HHV-8 and KS epidemiology and transmission

Atkinson (2004) identified AIDS-related KS cases in linked United States AIDS and cancer registries for 164,000 people. KS incidence was highest in gay men, lower for heterosexual men and lowest for women (5.7 vs. 0.7 vs. 0.4 per 100 person-years). Relative risk adjusted for age, race, location and year of AIDS onset for injecting drug use were 0.9 (95% CI 0.8, 0.9) for gay men, 1.1 (95% CI 0.7, 1.6) for heterosexual men and 1.3 (95% CI 0.9, 1.8) for women.

Phiri (2004) tested blood samples from children in Zambia. At months 12, 24 and 36, 681 (18%), 669 (27%) and 393 (25%) of the children were infected with HHV-8. Of the 269 children followed longitudinally, seroprevalence was 22, 23 and 27%. Children with HIV-1 or born to HIV-1-positive mothers were more likely to acquire HHV-8 infection.

Diamond surveyed 494 men (aged 15-22 years) of whom 26% were exclusively heterosexual (MSW) and 74% were gay/bisexual (MSM). Prevalence of HHV-8 was 6% in MSM and 5% in MSW. Risk factors for HHV-8 were unprotected anal sex with a man in the previous 6 months, CMV and amphetamine injection.

Hermans studied the incidence of KS from 1984-1997. Incidence dropped from 59 cases per 1000 patient years in 1984 to 21 cases/1000 patient years in 1990 to 3.2/1000 patient years in 1997.

Brewster analysed the epidemiology of KS in Scotland from 1976-1996. In men, KS increased from less than 0.09 cases per 100 000 prior to 1986, to 0.44 cases in 1991. Incidence then declined to 0.17 in 1996. AIDS-associated KS was associated with increased morbidity and shorter survival, compared with the traditional, non-AIDS KS seen in older men.

Brodt undertook a retrospective study of 241 HIV-infected men with KS and compared results with 241 HIV-infected men without KS. Controls were matched for age, previous antiretroviral treatment and stage of disease. Men with KS had a higher incidence of opportunistic infections (5.95 vs 3.88 per 100 person months) and a greater risk of death even when adjusting for age, OIs, baseline CD4 count and ART.

Gao reported that in the Multicenter AIDS Cohort Study, 80% of 40 men who developed KS were HHV-8 seropositive prior to the onset of KS. The median time from seroconversion to HHV-8 and development of KS was 33 months (range 6-75 months).

Martin enrolled 400 HIV-positive men and 400 HIV-negative men in the San Francisco Men's Health Study in 1984-5. HHV-8 antibodies were found in 223/593 (37.6%) who reported any homosexual activity in the previous 5 years and in no men who were exclusively heterosexual during this time. Men infected with HIV and HHV-8 at baseline had a 10-year probability of developing KS of 49.6%.

Osmond used data from the San Francisco City Clinic Cohort, the San Francisco Men's Health Study and the San Francisco Young Men's Health Study to determine the prevalence of HHV-8 at the start of the HIV epidemic in San Francisco. In 1978-79, a random sample of 235 showed 26.5% had HHV-8 while the prevalence was 6.9% in the City Cohort. In 1984/85, 29.6% of the Men's Healthy Study sample had HHV-8 while the Young Men's sample had a prevalence of 26.4% in 1995-96. HIV prevalence fell from 49.5% in 1984-85 to 17.6% in 1992-93. The authors document a dramatic fall in unprotected anal sex during this time but a high prevalence of unprotected oral sex continuing into the 1990s. O'Brien et al dispute the use of cross-sectional data to draw conclusions about the transmission of HHV-8 and the incidence of HHV-8.

Casper reported on HHV-8 seroprevalence in 578 gay men in Seattle. Baseline seroprevalence was 16%. HSV-2 antibody, hepatitis A or B and more than 4 sexual partners in the previous 6 months were each independently associated with HHV-8 infection at baseline. 15 new HHV-8 seroconversions occurred for an incidence of 3.8/100 person years. HHV-8 seroconversion was associated with an HIV-positive sexual partner, amyl nitrate use or lymphadenopathy in the past 6 months.

Melbye reported that in a group of 259 homosexual men studied from 1981 to 1996, 21% were KSHV-seropositive at baseline in 1981. The seroconversion rate was highest in 1981 (10 cases per 1000) and declined thereafter (five per 1000 in 1983-4, one per 1000 in 1989-92). Seroconversion to HHV-8 was associated with sexual exposure to homosexual men from the USA, and preceded the onset of KS symptoms in every case. Use of poppers was not related to HHV-8 seropositivity or to the onset of KS.

He conducted a cross-sectional study of 378 pregnant women in Zambia to investigate the seroprevalence of HHV-8. 183 (48.4%) had HHV-8 antibodies. 51.1% of the HIV-infected women were HHV-8-positive, compared to 47.3% of the HIV-negative women. Children of women with KS were not all born HHV-8-positive.

Transmission of HHV-8

Casper (2004) measured HHV-8 DNA in oropharyngeal swabs from 54 HHV-8- and HIV-infected patients. Median baseline CD4 cell count was 55 cells/mm³ and HIV VL was 6390 copies/ml. In patients not taking antiretroviral therapy, HHV-8 was detected on 34% of days. HHV-8 shedding was lowest on days when tenofovir (0 vs 29%), amprenavir (0 vs 30%), indinavir (0 vs 30%), lopinavir/ritonavir (0 vs 30%) or efavirenz (0 vs 32%) were taken). Nelfinavir (16 vs. 31%), saquinavir (15 vs 30%), abacavir (10 vs 31%), ddI (18 vs 30%), 3TC (16 vs 38%) and AZT (26 vs 29%) were associated with reductions in shedding. d4T, FTC and nevirapine were not associated with reduced shedding. Regimens containing drugs from all 3 classes were most effective (0%), followed by nucleoside analogue reverse transcriptase inhibitor (NRTI) and protease inhibitor (PI)-containing regimens (13%), NRTI and non-nucleoside reverse transcriptase inhibitor (NNRTI) regimens (35%) and NRTIs only (65%).

