|Home | About | Journals | Submit | Contact Us | Français|
To evaluate the safety and efficacy of topical cidofovir for treatment of high-grade squamous perianal and vulvar intraepithelial neoplasia (PAIN and VIN) lesions in HIV-positive individuals.
Phase IIa prospective multicenter trial conducted at eight clinical sites through the AIDS Malignancy Consortium (AMC)
HIV-positive patients with biopsy-proven high-grade PAIN that was ≥ 3 cm2 were enrolled. PAIN biopsy specimens were assessed for HPV using PCR and type-specific HPV probing. Subjects applied 1% topical cidofovir to PAIN and VIN (if present) for 6 two-week cycles. Results were designated as complete response (CR), partial response (PR) (> 50% reduction in size), stable disease (SD), or progressive disease (PD).
Twenty-four men and 9 women (8 with high-grade VIN as well) were enrolled. Mean age was 44 years, mean CD4+ count was 412 cells/μl. HPV DNA (most commonly HPV16) was detected in all pre-treatment study specimens. Twenty six (79%) subjects completed treatment per protocol—CR: 5 (15%); PR: 12 (36%), SD: 7 (21%); PD: 2 (6%) (1 with a superficially invasive cancer and 1 with new area of high-grade PAIN). Treatment was well tolerated with most common adverse events being mild to moderate affecting lesional skin: pain/burning/irritation (25 subjects) and ulceration (13 subjects).
Topical cidofovir had 51% efficacy in the short-term treatment of high-grade PAIN and VIN with acceptable toxicity in HIV-positive individuals. Randomized control studies with more prolonged treatment courses and longer follow-up to assess the durability of the response are needed.
HPV-associated cancers of the perianus and vulva in HIV-infected individuals continue to be a significant cause of morbidity in the era of highly active antiretroviral therapy (ART).  High-grade squamous perianal intraepithelial neoplasia (PAIN) and vulvar intraepithelial neoplasia (VIN), the precursors to invasive squamous cell cancer, are particularly difficult to treat due to their multifocal presentation and tendency for recurrence. [2-4]
Standard therapies for diffuse high-grade perianal or vulvar disease include laser ablation, cryotherapy, electrocautery, and surgical excision with possible skin grafting or skin flaps [2-4] all of which carry significant morbidity such as prolonged post-procedure pain, scarring, risk for infection, incomplete healing, incontinence, and sexual dysfunction. [5-8] Immunocompetent patients have recurrence rates after treatment of 25 to 50%  and HIV-positive patients have recurrence rates of 60 to 80%. [9-11]
Topical treatments such as imiquimod and 5-fluorouracil creams have some efficacy for treatment of high-grade PAIN and anal canal intraepithelial neoplasia (AIN) in HIV-infected men. [12-15] However, treatment with these agents have substantial skin and mucosal toxicity that often leads to discontinuation of treatment. There are few reports of topical treatments applied to high-grade VIN and PAIN in HIV-infected women. [7,16]
Cidofovir is a cytidine nucleotide analogue with in vitro and in vivo activity against a broad spectrum of herpesviruses as well as adenoviruses, human papillomaviruses (HPVs), polyomaviruses, and human poxviruses.  It is currently licensed for intravenous treatment of cytomegalovirus (CMV) infections. The mechanisms of action of cidofovir on HPV-associated neoplasia are not fully elucidated. [18, 19]
Several small studies have shown near-complete clearance rates of genital warts, with acceptable toxicity, using topical cidofovir in HIV-infected patients. [9, 20-23] Because of the clinical efficacy and tolerability for warts reported in these studies, we conducted a pilot study to evaluate the safety and efficacy of topical cidofovir in HIV-infected patients for treatment of high-grade squamous intraepithelial neoplasia of the perianus and vulva.
AIDS Malignancy Consortium (AMC) Protocol 046 was a multicenter, single-arm, open-label, pilot trial conducted at 8 AMC clinical trial sites (Boston University Medical Center, Boston, MA; Laser Surgery Care, New York, NY; Montefiore Medical Center, Bronx, NY; University of California, San Francisco, CA; Virginia Mason Medical Center, Seattle, WA; Weill-Cornell Medical College, New York, NY; Beth Israel Deaconess Medical Center, Boston, MA; University of California, Los Angeles, CA). Institutional review boards of the participating institutions approved the study. Each participant gave written informed consent (clinicaltrial.gov unique identifier: NCT00550589).
