phase iii vehicle-controlled, multi-centered study of topical alitretinoin gel 0.1% in cutaneous...

Phase III Vehicle-Controlled, Multi-CenteredStudy of Topical Alitretinoin Gel 0.1% inCutaneous AIDS-Related Kaposi’s SarcomaNeil J. Bodsworth,1 Mark Bloch,2 Mark Bower,3 David Donnell,4 Richard Yocum5 and TheInternational Panretin® Gel KS Study Group1 Taylor Square Private Clinic, Darlinghurst, New South Wales, Australia2 Holdsworth House General Practice, Darlinghurst, New South Wales, Australia3 Chelsea & Westminster Hospital, London, England4 Nelson-Tebedo Clinic, Dallas, Texas, USA5 Ligand Pharmaceuticals Inc., San Diego, California, USA

Abstract Objective: This randomized, double-blind and vehicle-controlled phase III study was conducted to evaluatethe efficacy and safety of alitretinoin gel 0.1% for the topical treatment of the cutaneous lesions of AIDS-relatedKaposi’s sarcoma (KS).Methods: Patients received treatment with alitretinoin gel (n = 62) or vehicle gel (n = 72) twice daily for 12weeks. The primary efficacy endpoint was the cutaneous KS tumor response rate according to AIDS ClinicalTrials Group (ACTG) objective criteria applied to topical therapy, with the patient as the unit of analysis.Results: Treatment of patients with alitretinoin gel resulted in a significant antitumor effect. The overall patientresponse rate (complete plus partial response) was 37% (23 of 62) for the alitretinoin-treated patients and 7%(5 of 72) for the vehicle-treated patients (p = 0.00003). The difference in response rates for the 2 treatmentgroups remained significant even after taking into consideration numerous variables, including age (p =0.00001), Eastern Cooperative Oncology Group (ECOG) status (p = 0.00002), CD4+ cell count (p = 0.00002),history of opportunistic infection (p = 0.00002), aggregate area of indicator lesions (p = 0.00005), number ofraised indicator lesions (p = 0.00002), prior therapy for KS (p = 0.00003), and number of drugs (p = 0.00002)used in concomitant antiretroviral therapy. Generally, treatment with alitretinoin gel was well tolerated. Theoverall incidence of adverse events was similar for the 2 treatment groups. Adverse events related to treatmentwith alitretinoin gel tended to be mild to moderate in severity and limited to the site of application. The mostfrequent adverse event occurring at the application site following alitretinoin gel treatment was irritation codedas rash (32%).Conclusions: The results of this study provide convincing evidence of the superiority of alitretinoin gel overvehicle gel for the treatment of the cutaneous lesions of AIDS-related KS.

ORIGINAL RESEARCH ARTICLE Am J Clin Dermatol 2001; 2 (2): 77-871175-0561/01/0003-0077/$22.00/0

© Adis International Limited. All rights reserved.

Kaposi’s sarcoma (KS) is the most common neoplasm asso-ciated with AIDS.[1,2] It is characterized by skin lesions that areoften disfiguring and can also involve gut mucosa, lung epithelia,lymph nodes, and visceral organs. A putative etiologic agent, hu-man herpesvirus type 8 (HHV8), has been described[3] and ap-pears to be sexually transmitted among homosexual men.[4] Safersexual practices among homosexual men and the introduction ofhighly active antiretroviral therapy (HAART), which preservesimmune competence, are both thought to contribute to a decliningincidence of AIDS-related KS.[5,6] However, the overall preva-

lence of AIDS-related KS remains high and the psychologic andphysiologic toll exacted by this disease presents a continuingchallenge for the clinician.

No known cure currently exists for KS. Present treatmentoptions include local therapy with radiation, cryotherapy, or in-tralesional therapy, and systemic therapy with liposomal anthra-cyclines (doxorubicin, daunorubicin), paclitaxel, other cytotoxicagents (vinblastine, vincristine, bleomycin), or interferon-α. Ra-diotherapy, cryotherapy and intralesional therapy, although con-sidered useful for patients with localized, cosmetically disturbing

KS lesions, often are limited by the number of office/clinic visitsrequired to achieve results. Moreover, local therapy often is as-sociated with significant mucositis (caused by radiation), painand blistering, and hypo/hyperpigmentation (all caused by cryo-therapy). The utility of systemic cytotoxic therapy is limited bydose-limiting toxicities such as neutropenia, thrombocytopenia,and anemia, together with the high cost of administration.

Several clinical trials suggest that retinoids may be useful inhealing KS.[7-9] Retinoids are retinol (vitamin A) derivatives thatexert numerous biologic effects, including modulation of cellproliferation, differentiation, and apoptosis, by binding to spe-cific intracellular retinoic acid receptors (RAR-α, RAR-β, RAR-γ) and retinoid X receptors (RXR-α, RXR-β, RXR-γ).[10] Theactivation of RAR receptors is associated with the regulation ofcellular proliferation and differentiation, whereas the activationof RXR receptors is associated with apoptosis.[11,12] Two endog-enous retinoids, tretinoin (all-trans-retinoic acid) and alitretinoin(9-cis-retinoic acid), are characterized by distinct receptor acti-vation profiles. Tretinoin interacts with the RAR subgroup ofretinoid receptors, whereas alitretinoin, a retinoid receptorpanagonist, interacts with both the RAR and RXR receptor sub-groups.[13] The ability of alitretinoin to also activate the RXRreceptor subgroup suggests that it may confer therapeutic benefitsdifferent from those observed in vitro with other retinoid prepa-rations.[14-17] Retinoids have been shown to inhibit KS cellgrowth in vitro.[10,18,19]

In an earlier, phase I-II clinical trial, alitretinoin gel 0.1%demonstrated significant antitumor activity when used to treat thecutaneous lesions of patients with AIDS-related KS.[20] A 27%response rate was observed for treated indicator KS lesions versusan 11% response rate for untreated control lesions. Moreover,treatment with the alitretinoin gel was well tolerated. In order toconfirm these promising phase I-II data, a larger, multi-centered,double-blind and vehicle-controlled phase III clinical trial wasconducted to evaluate the safety and efficacy of alitretinoin gel0.1% in the palliative topical treatment of AIDS-related cutane-ous KS.