Moore (Nov 2001) reviewed existing data on transmission of HHV-8 in an editorial in the New England Journal of Medicine. Unprotected anal sex, deep kissing, oral-penile contact, and organ transplant have all been associated with HHV-8 transmission. The high prevalence of HHV-8 in parts of Africa have led to suggestions that HHV-8 transmission may be associated with poverty and poor hygiene.

O'Brien reported on risk factors for HHV-8 infection in 259 HIV-positive men studied from 1982. Receptive anal intercourse OR 4.1); insertive anilingus (OR 5.5) and insertive fisting (OR 8.1) were strongly associated with HHV-8 infection, and men who were HHV-8-positive were 3.6 times more likely to develop KS than HHV-8 negative men. Men with HIV and HHV-8 had a 39% chance of developing KS within 10 years.

Grulich found that among 202 Australian men with AIDS, the factor most strongly associated with HHV-8 infection was insertive oral-anal intercourse.

Blauvelt reported that HHV-8 DNA was detected in PBMC samples from 8 of 35 adult people with AIDS in the USA, but none of 49 children with AIDS, indicating that in the USA transmission of HHV-8 occurs predominantly in the years following adolescence. However, this may also reflect the low incidence of HHV-8 in the general population and among HIV-positive women;

Broboski reported that HHV-8 could be detected in the semen of 18 of 36 HIV-infected men, at all stages of HIV infection. Monini reported that KSHV was present in 90% of genital tissues and semen from HIV-negative men in Italy, but many researchers believe that too sensitive a PCR screening tool was used, resulting in false positive results. Calabro detected HHV-8 in the cervicovaginal secretions of 1 of 13 women infected with both HHV-8 and HIV, and in 0 of 7 HIV-negative women with HHV-8. Similarly, HHV-8 was detected in the semen of only 1 of 6 HHV-8-positive men. The man with detectable HHV-8 in his semen had been diagnosed with KS.

La Duca reported on the isolation of HHV-8 by PCR from stool samples (isolated from 0%), skin lesions (92%), normal skin (23%), PBMCs (46%), plasma (7%), saliva (37% and semen (12%) of people diagnosed with KS. The average number of copies per mcg of positive target DNA was 64,000 copies (skin), 9,000 copies (PBMCs), 40 copies (plasma), 33,000 copies (saliva) and 300 copies (semen).

Prior to the discovery of HHV-8, many epidemiological studies had suggested that the causative agent was sexually transmitted, and that oral-anal contact might be a primary route of transmission (Archibald; Bary; Beral; Chu; Couterier; Darrow; Elford; Kaldor; Peterman). Some studies suggested an association with the use of nitrite inhalants (poppers), while others found no such association (Haverkos).

Several studies have found evidence of mother to child transmission, possibly via shedding of HHV-8 in saliva. Hladik reported a study of 406 blood samples (half HIV positive) from a pool of 3736 blood samples. HHV-8 was associated with HIV and HBV infection, but not with sex and did not vary by age, strongly suggesting infection during childhood. Vaithilingum studied 101 mother-infant pairs in South Africa. HIV prevalence among the mothers was 37% and HHV-8 prevalence 23%. HHV-8 infection among the infants was 19%. There was a strong correlation between maternal and infant HHV-8 infection. At 18 months of age, 45.8% of infants born to HHV-8 infected women were also infected, compared to 11.4% of infants born to uninfected women. Viera suggested that transmissible HHV-8 may be present in saliva. Yegorov reported genetic similarity of HHV-8 in 4 mother-infant pairs.

Mbulaiteye reported that blood transfusion was independently associated with HHV-8 infection in 609 Ugandan children with sickle cell anaemia.

Pauk added further weight to the case that HHV-8 is spread through saliva. He studied HHV-8 in mucosal samples and sexual fluids of 50 men who have sex with men. 30/50 (60%) were infected with HHV-8. Tissue samples taken from the mouths and throats of the HHV-8 infected men showed that 30% contained HHV-8 compared with 1% of anal and genital samples (p<0.001). In 39% of HHV-8 infected men, HHV-8 was detected in saliva on more than a third of the consecutive days when samples were taken. Letters by Biggar et al suggest HHV-8 shedding in HIV-infected people may play a part in HHV-8 transmission, while other correspondents criticised the cross-sectional study design.

HHV-8 and disease

Cannon (2003) recruited 91 men who were positive for both HHV-8 and HIV from two HIV clinics in Atlanta, Georgia, between 1996 and 1999. All the men were gay, and 57 had been diagnosed with KS. They also enrolled 70 controls including HIV-negative people and HIV-positive people at low risk of HHV8 infection. Among the men who were positive for both HIV and HHV-8, those with KS were likely more likely to have HHV-8 DNA in both PBMC (35%) and saliva (37%). HIV-positive men with HHV-8 but no KS lesions, had high prevalence of HHV-8 DNA in saliva (32.4%), but not PBMC (a little under 6%). When the risk factors for the development of KS lesions were examined, it was found that the most important factor was the presence of HHV-8 DNA in PBMC, and this remained the case when low CD4 count and high HIV viral load were controlled for. In addition, in men with pre-existing KS, HHV-8 in PBMC was strongly associated with the risk of developing new KS lesions, and HHV-8 DNA in saliva was predictive of the risk of new oropharyngeal lesions. This study also found some evidence that high levels of HHV-8 antibodies help to stabilise KS.

