Chapter 81: Hematologic Manifestations of Acquired Immunodeficiency Syndrome

HIV, the virus that causes AIDS, is a lentivirus that originated as a simian immunodeficiency virus (SIV) in chimpanzees and entered the human population in the early 20th century in equatorial Africa.^1,2 First isolated in 1983,^3,4 HIV-1 actually comprises four distinct viruses (types M, N, O, and P) that represent four separate transmission events that occurred between chimpanzees and humans, likely the result of predation of monkeys by humans and mucosal or nonintact skin contact with infected fluids. Group M, the viral type responsible for the HIV-1 pandemic, was detected in a tissue sample from 1959 and probably entered the human population in or around Kinshasa, Democratic Republic of Congo (then Leopoldville, Belgium Congo) between 1910 and 1930 based on phylogenetic analysis.² HIV-2 originated in West Africa, the result of cross-species transmission of SIV from sooty mangabeys to humans. Patients infected with HIV-2 progress more slowly and have lower plasma viral loads (often nondetectable) than those with HIV-1, reflective of the different virology and adaptation to humans of this SIV.⁵ Because of lower rates of replication and transmission, HIV-2 prevalence is declining and is being replaced by HIV-1 in countries where both viruses are endemic.^5,6 Among those infected with HIV-1, group M is the globally predominant viral strain and is further divided into nine subtypes and many more recombinant viruses (circulating recombinant forms [CRFs]) with some geographic localization. Subtypes A and D predominate in East Africa; subtype C in South Africa, India, and Asia; subtype B in the Caribbean, the Americas, and Western Europe; and CRF01 in Southeast Asia.¹

The development of Pneumocystis jiroveci (Pneumocystis carinii) pneumonia and Kaposi sarcoma in previously healthy men who have sex with men on both coasts of the United States in 1981 represented the first clinical manifestations of HIV and the onset of the HIV-1 pandemic.^7,8,9 Subsequent reports of similar illnesses in the sexual partners of index cases, injection drug users, patients with hemophilia and other transfusion recipients, infants born to infected mothers, and Haitian immigrants^{10,11,12,13,14,15,16,17} helped identify the routes of transmission as bloodborne, sexual, or vertical. With the discovery of HIV in 1983 and the subsequent development of serologic testing, more systematic detection of HIV infections became possible providing an understanding of the regional and global HIV epidemiology. While sexual contact between men was responsible for most infections in the United States, Northern Europe, Australia, and parts of Central and South America, heterosexual spread predominated in sub-Saharan Africa and injection drug use followed by sexual transmission was responsible for most infections in Southern and Eastern Europe and Southeast Asia.¹⁸ Transmission rates between individuals per incident/act is dictated by the viral load in the HIV-infected person,¹⁹ the presence of modifying factors such as concurrent ulcerative sexually transmitted diseases and the type of exposure.²⁰ Rates vary between 93 percent for blood transfusion from an infected person to less than 0.04 percent for oral sex. Estimated rates for mother-to-child transmission (in the absence of antiretroviral therapy [ART] prophylaxis) are 23 percent, for needle sharing 0.63 percent, for needle stick 0.23 percent, for receptive anal intercourse 1.38 percent, for insertive anal intercourse 0.11 percent, for receptive vaginal intercourse 0.08 percent, and for insertive vaginal intercourse 0.04 percent.²⁰ In 2012 there were an estimated 35.3 million people living with HIV, including 2.3 million newly infected persons.²¹ Although the global incidence of HIV is thought to have peaked in 1997, the prevalence of HIV is increasing because of ongoing new infections and the shrinking death rate of those already infected and on ART (2.3 million deaths in 2005 versus 1.6 million deaths in 2012). The majority of HIV-infected persons (approximately 23 million) now live in sub-Saharan Africa with 4 million in Asia and Southeast Asia and roughly 3 million in the Americas and Caribbean.

Eighty percent of HIV infections occur via mucosal transmission during sex²² when cell-free and cell-associated virions transverse the epithelium to gain access to macrophages, Langerhans cells, dendritic cells, and CD4-expressing T lymphocytes.²³ To infect most cells HIV must bind to CD4 and one of two major coreceptors (CCR5 or CXCR4); in most cases, CCR5-utilizing viral strains are those that are transmitted and predominate early in disease. Rare individuals who do not express CCR5 (homozygote for a 32 bp deletion mutation in the CCR5 gene) are highly resistant to HIV infection although they can be infected with isolates utilizing CXCR4. After transmission, low-level replication of HIV in tissue macrophages and dendritic cells can occur, but the key role these cells play is in trapping and trafficking virions and presenting them to CD4+ T lymphocytes within regional lymphatics (e.g., gut-associated lymphoid tissue and lymph nodes where the infection is amplified).²⁴ High-level viral replication proceeds within these local tissues leading to profound CD4+ T-cell depletion, establishment of a reservoir of latently infected memory T cells and eventually to high plasma levels of virus that are the hallmark of acute infection. The immune response to HIV is brisk but ineffective and may, in fact, fuel the infection because the expression of inflammatory cytokines²⁵ and migration of activated CD4+ T cells to the site of HIV-1 concentration provides additional activated cells that the virus coopts for its own replication.²⁶ The initial antibody response to HIV does not contain neutralizing antibodies; these develop only later, months after chronic infection is established. Furthermore, HIV escapes these antibodies by mutations within N-glycosylation sites that prevent antibody binding.²⁷ The CD8+ cytotoxic T-lymphocyte (CTL) cell response to HIV controls high-level viral replication during primary infection and establishes the viral “setpoint” or plasma level of HIV RNA in chronic infection. Evidence for the controlling anti-HIV effect of CD8+ T cells includes their detection immediately prior to peak viremia, the development of viral escape mutations^28,29,30 and the requirement for CD8+ T cells to control SIV infection in Rhesus macaques.³¹ The rate of viral escape mutations slows during chronic infection^32,33 and is not associated with further declines in viral load, reaching a stalemate where viral replication continues under the pressure of a slowly evolving CTL response leading to viral strains with reduced replication capacity.^34,35,36 As important as the direct cytolytic effect of HIV on CD4+ T cells, the virus induces a state of chronic immune activation of both the adaptive and innate immune systems that is central to disease pathogenesis.^{37,38,39,40,41} Because the immune response to HIV is defective and does not clear the virus, the immune system remains continually activated with high rates of T-cell turnover that eventually leads to T-cell exhaustion and depletion. This is particularly evident in gut-associated lymphoid tissue where early T-cell losses alter the integrity of the mucosal border leading to microbial translocation and leakage of lipopolysaccharide (LPS) into the blood which, in turn, amplifies the state of immune activation.⁴² This persistent, systemic inflammatory state leads to tissue fibrosis over time^43,44 that is partly responsible for immune failure and the increased frequency of nonimmune, nontraditional chronic diseases that now plague an aging HIV population.⁴⁵

Primary HIV infection that comes to medical attention presents as a febrile illness that may include headache, pharyngitis, lymphadenopathy, gastrointestinal symptoms, and rash and may be mistaken for mononucleosis or other nonspecific viral infections. Key to making the diagnosis is taking a history for HIV risk factors and obtaining appropriate laboratory testing (combined HIV Ag/Ab assays and plasma HIV RNA testing). However, in most cases, primary infection goes undiagnosed and patients are later identified in the chronic, asymptomatic phase of infection by routine screening or later still, after the development of symptoms that are often caused by opportunistic infections. Typically, the asymptomatic phase of chronic infection will last for 8 to 10 years, although there is great interindividual variation dictated by the effectiveness of the immune response in controlling HIV replication (the viral “setpoint,” see above). Long-term nonprogressors (those who maintain CD4+ T-cell counts >500 for 5 years without therapy) and elite controllers (those with low or nondetectable plasma HIV RNA without treatment) can live for decades with limited or no disease progression, while others with high viral setpoints in the range of 100,000 to >1,000,000 copies/mL can develop AIDS-defining illnesses quickly after primary infection. In untreated individuals CD4+ T-cell counts (typically at CD4 counts) typically decline by 50 to 100 cells/μL per year, taking 8 to 10 years before counts are in the range where symptoms develop (typically at CD4 count <500 cells/μL) or AIDS-defining illnesses occur (typically at CD4 count <200 cells/μL). Historically, opportunistic infections provided the first evidence for the existence of HIV and remain the most visible manifestation of infection in countries with limited access to ART and in individuals who are diagnosed late in the course of their disease. The development of opportunistic infections and AIDS-defining conditions is dependent on the virulence properties of the organism and the degree of host immune suppression. Pathogens with high virulence potential (e.g., Mycobacterium tuberculosis, Salmonella sp., the bacterial agents of community-acquired pneumonia) cause disease in patients without HIV and do so in HIV-infected persons regardless of CD4 count (although more severe and prolonged illness occurs with more profound immunodeficiency). Agents with more limited pathogenic potential typically cause disease at lower CD4 counts, for example, P. jiroveci at CD4 counts below 200 cells/μL, while those that rarely cause disease in immunocompetent persons, such as disseminated Mycobacterium avium complex, Toxoplasma gondii encephalitis, and JC virus (the agent of progressive multifocal leukoencephalopathy), typically occur only in those with very advanced HIV disease and CD4 counts less than 100 cells/μL or less than 50 cells/μL, respectively (Table 81–1 lists HIV staging; Table 81–2 lists AIDS-defining conditions; and Table 81–3 lists common HIV-associated conditions by CD4 count). Prophylaxis against the development of these infections is provided when the infection is common and significant and when the prophylaxis is effective, inexpensive and well tolerated. (Table 81–4 lists the organisms and medications used for primary prophylaxis.) Tumors classified as AIDS-defining malignancies are Kaposi sarcoma, Burkitt lymphoma, immunoblastic lymphoma, primary CNS lymphoma and cervical cancer, as these were first identified at high rates among infected persons early in the epidemic. Many other cancers also occur at increased rates among HIV-infected patients because of a higher rate of traditional cancer risk factors and the long-term effects of immune dysregulation leading to decreased tumor surveillance and chronic systemic inflammation. Outside of the immune defects levied by HIV, the virus can directly or indirectly cause specific organ or tissue damage including the nervous system (causing cognitive impairment, dementia and peripheral neuropathy), cardiovascular system (HIV cardiomyopathy), kidney (HIV nephropathy), gastrointestinal system (HIV enteropathy and cholangiopathy), and can accelerate disease progression as a result of other infections, such as hepatitides B and C.^46,47 Finally, the long-term effects of chronic immune activation and persistent inflammation is likely a factor in the development of coronary artery disease,⁴⁵ chronic liver disease,^47,48 and a hypercoagulable state^49,50,51 that is only partly corrected by the initiation of ART. These “aging effects” of HIV are likely to dominate the health issues for infected persons now that opportunistic infections are readily treated or avoided altogether through a combination of prophylaxis and ART.