Methods

Study Description

This study included 22 centers under 2 protocols, L1057-503and L1057-504. The L1057-503 protocol provided a 12-week,randomized, double-blind, vehicle-controlled study. Following12 weeks of double-blind treatment or at time of progressive dis-ease (whichever occurred sooner), patients could elect to partic-ipate in protocol L1057-504, a follow-on, open-label treatment

for 36 weeks that was designed to assess the same indicator le-sions evaluated during the first double-blind protocol. The open-label protocol L1057-504 ensured patients active study drug treat-ment and continued indicator lesion assessment for evaluability.

Study Population

Adult patients (≥18 years) with documented HIV infectionand biopsy proven cutaneous KS were eligible for enrolment pro-vided that they had: at least 3 localized lesions amenable to self-treatment, each with minimum diameters of 10mm × 2mm; East-ern Cooperative Oncology Group (ECOG) status performancestatus of 0-2 (capable self care); a life expectancy greater than 4months; and no hematologic, renal, or hepatic laboratory meas-ures indicating severe organ dysfunction. Nonpregnant women ofchildbearing potential and men with partners of childbearing po-tential were required to use adequate barrier contraception duringthe trial and for at least 3 months thereafter.

Patients were to be excluded from enrolment if they had re-ceived systemic treatment for KS within the previous 4 weeks;had been administered systemic treatment with retinol (>15 000IU or 5 mg/day) for KS at any time; had received prior localtherapy for any indicator lesion; had a known allergy or sensitiv-ity to retinoids; were women who were nursing; were concur-rently enrolled in any investigational drug trial for KS or hadparticipated in one within the previous 30 days; had suspected orknown poor compliance; had a concurrent uncontrolled infectionrequiring intravenous medication; or had a concomitant malig-nancy (including visceral KS) likely to require systemic treatmentduring the course of the trial.

All patients were to complete an approved written informedconsent process prior to study participation.

Treatments

For the initial, double-blind protocol, patients were random-ized in a one-to-one ratio to receive either alitretinoin gel 0.1%or matching vehicle gel, administered twice daily for 12 weeksor until confirmation of disease progression, whichever wassooner. Randomization in blocks of four was stratified by studysite only. At baseline, 3 to 8 indicator lesions were selected formeasurement and treatment. The investigator, for reasons of ac-cessibility as well as assessability, selected the indicator lesions.Instructions were given for patients to apply an even layer of gelover the entire surface of indicator lesions. Patients were free totreat any other cutaneous KS lesions present at enrolment orwhich developed during the study. For the open-label protocol,all patients received treatment with alitretinoin gel 0.1%, admin-istered twice daily for up to 36 weeks. Application frequency was

78 Bodsworth et al.

© Adis International Limited. All rights reserved. Am J Clin Dermatol 2001; 2 (2)

reduced to once daily in the event of grade 3 (‘severe’) localdermal irritation and discontinued if this dose reduction was nottolerated by the patient.

Efficacy Assessments

The primary efficacy endpoint was based on an objectiveassessment of cutaneous KS tumor response rate after 12 weeksof double-blind treatment, with the patient as the unit of analysis.At each visit, designated indicator lesions (minimum 3, maxi-mum 8), were measured in area and height. Response was asses-sed using AIDS Clinical Trials Group (ACTG) response criteriaapplied to topical therapy,[21] and designated as clinical completeresponse (CCR), partial response (PR), stable disease (SD), orprogressive disease (PD) (table I). Secondary analyses also wereperformed for individual indicator lesion response rates (table I).All responses required confirmation by a second objective eval-uation at least 4 study weeks after the initial observation of re-sponse.

Response rate was calculated based on best response, whichwas defined as the highest confirmed response classification (seetable I) for the patient during the study. However, if a confirmedPD classification occurred before any other response criteria, thenthe patient’s best response was defined as PD.

For responding patients (CCR or PR), the durability of re-sponse was calculated from the first day of treatment to the onsetof relapse.

The physician’s and patient’s subjective assessments oftreatment were prospectively defined secondary efficacy end-points. At each study visit, both patients and physicians wereasked to assess the extent of improvement or worsening of treatedlesions relative to baseline according to a fixed 7-point scale (ta-ble II). All treated lesions, both indicator and non-indicator, wereevaluated to provide a global, subjective assessment of response.Patients and physicians were independent with respect to the oth-ers assessment. These assessments required ‘100% improvement’for characterization of CCR and ‘50% or more improvement’ forPR.

Safety Assessment

Treatment-emergent adverse events were coded according tothe Coding System Thesaurus for Adverse Reaction Terms 5(COSTART 5) dictionary and classified as treatment-related(possibly, probably, definite), treatment-unrelated (none, un-likely), or unknown. Adverse event severity was categorized ac-cording to the National Cancer Institute (NCI) grades 1 through4. A serious adverse event was defined conventionally as anyuntoward medical occurrence that was life-threatening, requiredor prolonged hospitalization, or resulted in persistent or signifi-cant disability, cancer, congenital anomaly, or death. Hematol-ogy, biochemistry and urinalysis were performed at baseline andweeks 4 and 12. CD4+ lymphocyte count was measured at base-

Table I. Response definitions according to AIDS Clinical Trials Group objective criteria applied to topical therapy[21]

Criterion Individual lesion response Patient responseClinical completeresponse (CCR)

Disappearance of all clinical signs and symptoms of theindicator lesion, determined by 2 evaluations not less than4 weeks apart

Disappearance of all clinical signs and symptoms of all indicatorlesions, determined by 2 evaluations not less than 4 weeksapart

Partial response (PR) A 50% or greater decrease in the product of the longestperpendicular diameters of the indicator lesion,determined by 2 evaluations not less than 4 weeks apart,and no concurrent increase in the height of the lesion if flatat baseline ORcomplete flattening of the lesion raised at baseline withoutconcurrent increase in area by 25% or more from baseline ORcomplete flattening of the lesion raised at baseline withoutconcurrent increase in area by 25% or more from baseline

A 50% or greater decrease in the sum of the products of thelongest perpendicular diameters of all indicator lesions,determined by 2 evaluations not less than 4 weeks apart. Lessthan 25% increase from baseline in number of raised indicatorlesions ORat least 50% decrease in number of raised indicator lesionscompared with baseline, without increase in sum of the areas ofthe indicator lesions by 25% or more from baseline

Stable disease (SD) Any classification not meeting the criteria for CCR, PR, orPD

Any classification not meeting the criteria for CCR, PR, or PD.

