Sexual transmission is currently the main route of spread of HIV type 1 (HIV-1) . Anal intercourse, principally among MSM, constitutes the primary mode of HIV-1 acquisition in high-income countries [2,3]. Although receptive anal sex carries the highest per-act risk of HIV-1 transmission, insertive anal intercourse is responsible for a substantial number of new infections among MSM .
Combination antiretroviral therapy (ART) reduces significantly the risk of HIV-1 transmission. Several studies have shown the beneficial effects of ART in the prevention of HIV-1 infection in serodiscordant heterosexual couples [5,6]. However, the protective effect of ART has not yet been definitively established in MSM .
Although ART allows the suppression of blood plasma HIV-1 viral load (bpVL), circulating virus differs from virus in genital or rectal reservoirs, where local factors and ART penetration may also determine HIV-1 RNA levels . Moreover, the rectal reservoir has unique characteristics compared with male or female genital tract as the gut-associated lymphoid tissue (GALT) contains most of the total CD4+ T-cell pool, representing the largest HIV-1 reservoir that is capable of producing infectious viral particles whenever activated by the appropriate antigen [9,10]. In addition to HIV, the rectal mucosa of MSM is also a reservoir for multiple bacterial sexually transmitted infections (STIs), and most of them are also coinfected with human papillomavirus (HPV), that may additionally induce the development of anal intraepithelial neoplasia (AIN). The locally associated immune responses to such stimuli might promote HIV-1 replication [11–13].
Although the anal–rectal secretions constitute an important potential source of HIV transmission [10,14], to date, very limited and discordant information exists about anal–rectal HIV-1 shedding in PLWH. Data come from studies where the number of virologically suppressed participants receiving ART was low and/or were carried out in the earliest ART period [14–16]. In addition, the influence of the interaction between ART, STIs, HPV and the HPV-associated anal mucosal changes on HIV-1 anal–rectal shedding has not been characterized.
We aimed to describe the prevalence of persistent HIV-1 RNA shedding in the rectal secretions of stable, ART-suppressed HIV-1 positive MSM and to characterize the factors associated with viral shedding, including the impact of STIs and anal dysplasia, and their interaction with ART, within this population.
Study design and participants
A prospective study was conducted at the HIV cohort from the Hospital General Universitario de Elche (Spain) between November 2013 and February 2017. Eligible participants were MSM at least 18 years old under stable ART and with bpVL less than 50 copies/ml for greater than 6 months. The study included all persons routinely scheduled for high-resolution anoscopy (HRA), specifically, those with abnormal findings on anal cytology, or previously diagnosed of high grade (HG)-AIN and undergoing periodical revision. Individuals with normal cytology or those in whom the cytology had not yet performed were also invited to participate. Recent anorectal surgery was an exclusion criterion. The study was approved by the local Ethics Committee and all patients provided written informed consent.
Study participants were invited to be explored through HRA for screening and treatment of anal dysplasia, and for rectal secretion sampling for HIV-1 RNA and STI investigation. All HRAs were performed by a single trained infectious diseases specialist. After topical application of acetic acid and under the colposcope vision, the anoscope was passed through the squamocolumnar junction until reaching the rectal mucosa, where two flocked swabs (FLOQSwab; Copan flock technologies SRL, Brescia, Italy) were inserted and rotated against the rectal wall and the squamocolumnar junction several times. Swabs contaminated with faecal material were discarded and the procedure repeated. Each rectal swab was placed into 2 ml transport and preservation medium (eNAT tube; Copan Italia SpA, Brescia, Italy) and immediately frozen at −80 °C. The anal transformation zone and distal canal were afterwards explored using also lugol solution, and all suspicious areas were biopsied for histopathological assessment. Lesions were classified by the Bethesda system into three categories: without dysplasia, low-grade dysplasia (AIN-I) and high-grade dysplasia (AIN-II/III) . If multiple biopsy samples were taken at once, the one with the highest AIN grade was considered.
Participants were asked to collect a semen sample at home, the same day and as close as possible to the HRA visit. After liquefaction, the seminal plasma was separated from pellet by centrifugation at 970×g for 15 min and stored at −80 °C until use.
