Dominguez-Molina B et al. Journal of the International AIDS Society 2020, 23:e25607 http://onlinelibrary.wiley.com/doi/10.1002/jia2.25607/full | https://doi.org/10.1002/jia2.25607 RESEARCH ARTICLE Persistent HIV-controllers are more prone to spontaneously clear HCV: a retrospective cohort study s-Guisado1, Miguel Genebat2,3, Beatriz Dominguez-Molina1,2, Laura Tarancon-Diez1,2, Yusnelkis Milane Salvador Resino 4 , Carmen Rodriguez , Juan Gonzalez-Garcıa , Norma Rallo 5 6 n7,8, Maria Pernas9 , Concepcion Casado , Cecilio Lopez-Galındez , Agathe Leo 9 9 n , Jose M Benito , Felipe Garcıa10, Jorge Del Romero5, 10 8,9 Pompeyo Viciana , Luis F Lopez-Cortes , Manuel Leal *, Ezequiel Ruiz-Mateos1,§,* 1 1 2,11, and ECRIS integrated in the Spanish AIDS Research Network1 § Corresponding author: Ezequiel Ruiz-Mateos, Infectious Diseases Unit, Clinical Microbiology and Preventive Medicine, Institute of Biomedicine of Seville, CSIC, Virgen del Rocio University Hospital, Avd. Manuel Siurot s/n. CP 41013, Seville, Spain. Tel: +34955923107. (ezequiel.ruizmateos@gmail.com) *These authors equally contributed to this work a The clinical Centers which contribute to ECRIS are shown in the Annex S1. Abstract Introduction: HIV-controllers have the ability to spontaneously maintain viraemia at low or undetectable levels in the absence of antiretroviral treatment. Furthermore, HIV-controllers seem to have a superior capacity to spontaneously clear hepatitis C virus (HCV) compared to non HIV-controllers. Some of these subjects eventually lose HIV-controller status (transient con- trollers), whereas some HIV-controllers show a persistent natural HIV control (persistent controllers). We aimed to analyse whether persistent controllers have superior capacity to spontaneously clear HCV compared to transient controllers. Methods: We recruited HIV-controllers from January 1981 up to October 2016 with available antibodies to HCV (anti-HCV) data (n = 744). Factors associated with HIV spontaneous control in relation to HCV status were analysed in persistent and transient HIV-controllers with anti-HCV positive (n = 202 and n = 138 respectively) in comparison with 1700 HCV positive non HIV-controllers recruited from January 1981 up to March 2018, bivariate and multivariate analyses, following a logistic regression model, were applied. In addition, the factors related to the loss and time to lose HIV-controller status were explored (n = 744) using Log rank test and Kaplan–Meier curves, in this case the multivariate analysis consisted in a Cox regression model. Results: A higher frequency of HCV spontaneous clearance was found in persistent HIV-controllers (25.5%) compared to non- controllers (10.2%). After adjusting for potential confounders, as sex, age, HIV transmission risk, CD4+ T-cell nadir and time of follow-up, HCV clearance was independently associated with persistent HIV spontaneous control (p = 0.002; OR (95% CI) = 2.573 (1.428 to 4.633)), but not with transient spontaneous control (p = 0.119; 1.589 (0.888 to 2.845)). Furthermore, persistent HIV-controllers were more likely to spontaneously clear the HCV in comparison with transient controllers (p = 0.027; 0.377 (0.159 to 0.893). Finally, not to lose or lengthen the time of losing this control was independently associated with HCV spontaneous clearance (p = 0.010; 0.503 (0.297 to 0.850). Conclusions: This study shows an association between spontaneous persistent HIV-control and HCV spontaneous clearance. The study findings support the idea of preserved immune mechanisms in persistent HIV control implicated in HCV sponta- neous clearance. These results highlight persistent HIV-controllers but not transient controllers as a good model of functional HIV cure. Keywords: HIV; HCV; HIV-controllers; persistent; transient; HCV spontaneous clearance Additional information may be found under the Supporting Information tab for this article. Received 4 March 2020; Accepted 7 July 2020 Copyright © 2020 The Authors. Journal of the International AIDS Society published by John Wiley & Sons Ltd on behalf of the International AIDS Society. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. 1 | INTRODUCTION controllers) HIV viral load (VL) in the absence of antiretroviral treatment [1]. In large cohort studies [2,3], it is clear that this HIV-controllers are a small group of subjects (<1%) that main- phenotype is heterogeneous with some HIV-controllers even- tain low (viraemic controllers) or undetectable (elite tually losing the spontaneous viral control capacity, these 1 Dominguez-Molina B et al. Journal of the International AIDS Society 2020, 23:e25607 http://onlinelibrary.wiley.com/doi/10.1002/jia2.25607/full | https://doi.org/10.1002/jia2.25607 individuals are named transient controllers as opposed to per- than one year. Subjects without loss of HIV-controller status sistent controllers [4]. Understanding the mechanisms were defined as “Persistent HIV-controllers” (n = 411). HCV involved in this phenomenon is quite important. On one hand spontaneous clearers were defined as patients exposed to factors associated with the loss of HIV-controller status may HCV (anti-HCV positive) with qualitative negative HCV PCR assist giving an optimum clinical care to these individuals and in the absence of HCV specific treatment. These three data reciprocally, persistent controllers may serve as a good model (anti-HCV positive, HCV PCR and HCV specific treatment) for functional HIV cure. The mechanisms and attributes of this must be available to classify a patient as a HCV spontaneous extraordinary phenotype of exceptionally long-term control clearer. are still not completely understood [5-7]. Recent studies show For the analysis of the factors associated with persistent that persistent HIV-controllers in comparison with transient and transient HIV spontaneous control based on HCV