Article Text
PDF
Abstract
Objective Anal cancer is a common cancer among men who have sex with men (MSM); however, there is no standard screening protocol for anal cancer. We conducted a phase II clinical trial to assess the feasibility of teaching MSM to recognise palpable masses in the anal canal which is a common sign of anal cancer in men.
Methods A clinician skilled in performing digital anorectal examinations (DARE) used a pelvic manikin to train 200 MSM, aged 27–78 years, how to do a self-anal examination (SAE) for singles or a partner anal examination (PAE) for couples. The clinician then performed a DARE without immediately disclosing results, after which the man or couple performed an SAE or PAE, respectively. Percentage agreement with the clinician DARE in addition to sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were calculated for the SAE, PAE and overall.
Results Men had a median age of 52 years, 42.5% were African American and 60.5% were HIV positive. DARE detected abnormalities in 12 men while the men’s SAE/PAEs detected 9 of these. A total of 93.0% of men classified the health of their anal canal correctly (95% CI 89.5 to 96.5). Overall percentage agreement, sensitivity and specificity were 93.0%, 75.0% and 94.2%, respectively, while PPV and NPV were 45.0% and 98.3%, respectively. The six men who detected the abnormality had nodules/masses ≥3 mm in size. More than half of men (60.5%) reported never checking their anus for an abnormality; however, after performing an SAE/PAE, 93.0% said they would repeat it in the future.
Conclusion These results suggest that tumours of ≥3 mm may be detectable by self or partner palpation among MSM and encourage further investigation given literature suggesting a high cure rate for anal cancer tumours ≤10 mm.
- digital rectal examination
- HIV
- anal canal
- anus neoplasms
- homosexuality
- male
- early detection of cancer
Statistics from Altmetric.com
The incidence rate of anal cancer among males steadily increased from 1973 to 2006 in the USA with an average annual percentage increase of 3.3%.1 While there is no proven screening for anal cancer, various biomarkers and high-resolution anoscopy are currently being investigated to screen for precancerous anal lesions2; however, there are no efficacy data for these methods, infrastructure is poor and there is no efficacious treatment for precancerous lesions if they are found.3 Because anal cancer is a common cancer among HIV-positive gay, bisexual and other men who have sex with men (MSM),4 and because most squamous cell anal carcinomas are palpable,5 annual digital anorectal examination (DARE) is standard of care with this population.6 In addition, consensus expert opinion recommends DARE for detection of tumours among HIV-negative MSM.7 8 Even so, DARE is severely underused. For example, 77% of persons with HIV in the USA did not get an annual examination in 2011.9 Undiscovered tumours through lack of DARE and delayed DARE may substantially add to anal cancer morbidity, mortality and cost since these outcomes are highly related to stage of diagnosis.10 For example, for anal cancer tumours ≤1 cm, cure rates have been reported at >90% while tumours >5 cm have a much poorer prognosis.11 12
Given urgent calls to screen more MSM for anal cancer,13 we investigated if MSM can learn to examine their own anal canal or if partners can examine each other. If so, then detection and treatment of smaller malignant tumours among MSM may increase for those who do not or cannot access screening for precancers, due to, for example, lack of infrastructure, lack of efficacy or embarrassment.3 14 15 We posited that teaching MSM to do their own self-anal examinations (SAEs) or, within couples, a partner anal examination (PAE) was feasible for several reasons: most anal canal malignancies in men are palpable5 16; the full length of the anal canal (3–5 cm) is shorter than the male index finger (~7–9 cm in length) which makes the entire surface of the canal accessible to palpation17 18; lay persons have been trained to teach DARE to clinicians using didactic instruction and pelvic models19; lay persons already examine the anus for disease20; and self-anal swabbing for STD screening has been successful21 and acceptable among MSM.22 In addition, a recent qualitative study reported that HIV-positive MSM would find SAEs acceptable.23
Our objective was to evaluate the feasibility of the SAE and PAE by assessing their accuracy and acceptability. We also assessed factors associated with concordance of SAE/PAEs and a clinician’s DARE.
