Anal intraepithelial neoplasia (AIN) is the precursor to invasive anal squamous cell carcinoma (SCC) . Its incidence is rising in high-risk groups, particularly individuals infected with HIV [2,3]. Like cervical cancer, anal cancer progresses in a step-wise manner from low-grade dysplasia to high-grade dysplasia, which precedes micro-invasion. Anal SCC develops as a result of infection with oncogenic strains of the human papilloma virus (HPV), mainly HPV-16 and HPV-18 . However, the natural history of AIN is unclear, and management strategies are lacking.
Experience with treatment of cervical cancer shows that removal of intraepithelial neoplasia decreases the incidence of invasive disease. The same could be true of invasive anal SCC [5,6]. The most effective treatment for AIN, excision of the anal squamocolumnar junction, is associated with marked morbidity (anal stenosis, abscess formation, anal spasm, and dyschezia) [7,8]. Treatment of AIN in its early stages could be the best way to modify the natural history of the disease and thus avoid progression to invasive anal SCC. The most common approach to management of AIN during the last decade has largely centered on local ablation of individual lesions using laser therapy, cryotherapy, electrocautery, infrared coagulation (IRC), and topical agents (imiquimod, trichloracetic acid, and 5% 5-fluorouracil cream) [6,8–15].
IRC ablation is an effective treatment for high-grade squamous intraepithelial lesions (HSIL). However, recurrence of squamous intraepithelial lesions (SIL) has to be taken into account. In fact, few studies have been published on the efficacy or effectiveness of IRC [7,13–16]. Four of the five studies published [13–16] focused on HIV-positive men who have sex with men (MSM). Only one study included two HIV-positive women , and another only evaluated the long-term efficacy of IRC . Therefore, more data on the effectiveness of IRC are necessary before the real benefits of this treatment become clear. The aim of the present study was to assess the prevalence of AIN and the long-term effectiveness and safety of IRC in HIV-infected patients. We also investigated whether treatment results differed between HIV-infected men and HIV-infected women.
Patients and methods
The study was a single-center, retrospective cohort study based on data from a prospectively compiled database (electronic medical files) of outpatients who were attended in the Clinical Proctology Section of the HIV Unit of University Hospital Germans Trias i Pujol, Badalona, Spain. All patients were systematically screened for HPV-related anal disease. The study was performed according to the stipulations of the Declaration of Helsinki, and all patients gave their written informed consent for their medical information to be used for purposes of scientific research in accordance with the guidelines of the local ethics committee.
The patients included in this study had to fulfill the following criteria: age at least 18 years, positive HIV serology results, and biopsy-proven (histological diagnosis) AIN-2 or AIN-3 confirmed by high-resolution anoscopy (HRA).
The inclusion period ran from January 2005, when the Clinical Proctology HIV unit was created and the screening program implemented, through December 2011. The starting point of the study was January 2005. Following our standard operating procedure, all HIV outpatients who were attended in the HIV Unit for the first time from January 2005 onward were informed about the screening program for detection of HPV-related disease of the anal canal.
Patients whose HIV infection was already being managed before January 2005 were informed about this screening program by our reception staff when they came for their routine check-ups. They were also informed about the program by their HIV care providers. It is worth noting that all care providers for HIV-infected patients were reminded at staff meetings (every 2 months during the first 2 years) to inform patients about the program. Consistent with the usual procedure, patients were referred to the Clinical Proctology HIV Unit if their HIV doctor detected an anal condition.
The following data were gathered at baseline: date of birth, date of HIV diagnosis (time with HIV infection in years), date of the visit, CD4 cell count (the most recent value before cytology sample collection), nadir CD4 counts (the lowest CD4 value of each patient taken from medical records), plasma viral load (the most recent value before cytology sample collection), HAART before inclusion (yes/no), and time on HAART. The results of anal examinations were also recorded [anal cytology, HRA (number and localization of the lesions) and histology (biopsy) results], as were data on anal evaluations during follow-up. CD4 cell and nadir CD4 were counted using flow cytometry, and HIV viral load was determined using Nuclisens (detection limit 80 copies/ml; bioMérieux, Inc., Durham, North Carolina, USA).
