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

Kaposi sarcoma (KS) was one of the first conditions recognized as an opportunistic sequela of HIV infection, and remains the most common AIDS-associated neoplasm.(1) AIDS-associated KS occurs with increased frequency in all HIV transmission groups compared with the general population, but at a particularly high rate among men who have sex with men (MSM). The epidemiology of AIDS-associated KS has long suggested that an environmental or infectious sexually transmitted cofactor might contribute to the development of KS. The search for such a cofactor led, in 1994, to the discovery of a novel herpesvirus, human herpesvirus-8 (HHV-8), also known as the Kaposi sarcoma-associated herpesvirus (KSHV).(2) HHV-8/KSHV is found in all forms of KS, and infection with the virus appears to be necessary but not sufficient for KS to develop. Other factors believed to be involved in the development of AIDS-associated KS include altered expression and response to growth factors and cytokines, and modulation of KS growth by an HIV gene product, the Tat protein. AIDS-associated KS varies in its presentation from an indolent process with minimal clinical consequences to a disseminated, aggressive disease.

KS incidence in the United States has varied over time. The incidence rose steadily during the 1980s, but declined significantly over the next decade. The change in incidence was most dramatic in San Francisco, where the incidence in white men peaked at 32.1 per 100,000 in 1987 and fell to 2.8 per 100,000 in 1998.(3) Nationally, KS incidence in white men prior to the AIDS epidemic was 0.3 per 100,000 in 1973-78, but rose to a high of 8.9 between 1989 and 1991.(3) After the onset of the AIDS epidemic, 90% of people with KS were aged 20-54, whereas individuals in that same age range comprised only 11% of KS cases in 1973-1978. Although the decrease in KS incidence began prior to the introduction of effective antiretroviral therapy (ART),(4) the decline has been steeper since ART became more widespread in the mid-1990s.(5) In a large surveillance study that included 37,303 HIV-infected individuals, overall KS incidence declined by an estimated 8.8% per year from 1989 to 1998.(5) Although the incidence of KS as an AIDS-defining illness has decreased, its incidence as a secondary AIDS diagnosis increased from 23% in the 1980s to 50% in 1996-97.(6)

The mechanisms underlying the steep decline in KS incidence and the occasional regression of established KS lesions in the United States following the introduction of effective ART may be multifactorial. It is possible that these events were triggered by decreased extracellular release of Tat in patients with well-controlled HIV infection, a change in the cytokine milieu, more effective immune responses to HHV-8, or some combination of these events. Studies indicate that both CD8 T-cell responses against HHV-8-specific cytotoxic T-lymphocyte epitopes (7) and natural killer cell cytotoxicity against cells latently infected with HHV-8 (8) increased during successful ART--which included a protease inhibitor or nonnucleoside reverse transcriptase inhibitor (NNRTI)--in HHV-8-infected individuals, suggesting that immune restoration is at least partly responsible. In one study, however, development of a sustained increase in anti-HHV-8-specific CD8 T-cell responses in HHV-8-seropositive individuals required more than 12 months of ART,(9) whereas KS regression may begin within a few months of starting ART, suggesting that other mechanisms may be involved in ART-induced KS regression. Additionally, although some protease inhibitors may have direct inhibitory effects on angiogenesis,(10,11) there is evidence that both protease inhibitor- and NNRTI-containing ART regimens are equally effective in decreasing the incidence of subsequent KS development.(12)

The risk of KS varies in different HIV risk groups and has changed over time. Multistate AIDS-Cancer Match Registry data through 1990 (13) showed a relative risk of 106,000-fold in MSM and 13,000-fold in people exposed to HIV through injection drug use or heterosexual activity, compared with the general population. At the beginning of the epidemic, 40-50% of MSM with AIDS developed KS, whereas the proportion had fallen to <15% by the late 1980s. By contrast, only 10% of injection drug users, 4% of hemophiliacs, and 3% of children with AIDS had KS as a presenting diagnosis.(14) In MSM, the risk of KS has been associated with the number of sexual partners, receptive penile-anal or insertive oral-anal intercourse, and sexually transmitted diseases (STDs).