The story of ART from the first reports that zidovudine had activity against HIV to the current formulary of drugs, including single-tablet, fixed-dose, once-daily formulations, is one of the great achievements in medicine. Early studies of zidovudine monotherapy demonstrated a delay to the development of AIDS and short-term mortality benefits but no long-term effect on survival; zidovudine also carried significant toxicity.^{52,53,54,55,56,57} Combination therapy with other nucleoside reverse transcriptase inhibitors (nRTIs) proved slightly more effective than zidovudine alone, but not until combination nRTI was used with a third drug from another class, first nonnucleoside reverse transcriptase inhibitors (nnRTIs)⁵⁸ and then protease inhibitors,^59,60,61,62 were sustained viral suppression and substantive, dramatic improvements in survival realized.⁶³ The recommended time to initiate ART has evolved in response to studies demonstrating benefits of ART at high CD4 counts and improvements in drug tolerability and formulation. Current Department of Health and Human Services (DHHS) guidelines suggest that all HIV-infected persons be offered ART regardless of CD4 count, although the strength of evidence supporting this recommendation varies by CD4 count (Table 81–5 lists the criteria for initiating ART), while the World Health Organization sets a CD4 count threshold of 350 cells/μL for initiating ART in more resource-limited countries. The rationale for these expanded ART recommendations include the recognition that newer therapies are more convenient, have fewer adverse effects and are associated with lower rates of drug resistance. Furthermore, in addition to improved all-cause survival, ART preserves renal function in those with HIV-associated nephropathy,⁶⁴ slows the progression of hepatic fibrosis in those coinfected with hepatitis C,^65,66,67,68 decreases (but does not normalize) markers of chronic inflammation⁶⁹ and may be associated with reduced cardiovascular disease,⁷⁰ prevents the development of HIV-associated dementia,^71,72 and is highly effective in reducing mother-to-child^73,74 and sexual transmission.^19,22,75 Adverse effects related to ART do occur but are less common with current regimens and can usually be effectively managed with corrective treatments and by substitution of the offending drug with an alternative medication.^76,77 Similarly, the presence or development of drug resistance can usually be overcome by the use of secondary or salvage ART regimens that are fully suppressive. At the current time, only rare patients who are fully adherent to ART fail to control HIV replication. Treatment of early and primary infection provides a unique opportunity to intervene and possibly alter the course of HIV infection. Several studies have demonstrated that treatment in early or primary HIV lowers the rate of disease progression if treatment is subsequently interrupted^{78,79,80,81,82,83} and may also limit the size of the latent HIV reservoir,^84,85,86,87 the impediment to curing patients. One interesting group of 14 patients initiated ART during primary infection and stayed on treatment for a median of 3 years and then controlled HIV replication for a median of 7 years after ART interruption.⁸⁸ Finally, given the high viral loads typical of primary HIV, these patients are thought to be highly infectious; therefore identifying them and initiating treatment should prevent transmission to their uninfected partners. One consequence of initiating ART in the setting of a known or occult infection is the development of an acute inflammatory reaction as a result of reconstitution of the immune system in the presence of organisms or foreign antigens.^{89,90,91,92,93} The immune reconstitution inflammatory syndrome (IRIS) occurs in between 8 and 30 percent of patients who start ART, depending on the opportunistic infection and the timing of ART.⁹⁴ Risk factors for the development of IRIS include a low baseline CD4 count, more-severe disease and a short interval between treatment of the opportunistic infection and initiation of ART. The treatment of IRIS should include treatment of the underlying infection or condition, continued ART, and antiinflammatory medication, such as corticosteroids, depending on the severity of the reaction.⁹⁵

The future of the HIV epidemic will differ by region and be dictated by local public health responses, HIV testing rates, sociobehavioral prevention interventions, and access to ART. Expanded HIV testing is an essential element of any prevention campaign as an estimated 50 percent of all new infections originate from individuals unaware of their HIV status.⁹⁶ Behavioral interventions can have some preventative effect^97,98,99 but biomedical methods have emerged as the most effective means to prevent new infections. Male circumcision can reduce female-to-male sexual transmission by 51 percent¹⁰⁰ and is being implemented on a population level in some African countries. ART administered peripartum will prevent most mother-to-child transmissions^101,102 and fully suppressive ART provided throughout pregnancy essentially eliminates all infant infections.^73,74 A prospective randomized trial of HIV-discordant couples demonstrated that ART provided to the HIV-infected partner was almost 100 percent effective in preventing transmission²² and other studies have shown that elements of ART given to HIV-negative but at-risk persons (preexposure prophylaxis [PrEP]) can prevent HIV acquisition when subjects are adherent to treatment.^103,104,105 These studies point the way to the best strategies and interventions to curtail the HIV epidemic until an effective HIV vaccine is available.

The persistence of replication-competent but transcriptionally silent HIV proviral DNA in long-lived resting cells (the HIV latent reservoir) is the impediment to cure for nearly all HIV-infected people.^{87,106,107,108,109} Although combination ART is highly effective at controlling HIV replication in activated cells, it has no effect on the latent HIV reservoir, which will persist as long as the cells harboring HIV survive. Because most HIV genomes reside in central and effector memory T cells that decay at a negligible rate, there is no possibility of cure with ART alone. The only individual cured of chronic HIV infection (Timothy Brown, the Berlin patient) developed acute myelogenous leukemia that was treated with high-dose conditioning and transplantation of HIV-resistant (CCR5D32/D32) allogeneic blood stem cells.¹⁰⁹ While this case provides proof-of-principle that the latent HIV reservoir can be eliminated by allogeneic hematopoietic cell transplantation, the approach is impractical for widespread application because of the high toxicity associated with this procedure, the morbidity of graft-versus-host disease, and the scarcity of CCR5D32/D32 donors. To date, most HIV cure efforts have focused on the strategy of reversing latency with the notion that once resting cells begin producing HIV they will be targeted by the immune system or die from apoptosis. However, early studies suggest that activating latent cells to express HIV does not reliably lead to their death and additional treatments including vaccination to boost cytotoxic responses will be needed.^109A Gene therapy is also being pursued as a means to control or cure HIV; specifically, DNA editing enzymes that disrupt CCR5 have been used to eliminate CCR5 expression in cells that are then expanded ex vivo and reinfused into patients, creating a population of HIV-resistant CD4+ T cells.^110,111

When AIDS was first identified as a clinical syndrome it was quickly appreciated that these patients were at greatly increased risk for certain types of malignancies, including Kaposi sarcoma, various types of non-Hodgkin lymphoma (NHL), and invasive cervical cancer. Each of these AIDS-defining cancers is frequently associated with an oncogenic virus (Table 81–6). As effective ART was developed and patients began living into their 50s, 60s, and 70s,^112,113 it became apparent that many non–AIDS-defining malignancies were also more common in this population compared to HIV-uninfected patients. Anal cancer is 120-fold more common in people living with HIV, particularly among men who have sex with men. Hodgkin lymphoma incidence is increased approximately 20-fold, hepatocellular cancer fivefold, and the risk of lung cancer is increased twofold. In contrast, the risks of other common cancers, including breast cancer, prostate cancer, and colon cancer, are not increased in comparison to HIV-negative people.¹¹⁴ In the ART era, non–AIDS-defining malignancies comprise approximately half of the cancers in people living with HIV, and overall cancer causes approximately 25 to 33 percent of all deaths in HIV-infected patients, supporting the importance of age-appropriate standard cancer screening.¹¹⁵

The Centers for Disease Control estimates that 20 percent of HIV+ people in the United States do not know that they are HIV+,¹¹⁶ and HIV testing is strongly recommended in all patients who present to the hematologist with NHL, Hodgkin lymphoma, or idiopathic thrombocytopenic purpura (ITP), or other malignancies.¹¹⁷ This recommendation is made because approximately 5 percent of those with diffuse large B-cell lymphoma and 22 percent of patients with Burkitt lymphoma in the United States are HIV+ (Fig. 81–1). These proportions vary substantially by demographic group¹¹⁸: among men, 10 percent of those with diffuse large B-cell lymphoma are HIV+, in contrast to 1 percent of women, and approximately 40 percent of those 30 to 59 years old with Burkitt lymphoma are HIV+. It is important to diagnose HIV infection when present, as effective treatment of HIV is essential for successful treatment of the malignancy or ITP.

