July 25, 1994
Treatment should be individualized based on the prognostic parameters and the desired effect of therapy. Many researchers believe that immediate treatment is not necessary at the first episode of KS with a few small lesions and that a "watchful waiting" period would be in order to determine the course of the disease (Levine 1993).
During this initial observation period, a biopsy confirmation is crucial for many reasons other that histological confirmation (Tappero 1993 b). First, a biopsy confirmation will undoubtedly be needed when a patient is later referred for radiation therapy, chemotherapy, or volunteers for a clinical trial. Secondly, a tumor biopsy is in order to avoid confusion with bacillary angiomatosis. Lastly, biopsy reports from pathology laboratories go into tumor registries which are major sources for compiling surveillance data and tracking disease trends.
If the patient has rapidly progressing KS, the physician should take notice and thoroughly check for previously undiagnosed OIs. Sometimes KS appears aggressive and is followed by the occurrence of OIs (Donald Abrams, personal communication).
However, if the patient begins with only a few lesions, this does not mean that a primary care physician should do nothing. The trauma and disfigurement of the KS lesion, a social stigma and constant visual reminder of one's HIV disease, can take a serious psychological toll on patients. Indeed, one clinician contends, "a patient's psychological distress may bear no relationship to the actual extent of cutaneous involvement by KS." (Schwartz 1992) Physicians must explain to their patients the various courses the disease might take and assure them that there are various treatment options at the different stages that can be palliative, limit morbidity and possibly prolong survival. In fact, treating a patient with asymptomatic indolent lesions might be appropriate for some simply for its psychological benefit (Miles: in press).
Liquid nitrogen cryotherapy is often used by dermatologists for the palliation of minor, yet cosmetically unacceptable lesions on the face, neck and hands. Cryotherapy has been shown to be safe and relatively inexpensive. Patients generally receive two freeze-thaw cycles of liquid nitrogen, administered with cryospray for 30 to 60 seconds in each cycle.
In 1991, Tappero and colleagues (Tappero 1991) from San Francisco General Hospital conducted a 20 patient, non-controlled, phase II efficacy and toxicity trial of liquid nitrogen for cutaneous KS. Complete responses in 80% of the treated lesions were obtained and lasted from six weeks to six months. The best cosmetic results were seen in macular and papular lesions less than 1 cm in diameter. Treatment was repeated at three week intervals, and on the average, patients received three treatments per lesion with a mean follow up time of 11 weeks. Minor pain and blistering did occur, but there were no secondary infections. In addition, a response to cryotherapy had no bearing on the patients' CD4 counts, presence of "B" constitutional symptoms, or previous treatment with chemotherapy. Patients with prior OIs, however, did not respond as well to therapy.
Cryotherapy does give some cosmetic benefit, but this is often accomplished by superficial scarring. The scarring results because the liquid nitrogen basically freezes/burns the lesions. Some patients might find this better than nothing at all, but they should realize that KS is often persistent in the deep reticular dermis of the treated lesion (Tappero 1993 b). Cryotherapy is also mainly recommended for fair skinned individuals because liquid nitrogen causes hypopigmentation leaving a flat, white scar (Krown 1992 a). This white scarring may be more noticeable than the original KS lesion itself for individuals with dark skin.
Because lesions usually blister and then form a crust, it has been strongly advised (Tappero 1993 b) that patients keep the treated site covered until well healed because HIV has been found in blister fluid (Supapannachart 1991).
Intralesional vinblastine has also proven to an be effective treatment for the palliation of minor lesions. Unlike liquid nitrogen, vinblastine can be used for bulkier lesions and symptomatic oral lesions.
Newman and colleagues (Newman 1988) conducted an uncontrolled, 15 patient trial in which they injected vinblastine into a total of 190 lesions. Each lesion was injected with 0.1 - 0.2 mg of vinblastine every two weeks. There was a complete resolution (flattening or depigmentation) of 25 lesions (13%) in 4 patients, a partial resolution of 149 lesions (78%) in 8 patients, and no response was seen in 16 treated lesions (8%) in 3 patients. Local pain from the injections was noted in all patients, and 90% had minor skin irritation.
Recommended doses of intralesional vinblastine for cutaneous lesions are: 0.1 mg (0.5 ml of a 0.2 mg/ml solution) injected per square centimeter of lesion; and for those not responding to 0.1 mg/cm2, incremental doses to a maximum of 0.2 mg/cm2can be used (Tappero 1993 b).
