Vivax malaria

Epidemiology

Using national statistics reported by regional offices of the World Health Organization, the 1999 World Health Report estimated that there are between 72 to 80 million cases of malaria due to P. vivax each year with the greatest burden observed in South and East Asia (52%), Eastern Mediterranean (15%) and South America (13%) ². Recently these statistics have been challenged by an analysis using a combination of geographic information systems, malaria epidemiology, historical maps and information on population densities, environment, and vector limits ³. Revised estimates of the global malaria burden were found to be up to 2.5 fold higher than that derived from national statistics ⁴. In preliminary work, Hay and colleagues also found that non-falciparum malaria, predominantly P. vivax, accounts for 25-40% of the global malaria burden with between 132 - 391 million cases per year; table 1. When the population density of areas endemic for P. vivax is taken into consideration, the number of people at risk of infection reaches 2.6 billion, slightly greater than that for P. falciparum ​³. Although the debate over methodologies continues ^{5, 6}, it is likely that the true burden of vivax malaria has been grossly underappreciated and is probably in the region of a quarter of a billion clinical cases a year.

In south and south east Asia, where the majority of vivax malaria occurs, P. vivax accounts for up to 50% of malaria cases with prevalence rates between 1-6% of the population ^{7, 8, 9}. The proportional burden of vivax is even greater in Central and South America, reaching 71-81% of all malaria cases ¹⁰. In eastern and southern Africa only 5% of malaria infections are attributable to P. vivax, although this still accounts for between 6-15 million cases per year ². The emergence of chloroquine-resistant P. falciparum has tended to reduce the proportion of malaria cases due to P. vivax ​¹¹; nevertheless the absolute numbers of P. vivax remain high. Conversely, where successful malaria control strategies have been employed the ratio of P. falciparum to P. vivax infections has fallen ¹².

In most areas the burden of disease is greatest in young children and infants with immunity usually developing by 10-15 years of age ¹³. In a longitudinal study from Thailand, incidence rates varied from over 800 per 100 person -years in children under 5 yrs to 200 per 100 person years in older adults. In these low transmission settings premunition and asymptomatic carriage occur ¹⁴, although overall 82% of patients with P. vivax parasitaemia were still found to be symptomatic ⁷. Extrapolating this figure to estimated total number P. vivax infections presented in table 1, suggests that there are between 106 and 313 million clinical cases of vivax malaria each year.

Pathobiology

Compared to P. falciparum, P. vivax has a slightly longer incubation period (12 days to several months), a similar erythrocytic cycle (42-48 hours) and produces fewer merozoites per schizont ¹⁵. It is generally believed that P. vivax merozoites require a single cell receptor, the Duffy antigen, to invade host erythrocytes. Humans lacking this antigen are not susceptible to infection, explaining why P. vivax is largely absent from West Africa, a highly malarious region where the Duffy negative blood group is ubiquitous ¹⁶. Recently this paradigm has been challenged by observations in east Africa, demonstrating P. vivax transmission in a population found to be Duffy antigen negative, although further studies are needed to elucidate the epidemiology ¹⁷. P. vivax is further limited by its apparent preference for invading young red cells, whereas P. falciparum invades a broader range of erythrocyte ages ¹⁸. This selective preference for red cell invasion is an important factor limiting the growth of P. vivax both in vitro and in vivo. The latter was noted over 100 years ago by Ross who commented upon the lower levels of parasitaemia observed in febrile patients with benign tertian malaria (P. vivax) compared to those with malignant tertian (P. falciparum) ​¹⁹. P. vivax is capable of inducing fever at levels of parasitemia lower than those causing fever in P. falciparum infection. In western Thailand, a region of low endemicity, the pyrogenic density was 180 parasites /μl, compared to 1000 parasites/μl observed in P. falciparum infection ⁷. These estimates are similar to those observed in volunteers and malariotherapy patients ²⁰. In keeping with this the host inflammatory response is activated to a greater extent during infections with P. vivax, with plasma levels of fever-inducing cytokines such as TNF-α being higher in vivax malaria compared to P. falciparum infections with similar parasitemia ^{21, 22}.

