Malaria is a serious disease caused by protozoa called plasmodia. The infection is spread by the bite of Anopheles mosquitoes and is prevalent throughout the tropics. Malaria causes severe fever, and, in some cases, fatal complications affecting the kidneys, liver, brain, and blood.


There are five species of plasmodia that commonly cause malaria in humans:

  1. Plasmodium falciparum
  2. Plasmodium vivax
  3. Plasmodium ovale
  4. Plasmodium malariae
  5. Plasmodium knowlesi

When a mosquito carrying the infection bites a human, the plasmodia enter the bloodstream. They invade the liver and red blood cells, where they multiply. The red cells then rupture, releasing the new parasites. Some of them infect new red cells, and the others develop into forms that can infect more mosquitoes. Falciparum malaria infects more red cells than the other species and thus causes a more serious infection. Most cases of this form occur in Africa.


Symptoms of malaria include fever, shaking, and chills. There may also be severe headache, general malaise, and vomiting. The fever often develops in cycles, occurring every other day (in vivax and ovale infections) or every third day (in malariae infections). Falciparum malaria can be fatal within days. Infected red cells become sticky and block blood vessels in vital organs. The spleen becomes enlarged and the brain may be affected, leading to coma and convulsions. Destruction of red blood cells causes haemolytic anaemia. Kidney failure and jaundice often occur.

Diagnosis and treatment

A blood film examination is carried out to detect the parasites. Falciparum malaria is treated with quinine or mefloquine, or proguanil and atovaquone, or artemether with lumefantrine. Chloroquine is the usual treatment for species other than falciparum. People with vivax or ovale malaria must also take the drug primaquine to eradicate parasites in the liver. In severe cases, blood transfusions may be necessary.


People who live in or visit malarial areas should take precautions in order to avoid mosquito bites; such measures include keeping the arms and legs covered and using insect repellents and mosquito nets. In addition, travellers to such areas should take preventive anti-malarial drugs. The course of drugs must be started up to three weeks before the person enters the area, and should be continued for one to four weeks after the visit (exact timings depend on the type of drug). Doctors can provide up-to-date advice on the choice of drugs.

Malaria in detail - technical


Malaria has been eliminated from many countries but remains the most important human parasitic disease in sub-Saharan Africa and other tropical and subtropical zones. It causes more than 500 million cases of illness and a million fatalities each year in 100 countries. There is currently a renewed international effort to reduce its impact on populations still at risk.

Human malaria parasites, mosquitoes, and transmission of malaria

Malaria parasites and their impact on the human genome—five species of Plasmodium commonly cause malaria in humans: P. falciparum, P. vivax, P. ovale, P. malariae and P. knowlesi. The genome of P. falciparum, the most pathogenic species, has been completely sequenced. This parasite has exercised immense selection pressure on the human genome, as is evident from the global distribution of the many human genes that constrain malarial development, such as a point mutation in position 6 of the β-globin chain (sickle cell haemoglobin), and deletion of α-globin genes (α thalassaemia).

Biology of the parasite and mosquito vector—sporozoites are injected into humans during the female anopheles mosquito’s blood meal. They invade hepatocytes. Hepatic schizogony releases merozoites into the blood stream where they invade red blood corpuscles (RBCs) and undergo further asexual multiplications before gametocytes form. If these are ingested by mosquitoes, male and female gametes fuse, resulting in ookinetes that penetrate the mosquito’s midgut and develop into oocysts. Daughter sporozoites are released. They invade the mosquito’s salivary glands, ready to infect a new human host. Persistent latent forms (hypnozoites) of P. vivax and P. ovale remain in the liver to give rise to later relapses of parasitaemia and symptoms. All the stages express distinct antigen, repertoires excite different immune responses, and are equipped survive in different microenvironmants.

Mosquito biology—species of the Anopheles gambiae complex, the most effective malaria vectors, prefer to feed on humans to whom they are attracted by smell: other species are less particular. They vary in their choice of breeding habitats. MacDonald’s equation for vectorial capacity and the related basic reproduction number (R 0) allows prediction of the impact of vector control methods under different conditions. The genome sequence of An. gambiae is known. Important mosquito phenotypes that have a genetic basis include blood feeding preference, habitat choice, insecticide susceptibility, and vectorial capacity.

Other mechanisms of transmission—malaria can be transmitted by transfusion of blood products, marrow transplants, and contaminated needles.


In 2007, 2.4 billion people were exposed to P. falciparum infection across 87 countries, and 3.18 billion people were exposed to P. vivax across 63 countries. Intensity of malarial transmission depends on the varying efficiencies of the local anopheline vectors and their frequency of contact with humans.

Malarial endemicity expresses the amount or intensity of transmission in an area or community. Epidemic malaria implies a periodic or sharp increase in the amount of malaria. Stable transmission implies persistently high prevalence, insensitive to aberrations in climate or local habitats as in holoendemic areas of Africa; unstable malaria is characterised by great variability in space and time, as in South-East Asia. Prevalence of infection in children aged 2 to 9 years is described as hypoendemic (<10%), mesoendemic (11–50%), hyperendemic (51–75%), or holoendemic (>75%).

The epidemiological background to clinical malaria—is changing due to population growth, environmental changes (often human-induced, whether local or global), changing resistance of parasites to drugs, the HIV epidemic and the consequences of attempts at malaria control. An estimated 550 million clinical attacks of P. falciparum occurred worldwide in 2002: 71% in Africa, 23% in the low-transmission but densely populated countries of South-East Asia, and 3% in the Western Pacific. In Africa in 2005, P. falciparum is estimated to have caused 1.1 million deaths directly, 71 000 to 190 000 infant deaths following placental infection in utero, and over 3000 newly acquired persistent epilepsies through brain insults among patients surviving an episode of cerebral malaria in childhood.

Innate resistance and immunity

More human genetic polymorphisms have been associated with innate protection from malaria than for any other infectious disease. Duffy blood group negative RBCs are resistant to P. vivax infection, explaining the prevalence of the DARC(Fy) –46C/C genotype especially in West Africa, but there may be an associated susceptibility to HIV-1 infection.

In most stably endemic areas, acquisition of immunity, although never complete, ensures that death due to malaria is rare after the age of 5 years and hardly ever occurs in normally immune competent adults. Immunity allows tolerance of levels of parasitisation that would cause illness in a naive individual by neutralizing parasite toxins or down-regulating the cytokine response to challenge. However, a key aspect of immunity to malaria is control of parasite growth by interfering with parasites’ replication or accelerating their removal from the circulation. There is progressive acquisition of both ‘strain’-specific and cross-protective responses to a range of potential malarial epitopes. Immunity is stage-specific but probably acts predominantly at the blood stages. Antibody-mediated protection against blood-stage parasites is demonstrated by the relative protection of children in endemic areas during their first few months of life by passively transferred maternal antibody and by experimental amelioration of acute malaria by immune gammaglobulin. Malnutrition increases the risk of severe falciparum malaria in children.

HIV–malaria interaction—in pregnant women, HIV and P. falciparum infections are mutually synergistic. Consequences of malaria, especially anaemia, are more severe in HIV-positive women. In areas of unstable malarial transmission, HIV-positive nonimmune adults are at increased risk of severe and fatal malaria. In malaria endemic areas, HIV-positive children are at increased risk of severe malaria.

Molecular pathology, organ pathology, and pathophysiology

Molecular pathology—intravascular, asexual forms are responsible for all the pathological effects of malaria in humans. Fever and inflammation are probably initiated by interaction between parasite products and pattern recognition receptors on host cells, leading to cytokine release by macrophages. The relative virulence of P. falciparum is attributed to cytoadherence and sequestration of parasitized RBCs to venular endothelium, especially in the lungs, brain, intestines and muscles, resulting in reduced perfusion and tissue damage. Local release of potentially toxic/pharmacologically active compounds such as reactive oxygen species or nitric oxide may also be involved.

Organ pathology—the brain may be oedematous, especially in African children. Small blood vessels are congested with tightly sequestered parasitized RBCs (PRBCs) containing pigmented mature trophozoites and schizonts, making the brain slate-grey in colour. The cerebrovascular endothelium shows pseudopodial projections, closely apposed to electron-dense, knob-like protruberances on the surface of PRBCs. Other changes include petechial haemorrhages in the white matter, ring haemorrhages and Dürck’s granulomas. Among other organs and tissues, retina, bone marrow, lung, heart, liver, intestine, spleen, kidney, and placenta show variable evidence of PRBC sequestration and some other distinctive features.

Pathophysiology—anaemia results from destruction/phagocytosis of both normal red cells and PRBCs as well as from dyserythropoiesis; autoimmune haemolysis is rare. Thrombocytopenia is attributable to splenic sequestration, dysthrombopoiesis, and immune-mediated lysis. Cerebral malaria is associated with inappropriately low cerebral blood flow, increased cerebral anaerobic glycolysis and microcirculatory obstruction. In African children, plasma concentrations of TNF-α, IL-1α and other cytokines correlate with disease severity. Cytokines may be involved in hypoglycaemia, coagulopathy, dyserythropoiesis, and leucocytosis in falciparum malaria. Pulmonary oedema may result from fluid overload, but more often there is increased pulmonary capillary permeability associated with neutrophil sequestration in the pulmonary capillaries. In African children, a syndrome of respiratory distress is associated with metabolic acidosis and severe anaemia. Hypoglycaemia is caused by impaired gluconeogenesis, reduced hepatic glycogen or hyperinsulinaemia secondary to quinine/quinidine treatment. In malarial acute renal failure, there is evidence of PRBC sequestration, and pigment (haemoglobin and myoglobin) toxicity may contribute.

Clinical features

Classic periodic febrile paroxysms with afebrile asymptomatic intervals are uncommon unless treatment is delayed.

Severe falciparum malaria—this is defined by (1) clinical features—prostration, impaired consciousness, respiratory distress/acidotic breathing, multiple convulsions, circulatory collapse, pulmonary oedema (radiological), abnormal bleeding, jaundice, and haemoglobinuria; and (2) laboratory tests—severe anaemia, hypoglycaemia, acidosis, renal impairment, and hyperlactataemia, that are of proven prognostic significance.

Cerebral malaria is defined by impaired consciousness in patients with acute P. falciparum infection in whom other causes of coma, including hypoglycaemia and transient postictal coma, have been excluded. Convulsions, dysconjugate gaze, retinal changes, symmetrical upper motor neuron signs, and abnormal posturing are common. Neurological manifestations are different in adults and children. African children surviving cerebral malaria may suffer persistent neurological, cognitive, and learning defects.

So-called benign malarias, P. ovale, P. malariae, and particularly P. vivax, can cause even more severe feverish symptoms than falciparum malaria. Splenic rupture is more common with vivax malaria. P. knowlesi, one of the monkey malarias, has recently been recognized as an important and potentially fatal zoonosis in humans in several South-East Asian countries.

Malaria in pregnancy—malaria is an important cause of maternal anaemia and death, abortion, stillbirth, premature delivery, low birth weight, and neonatal death. RBCs infected with strains of P. falciparum expressing Var2CSA bind to chondroitin sulphate A expressed on the surface of the syncytiotrophoblast. Placental dysfunction, fever, and hypoglycaemia contribute to fetal distress.

Chronic immunological complications of malaria—these include quartan malarial nephrosis, tropical splenomegaly syndrome (hyper-reactive malarial splenomegaly) and endemic Burkitt’s lymphoma.


Repeated thick and thin blood smears and rapid antigen detection over a period of 72 h are necessary to confirm or exclude the diagnosis of malaria. Differential diagnoses include other acute febrile illness: falciparum malaria has been misdiagnosed as influenza, viral hepatitis, epilepsy, viral encephalitis, or traveller’s diarrhoea, sometimes with fatal consequences.

Laboratory investigation

In falciparum malaria, blood glucose must be checked frequently, especially in children, pregnant women, and severely ill patients, whether or not the patient is receiving quinine/quinidine treatment.


The efficacy of antimalarial chemotherapy is threatened by emerging resistance of P. falciparum to available drugs. The World Health Organization (WHO) now advocates the combination of two or more different classes of antimalarial drugs with unrelated mechanisms of action to delay emergence of resistance.

P. vivax, P. ovale, P. malariae, P. knowlesi malarias—these are treated with chloroquine. Resistant P. vivax (New Guinea, Indonesia) is treated by increasing the dose of oral chloroquine.

Uncomplicated P. falciparum malaria in malarious areas—WHO recommends the replacement of monotherapy with the combination of an artemesinin with another drug (artemisinin-based combination therapy, ACT), even in Africa, although this is more expensive and resistance to artemisinins has recently emerged in Cambodia. In South-East Asia, lumefantrine or mefloquine is added to artesunate. In Africa, lumefantrine, amodiaquine, or sulfadoxine–pyrimethamine might be added. For presumed nonimmune travellers returning to nonendemic areas, artemether–lumefantrine, atovaquone–proguanil, or quinine with doxycycline or clindamycin (pregnant women and children) are recommended.

Severe falciparum malaria—urgent appropriate, parenteral chemotherapy is necessary, initiated with a loading dose. Intravenous artesunate is the drug of choice. Intramuscular artemether, or quinine by intermittent or continuous intravenous infusion or intramuscular injection are less effective. Artemisinin by rectal suppository has proved effective. Resistance to artemisinins is emerging in Cambodia, Thailand, and Burma.

Supportive care—patients with severe malaria should be transferred to the highest possible level of care. Convulsions must be controlled; fluid, electrolyte, and acid–base homeostasis restored; and organ/tissue failure treated (e.g. haemofiltration for acute renal failure). Harmful ancillary remedies of unproven value, such as corticosteroids and heparin, have no role in the treatment of cerebral malaria.


Modern malaria control and prevention aims to limit human–vector contact by indoor residual spraying (IRS) and insecticide (pyrethroid) treated nets (ITNs). ITNs can reduce all-cause childhood mortality by 17%, averting 5.5 deaths for every 1000 African children protected, preventing over 50% of clinical cases, and reducing prevalence by 13%. Repellents such as diethyltoluamide (DEET) are used for personal protection. Vectors can also be controlled by environmental modification or manipulation, and human contact can be reduced by zooprophylaxis and by modifying human dwellings and behaviour.

Intermittent preventive treatment in pregnant women (IPTp) and infants (IPTi) with sulphadoxine–pyrimethamine—efficacy is likely to decrease because IPTp works less well in HIV-positive women and there is no proven safe alternative to sulphadoxine–pyrimethamine in areas where resistance to this combination is rapidly expanding.

Malarial vaccines—obstacles to developing a malaria vaccine are the multistage complexity of the parasite, polymorphism of potential immune targets, and the parasite’s capacity for evolving evasive strategies, such as antigenic variation and diversity. However, candidate pre-erythrocytic, blood-stage, and transmission-blocking vaccines have been developed. A subunit vaccine (RTS,S) comprising a fusion protein combining part of the circumsporozoite protein of P. falciparum with HBsAg and a complex adjuvant (AS02) has achieved 53% protective efficacy against malaria disease.

Travellers—prevention of malaria in people from nonmalarious areas who are visiting endemic regions, including those visiting their friends and relatives (VFRs), has become more difficult because of resistance to antimalarial drugs. Travellers are advised to (1) be aware of the risk; (2) prevent exposure to anopheline mosquitoes; (3) take chemoprophylaxis where appropriate—malarone, mefloquine, or doxycycline is appropriate in areas of chloroquine-resistant falciparum malaria; (4) seek immediate medical advice in case of any feverish illness developing while abroad, or within 3 or more months of returning, and to mention malaria as a possibility—regardless of the precautions taken—to any doctor who sees them. Up-to-date advice is important, as the global distribution and intensity of malarial transmission is changing. Pregnant women are best advised to avoid malarious areas.


Malaria is the most important human parasitic disease globally and has had large effects on the course of history and settlement in tropical regions. Following the discovery in the 19th century of both the causative protozoan parasite, Plasmodium, and its mosquito vector, the disease was brought under control in many countries through the application of antimalarial drugs, insecticides such as dichlorodiphenyltrichloroethane (DDT), and other environmental interventions including urbanization. In the United States of America, Europe, the Mediterranean region, the Middle East, most Caribbean islands, some South American countries, northern Australia, and most of China, elimination or a high degree of control was largely achieved. Even in Sri Lanka in the early 1960s, cases had fallen from 1.1 million annually to just 18, but failure to maintain surveillance and react to outbreaks resulted in a return to previous levels.

In recent years, malaria has been subject to increased control efforts, with varying degrees of success, but the disease was resurgent in the 1980s and 1990s. Malaria remains the dominant tropical vector-borne disease but, after decades of neglect, international interest in its control has recently revived. There is now a global effort to develop new methods to intervene against parasite dissemination, stimulated by the emergence of drug resistant parasites in South-East Asia and Africa, mosquitoes resistant to DDT and other insecticides, and by the recognition that malaria has a considerable economic impact: in Africa, total gross domestic product losses due to malaria amount to US$12 billion per year, and the global cost is US$18 billion. With over 500 million malaria cases annually and more than a million deaths in more than 100 countries, there remains a clear and urgent need for improved control and treatment.

Biology of the malaria parasite

Life cycle and parasite cell strategies

Five of the 147 known species of protozoan parasites genus Plasmodium that cause malaria (mal aria, Italian, literally ‘bad air’) commonly infect humans: P. falciparum, P. vivax, P. ovale, P. malariae, and P. knowlesi (Fig. 1). The biological organization and life cycle of P. falciparum, the species most pathogenic to humans, are distinct from that of all but P. reichenowi.

Evolutionary tree of malaria parasites

Above: Fig. 1 Evolutionary tree of malaria parasites inferred from 30 protein-encoding genes from the apicoplast genome. Four human parasites do not show a monophyletic relationship revealing host-switching during parasite evolution.

Genomic organization of Plasmodium and consequences for the human host

Plasmodium is diploid only until the first (meiotic) division of the genome following fusion of the male and female gametes and is haploid for the rest of its life cycle. The 23-Mb genome of P. falciparum has been completely sequenced: it contains from 5000 to 6000 genes distributed between 14 chromosomes ranging in size from about 1 Mb to 2.4 Mb. Considering its comparatively small number of genes, it is remarkable that it not only maintains a complex life cycle but survives in the face of the overwhelming number (20 000–30 000) of genes available to its human host. The parasite has exercised immense selection pressure on the human genome as is evident from the global distribution of the many human genes that constrain malarial development (Bullet list 1 below).

Development in the mosquito

After the blood meal, ingested intraerythrocytic gametocytes (Fig. 2)  in the midgut of the female mosquito are triggered to undergo gamete formation (exflagellation) by a drop in temperature of more than 5°C and the presence of raised concentration of mosquito-derived xanthurenic acid. The gametocytes escape out of the red blood cell (RBC) into the lumen of the midgut, where the female gamete can be fertilized within minutes by a microgamete released from microgametocytes (Fig. 2). Major zygote/ookinete surface proteins are detectable on the zygote surface within 1 h of fertilization, particularly P48/45, P230, P25, and P28, which are potential components of a transmission-blocking vaccine.

Within about 8 h of fertilization, the briefly diploid genome has undergone meiosis producing a single nucleus containing four haploid genomes. Over the ensuing 9 to 12 h, the zygote becomes a motile and invasive banana-shaped ookinete (Fig. 2 and 3). The extracellular gametes, zygotes, and ookinetes are exposed to potentially lethal components in the blood meal, such as complement to which the parasite is initially resistant, antibodies, and mosquito proteases. Then, 24 to 36 h after the blood meal was ingested, the ookinete penetrates the chitinous peritrophic membrane, newly secreted by the mosquito to defend itself against parasitic invasion, and the plasma membrane of midgut epithelial cells. Unlike sporozoites and merozoites, the ookinete lacks rhoptries and consequently does not form a parasitophorous vacuole (PV) when it invades the epithelial cells. On meeting the collagen-containing basal lamina on the outer wall of the midgut, the ookinete transforms into a vegetative replicating form, the oocyst, protected from the mosquito’s immune system by a thick proteinaceous wall containing proteins P380 and circumsporozoite protein (CSP). Commonly, fewer than five of the thousands of gametocytes originally present in the blood meal form oocysts.

Plasmodium life cycle

Above: Fig. 2 Plasmodium life cycle

Parasite movement

The motile ookinete, merozoite, and sporozoite are impelled by an unconventional actomyosin motor.

Above: Fig. 2 Plasmodium invasive stages, ookinete, sporozoite, and merozoite

Development of the oocyst into sporozoites

Depending partly on ambient temperature, the oocyst nucleus undergoes from 10 to 13 endomitotic divisions over a period of 10 to 25 days, until finally a single cytokinetic division results in the simultaneous production of between 2000 and 10 000 daughter sporozoites. Mature sporozoites secrete a protease (ECP-1) to digest the proteinaceous oocyst wall to escape into the mosquito’s haemocoelomic fluid. Only those capable of invading the salivary glands survive. They bind to salivary gland receptors via ligands such as CSP, TRAP, and MAEBL (apical membrane antigen/erythrocyte binding-like), penetrate the plasma membrane of the acinar cells, and come to lie in the salivary ducts. P. falciparum sporozoites can remain infectious in the glands for up to 55 days, many times longer than the natural lifetime of the infected mosquito, which delivers 10 to 100 sporozoites per bite.

Development of the exo-/pre-erythrocytic (liver) stages

CSP, the dominant surface protein on the sporozoite, is critical to this phase of the life cycle and has been the most popular vaccine candidate. Most of the sporozoites deposited in the dermis by the biting mosquito cross the capillary epithelium and are rapidly transported in the bloodstream to the liver, where they invade phagocytic Kupffer cells. A few sporozoites may enter the lymphatic system, where they may prime antigen-presenting cells in the lymph nodes. Kupffer cells tolerate microbes and their products (portal vein tolerance), but intracellular sporozoites also inhibit phagocytes’ oxidative burst. CSP inhibits fusion of lysosomes with the parasite-containing vacuoles (PV). Sporozoites then escape from the Kupffer cells into the space of Disse and invade adjacent hepatocytes. CSP, secreted from the micronemes, binds to heparin sulphate proteoglycans on the hepatocytes and, with TRAP and a perforin-like molecule, enables penetration of the hepatocyte membrane. The parasite may migrate through several hepatocytes, killing them in the process and inducing the production of hepatocyte growth factor that contributes to parasite nutrition. The sporozoites form PVs in hepatocytes where they differentiate into replicating exo-erythrocytic (EE) schizonts.

Parasites inhibit apoptosis, a host cell defence mechanism, so that hepatocyte mitochondria remain available for recruitment by the PV. Down-regulation of hepatocyte proteosomal activity reduces presentation of secreted parasite antigens to major histocompatability complex molecules, compromising recognition of the infected hepatocyte by cytotoxic T cells. However, EE-stage parasites remain a prime target for potential vaccines. Within the infected hepatocyte, sporozoites undergo schizogony, each producing from 10 000 to 30 000 daughter cells (Table 1), or merozoites. They are released in large cellular masses (merosomes) containing hepatocyte cytoplasm that are attacked by macrophages and neutrophils in the liver. However, individual merozoites escape into the circulation where they invade red blood cells (RBCs). Infection of the liver is without clinical consequences, possibly because only a small number of parasites complete development and little toxic waste is released. The cell cycles of P. vivax and P. ovale can become arrested with formation of quiescent hypnozoites that can persist in the hepatocyte for long periods, enabling these parasites to survive seasonal absences of mosquitoes. Reactivation of hypnozoites by unknown factors produces ‘relapse’ infection in the blood. Latencies and frequencies of hypnozoite relapses are very variable, suggesting that the strains responsible had their origins in both tropical and temperate zones (Table 1). Hypnozoites do not grow and are therefore susceptible only to ‘causal prophylactic’ drugs such as primaquine and atovaquone that target mitochondrial enzyme pathways responsible for essential energy metabolism. Relapse must not be confused with a recrudescence, which occurs as a result of the amplification of a chronic subpatent blood-stage infection of P. falciparum orP. malariae.