pro- controllers have higher levels of HIV-specific T-cell response, file, we selected HCV exposed persistent (n = 202) and tran- low viral evolution, low levels of inflammation and a peculiar sient (n = 138) HIV-controllers and non HIV-controllers proteomic and metabolomic profile [4,8,9]. (n = 1700). HCV exposed HIV non-controllers were subjects On the other hand, HIV-controllers as a whole seem to with consecutive visits to the Infectious Disease Unit at Vir- have a higher capacity to spontaneously clear hepatitis C virus gen del Rocıo University Hospital (Seville, Spain) from Jan- (HCV) compared to non HIV-controllers [10,11], although uary 1981 up to March 2018. These subjects were other studies have not found this enhanced capacity [12,13]. administered ART based on clinical criteria of contemporary Besides, it remains unknown whether persistent HIV-con- guidelines. trollers have a higher capacity to spontaneously clear HCV. For the analysis of factors associated with the loss of HIV According to the improved immunological capacities men- spontaneous control we used the complete HIV ECRIS cohort tioned earlier, we hypothesize that persistent HIV-controllers (n = 744) with available anti-HCV data. In addition, we per- may have a better ability to spontaneously clear HCV. These formed the analysis of variables associated with the time to findings could lead to significant advances on the chronic viral lose HIV-controller status. The time to lose HIV-controller sta- infections therapeutically designs as they act as an additional tus was considered the time period between the HIV diagno- attribute of the persistent HIV-controllers as a model for a sis and the loss of HIV-controller status. The subjects who did HIV functional cure. not lose HIV-controller status were censored in the last fol- In this study we have been able to investigate HCV status low-up point. profile in the large cohort of HIV-controllers of the Spanish All subjects gave written informed consent in accordance AIDS Research Network (ECRIS). Thus, the aim of this study with the Declaration of Helsinki. The protocol was approved was to analyse whether persistent HIV-controllers had supe- by the Comite  de Etica de la Investigacio n at the Hospital rior capacity to spontaneously clear HCV compared to tran- Universitario Virgen del Rocıo de Sevilla (2012PI/240). sient HIV-controllers. Routinely, the absolute numbers of CD4+ T cells were assayed from fresh whole blood samples, using the Epics XL- MCL flow cytometer (Beckman Coulter). Plasma HIV RNA 2 | METHODS load was measured in fresh samples by quantitative poly- merase chain reaction (PCR; COBAS Ampliprep/COBAS Taq- 2.1 | Patients and variables man HIV test; Roche molecular systems), according to the Patients’ clinical and epidemiological data were included in the manufacturer’s instructions. The detection limit was 50 HIV cohort of HIV-controllers of the AIDS Research Network RNA copies/mL. Qualitative and quantitative reverse tran- (ECRIS). This is an open, multicentre cohort of HIV controllers scription PCR was performed for plasma HCV RNA amplifica- whose data come from the Long-Term Non-Progressors tion (COBAS Amplicor; Roche Diagnosis, Barcelona, Spain), cohort and the Cohort of the Spanish AIDS Research Network with detection limit of 15 IU/mL. HCV genotype was deter- (CoRIS) [14,15] and other clinical centres across Spain. This mined using a reverse-hybridization assay (InnoLIPA HCV II; cohort has been previously published [3,16]. See Annex S1 for Innogenetics, Barcelona, Spain). participating centres. In the last update, October 2016, this cohort included 807 HIV-controllers. 2.2 | Statistical analysis and Covariates This is a retrospective cohort study. Inclusion criteria were patients with VL < 2000 HIV RNA copies/mL ((elite con- Continuous variables were expressed as median and trollers, VLs below the detection limit [<50 HIV-RNA copies/ interquartile range (IQR) and categorical variables were mL], n = 273) and viraemic controllers (controllers with VLs expressed as number and percentage. Bivariate and multi- between 50 and 2000 HIV-RNA copies/mL, n = 471)) in the variate analyses, following a logistic regression model, were absence of antiretroviral treatment for at least one year of applied for the analysis of factors associated with persistent follow-up and with available antibodies to HCV (anti-HCV) and transient HIV spontaneous control and for the analysis data (n = 744). HIV-controllers were split into “Transient HIV- of factors associated with loss of HIV-controller status. For controllers” (n = 333), subjects who lost HIV-controller status these two analyses the dependent variables were being HIV- defined as loss of virological control (first viral measurement controller (persistent and transient) and loss of HIV-control of a sustained increase of >2000 HIV RNA copies/mL during respectively. For the analysis of time to lose HIV-controller more than one year of follow-up) or initiation of an antiretro- status, the time to lose this condition was considered the viral treatment, as previously reported [16]. The same criteria dependent variable, Log rank test and Kaplan–Meier curves were applied for the loss of elite controller status but with a were used. In this case the multivariate analysis consisted in sustained increase of >50 HIV RNA copies/mL during more a Cox regression model, using the enter procedure. As 2 Dominguez-Molina B et al. Journal of the International AIDS Society 2020, 23:e25607 http://onlinelibrary.wiley.com/doi/10.1002/jia2.25607/full | https://doi.org/10.1002/jia2.25607 covariates these models were adjusted by sex, age at HIV 3.2 | HCV spontaneous clearance was diagnosis, time of follow-up, people who inject drugs as independently associated with persistent (PWID) as risk of HIV transmission, CD4+ T-cell