Chang and Moore initially isolated a small section of HHV-8 DNA from a KS specimen, using a PCR-based approach. Presence of HHV-8 in KS tissue was subsequently confirmed by research teams throughout the world. It has been found in 90% of studied cases, including AIDS-associated KS, classic Mediterranean KS, endemic KS from Africa and transplant-associated KS (Boshoff; Dupin 1995; Huang 1995; Rady; Su). HHV-8 has also been isolated from primary effusion lymphomas (Cesarman) and from 100% of HIV-positive and 50% of HIV-negative people with Castleman's disease (Ganem).

Nasti retrospectively compared 54 women and 108 men with AIDS-associated KS. Women with KS were more likely to be younger, with greater immune damage (lower CD4 count ) than men. When controlling for age and CD4 count, women with KS were more likely to have KS in their vital organs, particularly the lungs and other atypical sites. Survival following KS diagnosis was shorter in women than men (8.9 months vs 14.4 months, p=0.07). This suggests KS, although less common in women, may have a more aggressive course.

Engels studied 189,159 subjects from cancer registries in the US, including 26,972 infected with HHV-8. Incidences of specific cancers among people with and without HHV-8 were compared. Immunoblastic lymphoma was significantly associated with Kaposis sarcoma and HHV-8 is implicated in the immunoblastic tumour. None of the other 37 specified cancers were associated with HHV-8 or KS.

Campbell conducted a blinded, cross-sectional analysis of HHV-8 in the blood of 26 people with AIDS-related KS and 5 people with HHV-8 without symptoms of KS. HHV-8 DNA levels were not associated with KS but HHV-8 in peripheral blood mononuclear cells was strongly associated with clinical KS (p=0.05).

Fitzpatrick reported the development of a PCR test for HHV-8.

English and Danish retrospective cohort studies have found that infection with HHV-8 (KSHV) is associated with a lower incidence of HIV-related dementia. Baldewag found that the incidence of CNS disease among people diagnosed with KS was 10% in comparison to an incidence of 22% in people without KS. Researchers are exploring the theory that HHV-8 interferes with the way HIV binds to the microglial cells. Overall HIV disease progression was not affected (Liestoel).

A number of possible cofactors for the development of KS have been proposed, including the growth factor oncostatin M (Miles 1992; Nair), HIV's tat protein (Ensoli; Vogel) or human papilloma virus 16 (Huang 1992).

Soderberg injected mice with cancer cells and then compared mice who inhaled isobutyl nitrite (poppers) with those who had not. Mice were exposed to 900 parts per million for 45 minutes a day. 75% of mice exposed to poppers developed tumours compared with 21% of non-exposed mice. Other tests found that poppers suppressed immune functions which kept tumour cells in check.

Ascherl found increased expression of vascular endothelial cell growth factor-A (VEGF-A) from HIV-infected T cells in comparison to non-infected T cells. This promoted inflammatory cytokines such as tumor necrosis factor. VEGF-A levels were higher in people with HIV, both with and without KS, as compared with uninfected individuals. This provides evidence to support the theory that increased secretion of VEGF-A in HIV-infected individuals increases vascular leakage and stimulates proliferation of vascular endothelial cells, hallmarks of KS development.

KS in the age of anti-HIV therapies

Bourboulia (2004) recruited 27 gay men starting their first HAART regimen. 4 had symptomatic KS and 20 had detectable HHV-8 latent nuclear antigen (LANA) antibodies. 7 patients had detectable HHV-8 VL in blood plasma and 14 in peripheral blood mononuclear cells (PBMCs). After 23 months' follow up, there was no change in LANA antibodies, but more men acquired anti-HHSV-8 lytic antibodies. HHV-8 VL declined after 12 months in blood plasma (26% at baseline and 13% after 12 months) and in PBMCs (odds ratio 0.4; 95% confidence interval [CI] 0.26, 0.63). 3 of the men with symptomatic KS had complete remission of lesions, and an increase of HHV-8 CD8 T-cell responses. HHV-8 VL became undetectable in 3 men after 1 year and after 23 months in the 4th. 8 men were followed for 41 months: 7 had undetectable HHV-8 VL in PBMCs at the end of follow-up.

Martinez (2004) analysed data from 138 HIV-positive patients presenting with KS and treated with HAART. At months 6, 12, 18 and 24, remission was seen in 51, 62, 70 and 76% of patients. Chemotherapy was being received by 54, 15, 6 and 4% at these time points. There were no differences between patients in remission and progression in age, CD4 cell count and HIV VL at KS diagnosis. Progression was associated with stage S1 KS at month 6. Partial or complete remission was associated with suppressed HIV VL for >6 months but not CD4 cell count (p < 0.001). MCD was associated with progression of KS (p < 0.001). There was no difference in outcome between patients on protease inhibitor- and non-nucleoside reverse transcriptase inhibitor-based regimens.