Both studies -- and other practicing physicians -- have noted that while intralesional vinblastine is efficacious, there is often pain lasting one or two days after injection, and the post-inflammatory hyperpigmentation (brown spots) that occurs might be cosmetically unacceptable to light and fair skinned individuals (Timothy G. Berger, personal communication). Other toxicities not uncommon to intralesional vinblastine include a transitory flu-like syndrome, intense edema (swelling around the treated lesion), and erythema (blotchy red patches). Most recently, researchers warned against using intralesional vinblastine around thick hair-bearing sites such as the eyebrows and near the hairline because alopecia (baldness) might occur and also caution against intralesional use near large peripheral nerves to avoid mono-neuropathy (Tappero 1993 a).
Symptomatic oral lesions also respond well to intralesional vinblastine. Epstein and colleagues (Epstein 1989) completed a 10 person trial of intralesional vinblastine for KS of the palate. One lesion was treated while another lesion was used as a control. At baseline, most of the patients had concomitant oral infections, including hairy leukoplakia, candidiasis and ulcerated gingivitis.
Patients in this study received one or two intralesional injections of vinblastine (0.2 mg/mL, at 0.1 mL per 0.5 cm2of KS lesion). The mean reduction of the palatal lesion area after one or two treatments was 63%. All patients were noted to have responded to treatment. Four had a "significant" response, two had a partial response, and two had a "minor" response. There was increased size in seven out of the ten control lesions. There was no increase in the oral opportunistic infections. Mild to moderate pain was noted in all patients which lasted one or two days.
In a larger and more recent 42 patient non-controlled trial of intralesional vinblastine for oral KS, Epstein noted a greater that 50% reduction in the lesions of 74% of the patients (Epstein 1993). For patients not lost to follow-up, the mean duration of response was 4.25 months.
An alternative to intralesional vinblastine for oral KS may be the injection of 3% sodium tetradecyl sulfate (Sotradecol) which is commonly used in eliminating vascular lesions of the oral tissue (Goebel 1976). While this agent is only being used by a handful of physicians to treat oral KS, Drs. Lucatorto and Sapp of the School of Dentistry at UCLA recently published results of a preliminary study of injecting Sotradecol into oral KS lesions (Lucatorto 1993). In treating 15 lesions in 12 patients with one or two injections, they noted 8 complete responses and 4 partial response with no recurrence for as long as 18 months. Unlike intralesional vinblastine, initial injections of Sotradecol caused ulcerations in all treated lesions, but healing began by the middle of the second week. Best results were obtained in lesions that were 2.6 cm or less in size.
Lucatorto and colleagues are in the process of planning a comparative efficacy trial of 3% sodium tetradecyl sulfate (Sotradecol) vs. intralesional vinblastine for oral KS (Frank M. Lucatorto, DDS, MS, personal communication).
Initial in vitro studies of TNF-alpha demonstrated a vascular inhibitory effect on endothelial cells (Sat 1986) and also induced necrosis of solid transplantable tumors (Carswell 1975). TNF-alpha was then tried in the late 1980s as a therapeutic modality, both intravenously and intralesionally, for KS. Aboulafia and colleagues (Aboulafia 1989) at UCLA tested intravenous recombinant TNF-alpha in a five patient study. All five patients receiving TNF-alpha intravenously had some progression in their lesions. The noted toxicities were fatigue and arthralgia (pain in the joints).
Lesion reduction was seen in a 27 patient, double-blinded, randomized, placebo-controlled study of recombinant intralesional TNF-alpha conducted by Kahn and colleagues at San Francisco General Hospital (Kahn 1989). Each patient had one lesion injected with TNF-alpha and one lesion injected with saline as a control. Of 16 evaluable patients, 15 showed response (a reduction in the cross-sectional area of the TNF-injected lesion) to therapy compared with a 7% response in the placebo controlled lesions. Of the TNF-alpha treated lesions, there were 3 CRs, 5 PRs, and 8 minor responses. Because of the toxicities observed -- fever, rigors and chills -- and the lack of antitumor effects in non-injected lesions, the investigators said they doubted the usefulness of intralesional TNF-alpha as a treatment for KS.
Intralesional recombinant interferon-alpha is still considered an experiment treatment for cutaneous KS (Tappero 1993 b). In a 5 patient, placebo-controlled trial, Sullis and colleagues (Sulis 1989) injected 3 to 5 million units of recombinant IFN-alpha-2b three times a week for four weeks into skin and oral lesions. There was a complete resolution of all lesions treated, whereas no response was seen in the untreated lesions. It should be noted that two of the five treated patients had previously been receiving systemic IFN for four month with no results.
