Comprehensive, up-to-date information on HIV/AIDS treatment, prevention, and policy from the University of California San Francisco

There is clear clinical evidence that effective antiretroviral therapy (ART) can result in beneficial immune reconstitution for patients with advanced HIV disease. Examples include the decreased incidence of AIDS-defining opportunistic infections (OIs), the dramatic decline in mortality rates from AIDS, and the clinical resolution of OIs (eg, cryptosporidiosis and progressive multifocal leukoencephalopathy) without specific antimicrobial therapy that has been reported in areas where ART has become widely available. However, laboratory studies suggest that, in spite of the normalization of absolute CD4 counts that can result from long-term, virologically effective ART, immune reconstitution may not be complete, at least with the follow-up data available to date. In addition, among patients with advanced HIV disease who initiate ART, some have exuberant inflammatory responses to opportunistic pathogens that can lead to troublesome immune reconstitution disease (IRD) syndromes.

Effective ART that durably suppresses HIV replication results in rapid and sustained rises in absolute CD4 counts. Typically, a 2-phase increase in CD4 cells occurs after effective ART is initiated: a rapid initial rise in the first few months, primarily due to increases in memory T cells, followed by a slow, steady increase in naive T-cell counts that can continue for years with sustained suppressive ART.(1,2) The majority of patients whose plasma HIV RNA levels remain below the level of detection on ART eventually achieve absolute CD4 counts that are in the normal range. However, several recent long-term follow-up studies have reported that up to one third of patients with long term ART suppression of HIV RNA do not achieve absolute CD4 counts >500 cells/µL.(3,4)

Many studies have demonstrated that the loss of functional lymphocyte responses to many pathogen antigens that occurs in progressive HIV disease can be reversed by ART. For example, improvements in Candida-specific, cytomegalovirus-specific, and mycobacterial-specific lymphoproliferative responses have been observed within 3-6 months of initiating combination ART. Unfortunately, HIV-specific T-cell functional responses, which are lost early in the course of HIV disease, are not reconstituted by ART.

The thymus, which is the primary source of new naive T cells, atrophies with age, and HIV infection itself can destroy thymic tissue. However, computed tomography (CT) scan studies have demonstrated that some patients with untreated HIV disease have abundant thymic tissue. The amount of thymic tissue (measured by CT scan) at initiation of ART correlates with the magnitude of subsequent increase in naive CD4+ T cells.(5) For patients with low thymic volume, there is now evidence that IL-7-mediated peripheral expansion of naive T cells may drive T-cell restoration.(6) Another investigational tool is measurement of T-cell receptor gene rearrangement excision circles (TRECs), which are circular forms of DNA present in CD4+ T cells that have recently emigrated from the thymus. TRECs can be used as an indicator of thymic output. Several studies have demonstrated that ART increases TREC content in circulating T cells, indicating an increase in naive T-cell production.

Recent studies of the immunopathogenesis of HIV disease have emphasized the importance of T-cell activation and apoptosis in explaining the decline in CD4 T-cell count that occurs with HIV disease progression. Studies of patients initiating ART have consistently reported that virologically suppressive ART reduces abnormal activation of T cells (eg, decreases percentage of CD4+ and CD8+ T cells expressing activation markers such as CD38) and also downregulates T-cell apoptosis, leading to a net increase in T-cell production.

However, other laboratory studies suggest that, in spite of the normalization of absolute CD4 counts that can occur on ART, immune reconstitution may not be complete, even after years of virologically suppressive ART. There are reports that the lack of lymphoproliferative response to tetanus toxoid after immunization (a typical functional immune deficit caused by HIV) persists in many patients despite years of ART suppression of HIV replication.(7) Even though effective ART clearly reduces T-cell activation, it appears that T-cell activation status is not fully normalized by ART. In cohorts of patients who had plasma HIV RNA levels suppressed by ART to levels consistently <1,000 copies/mL for 2 or more years, the majority of patients still had abnormally elevated expression of the CD38 activation marker on CD4+ and CD8+ T cells.(7,8) In one cohort, the median frequencies of CD4+/CD38+ and CD8+/CD38+ T cells were 8% and 46%, respectively, in untreated HIV-infected controls and 4% and 11%, respectively, in ART-treated patients with viral loads consistently <1,000 copies/mL, but both frequencies were only 1% in healthy HIV-negative control subjects.(8) Of particular note in this study was the observation that for every 5% increase in CD8+/CD38+ T cells, 35 fewer CD4+ T cells/µL were gained after 2 years of ART.