Mocroft (2004) conduced a prospective cohort study of incidence and risk factors for KS in >7000 individuals from the EuroSIDA cohort between 1994 and 2002. 567 had KS at recruitment: these were more likely to be male, gay men and from northern Europe (p < 0.001). KS declined from 8% pre-HAART to 2% in 1998 onwards (p < 0.001). Individuals with KS had lower CD4 cell counts (103 vs. 245 cells/mm³, p < 0.001), were older (39 vs. 36 years, p < 0.001) and were recruited to EuroSIDA earlier (p < 0.001). VL was similar in those with and without KS. 4014 patients started HAART during follow-up. 1% of these developed KS: these had lower CD4 cell counts (median 77 vs. 210 cells/mm³, p < 0.001), higher VLs (median 150,000 vs. 15,000 copies/ml, p < 0.001), lower nadir CD4 cell counts (p = 0.003) and higher pre-HAART VLs (p = 0.042).

Thirlwell (2003) reported that 70 HIV-infected patients were diagnosed with KS between 2000-2002. At diagnosis, 9 had visceral (organ) involvement, average CD4 count was 212, 20 were on HAART and 11 had undetectable viral load. Of the 50 not on HAART at diagnosis, 35 were treated with HAART alone; at one year only five of these patients required chemotherapy or radiotherapy for HIV. Of the total cohort, 2 patients died of Castleman's disease and 2 due to PCP with an overall 2-year survival rate of 93%.

Little (2003) followed 211 people with KS for 3 years during the HAART era. Prognosis was not associated with virological response to HAART, HAART regimen or age. Only the combination of poor tumor stage and poor systemic disease was associated with unfavourable prognosis. Average survival among people with poor tumour stage and poor systemic disease was 38 months, whereas median survival among other patients has yet to be reached.

Martinez (2003) studied the evolution of KS in 61 HIV-infected people treated with combination anti-HIV treatments (HAART). Remission rates rose from 42% at six months to 64% at 2 years, following treatment with chemotherapy. Remission was associated with older age and viral suppression, while non-response was associated with being treatment-naive. Progression at 6 and 12 months predicted subsequent progression. The four patients with Castleman's disease had the worst prognosis.

Tirelli reported that cases of KS have fallen by 30-50% since the introduction of protease inhibitors. Highly active antiretroviral therapy (HAART) may be a useful alternative to immune response modifiers (e.g. interferon) during the less aggressive stages of KS and to systemic chemotherapy during long-term follow-up of KS patients.

Tam studied the survival of 387 men from the Multicenter AIDS Cohort Study (MACS) during 1990-1999. 287 men developed KS, of whom 43 (15%) received HAART. HAART was associated with improved survival (p=0.0001) and 81% reduced risk of death.

Ledergerber analysed data from over 6,000 people in the Swiss HIV Cohort Study, comparing new AIDS-defining illnesses between 1992-1994 and 1997-1998. AIDS-defining illnesses including KS declined by about 80%. Only non-Hodgkin's lymphoma did not decline significantly.

Holkova conducted a retrospective study of 37 HIV-infected people with KS of the lungs between 1994-1997 in New York City. Survival after diagnosis was significantly longer in the HAART era (1996-97) compared to 1995-96 (p=0.0025).

Renato reported on the response to HAART in 53 individuals with a prior KS diagnosis. Mean CD4 count was 174 (range 1-882) and mean viral load was 184,282 copies. Individuals received no other therapy for KS during the prospective study. The magnitude of CD4 increase after commencing HAART was associated with the likelihood of complete or partial remission of KS, but HIV RNA response was not predictive of remission. The rate of complete remission was 36%, and partial remission occurred in a further 36% of cases.

Bower conducted a retrospective review of 78 people diagnosed with KS who started HAART (49% on PI-containing triple regimen, 35% on NRTI/NNRTI triple regimen, and 16% of dual PI therapy). Median follow-up was 12 months. 24 (31%) required anti-KS treatment during this time. Median time to HAART failure as a treatment for KS in this group was 1.7 years, compared with half a year for previous KS treatments.

Sanatambrogio assessed the impact of HAART in nine people with KS. None had received prior KS treatment but all had taken NRTIs previously. Median baseline viral load was about 60,000 and median CD4 count was 50. At 16 months follow-up, 7 people had complete remission of KS, and one had a partial response.

Buchbinder studied AIDS-related cancers among 622 HIV-infected men in San Francisco. The incidence of KS fell from 3 per 100 person years in 1993-1995 to 0 per 100 person years in 1996. (p=0.07). Use of combination therapy rose from 13% at the start of 1993 to 79% in 1997. No decrease in lymphoma was found during this period.

Jacobson followed 5622 gay and bisexual men for over 10 years (MACS). Incidence of KS fell from 25.6 per 1000 person years in 1985, to 7.5 per 1000 person years in 1996-97.

Dupin (1998) reported complete responses in 3 of 5 people with pulmonary KS and 4 of 8 people with cutaneous KS after introduction of triple combination anti-HIV therapy that included a protease inhibitor.

Monticelli reported that remission of KS in 18 people with AIDS treated with combination anti-HIV therapy that included a protease inhibitor was more pronounced among individuals in the early stages of KS. These individuals also had low viral load and CD4 counts greater than 200.

Lyter reported that among 18 patients receiving a HAART regimen who experienced a rise in their HIV viral load, progression of KS occurred in each case after a median of 2 months. Improvement in KS after a fall in HIV viral load was much more variable.

Radiation and intralesional injections for KS

Chak and Quivey reported that radiation therapy can be an effective method of reducing facial oedema (accumulation of fluid below the skin or in the cavities of the body) and treating KS lymphoedema (accumulation of fluid because the lymphatic system is blocked).