Methods
Recruitment and protocol
We conducted a phase II clinical feasibility study among MSM in Harris County, Texas, USA of SAE and PAE. Inclusion criteria targeted men, aged 27–80 years, who had no current doctor’s diagnosis of anal condylomas, haemorrhoids, fissures or anal cancer, and who could speak English. Men were excluded if they reported a DARE in the prior 3 months. We recruited persons 27 years and older since the human papillomavirus vaccination catch-up programme in the USA ends at 26 years and because of low cancer incidence in younger men. Recruitment occurred in 2015–2016 and was stratified to enrol individuals and partners in couples who reported sex with men in their lifetime. Recruitment used flyers posted at three clinics (flyer text said the study was seeking ways to prevent anal cancer) and face-to-face recruitment by staff at the clinics. Recruitment also targeted the general MSM community using social media (eg, Facebook and Growlr), flyers posted in gay venues like bars, publicity at mass events and print media advertisements. Interested men called a study co-ordinator who made a preliminary eligibility assessment. The study protocol was approved (HSC-SPH-13–0671) by the Committee for the Protection of Human Subjects at the University of Texas Health Science Centre, and a written and informed consent was obtained from all participants.
All participants attended a single clinic visit for SAE/PAE training, a DARE performed by a clinician, an SAE/PAE performed by the participant and survey completion. The clinician was a nurse practitioner (NP) with 3.5 years of NP experience focused on HIV care including completion of more than 2000 DAREs. After consenting, the men completed an initial presurvey followed by brief anal cancer and SAE/PAE training provided by the clinician. The training included practice using two digital rectal examination manikins (Kyoto Kagaku, Kyoto, Japan) to contrast a normal anal canal and one with a 7 mm tumour. Verbal and written SAE/PAE instructions (Supplement) were provided. The clinician emphasised that the goal was to detect any palpable abnormality regardless of type (eg, wart, haemorrhoid, fissure or tumour); thus, we assessed SAE and PAE for multiple palpable anal conditions with the logic that if a mass or induration, regardless of aetiology, can be palpated, then malignant tumours may be palpated too. Prostate and distal rectal palpation was not taught.17 Finally, the clinician performed a DARE on the participant, recording a result of normal or abnormal, but did not communicate results to the participant at that time. For couples, both partners remained in the room during the DARE. For both individuals and partners, the clinician spoke to the men as she performed the DARE to describe what she was doing, for example ‘I’m now feeling 360° around your anal canal’.
The clinician then left the men in private to conduct the examination on themselves, or their partners. Single men were encouraged to view the perianal region with a mirror or in a selfie taken with a smartphone. Several body positions for both individuals and partners were suggested (eg, putting one leg on a chair, squatting or bending over a table) with participants encouraged to use whatever position was most comfortable for them (Supplement). Men in couples were encouraged to ask for permission from their partner before beginning the examination. Each participant recorded his result as normal or abnormal, which allowed comparison of the clinician DARE and participant SAE/PAE results.
Men then completed an extensive computer-assisted self-interview after which participants met with the clinician to receive DARE results. Men with abnormalities were scheduled for treatment or provided referrals. Procedure safety was assessed with a general question about any pain during the visit. This question was expanded during the study so that a subset of the final 54 men doing study procedures were asked more focused questions about pain. If the participant reported an abnormality not observed by the clinician, the clinician offered a second DARE to check the initial DARE results. A total of four focus groups were also conducted with the participants.24
To check the quality of the clinician’s DARE, an HIV oncologist (EYC) observed the initial clinician DAREs and a total of 10% of the initial recruits into the study were also examined by a board-certified colorectal surgeon (SM) using DARE and standard anoscopy generally within 2 weeks after the clinician’s DARE (high-resolution anoscopy was not performed since it detects precancerous lesions that are putatively non-palpable). In 19 of 20 cases, the DARE result of the surgeon and clinician were the same. Participants were compensated $40 for the clinic visit and $50 for optional standard anoscopy.