Detection of HHV-8 DNA in the blood of MSM is predictive of subsequent KS development.(15) The risk of KS also increases as immune function declines, as measured by CD4 counts and duration of HIV infection.(16,17) Furthermore, KS risk is correlated with HHV-8 viral load.(15) In general, the incidence of KS is higher in areas where the underlying rate of HHV-8 seropositivity is high,(18) but exceptions have been noted, particularly in Africa.(19-21)

HHV-8 transmission is not well understood. Although early epidemiologic studies implicated sexual transmission in MSM, recent studies indicate that HHV-8 DNA is detected more commonly in saliva than in genital secretions.(22) In regions where HHV-8 infection is endemic, horizontal infection appears likely. Although HHV-8 DNA can be detected in peripheral blood mononuclear cells,(23,24) and women who inject drugs are more likely to be HHV-8 seropositive than are women without this risk factor,(25,26) seroconversion after transfusion of HHV-8 seropositive blood has not been observed.(27,28)

Contrary to the experience in the United States, the incidence of KS in HIV-infected individuals is increasing in sub-Saharan Africa, where there is limited access to ART. Although KS was endemic in central and east Africa before AIDS, AIDS-related KS has become the most frequently diagnosed tumor in several African countries.(29-31) In Uganda, age-standardized incidence rates rose from 3.2 in men and 0.1 in women per 100,000 in 1960-66 to 39.3 and 21.8 per 100,000, respectively, in 1995-97,(29) and KS accounted for 2% of childhood tumors in the 1960s and 33% in the 1990s. At the same time, the male-to-female ratio of KS incidence in adults declined from 20:1 to 2:1. KS in HIV-infected people has been associated with higher socioeconomic status, history of STDs,(32) and a higher number of sexual partners.(33)

Racial, genetic, and hormonal factors all have been reported to influence the development of KS. Linked AIDS and cancer registry data indicate that the proportion of black MSM who develop KS is about one half that of white MSM,(34) but this finding has not been uniform.(35) It is possible that KS is more difficult to recognize clinically in black patients, or that access to medical attention is less likely for socioeconomic reasons,(34) resulting in artifactually low reporting rates.

Although results of early studies suggested that certain human leucocyte antigen (HLA) types were more commonly associated with KS in HIV-infected individuals,(36,37) subsequent studies have not confirmed these findings.(38) It has been shown, however, that variant genotypes of the Fc gamma receptor IIIA and certain IL-6 promoter polymorphisms are associated with the risk of KS in HIV-infected men,(39,40) suggesting that inherited factors that influence the immune response to HHV-8 may be involved.

All forms of KS are more common among men than among women. In the absence of immunosuppression, KS is currently estimated to be 3 to 4 times more common in men,(41,42) but even higher male-to-female ratios have been reported for the classic and African endemic forms of the disease. Among the iatrogenically immunosuppressed and among HIV-infected individuals other than MSM, reported KS incidence rates are only slightly higher in men.(34) The observations that a tumorigenic KS cell line could not be established in pregnant immunodeficient mice and that human chorionic gonadotropin inhibited KS growth in vitro (43) suggested a possible biologic basis for the lower KS incidence among women. Among Zambian women with HIV infection, however, pregnancy apparently afforded no protection from KS development or dissemination.(44) KS also has been reported to behave more aggressively in HIV-infected women than in men.(45)

Although all forms of KS are histologically similar, and include a spindle cell component, slitlike vascular spaces containing erythrocytes, and a variable inflammatory cell infiltrate, there is a wide range in the distribution and clinical manifestations of AIDS-associated KS. The disease usually presents initially as violaceous skin lesions, but oral, visceral, or nodal KS may precede cutaneous involvement. Biopsy for definitive diagnosis is recommended to distinguish KS from other pigmented skin conditions, which may include bacillary angiomatosis, non-Hodgkin lymphoma, and cutaneous fungal or bacterial infections.

Skin Lesions

KS usually manifests dermatologically as pigmented macules, plaques, papules, or nodules, but subcutaneous nodules may arise without visible skin pigmentation. They can range in size from a few millimeters to large confluent areas many centimeters in diameter and in color from pink (especially in newly developed macules or papules) to red or purple. Often, with time, the color of skin lesions darkens. In black or olive-skinned people, KS lesions may appear dark brown or black. In some patients, red or violaceous lesions may be preceded by, or surrounded with, a yellow to green "halo," probably representing extravasated erythrocyte pigments. With treatment, lesion color may fade to brown, gray, or tan, and pigmentation may persist even in the absence of residual evidence of KS on biopsy. Treatment frequently results in flattening of previously raised papules or nodules and also may be associated with central hypopigmentation of large KS plaques.

Oral Lesions

Lesions of the oral cavity occur in about one third of patients with AIDS-associated KS. Hard palate lesions are most common. These flat, red or purple plaques, either focal or diffuse, may be completely asymptomatic and easily overlooked. In other patients, however, larger nodular lesions involving the hard or soft palate, or both, may become exophytic and ulcerated, and may bleed. Other oral sites of KS involvement include the gingiva, tongue, uvula, tonsils, pharynx, and trachea. These lesions may interfere with eating and speaking, cause tooth loss, or compromise the airways.