T-cell diversity is another marker of immunologic health, corresponding to an individual's total CD4 antigen receptor repertoire which, in healthy individuals, has the potential to recognize between 10⁹ and 10¹¹ unique epitopes. T-cell diversity can be crudely assessed by using polymerase chain reaction or flow cytometry to measure the proportional expression of the 25 subfamilies of the T-cell receptor beta chain. Although healthy volunteers have more or less equal representation of all T-cell receptor subfamilies, patients with AIDS have overrepresentation of some subfamilies and underrepresentation of others. The results of several small studies that have examined the effect of ART on improving T-cell diversity have not provided convincing evidence that this HIV-induced skewed diversity is fully corrected by ART.

Another immunologic abnormality caused by HIV infection is abnormal B-lymphocyte activation. Studies conducted prior to the availability of effective ART reported that hypergammaglobulinemia and abnormally increased spontaneous B-cell production of immunoglobulin G (IgG) were present in the majority of untreated HIV-infected patients. Although several studies conducted following the advent of combination ART have reported significant reductions in serum IgG levels after short-term ART, serum IgG did not completely normalize in the majority of the patients described. A longer follow-up of ART effects on serum IgG levels was reported in a study of 29 HIV patients on chronic ART whose absolute CD4 counts had risen by a minimum of 150 cells/µL (median >200 cells/µL) over their pre-ART nadir values for an average of 28 months and whose viral loads were undetectable. Results were compared with those of 29 untreated HIV-infected patients, whose current CD4 counts matched each case's prior nadir CD4+ T-cell count.(9) After long-term ART suppression and CD4 T-cell reconstitution, serum IgG remained abnormally elevated in 45% of the treated group compared with 69% of untreated patients (p = .112).

If long-term ART does not eventually fully normalize the elevated activation status of T and B lymphocytes, nor completely correct skewed T-cell diversity, then long-term survivors of HIV may remain at risk for malignancies, such as non-Hodgkin lymphoma, and autoimmune diseases. Although ART has resulted in a dramatic reduction in OIs and incidence of Kaposi sarcoma in all population-based studies reported, it is notable that, in most natural history studies conducted in the setting of effective ART, the decrease in incidence of non-Hodgkin lymphoma has been of a lesser magnitude than decreases in the incidence of Kaposi sarcoma and of OIs. This raises the possibility that, for many patients, ART may not correct the immunologic abnormalities involved in lymphomagenesis (ie, antigen-driven, T-cell dependent activation and proliferation of B lymphocytes). Long-term follow-up of the incidence of malignancy and autoimmune disease in ART-treated patients will be key in determining the ultimate risk for these potential complications in patients on lifelong ART.

There is convincing evidence from prospective trials that primary prophylaxis for Pneumocystis pneumonia,(10-15) disseminated Mycobacterium avium complex (MAC) infection,(16-18) and toxoplasmosis (10,11,19,20) can be discontinued in patients whose absolute CD4 counts have increased and been sustained by ART for at least 6 months at a level above the CD4 threshold values for prophylaxis (ie, >200 cells/µL for Pneumocystis pneumonia and toxoplasmosis, and >50 cells/µL for MAC). There is also evidence that chronic suppressive antimicrobial therapy can be safely discontinued in most AIDS patients with a history of Pneumocystis pneumonia,(12,13,21-23) disseminated MAC,(21,24-27) cryptococcal meningitis,(26,28) cytomegalovirus (CMV) end-stage organ disease,(21,26,29,30) or toxoplasmic encephalitis (10,24,26) who have been similarly immunorestored by ART for >6 months and have completed an adequate course of pathogen-specific antimicrobial therapy: at least 6 months of antiviral therapy for CMV retinitis and at least 1 year of pathogen-specific antimicrobial therapy for these other OIs. More detailed information on discontinuing prophylaxis is available in the U.S. Public Health Service guidelines on preventing HIV-associated OIs.(31)