At the HRA visit, the patient completed a questionnaire about high-risk sexual behaviours.
Multiplex-PCR for sexually transmitted infections (sexually transmitted infection tests)
Rectal swabs inserted in the 2 ml collection tubes were warmed and thoroughly mixed by vortexing for 1 min. The nucleic acids were extracted from 200 μl of the preservation medium in an automated nucleic acid extractor (MagcoreHF16; RBC Bioscience, New Tapei, Taiwan). Final 60 μl nucleic acid elution sample was used to identify several pathogens by real-time PCR amplification with three commercially available kits. Specifically, the detection of Chlamydia trachomatis, Neisseria gonorrhoeae, Mycoplasma genitalium, Mycoplasma hominis, Ureaplasma urealyticum, Ureaplasma parvum and Trichomonas vaginalis was performed using the Anyplex II STI-7 Detection (Seegene, Seoul, Korea). Treponema pallidum, C. trachomatis L. serovar and herpes simplex virus (types 1 and 2) were detected by RealCycler ULCGEN-G (Progenie Molecular S.L., Valencia, Spain). The Anyplex II HPV28 (Seegene) was used for determination of 19 high-risk (16, 18, 26, 31, 33, 35, 39, 45, 51, 52, 53, 56, 58, 59, 66, 68, 69, 73, 82) and 9 low-risk (6, 11, 40, 42, 44, 54, 61, 70) human papillomavirus (HPV) genotypes.
All assay procedures were carried out in accordance with its manufacturer's protocol, in a CFX96 real-time thermocycler (Bio-Rad, Hercules, California, USA).
Quantification of HIV-1 RNA
HIV-1 RNA viral load in seminal plasma (spVL) and rectal swabs (rVL) were quantified using the COBAS AmpliPrep/COBAS TaqMan HIV-1 Test, version 2.0 (Roche Molecular Systems, Pleasanton, California, USA) with a detection threshold of 20 copies/ml. Quality controls, provided by the manufacturer and ready-to-use (human plasma COBAS TaqMan negative, low-positive and high-positive controls) were included in each run.
Rectal swabs in the 2 ml collection tubes were warmed and thoroughly mixed by vortexing for 1 min. Six hundred microlitres of the preservation medium was centrifuged at 3500 g for 15 min to remove faecal matter and suspended particles. Supernatant was diluted in plasma COBAS TaqMan Negative Control (dilution factor 1 : 2). The COBAS results were multiplied by four to obtain the correct viral load results (final limit of detection – LOD – for rVL at 80 copies/swab).
Frozen seminal samples were warmed, vortexed and diluted in plasma COBAS TaqMan Negative Control (dilution factor 1 : 2). The COBAS results were multiplied by two to obtain the correct viral load results (final LOD for spVL at 40 copies/ml).
Descriptive statistics for demographic and clinical variables were performed using frequency distributions for categorical variables and median and interquartile range (IQR) for continuous variables. Categorical variables were compared between patients with or without detectable rectal viral shedding using the chi-square or Fisher's exact tests, and continuous variables with the Mann–Whitney U-test. Random forest variable importance measures were used as a multivariate test to select the variables that best classified patients according to the presence of rectal viral shedding. This is an exploratory analysis that evaluates the discriminatory capacity of the variables and differs from the usual univariate or multivariate regression tests; therefore, it can lead to different results. Random forest interpretation was represented using the mean decrease accuracy (MDA), which estimates how much excluding or permuting each variable reduces the accuracy of the model during the out-of-bag error calculation phase. In this model, the greater the value of the MDA for a variable, the greater the discriminative capacity of that variable. The variables with a larger MDA were selected, and multivariable generalized linear mixed (GLM) models were used to identify predictors of rectal viral shedding by combining the statistically significant variables obtained in both multivariate and univariate tests. We included patient identification as a random effect in all models to control for repeated sampling. The proposed model was tested with the Hosmer–Lemeshow test.