Clinical examination and anal sample collection
A clinical examination (visual inspection) including a digital rectal examination was performed at the first clinical visit. A sample of tissue from the anal canal was obtained when deemed necessary by introducing a cytobrush (Eurogine SL, Sant Boi del Llobregat, Spain) 3 cm into the anal canal and softly rotating for 30–45 s. The cytobrush was introduced into 20 ml of PreservCyt/ThinPrep Pap test solution (Cytyc Iberia SL, Barcelona, Spain) and shaken for 30 s. This sample was used to carry out the cytology analysis (Papanicolaou test). The flow chart for detection of AIN is shown in Fig. 1.
High-resolution anoscopy and infrared coagulation
If the anal cytology result was positive [atypical squamous cells of unknown significance (ASCUS), low-grade squamous intraepithelial lesion (LSIL), or HSIL], the patient was informed and HRA was programmed. HRA was performed using a technique described elsewhere . If the HRA (after topical application of 3% acetic acid in the anal canal for 2 min) revealed a lesion, a biopsy was performed. If the result of the biopsy (histology) was AIN-2 or AIN-3, the patient was informed and treatment with IRC was scheduled.
The intra-anal lesions detected with HRA, namely, biopsy-proven AIN-2 or AIN-3, were treated with in-office IRC ablation (Redfield IRC 2100; Redfield Corporation, Rochelle Park, New Jersey, USA). Each lesion was identified, infiltrated with local anesthesia, and repeatedly coagulated with IRC in pulses of 1.5 s. IRC delivers short pulses of a narrow beam of visible infrared light through a small contact tip applicator that is applied directly to the target tissue. It results in thermal coagulation and tissue necrosis. The scab was removed, and the process was repeated until the submucosal vessels were coagulated. The depth of coagulation and tissue necrosis was adjustable. All anal lesions present were treated at the time of ablation. All procedures were performed by trained surgeons (M.P., F.G.-C. and S.Vela).
If anal canal condylomata were diagnosed during clinical examination, either by digital rectal examination or HRA, the patient was referred to the major ambulatory surgery unit for treatment.
After IRC ablation, patients underwent routine evaluations involving visual inspection, digital rectal examination, and anal canal cytology at 3-month to 6-month intervals. Postsurgical complications were also recorded. Patients with abnormal cytology results underwent another examination with HRA. If HRA revealed a lesion, a biopsy was performed. Patients with biopsy-proven AIN-2 or AIN-3 were advised to undergo further IRC ablation. Patients with AIN-1 were monitored with anal cytology at 3-month to 6-month intervals. Patients with normal cytology results 12 months after IRC were monitored using anal cytology at 12-month intervals. The screening and follow-up algorithm is shown in Fig. 1.
Anal cytological and histological assessment
Cytological changes were classified according to the Bethesda System as ASCUS and LSIL or HSIL. Generally, samples were independently assessed by two expert cytopathologists. Histological changes were classified according to the grade of AIN as AIN-1, AIN-2, and AIN-3. Samples were independently assessed by two expert pathologists.
Human papilloma virus detection and typing
DNA was extracted from cell suspensions (in ThinPrep Pap solution) using the QIAamp Viral DNA kit (QIAGEN, Hilden, Germany). HPV was detected and typed in all samples using a commercial in vitro diagnostic CE-marked assay [Multiplex Fluorescent-PCR Kit for HPV Genotyping (F-HPV) typing, Molgentix SL, Barcelona, Spain] in accordance with the manufacturer's instructions . The kit enables the detection of 13 high-risk HPV (HR-HPV) genotypes (16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68) and the two most frequent low-risk HPV genotypes (6, 11). A human short tandem repeat sequence included in the same multiplex reaction was amplified as an internal control for DNA integrity and the absence of polymerase chain reaction (PCR) inhibitors. Products were analyzed using capillary electrophoresis on an ABI 3130 XL genetic analyzer and GeneMapper 4.0 Software (Applied Biosystems). Each PCR run included HPV-positive and HPV-negative controls. Particular care was taken to prevent carry-over contamination by separating pre-PCR and post-PCR areas in the laboratory.