Epstein treated 42 people with oral KS with intralesional injections of vinblastine (0.2 mg/ml) under local anæ³´hesia. A greater than 50% reduction in the lesions was seen in 74% of patients. The mean duration of response was 3.52 months for all patients. For patients not lost to follow-up and observed until recurrence of the oral lesions, palliation was achieved for a mean of 4.25 months.

Staddon reported the preliminary analysis of a dose finding study of intralesional recombinant platelet factor-4 (rPF4). Results of the trial were based on data from seven patients with KS. When rPF4 was administered via intralesional injection, there were 2 complete responses, 2 partial responses, 2 with stable disease and 1 with progressive disease. No complete responses and only one partial response was reported in the non-injected, proximal and distant control lesions.

Boente described the sustained clearing of KS lesions on an individual's foot after local injections of 400 µg recombinant GM-CSF.

Interferon for KS

Krown and Groopman reported response rates of 30-50% when treating cutaneous KS in patients with no history of opportunistic infection or systemic symptoms with interferon alfa.

Mitsuyasu (1991) described objective responses in approximately 30% of people treated with high doses (above 20 MU/m²). There is a strong correlation between dose and response.

Rozenbaum reported on 120 patients with AIDS-related Kaposi's sarcoma treated by 18 million units of recombinant interferon alfa 2a daily and followed prospectively for 1-6 years. An overall complete response was observed in 35%; the figure was significantly higher in those who did not have a visceral localization or opportunistic infections. Total lymphocyte count, CD4 lymphocyte count, and CD4/CD8 ratio were significantly higher, and beta₂-microglobulin significantly lower, in the responders than in the non-responders.

Kovacs treated 39 HIV-positive patients with KS with AZT (250, 100, or 50mg orally every 4 hours) and 6 weeks later with interferon alfa (5 million units/day increased every 2 weeks until a maximum tolerated dose was determined). Patients then received the maximum tolerated dose of the combination for a minimum of 12 weeks. In the dose-escalation phase, those receiving a lower AZT dose could tolerate a higher dose of interferon alfa. Dose-limiting toxicities included neutropenia (57%), fatigue (16%), thrombocytopenia (14%), and hepatic dysfunction (10%). Of the 22 patients who received a stable dose of both drugs for 12 weeks, 11 showed a complete or partial tumour response and 8 showed an anti-HIV response.

Shepherd randomised 108 people with KS to receive AZT (500 mg/day) plus either high-dose (8 million units/day subcutaneously) or low dose (1 million units/day subcutaneously) interferon alfa. Response was measured by reduction in number and/or flattening of lesions. 31% of the high-dose arm had a complete or partial response, compared with 8% of the low-dose group (p=0.011). People with baseline CD4 counts above 150 had a better response to therapy. Dose alterations were more common in the high dose group, as were adverse events.

De Wit evaluated the effectiveness and antiretroviral activities of interferon alfa in AIDS-related Kaposi's sarcoma in a non-randomised, phase-II clinical trial. 28 participants were treated with high-dose (27-36 million units) human recombinant interferon alfa 2a subcutaneously every day for 8 weeks. In participants with stable disease or showing a response, treatment was continued three times weekly until a complete response was achieved or there was progression. 12 of the 26 evaluable participants achieved a major response; 5 of these showed histologically confirmed complete responses. There was a significant increase in CD4 cells in the responders and a significant decrease in HIV antigen (HIV-Ag) in the 7 responders with initially detectable HIV-Ag.

Krown (1992) reported that concomitant GM-CSF can maintain an adequate neutrophil count during combined treatment with interferon alfa and AZT. However, constitutional side-effects of the combination prevented significant increases in the dose of interferon alfa used.

Miles (1992b) treated 39 people with advanced KS with interferon beta (90 mu/day or 180 mu/day subcutaneously for 5 days per week). 6/39 (16%) had objective responses, including 2 complete responses. Responses were not dose-related. All recipients experienced fever, night sweats, fatigue and headache; no hæ­¡tological toxicities were seen.

Chemotherapy for Kaposi's sarcoma

Studies which used liposomal chemotherapies are summarised in 'Liposomal therapies for Kaposi's sarcoma' below.

Rosenthal retrospectively analysed 91 HIV-infected patients with KS who had received at least one cycle of daunorubicin between September 1996 and September 1997. Mean CD4 count was 114. Daunorubicin was given as a single chemotherapy in 70% of cases and the average number of cycles was 16. 90% were taking concurrent antiretroviral therapy. 26.5% had a partial response and 11.5% had a complete response.

Evans conducted a study of oral etoposide 50mg/d for 7 days, every second week, in 36 HIV-infected people with KS which had not resolved after treatment with chemotherapy or anthracycline therapy. In non-responders or partial responders, the dose of etoposide was escalated to 100 mg/d orally on days 1 to 7 of each 14-day cycle. Overall response rate was 36% with only one person having a complete response. 33% had stable KS. 5/14 who had a dose increase showed a response. Average time to response was 18 weeks and mean duration of response was 25 weeks. Severe neutropenia occurred in 11 people.

Laubenstein reported a 30% complete response and a 46% partial response among 41 people with KS and favourable prognostic characteristics treated with etoposide (150 mg/m² intravenously for 3 consecutive days every 4 weeks), with a median duration of response of 9 weeks. Myelosuppression and GI toxicity were tolerable.

Schwartsmann (1997) treated 25 people (1 woman) with AIDS-related KS with oral etoposide at a dose of 25 mg/m² twice daily for 7 days followed by a one week rest. Prior therapy included irradiation, interferon alfa, resection or cryotherapy. No participants had previous received chemotherapy. The average number of treatment cycles was six (range 4-27). Complete and partial responses were achieved in 2 and 6 patients, respectively (32%). 5 (20%) had stable KS while 12 (48%) did not respond. The most common adverse effects were leukopenia (36%), thrombocytopenia (20%) and nausea (50%).