Statistical methods
Tests of proportion were used to assess differences in characteristics between individuals and partners. The primary outcomes were the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of SAE and PAE in addition to percentage agreement between participant SAE/PAE and the reference standard of the clinician DARE. Clopper-Pearson confidence limits were calculated for each proportion. The sensitivity of DARE has been estimated at 90%.25 For couples, each partner performed digital insertion on the other partner; thus, partners provided two participant results while the clinician produced two DARE results.
Concordance was used as the outcome for assessment of factors associated with participant and clinician examination agreement. For individuals, concordance was coded as one if the result of both clinician and participant was the same after performing a DARE and SAE, respectively. Concordance was coded as zero if clinician and participant disagreed. For partners, concordance was coded as one if both clinician and the partner performing digital insertion agreed on the results for the other partner and zero if the performing partner and clinician disagreed. For analysis of factors associated with concordance, prevalence ratios (PR), adjusted PR and 95% CIs were calculated by bivariate and multivariable Poisson regression using a robust sandwich estimator of the variance.26 27 Factors chosen for multivariable analysis were included in modelling if they were not a variable in the causal pathway between exposure of interest and accuracy, and if they were significant in bivariate analysis at p<0.25. Age was included in all modelling as a potential confounder. Waist circumference was categorised as ≤102 cm and >102 cm in accordance with prior literature identifying increased health-related problems among persons with >102 cm waist size.28
A priori calculations indicated that 200 participants provided 97% power to test the primary hypotheses that SAE and PAE would have moderate or substantial agreement with a clinician’s DARE and that PAEs would have better agreement with the clinician DARE than would an SAE (assuming 30% prevalence of abnormalities, α=0.05, with two-sided tests).
Results
A total of 386 men expressed interest in the study through phone calls and email between February 2015 and March 2016. Of these, 27 (7.0%) declined participation and 105 (27.2%) men were ineligible or non-responsive. Of 254 men scheduling appointments, 54 (21.3%) did not show, were ineligible at the time of appointment or did not provide an outcome result leaving 200 men for analysis (figure 1) including 148 individuals and 52 partners. A plurality of these men (45%) were recruited through a clinic or community centre, 25% through friend referral, 19% through social apps and 11% through flyers (data not shown).
Recruitment and completed appointments for men conducting self and partner anal examinations, Houston, 2015–2016.
The median age was 52 years with 46.5% and 42.5% of participants Caucasian or African American, respectively. A total of 17.5% of men were Latino. Just under two-thirds of men were overweight/obese and 60.5% reported HIV infection (table 1).
Characteristics of men conducting SAE and PAE, Houston, 2015–2016
Most men (88.5%) agreed with the clinician’s observation of no abnormality (true negatives). The clinician palpated an abnormality in 6.0% of men (n=12) with nine of these men also reporting an abnormal palpation by SAE/PAE (true positives) (table 1). The clinician described palpated abnormalities as likely condylomas (n=4), scar tissue (n=3) and one each papule, nodule, haemorrhoid, skin tag and flat mass. The nine abnormalities palpated by participants were ≥3 mm in diameter according to the clinician. The three false negatives were characterised by condylomas and papules ≤3 mm in diameter (data not shown). A total of 11 men reported an abnormality not observed by the clinician and 8 of these agreed to a second confirmatory DARE during the same clinic visit. No abnormality was observed on the second DARE. In six of these eight men, the clinician palpated stool.
Sensitivity for the combined SAE/PAE was 75.0% (95% CI 42.8% to 94.5%), while specificity was 94.2% (95% CI 89.8% to 97.0%) (figure 2). The NPV of combined SAE/PAE was 98.3% (95% CI 95.2% to 99.7%), while PPV was substantially lower (45.0%, 95% CI 23.1% to 68.5%).
Accuracy estimates for all examinations, self-anal examinations (SAEs) and partner anal examinations (PAEs) with 95% CIs, Houston, 2015–2016. NPV, negative predictive value; PPV, positive predictive value.