In addition to lower peripheral parasitemias, patients with P. vivax infections also tend to present with all parasite stages visible on the peripheral blood film. In contrast the late stages of P. falciparum are usually absent or present in low numbers, reflecting cytoadherence in the postcapillary venules. The latter is fundamental to the etiology of severe and placental malaria in P. falciparum infections ²³. In contrast P. vivax-infected red blood cells become increasingly more deformable as they mature ²⁴ and are usually considered not to cytoadhere or sequester in the microvasculature. These characteristics underlie the reason why severe pathology in vivax infections is much less common than with P. falciparum infection. Interestingly recent studies have challenged this paradigm, by suggesting that P. vivax-infected red cells may sequester in organs such as the lung ²⁵.

One of the most important differences between the plasmodial species infecting humans lies in the ability of P. vivax and P. ovale to relapse following the cure of the original infection. A proportion of sporozoites do not undergo immediate development in the invaded hepatocytes. Instead they remain dormant in the liver, as hypnozoites, for prolonged periods of time before developing and causing recurrent infection. P. vivax strains from different geographic areas display widely different relapse patterns, reflecting evolutionary adaptation to local environmental conditions that optimize transmission potential of the parasite. Strains from tropical areas are characterized by early primary infection followed by frequent (10-14) relapses 3-6 weeks apart ²⁶. Whereas in temperate areas primary infection tends to occur later with intervals up to one year and fewer and later relapses to overcome prolonged periods of the year when transmission to the mosquito vector is unfavorable ²⁷. The ability to relapse renders P. vivax resilient to eradication from both the human host and the environment.

Clinical manifestations

Clinical disease is associated with chills, vomiting, malaise, headache, fever and myalgia. These symptoms are non-specific and cannot be distinguished reliably from other febrile illnesses or other malarias ⁷. High fever and rigors are more common in vivax than falciparum malaria, reflecting synchronicity of schizont rupture. The classic paroxysms of fever lasting 4-8 hours and occurring with 48-56 hour periodicity, take several asexual cycles to develop in primary infections, although relapses often start synchronously with rigors ²⁸. Plasmodium vivax has a reputation as a benign infection, when compared to the severe manifestations frequently observed with untreated P. falciparum. However in the pre-antibiotic era vivax malaria resulted in a chronic debilitating febrile illness which could last for years and was associated with hypoproteinemia and edema and dramatic weight loss akin to kwashiorkor ²⁶. Even today with access to effective antimalarials severe and fatal infections can occur with P. vivax. It has long been recognized that splenic rupture is a life-threatening, though rare, complication ²⁹. More recently P. vivax has been shown to cause severe anemia, respiratory distress, malnutrition ³⁰ and possibly coma ^{31, 32, 33}.

In areas endemic for P. vivax, prevalence usually reaches a peak in young infants and children. In impoverished communities anemia from malnutrition, worm infestations and falciparum malaria can be near-universal in this age group. Recurrent hemolysis (of both parasitized and non-parasitized red cells) ³⁴ and dyserythropoesis from relapsing vivax malaria exacerbate this multifactorial anemia. Although most studies have not quantified the proportion attributable to vivax malaria, in some areas P.vivax infections are likely to be the most important determinant of anemia in infants. A case in point comes from southern Papua, Indonesia where almost 38% of patients hospitalized with malaria are anaemic (Hb<7g/dl). In this region 40% of severe anaemia (Hb<5g/dl) in infants can be attributed to infections with P. vivax, a rate almost three fold higher than that for P. falciparum ​⁸. Similar findings in Venezuela also suggest that anaemia may be more severe and frequent in P. vivax infections than in P. falciparum ​³⁵. When compounded by even mild respiratory dysfunction and malnutrition such degrees of anaemia are likely to be an important contributor to morbidity, hospital stay and even mortality ³⁰.