Lassoued treated 60 people with disseminated KS and CD4 cell counts below 400 cells/mm³ with bleomycin, either intramuscularly (5 mg/day for 3 days every 2-3 weeks) or by continuous intravenous infusion (6 mg/m² over 4 days). Overall 48% responded to treatment with no difference between the two regimens. Toxicity was mild with occasional fevers and skin rashes.

Fischl treated 53 people with KS and no previous treatment with doxorubicin (intravenously at a dose of 15 mg/m²). 51 participants were evaluable for toxicity and 50 for tumour response. 5 people had a partial response (10%); 32 a minor response (64%); 12 no change (24%); and one progression (2%) as the best measurable response. Partial response durations ranged from 4 to 14 weeks. Fifteen patients subsequently showed progression while on treatment. A significantly greater number of people with cutaneous disease alone (20.1%) had a partial response compared with those with visceral involvement (0%). The major toxicities included nausea (37%), stomatitis (9.8%), mucositis (13.7%), and moderate to severe neutropenia (71%). Two participants developed cardiac toxicity.

Gill (1990) reported response rates of 45 to 88% and median survival of 9 months in patients with aggressive cutaneous or visceral KS treated with doxorubicin (10-20mg/m²), bleomycin and vincristine (ABV).

Gill (1991) reported that the combination of doxorubicin (20mg/m²) with bleomycin and vincristine produced significantly higher rates of complete and partial KS resolution than doxorubicin alone (88% vs 48%).

Saville conducted a Phase I trial of paclitaxel (Taxol) at a dose of 135mg/m² administered as a 3-hour IV infusion, with dose escalation each cycle to a maximum of 175mg/m². Over 65% of treated people had partial responses. All participants with pulmonary involvement experienced partial responses. Toxicities included neutropenia, fever, rash, elevated eosinophil count and mild kidney toxicity.

Gill (1999) reported on the treatment of 56 people with advanced KS with low dose paclitaxel (100 mg/m² intravenously given over 3 hours every 2 weeks until remission or progression of KS). 70% had tumour-associated edema and 45% had organ involvement. 71% had previously received systemic treatment and 31 were resistant to anthracycline. Median CD4 count was 20 (range 0-358). After an average of 10 cycles of treatment, one person had a complete response and 32 had partial responses (total response 59%). Median duration of response was 10 months. Toxicities included thrombocytopenia, anaemia, alopecia, rash, fatigue, nausea, vomiting and myalgia.

Welles enrolled 29 people with advanced HIV disease and KS into an open-label study of paclitaxel. Overall response rate was 71.4% (18 partial, 1 complete). 5 people with pulmonary KS all responded to treatment, as did 4 people who had previously received anthracycline.

Nasti reported the use of vinorelbine (30mg/m² IV every two weeks), a new semisynthetic vinca alkaloid, in the treatment of 35 patients with KS who had relapsed after first-line chemotherapy. 9% had complete responses and 34% had partial responses by ACTG criteria. Toxicities were mainly myelosuppression, but required no discontinuations. Median progression-free survival was 176 days.

Liposomal therapies for Kaposi's sarcoma

Martin-Carbonero (2004) randomised 28 patients with >9 cutaneous KS lesions or mucosal or visceral KS, but not life-threatening KS, to receive HAART with or without pegylated liposomal doxorubicin (PLD) in 3-weekly cycles of 20mg/m². Median baseline CD4 cell count was 97 cells/mm³ and VL was 39,800 copies/ml. After 48 weeks, more patients in the PLD group had partial or complete remission (76 vs. 20%; odds ratio 27; p = 0.008). 10 patients in the HAART alone group had to be rescued with PLD during the study. The commonest side-effects were anaemia and neutropenia.

Little (2003) treated 26 people with advanced KS with liposomal doxorubicin (20 mg/m² IV every 3 weeks) with IL-12 (300 ng/kg) for 18 weeks followed by maintenance IL-12 (500 ng/kg twice weekly) to assess the response rate and progression-free survival and toxicity. Of 24 evaluable patients to date, the overall response rate is 83%. At one year, the probability of being progression free is 75%.

Nunez conducted a prospective, open label study of 79 HIV-infected patients with KS (more than 10 lesions or organ disease). All received 20mg/m² liposomal doxorubicin every 3 weeks in addition to combination antiretroviral therapy. 40% had a complete response, 38% had a partial response, 6% had stable disease, 4% had progression and 9 individuals were not evaluable. No factors significantly predicted response.

Bennett compared clinical trial data from separate, randomised, phase III trials of pegylated liposomal doxorubicin and liposomal daunorubicin in treating KS. Clinical response rates were 59% for liposomal doxorubicin and 25% for liposomal daunorubicin. Toxicities and cost-effectiveness were similar for the 2 treatments.

Gill (1995) enrolled 227 people in a Phase III trial comparing liposomal daunorubicin (40 mg/m²) versus doxorubicin/bleomycin/vincristine (ABV) every two weeks as first-line therapy of advanced KS. Participants had at least 25 mucocutaneous lesions, and/or symptomatic visceral involvement, and/or tumour-associated oedema, and no prior chemotherapy. In the daunorubicin, 3 people had a complete response and 26 had a partial response (overall response rate 25%). In the ABV group, one person had a complete response and 30 had a partial response (overall 28%). Liposomal daunorubicin was associated with lower rates of alopecia and neuropathy, and with higher rates of severe neutropenia.