Bullet list 1 Some human genetic polymorphisms associated with resistance to malaria

  • α –Thalassaemia
  • β –Thalassaemia
  • Haemoglobin S
  • Haemoglobin E
  • Haemoglobin F
  • Haemoglobin C
  • South-East Asian ovalocytosis
  • Hereditary sphero-, ellipto-, pyropoikilocytosesa
  • G6PD deficiency
  • Pyruvate kinase deficiency
  • Duffy blood group
  • ABO blood groups
  • S-s-U blood group
  • Glycophorin B deficiency
  • Complement receptor-1
  • MHC class I
  • MHC class II
  • HLA Bw53
  • HLA DRB1*1302
  • TNF- α promoter
  • IFN- γ receptor

G6PD, glucose-6-phosphate deydrogenase; IFN, interferon; MHC, major histocompatibility complex; TNF, tumour necrosis factor (see also

a In vitro evidence only.

Development in the erythrocyte

The underlying mechanism of merozoite invasion of the RBC is highly conserved across Plasmodium species, whereas host cell recognition and binding is species limited. Different Plasmodium species invade RBCs of different ages. P. vivax invades reticulocytes and P. ovale also prefers younger RBCs; P. falciparum and P. malariae invade mature RBCs. Antibodies recognizing the various parasite ligands can inhibit merozoite invasion, thus offering potential targets for prophylactic vaccines. Invasion is a complex active process, taking less than a minute, that involves modification of both the merozoite surface and RBC membrane. Inside the RBC, a PV is created to contain the merozoite. After successful invasion of the RBC, proteins such as P. falciparum reticulocyte-like binding homologue proteins (PfRh) 1 to 4 are released into the PV (Table 2).

Life Cycle of Malaria Parasite; Sexual Phase, Sporogony; Asexual phase, Schizogony

Above: Fig. 4 Life Cycle of Malaria Parasite; Sexual Phase, Sporogony; Asexual phase, Schizogony

Following invasion, phagocytosis of RBC cytoplasm by the growing trophozoite occurs through a cytostome or micropore in the parasite’s plasma membrane, forming intracytoplasmic digestive vacuoles into which digestive enzymes such as plasmepsins and dipeptide aminopeptidase are secreted. Digestion of RBC proteins yields toxic haem that is sequestered as haematin crystals in membrane-bound vesicles, to be discarded when schizonts divide into merozoites. The rapid growth of the asexual parasite within the RBC demands new permeability pathways in erythrocyte membranes to facilitate entry of essential nutrients and the egress of toxic metabolites, especially lactic acid. P. falciparum builds membranous transport structures in the RBC cytoplasm including Maurer’s clefts, originally described in 1903, and the recently described tubulovesicular network. Parasite nutrient transporters in the infected RBC (iRBC) membrane are important targets for new antimalarial compounds. P. falciparum erythrocyte membrane protein 1 (PfEMP1) is the major parasite protein and is exposed on the RBC membrane as discrete warts (knobs). Energy metabolism of the asexual blood stages is critically dependent upon the mitochondrion. The recently discovered apicoplast is a vestigial chloroplast originating from red algae. It is responsible for pathways in lipid metabolism distinct from those of either vertebrates or mosquitoes and can, therefore, be targeted by drugs such as fosmidomycin, doxycycline, and clindamycin. These drugs are slow acting, taking two generations of parasite growth (96 h) before inhibition takes effect. Over the next 24 to 72 h, depending on the species, the parasite develops within the PV, eventually forming schizonts containing up to 30 merozoites (Table 1). Just before merozoite release, schizont volume increases and bursts the iRBC explosively. When blood-stage infections are synchronous, iRBC destruction, release of merozoites, and parasite toxic products into the bloodstream result in typical periodic patterns of fever in the human host.

Table 1 Distinguishing characteristics of malaria parasites infecting humans


Parasite P. falciparum P. vivax P. ovale P. malariae P. knowlesi
Development of liver stages (days) 5–7 6–8 9 14–16 5.5
Merozoite number in exo-erythrocytic schizont <30 000 <10 000 <15 000 <15 000 Unknown
Hypnozoite No Yes Yes No No
Maximum period to first relapse <3 yrs <100 days  
Blood parasites detected by microscopy (days) 9–10 11–13 10–14 15–16 9–11b
Days/years to first symptomsa 12 15/<1a 17/<4a 28 9–11b
RBC Cycle (hours) 48 48 49–50 72 24
Merozoite number in blood schizonts 16 16 8/16c 8–12 10
Distinguishing characteristics of species by microscopyd Commonly rings only; Maurer’s clefts; crescentic gametocytes, Schuffner’s dots, trophozoites irregular Large nuclei; Schuffner’s dots, trophozoites irregular Band forms Light stippling; sometimes band forms; resembles P. malariae
Maximum RBC infected (%) >60 0.01 <0.3 <0.2 12
Oocyst development at 28°C (days) 9–10 8–10 12–14 14–16 8–10
Oocyst size (µm) 55 50 45 40 65

a Exceptional cases of P. vivax and P. ovale took nearly 1 year and 4 years, respectively.

b From Chin, et al. (1968). Experimental mosquito-transmission of Plasmodium knowlesi to man and monkey. Am J Trop Med Hyg, 17, 355–8.

c Merozoite numbers in schizonts from relapses of P. ovale are increased.

d See Garnham PCC (1966). Malaria parasites and other haemosporidia. Blackwell, Oxford, and Coatney GR, et al. (1971). The primate malarias. US Department of Health, Education and Welfare, Bethesda, Maryland, USA.

Data from Garnham PCC (1966). Malaria parasites and other haemosporidia. Blackwell, Oxford; Coatney GR, et al. (1971). The primate malarias. US Department of Health, Education and Welfare, Bethesda, MD; and Bruce-Chwatt LJ (1985). Essential malariology, 2nd edition, Heinemann Medical, London.

Sexual development (gametocytes)

The asexual parasites themselves are a developmental dead end, but their expansive growth in the blood increases the potential for differentiation into the sexual forms that are responsible for continuing the life cycle by infecting female mosquitoes. Stress on the developing asexual forms, e.g. antimalarial drugs, immune pressure, and metabolic stress induced by the asexual population itself, stimulates gametocyte production. The progeny of each schizont are all asexual, all male, or all female. Males (microgametocytes) accumulate the proteins required for rapid DNA replication and flagellar motility during gametogenesis and then shut down protein synthesis with the loss of ribosomes and endoplasmic reticulum (ER). In contrast, the mature females (macrogametocytes) retain protein synthetic machinery (ER and Golgi), although shutting down active protein synthesis. Because mature gametocytes of both sexes have ceased protein synthesis, they are less susceptible than asexual parasites to many antimalarial compounds. However, they remain vulnerable to inhibitors of energy metabolism such as primaquine and artemisinin combination therapy. The sexual stages of most malarial parasites mature in the same time as the asexual parasites, but P. falciparum is atypical in requiring not 48 h but about 10 days to mature (Table 1). Like the asexual parasites, immature gametocytes of P. falciparum express PfEMP1, have knobs, and adhere to receptors. Normally, only mature gametocytes are released into the peripheral bloodstream, where they may persist for 22 days, with a population half-time of 2.2 to 7 days. Most gametocytes are not taken up by mosquitoes but are removed by the spleen, where they stimulate antibody responses to the ‘stored’ gametocyte proteins, some of which are subsequently expressed on the surface of the gametes (e.g. P230, P48/45) and are now considered possible targets for transmission-blocking vaccines.

Table 2 Some of the major proteins concerned with RBC attachment and merozoite invasion for P. falciparum and P. vivax
Species Parasite ligand Location Sialic acid RBC receptor for parasite ligand RBC receptor sensitivity to trypsina
P. falciparum PfMSP-142 MS Independent Band 3
PfMSP9 MS Band 3
EBA175 MN Dependent Glycophorin A Sensitive
EBA140/BAEBL MN Dependent Glycophorin C Resistant
EBA181/JESEBL MN NK Resistant
PfRh1–4 APM/Rhoptries/MN NK Resistant (1 and 2b)
P. vivax DBP Duffy antigen
PvRBP1 Rhoptries NK
PvRBP2 Rhoptries NK

AMA-1, apical membrane antigen 1; APM, apical pole of merozoite; DBP, Duffy binding protein; EBA175–181, erythrocyte binding antigen 175, 140, or 181 kDa (BAEBL and JESBEL are alternative names, this group are homologues of P. vivax Duffy binding protein); MN, microneme; MS, merozoite surface; NK, not known; PfMSP-142, P. falciparum merozoite surface protein-1 42 kDa fragment; PfMSP-9, P. falciparum merozoite surface protein-9; PfRh1–4, P. falciparum reticulocyte-like binding homologue proteins 1, 2,2a, 3, and 4; PvRBP1–2, P. vivax reticulocyte binding proteins 1 and 2.

This table summarizes: the main parasite ligands thought to be involved in merozoite attachment to, and invasion of, RBC; their location in the merozoite; whether or not they depend on sialic acid residues on the RBC proteins, and the sensitivity of the RBC receptors to trypsin. Other proteins are involved but less is known of their function.

Based on Oh SS, Chishti AH (2005). Host receptors in malaria merozoite invasion. Curr Top Microbiol Immunol, 295, 203–32, and Gaur D, et al. (2004). Parasite ligand-host receptor interactions during invasion of erythrocytes by Plasmodium merozoites. Int J Parasitol, 34, 1413–29.

Biology of the mosquito vector

The first indication that mosquitoes might be involved in human disease cycles was in 1876, when Patrick Manson found that culex mosquitoes transmitted filarial worms. Ross and Grassi’s discoveries of malaria transmission by anopheles mosquitoes followed in the 1890s. The Dipteran order of insects, to which the more than 3500 species and subspecies of mosquitoes belong, has many blood-feeding members and contains the insects of greatest medical and veterinary importance. Mosquitoes have coevolved with their vertebrate hosts, extending their feeding range from reptiles to mammals. The adaptation of malaria parasites to their mosquito hosts probably occurred about 20 000 years ago. From the human perspective, the most devastating link is that between P. falciparum and the African mosquito Anopheles gambiae, which is estimated to have been in place for as little as 10 000 years.

Blood feeding and host preference

Unlike many haematophagous insects, it is only adult female mosquitoes that have piercing and sucking mouthparts, adapted for taking a blood meal from vertebrate hosts to nourish the development of a single egg batch. Of the three groups of mosquitoes—anophelines, culicines, and aedines—only about 50 anophelines are malaria vectors. Many mosquitoes are part of complexes of sibling species, distinguishable only by modern molecular methods. The An. gambiae and An. funestus groups contain the most important African malaria vectors. Within the An. gambiae complex, An. gambiae sensu stricto, the best of the human malaria vectors, prefers to feed on humans, while An. quadriannulatus, a nonvector, feeds on cattle and An. arabiensis, a secondary vector in many parts of Africa, preferentially feeds on cattle but will take a human blood meal. Olfaction plays a vital role in the host-seeking behaviour of mosquitoes. The segmented antennae and, to a lesser extent, the maxillary palps have numerous sensillae, mostly olfactory, which are responsible for detecting stimuli and eliciting specific behaviour patterns from the mosquito. Feeding selectivity is based on attraction by warmth, moisture, carbon dioxide, and constituents of sweat. Human odour contains 33 chemical signalling compounds, but at least 5 are repellent to mosquitoes.

Preferred habitat for breeding

Mosquitoes have exploited a wide range of aquatic breeding habitats. In Africa, An. gambiae breeds extensively in any small, open, clean water body, including standing water in cattle hoofprints, while An. funestus breeds in larger, open, clean water bodies such as small ponds. Female mosquitoes lay one to three batches of 30 to 200 eggs during their lifespan, allowing an explosive increase in mosquito numbers from a relatively small number of females once breeding conditions become favourable.

Vectorial capacity and transmissibility

Mosquitoes are most efficient as vectors when the interval between parasite ingestion and its transmission to the next human host (extrinsic incubation period) coincides with the periodicity of female mosquito blood feeds associated with egg production. Vectorial capacity is defined as the average number of potentially infective bites delivered by all the mosquitoes feeding on a single host within 1 day. The numbers of mosquitoes feeding depends on mosquito density in relation to host density and the probability that the mosquito feeds on a host in any 1 day. The feeding frequency on humans is related to the proportion of meals taken on humans compared to other potential hosts. 

A measure of the proportion of mosquitoes taking a meal from an infected human that actually become infective is often added to this equation. This is a measure of the genetic and physiological competence of the mosquito. Small changes in the probability that the vector feeds on a host in 1 day (a), the duration of the extrinsic incubation period (n), and the probability that the mosquito will survive 1 day (p) produce large changes in vectorial capacity. This outcome led MacDonald to predict, as early as 1957, that adulticides would be more effective than larvicides in reducing malaria transmission rates, a lesson that has been relearned many times since by successive generations of entomologists. A related measure of the transmissibility or ability of an infectious agent to spread in a population is the basic reproduction number (R 0). R 0 is generally defined as the expected number of hosts who would be infected after one generation of the parasite by a single infectious person who had been introduced into an otherwise naive population. If R 0 is greater than 1, the number of people infected by the parasite increases; and if R 0 is less than 1, the number declines. The value of R 0 and the proportion of the mosquito population that is refractory to infection ultimately explain whether malaria will spread or be eliminated.

To monitor insect infection rates a number of different techniques can be deployed. The gold standard is the dissection and microscopical examination of blood-fed female mosquitoes, but the polymerase chain reaction (PCR) is used increasingly. The recent discovery of a much higher level of human transmission of P. knowlesi in Borneo also shows how reliant techniques such as microscopy are on the ability of microscopists to differentiate between what they believe they should see and what they actually see. New molecular techniques that track specific single nucleotide polymorphism (SNP) patterns now make it practicable to follow the emergence and spread of a disease outbreak and should reduce the level of parasite misclassification in both humans and mosquitoes. However, the specificity and sensitivity of such tests needs careful analysis and is often poorly understood by field practitioners. Sensitivity is the probability that a test will correctly identify an infected host; specificity is the probability that a test will correctly identify organisms. PCR-based tests are available for the four human malaria parasites and P. knowlesi. A valid concern with this type of molecular method is that, although they detect the presence of pathogen nucleic acids with great sensitivity, they may not be well correlated with the presence or abundance of viable pathogens. This may be complicated by issues of vector competence, when the mosquito is infected with the parasite but is incapable of transmitting it.

Distribution and density of mosquito populations

Mapping mosquito populations is important for guiding epidemiological activities within a study area. Advances in remote sensing should improve this process. In Kenya, multitemporal meteorological satellites have been used to predict periods when malaria transmission is likely to increase, based on correlations between advanced high-resolution satellite-derived indices of vegetation biomass and mosquito abundance. Use of such remote sensor data will allow studies of large, remote geographical areas to which access is difficult. Logistical growth models have been used to estimate the mosquito population carrying capacity of a given environment. Within this model, the extent to which mosquito births and deaths are conditioned by density is referred to as density dependence. However, the role of density dependence in natural populations is controversial. Other factors, such as predation, interspecies competition, and disease, may intervene to regulate population size long before it reaches its carrying capacity. Density-independent factors that influence population growth include environmental conditions such as food availability, adverse weather, extremes of temperature and relative humidity, and insecticide treatment programmes aimed at changing the age structure/size of mosquito populations. However, some insecticide-based interventions act only at a personal level, failing to reduce insect populations sufficiently to produce a herd or population protection effect on humans.

Mosquito genetics and insecticide resistance

The genome sequence of An. gambiae was published in 2002. Important mosquito phenotypes that have a genetic basis include blood feeding preference, habitat choice, insecticide susceptibility, and vectorial capacity. Blood feeding involves expression of salivary gland proteins that promote vasodilatation, inhibit platelet aggregation, and prevent blood coagulation in the host. Vector competence is a complex trait involving ingestion, replication, and transmission of plasmodium. The absence of any one of several structural or biochemical properties of the female mosquito could render her incapable of supporting successful completion of the parasite’s life cycle. An important advance towards genetic characterization of mosquito refractoriness to parasite invasion in An. gambiae was the construction of a microsatellite map for quantitative trait loci. Insecticide resistance is genetically inherited. DNA-based systems are now available for identifying species, determining whether they are infected, identifying the source of their blood meal, and finding the most common insecticide resistance mechanisms. Microarray technology has speeded up the process of identification of metabolic genes that are over- or underexpressed in resistant insects. Different target site resistances can be detected by simple PCR, e.g. a simple SNP-based PCR assay can be deployed to detect the kdr-type pyrethroid resistance mechanism, which results from a single nucleotide change in the sodium channel of the insect’s nervous system and results in phenotypic resistance to DDT and to all pyrethroids in homozygotes. Heterozygous insects are phenotypically susceptible to all the insecticides. The resistance can be selected by exposure to either DDT or pyrethroids. Retrospective analysis of specimens in laboratory or museum collections demonstrated that the kdr mutation was first selected in the 1940s to 1960s by the use of DDT in West and East Africa but remained completely undetected. In West Africa, resistance spread dramatically and was heavily reselected by the introduction of the pyrethroids in the late 1970s. Today, kdr in An. gambiae in West Africa has been selected almost to completion. It has managed to move between the sibling species of the An. gambiae complex and is still spreading. In East Africa, resistance has been confined to small areas of Kenya and has so far shown little evidence of increasing its range. In southern Africa, despite the extensive use of DDT in indoor residual spraying programmes over many years, there is no evidence of kdr-type resistance in either An. gambiae or An. funestus.

There has been spectacular progress in producing transgenic mosquitoes in recent years. At least five mosquito species (An. gambiae, An. stephensi, An. albimanus, Aedes aegypti, and Culex quinquefasciatus) can be transformed using at least four transposable elements. Effector genes can abolish the mosquito’s vectorial ability. Introduction of such mosquito strains might reduce malaria transmission by replacing wild-type mosquitoes.

Mosquito immunity to malaria

For many years, it was assumed that the insects did not possess an immune system. However, mosquitoes express several elements of vertebrate-specific immune responses. The ookinetes penetrating the mosquito midgut epithelial cells induce some Anopheles species to produce nitric oxide synthetase; defensin, a Gram-negative bacterial binding protein; and a thioester-containing protein TEP-1, and to initiate several other enzymatic pathways that may ultimately lead to parasite death. As a result, only a small percentage of mature ookinetes manage to reach the basal lamina to form oocysts, which are themselves vulnerable to melanization and destruction.


Spatial limits of malaria

The probable maximum preintervention distribution of malaria (c.1900) reached latitudinal extremes of 64 ° north and 32 ° south. Human efforts to control malaria have restricted its distribution dramatically during the 20th century as shown by the reported limits in 1946, 1965, 1975, 1992, 1994, and 2002. These distribution maps were compiled largely from country reports and expert opinion arising from the network of regional offices of the World Health Organization (WHO). Although they are imperfect representations of the distribution of global malaria infection risk in space and time, they do highlight the progress of malaria control in the 20th century. Between 1900 and 2002, the combined effects of development and control have halved the area of human malaria risk from 53 to 27% of the Earth’s land surface. The number of countries and territories with populations of over 100 000 inhabitants exposed to some level of malaria risk fell from 140 to 106 during this time. However, population growth has increased the total number of people exposed to malaria risk from approximately 1 billion in 1900 to approximately 3 billion in 2002.

Renewed interest in global malaria control has been associated with a renaissance in mapping malaria risks. The Malaria Atlas Project has synthesized all available medical intelligence on areas of the world reportedly free from malaria risk and adjusted these limits to other factors that would not support transmission of either P. falciparum or P. vivax, including human settlement patterns, climate, and altitude. It has been estimated that in 2007 2.4 billion people were exposed to some risk of infection with P. falciparum across 87 countries. The true biological and medical extent of P. vivax is harder to map and estimate; adaptations of work published in 2006 suggest that 3.18 billion people may be exposed to P. vivax in 2007 across 63 countries. No efforts have yet been made to map the distributions of the other two human malarias, P. malariae and P. ovale. P. malariae is widespread and often overlooked and P. ovale largely replaces P. vivax in West Africa, where the population is resistant.

The variation in intensity of malaria transmission worldwide

Mosquito-related factors

Only mosquitoes that become infected and then survive for longer than the duration of the extrinsic cycle of the parasite (say 10 days) can pass on the infection. As mosquitoes of a given species have a relatively constant probability of dying during a day, regardless of their age, the longevity may be described by the probability of surviving through 1 day. It varies greatly between mosquito species and environments. Rainfall, temperature, ecology, human settlement patterns, and prevalence of effective control measures largely govern the abundance of malaria mosquito vectors and the development of the parasite in their salivary glands. The behavioural characteristics of mosquitoes make some of them more efficient vectors of malaria than others. The ability of An. gambiae to breed opportunistically in small collections of water in rural areas, feed and rest indoors, take frequent blood meals, and live for a relatively long period makes it the world’s most efficient malaria vector. However, urban areas provide environments that are less suited to the breeding of An. gambiae and so malaria transmission in the rapidly expanding conurbations of Africa is very low. In Africa, An. arabiensis is better adapted to semiarid areas but is a less efficient vector than An. gambiae, accounting for the lower transmission of P. falciparum observed at the fringes of the Sahara, southern Africa, and the Horn of Africa. In South Asia, An. culicifacies may feed only every third day. As few as 10% of its meals may be from people, resulting in a human-biting habit that is 15-fold lower than An. gambiae. The diversity of vectors is driven by habitat preferences and adapted behaviours (‘bionomics’), e.g. An. sundaicus, An. maculatus, An. balabacensis, and An. subpictus occupy specific geographical niches in Java (Indonesia). Their varying efficiencies as vectors of malaria account for most of the diversity of malaria in that archipelago.

Transmission and malarial endemicity

Within the geographical ranges of dominant vector species, there are huge variations in the likelihood that a mosquito is infected with malaria and the frequency with which they feed on humans. Across the central belt of tropical Africa, individuals may be challenged from less than once to over 1500 times each year. In other parts of the world, where malaria vectors are less efficient than in Africa, an individual might expect to be infected from once a year to every 10 years. Exceptions are found in New Guinea, several states in India, and smaller foci at forest fringes of Thailand, Vietnam, Cambodia, and Burma (Myanmar).The frequency of contact between humans and malaria-infected vectors is a fundamental epidemiological concept that drives the health impact of malaria and the choice of control strategies. The frequency of malaria parasite encounters experienced by communities (transmission) is expressed using a variety of epidemiological terms and measured using field studies of contact and infection in mosquitoes and humans. The term ‘endemicity’ is a general expression of the amount or intensity of malaria transmission in an area or community. ‘Epidemic malaria’ indicates a periodic or sharp increase in the amount of malaria in a given indigenous community. Precise information about the degree of endemicity must be based on quantitative and statistical concepts. Malaria transmission is also classified as stable or unstable. ‘Stable’ implies equilibrium; the prevalence of infection is persistently high and endemicity is relatively insensitive to aberrations in climate or local habitats. Under stable endemic conditions, variation in transmission is minimal over many years although seasonal fluctuations still do occur and transmission can continue even with very few vectors. Conversely, ‘unstable’ malaria is characterized by great variability in space and time.