We enrolled adult men and women with documented HIV-1 infection and biopsy-proven high-grade PAIN within 12 weeks of study entry covering a surface area of at least 3 square centimeters. The perianus was defined as the skin within a 5 cm radius of the anal verge and thereby included women with high-grade VIN on the posterior perineum. Patients receiving ART had to be on a regimen for at least 12 weeks prior to entry. If they were not receiving ART, the CD4+ t-cell level needed to exceed 200 cells/mm3. For patients receiving ART, any CD4+ t-cell level was acceptable.
Cidofovir, was provided by Gilead Sciences, Inc (Foster City, California) and compounded by the study compound pharmacy (MasterPharm, LLP, Richmond Hill, New York) into a 1% topical cream packaged in 10 gm tubes, sufficient for 2 treatment cycles. The vanishing cream base is an emulsion of mineral oil, deionized water, cetyl alcohol, ceresin wax, beeswax, and sodium borate. Each batch underwent testing by a third party (Eagle Analytical Services, Houston, Texas) to confirm potency and uniformity. Approval of the compounded product by the Food and Drug Administration was obtained prior to initiating the trial (Investigational New Drug 74,757).
Lesion measurements were taken from the photo-documentation of high-grade PAIN and VIN with the skin on gentle stretch after application of 3-5% acetic acid. A software program (Second Opinion© Torrance, California or DermImage, New York, New York) was used to calculate the total lesion area from the digital images.
Pre-treatment specimens for histology and correlative studies included two 3-mm biopsies of the lesions. One specimen was placed in formalin. The other biopsy was flash-frozen for HPV DNA typing. All biopsy sites were photo-documented, so that biopsies collected during and after treatment could be taken from the same location(s).
The participants self-applied the study cream sparingly in a thin layer over the affected areas with a gloved finger once daily and then washed the cream off six to eight hours later. Women with high-grade VIN were instructed to apply the treatment medication to both the high-grade VIN and PAIN. Treatments occurred for 5 consecutive days followed by no treatment for 9 days (a 14-day treatment cycle). Participants received a treatment diary to record the timing of treatment and any side effects during each week treatment cycle.
Following each treatment cycle, a study clinician reviewed the participant’s treatment diary and examined the participant to assess local toxicity and clinical status of lesion(s). Pre-cycle laboratory assessments included serum creatinine, urine protein, and complete blood counts. Absent complications requiring dose delay, the participant received study medication for the next treatment cycle. Therapy continued until completion of six treatment cycles.
Study participants underwent a final evaluation six weeks after completion of the sixth treatment cycle. Lesion measurements and two biopsies for histology and molecular studies were taken from the residual lesion, if present, or from the original lesion site
Complete response (CR) was defined as the absence of any high-grade PAIN or VIN or cancer. A partial response (PR) was defined as no new lesions and a 50% or greater decrease in total lesion area of high-grade PAIN or VIN. Progressive disease (PD) was defined as either ≥ 25% increase in the total lesion area, or biopsy-proven invasive perianal or vulvar cancer. Stable disease (SD) included responses that were not defined in the other response definitions.
Adverse events were graded according to version 4.0 of the NCI Common Terminology Criteria for Adverse Events (CTCAE).
The formalin-fixed, paraffin-embedded biopsy specimens were sectioned for H&E histologic confirmation by the study central pathologist (TD). The frozen specimen biopsies were processed for histologic confirmation, followed by HPV DNA PCR testing. HPV DNA typing was performed using MY09/MY11 consensus HPV-L1 primers as well as primers for amplification of the human β-globin gene, as a control. 
All study specimens (formalin and frozen) were read by two pathologists. Histologic interpretations were categorized as (1) benign or low-grade PAIN or VIN, (2) high-grade PAIN or VIN, or (3) cancer. If there was a discrepancy in histologic interpretation of a specimen between the two central pathologists, then the specimen was reviewed by a third pathologist. Final determination of histologic interpretation was determined by agreement of at least 2 pathologists. Each study pathologist was blinded to the interpretation of the others.
33 participants were enrolled to detect a 50% combined PR and CR (with one-sided 0.05 significance level with power of 0.95) compared with an assumed natural regression rate of 20% (allowing a 10% participant drop-out/withdrawal rate).
Univariate and multivariate logistic regression analyses were used to compare study participants with a clinical response (PR or CR) to study participants without a significant response (PD or SD) for possible predictors of response (pre-treatment total lesion size, presence of HPV16, CD4+ cell count, and undetectable HIV viral load). All analyses were performed in SAS version 9.1, SAS Institute, Inc, Cary, NC. Final results are interpreted using an α = 0.05 level of significance.