Treatment success (percentage of responders) was defined as no evidence of cytology-proven SIL and no evidence of anal lesions on HRA or no evidence of anal lesions on clinical examination (visual inspection and digital rectal examination) if a follow-up HRA was not performed at 12 months after IRC. Overall treatment success was defined as no evidence of cytology-proven SIL and no evidence of anal lesions on HRA or no evidence of anal lesions on clinical examination if a follow-up HRA was not performed at any point during follow-up. Recurrence was defined as cytology-proven SIL or biopsy-proven AIN at the treated site or at a new site at 12 months after treatment. Overall recurrence was defined as identification of cytology-proven SIL or biopsy-proven AIN at any point during follow-up at the treated location or at a new site. Persistence was defined as the existence or recurrence of disease at the treatment site. A metachronous recurrence was defined as biopsy-proven AIN found at a previously untreated site.
No formal sample size was calculated. The sample was defined as all HIV-infected patients with biopsy-proven AIN-2 or AIN-3 (histological diagnosis) confirmed by HRA.
Baseline characteristics were summarized using standard descriptive statistics, and a descriptive analysis was carried out. The prevalence of anal canal SIL, AIN, and anal HPV-type specific infection was estimated, and 95% confidence intervals (95% CI) were calculated. The effectiveness of IRC (treatment success) was expressed as a percentage, and overall treatment success was calculated. The mean time to recurrence (range) was analyzed using the Kaplan–Meier method. Cohen's κ was used to assess agreement between the cytological diagnosis (Papanicolaou test) and histological diagnosis (biopsy). Data were analyzed using SPSS version 15.0 statistical software (SPSS, Inc., Chicago, Illinois, USA) and StatXact-8 (Cytel Inc., Cambridge, Massachusetts, USA).
We studied 69 HIV-positive patients (55 men and 14 women) with a diagnosis of AIN-2 or AIN-3 confirmed by HRA and biopsy who underwent treatment with IRC. Figure 1 shows the progress of patients through the study, from the screened population (patients at their first visit to the Clinical Proctology HIV Unit) to the study population (patients diagnosed with AIN-2 or AIN-3).
Of the 1776 patients initially screened, a cytological diagnosis was available for 1518 HIV-positive patients (1310 men and 208 women). The number of patients screened represents over 70% of the HIV-positive population attending our HIV unit. The baseline characteristics of the study population are shown in Table 1. Among HIV-positive men (n = 55), 51 (93%) were MSM and four (7%) were heterosexual. No patients were vaccinated against HPV. During follow-up, all patients were on HAART, which included at least two nucleoside reverse transcriptase inhibitors in combination with either a protease inhibitor or a nonnucleoside reverse transcriptase inhibitor.
Prevalence of anal disease
Based on the screened population and number of patients with a cytological diagnosis, the prevalence of SIL (low and high grade) was 49.5% (751/1518, 95% CI: 47–52%) (Fig. 1). Based on the histological diagnosis, the prevalence of AIN (AIN-1, AIN-2, and AIN-3) was 15% (223/1518, 95% CI: 13%, 17%) (Fig. 1).
The agreement between the cytological diagnosis (Papanicolaou test) and the histological diagnosis (biopsy) was evaluated in 288 patients. Table 2 shows the agreement between the anal cytological and histological results in patients with an abnormal anal cytology result and a lesion visualized using HRA. The agreement between cytological and histological diagnosis was poor [k = 0.23 (95% CI: 0.13,0.32), P < 0.001, including ASCUS in the analysis; k = 0.13 (95% CI: −0.01,0.26), P = 0.066, not including ASCUS in the analysis]. Sixty-one percent of patients with a cytological diagnosis of LSIL had a diagnosis of AIN-1, and only 50% of patients with a cytological diagnosis of HSIL had a diagnosis of AIN-2 or AIN-3. Furthermore, 60% of patients with a cytological diagnosis of ASCUS had a histological diagnosis of AIN.
