Method
We defined a case as a HIV+ MSM with acute hepatitis C occurring since 1 January 2001. Acute hepatitis C was defined as a documented HCV seroconversion within 6 months, or a positive HCV-RNA by polymerase chain reaction assay (PCR) following a negative assay within 6 months previously, or a positive HCV-RNA PCR and ?10-fold the normal limits of serum alanine aminotransferase level (ALT) with documented normal levels during the preceding year.
The three Parisian hospitals wards (two infectious disease wards and one hepatology ward) initially involved were asked to find all cases of acute hepatitis C that fitted the case definition. We reviewed the medical records of all identified cases to assess possible risk factors for HCV infection during the 6 months prior to acute hepatitis C, including patient demographics, HIV infection characteristics (date of diagnosis, group of transmission, clinical stage, CD4 count and viral load level, antiretroviral therapy), hepatitis C infection description (date of diagnosis, reasons for HCV screening, viral genotype, treatment, progression), sexual behaviour and risk factors for HCV transmission (IDU, professional or nosocomial exposure). An anonymous self-administered questionnaire on sexual behaviour was offered to cases at the next follow-up visit: it listed sexual practices including hardcore gay practices (e.g., fisting), number of sexual partners, attending gay venues and drug use, in the 6 months prior to acute hepatitis C onset. The study questionnaires were notified to, and approved by, the national commission for data protection.

Results
We found 29 acute hepatitis C cases that occurred between March 2001 and October 2004: two cases in 2001, 10 in 2002, nine in 2003 and eight in 2004 [FIGURE]. The median age at hepatitis C onset was 40 years (range: 28-54). The reasons identified for HCV testing were an ALT level increase (24), acute jaundice (3) or an STI (1). HCV seroconversion within 6 months was documented for 16 patients. A positive HCV RNA-PCR and > or =10-fold normal limits of ALT with normal levels for the preceding year were documented for 13 patients with a median time of HCV seroconversion of 19 months [range : 7 - 65]. Genotype 4 was detected in 15 (52%) patients, genotype 3 in seven (24%) and genotype 1 in six (22%); for one case, no genotype was available.

None of the common risk factors for HCV transmission (IDU, professional or nosocomial exposure) were found. All patients had had unprotected anal sex and six reported ‘hard” sexual practices, with two resulting anal perforations. A STI co-infection was recorded for 12 (41%) patients and specified for nine: syphilis (7), gonorrhoea (1) and genital herpes (1).
The median time between HIV and hepatitis C diagnosis was 6.5 years (interquartile range: 4-14 years). Two patients had HIV primary infection within 3 months prior to hepatitis C onset. Median age at HIV diagnosis was 29 years (range: 20-50). At the time of acute hepatitis C, 22 (76%) patients were clinically asymptomatic for HIV infection, and four (14%) had AIDS (HIV clinical stage unspecified for three). Most of the patients (86%) were taking antiretroviral therapy and four were untreated. All patients had more than 200/mm3 CD4 and 16 (55%) more than 500/mm3. HIV viral load was undetectable for 19 (65%) patients, detectable but <104 copies/ml for six (21%) and >104 copies/ml for four (14%).
The 11 anonymous self-administered questionnaires respondents had had unprotected anal sex with casual partners within the six months prior to hepatitis C onset, nine had ?10 partners and six had unprotected anal sex one or more times per week. All respondents met their casual sex partners at commercial gay venues (sex-on-site) or via the internet. A STI was reported by all but one, including syphilis (6), gonorrhoea (3), genital herpes (3), Chlamydia (3) and warts (anal or genital) infections (3). Eight patients reported “hard” sexual practices, including fisting for five. Bleeding during sex was reported by six patients. All patients inhaled poppers (nitrite inhalants) during sex, none used intravenous drugs and five reported consuming psychoactive drugs (ecstasy, cocaine, gamma hydroxybutyrate (GHB), ketamine or LSD). Tattoos or piercings within the six months prior to hepatitis C onset were reported by three patients.