As the variable of interest, namely, the rectal viral load, had a preponderance of undetectable values, a statistical model to deal with excess zero observations was also run, particularly, a zero-altered negative binomial (ZANB) model. In this model, the complete distribution of the outcome variable is represented by two separate components: a first part modelling the probability of excess zeros, and a ‘zero truncated’ second part accounting for all nonzero observations. When applied to our model, there is a process, which determines whether an individual is likely to have undetectable rectal viral load (viral load probability), and a second process determining the quantity of viral load among samples with detectable viral load levels (viral load intensity). To predict the viral load probability and intensity, a quantitative value was assigned to samples under LOD of the Roche assay. Specifically, samples undetectable by the assay (HIV-1 RNA not detected) were considered ‘real zero’ (rVL = 0 cp/swap) and samples detectable within the LOD (HIV-1 RNA <20 cp/ml, i.e. rVL <80 cp/swab) were assigned as rVL = 79 cp/swab.
Assuming an 8% prevalence of rectal HIV-1 RNA shedding among HIV-1 positive MSM on the basis of a previous study , a sample size of 114 patients would be required for a 95% confidence level (CI) and a margin of error of 5%. Rectal HIV-1 RNA shedding (rVL) was examined as a binary variable (detectable vs. undetectable), considering as undetectable HIV-1 values those under the LOD of the technique of HIV-1 RNA quantification. Statistical analysis was performed using Stata 8 (StataCorp Limited, College Station, Texas, USA).
One-hundred and fifty-five rectal samples from 132 virologically suppressed HIV-1-positive MSM were included in the study. The median (IQR) time between rVL and bpVL was 36 (16–57) days. Seven (4.5%) plasma viral loads were measured on the same day of the HRA procedure, and 72 (46.5%) within the same month. In 23 participants, a second rectal sample was analyzed, with median (IQR) time between samples of 16.0 (13.0–19.5) months. Two swab samples (1.29%) were directly discarded without further analysis because of abundant faecal material. Baseline characteristics of the study participants according to the presence of HIV-1 RNA rectal viral shedding are shown in Table 1. Median (IQR) age was 45 (37–54) years, and they had been virologically suppressed for 48.5 (20.7–91.0) months. Thirty-eight participants (29.9%) had previous abnormal cytology or HG-AIN in the anal biopsy.
Semen was collected from 71 patients. Among the samples, detectable levels of HIV-1 RNA were found in six (8.4 [95% CI 2.0–14.9]%), with median spVL of 96.4 (IQR 45.3–409.9) copies/ml.
HIV-1 RNA detection
HIV-1 RNA was detectable in 61 (39.3 [95% CI 31.7–47.0]%) rectal swab samples, with median (IQR) rVL of 295.8 (158.8–522.0) copies/swab. Factors associated with HIV-1 RNA rectal shedding (detectable rectal HIV-1 RNA levels) on univariate analysis were higher plasma HIV-1 RNA levels at diagnosis (P = 0.008) and lower concomitant CD4+ T-cell count (P = 0.017; Table 1). There was also an association of HIV-1 RNA rectal shedding with antiretroviral regimen composition, specifically with higher use of protease inhibitors. The frequency of recent (<6 months) unprotected anal intercourse was 54.7% (52 of 95 surveyed persons), with no differences between groups (Table 1).
Of the 23 individuals who were sampled twice, most (15/23) retained the same rVL results of the first sample; four of 23 changed from detectable to undetectable Rvl, and four of 23 changed from undetectable to detectable rVL.
For seminal HIV-1 RNA, the same statistical analyses as for rectal shedding were run, but no statistically significant associations were found (data not shown).
Bacterial sexually transmitted infection detection
In 39 of the 155 rectal swab samples (25.2%), one or more bacterial STIs were detected (Table 2). Ureaplasma urealyticum in 23 (14.8%) samples was the most frequently involved, followed by M. hominis in 22 (14.2%), and C. trachomatis in 11 (7.1%) samples. In 16 (10.3%) samples, only one pathogen was detected, in 17 (11.0%), two pathogens and in six (3.9%) samples there were three or more pathogens simultaneously detected. There was a higher proportion of several bacterial STIs among samples with detectable HIV-1 RNA (P = 0.016). No participant had symptoms or lesions suggesting the presence of an STI whenever HRA was performed. Anal HPV infection was found in 136 (87.7%) samples. High-risk HPV genotypes 16, 18 and 45 were detected in 39% of the positive samples, and other HR HPV types in 91.2%. Biopsy results showed AIN in 64.1% (84/131) of rectal samples, and HG-AIN in 27.5% (36/131). Of 19 individuals who were sampled twice, in those who already had some degree of dysplasia (LG-AIN or HG-AIN) in the first HRA (13/19), the majority (7/19) maintained the same degree, in one of 19 the degree increased, in one of 19 decreased and three of 19 had no dysplasia. The remaining six of 19 patients who did not have dysplasia in the first HRA, did show some degree of dysplasia in the second (1 HG-AIN and 5 LG-AIN).
