Approach to the Patient with HIV and Coinfecting Tropical Infectious Diseases

All … fall into one of two categories: Infected with HIV or at risk for HIV infection.

Mary E. Wilson¹

INTRODUCTION

The morbidity, mortality, and social disruption due to the global acquired immunodeficiency syndrome (AIDS) pandemic weigh disproportionately upon resource-poor areas of the tropics.² Because these are the very areas where “tropical” infectious diseases continue to hold greatest sway, the potential for interactions between human immunodeficiency virus (HIV) infection and other tropical infectious diseases is great.

Such interactions are marked by epidemiologic complexity. The AIDS pandemic is best described as the sum of discontinuous and overlapping epidemics of disease among populations of variable and varying risk (see Chapter 76). The predominant modes of transmission of HIV (perinatal, sexual, and parenteral) result in a bimodal distribution of disease, with peaks among young children and young adults. The risk of infection or disease due to tropical pathogens varies widely with differences in ecological, political, and socioeconomic conditions (including access to medical care); related specific host factors such as age of exposure, pregnancy, behavior, and nutrition; and host genetics. Disease due to a coinfecting pathogen may be due to primary infection, recurrent infection, or the reactivation of latent infection. For some pathogens, the risk factor responsible for the acquisition of HIV may also be the risk factor responsible for the acquisition of the coinfecting pathogen. As a consequence of this epidemiologic complexity, both the prevalence and the expression of coinfection are likely to be variable across ecological, economic, political, behavioral, and cultural divides.

There is also, of course, considerable biologic and clinical complexity in the interaction of agents of tropical diseases with HIV. Either pathogen has the potential for altering the natural history, immune response, or response to therapy of the other.^{3, 4}

EFFECTS OF HIV ON TROPICAL COINFECTIONS

Infection with HIV has the ability to influence the natural history of infection with other pathogens through (1) facilitating infection, (2) altering the incidence of disease by increasing the ratio of disease to infection, (3) changing the presentation of disease, or (4) exacerbating the course of disease.⁴ Such effects are presumed to be primarily the result of the immunosuppression associated with HIV infection.

Abnormalities of immune function are found in essentially every cellular and functional compartment of the immune system in AIDS, although profound defects in cell-mediated immunity (CMI) appear to be of greatest clinical importance.⁵ In vitro correlates include functional abnormalities of

• •CD4+ T cells, with progressive failure of proliferation, interleukin (IL)-2 and interferon-γ (IFN-γ) production,^{6, 7} dysregulated expression of molecules essential for T-cell/antigen presenting cell interactions,⁸ abnormal activation-induced apoptosis,^{9, 10} and expansion of a subset of CD4+ T cells (CD4+CD25+ regulatory T cells) that are potent inhibitors of immune responses both to self and to pathogens^{11, 12}
• •Monocyte/macrophages, with decreased chemotaxis and intracellular microbicidal activity and abnormal cytokine production^{5, 6}
• •Dendritic cells, with reduced ability to present antigen and activate T cells (along with efficient transfer of HIV infection to CD4+ T cells)¹³
• •CD8+ T cells, with decreased cytotoxic T-lymphocyte (CTL) function⁵ and abnormal activation-induced apoptosis¹⁰
• •Natural killer (NK) cells, with decreased proliferation and IFN-γ production¹⁴

Dysregulation of humoral immunity, marked by polyclonal B-cell activation, is also seen.⁵ Functionally, abnormalities of CD4+ T cell, monocyte/macrophage, and dendritic cell function have been thought to be paramount to the suppression of CMI and the opportunistic infections seen in patients with AIDS. In addition to the direct effects of HIV infection, the immune system of HIV-seropositive people may also be compromised in clinically significant ways by profound nutritional and metabolic derangements (e.g., wasting or “slim disease”), therapeutic interventions (e.g., corticosteroids used for the treatment of Pneumocystis pneumonia), and the immune abnormalities associated with secondary infections (e.g., the suppression of CMI seen in visceral leishmaniasis).

The list of infectious diseases that are exacerbated by HIV coinfection⁵ includes many that are predictable from data demonstrating the important role of CMI in protection from the etiologic agent. Several pathogens for which immunity has been presumed to depend on CMI do not appear to be exacerbated by HIV coinfection, however. The HIV “experiment of nature” has caused a reexamination of the immunology of such “missing infections” in AIDS.¹⁵

HIV infection can also influence the therapeutic response of patients with presumed tropical infection. The ability to diagnose and monitor coinfection may be compromised by aberrant serologic responses, including false positives due to polyclonal B-cell activation, false negatives due to blunted antigen-specific antibody responses to newly acquired pathogens in late HIV disease, and false serologic reversion after treatment in late HIV disease.⁵ Diagnosis may also be hindered by unusual presentations of disease with coinfecting pathogens. Finally, a plethora of intercurrent pathologic conditions may lead to a dulling of Ockham's razor during the evaluation of disease in AIDS patients. A single pathogen, multiple pathogens, HIV infection, side effects of therapeutic drugs, or a combination of these may be responsible for the presenting complaints. Given the suboptimal response to the chemotherapy of many infections in the presence of profound immunosuppression, drugs may need to be given in greater numbers or for a longer duration. With many pathogens, in the absence of immune reconstitution resulting from highly active antiretroviral therapy (HAART), lifelong suppressive therapy is necessary. Drug therapy in the HIV-infected patient may be complicated further by increased rates of drug allergy as well as by untoward drug interactions in the setting of polypharmacy.¹⁶ Prophylaxis against coinfections may be compromised by substandard vaccination responses.17, 18, 19 Finally, the presence of HIV coinfection can complicate the public health consequences of tropical diseases. AIDS may increase the transmissibility of secondary infections and provide fertile soil for the development of drug resistance. Public health resources devoted to the AIDS pandemic may divert resources away from the control and prevention of other infectious diseases.

The presence of HIV coinfection can lead to disease of markedly greater incidence or severity [the standard definition of an opportunistic infection (OI)]²⁰ with some tropical infectious diseases, such as leishmaniasis and American trypanosomiasis (Chagas' disease). Coinfection has also been demonstrated to have subtle effects on the course of disease with other tropical agents, such as Schistosoma mansoni. No identified alteration has been found in the natural history of many tropical infections, including most nematodes.

With organisms in the latter groups, the current absence of evidence of significant effects of HIV on the expression of disease or the response to treatment should not be construed as strong evidence for the absence of such effects. Most research resources have been spent on understanding the clinical and epidemiologic manifestations of the HIV pandemic in industrialized countries, where tropical infectious diseases are obviously underrepresented.^{4, 21} Where coinfections with HIV and endemic tropical diseases are marked by low prevalence, subtlety of interaction, diagnostic difficulty, or low research priority, the presence and significance of any interaction are likely to be missed. For example, despite the research priority among tropical infections accorded to malaria, a significant interaction with HIV infection—the lack of a benefit of increasing parity in the control of malaria in pregnant women—was only discovered 15 years after the AIDS epidemic was recognized.²² With less heavily studied pathogens, comparatively subtle interactions will likely emerge over time as research resources are appropriately directed.

Focusing on OIs may help to highlight some of the clearest data on the clinical expression of AIDS in the tropics. Of the more than 100 agents known to cause OIs in AIDS patients, several are classic tropical pathogens.²⁰ These are mostly found among the intracellular protozoans, bacteria, and endemic fungi; there is a marked absence of metazoans. Overall, the clinical expression of AIDS in many resource-poor areas of the tropics appears to involve a different spectrum of OIs than those common in North America and Europe. In place of the high incidences of Pneumocystis carinii pneumonia (PCP), disseminated Mycobacterium avium, and cytomegalovirus (CMV) found in the resource-rich north, the clinical expression of AIDS in much of the tropics has been marked by frequent tuberculosis (the most common serious AIDS OI in the world), chronic diarrhea, wasting, chronic fever without an obvious localizing source, and pulmonary disease.²³

The contribution of predominantly tropical pathogens to these latter common syndromes is unclear, which illuminates the problems with much of the available data on HIV disease in the tropics. Understanding the spectrum of AIDS-associated OIs in any given area depends on the presence of adequate surveillance systems, which are often lacking in resource-poor regions of the tropics. In the presence of inadequate infrastructures, limited financial resources, and difficult access to medical care on the part of the socially disadvantaged, surveillance is likely to be sporadic and to involve mainly the sampling of subgroups of AIDS patients at late stages of disease.²¹ Where resources are limited, diagnostic reporting is likely to be biased in favor of OIs that are inexpensive to diagnose (or misdiagnose).²¹ Even the common impression that the progression of AIDS is more rapid in sub-Saharan Africa than in industrialized countries24, 25, 26, 27 rests on data that are less than robust.²⁸ A more rapid observed course (presumed due to a higher frequency and virulence of coinfection and problems of nutrition and access to medical care) may represent in large part a systematic bias in favor of later initial diagnosis of HIV infection and AIDS.^{21, 23} Conversely, it has been suggested that the burden of illness and mortality in early HIV disease (often unrecognized as such) due to high-grade pathogens, such as Streptococcus pneumoniae, Mycobacterium tuberculosis, and the salmonellae, may rival that due to the OIs of late-stage AIDS in the tropics.^{29, 30}

HIV has shed light on many previously obscure human pathogens. Some, such as the enteric microsporidians, were unknown as agents of human disease prior to the AIDS epidemic. Others, such as Cryptosporidium parvum, were underappreciated as causes of disease in normal hosts until their prevalence in AIDS patients led to systematic study in normal hosts. The list of agents causing OIs in AIDS patients is bound to expand. It is reasonable to expect that tropical regions will be prime locations for the identification of further such agents.

EFFECTS OF TROPICAL INFECTIONS ON HIV COINFECTION

There are theoretical and experimental reasons to believe that coinfection can significantly alter the course of HIV pathogenesis. The central role of ongoing viral replication in HIV pathogenesis is firmly established, and the set point concentration of plasma viremia correlates well with long-term clinical outcome.³¹ It is presumed that any increases in viral replication have the potential for accelerating the course of disease. Efficient replication of HIV in CD4+ T cells is dependent on cellular activation. Similarly, activation of monocyte/macrophages and dendritic cells can stimulate HIV replication by increasing transcription factor binding to the HIV long terminal repeat region (LTR), enhancing LTR-directed transcription. Coinfecting pathogens can stimulate such immune cell activation either directly (e.g., stimulating signaling through pathogen recognition receptors such as Toll-like receptors³² or upregulating transcription factor transactivation in coinfected immune cells) or indirectly (e.g., promoting the generation of proinflammatory cytokines such as tumor necrosis factor-α [TNF-α] or activating CD4+ T cells as part of the adaptive immune response). Immune activation can also lead to upregulation of the expression of HIV coreceptors,³³ thereby facilitating the infection of fresh cells.

Immunologic responses to pathogens, as well as to purified vaccine antigens, clearly have the potential for enhancing the dynamic burden of HIV replication. In vitro studies with diverse pathogens have provided mechanistic support for this idea.34, 35, 36 Experimental evidence has also suggested that immune activation-driven augmentation of the HIV viral burden can occur in vivo.37, 38, 39, 40, 41, 42, 43, 44 There is also circumstantial evidence that such immune activation may enhance HIV pathogenesis.^{45, 46} Both points remain somewhat controversial, however.⁴⁷ Whether immune activation-related increases in viral load actually accelerate the pathogenesis of HIV may depend on whether the changes are transient (as with immunization or with treated acute infection) or chronic (as with untreated or untreatable infection, or through a resetting of the set point of plasma viremia by a particular coinfection).⁴⁸

Direct equation of immune activation with upregulation of HIV replication is simplistic, however. With CD4+ T cells, the mechanism of activation appears to be critical to whether viral replication is induced or suppressed.^{49, 50} Furthermore, activation of proinflammatory cytokine production with positive effects on HIV replication goes hand in hand with activation of anti-inflammatory cytokine production that can inhibit HIV replication.⁵¹ More generally, proinflammatory responses reliably induce counterregulatory responses that suppress subsequent immune activation.⁵² It thus should not be surprising that in vitro studies have provided mechanistic support for the ability of coinfecting pathogens to suppress HIV replication.53, 54, 55, 56 Indeed, the overall effect of acute coinfection with some pathogens, including measles virus, dengue virus, and Orientia tsutsugamushi, may be a decrease in HIV viral load.57, 58, 59

In addition to the viral sequelae of generalized immuno-logic activation, induction of specific alterations in the immunoregulatory environment of the host by ubiquitous tropical pathogens has been postulated to accelerate the course of HIV. Cross-regulating subsets of CD4+ T cells have been distinguished by their cytokine profiles and functional activities: Th1 cells (producing IFN-γ and IL-2, among other cytokines) are important in classical macrophage activation, the development of CMI, and the generation of humoral responses involving complement-fixing antibody isotypes; Th2 cells (producing IL-4, IL-5, and IL-13, among other cytokines) are important in alternative macrophage activation, the generation of immunoglobulin E (IgE) responses, eosinophilia, mast cell responses, and atopy. The immunologic response to most helminthic parasites is dominated by the production of Th2 cytokines. Evidence from murine systems shows that helminth-driven Th2 polarization can shift the immunologic response to heterologous antigens and pathogens from a Th1- to a Th2-dominant pattern, as well as significantly suppress CD8-mediated viral clearance.^{60, 61} Such responses have also been found to impair antigen-specific Th1 immune responses in both mice and humans.^{60, 62} Helminthic infection is chronic and widespread in the tropics. The resultant Th2 “priming” of the immune system may favor progression of HIV disease.⁶³ Several mechanisms have been postulated. First, a Th2-polarized immune system may directly suppress CD8 T cell–mediated anti-HIV responses.⁶¹ Second, HIV may preferentially replicate in Th2 cells.⁶⁴ Third, T cells from HIV-seropositive people undergo abnormal activation-induced apoptosis,^{9, 10} which is thought to play a role in the depletion of both CD4+ and CD8+ T cells over time. Th2 cytokines can amplify such activation-induced apoptosis.⁶⁵ Fourth, Th2 cytokines can upregulate HIV coreceptor expression by CD4+ T cells and monocyte/macrophages.^{66, 67} Of note, it has been demonstrated that peripheral blood cells from patients with intestinal helminth infection are more susceptible to in vitro infection with HIV than are cells from helminth-uninfected patients.^{68, 69}

Although intriguing, the hypothesis that endemic helminth coinfection leads to acceleration of the disease course of HIV remains unproven. A study in Ethiopia indicated that HIV viral load was significantly higher in individuals with various helminthic infections than in individuals without helminths, correlating positively with the parasite load as well as decreasing after elimination of the worms by antiparasitic treatment.⁷⁰ However, similar studies performed in Uganda and examining larger numbers of patients convincingly failed to replicate these findings.^{71, 72} These latter studies strongly suggest that helminth coinfection is not associated with faster progression of HIV disease.

Other effects of coinfection are perhaps more concrete. Tropical diseases may lead directly to an increased risk of infection with HIV. Treatment of the severe anemia induced by malaria has led to the HIV infection of countless children by transfusion.⁷³ Genital schistosomiasis, like other genital inflammatory conditions, may increase the efficiency of HIV transmission.⁷⁴

CLINICAL SUSPICION OF COINFECTION

Lewis Thomas wrote that infectious disease “usually results from inconclusive negotiations for symbiosis.”⁷⁵ In this light, most tropical pathogens lead to high infection:disease ratios. Whether disease is present or not, sterile immunity is not thought to be achieved after infection with most intracellular protozoans. This is as much the case with the tropical and subtropical agents of leishmaniasis and Chagas' disease as it is with the ubiquitous agent of toxoplasmosis. These are thus lifelong infections, with serious disease often resulting when an HIV-infected host defaults on his or her side of these immunologic “negotiations.” From the point of view of the clinician in a nonendemic area, this necessitates obtaining a lifetime travel history for all HIV-infected patients. From the perspective of the clinician from either endemic or nonendemic areas, it is equally necessary to have an appropriate index of suspicion for HIV coinfection when contemplating the possibility of a tropical infectious disease. In the process of giving this important virus its diagnostic due, however, it is critical to avoid letting the presence or suspicion of HIV infection distract from the careful consideration of etiologic agents that present and respond similarly in HIV-seronegative and HIV-seropositive people.

The basic biology of HIV; the progression, diagnosis, and treatment of HIV disease; and the epidemiology of HIV/AIDS in the tropics are discussed in Chapter 76. This chapter focuses on the natural history, diagnosis, treatment, and prevention of tropical diseases and OIs in the HIV patient. Distinguishing tropical from other pathogens has its artificial side. Measles, not usually thought of as a tropical disease, causes significant mortality only in the tropics today. Tuberculosis, a pandemic infection, is the principal AIDS-related OI in the tropics. Given the historical preoccupation of the field of tropical medicine with parasitic infections, such infections receive close attention here. For practical reasons, OIs that are common in the industrialized world are not discussed in depth unless there are compelling clinical or epidemiologic reasons for doing so. Multiple references are available via the Internet that discuss these cosmopolitan OIs in detail.76, 77, 78, 79 Further information on all of the specific organisms discussed here can also be found in the relevant chapters of this book.

The diagnostic, prophylactic, and therapeutic recommendations discussed here describe an approach to the HIV patient that is not limited by scarce medical resources. As such, like many strategies for dealing with HIV disease that have evolved in affluent industrialized countries (including HAART, high-technology diagnostics, and multidrug chemoprophylaxis), many of these recommendations may not easily be translated to resource-poor areas of the tropics.

PATHOGENS

Protozoan Infections

Malaria

Malaria (see Chapter 90) remains one of the most important infectious diseases in the world today, causing 100 to 200 million new cases and 1 to 2 million deaths each year. Evidence from both murine and human studies suggests an important role for CD4+ T cells in protective immunity to blood-stage malaria. With large areas of shared endemnicity and prevalence, a medically significant interaction between HIV and malaria was thus expected and feared.⁸⁰ Initial studies were negative; falciparum malaria did not appear be an OI or to accelerate the progression of HIV disease.80, 81, 82, 83, 84, 85 However, follow-up studies have revealed significant bidirectional interactions between Plasmodium falciparum and HIV.

HIV replication in peripheral blood cells is enhanced by exposure to P. falciparum antigens in vitro, in part through induction of TNF-α.⁸⁶ Increased HIV replication in dendritic cells has also been seen after the in vivo infection with Plasmodium chabaudi of mice transgenic for the HIV genome in a process that appears to be dependent on CD4+ T-cell activation.⁸⁷ The production of TNF-α during malarial paroxysms,⁸⁸ along with the antigenic exposure of parasitemia, might thus reasonably be expected to increase HIV viral load. Indeed, clinical studies from Malawi have shown that P. falciparum infection is associated with increased HIV viral burden in peripheral as well as placental blood.^{89, 90} Treatment was associated with a reduction in viral load, although it remained elevated compared to controls for the 4-week duration of the study.⁸⁹ Whether or not malaria-mediated increases in HIV replication accelerate the course of HIV disease remains to be determined.

The first significant clinical effect of HIV on malaria was found in the setting of pregnancy. In areas of high malarial endemnicity, the high degree of immunity that women of childbearing age have developed to severe malaria is compromised by pregnancy. The placental vasculature shields parasitized erythrocytes from the systemic immune response, allowing localized erythrocytic replication of the parasite. Placental parasitemia has been associated with low birth weight and, hence, increased infant mortality. Local uteroplacental immune responses do restrict parasite replication, however, and the effectiveness of these local responses increases in subsequent pregnancies under pressure of recurrent malarial exposure. In 1996, a study performed in rural Malawi demonstrated that the beneficial effects (maternal, placental, and neonatal) of parity in the control of parasitemia during pregnancy were markedly attenuated in the face of HIV coinfection.²² Since then, multiple studies performed in sub-Saharan Africa have probed the effects of coinfection on the outcome of pregnancy.90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100 Such studies have shown that (1) HIV infection is associated with increased rates and levels of peripheral and placental parasitemia, clinical malaria, and maternal anemia in pregnant women, and (2) coinfection is associated with a higher risk of low birth weight, preterm birth, intrauterine growth retardation, and postnatal infant mortality.¹⁰¹ Although placental parasitemia is associated with increased placental HIV viral loads in coinfected patients,⁹⁰ it remains unclear as to whether malaria infection increases the risk of mother-to-child transmission of HIV. Conflicting results (enhancement, protection, and no effect) have been published.^{96, 97, 100} A World Health Organization (WHO) technical consultation has recommended that HIV-infected pregnant women who are at risk for malaria should always have protection with insecticide-treated bed nets, along with (according to HIV stage) either intermittent preventive treatment with sulfadoxine/pyrimethamine or daily trimethoprim– sulfamethoxazole (cotrimoxazole; TMP–SMX) prophylaxis.¹⁰²

HIV infection is associated with an increased incidence of parasitemia, the risk of clinical malaria is significantly higher in HIV-infected adults, and this risk increases with decreasing CD4 cell count.103, 104, 105 Notably, HIV infection has also been found to be a risk factor for severe malaria in nonimmune populations (in an area of unstable transmission).¹⁰⁶ Finally, although early studies found no evidence that the treatment or prophylaxis of malaria was altered by HIV coinfection,^{81, 83} the risk of malaria treatment failure has recently been found to be higher in CD4-depleted HIV positive individuals.^{107, 108}

HIV protease inhibitors currently used for HAART have understudied interactions with a variety of drugs used for malaria prophylaxis and treatment. A variety of protease inhibitors, as well as the nonnucleoside reverse transcriptase inhibitors delaviradine and efavirenz, inhibit hepatic cytochrome P-450 enzymes. Principal effects are on the CYP3A4 isoform (with ritonavir being the most potent inhibitor), although the CYP2D6 isoform may also be affected.^{109, 110} Nevirapine and efavirenz cause secondary induction of CYP3A4, an effect of ritonavir and nelfinavir as well.^{109, 110} Most antimalarial agents are largely metabolized via P-450 enzymes. Mefloquine appears to be no exception, although the details remain poorly understood.¹¹¹ Proguanil and chloroquine metabolism appears to be largely by the CYP2C19 and CYP2D6 isoforms, respectively.112, 113, 114, 115 Chloroquine also undergoes appreciable renal excretion, whereas doxycycline largely avoids these pathways in vivo. A detailed understanding of atovaquone metabolism appears to be lacking. Actual published pharmacokinetic data suggest that (1) there are no significant drug–drug interactions between nelfinavir or indinavir and mefloquine¹¹⁶; (2) ritonavir has minimal effects on mefloquine pharmacokinetics, whereas mefloquine suppresses ritonavir plasma levels¹¹⁷; and (3) atovaquone increases serum zidovudine (AZT) levels by approximately 30%, although AZT has no effect on the pharmacokinetics of atovaquone.¹¹⁸ In summary, the actual pharmacokinetics are not easily predictable from theoretical considerations, and there is a paucity of data. Based on the current data, mefloquine, doxycycline, chloroquine, and malarone (atovaquone + proguanil) are likely to be safe and to retain efficacy for prophylaxis of sensitive strains of malaria.

