Pegylated liposomal doxorubicin plus highly active antiretroviral therapy versus highly active antiretroviral therapy alone in HIV patients with Kaposi's sarcoma

Twenty-eight HIV patients either naive or failing highly active antiretroviral therapy (HAART) with moderate–advanced Kaposi's sarcoma (KS) were randomly chosen to initiate a new HAART regimen plus pegylated liposomal doxorubicin (PLD) or the new HAART regimen alone. After 48 weeks, better response rates were observed in the HAART plus PLD group (76% versus 20%). In HIV-infected patients with moderate–advanced KS, HAART alone may not be enough for KS response. Since the introduction of highly active antiretroviral therapy (HAART), the incidence and mortality of Kaposi's sarcoma (KS) in HIV-infected individuals has declined dramatically [1,2], KS response rates have improved, and relapses have decreased when HAART has been administrated with chemotherapy [3]. However, the exact role of HAART alone in this tumour is controversial. Whereas KS lesions have improved in some patients when HAART was introduced [4], cases of disease progression have also been described [5]. The aim of this study was to compare the evolution of KS in HIV patients receiving pegylated liposomal doxorubicin (PLD) together with HAART with patients receiving HAART alone. This was a randomized, open-label, multicentre study conducted by the Caelyx/KS Spanish Study Group. HIV patients with KS were randomly selected to receive HAART plus PLD from the beginning of the study (group A) or to receive HAART alone (group B). The inclusion criteria were adult HIV–KS-affected patients (biopsy confirmed) without HIV-RNA control, whether they were naive, without HAART, or with failing treatment. Patients had to have at least 10 cutaneous lesions or mucosal or visceral involvement. Patients with life-threatening KS were excluded. Intravenous PLD was administrated at doses of 20 mg/m2 every 3 weeks. The total number of cycles was decided by each investigator. Patients allocated to group B who had KS progression or did not respond to HAART alone were considered to be non-responders and were allowed to receive PLD. Response rates were calculated as an intent-to-treat analysis, when missing cases were considered as failures, and as on-treatment analysis, when missing cases were excluded. Univariate and multivariate logistic regression analyses were performed for the identification of factors associated with response. Those factors with P < 0.2 in the univariate analysis were entered into the multivariate test. Twenty-eight patients were included in the study (group A, 13; group B, 15). Baseline characteristics are shown in Table 1 [6]. The KS stage was heterogeneous among the participants. The extremes were represented by five (18%) with all prognostic predictors being good (T0I0S0) and two (8%) with all being bad (T1I1S1). Five patients had visceral involvement: three with lung and two with gastric disease. The only statistical difference between the groups was that the prevalence of previous opportunistic infections was higher in patients allocated to group A.Table 1: Characteristics and response of study patients.At 48 weeks of follow-up two patients in each group were lost. A complete or partial remission was observed in 13 patients. Response rates were clearly different between the groups (Table 1); in intent-to-treat analysis, 76 versus 20% for groups A and B, respectively (P = 0.003); in on-treatment analysis, 91 versus 23% for groups A and B, respectively (P = 0.003). Ten patients in group B had to be rescued with PLD; in seven of them progression was identified during the first 3 months of therapy; in two a delayed progression was observed during months 5 and 6, and in the last one no amelioration was observed after 9 months of HAART. With regard to group B, the baseline characteristics of patients who had tumor progression were similar to those who achieved complete or partial remission. In the univariate analysis, factors related to response were the use of PLD, higher baseline HIV viral load, and being SI in KS stage. However, the only factor that remained related to response in the multivariate logistic regression analysis was the use of PLD [odds ratio (OR) 27; 95% confidence interval (CI) 2.3–307; P = 0.008]. Twenty-three patients received at least one cycle of PLD: 13 in group A and 10 in group B. Two of them were lost to follow-up. The median (interquartile) number of cycles received was 11 (7.2–14.7). One or more adverse event was identified in 33% of the patients. The most common events were anaemia and neutropenia, which appeared in three cases. Other toxicities were mucositis (two cases), enteritis, hepatotoxicity and fever (one case each). Most events were mild, and in no case did PLD have to be interrupted. In three subjects doses had to be reduced. Two patients died while they were receiving PLD. Their physicians did not relate these cases to PLD or KS. To our knowledge this is the first randomized study conducted to assess the direct effect of HAART on HIV-related KS regression. Our results indicate that HAART alone is not enough to produce tumour regression in most cases of moderate–advanced KS. However, the good response rate (76%) and tolerance of the combination of HAART plus PLD supports the fact that systemic therapy plays an important role in the HAART era. Our results do not fit with those of other authors, which have shown more cases of KS regression with HAART alone. However, most of those studies included few patients, with less advanced KS, were not randomized, and mixed patients treated with HAART alone with patients treated with specific KS therapy plus HAART [7–12]. All these reasons can explain the different results. Dupont et al. [4] observed more than a 50% complete remission rate in patients not receiving systemic tumour therapy. However, as it was not a randomized study, patients with more severe disease were probably receiving some type of chemotherapy. Besides, one third of the remaining patients received local therapy, which could also contribute to better complete remission rates. Two-thirds of the patients who initiated HAART alone had to be rescued with PLD. In all but one, KS progression was identified, and in most tumour progression occurred during the first 3 months of HAART. Zala et al. [5] observed some cases of KS that appeared after virological and immunological response to HAART, and suggested an immune recovery syndrome. Similarly, in two of our patients progression appeared with virological and immunological control. However, the rest of the cases occurred before an increase in the CD4 cell count was obtained. In this context, it is more probable that HAART was not able to control the natural progression of the disease than that there was an immune-mediated KS deterioration. In conclusion, although HAART alone may be enough for patients with less advanced KS, in cases of moderate–severe disease better response rates are observed when PLD is given with HAART. Members of the Caelyx/KS Spanish Group Coordinators: J. González-Lahoz, Instituto de Salud Carlos III, Madrid; D. Podzamczer, Hospital Universitari de Bellvitge, Barcelona. Participants: M. Alegre, Hospital Sant Pau, Barcelona; M. Alsina, Hospital Clínico, Barcelona; A. Barrios, L. Martín-Carbonero and E. Valencia, Instituto de Salud Carlos III, Madrid; A. Berrocal, Hospital General, Valencia; M.J. Galindo, Hospital Clínico, Valencia; J. Martínez-Lacasa, Hospital Mutua de Terrassa, Barcelona; A. Ocampo, Hospital Xeral Cies, Vigo; R. Palacios and J. Santos, Hospital Universitari Virgen de la Victoria, Málaga; J.D. Pedreira, Hospital Juan Canalejo, La Coruña; P. Saballs, Hospital del Mar, Barcercelona; G. Sirera, Hospital Germans Trias y Pujol, Barcelona. Acknowledgements The authors are very grateful to Carlos Toro and Victoria Jiménez Baena for their work on human herpesvirus 8 measures.