The few statistics illustrating the effect of condoms use on HIV
transmission in
homosexual acts do not give clear cut results.
R.Detels (1989)lxxxix observed a 2.9% drop in seroconversion when condoms
were used in homosexual intercourse, in a cohort of 2915 active
homosexuals initally
HIV sero-negative. But, in that statistic, the number of partners for each
member of the
cohort was almost as important a factor as the use or not of a condom.
L.Levin et al. (1995)xc found that most (71%) of the 140 active-duty young
men
with documented sero-conversion who were identified in various US army
installations
did use a condom during their homosexual or heterosexual relations.

Results seemed to indicate a paradoxal increased risk of seroconversion
proportional to condom use in
this series.