Progressive disease (PD) A 25% or more increase from baseline in the product ofthe longest perpendicular diameters of the indicator lesion ORa change in the character of the lesion from flat to raised ORa change in the character of the lesion from ‘flat’ to ‘raised’

A 25% or more increase from baseline in the sum of theproducts of the longest perpendicular diameters of all indicatorlesions ORAt least 25% increase from baseline in the number of raisedindicator lesions.

Topical Alitretinoin Gel in AIDS-Related KS 79


line and at week 12, and patients were assessed at each visit forany new opportunistic infections or other AIDS-related events.

Statistical Analyses

All statistical analyses were carried out using theSAS/STAT® statistical package (SAS Institute, Inc., Cary, NC).All analyses reported were performed on the intention-to-treatpopulation, defined as all randomized patients who were dis-pensed at least one dose of trial medication (n = 134). A 2-sidedtype I error ≤5% was considered to be statistically significant.

The primary efficacy endpoint of objective tumor responseaccording to ACTG criteria was compared between treatmentgroups using Fisher’s exact test for unadjusted analyses and theMantel-Haenszel test for adjusted analyses. Objective tumor re-sponse was analyzed with respect to baseline demographic andpotential prognostic factors, including ECOG performance score,CD4+ lymphocyte count at baseline, history of opportunistic in-fections, aggregate area of indicator lesions, number of raisedindicator lesions, previous therapies for KS, and concurrent anti-viral therapy.

Endpoints defined by time to event (time to first response anddurability of response) were analyzed using time-to-failure meth-ods for censored observations. The Kaplan-Meier product limitmethod provided the mean and median time to event and standarderror for each treatment arm.

Categorical secondary endpoints (physician’s subjective as-sessment, patient’s subjective assessment, and individual indica-

tor lesion response) were compared between treatment groupsusing Fisher’s exact test. The agreement between the objectiveACTG criteria and both physician’s and patient’s subjective as-sessments was examined using Kappa statistics and Spearman’sRho test. Patient demographic and baseline characteristics werecompared between the 2 treatment groups using the Chi-squaretest.

One interim analysis was planned, which was based on themethod of O’Brien and Fleming.[22] Based on the first 78 patients(39 in each treatment arm) completing the 12 weeks of double-blind treatment and an 80% power, a target alpha level of 0.00025(2-sided) was used.

Interim Analysis

At the time the interim analysis was performed, 82 patientshad been enrolled and dispensed at least 1 dose of study drug. Aresponse rate comparison revealed a calculated p-value of0.00027 using Fisher’s exact test and a p-value of 0.0001 usingthe Chi-square test. Based on the target significance level of0.00025 for the interim analysis (in accordance with O’Brien andFleming[22]), the results met the conditions of the protocol-spec-ified Chi-square test and indicated a statistically significant dif-ference between the 2 treatment groups. Consequently, in accord-ance with the early stopping rule, the study was terminated withfewer patients enrolled than the number originally proposed (n =270) in the protocol. During the conduction and verification ofthe interim analysis of 82 patients, an additional 52 patients wereenrolled, yielding a final sample size of 134 patients. The finalanalysis of the full data set from 134 patients confirmed and re-inforced the findings of the interim analysis.

Results

Intention-to-Treat Population

Of the 134 patients enrolled in the double-blind study anddispensed trial drug, 62 patients were randomized to receive al-itretinoin gel and 72 patients were randomized to receive vehiclegel. The apparent imbalance in treatment assignment was not sys-tematic and was caused by the randomization block of four, with11 of 22 study centers enrolling fewer than 4 patients. Patientwithdrawals during the double-blind protocol are described intable III.

Patient Characteristics at Baseline

The demographics of this study population were consistentwith patients with AIDS-related KS in the geographic locations

Table II. Physician’s and patient’s subjective assessment scale[21]

Score Definition Responsea

0 Completely cleared except for pigmentationfrom residual hemosiderin

CCR

1 Almost cleared; very significant clearance indisease, with only traces of disease remaining(approximately 90% improvement)

PR

2 Marked response; significant improvementwith some disease remaining (approximately75% improvement)

PR

3 Moderate response; intermediateimprovement, between slight and markedimprovement (approximately 50%improvement)

PR

4 Slight response; some improvement butsignificant disease remaining (approximately25% improvement)

SD

5 Condition unchanged SD6 Condition worsened PDa Response required confirmation over at least 4 study weeks. CCR = clinical complete response; PD = progressive disease; PR = partialresponse; SD = stable disease.

80 Bodsworth et al.


studied (table IV). The majority of patients enrolled were Whitemales aged 30 to 49 years; no women were enrolled. The activeand vehicle treatment groups were comparable with regard to allfactors assessed, including age, race, ECOG performance status,baseline CD4+ lymphocyte counts, and characteristics of KS dis-ease (table IV). Moreover, the 2 treatment groups were similarregarding history of prior anti-KS therapy and concurrent anti-retroviral (ARV) therapies (table IV). 35% of patients in bothtreatment groups had received some prior therapy for KS.