Pregnant women are particularly vulnerable to malaria. Plasmodium falciparum sequesters in the placenta resulting in pathological changes and compromises materno-fetal transport resulting in adverse outcomes in both mother and baby. The burden of P. vivax malaria in pregnancy is less well described. P. vivax does not appear to sequester in the placenta ³⁶ and is associated with less pronounced histological changes in the placenta, the main process being increased haemozoin deposition ³⁶. However P.vivax may still exert an adverse effect on the fetus through maternal anemia and the induction of a strong local inflammatory response, both of which interfere with utero-placental hemodynamics ³⁷. Clinical findings from Thailand and India have shown that vivax malaria during pregnancy causes maternal anemia and a significant reduction in mean birth weight (circa 110g) which is about 60% of that observed with falciparum malaria ^{38, 39}; this almost certainly contributes to infant mortality. Interestingly, in contrast with falciparum malaria, the birthweight reduction is greater in multigravidae than primigravidae.

There have now been at least 18 case reports of non-cardiogenic pulmonary edema/acute lung injury in vivax malaria, almost all of which have excluded concurrent infection with P. falciparum ^{25, 40, 41}. All of the cases developed after commencement of antimalarial chemotherapy, raising the possibility of a pulmonary inflammatory response to parasite killing ⁴². The possible sequestration of P. vivax in the pulmonary microvasculature ²⁵ may target this inflammatory response to the lung. Although further studies are needed, these findings and the observation that over 50% of patients with vivax malaria present with cough ^{25, 42} suggest that, as with P. falciparum ​⁴³, there may be an important overlap between the clinical features of acute respiratory infections and acute vivax malaria. There have also been several reports of coma associated with vivax malaria ^{31, 32, 33} though most have not excluded concurrent infection with P. falciparum or other pathogens with central nervous system manifestations.

These reports of severe manifestations following vivax malaria challenge the usual perception of P. vivax as a benign infection. The spectrum of disease caused by P. vivax in Papua New Guinea and Papua, Indonesia, appear particularly bad; indeed in southern Papua 16% of all cases of severe malaria admitted to hospital were attributable to P. vivax ​⁸. Reports from India and Sri Lanka also highlight the broad range of pathology associated with P. vivax ^{2, 33}. However these rates may not necessarily be extrapolated to other endemic areas. Differences in parasite virulence, host susceptibility, age of exposure as well as parasite factors such as antimalarial drug resistance may play a crucial role in modulating the clinical presentation. Alternatively the spectrum of disease in other areas may simply be under reported, particularly if severe pathology is mistakenly attributed to co-infection or misdiagnosis in speciation.

Mixed infections

In regions where P. falciparum and P. vivax coexist mixed-species infection are common and yet rarely reported. Surveys usually report rates less than 2% and yet careful clinical studies record rates up to 30% and this figure is even higher when PCR detection methods are used ⁴⁴. Concurrent infections with different plasmodium species may have important implications on the host response and development of cross-species immunity ⁴⁵. Indeed in Thailand where this has been particularly studied, mixed infections with P. vivax appear to attenuate the severity of disease with P. falciparum reducing the risk of severe malaria ⁴⁶, decreasing the risk of treatment failure ⁴⁷, lowering gametocyte carriage ⁴⁸ and reducing the prevalence of anaemia⁴⁹. However in higher transmission areas and areas of emerging drug resistance, the additive burden of severe malarial anemia and maternal malaria has not been addressed in detail. The potential for P. vivax to attenuate falciparum malaria requires further characterization, and has significant implications for vivax-only vaccination strategies, and the deployment of drugs such as chloroquine which have lost efficacy against P .falciparum but still retain it against P. vivax.

Socioeconomic Burden

The appreciation of the socio-economic burden of malaria has risen over the last decade and been a major factor in raising its global profile and the funds dedicated to combating the disease. Falciparum malaria alone is estimated to have reduced the Gross Domestic Product of 31 African countries by almost 10% ⁵⁰. Although neglected, a number of observations suggest that P. vivax is also likely to exert a considerable economic toll. The incidence of P. vivax varies between regions, and age groups, however since P. vivax relapses in 20-80% of patients ²⁷. This ensures that even in low transmission areas, up to 20% of the population can have a symptomatic infection each year with a cumulative experience of 10-30 episodes of malaria during a lifetime ².

