| |
Screening HIV-Infected Individuals for Anal Cancer Precursor Lesions: A Systematic Review
|
| |
| |
Clinical Infectious Diseases July 15, 2006;43:223-233
Elizabeth Y. Chiao,1,2 Thomas P. Giordano,1,2 Joel M. Palefsky,4 Stephen Tyring,3 and Hashem El Serag1,2
1Baylor College of Medicine, 2Houston Center for Quality of Care and Utilization Studies, Health Services Research and Development Service, Michael E. DeBakey Veterans Affairs Medical Center (152), and 3University of Texas Health Science Center at Houston, Houston, Texas; and 4University of California, San Francisco, San Francisco
ABSTRACT
Individuals with human immunodeficiency virus (HIV) infection are at increased risk for human papillomavirus-related squamous cell cancer of the anus. Screening HIV-infected patients for squamous cell cancer of the anus and human papillomavirus-related anal dysplasia may prevent excess morbidity and mortality. We have conducted a systematic review of the indirect evidence in the literature regarding the utility of anal Papanicolau (Pap) smear screening of HIV-infected individuals in the highly active antiretroviral therapy era. Although there are no published studies evaluating the efficacy of anal Pap smear screening for preventing squamous cell cancer of the anus or anal intraepithelial neoplasia, we reviewed data regarding the burden of disease, anal Pap smear sensitivity and specificity, the prevalence of anal dysplasia, and 1 cost effectiveness study. The available evidence demonstrates that HIV-infected individuals have an increased risk for squamous cell cancer of the anus and anal intraepithelial neoplasia. This review identifies important areas for further study before routine anal Pap smear screening can be recommended.
Over the past 20-30 years, the incidence of squamous cell cancer of the anus (SCCA) in the United States has increased by about 96% in men and about 39% in women [1, 2]. Among men who have sex with men (MSM), the incidence of anal cancer has been estimated to be 35 cases per 100,000 population, which is comparable to the incidence of cervical cancer before the introduction of routine cervical Papanicolau (Pap) screening [3-7]. Furthermore, the risk for SCCA among HIV-infected MSM has been estimated to be twice that among non-HIV-infected MSM [8].
On the basis of multiple epidemiologic and histopathologic associations, SCCA is thought to behave like cervical cancer. Both anal and cervical cancer share an etiologic link to high-risk types of human papillomavirus (HPV) infection [9-12], and they share similar cytologic grading systems for dysplasia [11]. The prognosis of SCCA, like that of many other cancers, is strongly associated with the stage of disease at diagnosis [13]. A recent analysis of anal cancer outcomes from 1972 through 2000 found that survival was significantly improved for patients who received a diagnosis of local disease (5-year survival rate, 78%), compared with those who received a diagnosis of regional disease (5-year survival rate, 56%) or distant disease (5-year survival rate, 18%) [1]. Thus, some practitioners have adopted anal Pap smear screening as a standard intervention in high-risk populations, including MSM and HIV-positive individuals.
Anal Pap smear screening involves inserting a swab blindly into the anal canal and fixing the cells either on a slide or in fluid for liquid cytological examination. Anal cytology is classified on the basis of the revised Bethesda System of cervical cytology classification: normal, atypical squamous cells of uncertain significance, low-grade squamous intraepithelial lesion, or high-grade squamous intraepithelial lesion. Abnormalities detected by anal Pap smear are further evaluated by high-resolution anoscopy (HRA), which is similar to cervical colposcopy [14].
No randomized clinical trials exist to support anal Pap smear screening. Thus, practitioners rely on indirect evidence to inform their decisions regarding anal Pap smear screening. Sackett et al. [15] proposed a set of principles to evaluate the utility of early cancer screening programs (table 1). We have performed a systematic review of the literature published in the HAART era using the principles of Sackett and colleagues to evaluate and summarize the indirect evidence concerning anal Pap smear screening for HIV-infected patients.