HUMAN IMMUNODEFICIENCY VIRUS–ASSOCIATED DIFFUSE LARGE B-CELL LYMPHOMA

Among HIV+ patients in the United States, diffuse large B-cell lymphoma is now more common than Kaposi sarcoma, although Kaposi sarcoma remains the most common malignancy in people living with HIV worldwide.¹¹⁸ The pathophysiology of diffuse large B-cell lymphoma in HIV has been reviewed.^119,120 In a recent case series, HIV+ patients presented with diffuse large B-cell lymphoma at a median age of 43 years, 2 decades younger than HIV– patients.¹²¹ Patients often present with a rapidly growing lymph node or extranodal mass, and frequently have B symptoms (drenching night sweats, fever, or loss of 10 percent of body weight). Involvement of extranodal sites, including the gastrointestinal tract, liver, CNS, lung, and other sites is common.^121,122 Diagnosis is most commonly made by excisional lymph node biopsy. Evaluation should include careful examination of all lymph nodes sites, and the oral cavity. Standard staging with positron emission tomography–computed tomography (PET-CT), marrow evaluation, and lumbar puncture for cerebrospinal fluid cytology and flow cytometry¹²³ should be performed. Patients should be evaluated for hepatitis B prior to initiation of chemotherapy. If active hepatitis B is found (hepatitis B DNA+), it must be managed in the context of the HIV treatment, as several commonly used medications for hepatitis B are also active against HIV. Although initial studies in the pre-ART era focused on low-dose chemotherapy,¹²⁴ it is now appreciated that full-dose multiagent systemic chemotherapy with appropriate supportive care using filgrastim or peg-filgrastim and prophylaxis against infectious complications, offers the best chance for permanent cure. Cohort studies show that the 5-year overall survival in the ART era is far better than the pre-ART era.¹²⁵ A National Cancer Institute study using six cycles of dose-adjusted etoposide, prednisone, vincristine, cyclophosphamide, and doxorubicin (EPOCH), in which the initial cyclophosphamide dose was adjusted based on the CD4 count, and subsequent cycles cyclophosphamide dosing was adjusted based on the neutrophil nadir, showed an overall survival of 60 percent at 53 months (39 patients, 79 percent had diffuse large B-cell lymphoma, 18 percent had Burkitt lymphoma, and none were on ART during chemotherapy).¹²⁶ CD4 counts dropped by 190 cells/μL during chemotherapy, but recovered to baseline by 6 to 12 months following chemotherapy. All patients received P. jiroveci prophylaxis, and if CD4 counts were less than 100 cells/μL, M. avium complex prophylaxis. All patients also received filgrastim following each cycle of chemotherapy. This key study demonstrated that EPOCH is safe and effective in HIV+ patients with aggressive lymphoma. Outcomes differed markedly depending on the initial CD4 count: patients with an initial CD4 count greater than 100 cells/μL had an 87 percent overall survival at 53 months, whereas those with CD counts of less than 100 cells/μL had a 16 percent overall survival at 53 months.¹²⁶ A larger multiinstitution study done by the AIDS Malignancy Consortium (AMC-034) randomized patients to receive EPOCH with concurrent rituximab versus EPOCH with sequential rituximab (given weekly for 6 weeks following completion of EPOCH) in 101 patients with HIV-associated NHL (approximately 75 percent of the patients had diffuse large B-cell lymphoma and 25 percent had mainly Burkitt lymphoma).¹²⁷ Administration of EPOCH with concurrent rituximab resulted in a high complete response rate (73 percent) in comparison to EPOCH with sequential rituximab (55 percent complete response rate). The National Cancer Institute evaluated short-course EPOCH with dose-dense rituximab (rituximab day 1 and day 5 of each cycle of EPOCH), and achieved an overall survival of 68 percent at 5 years in 33 patients with diffuse large B-cell lymphoma.¹²⁸ In this study, the initial cyclophosphamide dose was 750 mg/m² (even in those patients with a low CD4 count) and was dose adjusted, depending on the neutrophil nadir, in subsequent cycles of EPOCH. Patients were treated to complete response plus one additional cycle. The majority (79 percent) of patients received three cycles of short-course EPOCH with dose-dense rituximab, and ART was suspended during chemotherapy because of concern about alteration in the pharmacokinetics or pharmacodynamics of the chemotherapy agents or overlapping toxicity. CD4 counts dropped a median of 64 cells/μL, and recovered to baseline by 6 to 12 months. Consistent with other studies, patients with initial CD4 counts of 100 cells/μL or greater had a much better 5-year overall survival (90 percent) than did the patients with a CD4 count of less than 100 cells/μL (20 percent). Cyclophosphamide, doxorubicin, vincristine, prednisone (CHOP) has also been studied in HIV+ patients with diffuse large B-cell lymphoma. In an AMC phase III multiinstitution clinical trial (AMC 010), patients with HIV-associated NHL (diffuse large B-cell lymphoma in 80 percent, Burkitt lymphoma in 9 percent) were randomized to six cycles of CHOP (n = 50) versus rituximab plus CHOP (R-CHOP) (n = 99), with all patients on ART.¹²⁹ The R-CHOP group also received three monthly doses of rituximab following completion of chemotherapy. Of note, the median CD4 count at enrollment was 133 cells/μL and 24 percent of the patients had advanced HIV with CD4 counts less than 50 cells/μL, so this was a fairly immunocompromised group of patients. Overall survival was identical with CHOP or R-CHOP, unlike HIV-negative patients in whom the addition of rituximab significantly improves outcome. In the AMC 010 clinical trial, there were significantly more deaths from infection in the R-CHOP arm in comparison to the CHOP arm, which offset the trend toward better control of lymphoma with R-CHOP. The majority of these infectious deaths occurred in patients with a CD4 count of less than 50 cells/μL, suggesting that rituximab should be used cautiously in immunologically vulnerable patients. As in studies of EPOCH, patients with a CD4 count greater than 100 cells/μL had a better overall survival than those with a lower CD4 count. Other reports suggest that R-CHOP treatment in HIV+ patients with diffuse large B-cell lymphoma is safe and effective,^121,130 including in those with a low CD4 count.¹³¹ In a phase II study of modified R-CHOP in 40 patients with diffuse large B-cell lymphoma, pegylated liposomal doxorubicin was substituted for doxorubicin¹³² and the complete response rate was 48 percent, lower than what is reported with rituximab plus EPOCH (R-EPOCH) or R-CHOP. At the time of this writing, there are no phase III data comparing R-EPOCH to R-CHOP in HIV+ patients. Pooled analysis of two sequential AMC clinical trials, AMC 010 (99 patients on the R-CHOP arm) and AMC 034 (51 patients on the concurrent R-EPOCH arm) showed that the 2-year overall survival was approximately 50 percent with R-CHOP and approximately 65 percent with R-EPOCH (p <0.01), suggesting superiority of R-EPOCH.¹³³ Similarly, a pooled analysis of 1546 patients enrolled in 19 prospective clinical trials, concluded that EPOCH was associated with a better overall survival than CHOP in patients with HIV-associated diffuse large B-cell lymphoma (hazard ratio 0.33, p = 0.03).¹³⁴ However these observations require validation in prospective randomized studies. For patients with relapsed or refractory HIV-associated diffuse large B-cell lymphoma, salvage regimens such as gemcitabine/dexamethasone/cisplatin¹³⁵ or etoposide, methylprednisolone, high-dose cytarabine, cisplatin (ESHAP)¹³⁶ can provide response rates of approximately 50 percent.

HUMAN IMMUNODEFICIENCY VIRUS–ASSOCIATED BURKITT LYMPHOMA

HIV-associated Burkitt lymphoma is approximately one-third as common as HIV-associated diffuse large B-cell lymphoma in the Western world, and occurs at a higher CD4 count.¹³⁷ In a Surveillance, Epidemiology, and End Results (SEER) Registry–based study, there was an increase in the number of cases of Burkitt lymphoma in the United States in the late 1980s that is maintained to the present time (see Fig. 81-1), particularly in men, and is thought to be attributable to the HIV epidemic.¹³⁸ The pathogenesis of HIV-associated Burkitt lymphoma is similar to that of Burkitt lymphoma in HIV– people, and involves translocation of the Myc gene on chromosome 8 with one of the immunoglobulin genes on chromosomes, 2, 14, or 22, resulting in overexpression of Myc.¹³⁹ HIV-associated Burkitt lymphoma is an aggressive malignancy, and it is important to act decisively in these often very ill patients. More than 80 percent of patients with HIV-associated Burkitt lymphoma present with stage IV disease¹⁴⁰ and extranodal sites are often involved. Marrow, liver, gastrointestinal tract, kidney, and CNS involvement are common, with cranial nerve palsies a common feature of CNS involvement.¹⁴¹ The serum lactate dehydrogenase (LDH) is elevated in more than 80 percent of patients, often to high levels (greater than fivefold normal). A number of chemotherapy regimens have been studied in HIV+ patients with Burkitt lymphoma. As in the HIV– setting, CHOP is not adequate treatment for Burkitt lymphoma,^140,142 and should not be used. Recent data show excellent outcomes with a variant of the R-EPOCH regimen.¹⁴³ In this single-institution, small prospective clinical trial, a total of 30 patients with Burkitt lymphoma were treated, including 11 who were HIV+ with a median CD4 count of 322 cells/μL. The short-course EPOCH-RR used in this clinical trial included two doses of rituximab per cycle of EPOCH, and a total of three or four cycles of EPOCH (to complete response plus one additional cycle), and included prophylactic intrathecal methotrexate. With a median follow up of 6 years, the overall survival of the HIV+ patients was 90 percent. The major toxicity was neutropenia in 31 percent of cycles of EPOCH-RR, and hospital admission for febrile neutropenia was required in 10 percent of cycles. This study showed that low-intensity therapy administered primarily in the outpatient setting can be effective for HIV-associated Burkitt lymphoma. Other studies of R-EPOCH that included a subset of patients with HIV-associated Burkitt lymphoma also report excellent outcomes with R-EPOCH.¹²⁷ Other regimens reported in patients with HIV-associated Burkitt lymphoma include cyclophosphamide, vincristine, doxorubicin, and dexamethasone (HyperCVAD) alternating with high-dose methotrexate plus cytarabine.¹⁴⁴ In this study, patients were very immunocompromised with a median CD4 count of 77 cells/μL; nevertheless, complete remission was achieved in more than 90 percent of patients, with 48 percent of patients alive at 2 years. Severe myelosuppression was universal, but infectious complications were similar to HIV– patients. In this small study, those on ART had a better outcome than those not on ART. Cyclophosphamide, vincristine, doxorubicin, methotrexate/ifosfamide, mesna, etoposide, cytarabine (CODOX-M/IVAC) has also been employed to treat HIV-associated Burkitt lymphoma,^145,146,147 with 3-year overall survival of approximately 50 percent.¹⁴⁶ A retrospective review compared eight HIV+ patients who received CODOX-M/IVAC to 24 HIV– patients with Burkitt lymphoma¹⁴⁵: Patients had similar rates of myelosuppression, infection, and complete response regardless of HIV status. The LMB86 protocol (including high-dose methotrexate plus cytarabine) was used to treat HIV-related Burkitt lymphoma¹⁴¹ in a prospective single center study of 63 patients on ART. The complete response rate was 70 percent and the estimated disease-free survival at 2 years was 67 percent. This regimen was characterized by severe marrow toxicity, and more than 10 percent of patients died of regimen-related toxicity. Poor prognostic factors included a CD4 count of less than 200 cells/μL and an Eastern Cooperative Oncology Group (ECOG) performance status of greater than 2. Other intensive regimens have also been used, with 4-year overall survival of 70 percent, but with death in 11 percent from regimen-related toxicity.¹⁴⁸