R 0 is often the benchmark epidemiological measure of malaria transmission but it is rarely measured empirically in the field. Two related measures, more commonly used, are derived from sampling mosquitoes or young children. The entomological inoculation rate (EIR) measures the average number of infected bites that an individual might experience from local vectors in a unit of time (often expressed per year) and measured by catching mosquitoes inside and outside people’s houses and dissection of mosquito salivary glands to see if they are infected. The parasite rate (PR) represents the proportion of individuals (usually children aged 2–9 years) who have evidence of infection in their peripheral blood when sampled during a cross sectional study in the community. It is not strictly a rate but a proportional ratio of infected persons. The PR has been widely used to classify P. falciparum endemicity since the 1950s. Four commonly used terms indicate the prevalence of infection in children aged between 2 and 9 years: hypoendemic (< 10%); mesoendemic (11–50%); hyperendemic (51–75%); and holoendemic (>75%, when measured in infants but routinely measured in children aged 2–9 years). Most measures of malaria transmission are related, often nonlinearly. Classical epidemiological models of malaria transmission, based largely on infection and vectors, are gradually accommodating new concepts related to pathogenesis, virulence, disease outcomes, and heterogeneity of susceptibility and transmission. These new suites of mathematical models should provide a more elaborate framework for understanding the diversity of malaria as a public health problem and how best to tailor control methods to meet specific short and long-term transmission-dependent needs.

The changing epidemiology of malaria

Population growth and environmental change

In most parts of the world, the epidemiological background to clinical malaria is likely to change due to population growth, environmental changes (often the result of human activity, whether local or global), changing resistance of parasites to drugs, and the consequences of attempts at malaria control. Predicting the resources needed to meet international malaria control objectives in the near future must take account of increasing populations at risk of malaria and the changing pattern and intensity of land use. The rate of population growth is significantly higher in urban than rural areas; sometime before 2025, most Africans will live in cities. Urban growth will reduce malaria risk. The pressure on agricultural land as populations grow can lead to deforestation, while increases in irrigation and dams together with poor land management can lead to desertification. In Africa, deforestation rates in the 1990s exceeded those in South America and are projected to increase with the growing capacity of humans to exploit forest habitat. The impact of deforestation on malaria transmission depends on which vector is dominant locally. In Africa, deforestation could create a habitat favouring An. gambiae, a more efficient malaria vector than the forest mosquito An. moucheti. Deforestation is also benefiting An. darlingi, the most efficient malaria vector in the Americas, but in South-East Asia deforestation may reduce malaria transmission by An. dirus. In Africa, most of the 525 large and 45 594 small dams have been built since 1950. Their number will increase in the near future and may aggravate malaria transmission, e.g. the restoration of An. funestus to the Sahel, after a prolonged period of drought and desertification, has been attributed to irrigation.

Population movements

Human migration has been associated with malaria epidemics when population pressure in hilly areas drives the inhabitants down into malarious regions, when congregation of workers at new sites mixes infected with susceptible people, or when malnourished refugees, with impaired resistance to infection, camp where public health measures have collapsed. Regional conflicts result in large-scale population movements to avoid the ravages of war. During the mid-1990s, the exodus of nonimmune refugees from the nonmalarious highlands of Burundi to endemic areas of Tanzania resulted in an epidemic of severe malaria.

HIV epidemic

In sub-Saharan Africa, an HIV epidemic has been superimposed on an established malaria pandemic. Considering the wide geographical overlap and concurrent high prevalence of both infections, even a modest interaction could have substantial public health implications. HIV-infected adults in malaria-endemic areas and HIV patients of all ages in areas of unstable malaria transmission are at increased risk of malaria infection and death. In endemic areas, the case fatality of malaria is also higher in HIV-infected children. The impact of HIV on malaria depends on the level of malarial endemicity and, hence, the age patterns of clinical malaria and on the geographical distributions of HIV. Populations in southern Africa and urban areas of Africa are most vulnerable to this interaction.

The public health burden of malaria

Direct and indirect consequences of infection

The relationship between P. falciparum infection and disease outcome is complex. People born into areas of stable P. falciparum transmission frequently have periods when they are being infected with the parasite and periods when they remain uninfected. Most will, at some stage in their lives, develop a clinical response to infection, usually an attack of fever. This may resolve without any medical intervention, progress to severe disease with natural resolution, resolve through medical intervention, or end fatally. In areas where transmission is stable, less than 0.05% of infections prove fatal. This low case fatality is largely a result of combinations of innate genetic protection and acquired clinical immunity. There are, however, indirect consequences of malaria infection that are less effectively controlled by immunity. Chronic subclinical infections, e.g. due to incomplete parasite elimination with failing drugs, may lead to anaemia or other forms of malnutrition that independently increase susceptibility to severe effects of future infections. Subclinical infections may increase the severity of other infectious diseases. Asymptomatic infection of the placenta of a pregnant woman may significantly reduce the weight and hence the chances of survival of her newborn child. Patients who survive severe disease may be left with debilitating sequelae, such as spasticity or epilepsy, or more subtle consequences including behavioural disturbances or cognitive impairment. 

Epidemiological patterns of stable and unstable malaria

The epidemiological features of human malaria differ markedly even between endemic areas. At one extreme, as in holoendemic areas of tropical Africa, everyone is infected shortly after birth, parasitaemia is almost universal throughout childhood, and the brunt of mortality falls in early childhood; epidemics do not occur except at high altitude or during aberrant rainfall in semiarid areas.

Children living in the An. gambiae belt of tropical Africa will usually have their first malaria infection during their first 3 months of life when they still have maternal antibodies. The disease is very mild, consisting of just one or two peaks of fever that usually resolve without treatment. When they are between 4 and 6 months old, when maternal antibodies have waned, each new infection leads to more severe febrile illnesses. The repeated infection–fever cycle progressively increases the loss of RBCs to the parasite plus a suppressed ability to replace them. Combined with other severe pathological consequences such as metabolic acidosis and coma, approximately 1% of all children living in these areas will die before their third birthday. The survivors will continue to have frequent attacks of fever until about the onset of puberty. Many will experience between 20 and 50 malaria attacks before reaching their fifth birthday. Thereafter, the frequency of attacks of fever and the intensity parasitaemia slowly wanes until levelling out sometime in early adolescence.

From that point forward, people have an almost constant low parasitaemia despite an almost complete absence of symptoms. In adults, episodes of fever due to malaria occur only once in 2 to 10 years, and last only a few hours even without treatment. The one exception to this rule is malaria in pregnant women, especially those pregnant for the first time.

Malaria in pregnancy

It is estimated that, each year, over 30 million women become pregnant in malarious areas of Africa. In mesoholoendemic areas, pregnant women experience relatively little malaria-specific morbidity but have an increased risk of infection and higher density parasitaemia leading to anaemia and placental sequestration of parasites. Maternal anaemia is an important contributor to maternal mortality and it is estimated that 9% of the excess risk is directly attributed to P. falciparum infection. Prematurity and low birth weight (<2500 g) associated with maternal malaria have been reported indirectly to contribute to 3 to 8% of infant mortality. Interactions between malaria and HIV are particularly important among pregnant women who, when coinfected, have an increased risk of clinical attacks of malaria, anaemia and giving birth to a low birth weight baby. It is estimated that 500 000 women will develop clinical malaria in Africa as a result of their HIV infection. It remains uncertain whether malaria during pregnancy increases the vertical transmission of HIV.

Areas of unstable malaria

At the other end of the spectrum, as in parts of South-East Asia, where malaria is unstable or hypoendemic disease affects all ages, the risk of clinical illness is almost directly proportional to the risk of infection, and the risks of acquiring the parasite and developing a clinical event are time and space dependent throughout heterogeneous foci. Despite differences in the age patterns of disease with differing levels of endemicity, there is a much lower overall incidence of clinical disease and death in communities located in areas of low transmission intensity (hypomesoendemic) than in those experiencing moderate-to-high transmission (hyperholoendemic). However, during epidemics, disease burdens in lower transmission areas can be devastating, disrupting livelihoods. Overall, as the annual risk of new infections increases from zero, through to low-to-moderate risks of new infections (say one or two new infections per year), the rates of disease increase proportionately and probably linearly. As transmission intensity increases through hyper- to holoendemic conditions, the relationship between the annual risk of infection and the rates of disease is less clear, probably nonlinear and may reach a plateau after about 10 new infections per year. To understand the clinical spectrum of malaria seen in patients from a given locality, it is essential to understand the local epidemiology.

Global malarial morbidity and mortality

Although it is convenient to classify entire regions of the world by levels of endemicity, the low intensity transmission characteristics of most areas of South-East Asia resemble large swathes of Africa. Similarly, there are areas of South-East Asia that might be regarded as typical of the central An. gambiae belt of Africa. Precise estimates of the numbers of clinical cases and deaths due to malaria are notoriously poor in all malaria endemic countries. National data on malarial illness and death are characterized by gaps and inaccuracies due to under-reporting and misdiagnosis. Deaths usually occur outside the formal health sector and many clinical events are self medicated.

Given the weaknesses in international malaria disease reporting, estimations have been made of the public health burden using modified historical and climate-driven maps of malaria transmission, projected population human settlement counts, and endemicity-specific epidemiological survey data of clinical attacks and deaths due to P. falciparum. These approaches suggest that there were approximately 550 million clinical attacks of P. falciparum worldwide in 2002, distributed by WHO regions as follows: 71% in the Africa region, 23% in the low-transmission but densely populated countries of South-East Asia, 3% in the western Pacific, 0.7% in the Americas; 2.3% in the eastern Mediterranean (including Somalia and Sudan), and 0.1% in the European region. Less is known about global P. falciparum mortality and still less about the neglected disease burdens posed by P. vivax, despite its wider distribution. In Africa, application of similar epidemiological disease burden models to assess the wider public health consequences of P. falciparum suggested that in 2005 P. falciparum caused 1.1 million deaths directly, between 71 000 and 190 000 infant deaths following placental infection in utero, and over 3000 newly acquired persistent epilepsies through brain insults among patients surviving an episode of cerebral malaria in childhood.

Susceptibility to infection and innate resistance

In endemic areas, malaria is thought to account for around one-quarter of all childhood deaths so that this infection has been a major selective force in human evolution. More human genetic polymorphisms have been associated with innate protection from malaria than with any other infectious disease (Table 1). The best known is sickle cell haemoglobin, due to a point mutation in position 6 of the β-globin chain. Here, the mutant-gene frequency is stabilized because the enhanced survival of AS heterozygotes is counterbalanced by the lethal consequences of homozygosity (SS). The protection afforded to AS heterozygotes seems to act predominately on the development of clinical disease. Although infection rates are similar for AA and AS genotypes, the AS genotype is almost completely protected against life-threatening P. falciparum malaria. α-Thalassaemia, resulting from deletion of α-globin genes, is less strongly protective against P. falciparum malaria but, as most forms are virtually asymptomatic, very high gene frequencies have developed in some parts of the world such as Oceania. However, in Africa, where the selective pressure is greatest, the gene frequency is less (typically c.40%). This may be explained by the fact that when sickle cell trait and α-thalassaemia coexist, far from acting synergistically, they cancel out each other’s protection, a striking example of negative epistasis.

Many other genetic variants affecting the RBC have been associated with protection from malaria, including other variants of the β-globin gene (HbC and HbE), polymorphisms of the RBC enzymes glucose-6-phosphate dehydrogenase (G6PD) and pyruvate kinase, and variants of structural proteins (erythrocyte band 3), which cause South-East Asian ovalocytosis.

The Duffy (blood group) antigen receptor for chemokines (DARCFy), expressed on the surface of RBCs, is also a receptor for penetration of RBCs by merozoites of P. vivax. The extreme rarity of the DARC +46C/C genotype is responsible for the striking resistance to this parasite of people of West African origin. However, these same receptors influence plasma levels of HIV-1 suppressive and proinflammatory chemokines such as CCL5/RANTES and are the site of HIV-1 attachment to RBCs, affecting by HIV adsorption the transinfection of target cells. In African Americans, possession of the prevalent DARC –46C/C genotype was found to be associated with a 40% increase in the risk of acquiring HIV-1. If extrapolated to all Africans, approximately 11% of the HIV-1 burden in Africa may be linked to this genotype.

Many potentially protective polymorphisms affecting other key aspects of malaria–human interaction have now been identified. These include polymorphisms affecting endothelial receptors, cytokines, and other key molecules of the immune system. Any listing of putative protective polymorphisms, such as Table 1, is bound to become quickly out of date. There is now an international coordinated effort to apply whole genome scanning approaches to identify key polymorphisms conferring protection against malaria. Regularly updated information can be found at

Acquired resistance

The prevalence of asymptomatic infection remains high for many years and even in adulthood a substantial proportion of people are infected at a single time point. It is unlikely that anyone ever achieves sterilizing immunity. However, the most important aspects of immunity, the ability to avoid severe disease and death, develop much faster. In most stably endemic areas, death due to malaria is rare after the age of 5 and hardly ever occurs in normally immunocompetent adults.

A characteristic of immunity to malaria is the early acquisition of the ability to tolerate levels of parasitization that would cause illness in a naive individual. This presumably involves either, or both, a immune response to neutralize parasite toxins or a down-regulation of the host’s normal cytokine response to challenge. However, it is important to recognize that such ‘antitoxic’ responses cannot of themselves be the mainstay of protection against serious morbidity, for if the parasite population continued to expand, the host’s RBC population would soon be overwhelmed leading to severe anaemia. Thus, even at an early stage, the key aspect of immunity to malaria is an acquired ability to control parasite growth, either by interfering with the replication of parasite or by accelerating the removal from the circulation.

Emphasis is often given to the strain-specific nature of malarial immunity, with the idea that immunity depends on having to acquire a repertoire of responses to different ‘strains’ of the parasite. However, because the parasite population is constantly outbreeding, it is more accurate to think of polymorphism in key molecular targets of protective immune responses rather than of fixed ‘strains’. For some of these polymorphic targets, i.e. parasite-derived molecules expressed on the outside of the infected RBC membrane, children do build up a repertoire of responses. New infections tend to be those that the individual has not yet seen. However, there is also considerable evidence that some protective responses are cross-protective from an early stage. It is likely that the development of immunity involves the progressive acquisition of both strain-specific and cross-protective responses to a range of potential targets.

Potentially, effective immune responses could act at any point in the parasite’s life cycle. Immunity is largely stage specific, i.e. immunity induced by the sporozoite stage that can prevent infection probably has little effect on the blood stages. Similarly the targets and mechanisms responsible for immunity to gametocytes (which prevent transmission) are separate from those responsible for immune response to pre-erythrocytic and erythrocytic stages. Experimental infection of humans and animals with attenuated sporozoites leads under some circumstances to complete immunity to infection (see ‘Malarial vaccines’, below). However, there is little evidence that such responses play a major role in naturally acquired immunity to malaria and, in fact, adults in endemic areas become infected at similar rates to children (albeit without progressing to clinical illness).

Naturally acquired immunity to malaria probably acts predominantly at the blood stages of the parasite’s cycle. Here, the parasite is present briefly as a free merozoite before spending most of its time apparently hidden inside the host RBC. Effective responses are likely to depend either on antibody-mediated mechanisms or on indirect cellular effects involving the release of a range of mediators. The potential importance of antibody mediated responses to blood-stage parasites was demonstrated in classical studies in which gammaglobulin from immune adults was shown to markedly ameliorate attacks of malaria in children. Similarly, the relative protection of children in the first few months of life in endemic area is thought to be due in part to passively transferred maternal antibody. A number of potentially key targets for protective immune responses have been identified. These tend to be either molecules involved in the invasion of RBCs by merozoites or parasite-induced molecules inserted on the surface of the infected RBC during the second half of its blood stage. Most of these antigens show considerable antigenic polymorphism and those growing up in endemic areas develop a wide range of antibodies to many antigenic types. Increasingly it appears that protection from clinical malaria may stem from multiple mechanisms and that the breadth and quality of the immune response to a number of antigens may be more important than responses to any single antigen. Humans also make immune responses to several gametocyte specific antigens. However, although experimental approaches to interrupting transmission in animal models hold promise for transmission blocking vaccines, the evidence that such responses play an important role in nature is controversial.

Immunity is diminished after long periods of absence from endemic areas and in pregnancy. Previously immune pregnant women are at risk of parasitization of the placenta, resulting in spill-over peripheral parasitaemia, increasing anaemia, and impairment of placental function leading to low birth weight babies. This results not from a breakdown in previously established immunity but from the specific appearance of a new site for sequestration of parasites. Parasitized RBCs (PRBCs) recovered from the placenta express a specific subset of PfEMP1 molecules that bind to chondroitin sulphate A (CSA) on the syncytiotrophoblast, leading to the accumulation of sometimes massive numbers of metabolically active parasites. The PfEMP1 molecules responsible for this adhesion are rarely encountered in infections outside pregnancy and so women enter their first pregnancy without specific antibodies. These are acquired during the course of the first and subsequent pregnancies so that effects of placental parasitization and its consequences decrease progressively with ensuing pregnancies.

Malaria and HIV immunosuppression

HIV infection modifies the response to malaria under a number of situations. In pregnant women, the two conditions seem to be mutually synergistic with the prevalence and consequences (particularly anaemia) being more severe in HIV-positive women. In nonimmune adults in areas of unstable transmission, HIV positivity is associated with an increased risk of clinical malaria and with increased risk of death in those who develop it. Recently it has become clear that HIV positivity in children in endemic areas is strongly associated with increased risk of being admitted to hospital with severe malaria.

Malaria and malnutrition

For many years the relationship between malnutrition and malaria was contentious, with claims that it was associated with protection from severe disease. Although there may be situations where very severe nutritional deficiencies may be associated with reduced risk, in general malnutrition is an important risk factor for severe malaria in children in endemic areas.

Molecular Pathology

All the pathology associated with malaria infection is attributable to asexual parasite multiplication in the bloodstream. The consequences to the host of the intraerythrocytic multiplication of parasites range from a variety of severe, but not life threatening, symptoms common to all the species that infect humans, to the potentially lethal complications particularly associated with acute P. falciparum infection.

The characteristic symptom of malaria is fever and this is probably initiated by a combination of stimuli involving parasite products interacting with pattern recognition receptors, such as Toll-like receptors, and cell surface receptors, such as CD36, on host cells. Parasite products involved include variant parasite molecules expressed on the surface of PRBCs, such as PfEMP1, glycosylphosphatidylinositol anchors which are found on many plasmodium membrane proteins, and haemozoin. These interactions lead to the stimulation of both pro- and anti-inflammatory cytokine cascades, the balance of which may determine the relative outcome in terms of antiparasite effect and host pathology. At one end of the spectrum lies a relatively mild, self-limiting illness; at the other is an attack of severe disease that shares many of the pathological features of sepsis, in which overvigorous or disordered immune responses play a key role. Thus, while tumour necrosis factor (TNF) is protective against the parasite, many studies show an association between severe disease and exaggerated proinflammatory cytokine responses, including TNF, interleukin (IL) IL-1β, IL-6, IL-10, and interferon-γ as well as the macrophage inflammatory protein chemokines MIP1α and MIP1β.

Infection with all species of malaria induces fever, but the acute illness with P. malariae or P. ovale is relatively self limiting. Although P. vivax is traditionally considered a relatively benign parasite, the acute illness can be quite severe and it is increasingly realized that deaths due to P. vivax infection do occur. However, it is P. falciparum that is responsible for the majority of severe disease and death. The principal life-threatening complications of P. falciparum in African children are cerebral malaria and severe anaemia often associated with metabolic acidosis and respiratory distress. The clinical picture in nonimmune adults is more complex and can include single or multiple organ failure. A key difference in the biology of P. falciparum believed to play a central role in its enhanced virulence is its propensity to undergo sequestration. Only the younger developmental stages of the parasite circulate, as the more mature forms adhere to specific receptors on venular endothelium. Parasite sequestration occurs in many capillary beds and is often particularly intense in the lungs, brain, intestines, and muscles. The resultant reduction in, or obstruction of, local blood flow probably results in reduced perfusion and tissue damage. However it seems likely that this is just one part of a complex set of responses set in train by the interaction of sequestered cells and endothelial cells and the cells of the immune system leading to local release of a number of potentially toxic or pharmacologically active compounds (such as reactive oxygen species or nitric oxide).

Several endothelial receptors for infected RBC cytoadherence have been identified, including CD36 (formerly platelet glycoprotein IV), thrombospondin, ICAM-1, VCAM-1, and E-selectin. No clear correlation has yet emerged between the ability of parasites to bind to individual receptors and disease pattern (other than in the case of pregnancy-associated malaria), though there is suggestive evidence that severe disease in children maybe associated with the ability of parasites to utilize multiple receptors. Some parasite isolates show two other adhesive properties: the rosetting of uninfected erythrocytes around RBCs containing mature developmental forms of the parasite and autoagglutination of infected erythrocytes in the absence of immune serum. Both phenomena have been linked to severe malaria and it is presumed that the multicellular aggregates, if they occur in vivo, may exacerbate vascular obstruction caused by sequestration.

On the parasite side of the equation, PfEMP1 plays a key role in cytoadherent interactions, with binding to different receptors localized to different domains of the molecule. The case of pregnancy-associated malaria offers one example where a specific set of PfEMP1 molecules with the ability to bind to a specific receptor (chondroitin sulphate) forms the basis of organ specific biology and pathology. It remains to be seen whether similar subsets of PfEMP1 molecules will be identified associated with other clinical syndromes.

Severe anaemia is a common part of the picture of severe malaria, especially in young children. Although destruction of parasitized RBCs per se, especially in a heavy infection, may lead to a significant fall in haemoglobin, it has long been recognized that this cannot account for the often profound degree of anaemia. Two additional processes seem to be important: the sensitization of RBCs with immune complexes and activated components of the complement system leading to immune mediated removal of noninfected cells and a degree of dyserythropoiesis.

Although an episode of P. falciparum malaria is potentially life threatening, in endemic areas the large majority of clinical episodes resolve (albeit after an unpleasant illness) without producing severe disease and death. Clearly behavioural factors such as treatment-seeking behaviour are important, but it also seems likely that the wide range of outcomes represent a balance between host and parasite specific factors. In the end, it may be that severe disease represents the unfortunate coincidence of the wrong host with the wrong parasite.

Organ pathology


Probably only falciparum malaria causes cerebral pathology although P. vivax-infected RBCs may also be sequestered. At autopsy, the brain may be oedematous, especially in African children, but there is rarely any evidence of cerebral, cerebellar, or medullary herniation. Small blood vessels, including those of the leptomeninges, are congested with PRBCs. Many of the parasites are schizonts and mature trophozoites containing malaria pigment, giving the surface of the brain a characteristic leaden or plum-coloured appearance. Its cut surface is slate grey. In larger vessels, PRBCs form a layer along the endothelium (‘margination’). Up to 70% of RBCs in the cerebral vessels are parasitized and are more tightly packed than in other organs. The cerebrovascular endothelium shows pseudopodial projections, closely apposed to electron-dense, knob-like protruberances on the surface of PRBCs. Numerous petechial haemorrhages are seen in the white matter, the result of bleeding from end arterioles, proximal to occlusive plugs of PRBCs, and fibrin. Ring haemorrhages are centred on small subcortical vessels. They may organize, attracting small collections of microglial cells around an area of demyelination without inflammatory cells (Dürck’s granulomas).


Retinal whitening is associated with swelling of neurons secondary to local hypoxia and haemorrhages are caused by blockage of small retinal vessels with PRBCs and microthrombi.

Bone marrow

In the acute phase of falciparum malaria, there is iron sequestration, erythrophagocytosis, dyserythropoiesis, and cytoadherence with plugging of sinusoids. Maturation defects are present in the marrow for at least 3 weeks after clearance of parasitaemia. Increased numbers of large, abnormal-looking megakaryocytes have been found in the marrow and the circulating platelets may also be enlarged, suggesting dysthrombopoiesis. Malarial pigment and parasites may be found in monocytes and phagocytes in the marrow, even when they are not detectable in peripheral blood.