Objective Responses According to AIDS Clinical TrialGroup Criteria

Figure 1 illustrates the best response rates for patients duringthe initial double-blind protocol. The overall response rate (CCRplus PR) was 37% (23 of 62) for the alitretinoin gel group versus7% (5 of 72) for the vehicle gel group (p = 0.00003, Fisher’s exacttest). The only patient to achieve a CCR was in the alitretinoin-treated group. This difference in response rates for the 2 treatmentgroups remained highly significant after consideration of numer-ous variables using the Mantel-Haenszel test, including age (p =0.00001), ECOG status (p = 0.00002), CD4+ lymphocyte count(p = 0.00002), history of opportunistic infection (p = 0.00002),aggregate area of indicator lesions (p = 0.00005), number ofraised indicator lesions (p = 0.00002), and prior therapy for KS(p = 0.00003). Incidence of progressive disease was considerablylower for the alitretinoin gel group than for the vehicle gel group(19 vs 35%, respectively; p = 0.047, Chi-square test for 2 x 2comparison).

Without exception, the response rate was higher for the al-itretinoin gel group than for the vehicle gel group at each timepoint assessed during the 12-week double-blind protocol. Thesuperiority of alitretinoin gel treatment was apparent at week 2,

the first scheduled on-study lesion assessment (p = 0.0123, Fish-er’s exact test), and was maintained at week 4 (p = 0.0001, Fish-er’s exact test), and for all weeks combined (p = 0.0003, Fisher’sexact test). Ten alitretinoin gel-treated patients met objective re-sponse criteria by week 2, and responses to alitretinoin gel treat-ment continued to accrue with each successive time point duringthe double-blind protocol: 21 by week 4, and 23 by week 8. Re-

Table III. Patient withdrawals from treatment during double-blind protocolphase

Reason for withdrawal Number of patients withdrawn bytreatment groupalitretinoin gel 0.1%(n = 62)

vehicle gel (n = 72)

Protocol violation 0 1Grade 3 dermal irritationa 2 0Disease progression 5 15Lost to follow-up 7 3Patient’s interest or request 4 4Total (%) 18 (29b) 23 (32b)a Did not resolve following dose reduction to once daily.b Percentage of patients enrolled in that treatment group.

Table IV. Patient characteristics at baselinea

Variable Alitretinoin gel 0.1%recipients (n = 62)

Vehicle gelrecipients (n = 72)

Demographic characteristicsMedian age (y) 38 38Gender (% men) 100 100Race (% White) 90 88

HIV baseline characteristics Median CD4+ lymphocyte count(cells/mm3)

160 132

≤50 cells (%) 31 28History of opportunistic infections/AIDS-related illnesses (%) 1 11 10 2 21 17

≥3 32 35Concomitant ARV therapy (%) any ARV therapy 92 96

≥3 agents 73 82 any protease inhibitor 73 81

KS baseline characteristics ECOG performance status (%) 0 66 68 1 23 22 2 11 10Coexisting visceral KS (%) 16 13Median number of indicatorlesionsb

4 4

Proportion with raised indicatorlesions (%)

61 68

Aggregate indicator lesion area(mm2) median 490 658 range 95-5802 70-9116Prior KS therapy (%) none 63 67 systemic only 31 25 local only 2 4 both systemic and local 5 4a All differences between treatment groups were not statistically signifi-

cant.b At start of double-blind treatment.ARV = antiretroviral; ECOG = Eastern Cooperative Oncology Group; KS =Kaposi’s sarcoma.



sponses not yet confirmed by a second assessment by week 12,the final assessment in the 12-week double-blind protocol, werenot classified as responses.

When data from the follow-on open-label treatment protocolwere added to the double-blind data, the overall response rate forpatients receiving alitretinoin gel was 44% (44 of 99), with 4patients achieving a CCR best response. 41 of the 62 patients(66%) randomized to receive alitretinoin gel therapy during thedouble-blind protocol continued to receive this drug during open-label therapy. Of these 41 patients, 23 (56%) were categorized asresponders during open-label therapy. 58 of the 72 patients (81%)randomized to receive vehicle treatment during the double-blindprotocol received treatment with alitretinoin gel for the first timeduring open-label therapy. 21 (36%) responded to alitretinointreatment; and one of these patients achieved a CCR. The overallresponse rate for these same 58 patients during double-blind ve-hicle treatment had been 7%, with no CCRs (p = 0.00000,McNemar test).

Time to Objective Response and Durability of Response

The projected time to onset of response was estimated usingthe Kaplan-Meier method (fig. 2). Among patients receiving al-itretinoin gel during both the double-blind and the open-labelprotocols, the median time to onset of response was 29 days(range 13 to 171 days).

Durability of response was analyzed by the Kaplan-Meiermethod (fig. 3). These data are based on all 44 patients who re-sponded, including those patients randomized to alitretinoin geland those patients initially randomized to vehicle gel but latertreated with open-label alitretinoin gel. Only 7% (3 of 44) of

patients relapsed over a median 190 days of monitoring (range 1to 345 days). A median time to relapse could not be estimatedbecause of the low rate of relapse.

Effect of Prior Anti-Kaposi’s Sarcoma Treatment on

Objective Response

Positive responses to alitretinoin gel were observed for pa-tients who were considered refractory to prior KS treatment. Atotal of 43 patients had received at least one prior systemic anti-KS therapy. Four and 9 of these patients in the alitretinoin andvehicle groups, respectively, had been considered unresponsive(i.e., failed to achieve ≥50% improvement) to at least one suchtreatment. Response rates for these patients during the double-blind protocol were 2 of 4 (50%) for the alitretinoin group and 0of 9 (0%) for the vehicle group. In 4 alitretinoin gel recipients and7 vehicle gel recipients their condition was considered whollyrefractory (failed to achieve at least a partial response to all priorsystemic therapies); the response rates during the double-blindprotocol for these patients were 2 of 4 (50%) and 0 of 7 (0%),respectively.