Since 1996, when the use of triple-combination ART became widespread in the United States and Western Europe, there have been a number of published reports describing HIV-infected patients who, soon after initiating ART, developed localized inflammatory disease in association with OIs. Many of these cases occurred within a few months after ART was initiated, were self-limited in duration, and involved either exacerbation (often with unusual manifestations) of a previously treated OI or unusual clinical appearance of a previously subclinical infection. Several recent review articles have discussed this phenomenon, which is now generally designated as IRD.(32-34) Generally, IRD is a clinical diagnosis that is made by linking the typical clinical findings unique to the specific OI-related IRD syndrome (see following sections) to a temporally-related, ART-induced, increase in absolute CD4+ T-cell count.

Little is known about the specific immune mechanisms involved in the pathogenesis of IRD, and these mechanisms likely vary from one OI to another. Nevertheless, many IRD syndromes appear to result from enhanced antigen-specific immune responses that lead to increased production of inflammatory mediators. For example, in several small studies, plasma interleukin-6 levels were noted to be elevated in patients with IRD.

MAC IRD

MAC IRD usually presents as lymphadenitis (often hilar, retroperitoneal, or cervical), nearly always accompanied by high fever.(35) Sometimes, the lungs are involved, with pulmonary infiltrates apparent on chest X ray. There have been reports of localized bone, joint, skin, soft tissue, prostate and brain lesions (36). Hypercalcemia has been described in MAC IRD, probably related to overproduction of activated vitamin D by gamma-interferon-activated monocytes, tissue macrophages, and granulomatous tissue. Typically, the onset of this syndrome occurs 1 to 12 weeks after ART initiation, usually in the setting of substantially increased CD4 counts in a patient whose pre-ART absolute CD4 T-cell count was <50 cells/µL. Blood cultures for MAC are usually negative. Histologic examination of MAC IRD lesions shows well-formed granulomas and few MAC organisms. In one observational study from British Columbia, 3.5% of patients (with or without prior diagnosis of mycobacterial infection) who began ART at a CD4 T-cell count <100 cells/µL developed a nontuberculous mycobacterial IRD syndrome (most due to MAC).(36) In a retrospective observational study from Houston, 35 patients who had a diagnosis of disseminated MAC infection initiated ART, and approximately one third of them developed a MAC IRD syndrome.(37) In this study, starting ART within 30 days of beginning antimycobacterial therapy was associated with an increased risk of IRD occurring.

Tuberculosis IRD

Tuberculosis IRD is the most commonly occurring IRD syndrome worldwide.(38-40) It nearly always presents with fever. Other clinical findings may include worsening infiltrates or new pleural effusion on chest X ray, mediastinal and/or peripheral lymphadenopathy, skin or visceral abscesses, arthritis, and osteomyelitis. Hypercalcemia has been described in tuberculosis IRD, probably attributable to the same mechanism as that which occurs in MAC IRD. Typically, the onset of the syndrome occurs 1 to 6 weeks after ART is initiated in an HIV-infected patient who is already on treatment for active tuberculosis. The best available data on the incidence and severity of tuberculosis IRD comes from a prospective treatment trial conducted by the U.S. Centers for Disease Control and Prevention.(38) In this trial, 137 HIV-infected patients with active tuberculosis, and undergoing antimycobacterial therapy, began combination ART. Nineteen percent developed IRD, and 50% of these TB IRD cases required hospitalization. Median duration of IRD symptoms was 64 days. In this study, risk factors for IRD included the presence of extrapulmonary tuberculosis, lower pre-ART CD4 T-cell count, pre-ART plasma viral load >100,000 copies/mL, and earlier start on ART after beginning antimycobacterial therapy. In the retrospective observational study from Houston, 86 patients who had a diagnosis of tuberculosis initiated ART, and approximately one third of them developed a TB IRD syndrome.(37) In this study, starting ART within 30 days of beginning antimycobacterial therapy was associated with an increased risk of IRD occurring. Two other retrospective studies each reported a 15% incidence of IRD occurring in HIV-infected tuberculosis patients who initiated ART, but there was no association in these latter studies of IRD with the interval between initiating antimycobacterial therapy and initiating ART.(39,40)