Various kinds of laser therapy -- argon, carbon dioxide, and pulsed-dye (PDL) -- have been used for cutaneous and oral KS over the past years (Wheelard 1985; Tappero 1992). The moderate response rates achieved in treating AIDS-KS using these three types of lasers have never been acceptable in light of the cost and time consuming nature of the treatment. There have also been reports of infectious virus particles from laser smoke (Baggish 1991).
A recent 15 patient study of PDL conducted by Tappero and colleagues (Tappero 1992) at San Francisco General Hospital documented a complete or partial response of 44% (17 of 39) for treated lesions. There was an 18% response rate (7 out 39) for the untreated control lesions. After 12 weeks, all treated lesions recurred. Thus, the investigators concluded that PDL should not be recommended for AIDS-KS.
A new approach to laser therapy has come about with the use of photodynamic therapy (PDT) for AIDS-KS (Schweitzer 1990; Hebeda 1993). Safety and efficacy trials are currently underway at San Francisco General Hospital and Roswell Park Cancer Institute in Buffalo. The Buffalo group invented PDT 15 years ago and have been using it for a myriad of cancers.
Bernstein and colleagues (Bernstein 1994) at Roswell Park Cancer Institute recently completed a 10 patient, PDT pilot study. Of 6 evaluable patients, 5 had complete responses for a minimum period of 60 days. The remaining patient had a "mixed" response and was retreated to a complete response for a period of 80 days. No significant toxicities or change in CD4 counts was noted.
Roswell Park is currently running a phase I/II trial which consists of injecting Photofrin, a light sensitizing chemical into the vein, waiting two or three days, then zapping the lesions -- which have a 5 fold retention of Photofrin over normal skin -- with an argon laser. They are observing significant response rates that are lasting up to 8 months in some patients (Thomas Mang, personal communication). At higher doses, the hyperpigmentation goes away and has shown beneficial, cosmetic results in facial lesions.
The San Francisco group is substituting Tin Etiopurin (SnEP2), a newer light sensitizing chemical, in place of Photofrin. They are trying to determine the optimal waiting time, after injecting Tin Etiopurin, to initiate laser therapy (Margaret Whitfeld, personal communication). It is important to stress that Photofrin and Tin Etiopurin make the skin extremely sensitive to light. Thus, exposed normal skin can develop a severe sunburn-like reaction.
AIDS-KS lesions respond well to radiation therapy (RT) but response rates and time to remission are significantly less than those seen for classic KS. In 1981, Nisce and colleagues (Nisce 1982) at Memorial Sloan-Kettering in New York studied 20 patients with classic KS using 400 rads (100 rads3= 1 Gy) of subtotal or total skin electron beam therapy once a week for six weeks and noted an 85% CR with a median duration of 48 months. The duration of remission seen in AIDS-KS can be as short as 6 months and as long as 3 to 4 years (Nisce 1993).
RT may be necessary and most beneficial for certain KS lesions on various parts of the body including: large KS tumors or plaques of the face, periorbital edema (swelling around the eyes) and lesions on the eyelid, lesions on or around the penis, and painful lesions on the weight-bearing areas of the body, such as the sole of the foot. RT is commonly used to reduce painful lymphadenopathy (swelling of the lymph nodes) and edema (swelling) on legs and the soles of their feet that inhibit mobility. While in agreement with using RT in these settings, some researchers warn against believing that RT is "the answer" for KS-associated lymphadenopathy and edema. Late complications of RT , include fibrosis, ulceration, and superinfection. Close monitoring of the areas is also recommended and crucial because the lymph nodes might be filled with KS or have an underlying infection, and over radiating might aggravate the edema and cause further fibrosis (Patrick Swift, personal communication).
Attempting to identify the best dosing regimen, Bernson and colleagues (Bernson 1990) studied 187 AIDS-KS patients at UCSF comparing 8 Gy in a single fraction (one dose) to 15 to 40 Gy in 5 - 10 fractions (doses over a period of time). Responses were defined as a complete flattening of a lesion or a decrease in the dimension of the lesion to at least 50% of the pretreatment size with decrease in pigmentation. There was a response to treatment in over 90% of the treated lesions with a median time to progression of 21 months. There was no difference in the response rates regardless of fractionation regimen used. The severe reactions, such as blistering and pain, which occurred with 17% of the treated fields were mostly seen in the smaller fractionated doses. The results from this study, and the results from another study in Amsterdam (de Wit 1990), have led some researchers to conclude that single fraction of 8 Gy is most effective for large areas because it is more time efficient, cost effective and limits severe reactions.
