From January 2004 an outbreak of hepatitis A affecting predominantly MSM occurred in Copenhagen [9]. In April, active case finding in collaboration with laboratories was set up. Awareness of hepatitis A diagnosis and the need for reporting was raised among clinicians. Apart from ordinary precautions in a hepatitis A outbreak, such as increased hygiene and immunoglobulin for close contacts, vaccination was recommended for MSM not living in monogamous relationships. Further information for MSM was provided by a national STI campaign carried out by the Danish gay organisation STOP AIDS and the Danish National Association for Gays and Lesbians. The outbreak continued, and in August 2004, the Statens Serum Institut (SSI) and STOP AIDS carried out a case-control study to determine risk factors for hepatitis A infection in this outbreak in order to inform targeted preventive measures.

Methods
A case was defined as an MSM >17 years, living in Copenhagen, in whom hepatitis A infection with positive anti-HAV IgM was diagnosed between 1 June and 14 August 2004, and who did not have contact to a hepatitis A case in his household in the six weeks before onset of illness. Cases were selected from the notifications received.
Controls, frequency matched to the cases date of onset of illness, were selected at the annual Copenhagen Gay Pride Festival on 14 August 2004 from MSM resident in Copenhagen. Based on the result of a saliva antibody test (Methods: see [10]) only persons susceptible to hepatitis A were included as controls. Data on exposure was collected for a six week period before illness onset and for the same period in controls using piloted self-administered questionnaires. To protect privacy, patients were contacted by their physician who obtained informed consent before posting a questionnaire to them. Information collected included: eating in restaurants/cafes/bars, shellfish consumption, whirlpool use, travel abroad, contact with hepatitis A cases, number of regular and casual sexual partners, venues for sexual contact, sexual contacts abroad or arranged via internet, oral-anal and digital-anal sex practices, history of STIs. Additionally, controls were asked about their attitude towards vaccination.
Data analysis was performed in STATA 8.0. Matched odds ratios and 95% confidence intervals were calculated for each exposure factor. Adjusted odds ratios were calculated using conditional logistic regression analysis. Exposure factors with P values <0.20 and confounding variables such as age were included. The final model was build by backward elimination of variables above the threshold of P= 0.10.

Results
In 2004, 163 cases of hepatitis A in men >17 years were notified to the Department of Epidemiology, SSI. In the past five years the median number of annually reported cases of hepatitis A among men of this age group was 13 (range 7-25). Of the 163 cases, 107 were from Copenhagen, 56 from the rest of Denmark. The incidence rate in Copenhagen was 23 per 100 000 and declined with increasing distance from the capital [FIGURE 1].

The following results are restricted to cases in Copenhagen. Of the 107 patients, 68 (64%) were reported to be MSM and five to be heterosexuals. For 34 patients, the sexual orientation was not known [FIGURE 2].

Patient ages ranged between 19 and 73 years with a median of 41. Ninety seven (91%) patients were residents in Denmark. Thirty seven (35%) patients were admitted to hospital.. Forty nine (46%) cases were reported by general practitioners and 58 from hospital in- or outpatient departments.

Case-control study
The case-control study conducted among MSM included 18 cases and 64 controls. Physicians of 36 notified cases were asked to recruit their patients for the study; 30 patients agreed to participate and 24 of these (80%) returned questionnaires. Six patients did not fulfil the case definition (four self-identified as heterosexual, one was infected by a household contact and one did not live in Denmark during the exposure period). Saliva samples were taken from 105 MSM visiting the Copenhagen Pride Festival: 86 (82%) had no detectable antibodies against HAV; 15 (14%) were IgG positive, two of these were also IgM positive; four samples were inconclusive. Of the 86 without detectable antibodies, 17 reported having been vaccinated against hepatitis A, three were not in Copenhagen during the period required and two refused to participate. Therefore, 64 (61%) of the participants were included as controls.

Case patients and controls were similar with regards to residence within Copenhagen, but patients were older than controls [TABLE 1].The proportion of HIV-infected people was higher among patients than controls. Neither patients nor controls reported a previous syphilis infection.

Matched univariate analysis of exposure factors [TABLE 2] suggests that HAV infection was not associated with consumption of seafood or eating outside home. There was no evidence of a cluster of cases linked specific food outlets or restaurants. Travelling abroad was less frequent for case patients than controls. Patients had a median of three sexual partners in the six weeks before illness and controls, a median of two.

Patients were more likely to have had sex with casual partners than controls. One third of both patients and controls had sex with partners they met via the internet. HAV infection was associated with sex in gay saunas. This association was very strong in May and June, when 9 of 10 patients were exposed (OR_MH 129.2, 95% CI 7.6-2197.5), but not in the later part of the study. No single sauna was implicated. Sex at cruising grounds and toilets was not associated with infection. Sex at private homes appeared protective. Participation in group sex was reported by 17% of the patients and 5% of controls. High risk practices such as oral-anal and digital-anal sex were common among both cases and controls, and were not associated with increased risk of HAV infection.
After adjusting for confounding, sex with casual partners (adjusted odds ratio(aOR) 8.7; 95% CI 1.6-48.9) and sex in private homes (aOR 0.1; 95% CI 0.0-0.5) seem to respectively increase and decrease the risk of infection. Because of the time dependency, sex in gay saunas did not prevail as an independent risk factor in the multivariate analysis.