nadir, HLA- HIV-controller status B*57 typing, and finally by HIV elite-controller status and HCV spontaneous clearance (as describe earlier), as potential We analysed the whole HIV-controller subset in detail confounders. All variables with a p < 0.1 were included in (n = 744). Table 4 shows the differences between persistent the multivariate analysis, and p < 0.05 was considered statis- controllers and transient controllers. Persistent controllers tically significant. Statistical analyses were performed using had longer time of follow-up, higher CD4 T-cell nadir and SPSS software version 22 (IBM SPSS, Chicago, Illinois) and higher frequency of elite control, PWID and HLA-B*57 carri- graphs were generated with Prism, version 6.0 (GraphPad ers than transient controllers. We also categorized patients Software, Inc.). based on the calendar year of HIV diagnosis. Interestingly, while there were more persistent controllers HIV-diagnosed in the 80s decade, there were more transient controllers HIV- 3 | RESULTS diagnosed in the 2000s. It is important to note that there were no differences in HCV genotype distribution between 3.1 | Higher frequency of HCV spontaneous groups (Table 4). These results point out that the loss of con- clearance in persistent HIV-controllers, but not in trol was not associated with the different genotype distribu- transient HIV-controllers, compared to non- tion. controllers Regarding HCV profile, HIV persistent controllers had higher frequency of anti-HCV positive than transient con- First, we found that the rate of HCV spontaneous clearance trollers. Interestingly, there was a higher proportion of HCV in the overall HIV-controller population (n = 340) was higher spontaneous clearers in persistent controllers. In a multivari- compared with non-controllers (n = 1700) (19.4% vs. 10.2% ate analysis, the factors independently associated with loss of respectively) confirming previous results in small cohorts HIV-controller status were lower CD4 T-cell nadir, non-elite [10,11]. After adjusting for sex, age at HIV diagnosis, time of controller status and not achieving HCV spontaneous clear- follow-up and the risk of transmission  frequency of PWID ance (table 5). We also adjusted by calendar year of HIV posi- , as potential confounders, HCV spontaneous clearance tive instead of by the time of follow-up, as these were co- remained independently associated with HIV spontaneous linear variables and similar results were found (data not control (p < 0.001; 2.180 (1.472 to 3.229). Afterwards, we shown). We could not adjust simultaneously by HLA-B*57 and compared HCV-exposed persistent (n = 202) and transient spontaneous HCV clearance due to the low number of sub- HIV-controllers (n = 138) with the 1700 HCV positive HIV jects sharing both determinations. However, when we included non-controllers (Table 1). We found that in both groups there HLA-B*57 genotype in the previous model no association was was a higher proportion of women, higher nadir CD4+ T-cell found between this variable and the loss of HIV-controller sta- levels, lower frequency of patients at clinical stage CDC C, a tus (p = 0.384; OR (95%) = 0.650 (0.247 to 1.714)). higher representation of HCV genotype 3 and a lower repre- The same results were obtained when the patients were sentation of genotype 1 and, as expected, an overrepresenta- categorized based only on the loss of virological control, not tion of HLA-B57 + subjects compared to HIV-non-controllers on the loss of HIV controller status (n = 563 and 181 respec- (Table 1). However, persistent HIV-controllers presented tively). After adjusting by sex, time of follow-up and PWID, higher rates of HCV spontaneous clearance than HIV non-con- again: CD4 T-cell nadir, the frequency of elite controllers and trollers (25.5% vs. 10.2%), although this was not the case for HCV spontaneous clearance were independently associated transient HIV-controllers (14.5% vs. 10.2%). After adjusting with not having loss virological control ((p < 0.001; OR for potential confounders, HCV spontaneous clearance (95%) = 0.995 (0.993 to 0.997), p < 0.001; OR (95%) = 0.136 remained independently associated with persistent HIV spon- (0.068 to 0.274) and p = 0.027; OR (95%) = 0.377 (0.159 to taneous control but not with transient (Table 2). Due to the 0.893) respectively)). scarce data, we could not adjust for HLA-B*57 typing, but it was associated with both persistent and transient control (Table 2). 3.3 | HCV spontaneous clearance was associated When we restricted the analysis to persistent elite con- with the time to maintain HIV-controller status trollers (n = 143) and transient elite controllers (n = 43) Another way of analysing the results is to perform a time to (Table 3), again a higher frequency of HCV spontaneous event analysis. The event was the loss of HIV-controller sta- clearers was found in persistent elite controllers than in non- tus. The median time of follow-up was 7.2 [3.7 to 17.7] years. controllers (20.6% vs. 10.2%) but not in transient elite con- The analysis revealed that younger age at diagnosis, PWID as trollers ((11.1% vs. 10.2%) (p = 0.008; OR (95%) = 2.279 risk factor, high CD4+ T-cell nadir, being elite, HLA-B*57 carri- (1.240 to 4.189) and p = 0.860; OR (95%) = 1.099 (0.384 to ers and having cleared HCV spontaneously were associated 3.144) respectively). In the multivariate analysis, in a first with maintaining HIV-controller status overtime (Figure 1). model after adjusting by sex, again, HCV spontaneous clear- After adjusting by all these variables, but HLA-B*57, in a Cox ance was associated with persistent elite HIV control regression model, having cleared HCV spontaneously was (p = 0.016; OR (95%) = 2.127 (1.149 to 3.938). In a second independently associated with the time to maintain HIV-con- model after adjusting by sex and CD4+ T-cell nadir levels