At the end of 2007, Bill and Melinda Gates called for a sustained initiative to eliminate malaria. It has taken a long time for the world to think positively again about conquering malaria (24, 29, 30) after the painful failure of the first global eradication campaign to eliminate the disease from the tropics. Whether malaria can and will be eliminated from high-transmission areas, where the vectorial capacity is enormous, remains to be seen, but most agree that falciparum malaria elimination in many low and unstable transmission settings is feasible (these include much of Asia and South America and large areas of Africa). Providing highly effective treatment is an essential component of malaria control and is required for elimination.

ACTs are now recommended by the WHO as the first-line treatment for all falciparum malaria in malaria endemic countries. WHO recommends aiming for cure rates of 95%, assessed at 28 days, and changing antimalarial treatment policy if cure rates are less than 90% (31). Artesunate-mefloquine and artemether-lumefantrine are reliably efficacious everywhere except Western Cambodia (32). Artesunate-amodiaquine and artesunate-SP give cure rates over 90% when cure rates with the corresponding monotherapies exceed 80%. In low-transmission settings, where symptomatic infections constitute a major source of transmission, ACTs reduce gametocyte carriage and, if widely deployed, reduce the incidence of falciparum malaria (24, 29, 30). The ACTs are also highly effective against infections caused by P. vivax, P. malariae, and P. ovale (except for artesunate-SP against antifol-resistant P. vivax), but most people who need ACTs still do not receive them.

Severe Malaria

In the 1990s, several randomized trials were conducted that compared parenteral artemether with quinine. An individual patient data meta-analysis of these trials did not show an overall difference in mortality between the two treatments, although in the prospectively defined subgroup of Southeast Asian adults, mortality was significantly lower in artemether recipients (33). The oil-based artemether is absorbed slowly and erratically from intramuscular injection sites (Fig. 5), particularly in severely ill patients (34, 35). This pharmacokinetic disadvantage may have offset the intrinsic pharmacodynamic advantage of the artemether in killing malaria parasites. By contrast, the water-soluble artesunate can be given by bolus intravenous injection and is absorbed rapidly and reliably after intramuscular injection. In the largest prospective randomized study in severe malaria (which enrolled 1461 patients), conducted in Southeast Asia, artesunate reduced the mortality of severe malaria from 22 to 15%—a 35% reduction (36) (Fig. 6). Artesunate is now the treatment of choice for severe falciparum malaria in areas of low transmission (31). However, most deaths from severe malaria occur in or near home. If treatment were available close to home, then the lethal progression to severe malaria might be halted. A formulation of artesunate for rectal administration with ∼50% bioavailability in acute malaria has been developed (37) that can be administered easily by a village health worker. In very large field studies, it has proved safe, well tolerated, and effective.

Fig. 6.

Treatment with intravenous artesunate was associated with a 35% reduction in mortality from severe malaria compared with intravenous quinine in a randomized trial of 1461 adults and children in Southeast Asia (36). The vertical axis shows the proportion of patients surviving.

Safety and Efficacy

The artemisinins have proved to be well tolerated and safe drugs. They temporarily suppress erythropoeisis, but this does not cause significant anemia. The only potentially serious adverse effects are relatively unusual hypersensitivity reactions (∼1 in 3000 treatments) that manifest initially as urticaria. In experimental animals, sustained central nervous system exposure to high concentrations of the artemisinins produces an unusual selective pattern of damage to certain brain-stem nuclei, particularly those involved with hearing and balance (38). Experimental neurotoxicity is therefore greater after intramuscular administration of high-dose oil-based artemether and artemotil, which are released slowly from the injection site, than with the same doses given orally (despite much higher peak concentrations) or the water-soluble artesunate given by any route (39). With the exception of a much-debated case series (40), extensive clinical and neurophysiological studies and a small series of neuropathological evaluations have provided no evidence of neurotoxicity in humans (4143). Although concerns over potential neurotoxicity have receded, the animal data do argue against using artemisinins for prophylaxis or giving long courses of high-dose intramuscular artemether. The one unresolved safety concern is in the first trimester of pregnancy. When experimental animals are exposed to artemisinins during early pregnancy, fetal resorption may result. This results from temporary suppression of fetal erythropoeisis caused by depletion of fetal erythroblasts (44, 45). Whether artemisinins could be teratogenic in humans has not been established. So, although there is increasing evidence of safety for the artemisinins in the second and third trimesters, these drugs are not recommended in the first trimester except in severe malaria, where their life-saving benefit to the mother outweighs the putative risks to the fetus (31).