Predictors of viral shedding
To explore the classificatory value of the study variables as predictors of rectal viral shedding, random forest analysis was performed. The model revealed that the best classifiers of rectal viral shedding were CD4+/CD8+ T-cell ratio at HRA performance, protease inhibitor-based antiretroviral regimen, CD4+ T-cell count at HRA and the nadir CD4+/CD8+ T-cell ratio (Fig. 1).
Multivariable generalized linear mixed modelling was performed to disclose factors independently associated with rectal viral shedding (Table 3). The significant variables from the univariate analysis and the best classifiers from the multivariate random forest model were included. Results showed that the presence of HG-AIN in the rectal mucosa (OR 2.85; 95% CI 1.10–7.38; P = 0.031) and protease inhibitor-based regimen (OR 2.49; 95% CI 0.98–6.34; P = 0.056) resulted in an increased risk for rectal viral shedding. By contrast, higher nadir CD4+/CD8+ T-cell ratio (OR 0.18; 95% CI 0.04–0.93; P = 0.041) and marginally, duration >18 months with suppressed plasma viremia (OR 0.48; 95% CI 0.22–1.07; P = 0.073) were negatively associated with rectal viral shedding. The burden of STIs was not kept in the model regardless of whether M. hominis and U. urealyticum were or not included.
Because of the preponderance of zero observations in the variable rVL, a ZANB model was run to fit the data (Table 4). This model showed that the best predictors of having detectable rVL were the nadir CD4+/CD8+ T-cell ratio (OR 0.98 [0.96–0.99]; P = 0.007) and protease inhibitor-based regimens (OR 4.85 [1.29–18.24]; P = 0.019); as a result, participants with lower nadir CD4+/CD8+ T-cell ratio and protease inhibitor-based current ART were more likely to have rectal viral shedding. Predictors of rectal viral shedding intensity were the presence of HG-AIN (relative risk 2.50 [1.41–4.45]; P = 0.002); and a higher burden of STIs (relative risk 1.39 [1.03–1.85], P = 0.028; Table 4). In a sensitivity analysis excluding M. hominis and U. urealyticum, the burden of STIs lost the significant association with rectal viral-shedding intensity.
A significant proportion of MSM with plasma virological suppression while receiving stable ART have detectable levels of HIV-1 RNA on rectal secretions. This may be particularly relevant, as the frequency of unprotected sexual practices is high. Factors associated with rectal viral shedding are a lower CD4+/CD8+ T-cell count nadir, co-existing HG-AIN on rectal biopsy, a higher burden of bacterial STIs and protease inhibitor-based regimens, and possibly a shorter duration of sustained plasma virological suppression.
To date, limited information was available regarding anal–rectal HIV-1 RNA shedding in virologically suppressed PLWH. This study conducted in MSM participants on stable ART with undetectable plasma viremia showed that rectal secretions still harbour HIV-1 RNA in more than one-third of participants, suggesting locally produced viral shedding coming from rectal lymphoid cells. Another finding was that the presence of detectable rectal virus was inversely associated with nadir CD4+/CD8+ T cells. A low nadir CD4+/CD8+ reflects a late initiation of ART, and this would implicate a larger-size of the HIV-1 reservoir, which could translate into higher transfer of HIV-1 to rectal secretions. Previous small studies reported an inverse correlation of HIV-1 proviral DNA levels with nadir CD4+ T-cell count .