The socio-economic burden of P.vivax infection is not known. Based on the estimated number of clinical episodes per year (106 - 313 million), assuming full treatment coverage with drug costs of $0.1 per course of chloroquine and non-drug treatment costs of $8.3 ⁵¹, the direct cost of treatment is between US$0.9 - 2.7 billion per year. Since each symptomatic infection is associated with at least 3 days absence from school or work (unpublished data), the overall global cost of P. vivax infection to the individual in terms of lost productivity (assuming a daily wage of US$1.5 per day), healthcare costs and transport to clinics can be conservatively estimated as being between US$ 1.4 - 4.0 billion per year.

These immediate costs are likely to be augmented by recurrent P. vivax infections in children. Longitudinal studies from Sri Lanka suggest that repeated attacks of malaria have an adverse impact on the school performance of children ^{52, 53}. Intuitively, recurrent vivax malaria will also result in anemia, malnutrition, growth retardation, stunting of development, and impaired economic productivity later in life, though detailed studies are lacking.

The burden of vivax malaria is felt not only at the individual level but also by health care providers. In many parts of Asia and South America, where P. vivax accounts for more than half of the cases of malaria, institutional healthcare costs are appreciably higher due to high rates of clinic attendance and hospitalization. In Thailand and Indonesia almost 30% of patients hospitalized with malaria are associated with pure P. vivax infections ^{8, 12}. Again further studies are needed to quantify these costs more precisely.

Diagnosis and Management of Vivax Malaria

While microscopy remains the mainstay of diagnosis of vivax malaria, lack of access to good quality microscopy services in many endemic regions limits the reliability of diagnosis (as well as contributing to the underestimation of cases). HRP-2-based rapid diagnostic tests (RDTs) are usually sensitive for the diagnosis of P. falciparum, although aldolase-based and parasite lactic dehydrogenase-based RDTs have suboptimal sensitivity for P. vivax ^{54, 55}, limiting the utility of RDTs for vivax diagnosis ⁵⁶.

Chloroquine has been the mainstay of treatment for acute vivax malaria for more than 50 years with dosing recommendations generally similar to those for P. falciparum (25mg base/kg up to 1.55g, given in three or four doses over 3 days). After eradication of the asexual stages of P. vivax from the peripheral blood, patients are at risk of relapse from the dormant liver stage. In equatorial regions such relapses occur in 20-80% of patients ^{27, 57}. Chloroquine has potent activity against asexual stages, but no activity against hypnozoites. However its long terminal elimination ensures that plasma concentrations of chloroquine are sustained above the minimum inhibition concentration for at least 28 days and thus are capable of suppressing the first relapse which in tropical zones occurs at about 21 days ⁵⁸. The prevention of further relapses necessitates eradication of the dormant liver stages; however, only one antimalarial agent is registered with such activity, the 8-aminoquinoline primaquine. The standard recommendation for radical cure of P.vivax has been 15mg primaquine daily over 14 days ²⁷. In some areas, particularly the Indian subcontinent, five day courses are recommended despite there being no evidence that they prevent relapse. The efficacy of the 14 day regimen varies considerably with relative resistance documented in Oceania and parts of Asia ^{59, 60}. Several unsubstantiated reports also suggest that primaquine efficacy may be waning elsewhere ²⁷. The problem is particularly bad in Papua and Papua New Guinea where P. vivax demonstrates marked tolerance to primaquine ^{61, 62}. Higher doses or longer courses have been advocated and may be more effective ⁶³; however, in practice, long courses of unsupervised therapy are rarely adhered to and thus likely to be ineffective. Furthermore, primaquine is contraindicated in infancy and pregnancy and causes hemolysis in G6PD deficient individuals. In endemic settings shorter alternatives are needed urgently. Short course tafenoquine has potential ⁶⁴, but is not registered and concerns about its prolonged hemolytic potential has tempered enthusiasm for its use in endemic areas where routine screening for G6PD deficiency is not possible.