Two thirds of controls disclosed their sexual orientation to their general practitioner. Of 36 controls, who had had casual sex, 11 had been recommended hepatitis A vaccination (gay campaigns (5), an STI clinic (4), and general practitioners (2)). Among controls, 53 (83%) were willing to be vaccinated against hepatitis A. However, of these, only 14 (26%) were willing to pay for the vaccination.

Discussion
The study suggests that sexual activity was the major mode of transmission in this hepatitis A outbreak among MSM. Sex with casual partners and sex in gay saunas contributed to the spread of hepatitis A among MSM.
However, as the study period encompassed only part of the outbreak, we must be cautious about extrapolating results to the whole period of the outbreak. The study may be further limited by a small sample size. Controls selected at the festival may not be representative of the total MSM population in Copenhagen. However, since it is impossible to select controls directly from the study population, we consider our approach the best possible way to represent a broad spectrum of MSM. Protecting privacy was considered to be extremely important in this outbreak investigation, and therefore access to information on the cases was limited due to the sensitive nature of the subject.

This was the largest recorded outbreak of hepatitis among MSM in Denmark and the first one in more than a decade. In Europe, cases of hepatitis A among MSM are reported with increasing frequency: since 1995 there have been outbreaks in major European cities on an almost annual basis. Venues for casual sex, such as gay saunas and darkrooms are frequently implicated [7,8,11,12]. The increased risk of infection is presumably related to the possibility of having several partners within a short period of time. In Copenhagen, gay saunas are popular places for both Danes and visitors from abroad to have casual sex. There are at least seven saunas, they operate all year round and the largest sauna typically attracts approximately 700 to 1000 visitors per week. With increasing tourist traffic between European cities and increasing susceptibility to HAV in the population, gay saunas offer a perfect venue for endemic and epidemic spread of hepatitis A among MSM in Europe. The Copenhagen outbreak caused hepatitis A in at least 13 Swedish men [13]. Most of these men were from southern parts of Sweden, which is situated close to Copenhagen, and in some of them, infection could be directly linked to sex at gay saunas. The outbreak in Copenhagen is most likely also responsible for an outbreak among MSM in Norway. There, the same outbreak virus strain circulated among infected MSM that caused the Copenhagen and Swedish cases [14].

Hepatitis A is a relatively mild disease with a low case fatality. No death was recorded in Copenhagen. However, one third of cases were admitted to hospital. Costs associated with hospital admission and days off work could be avoided by vaccination. In HIV-infected individuals, HAV infection has been associated with prolonged HAV viraemia, which might lead to longer infectivity and increased risk of spread in this population [15]. Inactivated HAV vaccine is safe in HIV-infected individuals [16].

The prevalence of anti-HAV antibodies among MSM tested at the Copenhagen Pride Festival was 14%. This is low compared with a serological study in two gay saunas in Copenhagen conducted in 1984, where 36% of sauna attendees were immune to HAV – a figure three times higher than in the general Danish population at that time [17]. It is uncertain whether this reflects an overall decline of hepatitis A infections among MSM over the last 20 years, because sauna attendees may not be representative of the population of MSM attending the festival. However, it suggests that the population of MSM in Copenhagen is susceptible to hepatitis A infection and therefore need to be alerted of the risks of infection and how to prevent it.

Based on the results of the investigation we suggest recommending hepatitis A vaccination to all MSM who are not in a monogamous relationship, especially if they visit gay saunas or other places with frequent partner change.
Opportunities for vaccination could be visits to general practitioners (although not all MSM disclose their sexual orientation to their doctor), sexual health clinics or outreach campaigns at saunas or mobile clinics. Willingness to be vaccinated was high among MSM, but a considerable number were reluctant to pay for the vaccination. This attitude may be influenced by information about the importance of vaccination. As free hepatitis B vaccination is available for MSM in Copenhagen, exchanging the monovalent vaccine for the combined hepatitis A and B vaccine would make protection against hepatitis A available at little extra costs.
We further suggest that adequate hygiene should be ensured in saunas. An information campaign on risks and prevention of hepatitis A transmission should be targeted at sauna visitors (both Danish and international guests). To stop spread of hepatitis A among MSM in Europe, a European consensus on prevention and control measures may be required.

Acknowledgements
We thank all physicians for their cooperation in recruiting their patients for the study and the volunteers who helped with control selection at the Pride Festival. We thank John V Parry from the Sexually-Transmitted & Bloodborne Virus Laboratory, Health Protection Agency, London, for analysing the oral fluid samples and for his valuable support in interpreting the results.