There were few differences in characteristics and behaviours between individuals and partners. Partners used somewhat different body positions during the PAE than individuals’ SAE (table 1). Two-thirds of individuals said they used only one hand to perform the procedure. More than one-third of individuals (37.8%) and partners (40.4%) said they had used their fingers before to check their anus for disease. After being taught how to perform an SAE or PAE, 93.0% of men said they planned to do another in the future. After completing the examination, five men said they were not willing to do the SAE/PAE again due to lack of concern about anal cancer, dislike of the examination, forgetting how to do the examination, embarrassment, dislike of potentially toughing faeces and/or difficulty reaching the anus. Most men (92.5%) reported they would see a doctor if they detected an anus abnormality after doing an SAE/PAE. Similar proportions of men preferred the SAE/PAE and a doctor examination.
At the end of the clinic visit, the clinician asked the men ‘if the procedure hurt’ without reference to any particular procedure during the visit, for example, DARE or SAE/PAE. No men reported bleeding although 15.0% reported that a procedure hurt. During the study, this question was expanded so that the final 54 participants were asked more detailed questions about pain. A total of 4/54 (7.4%) of these men reported any SAE/PAE pain with an average score for these four men of 1.75 on a scale of 0–5, where 0 indicates no pain. All four said the pain would not make them avoid the procedure in the future (data not shown).
Among individuals, concordance between clinician result and participant was 91.2% (95% CI 86.7% to 95.8%) (figure 2). In multivariable analysis, individuals with a history of using fingers to check the anus for disease had increased concordance with clinician (adjusted PR 1.10, 95% CI 1.01 to 1.19, compared with persons who had never used their fingers to check their anus for disease) (table 2). In addition, individuals who said they worried about anal cancer a lot were more likely to be concordant with the clinician compared with men who said they did not worry about anal cancer (PR, 1.09, 95% CI 1.01 to 1.18). Among partners, concordance was 98.1% (95% CI 94.3% to 100.0%) which provided too little variance to assess factors associated with concordance.
Bivariate and multivariable factors associated with concordance between DARE and SAE among individuals (n=148), Houston, 2015–2016
Discussion
To our knowledge, this is the first report of an effort to teach MSM to recognise an abnormality at the anus. A large majority of men doing an SAE/PAE were concordant with the clinician; however, a low number of clinician-observed abnormalities limited the study’s ability to derive more robust estimates of sensitivity and PPV.
Nevertheless, of 12 abnormalities identified by clinician, 9 were also identified by participant. Most men (88.5%) matched the clinician result of no anal canal abnormality. While false negatives (1.5%) and false positives (5.5%) were rare, they could lead to missed disease, anxiety and unnecessary clinic visits for procedures like DARE. In most false-positive men, the clinician palpated stool which prompted changes to our educational protocol to differentiate the feeling between an anal canal-affixed mass and stool.
More than one-third of men reported checking their anus for disease in the past indicating their concern about anal disease and attempts to screen for disease, putatively without instruction on how to do an anal examination. Since lack of knowledge about anal cancer is associated with substantial anxiety after an abnormal screening result,29 increased anal cancer educational efforts are needed for MSM in addition to education about self-palpation if subsequent studies confirm a high proportion of MSM already self-palpate for disease.
However, the best way to conduct an SAE may not be fully understood. Two-thirds of individuals said they used only one hand to perform the procedure; nevertheless, a large majority of men reported palpating 360° of the anal canal. It may be possible that an index finger and opposable thumb allow complete self-palpation of the anal canal. It is also possible that individuals were inaccurate in their claim of full palpation since self-palpation is likely more difficult than partner palpation. Nevertheless, the SAE/PAE may be well tolerated since the large majority reported no SAE/PAE pain.
Prior self-palpation out of concern for disease was associated with concordance with clinician results; thus, repeat palpation may increase men’s familiarity with their anus, establish a baseline and increase recognition of subtle changes in this soft and smooth mucosal tissue. Given a report of a 100% disease-specific cure rate for anal cancer tumours ≤10 mm,11 self-recognition of small tumours (eg, 5–10 mm) may reduce anal cancer morbidity and mortality.