Between February 2008 and August 2009, 47 participants were screened and 33 were enrolled. Reasons for screen failures included: participants did not have biopsy-proven high-grade PAIN (6), failure to meet laboratory eligibility criteria (4), clinician judgment of an unacceptable risk for cancer (2), voluntary withdrawal prior to treatment (1) and one potential participant was screened in error. The median age at enrollment was 44 years (range 24 to 66), 24 (73%) of the participants were male,12 (36%) were African-American, and eight (24%) were Hispanic. Eight of the nine enrolled female participants had both high-grade VIN and high-grade PAIN. At the time of enrollment, median CD4+ count was 412 cells/mm3 (range 2-1152), median HIV viral load was <75 copies/mm3, and 97% of the participants were using ART. The median baseline total lesion size was 6.6 cm2 (SD 5.4, range 3.0-21.3) (Table 1).
The 33 enrolled study participants who completed at least one cycle of treatment constituted the intention-to treat (ITT) group. Twenty-six of the 33 (78.8%) participants completed the study per protocol and were included in the per-protocol analysis. Four participants were lost to follow-up, two withdrew because of mild adverse events (grade 1 bleeding and anal pain) and one participant was excluded from analysis because the consensus pathology review of the enrollment biopsy was low-grade. Overall, five of the 33 in the ITT group (15%) had a CR, 12 (36%) had a PR, 7 (21%) had SD, and 2 (6%) had PD. Response rates for the per-protocol group (n=26) were CR, 19%; PR, 46%; SD, (31%); and PD, (8%). Of the two participants with PD one was diagnosed with invasive cancer and the other with an increase in high-grade PAIN lesion size. The overall clinical benefit rate (CR+PR) for participants among the ITT population was 17/33 (51.5%); one-sided 95% confidence interval (36.1%, 100%). The response rate for participants in the per-protocol group was 17/26 (65.4%); one-sided 95% confidence interval (47.4%, 100%). The outcomes for the eight women treated for both high-grade VIN and PAIN were: CR,1 (13%); PR, 3 (38%); SD, 1 (13%); PD, 1 (13%); lost to follow-up, 2 (25%).
Thirty of 33 baseline specimens were available for HPV typing. The other specimens had insufficient tissue for analysis or significant freezing artifact limiting HPV testing. All 30 evaluable specimens were HPV positive. HPV 16 was detected in 16 (53%) of the evaluable baseline biopsy specimens. Nineteen study participants had pre- and post-treatment specimens evaluable for HPV typing. Comparing HPV types detected pre and post treatment: 5 patients had the same HPV types (4 with HPV16), 3 had the same HPV type with additional HPV types detected post-treatment, 9 had different HPV types detected, and only 1 patient had no HPV detected post-treatment (Table 2)
Risk factors including 1) median CD4+ cell count, 2) total lesional area, 3) detection of HPV16 in pretreatment specimen and 4) having undetectable HIV viral load were assessed for association with clinical response (CR or PR). No significant factors were associated with clinical response with univariate or logistic regression analysis, but this could be due to small numbers in each strata.
Adverse events (AEs) were reported for 32 (97.0%) participants.(Table 3). Most were localized to the skin at the lesions undergoing treatment—25 study participants complained of mostly mild and self-limiting pain/burning/irritation of the affected skin. Clinicians reported mild to moderate ulceration of the lesional skin in 13 study participants. Three patients experienced severe AEs. One participant had an invasive perianal squamous cell carcinoma at study completion which was treated with surgical excision. Another study participant had bacterial pneumonia and herpes zoster; and a third participant had Methicillin-resistant Staphylococcus aureus cellulitis on her thigh, distant from the site of treatment, that required parenteral antibiotics.
There are few reports of topical treatments applied to pre-invasive lesions of the vulva and perianus in persons with HIV [12-15] and no reports of topical cidofovir for this purpose in HIV-positive men or women. We report the first prospective, open-label, multi-site study evaluating topical cidofovir for the treatment of high-grade PAIN and VIN in HIV-infected patients. Our results show that over 65% of participants who completed the therapy had at least a PR, and although local skin reactions were very common, most study participants completed the treatment course.
We believe that our results are particularly important because treatment of extensive high-grade PAIN and VIN in HIV-infected patients is difficult. Surgical and ablative therapies are associated with significant morbidity. Often, the lesions recur and the repeat treatments cause additive morbidity. An effective topical treatment that is well tolerated has clear clinical benefits over the current surgical options. Even though complete clearance occurred in only 15%, the partial response in an additional one-third of treated patients is clinically beneficial. Many patients with flat low-grade disease may not require treatment and a reduction in high-grade lesion area may enable a targeted and less morbid excision or ablative procedure. 