In uncomplicated malaria, the safety, tolerability, and efficacy of ACTs is determined largely by the partner drug, because the artemisinin component is very well tolerated and provides a fixed antimalarial effect. The WHO currently recommends four ACTs (artesunate-sulfadoxine-pyrimethamine, artesunate-amodiaquine, artesunate-mefloquine and artemether-lumefantrine) and, to ensure adherence to combination treatment, encourages the use of fixed-dose combinations (FDCs) (31). Dihydroartemisinin-piperaquine is another highly effective FDC that is already used in several countries. Artesunate-pyronaridine is an FDC ACT in the later stages of development. Whereas ACTs containing mefloquine, lumefantrine, and piperaquine are effective nearly everywhere in the world, high levels of resistance to pyrimethamine, amodiaquine, and chlorproguanil, particularly in Asia and South America, limit the usefulness of ACTs containing these drugs.


Because current antimalarial treatment is so dependent on artemisinins, there has been considerable concern that resistance would emerge to this class of drugs as it has to other classes of antimalarials. Reassuringly, it has proved difficult to induce stable high-level resistance to artemisinins in the laboratory. Rodent malarias with reductions in susceptibility by factors of 5 to 10 have been selected, and similar reductions in susceptibility have been obtained by selection in P. falciparum (46). Amplification of Pfmdr1 is associated with relatively small but significant reductions in susceptibility to artemisinins. Most isolates of P. falciparum are extremely sensitive to artemisinin and its derivatives; in vitro median inhibitory concentration (IC50) values for dihydroartemisinin and artesunate are typically below 10 nmol/L, with values for artemether that are usually 2 to 3 times as high, and for artemisinin 5 to 10 times as high. Jambou et al. reported isolates from French Guiana with artemether IC50 values up to 117 nmol/L associated with mutations in PfATPase 6. Mutations in this gene are not associated with resistance in field isolates from elsewhere or in the laboratory lines selected for resistance (11). Yang et al. reported reduction by a factor of 3.3 in the susceptibility of P. falciparum to artesunate between 1988 and 1999 in Yunnan, Southwest China (47). This is a region where artemisinins have been used extensively for more than 20 years. More recently, studies from the Thai-Cambodian border have demonstrated reduced susceptibility to artemisinins both in vitro and in vivo (48). This is an area notorious as the epicenter from which chloroquine-resistant and later multidrug-resistant P. falciparum spread 50 years ago. Treatment failure rates after artesunate-mefloquine and artemether-lumefantrine in western Cambodia often exceed 10% (32), consistently higher than anywhere else in the world. Although P. falciparum is relatively resistant to mefloquine and lumefantrine in this area, and this certainly contributes to the high failure rates, there is concern that sensitivity to artemisinins may have declined as well. Artemisinins have been used in a variety of formulations and doses in western Cambodia for approximately 30 years, providing a cumulative exposure that may be more than anywhere else in the world. Although in vitro susceptibility tests indicate a relative reduction in artemisinin susceptibility, compared with most other regions of the world, the parasite isolates studied from this region are not highly resistant (i.e., reductions in susceptibility by no more than a factor of 10). They are not more resistant in vitro than parasites isolated on the eastern and western borders of Thailand, but in vivo the parasite clearance times in western Cambodia are consistently longer than elsewhere (49). Rapid parasite clearance is the pharmacodynamic hallmark of the artemisinins, reflecting their unique activity against young ring-stage parasites (14). Thus, there is a discrepancy between the in vitro and in vivo findings. Intensive studies are under way to assess the geographic extent of the problem, to characterize the in vivo and in vitro responses further and, if possible, to identify the molecular basis of the artemisinin-tolerant phenotype (49).

In assessing possible artemisinin resistance, specific features of the therapeutic response to these drugs needs to be considered. First, if the artemisinins are given alone for 7 days to nonimmune patients, ∼10% of patients fail treatment, yet when the recrudescent parasites are compared with parasites isolated from successfully treated infections, they are not more resistant. The persistence of temporarily growth-arrested intraerythrocytic merozoites or young trophozoites, which can “awaken” from dormancy days or weeks later, provides a plausible explanation for this result (50). The biological basis for this phenomenon is not understood. Second, apparently slow parasite clearance has been observed in a small proportion of patients in most trials conducted in nonimmune patients. In patients who have undergone a previous splenectomy, dead intraerythrocytic parasites may persist for up to a month in the circulation, so reduced splenic phagocytic function might explain slow parasite clearance in some cases (51). Third, current in vitro susceptibility tests assess maturation to mature schizonts, or the production of proteins or nucleic acid, which is maximal in mature trophozoites, whereas accelerated parasite clearance reflects drug effects on young circulating ring stages of the parasite. Current in vitro tests may be insensitive to the early development of artemisinin resistance.