Testicular self-examinations are not currently recommended, in part, because scrotal tumours are often found accidentally.30 The relative inaccessibility of the anal canal suggests accidental findings of small tumours may be less likely. For example, 77.5% of this study’s participants reported that during bathing they either do not insert a finger into their anal canal or do not insert past the first knuckle (data not shown) and therefore would fail to reach the full length of the anal canal.
About one-half of men preferred the SAE or PAE rather than a doctor-performed examination which may speak to embarrassment associated with DARE31 and stigma regarding the anus15; thus, SAE/PAEs may increase anal cancer screening access by providing an option to detect an abnormality that motivates a clinic visit.
Future studies need a larger sample to detect more abnormalities and involve multiple clinicians to assess replicability. The impact and predictors of false-positive results, for example, for benign results like haemorrhoids, need more study given potential anxiety and economic implications.
The study assessed the men’s ability to palpate an anal canal abnormality, not a malignant tumour; however, a properly powered study to palpate malignant tumours would require >10 000 participants even in this high-risk population. Also, it may be of value if MSM can palpate other pathology like anal condylomas since they increase risk for high-grade squamous intraepithelial lesions.32 If found effective, these examinations could be used in low-resource settings that lack the option of high-resolution anoscopy and to overcome barriers to clinic DAREs, for example, due to stigma.15
Given high anal cancer incidence among MSM, no uniform screening standard for anal cancer and lack of infrastructure for proposed screening modalities (in both high and low-resource settings),14 these results suggest the SAE/PAE be further explored for detecting early-stage anal cancer when treatment is more successful.11 12 Larger studies are needed to assess harms and benefits of SAE/PAE and to provide robust estimates of sensitivity and PPV before either examination could be incorporated into an anal cancer screening programme. If SAE/PAE benefits outweigh harms, translational and cost-effectiveness studies can answer questions about SAE/PAE frequency and optimal target populations which may be defined by HIV status, sexual behaviour, local resources and patient preference. Nevertheless, it could be a valuable community-led tool for both raising awareness and for screening for anal cancer.
Key messages
Self-anal examinations (SAE) and partner anal examinations (PAE) taught by a clinician might support reductions in anal cancer morbidity and mortality.
In this feasibility study, a large majority of men doing an SAE or PAE were concordant with the clinician’s digital anorectal examination.
These examinations may address screening barriers associated with stigma regarding the anus and embarrassment with having a physician perform digital anorectal examinations.
Supplementary file
![[SP1.jpg]](https://sti.bmj.com/content/sextrans/94/2/124/DC1/embed/inline-supplementary-material-1.jpg?download=true){kind=link}
Acknowledgments
We thank the men who volunteered for the study. In addition, thanks to Oluwatobi Adegboyega, Jie Deng, Mohan Yedhoti, Shadi Al Sayyed, Anshu Khanna, Wen Li, Mihyun Chang, Anand Narayanan, Dr Gordon Crofoot, Charles Sydnor, Paul Simmons, Fred Reninger, Jeffrey Benavides, Dr Anna Giuliano, Dr David Shibata, Dr Pratik Doshi, Dr Christopher Fairley, Dr Jason Farr, Hologic and Roche.
References
Footnotes
Handling editor Joseph D Tucker
Contributors AGN, L-YH, MWR and EYC conceptualised the study and formulated the hypotheses. AGN, JTH, NO, MW and SM conducted the investigation. AGN, XZ and SB did the statistical analysis. XZ provided data curation. AGN was responsible for funding acquisition, determining methodology and writing the paper. JTH, L-YH, SB, MW, SM, NO, XZ, ELB, MWR and EYC critically reviewed and edited subsequent drafts.
Funding This work was supported by the National Cancer Institute at the National Institutes of Health [grant number NCI 1R21CA181901-02) to AGN; ClinicalTrials.gov number, NCI-2015-00248]. Publication and report contents are solely the responsibility of the authors and do not necessarily represent the official views of the NCI.
Competing interests The institutions of AGN, JTH, and EYC received NIH grants or subcontracts for the submitted works and/or grants from the NIH outside of the submitted work.
Ethics approval University of Texas Health Sciences Center Committee for the Protection of Human Subjects.
Provenance and peer review Not commissioned; externally peer reviewed.