The liver is most severely affected in P. falciparum malaria. It becomes enlarged and oedematous and is coloured brown, grey, or even black from deposition of malaria pigment. Hepatic sinusoids are dilated, containing hypertrophied Kupffer cells and PRBCs obstructing the circulation. Parasitized and uninfected RBCs are phagocytosed by Kupffer cells, endothelial cells, and sinusoidal macrophages. The small areas of centrilobular necrosis present in severe cases may be attributable to shock or disseminated intravascular coagulation. Hepatocytes appear only mildly abnormal but are depleted of glycogen in some hypoglycaemic patients. Lymphocytic infiltration of portal tracts has been described in some cases of tropical splenomegaly syndrome, a chronic immunological complication of malaria.

Gastrointestinal tract

Cytoadherent, sequestered, PRBCs may be found in the small and large bowel, especially in capillaries of the lamina propria and larger submucosal vessels. The bowel may appear congested, with mucosal ulceration and haemorrhage.


Renal failure, with or without ‘blackwater fever’, is a common and serious complication of severe falciparum malaria in some populations. It is usually associated with acute tubular injury rather than glomerulonephritis. Glomerular lesions consist of mild accumulations of monocytes within glomerular capillaries (acute transient glomerulonephritis) without immune complex deposition. There is PRBC sequestration in glomerular and tubulointerstitial vessels, with fibrin thrombi and pigment-laden macrophages. Tubular pigment casts are prominent in cases of blackwater fever and severe rhabdomyolysis.

Levels of parasite sequestration in the kidney are usually relatively low. They correlates with premortem renal failure, and are significantly higher in malaria-associated renal failure than in fatal cases without renal failure. In quartan malarial nephrosis, a chronic immunological complication of malaria, a distinctive chronic glomerulonephritis develops.


At autopsy, the lungs are found to be oedematous in almost every case. Pulmonary capillaries and venules are packed with PRBCs and inflammatory cells: neutrophils, plasma cells, and pigment-laden macrophages. The capillary lumen is narrowed by oedema of vascular endothelium and there is interstitial oedema and hyaline-membrane formation. Secondary bronchopneumonia is commonly found.


The spleen is enlarged, engorged, and coloured dark red or grayish black. The red and white pulp is congested and hyperplastic, and the splenic cords and sinuses are filled with phagocytic cells containing pigment, PRBCs, and noninfected RBCs. Macrophages may extract the parasites from PRBCs, a process known as ‘pitting’. Tropical splenomegaly syndrome is a chronic immunological complication of malaria (see below).


Myocardial capillaries are congested with pigment-laden macrophages, lymphocytes, plasma cells, and PRBCs but these are not tightly packed or cytoadherent. Subendocardial and epicardial petechial haemorrhages are unusual and there is no myocarditis.


Sinusoids are packed with PRBCs and pigment-laden macrophages, giving the placenta a black or slate-grey hue. Necrosis of the syncytiotrophoblast, fibrinoid necrosis, loss of villi, proliferation of cytotrophoblastic cells, and thickening of the trophoblastic membrane may explain impaired fetal nutrition. Although transmission of infection across an intact placenta is considered uncommon, PRBCs are sometimes visible in fetal–placental vessels.


Anaemia and thrombocytopenia

Malarial anaemia results from destruction/phagocytosis of PRBCs and dyserythropoiesis. Hyperferritinaemia, an acute-phase reaction, explains the initial iron sequestration and hypoferraemia. Immune-mediated haemolysis occurs in some populations. Erythrocyte survival is reduced even after the disappearance of parasitaemia and there is increased splenic clearance of nonparasitized RBCs and PRBCs. Evidence of Coombs’ test-positive haemolysis was found in the Gambia. Intravascular haemolysis occurs in patients whose erythrocytes are congenitally deficient in enzymes such as G6PD in response to oxidant drugs such as primaquine. In classic blackwater fever, G6PD levels are, by definition, normal and the mechanism of haemolysis is unknown. Quinine mediated haemolysis is suspected but has never been satisfactorily demonstrated. Thrombocytopenia is attributable to splenic sequestration, dysthrombopoiesis, and immune-mediated lysis.

Cerebral malaria

In Thai adults with cerebral malaria, global cerebral blood flow was inappropriately low and there was evidence of cerebral anaerobic glycolysis with increased lactate concentrations in the cerebrospinal fluid. Recently, in Bangladeshi patients with severe malaria, orthogonal polarization spectral imaging was used directly to observe the rectal mucosa, as a surrogate for the cerebral microcirculation. Microcirculatory obstruction (proportion of vessels involved and the degree of obstruction) correlated with disease severity and decreased on clinical recovery. Vessels with little or no blood flow were often seen adjacent to vessels with hyperdynamic blood flow.

In African children with cerebral malaria, plasma concentrations of TNFα, IL-1α, and other cytokines correlate closely with disease severity, as judged by parasitaemia, hypoglycaemia, case fatality, and the incidence of neurological sequelae. Cytokines may have other effects on cerebral function, perhaps by releasing nitric oxide, which interferes with neurotransmission, or by leading to the generation of free oxygen radicals. Cytokines may also cause fever, hypoglycaemia, coagulopathy, dyserythropoiesis, and leucocytosis in falciparum malaria.

In South-East Asian adults, the opening pressure of cerebrospinal fluid at lumbar puncture was usually normal. Cerebral oedema was demonstrable by CT during life in only a small minority, usually as an agonal phenomenon. In these patients, there was little evidence of increased blood–cerebrospinal fluid barrier permeability or that brain swelling was responsible for coma. However, in African children with cerebral malaria, intracranial pressure, as reflected by cerebrospinal fluid opening pressure at lumbar puncture, is usually elevated and the majority have swollen brains. In fatal cases, the brain shows evidence of increased vascular permeability. Ischaemic damage, resulting from a critical reduction in cerebral perfusion pressure, hypoglycaemia, and status epilepticus, probably contributes to brain damage in these children.

Pulmonary oedema

This may be provoked by fluid overload, in which case, central venous and pulmonary artery wedge pressures will be elevated. More commonly, the clinical picture is of acute respiratory distress syndrome (ARDS), with normal or low hydrostatic pressures in the pulmonary vascular bed. In these cases, the mechanism is likely to be increased pulmonary capillary permeability resulting from leucocyte products and cytokines, consistent with the histological appearances of neutrophil sequestration in the pulmonary capillaries, increased permeability, and hyaline membrane formation.

Hypoglycaemia and other metabolic disturbances

Cinchona alkaloids (quinine or quinidine) are potent stimulators of insulin secretion by the pancreatic β-cells, causing hyperinsulinaemia. The resulting reduction in hepatic gluconeogenesis and increased peripheral glucose uptake by tissues causes hypoglycaemia. In malaria, glucose consumption is increased by fever, infection, anaerobic glycolysis, and the metabolic demands of the malaria parasites. Glycogen reserves may be depleted, especially in children and pregnant women, as a result of fasting and ‘accelerated starvation’. In African children with severe malaria, adult patients with severe disease, and pregnant women, hypoglycaemia develops spontaneously (without treatment with cinchona alkaloids) and is associated with appropriately low plasma insulin concentrations. Plasma lactate and alanine concentrations are elevated and ketone bodies are moderately increased. Counter-regulatory hormone levels are usually very high. The mechanism of hypoglycaemia in these cases may be inhibition of hepatic gluconeogenesis by TNFα and other cytokines. In African children, severe anaemia, tissue hypoxia, hypoperfusion, and increased anaerobic glycolysis by host and parasites contribute to profound metabolic (lactic) acidosis, manifesting clinically as respiratory distress.

Acute renal failure

Oliguria and renal dysfunction reversible by fluid replacement are attributable to hypovolaemia resulting from dehydration. Hyperparasitaemia, jaundice, and haemoglobinuria are risk factors for acute tubular necrosis. Renal cortical perfusion is reduced during the acute stage of the disease but renal cortical necrosis is rare and survivors rarely show evidence of chronic renal impairment. Cytoadherence of PRBCs in the renal microvasculature, deposition of fibrin microthrombi, prolonged hypotension (‘algid malaria’), haemoglobinuria in ‘blackwater fever’ and myoglobinuria may contribute to acute renal failure. Quartan malarial nephrosis is a chronic immunological complication of malaria (see below).


In patients with relatively normal plasma osmolalities, hyponatraemia has been attributed to the inappropriate secretion of ADH triggered by fever or reduced effective plasma volume. However, in Thai patients, ADH levels were appropriate to their gross hypovolaemia. This has been confirmed in Bangladeshi patients. In those who are salt depleted and dehydrated, mild hyponatraemia is often attributable to intravenous therapy with 5% dextrose.

Hypovolaemia and ‘shock’ (‘algid malaria’)

Hypotension may be explained by hypovolaemia (dehydration and, rarely, haemorrhagic shock following splenic rupture or gastrointestinal haemorrhage) but is most often associated with a secondary Gram-negative bacteraemia. The source may be an intravenous cannula, urethral catheter, aspiration pneumonia, or invasion of the bloodstream by an enteric pathogen such as salmonella. Transient immunosuppression, impaired macrophage function, or ‘blockade’ of the reticuloendothelial system may increase the susceptibility of patients to severe secondary bacterial infections.

Clinical features in adults and children

Malaria is typically an acute febrile illness that, if incompletely treated, tends to recrudesce or relapse over periods of months or even years. The classic periodic febrile paroxysms—occurring every 24 h (quotidian), 36 h (subtertian), 48 to 50 h (tertian), or 72 h (quartan) with afebrile asymptomatic intervals—are rarely observed unless treatment is delayed. Severity depends on the species and strain and, hence, on the geographical origin of the infecting parasite, on the age, genetic constitution, state of immunity, general health, and nutritional state of the patients, and on the speed and appropriateness of antimalarial treatment.

Falciparum malaria (‘malignant’ tertian or subtertian malaria)

The ‘prepatent period’, the shortest interval between an infecting mosquito bite and detectable parasitaemia, is usually 9 or 10 days but may be as short as 5 days (Table 1. The incubation period, the interval between infection and the first symptom, usually ranges from 7 to 14 days (mean 12 days) but may be prolonged by immunity, chemoprophylaxis, or partial chemotherapy. In Europe and North America, 98% of patients with imported falciparum malaria present within 3 months of arriving back from the malarious area. A few present up to 1 year later, but none after 4 years.

Several days of prodromal symptoms such as malaise, headache, myalgia, anorexia, and mild fever are interrupted by the first paroxysm. Suddenly the patient feels inexplicably cold (in a hot climate) and apprehensive. Mild shivering quickly turns into violent shaking with teeth chattering. There is intense peripheral vasoconstriction and gooseflesh. Some patients vomit. The rapid increase in core temperature may trigger febrile convulsions in young children. The rigor lasts up to 1 h and is followed by a hot flush with throbbing headache, palpitations, tachypnoea, prostration, postural syncope, and further vomiting while the temperature reaches its peak. Finally, a drenching sweat breaks out and the fever defervesces over the next few hours. The exhausted patient sleeps. The whole paroxysm is over in 8 to 12 h, after which the patient may feel remarkably well. These symptoms are typical of a classical ‘endotoxin reaction’ produced by infection with Gram-negative bacteria or the release of TNFα and other cytokines by other agents. Classic tertian or subtertian periodicity is rarely seen with falciparum malaria. A high irregularly spiking, continuous, or remittent fever or daily (quotidian) paroxysm is more usual. Other common symptoms are headache, backache, myalgias, dizziness, postural hypotension, nausea, dry cough, abdominal discomfort, diarrhoea, and vomiting. Nonimmune patients with falciparum malaria are usually severely unwell. Commonly, there is anaemia and mild jaundice, with moderate tender enlargement of the spleen and liver. Useful negative findings are the lack of lymphadenopathy, rash (apart from herpes simplex ‘cold sores’), and focal signs.

Severe falciparum malaria

WHO (2000) has defined severe disease by the clinical and laboratory features shown in Bullet list 2.

Cerebral malaria

The global average case fatality of falciparum malaria is about 1%, or 1 to 3 million deaths per year. Cerebral malaria is an important severe manifestation of P. falciparum infection, accounting for a large proportion of adult deaths. Patients who have been feverish and ill for a few days may have a generalized convulsion from which they do not recover consciousness, or their level of consciousness may decline gradually over several hours. High fever alone can impair cerebral function causing drowsiness, delirium, obtundation, confusion, irritability, psychosis, and, in children, febrile convulsions. The term ‘cerebral malaria’, implying encephalopathy specifically related to P. falciparum infection, should be restricted to patients: (1) who are unrousable and comatose, showing no appropriate verbal response and no purposive motor response to noxious stimuli (Glasgow Coma Scale ≤9/14); (2) who have evidence of acute P. falciparum infection; and (3) in whom other encephalopathies, including hypoglycaemia and transient postictal coma, have been excluded. Mild meningism may be elicited but neck rigidity and photophobia are rare. Retinal abnormalities (Fig. are best seen with the pupils dilated with 0.5 to 1% tropicamide and 2.5% phenylephrine. Haemorrhages like Roth spots, papilloedema, or exudates are present in about 15% of South-East Asian adults with cerebral malaria. In African children with cerebral malaria, retinal changes include macular and peripheral retinal whitening, vessel changes (orange vessels, tramlining, capillary whitening), retinal haemorrhages, papilloedema, and cotton wool spots. Of these changes, the whitening and vessel changes are specific (‘malarial retinopathy’) and are associated with a case fatality of about 18%, compared with 44% in children with papilloedema and 7% in those with normal fundi.

Bullet list 2 Features of severe falciparum malaria

Clinical manifestations
  • Prostration
  • Impaired consciousness
  • Respiratory distress (acidotic breathing)
  • Multiple convulsions
  • Circulatory collapse
  • Pulmonary oedema (radiological)
  • Abnormal bleeding
  • Jaundice
  • Haemoglobinuria

Laboratory tests

  • Severe anaemia (haemoglobin <5 g/dl or haematocrit <15%)
  • Hypoglycaemia (blood glucose <2.2 mmol/litre or 40 mg/dl)
  • Acidosis (plasma bicarbonate <15 mmol/litre, or base excess more than –10, or arterial pH <7.35)
  • Renal impairment (urine output <12 ml/kg/h, or plasma creatinine above age-related normal range; persisting after rehydration)
  • Hyperlactataemia (plasma lactate >5 mmol/litre)

In adult patients, pupillary, corneal, oculocephalic, and oculovestibular reflexes are normal. Dysconjugate gaze is common. Muscle tone and tendon reflexes are usually increased and there is ankle clonus. Plantar responses are extensor and abdominal reflexes are absent. In African children, brainstem reflexes may be abnormal and there may be neurological evidence of severe intracranial hypertension with rostrocaudal progression suggesting cerebral, cerebellar, and medullary herniation. Hypotonia is more common than in adults. Patients of all ages may show abnormal flexor or extensor posturing (decerebrate or decorticate rigidity), associated with sustained upward deviation of the eyes, pouting, and grunting respiration. Hypoglycaemia must be excluded. Most children with cerebral malaria and about half the adult patients experience generalized convulsions. In children, seizures may be covert and difficult to detect. Twitching of the facial muscles or the corner of the mouth, deviation of gaze with nystagmus, irregularities of breathing, and stereotyped posturing of one arm may provide the only clue. Fewer than 5% of adult survivors have persisting neurological sequelae: these include cranial nerve lesions, extrapyramidal tremor, and transient paranoid psychosis. However, more than 10% of African children who survive an attack of cerebral malaria have sequelae such as hemiplegia, cortical blindness, epilepsy, ataxia, or cognitive and learning disabilities.

Other severe manifestations and complications

Anaemia (see above) is an inevitable consequence of all but the mildest infections. It is most common and severe in pregnant women, children, and in patients with high parasitaemia, schizontaemia, secondary bacterial infections, and renal failure. Spontaneous bleeding from the gums and gastrointestinal tract is seen in fewer than 5% of adult patients with severe malaria and is rare in children. Jaundice is common in adults but rare in children. Biochemical evidence of severe hepatic dysfunction is most unusual and hepatic failure suggests concomitant viral hepatitis or another diagnosis.

Hypoglycaemia is an important complication. Quinine or quinidine treatment can cause hypoglycaemia in pregnant women with severe or uncomplicated falciparum malaria and in any patient with severe disease. This develops a few hours to 6 days after starting this treatment, even after the parasitaemia has cleared. However, even in the absence of quinine or quinidine treatment, pregnant women and children with falciparum malaria and patients with hyperparasitaemia or complicating bacteraemia may become hypoglycaemic early in their illness. Clinical features of hypoglycaemia include anxiety, tachycardia, breathlessness, feeling cold, confusion, sweating, light-headedness, restlessness, fetal bradycardia or other signs of fetal distress, coma, convulsions, and extensor posturing. All may be misinterpreted as manifestations of malaria per se.

Hypotension and shock (‘algid malaria’) is a consequence of pulmonary oedema, metabolic acidosis, gastrointestinal haemorrhage, and complicating Gram-negative bacteraemias. Mild supine hypotension with postural drop in blood pressure is caused by vasodilatation and relative hypovolaemia. Cardiac arrhythmias are rare but may be precipitated by too rapid intravenous infusion or excessive doses of chloroquine, quinine, or quinidine. Patients with coronary insufficiency may develop angina during febrile crises of malaria.

Renal dysfunction, indicated by oliguria and increased blood urea and serum creatinine concentrations, occurs in about one-third of adults with severe malaria but is uncommon in children. Most patients respond to cautious rehydration, but 10% develop renal failure requiring dialysis.

In patients whose RBCs are congenitally deficient in G6PD or other enzymes, intravascular haemolysis and haemoglobinuria (Fig. may be precipitated by oxidant antimalarial drugs such as primaquine and tafenoquine, whether or not they have malaria. Classic blackwater fever is the association of haemoglobinuria from massive intravascular haemolysis, not explicable by G6PD deficiency, with severe manifestations of falciparum malaria, such as renal failure, hypotension, and coma, in a nonimmune patient.

Metabolic acidosis is seen in association with hyperparasitaemia, hypoglycaemia, and renal failure. It is usually lactic acidosis. In African children, respiratory distress with deep (Kussmaul) breathing, associated with severe anaemia and metabolic acidosis, is a syndrome that carries an even higher case fatality than cerebral malaria.

Pulmonary oedema is the terminal event in many adults dying of falciparum malaria. It may be precipitated by fluid overload late in the clinical course but pulmonary oedema can also develop in patients with severe disease in normal fluid balance, in which case jugular venous, central venous, or pulmonary artery wedge pressures are normal, as in ARDS. In pregnant women, pulmonary oedema may evolve suddenly after delivery. The earliest sign is an increase in respiratory rate. Without a chest radiograph, pulmonary oedema may be difficult to differentiate from aspiration pneumonia, a common complication in comatose patients, or metabolic acidosis.

Cerebellar dysfunction

A rare presentation of falciparum malaria is cerebellar ataxia with unimpaired consciousness. Similar signs may be seen in patients recovering from cerebral malaria. In Sri Lanka, a syndrome of delayed cerebellar ataxia has been described 3 to 4 weeks after an attack of fever attributed to falciparum malaria. Complete recovery is the rule.

Malarial psychosis

The term ‘malarial psychosis’ has been uncritically applied, often without proven aetiology. Acute psychiatric symptoms in patients with malaria may be attributable to their drug treatment, including antimalarial drugs such as chloroquine, mefloquine, and the obsolete mepacrine, and to exacerbation of pre-existing functional psychoses. However, in some patients, organic mental disturbances associated with malaria infection have been the presenting feature or, more often, have developed during convalescence after attacks of otherwise uncomplicated malaria or cerebral malaria. Depression, paranoia, delusions, and personality changes associated with malaria are classified as brief reactive psychoses. These symptoms rarely last for more than a few days.

Vivax, ovale, and malariae malarias

The prepatent and incubation periods are given in Table . Some strains of P. vivax, especially those from temperate regions (P. v. hibernans from Russia, P. v. multinucleatum from China) may have very long incubation periods (250–637 days). Only about one-third of imported cases of vivax malaria present within a month of returning from the malarious area; between 5 and 10% will present more than 1 year later.

The inappropriately termed ‘benign’ malarias cause paroxysmal, feverish symptoms even more hectic and distressing than those of falciparum malaria. Prodromal symptoms may be more severe with P. malariae infection. The characteristic tertian interval between fever spikes in P. vivax and P. ovale infections and the quartan pattern in P. malariae infections is established after several days of irregular fever if treatment is delayed. Vivax and ovale malarias have a persistent hepatic cycle, which may give rise to relapses every 2 or 3 months for 5 to 8 years in untreated cases. P. malariae does not relapse but a persisting, undetectable parasitaemia may cause recrudescences for more than 50 years.

Vivax malaria

People of West African origin are inherently resistant to P. vivax infection. Although symptoms may be severe and temporarily incapacitating, especially in nonimmunes, the acute mortality of vivax malaria is very low. During the 1967–9 Sri Lankan epidemic of predominantly vivax malaria, there were more than 500 000 cases with a case fatality of only 0.1%. Acutely, vivax malaria can cause anaemia, thrombocytopenia, and mild jaundice with tender hepatosplenomegaly. Rarely, the anaemia may be severe enough to be life threatening in debilitated patients and it may contribute to chronic malaise, wasting, malnutrition, and underperformance. Splenic rupture, which carries a mortality of 80%, may be more common with vivax than falciparum malaria. It results from acute, rapid enlargement of the spleen, with or without trauma. Chronically enlarged spleens are less vulnerable. Splenic rupture presents with abdominal pain and guarding, signs of haemorrhagic shock, fever, and a rapidly falling haematocrit. These features may be misattributed to malaria itself. In pregnancy, vivax malaria contributes to maternal anaemia and reduced birth weight. Cerebral vivax malaria has occasionally been reported especially with P. v. multinucleatum in China. Mixed P. falciparum infection or another encephalopathy must be adequately excluded in such cases. Acute noncardiogenic pulmonary oedema is an increasingly recognized complication of vivax malaria in nonimmune people. Clearly, the pathogenicity and clinical consequences of vivax malaria deserve re-evaluation.

Ovale and malariae malarias

Acute symptoms may be as severe as those of vivax infection, but anaemia is less severe and the risk of splenic rupture is lower although splenomegaly may be particularly gross in areas where P. malariae is prevalent. P. ovale causes negligible mortality, but P. malariae causes chronic morbidity and mortality from nephrotic syndrome and tropical splenomegaly syndrome. The same strictures apply to cerebral malariae malaria as to cerebral vivax malaria, especially as P. malariae coexists with P. falciparum throughout most of its range.

Monkey malarias

Knowlesi malaria

Recently, P. knowlesi infection in humans has been recognized as an important zoonosis in several South-East Asian countries. It is probably not new but has been overlooked. It is transmitted among long-tailed macaques Macaca fascicularis and related cercopithecine monkeys by jungle mosquitoes of the An. leucosphyrus group (notably An. latens) and causes fatal malaria in rhesus monkeys M. mulatta. P. knowlesi was first identified as an important cause of human malaria in Kapit Division, Sarawak (Malaysia), in 2000–2, where 120 (58%) of malaria cases were found to be infected. It had been confused microscopically with P. malariae because early trophozoites may appear as band forms, although these are not always seen. Using P. knowlesi-specific PCR primers, about 30% of human cases of malaria in Sarawak, together with cases in Sabah, Palawan (Philippines), Pahang (peninsular Malaysia), Thailand, Burma, and Singapore, have been identified. Increasing human encroachment into the jungle habitat in South-East Asia and possibly an adaptive switch in pathogenicity suggest that P. knowlesi infection may become more important. It is currently regarded as a zoonosis because human-to-human transmission has not yet been demonstrated.

