Discussion
Prevalence of active HCV infection among HIV patients in France is globally projected to decrease from 5.09% to 1.08% within the next 10 years under current treatment rates. This decrease should result in approximately 2100 patients with active HCV infection in 2026, 34% of whom will remain outside of the healthcare system. In our model, a decrease of active HCV infection prevalence is expected in almost all risk groups, except for high-risk MSM in whom HCV prevalence would remain almost stable, unless a minimum of 50% treatment coverage rate is reached.
To our knowledge, this is the first study to model HCV epidemic among HIV-infected patients in a whole country and within all risk groups. Since incidence and prevalence data of HIV-HCV coinfection were heterogeneous among the different risk groups in the Dat’AIDS cohort, we considered eight distinct risk groups, including high-risk MSM, who represent 18% of all HIV-monoinfected MSM. This estimated proportion is close to the 25% estimate reported within the Swiss HIV cohort using patients’ unsafe sex reports as reference [
14], while a lower estimate of 7% was reported in the UK Collaborative HIV cohort [
13].
HIV-HCV coinfection was historically associated with intravenous drug use in France and this risk group remained the largest group of patients at the beginning of 2016. However, very few new HCV infections were observed in this risk group, resulting in a drastic decrease in our projections. On the other hand, HCV infection incidence rate in high-risk MSM increased in the cohort since the early 2000s and nearly doubled in this group between 2012 and 2015 [
2]. This trend was also observed in the Swiss cohort [
14,
22], as well as in cohorts from Netherlands [
23] or Japan [
24], with similar estimates, but was not observed in the UK [
13]. The reinfection rate observed in our cohort was similar to a pooled reinfection rate estimate of 32/1000 person-years reported in a meta-analysis [
12], while higher rates were recently reported in the UK (7.8/100 person-years) and within the European NEAT cohort (7.3/100 person-years) [
9,
11]. High-risk MSM could therefore drive an HIV-HCV epidemic over the next years in France and, potentially, in other high-income countries. In order to control such a future epidemic, the targeting of high-risk behavior MSM patients will be crucial.
In our study, we considered that all HIV patients after 6 months of HCV infection were eligible for DAA treatment as recommended in France. However, we also demonstrated a marginal effect of treating high-risk MSM from the third month of infection. Martin et al. [
13] explored the benefit of HCV treatment at both the chronic and acute phases of HCV infection, but defined the acute phase after 1 year of infection. Patients after 6 months of HCV infection are eligible for DAA treatment in France. It is therefore likely that a high treatment coverage for all patients after 6 months of HCV infection would have a similar effect on HCV prevalence than a combined strategy of treating HCV patients during the acute phase (before 6 months) and after 6 months for other patients, but with a lower treatment coverage for the latter. Indeed, in our projections, increasing the acute infection treatment coverage rate to over 70% in high-risk MSM slightly strengthened a decrease in HIV-HCV coinfection prevalence. Increasing HCV treatment coverage after 6 months of infection in high-risk MSM consequently appears as an effective solution to significantly decrease HCV prevalence in this population over the next 10 years.
We used estimates of the undiagnosed HIV population to analyze the impact of HCV treatment on the whole HIV-HCV epidemic in France. Our projections show that the proportion of HIV-HCV undiagnosed patients could increase to 35% in 2026 considering a 30% treatment coverage rate. This proportion could increase to 64% and 79% in 2026 with treatment coverage rates of 50% and 70%, respectively. Although our estimates of coinfection in this undiagnosed population are relatively low, undiagnosed HIV-HCV patients could fuel an HCV epidemic in the future if no specific interventions are undertaken to identify and enroll undiagnosed patients in care.
Our model does have some limitations. First, we considered that no mixing occurred between MSM, heterosexuals, IVDU, and other risk groups. While HCV transmission among heterosexual couples is rare [
25], the source of HCV infection in MSM is often difficult to establish due to concomitant use of intravenous or nasal drugs, and sexual risk behavior [
26]. However, drug use in MSM appears to be mostly driven by consumption during sexual intercourse, and there is no evidence that former opiate users could be a significant source of HCV infection in MSM [
27‐
29].