The treatment of symptomatic and asymptomatic oral lesions has been debated for a number of years. Nisce, who goes as far as saying, "RT of the oral mucosa had rather disastrous results," (Nisce 1993) was noted for advocating against using RT for oral KS (Lourdes Nisce, personal communication). The dead tissue that falls off after oral RT can cause severe pain and malnutrition because adequate normal tissue does not grow back quickly enough. It often takes 2 to 3 weeks to resume normal eating which can result in weight loss of 10% to 15% (Nisce 1993). Also RT of oral mucosal lesions often induces severe mucositis (Krown 1992 a).
Dr. Nisce has changed her mind about the use of RT for oral KS because of preliminary findings from a study she is conducting in which the inflamed oral lesions are premedicated with Oratec Gel (MGI-209), a topical benzocaine anesthetic (Lourdes Nisce, personal communication). Nisce has seen excellent results in the patients treated with MGI-209 before RT, and is then able to give daily low doses of 1.6 to 1.8 Gy until maximum doses of 25-30 Gy are reached. Patients are able to eat after treatments and there is no apparent weight loss (Nisce 1993).
Interferons have been extensively studied and used to treat AIDS-KS since the early 1980s (Krown 1983). We have learned more about interferons -- their effectiveness in certain patient populations, toxicity profile and possible use in combination with nucleoside analogues -- than any other anti-KS medication. Interferons became a logical therapeutic candidate to treat KS because early trials demonstrated its immunoregulatory, anti-proliferative and antiviral activity, (Borden 1981, 1982). Interferons can inhibit the replication of HIV in vitro (Poli 1989) and there is evidence that suggests interferons might inhibit lesion-associated angiogenesis (Sidky 1987).
IFN-alpha-2a (Roferon-A made by Hoffmann-La Roche) and IFN-alpha-2b (Intron A made by Schering-Plough) are approved by the Food and Drug Administration (FDA) for KS in patients with CD4 counts over 200/mm3. No other medication has been "specifically" approved for the treatment of KS. The two agents were reviewed by the FDA's Center for Biologics, and not by its Divisions of Antiviral or Oncologic Drugs.
It is difficult to compare the available data from the many interferon KS trials over the years due to the varying sizes of cohorts, the absence of uniform staging systems, the different preparations of interferons and the myriad of various dosing regimens (Rozenbaum 1990). Moreover, there has not yet been a large, randomized, placebo-controlled trial of IFN-alpha as a single agent for KS (Miles, in press). Except for one French study (Rozenbaum 1990) that followed patients from 1 to 6 years, follow-up on most trials has been short. Thus, we have little if no long-term efficacy data on the use of interferons for KS.
In most studies, IFN-alpha, as a single agent, given intramuscularly or subcutaneously in moderate to high doses (at least 20 million units per square meter body-surface area) has been shown to induce regression of KS lesions in roughly 30% of patients (Lane 1988; Real 1986; Evans 1991; de Wit R 1988). Response rates under 10% were seen with a low-dose regimen of 1.5-2.0 MU/m2(about 3 MU) (Real 1986).
In many of these trials, IFN-alpha response rates were higher for those patients who had CD4 counts of over 200, no previous OIs and no "B" symptoms (Lane 1988; Evans 1991; de Wit 1988). A patient's CD4 count is undoubtedly the major predictor of IFN-alpha response. An IFN-alpha trial conducted by Lane and colleagues (Lane 1988) revealed that the 5 patients with over 400 CD4 cells had substantial reductions in tumors whereasnoneof the 7 patients with under 150 CD4 cells showed a response.
There is little published information on the duration of response with IFN-alpha. A study conducted by Real and colleagues (Real 1986) determined that the median response duration was 24 month for complete responders and 11.5 months for partial responders. In a long-term follow-up study (followed 1 to 6 years) of 120 AIDS-KS patients treated with IFN-alpha, Rozenbaum and colleagues (Rozenbaum 1990) concluded that the median survival time for all patients from onset of treatment was 794 days; 1,723 days for responders and 375 for non-responders. Both studies claim that "responders" (patients with a CD4 count of over 200 at baseline) to IFN-alpha had a greater probability of survival. Moreover, Evans and colleagues (Evans 1991) stipulate that even among patients with a CD4 count of over 200, those who responded to therapy (had a regression of their tumors) had a distinct survival advantage.