Chloroquine resistance (CQR)

Whereas chloroquine-resistant P. falciparum was first documented over 50 years ago, drug resistant strains of P. vivax have taken much longer to emerge. This may reflect species differences in the underlying molecular mechanisms of resistance or lower selection pressure because of differences in the biology of gametocytogenesis. Gametocytogenesis in P. falciparum occurs after the appearance of symptoms and is refractory to schizontocidal drugs, thus favoring the transmission of resistant genotypes. In contrast gametocytes occur earlier in P. vivax compared to P. falciparum infections and are eliminated rapidly by schizontocidal agents, ensuring early transmission prior to selection.

The first case of CQR P. vivax (CQRPV) was reported in 1989 from Papua New Guinea ⁶⁵ and was soon followed by reports from northern Papua, Indonesia ^{66, 67, 68}. By 1995, in some parts of Papua almost all patients with vivax malaria treated with chloroquine monotherapy failed therapy within 28 days ⁶⁹. Other studies subsequently confirmed a high prevalence of chloroquine-resistance elsewhere in Papua ⁷⁰ and other parts of Indonesia ^{71, 72}. Sporadic cases have also been reported from Burma ⁷³, South America ^{74, 75}, Viet Nam ⁷⁶, and Turkey ⁷⁷; see figure 1. In contrast chloroquine mostly retains its efficacy against P. vivax in Thailand and India, ^{57, 78, 79} Despite these reports, the global prevalence of CQRPV is still poorly defined. In a recent systematic review only 11% (47/435) of published antimalarial drug trials between 1966 and 2002 addressed the efficacy of antimalarial drugs against P. vivax ​⁸⁰. This reflects both the inherent difficulties associated with the in vivo test for relapsing malarias and, in the absence of any empirical evidence, the belief that established protocols continue to be effective. The latter was a major reason why policies were so slow to respond to the collapse of chloroquine efficacy against P. falciparum in sub-Saharan Africa a decade ago. The WHO 1999 Guidelines recommended a 14 day follow up for the in vivo diagnosis of CQR P. vivax and such a short duration is incapable of detecting early decline in antimalarial therapeutic efficacy ⁸¹. The 2005 WHO guidelines have extended follow up to 28 days, which improves the sensitivity of the test, but at the expense of increasing confounding results from reinfection and relapse. Parasite genotyping has been useful in discriminating reinfection and recrudescence in P. falciparum, but the use of a similar strategy in P. vivax is more complex ⁸². Relapses can occur either from the primary infection (in which case they are identical to recrudescent infections) or from a previous infection (in which case they maybe be misclassified as a reinfections): allocating a treatment outcome to a recurrent parasitaemia is therefore not possible. In practice recurrent infections occurring within 28 days of supervised chloroquine administration will be growing in the presence of chloroquine concentrations normally considered above those required to inhibit parasite multiplication and thus can be considered resistant ⁸³. Monitoring the uncorrected cure rate at day 28 is therefore a reasonable indicator of declining chloroquine efficacy which can be confirmed by analysis of whole blood chloroquine concentrations to ensure adequate drug absorption.

Open in a separate window

Figure 1

Published reports of treatment failure following chloroquine monotherapy (25mg base/kg up to 1.55g over three days) for P. vivax infections. Large white stars = clinical trials of chloroquine monotherapy with >10% recurrence rate by day 28; small black stars = case series with <5 recurrences before day 28 (with or without chloroquine plasma drug levels); black circles = clinical trials after 2000, with no recurrences by day 28.

In vitro susceptibility assays provide another means of assessing drug susceptibility of Plasmodium spp. free from the confounders of host immunity, relapse and reinfection. Although these tests have been used in P. falciparum for more than 30 years, their application in P. vivax has been more difficult to develop because of the inability to culture the parasite through schizogeny in an ex-vivo assay. Recently several groups from Thailand where P.vivax responds well to chloroquine have reported successful adaptation of schizont maturation assays deriving IC50 values for P. vivax to chloroquine in the range of 30-120nM ^{79, 84, 85}. Preliminary studies from Papua, Indonesia, where clinical studies have established the presence of high grade CQR, demonstrate that chloroquine susceptibility of P. vivax is five fold lower than that reported in Thailand ⁸⁶. These findings suggest that in vitro testing may prove useful in monitoring for the emergence of resistance.