During the open-label protocol, positive responses to al-itretinoin gel were observed in 4 of 13 (31%) patients whosecondition was considered unresponsive or wholly refractory toprior systemic chemotherapy. This included new responses in 2of 9 (22%) patients who had received vehicle in the double-blindprotocol but subsequently were treated with alitretinoin gel forthe first time on the open-label protocol.

23

51

22

5

27

42

12

25

00

10

20

30

40

50

60

Patie

nts

(%)

Overallresponse

CCR PR SD PD

Best response

Alitretinoin gel 0.1%Vehicle gel

Fig. 1. By-patient best response (according to AIDS Clinical Trials Group objectivecriteria applied to topical therapy)[21] achieved during double-blind treatment (12weeks) with alitretinoin gel 0.1% (n = 62) or vehicle gel (n = 72). p = 0.00003(Fisher’s exact test for a 2 × 2 comparison). CCR = complete clinical response;PD = progressive disease; PR = partial response; SD = stable disease.

00 50 100 150 200 250 300 350 400

10

20

30

40

50

60

70

Pro

ject

ed re

spon

se ra

te (%

)

Time to first response (d)

Alitretinoin gel 0.1% Vehicle gel

Fig. 2. Projected time to onset of response during double-blind and open-labeltreatment with alitretinoin gel 0.1% (n = 62) or vehicle gel (n = 72). p = 0.0001 (logrank test).

82 Bodsworth et al.


Effect of Concurrent Antiretroviral Treatment onObjective Response

Treatment with alitretinoin gel was consistently superior totreatment with vehicle gel, regardless of the number of concom-itant antiretroviral therapies (p = 0.00002, Mantel-Haenszel test)[fig. 4].

Individual Indicator Lesion Response

Response of individual lesions rather than patients as the unitof analysis was a secondary endpoint. The response rate wasagain strongly in favor of alitretinoin gel (p ≤0.00001, Fisher’sexact test, table V) and reinforced the positive primary efficacyendpoint finding. 34% of all indicator lesions responded in thealitretinoin treatment group during double-blind treatment com-pared with 11% of indicator lesions in the vehicle control group.

Physician and Patient Subjective Assessments

There was a very strong correlation between the physician’sand patient’s subjective assessments, which included an evalua-tion of treated non-indicator lesions as well as indicator lesions(table VI). Both physicians and patients reported overall responserates of 40% for the alitretinoin gel group and 10% for the vehiclegel group (p = 0.00004). These assessments were consistent withthe objective response rates based on indicator lesions assessedaccording to ACTG criteria (fig. 1).

Safety and Tolerability

Alitretinoin gel was well tolerated throughout this study. 120patients were exposed to alitretinoin gel during either the double-

blind or open-label protocols of the study (median duration 186days, range 1 to 506). Only 4 patients (3%) withdrew because ofan adverse event (AE) attributed to study medication (all werecases of grade 3 dermal irritation that did not resolve with dosereduction). The irritation resolved after discontinuation for 2 pa-tients; the other 2 patients were lost to follow-up.

The most commonly reported (occurring at ≥ 5% incidence)alitretinoin gel–related AEs, among the 120 patients exposed toalitretinoin gel during the protocols are listed in table VII.

During the double-blind protocol, 53 of 62 alitretinoin gelrecipients (86%) and 59 of 72 vehicle gel recipients (82%) expe-rienced at least one AE (p = 0.58). Generally, the AEs were mildto moderate in severity and associated with the application site(table VIII). Application site AEs were significantly more com-mon among alitretinoin gel recipients than among vehicle gelrecipients (47 vs 19%, respectively; p = 0.0007, Chi-square test).Rash was the most frequent application site AE occurring in thealitretinoin gel group (32% incidence vs 11% incidence in thevehicle gel group). As a consequence of local toxicity, 9 of 62alitretinoin gel recipients (15%) reduced application to once dailyat some time during the double-blind protocol, compared with 1of 72 vehicle gel recipients (1.4%).

Serious AEs (SAEs) that emerged during the double-blindprotocol were significantly more common among vehicle gel re-cipients than alitretinoin gel recipients (22 vs 7%, respectively; p= 0.01) and were consistent with the patient’s underlying disorder(AIDS), and its treatment and complications. All SAEs (cachexia,cryptococcosis, infection, diarrhea, nausea, and vomiting) re-ported for the alitretinoin gel group during the double-blind pro-

100

90

80

70

60

50

40

30

20

10

0

Patie

nts

who

rela

psed

(%)

0 25 50 75 100 125 150 175 200 225 250 275 300 325 350Time to relapse (d)

Fig. 3. Projected time to relapse for all responders during double-blind and open-label treatment with alitretinoin gel 0.1% using Kaplan-Meier statistics (n = 44).

1/6

0/4

1/9

4/11

18/45

4/59 4/58

5/17

18/45

1/14

0

5

10

15

20

25

30

35

40

45

Res

pond

ing

patie

nts

(%)

0-1 2 ≥3 Any PI No PINumber of antiretroviral drugs

Concurrent antiretroviral therapy

Alitretinoin gel 0.1% (n = 62)Vehicle gel (n = 72)

Fig. 4. Best response to alitretinoin gel 0.1% versus concurrent antiretroviral ther-apy during masked treatment (12 weeks). p = 0.00002 (Mantel-Haenszel test 3× 2 × 2 of response versus treatment, adjusting for number of concurrent anti-retroviral therapies). PI = therapy with any HIV-1 protease inhibitor.



tocol occurred in individual patients at non-application sites, andwere considered by the investigator to be unrelated to study med-ication.

Three SAEs occurring at non-application sites, including in-dividual cases of pain, peripheral edema, and pharyngitis, werereported during the study. These non-application site SAEs weremore consistent with the patient’s underlying disease and its treat-ment than to treatment with study drug.

Five deaths (1 suicide and 4 cases of advanced HIV disease)occurred during this study; none were considered by the investi-gator to be related to study medication. Only one death occurredduring the double-blind protocol in a patient who had receivedvehicle gel.