Presant reported the results of Phase II studies of liposomal daunorubicin in 135 people with poor risk KS. Out of 123 evaluable participants who received 40 mg/m² every two weeks, 65% achieved partial response. Partial response was seen in 58% of participants with prior chemotherapy, 35% of those with prior doxorubicin and 71% of those with no prior chemotherapy. 1.6% achieved complete response. Complete or partial response was seen in 64% of participants with CD4 counts below 200, rising to 75% in those with CD4 counts above 200, and in 59% of participants with visceral KS. The median time to progression in responders was 172 days. Dose-limiting toxicity was related to myelosuppression, with grade 3 or 4 leukopenia occurring in 21.4% of cycles. Other significant toxicities (grade 2, 3 or 4) included fatigue (10%), nausea (4%) , alopecia (1%) and vomiting (2%). No cardiac problems were reported.

Uthayakumar enrolled 29 people with less than 20 discrete cutaneous KS lesions and no prior chemotherapy in a randomised cross-over study comparing 12 weeks of treatment with liposomal daunorubicin (40 mg/m² once every fortnight) versus no treatment. Analysis of response for the initial 12 weeks showed that in the treated group time to onset of progressive disease was prolonged. 6 of 15 treated participants had partial responses, versus none in the observed group. 40% in the treatment arm progressed compared to 72% of the observation arm. Neutropenia was documented following 14% of treatment cycles.

Northfelt reported the pooled results of two studies that enrolled people with biopsy-proven extensive, progressive mucocutaneous and/or visceral AIDS-related KS, and who had received no prior KS therapy. One study compared liposomal doxorubicin (20 mg/m²) every two weeks versus ABV (20 mg/m² doxorubicin; 10 units/m² bleomycin; 1.0 mg vincristine) every two weeks for a maximum of seven cycles. The second study compared liposomal doxorubicin every three weeks versus BV. A total of 254 participants received liposomal doxorubicin. An overall response rate of 52% was documented among the liposomal doxorubicin recipients, versus 25% with ABV and 23% with BV. Mucositis was more frequent in liposomal doxorubicin recipients, but otherwise there were fewer and less severe toxicities than among ABV or BV recipients.

Northfelt (1998) then conducted a randomised phase III trial comparing pegylated liposomal doxorubicin (20mg/m2) with doxorubicin (20mg/m2), bleomycin (10mg/m2) and vincristine (1mg) every 14 days for 6 cycles. 60 of 133 on the doxorubicin arm achieved a partial response and one person had full remission. In contrast, 31 of 125 had a partial response in the chemotherapy arm. The difference was statistically significant. Side-effects were fewer in the pegylated liposomal doxorubicin group.

Stewart randomised 241 people to receive six 3-weekly cycles of either BV (bleomycin 15 mg/m² plus vincristine 2 mg) or liposomal doxorubicin (20 mg/m²). Using ACTG criteria, participants receiving liposomal doxorubicin had a higher rate of complete or partial responses (58.7% versus 23.3%), and a shorter mean time to response. Liposomal doxorubicin was better tolerated in terms of nausea/vomiting, alopecia and neuropathy, but was more myelosuppressive. Discontinuations due to adverse events occurred in 26.7% of the BV group, versus 10.7% of the liposomal doxorubicin group, and fewer BV recipients completed the full six cycles of treatment (30.8% v 55.4%).

Goebel conducted a Phase II trial in which participants with moderate-to-severe KS were treated biweekly with liposomal doxorubicin. Almost all received doses of 10 or 20 mg/m², rising to 40 mg/m² in non-responders. Best response was determined for 238 participants and was achieved after a mean of 2.3 cycles. Using ACTG criteria, 15 (6.3%) had a complete response, 177 (74.4%) had a partial response, 44 (18.5%) had stable disease and 2 (0.8%) had disease progression. Degree of response did not seem to be correlated with dose. Ten participants discontinued therapy because of adverse events. Grade 3 or 4 neutropenia occurred after 281 of 2023 cycles (13.9%) but involved 137 of 240 participants (57.1%) for whom data were available. Updated results presented by the manufacturer to the European Agency for the Evaluation of Medicinal products (EMEA) found that the mean response duration is 117 days.

The manufacturer also presented data from a open label study in which 137 people with KS received liposomal doxorubicin at a median dose of 20 mg/m², repeated three weekly. 793 cycles were administered with a median cumulative dose of 110 mg/m². Response rates were: no complete responses, 61.8% partial responses, 26.5% stable and 11.8% progression. Median duration of response was 92 days.

Newell reported that among 22 patients whose KS had been successfully stabilised by liposomal doxorubicin (20 mg/m² every two weeks), 77% successfully maintained their response with a reduced dose of 10 mg/m².

Anti-herpes drugs for Kaposi's sarcoma

Morfeldt (1994, 1995) treated 5 people with AIDS who had recently developed KS with foscarnet (60 mg/kg every 8 hours for the first 5 days followed by 90 mg/kg every 12 hours for the next 5 days). In the first case, lesions became thinner and nearly disappeared over a 12-month period, but all reappeared 17 months after treatment. In the second case, all 15 lesions clearer and remained in remission 20 months after treatment. In the third case, all 29 lesions faded and the subject remained free from lesions 22 months after treatment. The remaining two cases had progressive lesions.