METHODS
We performed a systematic review of the English-language literature published from July 1996 through July 2005. We searched the National Library of Medicine PubMed database, meeting abstract databases, relevant journals, and reference lists to obtain publications on anal Pap smear screening. The search terms used were "anal intraepithelial neoplasia," "anal cancer," and "HIV." Abstracts (including those from case series and reports) were reviewed from potentially relevant titles if the study addressed any of the following issues: (1) evaluation of screening programs, (2) estimation of the number of patients needing evaluation, (3) compliance with and tolerability of screening, (4) effectiveness of the screening test and treatments, (5) evaluation of the burden of disability from anal cancer, and (6) the cost of the screening test and treatment. Studies were excluded if they did not include HIV-infected individuals, did not include original data, or did not address outcomes related to the issues described above. One reviewer assessed all potential abstracts and publications.
RESULTS
The initial search yielded 181 abstracts. After applying the selection criteria, 63 articles were found to contain information evaluating anal Pap smear screening (figure 1). The distribution of articles addressing each of the criteria is shown in table 1.
Does Screening HIV-Positive Individuals with Anal Pap Smears Lead to Improved Outcomes?
The effectiveness of screening with anal Pap smears for anal cancer prevention among HIV-positive individuals is not known. No randomized or cohort studies exist to determine if there are improved survival or outcomes for those who have been screened with anal Pap smears. Furthermore, there have been no ecologic studies to correlate use of screening and the incidence or outcomes of SCCA.
What is the Burden of Disability from Anal Cancer among HIV-Positive Individuals?
The incidence of SCCA among men in the general population is 0.8 cases per 100,000 population [1]. However, the incidence of SCCA among HIV-infected MSM is 70 per 100,000 population [8]. Recent studies (table 2) have examined either HIV-infected cohorts or AIDS-Cancer Registry-matched data to evaluate the incidence and risk of SCCA among HIV-infected individuals. Retrospective AIDS-Cancer Registry match studies found that standardized incidence ratios (generally standardized for age and sex) varied from 3.3 to 352.0 [7, 16-18]. The standardized incidence ratios for anal cancer among women in these studies ranged from 3.0 to 6.8 [7, 16]. In comparison, similar studies have found that the standardized incidence ratios for cervical cancer among HIV-positive women range from 8.0 to 21.8 [7, 16, 19-69]. There are no data evaluating HIV-infected MSM versus HIV-infected men who do not have sex with men.
Cohort studies involving HIV-positive individuals found that the incidence of SCCA ranged from 3.9 to 92 cases per 100,000 population and that the relative risks, compared with the general population, ranged from 33.4 to 222 [19, 20-22]. None of the studies found a statistically significant difference between SCCA incidences in the pre-HAART era versus the HAART era.
In addition to the increased risk, the SCCA-related morbidity and mortality among HIV-infected individuals is substantial. Combined chemoradiotherapy is the primary treatment of choice, with surgical treatment reserved for relapsed disease. The 5-year survival rate in the general population is 70%-80% [70, 71]. Only 8 small case series (ranging in size from 4 to 26 patients) describe outcomes of HIV-associated SCCA, with 5-year survival rates ranging from 47%-60% (table 3) [72-79]. The majority of cases were diagnosed at either stage 1 or stage 2 (localized tumor without lymph node invasion) and were treated with combined chemoradiotherapy. Most series reported some toxicity associated with therapy, and 1 series reported that up to 50% of patients were unable to complete planned therapy [23-29]. Among 3 studies that specifically compared survival among patients with SCCA in the pre-HAART era versus the HAART era, there was a nonsignificant trend toward improved survival, better tolerability of chemoradiotherapy, and improved local tumor control in the HAART era [22, 23, 24, 29].
Has the Effectiveness of the Components of an Anal Pap Smear Screening Program Been Demonstrated?
The main components of anal Pap smear screening are (1) anal Pap smear screening followed by HRA-guided biopsy and (2) anal dysplasia treatment.