The other important tool for monitoring antimalarial drug resistance is tracking the prevalence of genetic markers of resistance. However our current understanding of the molecular mechanisms of CQR in P. vivax is limited and preliminary studies suggest that resistance mechanisms may be different from that seen in P. falciparum ​⁸⁷. Renewed interest in identifying putative molecular markers of CQRPV has been stimulated by the recent completion of the vivax genome project.

Treatment of resistant P. vivax

In vivo drug studies on strains of P. vivax sensitive to chloroquine have demonstrated the efficacy of artemisinin derivatives, mefloquine, quinine and halofantrine ⁸⁸, however few studies have addressed suitable treatment regimens for CQR vivax malaria ⁸³. P. vivax has previously been regarded as intrinsically resistant to antifols. It is not; but resistance develops readily ^{68, 89}, through the sequential acquisition of mutations in Pvdhfr ^{68, 90, 91}. The prevalence of antifol resistance is similar to that of P.falciparum with high levels of resistance in much of SE Asia ⁹² (Laos is a notable exception) and lesser degrees of resistance in the Indian subcontinent.

In Indonesia the efficacy of amodiaquine is superior to that of chloroquine, but in regions where high levels of CQR predominate, day 28 failure rates exceed 25% suggesting amodiaquine monotherapy is not a suitable alternative once high grade resistance has been reached ⁹³. Better success in the same region has been obtained with mefloquine and halofantrine which retain almost 100% cure rates by day 28 ^{94, 95}. Atovaquone-proguanil plus primaquine also showed excellent efficacy against CQR P. vivax ​⁹⁶, however the high cost of this regimen (∼$40) virtually precludes its use in endemic communities.

The artemisinin derivatives are highly potent antimalarials and are even more active against P. vivax than P. falciparum ​⁹⁷. Artemisinin combination therapy (ACT) is now widely advocated for the treatment of P. falciparum infections. In practice only a minority of infections have correct species identification prior to treatment and therefore in Asia and South America it is highly likely that other species will also be treated with ACT. Despite this, only three studies have addressed the efficacy of these combinations against P. vivax. In the first, a three day course of artesunate plus sulfadoxine-pyrimethamine resulted in a recurrence rate of 10.5% by day 28 ⁶⁸. In two subsequent studies of patients with pure P. vivax infections, recurrence rates by day 28 were significantly higher after artemether-lumefantrine (23%) and amodiaquine-artesunate (12%) than after dihydroartemisinin piperaquine (3.6%) ^{93, 98}. None of these antimalarial drugs have any efficacy on the hypnozoite stages of P. vivax and hence the difference in cure rates is likely to be a consequence of the difference between the terminal elimination half life of lumefantrine (∼4 days), amodiaquine (18 days) and that of piperaquine (28-35 days). Piperaquine, like chloroquine, suppresses the first relapse whereas lumefantrine and amodiaquine do not. In areas where primaquine can not be supervised such post treatment prophylaxis afforded by drugs with long half lives will not prevent subsequent relapses, but will provide the only practical means currently available of delaying the timing of clinical illness from these relapses. The latter affords patients a longer period free from symptomatic malaria, decreases the risk of anaemia and reduces significantly gametocyte carriage and thus the transmission potential to the mosquito vector ⁹³. Although longitudinal studies are needed to confirm these findings, when choosing an alternative to chloroquine in communities where CQR vivax malaria is endemic the relevance of post treatment prophylaxis should not be overlooked.

We thank Simon Hay, Kathryn Maitland, Bob Snow, and Shunmay Yeung for their helpful comments on the manuscript. RP is funded by a Wellcome Trust Career Development Award (074637). CAG is supported by a Wellcome Trust Project grant (#076951) attached to the Malaria Atlas Project (MAP, http://www.map.ox.ac.uk). NJW is a Wellcome Trust Principal Fellow. NA is supported by an NHMRC Practitioner Fellowship.