Highly active antiretroviral therapy (HAART) has profoundly influenced the natural history of AIDS-associated Kaposi's sarcoma (KS). Not only has the incidence of KS declined sharply since HAART became widely available [1-10], but HAART has also been associated with lengthening of the time to treatment failure [11] and longer survival among patients with pulmonary KS who have received chemotherapy [12]. In addition, regression of KS after institution of HAART is well documented [13-36]. The responsible mechanisms are not well elucidated, but restored immunity to the KS-associated herpesvirus, decreased levels of angiogenic factors that stimulate KS proliferation, and direct angiogenesis-inhibitory effects of HIV-1 protease inhibitors could all be involved.

The benefits of HAART notwithstanding, KS has not disappeared as a clinical problem. Because HAART can induce tumor regression, and because appropriate HIV management generally requires administration of HAART, distinguishing the antitumor effects of HAART from those induced by an agent under study for KS treatment presents difficulties. The importance of this issue was brought to the forefront by a recent US Food and Drug Administration (FDA) decision concerning pegylated liposomal doxorubicin (Doxil, Ben Venue Laboratories, Bedford, OH), which received initial marketing approval for treatment of KS in 1995 under the accelerated approval mechanism. Full approval required demonstration of clinical benefit in a phase IV clinical trial, the design of which was developed in consultation with the FDA. Eligible patients had advanced AIDS-associated KS and one or more specific tumor-related symptoms, the progress of which was monitored for evidence of clinical benefit. Of 60 pegylated liposomal doxorubicin-treated patients, 48 (80%) showed clinical benefit [37]. Because the start of this trial coincided with the widespread introduction of HAART for HIV treatment, before its effects on KS had been described, some patients began HAART either shortly before or shortly after chemotherapy was begun. The FDA concluded that these changes in antiretroviral therapy confounded assessment of the results, which led to a regulatory decision denying full approval. In subsequent discussions regarding potential trial designs, the results of which would be unambiguous, it was suggested that patients must either have been receiving stable and effective HAART therapy for an extended period before study enrollment, or should be randomly assigned to receive either HAART alone or HAART plus chemotherapy.

Although such a plan initially appears quite reasonable, there are several obstacles to its implementation. For pegylated liposomal doxorubicin, the fact that the drug is both available and preferentially prescribed as first-line therapy for advanced KS would make it difficult to enroll patients onto such a trial in a timely fashion. A more generic issue is the shrinking numbers of patients with KS that is sufficiently advanced to justify systemic chemotherapy. In addition, although there are a few reports of KS development or progression during HAART, despite effective HIV suppression and improved CD4 counts [38-42], many patients with the most advanced KS are individuals in whom HAART was ineffective, so requiring effective HAART as a precondition of study entry would be inappropriate.

An even more serious concern, however, is whether the available data support the FDA's position that advanced, symptomatic KS frequently responds to HAART alone (and, by implication, results in clinical benefit). To address this question, a search of the English-language literature was conducted, yielding 24 primary references (14 patients in 12 case reports and 177 patients in 12 prospective series) that described the effects of HAART on KS [13-36]. For each article, data on concomitant KS therapy, KS stage, response criteria, response proportion, and time to response were abstracted.

Inconsistencies in study design hindered the ability to reach firm conclusions: KS staging was not uniform, response was not consistently measured, some patients did not have biopsy-confirmed KS, entry criteria were not always specified, one study specifically excluded patients with life-threatening KS [26], and it was not always possible to determine the response of patients according to the severity of their KS. Nonetheless, an attempt was made to classify patients according to their AIDS Clinical Trials Group tumor stage, where T0 denotes good risk (ie, KS confined to the skin, and/or lymph nodes, and/or minimal oral disease) and T1 denotes poor risk (ie, symptomatic lymphedema, tumor ulceration, extensive oral disease, and/or visceral involvement) [43]. The T1 and T0 groupings have been associated with distinct survival distributions [44,45]. Although in general, T1 denotes advanced KS and T0 denotes less-advanced disease, some patients with T0 KS may have extensive, disfiguring skin disease that interferes with function and requires intervention, whereas some patients with T1 KS (eg, those with asymptomatic gastrointestinal lesions found incidentally, small oral nodules, or mild lymphedema) may not need immediate treatment.