There were no trends in clinically significant, treatment-emergent laboratory abnormalities in this study. This was pro-jected based on pharmacokinetic data from an earlier, phase I-IIclinical program showing that the range and frequency of detec-tion of 9-cis retinoic acid in patients applying alitretinoin gel wasnot different from those of untreated individuals.

At least one opportunistic infection or AIDS-related eventwas reported for 8 of 62 alitretinoin gel recipients (13%) and 6 of72 vehicle gel recipients (8%). There were no treatment-emergenttrends in CD4+ lymphocyte counts.

Discussion

The results of this randomized study demonstrate the supe-rior efficacy of alitretinoin gel compared with vehicle gel for thetreatment of cutaneous AIDS-related KS. In this double-blindtrial of 134 patients, 37% of the patients in the alitretinoin gelgroup responded favorably to topical therapy within 12 weeks,compared with only 7% of the patients in the vehicle gel group.Including data from the follow-on open-label protocol, 44% ofpatients responded to the treatment with alitretinoin gel, includ-ing 36% of the 58 patients who initially were randomized to re-ceive vehicle treatment. Significant responses were seen as earlyas 2 weeks in 10 alitretinoin gel recipients and the response ratefor the alitretinoin treatment group was significantly higher thanthat of the vehicle group for all time points assessed.

Alitretinoin gel is an effective topical medication for thetreatment of cutaneous AIDS-related KS.

Responses were observed after 2 to 4 weeks of active treat-ment with a median time to onset of response of 29 days and amedian duration of response of 217 days (range 64 to 372). Thevariation in time to response may conceivably be influenced bythe age of lesions treated, extent of KS disease, and patient com-pliance with the treatment regimen. However, the response rateand durability among patients makes alitretinoin gel a potentiallyeffective, self-administered, home treatment. The response data

Table V. Best responsea by individual indicator lesion during double-blind protocol

Treatment group No. of lesions (%) No. of CCR + PR (%) No. of CCR (%) No. of PR (%) No. of SD (%) No. of PD (%)

Alitretinoin gel 271 (47) 91 (34)b 17 (6) 74 (27) 121 (45) 59 (22)

Vehicle gel 303 (53) 34 (11)b 3 (1) 31 (10) 1637 (55) 102 (34)

a Based on AIDS Clinical Trials Group objective criteria applied to topical therapy[21] as described in table I.

b p < 0.00001 for alitretinoin gel versus vehicle (Fisher’s exact test based on 2 × 2 comparison).

CCR = clinical complete response; PD = progressive disease; PR = partial response; SD = stable disease.

Table VI. Best response according to physician’s and patient’s subjective assessments during double-blind protocol

Assessment No. of patients No. of CCR + PR (%) No. of CCR (%) No. of PR (%) No. of SD (%) No. of PD (%)

Physician’s subjective assessmentAlitretinoin gel 62 25 (40) 1 (2) 24 (39) 34 (55) 3 (5)Vehicle gel 72 7 (10) 0 (0) 7 (10) 61 (85) 4 (6)p = 0.0004a

Patient’s subjective assessmentAlitretinoin gel 62 25 (40) 1 (2) 24 (39) 35 (57) 2 (3)Vehicle gel 72 7 (10) 0 (0) 7 (10) 63 (88) 2 (3)p = 0.0004a

a Fisher’s exact test for 2 × 2 comparison of CCR + PR.CCR = clinical complete response; PD = progressive disease; PR = partial response; SD = stable disease.

84 Bodsworth et al.


from this international study are supported by a second pivotalstudy of similar design conducted in the US and Canada,[23] aswell as the phase I-II study.

Patients who were refractory to prior systemic or local anti-KS therapies responded to the treatment with alitretinoin gel. Thecomparative response rates during double-blind treatment for pa-tient’s refractory to prior systemic treatment were 2 of 4 (50%)for alitretinoin gel and 0 of 9 (0%) for vehicle gel. During theopen-label study, 4 of 13 (31%) patients considered refractory toprior systemic chemotherapy responded to treatment with al-itretinoin gel. The prior systemic treatments for the patients whoresponded to alitretinoin gel were systemic vinblastine and tha-lidomide therapies.

There were also a substantial number of patients (77%) whowere taking antiretroviral therapies concomitantly with the studydrug. The response rate for these patients treated with alitretinoingel during the double-blind protocol was significantly higher thanfor the group of these patients treated with the vehicle gel. Thesuperior response rate occurred with alitretinoin gel despite atrend of more frequent and intensive concomitant antiretroviraltherapy in the vehicle-treated patients. Protease inhibitors havebeen suspected to diminish KS lesions in some patients withAIDS.[24] However, it does not appear that the use of concomitantantiretroviral therapy significantly influenced the response rateobserved with alitretinoin gel treatment.

Both patients and their treating physicians were asked tomake a global assessment of all treated lesions. These subjectiveassessments demonstrated the superiority of alitretinoin gel overvehicle and satisfaction with the treatment. There was a remark-able correspondence between the evaluations made by the pa-tients and their physicians (table VI), suggesting that the re-sponses were not biased by either group of evaluators.

Treatment with alitretinoin gel was well tolerated for pa-tients. Attributed toxicity was primarily restricted to the applica-tion site and generally mild to moderate in severity.. The most

common adverse effect was local irritation, the severity of whichcould be limited by patients reducing the frequency of topicalapplications. No clinically significant, treatment-emergent trendsin any laboratory parameters were observed. Prior studies withalitretinoin gel, including a phase I-II clinical program of 115

Table VII. Most common treatment-related adverse events for alitretinoin gel during double-blind and open-label protocols mapped to COSTART 5 dictionarypreferred terms (n = 120)

COSTART 5 preferred term Representative investigator verbatim terms Overall number of patientsexperiencing the reaction (%)

Rash Erythema, scaling, irritation, redness, rash, dermatitis 59 (49)Pruritus Itching, pruritus 24 (20)Skin disorders Flaking, peeling, desquamation, exfoliation 17 (14)Pain Burning, pain 15 (13)Skin discoloration Skin discoloration 15 (13)Exfoliative dermatitis Excoriation, crusting scab, cracking, drainage, eschar, fissure, oozing 14 (12)Paresthesia Tingling, stinging, numbness 9 (8)COSTART = Coding System Thesaurus for Adverse Reaction Terms.