Operating characteristics of anal Pap smear screening. Eight studies evaluated the effectiveness of anal Pap smears by calculating the sensitivity and specificity of the test through a comparison between Pap smear results and results of histological examination of HRA-directed biopsy samples (table 4). Excluding 2 studies [30, 31] in which the methodology differed from that of the others, the sensitivity ranged from 69% to 93%, and the specificity ranged from 32% to 59%. In a study involving the largest prospective cohort, Palefsky et al. [32] reported that the positive predictive value increased from 38% to 78% and that the negative predictive value increased from 46% to 79% after 2 years of recurrent screening. In addition, agreement between concurrent cytological examination and biopsy results has been reported ato be 74.7% with a weighted k1 statistic of 0.36 [33]. These findings are comparable to those of studies comparing cervical cytological examination and cervical biopsy results, which reported that the absolute agreement and the weighted k1 agreement ranged from 64% to 91% and from 0.18 to 0.65, respectively [80, 81].
The intraobserver and interobserver agreement among 3 pathologists have been reported as moderate [82]. Lytwyn et al. [83] reported that the weighted k1 statistic ranged from 0.72 to 0.92 between each pathologist and the consensus interpretation of all 4 pathologists. This range of interobserver reliability for anal Pap smears is similar to the reliability cited for cervical cytological and histological findings [84-86].
Self-collected anal Pap smears have also been evaluated [87, 88]. Although the sensitivity of self-collected specimens was 68%, compared with a sensitivity of 70% for clinician-collected samples, larger studies are necessary to confirm the performance characteristics of this screening method [88].
What is the efficacy of treatment? Treatment outcomes for anal dysplasia have only been reported for small case series. There are no efficacy studies or randomized clinical trials. Large loop excision of the transformation zone has been shown to effectively and safely treat cervical intraepithelial neoplasia lesions, with a reported 95% cure rate at 1 year among HIV-negative women [89]; however, in HIV-positive women, the recurrence rate has been shown to be significantly higher than that for HIV-negative women (73% vs. 27%; P = .019) [90]. In addition, surgery for anal intraepithelial neoplasia may not be as effective as loop excision of the cervical transformation zone, because it is not possible to excise the entire transformation zone in the anus. In a recent series of 29 HIV-positive patients who underwent electrocautery of dysplastic lesions, 80% of patients with high-grade anal intraepithelial neoplasia had persistence or recurrence of lesions, with a mean time of 1 year to recurrence. Thus, the authors recommended that HIV-positive patients undergo multiple staged procedures and continued surveillance [38]. There is also concern that HIV-positive patients may have impaired healing after anal surgeries [91]. Therefore, some surgeons also recommend that an initial conservative surgical approach, confined to ruling out invasive cancer and removing foci of symptomatic lesions, may be preferable to wide excision [92].
In an uncontrolled case series, Goldstone et al. [39] found that infrared coagulation adequately and safely treated high-grade anal dysplasia in a group of 68 HIV-positive men. After 3 treatments, Goldstone et al. [39] found a 72% probability of cure in retreated lesions after a median follow-up period of 413 days. No patient developed squamous-cell carcinoma, and there were no serious adverse events, including stenosis.
Topical therapies, including imiquimod and topical 5-fluorouracil, have also been used. Kreuter et al. [40] reported on the use of 5% imiquimod therapy for different grades of dysplasia in 10 HIV-positive MSM. After 16 weeks of self-applied therapy, histologic studies showed regression of disease by at least 2 grades and no HPV-16 detected by PCR. Graham et al. [41] found that, in a cohort of 11 patients, 5% topical 5-fluorouracil may have some activity in anal Bowen's disease. Finally, Pehoushek et al. [42] reported a combined approach with topical imiquimod and 5% topical 5-fluorouracil for a period of 16 weeks in 1 HIV-infected man with high grade dysplasia who, despite developing significant skin irritation, had no residual disease at 3 months.