The FDA approved recommended dose of recombinant IFN-alpha-2a (intramuscularly or subcutaneously daily) as a single agent is 36x106 U daily. And for IFN-alpha-2b, the recommended dose is 30x106 U/m2 subcutaneously three times a week.
IFN-alpha has also been noted for its toxicities which include: a "flu-like" syndrome characterized by myalgia, fever, fatigue and chills, anorexia, weight loss, nausea, diarrhea, confusion, hematological toxicities such as neutropenia and anemia as well as elevated liver enzymes (Pluda 1993 a). These symptoms are germane to IFN-alpha treatment for KS but might be exacerbated by overlapping or undiagnosed OIs. These symptoms can sometimes be reduced (or eliminated) by gradual dose-escalation, or in some cases, a temporary discontinuation of treatment and subsequent dose reduction (Krown 1992).
Zidovudine (AZT) alone is not an effective drug for the treatment of KS (Valentine F, et al., unpublished ACTG 001 data; Lane 1989; de Wit 1989). Even though AZT induced a significant rise in CD4 cells, there were few or no tumor responses (de Wit 1989). Possible reasons for AZT's lack of efficacy on tumor regressions are (1) AZT might not suppress HIV replication sufficiently to inhibit the release of the many cytokines involved in tumor growth and (2) various co-factors other than HIV play a more substantive role in the pathogenesis of KS (Krown 1992 a).
In vitro studies, however, have demonstrated a synergistic suppression of HIV replication when IFN-alpha and AZT were combined (Hartshorn 1987) as well as a synergistic antiviral activity in mice infected with a murine retrovirus (Ruprecht 1987, 1990). These findings gave rise to several clinical trials evaluating this combination.
Phase I and II combination AZT/interferon alpha trials which demonstrated response rates between 40% and 50% (Kovacs 1989; Krown 1990; Fischl 1991) were substantially higher than those seen with high dose IFN-alpha alone. A significant percentage of patients with CD 4 counts below 200 (as high as 30%) responded, and for the first time, low dose IFN-alpha (under 20 million unit) demonstrated antitumor activity. Because of the overlapping myelotoxicities of these agents, it was no surprise that neutropenia (as high as 57%) (Kovacs 1989) was the major dose limiting toxicity.
The use of colony stimulating factors (G-CSF and GM-CSF) has helped to considerably lessen the neutropenia seen when IFN-alpha and AZT are used in combination (Scadden 1991; Krown 1992 b). One study conducted by Scadden and colleagues (Scadden 1991) administered GM-CSF to 19 of the 26 patients (66%) who developed an absolute neutrophil count (ANC) below 1,000 ml. (This is not so shocking considering patients were receiving a high dose of 1,200 mg a day of AZT.) All patients who went onto GM-CSF had a prompt ANC increase, but there was no difference in tumor response or CD4 count for those who went onto GM-CSF versus those who didn't. Krown and colleagues (Krown 1992 b) noted that GM-CSF helped maintain the patients' ANCs. The major side affects were "flu-like" symptoms including malaise, fever, and fatigue. These non-hematologic toxicities prevented major dose increases of IFN-alpha and AZT.
Most oncologists -- in every day practice -- use G-CSF over GM-CSF to ameliorate the neutropenia caused by IFN-alpha. Although G-CSF's effects have never been proven in an IFN-alpha clinical trial, oncologists prefer it because the of the impression that "flu-like" symptoms from GM-CSF and IFN-alpha combined are sometimes too overwhelming for patients (Susan E. Krown, personal communication).
IFN-alpha is currently being studied in combination with ddI. ACTG 206 is testing ddI with low dose (1 MU) or moderate dose (10 MU) IFN-alpha-2b. The principal Investigator noted that responses are being seen when 1 MU daily of IFN-alpha-2b is used in combination with 200 mg of ddI twice a day (Susan E. Krown, personal communication).
More recently, studies have concluded that interferon may in fact alter certain cytokines and growth factors involved in the pathogenesis of KS. Fidler and colleagues (Fidler 1994) found the IFN-alpha and IFN-beta decrease bFGF in cultured tumor cells at the protein and message level. Tilg and colleagues (Tilg 1993) have demonstrated that IFN-alpha increases the circulating levels of the IL-1 receptor without an increase in IL-1 in the plasma of healthy patients and those with chronic hepatitis C. IFN-alpha can also reduce the accumulation of mRNA of TNF-alpha and IL-6 and thus interrupt their autocrine production in hairy cell leukemia and B-chronic lymphocytic cells (Heslop 1990).
