Interactions between HIV and malaria in non-pregnant adults: evidence and implications

Abstract

Introduction Malaria and HIV are two of the most common and important health problems facing developing countries. It is estimated that over 40 million people are living with HIV globally [1] and there are 350–500 million clinical malaria episodes annually [2]. Even modest interactions between the two infections would have substantial public health implications in resource-constrained countries, especially in subSaharan Africa, where both infections are highly prevalent. Our current immunological knowledge suggests potential for a detrimental interaction in both directions. HIV infection impairs T-cell immunity, which is of crucial importance for antimalarial responses [3]. Therefore, in theory, HIV immunosuppression should increase the risk and severity of malarial infection. In addition, malaria infection activates T-cells, promoting HIV replication [4,5]. Since increased HIV RNA levels are associated with accelerated disease progression [6,7], malaria could potentially facilitate faster progression to AIDS and death [8]. Malaria–HIV interactions have been clearly demonstrated in young children, in whom malaria-induced anaemia leads to blood transfusions, which may transmit HIV [9–11]. Also, in pregnant women, HIV contributes to higher malaria infection rates, higher parasite density, more clinical illness, more anaemia, and diminished response to treatment [12–16]. Malaria infection also contributes to higher maternal HIV viral load, but whether this leads to increased mother-to-child HIV transmission is uncertain. Most studies have not found any evidence of an effect [12,17–21], although one study found a substantial increase in HIV transmission associated with malaria infection, and another with the rainy season when malaria infection rates are high [22,23]. This study reviews the evidence regarding HIV–malaria interactions in non-pregnant adults, for whom the likely overall public health impact is less clear. It focuses on how HIV alters the clinical presentation and outcome of malaria, what effect malaria may have on the natural history of HIV infection, and the implications of an interaction for surveillance systems, burden estimates and for the prediction of future trends. The implications of the interactions on policies for prevention and control of both diseases are considered as are priority areas for research. The data were obtained by searching PubMed for English publications using the keywords HIV, malaria, and adults; further information was obtained from abstracts of scientific meetings, the Internet and personal communications with scientists. Malaria in this review refers to infection with Plasmodium falciparum and HIV to HIV-1 unless otherwise specified. Effect of HIV on malaria The effect of HIV on malaria infection and clinical illness Early studies Early studies of the interaction between HIV and malaria were reviewed in 1992 [24] and again in 1998 [25]. Many of the studies available for those reviews had several features in common: they did not report CD4 cell counts and so did not distinguish between HIV infection and the associated immunodeficiency; and most were facility-based case–control or cross-sectional designs with limited ability to assess any longitudinal population-based impact of the two infections and their interaction [26]. Only one cohort study (involving both children and adults) had been published at the time [27], and this reported increased rates of non-severe malaria and significantly higher fever rates among HIV-seropositive individuals compared with seronegative individuals in Kinshasa. However, there were no significant differences in malaria slide positivity rate or parasite densities and the authors concluded that the higher rates of malaria were a consequence of ascertainment bias. Several other studies raised the prospect of an association between malaria and HIV. A cross-sectional study in rural Tanzania reported increased prevalence of parasitaemia, and non-significant increases in rates of non-severe malaria, among HIV-seropositive adults compared with those who were seronegative [28]. Hospital-based studies in Zambia and Burundi found increased malaria case fatality ratios in patients with HIV compared with HIV-seronegative individuals, but the small sample sizes did not allow firm conclusions to be drawn [29,30]. Overall, these early studies and reviews concluded that there was little evidence of substantial interactions between HIV and malaria in non-pregnant adults. In contrast to similar reviews of HIV and tuberculosis or other opportunistic infections, it was clear that malaria was not going to be a florid opportunistic infection affecting many HIV-infected adults soon after immunodeficiency developed. However, more subtle interactions and programmatic considerations were generally ignored and a period of relatively little scientific investigation on HIV–malaria interactions ensued. Recent studies Studies published since