HUMAN IMMUNODEFICIENCY VIRUS–ASSOCIATED PRIMARY CENTRAL NERVOUS SYSTEM LYMPHOMA

Primary CNS lymphoma in HIV+ patients is an Epstein-Barr virus (EBV)–related diffuse large B-cell lymphoma occurring in the brain, typically in patients with very advanced HIV: These patients usually have a CD4 count of less than 50 cells/μL, and often of less than 20 cells/μL.^{137,149,150,151} The epidemiology of primary CNS lymphoma illustrates the concept of specific types of lymphoma occurring at different levels of immunodepletion. The incidence of primary CNS lymphoma has declined markedly since the availability of ART (see Fig. 81–1).^150,152 The pathophysiology of HIV-associated primary CNS lymphoma is related to EBV which is detectable in virtually all cases.¹³⁹ Primary CNS lymphoma should be considered in an HIV+ patient who presents with neurologic symptoms (confusion, cognitive decline, memory loss), headache, seizures, or ataxia. In one series, the most common symptom was headache, followed by memory loss, ataxia, and seizure.¹⁵³ Characteristic features on magnetic resonance imaging (MRI) of the brain include a single to several mass lesions in the subcortical white matter.¹⁵⁴ Anatomic sites commonly involved are predominantly the cerebral cortex and periventricular area, but the basal ganglia can be involved in up to one-third of cases. Cerebellar or brain stem involvement is rare.¹⁵⁰ A thorough physical exam for signs of systemic lymphoma is key, including a testicular exam as testicular lymphoma frequently involves the CNS. A slit-lamp exam to assess for vitreous disease should be done; this may assist in diagnoses of primary CNS lymphoma and may affect therapy. Evaluation with a chest, abdomen, and pelvic CT and a marrow aspirate and biopsy should be performed. If a lumbar puncture can be safely done, cerebrospinal fluid (CSF) should be sent for cytology and flow cytometry to evaluate for leptomeningeal involvement with lymphoma, and also for polymerase chain reaction (PCR) for EBV. Detection of EBV in the CSF supports, but does not confirm, the diagnosis of primary CNS lymphoma in these patients.¹⁵⁵ PET-CT of the brain can also help distinguish primary CNS lymphoma from other common causes of focal brain lesions in profoundly immunosuppressed patients with HIV, namely, cerebral toxoplasmosis and other infections.^156,157 Evaluation of HIV+ patients with CNS mass lesions should include blood serology for toxoplasmosis, although a small percentage of patients with CNS toxoplasmosis will have negative serologies. Stereotactic brain biopsy should be performed if possible, but some lesions are not readily accessible to biopsy; in these cases, the diagnosis of primary CNS lymphoma may rest on CSF cytology, detection of EBV in the CSF, and the results of PET-CT. Because of the rarity of primary CNS lymphoma in the ART era, there are no large prospective clinical trial data to define optimal therapy. Case reports document long-term responses to initiation of ART as a sole intervention in a small number of patients who refused other therapy. All HIV+ patients with primary CNS lymphoma should be on effective ART. Systemic glucocorticoid treatment can temporarily ameliorate symptoms. Small retrospective series report that whole-brain radiation therapy can result in improved survival,¹⁴⁹ but approximately one-third of these patients had detectable leukoencephalopathy on followup. A large retrospective study found that treatment with whole-brain radiation therapy and/or chemotherapy was associated with a decreased risk of death,¹⁵¹ but this analysis is confounded by lack of information on performance status. Small numbers of patients have been treated with multiple cycles of high-dose methotrexate with leucovorin rescue, without radiation therapy, with prolonged survival and no cognitive dysfunction,^158,159 and this may be a reasonable option in patients with good performance status. In the pre-ART era, median survival of the HIV+ patient with primary CNS lymphoma was approximately 2 months. The outcome has improved in the present era, but remains inferior to that of patients without HIV.¹⁵³ The Center for AIDS Research database for 1996 to 2010 shows that the 2-year survival of HIV+ patients with primary CNS lymphoma was 24 percent, inferior to other major types of HIV-associated lymphoma (diffuse large B-cell lymphoma, Burkitt lymphoma, Hodgkin lymphoma).¹³⁷ Prior CNS opportunistic infection¹⁵¹ and poor performance status^149,150 confer an increased risk of death.

HUMAN IMMUNODEFICIENCY VIRUS–ASSOCIATED PLASMABLASTIC LYMPHOMA

Described in 1997,¹⁶⁰ plasmablastic lymphoma is a rare and very aggressive B-cell NHL with plasmacytic differentiation that often involves the oral cavity, typically the gingiva and the palate. In the original report, 15 of the 16 cases were HIV+, and subsequent studies showed that plasmablastic lymphoma comprises approximately 2 to 3 percent of NHL in people living with HIV.¹⁶¹ A review of 112 cases of HIV-associated plasmablastic lymphoma showed that the median age of presentation was approximately 40 years, and the median CD4 count was approximately 180 cells/μL. Of the 112 patients, 58 percent had primary oral involvement. Other common sites of involvement were the gastrointestinal track, the lymph nodes and skin, among other sites.¹⁶² In a recent series of 50 cases of plasmablastic lymphoma,¹⁶³ approximately 25 percent of the patients had oral cavity involvement, and extraoral involvement was common. Diagnosis requires biopsy. The pathology shows a monomorphic diffuse lymphoid infiltrate with cells resembling plasmablasts. The cells have a high proliferative rate with a Ki-67 often exceeding 90 percent, and are positive for plasma cell markers. CD20 is expressed in 2 percent or fewer of the cases and the majority of the cases (>80 percent) are EBV+. The differential diagnosis of an oral cavity lesion in a person with HIV includes odontogenic infection, squamous carcinoma, Kaposi sarcoma, and diffuse large B-cell lymphoma or Burkitt lymphoma.

Many of the patients with plasmablastic lymphoma have been treated with CHOP or with EPOCH, with poor outcome. In one retrospective series,¹⁶³ median overall survival was 11 months with no difference in outcome between CHOP versus more intensive chemotherapy (EPOCH, hyperCVAD, or other regimens). Data from the German AIDS Related Lymphoma Cohort Study in the ART era confirmed the poor outcome of these patients, with a median survival of 5 months.¹⁶⁴ There are no prospective clinical trials to define optimal treatment for patients with HIV-associated plasmablastic lymphoma. Case reports of individual patients suggest that bortezomib may have activity in these patients, and this should be explored in future clinical trials.^165,166

HUMAN IMMUNODEFICIENCY VIRUS–ASSOCIATED PRIMARY EFFUSION LYMPHOMA

Primary effusion lymphoma is an aggressive B-cell lymphoma characterized by lymphomatous effusions in body cavities, most commonly pleural effusion,^167,168 followed by ascites and pericardial effusion or multiple body cavities; lymph nodes, marrow, and skin can also be involved. A solid variant of primary effusion lymphoma presents without effusion, but with lymph node, gastrointestinal, skin, or liver involvement has been reported.¹⁶⁹ Primary effusion lymphoma comprises approximately 4 percent of HIV-associated NHL¹⁷⁰ and occurs much more frequently in men than in women (10:1 ratio), usually associated with low CD4 counts (50 to 200 cells/μL).¹⁷¹ Of primary effusion lymphoma cases, 100 percent are human herpesvirus-8+ (HHV8+), and approximately 80 percent are EBV+. HHV8 plays a key pathophysiologic role, possibly by elaboration of a viral homologue of FLICE inhibitory protein and a viral homologue of interleukin (IL)-6.¹²⁰ Other HHV8-related disorders include Castleman disease and Kaposi sarcoma, both of which may coexist with primary effusion lymphoma in a substantial proportion of patients.¹⁶⁸ Patients may present with dyspnea from pleural effusions or new-onset ascites. A high index of suspicion for lymphoma is needed so that appropriate samples are sent to Hematopathology for analysis. Primary effusion lymphoma cells have an immunoblastic, plasmablastic, or anaplastic appearance and are CD45+ and CD30+; CD20 is expressed less than 5 percent of the time. The malignant cells are latently infected with HHV8, which is detectable by immunocytochemistry. There are no large prospective studies of treatment of primary effusion lymphoma, a consequence of its rarity, and the majority of the available information is derived from retrospective case series. There are a few case reports of complete remission following initiation of ART without chemotherapy, and ART should be a component of the treatment plan. Patients have been treated with CHOP, EPOCH, and other regimens. Approximately 50 percent of patients with primary effusion lymphoma achieve a complete response, but relapse within the next few months is common, and the median survival is approximately 6 months, with most deaths a result of progressive lymphoma. In one series, poor prognostic features included an ECOG performance status greater than 2 and ART noncompliance. Promising preclinical data show that treatment with the anti-CD30 agent brentuximab vedotin¹⁷² or bortezomib with or without vorinostat¹⁷³ can decrease growth of primary effusion lymphoma cell lines and prolong survival in a mouse xenograft model.

Prognosis

As ART improves, the prognosis for patients with HIV-associated NHL is defined mainly by lymphoma-related features, and less by HIV.¹²¹ A retrospective review of patients with HIV-associated diffuse large B-cell lymphoma diagnosed in the pre-ART era (120 patients), and in the ART era (72 patients) showed a median survival of 8 months in the pre-ART era and 43 months in the ART era; this held true for each of the different International Prognostic Index groups.¹²⁵ Pooled data for 1546 patients with HIV-associated diffuse large B-cell lymphoma or Burkitt lymphoma who had been enrolled in phase II or phase III clinical trials was evaluated to identify treatment-related factors associated with overall survival. The use of rituximab was significantly associated with improved overall survival (hazard ratio 0.55, p <0.001) for patients with a CD4 count of greater than 50 cells/μL but not for those patients with CD4 counts of less than 50 cells/μL. A focus on the 1059 patients with diffuse large B-cell lymphoma suggested that treatment with EPOCH resulted in a better overall survival (hazard ratio 0.33, p = 0.031) than treatment with CHOP. On multivariant analysis of R-EPOCH versus R-CHOP, the hazard ratio for overall survival was 0.34 favoring R-EPOCH, although this did not achieve statistical significance. An enhanced internal prognostic index based on 650 adults with de novo diffuse large B-cell lymphoma treated at seven National Comprehensive Cancer Network Cancer Centers in the rituximab era included a small portion of HIV+ patients.¹⁷⁴ Patients were risk stratified on an 8-point scale. Patients with a low score (0 to 1 points) had a 5-year overall survival of 96 percent; patients with a low intermediate score (2 to 3 points) had a 5-year overall survival of 82 percent; patients with a high intermediate score (4 to 5 points) had a 5-year overall survival of 64 percent, and patients with high risk (6 to 8 points) had an overall 5-year survival of 33 percent. This scale offered better risk stratification than the original International Prognostic Index, which was developed in the prerituximab era. It is recommended that this enhanced International Prognostic Index be used as a guide, in addition to the very robust data that patients with CD4 counts of less than 100 cells/μL at the time of diagnosis of lymphoma have a much worse outcome than those with higher CD4 counts.

HUMAN IMMUNODEFICIENCY VIRUS–ASSOCIATED HODGKIN LYMPHOMA

Hodgkin lymphoma tends to occur at moderate levels immunosuppression in HIV+ patients, unlike NHL, where the risk increases as the CD4 count decreases.¹⁷⁵ A retrospective cohort study from the Veterans Administration Clinical Case Registry from 1985 to 2010 showed that Hodgkin lymphoma was most common in patients with CD4 counts of 200 to 350 cells/μL. The risk was highest in the first year after starting ART, and was lower in people with a greater percent of time with an undetectable viral load.¹⁷⁶ Data from 14 U.S. Cancer Registries representing 25 percent of the U.S. population was used to compare the clinical features of HIV+ and HIV– patients with Hodgkin lymphoma in the ART era.¹⁷⁷ In this study, Hodgkin lymphoma occurring in HIV+ people was shown to be a clinically aggressive disease. Of the 22,355 patients with Hodgkin lymphoma, 3.8 percent were HIV+. However this percentage varied depending on sex, ethnicity, and age. Prevalent HIV infection was higher in men (6 percent) than in women (1.2 percent), and among men in the 40- to 59-year-old age group, those newly diagnosed with Hodgkin lymphoma had a 14.2 percent chance of being HIV+. Non-Hispanic blacks with newly diagnosed Hodgkin lymphoma had a 16.9 percent chance of being HIV+ and Hispanics with newly diagnosed Hodgkin lymphoma had a 9.9 percent chance of being HIV+. Unlike NHL, the incidence of Hodgkin lymphoma is comparable in the pre- and post-ART era. The pathology of Hodgkin lymphoma in HIV+ cases differs from that of HIV– cases, with a higher percent of HIV+ patients having a more aggressive histology (mixed cellularity or lymphocyte depleted; Table 81–7). An HIV-AIDS cancer match study that linked HIV and cancer registry data found that an even higher percent of patients with HIV-associated Hodgkin lymphoma had mixed cellularity on biopsy (53.7 percent).¹⁷⁵ The Ann Arbor stage at diagnosis is higher in HIV+ Hodgkin patients than in HIV– cases, with 41.5 percent of those with HIV+ Hodgkin lymphoma having stage IV disease at presentation, compared to 17 percent of those with HIV– Hodgkin lymphoma (Table 81–7). B symptoms (drenching night sweats, fever, or loss of 10 percent of body weight) are also more common in HIV+ patients with Hodgkin lymphoma (Table 81–7).¹⁷⁷