this troller status together with the rest of the variables men- correlation was lost (p = 0.123; OR (95%) = 1.749 (0.859 to tioned earlier (Table 6). However, when we included HLA- 3.558). B*57 genotype in the previous model no association was 3 Table 1. Characteristics of persistent and transient HIV-controllers and HIV/HCV positive non-controllers Persistent HIV-controllers vs. HIV Transient HIV-controllers vs. HIV Persistent HIV-controllers Transients HIV-controllers HIV non-controllers non-controllers non-controllers (n = 202) (n = 138) (n = 1700) p; OR (CI 95%) p; OR (CI95%) Female sex (n) % 66/202 (32.7) 49/138 (35.5) 235/1700 (13.8) <0.001; 3.025 (2.187 to 4.186) <0.001; 3.432 (2.359 to 4.993) Age at HIV diagnosis 28 [24 to 33] 28 [23 to 33] 27 [23 to 33] 0.360; 1.008 (0.991 to 1.025) 0.667; 1.004 (0.984 to 1.025) HIV transmission risk, 162/202 (80.2) 102/138 (73.9) 1379/1700 (83.6) 0.226; 0.796 (0.550 to 1.152) 0.004; 0.557 (0.373 to 0.832) PWID (n) % PWID, female (n) % 47/66 (71.2) 32/49 (65.3) 161/235 (70.9) 0.964; 1.074 (0.554 to 1.857) 0.437; 0.772 (0.401 to 1.484) CD4 nadir 520 [388 to 675] 285 [213 to 396] 111 [31 to 224] <0.001; 1.009 (1.008 to 1.010) <0.001; 1.006 (1.005 to 1.007) Time of follow-up 18.1 [11.7 to 23.7] 15.8 [8.0 to 21.0] 18.6 [11.8 to 24.9] 0.578; 0.995 (0.978 to 1.012) <0.001; 0.961 (0.942 to 0.981) (years)a Clinical stage C (n) %b 7/107 (6.5) 10/121 (8.3) 275/1674 (16.4) 0.009; 0.356 (0.164 to 0.775) 0.020; 0.458 (0.237 to 0.887) HCV PCR+ (n) %c 99/123 (80.5) 113/130 (86.9) 1521/1694 (89.8) 0.002; 0.469 (0.292 to 0.753) 0.304; 0.756 (0.443 to 1.289) Dominguez-Molina B et al. Journal of the International AIDS Society 2020, 23:e25607 HCV genotype detected (n) %d Genotype 1 28/132 (40.0) 37/84 (44.0) 616/1075 (57.3) 0.005; 0.497 (0.303 to 0.813) 0.019; 0.587 (0.375 to 0.918) Genotype 2 2/132 (2.9) 3/84 (3.6) 9/1075 (0.8) 0.115; 3.484 (0.738 to 16.440) 0.029; 4.387 (1.165 to 16.521) http://onlinelibrary.wiley.com/doi/10.1002/jia2.25607/full | https://doi.org/10.1002/jia2.25607 Genotype 3 23/132 (32.9) 31/84 (36.9) 237/1075 (22) 0.038; 1.730 (1.030 to 2.908) 0.002; 2.068 (1.298 to 3.296) Genotype 4 17/132 (24.2) 13/84 (15.5) 213/1075 (19.8) 0.367; 1.298 (0.737 to 2.287) 0.335; 0.741 (0.403 to 1.364) HCV spontaneous 24/94 (25.5) 17/117 (14.5) 173/1694 (10.2) <0.001; 3.014 (1.847 to 4.918) 0.143; 1.495 (0.873 to 2.559) clearers (n) %e HLA-B*57 (n) %f 25/79 (31.6) 12/56 (21.4) 41/818 (5) <0.001; 8.774 (4.968 to 15.45) <0.001; 5.169 (2.538 to 10.53) p values < 0.05 are shown in bold. CI, confidence interval;.OR, odd ratio; PWID, people who inject drug. a Time period between HIV diagnosis and loss of HIV-controller status or last monitored visit point; bclinical stage C data from 107 persistent HIV-controllers, 121 transient HIV-controllers and 1674 non-controllers; cHCV PCR + was considered before HCV treatment onset in those who were treated; dHCV genotype detected from 132 persistent HIV-controllers, 84 transient HIV- controllers and 1075 non-controllers; efor HCV spontaneous clearance calculation we only considered patients with available data of anti-HCV, HCV PCR and HCV treatment. 94 persistent HIV-controllers, 117 transient HIV-controllers and 1694 non-controllers accomplished all data. ; fHLA-B*57 data from 79 persistent HIV-controllers, 56 transient HIV-controllers and 818 non- controllers. 4 Dominguez-Molina B et al. Journal of the International AIDS Society 2020, 23:e25607 http://onlinelibrary.wiley.com/doi/10.1002/jia2.25607/full | https://doi.org/10.1002/jia2.25607 Table 2. Factors associated with persistent and transient HIV spontaneous control Persistent HIV-control Transient HIV-control Variables Adjusted p; OR (IC 95%) Adjusted p; OR (IC 95%) HCV spontaneous clearance 0.002; 2.573 (1.428 to 4.633) 0.119; 1.589 (0.888 to 2.845) Female sex 0.006; 2.176 (1.250 to 3.790) <0.001; 2.854 (1.823 to 4.467) CD4 nadir <0.001; 1.009 (1.008 to 1.010) <0.001; 1.006 (1.005 to 1.007) Time of follow-up NA 0.002; 0.961 (0.938 to 0.985) HIV transmission risk, PWID NA 0.530; 0.855 (0.523 to 1.396) p values < 0.05 are shown in bold. CI, confidence interval; NA, not applicable; OR, odd ratio, PWID, people who inject drug. Table 3. Characteristics of persistent and transient HIV elite-controllers and HIV/HCV positive non-controllers Persistent elite Transients elite HIV non- Persistent HIV-controllers vs. Transient HIV-controllers vs. controllers controllers controllers HIV non-controllers HIV non-controllers (n = 143) (n = 43) (n = 1700) p; OR (CI95%) p; OR (CI95%) Female sex (n) % 50/143 (35.0) 17/43 (39.5) 235/1700 <0.001; 3.352 (2.314 to 4.854) <0.001; 4.076 (2.178 to 7.628) (13.8) Age at HIV 28 [23 to 33] 29 [24 to 35] 27 [23 to 33] 0.397; 1.009 (0.989 to 1.029) 0.080; 1.028 (0.997 to 1.060) diagnosis HIV transmission 115/143 (80.4) 32/43 (74.4) 1379/1700 0.332; 0.807 (0.523 to 1.245) 0.116; 0.572 (0.285 to 1.148) risk, PWID (n) (83.6) % PWID, female (n) 47/66 (71.2) 32/49 (65.3) 161/235 (70.9) 0.964; 1.074 (0.554 to 1.857) 0.437; 0.772 (0.401 to 1.484) % Time of follow- 18.5 [13.0 to 17.4 [8.1 to 22.6] 18.6 [11.8 to 0.491; 1.007 (0.987 to 1.028) 0.094; 0.971 (0.938 to 1.005) up (years) 24.7] 24.9] CD4 nadir 510 [363 to 680] 292 [220 to 420] 111 [31 to <0.001; 1.009 (1.008 to 1.011) <0.001; 1.005 (1.004 to 1.007) 224] Clinical stage C 7/72 (9.7) 3/41 (7.0) 275/1674 0.136; 0.548 (0.249 to 1.208) 0.131; 0.402 (0.123 to 1.310) (n) %a (16.4) HCV PCR+ (n) 76/90 (84.4) 38/42 (90.5) 1521/1694 0.110; 0.617 (0.342 to 1.115) 0.884; 1.081 (0.381 to 3.064) %b (89.8) HCV genotype detected (n) %c Genotype 1 23/54 (42.6) 11/28 (39.3) 616/1075 0.036; 0.553 (0.318 to 0.961) 0.063; 0.482 (0.224 to 1.039) (57.3) Genotype 2 1/54 (1.9) 0/28 (0) 9/1075 (0.8) 0.450; 2.235 (0.278 to 17.956) 0.999 Genotype 3 16/54 (29.6) 13/28 (46.4) 237/1075 (22) 0.195; 1.489 (0.816 to 2.717) 0.004; 3.064 (1.438 to 6.530) Genotype 4 14/54 (25.9) 4/28 (14.3) 213/1075 0.276; 1.416 (0.757 to 2.651) 0.470; 0.674 (0.232 to 1.965) (19.8) HCV 14/68 (20.6) 4/36 (11.1) 173/1694 0.008; 2.279 (1.240 to 4.189) 0.860; 1.099 (0.384 to 3.144) spontaneous (10.2) clearers (n) %d HLA-B*57 (n) %e 25/64 (39.1) 3/20 (15.0) 41/818 (5) <0.001; 12.15 (6.72 to 22.00) 0.062; 3.344 (0.942 to 11.87) p values < 0.1 are shown in bold. CI, confidence interval; NA, not applicable; OR, odd ratio; PWID, people who inject drug. a Clinical stage C data from 72 persistent HIV-controllers, 41 transient HIV-controllers and 1674 non-controllers; bHCV PCR+ was considered before HCV treatment onset in those who were treated; cHCV genotype detected from 54 persistent HIV-controllers, 28 transient HIV-con- trollers and 1075 non-controllers. 