Table VIII. Adverse events recorded during double-blind protocol regard-less of relatedness to study compound and dosage adjustments

Adverse event Alitretinoin gel0.1% (n = 62)

Vehicle gel (n= 72)

Any adverse event (%)a 86 82At application site (%)Rash 32 11Parasthesia 6.5 28Itch 9.7 8.3Pain 3.2 13Peeling 6.5 2.8At other sites (%)Infection 13 22Rash 15 6.9Nausea 6.5 14Headache 6.5 2.8Rhinitis 3.2 5.6Herpes simplex 6.5 1.4Asthenia 1.6 5.6Abdominal pain 1.6 5.6Bodyweight loss 1.6 5.6

Serious adverse events [n (%)] 4b (6.5) 16 (22)Dose alterations due to adverse events [n (%)]Reduction to once-daily 9 (15) 1 (1.4)Discontinuationc 2 (3.2) 0 (0)Deaths (n) 0 1d

a Adverse events with ≥5% incidence in either treatment group.b All different adverse events occurring at non-application sites in differ-

ent patients.c Two cases of grade 3 dermal irritation.d Related to advanced HIV disease.



patients[20] and a second phase III trial of 268 patients,[23] reportedprimary local adverse effects at least possibly related to the useof the drug. The phase I-II program did not detect any differencein systemic levels of endogenous 9-cis retinoic acid during topicalalitretinoin gel treatment compared with untreated individuals.[20]

The cosmetic effects of topical irritation with alitretinoin gel aretemporary and reversible upon cessation of treatment. The verylow risk of adverse effects with this treatment is supportive of itspotential for self-administration.

Alitretinoin gel offers several advantages to the KS patient.As a topical agent, it does not require frequent clinical visits. Theactive agent, 9-cis retinoic acid, is a natural, endogenous com-pound. There is a low risk of drug-related systemic adverse ef-fects. By contrast, other local therapies for KS may have moreserious consequences. Radiation therapy can be complicated bymucositis, nausea, or fibrosis.[25] Cryotherapy may cause scarringor skin depigmentation.[26] Intralesional therapy may lead to pain,blistering, alopecia, and prolonged healing.

Results of this international, multicenter, randomized, vehi-cle-controlled, double-blind phase III trial complement the re-sults of the North America phase III trial[23] in demonstrating thatalitretinoin gel is an effective and well-tolerated topical treatmentfor cutaneous AIDS-related KS. Treatment with alitretinoin gelcan be efficacious regardless of immune status (as reflected bybaseline CD4+ lymphocytes counts), history of prior anti-KStherapy (including refractory patients), or concurrent antiretrovi-ral therapy. The acceptable, low-risk safety profile of this topicaltreatment also contributes to a favorable risk-benefit ratio for itsuse in the treatment of cutaneous KS. Alitretinoin gel may alsoprovide AIDS-related KS patients with a convenient, non-inva-sive alternative treatment for KS lesions. Currently, alitretinoingel is approved in the US and Canada for the topical treatment ofcutaneous AIDS-related KS.

Acknowledgments

This study was supported by Ligand Pharmaceuticals Inc., San Diego,California, USA.

The following members of the International Panretin® Gel KS study groupparticipated in this study, enrolled study participants, or both.

Jonathan Anderson, The Carlton Clinic, Carlton, VIC, Australia; C.Aquilina, Hópital de la Grave, Toulouse, France; Philippe Canton, CHU deBrabois, Vandoeuvre-les-Nancy, France; Timothy Cooley, Boston MedicalCenter, Boston, MA, USA; Kenneth Fife, University Hospital, Indianapolis,IN, USA; Martin Fisher, Claude Nichol Centre, Brighton, Sussex, England;Stephen Follansbee, Institute for HIV Research and Treatment, San Fran-cisco, CA, USA; Jeff Goodgame, Central Florida Research Initiative, Mait-land, FL, USA; Stephen L. Green, Hampton Roads Medical Specialists,Hampton, VA, USA; Margaret Johnson, Royal Free Hospital, London, En-gland; Jean-Marie Lang, Hôpitaux Universitaire de Strasbourg, Strasbourg,France; Celeste Lebbe, Hopital Saint-Louis, Paris, France; Sam Milliken, St.Vincent’s Hospital, Darlinghurst, NSW, Australia; Peter Piliero, Albany Med-

ical College, Albany, NY, USA; Donald M. Poretz, Infectious Diseases Phy-sicians, Inc., Annondate, VA, USA; A. Pozniak, The Caldecot Centre, London,England; Harry Rosado-Santos, University of Utah, Salt Lake City, UT, USA;Eric Rosenthal, Hópital Cimiez, Nice, France; Peter Ruane, Tower InfectiousDisease, Los Angeles, CA, USA; Thierry Saint-Marc, Hópital Edouard Her-riot, Lyon, France; Margaret A. Slade, Taylor Square Private Clinic,Darlinghurst, NSW, Australia; Denis Spelman, Alfred Health Care Group,Prahran, VIC, Australia; M. Spittle, Clinical Oncology Dept., London, En-gland; Victor Stevens, Ligand Pharmaceuticals, Inc., San Diego, CA, USA;Alan M. Sugar, Hyannis, MA, USA; Daniel Vittecoq, Hópital Paul Brousse,Villejuif, France; E. Wilkins, North Manchester General Hospital, Manches-ter, England; Allergan Pharmaceuticals, Inc, Irvine, CA, USA