Effectiveness of infrared coagulation
A total of 69 HIV-infected patients with a diagnosis of biopsy-proven AIN-2 or AIN-3 were candidates for treatment with IRC: 68 (99%) patients presented only one confirmed lesion and only one patient (woman) presented two confirmed lesions (total number of lesions treated, 70). Lesions were mainly located at 7 o’clock (12 patients), 11 o’clock (11 patients), 6 o’clock (10 patients), 9 o’clock (8 patients), and 12 o’clock (7 patients). In the patient with two lesions, both were at 3 o’clock.
With regard to sex, the percentage of HIV-positive women candidates for treatment with IRC (7%, 14/208) was slightly higher than that of HIV-positive men [4% (55/1310), P = 0.103], MSM [5% (51/996)], and heterosexual men [1% (4/314)].
Three patients (4%) rejected treatment with IRC and 10 (14%) had a follow-up of less than 12 months. Therefore, the effectiveness of IRC was evaluated in 56 patients (45 men, 11 women). No evidence of cytology-proven SIL or anal lesions was detected at 12 months after IRC. Therefore, treatment was successful in all cases. The 10 patients with a follow-up of less than 12 months did not have evidence of cytology-proven SIL or anal lesions after IRC.
The overall success of treatment was 87.5% (49/56), with a mean follow-up of 25 (range: 12–60) months. After the 12 months of post-IRC follow-up, only seven patients [12.5%: two women (18%), five men (11%)] had abnormal cytology findings, and recurrent disease (AIN) was biopsy-proven in all of them (overall recurrence). The mean time to recurrence (range) was 30 (18–43) months. The recurrence of disease was at the treatment site (persistence) in one patient (14%: AIN-1) and at a previously untreated site (metachronous recurrence) in the remaining six patients (86%: AIN-1, three patients; AIN-2, three patients).
Human papilloma virus infection
Data on HPV infection were available for 58 out of 69 patients with AIN-2 or AIN-3; all 58 patients presented detectable infection by HPV. Single HPV infection was found in only seven patients (12%) and multiple infection (more than one HPV) in 51 (88%). The most prevalent HR-HPV types detected at baseline were HPV-16 (65.5%), HPV-58 (41.4%), HPV-33 (34.5%), HPV-51 (27.6%), and HPV-18 (25.9%). The prevalence of HPV-6 and HPV-11 (HPV related to nonmalignant diseases) was 20.7 and 3.4%, respectively. Figure 2 shows the genotype-specific prevalence of HPV in the anal canal of patients with AIN-2 (n = 37) and AIN-3 (n = 21). In the group of patients with AIN-2, infection by a single genotype was found in six patients (16%) and by multiple genotypes in 31 patients (84%). However, in the group of patients with AIN-3, only one (5%) patient presented infection by a single genotype; the remaining 20 patients (95%) presented infection by multiple genotypes. Data on HPV information were only available in one out of seven patients with a recurrent lesion. The same HR-HPVs (HPV-33, HPV-58) seen at baseline were detected in the recurrence.
No patients developed serious adverse events (anal stricture, persistent bleeding, significant postoperative hemorrhage, failure to heal, or infection requiring antibiotic therapy) following IRC ablation. Three patients presented mild bleeding during the first few hours after IRC. Pain was the most common post-IRC event until 72 h after the IRC [nine patients (13%) scored greater than 3 in the verbal numerical rating pain intensity scale] and was adequately controlled with mild analgesics. Of note, no patient developed anal SCC during study follow-up.
Few data are available on the efficacy or effectiveness of IRC [7,13–16] in high-grade anal dysplasia. Studies have generally focused on HIV-positive MSM, and findings for HIV-infected women are practically nonexistent. The high long-term effectiveness of IRC found in our sample, which included HIV-infected men and women without concomitant condylomata in the anal canal, suggests that IRC might help to prevent anal SCC and progression to invasive disease.