Our results also support a central role played by local factors on anal–rectal HIV-1 RNA shedding. One of the main findings of the study was the association between HG-AIN on biopsies and rectal HIV-1 RNA shedding, which had not been previously described. Different factors might contribute to explain such an association. HPV infection causes local increased immune activation and inflammation , which has been associated with anal–rectal HIV RNA and DNA levels . Another explaining factor could be related to changes specifically related to HG-AIN lesions. In HIV-1 infected women with cervical intraepithelial neoplasia (CIN), a study showed increased HIV-1 shedding in the genital tract . The increased levels of T-cell infiltration and immune activation described in HG-CIN [21–23] might be potential implicated factors. These factors might also contribute to rectal HIV-1 replication, and could explain the higher rectal viral shedding associated with HG-AIN. Interestingly, increased levels of HIV-1 RNA and DNA were described in the biopsies obtained under anoscopic vision of the suspected HG-AIN lesions of three virologically suppressed MSM recently diagnosed with HG-AIN lesions, when compared with normal-appearing, nonlesional biopsied anal mucosa .
In addition to HPV, other STIs were also frequent in our patients. We used multiplex PCR for the diagnosis of STIs, which allowed the investigation of less well characterized STIs, and the unexpected finding of the predominance of M. hominis and U. urealyticum among the bacterial STIs detected in the anal–rectal secretions of the study participants, which had not been previously described. The multiplex PCR also permitted evaluating a high number of STIs simultaneously, and the novel finding of the positive association observed between the levels rVL and the burden of rectal STIs. Like HPV, bacterial STIs have also been associated with genital HIV-1 shedding through local increase in leukocyte concentrations and the release of proinflammatory cytokines . A higher number of STIs may be associated with a higher recruitment of leukocytes and other innate immune factors, and therefore, a potentially higher viral shedding .
Duration of virological suppression would be expected to be an important factor contributing to genital viral shedding. Our results showed a marginal association between suppressed plasma viremia for longer than 18 months and absence of rectal viral shedding, suggesting that this time frame could be an additional protective factor against persistent HIV-1 replication in rectal mucosa.
Compared with the rectal mucosa, detection of HIV-1 RNA in semen was infrequent. Factors that influence the rVL levels may differ from those for seminal viral load. The size of the rectal reservoir as part of the GALT could be the major reason for higher prevalence of rVL because of local HIV-1 production. On the other hand, the impact of ART in semen is rapid and effective, and substantial reductions in HIV RNA have been described soon after starting ART . There is less available information regarding ART levels in rectal mucosa, and varies widely within-drug and between-drug classes. Interestingly, protease inhibitors and nonnucleoside reverse-transcriptase inhibitors (NNRTIs) have demonstrated lower penetration into male and female genital tracts, and in colorectal tissue . Our results might suggest that therapy with protease inhibitors could be associated with higher rectal viral shedding than with other classes; however, this is a study with a cross-sectional design, and findings must be interpreted accordingly.
Our HIV-1 RNA assays measured total cell-free RNA, and hence cell-associated RNA (e.g. mRNA, spliced, unspliced) was not assessed. Although most data about HIV transmission come from studies of cell-free virus, and it has been suggested that ART has a higher efficacy against cell-free spread, the relative importance of cell-associated viral transmission has not yet been determined , and therefore, this would constitute a limitation of the study. Likewise, it is not possible to determine if the amount of HIV-1 RNA represents viral particles, free HIV-1 RNA or HIV-1 RNA in exosomes or cell fragments, nor whether it is infectious. Additionally, the inoculum necessary for transmission has not yet been defined; given the strong evidence for the preventive effect of ART, the significance on the risk of transmission of the detectable rectal HIV-1 RNA levels found in our study, most of which were relatively low, is largely unknown. The strengths of the study are the use of HRA to collect the cleanest and most suitable swab samples for the study of anal–rectal secretions, which also allowed the simultaneous histopathological study of HG-AIN lesions obtained through biopsy . The study included multiplex real-time PCR, which is probably the most sensitive test for diagnosing STIs, and allowed the simultaneous detection of numerous and less well known STIs. We studied HIV-1 RNA detection in two genital compartments on ART-suppressed patients, semen and rectal secretion. No studies have made comparisons in male genital and anal–rectal tract in ART-suppressed population.
