Faculty Bibliography

OBJECTIVES/OBJECTIVE:This study assessed trends in HIV risk behaviors among injection drug users in New York City from 1990 to 1997. METHODS:Injection drug users were recruited continuously from a large drug detoxification treatment program (N = 2588) and a research storefront located in a high-drug-use area (N = 2701). Informed consent was obtained, and a trained interviewer administered a structured interview covering sociodemographics, drug use history, HIV risk behavior, and participation in syringe exchange. RESULTS:Trends were assessed for 5 risk behaviors in the 6-month period before the interview. The 3 injection risk behaviors declined significantly over time at each site (all P < .01). When data were pooled across sites, all 5 risk behaviors declined significantly over time (all P < .01). Participation in syringe exchange programs and in HIV counseling and testing increased greatly from 1990 to 1997. CONCLUSIONS:The continuing risk reduction among injection drug users indicates a "declining phase" in the large HIV epidemic in New York City. HIV prevention programs appear to be making an important contribution to the declining phase.

In many cities, human immunodeficiency virus (HIV)-1 seroprevalence among drug injectors stabilizes at 30-70% for many years without secondary outbreaks that increase seroprevalence by 15% or more. The authors considered how HIV-1 incidence can remain moderate at seroprevalence levels that would give maximum incidence. Previously suggested answers include behavioral risk reduction and network saturation within high-risk subgroups. Among 767 drug injectors studied in 1991-1993, during a period of stable high seroprevalence in New York City, risk behaviors remained common, and networks were far from saturated. The authors suggest a different network-based mechanism: in stable high-prevalence situations, the relatively small sizes of subnetworks of linked seronegatives (within larger networks containing both infected and uninfected persons) may limit infectious outbreaks. Any primary infection outbreak would probably be limited to members of connected subcomponents of seronegatives, and the largest such subcomponent in the study contained only 18 members (of 415 seronegatives). Research and mathematical modeling should study conditions that may affect the size and stability of subcomponents of seronegatives. Finally, if the existence of small, connected components of seronegatives prevents secondary outbreaks, this protection may weaken, and vulnerability to new outbreaks increase, if HIV-1 seroprevalence falls. Thus, in situations of declining prevalence, prevention programs should be maintained or strengthened.