Among other malaria treatment options, quinidine, quinine, and β-artemether are all predominantly metabolized through CYP3A4 isoforms.¹¹⁰ Large (more than threefold) increases in the area under the curve (AUC) for quinidine are expected for ritonavir.¹¹⁹ As a result, quinidine has been considered to be contraindicated for those on ritonavir,¹¹⁹ and this likely applies to quinine as well. There are no actual data, however. Whether there are clinically relevant effects of these protease inhibitors on the metabolism of artemisinin compounds (dependent at least in part on CYP3A4) remains unknown. Uncomplicated malaria can probably be safely treated with chloroquine (if sensitive), pyrimethamine/sulfadoxine, or mefloquine. Use of quinine, quinidine, or artemisinin compounds remains essential for the parenteral therapy of severe chloroquine-resistant malaria. For those on ritonavir (and/or other protease inhibitors, delaviridine, or efavirenz), the normal loading dose of quinine or quinidine should probably be given, along with some reduction of the maintenance infusion dose. Obviously, careful monitoring needs to be done for the potentially fatal arrhythmic consequences of quinine/quinidine overdosage in this setting. Given the lack of data, however, underdosing may also be a potential problem. The “washout” period for the metabolic effects of ritonavir is thought to be 24 to 48 hours.

A higher incidence of allergic responses to sulfonamides makes pyrimethamine–sulfadoxine (Fansidar) less attractive as a malaria therapy in HIV-seropositive patients, at least in North American populations.¹²⁰ Furthermore, Stevens–Johnson syndrome and related adverse mucocutaneous reactions to the long-acting sulfa compound, sulfadoxine, have contraindicated its benefits for use in malaria prophylaxis in developed countries.¹²¹

The presence of HIV infection alters the predictive value of fever in the empirical diagnosis of malaria. In areas with a high prevalence of both HIV and malaria, the common practice of empirically treating febrile adults for malaria leads to gross overestimation and overtreatment of malaria.¹⁰³ Finally, treatment of severe anemia due to malaria is one of the most common reasons for blood transfusion in sub-Saharan Africa. Malaria thus provides an indirect but very important risk factor for the acquisition of HIV infection by children where the blood supply is not well screened.⁷³

Babesiosis

The genus Babesia contains more than 70 known species of tick-borne intraerythrocytic protozoans that parasitize wild and domestic vertebrates, predominantly in tropical and subtropical areas. Human babesial disease has been reported mostly from temperate climates (see Chapter 91). A significant clinical interaction with HIV infection has been suggested for Babesia microti, raising the possibility that disease with tropical babesial species may be a risk for AIDS patients and providing the rationale for the current discussion. There are five reported cases of babesiosis due to B. microti in HIV-infected people.122, 123, 124, 125, 126 Two cases occurred in splenectomized patients; in one, chronic low-level hemolysis due to Babesia prior to splenectomy was likely. Patients with intact spleens presented with fevers of unknown origin (FUOs) in the face of CD4 counts less than 200/μL. In one, the FUO lasted for months and was associated with night sweats, dry cough, weight loss, and dyspnea on exertion. Persistent parasitemia after clinically successful chemotherapy led to the need for chronic suppressive therapy. In another patient, recurrent disease led to retreatment 8 months after initial therapy. Quinine plus clindamycin and atovaquone plus azithromycin both have therapeutic efficacy in acute disease.¹²⁷ In HIV-infected patients, chronic suppressive therapy appears to be indicated. As with all vector-borne diseases, vector avoidance is the most efficient way to prevent disease. Significant interactions with HIV infection remain to be described for European bovine Babesia species (Babesia bovis and Babesia divergens) and the emerging agents of human babesiosis (WA₁, CA₁, MO₁) in North America.

Leishmaniasis

With the exception of Toxoplasma gondii, Leishmania is the most common tissue protozoan causing OI in patients with AIDS (see Chapter 94). This is not surprising because cellular immune responses (in particular, Th1-mediated immune responses) are critical for protection from Leishmania. The competence of the Th1 axis of cellular response becomes increasingly compromised during the progression of HIV-related immunosuppression, providing a favorable environment for disease due to Leishmania species. Furthermore, in vitro evidence suggests that coinfection of macrophages with HIV and Leishmania can directly upregulate parasite replication.¹²⁸ In vivo, the overall loss of immunological control of parasite infection is reflected by often aberrant manifestations of visceral leishmaniasis (VL) in AIDS, including peripheral parasitemia (found in more than 50% of coinfected people) and parasite dissemination to unusual body compartments.¹²⁹ An AIDS-related OI occurring at low CD4+ T-cell counts, leishmaniasis may be due either to primary Leishmania infection or to the reactivation of clinically latent infection.^{130, 131} Although the published data on the interaction of HIV and Leishmania focus largely on the effects of HIV on leishmanial infection and disease, it should be noted that there is also both in vitro and in vivo evidence that Leishmania can augment HIV replication.132, 133, 134, 135

Although leishmaniasis has a worldwide distribution in the tropics and subtropics, it normally requires an arthropod vector, the sandfly, to move the organism from its sylvatic (zoonotic) cycle to the human host. With certain species of Leishmania (Leishmania tropica and Leishmania donovani) and in some locations (e.g., Syria and India, respectively), an anthroponotic human-to-human cycle via the sandfly can exist. In situations in which intravenous drug use is practiced, transmission is simplified even further by direct person-to-person transfer via contaminated needles and syringes. Generally, however, leishmaniasis is a rural or periurban zoonosis.

The experience with VL complicating HIV/AIDS in Mediterranean countries indicates that many, perhaps most, of the leishmanial infections are acquired with HIV or after HIV infection has already occurred. The transmission of both agents that occurs by sharing of needles and syringes by intravenous drug users could theoretically be reduced by an aggressive program of education and provision of clean needles and syringes. An effective program of sandfly vector control will interrupt transmission from heavily infected human reservoirs to other humans as well as the more usual cycle of infected dogs to humans. Vector control is also the only way to prevent coinfection with Leishmania in those who acquire HIV sexually.

From the relatively high prevalence of latent leishmanial infections, it would appear that reactivation of latent infections could account for the increasing numbers of HIV–Leishmania coinfections; however, this concept is not always supported by epidemiologic evidence. A greater variability in zymodemes (enzyme markers) has been found in parasite isolates from HIV-infected than -uninfected patients. In one series, five isolates were recovered from HIV-infected patients that had previously not been encountered in immunocompetent people with either VL or cutaneous leishmaniasis (CL).¹³⁶ The finding that certain strains of Leishmania typically causing cutaneous disease are being recovered from the bone marrow of coinfected patients could support either the primary or the reactivation hypothesis.¹³⁷

Normally, the age distribution of VL caused by L. donovani includes adults as well as children. In contrast, VL due to Leishmania infantum affects children predominantly, often age 5 years or younger. In Spain, where intravenous drug use accounts for the majority of HIV–Leishmania coinfections, the age distribution of VL has been reversed, with most cases occurring in young adult males.¹³⁸ The fact that 50% of coinfected patients have demonstrable organisms in peripheral blood smears¹³⁹ and the fact that sandflies can readily be infected by feeding on coinfected patients¹⁴⁰ provide evidence for an additional anthroponotic cycle of transmission in this setting.¹³⁷ In summary, although reactivation of latent leishmanial infection is difficult to exclude, increasing evidence—in southern Europe, at least—favors primary infection by certain strains of Leishmania as the main mechanism for coinfection with HIV/AIDS.

With the spread of the HIV pandemic, there is increasing epidemiological overlap of areas in which HIV and leishmania occur, particularly in eastern Africa, India, Brazil, and Europe. Cases have been reported from approximately 40 countries, although the bulk of cases have been reported from southern Europe.129, 130, 131 ^{137, 141, 142} Of note, relatively few cases of American mucocutaneous leishmaniasis have been recognized in HIV-infected subjects.^{143, 144} The propensity for disseminated disease in the presence of HIV appears to be limited to certain species of Leishmania. The bulk of the information on VL complicating HIV infection involves L. infantum in the Mediterranean region. Presumably, the ability to visceralize under the influence of HIV also applies to L. donovani in southern Asia and Africa and to L. chagasi in Latin America; however, documentation for this is still somewhat meager, one of the possible reasons being the poor overlap between geographic distribution of leishmaniasis caused by these species and the distribution, as well as prevalence, of HIV infection. The species of Leishmania that cause CL have been implicated only rarely as OIs in HIV/AIDS. In one instance, L. braziliensis was recovered from the bone marrow of a patient with a CD4+ T-cell count of less than 10/μL,¹⁴⁵ but the main clinical picture in this case, as well as in others,146, 147, 148 including a patient infected with Leishmania major, ¹⁴⁹ has been one of multiple cutaneous lesions resembling diffuse CL.

A febrile illness of longer than 2 weeks' duration in an HIV-infected person with a lifetime history of travel to Leishmania-endemic regions of the world should certainly raise suspicion of leishmaniasis complicating HIV infection. If the patient is an intravenous drug user, travel to southern Europe, especially Spain, France, and Italy, would be particularly pertinent. Clinical diagnosis of VL in leishmania–HIV coinfected people may be difficult. Only 75% of HIV-infected patients, as opposed to 95% of non-HIV-infected patients, exhibit the characteristic clinical pattern, namely fever, splenomegaly, and hepatomegaly.^{130, 131, 137, 142, 150} With increasing immunosuppression, clinically evident ectopic localization of parasites becomes common.¹⁵¹ Gastrointestinal, laryngeal, pulmonary, and peritoneal involvement has been reported.151, 152, 153, 154, 155, 156, 157, 158 Single and multiple cutaneous forms and/or mucosal and mucocutaneous lesions have also been described in AIDS patients worldwide.^{148, 153, 159}

In immunocompetent people, tests for antileishmanial antibodies have been very useful in the diagnosis of VL because B cell activation is prominent, with large amounts of both specific and nonspecific antibody being produced. In contrast, approximately 50% of coinfected patients lack detectable antibody levels.^{130, 131, 150, 160} The situation may be different in instances in which leishmanial infection has preceded HIV infection and the impaired immune response that ensues. Gradoni and associates¹⁶¹ suggested that this type of serologic data could be used as an indicator of the sequence of acquisition of the two infections. Support for this concept is provided by a report from Ethiopia of seven cases of VL with HIV coinfection, all with highly elevated antileishmanial antibody titers.¹⁶² All patients had lived for many years in a leishmaniasis-endemic area of Ethiopia.¹⁶² The recombinant antigen rK-39 appears to be highly sensitive and specific for immunodiagnosis of VL due to L. donovani and L. chagasi in patients without complicating HIV infection; however, the sensitivity of rK-39 for immunodiagnosis of cutaneous cases from Turkey was greatly reduced compared with most cases of VL.¹⁶³ The utility of rK-39-based diagnostics is not clear in HIV-seropositive people. The peripheral parasitemia displayed by many HIV coinfected individuals allows the detection of parasites from the blood in approximately 50% of cases. Cultures and polymerase chain reaction (PCR) of buffy-coat preparations are positive in 70% and up to 100%, respectively.^{130, 131, 164}

There is abundant evidence that successful treatment of leishmanial disease, regardless of the drugs used, ultimately requires intact CMI. The coinfected patient is the victim of a double insult to the immune system. VL is associated with antigen-specific T-cell unresponsiveness¹⁶⁵ and dysfunctional cytokine responses.¹⁶⁶ This situation is further compounded by the immunologic abnormalities associated with HIV infection.

Therapy for VL in the face of HIV coinfection remains controversial, largely due to a lack of firm data. The same drugs used for treatment of VL in normal hosts (including pentavalent antimonials and amphotericin B preparations) have utility in the treatment of HIV coinfected patients, albeit with significantly less efficacy.¹⁵⁰ Amphotericin B is a conventional drug for all forms of leishmaniasis, including visceral disease. Liposomally encapsulated amphotericin has the theoretical advantage of being targeted to monocyte/macrophages, host cells for leishmanial parasites. Between 40% and 65% of coinfected patients have initial parasitological cure after treatment with pentavalent antimonials, amphotericin B deoxycholate, or amphotericin B lipid complex.^{150, 167, 168} Among these options, treatment with lipid formulations of amphotericin B appears to have similar efficacy, but less severe toxicity, than the other drugs. However, the experience with lipid formulations of amphotericin B in coinfected patients remains meager.¹⁶⁸ These lipid formulations are also quite expensive. Trials aimed at optimizing the therapy of VL in AIDS are clearly needed.¹⁶⁸ Even with initial cure, relapse is predictable over time, occurring in up to 80% of coinfected individuals within 1 year.^{150, 167, 169} The optimal drug for secondary prophylaxis remains unclear. Pentamidine given once every 3 or 4 weeks¹⁷⁰ and liposome-encapsulated amphotericin every 2 weeks¹⁷¹ or 3 weeks¹⁶⁹ have been used for secondary prophylaxis.

Miltefosine, an oral agent that is safe and effective for the treatment of Indian patients with VL,¹⁷² has shown some promise in early compassionate-use treatments for VL in HIV-infected subjects.¹⁷³ Further optimization of treatment and suppressive regimens of miltefosine in HIV patients may establish roles for this new antileishmanial agent for therapy and suppressive prophylaxis of VL in HIV-infected patients.

The fact that significant reductions in the incidence of AIDS-related VL have been seen in southern Europe after the advent of HAART,^{174, 175} along with the fact that HAART-related immune reconstitution has allowed the discontinuance of secondary prophylaxis for other OIs, has raised hope that HAART will allow for safe discontinuance of secondary prophylaxis for VL. Details of the levels of immunological and virological responses needed for termination of such secondary prophylaxis remain to be determined.176, 177, 178

American Trypanosomiasis (Chagas' Disease)

American trypanosomiasis, or Chagas' disease (see Chapter 93), is a well-recognized OI in AIDS.¹⁷⁹ The causative organism, Trypanosoma cruzi, and the blood-sucking vector (triatomine) bugs that transmit this protozoan are restricted to the Western Hemisphere but are widely distributed from the United States to Chile and Argentina. Because the HIV-related Chagas' disease reported to date largely represents reactivation of chronic infection during the course of HIV-induced immunosuppression and not primary infection in the face of AIDS (which is not surprising given the differing patterns of epidemiological risk for these infections: largely rural for T. cruzi and largely urban for HIV), this OI can be expected to appear outside these geographic bounds. It should be noted that activation of latent T. cruzi infection, as well as exacerbated primary infection (transmitted by blood transfusion), is also well described in the face of the iatrogenic immunosuppression used for solid organ transplantation and therapy for hematological malignancies.

Available data suggest that clinical T. cruzi reactivation in the face of HIV coinfection occurs largely in those with CD4+ T-cell counts less than 200/μL. Clinically, such reactivation most commonly involves the central nervous system.^{179, 180} Trypanosoma cruzi was probably late in being recognized as an opportunistic pathogen in those with HIV infection because the most prominent features of central nervous system disease are similar to those of toxoplasmic meningoencephalitis. Enlargement of hemorrhagic foci can produce mass effects simulating brain tumors. Lesions are often multiple, with computed tomographic (CT) scans and magnetic resonance imaging (MRI) showing ring enhancement and preferential involvement of the white matter. Toxoplasmic encephalitis may also be present in the same patient.¹⁸¹ The cerebrospinal fluid (CSF) findings include a slight pleocytosis, increased protein, slightly decreased glucose in some patients, and the presence of trypanosomes. Histologically, the brain lesions show necrotic foci with hemorrhage and infiltration of inflammatory cells. Amastigote forms of the parasite are abundant in glial cells and macrophages and only occasionally in neuronal cells. Myocarditis is a common autopsy finding in those dying of AIDS-related T. cruzi meningoencephalitis.¹⁷⁸ Such myocarditis is often clinically silent. Clinical manifestations, when present, involve arrhythmias and congestive heart failure.^{179, 182, 183} Correct diagnosis of reactivated T. cruzi infection depends, first, on considering the possibility based on the geographic origin of the patient and on an appreciation of the clinical picture. If neurologic signs are present, performing a CT scan or MRI is key.¹⁸⁴ The imaging pattern of central nervous system (CNS) T. cruzi infection is indistinguishable from that of toxoplasmic encephalitis. Direct microscopic examination of centrifuged sediment of CSF will often show motile trypanosomes. If fever and other systemic signs are present, direct examination of the buffy coat from the microhematocrit tube may also show motile trypanosomes. Since serum antibodies to T. cruzi indicate previous infection with the parasite, this test is only useful for ruling out reactivated infection if it is negative. If other tests are inconclusive, biopsy of a brain lesion to demonstrate characteristic organisms can be done. PCR on blood or CSF requires research laboratory facilities.

Clinically, differentiating HIV-related reactivation of Chagas' disease reactivation from chronic chagasic disease may be difficult. HIV-related reactivation is associated with high parasitemia, however, whereas the parasitemia of chronic disease is very low.¹⁸⁵ Indeed, even in the absence of overt, clinical reactivation, chronic Chagas' disease is associated with a higher percentage and level of parasitemia in those coinfected with HIV (independent of CD4 count) than in HIV seronegatives.¹⁸⁶ The effects of coinfection appear to be bidirectional. HIV viral load was carefully documented to increase simultaneously with an asymptomatic increase in T. cruzi parasitemia, subsequently returning to baseline in the face of successful antiparasitic treatment.¹⁸⁷ Nifurtimox and benznidazole, both of which have moderate antitrypanosomal activity, are the standard drugs recommended for treatment of Chagas' disease. There simply is not enough experience to evaluate the effectiveness of these drugs in the treatment of T. cruzi infections complicating HIV or AIDS, especially in cases with meningoencephalitis. No information is available on the penetration of these drugs into the CNS, and the survival time of reported cases of coinfection has been short. A patient reported by Nishioka and coworkers¹⁸⁸ survived for 92 days, with disappearance of trypanosomes from the blood and CSF as well as clearance of a brain lesion while being treated with benznidazole at a dose of 8 mg/kg/day for 80 days. Clinical improvement and reduction in size of a brain lesion were attributed to treatment with benznidazole plus, later, itraconazole and fluconazole in another patient with coinfection who survived for at least 6 months.¹⁸⁹ Although there is no other reported experience with the use of itraconazole or fluconazole in the treatment of American trypanosomiasis in humans, itraconazole was reported to be very effective in experimental infections.¹⁹⁰ Infected mice given as little as 15 mg/kg/day were protected against death, and concentrations of itraconazole as low as 0.001 μg/mL inhibited replication of the parasites in macrophages. It has been recommended that treatment of T. cruzi infection in HIV-positive individuals be started early in the reactivation process, when parasitemia is detectable, but before irreversible end-organ damage has occurred.¹⁸⁷ Such a strategy would hinge on serological identification of those at risk, something indicated in all HIV-infected individuals with appreciable risk of T. cruzi infection. Although data are lacking, it should be noted that immunological reconstitution through HAART therapy is likely to provide considerable prophylactic and therapeutic benefit in this disease.

African Trypanosomiasis

No significant interactions between the agents of African trypanosomiasis (see Chapter 92) and HIV have been delineated. Although T cell and macrophage responses are not thought to be important in the protective host response to trypanosomiasis, trypanosomiasis can suppress cellular immune responses, so a biologic interaction between the two is plausible. No significant epidemiologic association between Trypansoma brucei gambiense and HIV has been found.191, 192, 193, 194 Whether HIV alters the clinical course of either West or East African trypanosomiasis is unclear.¹⁹³ There is anecdotal evidence that HIV may complicate the therapy of West African trypanosomiasis, however. Of 18 patients treated with melarsoprol in a rural hospital in the Congo, all 14 HIV-negative patients recovered, whereas 3 of 4 HIV-positive patients died during treatment (likely due to treatment-related encephalopathy) and the fourth failed to respond to therapy.¹⁹⁵

Other Trypanosomatids

In addition to the two genera, Leishmania and Trypanosoma, known to cause disease in humans, the Trypanosomatidae family includes other genera of protozoa that parasitize other vertebrates, insects, and plants. There have been three reports of HIV-infected individuals presenting with symptoms typical of visceral leishmaniasis in which ultrastructural, isoenzyme, and/or kinetoplast DNA analyses of the isolated lesional parasites have indicated that the responsible organism actually belongs to one of these latter genera.¹⁹⁶ The strong implication is that HIV-related immunosuppression can render humans vulnerable to normally nonpathogenic lower trypanosomatids.