Prior reports of locally-applied cidofovir have focused on the treatment of genital and/or perianal warts (condylomas) in HIV-infected patients and VIN in immunocompetent women. In studies of HIV-infected participants with genital warts treated with 1% cidofovir cream or gel, overall response rates of 70-90% were reported, with significantly less pain and a lower relapse rate compared with ablative procedures. [9,10, 17, 20-23] A series of 12 immunocompetent women with VIN3 treated with topical cidofovir found that four women had a CR, and three had PR. 
The diagnosis of cancer in one of our participants is concerning, and may represent an occult lesion that was present prior to treatment and not sampled by biopsy. However, it could also be failure of the study agent to prevent progression of the precancerous lesion to invasive cancer or non-adherence to treatment. Since these are typically heterogeneous and multifocal lesions, one small biopsy as indicated per protocol might not have identified the cancer. Wide local excision of high-grade VIN lesions is associated with detection of occult cancer in 3-5% of immunocompetent patients. In a placebo-controlled study of imiquimod 5% treatment of immunocompetent patients with high-grade VIN, 1 of 26 (4%) subjects in the treatment arm and 2 of 26 (8%) subjects in the placebo arm were diagnosed with cancer of the vulva within 7 months and 12 months, respectively.  Similar progression rates have been seen in other studies. [13, 25, 27] As in our study, the cancers detected in these studies could represent occult disease or true progression to cancer and highlight the inherent diagnostic limitations that should be considered when designing future studies.
This is the first study to evaluate HPV genotyping in HIV-infected patients with high-grade PAIN. HPV was detected in all lesions in the study, and HPV-16 was detected in 50% of these lesions; Duvuyst reports similar HPV genotypes in high-grade AIN in HIV-positive individuals where HPV was identified in 97% of high-grade AIN and HPV16 was identified in 55%.  It is unclear why 13 of 19 (68%) study participants had a new HPV type detected post-treatment. This may represent acquisition of a new HPV infection, activation of a previously subclinical HPV infection, or detection of a concurrent HPV type that was not identified in the pretreatment specimen.
Strengths of our study include its prospective multi-center trial design, inclusion of HIV-infected men and women, uniformity of the compounded study drug, HPV data analysis, colposcopic evaluation of lesion size with validated measurement software to better delineate lesions and response, and independent pathology review of specimens to determine inclusion and response in the trial.
Study limitations include the small sample size, lack of placebo arm, limited duration of treatment and short-term follow-up. In the absence of a placebo arm we were not able to distinguish the “true” response rate from spontaneous regression. However, prior studies of anal, perianal, and vulvar lesions have shown very low rates of natural regression, especially in the setting of HIV. None of the 26 immunocompetent participants in the placebo arm of the imiquimod study of high-grade VIN  and only 1 of 26 (4%) HIV-infected subjects in the placebo arm of the imiquimod study of AIN2-3  demonstrated regression of high-grade intraepithelial disease. Therefore, it is likely that the rate of spontaneous regression in HIV-infected individuals is low and thus the short-term responses to topical cidofovir are likely not due to natural regression, but due to the study drug. The standardization of treatment cycles in this study did not allow for prolonged treatment courses for patients with partial responses. In addition, the short term follow-up (6 weeks after completion of treatment) prevents our commenting on the durability of treatment response. High recurrence rates (>50%) following other forms of treatment for high-grade AIN have been documented for HIV-infected patients, even in the ART era.  Lastly, long-term toxicity or toxicity with repeated use of topical cidofovir are not known.
In summary, topical cidofovir had 51% efficacy in the short-term treatment of high-grade PAIN and VIN in HIV-infected individuals. Phase 3 trials should be conducted with more prolonged treatment courses and longer follow-up to assess the durability of the response.
We thank the following individuals who made substantial contributions to this investigation and the drafting of this manuscript but who did not meet the criteria for authorship: Oscar Lin, MD and Mark Stoler, MD, Paul Booth, Ronald Mitsuyasu. We would also like to thank the research nurses and staff at each of our respective institutions as well as the staff at EMMES, the operations center for the AMC and the patients who participated in this study.
This work was supported by the National Cancer Institute at the National Institute of Health [R21 CA117081 to EAS and UO1 CA 121947 to Ronald T Mitsuyasu, UCLA] Clinicaltrial.gov unique identifier: NCT00550589
Gilead Sciences Inc (Foster City, CA) provided the intravenous formulation of the study drug.
Potential conflicts of interest. All authors: no conflicts.
Contributions of authors.) Substantial contributions to conception and design (EAS, SEG, MHE, JYL, JMP); acquisition of data (EAS, SEG, MHE, NJ, JMB, LP, DA, JMP); pathology review (TMD); HPV analyses (MDC and JMP); analysis and interpretation of data (JYL, TMD, MDC, JMP); drafting the article (EAS); revising it critically for important intellectual content (all other authors)