Second, we modeled HCV transmission among HIV-infected patients only, without considering any other route of transmission such as from the monoinfected HCV population. In a sensitivity analysis, we considered an external force of infection for IVDU to investigate the impact of a potential risk of HCV transmission between HIV-negative and HIV-positive IVDU; the model projections under this hypothesis were similar to our main analysis projections. However, it is likely that HCV transmission between HIV-negative and HIV-positive IVDU could fuel the HCV epidemic in this risk group in other epidemiological contexts, more particularly in countries where HCV incidence and prevalence remain high in this risk group [
30,
31]. Moreover, the observed reinfection rate among IVDU in the Dat'AIDS cohort was lower than the reinfection rate observed in other risk groups. Most of IVDU (90%) included in the cohort were former drug users with a median age of 55 years, which could explain why the observed reinfection rate in this risk group was relatively low. It is potentially likely that HCV reinfection rate in HIV-positive IVDU could be higher in younger, active IVDU, who may not be well represented in the Dat'AIDS cohort compared with other epidemiological context such as in the US.
On the other hand, a recent study from the Netherlands reported that similar HCV strains were circulating among HIV-HCV coinfected MSM and among MSM with high-risk behaviors engaged in a pre-exposure prophylaxis program and acutely infected by HCV [
32]. Since HCV infection incidence is usually lower in HIV-negative than in HIV-positive MSM, this study suggests that HCV infection is now spreading from HIV-positive to HIV-negative MSM. Another recent study from the Netherlands showed a significant decrease in the number of acute HCV infections among HIV positive MSM since the commencement of universal access to DAA treatment for this population [
33]. These results are in favor of a limited HCV transmission from HIV-negative to HIV-positive patients. In any case, a recent increase in unprotected sexual intercourse and sexually transmitted infection incidence in HIV-negative MSM as well as acute HCV infections in MSM enrolled in a pre-exposure prophylaxis program warn us to increase regular HCV screening of all high-risk MSM regardless of HIV infection [
34‐
37].
Third, we estimated the number of undiagnosed HIV-HCV patients in France using the hypothesis of a similar HCV prevalence in this population compared with new patients entering the Dat’AIDS cohort in recent years. Therefore, we added these patients as a constant over time in the projections and no interaction between this population and HIV-diagnosed patients was considered. It is also possible that the size of this population will change in the future due to potential interventions for HIV-diagnosed patients.
Fourth, the model neglected international migration and did not consider the potential risk of HCV transmission related to MSM international networks, as previously described [
38,
39], since no clear data were available to integrate this risk of HCV transmission in our model.
Fifth, we estimated the proportion of HIV-monoinfected high-risk MSM as a constant over time in our main analysis. A specific definition of this population is, to date, not globally approved. We assessed the impact of an increase in the proportion of HIV monoinfected high-risk MSM on HIV-HCV prevalence in the model projections but we did not assess the impact of effective behavioral interventions in this population because no study has thus far proved an effect of such an intervention for HCV infection on the size of the high-risk population.
Sixth, we considered the reinfection rate as a constant force of infection in each risk group (i.e., independent of HCV prevalence), as the proportion of low- and high-risk MSM could not be determined among HIV-HCV coinfected MSM, except at the beginning of the calibration process, i.e. on January 1
st, 2012 (Additional file
1: Appendix). Finally, our projections are promising as most HIV-infected individuals in France are under care and DAA treatment access is universal, i.e., not restricted to a specific fibrosis score and/or comorbidities. These projections are likely to be different if our model was fitted to another country’s settings without universal DAA treatment access and/or with a different spectrum of engagement in HIV care.
Our study also has a number of strengths. The model is based on the largest and most exhaustive HIV-HCV database reported in the literature to date, with yearly data available for all included patients. Moreover, all risk groups were considered in the model for calibration and projections, and reliable HCV risk of transmission for the first infection in each risk group was thus estimated during the calibration by extending the observed data to the whole HIV-diagnosed population in France.