CMV IRD

Three IRD syndromes associated with CMV end-stage organ disease have been described. CMV retinitis IRD presents as a new opacified retinal lesion, often at the site of a previously noted retinal lesion in patients with a prior diagnosis of CMV retinitis, usually 1-2 months after ART has been initiated and the absolute CD4 T-cell count has risen to >50 cells/µL from a pretreatment value of <50 cells/µL.(41) CMV IRD retinitis appears identical on exam to the active OI, CMV retinitis (caused by uncontrolled CMV replication with retinal cytopathic effect). The diagnosis of CMV retinitis IRD is therefore a clinical one, based on appropriate setting (a recent increase in CD4 T-cell count) and supported by frequent ophthalmologic exams (eg, every 2 weeks) that reveal clearing of the lesions without introduction of or change in anti-CMV therapy. In contrast, retinal lesions caused by uncontrolled CMV replication (eg, due to antiviral drug-resistant CMV) will generally increase in size, or new lesions will occur, within a month of follow-up if there is no change in anti-CMV therapy. CMV vitritis IRD is a benign, but frightening, syndrome. Patients with a preexisting diagnosis of CMV retinitis who are receiving anti-CMV therapy present with acute onset of visual blurring, usually 1-2 months after ART has been initiated and the absolute CD4+ T-cell count has risen to >50 cells/µL from a pretreatment value of <50 cells/µL.(42) Ophthalmologic exam reveals extensive infiltration of the vitreous humor with inflammatory cells, which is the cause of the blurred vision. The condition usually resolves without any specific intervention within 1 month, without residual visual morbidity. CMV uveitis IRD is a late complication of immune reconstitution that occurs a median 3 years after ART is initiated in patients who have a prior history of CMV retinitis.(43) The uveitis primarily involves the posterior pole of the eye and is usually painless. This IRD frequently leads to macular edema, epiretinal membrane formation, or cataracts, and usually results in permanently impaired visual function. Larger area of retinitis involvement may be a risk factor. There is no evidence that reinitiation of anti-CMV therapy or intraocular injection of corticosteroids is beneficial in preventing or altering the course of CMV uveitis IRD.

Cryptococcal IRD

Cryptococcal meningitis IRD typically presents with headache and new meningeal signs and symptoms in a patient with previously diagnosed cryptococcal meningitis who has initiated ART and has had a substantial rise in CD4 T-cell count.(44) The onset has been reported to occur from 1 week to 11 months after initiating ART. Many patients have had increased white blood cell counts (primarily lymphocytes) in their cerebrospinal fluid (CSF) at the time of IRD diagnosis. (CSF white blood cell counts are not typically elevated in AIDS-related cryptococcal meningitis.) Lymphadenitis, particularly involving the mediastinum, has been reported in cryptococcal IRD, and granulomatous inflammation leading to hypercalcemia also can occur. In one study, a prior history of CSF cryptococcal antigen titer higher than 1:1,024 was associated with increased risk of developing cryptococcal IRD. In the retrospective observational study from Houston, 59 patients who had a diagnosis of cryptococcal meningitis initiated combination ART, and approximately one third of them developed an IRD syndrome.(37) In that study, starting ART within 30 days of beginning antifungal therapy was associated with an increased risk of IRD.