Toxoplasmosis

Toxoplasma gondii is a ubiquitous parasite of mammals throughout the world (see Chapter 97). Latent infection lasts for the lifetime of the host. Maintenance of latency is dependent on CMI responses. Reactivation of latent infection is common with increasing immunosuppression in AIDS. The principal manifestation of such reactivation, toxoplasmic encephalitis (TE), is thus a common OI in AIDS patients throughout the world. The incidence of TE is proportional to the prevalence of latent infection in the population at risk of or with AIDS.¹⁹⁷ In the United States, the rate of latent infection varies between 10% and 40%; in Paris, the rate is 90%.¹⁹⁷ Acquisition of Toxoplasma infection is age dependent, but there is wide variation in infection rates even over narrow geographic areas.^{198, 199} Prevalence rates in the tropics vary from 0% to 90%, with most measured communities falling in a broad middle range.200, 201, 202, 203, 204, 205, 206

In the United States, prior to the advent of HAART, one-third of Toxoplasma-seropositive AIDS patients developed TE in the absence of prophylaxis,²⁰⁷ 90% of such cases were in patients with less than 200 CD4+ T cells/μL and 70% in those with less than 100 CD4 T cells/μL.²⁰⁸ The prevalence of TE in AIDS patients in the tropics is unclear, but the burden is thought to be immense and underdiagnosed. Autopsy series that have included examination of the brain have suggested disease prevalence rates in late-stage AIDS patients of 15% in Abidjan, Côte d'Ivoire,²⁰⁹ 25% in Mexico City,²¹⁰ and 36% in Kampala, Uganda.²¹¹

The presumptive diagnosis of TE is based on clinical presentation, positive Toxoplasma serologies, and characteristic neuroradiologic features.²¹² A final clinical diagnosis is made based on the clinical and radiographic response to specific chemotherapy. Less common manifestations of toxoplasmosis in AIDS include pneumonia, retinochoroiditis, myocarditis, orchitis, and gastrointestinal involvement. The reader is referred to one of many excellent reviews on Toxoplasma in AIDS for information on the clinical management of this cosmopolitan OI.76, 77, 78, 79

Five percent of TE occurs not as reactivation but as an acute infection.²⁰⁷ Preventing the transmission of T. gondii to Toxoplasma-seronegative, HIV-infected people has two facets: (1) avoiding the ingestion of tissue cysts of other intermediate mammalian hosts (i.e., cooking meat well) and (2) avoiding the oocysts of the definitive host, the cat. Avoiding cat feces in and around dwellings is probably not sufficient because the oocysts are viable for up to 18 months in moist soil. Contamination of fresh vegetables may be a common method of human infection, and such foodstuffs should probably be washed well or cooked or both.

Primary prophylaxis (TMP–SMX is preferred)²¹³ should be taken by all Toxoplasma-seropositive HIV patients with a CD4+ T-cell count less than 100/μL. It is safe to discontinue both primary and secondary prophylaxis after HAART-related immune reconstitution (sustained CD4+ T-cell counts >200/μL).76, 77, 78, 79

Free-Living Amebae

Free-living amebae of the Acanthamoeba and Balamuthia genera (see Chapter 95) are rare causes of opportunistic encephalitis and cutaneous disease in late-stage AIDS. Most case reports have been from the United States, but the worldwide environmental distribution of these ubiquitous protozoans and the fact that diagnosis is often postmortem suggest that underdiagnosis is widespread in the tropics and elsewhere.

Granulomatous amebic encephalitis (GAE), a subacute to chronic disease of compromised hosts caused by multiple species of Acanthamoeba as well as Balamuthi mandrillaris, generally causes death in weeks to months. Clinical and pathologic data, as well as animal models, suggest that the pathogenesis of GAE involves hematogenous dissemination to the brain from initial upper or lower respiratory (or perhaps cutaneous) sites of infection.²¹⁴ Pathologic changes, in the form of necrotizing granulomatous inflammation, are found predominantly in the posterior neuraxis.

Acanthamoeba and Balamuthia have been isolated from soil, water (including tap water, bottled water, chlorinated pools, and natural sources of fresh- and seawater), and air throughout the world.²¹⁵ The isolation of Acanthamoeba from the nasopharynx of healthy adults indicates that these organisms may be a common constituent of normal flora.²¹⁶ Cellular immunity, along with antibody and complement, appears to be critical to protective immunity to Acanthamoebae. ²¹⁷ Invasive disease occurs in the immunocompromised and debilitated.²¹⁴ Occasionally, encephalitis with Balamuthia mandrillaris has occurred in apparently normal hosts.^{218, 219}

More than 20 cases of GAE have been reported in AIDS patients.^214, 220, 221, 222, 223, 224, 225, 226, 227, 228, 229 Implicated Acanthamoeba organisms include Acanthamoeba castellani, Acanthamoeba culbertsoni, Acanthamoeba polyphaga, Acanthamoeba rhysodes, and Acanthamoeba divionensis. Disseminated cutaneous disease [subacute granulomatous dermatitis (SGD)] has been a feature of many of these cases and has preceded clinical cerebral involvement by weeks or months in some. SGD has been the sole manifestation of invasive disease in some patients.230, 231, 232, 233, 234 Reported CD4+ T-cell counts have been less than 250/μL (median, 24/μL) at the time of presentation. Where CD4 counts have not been reported, the histories reveal clinical evidence of late-stage AIDS.^{214, 234}

GAE in AIDS patients is marked by a more rapid course (with death in 3 to 40 days)²¹⁴ and a paucity of well-formed granulomas in comparison to other hosts with the disease.^{214, 233} Symptomatic involvement of the nasopharynx, paranasal sinuses, or the skin prior to development of GAE is common in AIDS patients.²¹⁴ Cutaneous lesions are usually nodular, with subsequent enlargement, ulceration, and metastatic spread. Such lesions can be quite pleomorphic (pustules, plaques, eschars, and cellulitis), however, and have been confused with cat-scratch disease, cryptococcosis, sporotrichosis, bacillary angiomatosis, mycobacterial infections, and vasculitis.²¹⁴ The most common presentation of cerebral disease is that of fever and headache.^{214, 232} Focal neurologic deficits and profound changes in mental status are also frequent. Neuroradiologic findings mimic those of toxoplasmic encephalitis, with multiple enhancing mass lesions and surrounding edema. CSF findings are quite variable.^{214, 232}

A high index of suspicion and tissue or microbiologic diagnosis is key to the antemortem identification of disseminated Acanthamoeba infection. Wet mounts of CSF are occasionally useful. Both trophozoites and cysts can be found in tissue biopsies. Cysts have been mistaken for the sporangia of Rhinosporium or Prototheca or for cryptococci; trophozoites have been mistaken for tissue macrophages.²¹⁴ Acanthamoebae can be isolated by culture on Escherichia coli–seeded nonnutrient agar or in tissue culture medium.^{214, 232} Identification of species (and even differentiation of Acanthamoeba from Balamuthia) is not possible morphologically. Immunofluorescence techniques can differentiate Acanthamoeba to the group level in tissue section or with cultured organisms. Treatment of disseminated disease due to these organisms is difficult. No chemotherapeutic regimen is clearly efficacious. Agents with possible clinical utility in combination therapy include pentamidine, 5-fluorocytosine, sulfamethazine, sulfadiazine, fluconazole, itraconazole, ketoconazole, macrolides, phenothiazines, and rifampin.^{214, 228, 229, 231} There may be value in testing clinical isolates for drug sensitivities. With isolated cerebral lesions, there may be a role for surgical excision.²²⁹

In a possible foreshadowing of a newly emerging OI, a case of primary amebic meningoencephalitis due to an apparently newly recognized ameba and not associated with thermally polluted water was reported in a patient with late-stage AIDS in Spain.^{235, 236}

Enteric Coccidiosis (Isospora, Cryptosporidium, and Cyclospora)

A trio of coccidian protozoa—Isospora belli, Cryptosporidium spp., and Cyclospora (Eimeria) cayetanensis—are all prominent causes of self-limited, small bowel diarrhea in immunologically normal hosts as well as causes of chronic, severe disease in the face of HIV coinfection. All are cosmopolitan infections. Infection with a fourth organism, Sarcocystis hominis, responsible for both enteric and disseminated coccidioisis in humans, does not appear to have been reported in HIV-infected individuals.

Cryptosporidium spp. (see Chapter 88).

In addition to the most common human pathogen Cryptosporidium hominis (previously Cryptosporidium parvum human genotype, or genotype 1), a variety of zoonotic species also infect humans, including Cryptosporidium parvum (previously bovine genotype, or genotype 2), Cryptosporidium canis, Cryptosporidium felis, Cryptosporidium meleagridis, and Cryptosporidium muris.237, 238, 239 Zoonotic species may cause more severe human disease and may occur more commonly in immunocompromised people. Because of the high prevalence of disease and the lack of effective specific treatment, cryptosporidiosis is a particularly common and severe problem as an OI throughout the world. Chronic infection and disease are most frequent with CD4+ T cell counts less than 180/mL²⁴⁰ and are associated with increased mortality.241, 242, 243 The use of HAART therapy has led to a decreasing prevalence of cryptosporidial disease in HIV-infected individuals.^{244, 245}

Four clinical syndromes of cryptosporidial diarrheal disease in patients with AIDS have been limned: chronic diarrhea (36%), cholera-like disease (33%), transient diarrhea (15%), and relapsing illness (15%). The severe end of the spectrum is seen largely in those with CD4+ T-cell counts less than 180/μL.^{246, 247} Less commonly, extraintestinal sites are secondarily involved, including biliary tract, stomach, pancreas, lung, paranasal sinuses, and middle ear.246, 247, 248, 249 Of these, biliary tract involvement (presenting with right upper quadrant pain, nausea, vomiting, and fever) represents the most common, clinically important site, being found in up to one-fourth of patients with AIDS-related intestinal disease prior to the use of HAART.²⁵⁰ Individuals with CD4+ T-cell counts less than 50/μL are at a particular risk for development of symptomatic biliary disease.²⁵⁰

No antimicrobial agent has demonstrable, consistent efficacy in HIV-related cryptosporidiosis. Immune reconstitution with HAART should be pursued.^{251, 252} If HAART fails or is not available, a variety of antimicrobial agents (including nitazoxanide, azithromycin, paromomycin, and atovaquone) may be tried. Supportive treatment with fluids, nutrition, and antimotility agents plays an obvious therapeutic role.76, 77, 78, 79

Isospora belli (see Chapter 88).

Disease due to Isospora is less cosmopolitan than that due to Cryptosporidia, being most common in tropical and subtropical areas.²⁵³ Isosporiasis usually presents with chronic watery diarrhea and weight loss, with or without vomiting, abdominal pain, and fever.²⁵⁴ Invasion of gallbladder tissue, similar to that described with Cryptosporidium, has been described, along with disseminated involvement of mesenteric and tracheobronchial lymph nodes, in the setting of HIV coinfection.^{254, 255} Prominent tissue eosinophilia of the involved lamina propria is often present.²⁵³ Diagnostic and therapeutic issues are covered in Chapter 88. TMP–SMX provides effective therapy.²⁵⁶ Pyrimethamine (with leucovorin) provides a second option.²⁵⁷ Clinical response is usually rapid, but relapses are common. In the absence of immune reconstitution, suppressive therapy is indicated.76, 77, 78, 79

Cyclospora (Eimeria) cayetanensis (see Chapter 89).

The clinical picture of enteric infection with C. cayetanensis in AIDS appears to be similar to that due to other coccidia.²⁵⁸ It is of interest that biliary tract involvement—as evidenced by right upper quadrant pain, elevated alkaline phosphatase, and thickened gallbladder by ultrasound—has also been described in Cyclospora infection.²⁵⁹ Thus, all three of the enteric coccidia of humans are capable of invading the gallbladder. Diagnostic and therapeutic issues are covered in Chapter 89. As with isosporiasis, cyclosporiasis in AIDS is treatable with TMP/SMX.²⁵⁸ Subsequent suppressive therapy is indicated.76, 77, 78, 79

Microsporidiosis

Microsporidia are intracellular protozoans that, due to HIV and AIDS, have emerged from their relative obscurity as pathogens of insects, fish, and laboratory animals to occupy a new role as important OIs of humans. These cosmopolitan emerging pathogens of the immunosuppressed (including Enterocytozoon bienusi, Enterocytozoon [Septata] intestinalis, Entero-cytozoon cuniculi, Enterocytozoon hellem, as well as pathogens from several other genera) are considered in Chapter 96.

Other Protozoan Infections

Entamoeba histolytica.

This intestinal parasite (see Chapter 86) was initially associated with HIV because of its high prevalence in men who have sex with men (MSM). Despite considerable evidence that immunity in amebiasis requires the participation of CMI, there is no evidence that patients with HIV infection or AIDS are more likely to develop invasive disease.

Giardia lamblia.

As with E. histolytica, a high prevalence of infection with G. lamblia (see Chapter 87) was found in the 1980s among MSM.²⁶⁰ A study of MSM performed at that time revealed no increased prevalence or severity of giardiasis in patients with AIDS.²⁶¹ Since then, no evidence has been found of a significant effect of HIV coinfection. Although some studies have indicated a higher prevalence of giardiasis in HIV seropositives, this has not been a consistent finding. Therapy of giardiasis in people with AIDS is usually successful. Some patients, immunocompromised as well as immunocompetent, are refractory to standard therapeutic regimens for giardiasis. It may well be that such refractoriness to standard therapy is found more commonly in the face of HIV coinfection.²⁶²

Blastocystis hominis.

Controversy continues to exist as to the role of this organism as a cause of diarrheal disease in either immunocompetent patients or HIV-infected people.²⁶³ Blastocystis hominis has a cosmopolitan distribution; there is no association with the tropics.

Balantidium coli.

No information is available as to whether this organism can serve as an OI in HIV-infected people.

Helminthic Infections

Trematodes

There is no evidence that any trematode infection is more severe or difficult to treat in HIV-infected people. More subtle interactions have been explored in schistosomiasis (see Chapter 116). Study of car washers working on the shores of Lake Victoria in Kenya, a population with a high intensity of exposure to S. mansoni and an HIV seroprevalence of approximately 30%, has provided insights into the bidirectional effects of coinfection.²⁶⁴ The CD4+ T cell–dependent granulomatous response to schistosome eggs has been shown to be important in egg migration from venules to the lumen of the intestine in mouse models of disease.²⁶⁴ As might thereby be expected, a significant suppression of egg excretion efficiency, controlled for the degree of infection, was found in S. mansoni–infected patients in the presence of HIV coinfection and low CD4+ T-cell counts.²⁶⁵ Although successful therapy of S. mansoni infection with praziquantel may depend on the host antibody response, praziquantel was efficacious in treating schistosomiasis in this HIV-infected cohort.²⁶⁶ Given that schistosome infection likely preceded HIV infection in these individuals, whether praziquantel will have equal efficacy in individuals infected with HIV first remains an open question. Notably, despite similar responses to therapy, individuals with HIV coinfection and low CD4+ T cell counts showed increased susceptibility to reinfection after therapy,²⁶⁷ a response that appears to correlate with blunted immunological responses to successful drug therapy (and the resultant release of parasite antigen) in such individuals.²⁶⁸ A study of HIV/S. haematobium coinfection in Zambia led to findings that mirror these findings with S. mansoni: (1) Coinfected individuals had lower egg excretion, and (2) praziquantel retained efficacy in the face of HIV coinfection.²⁶⁹ No alteration in resistance to reinfection with S. haemotobium was seen in the face of HIV infection, but CD4+ T-cell counts were not performed in this cohort lacking evident HIV-related disease. It thus remains possible that, as with S. mansoni infection, resistance to reinfection with S. haematobium is decreased with progression of HIV/AIDS.

As for effects of schistosomiasis on HIV, there is evidence suggesting that genital schistosomiasis due to S. haematobium infection is a risk factor for HIV transmission.269, 270, 271 As with a variety of sexually transmitted diseases, the mucosal (vulval, vaginal, and cervical) inflammatory lesions associated with female genital schistosomiasis are likely to compromise the antiviral barrier of the mucosa, provide a cellular milieu that allows for efficient viral transmission and replication, and enhance viral shedding.^{270, 271} Correspondingly, male genital schistosomiasis is also likely to be associated with an increased risk of HIV transmission. In Madagascar, S. haematobium caused inflammation of the prostate and of the seminal vesicles in most adolescent and adult male patients.²⁷² By analogy to bacterial urethritis, such chronic inflammation is likely to be associated with increased viral shedding in the semen in HIV-coinfected individuals.

Schistosomiasis is a prime example of a chronic tropical infection that has been postulated to enhance the pathogenesis of HIV. Schistosoma mansoni infections act as powerful inducers of Th2 polarization in both murine models and humans.60, 61, 62 The type 2 cytokine environment induced by S. mansoni eggs can significantly suppress CD8 T cell–mediated viral clearance in experimental models⁶¹ and has been found to impair antigen-specific Th1 immune responses in both mice and humans.^{61, 66} Type 2–dominant immunologic responses have also been postulated to favor HIV progression due to preferential replication of HIV in Th2 cells,⁶⁴ amplification of activation-induced apoptosis in lymphocytes by type 2 cytokines,^{9, 10} and upregulation of expression of the HIV coreceptors CCR5 and CXCR4 by CD4+ T cells and monocyte/macrophages.^{66, 67} Finally, the antigenic exposure of schistosomiasis may lead to sustained upregulation of HIV replication via the effects of chronic immune activation.⁶⁰ These considerations remain theoretical. Indeed, treatment of S. mansoni infection does not appear to reduce plasma HIV load in coinfected individuals.^{273, 274}

Cestodes

A few unusual manifestations of cestode infection have been reported in AIDS patients. A rapidly expanding, invasive, and ultimately lethal abdominal mass in a patient with a CD4+ T-cell count less than 100/μL was found by ribosomal DNA amplification and sequencing to be due to an as yet uncharacterized cestode.²⁷⁵ Whether this represents merely the fortuitous concurrence of an unusual pathologic finding with dramatic improvements in diagnostic technology (previous rare cases in normal hosts having occurred in the absence of diagnosis) or the recognition of a new disease because its expression is facilitated or dependent on immunosuppression (AIDS patients serving as “sentinel chickens” for the population at large) is unclear. The latter interpretation is favored by a previous similar case report of presumably disseminated cestode infection in the face of immunosuppression due to Hodgkin's disease and its therapy.²⁷⁶

Four cases of exuberant subcutaneous disease due to the larval form of Taenia crassiceps have been reported in AIDS patients.277, 278, 279, 280 As this bests by one the previous number of case reports of human infection with T. crassiseps, the suggestion is that HIV infection is a risk for disease with this cestode. A case of hepatic alveolar echinococcal disease in a 6-year-old child with AIDS has been described.²⁸¹ Uncommon features of this case include the remarkably young age and hence short incubation period for disease and the complete lack of demonstrable parasite-specific humoral or cellular immune responses.

Finally, 10 cases of neurocysticercosis have been reported in HIV-infected patients,282, 283, 284, 285, 286 a number that is sure to increase given the increasing rates of HIV infection in endemic areas. The frequency of giant cysts and racemose forms of disease is remarkably elevated in these reported cases, again perhaps a reflection of the role of CD4+ T cells in tissue immunity to T. solium. Further clinical data on the interaction between HIV and cestode infections are awaited.

Nematodes

Strongyloides stercoralis.

The only nematode implicated as a cause of an OI in the presence of HIV coinfection is the intestinal parasite, Strongyloides stercoralis (see Chapter 111). S. Stercoralis appeared to qualify as an OI because it is one of the few nematodes capable of actually multiplying in, especially immunocompromised, human hosts. Capillaria philippinensis, for which humans are not the natural host, can also multiply internally, but C. philippinensis is rare, quite limited in its geographic distribution, and has not been reported as a coinfection in HIV-seropositive people (see Chapter 106).

One way in which immunosuppression enhances Strongyloides infection is by permitting or stimulating an increased degree of the normal process of autoinfection.²⁸⁷ In this process, first-stage rhabditiform larvae (L1) produced by the adult female worm in the upper small bowel are transformed into infective filariform larvae (L3) that can reinvade the intestinal wall of the colon or the perianal or perineal areas. Massive upregulation of the autoinfective process results in the hyperinfection syndrome, with the development of many more adult worms, and the production of large numbers of larvae that disseminate to all organs. The clinical picture is dominated by gram-negative bacterial sepsis, meningitis, or pneumonia.

Hyperinfection is usually associated with immunosuppression, particularly the administration of corticosteroids. In fact, the list of immunosuppressive diseases, both malignant and benign, associated with hyperinfection is unified by having corticosteroids as a common denominator of treatment,²⁸⁷ but steroids do much more than alter T-cell function, both in the host and in the organism. The defense mechanisms necessary for control of S. stercoralis in humans have not been identified. Even direct effects of steroids on the female worms' reproductive efficiency have been postulated.^{287, 288} The data are meager.