Clinical features

There are daily spikes of fever (quotidian periodicity). Four fatal cases in Sarawak had fever and chills, abdominal pain and other gastrointestinal and pulmonary symptoms, jaundice, hypotension, acute renal failure, and hyperparasitaemia (764 720 parasites/μl in one case). In Malaysian Borneo, about 10% of patients develop potentially fatal complications and the case fatality is 1.8%.

Other monkey malarias

Human erythrocytes can be infected with at least five other species of simian plasmodia. There have been rare cases of natural infections or accidental laboratory infections by P. brasilianum, P. cynomolgi, P. inui, P. schwetzi, and P. simium. Severe feverish and systemic symptoms have been described, but no cerebral or other severe complications. No patient has died. Parasitaemia may remain undetectable for 2 to 6 days after the start of symptoms. Periodicity is tertian in P. simium and P. cynomolgi infections. Infectivity and virulence may be enhanced by repeated passage in humans.

Malaria in pregnancy and the puerperium

Malaria is an important cause of maternal anaemia and death, abortion, stillbirth, premature delivery, low birth weight, neonatal death, and congenital malaria in areas of unstable malaria transmission where women of reproductive age have little acquired immunity. In nonimmune women, hyperpyrexia, hypoglycaemia, anaemia, cerebral malaria, and pulmonary oedema are more common in pregnancy. During the great epidemic of falciparum malaria in Sri Lanka in 1934–5, case fatality among pregnant women was 13%, twice that in nonpregnant women. In Thailand, where malaria was at one time the leading cause of maternal mortality, cerebral malaria in late pregnancy had a case fatality of 50%. In some parts of Africa, one-quarter to one-half of all placentas are parasitized. The incidence is highest in primiparae. Changes in humoral and cell mediated immunity in pregnancy do not explain this vulnerability. It is clear that the placenta is a privileged site for parasite multiplication. RBCs infected with strains (genotypes) of P. falciparum expressing Var2CSA, a member of the PfEMP1 family, bind to chondroitin sulphate A, a receptor expressed on the surface of the syncytiotrophoblast. Other host receptors, such as hyaluronic acid and the neonatal Fc receptor, may also support placental binding. This may explain sequestration in the placenta. Placental dysfunction, fever, and hypoglycaemia contribute to fetal distress, which is common when malaria strikes in the last trimester of pregnancy. Painless uterine contractions are often detectable by monitoring. They may subside as the patient is cooled.

Special risks to the mother of malaria during pregnancy

Severe anaemia, exacerbated by malaria, is an important complication of pregnancy in many tropical countries that may persist into the puerperium and beyond. Especially in communities where chronic hookworm anaemia is prevalent, high-output anaemic cardiac failure may develop in late pregnancy.

Asymptomatic hypoglycaemia may occur in pregnant women with malaria, without provocation by cinchona alkaloids, and pregnant women with severe uncomplicated malaria are particularly vulnerable to quinine-induced hypoglycaemia. There is an increased risk of pulmonary oedema, precipitated by fluid overload and the sudden increase in peripheral resistance and autotransfusion of hyperparasitaemic blood from the placenta that occurs just after delivery.

Interaction between malaria and HIV in pregnancy

In HIV-infected pregnant women, the beneficial effects of parity on severity of malaria are attenuated and peripheral and placental parasitaemia and risk of having an episode of malaria and anaemia during pregnancy are increased. Malaria–HIV coinfection is associated with an increased risk of low birth weight, preterm birth, intrauterine growth retardation, and postnatal infant mortality. Malaria transiently increases peripheral blood and placental HIV viral load but whether this affects the risk of vertical transmission of HIV infection or accelerates HIV disease progression is unknown.


Whenever possible, pregnant women should avoid living in and, especially, sleeping in malarious areas. Otherwise, they should sleep under insecticide-treated bed nets, should be monitored for infection in an antenatal clinic and should receive intermittent preventive treatment with sulphadoxine–pyrimethamine or antimalarial prophylaxis extending into the early puerperium.

Congenital and neonatal malaria

Congenital or vertically transmitted malaria is diagnosed by detecting parasitaemia in the neonate within 7 days of birth, or later if there is no possibility of postpartum mosquito-borne infection. Save for a few discordant reports, most evidence from malarious parts of the world indicates that congenital malaria is rarely symptomatic, despite the high prevalence of placental infection. This confirms the adequacy of protection provided by IgG from the immune mother, which crosses the placenta, by active immunization from exposure to soluble malarial antigens in utero and by the high proportion of fetal haemoglobin in the neonate, which retards parasite development. Congenital malaria is, however, much more common in infants born to nonimmune mothers. Its incidence increases during malaria epidemics and it can cause stillbirth or perinatal death. All four species can produce congenital infection, but, because of its very long persistence, P. malariae causes a disproportionate number of cases in nonendemic countries. Fetal plasma quinine and chloroquine concentrations are about one-third of the simultaneous maternal levels. Thus, antimalarial concentrations adequate to cure the mother may be subtherapeutic in the fetus. Quinine and chloroquine are excreted in breast milk, but the suckling neonate would receive only a few milligrams per day. Maternal hypoglycaemia, a common complication of malaria, or its treatment with quinine may produce marked fetal bradycardia and other signs of fetal distress.

Differential diagnosis

Clinical features of congenital malaria are nonspecific: irritability, feeding problems, hepatosplenomegaly, anaemia, and jaundice. Unless parasites are found in a smear from a heel prick or cord blood, the patient may be misdiagnosed as having rhesus incompatibility or another congenital infection such as cytomegalovirus, herpes simplex, rubella, toxoplasmosis, or syphilis.

Transfusion malaria, ‘needlestick’, and nosocomial malaria

Malaria can be transmitted in blood from apparently healthy donors. Exceptionally, donors may remain infective for up to 5 years with P. falciparum and P. vivax, 7 years with P. ovale, and 46 years with P. malariae. Because the infecting parasites are erythrocytic forms (not sporozoites), no exoerythrocytic (hepatic) cycle will be established and so vivax and ovale malarias will not relapse. Theoretically, parasitaemia might be detectable immediately and, hence, the incubation period should be shorter than with mosquito-transmitted malaria. However, the incubation period tends to be longer because of the time needed to build up parasitaemias sufficient to cause symptoms. Mean incubation periods are 12 days (range 7–29 days) for P. falciparum, 12 days (range 8–30 days) for P. vivax, and 35 days (range 6–106 days) for P. malariae. Whole blood, packed cells (blood products), leucocyte or platelet concentrates, fresh plasma, marrow transplants, and haemodialysis have been responsible for transfusion malaria. As patients requiring transfusion are likely to be debilitated and may be immunosuppressed, and there may be a long delay before making the diagnosis because malaria is not suspected, unusually high parasitaemias may develop with P. falciparum and P. malariae, but with P. ovale and P. vivax infections, the parasitaemia is usually limited to 2% because only reticulocytes are invaded. Severe manifestations are common and mortality may be high, e.g. 8 out of 11 infections in a group of heroin addicts and even acute P. malariae infections may prove fatal.

Nosocomial malaria has resulted from contamination of saline used for flushing intravenous catheters, contrast medium, and intravenous drugs. Malaria has complicated parenteral drug abuse.


Outside the malaria endemic area, donors who have been in the tropics during the previous 5 years should be screened for malarial antibodies (indirect fluorescent antibody). In the endemic area, recipients of blood transfusions can be given antimalarial prophylaxis, or at least they should be watched carefully for evidence of infection. Addition of antimalarial drugs to stored blood is not justified.


Since malaria can present with a wide range of symptoms and signs, none of them diagnostic, it must be excluded by repeated thick and thin blood smears and rapid antigen detection in any patient with acute fever who has history of possible exposure. Until malaria is confirmed or an alternative diagnosis emerges, at least three smears should be taken over a period of 72 h. However, if the patient has symptoms compatible with severe malaria, a therapeutic trial of antimalarial chemotherapy must not be delayed. Antimalarial drugs may make microscopical diagnosis more difficult and so chemoprophylaxis should be stopped while the patient is under investigation for malaria. A history of travel to malarious areas during the previous year must be obtained. Malaria cannot be excluded just because the patient took prophylactic drugs, for none is completely protective. Short airport stopovers, even on the runway, or working in or living near an international airport may allow exposure to an imported infected mosquito. Small outbreaks of autochthonous malaria (transmission of malaria imported into areas from which malaria has been eliminated but where competent vector mosquitoes exists) have been reported in Europe, North America, and elsewhere. The possibility of transmission by blood transfusion, ‘needlestick’, or nosocomial infection should be kept in mind. Those who grew up in an endemic area will probably lose their immunity to disease after living for a few years in a temperate zone and they become newly vulnerable on return home to visit friends and relations, especially in rural areas. In malaria endemic regions, a large proportion of the immune population may have asymptomatic parasitaemia and it cannot be assumed that malaria is the cause of a patient’s symptoms even if parasitaemia is detected. In malarious countries, the diagnosis of malaria may be missed in the heat of an epidemic of some other infection such as meningitis, pneumonia, or cholera.

Differential diagnosis (Table 3)

Malaria must be included in the differential diagnosis of any acute febrile illness unless it can be excluded by: (1) impossibility of exposure, (2) repeated negative blood smears, and (3) a therapeutic trial of antimalarial chemotherapy. In Europe and North America, imported malaria has been misdiagnosed as influenza, viral hepatitis, viral haemorrhagic fever, epilepsy, viral encephalitis, or travellers’ diarrhoea, sometimes with fatal consequences. Cerebral malaria must be distinguished from other infective meningoencephalitides. Examination of the cerebrospinal fluid will identify most of these infective causes. Abdominal reflexes are absent in cerebral malaria but are brisk in patients with psychotic stupor and hysteria. Overdose of antimalarial drugs (chloroquine and quinine) has been confused with cerebral malaria. Intravenous drug abusers are at risk from both severe malaria and drug overdose. Alcoholism may be confused with cerebral malaria, whether the patient presents inebriated, with delirium tremens, or with Wernicke–Korsakoff syndrome.

Table 3 Differential diagnosis of malaria


Symptom Diagnosis
Acute fever Heat stroke, hyperpyrexia of other causes, other infections, other causes of fever
Fever and impaired consciousness (cerebral malaria) Viral, bacterial, fungal, protozoal (e.g. African trypanosomiasis, amoebic) or helminthic meningoencephalitis, cerebral abscess. Head injury, cerebrovascular accident, intoxications (e.g. insecticides), poisonings (e.g. antimalarial drugs), metabolic (diabetes, hypoglycaemia, uraemia, hepatic failure, hyponatraemia). Septicaemias, cerebral typhoid
Fever and convulsions Encephalitides, metabolic encephalopathies, hyperpyrexia, cerebrovascular accidents, epilepsy, drug and alcohol intoxications, poisoning, eclampsia, febrile convulsions, and Reye’s syndrome (children)
Fever and haemostatic disturbances Septicaemias (e.g. meningococcaemia), viral haemorrhagic fever, rickettsial infection, relapsing fevers, leptospirosis
Fever and jaundice Viral hepatitis, yellow fever, leptospirosis, relapsing fevers, septicaemias, haemolysis, biliary obstruction, hepatic necrosis (drugs, poisons)
Fever with gastrointestinal symptoms Travellers’ diarrhoea, dysentery, enteric fever, other bacterial infections, inflammatory bowel disease
Fever with haemoglobinuria (‘blackwater fever’) Drug-induced haemolysis (e.g. oxidant antimalarials in glucose 6-phosphate- dehydrogenase-deficient patient), favism, transfusion reaction, dark urine of other causes (e.g. myoglobinuria, urobilinogen, porphobilinogen)
Fever with acute renal failure Septicaemias, yellow fever, leptospirosis, drug intoxications, poisonings, prolonged hypotension
Fever with shock (‘algid malaria’) Septicaemic shock, haemorrhagic shock (e.g. massive gastrointestinal bleed, ruptured spleen), perforated bowel, dehydration, hypovolaemia, myocarditis

Suspicion of viral haemorrhagic fever may lead to patients with imported fevers being isolated in a high-containment unit where basic investigations, such as examination of a blood smear, may be delayed for fear of infection. Jaundice is a common feature of yellow fever but unusual in other viral haemorrhagic fevers.

Malaria in pregnancy may be confused with viral hepatitis, acute fatty liver with liver failure or eclampsia, and in the puerperium with puerperal sepsis or psychosis.

Laboratory diagnosis


Parasites may be found in blood smears taken by venepuncture, finger-pulp or earlobe stabs, and from the umbilical cord and impression smears of the placenta. In fatal cases, cerebral malaria can be confirmed rapidly as the cause of death by making a smear from cerebral grey matter obtained by needle necropsy through the superior orbital fissure, the foramen magnum, the ethmoid sinus via the nose, or a fontanelle in young children. Sometimes no parasites can be found in peripheral blood smears from patients with malaria, even in severe infections. This may be explained by partial antimalarial treatment or by sequestration of parasitized cells in deep vascular beds. In these cases, parasites or malarial pigment may be found in a bone marrow aspirate. Pigment may be seen in circulating neutrophils. A number of Romanowski stains, including Field’s, Giemsa, Wright’s, and Leishman’s, are suitable for malaria diagnosis. The rapid Field’s technique, which can yield a result in minutes, and Giemsa are recommended. Smears may be unsatisfactory for any one of a number of reasons: the slides are not clean; stains are unfiltered, old, or infected; the buffer pH is incorrect (it should be pH 7.0–7.4); drying is too slow, especially in a humid climate (producing heavily crenated erythrocytes); or the blood has been stored in anticoagulant causing lysis of parasitized erythrocytes. It is difficult to make a good smear if the patient is very anaemic. Common artefacts resembling malaria parasites are superimposed platelets, particles of stain and other debris, and pits in the slide. Other erythrocyte infections such as bartonellosis and babesiosis may be misdiagnosed as malaria. Parasites should be counted in relation to the total white cell count (on thick films when the parasitaemia is relatively low) or erythrocytes (on thin films). An experienced microscopist can detect as few as 5 parasites/µl (parasites in 0.0001% of circulation RBCs) in a thick film and 200/µl (0.004% parasitaemia) in a thin film.

Fluorescent microscopy

The quantitative buffy coat method involves spinning blood in special capillary tubes in which parasite DNA is stained with acridine orange and a small float presses the PRBCs against the wall of the tube where they can be viewed by ultraviolet microscopy. In expert hands, the sensitivity of this method can be as good as with conventional microscopy of thick blood films but species diagnosis is difficult and the method is much more expensive.

Rapid malarial antigen detection

Malaria dipstick antigen-capture assays employ monoclonal antibodies to detect P. falciparum histidine-rich protein 2 (PfHRP-2) or parasite-specific lactate dehydrogenase or aldolase from the glycolytic pathway found in all species. They are a convenient addition or alternative to microscopy as they are quick (taking about 20 min), sensitive (detecting >100 parasites/μl or 0.002% parasitaemia), and species specific. However, false positivity has been a problem and only the parasite-specific lactate dehydrogenase tests detect P. ovale and P. malariae. The NOW malaria test (Inverness Medical) is available in the United Kingdom. Currently, Paracheck Pf (Orchid Biomedical Systems, Goa, India) and SD Bioline malaria antigen test (Standard Diagnostics, South Korea) are not available in the United Kingdom but are recommended. ParaSight F (Becton-Dickinson), ICT Malaria Pf (ICT Diagnostics), and OptiMAL (Flow Laboratories) are also available.

Other methods

Enzyme- and radioimmunoassays, DNA probes (using chemiluminescence for detection), and PCR now approach the sensitivity of classical microscopy. They take much longer (up to 72 h), are much more expensive, and are unlikely to replace microscopy for routine diagnosis. However, some of these newer methods could be automated for screening blood donors or for use in epidemiological surveys. PCR can distinguish parasite strains.

Serological techniques

Malarial antibodies can be detected by immunofluorescence, enzyme immunoassay, or haemagglutination, for epidemiological surveys, for screening potential blood donors, and occasionally for providing evidence of recent infection in nonimmune individuals. These tests are not useful in making an acute diagnosis of malaria. In future, detection of protective antibodies will be important in assessing the response to malaria vaccines.

Other clinical laboratory investigations

Anaemia is usual, with evidence of haemolysis. Serum haptoglobins may be undetectable. The direct antiglobulin (Coombs’) test is usually negative. Neutrophil leucocytosis is common in severe infections whether or not there is a complicating bacterial infection, but the white cell count can also be normal or low. The presence of visible malarial pigment in more than 5% of circulating neutrophils is associated with a bad prognosis. Thrombocytopenia is common in patients with P. falciparum and P. vivax infections; it does not correlate with severity. Prothrombin and partial thromboplastin times are prolonged in up to one-fifth of patients with cerebral malaria. Concentrations of plasma fibrinogen and other clotting factors are normal or increased, and serum levels of fibrin(ogen) degradation products are normal in most cases. Fewer than 10% of patients with cerebral malaria have evidence of disseminated intravascular coagulation. However, antithrombin III concentrations are often moderately reduced and have prognostic significance. Total and direct (unconjugated) plasma bilirubin concentrations are usually increased, consistent with haemolysis, but in some patients with very high total bilirubin concentrations there is a predominance of conjugated bilirubin, indicating hepatocyte dysfunction. Some patients have cholestasis. Serum albumin concentrations are usually reduced, often grossly. Serum aminotransferases, 5′-nucleotidase, and, especially, lactic dehydrogenase are moderately elevated, but not nearly as much as in viral hepatitis. Hyponatraemia is the most common electrolyte disturbance. There may be mild hypocalcaemia (after correction for hypoalbuminaemia) and hypophosphataemia, especially after patients have been given blood or a glucose infusion. Biochemical evidence of generalized rhabdomyolysis (elevated serum creatine kinase concentration, myoglobinaemia, and myoglobinuria) is sometimes found. In about one-third of patients with cerebral malaria, the blood urea concentration is increased above 80 mg/dl (13 mmol/litre) and serum creatinine above 2 mg/dl (176 µmol/litre). Lactic acidosis is common in severely ill patients, especially those with hypoglycaemia and renal failure. It may be suspected if there is a wide ‘anion gap’, i.e. [Na+] – [Cl] + [HCO3] is greater than 12 meq/litre.

Blood glucose must be checked frequently, especially in children, pregnant women, and severely ill patients, even if the patient is not receiving quinine treatment and is fully conscious. A bedside dipstick method, with or without photometric quantification, is rapid and convenient. Microscopy and culture of cerebrospinal fluid is important in patients with cerebral malaria to exclude other treatable encephalopathies. In cerebral malaria the cerebrospinal fluid may contain up to 15 lymphocytes/µl and an increased protein concentration. Pleocytosis of up to 80 cells/µl, mainly leucocytes, may be found in patients who have had repeated generalized convulsions. The cerebrospinal fluid glucose level will be low in hypoglycaemic patients and this result may be the first hint of hypoglycaemia. In view of the finding of cerebral compression and high opening pressures in many African children with cerebral malaria, some paediatricians prefer to delay lumbar puncture, while covering the possibility of bacterial meningoencephalitis with empirical antimicrobial treatment. Blood cultures should be performed in patients with a high white cell count, shock, persistent fever, or an obvious focus of secondary bacterial infection. Gram-negative rod bacteria (Escherichia coli, Pseudomonas aeruginosa, etc.) have been cultured from the blood of adult patients with ‘algid’ malaria and in African children an association was found between malaria and nontyphoid salmonella septicaemia.

Urine should be examined by microscope and dipsticks. Common abnormalities are proteinuria, microscopic haematuria, haemoglobinuria, and RBC casts. The urine is literally black in patients with severe intravascular haemolysis. Urine specific gravity should be measured: the optical method is most convenient when urine output is small. Monitoring plasma concentrations of antimalarial drugs such as quinine is rarely possible but can be useful.


Antimalarial chemotherapy

Classes of drugs that have antimalarial activity are shown in Bullet list 3.

Stage specificity

Among blood schizonticides, artemisinin derivatives can prevent the development of rings or trophozoites, but quinine and mefloquine cannot stop development before the stage of mature trophozoites and pyrimethamine–sulphadoxine combinations do not prevent the development of schizonts.

Bullet list 3 Classes of antimalarial drugs

  • Arylaminoalcohols—quinoline methanols (quinine and quinidine extracted from the bark of the cinchona tree), mefloquine, and lumefantrine
  • 4-Aminoquinolines—chloroquine and amodiaquine
  • Bisquinolines—piperaquine
  • Folate-synthesis inhibitors—type 1 antifolate drugs that compete for dihydropteroate synthase (sulphones and sulphonamides); type 2 antifolate drugs that inhibit malarial dihydrofolate reductase (the biguanides proguanil and chlorproguanil and the diaminopyrimidine pyrimethamine)
  • 8-Aminoquinolines—primaquine and tafenoquine (Etaquine, WR-238605, or SB-252263)
  • Antibiotics—tetracycline, doxycycline, clindamycin, azithromycin, and fluoroquinolones
  • Peroxides (sesquiterpene lactones)—artemisinin (qinghaosu) derivatives and semisynthetic analogues (artemether, arteether, artesunate, and artelinic acid)
  • Naphthoquinones—atovaquone
Antimalarial drugs

The antimalarial properties of cinchona alkaloids were discovered in Peru around 1600 but their mode of action remains unknown. Quinine became the first-line treatment of severe falciparum malaria after the emergence of chloroquine resistance but is now being replaced by artemisinin derivatives. Intravenous injection of quinine is dangerous as high plasma concentrations may result during the distribution phase, causing fatal hypotension or arrhythmias. However, quinine can be given safely if it is diluted and infused intravenously over 2 to 4 h, or, if intravenous infusion is not possible and parenteral treatment is needed, it may be given by intramuscular injection divided between the anterolateral parts of both thighs. For intramuscular injection, the stock solution of quinine dihydrochloride (300 mg/ml) should be diluted to 60 mg/ml. It is well absorbed from this site. Historically, intramuscular quinine carried the risk of tetanus but it is safe provided that strict sterile precautions are observed. Because most deaths from severe falciparum malaria occur within the first 96 h of starting treatment, it is important to achieve parasiticidal plasma concentrations of quinine as quickly as possible. This can be accomplished safely with an initial loading dose of twice the maintenance dose (Table 4). The initial dose of quinine should not be reduced in patients who are severely ill with renal or hepatic impairment, but in these cases the maintenance dose should be reduced to between 3 and 5 mg/kg if parenteral treatment is required for longer than 48 h.

Table 4 Antimalarial chemotherapy in adults or children with uncomplicated malaria who can swallow tablets


Option and age of patient Chloroquine-resistant P. falciparum or where the origin of the infection is unknown Chloroquine-sensitive P. falciparum or P. vivax, P. ovale, P. malariae, or monkey malarias
1 Artemether with lumefantrine Chloroquinea
Adult 4 tablets (each containing 20 mg artemether and 120 mg lumefantrine) 600 mg base on the 1st and 2nd days; 300 mg on the 3rd day
Twice daily for 3 days  
Child <15 kg body weight, 1 tablet c.10 mg base/kg on the 1st and 2nd days; 5 mg base/kg on the 3rd day
15<25 kg, 2 tablets  
25<35 kg, 3 tablets  
All twice daily for 3 days  
  For radical cure of P. vivax/P. ovale (except pregnant and lactating women or G6PD-deficient patients), add:
2 Proguanil with atovaquone Primaquine
Adult 4 tablets (each containing 100 mg proguanil and 250 mg atovaquone) 15 mg base/day on days 4–17; or 45 mg/week for 8 weeksb
Once daily for 3 days  
Child 11–20 kg, one tablet 0.25 mg/kg per day on days 4–17; or 0.75 mg/kg per week for 8 weeksb
21–30 kg, 2 tablets  
31–40 kg, 3 tablets  
All once daily for 3 days  
3 Quinine  
Adult 600 mg salt, 3 times daily for 7 daysc  
Child Approx. 10 mg salt/kg, 3 times daily for 7 daysc  

a For chloroquine-resistant P. vivax, repeat the course.

b For Chesson-type strains (South-East Asia, western Pacific), use double the dose or double the duration up to a total dose of 6 mg base/kg in daily doses of 15–22.5 mg for adults.

c In areas where 7 days of quinine is not curative (e.g. Thailand), add tetracycline 250 mg four times each day or doxycycline 100 mg daily for 7 days, except for children under 8 years and pregnant women, or add clindamycin 5 mg/kg three times daily for 7 days.

