By the end of 2005, the number of people living with HIV (PLWHIV) in Germany
was estimated at about 49 000, among a total population of 82.5 Mio.
Around two thirds of all PLWHIV in Germany are estimated to be men who
have sex with men (MSM) . Since 2001, there has been a continuous
increase in the number of newly diagnosed HIV infections in MSM. Over
the same time period, similar increases of HIV diagnoses in MSM have
been reported from the majority of western European countries in which
HIV surveillance systems are established. These simultaneous increases
of newly diagnosed HIV infections were preceded, accompanied or followed
by increasing incidences of other sexually transmitted diseases in MSM,
such as syphilis, gonorrhea and lymphogranuloma venereum [2,3].
Among the first consequences of the growing awareness of the HIV epidemic
in gay communities in the early to mid-1980s was a dramatic reduction
of partner numbers, followed in the late 1980s and early 1990s by the
increasing use of condoms within the context of casual sexual encounters.
This decrease in sexual risk-taking resulted in reductions of transmission
events and a shift of HIV transmission in MSM from casual sexual encounters
towards more long-term partnerships, because internal relationship
dynamics within such partnerships favour unsafe sexual behaviour [FIGURE
It is clear that something changed in the second half of the 1990s,
because the incidence of sexually transmitted infections in MSM began
to increase again, not just in selected countries or regions, but in
all countries of the developed and developing world (such as Thailand
and Brazil) that had large, organised and visible MSM communities.
Based on the findings of national and international behavioural surveillance
data in MSM from recent years, and on the HIV surveillance data from
Germany, we propose a model for the HIV epidemic in MSM in Germany that
could explain the trends observed and identify relevant areas for future
epidemiological and behavioural research.
The construction of a model for the MSM HIV epidemic in Germany is based
on a review of national and international behavioural surveillance
studies of MSM and HIV surveillance data.
The German HIV surveillance system has been described in detail elsewhere
; briefly, newly diagnosed HIV infections must be reported by laboratories
with complementing patient history and clinical data provided by the
primary care physician on a duplicate of the laboratory reporting form.
For Germany, the main data source for sexual behaviour changes in MSM
are the national surveillance studies that have been performed every
2-3 years since 1987 .
Behavioural surveillance studies of MSM in western developed countries
were identified by a Medline search using the search criteria ‘MSM’ and ‘sexual
risk behaviour’, including studies published since 1998 from western
Europe, Australia and North America.
Behaviour changes in MSM
Recent studies on sexual risk behaviour of MSM have described a range
of changes in sexual risk-taking behaviours in MSM in recent years. Many
MSM report increased numbers of partners, there has been an apparent
revival of anogenital practices and an increase in the proportion of
unprotected episodes of anogenital intercourse. Other behavioural features
are frequent HIV testing (in a recent internet-based survey in which
over 45 000 MSM from Germany participated, 70% reported that they had
been tested for HIV at least once, and 53% reported having had between
two and five tests ), HIV ‘sero-sorting’, and HIV ‘sero-positioning’ ,
and the growing importance of the internet for seeking and selecting
sexual partners [3,5-12].
Number of sexual partners
German behavioural surveillance of MSM has found that MSM have been having
increasing numbers of sexual partners since the early 1990s, after
a short period during which MSM had decreasing numbers of partners,
from approximately 1984 until 1989 .
While there was no clear evidence from the HIV surveillance system that
increasing numbers of partners in MSM resulted in accelerated HIV transmission
in subsequent years, the incidence of some sexually transmitted infections,
which are more easily transmissible than HIV, may already have increased
during this period, despite widespread condom use, because of transmission
through sexual practices, which were usually performed without protection
(e.g. genital-oral sex: no data available for Germany, but see ).
Increase of unsafe sex
Sexual behaviour surveys in MSM in Germany  have repeatedly found
evidence of increasing unsafe sexual contact in MSM since 1996, after
a decade of continuously declining occurrence of unsafe sexual behaviour.
The proportion of men reporting unsafe anal intercourse with partners
of unknown HIV status, and the reported number of partners with whom
unsafe sex is practised, have been increasing since 1996 [FIGURE 2].
Urban-rural differences, the internet, and the formation
of high risk networks
The wish to fulfil sexual desires, unconstrained by social and family
supervision, results in a concentration of MSM in large cities, where
gay bars, discotheques, bath houses, and a wide range of gay community
organisations make it easier for MSM to seek and find sexual partners.
Therefore, MSM living in metropolitan areas usually report higher numbers
of sexual partners and higher incidences of sexually transmitted infections
than MSM living in rural areas. Since the late 1990s, the increasing
availability of the internet has led to more opportunities to seek and
meet new sexual partners. The decreasing gap between MSM residing in
metropolitan and rural areas in terms of partner numbers and (self-reported)
STI incidences observed in the German behavioural surveys (data not shown)
probably reflects the increased use and availability of the internet.