A retrospective study of Adriamycin, bleomycin, vinblastine, and dacarbazine (ABVD) chemotherapy in 62 HIV+ patients newly diagnosed with advanced-stage Hodgkin lymphoma showed that ABVD and ART could be given safely together.¹⁷⁸ In this study, the median CD4 count at diagnosis was 129 cells/μL, and all patients had stage III or stage IV Hodgkin lymphoma. Patients received ABVD with filgrastim support, as well as trimethoprim-sulfamethoxazole or pentamidine for P. jiroveci prophylaxis. The overall survival was 76 percent at 5 years, with treatment-related mortality of 10 percent. In a large retrospective study¹⁷⁹ of 93 HIV+ patients and 131 HIV– patients with Hodgkin lymphoma who were treated with ABVD, (HIV+ patients also received concomitant ART), those with stage I or II nonbulky disease received four cycles of ABVD plus involved field radiation therapy; the rest received six cycles of ABVD with involved field radiation therapy if bulky disease was present. All patients received prophylaxis for opportunistic infections. The HIV+ patients had more advanced-stage Hodgkin lymphoma and a worse International Prognostic Score¹⁸⁰; despite this the 5-year overall survival was 81 percent for the HIV+ patients compared to 88 percent for the HIV– patients, a nonsignificant difference. This retrospective case series demonstrated comparable overall survival in HIV+ and HIV– patients with Hodgkin lymphoma, and that ART could be given safely with ABVD chemotherapy.

The German HIV Related Lymphoma Study Group¹⁸¹ conducted a prospective multicenter study in which HIV+ patients with early stage favorable Hodgkin lymphoma were treated with two to four cycles of ABVD plus 30 Gy of involved field radiation therapy, while patients with early stage unfavorable Hodgkin lymphoma received four cycles of bleomycin, etoposide, Adriamycin, cyclophosphamide, vincristine, procarbazine, and prednisone (BEACOPP) baseline or four cycles of ABVD plus 30 Gy of involved field radiation therapy. More advanced-stage Hodgkin lymphoma patients received six to eight cycles of BEACOPP. Patients with advanced HIV infection (defined as two of the following: CD4 count <50 cells/μL, prior AIDS-defining opportunistic infection, performance status >2) were treated with six to eight cycles of ABVD. ART was used with chemotherapy. CD4 count decreased with chemotherapy, but recovered over the subsequent 6 to 9 months and treatment-related mortality was 5.6 percent, mainly from sepsis. Using this risk-adapted approach, the 2-year overall survival was 90.7 percent, comparable to HIV– patients with Hodgkin lymphoma.

Areas of controversy in treatment of Hodgkin lymphoma in HIV– patients include whether BEACOPP escalated is better than ABVD for advanced Hodgkin lymphoma.^182,183,184 Until this controversy is settled, most authorities recommend using ABVD for patients with HIV-associated Hodgkin lymphoma. Brentuximab vedotin, an antibody–drug conjugate that targets CD30, has been combined with Adriamycin, vinblastine, and dacarbazine (AVD) to treat HIV- patients with Hodgkin lymphoma.¹⁸⁵ Whether brentuximab vedotin plus AVD will be effective for HIV+ patients with Hodgkin lymphoma is being addressed in a prospective phase II clinical trial being done by the AMC (AMC 085).

In the ART era, HIV+ patients with Hodgkin lymphoma have similar excellent outcomes with systemic chemotherapy as do HIV– patients. This is corroborated by the SEER Study¹⁷⁷ in which the 5-year risk of death from Hodgkin lymphoma was 6.2 percent in the HIV+ patients and 9 percent in the HIV– patients. However, this interpretation is confounded by competing risks of death: the overall risk of death was higher in the HIV+ cohort, and the majority of the deaths were from HIV. Many¹⁷⁹ but not all¹⁸¹ studies conclude that the International Prognostic Score has predictive value for HIV+ patients with Hodgkin lymphoma.

Stem Cell Transplant

There is more than a decade of experience in high-dose therapy followed by autologous blood stem cell transplantation showing that this technique is feasible, safe, and effective in patients with HIV and NHL or Hodgkin lymphoma.^186,187 Adequate numbers of blood stem cells can be mobilized in the majority of patients with HIV-associated NHL or Hodgkin lymphoma. A retrospective study of 155 patients with either NHL or Hodgkin lymphoma in the ART era showed that mobilization of greater than 2 × 10⁶ CD34+ cells/kg was achieved in 73 percent of patients and greater than 5 × 10⁶ CD34+ cells/kg was achieved in 48 percent. Factors reducing the chance of optimal mobilization included a low platelet count, a low CD4 count, and use of filgrastim alone, rather than chemotherapy plus filgrastim, to mobilize blood stem cells.¹⁸⁸

Studies of autologous blood stem cell transplantation in HIV+ patients with NHL (mainly diffuse large B-cell lymphoma) or Hodgkin lymphoma show that outcomes are similar to those reported for patients without HIV infection.^189,190 The median time to neutrophil and platelet engraftment is 11 and 14 days, respectively. The risk for nonrelapse mortality is in the 5 to 8 percent range and the overall survival at 3 to 4 years after transplantation is approximately 50 percent.^191,192,193 These patients were supported throughout the transplantation with ART, as well as with antibacterial, antifungal, and antiviral prophylaxis. The European group for Blood and Marrow Transplant performed a case-control study that included 53 HIV+ patients with either NHL (two-thirds of patients) or Hodgkin lymphoma (one-third of patients) that matched patients on the basis of histology, Ann Arbor stage, International Prognostic Index and disease status. The overall survival, progression-free survival, and relapse rates were similar in the HIV+ and the HIV– groups across all histologies and disease states. Nonrelapse mortality at 1 year was 8 percent in the HIV+ group and 2 percent in the HIV– group; the difference mainly was a result of early bacterial infections. Overall survival at 30 months was 59 percent for both HIV+ and HIV– patients with NHL, and the main cause of death in both HIV+ and HIV– patients was relapse. A single-institution matched case-control study of HIV+ and HIV– NHL patients showed nonrelapse mortality of 11 percent in the HIV+ group, and 4 percent in the HIV– group. Overall survival at 2 years was 75 percent for both the HIV+ and HIV– groups.¹⁹⁰ Thus, HIV status did not affect the outcome of autologous transplantation for NHL although there were more viral opportunistic infections, particularly cytomegalovirus (CMV), adenovirus, and varicella in the HIV+ population. The main cause of death was relapse and the major predictor of outcome in both groups was disease status at time of transplantation. Together these data suggest that long-term outcomes are similar in the HIV+ and HIV– patients who undergo autologous blood stem cell transplant in the setting of NHL or Hodgkin lymphoma and the reconstitution of neutrophils and platelets are similar in both groups.¹⁹⁴ Experience with allogeneic stem cell transplantation HIV+ patients is less extensive.^195,196,197 This remains an area of intensive research, in part because of the report of cure of HIV in a patient who underwent allogeneic transplant to treat acute myeloid leukemia and received a donor graft that was homozygous for a deletion mutation eliminating the chemokine and HIV coreceptor CCR5.¹⁰⁹

Antiretroviral Therapy and Chemotherapy

Controversy persists regarding whether ART should be initiated or continued during multiagent systemic chemotherapy. In the pre-ART era, the outcomes for treatment of lymphoma in HIV+ patients were markedly inferior to the present era.¹²⁵ Some clinical trial groups stop or do not initiate ART during systemic chemotherapy,¹⁹⁸ whereas others continue ART or initiate ART after the first cycle of chemotherapy in those patients not on ART at time of lymphoma diagnosis.^199,200 Certain classes of ART medications, notably protease inhibitors and nnRTIs can alter the metabolism of other medications primarily though inhibition or induction of the cytochrome P450 enzymes, potentially leading to higher or lower concentrations of other drugs (including chemotherapy agents).²⁰¹ Unfortunately, there are limited pharmacokinetic data from clinical trials to elucidate these potential drug interactions and their effects on drug levels. Chemotherapy agents eliminated by non-P450 routes are less likely to be affected by ART. A study of 154 HIV+ patients with a broad range of malignancies (many with hematologic malignancies) treated at the MD Anderson Cancer Center showed that clinically relevant drug interactions between ART and chemotherapy drugs were rare (4 percent of cases) and confined to the protease inhibitor class.²⁰² The nnRTIs and the integrase strand-transfer inhibitors had a better safety profile when combined with chemotherapy than did the protease inhibitors, in this retrospective series. Others have also found that protease inhibitors may exacerbate chemotherapy-induced neutropenia,²⁰³ although this is controversial.^204,205 A retrospective review of 34 patients with diffuse large B-cell lymphoma showed similar complete response rates and toxicities in those receiving CHOP while also on an HIV protease inhibitor as those who were receiving a nonprotease inhibitor ART regimen.²⁰⁶ In another study of HIV+ patients with Hodgkin lymphoma receiving ABVD chemotherapy, the use of the protease inhibitor ritonavir appeared to increase the risk of peripheral neuropathy.²⁰⁷ More accurate information regarding drug interactions between ART and chemotherapy agents will become available through the AMC, which is currently conducting clinical trials in which patients are stratified by the type of ART and its potential for inducing or inhibiting cytochrome P450 enzymes, and obtaining detailed pharmacokinetic data. With the growing number of available ART agents, including integrase strand-transfer inhibitors, it is usually possible to adjust an ART regimen that minimizes potential drug interactions but still suppresses HIV replication throughout the course of chemotherapy.

A second issue is overlapping toxic effects between ART and chemotherapy drugs. Zidovudine causes marrow suppression, and should be avoided in patients receiving myelosuppressive chemotherapy. Didanosine and stavudine (older, infrequently used agents) cause peripheral neuropathy, and their use should be reconsidered in patients who will receive potentially neurotoxic chemotherapy. Atazanavir, a commonly prescribed protease inhibitor, causes unconjugated hyperbilirubinemia similar to Gilbert syndrome and can complicate the administration of some chemotherapy agents. However, if there is no other evidence of hepatotoxicity, chemotherapy can be given at standard doses. Earlier studies suggested some ART drugs (saquinavir, efavirenz, nelfinavir) might cause cardiac conduction abnormalities (prolongation of the electrocardiographic “QTc interval”), but this has not been confirmed in subsequent studies. If the patient is stable on ART with excellent HIV virologic control, a reasonable approach is to consult the infectious diseases physician who is managing the patient and also the chemotherapy pharmacist to identify potential drug interactions between ART and the planned chemotherapy, to adjust ART as necessary, and then proceed with chemotherapy with concurrent ART. If the patient is not on ART, it is reasonable to give one cycle of chemotherapy and then start ART with the second cycle of chemotherapy following multidisciplinary consultation with the pharmacist and infectious diseases physician.