Sufficient information was given to divide patients by tumor stage in 11 prospective studies that included 156 patients [25-33,35,36]. Of these patients, 102 (65%) had good risk (T0) and 54 (35%) had poor risk (T1) KS. Of these 11 studies, there were nine [25,27-32,35,36] that included 122 patients for whom there was sufficient information to determine how many individuals with T0 or T1 disease, respectively, received HAART alone or HAART with specific anti-KS therapy. Of these 122 patients, 83 had T0 KS, of whom 62 (75%) received HAART alone and 21 (25%) received HAART with KS therapy. Of the 39 patients with T1 KS, only nine (23%) received HAART alone, whereas 30 (77%) received HAART together with KS therapy. However, the largest of these nine studies [31], which included 39 patients, gave insufficient information to determine response rates for patients in each of the four subgroups (ie, T0 or T1; HAART alone or HAART with KS therapy). Thus, response rates according to both KS stage and treatment could only be ascertained for 83 patients, of whom 62 had T0 and 21 had T1 KS. Forty-eight patients had T0 KS that was treated with HAART alone, of whom 39 (81%) showed KS regression. Of the 21 T1 patients, however, only four received HAART without concomitant KS therapy, of whom three responded. None of these four patients had visceral KS and all were classified as T1 on the basis of lymphedema or mucosal involvement. Furthermore, in only two single-patient case reports [20,22] was unequivocal regression of well-documented visceral KS described after treatment with HAART alone. Thus, the entire literature documents only five cases in which patients with T1 KS responded to HAART in the absence of concomitant KS therapy.

Time to response was given in only six of the 12 prospective studies [27,30,32-34,36], with medians ranging from 3 to 9 months. In one study [33], the median time to response was significantly shorter (P = .003) for patients not receiving chemotherapy at baseline (123 days) than for those with more advanced KS receiving chemotherapy (314.5 days). In another study [34], the authors remarked that patients whose KS responded “tended to have less severe disease at baseline...than did those who had progressive disease,” and documented only one partial response among eight patients who received concomitant KS therapy. One of the eight patients, who received concomitant radiation therapy to cutaneous KS lesions, died from rapidly progressive pulmonary KS “before chemotherapy could be instituted” [34]. In yet another study [31] in which patients had their KS treatment tapered during HAART according to their KS response, 20 of 39 patients were receiving chemotherapy at baseline; chemotherapy was classified as minor (bleomycin or vinblastine) in 13 patients, and major (the combination of doxorubicin, bleomycin, and vincristine, or a liposomal anthracycline, or paclitaxel) in seven patients. KS response was recorded at 6-month intervals. At month 6, there were 10 patients receiving major chemotherapy, nine receiving minor chemotherapy, and two patients had started to receive miscellaneous treatments (interferon alfa and all-trans-retinoic acid). At month 12, 12 patients were still receiving chemotherapy (classified as major in nine patients). Of the 20 patients receiving chemotherapy at baseline, only 35% had achieved a partial response after 6 months of HAART, whereas the 19 patients who were not receiving systemic chemotherapy at baseline showed a 64% response rate at the same time point. The semiannual timing of response assessments precludes any statement about the median time to achieve a response. However, although the eventual response rate was high, responses occurred slowly among patients who had KS of sufficient severity to warrant systemic chemotherapy.

Even if we accept that there is likely to be significant publication bias favoring reports of KS regression during HAART (versus reports to the contrary), the available evidence suggests that HAART often leads to regression of limited KS in patients who have not received HAART previously. For such patients, it is appropriate to require an extended prior period of stable HAART before initiating new KS treatments. There are, however, insufficient data on which to base conclusions about the activity of HAART in patients with advanced, symptomatic disease. In most prospective studies, patients with advanced KS received chemotherapy either from the outset of HAART or had it added after KS failed to respond to HAART alone. The KS lesions of such patients regressed less often, and more slowly, than the lesions of patients with less severe KS who received only HAART. Once KS regression was achieved with a combination of HAART and systemic KS therapy, however, chemotherapy could often be withdrawn. In an instructive case report, Vaccher et al [38] described a patient whose T1 KS progressed during initial HAART therapy despite suppression of HIV viremia and increasing CD4 counts. Chemotherapy was then added, and the patient achieved a partial response. On discontinuation of chemotherapy, residual KS lesions regressed completely when the patient received HAART alone during the next 5 months.

Thus, although HAART is an important component of treatment for all patients with KS, there is scant evidence that HAART induces regression of advanced, symptomatic KS without concomitant KS therapy, nor have any of these studies addressed clinical benefit. Therefore, the available data do not justify clinical practice or study designs in which KS therapy is withheld from symptomatic patients while awaiting a response to HAART. This presents a challenge in designing clinical trials to assess the benefits of chemotherapy, which is usually reserved for patients in whom KS is most advanced. In such patients, initiation or modification of HAART may be needed at the same time KS-specific therapy is started. Although concurrent initiation of HAART and KS therapy may make it more difficult to determine their relative contributions to KS response, on the basis of what is already published, it should be possible to develop mutually agreed on benchmarks for time to response and clinical benefit. These benchmarks would fulfill the needs of investigators and regulatory agencies, while at the same time giving patients with advanced KS access to treatment.

Author's Disclosures of Potential Conflicts of Interest

Section:

ChooseTop of pageAuthor's Disclosures of P... <<REFERENCES

The following authors or their immediate family members have indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. Acted as a consultant within the last 2 years: Susan E. Krown, Ortho Biotech Products LP. Received more than $2,000 a year from a company for either of the last 2 years: Susan E. Krown, Ortho Biotech Products LP.

REFERENCES

Section:

ChooseTop of pageAuthor's Disclosures of P...REFERENCES <<