Strongyloides was initially designated as an AIDS OI on its past record of causing hyperinfection in the immunosuppressed.²⁸⁹ Five years later, when it became apparent that hyperinfection syndrome was not being encountered frequently in patients with AIDS, it was removed from the list of OIs indicative of AIDS.²⁹⁰ Given the low but appreciable rate (3.9%) of strongyloidiasis among men attending a venereal disease clinic in New York City in 1981,²⁹¹ the AIDS epidemic in the United States that developed in the 1980s should have provided some clinical evidence of any predisposition of AIDS patients to hyperinfection. This did not occur. The available evidence makes it extremely unlikely that misdiagnosis or underreporting are the relevant factors here; severe strongyloidiasis or hyperinfection syndrome has prominent clinical features and is often fatal if untreated, and it is not likely that the association would escape notice. Few cases of hyperinfection syndrome have been reported in the English-language literature.292, 293, 294, 295, 296, 297, 298, 299, 300, 301 Even among these cases, the presence of hyperinfection is poorly documented in many.

Diagnosis of (as opposed to suspicion of) hyperinfection syndrome depends on the demonstration of markedly increased numbers of filariform larvae in the stool or multiple such larvae in the sputum. The mere presence of filariform larvae in the sputum only indicates the existence of autoinfection. (It should also be noted that the presence of rhabditiform larvae in the sputum points to neither autoinfection nor hyperinfection but to the presence of adult female worms in the lung.) Unfortunately, confusion of gastrointestinal disease with hyperinfection syndrome is embedded in the literature.

It is possible that the frequency of severe strongyloidiasis complicating HIV infection is much higher in certain areas of the tropics where both infections are prevalent and medical facilities are lacking; however, an absence of such an association has been noted from just such areas.¹⁵ Petithory and Derouin³⁰² pointed out that clinical studies of AIDS patients in central Africa, where the prevalence of strongyloidiasis varies from 26% to 48%, did not mention extraintestinal strongyloidiasis. Similarly, a report from Brazil estimated a 1% or 2% prevalence of Strongyloides infection in the population of São Paulo, finding the parasite in 10% of 100 AIDS patients, who showed no evidence of systemic strongyloidiasis.³⁰³ Similar results have been found in Zambia.^{304, 305} A survey of urban adults in Kinshasa (Congo, formerly Zaire) detected S. stercoralis in 20% by intensive fecal examinations of single specimens, and it estimated a 50% infection rate in the same population on the basis of positive serologies. There were no significant differences in infection rates in those seropositive or seronegative for HIV (F. Neva, unpublished observations).

Taken together, these data suggest that the presence and severity of clinical disease due to S. stercoralis are not significantly increased in patients with HIV infection or AIDS alone. A recent study has shed light on the subject.³⁰⁶ Careful quantitation of the numbers and proportions of free-living adult worms and directly developing L3 larvae in stool cultures revealed a surprising negative correlation between CD4+ T-cell count and the proportions of adult worms in individuals infected with HIV. Thus, advancing immunosuppression due to HIV is associated, paradoxically, with suppression of the direct development of L3 larvae.³⁰⁶

More subtle interactions, such as an increased mean gastrointestinal parasite burden or slower response to therapy, may have been missed. Also, some conditions that cosegregate with HIV/AIDS are known to predispose to the hyperinfection syndrome, including the use of steroids (given for pneumocystis pneumonia and lymphoma in AIDS), inanition (seen in patients with chronic diarrhea, untreated oropharyngeal or esophageal candidiasis, and slim disease), and coinfection with human T-cell lymphotropic virus type I (HTLV-I; see Chapter 76). Strongyloidiasis is an important OI in individuals infected with HTLV-1, and Strongyloides infection has been suspected to be a cofactor in the development of acute T-cell leukemia and tropical spastic paraparesis in asymptomatic carriers of HTLV-I.^{307, 308} Notably, intravenous drug use is a risk factor for infection with both HTLV-I and HIV.

Intestinal helminthiasis.

Intestinal helminth infection is ubiquitous in low-income tropical countries. Although such helminths do not appear to act as OIs in AIDS, it has been hypothesized that the immune dysregulation associated with geohelminthic infections may alter the natural history of HIV infection in an unfavorable manner. Such hypotheses are founded on the presence of chronic immune activation and Th2 polarization during chronic helminthic infection. Indeed, it has been demonstrated that peripheral blood cells from patients with intestinal helminth infection⁶⁸ (and filarial infection⁶⁹) are more susceptible to in vitro infection with HIV than are cells from helminth-uninfected patients. The overall hypothesis remains unproven, however. An initial study from Ethiopia indicated that HIV viral load was significantly higher in individuals with various helminthic infections than in individuals without helminths, correlating positively with the parasite load as well as decreasing after elimination of the worms by antiparasitic treatment.⁷⁰ However, similar studies performed in Uganda and examining far larger numbers of patients have convincingly failed to replicate these findings.^{71, 72} These latter studies strongly suggest that helminth coinfection is not associated with faster progression of HIV disease.

Onchocerca volvulus.

Among the filaria, the effect of HIV coinfection has been studied in a large cohort of patients with Onchocerca volvulus infection. No significant epidemiological association was found between the two infections, nor was there any difference in the efficacy of ivermectin treatment in HIV-infected compared with -uninfected patients.³⁰⁹

Arthropods

Sarcoptes scabiei var. hominis stands alone among the arthropod and crustacean infestations of humans as a cause of exacerbated disease in the presence of HIV infection. In normal hosts, scabies is usually manifest as a markedly pruritic, papular, and vesicular dermatitis, with pathognomonic burrows harboring gravid females. Excoriations, nodules, and eczematous or impetiginized plaques may also be found. Relatively few adult mites are normally present.

Norwegian or crusted scabies is seen in neurologically impaired or immunosuppressed patients. Pruritus is often absent or mild. Lesions consist of widespread hyperkeratotic, crusted, scaling, fissured plaques. The nails are frequently involved. Patients tend to be heavily infested, with thousands of adult mites (see Chapter 118). Crusted scabies has been reported as a complication of HIV infection. CD4+ T-cell counts in reported cases have been less than 500/μL.310, 311, 312 Both typical and atypical presentations are seen, the latter including the “pruritus of AIDS,” crusting with pruritus, pruritic papular dermatitis, and mimics of Darier's disease and psoriasis.³¹¹ Secondary sepsis and death have been reported.³¹³ In the face of this clinical variability, the diagnosis of crusted scabies in HIV-seropositive people rests on appropriate clinical suspicion and the demonstration of heavy infestation by microscopic examination of skin scrapings. With such extraordinary mite loads, these patients are remarkably contagious.^{314, 315} Combination therapy with ivermectin 200 μg/kg and topical benzyl benzoate (or perhaps permethrin) appears be the treatment of choice.³¹² Single-dose ivermectin is also effective at preventing transmission in close contacts. Despite speculation early in the AIDS pandemic, there is no evidence of transmission of HIV by arthropod vectors.

Pruritic papular eruptions associated with HIV infection are common in sub-Saharan Africa. The etiology of these intensely pruritic lesions has been attributed to exaggerated immune responses to arthropod bites in HIV-infected individuals.³¹⁶

Fungal Infections

Penicillium marneffei

Disseminated infection with P. marneffei, a dimorphic fungus endemic to Southeast Asia and southern China (see Chapter 82), has emerged as an important OI in AIDS patients. It is the third most common OI in HIV disease in northern Thailand, after extrapulmonary tuberculosis and cryptococcal meningitis.³¹⁶ First isolated in 1956, infection with P. marneffei was a rare event before the arrival of the AIDS pandemic in Southeast Asia.³¹⁶ Since then, thousands of cases have been diagnosed, primarily in southern China, northern Thailand, Hong Kong, Taiwan, Malaysia, Vietnam, Singapore, Indonesia, and Myanmar.316, 317, 318, 319, 320 The overwhelming majority of cases have been in AIDS patients, although normal hosts are also known to develop systemic disease with this fungus.^{316, 318, 321} There is a pronounced intracountry variation in infection rates. In northern Thailand, up to one-fourth of AIDS patients suffer disease with it, whereas in southern Thailand the prevalence is 10-fold less.³²⁰ The environmental reservoir for P. marneffei is unknown, but the organism has been isolated from the organs, feces, and burrows of three species of bamboo rats. The geographic range of these rodents overlaps the previously mentioned known areas of endemicity for disease with P. marneffei 317, 318, 319 and suggests the likelihood that this fungus is also endemic in Laos, Cambodia, and Malaysia.³¹⁶ Whether bamboo rats are important reservoirs for human infection or just another natural host is unclear. There is no evidence of transmission between rats and humans. The seasonal distribution of the diagnosis of disseminated disease in AIDS patients suggests that the reservoir for P. marneffei expands during the rainy season.³¹⁷ Exposure to soil appears to be a key factor.³²²

The pathogenesis of disease due to P. marneffei is presumed by analogy with other endemic systemic mycoses to involve transmission by inhalation, with secondary systemic dissemination. Like Histoplasma capsulatum, P. marneffei is an intracellular parasite of monocyte/macrophages.³¹⁸ A murine model of pulmonary and disseminated infection shows that T cells play a central role in controlling infection.³²³ In AIDS patients, disseminated disease is associated with CD4+ T-cell counts less than 100/μL.^{316, 324}

The largest clinical series reported to date of AIDS patients with disseminated P. marneffei infection provided detailed information on 80 patients.³¹⁶ The onset of symptoms was generally sudden and intense. The most common presenting symptoms and signs were fever (92%), anemia (77%), weight loss (76%), and skin lesions (71%). Other frequent signs and symptoms included cough (49%), generalized lymphadenopathy (58%), hepatomegaly (51%), and diarrhea (31%). The most common cutaneous manifestation (87%) was a generalized papular rash with central umbilication that resembled the lesions of molluscum contagiosum. These were predominantly found on the face, scalp, and upper extremities but occurred throughout the body, including the palate. Other cutaneous lesions included papules without umbilication, a maculopapular rash, subcutaneous nodules, acne-like lesions, and folliculitis. Chest films were frequently abnormal, with diffuse reticulonodular or localized alveolar infiltrates the most common.

The mean duration of illness prior to presentation in this study was 4 weeks. The incubation period for disseminated disease is unclear, as is the percentage of patients whose disease is a result of reactivation of latent infection, as opposed to new infection or reinfection. The fact that reactivation with increasing immunosuppression occurs is supported by the several cases of disseminated disease reported from nonendemic areas in AIDS patients who had a distant history of travel to endemic areas.^{324, 325} Many such patients had spent little time in endemic areas, indicating that infection with P. marneffei can occur rapidly. The development of clinically active disease within weeks of exposure in endemic areas³²⁶ and the reports of children with vertically transmitted HIV infection developing disease in the first months and years of life³²⁷ demonstrate that primary infection can quickly lead to disseminated disease. Finally, the pronounced seasonal variation in disease incidence implies an important role for exogenous reinfection in the expression of disease with P. marneffei in AIDS patients in endemic areas.³¹⁷

The mortality rate of patients with disseminated P. marneffei infection is very high in the absence of prompt treatment. Diagnosis depends on a high index of suspicion, including a careful history to assess possible residence or travel in an endemic area. The differential diagnosis includes tuberculosis, other endemic fungi, and cryptococcosis. Cutaneous lesions may mimic those of AIDS-related molluscum contagiosum, Histoplasma capsulatum and Cryptococcus neoformans. An absence of cutaneous lesions may retard diagnosis. In this regard, a characteristic syndrome of hepatic disease in the absence of skin lesions (fever, hepatomegaly, and markedly elevated serum alkaline phosphatase levels) should be noted.³²⁸ A presumptive diagnosis can be made by the examination of a Wright's-stained bone marrow aspirate, lymph node aspirate, or touch preparations of skin biopsy specimens.^{316, 318, 329} Intracellular and extracellular basophilic elliptic yeast-like organisms with central septation (as opposed to the budding of H. capsulatum) are characteristic. Indirect fluorescent antibody reagents have been developed that may prove useful for differentiating P. marneffei from H. capsulatum and C. neoformans in tissue.³³⁰ Characteristic intracellular organisms have been detected on routine blood smears.³³¹ In the previously discussed series, definitive diagnosis was performed by culture of P. marneffei from blood (76%, even in the absence of routine lysis-centrifugation culture), skin biopsy (90%), bone marrow (100%), and sputum (34%). Diagnostic antigenemia tests that may prove valuable for rapid diagnosis have been developed.^{332, 333} Quantitation of urinary antigen by enzyme immunoassay (employing rabbit hyperimmune IgG) is especially promising: High sensitivity and specificity were demonstrated in an area of high endemnicity.³³⁴ Of note, P. marneffei infection is a known cause of false-positive reactions in the H. capsulatum polysaccharide antigen immunoassay.³³⁵ Current serologic assays are unlikely to be helpful in the diagnosis of AIDS patients but, with improved sensitivity, may provide a useful index of infection.^{330, 336}

Amphotericin B, 0.6 mg/kg/day for 2 weeks, followed by itraconazole, 200 mg twice a day for 10 weeks, is safe and effective.³³⁷ In mild to moderately ill patients, primary therapy with itraconazole may be reasonable. Secondary prophylaxis is mandatory, given relapse rates of 50% within 6 months in its absence.³³⁸ A placebo-controlled, double-blind randomized trial showed that secondary prophylaxis with itraconazole (200 mg once daily) is safe and effective.^{78, 339} With immune reconstitution as a result of a successful response to HAART, discontinuation of secondary prophylaxis is probably safe.³⁴⁰ A controlled, double-blind trial of primary prophylaxis with itraconazole (200 mg once daily) in Thai patients with AIDS and CD4+ T-cell counts less than 200/μL showed that the regimen was well tolerated and effective at preventing both cryptococcosis and penicilliosis.³⁴¹ No survival benefit was found, but the study was not powered to detect a survival advantage.³⁴¹

Paracoccidioides brasiliensis

The dimorphic fungus P. brasiliensis is the cause of the most common systemic mycosis in Latin America (see Chapter 81). Two clinical forms are distinguished in normal hosts: an acute or subacute “juvenile” form and a chronic “adult” form. Acute, juvenile disease, occurring in children and young adults and accounting for a small minority of cases (3% to 5%), is marked by a rapid course, disseminated involvement of monocyte/macrophages and lymphoid tissue, and severe suppression of CMI. Chronic, adult disease, accounting for the vast majority of cases, is a slowly progressive disease, predominantly of older men. In most patients, the primary clinical and pathologic manifestations are pulmonary, with nodular, infiltrative, or cavitary lesions progressing to fibrosis. Other frequent manifestations of adult disease include infiltrative and ulcerative mucosal lesions of the oro- and nasopharynx, polymorphic cutaneous lesions, lymphadenopathy, and adrenal infiltration. Most infections are subclinical. Long latency has clearly been demonstrated, with a mean of 15 years between leaving an endemic area and presentation.³⁴²

It is thought that CMI responses are critical to the host defense from disease with P. brasiliensis. ^{342, 343} Clinical and experimental evidence indicates that paracoccidioidomycosis is associated with marked abnormalities of immune function, with suppression of CMI responses, polyclonal B-cell activation, and elevation of plasma IgE levels.342, 343, 344 These immunologic perturbations are more common and severe in juvenile disease and are reversed with successful therapy.³⁴²

Given the immunology of paracoccidioidomycosis, one might expect it to be a prominent OI in South America and among HIV patients with a history of travel there. In fact, fewer than 100 cases have been reported, despite the presumed wide prevalence of infection or coinfection in areas such as urban Brazil.345, 346, 347, 348 Possible reasons for the low number of cases in HIV-seropositive patients include (1) prophylaxis with TMP– SMX, which has activity against P. brasiliensis; (2) the use of ketoconazole for oropharyngeal candidiasis; (3) misdiagnosis as PCP, with a therapeutic response to TMP–SMX; (4) lack of diagnosis; and (5) the presence of a particularly subtle interaction between HIV and P. brasiliensis. ^{345, 348}

Paracoccidioidomycosis in HIV-seropositive people has been primarily of the “acute” form, with prominent involvement of the reticuloendothelial system. However, pulmonary and oral mucosal involvement, more typical of the “chronic” form, often coexists.³⁴⁶ Although published reports have suggested that this disseminated disease may occur across a broad range of HIV-associated immunosuppression, CD4+ T-cell counts less than 200/μL have been the reported norm.^{346, 347} More than one-third of patients with paracoccidioidomycosis have presented with another opportunistic coinfection, most frequently oral/esophageal candidiasis or tuberculosis.³⁴⁶ Reported clinical presentations span a wide spectrum, from relatively indolent to rapidly progressive disease. Clinical manifestations have included prolonged fever, weight loss, cough, dyspnea, generalized lymphadenopathy, hepatosplenomegaly, skin lesions (localized or disseminated maculopapular, nodular, or ulcerative), oral lesions (ulcerative and/or nodular), osteoarticular lesions, and meningitis.^{345, 346}

Diagnosis in these patients was made by direct examination or culture of clinical specimens, including skin biopsies, lymph node aspirates or biopsies, bone marrow aspirates, CSF, or blood.^{345, 346} Sputum should also be examined using potassium hydroxide preparations, calcofluor stains, or immunofluorescence. The “pilot wheel” cell, consisting of numerous small buds surrounding the mother cell, is characteristic. Serologies have not been diagnostically helpful.

Mortality in the reported cases of disease in HIV-seropositive people was 30%.³⁴⁵ No randomized clinical trials have been performed with any of the drugs commonly used for the treatment of P. brasiliensis infection (sulfonamides, amphotericin B, ketoconazole, and itraconazole), even in normal hosts. Treatment recommendations are based on data from case series and comparison with historical controls.³⁴⁸ However, the data are fairly compelling that itraconazole (100 mg/day) is the drug of choice in normal hosts.³⁴² Published reports of itraconazole treatment in the face of HIV coinfection are scant.^{346, 347} Although amphotericin B and intraconazole may both have therapeutic roles to play, amphotericin B should probably be used for initial treatment in HIV coinfected patients. Lifelong suppressive therapy is necessary; itraconazole seems to be a reasonable choice.

Histoplasma capsulatum var. duboisii

The endemic dimorphic fungus H. capsulatum var. duboisii is localized to western and central Africa and Madagascar (see Chapter 88). In normal hosts, it tends to cause chronic necrotizing cutaneous and skeletal infections. Disseminated disease is unusual. It may be an emerging OI in AIDS patients. No increases in the incidence of African histoplasmosis were reported in a study from the People's Republic of the Congo (now Congo Republic) in the 1980s, despite a rapid increase in the AIDS-related incidence of cryptococcal disease.³⁴⁹ Disease manifestations reported in the handful of HIV–H. capsulatum var. duboisii coinfections described in the literature, however, suggest that AIDS patients are at risk of more severe, disseminated disease.350, 351, 352, 353, 354, 355 Diagnosis is by direct examination of clinical specimens and culture. The yeast form is larger and has a thicker wall than H. capsulatum var. capsulatum. Amphotericin B and itraconazole have therapeutic efficacy.

Sporothrix schenckii

The dimorphic fungus S. schenckii has a worldwide distribution, although most reports have been from tropical and subtropical areas of the Americas (see Chapter 84). The highest incidence of disease is thought to be in the highlands of Mexico and in southern Brazil. Cutaneous and lymphocutaneous disease is most common. Extracutaneous involvement, including osteoarticular disease, pneumonia, and meningitis, has been described in both normal and immunosuppressed hosts. A handful of cases of severe, disseminated sporotrichosis in late-stage AIDS have been described.356, 357, 358 Diffuse cutaneous involvement is the norm. Some patients have also presented with CNS, ocular, osteoarticular, splenic, bone marrow, and/or mucosal involvement. It appears likely that disseminated Sporothrix will become a more prominent OI in heavily endemic areas. The response to therapy (with amphotericin B, potassium iodide, itraconazole, ketoconazole, and 5-fluorocytosine) has been variable and problematic. Amphotericin B should probably be used for initial treatment, followed by lifelong suppressive therapy with itraconazole.³⁵⁹

Other Endemic, Systemic Mycoses

Histoplasma capsulatum, Blastomyces dermatitidis, and Coccidioides immitis are systemic mycoses endemic to the United States that cause OIs in AIDS patients. As such, they are obviously not distinctly tropical diseases and have been covered in-depth elsewhere.76, 77, 78, 79 360, 361, 362 The tropical extent of their respective areas of endemicity deserves brief mention, however.

Histoplasma capsulatum var. capsulatum (see Chapter 80) is found in distinct river basin systems worldwide between 45° N and 30° S of the equator.³⁶⁰ Progressive disseminated histoplasmosis is common in AIDS patients in endemic areas.³⁶⁰ In addition to the most prominent worldwide focus (the Ohio and Mississippi River valleys of the United States), cases have been reported from Central and South America, the Caribbean, Africa, Southeast Asia, and Europe.^{363, 364}

Disease caused by B. dermatitidis was originally named North American blastomycosis. It is now clear, however, that the distribution of this fungus is far more cosmopolitan (see Chapter 80). Blastomycosis has been reported in all the major regions of Africa, with a concentration in southern Africa. It is likely underreported.³⁶⁵ Occasional cases have been reported from Central and South America, the Middle East, and India.³⁶⁶ African strains of B. dermatitidis appear to be antigenically distinct from North American strains. The clinical spectrum likewise appears to be different in African cases, with prominent involvement of bone and chronic draining sinuses. Disseminated blastomycosis is an uncommon, late, frequently fatal OI in patients with AIDS in the United States.^{367, 368} Cases in Africa are to be expected in the future.