5 Dominguez-Molina B et al. Journal of the International AIDS Society 2020, 23:e25607 http://onlinelibrary.wiley.com/doi/10.1002/jia2.25607/full | https://doi.org/10.1002/jia2.25607 Table 4. Characteristics of HIV-controllers stratified according to the loss of HIV control status Persistent controllers (411) Transient controllers (n = 333) p; OR (IC 95%) Female sex (n) % 114/417 (27.7) 111/333 (33.3) 0.099; 1.303 (0.952 to 1.783) Age at HIV diagnosis 29 [24 to 35] 30 [25 to 36] 0.334; 1.008 (0.992 to 1.024) Calendar year of HIV diagnosis 1980 to 1989 91/411 (22.1) 52/332 (15.7) 0.027; 0.653 (0.448 to 0.952) 1990 to 1999 99/411 (24.1) 94/332 (28.3) 0.192; 1.245 (0.896 to 1.729) 2000 to 2009 157/411 (38.2) 174/332 (52.4) <0.001; 1.782 (1.329 to 2.389) 2010 to 2015 64/411 (15.6) 12/332 (3.6) <0.001; 0.203 (0.108 to 0.384) Time of follow-up (years)a 9.4 [4.1 to 19.1] 5.7 [3.0 to 14.7] <0.001; 0.960 (0.942 to 0.977) HIV transmission risk, PWID (n) % 167/411 (40.6) 106/333 (31.8) 0.013; 0.682 (0.504 to 0.924) CD4 T-cell nadir 550 [419 to 710] 327 [253 to 443] <0.001; 0.994 (0.993 to 0.995) Clinical stage C (n) %b 7/296 (2.4) 12/305 (3.9) 0.277; 1.691 (0.656 to 4.356) Elite HIV-controller status (n) % 207/411 (50.4) 66/333 (19.8) <0.001; 0.244 (0.175 to 0.339) Anti-HCV+ (n, %) 202/411 (49.1) 138/333 (41.4) 0.036; 0.732 (0.547 to 0.980) HCV PCR+ (n,%)c 99/332 (29.8) 113/325 (34.8) 0.175; 1.254 (0.904 to 1.741) Log10 HCV VL (IU/mL)d 5.85 [5.40 to 6.40] 6.07 [5.41 to 6.60] 0.289; 1.166 [0.878 to 1.550] HCV genotype detected (n) %e Genotype 1 28/70 (40) 37/84 (44) 0.613; 1.181 (0.620 to 2.248) Genotype 2 2/70 (2.9) 3/84 (3.6) 0.804; 1.259 (0.204 to 7.756) Genotype 3 23/70 (32.9) 31/84 (36.9) 0.600; 1.195 (0.613 to 2.329) Genotype 4 17/70 (24.3) 13/84 (15.5) 0.172; 0.571 (0.255 to 1.277) HCV clearance (n%)f 24/94 (25.5) 17/117 (14.5) 0.047; 0.496 (0.248 to 0.991) HLA-B*57 (n) %g 31/105 (29.5) 15/96 (15.6) 0.021; 0.442 (0.221 to 0.883) p values < 0.1 are shown in bold. CI, confidence interval; NA, not applicable; OR, odd ratio; PWID, people who inject drug. a Time period between HIV diagnosis and loss of HIV-controller status or last monitored visit point; bclinical stage C data from 296 persistent con- trollers, and 305 transient controllers; cHCV PCR data available for 332 persistent controllers and 325 transient controllers. HCV PCR+ was con- sidered before HCV treatment onset in those who were treated; dHCV VL data from 96 persistent controllers and 101 transient controllers; e HCV genotype detected in 70 persistent controllers, and 84 transient controllers; ffor HCV spontaneous clearance calculation we only considered patients with anti-HCV+ and available data of HCV PCR and HCV treatment. HCV spontaneous clearance was calculated in 94 persistent con- trollers and 117 transient controllers; gHLA-B*57 data from 105 persistent controllers, and 96 transient controllers. Table 5. Factors associated with loss of HIV-controller status time to lose HIV-controller status was associated with HCV spontaneous clearance. Adjusted HIV-controller is a heterogeneous phenotype [2,3], some p; OR (IC 95%) HIV-controllers eventually lose HIV-controller status [17–19]. It is a key to detect these patients before the loss of control Female sex 0.373; 1.382 (0.679 to 2.815) takes place and differentiate them of the persistent HIV-con- Time of follow-up 0.328; 0.978 (0.936 to 1.022) trollers. This will enable a correct treatment in transient HIV- HIV transmission risk, PWID 0.149;1.767 (0.816 to 3.827) controllers to avoid non AIDS events, as cardiovascular dis- CD4 T-cell nadir <0.001; 0.995 (0.993 to 0.997) eases [16,20,21] and, on the other hand, the persistent con- Elite HIV-controller status <0.001; 0.136 (0.068 to 0.274) trol phenotype will become a good model of persistent HIV- remission in order to design immunotherapeutic strategies [7]. HCV spontaneous clearance 0.027; 0.377 (0.159 to 0.893) In this manuscript we describe a new characteristic of persis- p values < 0.05 are shown in bold. CI, confidence interval; OR, odd tent HIV-controllers as is a higher ability to spontaneously ratio; PWID, people who inject drug. clear HCV unlike transient controllers. Cohort studies are needed to validate biomarkers associated with the loss of HIV found between this variable and the time to lose HIV-con- control, independently of HCV clearance status, to offer these troller status (p = 0.229; OR (95%) = 0.703 (0.396 to 1.249)). subjects therapeutic options, not necessarily cART [22]. HIV persistent controllers have been recently characterized as having low levels of proviral diversity, low reservoir size 4 | DISCUSSION and high levels of HIV-specific T-cell polyfunctionality [4] asso- ciated with a specific proteomic and metabolomic profile [8,9]. Results presented herein show that persistent HIV-controllers, Many of these subjects were PWID and had multiple re-expo- as opposed to what happened to transient HIV-controllers, sures to HCV during the first years of being positive that had a better ability