HUMAN IMMUNODEFICIENCY VIRUS–ASSOCIATED CASTLEMAN DISEASE

Castleman disease is a rare polyclonal lymphoproliferative disorder characterized by periodic flares (an inflammatory illness accompanied by lymphadenopathy and splenomegaly), and a high risk of progression to lymphoma.²⁰⁸ Castleman disease occurs with higher frequency in people living with HIV.²⁰⁹ The pathophysiology of Castleman disease is related to IL-6 (reviewed in Ref. 210). Virtually all patients with HIV-associated Castleman disease are infected with HHV8, whose viral genome encodes a viral homologue of IL-6. Sequential analysis of human IL-6, viral IL-6, and a variety of other cytokines in patients with HIV-associated Castleman disease during and after flares showed that either human IL-6, viral IL-6, or both were elevated during the flare and diminished during remission.²¹¹ IL-6 is a proinflammatory cytokine and stimulates polyclonal proliferation of B lymphocytes. Serologic evidence of HHV8 infection is present in 7 percent of healthy U.S. blood donors,²¹² but few healthy adults or HIV+ patients without Castleman disease, Kaposi sarcoma, or primary effusion lymphoma have detectable HHV8 virus in the blood. More than 80 percent of HIV+ patients with symptomatic multicentric Castleman disease have detectable HHV8 viremia, often at a high level.²⁰⁹ A retrospective analysis of 52 HIV+ patients with multicentric Castleman disease at time of presentation²¹³ showed that more than 50 percent of the patients had a prior diagnosis of Kaposi sarcoma, and 80 percent were on ART. The mean CD4 count at time of diagnosis of Castleman disease was 287 cells/μL, indicating that Castleman disease occurs in moderately immunocompromised patients. Symptoms at presentation included fever, lymphadenopathy, and splenomegaly in 98 to 100 percent, pulmonary symptoms in 60 percent, and edema in 40 percent of patients. A single-institution study of 113 patients with Castleman disease²¹⁴ also reported that approximately half of the patients had a prior diagnosis of Kaposi sarcoma. The median CD4 count in this study was 188 cells/μL. The ECOG performance status at diagnosis of Castleman disease was greater than 2 in 45 percent of patients, 36 percent had hemophagocytosis, and 30 percent required an ICU stay, illustrating how ill these patients often are. During a Castleman disease flare, patients may rapidly develop cytopenias: half the patients in this series had a hemoglobin of less than 8 g/dL and 29 percent had a platelet count of less than 150,000/μL. High C-reactive protein (often 10-fold higher than normal) and plasma HHV8 levels (median 30,000 copies/mL, range: 60 to 1 million copies/mL) are found during a flare,^215,216 and these values can be helpful. Diagnosis requires a lymph node biopsy that is consistent with Castleman disease, ideally an excisional biopsy rather than a needle biopsy. The morphology of multicentric Castleman disease in HIV+ patients shows plasmablastic cells in the mantle zone of the follicles; HHV8 is detectable in the plasmablasts.²¹⁷ Often, lymph nodes diagnostic of Castleman disease also contain malignant appearing spindle cells characteristic of Kaposi sarcoma.

Castleman disease is rare, and treatment recommendations are largely based on case series and small clinical trials. Several clinical trials show that the anti-CD20 monoclonal antibody rituximab is very effective for treatment of Castleman disease. Prospective studies^213,218,219 demonstrate that 375 mg/m² of rituximab administered intravenously weekly for 4 doses resulted in rapid symptomatic improvement, and 90 percent of patients achieved a response that lasted for more than 6 months. Clinical improvement with rituximab may begin within hours to days and is associated with diminished levels of serum IL-6.²²⁰ In one series of 113 HIV+ patients with multicentric Castleman disease, some of whom received rituximab, the overall median survival was longer than 12 years with therapy; this is much improved in comparison to 20 years ago when Castleman disease was rapidly fatal.

Incipient flares of Castleman disease may be heralded by cytopenias, increased C-reactive protein, or by increased levels of HHV8 in the blood.²¹⁶ Serial studies of 52 HIV+ patients with multicentric Castleman disease showed that approximately 75 percent had elevated C-reactive protein and 90 percent had detectable plasma HHV8 during a flare, with a median HHV8 viral load of 27,000 copies/mL. HHV8 was detectable in 38 percent of patients during remission. If relapse occurs, patients can safely be retreated with rituximab.²²¹ Up to 20 percent of HIV+ patients with Castleman disease will develop lymphoma,^208,214 often primary effusion lymphoma, plasmablastic lymphoma, or diffuse large B-cell lymphoma, but occasionally Hodgkin lymphoma, so it is important to consider this during followup. Use of rituximab to treat Castleman disease may decrease the risk of subsequent NHL.²¹⁴ In HIV+ patients with both Castleman disease and Kaposi sarcoma, Kaposi sarcoma may progress following treatment with rituximab.^218,219,222 When Kaposi sarcoma involves a key clinical location such as the lungs, where progression might result in significant respiratory compromise, a regimen of rituximab plus liposomal doxorubicin can be considered to address both the Castleman disease flare and Kaposi sarcoma.²²³ R-CHOP or other multiagent systemic chemotherapy has also been used to treat Castleman disease, but given the excellent response to rituximab alone, R-CHOP is usually not necessary. An alternative therapeutic approach focuses on the central role of IL-6. A multiinstitution, randomized, placebo-controlled clinical trial of an anti–IL-6 monoclonal antibody siltuximab in 79 HIV– patients with Castleman disease showed a 34 percent response rate.²²⁴ Siltuximab was given IV every 3 weeks and can be administered long-term (up to 60 months).²²⁵ Siltuximab is FDA-approved for treatment of Castleman disease in HIV– patients but not in HIV+ patients. Tocilizumab, an anti–IL-6 receptor monoclonal antibody has been FDA-approved for use in rheumatoid arthritis. Tocilizumab had short-term effects in two HIV+ patients with Castleman disease, but both patients relapsed within 6 months and required subsequent treatment with rituximab.²²⁶ Tocilizumab is not FDA-approved for treatment of Castleman disease. Another approach to treatment of Castleman disease is to use antiviral therapy, such as ganciclovir, valganciclovir, foscarnet, cidofovir, or zidovudine, to suppress HHV8 replication.^227,228 When 14 HIV+ patients with Castleman disease were treated with high-dose zidovudine plus valganciclovir, the C-reactive protein level, human IL-6 levels and HHV8 viral load improved. The median progression-free period was 6 months and the major toxicity was marrow suppression. However, until additional information is available from clinical trials, the data at this time most strongly support the use of rituximab as the backbone of treatment of Castleman disease.

KAPOSI SARCOMA–ASSOCIATED HERPESVIRUS–ASSOCIATED INFLAMMATORY CYTOKINE SYNDROME

A new syndrome termed Kaposi sarcoma-associated herpesvirus (KSHV)-associated inflammatory cytokine syndrome (KICS) is characterized by an inflammatory illness similar to a flare of Castleman disease, but without the pathologic diagnosis of Castleman disease.^229,230,231 Patients present with fever, sweating, anorexia, leukopenia, anemia, thrombocytopenia, hypoalbuminemia, and hyponatremia, and may also have dyspnea or abdominal pain. Four of the six patients in the original report also had severe manifestations of Kaposi sarcoma. Patients with KICS have a very high HHV8 viral load, similar to that seen in patients during a flare of multicentric Castleman disease. Levels of other cytokines, including human IL-6, viral IL-6, and IL-10, are similar in KICS and multicentric Castleman disease patients, and much higher than in patients with isolated Kaposi sarcoma. Whether KICS represents a form fruste of multicentric Castleman disease, part of the clinical spectrum of Castleman disease, or a distinct clinical syndrome requires more investigation. There is scant information on treatment of KICS.

HUMAN IMMUNODEFICIENCY VIRUS–ASSOCIATED HEMOPHAGOCYTIC SYNDROME

Hemophagocytic syndrome is a rare and potentially fatal disorder characterized by excess immune activation that occurs with increased frequency in people living with HIV. An autopsy study in the pre-ART era identified hemophagocytosis in 20 percent of 56 patients who died of AIDS.²³² The pathogenesis of the hemophagocytic syndrome results from failure to regulate the immune response, resulting in excess activation of T lymphocytes, increased cytokine secretion, and hyperactivation of macrophages. There are primary and secondary forms of hemophagocytic syndrome. Secondary forms can occur in the setting of infection, hematologic malignancy, or autoimmune disease.²³³ In HIV+ patients, the hemophagocytic syndrome is usually secondary to a viral infection, either HIV itself or, more commonly, a member of the herpes virus family (EBV, CMV, HHV8), but can also occur in patients with Hodgkin lymphoma, NHL, or Castleman disease.²³⁴ A study of 58 HIV+ patients with hemophagocytic syndrome performed in the ART era²³⁴ showed that the median duration of HIV at the time of diagnosis was 4 years. Most patients were on ART, and the median CD4 count was 91 cells/μL, demonstrating advanced HIV disease. Patients with hemophagocytic syndrome are typically very ill: greater than 40 percent of the patients in this series required ICU care and 31 percent died within a year of diagnosis. The clinical features of hemophagocytic syndrome in HIV+ patients include fever, hepatosplenomegaly, and cytopenias. Patients often have a markedly elevated ferritin (greater than 10-fold the upper range of normal), coagulopathy, and increased triglycerides.²³⁴

The differential diagnosis of hemophagocytic syndrome in HIV+ patients includes a Castleman disease flare or KICS, and the clinical features of these syndromes may overlap. Table 81–8 shows the diagnostic criteria for hemophagocytic syndrome. If an underlying trigger for the hemophagocytic syndrome is identified, specific therapy should be directed at the trigger. Treatment for hemophagocytic syndrome is standardized using the Histiocytosis Society 1994 protocol²³⁵ or the Histiocytosis Society 2004 protocol²³⁶ (Chapter 71). Caution regarding interaction of cyclosporine with specific ART medications is warranted.