In addition, candidate therapeutic HPV vaccines have been evaluated as immunomodulators to treat anal dysplasia progression. None of the studies that we reviewed were placebo controlled, and trials involving HIV-positive patients are pending. A phase-1 trial involving HIV-negative patients of HLA-A2-restricted epitopes derived from the HPV-16 E7 protein showed that the vaccine was well tolerated at all dose levels. There were partial histological responses of high-grade anal intraepithelial neoplasia in 25% of subjects [43]. Palefsky et al. [44] evaluated a vaccine using heat shock protein Hsp65 from bacilli Calmette-Guerrin fused to E7 protein from HPV-16. Among 82 HIV-negative patients with high-grade dysplasia, one-half showed a relatively durable regression from high-grade dysplasia to normal.
Effect of HAART on anal intraepithelial neoplasia. Studies evaluating HAART initiation for high-grade dysplasia treatment have yielded conflicting results. Table 5 summarizes the results of 2 cohort studies and 3 cross-sectional studies evaluating the effect of HAART on HPV-related anal diseases. Palefsky et al. [45] found that, after 6 months of HAART, the likelihood of lesion progression or regression was not affected, but they noted that, among the patients starting HAART at higher CD4+ cell counts, HAART was associated with a nonsignificant benefit with regard to anal dysplasia lesions. In a subsequent cross-sectional study, Palefsky et al. [46] found that men receiving HAART had a 12-fold increased risk (OR, 12; 95% CI, 2.4-64) for anal intraepithelial neoplasia grade 2 or grade 3 after adjustment for CD4+ cell count and length of time that the patient had been HIV positive. Another cross-sectional study of 45 HIV-positive men in France found that the prevalence of anal dysplasia and HPV infection were similar in patients with and patients without significant CD4+ cell count increases attributable to HAART [47]. However, cross-sectional studies done by Kiviat et al. [48] and Wilkin et al. [49] showed decreasing anal dysplasia prevalence after HAART initiation. In a multivariable analysis, Wilkin et al. [49] found that HAART was significantly protective for anal dysplasia (OR, 0.09; P < .01) but was not significantly protective for HPV infection (OR, 0.48; P = .3). Carefully designed prospective cohort studies will be necessary to determine the effect of HAART on anal dysplasia.
Is the Clinical Burden Manageable?
Seventeen screening studies gave estimates of the percentage of HIV-positive individuals who had dysplasia detected by anal Pap smear; 8 studies were restricted to HIV-positive men and gave estimates of 41% to 97%, and 4 studies were restricted to HIV-positive women and gave estimates of 14%-28% (table 6). Of note, one study showed comparable rates of high-grade dysplasia in MSM and men who denied receptive anal intercourse [50].
The large variability between studies is likely related to bias associated with referral patterns to anal dysplasia clinics. Using the median prevalence of anal dysplasia reported in these studies and using the number of persons estimated to be receiving treatment for HIV infection in the United States [93], we can estimate the clinical burden related to anal Pap smear screening. Assuming that 70% of the 444,900 patients in HIV care per year are male, we project that about 146,372 HIV-infected men and about 18,686 HIV-infected women throughout the United States receiving ambulatory HIV care would have abnormal anal Pap smears. These patients would need ongoing evaluation for anal dysplasia, and these numbers do not take the incidence of new lesions into account. Palefsky et al. [65] found that 32% of the participants with normal cytology and 52% of the participants with low-grade dysplasia at baseline developed subsequent high-grade lesions during a 4-year follow-up period. Thus, a widespread anal Pap smear screening program in the United States would require substantial additional physician resources and other practitioner resources to train for and conduct HRA-related activities.
How Compliant Are HIV-Positive Patients with Post-Anal Pap Smear Follow-Up?
Screening tests are only effective if subjects comply with health recommendations made on the basis of the screening test result. There are few studies involving select cohorts evaluating the willingness of HIV-positive patients to comply with necessary follow-up studies and treatments resulting from positive anal Pap smear results. Palefsky et al. [65] reported that about 80% of 608 subjects were seen after their initial baseline visit. Matthews et al. [33] reported that, among a cohort of 1732 HIV-positive patients receiving anal Pap smears outside of a clinical trial, 642 (40%) had additional testing within 3 years of the initial anal Pap smear within the same health care system. The low rate of follow-up anal Pap smears may reflect low compliance related to the relative novelty of the test, the lack of consistent guidelines, or the lack of capacity to manage the additional clinical burden resulting from testing.