Reducing Costs

Artemisinin is extracted from the leaves of Artemisia annua, which must be planted each year. China and Vietnam are major producers, and in recent years production in Africa has been increasing. Temperature, humidity, and soil type all affect yields considerably. Artemisinin is reduced in a potentially explosive reaction to dihydroartemisinin, which can then be converted to more heat-stable ethers (artemether, artemotil) or the succinate derivative (artesunate). While the market was small, prices were volatile, with prices for artemisinin ranging from US $350 to $1700 per kilogram on world markets. In recent years, there has been an increase in agricultural production and improvements in horticulture and yields. The number of companies extracting artemisininhas risen from less than 20 to more than 100. This has resulted in a reduction in the market price of artemisinin (52) and in retail costs, which can be less than US $1 for an adult treatment course. This is still approximately ten times as expensive as its synthetic predecessor chloroquine, and even chloroquine was unaffordable for many, so there have been considerable price barriers to widespread use. Current initiatives to reduce cost include (i) improvement in horticulture by fast-track selective breeding without genetic modification (53), (ii) microbial synthesis of the artemisinin precursor artemisinic acid followed by chemical synthesis of artemisinins (54), and (iii) development of fully synthetic antimalarial peroxides (55).

The microbial production of artemisinin is expected to begin in 2010 with goals of producing artemisinin at US $100 per kilogram and enough medicine for 200 million treatments each year. A large number of antimalarial peroxides have been synthesized over the past 15 years, but only one (OZ 277/RBx-11160, Arterolane) entered clinical development (55). This trioxolane compound was orally bioavailable and highly potent, but its development was discontinued before phase 3 clinical trials because of its instability in blood. More stable peroxides are now in development, and hybrid trioxane-aminoquinoline molecules have been developed (trioxaquines), which combine two mechanisms of action in a single molecule (56). These are undoubtedly promising developments, and it is likely that one or more of these candidates will eventually realize the objective of an extremely low-cost, fully synthetic peroxide antimalarial.

Increasing Access

Access to health care through the public sector in many tropical countries is inadequate, and for people with little or no money, malaria medicines may be unaffordable. Most febrile illnesses are treated empirically without any diagnostic procedure. Antimalarials are widely used commodities that can usually be purchased from shops, in the market, or from itinerant drug sellers, although in remote communities access to drugs is usually poor. The doses recommended are commonly incorrect, the quality of the drugs is often poor, and counterfeits are a major problem (57). Ineffective drugs are often purchased because they are less expensive, or an incomplete course of treatment is taken so that some tablets can be retained for the next febrile illness. The net result is both over- and undertreatment and increased selection pressure for the emergence and spread of resistance. Strengthening health service infrastructures and increasing the provision of diagnostics and effective treatment are vital components of current control efforts, but to reach most people in need, the availability of effective, affordable ACTs through the private sector also needs to be strengthened. In many malaria-endemic countries, this is the main source of treatment. The Institute of Medicine reviewed the economics of antimalarial drugs and in 2004 concluded that the most effective way of increasing access to antimalarial drugs would be to ensure that quality-assured effective drugs (currently ACTs) were provided free to the public sector and were also made available in the market place at a price close to that of the now largely ineffective chloroquine (about US $.10 to $.20 for a treatment) (58). This would have the added advantages of reducing the financial incentives for incomplete treatment, outcompeting ineffective drugs while driving out more costly and less effective artemisinin monotherapies, and reducing considerably the incentives to increasingly sophisticated and pervasive counterfeiters (59). It was proposed that a central subsidy be provided by international donors. This concept was taken forward by the Roll Back Malaria (RBM) Finance and Resource Working Group and the World Bank. In November 2007, the RBM board endorsed the further development of an Affordable Medicines Facility for malaria (AMFm), a groundbreaking initiative to improve access to safe, effective, and affordable malaria medicines (60). It is expected that the AMFm will require US $1.4 to 1.9 billion in funding over 5 years. This can be compared with estimates of the costs of malaria: US $12 billion a year in lost productivity in Africa alone and, in countries with a very heavy malaria burden, as much as 40% of public health expenditure, 50% of inpatient admissions, and up to 60% of outpatient visits (61). It is definitely a price worth paying.

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article