There are many myths and truths about systemic chemotherapy for the treatment of KS. Many of the preconceptions people have about chemotherapy are based on the experiences in generally older, sicker people who received high dose chemotherapy (often different agents) without the benefit of colony stimulating factors. It should be noted that the various cytotoxic (cell killing) chemotherapeutic agents used -- as single agents or in combination -- to treat KS have been around longer than AIDS and have been used for many years to treat the myriad of cancers.
It took oncologists a few years to understand what the safe and efficacious doses were for patients with HIV. Hence, patients with KS early on in the epidemic were receiving doses of chemotherapy that were highly toxic (Alvin Friedman-Kien, personal communication). It is now 1994. Oncologists who treat the various AIDS-related malignancies now know -- and basically tend to agree on -- the doses that are safe and efficacious for each cytotoxic agent used to treat KS.
For patients with widespread cutaneous or visceral KS, systemic (whole body) treatment with cytotoxic agents is the treatment of choice (Miles: in press). Excellent response rates with manageable side effects have been seen with various single agent regimens or combinations of: vincristine, vinblastine, doxorubicin (adriamycin), bleomycin and etoposide (VP 16). Having these agents to choose from has led oncologists, such as Steve Miles, to proclaim, "the [mere] observation of patients with progressive KS in the setting of HIV infection isno longer acceptable." (Miles: in press)
As with most other treatment options for KS, it is difficult to compare the advantages of one cytotoxic agent over another because randomized, controlled trials have not been conducted and most did not use a uniform response or staging criteria (Krown 1992 a; Miles: in press). There are two reasons for this: 1) researchers fully understand that it would be unethical to attempt -- and impossible to recruit patients for -- a placebo-controlled trial for patients with symptomatic, visceral KS; 2) most of these single agent and combination cytotoxic agent KS trials were completed and published before it was standard practice to use the 1989 ACTG staging and response criteria.
In a 38 patient non-randomized, non-controlled, dose-escalating, intravenous vinblastine study, Volberding and colleagues (Volberding 1985) noted a 30% response rate; one complete response, nine partial responses, 19 with stable disease, and 9 who progressed. Administered once weekly at a starting dose of 4 mg. (median dose of 6 mg), the median time to response was 5 weeks with a median response duration of 13 weeks. The hematological toxicities were minor and some nausea was noted. 30% of the patients developed OIs during the trial.
Single agent vincristine is also sometimes used to treat symptomatic cutaneous and visceral KS, particularly in patients with low platelet counts. Intravenous vincristine was studied in 23 patients in an open, non-controlled trial conducted by Mintzer and colleagues (Mintzer 1985). Of 18 evaluable patients, 11 (61%) obtained a partial response and 7 had a "minor" response to therapy. The median duration of the partial responders was over four months. The major toxicity reported was peripheral neuropathy, which required dose reduction in 6 patients and a discontinuation of treatment in 2 patients. 5 patients developed OIs during the trial.
Studies of vinblastine (2-4 mg) alternating weekly with vincristine (2 mg) have been conducted to attenuate the toxicity of each drug. Kaplan and colleagues (Kaplan 1986) developed an alternating weekly vinblastine (0.1 mg/kg) and vincristine (2 mg) regimen for their study that modified doses for myelosuppression or peripheral neuropathy. Of 21 evaluable patients, they noted: one complete response, eight partial responses, seven patients with stable disease, and five with progressive disease. The median time to response was 13 weeks with a median response duration of over nine months. Six patients required dose modification due to peripheral neuropathy. Hematologic toxicity was minimal.
Lassoued and colleagues (Lassoued 1990) used single agent bleomycin either intramuscularly or intravenously to treat 60 patients in an uncontrolled trial. 30 patients received 5 mg/day intramuscularly for 3 days every 2 or 3 weeks, and 30 patients received bleomycin by slow continuous infusion (6 mg/m2/day for 4 days every four weeks). 40 of the patients also received AZT (200-1200 mg/day) in combination with bleomycin. 29 (48%) achieved a partial response and 18 (30%) had stable disease. No significant differences were seen in response rates when comparing the intramuscular and intravenous routes of administration. 23 (38.3%) of the patients died while on the study. Toxicities consisted of fevers, mild myelosuppression and some cutaneous toxicity.