the late 1990s have begun to clarify the nature and scale of the interaction. The underlying epidemiology and intensity of malaria transmission and the duration of HIV infection and consequent immunosuppression appear to be critical in determining the consequences of coinfection. In areas of stable malaria, with intense continuous transmission, adults acquire substantial immunity against malaria through repeated infections in early life such that infection and clinical illness are uncommon. In contrast, young children who have not yet acquired malarial immunity and pregnant women, who transiently lose some of their acquired immunity, are at particular risk of malaria infection and its complications. In areas of unstable malaria, with intermittent unpredictable transmission, the entire population has little or no acquired immunity to malaria. While some HIV-infected individuals have a rapid course of immunosuppression and associated illnesses, most will develop gradual loss of immune responsiveness over a number of years. This leads to increased susceptibility to opportunistic infections and evidence is accumulating that this includes increased vulnerability to malaria infections and their consequences. Table 1 summarizes the findings of 12 studies on the impact of HIV on malaria infection and illness. In areas of stable malaria transmission, malaria infection rates and the frequency of clinical illness (fever and other findings) appear to be increased in HIV-infected adults, particularly those with immunosuppression, as measured by low CD4 T cell counts. In Ugandan adults, the odds of parasitaemia, and risk of malarial fever increased with decreasing CD4 cell count, such that individuals with CD4 cell counts < 200 cells/μl were more than twice as likely to suffer malarial fever as individuals with > 500 cells/μl [31,32]. Similarly, a hospital-based case–control study in northern Uganda found an association between HIV infection and symptomatic malaria [33]. A recent study in a cohort of Malawian adults reported a moderate association between malaria parasitaemia and HIV infection. High parasite densities with fever were associated with low CD4 cell counts [34]. Another study of 660 adults in a holoendemic area in urban Malawi found increasing incidence of clinical malaria, and clinical malaria with significant parasitaemia, with decreasing CD4 cell counts and an inverse relationship between parasite density and CD4 cell count [35]. Put together, these studies suggest that HIV-associated immunosuppression interferes with parasite control and may cause some loss of the specific disease-protective immune response that protects parasitaemic persons from developing clinical illness.Table 1: Effect of HIV on malaria infection and disease in adults, by regions of stable and unstable malaria transmission.Table 1: (Continued)Severe or complicated malaria has not been a feature of reports from regions of stable transmission. This may be because those with a high degree of acquired immunity may be able to sustain sufficient antitoxic immunity to avoid severe disease despite HIV-associated immunosuppression [36]. However, in regions of unstable malaria transmission, HIV-infected adults suffering from malaria are at increased risk of complicated and severe malaria and of death. This is particularly alarming in the light of recent estimates that more than half of the world's malaria is in hypo- and mesoendemic areas [37], indicating huge numbers of adults at risk of severe malaria and coinfection in regions and cities with generalized HIV epidemics. Studies in South Africa reported a doubling, or more, of the risk of severe or complicated malaria, and a five-fold increased risk of death, among HIV-infected adults compared with those who were uninfected [38,39]. In a study in a rural hospital in Zimbabwe, HIV-infected adults were more likely than uninfected patients to develop malaria complications and were more than twice as likely to die [40]. In a hospital-based study in urban Burkina Faso, where HIV prevalence in adults is reported to be 5–6% [41,42], more than 30% of adults with severe malaria were infected with HIV [43]. However, the number of cases involved was small, and HIV testing was not performed on all patients with severe malaria. Outside Africa, reports include a small study in Mumbai where higher HIV prevalence was found among patients with severe malaria than in the general population [44], studies demonstrating coinfection with HIV and malaria in injecting drug users in Brazil and Vietnam, and outbreaks of malaria in injecting drug users in previously malaria-free regions owing to needle sharing [45,46]. Effect of HIV on diagnosis and treatment of malaria Generally, malaria is clinically indistinguishable from many other causes of fever in HIV-infected individuals [38,39,43,47]. In many regions, malaria is diagnosed simply on the basis of fever, and drugs are administered without confirmatory testing [48,49]. Consequently, malarial fever may be overestimated and inappropriately