Gastrointestinal Tract

Reports describe involvement of the gastrointestinal (GI) tract of 40% of patients with AIDS-associated KS at the time of initial KS diagnosis, and of 80% at autopsy. Gastrointestinal KS may occur without cutaneous KS.(46-48) Although GI KS is common, the presence of asymptomatic GI KS does not appear to influence prognosis substantially. Therefore, formal evaluation of the GI tract is recommended only in those patients who are symptomatic. KS may occur throughout the GI tract, and although most patients are asymptomatic, some experience pain, obstruction, or bleeding.(49) Digital examination may reveal rectal KS, but diagnosis generally requires endoscopy, because most lesions are submucosal and are not visualized readily on contrast-enhanced radiographs.

Pulmonary Involvement

KS may involve the lung parenchyma, bronchial tree, and pleural surfaces. In general, patients with pulmonary KS have advanced HIV disease. Although pulmonary involvement is occasionally asymptomatic and found incidentally on radiographs or during bronchoscopy performed for other purposes, patients typically are symptomatic. Possible physical signs include shortness of breath, cough, wheezing, and hemoptysis, and respiratory failure may occur. Bronchoscopy may reveal endobronchial KS in the absence of radiographic findings, but chest radiographs often show ill-defined nodules, interstitial or alveolar infiltrates (which may be indistinguishable from infections such as Pneumocystis jiroveci pneumonia), or pleural effusions. Computed tomography (CT) scanning provides better-defined images of pulmonary KS than does plain radiography (50) and is helpful for following the response to treatment. Typically, KS nodules show a peribronchovascular distribution on CT and are larger than 1 centimeter in diameter.(51) Although gallium scans are generally negative in KS, whereas lesions have been described as thallium avid,(52) these diagnostic methods largely have been supplanted by CT. Endobronchial lesions generally are diagnosed by appearance and pulmonary specialists rarely biopsy such lesions because of concern about bleeding. Pleural effusions associated with KS often are sanguineous exudates that are cytologically nondiagnostic, may be very large, and may require chest tube drainage.

Lymph Nodes

As with other HIV-infected individuals, patients with AIDS-associated KS often have modestly enlarged lymph nodes. Routine lymph node biopsy often reveals focal KS involvement, although, as with asymptomatic GI KS, this finding appears to have little clinical consequence. Thus, routine biopsy of small lymph nodes is not recommended for the purpose of diagnosing nodal KS. Occasionally, however, massive nodal enlargement may occur, and lymph nodes may be replaced by KS. This presentation may occur in the absence of KS elsewhere, and may be associated with edema. Because the causes of massive or asymmetric nodal enlargement include lymphoma or various HIV-associated infections, diagnostic biopsy is warranted in such cases.

Other Visceral Sites

KS has been reported to involve many visceral organs, including liver, spleen, heart, pericardium, bone, and bone marrow,(49) but involvement of these sites is rarely diagnosed.

KS-Associated Edema

Lymphedema is a frequent complication of AIDS-associated KS, and its severity may be disproportionate to the extent of cutaneous KS. The edema generally is nonpitting. The feet and legs are most commonly affected, but other common sites include the groin, external genitalia, and the periorbital tissues. Less commonly, edema may involve the upper extremities and trunk. CT scans to investigate proximal lymphatic obstruction as a cause of lower extremity edema generally fail to show dramatic enlargement of inguinal or pelvic nodes. The cause of KS-associated edema is not entirely clear, but it may result from tumor involvement of dermal lymphatics or, perhaps, from the production by KS cells of growth factors that increase vascular permeability.(53) Severe edema, particularly of the legs, may be complicated by reduced mobility, contractures, diffuse serous drainage with protein loss, skin ulceration, and cellulitis, often caused by gram-negative bacteria.