Coccidioides immitis is endemic to lower Sonoran life zones in the United States, Mexico, Guatemala, Honduras, Colombia, Venezuela, Bolivia, Paraguay, and Argentina (see Chapter 80). Coccidioidomycosis is a severe, often fatal disease in patients with AIDS and low CD4+ T-cell counts.^{362, 369, 370} Most have presented with diffuse or focal pulmonary disease; extrapulmonary dissemination is not uncommon. In some endemic areas, it is the third most common OI in AIDS patients.^{362, 369, 370}

Cryptococcus neoformans

Cryptococcosis (see Chapter 80) is a common life-threatening fungal infection in AIDS patients.³⁷¹ Although the dissemination of C. neoformans can affect almost any organ system in HIV-infected people, meningitis is the most frequent manifestation. Other relatively common manifestations include pneumonia and cutaneous lesions. Occurring most commonly when CD4 counts fall well below 200/μL, cryptococcosis is a frequent presenting diagnosis in AIDS.³⁷¹ Excellent reviews on cryptococcosis in AIDS are available for detailed information on the clinical approach to this ubiquitous OI.76, 77, 78, 79 ³⁷²

Cryptococcus neoformans is distributed globally. The distribution of cryptococcus as an OI in AIDS is global as well. Regional differences exist in the prevalence of disease as defined by clinical or autopsy series. The prevalence of cryptococcosis in AIDS patients in the United States was estimated to be 7% or 8% in the 1980s.³⁷³ In Thailand, it is the second most common OI (after tuberculosis), with a prevalence of 13% to 44% in different clinical series.^{21, 374, 375} In Africa, the case series prevalence has been variable, from 1% in Soweto, South Africa,³⁷⁶ to 6% to 13% in Kinshasa (the former Zaire).377, 378, 379 The prevalence in autopsy series has similarly varied from 3% in Abidjan (Côte d'Ivoire)²⁰⁹ to 29% in Uganda.²¹¹ Overall, the rates of disease in Africa appear to be higher than those in North America or Europe. Interestingly, a large retrospective case study in London found a significantly higher rate of extrapulmonary cryptococcal disease in Africans attending an HIV clinic than in non-Africans attending the same clinic.³⁸⁰ Data from case series estimated the prevalence of cryptococcosis in Mexico to be 8% to 12%³⁸¹ and in Haiti 13%.³⁸² In Brazil, from 1980 to 2002, 6% of patients had cryptococcus as an AIDS-defining diagnosis.³⁸³

Cryptococcus neoformans exists in two varieties: C. neoformans var. neoformans and C. neoformans var. gatti. They inhabit different ecological niches, with C. neoformans var. neoformans being associated with soil contaminated with bird excrement and C. neoformans var. gatti having a unique, if poorly understood, association with the tree Eucalyptus camaldulensis. ^{384, 385} Whereas C. neoformans var. gatti has a predominantly tropical and subtropical distribution, C. neoformans var. neoformans occurs worldwide.^{384, 385} Of note, although cryptococcosis due to C. neoformans var. gatti occurs with some regularity in normal hosts in regions where this variety is endemic, cases of cryptococcosis in AIDS patients have been almost exclusively due to C. neoformans var. neoformans. ^{385, 386}

Pneumocystis jiroveci (Previously P. carinii f. spp. hominis)

Throughout the world, there is almost universal serologic evidence of exposure by the age of 2 years to P. jiroveci, a ubiquitous fungus³⁸⁷ (see Chapter 85). The prevalence of antibodies to specific P. jiroveci antigens varies, however, suggesting exposure to antigenically different strains in different areas of the world,³⁸⁸ which is mirrored by genetic studies revealing strain differences in this organism.³⁸⁹ PCP (Pneumocystis carinii pneumonia, based on the prior terminology) remains the most frequent serious OI in the United States and Europe, despite dramatic decreases in incidence due to the introduction of HAART.^{390, 391} Prior to HAART, it occurred in 40% to 50% of patients with a CD4 count less than 100/μL per year, and in 60% to 80% of patients overall, in the absence of prophylaxis.

HAART is far from widely available in much of the tropics, and PCP prevalence appears to be high, as expected, among AIDS patients in Central and South America and in Asia.³⁹¹ Interestingly, however, the incidence of PCP in adult AIDS patients in Africa is thought to be far lower than was seen in the pre-HAART era in industrialized countries.³⁹¹ Adult clinical series in Africa have shown prevalence rates of 0% to 22%.^{376, 382,} 392, 393, 394, 395 Studies including bronchoscopy for diagno-sis have described rates of 0% to 39% (the highest figures being obtained as a percentage of acid-fast bacillus-negative pneumonias).^{382, 392,} 396, 397, 398 Autopsy series have had rates of 0% to 11%.^{210, 394, 399}

The reasons for these lower rates of PCP in Africa are unclear. Possible explanations include less environmental exposure to P. jiroveci, exposure to differing strains of P. jiroveci, differences in host susceptibility, earlier deaths in tropical patients with AIDS due to exposure to more virulent organisms, diagnostic difficulties, and host-specific differences in susceptibility.^{382, 400} Exposure to P. jiroveci appears to be similar worldwide.^{387, 388} As noted previously, the existence of genetically and antigenically distinct human strains is likely; however, pediatric PCP rates in AIDS patients in Africa are quite similar to those in the industrial north.^{391, 401} Indeed, approximately one-third of HIV-infected infants in Africa die during the first year of life, and PCP is thought to be responsible for 30% to 50% of such deaths.^{402, 403} Demise from more virulent pathogens prior to clinical PCP may well occur in adults (and PCP does tend to occur early in the course of HIV disease in North American infants,⁴⁰⁴ perhaps with initial exposure⁷⁹). The high prevalence of cryptococcal disease in these same series, which is thought to occur at similar levels of immunosuppression, suggests that this is not the complete answer.³⁸² Diagnostic difficulties may also play a role, but these have been well addressed in several of the cited studies.

The clinical presentation of PCP in the tropics appears to be similar to that in the industrial north.³⁹⁸ Frequent coinfection with tuberculosis may obscure the diagnosis. Multiple reviews of the clinical approach to PCP in AIDS are available.76, 77, 78, 79 ⁴⁰⁵

Other Fungi

Other predominantly tropical fungi, such as the agents of maduromycosis, lobomycosis, rhinosporidiosis, and subcutaneous zygomycosis, may prove to cause opportunistic infection in AIDS patients but have not been reported as such. Isolated case reports of infection due to a variety of unusual fungi in AIDS patients have been published (reviewed in Kaplan and colleagues,²⁰ Vartivarian and associates,⁴⁰⁶ and Perfect and Schell⁴⁰⁷). Some may indeed prove to be OIs, even predominantly tropical OIs, but firm data are lacking. The common occurrence of superficial and invasive infections with Candida and the growing problem of Aspergillus infection in neutropenic long-term survivors of late-stage AIDS are beyond the scope of this chapter.

Mycobacterial Infections

Mycobacterium tuberculosis

Although tuberculosis (see Chapter 36) is not usually perceived as a tropical disease, the prevalence and mortality of the disease in the tropics far surpass those in the industrial countries of the temperate zones. Approximately one-third of the 39.4 million people living with HIV worldwide are coinfected with M. tuberculosis, 70% of whom live in sub-Saharan Africa.² In developing countries, 50% of patients with HIV infection will develop active tuberculosis; in contrast, in the United States, only 4% of patients with AIDS have had tuberculosis.^{2, 408} In some countries in sub-Saharan Africa, more than 70% of tuberculosis patients are HIV-seropositive. Tuberculosis is the leading cause of death among people with HIV infection, accounting for one-third of AIDS deaths worldwide.⁴⁰⁹ The introduction of HAART has decreased death and OIs such as tuberculosis by 60% to 90% among people living with HIV worldwide in affluent countries⁴¹⁰; in developing countries, however, HAART remains available only to a small minority of those who need it.

Tuberculosis was one of the earliest OIs to be linked to HIV infection and in many developing countries is the most common serious OI associated with HIV.⁴⁰⁸ Tuberculosis is a relatively early complication of HIV, occurring before other AIDS-defining illnesses in 50% to 67% of HIV-infected patients.⁴¹¹ Several relevant CMI functions, including lymphocyte proliferation and cytolytic T-cell activity, have been shown to be significantly suppressed in HIV-infected patients with tuberculosis. Additional host responses that may be impaired include the elaboration of cytokines such as IFN-γ and IL-2. CD4+ T-cell counts are suppressed in tuberculosis patients, both with and without HIV coinfection, and rise with therapy for tuberculosis.⁴¹²

Although the consequences of coinfection appear to adversely affect the tuberculous process mainly, there is also evidence, albeit somewhat controversial, of a deleterious effect of tuberculosis on the course of HIV disease. Several studies have shown that in vitro HIV replication is enhanced in blood monocytes from patients with active pulmonary tuberculosis and in lymph node mononuclear cells and CD4+ T cells from HIV-infected, purified protein derivative (PPD) skin test–positive patients after stimulation with PPD.⁴¹³ On the other hand, both enhancement and suppression⁵⁵ of HIV replication in monocytes have been reported after in vitro infection with M. tuberculosis. In vivo, plasma viral load has been shown to be higher in HIV-infected patients with active tuberculosis than in HIV-infected people without active tuberculosis, remaining high throughout the course of treatment.413, 414, 415, 416, 417 However, whether tuberculosis increases HIV load, or whether higher HIV load is really a marker of increased risk for tuberculosis in coinfected patients, remains unclear.⁴⁷ Similarly, although tuberculosis has been associated with reduced survival in HIV-infected patients,^46, 418, 419, 420 direct and indirect lines of evidence have suggested the relation may well not be a directly causal one.^{28, 421, 422}

The prophylaxis, diagnosis, and treatment of tuberculosis in the presence of HIV coinfection have been dealt with in depth elsewhere.76, 77, 78, 79 Certain issues of particular relevance to the tropics are explored further here.

Primary preventive therapy against tuberculosis with isoniazid has been shown be effective in HIV-infected individuals, regardless of tuberculin status.^{423, 424} Meta-analyses have suggested a reduction in tuberculosis incidence of 60% to 68% in those with a positive tuberculin test and a reduction of approximately 42% among all treated individuals.^{423, 424} WHO/UNAIDS recommendations are for primary preventive therapy to be given to PPD-positive, HIV-infected individuals who do not have active tuberculosis.⁴²⁵ In settings where it may not be feasible to do PPD testing, WHO/UNAIDS recommendations are for primary preventive therapy to be considered for those living in populations with a prevalence of tuberculous infection estimated to be more than 30%, health-care workers, household contacts of tuberculosis patients, prisoners, miners, and other groups at high risk of acquisition or transmission of tuberculosis. Although studies have shown that “short-course” therapy with rifampin plus pyrazinamide for 2 months is as efficacious as 6 to 12 months of isoniazid in preventing active tuberculosis in HIV-infected adults with positive tuberculin skin tests,426, 427, 428 the side effects associated with this regimen (in the general population and not, apparently, in the HIV-infected population) have left isoniazid the prophylactic agent of choice.^{429, 430}

Treatment of active, susceptible tuberculosis with first-line drugs is as effective for curing tuberculosis in HIV-infected as in HIV-uninfected individuals. In the absence of HAART, however, the death rate for those under treatment for tuberculosis is higher for people with HIV infection alone, mainly due to other OIs. Conflicting reports on increased rates of tuberculosis recurrence in the face of HIV coinfection⁴³¹ have not provided sufficient evidence for increasing the duration of treatment. Combining HAART with tuberculosis treatment is difficult for several reasons: overlapping toxicity profiles of some antituberculosis and antiretroviral drugs, drug interactions, and nonadherence with complicated treatment regimens.⁴³² An important problem is the possibility of paradoxical reactions. Such reactions include the transient worsening or appearance of new signs, symptoms, or radiographic manifestations of tuberculosis within days to weeks after initiating antiretroviral treatment. These reactions, likely due to immune reconstitution, may be particularly severe when HAART is started soon after initiating treatment for active tuberculosis. The Centers for Disease Control and Prevention (CDC)/American Thoracic Society recommendations for tuberculosis treatment are to (1) continue previously started HAART; (2) avoid initiating HAART and tuberculosis therapy at the same time; and (3) always start tuberculosis therapy first, delaying HAART initiation until the first 1 or 2 months of tuberculosis therapy have been completed.⁴³²

Many of the world's children are vaccinated with bacille Calmette–Guérin (BCG). The risk:benefit ratio of vaccination is surely altered by the presence of HIV infection. On the risk side, there are a few case reports of localized or disseminated disease due to BCG in children and adults.433, 434, 435, 436 Disseminated disease in most hosts is a devastating event, with an overall mortality of 70%, and usually occurs in the immunosuppressed patient. Whether local or disseminated disease was due to HIV coinfection in the cited cases remains unclear, however.433, 434, 435, 436, 437, 438, 439 On the side of potential benefit, the efficacy of BCG in HIV-infected populations is unclear. Data supporting both a lack of benefit of vaccination in HIV-seropositive children and a benefit of childhood vaccination in HIV-seropositive adults in protection from disease due to M. tuberculosis have been published.^{440, 441} Any such benefit would likely be multiplied by the much higher risk of tuberculosis in the face of concurrent HIV infection. Furthermore, data suggesting a beneficial effect of early BCG vaccination on mortality from all causes in HIV-uninfected children suggest that measures of benefit in HIV-seropositive patients need to be broader than mere prevention of tuberculosis.⁴⁴²

BCG vaccination remains in the WHO Expanded Programme on Immunization (EPI).⁴⁴³ WHO recommendations are that BCG be given to children with asymptomatic HIV infection in areas with a high risk of tuberculosis infection. BCG is not recommended for those with symptomatic HIV infection (defined as AIDS). In areas where the risk of tuberculosis is minimal, BCG is not recommended for people known or suspected of being infected with HIV.⁴⁴³ The use of BCG in HIV-seropositive patients is considered to be contraindicated under U.S. Public Health Service–Infectious Disease Society of America (USPHS–IDSA) guidelines²¹³ and by the Advisory Committee on Immunization Practices (ACIP).⁴⁴⁴

Mycobacterium avium

The M. avium complex (MAC) (see Chapter 36) consists of 28 serovars of two Mycobacterium species, Mycobacterium avium and Mycobacterium intracellulare. MAC bacteria are ubiquitous, with organisms commonly being isolated from soils, natural sources of water, tap water, and domestic and wild animals worldwide.^{445, 446} Most MAC isolates from AIDS patients are M. avium; more than 90% are of serovars 1, 4, and 8.^{447, 448} Disseminated disease due to M. avium is the most common systemic bacterial infection in AIDS patients in the industrial north, occurring in up to 43% of AIDS patients in the United States.^{449, 450} Disease occurs almost exclusively in those with CD4+ T-cell counts less than 100/μL, most frequently in those with CD4+ T-cell counts less than 50/μL.⁴⁵⁰ The pathogenesis of disseminated M. avium infection in AIDS is thought to involve primary infection (or reinfection) as opposed to reactivation, with initial colonization of the respiratory or gastrointestinal tracts followed by widespread dissemination.451, 452, 453 Systemic disease is marked by high-grade mycobacteremia (almost exclusively in monocytes) and impressive tissue burdens of bacteria.⁴⁵⁴

The remarkable feature of M. avium in the tropics is the apparent virtual absence of disseminated disease in AIDS patients in many areas, predominantly in sub-Saharan Africa. None of 95 blood cultures from severely ill patients with advanced AIDS in Uganda were positive for M. avium, nor were any of 165 mycobacterial sputum cultures from HIV-seropositive and -seronegative patients at the same hospital found to be positive for M. avium. ^{455, 456} None of 202 blood cultures from HIV-positive adult inpatients in Côte d'Ivoire grew M. avium (whereas 4% grew M. tuberculosis).⁴⁵⁷ None of more than 200 diagnostic lymph node biopsies in HIV-seropositive African patients had histology characteristic of disseminated M. avium infection.²¹¹ Intestinal biopsies from 98 Ugandan, Zairian, and Zambian patients with chronic HIV-related enteropathy yielded histology suggestive of M. avium infection in only 1 patient.^{304, 458, 459} Autopsies on 78 HIV-seropositive children in Côte d'Ivoire revealed no evidence of M. avium infection, whereas autopsies on 247 adult HIV patients in Côte d'Ivoire revealed a 3% prevalence of pathologic changes “indicative of atypical mycobacteriosis.”²⁰⁹ In contrast, 3 of 48 (6%) patients hospitalized in Kenya with late-stage HIV disease had M. avium bacteremia.⁴⁶⁰ Clinical and autopsy series from Mexico have revealed a prevalence of disseminated disease due to M. avium of 4% to 6%,^{210, 461} whereas 18% of 125 hospitalized patients with AIDS in Brazil had M. avium cultured from bone marrow.⁴⁶² Few data are available from India and Southeast Asia.

The reasons for the apparent absence of disseminated disease due to M. avium in areas of the tropics are unclear. As with the decreased prevalence of PCP, many explanations have been proposed, including less exposure to M. avium, exposure to different (less pathogenic) variants of M. avium, differences in host susceptibility, greater acquired immunity to mycobacteria, earlier death by more virulent pathogens, and diagnostic difficulties. Overall exposure to MAC organisms is likely to be similar. Environmental isolation of MAC occurs with similar or greater frequency in Congo and Uganda than in the United States,^{446, 456} and skin test surveys suggest a similar frequency of exposure to MAC in economically developed and developing countries.⁴⁶³ Piped water systems in the United States and Europe have a higher frequency of MAC isolation, however, and economic conditions may lead to greater exposure to MAC-containing droplets via showerheads in economically developed countries^{445, 446, 464}; such differences, however, are unlikely to lead to an essentially total absence of disease in countries such as Uganda. Exposure to different M. avium serovars or strains may well be important. Data on serotyping of African clinical strains are scarce. Preliminary data suggest that African clinical isolates are distinguishable from European and American isolates by restriction fragment length polymorphism analysis.⁴⁶⁵ The possibility of underlying genetic differences in host susceptibility is belied by the similar rates of disseminated M. avium infection as a presenting diagnosis in African and non-African AIDS patients in a London clinic.³⁸⁰ Greater acquired immunity to mycobacterial disease through BCG vaccination⁴⁴¹ or prior infection with M. tuberculosis may exist, but the reported BCG coverage (50%) and PPD reactivity (82%) in Uganda seem unlikely to explain the lack of any disseminated MAC disease.⁴⁵⁶ Earlier death due to a greater environmental presence of, or greater latent infection with, more virulent pathogens may occur. This is unlikely to be the entire explanation because patients in many of the previously mentioned studies had clinical late-stage AIDS. Finally, the design of several of the previous studies makes the assertion that the central problem is one of a lack of diagnostic sophistication untenable. Further data are awaited.

Mycobacterium leprae

Mycobacterium leprae, the causative agent of leprosy (see Chapter 38), is an incredibly slow-growing parasite of monocyte/macrophages. For comparative purposes, it may be useful to recall that Leishmania, which infects similar cells, is a prominent opportunistic pathogen for patients with HIV infection. The importance of CMI in leprosy and leishmaniasis was emphasized by Turk and Bryceson⁴⁶⁶ in their detailed comparison of skin lesions and histopathology of both diseases. Moreover, the immunopathology of both infections is very similar. In view of the widespread prevalence of leprosy in the tropics and subtropics, the immunosuppressive effects of HIV or AIDS on leprosy would be expected to become readily apparent, but there is little or no evidence of this interaction. In fact, in a comprehensive analysis of the possible interaction between HIV/AIDS and leprosy, Lucas concluded that leprosy appears to be another “missing infection in AIDS.”¹⁵

There is no quantitative measure of immune unresponsiveness in leprosy, such as the CD4 count in AIDS, as an indicator of clinical progression; however, there is general agreement that the tuberculoid form of disease is characterized by a well-organized granuloma on biopsy of a skin lesion or the lepromin reaction, with very few or no organisms (paucibacillary). Patients with lepromatous leprosy have a negative lepromin skin test, and biopsies of their skin lesions lack a granulomatous response and show large numbers of organisms (multibacillary). Some of the ways in which the adverse effects of HIV infection on leprosy could be manifested include an increase in the disease:infection ratio, a shift toward multibacillary disease, or an altered response to antileprosy chemotherapy. Of course, the presence of leprosy may also enhance HIV infection.

Several studies have examined positive serology for HIV in newly diagnosed leprosy cases. One report from a rural hospital in Zambia found a higher prevalence of reactors compared with blood donors and surgical patients (6 of 18 vs. 9 of 105), but the numbers were small and the controls were not adequately matched.⁴⁶⁷ A larger study of HIV seroprevalence in northwest Tanzania of 93 new leprosy cases compared with more than 4000 controls found that the presence of HIV antibody was significantly associated with multibacillary disease.⁴⁶⁸ The fact that this association was based on only five HIV-positive cases with multibacillary disease illustrates the complexity of epidemiologic analysis in a disease such as leprosy. Another comparison of seropositivity for HIV was carried out among 189 new cases of leprosy matched for age, sex, and district of residence with 481 controls in Uganda. No significant difference in overall positive rates was found (12% in cases vs. 18% in controls), but again, positive HIV reactions were more frequent among multibacillary cases.⁴⁶⁹ A different clinical association was noted in Zambia in leprosy patients with active neuritis, which suggested that HIV- positive cases had poorer recovery of nerve function than controls after treatment with steroids.⁴⁷⁰

A factor that should be considered in evaluating reports of HIV/AIDS in leprosy patients is the greater likelihood of false-positive serologic reactions. One or more positive bands to HIV antigens in Western blots were commonly found in several hundred sera from northern India in the absence of positive enzyme-linked immunosorbent assays (ELISAs).⁴⁷¹ Another report claimed that 3 of 75 (4%) sera from Indonesia and 6 of 100 (6%) sera from Somalia gave positive HIV ELISAs but negative Western blots.⁴⁷² These were attributed to leprosy.