treated if diagnosed without blood film examination. A study in 10 Tanzanian hospitals found that only 46% of patients clinically diagnosed with severe malaria had a positive P. falciparum blood slide result, yet 95% were given quinine [50]. The overdiagnosis of malaria increases the potential for severe outcomes of other causes of fever when incorrect treatment is given [50]. Although rates of clinical malaria increase in HIV-infected adults, the much increased rates of fever as a consequence of HIV infection, and its associated complications, mean that the proportion of fever that is attributable to malaria decreases as immunosuppression progresses [32,35,51]. This reduces further the specificity of a diagnosis of malaria based on febrile symptoms alone. It can be predicted that responses to antimalarial therapy will be decreased in immunosuppressed individuals coinfected with HIV and malaria living in regions of stable transmission. This may be either because of increased susceptibility to malaria reinfection, or because of recrudescence of infection, since antimalarial therapy is most effective in individuals who already have some acquired immunity. Results of seven studies of response to antimalarial therapy in HIV-infected adults are summarized in Table 2. Early studies did not suggest any impact of HIV on success of therapy with quinine, chloroquine and sulfadoxine–pyrimethamine, which were highly effective in all who were treated [27,52]. A more recent body of evidence suggests treatment failure may be common in HIV-infected individuals. In Ethiopia, parasite and fever clearance times were prolonged in HIV-infected adults treated with artemisinin for uncomplicated malaria [53]. In Kenya, significantly lower rates of parasite clearance at 28 days post-treatment with sulfadoxine–pyrimethamine were noted in patients with CD4 cell counts < 200 cells/μl compared with HIV-uninfected adults [54]. In a randomized controlled trial of artemether–lumefantrine versus sulfadoxine–pyrimethamine for the treatment of uncomplicated adult malaria in Zambia, the frequency of malaria treatment failure, for either therapy increased significantly with advancing immunosuppression [55]. A study in Uganda concluded, on the basis of molecular genotyping, that the increased risk of clinical treatment failure was a result of new infections rather than recrudescence [56]. There are no reports of increased adverse events associated with antimalarial therapy in HIV-infected individuals.Table 2: Summary of studies recording response to antimalarial therapy in HIV-infected adults.Effect of HIV on transmission of malaria No direct evidence has been found to suggest that HIV alters malaria transmission. Investigators have speculated that because HIV infection increases parasitaemia and reduces the response to therapy, it will increase the reservoir of infection in the human population and hence increase transmission [8,57]. In an area with high HIV prevalence (30%), it has been estimated that 20% of malaria infection could be attributable to HIV, assuming gametocytaemia parallels asexual parasitaemia [31]. However, the relationship between asexual parasitaemia and gametocytaemia is not straightforward and the effect of HIV infection on gametocytes remains unclear. The use of sulphonamide preparations, sulfadoxine–pyrimethamine for antimalarial treatment and cotrimoxazole for opportunistic infection prophylaxis, has been shown to increase the production of gametocytes [58]. However, a recent study reports a 38% decline in malaria incidence in household members of HIV-infected persons taking regular cotrimoxazole prophylaxis [59]. Nonetheless there may be an advantage from using effective artemisin-based combination therapy, particularly by rapidly lowering the burden of asexual parasites and transmissible gametocytes [60,61]. Effect of HIV treatment and prophylaxis on malaria A recent study in Uganda in HIV-infected adults found a reduction in febrile parasitaemia of 76% [95% confidence interval (CI), 62–85) associated with cotrimoxazole prophylaxis, of 92% (CI, 83–96%) associated with cotrimoxazole and antiretroviral treatment (ART), and of 95% (CI, 92–97) when insecticide-treated bed nets were added to cotrimoxazole and ART. This study demonstrates that the effects of these three interventions are cumulative, and that malaria infection can be successfully controlled in HIV-infected adults [62]. Effect of malaria on HIV Effect of malaria infection on HIV viral load Upregulation of HIV viral transcription has been described for coinfection with several pathogens [63–66] and appears to be the case for malaria, as summarized in Table 3. Investigators in Malawi observed nearly a doubling of HIV viral load in 77 subjects following malarial infection and an increase of almost 1 log copies/ml in 13 with CD4 cell count > 300 cells/μl who fulfilled a stricter definition (fever and high parasite density) of clinical malaria [8]. This study confirmed earlier reports from both Malawi [57] and Uganda [67] of an