Rectal HIV-1 RNA shedding is common in virologically suppressed PLWH receiving ART, who have frequently unprotected sexual practices. Local factors, including HG-AIN and the burden of anal–rectal STIs, as well as systemic conditions, like the level of immune impairment before ART initiation or the class of antiretroviral drugs received, predict rectal HIV-1 shedding. More information is needed to assess whether these factors could be additive. Although the significance of this residual shedding is undetermined, it suggests that plasma HIV-1 RNA measurements do not necessarily reflect the viral concentrations at the rectal reservoir, and additional factors, including those related to local environment, may account for the persistence of HIV-1 RNA shedding in rectal mucosa in HIV-1 positive MSM.
F.G. and M.M. contributed to the conception and design of the study. F.G., M.M. and S.P. contributed to data collection. E.G., M.F., C.R. and V.O. contributed to microbial analyses. J.A.G. and S.P. contributed to data analysis. J.A.G., S.P, F.G. and M.M. contributed to data interpretation. E.G., M.F., F.G. and M.M contributed to writing and editing the manuscript. All authors revised and approved the final version of the manuscript
This work was supported by the Plan Nacional R + D + I as a part of project [RD12/0017/0023] and cofinanced by ISCIII – Subdirección General de Evaluación y el Fondo Europeo de Desarrollo Regional (FEDER): FIS [PI08/893], FIS [PI13/02256], FIS [PI16/01740], Contrato de Intensificación de la Actividad Investigadora [INT 14/00207], Contrato Río Hortega [CM15/00187] and by FISABIO: FISABIO [UGP-14–197] and Contrato Predoctoral FISABIO 2015 [UGP-15–152].
Funding: Instituto de Salud Carlos III, Ministerio de Economia, Industria y Competitividad, Spain, programme FIS PI13/02256.
Conflicts of interest
There are no conflicts of interest.
1. People living with HIV. UNAIDS 2014. Available at: http://aidsinfo.unaids.org/
2. Sullivan PS, Jones JS, Baral SD. The global north: HIV epidemiology in high-income countries
. Curr Opin HIV AIDS
3. Fettig J, Swaminathan M, Murrill CS, Kaplan JE. Global epidemiology of HIV
. Infect Dis Clin North Am
4. Patel P, Borkowf CB, Brooks JT, Lasry A, Lansky A, Mermin J. Estimating per-act HIV transmission risk: a systematic review
5. Cohen MS, Chen YQ, McCauley M, Gamble T, Hosseinipour MC, Kumarasamy N, et al. Antiretroviral therapy for the prevention of HIV-1 transmission
. N Engl J Med
6. Anglemyer A, Horvath T, Rutherford G. Antiretroviral therapy for prevention of HIV transmission in HIV-discordant couples
7. Rodger AJ, Cambiano V, Bruun T, Vernazza P, Collins S, van Lunzen J, et al. Sexual activity without condoms and risk of HIV transmission in serodifferent couples when the HIV-positive partner is using suppressive antiretroviral therapy
8. Galvin SR, Cohen MS. Genital tract reservoirs
. Curr Opin HIV AIDS
9. Archin NM, Sung JM, Garrido C, Soriano-Sarabia N, Margolis DM. Eradicating HIV-1 infection: seeking to clear a persistent pathogen
. Nat Rev Microbiol
10. Stefano M Di, Favia A, Monno L, Lopalco P, Caputi O, Scardigno AC, et al. Intracellular and cell-free (Infectious) HIV-1 in rectal mucosa
. J Med Virol
11. Kiviat NB, Critchlow CW, Hawes SE, Kuypers J, Surawicz C, Goldbaum G, et al. Determinants of human immunodeficiency virus DNA and RNA shedding in the anal-rectal canal of homosexual men
. J Infect Dis
12. Zuckerman R a, Lucchetti A, Whittington WLH, Sanchez J, Coombs RW, Zuñiga R, et al. Herpes simplex virus (HSV) suppression with valacyclovir reduces rectal and blood plasma HIV-1 levels in HIV-1/HSV-2-seropositive men: a randomized, double-blind, placebo-controlled crossover trial