to spontaneously clear HCV. In fact, the mainly occurred in the 80s and early 90s and anti-VHC and 6 Dominguez-Molina B et al. Journal of the International AIDS Society 2020, 23:e25607 http://onlinelibrary.wiley.com/doi/10.1002/jia2.25607/full | https://doi.org/10.1002/jia2.25607 Figure 1. Kaplan–Meier curves for time to loose HIV-controller status analysis.Percent of subjects not losing HIV controller status are shown for (A) Age dichotomized using the median (29.4 [24.6–35.4]) as the cut-off value, (B) people who inject drugs (PWID), (C) Nadir CD4+ T- cell counts dichotomized using the median (447 [304–622]) as the cut-off value, (D) elite controller, (E) HCV spontaneous clearance and (F) carri- ers of HLA*B57. Bivariate analysis using a Log-rank test was performed to assess significant differences among the curves. Table 6. Factors associated with the time to lose HIV-controller status Univariate Adjusted p; OR (IC 95%) p; OR (IC 95%) Female sex (n) % 0.831; 1.025 (0.816 to 1.288) Age at HIV diagnosis >29.4 yearsa <0.001; 1.929 (1.541 to 2.416) <0.001; 2.301 (1.545 to 3.427) HIV transmission risk, PWID (n) % <0.001; 0.262 (0.206 to 0.334) 0.016; 0.587 (0.380 to 0.906) CD4 T-cell nadir >447 cells/mma <0.001; 0.337 (0.262 to 0.432) 0.003; 0.478 (0.296 to 0.772) Elite HIV-controller status (n) % <0.001; 0.199 (0.151 to 0.263) <0.001; 0.264 (0.172 to 0.406) HCV spontaneous clearance (n) % 0.029; 0.562 (0.335 to 0.943) 0.010; 0.503 (0.297 to 0.850) HLA-B*57 (n) % 0.004; 0.446 (0.256 to 0.779) p values < 0.05 are shown in bold. CI, confidence interval; OR, odd ratio; PWID, people who inject drug. a Patients were categorized according to the median age at diagnosis; 29.4 [24.6 to 35.4]; bpatients were categorized according to the median CD4+ T-cell nadir; 447 [304 to 622]. HCV RNA measurements were later available, if after all this showed higher levels of HCV-specific CD4+ T-cell polyfunc- re-exposures the individual remained with negative ARN for tionality, higher levels of CD8 + CD161 T-cells together with HCV tell us about the preserved ability of those subjects to low T-cell exhaustion associated with preserved innate immu- spontaneously clear the virus. We believe that independently nity features [23]. This sustained immunity is also reflected in the individuals became positive with one virus before the the rates of HCV spontaneous clearance in these subjects other, normally HCV precedes HIV positivity and the potential (25.5%) which is similar to that found in HCV monoinfected re-exposures to HCV, it will be interesting to extent immuno- patients in Spain (22%) [24]. Hence, supercontrollers, who are logical correlates [23] of persistent HIV control and HCV enriched in persistent controllers, contribute with these mech- spontaneous clearance to deeply dissect the intrinsic host anisms of sustained immunity to very long-term persistent mechanisms involved in this extraordinary phenotype. We HIV remission. have recently found that these subjects, able to clear HCV Based on this study, the higher rates of HCV spontaneous and spontaneously control HIV, referred as “supercontrollers,” clearance among HIV-controllers, found previously in small 7 Dominguez-Molina B et al. Journal of the International AIDS Society 2020, 23:e25607 http://onlinelibrary.wiley.com/doi/10.1002/jia2.25607/full | https://doi.org/10.1002/jia2.25607 cohorts [10,11], were mainly due to the contribution of persis- not transient controllers as a good model of functional HIV tent but not transient HIV-controllers. This was true even cure. after adjusting by elite controller status subset that was enriched in persistent controllers and when considering the AUTHORS’ AFFILIATIONS loss of virological control instead of the loss of HIV-controller 1 Enfermedades Infecciosas, Microbiologıa Clınica y Medicina Preventiva. Insti- status. As expected, we had an overrepresentation of HLA- tuto de Biomedicina de Sevilla/Hospital Universitario Virgen del Rocıo/CSIC/ B*57 carriers among HIV-controllers, which has been exten- Universidad de Sevilla, Sevilla, Spain; 2Laboratorio de Inmunovirologia, Instituto sively reported [25–27]. Some authors have previously de Biomedicina de Sevilla (IBiS), Hospital Universitario Virgen del Rocıo/CSIC/ Universidad de Sevilla, Sevilla, Spain; 3Servicio de Urgencias, Hospital Universi- observed higher HCV clearance among patients who carried tario Virgen del Rocıo, Sevilla, Spain; 4Unidad de Infeccion Viral e Inmunidad, HLA-B*57 alleles [28]. Nevertheless, we did not find an asso- Centro Nacional de TOUCH FOR FMicrobiologıa, Instituto de Salud Carlos III, ciation of HLA-B*57 with HCV spontaneous clearance neither Madrid, Spain; 5Centro Sanitario Sandoval, IdlSSC, Madrid, Espa~ na; 6Unidad de in HIV-controllers (persistent or transient, p = 0.2; 2.6 [0.6 to VIH, Hospital Universitario La Paz-IdiPaz, Madrid, Spain; 7Instituto de Investi- gaci on Sanitaria-Fundacion Jimenez Dıaz, Universidad Autonoma de Madrid (IIS- 11.8] and p = 0.6; 1.6 [0.3 to 7.4] respectively) nor HIV non- stoles, Spain; FJD, UAM), Spain; 8Hospital Universitario Rey Juan Carlos, Mo controllers separately (p = 0.3; 1.6 [0.7 to 3.6]). Asher et al., 9 Unidad de Virologia Molecular, Laboratorio de Referencia e Investigacio n en have communicated that other factors besides HLA-B*57, Retrovirus, Centro Nacional de Microbiologıa, Instituto de Salud Carlos III, such as HIV spontaneous control, were associated with HCV Madrid, Spain; 10Departmento de Enfermedades Infecciosas, Hospital Clinic de spontaneous clearance [29] in agreement with our results. Barcelona, Hospital Clinic de Barcelona, Barcelona, Spain; 11Servicio de Medic-  ina Interna, Hospital Viamed, Santa Angela de la Cruz, Seville, Spain In the present study we found no differences in HCV geno- type distribution between persistent and transient HIV-con- COMPETING