ANEMIA AND HUMAN IMMUNODEFICIENCY VIRUS

Anemia is common in people living with HIV. The multisite Adult and Adolescent Spectrum of HIV Disease Surveillance Project²³⁷ included approximately 32,000 HIV+ people living in the United States in the pre-ART era, 1990 to 1996. In this study, the first hemoglobin measured was less than 10 g/dL in 37 percent of men and 43 percent of women with clinical AIDS (an AIDS-defining illness). Even patients with a CD4 count of greater than 200 cells/μL and no AIDS-defining illness had a high prevalence of anemia: 28 percent of men and 31 percent of women had anemia defined as hemoglobin less than 14 g/dL for men or less than 12 g/dL for women. At the 1-year followup, the incidence of anemia was 3.2 percent for HIV+ people without immunologic or clinical AIDS, 12.1 percent for those with immunologic AIDS, and 36.9 percent for those with clinical AIDS. Notably, anemia was associated with a 1.5- to 2.5-fold increased risk of death during followup.²³⁷ One prospective study showed a moderate correlation between the hemoglobin and CD4 count at entry into the cohort, and found that initiation of ART could increase the hemoglobin value.²³⁸ Similarly, interruption of ART increased the risk of anemia.²³⁹ An investigation of 2056 HIV+ women at six sites in the United States showed that use of ART for as little as 6 months was associated with improvement of anemia,²⁴⁰ while several factors, including CD4 count less than 200 cells/μL, HIV viral load greater than 50,000 copies/mL, and mean corpuscular volume less than 80 fL were associated with decreased ability to correct the anemia. Anemia was identified as a key prognostic factor in large European and South African HIV+ cohorts.^241,242 The Veterans Aging Cohort Study evaluated HIV+ patients who had been on ART for 1 year, to identify factors predictive of mortality.²⁴³ This study identified age, CD4 count, HIV viral load, hemoglobin, glomerular filtration rate, presence or absence of hepatitis C, and a composite measure of liver function as predictive factors for mortality. Even a mild decrease in the hemoglobin contributed significantly to mortality in this index. The Veterans Aging Cohort Study Index was subsequently validated in an independent cohort.²⁴⁴ Thus, even in the ART era, anemia is associated with a worse prognosis, independent of the traditional risk markers such as the CD4 count and the HIV viral load.

Pathophysiology of Anemia in Human Immunodeficiency Virus

There are many pathophysiologic causes of anemia in HIV+ people (Table 81–9), and often anemia in an individual is multifactorial in origin. Unique to HIV are the effects of the virus on hematopoiesis, including direct infection of hematopoietic progenitor cells (impairing their survival, proliferation, differentiation, and maturation), infection of marrow stromal cells (affecting their ability to support hematopoiesis in the marrow microenvironment²⁴⁵), and alterations in hematopoietic growth factor production or function. Additionally, HIV infection is associated with a chronic inflammatory state³⁸ that may also contribute to ineffective erythropoiesis. Markers of inflammation, including IL-6, are higher in HIV+ individuals than in their HIV– counterparts, and this holds true even for those on ART with an undetectable viral load.⁶⁹ Inflammation, in part mediated by IL-6, upregulates hepcidin, a key regulator of iron trafficking, and serum hepcidin levels are inversely correlated with CD4 counts.^246,247

Attempts to infect normal hematopoietic progenitor cells in vitro with HIV and studies of hematopoietic progenitor cells obtained from HIV+ patients initially suggested that HIV infection of hematopoietic progenitor cells was an infrequent event. However, subsequent studies of HIV subgroup C (the subtype that predominates in African populations) showed that a proportion of burst-forming unit–erythroid (BFU-E), granulocyte-macrophage colony-forming units (CFU-GM), and granulocyte-erythrocyte-monocyte and megakaryocyte colony-forming units (CFU-GEMM) can be infected with HIV in vitro.²⁴⁸ The proportion of hematopoietic progenitor cells that could be infected with HIV subtype B (the predominant subtype in the United States and in Europe) was smaller (in the 1 percent range). Proviral HIV was detected by PCR in a small fraction of isolated CD34+ cells from some but not all HIV+ patients on ART with an undetectable viral load,²⁴⁹ suggesting that HIV can latently infect hematopoietic progenitor cells in vivo; similar results were obtained using CD133+ sorted marrow cells.²⁵⁰ Recent studies using a specific strain of humanized mice showed that some human CD34+, CD38+ intermediate progenitor cells could be infected with HIV in vivo, and that these progenitor cells when cultured in vitro had impaired growth, particularly in the erythroid and megakaryocytic lineages.²⁵¹

The normal response of the kidneys to a decline in hemoglobin is to produce more erythropoietin. However in HIV+ people, the incremental increase in erythropoietin is blunted.^252,253 Additionally, antierythropoietin antibodies can be detected in a portion of HIV+ patients and are associated with increased risk of anemia.²⁵⁴ The antierythropoietin antibodies recognize a peptide that is needed for erythropoietin binding to the erythropoietin receptor. This peptide has sequence homology to an HIV protein, suggesting molecular mimicry as a potential mechanism for development of the antierythropoietin antibodies.²⁵⁵

A cross-sectional study of 200 HIV+ patients in San Francisco during the ART era suggested an association between low testosterone levels and anemia in men living with HIV.²⁵⁶ Hypogonadism is associated with weight loss, osteoporosis, and AIDS wasting syndrome, but whether testosterone treatment can correct anemia in these patients awaits further study.

Another, occasionally dramatic, cause of anemia in those living with HIV is pure red cell aplasia caused by parvovirus B19. Parvovirus B19 is a small DNA virus that infects and lyses late human erythroid precursor cells,²⁵⁷ causing transient reticulocytopenia in immunocompetent patients. The clinical spectrum of parvovirus B19–related disease includes “slapped cheek rash” in children and arthralgias in adults, but most immunocompetent patients are asymptomatic. Patients with hemolytic anemia who depend on high production of reticulocytes may develop transient aplastic crisis following infection with parvovirus B19. Immunoincompetent individuals may not be able to mount an adequate antibody response to clear parvovirus B19 from the blood and marrow, resulting in sustained infection with reticulocytopenia and anemia. In a series of HIV+ patients with parvovirus B19–related anemia, the median CD4 count was 42 cells/μL, indicating that parvovirus B19–associated pure red cell aplasia is associated with advanced immunodeficiency.²⁵⁸ Characteristically, these patients have a normocytic anemia with virtual absence of reticulocytes, but normal white blood cells and platelets. Parvovirus immunoglobulin (Ig) G and IgM levels are generally not helpful in making the diagnosis, but PCR to detect parvovirus in the blood can be a reliable diagnostic test.²⁵⁹ Marrow evaluation may show characteristic giant proerythroblasts and markedly diminished late erythroid precursor cells, with full maturation of the myeloid and megakaryocytic lineages. Some patients respond to initiation of ART in addition to transfusion support, but the majority of patients require additional therapy. A high prevalence of antibodies to B19 parvovirus exists in the general population; therefore, pooled human IgG preparations are a rich source of antiparvovirus antibody. Infusion of human IgG results in clearance of parvovirus B19 from the blood and improved reticulocyte production and resolution of anemia in HIV+ patients with parvovirus B19–induced pure red cell aplasia.^259,260 However, patients with a CD4 count of less than 80 cells/μL are prone to relapse within 6 months and may require retreatment.

Medications and Anemia

An important cause of anemia in people living with HIV is medications; both ART and medications used for prophylaxis against infections are implicated. Zidovudine can cause macrocytic anemia and leukopenia, and is not commonly used at present because of its toxicity. Stavudine is associated with anemia, although at lower rates than zidovudine.²⁶¹ Dapsone or trimethoprim-sulfamethoxazole, used for P. jiroveci prophylaxis, can cause hemolytic anemia in patients with glucose-6-phosphate dehydrogenase (G6PD) deficiency. Ganciclovir or valganciclovir, used to treat herpes simplex virus (HSV), Varicella-zoster, CMV, and sometimes HHV8, often cause pancytopenia. Trimethoprim-sulfamethoxazole can also cause myelosuppression and pancytopenia.

Treatment of Anemia

Initiation of ART is the optimal treatment for HIV-associated anemia. If a specific cause of anemia is identified, such as parvovirus B19, specific treatment can be provided. Early clinical trials demonstrated that erythropoietin treatment in anemic HIV+ patients taking zidovudine can improve the hemoglobin and decrease red blood cell transfusion requirements,²⁶² although responses were confined to patients with an endogenous erythropoietin level of 500 mU/mL or less.^262,263 Since zidovudine is no longer commonly used as a component of ART, the use of erythropoietin has declined. Nevertheless, treatment of anemia caused by zidovudine in patients with HIV and an erythropoietin level of 500 mU/mL or less is an FDA-approved indication for erythropoietin. However, a Cochrane analysis²⁶⁴ that included six studies and more than 500 anemic HIV+ patients treated with erythropoietin or placebo concluded that erythropoietin did not reduce the risk of death and did not have a clear effect on quality of life. Additionally, erythropoietin has side effects including the potential to exacerbate hypertension, increased risk of thrombosis, a very rare risk of pure red cell aplasia, and increased risk of tumor progression or recurrence in specific types of cancer, as well as high cost.

HUMAN IMMUNODEFICIENCY VIRUS–ASSOCIATED THROMBOTIC MICROANGIOPATHY

Early in the HIV epidemic, it was reported that HIV+ people are at increased risk of thrombotic microangiopathy, either thrombotic thrombocytopenic purpura (TTP) or hemolytic-uremic syndrome (HUS). In a retrospective survey of 1223 patients with clinical AIDS in the pre-ART era, 1.4 percent met criteria for thrombotic microangiopathy. However, a prospective cohort of 347 patients followed from 1997 to 2000 identified no patients with thrombotic microangiopathy, suggesting that the incidence may have declined as a result of effective ART.²⁶⁵ In support of this, a study done in the ART era followed 6022 HIV+ patients and found that 0.3 percent of patients developed thrombotic microangiopathy.²⁶⁶ The group with thrombotic microangiopathy had a lower mean CD4 count, a higher plasma viral load, and a higher incidence of clinical AIDS than did HIV+ patients without microangiopathy. Registries of patients with TTP have been used to investigate what fraction of the patients are HIV+; of 362 patients in the Oklahoma TTP/HUS registry from 1989 to 2007,1.84 percent were HIV+ while the prevalence of HIV in the local community was 0.3 percent.²⁶⁷ In this study, several of the HIV+ patients were thought to have alternative causes of their microangiopathic hemolytic anemia, reinforcing the importance of a thorough consideration of the full spectrum of illnesses causing thrombotic microangiopathy, which may, in turn, affect the decision to use plasmapheresis/plasma exchange.²⁶⁷ One series of 24 HIV+ patients with TTP reported that these patients presented with a mean hemoglobin of 6.8 g/dL, a mean platelet count of 20,000/μL, a mean LDH of 4.5-fold normal, and a mean CD4 count of 236 cells/μL, and the majority had an HIV viral load of greater than 10,000 copies/mL. Patients were treated with prompt plasma exchange (mean of 13 treatments) and initiation of ART and achieved excellent outcomes comparable to HIV– patients with TTP, with death occurring in a single patient. However six of the 24 patients suffered relapse of TTP, often in the setting of increased viral load, and required retreatment.²⁶⁸ A study of patients enrolled in the French thrombotic microangiopathy (TMA) network revealed that 29 of 236 patients were HIV+. Of the HIV+ patients who had a disintegrin and metalloproteinase with a thrombospondin type 1 motif, member 13 (ADAMTS 13) activity of less than 5 percent, outcome was very similar to HIV– patients with a low ADAMTS 13 activity. However the HIV+ patients with an ADAMTS 13 activity of greater than 5 percent tended to have far advanced HIV with a median CD4 count of 30 cells/μL, more infectious complications present at the diagnoses of thrombotic microangiopathy, multiple AIDS-associated complications, and a high mortality.²⁶⁹ Thus the outcome of TTP in HIV+ patients can be excellent in patients without advanced HIV when treated promptly with plasmapheresis/plasma exchange and ART.

