Background The long average incubation time from HIV infection to AIDS make
s it difficult to estimate recent HIV transmission from AIDS incidence data
. Age-period-cohort (APC) analysis can separate out the effects of age, cal
endar time and birth cohort to provide a clearer picture of transmission tr
ends.
Methods AIDS incidence data from 1981 to 1994 drug users (IDU) for 12 Weste
rn European countries were used. Yearly incidences per 100 000 population o
r 100 000 person-years were calculated by age at diagnosis and 5-year birth
cohort (1950-1954, 1955-1959, 1960-1964, 1965-1969 and 1970-1974), and cor
rected for reporting delay. Incidence patterns were compared between birth
cohorts and countries.
Results For most countries the impact was greatest on the cohort born 1960-
1964. Comparing incidence patterns in the 1965-1969 to 1960-1964 cohorts su
ggest the epidemic has plateaued at low to intermediate levels in Austria,
Greece and the North-Western European countries, and at high levels in Fran
ce, Italy and Switzerland. For most countries transmission amongst the 1970
-1974 as compared to the 1965-1969 cohorts could not be assessed due to sma
ll numbers and short follow-up time. In Spain the epidemic was uncontrolled
with a high incidence among recent birth cohorts. In Portugal the epidemic
was still at an early and expanding phase.
Conclusions The APC analysis revealed large country differences in the dyna
mics of the HIV/ AIDS epidemic among IDU. Full interpretation of these diff
erences is dependent on information from other sources about the local publ
ic health response and trends in drug injecting behaviours. Earlier introdu
ction of the virus and higher prevalence of injecting drug use may explain
some of the generally higher incidence in Southern European countries, but
the larger part of it is most likely explained by local characteristics of
drug users, such as younger age and more frequent sharing of needles and sy
ringes, and a less effective public health response.