Minimum inhibitory concentrations of quinine for P. falciparum in South-East Asia and other areas of the tropics have increased. Longer courses of quinine and combination with pyrimethamine–sulphonamide combinations, tetracycline, or clindamycin have been required for cure. Quinine need not be withheld or stopped in patients who are pregnant. In therapeutic doses, it does not stimulate uterine contraction or cause fetal distress. Hypoglycaemia is the most important complication of quinine treatment. Plasma quinine concentrations above 5 mg/litre cause ‘cinchonism’: transient high-tone deafness, giddiness, tinnitus, nausea, vomiting, tremors, blurred vision, and malaise. Rarely, quinine may give rise to haemolysis, thrombocytopenia, disseminated intravascular coagulation, hypersensitivity reactions, vasculitis, and granulomatous hepatitis. Self-poisoning with quinine causes blindness, deafness, and central nervous depression. These features are rarely seen in patients being treated for malaria, even though their plasma quinine concentrations may exceed 20 mg/litre. This may be explained by the increased binding of quinine to α1-acid glycoprotein (orosomucoid) and to other acute-phase reactive serum proteins associated with acute infection.

Quinidine, the dextrorotatory stereoisomer of quinine, is more effective against resistant strains of P. falciparum but is more cardiotoxic than quinine. Because it was widely stocked for treating cardiac arrhythmias, quinidine gluconate injection was once more generally available than parenteral quinine. The Centers for Disease Control in the United States of America formerly supplied it for the parenteral treatment of malaria, but it has now been replaced by artesunate.

Mefloquine is a synthetic drug, effective against some P. falciparum strains resistant to chloroquine, pyrimethamine–sulphonamide combinations, and quinine. It cannot be given parenterally, but is well absorbed when given by mouth, reaching peak plasma concentrations in 6 to 24 h, with an elimination half-time of 14 to 28 days. The drug can be given as a single dose but, to reduce the risk of vomiting and other gastrointestinal side effects, the dose is best divided into two halves given 6 to 8 h apart. Gastrointestinal symptoms occur in 10 to 15% of patients but are usually mild. Less frequent side effects include nightmares and sleeping disturbances, dizziness, ataxia, sinus bradycardia, sinus arrhythmia, postural hypotension, and an ‘acute brain syndrome’ consisting of fatigue, asthenia, seizures, and psychosis. These unpleasant symptoms, whose incidence has probably been exaggerated, have made the drug unpopular. Those taking β-blockers or with a past history of epilepsy or psychiatric disease should avoid the drug. Unfortunately, in vitro resistance to mefloquine and treatment failures have now been reported in South-East Asia, Africa, and South America. One large observational study in Thailand suggested an increased risk of stillbirth associated with mefloquine but this was not found in Malawi. Mefloquine treatment should be avoided in pregnant women, especially during the first trimester, and pregnancy should be avoided within 3 months of stopping mefloquine.

Lumefantrine (benflumetol) is an arylaminoalcohol that, despite its structural similarity to halofantrine (now withdrawn), is not cardiotoxic. It is combined with artemether as a co-artemether (see below).


Chloroquine, a synthetic antimalarial, is concentrated in the parasite’s lysosomes where haemoglobin is digested, and may act by inhibiting the haempolymerase that converts toxic haemin into insoluble haemozoin (malarial pigment). Alternatively, the drug may interfere with parasite feeding by disrupting its food vacuole. From the original foci in Thailand and Colombia, chloroquine-resistant P. falciparum has spread to most parts of the tropics. The observation that chloroquine resistance could be reversed in vitro by high concentrations of calcium channel blockers, which in other situations could reverse the multidrug resistance (mdr) phenotype acquired by tumour cells, focused attention on a malarial homologue of the human MDR gene. Genes involved in the development of resistance include P. falciparum chloroquine resistance transporter on chromosome 7 and loci on chromosome 5, which harbours the MDR gene homologue PfMDR1, and chromosome 11. Despite the widespread resistance of P. falciparum to this drug and the recent emergence of chloroquine-resistant P. vivax in New Guinea and adjacent areas of Indonesia, chloroquine remains the most widely used antimalarial drug worldwide. It is the treatment of choice for P. vivax, P. ovale, P. malariae, and P. knowlesi infections and for uncomplicated falciparum malaria acquired in the few areas where the parasite remains sensitive to this drug (Central America north-west of the Panama Canal, Hispaniola (Haiti and the Dominican Republic, and parts of the Middle East). Elsewhere, the emergence of chloroquine resistance has had a devastating effect on malarial morbidity and mortality. In Senegal, mortality from malaria in children under 5 years old increased up to 11-fold between 1984 and 1995. Absorption of chloroquine after intramuscular or subcutaneous injection is very rapid. Unless small doses are given frequently, this can produce dangerously high plasma concentrations, probably accounting for the deaths of some children soon after they had received intramuscular injections of chloroquine. Therapeutic blood concentrations persist for 6 to 10 days after a single dose. Plasma concentrations above about 250 ng/ml cause dizziness, headache, diplopia, disturbed visual accommodation, dysphagia, nausea, and malaise. Chloroquine, even in small doses, may cause pruritus in dark-skinned races. It may exacerbate epilepsy and photosensitive psoriasis. Cumulative, irreversible retinal toxicity from chloroquine has been reported after lifetime prophylactic doses of 50 to 100 g base (i.e. after 3–6 years of taking 300 mg of base per week), although this is most unusual. Chloroquine overdose is described in Chapter 9.1. Chloroquine is safe during pregnancy and lactation.

Amodiaquine, although structurally similar to chloroquine, retains activity against chloroquine-resistant strains of P. falciparum in some geographical areas. Unlike chloroquine, it is metabolized to a toxic quinoneimine capable of causing toxic hepatitis and potentially lethal agranulocytosis (which occurred in up to 1 in 2000 people taking amodiaquine prophylactically). Amodiaquine is still used, but, because of its risks and the limited therapeutic advantage over chloroquine, its use for prophylaxis and repeated treatment is now discouraged by WHO.


Piperaquine was used extensively as prophylaxis and treatment in China and Indochina from the 1960s until emergence of piperaquine-resistant strains of P. falciparum during the 1980s. More recently, it has been combined with artemisinins, as an artemisinin-based combination therapy (ACT) as: dihydroartemisinin (DHA), trimethoprim, piperaquine phosphate, and primaquine phosphate (China-Vietnam (CV), CV4 and CV8); DHA, trimethoprim, and piperaquine phosphate (Artecom), and DHA and piperaquine phosphate only (Artekin, Duo-Cotecxin), which have proved effective and safe in mainland South-East Asia. Piperaquine is highly lipid soluble with a large volume of distribution at steady state/bioavailability, long elimination half-life, and a clearancethat is markedly higher in children than in adults. Its tolerability, efficacy, pharmacokinetic profile, and low cost make it suitable as a constituent of ACT.

Folate-synthesis inhibitors

Pyrimethamine–sulphonamide combinations

The mode of action of the antifolate drugs is well understood. Pyrimethamine 75 mg and sulphadoxine 1500 mg (Fansidar), once valuable in the treatment of chloroquine-resistant falciparum infections worldwide, is no longer marketed in the United Kingdom, but it and other pyrimethamine combinations such as with dapsone (Maloprim) and with sulphalene (Metakelfin) are still used elsewhere. Unfortunately, resistance to these synergistic combinations has spread to most malarious continents, resulting from mutations at residues 108, 51, 59, 16, and 164 of the parasite’s dihydrofolate reductase gene. Pyrimethamine is a folate inhibitor that may cause folic acid deficiency in pregnant women and others unless folinic acid supplements are given. The sulphonamide components of these combinations may cause systemic vasculitis (Stevens–Johnson syndrome), or toxic epidermal necrolysis. Fansidar caused fatal reactions in 1 in 18 000 to 26 000 prophylactic courses. Aplastic anaemia and agranulocytosis can also occur. Both pyrimethamine and sulphonamide cross the placenta and are excreted in milk. In the fetus and neonate, sulphonamides can displace bilirubin from plasma protein-binding sites, thus causing kernicterus. For these reasons, pyrimethamine–sulphonamide combinations are not recommended for treatment during pregnancy or lactation unless no alternative drug is available, and should never be used for prophylaxis.


This combination was developed as an alternative to pyrimethamine–sulphonamide combinations to replace chloroquine for the treatment of uncomplicated falciparum malaria in Africa. It proved more effective than pyrimethamine–sulphonamide combinations in treating parasites with DHFR mutations at bases 108, 51, and 59, but should probably be further combined with an artemisinin to extend its useful therapeutic life.


Primaquine is the only readily available drug effective against hepatic hypnozoites of P. vivax and P. ovale. It is essential for the radical cure of these infections. Primaquine is gametocytocidal for all species of malaria. Mass treatment of patients with P. falciparum infection could eliminate the sexual cycle in mosquitoes by sterilizing gametocytes. Its elimination half-time is 7 h. Primaquine causes haemolysis in patients with congenital deficiencies of erythrocyte enzymes, notably G6PD, but severe intravascular haemolysis is unusual except in areas such as the Mediterranean (e.g. Sardinia) and Sri Lanka. Primaquine can cross the placenta and cause severe haemolysis in a G6PD-deficient fetus, most commonly a boy, and is also excreted in breast milk. It should therefore not be used during pregnancy or lactation in areas where G6PD deficiency is prevalent. Like sulphonamides and sulphones (i.e. dapsone), primaquine can produce severe haemolysis and methaemoglobinaemia in patients with congenital deficiency of NADH methaemoglobin reductase. Those affected quickly develop dusky cyanosis, noticed first in the nail beds. In patients with G6PD deficiency, weekly dosage with 45 mg of primaquine is better tolerated than the usual daily dose of 15 mg. In the Solomon Islands, Indonesia, Thailand, and Papua New Guinea, a total dose of 6.0 mg/kg (twice the usual dose) or even more may be needed to eliminate the primaquine-resistant Chesson-type strain of P. vivax. This is usually given as 15 mg base/day for 28 days.

Tafenoquine is a newer 8-aminoquinoline; it has a longer half-life (2 weeks) than primaquine and is over 10 times more active as a hypnozoiticide. It is also a potent schizonticide.

Peroxides (sesquiterpene lactones)


Artemisinin or qinghaosu (pronounced ‘ching-how-soo’) from the Chinese medicinal herb Artemisia annua (sweet wormwood), family Compositae, has been used to treat fevers in China for more than 1000 years. It is a sesquiterpene lactone, with an endoperoxide (trioxane) active group that was isolated in China in 1971–2. Iron within the parasite probably catalyses the cleavage of the endoperoxide bridge leading to the generation of free radicals, which then form covalent bonds with parasite proteins (alkylation). Artemisinins destroy the blood stages of P. falciparum from trophozoite to schizont, including those of multiresistant strains. They clear parasitaemia more rapidly than other antimalarial drug. Resistance is emerging in SE Asia.


Dihydroartemisinin (DHA) is the active metabolite, which has a short half-life. In severe falciparum malaria, intravenous artesunate is the treatment of choice but this drug can also be given by intramuscular injection. Multicentre trials enrolling 1461 patients in South-East Asia demonstrated a case fatality of 15% in patients treated with intravenous artemether, compared to 22% in those treated with intravenous quinine, a reduction in mortality of 34.7% in the artesunate-treated group.


Systematic review of 11 randomized controlled trials comparing intramuscular artemether with parenteral quinine showed lower mortality with artemether, but this was not significant in an analysis of adequately blinded trials. Within these, in an individual patient data analysis of 1919 adults and children, the odds ratio for deaths in artemether recipients was 0.8. In the prospectively defined subgroup analysis of adults with multisystem failure, there was a significant difference in mortality in favour of artemether but intramuscular artemether is erratically absorbed in patients with severe malaria especially those with shock. Artemotil (arteether) is similar to artemether but has been far less used.

Artemisinin suppositories

Rectal artesunate proved superior to intravenous/intramuscular quinine in reducing parasite densities 12 and 24 h after administration. Suppository formulations of artemisinin should prove particularly valuable in treating children at peripheral levels of the health service.

Artemether with lumefantrine combination (co-artemether) is effective for the oral treatment of multiresistant falciparum malaria.

The severe neurotoxicity reported in animals given large doses of artemisinins has not been detected in any of the tens of thousands of human patients treated with these compounds. Artemisinins have proved safe in the second and third trimesters of pregnancy but there are insufficient data to support their use in the first trimester.


Naphthoquinones, such as atovaquone, act on the electron-transport chain in malarial mitochondria through their structural similarity to coenzyme Q. Atovaquone is marketed in combination with proguanil for the treatment and prevention of multiresistant P. falciparum. It inhibits the parasite’s mitochondrial respiration by binding to the cytochrome bc complex. The drug is poorly and variably absorbed, but bioavailability is greatly enhanced by a fatty meal. Its elimination half-life is between 50 and 70 h.


All antimalarial antibiotics inhibit ribosomal protein synthesis and probably act on the parasite’s mitochondria. Tetracycline, clindamycin, azithromycin, quinolones, and sulphonamides such as co-trimoxazole have some antimalarial activity. They kill parasites too slowly to be used alone but are useful in combination for the treatment of uncomplicated P. falciparum malaria.

Treatment of falciparum malaria

Despite discovery of the rapidly effective, easily used, and safe artemisinin derivatives, treatment of falciparum malaria remains challenging in many parts of the world. The use of antimalarial drugs is poorly controlled, there are supply problems resulting from expense, inadequate distribution and erratic and incomplete dosing. Fake antimalarial drugs are penetrating increasingly into the markets in Africa and Asia. A worrying recent development has been the documentation of reduced in vivo susceptibility to artemisinins with delayed parasite clearance in Pailin, western Cambodia, and possibly in adjacent countries.

Combination antimalarial treatment

The combination of two or more different classes of antimalarial drugs with unrelated mechanisms of action to delay emergence of resistance was proposed by Wallace Peters in the early 1970s but was not effectively implemented because of the difficulty of identifying drugs that were still active against multidrug resistant P. falciparum and whose elimination half-times were similar. To counter the threat of resistance of P. falciparum to monotherapies, and to improve treatment outcome, combinations of antimalarials are now recommended by WHO (2006) for the treatment of falciparum malaria.

Artemisinin-based combination therapy (ACT)

The rapid clearance of parasitaemia and resolution of symptoms by artemisinin derivatives provides strong theoretical support for their use in combination with drugs such as mefloquine, amodiaquine, or pyrimethamine–sulphonamide. A meta-analysis of 11 randomized controlled trials confirmed that, in patients with uncomplicated malaria, addition of 3 days of artesunate to these drugs significantly reduced treatment failure, recrudescence, and gametocyte carriage. This lead to WHO’s 2006 recommendation to replace monotherapy with ACT. This has proved effective, except in South-east Asia (Cambodia and possibly in Thailand and Burma), where resistance to artemisinins has recently been reported.

Treatment of uncomplicated P. falciparum malaria (Table 4)

For treating adults, infants, and children in malarious areas, WHO (2006) recommends ACTs even in Africa, where the deployment of artemisinins cannot yet be justified by published evidence. The drug combined with artesunate depends on the resistance of local strains of P. falciparum. In South-East Asia, lumefantrine or mefloquine might be added. In Africa, lumefantrine, amodiaquine, or sulphadoxine–pyrimethamine might be added. For presumed nonimmune travellers returning to nonendemic areas, WHO recommends artemether–lumefantrine, atovaquone–proguanil, or quinine + doxycycline or clindamycin. Doxycycline should not be given to pregnant women or children under 8 years old.

Patients with uncomplicated malaria can usually be given antimalarial drugs by mouth. However, feverish patients may vomit the tablets. The risk of vomiting can be reduced if the patient lies down quietly for a while after taking an antipyretic such as paracetamol. Otherwise, the initial dose of antimalarial drug may have to be given by injection or suppository.

Treatment of severe falciparum malaria (Table 5)

Urgent parenteral chemotherapy, initiated with a loading dose, is the priority as there is a highly significant relationship between delay in chemotherapy and mortality. Severely ill or deteriorating patients who have been exposed to malaria should be given a therapeutic trial even if the initial smears are negative. Dosage should be calculated according to the patient’s body weight and drugs should be administered parenterally both to patients with severe falciparum malaria and to those who are vomiting and unable to retain swallowed tablets. The treatment of choice is artesunate given by intravenous bolus injection. It can also be given by intramuscular injection. Artemether by intramuscular injection or quinine by intermittent or continuous intravenous infusion or intramuscular injection is less effective. If patients with severe malaria cannot swallow and retain tablets and antimalarial treatment by intramuscular/intravenous injection/infusion is likely to be delayed for several hours, insertion of a single artesunate suppository substantially reduces the risk of death or permanent disability.

Therapeutic response and vital signs must be monitored clinically (temperature, pulse, blood pressure) and by examination of blood films. Patients should be switched to oral treatment as soon as they are able to swallow and retain tablets. They must be watched carefully for signs of drug toxicity. In the case of quinine, the most common adverse effect is hypoglycaemia. Therefore, the blood sugar should be checked frequently.

General management

Patients with severe malaria should be transferred to the highest level of care available, preferably a high dependency area or intensive care unit. They must be nursed in bed because of their postural hypotension and because of the risk of splenic rupture were they to fall. Body temperatures above 38.5°C are associated with febrile convulsions, especially in children, and between 39.5 and 42°C with coma and permanent neurological sequelae. In pregnant women, hyperpyrexia contributes to fetal distress. Therefore, temperature should be controlled by fanning, tepid sponging, a cooling blanket, or antipyretic drugs, such as paracetamol (15 mg/kg in tablets by mouth, or powder washed down a nasogastric tube, or as suppositories) and ibuprofen (tablets or parenteral).

Cerebral malaria

Convulsions, vomiting, and aspiration pneumonia are common, so patients should be nursed in the lateral position with a rigid oral airway or endotracheal tube in place. They should be turned at least once every 2 h to avoid bedsores. Vital signs, Glasgow Coma Score, and convulsions should be recorded. Convulsions can be controlled with diazepam given by slow intravenous injection (adults 10 mg, children 0.15 mg/kg) or intrarectally (0.5–1.0 mg/kg), or with midazolam given initially by intravenous injection of small doses every 2 min and then, when the seizure is controlled, by continuous intravenous infusion diluted in 5% dextrose or normal saline (dosage is adjusted for age and response, in the range of 30 to 300 μg/kg per h). Prophylactic use of phenobarbital was associated with increased case fatality in a placebo-controlled study in African children and is not recommended. Stomach contents should be aspirated through a nasogastric tube to reduce the risk of aspiration pneumonia. Elective endotracheal intubation is indicated if coma deepens and the airway is jeopardized. Deepening coma with signs of cerebral herniation is an indication for CT or MRI, or a trial of treatment to lower intracranial pressure, such as an intravenous infusion of mannitol (1.0–1.5 g/kg of a 10–20% solution over 30 min) or mechanical hyperventilation to reduce the arterial P CO 2 to below 4.0 kPa (30 mmHg).

Table 5 Antimalarial chemotherapy in adults or children with severe malaria who cannot swallow tablets


Chloroquine-resistant P. falciparum or the origin of the infection is unknown Chloroquine-sensitive P. falciparum or P. vivax, P. ovale, P. malariae, or monkey malarias
1. Artesunate a 1. Chloroquine b
2.4 mg/kg (loading dose) IV on the first day, followed by 1.2 mg/kg daily for a minimum of 3 days until the patient can take oral therapy or another effective antimalarial 25 mg base/kg diluted in isotonic fluid by continuous IV infusion over 30 h (or 5 mg base/kg over 6 h every 6 h)
2. Artemether 2. Quinine (see left-hand column below)
3.2 mg/kg (loading dose) IM on the first day, followed by 1.6 mg/kg daily for a minimum of 3 days until the patient can take oral treatment or another effective antimalarial. In children, the use of a 1 ml tuberculin syringe is advisable since the injection volumes will be small  
3. Quinine  
Adults: 20 mg salt/kg (loading dose)c diluted in 10 ml/kg isotonic fluid by IV infusion over 4 h, followed 8 h after the start of the loading dose with 10 mg salt/kg over 4 h, every 8 h until patients can  
Children: 20 mg salt/kg (loading dose)c diluted in 10 ml/kg isotonic fluid by IV infusion over 2 h, followed 12 h after the start of the loading dose with 10 mg salt/kg over 2 h, every 12 h until patients can swallowd  
The 7-day course should be completed with quinine tablets, approximately 10 mg salt/kg (maximum 600 mg) every 8–12 he  
4. Quinine (in intensive care unit)  
7 mg salt/kg (loading dose)c IV by infusion pump over 30 min, followed immediately with 10 mg salt/kg (maintenance dose) diluted in 10 ml/kg isotonic fluid by IV infusion over 4 h, repeated every 8 h until patient can swallow, etc.d,e  
5. Quinidine (in intensive care unit)  
15 mg base/kg (loading dose)c IV by infusion over 4 h, followed 8 h after the start of the loading dose with 7.5 mg base/kg over 4 h every 8 h, until the patient can swallow,d then quinine tablets to complete 7 days of treatmente  
If it is not possible to give drugs by intravenous infusion  
1 Artesunate 1. Chloroquine b
Same dosage as for IV above given IM Total dose 25 mg base/kg given as either:
IM or SC 2.5 mg base/kg, every 4 h
OR IM or SC 3.5 mg base/kg, every 6 h
2 Artemether  
As above  
3. Quinine 2. Quinine
20 mg salt/kg diluted to 60–100 mg/ml (loading dose)c IM into anterolateral thigh (half given into each leg), followed by 10 mg salt/kg, every 8 h until patient can swallow etc.d,e IM (see above left-hand column)
If it is not possible to give drugs by injection (IM/IV) or infusion  
1. Suppositories 1. Chloroquine
Artemisinin f  
40 mg/kg loading dose as suppositories intrarectally, followed by 20 mg/kg at 4, 24, 48, and 72 h followed by an oral antimalarial drugg 10 mg base/kg of body weight as tablets/syrup by mouth or nasogastric tube, then refer the patient to a higher level of healthcare for parenteral treatment
Continue 5 mg base/kg 5, 24, and 48 h laterg
Artesunate f  
One 200 mg suppository intrarectally at 0, 4, 8, 12, 24, 36, 48, and 60 h followed by an oral antimalarial drugh  
  • 2. Tablets of artemisinin (artesunate, artemether, artemether with lumefantrine), quinine, mefloquine, or other appropriate antimalarials g
  • Given by mouth or crushed and given via naso-gastric tube
2. Suppositories of artemisinin or artesunate, oral quinine, mefloquine, or sulphadoxine/pyrimethamine (see left-hand column) g

IM, intramuscular; IV, intravenous; SC, subcutaneous.

a Artesunic acid 60 mg is dissolved in 0.6 ml of 5% sodium bicarbonate diluted to 3–5 ml with 5% (w/v) dextrose and given immediately by intravenous (‘push’) bolus injection.

b Parenteral chloroquine should be used with great caution in young children.

c Loading dose must not be used if the patient has received quinine, quinidine, or halofantrine within preceding 24 h.

d In patients requiring more than 48 h of parenteral therapy, reduce the dose to 5.7 mg salt/kg every 8 h or 3.75 mg quinidine base/kg every 8 h.

e In areas where 7 days of quinine is not curative (e.g. Thailand), add tetracycline 250 mg four times each day or doxycycline 100 mg daily for 7 days except for children under 8 years and pregnant women, or add clindamycin 5 mg/kg three times daily for 7 days.

f Artemisinin and artesunate suppositories are registered for use in a few countries. If suppository formulations are not available, tablets of artemisinins should be given orally if possible, or crushed and given by nasogastric tube.

g Transfer the patient to hospital as soon as possible after initiating chemotherapy.

h In Vietnam, 4 mg/kg of artesunate in suppository form (China) intrarectally as a loading dose, followed by 2 mg/kg at 4, 12, 48, and 72 h followed by an oral antimalarial drug, proved as effective as artemisinin suppositories.