INTEREST trollers, while we found a different HCV genotype distribution in the overall HIV-controller population vs. non-controllers The authors declare no conflicts of interest exist. [12]. Therefore, the most plausible explanation for this fact is that HIV-controllers spontaneously cleared genotype 1 more AUTHORS’ CONTRIBUTIONS easily than non-controllers, as it occurs in general population ER-M and ML designed and conceived the study. BD-M and ER-M performed [30]. Unfortunately we did not have enough subjects in acute the data analyses. BD-M and ER-M wrote the manuscript. LT-D coordinated the HCV infection in order to know HCV-genotype preferential management of the data from ECRIS cohort. PV, LL-C, YM-G, MG and ML con- tributed with the data management of the comparison cohort. MG, SR, CR, JG, clearance in HIV-controllers. L-GC, RN, BJM, PM, CC, LA, FG, JDR, PV, LL-C and ML contributed with This study describes a new attribute of persistent HIV-con- patient’s supervision and data collection. All the authors critically reviewed, edi- trollers, which is the enhanced capacity in people living with ted and approved the final manuscript. HIV to clear other chronic viral infection as HCV. Persistent controllers, that have special characteristics, in the case of ACKNOWLEDGEMENTS supercontrollers, consisting on an enhance T-cell response This work was supported by the Instituto de Salud Carlos III (research contracts [23], a specific proteomic and metabolomic profile [8,9], low CPII014/00025 to E.R.-M., and FI14/00431 to L.T.-D. and research projects levels of inflammation [4] and very low viral diversity and PI12/02283, PI16/00684, PI19/01127 to E.R.-M.) and Red Tematica de Investi- reservoir, absence of viral evolution [4] make these extraordi- gacion Cooperativa en SIDA (Projects RD12/0017/0029, RD12/0017/0031, and RD16/0025/0020 and RD16/0025/0013), which is included in the Accio n nary subjects an exceptional model of persistent viral remis- Estrategica en Salud, Plan Nacional de Investigacio n Cientıfica, Desarrollo e sion, that overlap largely with the described exceptional elite Innovacion Tecnologica 2008 to 2011 and 2013 to 2016, Instituto de Salud controllers [31–33] or even with a model of HIV eradication. Carlos III, Fondos FEDER. E.R.-M. was supported by Consejerıa de Salud y Bien- In fact, the provocative hypothesis of whether some of these estar Social of Junta de Andalucıa through the Nicolas Monardes program (C- 0032/17), N Rall on is a Miguel Servet investigator from the Spanish Carlos III subjects have been able to eradicate HIV needs to be proven. Institute of Health (ISCIII), grant CP14/00198, Madrid, Spain and B.D.M. We encountered several limitations while performing this received a grant from The Spanish Ministry of Education (FPU13/02451). Work work such as the low number of HLA-B*57 typed patients, in CL-G’s laboratory was supported by grants SAF (2010 to 17226) and (2016- responsible for the loss of statistical power in multivariate 77894-R) from MINECO (Spain) and FIS (PI 13/02269, ISCIII) and in part by analysis and that we were unable to include data on IL28B the RIS-RETIC grants RD06/006/0036 and RD12/0017/0028 funded by the ISC III-FEDER. MP has a contract of RIS-RETIC RD12/0017/0036. This study genotype and liver fibrosis. However, this is the largest HIV- would not have been possible without the collaboration of all the patients, medi- controller cohort analysed in relation to HCV exposition pro- cal and nursery staff, and data managers who have taken part in the ECRIS pro- file and that differentiates between persistent and transient ject. We acknowledge Alicia Gutierrez-Valencia for critical reading of the HIV-controllers that have allowed us to analyse the associa- manuscript. Part of these results has been published at XI GESIDA national meeting as oral presentation. Part of this work has been accepted in CROI2020 tion of spontaneous HCV clearance with persistent HIV-con- as poster presentation. troller status. These analyses were not possible in other published HIV-controller cohorts. REFERENCES 1. Lambotte O, Boufassa F, Madec Y, Nguyen A, Goujard C, Meyer L, et al. HIV 5 | CONCLUSIONS controllers: a homogeneous group of HIV-1 – infected patients with sponta- neous control of viral replication. Clin Infect Dis. 2005;41:1053–6. This study shows an association between spontaneous persis- 2. Chereau F, Madec Y, Sabin C, Obel N, Ruiz-Mateos E, Chrysos G, et al. Impact of CD4 and CD8 dynamics and viral rebounds on loss of virological con- tent HIV-control and HCV spontaneous clearance. The study trol in HIV controllers. PLoS One. 2017;12:e0173893. findings support the idea of preserved immune mechanisms in 3. Leo n A, Perez I, Ruiz-Mateos E, Benito JM, Leal M, Lopez-Galindez C, et al. persistent HIV control implicated in HCV spontaneous clear- Rate and predictors of progression in elite and viremic HIV-1 controllers. AIDS. ance. These results highlight persistent HIV-controllers but 2016;30:1209–20. 8 Dominguez-Molina B et al. Journal of the International AIDS Society 2020, 23:e25607 http://onlinelibrary.wiley.com/doi/10.1002/jia2.25607/full | https://doi.org/10.1002/jia2.25607 4. Pernas M, Taranc on-Diez L, Rodrıguez-Gallego E,G omez J,Prado JG,Casado 21. Li JZ, Segal FP, Bosch RJ, Lalama CM, Roberts-Toler C, Delagreverie H, C, et al. Factors leading to the loss of natural elite control of HIV-1 infection. et al. Antiretroviral therapy reduces T-cell activation and immune exhaustion J Virol. 2018;92:e01805-17. markers in human immunodeficiency virus controllers. Clin Infect Dis. 5. Shasha D, Walker BD. Lessons to be learned from natural control of HIV - 2019;2139:1–7. future directions, therapeutic, and preventive implications. Front Immunol. 22. Ruiz-Mateos E, Poveda E, Lederman MM. Antiretroviral treatment for HIV 2013;4:162. elite controllers? Pathog Immun. 