acute rise in HIV viral load at the time of malaria infection. In each study, convalescent viral loads generally returned to preinfection levels following malaria treatment.Table 3: Effect of malaria on HIV infection in adults.By contrast, in one small study in rural Guinea-Bissau, mostly involving HIV-2-infected individuals, no increase in viral load was found in HIV-infected patients during the wet (malaria) season compared with the dry season [68]. Effect of malaria infection on HIV disease progression and treatment No direct evidence has been found to suggest that malaria alters HIV disease progression. Investigators have speculated that transient increases in HIV viral load associated with malaria could accelerate disease progression [8,57]. However, other reports have suggested minimal differences between untreated median survival time in developed (non-malarious) and developing (malarious) countries [69,70]. Evidence to implicate these transient increases in viral load as clinically important is scarce. Indeed, a report from a cohort of HIV-infected individuals and uninfected controls in rural Uganda concluded that there was no strong detrimental effect of malaria on all-cause mortality in HIV-seropositive adults [71] (Table 3). However, another study in Uganda found significantly faster CD4 cell declines in individuals following episodes of malaria [72]. Indirect evidence of a possible role for malaria prevention in limiting HIV disease progression comes from studies in Côte d'Ivoire, Uganda and Kenya examining the effect of daily cotrimoxazole as prophylaxis against opportunistic infections, which consistently reported better prognosis and reduced rates of malaria [73–75]. In one study, failure to treat undiagnosed fevers with antimalarial drugs was also associated with death [51]. Successful ART is expected to reduce the risk of serious opportunistic infections dramatically, and this appears to include clinical malaria [62]. Additionally there are documented in vitro effects of ART on malaria parasites [76,77]. It would be prudent to ensure that malaria surveillance is part of any broader monitoring of ART outcomes in areas where ART use is increasing and HIV and malaria coexist. The potential for interactions between antimalarial and antiretroviral drugs has recently been reviewed [78]. Effect of malaria infection on HIV transmission No direct evidence has been found to suggest that malaria alters HIV transmission in adults. High HIV viral loads have been shown to be associated with increased potential for HIV transmission [79,80], but there is no empiric evidence to implicate the transient increases in HIV viral load associated with malaria infection with increases in HIV transmission [81]. A primary concern is the transmission of HIV through injections and transfusions for the management of malarial fever or malaria-induced anaemia [9–11]. While this is a documented problem in young children in areas of high malaria transmission [9], malaria-induced anaemia requiring transfusion is much less common in adults and so may not constitute a major means of HIV transmission. HIV and malaria programme overlap Current prevention and control programmes for HIV and malaria are defined and separate from one another, despite having a number of common features and goals. There may be substantial overlap in risk population groups, with common infrastructure needs. Many control efforts utilize the same health personnel and sites for service delivery, often with common intervention tools, programme and resource mechanisms. The overlapping characteristics of these programmes are shown in Table 4. In areas where both infections are common, specific coordinated prevention and treatment measures for adults that would have immediate impact on both diseases include:Table 4: Common features of HIV and malaria programmes and infrastructure needs. strengthening diagnostics and syndromic treatment protocols [82] for both HIV and malaria so that expensive drugs can be directed to those in need encouraging all identified HIV-infected persons (and especially those under HIV care and treatment programmes) to sleep under an insecticide-treated bed net [62] and use other malaria-prevention measures (indoor residual spraying, other personal protection measures) as appropriate providing all HIV-infected persons with highly effective antimalarial drugs for the treatment of any malaria illness providing all symptomatic HIV-infected persons with cotrimoxazole as prophylaxis for opportunistic infections; this is also likely to prevent malaria infection [62,73–75] providing voluntary counselling and testing sites for persons with unknown HIV status at places where malaria interventions, such as free or highly subsidized insecticide-treated bed nets, might be distributed, thus serving as an incentive for many to attend including interventions for both infections for pregnant women, and anaemia prevention and safe use of blood products for young children [12], as part of a comprehensive strategy for the two