Acknowledgments
The authors gratefully thank Dr. Pierre Pradat and the members of the Dat’AIDS study group: L. Cotte, C. Chidiac, T. Ferry, F. Ader, F. Biron, A. Boibieux, P. Miailhes, T. Perpoint, I. Schlienger, J. Lippmann, E. Braun, J. Koffi, C. Longuet, V. Guéripel, C. Augustin-Normand, C. Brochier, S. Degroodt (Lyon); P. Pugliese, C. Ceppi, E. Cua, J. Cottalorda, J Courjon, P. Dellamonica, E. Demonchy, A De Monte, J. Durant, C. Etienne, S. Ferrando, J.G. Fuzibet, R. Garraffo, A. Joulie, K. Risso, V. Mondain, A. Naqvi, N. Oran, I. Perbost, S. Pillet, B. Prouvost-Keller, S. Wehrlen-Pugliese, E. Rosenthal, S. Sausse, V. Rio, P.M. Roger (Nice); S. Brégigeon, O. Faucher, V. Obry-Roguet, M. Orticoni, M.J. Soavi, P. Geneau de Lamarlière, H Laroche, E. Ressiot, M. Carta, M.J. Ducassou, I. Jacquet, S. Gallie, A. Galinier, A.S. Ritleng, A. Ivanova, C. Blanco-Betancourt, C. Lions, C. Debreux, V. Obry-Roguet, I. Poizot-Martin (Marseille); R. Agher, C. Katlama, M.A. Valantin, C. Duvivier, O. Lortholary, F. Lanternier, C. Charlier, C. Rouzaud, C. Aguilar, B. Henry, D. Lebeaux, G. Cessot, A. Gergely, P.H. Consigny, F. Touam, C. Louisin (Paris); M. Alvarez, N. Biezunski, L. Cuzin, A. Debard, P. Delobel, C. Delpierre, C. Fourcade, B. Marchou, G. Martin-Blondel, M. Porte, M. Mularczyk, D. Garipuy, K. Saune, I. Lepain, M. Marcel, E. Puntis (Toulouse); N. Atoui, M.L. Casanova, V. Faucherre, J.M. Jacquet, V. Le Moing, A. Makinson, C. Merle De Boever, A. Montoya-Ferrer, C. Psomas, J. Reynes (Montpellier); F. Raffi, C. Allavena, E. Billaud, C. Biron, B. Bonnet, S. Bouchez, D. Boutoille, C. Brunet, T. Jovelin, N. Hall, C. Bernaud, P. Morineau, V. Reliquet, O. Aubry, P. Point, M. Besnier, L. Larmet, H. Hüe, S. Pineau, E. André-Garnier, A. Rodallec (Nantes); Ph. Choisy, S. Vandame, Th. Huleux, F. Ajana, I. Alcaraz, V. Baclet, T.H. Huleux, H. Melliez, N. Viget, M. Valette, E. Aissi, Ch. Allienne, A. Meybeck, B. Riff (Tourcoing); F. Bani-Sadr, C. Rouger, J.L. Berger, Y. N’Guyen, D. Lambert, I. Kmiec, M. Hentzien, D. Lebrun, C. Migault (Reims); D. Rey, M.L. Batard, C. Bernard-Henry, C. Cheneau, E. de Mautort, P. Fischer, M. Partisani, M. Priester (Strasbourg); F. Lucht, A. Frésard, E. Botelho-Nevers, A. Gagneux-Brunon, C. Cazorla, C. Guglielminotti, F. Daoud, M.F. Lutz (Saint Etienne); C. Jacomet, H. Laurichesse, O. Lesens, M. Vidal, N. Mrozek, V. Corbin, C. Aumeran, O. Baud, S. Casanova, D. Coban (Clermont Ferrand); L. Hustache-Mathieu, M.C. Thiebaut-Drobacheff, A. Foltzer, V. Gendrin, F. Bozon, C. Chirouze (Besançon); S. Abel, A. Cabié, R. Césaire, G. Dos Santos, L. Fagour, F. Najioullah, M. Ouka, S. Pierre-François, M. Pircher, B. Rozé (Fort-de-France); B. Hoen, R. Ouissa, I. Lamaury (Pointe-à-Pitre).