Interrupting the natural history of SCC is challenging. In the last decade, several approaches have been used to manage AIN (cryotherapy, electrocautery, topical agents, IRC), although solid data are scant and randomized clinical trials are lacking [6,8–15]. High rates of persistence and recurrence are reported with most approaches, including IRC. However, our results revealed a low rate of recurrence, possibly because of the characteristics of our population: our patients had only one lesion per person (single lesion without condylomata) and a good immunological status and were receiving HAART during follow-up. Nevertheless, it seems that HAART has not been able to curb rising incidence . Furthermore, metachronous lesions (i.e. at previously untreated sites) have been considered a major cause of recurrence (82%) in HIV-positive patients after their first IRC treatment . Although the low number of patients with recurrent lesions in our study prevents us from drawing firm conclusions, metachronous lesions recurred in six (86%) out of seven patients. Given the lack of clinical trials and the high rate of persistence or recurrence in HIV-positive patients and the high rate of recurrence due to metachronous lesions, continued surveillance after IRC is necessary until more data become available.
No patient treated with IRC progressed to anal SCC during follow-up. However, three MSM from the screened population were diagnosed with in situ anal SCC and treated with surgery. One of them had a recurrence (invasive anal SCC) 18 months after surgery. Among the nonscreened population (∼30% of our current population of HIV-positive patients), one case of invasive anal cancer (female IDU in 2006) was diagnosed during the same time period. Furthermore, two more cases of invasive anal cancer (one heterosexual male IDU in 2010 and one transsexual female IDU in 2012) were referred for treatment to our center, which is a referral hospital in our geographic area. From 1983 (the year the hospital opened) to December 2004, six cases of invasive anal SCC in HIV-positive patients were diagnosed and treated [one man in 1999 (no information on sexual practices); three MSM in 1999, 2000, and 2002; and two heterosexual male IDUs in 2004]. The first of these aforementioned 10 cases was diagnosed and treated in 1999 and the last one in 2012, that is, during the HAART era.
Despite the poor agreement between cytological and histological diagnosis of anal dysplasia when different grades of dysplasia are matched, the high prevalence of anal dysplasia (SIL or AIN) alerts us to the high probability of developing anal SCC in HIV-infected patients. Consequently, it is worth considering that around 10% of screened HIV-infected patients presented biopsy-proven AIN-1, and around 5% presented biopsy-proven AIN-2 or AIN-3. In other words, one out of every 20 HIV-infected patients (∼1 of every 15 women and ∼1 of every 24 men) needed treatment with IRC. These findings imply that physicians should be aware of anal SCC in their approach to HIV-infected patients. HIV-infected MSM are at increased risk of developing anal SCC with an incidence rate of ∼131 per 100 000 person-years . However, AIN was also diagnosed among heterosexual HIV-infected men and HIV-infected women. Both HIV-positive heterosexual men (∼46 per 100 000 person-years) and HIV-positive women (∼30 per 100 000 person-years) also have a clinically relevant incidence rate when they are compared with the general population (1 per 100 000 person-years) [19–21]. Faced with these incidence rates, sexual practices predispose to increased risk but should not be considered the only factor. It is likely that HIV infection itself plays a major role in the natural history of HPV-related diseases.
The poor correlation between cytology and histology findings when the different grades of dysplasia are matched is very concerning, since a number of patients with normal anal cytology results could have had anal intraepithelial disease. We were unable to establish this figure in our study because patients with normal cytology did not undergo HRA. However, our results show that 59% (47/59) of patients with a cytological diagnosis of ASCUS had a histological diagnosis of AIN. Furthermore, 19% (56/288) of patients with abnormal cytology findings (ASCUS, LSIL, or HSIL) had a normal histology result. These findings reinforce the possibility that a number of patients with a normal anal cytology result might have AIN. In fact, 86% (59/69) of patients with a cytological diagnosis of HSIL had a histological diagnosis of AIN. Likewise, 81% (114/141) of patients with a cytological diagnosis of LSIL had a histological diagnosis of AIN. If we compare normal and abnormal cytology results (independently of the grade of dysplasia) with normal and abnormal histology results, agreement between both diagnostic techniques is 81% (232/288). Then, if we take into account that an abnormal cytology result has a high probability of being AIN and that disease progression is slow, the cytological findings could prove sufficiently useful for clinical management. Consequently, the yearly anal cytological assessment could prove to be extremely important, despite the poor agreement between cytology and histology results when the different grades of dysplasia are matched.