There appears to be no striking evidence that HIV infection has an adverse effect on the course of leprosy. There is a suggestion from several of the reports cited previously that multibacillary disease may develop under the influence of HIV infection; however, there are several features of leprosy that may tend to obscure an interaction with HIV infection. Both infections are chronic and slow in their progression, so it may simply take more time to recognize an influence of one on the other. Leprosy is predominantly a rural infection involving people not yet caught up with the ravages of HIV and AIDS. Patients with AIDS in the tropics may not survive long enough to display interactions with leprosy. Finally, patients in the early stages of leprosy have relatively subtle clinical manifestations with which physicians in the urban environment may not be familiar. Therefore, there may be more going on out there than we realize. It has been noted that HAART-associated immune reconstitution may trigger injurious inflammatory reactions in treated patients coinfected with leprosy.473, 474, 475

Other Nontuberculous Mycobacteria

Disease due to Mycobacterium genavense (see Chapter 36) mimics that due to M. avium, causing disseminated disease in AIDS patients with very low CD4+ T-cell counts (mean, <50/μL). The pathogenesis appears to involve initial gastrointestinal colonization followed by dissemination.^{476, 477} The important environmental reservoirs are unclear, but pet birds can have extensive gastrointestinal tract involvement. The organism may also be present in tap water.⁴⁷⁸ The geographic range of disease is only beginning to be defined. Cases have been reported from North America, Europe, and Australia.⁴⁷⁶ The most common presenting symptoms and signs are fever, weight loss, abdominal pain, chronic diarrhea, lymphadenopathy, hepatosplenomegaly, “pseudo-Whipple's disease,” and anemia.^{476, 479, 480} Imaging of the spleen may suggest splenic abscesses⁴⁸¹; diffuse nodular infiltrates may be seen in the lung.⁴⁸² Pathologically, involved organs are filled with histiocytes that are packed with acid-fast bacilli. The diagnosis can be established by the isolation of M. genavense from normally sterile sites (blood, bone marrow, lymph node, and spleen). Specific diagnosis is confounded by the fastidious growth requirements of the organism. Primary isolation on solid media is difficult, and growth in liquid broth may have only 50% sensitivity. Definitive identification demands PCR techniques.⁴⁷⁶ The clear implication is that a significant percentage of cases ascribed to disseminated M. avium are likely due to M. genavense. ⁴⁷⁹ Data on treatment are all retrospective, but therapy appears to be associated both with improvement in symptoms and with survival. Multidrug regimens that include clarithromycin appear to be associated with the best clinical responses.^{483, 484}

Before the AIDS pandemic, Mycobacterium kansasii was known primarily as a cause of chronic pulmonary disease, resembling tuberculosis in lungs with underlying damage. Mycobacterium kansasii is second to MAC among nontuberculous mycobacteria as a cause of disease in HIV-infected patients in the United States.⁴⁸⁵ Although M. kansasii has been reported as a cause of pulmonary disease in most areas of the world, most case reports of HIV coinfection are from North America and Europe. However, coinfection may be especially prevalent in the gold mines of the Transvaal in South Africa.⁴⁸⁶ In HIV-infected individuals, disease due to M. kansasii and M. tuberculosis has very similar clinical and radiological characteristics.⁴⁸⁷ A major difference (with epidemiological and prognostic implications) is that M. kansasii disease tends to occur later in the course of HIV infection. The mean CD4+ T-cell count at the time of presentation is approximately 50 to 60/μL487, 488, 489; 60% to 90% present with pulmonary disease alone and 20% to 35% with disseminated disease.^{489, 490} The incidence in industrialized countries has plummeted in the HAART era.⁴⁹¹ With all nontuberculous mycobacteria, differentiation between colonization, contamination, and disease can be problematic. Mere colonization of the respiratory tract with M. kansasii appears to be infrequent in AIDS patients, however. All pulmonary isolates should be taken seriously.^{488, 490} Despite relative in vitro resistance to isoniazid,⁴⁹¹ the recommended therapy is isoniazid, rifampin, and ethambutol. Therapy clearly alters survival in patients with pulmonary disease.^{489, 490, 492} Disseminated disease has a particularly poor prognosis.

First described in 1977, Mycobacterium malmoense is an uncommon cause of pulmonary disease resembling tuberculosis. The environmental reservoir is unknown. Person-to-person transmission has never been documented. Multisystem disease with bacteremia has rarely occurred in the presence of profound immunosuppression, including several patients with AIDS and low CD4+ T-cell counts. Pulmonary and gastrointestinal disease, along with bacteremia, is usual.^{493, 494} In vitro susceptibility testing does not correlate well with clinical response.⁴⁹⁵ The best regimen for pulmonary disease in the nonimmunosuppressed patient appears to be isoniazid, rifampin, and ethambutol.⁴⁹⁵ Optimal therapy in AIDS patients is unclear.

Mycobacterium haemophilum causes localized lymphadenitis in immunologically healthy children and cutaneous, osteoarticular, and, more rarely, pulmonary or disseminated disease in immunocompromised patients. Several cases have been reported in AIDS patients.496, 497, 498, 499, 500 Cutaneous lesions include furuncles, abscesses, papules, vesicles, and deep ulcers. Such lesions are usually diffuse, most often on the extremities. Culture (at 30° to 32°C) demands supplementation of media with an iron source. In vitro susceptibility data and scattered clinical reports suggest that rifampin plus ciprofloxacin is reasonable empirical therapy. Other agents with good activity include amikacin, ciprofloxacin, and clarithromycin.⁴⁹⁸ The environmental source and mode of infection are unclear.

Several other mycobacteria have been demonstrated or suspected to cause opportunistic disease in AIDS patients, including Mycobacterium fortuitum (primary pulmonary disease,⁵⁰¹ disseminated disease,⁵⁰² cervical lymphadenitis,⁵⁰² and meningitis⁵⁰³), Mycobacterium marinum (disseminated cutaneous and systemically disseminated disease^{504, 505}), Mycobacterium celatum (pulmonary and disseminated disease^{506, 507}), Mycobacterium xenopi (disseminated disease, pulmonary disease, and pulmonary colonization508, 509, 510), Mycobacterium gordonae (pulmonary, cutaneous, and disseminated disease511, 512, 513), Mycobacterium scrofulaceum (disseminated disease⁵¹⁴), Mycobacterium bovis (disseminated disease), and Mycobacterium simiae (disseminated disease^{515, 516}). Although a smattering of case reports have suggested that HIV does not exacerbate disease due to M. ulcerans, the causative agent of Buruli ulcer,⁵¹⁷ a case report has called this into question.⁵¹⁸

Spirochetal Infections

Along with other genital inflammatory or ulcerative diseases, syphilis (see Chapter 44) has been implicated as a cofactor in HIV transmission. Many case reports have suggested that HIV infection can alter the course of disease with Treponema pallidum. In the presence of concurrent HIV infection, syphilis has been thought to (1) progress more frequently and rapidly to neurosyphilis,^{519, 520} (2) lead to an increased incidence of meningitic manifestations of neurosyphilis,⁵²¹ (3) lead to an increased frequency of “malignant secondary syphilis” with ulcerating lesions and prominent systemic symptoms,⁵²² and (4) be less amenable to successful therapy with standard regimens as assessed by clinical or serologic measures (including a lack of appropriate nontreponemal titer reduction or a serologic relapse).523, 524, 525, 526, 527 Such concerns, based largely on case reports and retrospective studies, were amplified by the disconcerting finding of T. pallidum invasion of the CNS in early syphilis in HIV-infected patients. Such early invasion of the CNS occurs equally frequently in HIV-seropositive and -seronegative people, however,⁵²⁸ and the “atypical” courses of syphilis described previously were well-known in the pre-AIDS era. Knowledge of the actual frequency and relative significance of such events has awaited well-designed prospective studies. Three studies now provide evidence that the clinical presentation and clinical and serologic responses to treatment of syphilis may not be appreciably altered by HIV coinfection.529, 530, 531 A major caveat of these studies is that the mean level of immunosuppression in the patients in these studies, as assessed by CD4+ T-cell counts, was not severe. Furthermore, the number of patients involved was relatively small. Thus, the clinical course of syphilis in the face of severe HIV-induced immunosuppression may in fact be exacerbated, and the response to conventional therapy may lead to the infrequent occurrence of serious adverse treatment outcomes.⁵³¹ The clinical approach to the HIV patient infected with this cosmopolitan sexually transmitted disease (STD) has been thoroughly discussed elsewhere76, 77, 78, 79 ^{532, 533} and will likely continue to evolve.

Whether HIV infection has a deleterious effect on the course of the nonvenereal, endemic treponematoses (see Chapter 44) is unknown, but such an effect has been postulated by analogy with syphilis.⁵³⁴ No effects of HIV on concurrent infection with the Borrelia species that cause relapsing fever have been reported (see Chapter 45). Whether Lyme disease follows an unusual course in HIV-infected people is unclear.^{535, 536} Finally, initial observations suggest that leptospirosis (see Chapter 46) runs a similar course in patients coinfected with HIV.^{537, 538}

Rickettsial and Ehrlichial Infections

Among the rickettsiae and related organisms, only Coxiella burnetii and Ehrlichia organisms have been suspected of being exacerbated by concurrent HIV infection. Notably, these pathogens are obligate intracellular parasites of monocyte/macrophages.

Q fever (see Chapter 54) has a worldwide distribution. The responsible pathogen, C. burnetii, lives and multiplies in the phagolysosomes of monocyte/macrophages. As with other such parasites, host defense against infection appears to depend on specific T-cell activation of the microbicidal effector functions of infected cells.⁵³⁹ Radiation, cyclophosphamide, corticosteroids, and pregnancy have led to reactivation of disease in animal models.⁵³⁹ Case series have suggested that patients with immunocompromise due to a variety of causes (including leukemia, Hodgkin's disease, bone marrow and renal transplantation, and alcoholism) are more susceptible to both symptomatic acute and relapsing or chronic disease with C. burnetii. 540, 541, 542 The rationale for expecting more frequent or serious disease in the HIV-infected patient is clear. The data are less so. A study from southern France demonstrated a threefold higher prevalence of antibodies to C. burnetii in HIV-seropositive people.⁵⁴³ This suggestion of an increased rate of transmission in HIV-infected people has not been found in other seroprevalence studies from Paris, Spain, or the Central African Republic.544, 545, 546 Given the differential prevalence of risk factors for the acquisition of HIV infection between the studies, the contradictory data strongly suggest that C. burnetii can be blood-borne and that intravenous drug use is a risk for its transmission. Two studies from southern Europe have further suggested that HIV infection leads to a higher disease-infection ratio with C. burnetii.^{543, 547} A retrospective serologic study of 520 patients with acute Q fever from an area of Spain with a high incidence of both HIV and C. burnetii infection revealed no overrepresentation of HIV-infected people, however.⁵⁴⁶ The clinical features of Q fever do not appear to vary between HIV-infected and -uninfected hosts.^543, 545, 546, 547 However, definitive statements await prospective studies in severely immunosuppressed AIDS patients.

Human monocytic ehrlichiosis (HME), caused by the tick-borne Rickettsia-like agent E. chaffeensis, is an acute febrile illness associated with leukopenia, thrombocytopenia, and hepatic enzyme abnormalities. Most case reports of infection with E. chaffeensis (see Chapter 53) have been from the United States. A report of infection in Mali supports a much wider distribution of disease, however.⁵⁴⁸ HME appears to be an AIDS-related OI.⁵⁴⁹ Reported hospitalized cases have had a high rate of complications and a mortality of approximately 30%⁵⁴⁹ (compared with an estimated case fatality rate for HME in the absence of HIV of <3%⁵⁵⁰). Patients with fatal disease had CD4+ T-cell counts less than 200/μL; in patients with less than 100/μL, the mortality rate was more than 50%. Of eight reported cases of disease caused by E. ewingii, a related tick-borne agent, seven occurred in patients with immune deficiencies, including four with HIV infection.^{549, 550} The suspicion is thus strong that E. ewingii is an opportunistic pathogen in the setting of HIV infection.⁵⁵¹ No cases of infection with the tick-borne agent of human granulocytic ehrlichiosis⁵⁵² in the face of HIV infection appear to have been reported.

There is no evidence implicating any of the spotted fever or typhus group of rickettsiae as having a clinically significant interaction with HIV. Although neither the mortality nor the morbidity of Rocky Mountain spotted fever are changed by HIV coinfection, this infection should not be overlooked in favor of defined OIs in the differential diagnosis of febrile illness in HIV-seropositive patients. A prospective study on scrub typhus [due to Orientia (formerly Rickettsia) tsutsugamushi] revealed no increase in clinical severity at time of presentation in HIV-infected patients with a median CD4+ T-cell count of 70/μL.⁵⁵³ Interestingly, rickettsemia occurred significantly less often in the HIV-seropositive patients. Neither the relative prevalence of infection nor the response to treatment was addressed in this study.

Bacterial Infections

Brucella

A cause of systemic disease worldwide, Brucella species are facultative intracellular parasites that infect and multiply in macrophages (see Chapter 41). CMI responses, particularly the activation of monocyte/macrophages by antigen-specific T cells, are important in host resistance. Despite this, the meager published data on coinfection do not support a significant effect of HIV on infection with Brucella and, in fact, prior to the AIDS pandemic, only two cases of brucellosis in immunocompromised hosts (hairy cell leukemia and IgM deficiency) had been reported.^{554, 555} A retrospective seroprevalence study found no significant association between Brucella serology and HIV serology in a cohort of female sex workers in Kenya.⁵⁵⁶ The prevalence of antibodies to each pathogen was high (HIV, 65%; Brucella, 35%). The clinical course of brucellosis in the 18 reported cases with concurrent HIV infection was not outside the spectrum of disease seen in normal hosts.556, 557, 558, 559, 560 Definitive data on the interaction between these two pathogens await careful prospective studies.

Burkholderia pseudomallei

Melioidosis (see Chapter 34) does not appear to behave as an AIDS-related OI. The disease is endemic in Southeast Asia, particularly in northern Thailand, where the prevalence of AIDS is high. Only one case of fatal, recrudescent, bacteremic disease in an HIV-seropositive person has been reported, however.⁵⁶¹ Clinical series from Thailand are silent with regard to the presence of melioidosis in AIDS patients,^{374, 375,} 561, 562, 563 and a 10-year study of bloodstream infections in a hospital in northern Thailand reported a similar proportion of B. pseudomallei isolates in HIV-infected and -uninfected patients.⁵⁶⁴

Enteric Bacteria

Several enteric bacterial infections have been reported to cause disease of greater severity, invasiveness, chronicity, or recurrence in the presence of HIV coinfection. Enterotoxigenic E. coli has not been described as causing more severe disease in HIV-seropositive patients, but Shigella species, Salmonellae, Campylobacter, and Listeria monocytogenes have all been implicated as causes of more severe or relapsing disease in the presence of HIV. Data from the tropics on enteric bacterial pathogens are scant. Studies of slim disease (enteropathic AIDS) have not revealed an enteric bacterial cause in most cases.^{304, 458, 459} The prevalence of certain enteric pathogens such as Campylobacter, vibrios, and enteropathogenic E. coli in the tropics has not been accurately assessed because their detection requires the use of special media and experienced laboratory personnel. In the case of less fastidious organisms that are easier to detect, such as Salmonella spp., the phenomenon of bacteremia with nontyphoid organisms has been noted in tropical Africa.²⁹ Since bacteremia with Shigella spp. probably occurs more commonly in patients with HIV disease,⁵⁶⁵ this association may be expected to occur in the tropics as well. There is no reported evidence to suggest that cholera is altered in the presence of HIV, although the gastric secretory failure that occurs commonly in AIDS may lead to a greater susceptibility to infection to Vibrio cholerae. ^{566, 567} It should be noted that although the live oral cholera vaccine is considered to be contraindicated in people with HIV infection by the USPHS–IDSA working group,²¹³ it has been shown to be safe and immunogenic in HIV-infected adults in Mali.⁵⁶⁸ Further details on the association of HIV and enteric bacterial infections, including therapy, are presented in Chapters 16 through 21Chapter 16Chapter 17Chapter 18Chapter 19Chapter 20Chapter 21 and in reviews elsewhere.^{76, 77}

Other Bacteria

Although the data on HIV infection and epidemic meningococcal meningitis have not provided evidence for a significant interaction,⁵⁶⁹ studies suggest that HIV infection may be a risk factor for sporadic meningococcal disease.^{570, 571} Interactions between HIV and Bacillus anthracis or Yersinia pestis have not been reported. The globally endemic Bartonella species, Bartonella henselae and Bartonella quintana (see Chapter 40), cause acute and persistent bacteremia as well as localized tissue infection (including bacillary angiomatosis, bacillary peliosis, microscopic abscess formation, and lymphadenitis), primarily in AIDS patients and other immunocompromised people.⁵⁷² The closely related species Bartonella bacilliformis, which is geographically restricted to Andean river valleys, causes a similar spectrum of disease (including acute and persistent bacteremia and hemangiomatous nodules resembling those seen in bacillary angiomatosis) in immunologically normal hosts (see Chapter 40). Cases of coinfection with HIV and B. bacilliformis do not appear to have been reported.

Viral Infections

Hemorrhagic Fever Viruses, Arboviruses, and Others

None of the viruses that, along with parasites, have formed the traditional focus of the Anglo-American specialty of tropical medicine have been reported to cause uniquely prevalent, severe, or unusual disease in people infected with HIV. No significant interactions have been well documented between HIV and bunyaviruses, hantaviruses, phleboviruses, arenaviruses, alphaviruses, or filoviruses. In part, of course, this may be a function of a lack of sufficient experience with coinfection with these agents. Among the flaviviruses, (1) two uncontrolled series of patients with St. Louis encephalitis in Texas have suggested the possibility that the ratio of disease to infection, but not the course of symptomatic disease, is worsened in the presence of HIV infection,^{573, 574} and (2) there are insufficient data to determine whether HIV alters the course of yellow fever (no case reports; 20% to 50% mortality in the absence of coinfection⁵⁷⁵), West Nile virus infection (single case report⁵⁷⁶), or dengue infection (single case report⁵⁷⁷). Interestingly, the case report of dengue coinfection suggests that dengue fever, like acute measles and scrub typhus, may lead to a reversible suppression of HIV replication.⁵⁷⁷

Should the live-attenuated yellow fever vaccine be given to those infected with HIV? There are theoretical risks of vaccine-induced encephalitis and/or hepatic damage due to prolonged viremia in the immunodeficient.⁵⁷⁵ A handful of cases of postvaccinial encephalitis or multiple organ failure (yellow fever vaccine-associated viscerotropic disease [YEL-AVD]) have been reported in presumably immunocompetent patients (against a denominator of approximately 400 million people vaccinated).⁵⁷⁵ Of note, 4 of 23 vaccinees who developed YEL-AVD had undergone thymectomy for thymomas, raising the concern that deficient thymic function may permit fatal vaccine-induced viral infections.⁵⁷⁸ A single case has also been reported of fatal myeloencephalitis after vaccination in a Thai man with asymptomatic HIV infection, albeit a low CD4+ T-cell count and a high viral load.⁵⁷⁹ Approximately 100 asymptomatic HIV-seropositive U.S. military personnel received yellow fever vaccination prior to the introduction of routine HIV screening; no adverse effects were detected (R. Redfield, personal communication, 1997). Small published series of travelers have suggested safety and variable efficacy of the 17D yellow fever vaccine in HIV seropositives without severe immunosuppression.580, 581, 582 The immunogenicity of yellow fever vaccination has been noted to be severely reduced, again in the absence of significant adverse events, in HIV-infected children in Côte d'Ivoire⁵⁸³ (T. Tsai, personal communication, 1997). WHO recommendations are to use yellow fever vaccine in HIV-seropositive patients who are asymptomatic. It remains a part of the WHO EPI.⁴⁴³ Pending further studies, yellow fever vaccine is not recommended for symptomatic HIV-infected patients by WHO.⁴⁴³ The ACIP recommends that HIV-infected people without AIDS or other symptomatic manifestations of HIV infection, who have laboratory-established verification of adequate immune function, and who cannot avoid potential exposure to yellow fever be offered the choice of vaccination.⁵⁸⁴ Given apparent reduced vaccination efficiency, neutralizing antibody titers should probably be measured prior to travel. If travel requirements (as opposed to actual risk of infection) are the only reason for vaccination of an asymptomatic HIV-infected person, a vaccination waiver letter (which may not be accepted at some borders) should be obtained.⁵⁸⁴ For all travelers, avoidance of areas of transmission and, if travel to such areas is essential, conscientiously avoiding mosquito exposure is prudent.