HUMAN IMMUNODEFICIENCY VIRUS–ASSOCIATED THROMBOCYTOPENIA

In the pre-ART era, approximately 10 to 30 percent of people living with HIV had thrombocytopenia, defined as a platelet count of less than 150,000/μL, and thrombocytopenia was the initial manifestation of HIV in approximately 10 percent of patients. Those whose risk factor for HIV was intravenous drug abuse had a higher incidence of thrombocytopenia, attributed to increased coinfection with hepatitis C and hepatic cirrhosis. A British study evaluated selective testing for HIV in patients who presented with specific medical conditions, including thrombocytopenia. They found an increased rate of HIV infection in patients presenting with thrombocytopenia,²⁷⁰ providing a rationale for including HIV testing in the evaluation of patients presenting with isolated thrombocytopenia. In the ART era, 26 percent of 5290 patients followed at the British Columbia Center for Excellence in HIV/AIDS had at least one platelet count less than 100,000/μL, and 3 percent had at least one platelet count less than 20,000/μL.²⁷¹ A study of the frequency and severity of thrombocytopenia in a large cohort of patients in the Collaboration in HIV Outcomes Research/U.S. study (CHORUS) included 6300 HIV+ people from 1997 to 2006²⁷² and found a prevalence of thrombocytopenia (platelet count <150,000/μL) of 14 percent. However, this cohort excluded patients who had hepatitis C or hepatitis B infection, so the prevalence of thrombocytopenia would likely be higher if these patients had been included.²⁷² In this study, 3.1 percent of patients had a platelet count of 50,000/μL or less, and 1.7 percent of patients had a platelet count of 30,000/μL or less. The majority of patients with severe thrombocytopenia who had not started ART had a CD4 count greater than 200 cells/μL, demonstrating that thrombocytopenia can occur prior to severe immunodepletion. These data are consistent with the findings of the British Columbia Center for Excellence in AIDS/HIV study, in which the CD4 count at diagnosis of ITP was 200 or greater in 72 percent of patients.²⁷¹

Mild thrombocytopenia also occurs during primary HIV infection, a time of unfettered HIV replication and intense immune activation. In one study of 957 patients evaluated during primary HIV infection, 9.7 percent had a platelet count of less than 150,000/μL, 2.3 percent had a platelet count of less than 100,000/μL, and none had a platelet count of less than 50,000/μL.²⁷³ Of those who started ART, the time to platelet recovery was approximately 1 month. Of those who did not start ART, the time to platelet recovery was just under 2 months. Those who developed thrombocytopenia during acute HIV infection had a threefold higher incidence of developing a low platelet count in the next 3 years (13.3 percent) in comparison to those who maintained a normal platelet count throughout. Thus, most patients who develop thrombocytopenia during primary HIV infection recover quickly, even if ART is not started immediately. A study to evaluate the risk factors for HIV-associated thrombocytopenia included 73 HIV+ people with a platelet count of less than 100,000/μL for 3 months matched to 73 nonthrombocytopenic controls. Identified risk factors were an HIV viral load of greater than 400 copies/mL, hepatitis C coinfection, and cirrhosis.²⁷⁴ The platelet count correlated inversely with the viral load in a study of 207 patients naïve to ART.²⁷⁵

Platelet kinetic studies demonstrate shortened platelet survival in HIV+ patients not on ART (92 hours) compared to HIV– healthy volunteers (198 hours).²⁷⁶ Even HIV+ patients with normal platelet counts had modestly diminished platelet survival in these studies as well as decreased platelet production in comparison to HIV– normal volunteers. As discussed above, some human hematopoietic progenitor cells can be infected with HIV in vivo, and these cells demonstrate impaired megakaryopoiesis in vitro.²⁵¹ Additionally megakaryocytes can be infected with HIV²⁷⁷ and these observations may contribute to the decreased production of platelets in the marrow of HIV+ individuals. The more frequent and severe thrombocytopenia seen in patients with HIV and hepatitis C coinfection can be explained by the increased risk of cirrhosis in people coinfected with HIV and hepatitis C. The combined effects of diminished production of thrombopoietin, the major thrombopoietic growth factor, together with portal hypertension, splenomegaly, and sequestration of platelets in the enlarged spleen can result in severe thrombocytopenia. Patients with severe liver failure may also have a component of low-grade disseminated intravascular coagulation.

Evaluation of thrombocytopenia in HIV+ patients is similar to HIV– patients and should include a thorough history and physical exam looking for symptoms and signs of platelet-type bleeding, to assess the clinical severity of the thrombocytopenia. The blood film should be reviewed to confirm that the patient does have low platelets, rather than platelet clumping, and to evaluate for abnormalities in red blood cell and white blood cell numbers and morphology. If not already done, the patient should be tested for hepatitis C. The HIV viral load and CD4 count should be determined, as noncompliance or development of resistance to the current ART regimen can exacerbate HIV-associated thrombocytopenia. The patient should be asked what percent of the patient’s HIV medications are taken, or alternatively how many missed doses the patient has had in the past month. The medication list, including nonprescription medications, naturopathic medications, and dietary supplements, should be thoroughly reviewed. The differential diagnosis for isolated thrombocytopenia includes HIV-associated ITP, hepatitis C-associated ITP, Helicobacter pylori–associated ITP, medication side effect, or antiphospholipid antibody syndrome. If the patient also has anemia, immunohemolytic anemia with ITP (Evans syndrome) or TTP should be considered. In a febrile and ill patient who has additional cytopenias, Castleman disease and hemophagocytic syndrome should be included in the differential diagnosis (Table 81–10).

Treatment of Human Immunodeficiency Virus–Associated Idiopathic Thrombocytopenic Purpura

ART improves the platelet count in patients with HIV-associated ITP over a period of approximately 3 months in the majority of patients.^278,279,280 The primary treatment of HIV-associated ITP is initiation of ART if the patient is not on ART, and assessment of the effectiveness of ART if the patient is taking ART. Reasons for failure of ART include suboptimal compliance with the medications, as the ability of ART to control the HIV viral load is related to adherence.²⁸¹ Alternatively, the patient’s HIV may have developed ART resistance which can be detected by resistance testing. Close communication between the hematologist and the infectious disease/HIV physician is essential for optimal management of these issues.

Because ART typically takes 3 months to improve the platelet count, additional interventions are needed if the patient is experiencing severe thrombocytopenia (platelets <20,000/μL) or has platelet-type bleeding. If the patient is Rh+ and has an intact spleen, intravenous anti-D can be very effective.²⁸² In a study that included both HIV+ and HIV– patients, treatment with intravenous anti-D increased the platelet count by an average of 45,000/μL in adults in both groups. In the majority of patients, the platelet increment lasted more than 21 days. The major side effect of anti-D treatment was a drop in the hemoglobin (average decrease in hemoglobin of 1 g/dL). Another approach is to use intravenous immunoglobulin although one study of immunoglobulin treatment reported a smaller increase in the platelet count (average of 29,000/μL), and a shorter duration of response (19 days) than was obtained with intravenous anti-D.²⁸³ Furthermore, intravenous anti-D may be less expensive than intravenous immunoglobulin. However, anti-D has the potential to cause significant hemolysis and has a “Black Box” warning because of this rare complication. Alternative approaches include the use of standard doses of prednisone or pulse dexamethasone, as in HIV– patients. Steroids are less attractive in HIV+ patients because of the potential to decrease CD4 counts, increase the risk of infection, and increase risk of Kaposi sarcoma progression. Splenectomy can be effective in HIV+ patients,²⁸⁴ but is rarely required. Case reports document the use of romiplostim or eltrombopag in HIV+ patients.^285,286

HIV+ patients with ITP should be counseled not to take aspirin or nonsteroidal antiinflammatory drugs or fish oil while severely thrombocytopenic, and educated on the symptoms of relapse of ITP (spontaneous nose bleeds, spontaneous large bruises, bleeding gums), and encouraged to report promptly for medical care should any of these occur. Most patients with HIV-associated ITP can be started on ART and treated with 1 or 2 doses of anti-D at 3-week intervals and will subsequently experience sustained platelet count response with effective control of HIV.

HUMAN IMMUNODEFICIENCY VIRUS–ASSOCIATED NEUTROPENIA

The prevalence of neutropenia (defined as a neutrophil count <1300/μL) in patients naïve to ART in Africa, Asia, the Americas, and the Caribbean was 14.3 percent,²⁸⁷ with the U.S. cohort having a slightly higher prevalence of approximately 16 percent. A multicenter prospective study of HIV+ women in the United States followed 1729 women²⁸⁸ and found that 7 percent had a pre-ART absolute neutrophil count of less than 1000/μL, and during 7.5 years of follow up 31 percent had an absolute neutrophil count of less than 1000/μL. Low CD4 count and high viral load were significant risk factors for development of neutropenia whereas initiation of ART correlated with resolution of neutropenia. Patients living with HIV may have neutropenia for a variety of reasons and neutropenia may be a component of pancytopenia (see Table 81–10) or occur as an isolated finding. As discussed in the section “Pathophysiology of Anemia in HIV” above, HIV has a multitude of effects on hematopoiesis, and the effects on progenitor cell survival, growth and differentiation and on the marrow microenvironment pertain to myelopoiesis as well as erythropoiesis and megakaryopoiesis. A study of 87 consecutive patients attending an HIV clinic who developed an absolute neutrophil count of less than 1000/μL showed that the vast majority were receiving at least one known myelosuppressive medication, and 59 of the 87 patients were receiving three or more myelosuppressive drugs. Many other medications are rare causes of neutropenia,²⁸⁹ illustrating the importance of drug-induced neutropenia in this population of patients.

Other causes of neutropenia reported in HIV+ patients include deficiency of vitamins such as folate or vitamin B₁₂, hepatitis C-associated autoimmune neutropenia, alcohol toxicity, and levamisole-adulterated cocaine, which can cause severe neutropenia (neutrophil counts as low as 0). Patients should be specifically asked about cocaine use if they are found to have severe neutropenia with normal hemoglobin and platelet count.²⁹⁰