Jones and colleagues at the CDC, investigating the hypothesis that KS is caused by a herpes virus, conducted a retrospective analysis of 20,288 people with HIV or AIDS, excluding people with a previous diagnosis of KS. During the 14 months following January 1990, 1033 people developed KS. People who had taken foscarnet for any reason had a 70% reduced risk of developing KS compared with those who had not received foscarnet. No decrease in the rate of KS was seen among people treated with ganciclovir. People treated with acyclovir had a 40% increased risk of developing KS. The researchers speculate that the low doses of acyclovir used for maintenance therapy against herpes simplex could result in the development of acyclovir-resistant strains of the proposed KS herpes virus.

Glesby reported that in the Multicenter AIDS Cohort Study, a trend towards a reduced risk of developing KS was seen among participants who received foscarnet or ganciclovir, but not among those who received aciclovir.

Mocroft retrospectively analysed the relationship between treatment with acyclovir, foscarnet and ganciclovir and risk of Kaposi's sarcoma among 3688 HIV-positive people followed for a median of 4.2 years. 598 of those studied developed KS (16.2%). There was a decreased risk of developing KS among those who received foscarnet or ganciclovir, but not with acyclovir.

Costagliola conducted a review of 16,229 French people with HIV who did not have KS at the time of their first recorded CD4 count and had not participated in trials of antiretroviral therapy and/or acyclovir, foscarnet or ganciclovir. KS was diagnosed in 7% at some point during follow-up. The relative hazard for KS was slightly increased among people treated at any point with acyclovir (1.4); no association was seen between use of foscarnet or ganciclovir and modified risk of developing KS.

Angiogenesis inhibitors for Kaposis sarcoma

Tulpule conducted an open label trial of IM862 nasal drops at a dose of 5mg, randomised to 1 of 2 dosing schedules: 5 days on, 5 days off for 90 days (n=18), or every second day until disease progression or unacceptable toxicity (n=26). 44 people were enrolled. 47% had more than 50 lesions, 32% had lymphoedema and none had organ disease. 75% had previously had systemic chemotherapy. All but 5 were on concurrent PI-containing therapy. Response rate was 36% (5 complete, 11 partial), with a median response time of 6 weeks. 21 had stable disease for 7-72 weeks. Side-effects included: headaches, fatigue, tingling and nausea.

Pluda (1999) tested IL-12 - an immune modulator that has anti-angiogenic activity. 15 people with KS on stable ART were enrolled. IL-12 was given subcutaneously twice a week at the following doses: 100ng/kg (5 people), 300 ng/kg (6 people), 500 ng/kg (4 people). Most participants had had extensive prior treatment for KS. Three of 5 people on the 300 dose and 1 of 4 on the 500 dose had partial remission. Side-effects were common: fever, sweating, fatigue, aches and pains; 4 developed the blood disorder leukopenia, 1 person developed liver toxicity, and 1 person developed toxoplasma encephalititis.

Pluda (1997) reported preliminary results from a Phase I study of TNP-470, an analogue of fumagillin. 14 patients have been treated at doses of 4.6, 9.3, 15.4 and 23.2 mg/m² intravenously every other day. No dose-limiting toxicities were observed, although two participants developed small, asymptomatic retinal flare haemorrhages.

Nakamura reported that SP-PG, a sulphated polysaccharide-peptidoglycan compound produced by bacteria, controlled the in vitro growth of AIDS-associated KS-derived spindle-shaped cells at non-cytotoxic concentrations. Angiogenesis induced by KS cells in the chicken corioallantoic membrane assay was blocked by SP-PG, as was the angiogenesis associated with the induction of KS-like lesions by subcutaneous inoculation of human KS cells into nude mice.

Wellstein reported that pentosan polysulphate (PPS) showed strong inhibition of autocrine and paracrine growth stimulated by AIDS-KS cells. However, in a pilot trial conducted by the US National Cancer Institute it caused significant toxicity (hepatitis and thrombocytopenia) with no signs of efficacy.

Sidky reported that interferons are able to inhibit angiogenic lesions induced in mice and modulate growth factor production.

Sgadari reported that protease inhibitors are potent anti-angiogenic and anti-tumour molecules.

Retinoic acid

Other treatments for Kaposi's sarcoma

Multicentric Castleman's disease

Oksenhendler (2003) conducted a prospective study of multicentric Castleman's disease (MCD) among HIV-infected patients attending a Paris hospital. 70 cases were identified between 1992-2002. Median CD4 count at MCD diagnosis was 167 (range 1-1567) and 11 of 44 had undetectable virus at that time. 31 had pre-existing KS and a further 17 subsequently developed KS. High levels of HHV8 were detected in blood cells during attacks of MCD. Average survival was 60 months, with 9 deaths due to MCD and 25 deaths due to other opportunistic illnesses or cancers. 21 patients developed HHV8-associated non-Hodgkin's lymphoma and average survival once NHL occurred was less than 6 months.

Guihot (2003) studied pulmonary (lung) symptoms of Castleman's disease in 11 patients. Main manifestations were interstitial pneumonitis with fever, weight loss and malaise. Frequent recurrence occurred despite initially good responses to treatment with etoposide and vinblastine. Five people developed non-Hodgkin's lymphoma and four deaths occurred due to lymphoma.

Neuville (2003) treated two HIV-infected people with long-term Castleman's disease, who had been on chemotherapy for an average of four years, with rituximab. Both patients were taking antiretroviral therapy and had undetectable viral load. One had a CD4 count of 169, and the other had a CD4 count over 1000. Despite an initial response, clinical failure occurred at 4 and 20 weeks and chemotherapy had to be resumed. Furthermore, KS worsened in both patients requiring treatment with paclitaxel.