HPV infection caused by high-risk genotypes was present in all patients with AIN-2 and AIN-3. Likewise, practically all patients (88%) presented HPV infection caused by more than one genotype. As expected, HPV-16 was the most prevalent. Others, such as HPV-58, HPV-33, and HPV-51, were more prevalent than HPV-18. Low-risk genotypes (HPV-6 and HPV-11) were also detected which are associated with HPV-related nonmalignant diseases (condylomata), although it is noteworthy that patients with condylomata were not included in this study. Given the low number of patients with recurrent lesions, no association can be established between HPV infection and recurrence, even though high-risk HPV-infection was detected in all cases of recurrence of AIN.
Our study is subject to a series of limitations. Its retrospective cohort design, although based on data from a prospectively compiled database (electronic medical files) might underestimate or overestimate the generalizability of the results beyond the population and conditions studied. Likewise, the sample size might underestimate or overestimate the generalizability of the results. HRA (after topical application of 3% acetic acid) is currently the best procedure for the detection of anal canal lesions; however, as it is subject to interobserver bias (despite thorough training), primary lesions or recurrent lesions could go undiagnosed. In addition, during follow-up, patients with benign lesions according to cytology and negative standard anoscopy results were considered nonrecurrent, thus potentially leading us to further underestimate recurrence rates. Moreover, as the study was only performed in patients with SIL and no evidence of condylomata, its applicability to patients with SIL and condylomata could be limited. Finally, adverse events after IRC were recorded during an interview held at subsequent visits; consequently, patients may have underestimated the number and intensity of the events.
In conclusion, AIN was highly prevalent among HIV-infected men and women. Although randomized clinical trials are lacking, IRC ablation of AIN-2 and AIN-3 lesions without concomitant condylomata in the anal canal could help prevent anal SCC.
We are grateful to Thomas O’Boyle for editorial assistance. We thank Eric Goode for the English correction of this article. We are also indebted to the patients of our Clinical Proctology HIV Unit.
Author contributions: S.V., G.S., J.C. B.C. designed and wrote the study protocol. G.S., J.C., F.G.-C., M.P., S.V. and P.C. visited and treated the patients and collected the anal samples. M.P.C. and L.D. performed HPV detection and genotyping using multiplex PCR. S.G. and N.P. were responsible for data management and statistical analysis. S.V., G.S., J.C. and B.C. wrote the article. All the authors read and approved the final version of the article.
Funding: This work was supported by grants from Red de Investigación en SIDA (RIS), ART AIDS Foundation, Gilead Sciences, Obra Social Caixa Sabadell, and Gala contra la SIDA-Barcelona 2011.
HIV-HPV Study Group: University Hospital Germans Trias i Pujol, Badalona (Barcelona), Autonomous University of Barcelona: Department of Pathology: Dr E. Castellà, Dr M. Llatjós
HIV Clinical Unit and Internal Medicine Department: Dr A. Bonjoch, Dr P. Echevarría, Dr A. Jou, Dr JM. Llibre, Dr J. Moltó, Dr E. Negredo, Dr C. Rey-Joly, Dr J. Romeu, Dr JR. Santos, Dr C. Tural
Nurses of HIV Clinical Unit: Ms. C. Alcalde, Ms. R. Guerola, Ms. A. Salas
Conflicts of interest
S.V. has received honoraria for collaborating with Laboratorios Dr Esteve on work unrelated to HPV/HIV. B.C. has received honoraria for speaking and participating on advisory boards from Abbott, Bristol-Myers Squibb, Boehringer-Ingelheim, Gilead Sciences, GlaxoSmithKline, Pfizer, Merck, Janssen-Tibotec, and Siemens. The remaining authors have no conflicts of interest.