. J Infect Dis
13. Johnson LF, Lewis DA. The effect of genital tract infections on HIV-1 shedding in the genital tract: a systematic review and meta-analysis
. Sex Transm Dis
14. Kelley CF, Haaland RE, Patel P, Evans-Strickfaden T, Farshy C, Hanson D, et al. HIV-1 RNA rectal shedding is reduced in men with low plasma HIV-1 RNA viral loads and is not enhanced by sexually transmitted bacterial infections of the rectum
. J Infect Dis
15. Zuckerman RA, Whittington WLH, Celum CL, Collis TK, Lucchetti AJ, Sanchez JL, et al. Higher concentration of HIV RNA in rectal mucosa secretions than in blood and seminal plasma, among men who have sex with men, independent of antiretroviral therapy
. J Infect Dis
16. Lampinen TM, Critchlow CW, Kuypers JM, Hurt CS, Nelson PJ, Hawes SE, et al. Association of antiretroviral therapy with detection of HIV-1 RNA and DNA in the anorectal mucosa of homosexual men
17. Durragh TM, Colgan TJ, Cox T, Heller D. The lower anogenital squamous terminology standardization project for HPV-associated Lesions
. Arch Pathol Lab Med
18. Boulassel MR, Chomont N, Pai NP, Gilmore N, Sékaly RP, Routy JP. CD4 T cell nadir independently predicts the magnitude of the HIV reservoir after prolonged suppressive antiretroviral therapy
. J Clin Virol
19. Behbahani H, Walther-Jallow L, Klareskog E, Baum L, French AL, Patterson BK, et al. Proinflammatory and type 1 cytokine expression in cervical mucosa during HIV-1 and human papillomavirus infection
. J Acquir Immune Defic Syndr
20. Spinillo A, Zara F, Gardella B, Preti E, Gaia G, Maserati R. Cervical intraepithelial neoplasia and cervicovaginal shedding of human immunodeficiency virus
. Obstet Gynecol
2006; 107 (2 Pt 1):314–320.
21. Garcia-Chagollan M, Jave-Suarez LF, Haramati J, Sanchez-Hernandez PE, Aguilar-Lemarroy A, Bueno-Topete MR, et al. Substantial increase in the frequency of circulating CD4+NKG2D+ T cells in patients with cervical intraepithelial neoplasia grade 1
. J Biomed Sci
22. Coleman N, Greenfield IM, Hare J, Kruger-Gray H, Chain BM, Stanley MA. Characterization and functional analysis of the expression of intercellular adhesion molecule-1 in human papillomavirus-related disease of cervical keratinocytes
. Am J Pathol
23. Coleman N, Stanley MA. Characterization and functional analysis of the expression of vascular adhesion molecules in human papillomavirus-related disease of the cervix
24. Pollakis G, Riche O, Vis JD, Prins JM, Paxton WA, De Vries HJC. Increased HIV-1 activity in anal high-grade squamous intraepithelial lesions compared with unaffected anal mucosa in men who have sex with men
. Clin Infect Dis
25. Cummins JE, Christensen L, Lennox JL, Bush TJ, Wu Z, Malamud D, et al. Mucosal innate immune factors in the female genital tract are associated with vaginal HIV-1 shedding independent of plasma viral load
. AIDS Res Hum Retroviruses
26. Sheth PM, Kovacs C, Kemal KS, Jones RB, Raboud JM, Pilon R, et al. Persistent HIV RNA shedding in semen despite effective antiretroviral therapy
27. Trezza CR, Kashuba ADM. Pharmacokinetics of antiretrovirals in genital secretions and anatomic sites of HIV transmission: Implications for HIV prevention
. Clin Pharmacokinet
28. Ronen K, Sharma A, Overbaugh J. HIV transmission biology: translation for HIV prevention
29. Salit IE, Lytwyn A, Raboud J, Sano M, Chong S, Diong C, et al. The role of cytology (Pap tests) and human papillomavirus testing in anal cancer screening
* Félix Gutiérrez and Mar Masiá contributed to this work equally.