A number of potentially harmful ancillary remedies of unproven value have been recommended for the treatment of cerebral malaria. Two double-blind trials of dexamethasone (2 mg/kg and 11 mg/kg intravenously over 48 h) in adults and children in Thailand and Indonesia showed no reduction in mortality but prolongation of coma and an increased incidence of infection and gastrointestinal bleeding. Low-molecular-weight dextrans, osmotic agents, heparin, adrenaline (epinephrine), ciclosporin A, prostacyclin, pentoxifylline, malarial hyperimmune globulin, anti-TNFα monoclonal antibodies, desferrioxamine, dichloroacetate, and N-acetyl cysteine have proved ineffective in the treatment of cerebral and other forms of severe malaria. Some of these interventions were associated with serious side effects. There is some evidence that levamisole inhibits sequestration of PRBCs in patients with falciparum malaria and clinical trials are planned.


Indications for transfusion include a low (<20% or rapidly falling) haematocrit, severe bleeding, or predicted blood loss (e.g. imminent parturition or surgery), hyperparasitaemia, and failure to respond to conservative treatment with oxygen and plasma expanders. When the screening of transfused blood is inadequate and infections such as HIV, HTLV-1, and hepatitis viruses are prevalent in the community, the criteria for blood transfusion must be even more stringent. Exchange transfusion is a safe way of correcting the anaemia without precipitating pulmonary oedema in those who are fluid overloaded or chronically and severely anaemic. The volume of transfused blood must be recorded on the fluid balance chart. Transfusion must be cautious, with frequent observations of the jugular or central venous pressure and auscultation for pulmonary crepitations. Survival of compatible donor RBCs is greatly reduced during the acute and convalescent phases of falciparum malaria.

Disturbances of fluid and electrolyte balance

Fluid and electrolyte requirements must be assessed individually in patients with malaria. Circulatory overload with intravenous fluids or blood transfusion may precipitate fatal pulmonary oedema, but untreated hypovolaemia may lead to fatal shock, lactic acidosis, and renal failure. Hypovolaemia may result from salt and water depletion through fever, diarrhoea, vomiting, insensible losses, and poor intake. The state of hydration is assessed clinically from the skin turgor, peripheral circulation, postural change in blood pressure, peripheral venous filling, and jugular or central venous pressure. The history of recent urine output and measurement of urine volume and specific gravity may be useful. In tropical climates, adult patients with severe falciparum malaria may require 1 to 3 litres of intravenous fluid during the first 24 h of hospital admission. Fluid replacement should be controlled by observations of jugular, central venous, or pulmonary artery wedge pressures. Hyponatraemia (plasma sodium concentration 120–130 mmol/litre) usually requires no treatment, but these patients should be cautiously rehydrated with isotonic saline if they are clinically dehydrated, have low central venous pressures, a high urinary specific gravity, and a low urine sodium concentration (<25 mmol/litre).

Renal failure

Patients with falling urine output and elevated blood urea nitrogen and serum creatinine concentrations can be treated conservatively at first, but established acute renal failure must be treated with haemofiltration or dialysis. Hypovolaemia is corrected by the cautious infusion of isotonic saline until the central venous pressure is in the range +5 to +15 cmH2O. If urine output remains low after rehydration, increasing doses of slowly infused intravenous furosemide up to a total dose of 1 g and finally an intravenous infusion of dopamine (2.5–5 µg/kg per min) can be tried. If these measures fail to achieve a sustained increase in urine output, strict fluid balance should be enforced with particular emphasis on fluid restriction. Indications for haemoperfusion/dialysis include a rapid increase in serum creatinine level, hyperkalaemia, fluid overload, metabolic acidosis, and clinical manifestations of uraemia (diarrhoea and vomiting, encephalopathy, gastrointestinal bleeding, and pericarditis). Haemofiltration is the most effective technique in malaria but haemodialysis or peritoneal dialysis is also effective. The initial doses of antimalarial drug should not be reduced in patients with renal failure but, after 48 h of parenteral treatment, the maintenance dose should be reduced by one-third or one-half.

Metabolic acidosis

Lactic acidosis is an important life-threatening complication, especially in anaemic children. It should be treated by improving perfusion and oxygenation by blood transfusion and correcting hypovolaemia, clearing the airway, increasing the inspired oxygen concentration, and by treating septicaemia, a frequently associated complication.

Pulmonary oedema

This must be prevented by propping the patient up at an angle of 45° and controlling fluid intake so that the jugular or central venous pressure is kept below +5 cmH2O. Those who develop pulmonary oedema should be propped upright and given oxygen to breathe. In a well-equipped intensive care unit, the judicious use of vasodilator drugs can be controlled by monitoring haemodynamic variables, fluid overload can be corrected by haemoperfusion, and oxygenation can be improved by mechanical ventilation with positive end-expiratory pressure.

Hypotension and ‘shock’ (‘algid malaria’)

This should be treated as for bacteraemic shock. The circulatory problems should be corrected with blood transfusion (e.g. in anaemic children with respiratory distress and acidosis), plasma expanders, dopamine, and broad-spectrum antimicrobial treatment (such as gentamicin with ceftazidime or cefuroxime plus metronidazole) should be started immediately, bearing in mind that likely routes of infection include the urinary tract, lungs, and the gut. Other causes of shock in patients with malaria include dehydration, blood loss (i.e. following splenic rupture), and pulmonary oedema.


This may be asymptomatic, especially in pregnancy, and its clinical manifestations may be confused with those of malaria. Blood sugar must be checked every few hours, especially in patients being treated with cinchona alkaloids. Hypoglycaemia may arise despite continuous intravenous infusions of 5 or even 10% dextrose. A therapeutic trial of dextrose (1 ml/kg by intravenous bolus injection) should be given if hypoglycaemia is proved or suspected. This should be followed by a continuous infusion of 10% dextrose. Glucose may be given by nasogastric tube to unconscious patients or by peritoneal dialysis in those undergoing this treatment for renal failure. Among agents that block insulin release, diazoxide was ineffective, but octreotide, a synthetic somatostatin analogue, proved effective in some severe cases of quinine-induced hypoglycaemia.

Hyperparasitaemia and exchange blood transfusion

In nonimmune patients, case fatality exceeds 50% with parasitaemias above 500 000/µl. In Western countries, exchange blood transfusion, haemopheresis (and even plasmapheresis) have been used in presumed nonimmune patients with ‘hyperparasitaemia’ variously defined as parasitaemias exceeding 5 to 10%. Apart from reducing parasitaemia, these procedures might remove harmful metabolites, toxins, cytokines, and other mediators and restore normal RBC mass, platelets, clotting factors, and albumin. Potential dangers are electrolyte disturbances (e.g. hypocalcaemia), cardiovascular complications including ARDS, and infection from the blood or through infection of intravascular lines. Among more than 100 patients reported, some improved clinically soon after the procedure and most survived. However, there was reporting bias and a meta-analysis discovered no advantage. The efficacy of exchange transfusion is never likely to be put to the test of a randomized controlled trial. Artemisinins clear parasitaemia so rapidly that additional reduction of parasite load by exchange transfusion may not be important.

Splenic rupture

Acute abdominal pain and tenderness with left shoulder-tip pain and haemodynamic deterioration in patients with vivax and falciparum malaria suggests splenic rupture, especially if there is a history of recent abdominal trauma. Free blood in the peritoneal cavity and a torn splenic capsule can be detected by ultrasound or CT examination and confirmed by needle aspiration of the peritoneal cavity, laparoscopy, or laparotomy. Conservative management with blood transfusion and close observation in an intensive care unit is sometimes successful but access to surgical help is essential in case of sudden deterioration.

Disseminated intravascular coagulation

Patients with evidence of a coagulopathy should be given vitamin K (adult dose 10 mg by slow intravenous injection). Prothrombin complex concentrates, cryoprecipitates, platelet transfusions, and fresh-frozen plasma should be considered. Anticoagulants such as heparin and dalteparin are absolutely contraindicated.

Management of pregnant and lactating women with malaria


Unjustified fears of abortifacient and fetus-damaging effects of antimalarial drugs have led to the delay or even withdrawal of treatment, but experience since the 19th century has confirmed the safety of quinine in pregnancy. Chloroquine, proguanil, pyrimethamine, and sulphadoxine–pyrimethamine are also considered safe in the first trimester of pregnancy. Inadvertent exposure to other antimalarials in pregnancy is not an indication for termination of the pregnancy. Concerns about mefloquine in pregnancy have not been confirmed but doxycycline and other tetracyclines should be avoided during pregnancy and breastfeeding. Artemisinin derivatives have proved safe in the second and third trimesters of pregnancy, but there are insufficient safety data about their use in the first trimester.

For pregnant women with uncomplicated falciparum malaria, WHO (2006) recommends quinine ± clindamycin in the first trimester and ACTs in the second and third trimesters with artesunate + clindamycin or quinine + clindamycin as alternatives.

For severe falciparum malaria, WHO (2006) recommends artesunate and artemether above quinine in the second and third trimesters of pregnancy because they do not cause recurrent hypoglycaemia. However, blood glucose must be checked at least once a day in pregnant women with malaria, whether or not they are receiving quinine. In the first trimester, the risk of hypoglycaemia associated with quinine is lower, and uncertainties over the safety of the artemisinin derivatives are greater.

In lactating women, only tetracyclines, dapsone-containing antimalarials and possibly primaquine are contraindicated.

Ancillary treatments

Maternal fever should be reduced as soon as possible. Induction of labour, caesarean section, or accelerating the second stage of labour with forceps or vacuum extractor should be considered in patients with severe falciparum malaria. Fluid balance is particularly critical in these patients. If possible, the central venous pressure should be monitored. Exchange transfusion of 1000 to 1500 ml of blood in late pregnancy proved an effective way of managing severe anaemia with high-output cardiac failure in Nigeria. Circulating volume could be reduced and the risk of postpartum pulmonary oedema lessened by replacing exfused blood with a smaller volume of packed cells.

Prevention of malaria during pregnancy

As malaria during pregnancy can result in severe consequences for both the mother and child, therapeutic courses of antimalarials are effective as an intermittent preventive treatment and can be considered (see ‘Control and prevention’, below).

Treatment of P. vivax, P. ovale, P. malariae, and P. knowlesi malarias (Tables 4 and 5)

Chloroquine is the treatment of choice for vivax, ovale, malariae, knowlesi, and uncomplicated falciparum malarias in the few geographical areas where this drug can still achieve a satisfactory clinical response. Severe infections will require parenteral treatment (Table 5). Chloroquine-resistant P. vivax (New Guinea, Indonesia) is treated by increasing the dose of oral chloroquine. The usual 3-day course of chloroquine is well tolerated. In patients with vivax or ovale malarias, this should be followed by radical cure with primaquine to destroy hepatic hypnozoites, but caution is needed if there is a risk of a congenital enzyme deficiency (see ‘8-Aminoquinolines’, above), especially in pregnant or lactating women.


Case fatality of acute vivax, ovale, and malariae malarias is negligible except in the circumstances mentioned above. In knowlesi malaria, case fatality appears to be about 2%. Strictly defined cerebral malaria has a mortality of about 10 or 15% when medical facilities are good and may be less than 5% in Western intensive care units. Antecedent factors that predispose to severe falciparum malaria include the lack of acquired immunity or lapsed immunity, splenectomy, pregnancy, and immunosuppression (e.g. HIV infection). There is a strong correlation between the density of parasitaemia and disease severity. Severe falciparum malaria is defined by clinical criteria such as impaired consciousness, renal failure, hypoglycaemia, haemoglobinuria, metabolic acidosis, and pulmonary oedema. The case fatality of pregnant women with cerebral malaria, especially primiparae in the third trimester, is several times greater than in nonpregnant patients. The following laboratory findings carry a poor prognosis: peripheral schizontaemia, malarial pigment in more than 5% of circulating neutrophils, high cerebrospinal fluid lactate or low glucose, low plasma antithrombin III, serum creatinine exceeding 265 µmol/litre or a blood urea nitrogen of more than 21.4 mmol/litre, haematocrit less than 20%, blood glucose less than 2.2 mmol/litre, and elevated serum enzyme concentrations (e.g. aspartate and alanine aminotransferases, lactate dehydrogenase).

Chronic immunological complications of malaria

Quartan malarial nephrosis

In parts of East and West Africa, South America, India, South-East Asia, and Papua New Guinea, epidemiological evidence links P. malariae infection to immune-complex glomerulonephritis that leads to nephrotic syndrome. Few of those exposed to repeatedP. malariae infections develop nephrosis, suggesting that additional factors are involved. Histological changes, which are not entirely specific, are progressive focal and segmental glomerulosclerosis with fibrillary splitting or flaking of the capillary basement membrane, producing characteristic lacunae. Electron microscopy reveals electron-dense deposits beneath the endothelium. Immunofluorescence confirms glomerular deposits of immunoglobulins, C3, and P. malariae antigen in about 25% of cases. More than half the patients present by the age of 15 years with typical features of nephrotic syndrome. P. malariae is frequently found in blood smears and P. malariae antigen in renal biopsies in children but not in adults. The renal lesions may be perpetuated by autoimmune mechanisms. The pattern of immunofluorescent staining has some prognostic significance. Few patients respond to corticosteroids, but some are helped by azathioprine and cyclophosphamide, especially those whose renal biopsies show the coarse or mixed patterns of immunofluorescence. Antimalarial treatment is not effective. This condition could be prevented by antimalarial prophylaxis and has disappeared in countries such as Guyana following malaria eradication.

Tropical splenomegaly syndrome (hyperreactive malarial splenomegaly)

Transient splenomegaly is a feature of acute attacks of malaria in nonimmune or partially immune patients, while progressive splenomegaly is seen in children resident in malarious areas while they acquire immunity. However, a separate entity has been described in Africa (especially Nigeria, Uganda, and Zambia), the Indian subcontinent (Bengal, Sri Lanka), South-East Asia (Vietnam, Thailand, and Indonesia), South America (Amazon region), Papua New Guinea, and the Middle East (Aden). Defining features are: (1) residence in a malarious area, (2) chronic splenomegaly, (3) persistently elevated serum IgM and malarial antibody levels, (4) hepatic sinusoidal lymphocytosis, and (5) a clinical and immunological response to antimalarial prophylaxis. This condition is thought to result from an aberrant immunological response to repeated infection by any of the species of malaria parasite. Though requiring exposure to malaria and responding to antimalarial therapy, there is no association with the actual level of malarial endemicity. However, major differences in incidence in different ethnic groups suggest genetic predisposition.


The essential feature is the dysregulation of IgM production leading to the formation of macromolecular aggregates of IgM (cryoglobulins), the clearance of which leads to progressive splenomegaly and hepatomegaly. This may stem in part from the production of lymphocytotoxic antibodies specific for suppressor T lymphocytes, which normally control B cell production of IgM. In African patients, there is often an increase in circulating B lymphocytes. Distinction from chronic lymphatic leukaemia may be difficult. In Ghana, clonal rearrangements of the JH region of the immunoglobulin gene were found in patients with tropical splenomegaly who failed to respond to proguanil chemoprophylaxis, suggesting that the syndrome may evolve into a malignant lymphoproliferative disorder. Some of these patients had features of splenic lymphoma with villous (hairy) lymphocytes.

Clinical features

In malaria endemic areas, patients with tropical splenomegaly syndrome are distinguishable by their progressive splenic enlargement persisting beyond childhood. The spleen may be enormous, filling the left iliac fossa, extending across the midline and anteriorly, and producing a visible mass with an obvious notch. The liver is usually enlarged, especially the left lobe. The massive splenomegaly causes a vague dragging sensation and occasional episodes of severe pain with peritonism, suggesting perisplenitis or infarction. Anaemia may become severe enough to cause the features of high-output cardiac failure. Acute haemolytic episodes are described. These patients are vulnerable to infections, especially of the skin and respiratory system, and most deaths are attributable to overwhelming infection.

Laboratory findings

Severe chronic anaemia is the result of destruction and pooling in the spleen and dilution in an increased plasma volume. Thrombocytopenia may also be caused by splenic sequestration; it rarely causes bleeding. There is neutropenia and, in African patients, peripheral lymphocytosis and lymphocytic infiltration of the bone marrow. Serum IgM is greatly elevated (>2 standard deviations above the population mean, and often very much higher).

The essential histopathological feature is lymphocytosis of the hepatic sinusoids with Kupffer-cell hyperplasia. In some cases, round-cell infiltration of the portal tracts is associated with fibrosis, leading to portal hypertension. In the spleen there is dilatation of the sinusoids, hyperplasia of the phagocytic cells with erythrophagocytosis, and infiltration with lymphocytes and plasma cells. In patients with splenic lymphoma and villous lymphocytes, more than 30% of circulating lymphocytes are villous. These cells can be distinguished from hairy-cell leukaemia by their lack of CD25, CD11c, and tartrate-resistant acid phosphatase markers.

Differential diagnosis

Tropical splenomegaly syndrome must be distinguished from other causes of chronic, painless, massive splenomegaly, including leukaemias, lymphomas, myelofibrosis, thalassaemias, haemoglobinopathies, visceral leishmaniasis (by examination of bone marrow or splenic aspirates), and schistosomiasis (by liver biopsy, rectal snip, and stool examination). Lymphomas (especially chronic lymphatic leukaemia and follicular lymphoma) and even leukaemias may develop in patients with tropical splenomegaly syndrome. Nontropical idiopathic splenomegaly (normal serum IgM) and Felty’s syndrome produce a similar histological picture in the liver. Many cases of splenomegaly in the tropics remain undiagnosed.


Prolonged antimalarial chemoprophylaxis is the most important element of treatment. The majority of patients improve within 12 months of chemotherapy. The choice of drug will depend on the local sensitivity of whichever species or group of species of malaria parasite are thought to be responsible for this syndrome. The short- and long-term dangers of splenectomy rule out this procedure in the rural tropics. Folic acid may be needed. Diagnosis of patients with splenic lymphoma with villous lymphocytes (Ghana) is important as, in this condition, the risks of splenectomy are outweighed by the benefits.

Endemic Burkitt’s lymphoma 

Endemic Burkitt’s lymphoma, a tumour of the jaw, abdomen, and other areas that spreads to the bone marrow or meninges, is the most common type of childhood malignant disease in many parts of East and West Africa and Papua New Guinea. It has also been reported from Brazil, Malaysia, and the Middle East. Burkitt noticed that its distribution (by altitude, temperature, and rainfall) and even its seasonal incidence followed that of falciparum malaria. Outside malaria endemic areas, Burkitt’s lymphoma occurs sporadically. There is a suggestion that the B-cell line in cases in whites comes from lymphoid tissue, whereas in cases in Africans it comes from the bone marrow. Epstein–Barr virus (EBV) produces a lifelong infection of B lymphocytes. Normally this is controlled by specific, HLA-restricted, cytotoxic T cells, which recognize a virus-induced, lymphocyte-detected membrane antigen on B cells. Immunosuppression, as in recipients of renal allografts, allows uncontrolled proliferation of the EBV-infected B-cell line, which may give rise to one of the three chromosomal translocations [t(8;14), t(2;8), t(8;22)] that activate the c-myc oncogene on chromosome 8 responsible for malignant transformation. AcuteP. falciparum infection leads to a reduction in the numbers of suppressor T (CD8) lymphocytes allowing proliferation and increased immunoglobulin secretion by EBV-infected B cells. In highly malaria endemic areas of Kenya, well children aged between 5 and 9 years (the age of maximum incidence of the tumour) have reduced EBV-specific interferon-γ responses. These tumours may grow so rapidly that massive local tissue destruction results in urate nephropathy and acute renal failure. Cyclophosphamide, vincristine, methotrexate, and prednisolone are used in chemotherapy, producing remissions in 80 to 90% of patients and a long-term survival of 20 to 70%. Breakdown of large tumours during the first week of chemotherapy may be so dramatic that the acute tumour lysis syndrome may be precipitated. This consists of metabolic acidosis, hyperuricaemia, hyperphosphaturia, hyperphosphataemia, hyperproteinaemia, and hyperkalaemia, which may result in fatal cardiac arrhythmia and acute uric-acid nephropathy with renal failure.

Control and prevention

General principles of control

The intensity of malaria parasite transmission is spatially heterogeneous. This has important implications for overall risks of disease and the age patterns of disease, disability, and death. Endemicity is a measure of the intensity of malaria transmission in a human population and determines the average age of first exposure, the rate of development of immunity, and, thus, the expected clinical spectrum of disease. It follows that interventions should be tailored to these basic epidemiological foundations, e.g. intermittent preventive treatment in infants (IPTi) is likely to have little impact on the incidence of clinical malaria and anaemia in areas of exceptionally low transmission. Optimizing the introduction of diagnostics to rationalize the use of new, expensive therapies will require better tools to target where this is cost-efficient and where presumptive treatment remains appropriate. Deciding the strategy and optimal mixture of interventions depends on an understanding of the epidemiological patterns in a given area: one size will not fit all.

Across the central belt of sub-Saharan Africa, interventions that minimize loss of life must be directed to young children and their mothers. In addition, careful thought must be given to measures that have a profound impact on the burden of malaria, particularly in the few areas where people might receive one new infection every night and immunity is acquired very early in life. Perhaps reducing human–vector contact might compromise the natural immunity prevalent in the community. In communities infrequently exposed to malaria, a focus on case management alone is a dangerous strategy where the lack of immunity means that each infectious bite carries a far greater risk of severe disease and death compared to many areas of Africa. Preventing the infectious bite carries little risk of compromised immunity and delivers benefit to all in the community.

The cornerstones of contemporary malaria control, case management, indoor residual house spraying (IRS), and insecticide (pyrethroid) treated nets (ITNs) will remain effective only as long as the drugs and pesticides remain effective. Case management is undermined by the evolution of parasite resistance to antimalarial drugs. IRS and ITNs lose their effectiveness as the vectors evolve behavioural or physiological resistance. How we use new chemical agents to minimize resistance is key but there is always a risk that, as resistance develops (or donors lose interest), malaria control will fail and populations that have lost their functional immunity will be more vulnerable to malaria.

Contemporary malaria control has a fundamentally different mission from the Global Malaria Eradication Campaign (1955–69) that was coordinated by WHO. That campaign did not eradicate malaria everywhere, as planned, but did reduce malaria morbidityand mortality and the global extent of the disease. The focus of eradication was on the use of indoor residual spraying with DDT (dichlorodiphenyltrichloroethane) accompanied by effective case detection and management with effective drugs. In some places, malaria remains absent today. In other places, elimination was only a remote possibility because the starting basic reproductive number of infection was so high. After 1974, when resistance began to emerge to widely used insecticides and drugs, the international political and financial commitment to global control waned. In areas where malaria persisted, populations had reduced functional immunity as malaria transmission increased, so malaria morbidity and mortality rebounded. Resistance of P. falciparum to most antimalarial drugs had reached epidemic proportions in South-East Asia. In Africa, mortality from malaria in children doubled from the 1980s through to the mid-1990s coincidentally with the rapid expansion of P. falciparum resistant to chloroquine and sulphadoxine–pyrimethamine.