A 1992 open, non-controlled, French study (Caumes 1992) evaluating the efficacy of intramuscular bleomycin on mucocutaneous KS in 70 patients noted a 74% overall response rate; two complete responders and 50 patients achieving a partial response. Using 5 mg/day for three consecutive days every two weeks, the median time to treatment response was 4 weeks, but the median time to relapse was 10 weeks.
Adriamycin (doxorubicin) has shown some efficacy used as a single agent. ACTG 006 was a phase II trial (Fischl 1993) of Adriamycin given intravenously to 53 chemotherapy naive patients at weekly doses of 15 mg/m2. Of 50 patients evaluable for tumor response, there were 5 (10%) partial response, 32 patients (64%) had a "minor response,"4 12 patients (24%) had no change, and one patient (2%) progressed. Response duration was short, ranging from 4 to 14 weeks. The major toxicity, seen in 71% of the patients, was moderate to severe neutropenia. 18% of the patients had to discontinue therapy because of severe neutropenia.
Etoposide (VP-16) has also shown some efficacy and is often used because it can be taken orally. Laubenstein and colleagues (Laubenstein 1984) studied intravenous etoposide (150 mg/m2 for three days every four weeks) on 41 chemotherapy naive patients with limited KS and no prior OIs. Response rates were high (12 CR and 19 PR), however, neutropenia and gastrointestinal toxicities were common and alopecia (baldness) was observed in all patients.
A 29 patient, phase I, dose-escalating, oral VP-16 trial (ACTG 110) was recently completed. Of 25 patients evaluable for response, 9 patients (36%) achieved a partial response, 14 (56%) had stable disease, and 2 patients (8%) progressed (Josephine Paredes, in press). The major toxicity was mild to moderate neutropenia. Peripheral neuropathy, nausea, and alopecia were also documented.
Very little efficacy data exists on the use of 4'-epirubicin for the treatment of KS. One phase II study of 4'-epirubicin (Shepherd 1991) in 26 patients with poor-risk noted an overall 42.3% response rate (1 CR and 10 PR) with a median time to relapse of 22 weeks. The dose-limiting toxicity was neutropenia.
Combination chemotherapy regimens can be tailored for each patient individually to enhance efficacy and reduce toxicity. In an attempt to test the efficacy and toxicity of single agent versus combination therapy, Gill and colleagues (Gill 1991 c) conducted a randomized study of 61 patients comparing low-dose adriamycin alone (31 patients) versus adriamycin, bleomycin and vincristine (ABV) (31 patients). 29 patients were evaluable for response in the Adriamycin arm as were 24 patients from the ABV arm. An overall response rate of 88% was obtained in the ABV arm; 9 (38%) complete responses and 12 (50%) partial responses. The Adriamycin arm yielded a lower overall response rate of 48%; one (3%) complete response and 13 (45%) partial responses. The median survival was 9 months for both arms. Neutropenia occurred in 34% of the patients receiving Adriamycin and in 52% of the patients in the ABV arm.
19 of the 31 patients who failed or did not respond to Adriamycin alone were able to roll-over into the ABV arm or receive BV. Four of the eight patients receiving BV responded, as did nine of the 11 patients receiving ABV.
To avoid the myelotoxicity associated with Adriamycin, G-CSF is now frequently used to bolster a patient's ANC. The use of G-CSF has led some oncologists to believe that "neutropenia is a thing of the past." (Donald Northfelt, personal communication)
The use of a BV combination regimen is also considered a safe and effective alternative for those patients who can not tolerate Adriamycin. Gompels and colleagues (Gompels 1992) completed a retrospective case analysis studying bleomycin (30 mg over an 18 hour infusion) and vincristine (2 mg) (or vinblastine if peripheral neuropathy developed) in 46 chemotherapy naive patients with poor risk factors (72% had previous OIs and all had CD4 counts under 200). 26 patients (57%) achieved a partial response, 16 (35%) had stable disease and 4 (9%) progressed. The mean duration of response was only 2 months and survival was 8 months from start of therapy. Peripheral neuropathy was observed in 13 (28%) of the patients.
A non-randomized study conducted by Ireland-Gill and colleagues (Ireland-Gill 1992) compared ABV versus BV in 99 patients. 30 patients received BV and 69 received ABV. The BV arm had an overall response rate of 76% (7 CRs, 16 PRs) compared with a response rate of 81% (20 CRs, 36 PRs) for patients who received ABV. While high response rates were seen, most patients relapsed 2 months after treatment was stopped.5 The median survival time for these patients was just under one year.