Is Anal Pap Smear Screening Cost-Effective?
The final criterion for implementation of anal cancer screening is whether the cost-taking into account the accuracy, acceptability and effectiveness of the anal Pap smear-is adequate for the purpose of screening. Goldie et al. [66] developed a state-transition Markov model to calculate lifetime costs, life expectancy, and quality-adjusted life expectancy for several different screening strategies using the anal Pap smear versus no screening for anal dysplasia and anal squamous cell cancer among HIV-positive MSM. They found that the screening of HIV-positive MSM with anal Pap smears once per year provided an incremental cost-effectiveness ratio of $16,600 per quality-adjusted life year saved. This ratio is similar to that for other widely accepted screening procedures and for preventive interventions, such as screening for colon cancer in the general population [94] and instituting Pneumocystis jiroveci prophylaxis at CD4+ cell counts of <200 cells/mL among HIV-positive individuals [66, 95]. A sensitivity analysis found that these results were most sensitive to the rate of progression of dysplasia to SCCA and to the effectiveness of treatment of precancerous lesions. Although the authors used rates of progression to high-grade dysplasia obtained from published anal dysplasia cohort studies, it should be emphasized that neither rates of progression from high-grade dysplasia to cancer nor efficacy of anal dysplasia treatments have been well characterized in the literature.
DISCUSSION
We have performed a systematic review of the literature to evaluate the utility of anal Pap smear screening among HIV-positive patients by applying the criteria developed by Sackett et al. [15] for determining screening and surveillance efficacy. Anal dysplasia is prevalent in the HIV-positive population, and although the incidence of SCCA is low in the general population, the incidence is increasing among men [2, 96]. The operating characteristics of the anal Pap smear have been evaluated and are of modest accuracy, with accuracy similar to that of the cervical Pap smear.
However, the data published to date highlight limitations to current anal Pap smear screening-related research. First, there has been no randomized trial showing improved survival with anal dysplasia screening. A prospective, randomized, controlled trial evaluating anal Pap smears, however, would need to be very large and would take many years to complete. Indeed, there have been no randomized studies evaluating the outcomes of cervical Pap smear screening, for similar reasons; however, ecologic studies have shown significant decreases in cervical cancer rates in areas of widespread and prolonged screening [97]. Although anal Pap smear screening is not being performed on a large enough scale to conduct ecologic studies, future anal Pap smear screening efficacy studies may instead focus on comparing different types and intensity of screening methods.
Second, few trials have evaluated treatment strategies for high-grade dysplasia. Studies demonstrate that current treatments for HIV-related anal dysplasia have, at best, moderate efficacy. In addition, commonly used therapies, such as treatment with 80%-90% trichloroacetic acid or liquid nitrogen, which have been described for the treatment of high-grade anal dysplasia [98], have not been studied. Other techniques, such as the use of topical cidofovir [99] and IFN [100], are currently being evaluated for the treatment of high-grade anal dysplasia, but published studies are again lacking.
Third, the natural history of progression from anal intraepithelial neoplasia to anal cancer is unknown. Although low-grade anal dysplasia has been shown to progress to high-grade dysplasia in a majority of HIV-infected men within 2 years after initial diagnosis, the true rate of progression from high-grade dysplasia to invasive anal cancer remains unclear.
Our review of the literature regarding the utility of anal Pap smear screening for anal cancer among HIV-infected individuals is limited, because there are no published randomized clinical trials evaluating many aspects of anal Pap smear screening. Until a consensus regarding anal Pap smear screening can be reached, a heightened awareness of anal cancer among HIV-infected individuals is warranted. Further research is needed to identify improved methods for preventing, detecting, and treating anal dysplasia.
|
|
| |
| |
|
|
|