Measles Virus

Measles virus (see Chapter 55) causes an annual mortality in the tropics far in excess of that due to the “traditional” tropical disease viruses. In fact, the worldwide yearly mortality due to measles is rivaled among single pathogens only by falciparum malaria, tuberculosis, and AIDS. This mortality is predominantly in sub-Saharan Africa. Similar to AIDS, infection with measles virus is accompanied by marked abnormalities of CMI that contribute to the increased susceptibility to secondary infections that account for much of the morbidity and mortality of the disease.585, 586, 587, 588, 589

Measles is exacerbated in the presence of HIV coinfection.590, 591, 592, 593, 594, 595, 596, 597, 598, 599, 600, 601, 602 The mortality rate in North American case series and reports of measles in HIV-positive children and adults has been 40%, far higher than the usual 0.1% case fatality rate seen in the United States.509, 510, 511, 512, 513, 514, 515, 516, 517, 518, 519, 520, 521, 522, 523, 524, 525, 526, 527, 528, 529, 530, 531, 532, 533, 534, 535, 536, 537, 538, 539, 540, 541, 542, 543, 544, 545, 546, 547, 548, 549, 550, 551, 552, 553, 554, 555, 556, 557, 558, 559, 560, 561, 562, 563, 564, 565, 566, 567, 568, 569, 570, 571, 572, 573, 574, 575, 576, 577, 578, 579, 580, 581, 582, 583, 584, 585, 586, 587, 588, 589, 590, 591, 592, 593, 594, 595, 596, 597, 598, 599 Although the presentation of disease has been normal in many, up to 40% have had no rash. In these reports, giant cell pneumonitis has been the principal complication and the prime cause of death, although fatal subacute measles encephalitis has also been described. CD4+ T-cell counts have not been reported in many of these cases, but where they have been reported they have generally been less than 500/μL. Such case reports and series are obviously likely to be biased toward the severe end of the spectrum of disease, however. Three substantial studies have investigated HIV–measles coinfection in sub-Saharan Africa.600, 601, 602 A study of children hospitalized with measles in Kinshasa showed similar mortality rates among HIV-seropositive (31.3%) and -seronegative (28%) children.⁶⁰⁰ The fact that only severely ill patients, with complications, were hospitalized likely obviated the ability to detect differential mortality in this study. An initial study of children with measles in Lusaka revealed a significantly higher mortality rate in HIV seropositives (28%) than seronegatives (8.3%).⁶⁰¹ A second prospective study in hospitalized children in Lusaka that distinguished between HIV infection and HIV seropositivity found few differences in the clinical presentation, complications, or mortality of HIV-infected compared to -uninfected children with measles.⁶⁰² However, enrollment was based on a clinical diagnosis of measles, which would be expected to minimize the ability to detect differences in clinical presentation; there was a bias against enrollment of critically ill children and those dying soon after admission; and there was significantly greater mortality among HIV-infected compared with -uninfected children among those with clinically diagnosed as opposed to confirmed measles.⁶⁰² Early death prior to mounting diagnostic measles IgM titers in the face of severe immunosuppression was suspected to be confounding. Other positive findings in this study included a higher proportion of HIV infection among children hospitalized with measles than expected from (maternal) population prevalence rates, a greater proportion of coinfected patients hospitalized with measles younger than the age of 9 months, and a longer duration of illness before hospitalization and longer hospitalization in coinfected children.⁶⁰² Follow-up studies have shown that coinfected patients have a higher risk (90.9% vs. 52.8%) for prolonged (30 to 61 days after rash onset) shedding of measles virus⁶⁰³; and that in vivo HIV replication appears to be suppressed during acute measles.⁶⁰⁴

No therapies have been rigorously studied. Vitamin A, which has been shown to be protective in severe measles in malnourished children,⁶⁰⁵ may be of benefit, especially given the marginal nutritional status of many with HIV infection. Ribavirin has been shown to reduce the severity of measles in normal hosts.⁶⁰⁶ Reports of its use in HIV-positive patients with measles pneumonitis have suggested some efficacy, although rigorous data are lacking.^{591, 592, 594, 595, 598} Intravenous use is probably most effective. Intravenous immune globulin (IVIG) may also be of benefit.⁵⁹²

Given the severity of measles in HIV patients, prevention is key. Postexposure prophylaxis with intramuscular immune globulin attenuates disease in normal hosts. It is recommended by the ACIP⁶⁰⁷ in symptomatic HIV patients (and in those with CD4+ T-cell counts <200/μL) regardless of measles serostatus, but it may have limited efficacy in these and other immunosuppressed patients.^{595, 608} The recommended dose is 0.5 mL/kg (15 mL maximum), given intramuscularly within 6 (or, better, 3) days. Such postexposure prophylaxis is also recommended by the American Academy of Pediatrics (AAP) for all HIV-infected children and adolescents, and for all children of unclear infection status born to HIV-infected women, regardless of measles immunization status or degree of immunosuppression.⁶⁰⁹ Preexposure prophylaxis with monthly IVIG has been advocated for HIV-positive children with documented measles vaccine nonresponsiveness during community outbreaks of measles,⁵⁹⁵ but this is not likely to be an economically viable option in the resource-poor areas of the tropics where measles is heavily endemic.

Vaccination remains the principal strategy for preventing measles in HIV-infected people. In normal hosts, the protective efficacy of measles vaccination is greater than 95%.⁶¹⁰ Vaccination efficacy data in HIV-seropositive people is lacking, but seroconversion data are available. This is a less than ideal surrogate. In general, there is a strong correlation between levels of neutralizing antibody and protection, but the failure of postexposure prophylaxis with immune serum globulin in preventing fatal giant cell pneumonia in patients with the cellular immunodeficiencies noted previously provides compelling evidence that CMI mechanisms, whether specific or nonspecific, are important even in protection from initial infection. In adults with HIV infection, there appears to be no waning of measles antibody titers with increasing immunosuppression.611, 612, 613 Unfortunately, there are no clear data on the response to vaccination in those adults who lack antibodies to measles.⁶¹⁴ In children, the situation is different. HIV-infected infants have a markedly lower rate of seroconversion after measles vaccination, generate lower titers of antibody on seroconversion, and have a high rate of secondary vaccine failure, with antibody titers that decrease with time and with increasing immunosuppression.⁶¹⁵ Variable, but generally poor, responses to second doses of vaccine have been reported.⁶¹⁵

There may be a benefit to vaccinating early (at 6 to 9 months of age) to take advantage of the fact that there is less HIV-related immunosuppression at this early age. Early vaccination may also be valuable because both HIV-positive and HIV-negative infants born to HIV-infected mothers have lower titers of maternal antimeasles antibody.⁶¹⁶ In a study in Kenya, the risk of acquiring measles before vaccination at 9 months of age (the standard age of vaccination in Africa at the time) was 3.8 times higher (95% confidence interval, 1.2 to 13.2) in infants born to HIV-seropositive mothers.⁶¹⁶ High-titer measles vaccines may be more immunogenic in HIV-infected children^{617, 618} but have been discontinued for safety reasons in all children.

Safety concerns have obviously been of great importance in the use of this live-attenuated vaccine in HIV patients. Although the use of measles vaccine had appeared to be quite safe in HIV-infected children and adults,⁶¹⁹ two reports have emphasized the need for some caution. A 20-year-old man with no HIV-related symptoms but a CD4+ T-cell count “too few to enumerate” received a second dose of measles vaccine prior to entry into college. One year after vaccination, he developed progressive, vaccine-associated measles pneumonitis.⁶²⁰ A study of the effect of HIV on measles mortality in 356 children hospitalized with measles in Lusaka, Zambia, is also troubling.⁶⁰¹ Previous studies have suggested that when prior measles vaccination does not prevent disease, it can reduce the severity of infection.⁶⁰⁰ In HIV-seronegative boys hospitalized for measles, prior measles vaccination did lead to a significantly lower case fatality rate.⁶⁰¹ Although case fatality rates were not significantly lower in vaccinated HIV-seropositive boys or HIV-seronegative girls than in their unvaccinated controls, there was a trend toward a lower case fatality rate in the vaccinees. Surprisingly, however, the case fatality rate was higher in measles-vaccinated than in unvaccinated HIV-seropositive girls. Although this did not reach statistical significance, it is reminiscent of the experience with the high-titered Edmonston–Zagreb (EZ) vaccine. Use of high-titered EZ vaccine at less than 9 months of age was associated with a delayed excess mortality in several study sites.621, 622, 623 This occurred exclusively in female infants for reasons that remain unclear. It is notable that in the Zambian study noted previously, the highest mortality was seen in the youngest, vaccinated HIV-seropositive girls. However, in regions where there is measles transmission, risk–benefit analysis clearly favors measles immunization of all children regardless of HIV status.⁶¹⁵ In regions where measles transmission does not occur and where immune status can be monitored, withholding of measles vaccine from HIV-infected children with severe immune compromise is wise.⁶¹⁵ WHO recommends measles vaccination for all children in developing countries regardless of HIV infection or symptom status because of the high risk and severity of measles in general in such countries.^{443, 624}

In the United States, the USPHS–IDSA working group and the ACIP recommend measles vaccination for HIV-infected people according to the schedule and conditions for normal hosts if they are not severely immunocompromised.^{213, 610} In addition, the risks and benefits of vaccination or immune globulin prophylaxis should be weighed in severely immunocompromised patients who are at increased risk due to travel or outbreaks.^{213, 607} AAP recommendations for HIV-infected infants to young adults in the United States include

• •No immunization in the face of severe immunosuppression
• •Use of the measles–mumps–rubella (MMR) vaccine at 12 months of age, with a second dose given as soon as 28 days after the first dose
• •With measles transmission in the community, vaccination of infants as young as 6 months old with MMR or monovalent measles vaccine, and revaccination with MMR at 12 months
• •Vaccination of all measles-susceptible household members of an HIV-infected person
• •Use of immune globulin prophylaxis as noted previously⁶⁰⁹

Few data are available on other paramyxoviruses in AIDS. With respiratory syncytial virus infection, pneumonia may be more common than bronchiolitis with wheezing, and viral carriage may also be prolonged.625, 626, 627, 628 Copathogens may occur more frequently than in normal hosts.⁶²⁶ Ribavirin appears to be efficacious in both children and adults.

Rabies

Most human rabies (see Chapter 75) occurs in tropical countries where canine rabies is still endemic. The presentation of rabies does not appear to be altered by HIV infection.⁶²⁹ HIV-infected children and adults clearly have substandard responses to rabies vaccination, however.630, 631, 632 WHO recommendations for postexposure prophylaxis in the face of HIV infection include mandatory use of rabies immune globulin, use of intramuscular vaccine, and monitoring of neutralizing antibody titers.⁶³³ Revaccination may be necessary. Even multiple-site, double-dose postexposure vaccination has led to poor responses in the face of HIV coinfection.⁶³²

Poliomyelitis

Great strides have been made in worldwide polio eradication (see Chapter 60). Most of the remaining burden of poliomyelitis is in the tropics outside the Western Hemisphere. There is no evidence that HIV infection alters the outcome of infection with poliovirus. It has been estimated that more 500,000 HIV-infected children have received live oral polio vaccine (OPV).⁶¹⁵ Only two cases of vaccine-associated paralytic poliomyelitis in HIV-infected children have been reported.⁶¹⁵ If there is greater risk of vaccine-associated disease in the face of HIV infection, the attributable risk is thought to be very low.⁶¹⁵ OPV remains part of the WHO EPI for all children.⁴⁴³ In the United States, the OPV is supported by the ACIP as the only vaccine recommended for polio eradication where there still is transmission of wild-type polio.⁶³⁴ ACIP recommendations have replaced OPV with inactivated polio vaccine for all vaccinees in the United States.⁶³⁴

Other Enteric Viruses

Chronic diarrhea is a common problem in AIDS patients throughout the world. Although no definite pathogenic role has been ascribed to enteric viruses (small round structured viruses, enteric adenoviruses, and coronaviruses) in AIDS-related diarrhea in either North America or Africa,^{635, 636} there are some preliminary data suggesting greater disease severity in children coinfected with HIV and astroviruses⁶³⁷ and an association of picobirnaviruses with diarrhea in HIV-coinfected patients.⁶³⁸ In non-cholera-endemic areas of the tropics, rotavirus is probably the principal cause of diarrheal deaths in HIV-uninfected infants.⁶³⁹ Rotavirus diarrhea does not appear to be an opportunistic pathogen in children with HIV coinfection. In a large study in Malawi, no differences in the severity of rotavirus gastroenteritis were found between HIV-infected and -uninfected children.⁶⁴⁰ Interestingly, rotavirus was less frequently detected in HIV-infected children with gastroenteritis. Despite equal resolution of clinical disease, however, the frequency of death after hospital discharge was significantly greater in coinfected children.⁶⁴⁰

Hepatitis Viruses

Infection with hepatitis A virus (HAV) (see Chapter 64) occurs worldwide. In resource-poor countries, especially in the tropics, HAV is hyperendemic, and exposure to the virus (usually subclinical) is essentially universal by the age of 10 years. Virtually all adults are immune. HIV/HAV coinfection appears to be associated with a higher HAV serum viral load, a longer duration of viremia, and lower elevations in serum alanine aminotransferase levels⁶⁴¹ but a similar disease course.641, 642, 643 Vaccination against HAV is safe in HIV-infected patients.644, 645, 646 Efficacy wanes with increasing immunosuppression.^{645, 646} Some recommend vaccination for all those at risk (defined by a negative serology).⁷⁶

Infection with hepatitis E virus (HEV) (see Chapter 64) is more localized, with sporadic and epidemic disease being reported in Mexico; North and West Africa; the Middle East; and South, Southeast, and East Asia. Clinical disease occurs in both adults and children. HEV has not been implicated as having any significant interaction with HIV.

With the cosmopolitan hepatitis viruses that are capable of causing chronic disease, however, several potentially important interactions with HIV have been described. The prevalence of coinfection is furthered by the fact that these viruses share routes of transmission with HIV. For hepatitis B virus (HBV) (see Chapter 64), CMI responses are thought to be important both for the resolution of acute disease and for the production of hepatic inflammation in chronic disease. HIV infection has been reported to lead to

• •At least a threefold increase in risk of the development of a chronic HBV carrier state (with an inverse correlation with the CD4 count), without a significant change in the severity of acute disease^{647, 648}
• •Potential reactivation of quiescent HBV infection^{649, 650}
• •Increased HBV replication and decreased inflammation in those with^{647, 649} or without⁶⁵¹ chronic hepatitis
• •A decreased response to HBV vaccination⁶⁵²
• •Loss of HBV antibody over time^{652, 653}
• •A 15-fold increase in liver-related mortality in HBV-infected individuals⁶⁵⁴

Vaccination against HBV has reasonable efficacy in HIV-seropositive populations. Interestingly, however, the highest reported rate of development of the chronic carrier state in adults (56% to 80%) occurred in HIV-infected people who were vaccinated at the same time that they developed HBV infection.⁶⁴⁸ HBV does not appear to have a significant effect on the clinical course of HIV infection.⁶⁵⁵

With hepatitis C virus (HCV) infection (see Chapter 64), HIV coinfection is associated with an increased risk of progression to chronic infection.⁶⁵⁶ Documented interactions between the two viruses⁶⁵⁷ include (1) an increase in HCV viral load^{658, 659}; (2) an increase in vertical⁶⁶⁰ and male-to-female sexual transmission⁶⁵⁹ of HCV, probably due to higher viral load; (3) a more severe clinical and histopathologic course^{661, 662}; and (4) a high percentage of indeterminate HCV immunoblot assays, with frequent HCV seroreversion.⁶⁵⁶ Whether HCV infection has a significant effect on the natural history of HIV infection or the response to HAART therapy remains controversial.663, 664, 665, 666, 667 Maternal HCV infection appears to be associated with an increased risk of HIV vertical transmission.⁶⁶⁸ Treatment of both HIV/HCV and HIV/HBV infection has been reviewed recently.⁶⁵⁷

The data on hepatitis D virus (HDV) are more meager. HDV replication may be prolonged or reactivated in the presence of HIV coinfection.^{669, 670}

Herpesviruses

As in the industrial world, herpes zoster is quite common in adult AIDS patients in the tropics. A history of shingles is reported by more than 10% of AIDS patients in Africa.⁶⁷¹ In areas of Africa with a high seroprevalence for HIV, the positive predictive value of a history of shingles for HIV infection is greater than 90%.⁶⁷¹ As elsewhere, zoster tends to develop early in HIV disease, and recurrence is common.^{671, 672}

Chronic genital herpes simplex lesions are common throughout the world in patients with sexually transmitted HIV infection. CMV infection is ubiquitous in most of the tropics.⁶⁷³ The reported incidence of severe disease due to CMV in AIDS in Africa and Asia (although not Latin America and the Caribbean) has lagged behind that of the industrial north, however, likely because of greater mortality at earlier stages of disease.^{674, 675} The natural history, diagnosis, and therapy of disease due to coinfection with these cosmopolitan herpesviruses have been discussed elsewhere.76, 77, 78, 79

Endemic (central Africa), AIDS-related, “classic,” and post-transplant Kaposi's sarcoma are all closely associated with human herpesvirus 8 (HHV8) (see Chapter 57).⁶⁷⁶ In the AIDS era, Kaposi's sarcoma has become one of the leading malignancies in areas of sub-Saharan Africa.⁶⁷⁷ The epidemiology, clinical manifestations, and therapy of HHV8 infection and Kaposi's sarcoma (a disease that the industrialized world had considerable experience with early in HIV pandemic) have been reviewed elsewhere.^{76, 77, 678, 679}

Other Viruses

Although data remain somewhat limited, the cosmopolitan influenza virus appears to cause more severe disease in HIV coinfected patients, with prolongation of symptoms and a higher risk for complications, hospitalization, and death.⁶⁸⁰ Influenza vaccination is safe and recommended by the ACIP for HIV-infected patients.⁶⁸⁰ Such vaccination has been shown to have good efficacy in a study in individuals with a mean CD4+ T-cell count of 200/μL.⁶⁸¹ With advanced HIV, vaccination may not generate protective antibody titers.⁶⁸²

The single case report of severe acute respiratory syndrome (SARS) coronavirus infection in the face of HIV coinfection was within the described clinical spectrum of this newly emerging virus.⁶⁸³

SPECIAL ISSUES

HIV and Eosinophilia

The association of blood (or tissue) eosinophilia with tissue exposure to helminthic parasites can be of considerable diagnostic utility in non-HIV-infected populations (see Chapter 125). Data suggest that this utility may be lessened by the presence of HIV infection. Both relative and absolute eosinophilia occur for nonparasitic reasons with considerable frequency in HIV-infected populations. A relative eosinophilia commonly occurs with progressive HIV disease, at least in North American adult AIDS patients,^{684, 685} due to a sparing of absolute blood eosinophil concentrations in the face of decreasing concentrations of all other granulocytic and mononuclear leukocyte populations.⁶⁸⁵ This effect becomes marked with CD4+ T-cell counts less than 200/μL.⁶⁸⁵ Nonhelminth-related absolute eosinophilia also occurs with regularity with advancing HIV disease. Absolute eosinophilia in North American HIV-seropositive adults has most frequently been associated with pruritic cutaneous disease, particularly the diagnoses of eosinophilic pustular folliculitis, atopic dermatitis, and prurigo nodularis.⁶⁸⁶ Most such patients have CD4 counts less than 100/μL.⁶⁸⁶ It remains unclear in these patients whether the eosinophilia is primary with secondary cutaneous manifestations or an allergic response to a predominantly cutaneous antigen. A few AIDS patients have presented with what appears to be a variant of the hypereosinophilic syndrome with prominent cutaneous manifestations^{687, 688} or a hyper-IgE-like syndrome.⁶⁸⁹ Rare causes of tissue or blood eosinophilia in AIDS patients have included acute eosinophilic pneumonia⁶⁹⁰ and high-grade B-cell lymphomas.⁶⁹¹ Adverse reactions to various prophylactic and therapeutic agents may also elicit eosinophilia in AIDS patients. Other causes of secondary blood eosinophilia appear to be quite uncommon in North American AIDS patients.⁶⁹¹ This includes adrenal insufficiency, despite the high prevalence of adrenalitis at autopsy in AIDS patients.⁶⁹² The upshot of these reports is that the diagnostic specificity of blood eosinophilia for tissue helminthic infection is probably poorer in HIV-infected than in HIV-free adult patients in the industrial north (a conclusion somewhat undermined by the low population prevalence of helminthic disease in the studies cited previously). It is probably reasonable, however, to avoid an extensive workup for parasitic and other causes of eosinophilia in adult eosinophilic AIDS patients with cutaneous disease from the industrial world; however, in populations at higher risk of helminthiasis, such a conclusion is unwarranted in the absence of prospective studies of eosinophil levels in AIDS patients with tissue helminthiasis.⁶⁹³ Such eosinophilia with advancing HIV disease has not been noted in children in the industrial world.^{694, 695}

The data from the tropics are dramatically at variance with the reports from the industrial world. A study of patients with tuberculosis in Burkina Faso reported lower eosinophil levels in HIV-seropositive people compared with seronegative people, with a strong correlation between CD4+ T-cell and eosinophil counts.⁶⁹⁶ A follow-up study in Faso reported that in helminth noninfected HIV patients, eosinophil counts were higher in CDC stage B patients than in controls but were severely decreased in CDC stage C patients.⁶⁹⁷ Mean eosinophil counts in a cohort of 611 HIV-infected patients in Cote d'Ivoire, with a mean CD4+ T-cell count of 115 cells/μL, were in the normal range for HIV-uninfected populations in areas of the world with a low prevalence of helminth infection.⁶⁹⁸ Finally, a case report from England of HAART-related immune reconstitution in a schistosome-infected woman from southeast Africa is intriguing.⁶⁹⁹ The patient presented with a CD4+ T-cell count less than 200/μL and an eosinophil count of 300/μL. Immune reconstitution was accompanied by the development of robust eosinophila (1500/μL) that decreased with praziquantel treatment. Taken together, these findings suggest that late HIV disease in the tropics, as opposed to the industrial world, is associated with a suppression of blood eosinophil levels. The mechanism(s) responsible for these divergent findings remains unclear.