References1. Miles SA, Mitsuyasu RI, Aboulafia DM. AIDS-related malignancies. In: Jr

DeVita VT, Hellman S, Rosenberg SA, editors. Cancer: principles & practiceof oncology. 5th edition. Philadelphia: Lippincott-Raven Publishers, 1997:2445-67

2. Tappero JW, Conant MA, Wolfe SF, et al. Kaposi’s sarcoma. Epidemiology,pathogenesis, histology, clinical spectrum, staging criteria and therapy. J AmAcad Dermatol 1993; 28: 371-95

3. Chang Y, Cesarman E, Pessin MS, et al. Identification of herpesvirus-like DNAsequences in AIDS-associated Kaposi’s sarcoma. Science 1994; 266: 1865-9

4. Martin JN, Ganem DE, Osmond DH, et al. Sexual transmission and natural historyof the human herpesvirus 8 infection. N Engl J Med 1998; 338: 948-54

5. Lebbe C, Blum L, Pellet C, et al. Clinical and biological impact of antiviraltherapy with protease inhibitors on HIV-related Kaposi’s sarcoma. AIDS1998; 12: 45-9

6. Aboulafia DM. Regression of acquired immunodeficiency syndrome-related pul-monary Kaposi’s sarcoma after highly active antiretroviral therapy. Mayo ClinProc 1998; 73: 439-43

7. Bower M, Fife K, Landau D, et al. Phase II trial of 13-cis-retinoic acid for poorrisk HIV-associated Kaposi’s sarcoma. Int J STD AIDS 1997; 7: 518-21

8. Bonhomme L, Fredj G, Ecstein E, et al. Treatment of AIDS-associated Kaposi’ssarcoma with oral tretinoin. Am J Hosp Pharm 1994; 51: 2417-9

9. Von Roenn J, von Gunten C, Mullane M, et al. All-trans-retinoic acid (ATRA) inthe treatment of AIDS-related Kaposi’s sarcoma. IX International ConferenceAIDS; Berlin, 1993. Abstract PO-B12-1571

10. Nagpal S, Chandraratna RAS. Retinoids as anti-cancer agents. Curr Pharm Des1996; 2: 295-316

11. Angulo A, Suto C, Boehm MF, et al. Retinoid activation of retinoic acid receptorsbut not of retinoid X receptors promotes cellular differentiation and replicationof human cytomegalovirus in embryonal cells. J Virol 1995; 69: 3831-7

12. Szondy Z, Reichert U, Fesus L. Retinoic acids regulate apoptosis of T lymphocytesthrough an interplay between RAR and RXR receptors. Cell Death Differ 1998;5: 1-3

13. Levin AA, Sturzenbecker LJ, Kazmer S, et al. 9-cis-retinoid acid stereoisomerbinds and activates the nuclear receptor RXRα. Nature 1992; 355: 359-61

14. Zhang x, Hoffman B, Tran PVB, et al. Retinoid X receptor is an auxiliary proteinfor thyroid hormone and retinoic acid receptors. Nature 1992; 355: 441-6

15. Yu VC, Delsert C, Andersen B, et al. RXBβ: a coregulator that enhances bindingof retinoic acid, thyroid hormone, and vitamin D receptors to the cognate re-sponse elements. Cell 1991; 67: 1251-66

16. Kliewer SA, Umesono K, Mangelsdorf DJ, et al. Retinoid X receptor interactswith nuclear receptors in retinoic acid, thyroid hormone and vitamin D3signal-ing. Nature 1992; 355: 446-9

17. Bissonette RP, Brunner T, Lazarchik SB, et al. 9-cis-retinoic acid inhibition ofactivation-induced apoptosis is mediated via regulation of Fas ligand and re-

86 Bodsworth et al.


quires retinoic acid receptor and retinoid X receptor activation. Mol Cell Biol1995; 15: 5576-85

18. Smith MA, Parkinson DR, Cheson BD, et al. Retinoids in cancer therapy. J ClinOncol 1992; 10: 839-64

19. Boehm MF, Heyman RA, Patel S, et al. Retinoids: biological function and usein the treatment of dermatological diseases. Exp Opin Invest Drugs 1995; 4:593-612

20. Duvic M, Friedman-Kien AE, Looney DJ, et al. Topical treatment of cutaneouslesions of AIDS-related Kaposi’s sarcoma using 9-cis-retinoic acid (alitretinoin):results of phase I-II trials. Arch Dermatol. 2000; 136: 1461-9

21. Krown SE, Metroka C, Wernz JC. Kaposi’s sarcoma in the acquired immune defi-ciency syndrome: a proposal for uniform evaluation, response, and staging cri-teria. J Clin Oncol 1989; 7: 1201-7

22. O’Brien PC, Fleming TR. A multiple testing procedure for clinical trials. Biometrics1979; 35: 549-56

23. Walmsley S, Northfelt DW, Melosky B, et al. Treatment of AIDS-related cutaneousKaposi’s sarcoma with topical alitretinoin (9-cis-retinoic acid) gel. J Acquir Im-mune Defic Syndr 1999; 22: 235-46

24. Conant MA, Ope KM, Poretz D, et al. Reduction of Kaposi’s sarcoma lesions fol-lowing treatment of AIDS with ritonavir. AIDS 1997; 11: 1300-1

25. Nisce L, Kaufman T. Radiation therapy for Kaposi’s sarcoma. Infect Med 1993; 10:54-8

26. Tappero, JW, Berger TG, Volberding PA, et al. Cryotherapy for cutaneous Kaposi’ssarcoma (KS) associated with acquired immune deficiency syndrome (AIDS): aphase II trial. J Acquir Immune Defic Syndr 1991; 4: 839-46

Correspondence and offprints: Dr Neil Bodsworth, Taylor Square PrivateClinic, 302 Bourke Street, Darlinghurst NSW 2010, Australia.E-mail: [email protected]



phase iii vehicle-controlled, multi-centered study of topical alitretinoin gel 0.1% in cutaneous...

Documents