diseases. Implications, summary and conclusions HIV and malaria coexist in regions with the most poorly developed health surveillance systems, making the public health scale of any interaction difficult to determine. In general, the national malaria estimates by the World Health Organization and the UNAIDS national HIV estimates are derived with much extrapolation from limited data and thus the estimates are imprecise [2,36,83–87]. Large sections of the populations of countries with generalized HIV epidemics (> 1% HIV prevalence in the general population) and a high malaria burden are at particular risk of coinfection. In low-level and concentrated HIV epidemics (< 1% HIV prevalence in the general population), specific population subgroups are likely to be most at risk of coinfection by virtue of their behaviour. For example, expanding HIV epidemics are being fuelled by injecting drug use and heroin trafficking in malaria-endemic areas in Asia and Latin America [88,89] and increasing incidence of coinfections are to be expected in the future. The two diseases have the greatest overlap in subSaharan Africa, where 70% of the global burden of P. falciparum infections [85] and 60% of HIV infections occur [1] in an area with only 12% of the world's population. However, HIV is predominantly a disease of sexually active adults, worse in urban environments and in southern Africa, while malaria is predominantly a disease of young children and pregnant women, with the greatest burden in rural areas and in west and eastern Africa [90]. This will reduce the actual infection overlap; indeed a recent modelling study estimated that the HIV-1 epidemic has increased the incidence of clinical malaria by only 1.3%, and malaria deaths by 4.9% in subSaharan Africa as a whole [91]. However, in regions of unstable malaria transmission, where adults are affected by both diseases, and especially in southern Africa, the HIV-attributable increase in clinical malaria incidence may rise up to 28%, and for malaria deaths up to 114%. However, more work is needed at subnational level to obtain an accurate estimate of the impact of each infection on the other. There is increasing clarity on the spectrum of interaction between HIV and malaria in non-pregnant adults (Table 5). The HIV–malaria biological interaction is apparently driven by substantial immune compromise, which contributes to more malaria infection, somewhat higher parasite density, more malaria clinical illness, and possibly a diminishing response to treatment when drugs may be starting to fail. More severe malaria disease and death are seen specifically among persons with little or no acquired immunity. The effect of malaria on HIV is less clear. Malaria contributes to transient HIV viral load increases, but these are reversible with effective therapy. Much remains to be discovered about the interaction between HIV and malaria, and some of the most important gaps in our knowledge are priorities for research:Table 5: Specific biological, clinical and transmission effects of HIV–malaria interactions. the effect of HIV infection on malarial infection and disease across the full range of malaria transmission settings the effect of HIV-associated immunosuppression on malaria transmission the clinical effect of acute malaria episodes on HIV prognosis and transmission interactions between HIV and non-falciparum plasmodium species details at subnational level of HIV and malaria distribution to identify geographical and age and sex group overlaps levels of antimalarial drug use in relation to HIV status the contribution of inaccurate malaria case definition and diagnosis to the overuse of drugs and clinical outcome the effects of ART and cotrimoxazole on malaria diagnosis, rates of clinical disease and the development of drug resistance in different settings interactions between antimalarial drugs, ART and cotrimoxazole the effect of intermittent presumptive treatment of malaria for HIV-infected individuals the impact on HIV infection of changing antimalarial drug regimens the contribution of HIV to overestimates of malaria-attributable mortality. The HIV–malaria programme interaction is extensive and driven by the common features of the clinical illnesses, the similar sites and health staff involved, and the similarity in services that must be provided, resourced and coordinated. Effective prevention and management of HIV and malaria, and indeed of tuberculosis, will jointly reduce their impact [62,92,93]. Malaria interventions are relatively inexpensive (insecticide-treated bed nets cost US$2.50–5.00 for quality products; antimalarial adult curative treatment doses are less than US$2.00) and have been shown to be highly cost effective [94]; this will be particularly true for the HIV-infected community. Collaboration in joint programming is often not achieved, but in this case and particularly for the benefit of the HIV-infected populations, joint advocacy and receptivity to combined service delivery will improve survival and the quality of many lives.