2020;5:121. 6. Deeks SG, Walker BD. Human immunodeficiency virus controllers: mecha- 23. Dominguez-Molina B, Ferrando-Martinez S, Tarancon-Diez L, et al. Immune nisms of durable virus control in the absence of antiretroviral therapy. Immunity. correlates of natural HIV elite control and simultaneous HCV clearance-Super- 2007;27:406–16. controllers. Front Immunol. 2018;9:1–13. 7. Autran B, Descours B, Avettand-Fenoel V, Rouzioux C. Elite controllers as a 24. Martınez-Bauer E, Forns X, Armelles M, Planas R, Sola R, Vergara M, et al. model of functional cure. Curr Opin HIV AIDS. 2011;6:181–187. Hospital admission is a relevant source of hepatitis C virus acquisition in Spain. 8. Rodrıguez-Gallego E, Taranco n-Diez L, Garcia F,del Romero J, Benito JM, J Hepatol. 2008;48:20–7. Alba V, et al. Proteomic profile associated with loss of spontaneous human 25. Pereyra F, Jia X, Mclaren PJ, et al. The major Genetic Determinants of HIV- immunodeficiency virus type 1 elite control. J Infect Dis. 2019;219:867–76. 1 Control Affect HLA Class I Peptide Presentation. Science. 2010;330: 9. Tarancon-Diez L, Rodrıguez-Gallego E, Rull A, Rull A, Peraire J, Vilades C, 1551–7. et al. Immunometabolism is a key factor for the persistent spontaneous elite 26. Fellay J, Shianna KV, Ge D, Colombo S, Ledergerber B, Weale M, et al. A control of HIV-1 infection. EBioMedicine. 2019;42:86–96. whole-genome association study of major determinants for host control of HIV- 10. Sajadi MM, Shakeri N, Talwani R, Redfield RR. Hepatitis C infection in HIV- 1. Science. 2007;317:944–7. 1 natural viral suppressors. AIDS. 2011;24:1689–95. 27. McLaren PJ, Coulonges C, Bartha I, Lenz TL, Deutsch AJ, Bashirova A, 11. Salgado M, Kirk GD, Cox A, Rutebemberwa A, Higgins Y, Astemborski J, et al. Polymorphisms of large effect explain the majority of the host genetic con- et al. Protective interleukin-28B genotype affects hepatitis C virus clearance, tribution to variation of HIV-1 virus load. Proc Natl Acad Sci USA. 2015;112 but does not contribute to HIV-1 control in a cohort of African-American elite (47):14658–63. controllers/suppressors NIH Public Access. AIDS. 2011;25:385–7. 28. Kim AY, Kuntzen T, Timm J, Nolan BE, Baca MA, Reyor LL, et al. Sponta- 12. Ruiz-Mateos E, MacHmach K, Romero-Sanchez MC, Ferrando-Martinez S, neous control of HCV is associated with expression of HLA-B 57 and preserva- Viciana P, Viciana P, et al. Hepatitis C virus replication in caucasian HIV con- tion of targeted epitopes. Gastroenterology. 2011;140(2):686–96.e1 trollers. J Viral Hepat. 2011;18(7):e350–7. 29. Asher AK, Santos G-M, Evans J, Dokubo EK, Lee T-H, Martin JN, et al. 13. Machmach K, Abad-Molina C, Romero-Sanchez MC, Abad MA, Ferrando- Human leukocyte antigen B*57 does not fully explain hepatitis C clearance in Martınez S, Genebat M, et al. IL28B single-nucleotide polymorphism HIV controllers. AIDS. 2013;27:2691–6. rs12979860 is associated with spontaneous HIV control in white subjects. 30. Grebely J, Page K, Sacks-Davis R,van der Loeff MS, Rice TM, Bruneau J, J Infect Dis. 2013;207:651–5. et al. The effects of female sex, viral genotype, and IL28B genotype on sponta- 14. Caro-Murillo AM, Castilla J, Pe rez-Hoyos S, Miro  JM, Podzamczer D, Rubio neous clearance of acute hepatitis C virus infection. Hepatology. 2014;59:109–20. R, et al. Cohorte RIS de pacientes con infecci on por VIH sin tratamiento antir- 31. Casado C, Galvez C, Pernas M, Tarancon-Diez L, Rodriguez C, Sanchez- retroviral previo (CoRIS): metodologıa y primeros resultados. Enferm Infecc Merino V, et al. Permanent control of HIV-1 pathogenesis in exceptional elite Microbiol Clin. 2007;25:23–31. controllers: a model of spontaneous cure. Sci Rep. 2020;10:1–11. 15. Sobrino-Vegas P, Guti errez F, Berenguer J, Labarga P, Garcia F, Alejos-Fer- 32. Canou€ı E, Lecuroux C, Avettand-Feno€ el V, Gousset M, Rouzioux C, Saez- reras B, et al. The Cohort of the Spanish HIV Research Network (CoRIS) and its Cirion A, et al. A subset of extreme human immunodeficiency virus (HIV) con- associated biobank; organizational issues, main findings and losses to follow-up. trollers is characterized by a small hiv blood reservoir and a weak T-cell activa- Enferm Infecc Microbiol Clin. 2011;29:645–53. tion level. Open Forum Infectious Diseases. 2017;4(2):ofx064. [cited 2017 Jul 16. Dominguez-Molina B, Leon A, Rodriguez C, Benito JM, Lopez-Galindez C, 5]. Available from: https://academic.oup.com/ofid/article-lookup/doi/10.1093/of Garcia F, et al. Analysis of non-AIDS-defining events in HIV controllers. Clin id/ofx064 Infect Dis. 2016;62:1304–9. 33. Mendoza D, Johnson SA, Peterson BA, Natarajan V, Salgado M, Dewar RL, 17. Okulicz JF, Marconi VC, Landrum ML, Wegner S, Weintrob A, Ganesan A, et al. Comprehensive analysis of unique cases with extraordinary control over et al. Clinical outcomes of elite controllers, viremic controllers, and long-term HIV replication. Blood. 2012;119:4645–4655. nonprogressors in the Us Department Of Defense HIV Natural History Study. J Infect Dis. 2009;200:1714–23. 18. Noel N, Lerolle N, L ecuroux C,Goujard C, Venet A, Saez-Cirion A, et al. Immunologic and virologic progression in HIV controllers: The role of viral ‘blips’ and immune activation in the ANRS CO21 CODEX study. PLoS One. SUPPORTING INFORMATION 2015;10:1–11. 19. Madec Y, Boufassa F, Porter K, Prins M, Sabin C, Monforte Ad, et al. Natu- Additional information may be found under the Supporting ral history of HIV-control since seroconversion. Aids. 2013;27:2451–2460. 20. Crowell TA, Gebo KA, Blankson JN, Korthuis PT, Yehia BR, Rutstein RM, Information tab for this article. et al. Hospitalization rates and reasons among HIV elite controllers and persons Annex S1. Clinical Centers and research groups which con- with medically controlled HIV infection. J Infect Dis. 2015;211:1692–702. tribute to ECRIS. 9