Against a rising malaria disease burden, new global programmes such as WHO’s Roll Back Malaria (RBM) initiative have emerged and aspire to reduce malaria morbidity and mortality by 75% by 2015. Because of malaria’s intrinsic links to development and poverty, malaria also forms part of the United Nations Millennium Development Goal that aims to halt and then reverse the rising incidence of malaria by 2015.

A number of strategies can be combined to reduce the burden of malaria effectively. Most are regarded as cost-effective solutions in resource-poor counties and affordable within the constraints of international financial support. The interventions can be grouped into those that limit human–vector contact (including indoor residual house spraying and insecticide-treated nets), those that aim to reduce vector abundance by targeting breeding sites or adult vector populations and those that target the parasite (including intermittent presumptive treatment and prompt case management).

Limiting human–vector contact

So far, two methods for large-scale operational vector control—indoor residual spraying and long-lasting insecticide-containing nets (LLINs)—have proved capable of reducing malaria transmission. Both are adulticide measures, targeted at reducing the number of adult infective female mosquitoes. Unfortunately, uptake and acceptance of both interventions is poor among local populations in malaria endemic regions, although the situation may be improving. Perversely, coils, aerosols, and insecticide-impregnated mats sold through the private sector have greater acceptance rates, but have little or no demonstrable affect on disease transmission.

Insecticide-treated nets (ITNs)

The use of bed nets impregnated with pyrethroids such as permethrin, cyhalothrin, or deltamethrin gives substantial protection against malaria in endemic areas. In some areas, insecticide-treated door and window curtains are used instead of bed nets. The current combined evidence indicates that ITNs can reduce all-cause childhood mortality by 17%, averting 5.5 deaths for every 1000 African children protected. Over 50% of clinical attacks can be prevented through the wide-scale use of ITNs and infection prevalence can be reduced by 13% in areas of stable endemic transmission. The effect is due to a combination of reduced access of mosquitoes to people because of the net, a repellent and lethal effect of the insecticide on the mosquitoes trying to bite, and, sometimes, an effect on local mosquito densities so that even those outside the nets may get some protection. Thus, health impacts are maximized when large sectors of a community are using ITNs, thereby providing a ‘public good’ by reducing local transmission. Nets are obviously most effective when mosquito biting is concentrated late at night and indoors.

ITNs appear to be one of the most promising means of control while the development of an operational vaccine is awaited. Although initial coverage in many malaria endemic countries was poor, it has increased in recent years through free distribution linked to mass vaccine campaigns or availability of heavily subsidized nets at clinics. This is crucial for reaching vulnerable and impoverished rural populations. Retreatment of nets has proved difficult to maintain in many parts of Africa. However, the recent launch of two registered brands of permanently treated nets aims to circumvent this problem. LLINs retain 50% of their original anopheline knock-down efficacy after 2 years and cost approximately US$4.80 each in 2005. Currently only one class of insecticides, pyrethroids, are recommended by the WHO Pesticide Evaluation Scheme (WHOPES) for use on nets because of their rapid action, low mammalian toxicity relative to their insect toxicity, and lack of odour. The efficacy of pyrethroids on LLINs is, however, threatened by the rapid increases in pyrethroid resistance among mosquito vectors in many parts of the world. In any insecticide-based vector control activity, insecticide resistance should be monitored at least annually as resistance is dynamic and can evolve rapidly.


Repellents are used for personal protection. Compounds such as diethyltoluamide (DEET) applied directly to the skin or to clothes can reduce the amount of mosquito–human biting contact for several hours after they are applied. However, their main use is against day-biting mosquitoes such as Aedes aegypti rather than malaria vectors. Some insecticides, such as DDT, have strong repellent properties.

Indoor residual house spraying (IRS)

During the global eradication era, IRS was instrumental in breaking malaria transmission in many parts of the world including Sri Lanka and South America. DDT at 2 g/m2 will remain toxic to endophilic anophelines for 6 months or more on nonabsorbent wall materials, with cyhalothrin or deltamethrin at a much lower dosage giving up to 4-month protection, while organophosphorus insecticides such as malathion, propoxur, and fenitrothion at the same dosage last about 3 months. This approach relies on killing the mosquito after it has fed and is thus a more community-focused intervention than ITN, requiring coverage of all houses and shelters. It requires a strong national organization to manage the routine spraying of houses and the compliance of householders to allow spray teams access and to remove their possessions from the house before spraying. There are few community-wide randomized controlled trials of IRS across Africa upon which to base an informed choice about the likely benefits of this approach in reducing morbidity and mortality. Most studies have been undertaken in areas of low transmission, with results similar to those with ITN. It is accepted, however, that interruption of transmission is harder to achieve under conditions of intense perennial transmission. In remote rural areas, the logistics of IRS are more difficult, but community cooperation has been achieved and mean parasitaemia was reduced by 80% in one remote area of Mozambique.

Reducing vector abundance

Mosquitoes are highly selective in their choice of larval habitat. The World Health Organization defines environmental management as the implementation of activities related to the modification or manipulation of environmental factors to minimize vector propagation in order to reduce human-vector interaction. Three accepted strategies are as follows:

  • Environmental modification—making sites unsuitable for vector breeding by draining, changing the rate of water flow, and adding or removing shade, cutting emergent vegetation, and altering the margins of bodies of water. Near the sea, salinity changes may be relevant. For small reservoirs and irrigation canals, cyclical changes in water level by means of a large siphon may control larvae by alternately stranding and flushing. Intermittent drying out of irrigation channels may also be of value
  • Environmental manipulation—filling holes, e.g. with polystyrene beads or soil, or using the Bti toxin derived from the bacteria Bacillus thuringiensis
  • Reducing human contact—using infective vectors by zooprophylaxis, modifying of human habitations, or purposely changing human behaviour

The basic epidemiology of local mosquito populations must be understood before control programmes are initiated. Mosquitoes cannot be eradicated because of the cost, ecological impact, or logistics and so the target threshold for the vector population is set at or below acceptable levels of potential disease transmission. Where habitats cannot be drained or rendered structurally unsuitable for mosquito breeding, chemical larvicides may be used. Historically, diesel oil, at 40 litre/ha of water surface with or without the addition of insecticides, prevented the larvae breathing when it was applied to the water surface with a spreading agent. Paris Green (1 kg/ha), temephos granules (2–20 kg/ha), or less than one-tenth of the amount of pyriproxyfen are effective and safer.

Intermittent preventive treatment

As an intermittent preventive treatment in pregnancy (IPTp), it was shown during the late 1990s that a therapeutic course of sulphadoxine–pyrimethamine given on two or three occasions during the second and third trimesters of pregnancy was effective in preventing infection of the placenta, reducing the incidence of anaemia in pregnant mothers, and increasing the birth weights of newborn children. This strategy continues to be implemented in many African countries but it is likely to become decreasingly effective because IPTp works less well in HIV-positive women and there is no proven safe alternative to sulphadoxine–pyrimethamine for IPTp in areas where resistance to this combination is rapidly expanding. DHA–mefloquine, DHA–piperaquine, and mefloquine–azithromycin combinations have been proposed as potential replacements subject to successful trials.

The concept of IPT has recently been extended to target infants (IPTi) by providing therapeutic courses of antimalarials at the same time as vaccination. Studies in Africa have shown that sulphadoxine–pyrimethamine or amodiaquine reduces the incidence of malaria and severe anaemia during the first year of life by 50 to 67%. The combined effects of IPTi and ITNs are currently being investigated.

Access to effective medicines

Even though preventive strategies may halve the incidence of clinical malaria, effective and prompt case management remains a fundamental adjunct to control, particularly among African children who experience at least 20 clinical attacks during their first 5 years of life. Cheap drugs such as chloroquine and sulphadoxine–pyrimethamine was the basis of malaria management in Africa and, during the early 1990s, resistance to both drugs increased rapidly and malaria mortality rose to levels similar to those witnessed in the colonial era despite a general decline in childhood deaths not attributable to malaria. Possible replacements for failing drugs are existing drugs combined with artemisinin derivatives. ACTs have the additional public health benefit of reducing transmission, like wide-scale use of ITNs. Introduction of combination mefloquine–artesunate therapy among refugees on the Thai–Burmese border was associated with an 18.5-fold reduction in gametocyte carriage rates, halving frequency of P. falciparum transmission in the area.

Control strategies based on case management depend on prompt treatment of appropriate patients with effective medicines. Since clinical criteria for diagnosing malaria have low specificity, parasitological diagnosis remains the gold standard. However, in many remote rural communities diagnostic facilities are inadequate and WHO’s Integrated Management of Childhood Illnesses (IMCI) recommends that all children living in malaria endemic areas (where >5% of fevers are due to malaria) should be given presumptive antimalarial treatment if they are febrile. The logic is that misdiagnosis, even in a small proportion of febrile children with malaria, can be serious because of the rapidity of progression to severe disease. Recommendations for older children and adults in malaria endemic areas are more ambiguous, posing a problem at a time when new, more expensive drugs are being deployed and malaria continues to be a diagnosis of convenience or a diagnosis by default, leading to massive overdiagnosis and overtreatment of malaria in these age groups.

Access to medicines remains poor in many malarious countries. In Kenya, less than 5% of children received an antimalarial within 24 h of the onset of symptoms. Increasing accessibility to new ACTs for the most vulnerable groups, largely African children, will require innovative methods of delivery. Operational approaches to improve access to effective medicines include training mothers or community-resource persons to administer medicines in the home, better training of shopkeepers to deliver advice when selling over-the-counter antimalarials, and improving community awareness of the need to get children to clinic early.

The aim of wide-scale use of ACT in Africa by 2015 is confronted by several problems. First, new ACT drugs cost 10 times as much as current failing drugs, putting them beyond the essential drugs budgets of most poor countries. International funding has been assembled through the Global Fund to Fight AIDS, Tuberculosis and Malaria but this will be effective only if there are guarantees of long-term sustainable financing. Secondly, the agricultural sector must produce sufficient Artemisia annua, the source of artemisinins. In the longer term, synthetic artemisinin compounds might eventually alleviate the dependence on natural products.

Malarial vaccines

Difficulties facing vaccine development

Vaccines offer one of the most effective public health tools for controlling infectious diseases. The obstacles to developing a malaria vaccine are formidable: the malaria parasite is complex and multistaged, with a large genome (25–30 Mb with 5000–6000 genes). Many of the potential immune targets are polymorphic and the parasite has a large capacity for evolving evasive strategies. Most effort has gone into developing subunit vaccines based usually on single antigens thought to be critical in the parasite’s biology. There are now a large number of potential vaccine candidates and one of the barriers to moving from concept to vaccine is the lack of appropriate animal models or in vitro correlates of immunity against which to select candidates for field trials. In the case of pre-erythrocytic vaccines, the development of centres for experimental immunization and challenge has offered a way of identifying effective candidates and it is likely that this approach will be extended to blood-stage vaccines.

Pre-erythrocytic vaccines

The aim of a pre-erythrocytic vaccine is either to block the establishment of an infection by preventing sporozoites from invading hepatocytes or to target the intrahepatic parasite to prevent progression to a blood-stage infection. Such infection-blocking immunity can be established in murine and human malaria by the repeated injection of irradiated sporozoites. Although establishing an important proof of principle, the apparent impracticality of using live attenuated sporozoites led to a focus on achieving the same effect using subunit vaccines. This led to the cloning in the 1980s of the first malaria gene, for the circumsporozoite protein, a major component of the coat of the sporozoite. This provided the basis for the development of a subunit vaccine, RTS,S, a fusion protein combining part of the circumsporozoite protein of P. falciparum with HBsAg with a complex adjuvant (AS02). RTS,S has consistently provided 30 to 40% protection against sporozoite challenges in nonimmune volunteers. Recent trials in infants and young children in Kenya and Tanzania showed a protective efficacy of 53% against malaria disease and large-scale phase III trials in infants across Africa are underway with results expected by 2011. The demonstration of efficacy (albeit not complete) of a subunit vaccine (RTS, S) based on one part of a single molecule from such a complex parasite has been important in giving confidence to the idea of developing antimalarial vaccines. Many investigators feel that given the complexity of the parasite and the high degree of antigenic polymorphism, the eventual ideal vaccine will involve combinations of antigens, probably from both pre-erythrocytic and erythrocytic stages. An extension of the same idea has recently seen a return of interest in the possibility of whole parasite vaccines. Remarkably, it has been shown that many of the apparent logistic objections to the production of attenuated sporozoite vaccines can in fact be overcome and plans are advanced for first experimental challenges of volunteers.

Blood-stage vaccines

The aims of blood-stage vaccines are to limit parasite replication and prevent clinical disease. A number of candidate vaccines based on key molecules on the merozoite surface or released from the merozoite at the time of RBC invasion (e.g. MSP1, MSP2, MSP3, AMA1) are at various stages of development and several candidate vaccines are in early field trials. Particular problems likely to be faced involve the high degree of antigenic polymorphism shown by most of the candidate vaccine molecules. This has also stood in the way of developing vaccines based on the other obvious target for protective immune responses, the parasite-derived antigens expressed on the infected RBC surface. However, in the particular case of pregnancy related malaria, the RBC expressed parasite molecules are of much more limited diversity and efforts are under way to develop a vaccine based on this subset of PfEMP1 antigens. These efforts have not yet reached the stage of human trials.

Transmission-blocking vaccines (TBVs)

TBVs aim to prevent the transmission of malaria by blocking the parasite’s development in the mosquito by inducing antibodies targeting either antigens present on the sexual stages of the parasites or mosquito antigens that are required for the successful development of the parasite in the midgut of its vector. Candidate vaccines have been shown to induce antibodies that completely block transmission of P. falciparum and P. vivax.

The malaria vaccine field is rapidly changing and updated information on the range of candidate vaccines can be obtained at and

Prevention of malaria in travellers

Advice to travellers

The prevention of malaria in travellers, particularly those usually resident in nonmalarious areas but visiting endemic regions, including those visiting their friends and relatives (‘VFR’s), has become more difficult because of resistance to antimalarial drugs. As a result, prevention can never be complete. Travellers must be advised to: (1) be aware of the risk; (2) reduce exposure to being bitten by anopheline mosquitoes; (3) take chemoprophylaxis where appropriate; and (4) seek immediate medical advice in the event of any fever or influenza-like illness developing while in the area, or within 3 months or more of leaving it, and to mention malaria as a possibility regardless of the precautions taken.

Preventive advice is subject to uncertainty because unequivocal data on efficacy are often unavailable, published studies are conflicting, the distribution of resistance to many prophylactics is changing and not well mapped, and experts disagree on the balance between the risk of malaria and the risk of side effects. Travellers may obtain conflicting opinions from different sources, jeopardizing their adherence to any one regimen. The WHO list of malarious areas, updated annually, and other publications are inevitably directed towards prophylaxis for areas of greatest transmission. Advice from someone who knows the country and the traveller’s itinerary is more specific and therefore more reliable.

No prophylactic regimen will give total protection, but many will reduce substantially the risk of malaria. Strict adherence, even to a suboptimal prophylactic regimen, is more important than vacillation in search of the ideal.

Diagnosis of imported malaria is a medical emergency as, exceptionally, falciparum malaria can be fatal within 24 h of the first symptom and the disease is often misdiagnosed (see Table 3). Several fatal cases are reported in the United Kingdom and the United States of America each year. Expert diagnosis and appropriate drugs may not be readily available. Useful guidelines are published by Centres for Disease Control in the United States of America and the Health Protection Agency and National Travel Health Network and Centre in the United Kingdom.

When falciparum malaria is diagnosed in a traveller, the rest of their tour group should be screened as a matter of urgency, as they can be presumed to have shared the same exposure risk.

Prevention of mosquito bites

Bed nets without tears or other holes through which mosquitoes might enter, impregnated with a pyrethroid insecticide such as permethrin, deltamethrin, or cyhalothrin, should be used and properly tucked in. These also afford protection against other arthropod vectors, ectoparasites and even night-biting kraits (snakes). A well-screened bedroom and other accommodation, combined with use of a knock-down insecticide after the doors have been closed before dusk, gives substantial protection. Clothes (long sleeves and trousers) that deter mosquito bites, repellent sprays and soaps (containing DEET or permethrin), and avoiding exposure to bites in the evenings will also help.

Choice and dosage of chemoprophylaxis

Detailed maps of the distribution of malaria in different countries and the recommended chemoprophylaxis for each area are listed in ‘Further reading’. Where there is a substantial risk of chloroquine-resistant falciparum malaria, atovaquone–proguanil, mefloquine, or doxycycline are appropriate (Bullet list 4). Of these, mefloquine and atovaquone–proguanil are licensed for children and doxycycline should not be given to children under 8 years old (British National Formulary: 12 years) (Table 6). Pregnant women are best advised to avoid malarious areas. Apart from proguanil–chloroquine, no drug has been proved safe for prophylaxis during pregnancy but, if exposure is unavoidable in a high-risk area, mefloquine is recommended.

Bullet list 4 Recommended malaria prophylaxis (adult dose) in addition to general measures specified in text

  • Where chloroquine-resistant P. falciparum is absent:
    • • Chloroquine 300 mg base weekly (best for short-term visitors)
    • • Proguanil 200 mg daily (best for long-term residents)
  • Where chloroquine-resistant P. falciparum is not widespread and is predominantly of low degree:
    • • Chloroquine 300 mg base weekly plus proguanil 200 mg daily
  • Where highly chloroquine-resistant P. falciparum occurs:a
    • (1) Atovaquone–proguanil 1 tablet daily
    • or
    • (2) Mefloquine 250 mg weekly
    • or
    • (3) Doxycycline 100 mg daily
    • or
    • (4) [Chloroquine 300 mg base weekly plus proguanil 200 mg daily]

a Regimens (1), (2), and (3) are more effective in some areas of South-East Asia, Africa, and South America, but there is a low but significant risk of severe side effects with (2) and (3). Regimen (4) will give only limited protection but is the least likely of the four regimens to cause toxic side effects and is preferred for pregnant women and, at reduced dosage, for young children (Table 7).


This combination has two great advantages: adverse effects are less frequent and less serious than for mefloquine and doxycycline; and it is a causal prophylactic, attacking pre-erythrocytic stages of malarial parasites. Consequently, it need be continued for only 7 days after leaving the malarious area, improving the chance of adherence. However, resistance is emerging and atovaquone–proguanil is expensive. The cost for short visits is similar to that of mefloquine or doxycycline but the differential cost rises greatly for longer visits.

Table 6 Doses of prophylactic antimalarial drugs for children


Age Weight (kg) Drug and tablet size
Chloroquine (150 mg) weekly with proguanil (100 mg) daily Mefloquine 250 mg Doxycycline 100 mg
Fraction of tablet
Term to 12 weeks <6.0 1/4 NR NR
3–11 months 6.0–9.9 1/2 1/4 NR
1–3 years 11 months 10.0–15.9 3/4 1/4 NR
4 –7 years 11 months 16–24.9 1 1/2 NR
8–12 years 25–44.9 1 1/2 3/4 NR
>12 years >44.9 2 1 1

NR, not recommended.

For children aged under 2 years in areas of chloroquine resistance, the appropriate medication is chloroquine plus proguanil. Chloroquine is available as a syrup but the proguanil has to be powdered on to jam or food. Measures against mosquito bites are specially important.


Mefloquine has a long half-life and on a weekly dosage schedule the blood level rises to a plateau from about 7 weeks. Most of its side effects, the main problem with its use, are associated with the initial three doses. The drug should therefore be started 2.5 weeks before departure to a malarious area, to allow a switch to an alternative if side effects prove troublesome. It can be used safely for at least 2 or 3 years. The most serious early side effects of mefloquine are neuropsychiatric: anxiety, depression, delusions, fits, and psychotic attacks. Their incidence is disputed. Airline passenger surveys have shown a frequency of 1:10 000, but experienced doctors in the United Kingdom assert a much higher frequency. It is not recommended for those in the first trimester of pregnancy or at risk of pregnancy during the 3 months after the end of chemoprophylaxis. In later pregnancy, the uncertain risk of stillbirth rate must be balanced against the considerable risks of malaria. Mefloquine is contraindicated in people with a history of epilepsy or psychiatric disease. Mefloquine resistance is reported from Africa, South-East Asia, and the Amazon region.


Doxycycline proved to give good protection against drug-resistant falciparum malaria in trials in Oceania and it is being increasingly used, especially for those who cannot or are unwilling to take mefloquine. It should not be used in children or pregnant women. The main side effects are photosensitization, which occurs in up to 3% of users, a tendency to precipitate vaginal thrush in women (preventable with a one-dose therapy for candidal infections), and the rare risk of Clostridium difficile diarrhoea. However, doxycycline may reduce the risk of travellers’ diarrhoea. ‘Heartburn’ and gastrointestinal discomfort from doxycycline itself is not uncommon. The drug is taken daily with food, taking care not to miss any days and avoiding lying down too soon after taking it to avert a real risk of acute pain from ulceration of the lower oesophagus. To get accustomed to taking daily medication, it should be taken a few days before departure.

Chloroquine and/or proguanil

In malarious areas where chloroquine-resistant P. falciparum is rare or absent, mainly in western Asia, North Africa, and Central America, chloroquine 300 mg (base), two tablets taken once a week, gives good protection. Since it suppresses only the blood forms, it will not prevent relapses of P. vivax or P. ovale. Continuous chloroquine prophylaxis is limited to 6 years because of a low risk of irreversible retinopathy. Beyond this, proguanil may be substituted. Proguanil 200 mg daily will act as a true causal prophylactic but is poorly protective against P. vivax. The extremely low incidence of adverse effects from proguanil makes it acceptable to long-term residents in endemic areas.

Where the prevalence and degree of chloroquine resistance is low, in parts of India and the rest of South Asia, the combination of chloroquine and proguanil (Table 7) remains effective and has the advantage of low toxicity and safety in pregnant women and in young children. However, it no longer provides adequate protection in sub-Saharan Africa, parts of India, South-East Asia, or the Amazon region.

Continuation after leaving the malarious area

All antimalarial agents except atovaquone–proguanil must be continued for 4 weeks after leaving the malarious area so that merozoites are killed when they emerge late from the liver into the blood stream.

Chemoprophylaxis in people with epilepsy

Proguanil, atovaquone–proguanil, and doxycycline do not increase the risk of fits in people with epilepsy.

Rejected chemoprophylactic drugs

The following drugs are unsuitable for chemoprophylaxis: amodiaquine because of the high risk of agranulocytosis; Fansidar (25 mg pyrimethamine and 100 mg sulphadoxine per tablet) because of the frequency of severe skin reactions; pyrimethamine on its own because it is ineffective in most malarial areas; and halofantrine because of its cardiotoxicity.

Geographical risk of malaria

The risk of malaria is much higher in sub-Saharan Africa than elsewhere. Prophylaxis must be taken except where the altitude is too great for transmission to occur or in the nonendemic southern parts of the continent. In Asia, the risk is usually much lower. Visitors to the air-conditioned hotels of the larger cities of South-East Asia do not need prophylaxis but elsewhere in Asia there may be urban malaria. Mefloquine does not protect adequately against malaria in South-East Asia; travellers to areas of higher transmission will need regimens (1) or (3) in Bullet list 4. Those living for long periods in such areas may prefer to adopt vigilance and the early treatment of fevers, but awareness of the risk is essential. Freedom from malaria in Asia by travellers does not mean that they will escape infection in Africa!

Travellers in remote areas away from prompt medical assistance should carry a therapeutic dose of atovaquone–proguanil, mefloquine, or lumefantrine-artemether in case they develop an acute fever.