To investigate evidence for resistance to HIV-1 infection associated with the heterozygous genotype CCR5-+/Delta32 and with the homozygous genotype CCR5-Delta32/Delta32, which results in a nonfunctional CCR5 receptor.Cohort study of initially HIV-seronegative high-risk individuals from eight different cities. Enrollment data were analyzed to investigate the association of demographic factors and risk behaviors with CCR5 genotypes on the assumption that increased genotype prevalence among persons with histories of longer or more intensive exposure to HIV would indicate HIV resistance associated with that genotype. Longitudinal data were analyzed to investigate the association of HIV seroincidence with CCR5 genotypes. The cohort of 2996 individuals included 1892 men who have sex with men (MSM), 474 male injection drug users (IDUs), 347 women at heterosexual risk, and 283 female IDUs.CCR5 genotype, HIV serostatus, demographic factors, and risk behaviors during the 6 months before enrollment, followed by measurement of HIV seroincidence during the subsequent 18 months (for men) and 24 months (for women).Forty (1.3%) subjects were homozygous CCR5-Delta32/Delta32 and 387 (12.9%) were heterozygous CCR5-+/Delta32. All but 1 CCR5-Delta32/Delta32 individuals and 51 CCR5-+/Delta32 individuals were Caucasian. Among 1531 Caucasian MSM, CCR5-+/Delta32 individuals were present more frequently (22.3%) among those reporting unprotected receptive anal intercourse than among those not reporting this risk (15.9%) (p =.002), suggesting a selective advantage of the heterozygous genotype. CCR5-+/Delta32 individuals also had a significantly reduced relative risk of HIV seroconversion adjusted for unprotected receptive anal intercourse compared with CCR5-/+ individuals (relative risk = 0.30, 95% confidence interval [CI]: 0.08-0.97). CCR5-Delta32/Delta32 prevalence among Caucasian MSM was significantly associated with age among subjects recruited from high HIV seroprevalence cities (New York City and San Francisco) (odds ratio [OR] for each decade increase in age = 2.57, CI: 1.56-4.21) but not among those recruited from lower HIV prevalence sites (Boston, Chicago, Philadelphia, Seattle, and Providence/Pawtucket, Rhode Island) (OR = 1.20, CI: 0.75-1.89).Cross-sectional and longitudinal analyses indicated that among high-risk HIV seronegative MSM, CCR5-+/Delta32 and CCR5-Delta32/Delta32 are associated with protection against HIV infection. These findings imply that strategies aimed at reducing susceptibility to HIV infection by blocking CCR5 receptor sites need not seek blockage of all receptor sites to achieve an imperfect but substantial degree of protection.