Other IRD Syndromes

Localized herpes zoster has been reported to occur with increased frequency within the first 4 months after initiating ART.(45) The risk of zoster occurring in this setting has been positively correlated with patients' subsequent increase in CD8+ T-lymphocyte percentage on ART, suggesting that zoster in this setting may represent an IRD syndrome.(45) To date, none of these reported cases have been disseminated disease, and patients generally have responded well to acyclovir therapy. Published case reports of encephalitis have been suggestive of an IRD syndrome based on temporal association with ART initiation and subsequent rise in CD4 counts and, in some cases, biopsy results. Progressive multifocal leukoencephalopathy, parvovirus B19 infection, central nervous system CMV or herpes simplex virus infection, and disseminated MAC infection each have been implicated in 1 or more cases.(46,47) One case series has described leprosy IRD, consisting of progressive skin ulcers occurring after ART-induced immune reconstitution in patients with a prior diagnosis of leprosy.(48) Case series of sarcoid occurring with CD4 T-cell increases in patients on ART have been reported.(49) The clinical presentation was similar to that of sarcoid in non-HIV-infected patients. As Th1-type CD4 T-cell-mediated granuloma formation is central to the immunopathogenesis of sarcoid, it is reasonable to consider such ART-related sarcoid cases as a form of IRD. There have been a few small series of cases in which patients with severe, active Pneumocystis pneumonia who were improving on antipneumocystis therapy and adjunctive corticosteroids developed acute respiratory failure with fever after early introduction of ART.(50) Because these cases of clinical deterioration have generally occurred after steroids were stopped, it is difficult to sort out whether they are true ART-induced IRD cases or whether they simply reflect a rebound in inflammatory response following withdrawal of corticosteroids. One recently reported series describes 4 cases of IRD that occurred in patients with disseminated histoplasmosis after initiating ART and presented as abscesses, lymphadenitis, uveitis, or arthritis.(51) In 3 of these patients, biopsy revealed well-formed granulomas with caseation. Finally, there are reports of local inflammation, involving lymphadenopathy and tissue edema, in patients with Kaposi sarcoma lesions who recently initiated ART.(52)

There is no evidence from randomized prospective trials on which to base recommendations for the prevention or management of ART-related IRD syndromes. Prior to initiating ART in patients with CD4 counts <50 cells/µL, it is prudent to initiate MAC prophylaxis, which is considered the standard of care for such patients. It may also be useful to obtain blood for mycobacterial culture and to have an ophthalmologist perform a dilated, indirect funduscopic exam, so that clinical occult disseminated MAC or CMV retinitis can be diagnosed. However, it is unknown whether these or any other intervention may reduce the risk of IRD developing.

It is important to note that most reported IRD cases have resolved within several weeks simply by continuing ART and treatment of the opportunistic pathogen. Unless clinical presentation of the IRD is immediately life threatening, there is no rationale for discontinuing ART. Occurrence of IRD does not require reinitiating antimicrobial treatment, or changing existing maintenance therapy, for the infection in question. For example, if a full course of TB treatment already has been completed when IRD occurs, retreatment for TB is not indicated. Similarly, if chronic suppressive treatment is being given (eg, for CMV or cryptococcus), the occurrence of IRD does not mean that the dosage of suppressive therapy should be increased. Invasive diagnostic procedures can be avoided if the clinical setting and presentation are typical for IRD, in which case the clinician can observe the patient to see if the syndrome resolves. It should be noted, however, that both MAC and tuberculosis IRDs can be severe and prolonged clinical syndromes. In several reports, a brief course of systemic prednisone (eg, 1 mg/kg/day for 1-2 weeks followed by a slow taper) appeared to be of benefit for severe cases of mycobacterial IRD.

There are some clinical management questions for which there is little or no evidence to help with decision making. Should ART be delayed or administered immediately to patients who have acute, severe OIs with associated high mortality risk (eg, severe Pneumocystis pneumonia requiring intubation or cryptococcal meningitis with altered mental status and high CSF pressure)? It is possible that a modicum of immediate immune restoration might have a beneficial effect on survival, but it is equally plausible that such immune restoration could lead to an increased inflammatory response that could increase mortality. In medically stable patients who are responding to therapy for recently diagnosed OIs, it is not known whether ART should be initiated immediately to reduce risk of subsequent mortality and development of other OIs or delayed by 4, 8, or 12 weeks with the intention of reducing the risk of IRD. For CMV disease, there appears to be no rationale for delaying ART, because the early CMV IRD syndromes (retinitis and vitritis) are benign and do not lead to long-term visual morbidity. On the other hand, there is some data from observational studies of disseminated MAC, tuberculosis, and cryptococcal meningitis suggesting that delaying ART for 4-8 weeks after initiating antimicrobial therapy for the OI is associated with a decreased risk of IRD. As mycobacterial IRD, in particular, can be clinically severe, delaying ART for 4-8 weeks after starting antimycobacterial therapy may be prudent. It is hoped that future randomized trials will answer some of these questions.