Discussion
We retrospectively identified 29 cases of acute hepatitis C occurred between March 2001 and October 2004 in HIV+ MSM from the medical records of three major infectious diseases or hepatology wards. A STI was simultaneously present in 41% of the cases. All patients had high-risk sexual behaviour for transmitting HIV (unprotected sex with multiple casual partners). Nearly a third reported having had “hard” sexual practices with their casual partners. Most patients were being treated for a long-time HIV infection that was immunologically and virologically well-controlled.
The report of these acute hepatitis C cases could be linked to an increase in routine HCV testing in HIV+ patients as recommended in 2002 [2] in the national management guidelines for HIV infection. This is unlikely, however, since most of the patients had been tested for HCV because of an increase in ALT levels. Furthermore, an increase in the incidence of HCV seroconversion in HIV+ MSM has also been reported since 2000 in Switzerland [3] and the United Kingdom [4].
HCV sexual transmission is extremely rare among monogamous heterosexual couples, and the incidence is low (0.37‰ person-years in an Italian study [5]). HCV prevalence in genitourinary medicine clinic attendees is low in those who do not report IDU (0.65%; 95% CI: 0.51-0.78 [6]). Furthermore, the low estimated incidence of hepatitis C in cohort studies of MSM [7,8] suggests that hepatitis C is not readily spread by sex between men. However, this route of transmission may be facilitated by HIV infection [9,10], which may promote viral receptivity in individuals sexually exposed to HCV and enhance HCV infectivity in genital secretions.

High risk sexual behaviour for STI transmission (multiple sexual partners), sexual practices (unprotected anal sex, rimming, fisting) and STI have been discussed [10,11] and are still debated as factors that may facilitate HCV sexual transmission [10,12]. In our study, unprotected anal sex was recorded for all patients, and fisting for 21%. Bleeding (visible or not) during sex, which was mentioned in half of the self-administered questionnaires, may be a likely route for HCV transmission in these cases. Furthermore, a co-infection with an STI, which has been described as potential co-factor for HCV transmission [10-12], may have facilitated HCV infection: co-infection with an STI and hepatitis C was documented in 41% of the medical records and most of the respondents to the questionnaire reported a STI within 6 months prior to HCV infection. These observations are consistent with the recent upsurge in recent syphilis infections [13], the lymphogranuloma venereum outbreaks in MSM [12,14] and the increase of high-risk sexual behaviours described among French MSM in 2000 and 2002 [15,16].
We documented no intravenous drug use (IDU) in any of the cases, but this risk factor could have been under-reported. However, the respondents to the self-administered questionnaire mentioned neither IDU nor heroin use, but did mention use of other drugs or psychoactive substances. Our study suggests that HIV+ MSM who acquired HCV may belong to a specific group of men who are engaged in sensation seeking and sexual experiments including, “hard” sex [17] with high-risk sexual practices (multiple partners, no condom for anal sex, fisting without protective gloves.) which result in multiple STIs. In this group, psychoactive substance use may be used to facilitate “hard” sexual practices by lowering inhibitions [18] and therefore could favour bleeding.

The proportion of genotype 4 (52%) among this group is much greater than usually observed in France (nearly 11% [19]). This finding may indicate that the emergence of acute hepatitis in HIV+ MSM living in the Paris area could be related to a transmission of selected strains in a social network of people with specific behaviours. Phylogenetic studies are necessary and are underway [20] to further document this hypothesis.
Because our study was limited to three of the 12 hospital wards that provide care to HIV+ patients in the Paris area, the number of acute HCV infections in HIV+ MSM is likely to be underestimated. Although these hospital wards give treat a large number of HIV+ MSM, we cannot extrapolate our findings to all cases that occurred during the same time period. Furthermore, the sexual behavioural self-administered questionnaires were completed by only 38% of the cases, and so we cannot use our findings to make generalisations about all HIV+ MSM with acute HCV infection. A case-control study would have been useful to better document risk factors for acquiring HCV infection among HIV+ MSM. However, we chose to do an exploratory descriptive study first, to document the phenomenon and propose hypotheses in a timely fashion that may be evaluated further by analytical epidemiological or qualitative studies. This descriptive study also allows the timely informing of all clinicians involved in HIV patient care and gay advocacy groups.

We conclude that HCV transmission can occur through bleeding during traumatic unprotected anal intercourse and “hard” practices which may be facilitated by mucosal lesions linked to STIs. In the context of the recent resurgence of STIs and the drop in infection prevention practices by MSM, this report stresses the urgent need for continuing strong advocacy of safer sex among MSM. There is also a need for social and behavioural studies on sexual practices that enhance HCV transmission in gay men.

Acknowledgements
We acknowledge Corinne Pioche for her excellent data assistance.