However, the internet not only makes the search for new partners easier,
it also allows for a more effective selection of partners, based on sexual
preferences, HIV status, and also on the willingness to practise unsafe
sex. This results in the formation of sexual networks suitable for fast
and efficient spread of STIs [DIAGRAM] [14-16].
Improvement of antiretroviral therapy and changing
Since the availability of highly active antiretroviral therapies (HAART)
from 1996, the number of people being treated with HAART (calculated
based on antiretroviral drug sales) increased rapidly until 1998, and
then levelled off, when problems associated with long term toxicities
of HAART began to emerge in 1999/2000. HAART is nowadays initiated later
in the course of HIV infection. The numbers of people living with HIV,
but being not treated with HAART, decreased sharply from 1996 to 1999,
but has since begun to increase slowly [FIGURE 3].
After 1996, HIV transmission risks were reduced within stable partnerships
between serodiscordant partners (where one partner is infected with
HIV, and the other partner is not infected) by widespread use of highly
active antiretroviral therapies (HAART), because of the reduced infectivity
of those being treated. However, prolonged survival and improved quality
of life for HIV-infected people resulted in a growing number of people
living with HIV, and an increase in their sexual activity. Since the
late 1990s, the change of HIV treatment strategies has led to treatment
interruptions and delayed initiation of antiretroviral treatment in
people newly diagnosed with HIV [DIAGRAM].
Therefore, by the early 2000s, increasing risk behaviour and increased
transmission risk were no longer being compensated for by the effects
of antiretroviral therapy. New infections with HIV increased in frequency
in the partners of chronically HIV-infected people, and mostly in the
age group in which most sexually active HIV-infected men are found
(35-40 years). However, an increase in risk-taking behaviour by uninfected
MSM and the formation of subpopulations of MSM with high partner numbers
and a high prevalence of unsafe sexual behaviour introduced new opportunities
for HIV to spread quickly and efficiently within high risk sexual networks,
where high incidences of other STIs further enhanced the spread 
[DIAGRAM]. These high risk networks are found within and across all
age groups of MSM between the ages of 25 and 45-50. The diffusion of
HIV from the core age group of 35-40 into the neighbouring age groups
via sexual networks occurs with some smaller time delays, which are
reflected in the successive increase of HIV diagnoses incidence in
different age groups [FIGURE 4].
The increase in number of partners in the early 1990s probably had few
consequences for HIV incidence, because during that time period, an overwhelming
proportion of penetrative sexual contacts within the context of casual
sex were protected by condom use. A few exceptions may have permitted
isolated infection clusters, but these did not spread extensively, because
of widespread condom use. During this period, a comparatively high proportion
of HIV transmission in MSM occurred within committed, short or long-term
partnerships, and therefore increased emphasis on HIV testing was a highly
appropriate part of prevention strategies.
Risk minimisation strategies based on better knowledge or assumptions
on HIV status developed during the 1990s and were appropriate for the
above described epidemiological situation, but they tended to fail in
the situation that developed after 1996, in which an increasing proportion
of new infections are transmitted during acute HIV infection, and outside
of stable partnerships. Emphasising HIV testing as a cornerstone of HIV
prevention in the current situation in the mentioned high risk subgroups
of MSM may even be counterproductive if it encourages men with a high
number of partners and frequent practice of unsafe sex to be overly confident
of a negative HIV status and to abandon condom use .
While surveillance of newly diagnosed HIV infections allows for a more
up to date monitoring of the HIV epidemic than AIDS case surveillance,
it cannot provide accurate information on HIV incidence. The challenges
are to apply methods to determine HIV incidence and to monitor risk
behaviour to detect changes in risk behaviours early enough to modify
prevention messages and prevention strategies before these changes
have a large impact on HIV incidence.
Our goal therefore is to build a system of second generation surveillance
in Germany. Currently, the following elements of such a system are
being implemented or planned:
1) A pilot study to establish a surveillance system and laboratory
assays to define the proportion of recent infections among newly diagnosed
HIV infections (ongoing).
2) Extension of behavioural surveillance of MSM. In addition to the repeated
behavioural surveillance studies that have been conducted in the past
and that focused on HIV-related risks, a new survey on MSM’s knowledge,
attitudes and behaviours regarding STIs has been executed in 2006. In
a pilot sub-study within this survey we tested the feasibility of self-collected
blood filter samples, with the aim of connecting behavioural and serological
data (results are currently analyzed).
3) Qualitative research on risk factors of incident HIV infections and
motives for unsafe sex in newly diagnosed HIV-infected MSM (ongoing).
4) A study on partner-seeking and risk communication by internet users
(results are currently analyzed).
While we are aware that not all aspects of our proposed model are
strongly corroborated by epidemiological and behavioural data, we think
the model represents a reasonable interpretation of epidemiological
trends based on available data, and as such, may be helpful for generating
research questions for further studies.