Lesion Distribution, Clinical Course, and Complications

KS can manifest at any time in the course of HIV infection, but becomes more common as immunocompetence declines.(54-56) Its distribution, rate of progression, and secondary complications vary tremendously. In some patients, only a few, slowly progressive, inconspicuous lesions are present. In others, however, KS is widespread, rapidly progressive, and cosmetically or functionally debilitating. Sudden progression of KS during acute opportunistic infections may be related to increased production of inflammatory cytokines or HIV proteins that stimulate KS growth. Prior to the introduction of effective ART, spontaneous KS regression was rare,(57) usually incomplete, and temporary. Beginning soon after the widespread introduction of effective ART, there appeared many reports of KS regression among patients on ART.(58) Although some patients were receiving no other antineoplastic therapy and KS regression occurred with ART alone, in others, particularly those with advanced KS, chemotherapy was administered concomitantly. The extent and severity of KS in patients who showed regression versus those who did not is not well delineated, and among those who showed regression, the time to achieving a response was usually many months, and in some cases response occurred more than a year after institution of ART. Thus, it is not possible to state with certainty what proportion of patients with KS will benefit from ART alone, or what are the precise characteristics of such patients. In a small proportion of patients, KS lesions may initially develop or progress rapidly after starting effective ART; this is considered a manifestation of the immune reconstitution inflammatory syndrome.(59) Distinctive cutaneous lesion distributions may occur, such as on the nose, on or behind the ears, or involving the periorbital tissues; on the lower legs or feet (with large plaques on the soles of the feet in some patients); large clusters of lesions in the upper medial thighs or the suprapubic and genital areas; and widespread and scattered over the trunk, but sparing the face and extremities. Lesions may be symmetric and linear, and distributed around skin folds. The factors that predispose to these varied distributions are unknown. The consequences of KS and its impact on quality of life are varied. Although most KS lesions are not painful, some are tender, particularly those on the soles of the feet, and edema of the lower extremities may cause stiffness and pain. Edema of the extremities, genitalia, or periorbital tissues can interfere with walking, urination, and vision, and may be complicated by local infection or ulceration. Oral or GI tract KS may be complicated by nutritional deficiencies. Lung or pleural KS can result in respiratory compromise. KS also has serious social and emotional consequences, including isolation because of obvious, disfiguring lesions, and depression and anxiety from the constant visible reminder of illness.

Various clinical and laboratory features of HIV-associated KS are associated with survival and have been used to develop KS staging systems. A particular difficulty in evaluating prognosis is that KS in HIV-infected individuals is a "disease within a disease." With advances in the therapy of HIV infection and its nonneoplastic complications, the natural history of HIV infection is modified and overall survival is prolonged. Although these advances clearly have been associated with a decreased incidence of KS and a better prognosis for patients with this tumor, the possibility has been raised that ART may enable some patients with KS to live long enough to develop late KS complications, whereas others without KS may survive long enough to develop KS later in the course of their HIV infection, particularly if resistance to ART develops. Despite a trend toward lower median CD4 counts at the time of KS presentation, the results of one study performed prior to the advent of effective ART showed a significant reduction in mortality in patients with AIDS-associated KS over time,(60) suggesting that such patients shared in the improvements in survival seen for other patients with advanced HIV disease. A more recent analysis of data from the Multicenter AIDS Cohort study showed an 81% reduced risk of death in KS patients who received effective ART.(61)

Laboratory tests reported to be positive or negative predictors of survival in HIV-infected patients with KS include the CD4 lymphocyte count, CD4 to CD8 cell ratios, hematocrit, cutaneous anergy, lymphocyte proliferative responses, and serum levels of HIV p24 antigen, endogenous interferon-alfa, beta-2 microglobulin, and neopterin.(60,62-66) Clinical features of patients with AIDS-associated KS, including the extent and distribution of KS lesions, the anatomic site of initial KS presentation, the presence of lymphoma-like "B" symptoms (ie, unexplained fever, weight loss, or night sweats), body mass index, and a history of HIV-defining opportunistic infections (OIs), all have been associated with survival in univariate analyses.(60,62,66) Many of these analyses of prognostic factors were conducted before the routine use of OI prophylaxis and ART.

The most widely used staging system for KS was proposed in 1988 by the Oncology Committee of the AIDS Clinical Trials Group (ACTG).(67) This system takes tumor distribution, CD4 count, HIV-associated symptoms, and opportunistic complications into account, and separates patients into good- and poor-risk groups for each of these 3 variables (see Table 1). Subjecting this classification to prospective validation revealed each of the 3 variables to be significantly associated with survival.(68) In multivariate analysis, only the CD4 count and tumor extent were significantly associated with survival, however, and additional analyses indicated that a lower CD4 count (ie, 150 cells/µL) was a better prognostic discriminant for survival than was the originally proposed 200 cells/µL cutoff. This study was performed on patients treated prior to the introduction of effective ART. A more recent survival analysis conducted after the introduction of effective ART showed that tumor extent and HIV-associated symptoms and complications were the most important predictors of survival, whereas the CD4 count did not add significant predictive value.(69) Of the 211 patients included in this analysis, 76% had not received ART at the time KS was diagnosed, but all received ART subsequently. The level of HIV viremia at KS diagnosis was not a significant predictor of survival in this analysis.

It should be noted that these staging classifications have been validated only for survival, but not for response to KS treatment. Further studies are ongoing to determine whether other variables more specific to KS (eg, measures of HHV-8 viral burden) add to the predictive value of the existing staging classification.