Prevention of Opportunistic Infections

The USPHS–IDSA working group has updated guidelines for preventing OIs in HIV-infected people in industrialized countries.²¹³ The data on 19 OIs frequent in North American HIV-infected populations were reviewed. Factors considered in these evaluations were

• •Disease incidence
• •Disease severity in terms of morbidity and mortality
• •The level of immunosuppression at which disease is most likely to occur
• •The feasibility, efficacy, and cost of preventive measures
• •The impact of intervention on quality of life
• •Toxicities, drug interactions, and the potential for drug resistance
• •The quality of the evidence supporting each recommendation²¹³

The prevention of each OI was evaluated from the standpoint of prevention of exposure, prevention of the first episode of disease by immuno- or chemoprophylaxis (primary prophylaxis), prevention of disease recurrence (secondary prophylaxis), and discontinuance of prophylaxis in those whose CD4+ T-cell counts had risen in response to HAART.²¹³

Specific recommendations for prevention of exposure were made for several agents, including T. gondii, C. parvum, M. tuberculosis, bacterial enteric agents, Bartonella, herpes simplex virus (HSV), varicella zoster virus (VZV), CMV, HHV8, human papillomavirus, and HCV.²¹³ For adults and adolescents, prophylaxis to prevent the first episode of opportunistic disease was strongly recommended as a standard of care for P. carinii, M. tuberculosis (in the face of tuberculin skin test reactivity or contact with a case of active tuberculosis), T. gondii, M. avium complex, and VZV (with exposure to chickenpox or shingles in patients without a history of such, or with negative serologies for VZV [varicella Zoster immunoglobulin (VZIG)]. Primary prophylaxis against HBV, HAV, influenza virus, and S. pneumoniae was generally recommended. Although evidence exists for efficacy, primary prophylaxis against C. neoformans, H. capsulatum, CMV, and bacterial infection (in the face of neutropenia) was not routinely recommended. For two of these agents (P. carinii and M. avium), primary prophylaxis has been shown to confer a survival benefit.^{700, 701} Secondary prophylaxis to prevent recurrent disease was strongly recommended as the standard of care for P. carinii, T. gondii, M. avium complex, CMV, C. neoformans, H. capsulatum, C. immitis, and nontyphi Salmonella species. Such prophylaxis was recommended for HSV and Candida only if subsequent episodes are frequent or severe.²¹³

For children, primary prophylaxis was strongly recommended as a standard of care for P. carinii, M. tuberculosis, M. avium complex, and VZV (VZIG); generally recommended for T. gondii, VZV (immunization in the absence of immunosuppression), and influenza virus; and recommended only in unusual circumstances for invasive bacterial infection (hypogammaglobulinemia, IVIG), C. neoformans (severe immunosuppression), H. capsulatum (severe immunosuppression, endemic geographic area), and CMV (CMV antibody positivity and severe immunosuppression). Primary prophylaxis was also addressed through review of the recommendations for routine immunization schedules in HIV-infected children. In addition to standard schedules for immunization against HBV, HAV, polio, H. influenzae type b, diphtheria, tetanus, and pertusis, altered schedules for vaccination against S. pneumoniae (use of the heptavalent pneumococcal conjugate vaccine beginning at 2 months, followed by 23-valent pneumococcal polysaccharide vaccine at 2 years), influenza (yearly dose recommended), MMR (no administration to severely immunosuppressed children), and VZV (administration only to asymptomatic, nonimmunosuppressed children) were reviewed. Recommendations for secondary prophylaxis were similar to recommendations for adults, with the addition of a recommendation for use of TMP–SMX or IVIG to prevent invasive bacterial infection in the presence of more than two such infections in a 1-year period.²¹³

These recommendations are likely to find broad applicability in the industrial world, where the OI spectrum, health-care priorities, and available prevention options are similar. The applicability to much of the tropics is less clear, however, given differing spectra of OIs, differences in antibiotic resistance patterns, and differences in sociocultural acceptability or feasibility of preventive measures. Limits in the availability of health-care resources (including not just an inability to support the cost of many prevention regimens but also an inability to diagnose HIV disease early enough for preventive measures to be effective, to reliably stage the degree of HIV-associated immunosuppression, and to definitively diagnose OIs) also directly influence the range of prevention options and priorities. It should also be noted that compared with HAART therapy (or prevention of HIV infection), the benefit of OI prevention in reducing HIV-related morbidity and mortality may be somewhat modest.⁷⁰² Adequate global provision of HAART represents an ongoing, immense challenge, however, and wide implementation of simple, cheap, effective OI prevention strategies provides an opportunity to rapidly and widely reduce morbidity and mortality.⁶⁷⁵ The wide availability of affordable and effective regimens may also encourage people to seek HIV testing.⁶⁷⁵

In 1996, Kaplan and colleagues argued that effective research on OI prevention strategies in the tropics will require an integrated approach, including the area-specific determination of OI spectra, determination of environmental reservoirs of opportunistic pathogens and feasible ways to reduce exposure, assessment of immuno- and chemoprophylaxis against such pathogens, and improvement in the ability to identify and inexpensively stage HIV infection.²⁰ Since then, data on the efficacy of some OI prevention strategies in resource-poor countries of the tropics have accrued.

First, as noted previously, primary preventive therapy against tuberculosis (TB) with isoniazid has been shown be effective in HIV-infected individuals, regardless of tuberculin status.^{423, 424, 675} WHO and UNAIDS recommendations are for primary preventive therapy to be given to PPD-positive, HIV-infected individuals who do not have active TB.⁴²⁵ In settings where it may not be feasible to do PPD testing, the recommendations are for primary preventive therapy to be considered for those living in populations with a prevalence of tuberculous infection estimated to be more than 30%, healthcare workers, household contacts of TB patients, prisoners, miners, and other groups at high risk of acquisition or transmission of TB.⁴²⁵

Second, TMP–SMX—a cheap, widely available antibiotic with activity against a plethora of OIs (including PCP, nontyphoid salmonellosis, pneumococcal disease, and toxoplasmosis) and malaria—has been shown to reduce morbidity and mortality in HIV-infected adults^675, 702, 703, 704, 705 and children.^{401, 706} WHO/UNAIDS recommendations are that TMP–SMX should be used for prophylaxis in adults and children living with HIV/AIDS in Africa as a minimum package of care.⁷⁰⁷ For adults in Africa (defined as those older than the age of 13 years), such prophylaxis should be offered to all people with symptomatic HIV disease, those who are asymptomatic with a CD4 count of 500/μL or less (or total lymphocyte equivalent), and pregnant women after the first trimester.⁷⁰⁷ WHO/UNAIDS/UNICEF recommendations are that all HIV-exposed children (born to HIV-infected mothers) should get TMP–SMX from the age of 4 to 6 weeks, as should any child identified as HIV infected with any clinical signs or symptoms suggestive of HIV, regardless of age or CD4+ T-cell count.⁷⁰⁸ It is further recommended that TMP–SMX should be discontinued (1) in HIV-exposed children only once HIV infection has confidently been excluded (and the mother is no longer breastfeeding); (2) in HIV-infected children on antiretroviral therapy only when evidence of immune restoration has occurred; and (3) in those with severe cutaneous, renal, hepatic, or hematological toxicity.⁷⁰⁸ It has been noted that such mass prophylaxis strategies entail some as yet to be quantified risks, principally of increasing rates of drug-resistant bacteria and malaria.⁷⁰⁹ Studies to examine such risks should be performed.

Third, a large, randomized, double-blind, placebo-controlled trial of the 23-valent pneumococcal polysaccharide vaccine in HIV-infected Ugandan adults showed no protective effect (with invasive pneumococcal disease as the primary endpoint).⁷¹⁰ Surprisingly, increased rates of pneumococcal disease were seen in vaccine recipients. The potential mechanism(s) remains unclear; not surprisingly, the study remains controversial.

Fourth, among the specifically tropical OIs, a controlled, double-blind trial of primary prophylaxis with itraconazole in Thailand showed efficacy in preventing penicilliosis (and cryptococcosis).³⁴⁰ There are little or no data available on the prophylaxis of other tropical OIs, such as leishmaniasis and American trypanosomiasis, although the utility of avoiding exposures to the vectors of the agents of these diseases should be clear.

More generally, certain options for preventing or reducing exposure to opportunistic pathogens are likely to be broadly useful, including avoiding unboiled water, raw or undercooked foods, and unpasteurized milk to prevent enteric bacterial and protozoan infections as well as T. gondii exposure and also avoiding contact with patients with TB, for example, in patient-care settings. Avoidance of exposure to the opportunistic agents of disseminated fungal disease is likely to be impractical in most settings.

The HIV-Infected Traveler in the Tropics

It is clear that for many people, HIV infection is more an inducement than a hindrance to travel.⁷¹¹ For HIV-infected patients from the industrial world, travel to the tropics leads to an increased risk of exposure to a variety of ubiquitous as well as geographically focal infectious diseases (see Chapter 120). The principal concerns are enteric, respiratory, and vector-borne infections. STDs deserve special mention, however. Despite the AIDS pandemic, studies continue to document a high frequency of unprotected sexual encounters among travelers to the tropics.^{712, 713} For the HIV-infected traveler, it bears remembering that spread of the AIDS pandemic has been along the routes of human travel and migration.⁷¹² HIV-infected travelers should notify sexual partners of their HIV status, observe safe sex practices, refrain from donating fluids or tissue, inform health-care providers of the need for blood precautions, and avoid the sharing or recreational use of needles. For travelers who are not infected with HIV, pretravel counseling should be given on strategies to minimize the risk of HIV acquisition during travel, including

• •Avoiding sex with new partners
• •Using latex condoms if having sex (remembering that condom quality varies throughout the world)
• •Taking a supply of sterile needles and syringes if routine or frequent injections are required (e.g., for insulin-dependent diabetes) or if a prolonged stay in remote areas is planned
• •Avoiding all skin-piercing procedures (e.g., tattooing, acupuncture, and shaving by barbers) and avoiding invasive medical and dental procedures if possible
• •Being cognizant of the risks of motor vehicle travel because motor vehicle injuries are not just among the greatest health threats to travelers in the tropics but also provide a major risk factor for needing emergency blood transfusion in countries where HIV screening of the blood supply is not routine

These are also important educational points for HIV-infected travelers.

There are recent reviews of medical issues in the HIV-infected tropical traveler from the industrial world^{714, 715} and information is available from www.cdc.gov/travel/hiv.trav.htm. Careful, timely pretravel advice and planning can help HIV-infected patients minimize the unavoidable risks of travel to the tropics. Issues to cover include

• •Providing an adequate supply of currently used medications
• •Identifying optimal sources of medical care in planned destinations (and obtaining adequate medical insurance to cover such care)
• •Discussing potential legal restrictions on travel
• •Providing education on avoiding food-, water-, and vector-borne disease
• •Providing vaccination, chemoprophylaxis, and antimicrobial agents as appropriate reviewing the medical geography of the route and planned activities to identify any special risks
• •Providing for adequate medical follow-up on return from travel

Prior to travel, those at risk of HIV infection should be screened for infection as part of the routine pretravel evaluation. Those known or found to be HIV infected should have their disease staged by CD4 count and viral load because the level of immunosuppression will influence the travel medical recommendations given. Therapy with HAART can complicate tropical travel.⁷¹⁵ For those already on HAART, concerns include

• •The problem of appropriate clinical and laboratory monitoring—for those planning extended residence in places lacking appropriate clinical or laboratory services, it would be prudent to schedule two or three trips per year for medical follow-up to a country where such services are available
• •Difficulties with compliance with complicated HAART regimens during the dislocations and changing daily schedules of travel
• •The fact that approximately 50 countries currently restrict the entry of those infected with HIV, especially long-term visitors such as students and workers, despite compelling arguments against the utility and for the counterproductive nature of such policies

An unofficial listing of country requirements is available at http://travel.state.gov/trvel/HIVtestingregs.html. Because possession of HAART medications may reveal the traveler's HIV infection, a potential solution is to remove or cover drug labels. Clearly, this is not without its own risks, which demand that careful records (medication names, dosages, frequencies, and physical appearance) be kept in more than one safe place by the traveler. Finally, there is the problem of untoward drug interactions with protease inhibitors. When initial HIV infection diagnosis and/or staging occurs during pretravel evaluation, the traveler may be newly recognized to fall into a group for whom HAART therapy is recommended. For those planning short-term travel, initiation of HAART just prior to departure is probably unwise due to the risk of developing significant side effects—many of which tend to be worse with initiation of therapy and some of which are intensified by stress—in the absence of close contact with a knowledgeable provider. For those in this situation who are planning long-term residence in the tropics, departure should probably be postponed until the traveler is stably on HAART therapy, unless close, timely follow-up can be achieved in the destination country. A critical variable, as ever in the rapidly changing field of HIV clinical care, is adequate, ongoing experience. Travel physicians without extensive, current experience in the care of those with HIV/AIDS should be in close consultative contact with physicians who possess this expertise: Today's verities rapidly become out of date. Textbooks of HIV care with frequent Web-based updates are also valuable resources.^{76, 77}

Sufficient quantities of currently used medications, along with replacement prescriptions, are essential. Patients should be counseled to seek medical attention promptly when ill. Although the current communication revolution allows a growing number of traveling patients to stay in close contact with their HIV and travel medicine physicians at home via e-mail, the prior identification of physicians with significant HIV experience in countries of planned travel is prudent. Obtaining medical insurance to cover such care is wise. Another basic consideration, especially for those with advancing HIV-related disease, is the buying of trip cancellation and/or repatriation insurance.

Enteric pathogens are a major threat to the HIV-infected traveler to the tropics. In addition to the fact that many tropical travel-associated enteric bacterial infections are more severe in the presence of HIV coinfection, the achlorhydria that occurs in AIDS may markedly lower the amount of inoculum needed for the establishment of infection.^{566, 567} Precautions recommended for all travelers should be followed assiduously. Foods and beverages may be contaminated, particularly raw, unpeeled fruits and vegetables; raw or undercooked eggs; meats; seafood; tap water; ice; unpasteurized dairy products (beware of soft cheeses); and food purchased from street vendors.²¹³ Steaming hot foods, meat cooked until brown throughout, water brought to a boil for 1 minute, bottled (especially carbonated) beverages, very hot beverages, beer and wine, and fruits and vegetables peeled by the traveler are generally safe.²¹³ In addition to preventing enteric disease, following these precautions will also lower the risk of infection with T. gondii. Treating water with iodine or chlorine may be done, preferably in conjunction with filtration, when boiling is not practical. Filtration of water, using reverse osmosis or sub 1 micron filters, is efficacious in removing Cryptosporidia from water.⁷¹⁶ All tap water should be avoided, even small amounts used for brushing teeth. Recreational water exposure also carries a risk. Obviously, fecally contaminated water should be avoided altogether. Swallowing water while swimming should be avoided.²¹³ Direct contact with soil and sand should be avoided in places where fecal contamination is likely; hands should be washed frequently. Antimicrobial prophylaxis for enteric pathogens is not routinely recommended for HIV-uninfected travelers. For HIV-infected travelers, such prophylaxis may be considered, however, depending on the level of immunosuppression, the risk of infection, and the duration of travel.²¹³ Fluoroquinolones are probably the drugs of choice. All HIV-infected travelers should carry antimicrobials (with or without antimotility agents) for an empirical treatment course in case diarrhea develops. Ciprofloxacin 500 mg twice a day for 3 to 7 days is a reasonable choice. Alternatives such as TMP–SMX or azithromycin should be used for children and pregnant women. Medical attention should quickly be sought if empirical therapy fails, if shaking chills or hematochezia occur, or if dehydration develops.²¹³

Contact with arthropod vectors should be reduced through the careful use of insect repellents, wearing clothes that cover the arms and legs when outdoors, sleeping in well-screened areas or under a bed net, and avoiding outdoor exposure between dusk and dawn (see Chapter 120). The threat of sandfly transmission of visceral leishmaniasis (see Chapter 94) deserves special mention and education.

Malarial chemoprophylaxis is essential, where indicated, for all travelers. The HIV protease inhibitors currently used for HAART have understudied interactions with a variety of drugs used for malaria prophylaxis and treatment. Based on current data, mefloquine, doxycycline, chloroquine, and malarone (atovaquone + proguanil) are likely to be safe and to retain efficacy for malaria prophylaxis. The provision of empiric, standby treatment doses (e.g., of malarone) seems prudent.

Recommendations for the immunization of HIV-seropositive travelers to the tropics are summarized in Table 133-1 . Preparation for travel should include a careful review of routine vaccinations in addition to both routine and special travel vaccinations. Vaccine response rates, as measured by antibody titers, tend to decline with increasing HIV-related immunosuppression. This gives impetus, especially for “routine” vaccinations, to early identification of the need for vaccination. Essentially all live bacterial and viral vaccines are considered to be contraindicated in the presence of HIV infection,²¹³ including oral polio vaccine, oral typhoid vaccine (Ty21a), and BCG. The special cases of measles and yellow fever vaccines are considered in the “pathogen-centered” discussions earlier in the chapter and in Table 133-1.

Table 133-1.

Immunizations for HIV-Infected Adults Who Are Traveling to the Tropics

Vaccine	Recommendation	Comments
Routine immunizations*
Tetanus–diphtheria	Boost every 10 yr
Pneumococcus	Use 23-valent polysaccharide vaccine; boost at 5 yr	Recommended by USPHS–IDSA
Influenza		Recommended by ACIP, USPHS–IDSA
		Year-round infection in the tropics
		Southern Hemisphere: April through
		September; repeat annually
Hepatitis B	3 doses of Recombivax HB or Engerix B
Measles†	Single dose of measles vaccine or MMR for susceptible persons who are not severely immunosuppressed; in the face of severe immunosuppression, consider immunoglobulin
Standard travel immunizations
Hepatitis A	Single dose 2 wk prior to travel; boost at 6–12 mo
Poliomyelitis	Single dose of enhanced inactivated vaccine	Oral (live) vaccine is contraindicated and discouraged in close contacts
Typhoid	Single dose of polysaccharide vaccine at least 2 wk prior to travel; boost at 2 yr Alternately, two doses (separated by at least 1 mo) of inactivated vaccine; boost at 3 yr	Side effects lessened with polysaccharide vaccine; live oral vaccine is contraindicated
Yellow fever†		Contraindicated by the ACIP; recommended for those with asymptomatic HIV infection by WHO; considered for those with asymptomatic HIV infection who cannot avoid potential exposure, by USPHS, IDSA
Special travel immunizations
Cholera	Two doses, at least 1 wk apart; boost at 6-mo intervals	Rarely indicated, given low risk of disease and limited effectiveness of vaccine; live vaccine contraindicated
Meningococcus	Single dose (A/C/Y/W-135)
Plague	Three doses (one each at 0, 1, and 3–6 mo), with boost at 1 or 2 yr
Rabies	See text	Avoid giving HDCV given intradermally, given potentially weaker response
Japanese encephalitis	Three doses, one each on days 0, 7, and 30; boost, based on antibody levels, at 1–3 yr

^*Routine primary series of immunization for diphtheria–tetanus, MMR, and polio is assumed.

^†Absolutely or relatively contraindicated in some circumstances.

The risk of acquisition of tuberculosis is much higher in most tropical countries than in the United States, although the risk of becoming infected during short-term travel is low. As noted previously, the ACIP and the USPHS–IDSA working group consider the administration of BCG to be contraindicated because of the risk of dissemination, although it is recommended by WHO for those with asymptomatic HIV infection in areas with a high prevalence of TB infection. All HIV-infected people should receive PPD testing. Follow-up testing after return from the tropics (or during travel or residence in the tropics, if extended) should be performed. With a positive PPD (>5 mm induration) or a known high-risk exposure, HIV-infected people with no evidence of active TB, or history of treatment for latent or active TB, should be treated for latent TB.²¹³ Whether primary chemoprophylaxis should be recommended for travel in areas with a high population prevalence of TB is unclear.

The endemic systemic fungi carry considerable risk for HIV-infected travelers. Where the environmental sources are known, the risk of infection may be reduced somewhat by attempting to minimize exposure. Such measures include avoiding soil exposure, especially during the rainy season, in southern China and Southeast Asia (P. marneffei); avoiding exposure to disturbed soil and dust in lower Sonoran life zones of the Americas (C. immitis); and avoiding caves and soil and dust exposure in areas with heavy avian and bat excrement (H. capsulatum). Most often, the reservoirs are either unknown or too widespread to allow for avoidance. Primary prophylaxis with itraconazole (200 mg daily) may have merit for short-term prophylaxis during travel in areas with a high incidence of systemic fungal infection.^{340, 717}

Finally, it may be wise to avoid some areas of high potential risk by changing the planned travel itinerary or activities, especially for those people with advanced HIV disease. In this regard, it should be noted that many of the tropical infectious diseases of special concern here, such as yellow fever, have a fairly focal pattern of risk within endemic countries.