28 out of the 99 patients underwent pulmonary function testing while on study to ascertain if cumulative doses of bleomycin induced pulmonary dysfunction. Investigators noted that there was significantly greater decline in the carbon monoxide diffusion capacity (DLCO) in patients who received more than 100 cumulative units of bleomycin. While none of these patients (who had received cumulative doses between 10 and 313 units of bleomycin) developed clinically significant pulmonary toxicity, investigators still concluded than any patient receiving over 100 cumulative units of bleomycin should have their pulmonary functions closely monitored.
Combination cytotoxic agents yield the most promising results for patients with pulmonary KS. Gill and colleagues (Gill 1989) evaluated several drug regimens (A versus BV versus ABV) in 20 patients with pulmonary KS. The 10 patients receiving ABV had an overall 80% response rate (4 CRs, 4 PRs); the 3 patients receiving BV had a 100% response rate (3 CRs); and the 7 patients receiving single agent Adriamycin only had a 14% response rate (1 PR). Out of the 20 patients, the median survival for the 12 responders was 10 months (range three to 31+ months) versus 6 months for the 8 non-responders. (It is important to note that 6 of the 8 non-responders were receiving only single agent Adriamycin.)
Gill and colleagues noted that pleural effusion (fluid surrounding the lungs, in the pleural space) and a CD4 count below 100 were the most significant predictors of shorter survival. It is also interesting that the DLCO level in patients increased while on study which was not the case in the later trial conducted by Ireland-Gill.
The median survival of 10 months (for responders) for patients in this trial is significantly higher than previously seen in other trials. In an older systemic chemotherapy trial of patients with pulmonary KS, Garay and colleagues (Garay 1987) noted a median survival of 6 months. And, in a trial that administered either radiation, IFN-alpha, or chemotherapy to 11 patients with pulmonary KS, Meduri and colleagues (Meduri 1986) only noted a median survival of 3.8 months.
Many researchers have hoped that IFN-alpha would be suitable for maintenance therapy after cytotoxic chemotherapy. At least two studies (Gill 1990, 1991 a) have shown that IFN-alpha did not provide additional antitumor effect or prolong response duration. IFN-alpha's lack of activity here should not be surprising considering these patients' studied mostly had a CD4 count of less than 200 and prior OIs.
GM-CSF and G-CSF have become an adjunctive part of various cytotoxic regimens to ameliorate neutropenia. Trials have been conducted -- primarily with GM-CSF -- to determine whether such agents can limit neutropenia and allow patients to receive full doses of cytotoxic agents on schedule. In a phase I trial conducted by Gill and colleagues (Gill 1992), GM-CSF at a dose of 250 mcg/m2 produced a cyclic increase in granulocyte counts in patients receiving ABV. At doses of 500 mcg/m2, however, GM-CSF induced fever, fatigue and diarrhea.
Such toxicities have not been found with G-CSF. It may be possible to increase the maximum tolerated doses of ABV if G-CSF can prevent neutropenia without inducing additional non-hematologic toxicities (Miles, in press). In a recent trial of DOX-SL (liposomal encapsulated doxorubicin) (Sandor 1993), G-CSF was administered to 24 of the 33 patients whose ANC dropped below 1,000. G-CSF boosted the ANC back up, and the rate of bacterial, viral and fungal infections was also significantly reduced: 18 infections in 106 cycles with G-CSF (17%) vs 35 infections in 120 cycles without G-CSF (29%) (p< 0.05).
Most trials evaluating the combination of single agent cytotoxic therapy and AZT have noted adequate responses but high hematologic toxicity. There are studies evaluating the efficacy and toxicity of AZT with combination cytotoxic agents, however, few have been published. Rarick and colleagues (Rarick 1990) in a 12 person trial of BV with varying doses of AZT (400 mg a day or 800 mg a day) noted an 83% response rate (4 CRs, 6 PRs). The investigators noted that the 400 mg/day dose of AZT (compared with 800 mg/day) was as active and less toxic.
Preliminary results from a phase I/II trial of ABV with ddI or ddC (ACTG 163) (Mitsuyasu 1993) revealed that toxicities are minor with the nucleoside analogues (and probably due to ABV therapy). From the first phase of the trial, which accrued 29 patients (15 ddC and 14 ddI), two complete responses and 13 partial responses have been achieved so far. Complete results will be available in late 1994 or early 1995.