Rationale: Reactivation tuberculosis (TB) occurs as a result of reactivation of latent TB infection (LTBI), and was reported to occur in the United States at a rate of 0.10 to 0.16 cases per 100 person-years in the 1950s; it has not been measured since. Objectives: To calculate the rate of reactivation TB in a U.S. community. Methods: A population-based tuberculin skin test survey for LTBI was performed in western Palm Beach County, Florida, from 1998 to 2000 along with a cluster analysis of TB case isolates in the same area from 1997 to 2001. Reactivation (unclustered) TB was presumed to have arisen from the population with LTBI. Measurements and Main Results: The rate of reactivation TB among persons with LTBI without HIV infection was 0.040 cases per 100 person-years (95% confidence interval ½CI, 0.024-0.067) using the n method and 0.058 cases per 100 person-years (95% CI, 0.038-0.089) using the n-1 method. HIV infection was the strongest risk factor for reactivation (rate ratio ½RR, 57; 95% CI, 27-120; P , 0.001). Among persons without HIV infection, reactivation was increased among those older than 50 years (RR, 3.8; 95% CI, 1.3-11) and among those born in the United States (RR, 3.2; 95% CI, 1.1-9.3). Conclusions: Rates of reactivation TB in this area have declined substantially since the 1950s. The greatest part of this decline may be attributed to the disappearance of old, healed TB in the population. If similar declines are seen in other areas of the United States, the cost-effectiveness of screening and treatment of LTBI may be substantially less than previously estimated.

Infection with Mycobacterium avium complex is acquired from the environment, but risk factors for M. avium complex infection and disease are poorly understood. To identify risk factors for infection, the authors performed a 1998-2000 cross-sectional study in western Palm Beach County, Florida, using a population-based random household survey. M. avium complex infection was identified by use of the M. avium sensitin skin test. Of 447 participants, 147 (32.9%) had a positive test reaction, 186 (41.6%) had a negative test reaction, and, for 114 (25.5%), test results were indeterminate. Among the 333 participants with positive or negative M. avium sensitin skin tests, age-adjusted independent predictors of M. avium complex infection in a multivariate model included Black race (odds ratio = 3.8, 95% confidence interval: 2.2, 6.6), birth outside the United States (odds ratio = 2.1, 95% confidence interval: 1.1, 3.9), and more than 6 years' cumulative occupational exposure to soil (odds ratio = 2.7, 95% confidence interval: 1.3, 6.0). Exposure to water, food, or pets was not associated with infection. Results indicate that soil is a reservoir for M. avium complex associated with human infection and that persons whose occupations involve prolonged soil exposure are at increased risk of M. avium complex infection.

In 1986, a population-based survey of human immunodeficiency virus (HIV) infection in a rural Florida community showed that HIV prevalence was 28/877 (3.2%, 95% confidence interval (CI): 2.0, 4.4). In 1998-2000, the authors performed a second population-based survey in this community and a case-control study to determine whether HIV prevalence and risk factors had changed. After 609 addresses had been randomly selected for the survey, 516 (85%) residents were enrolled, and 447 (73%) were tested for HIV. HIV prevalence was 7/447 (1.6%, 95% CI: 0.4, 2.7) in western Palm Beach County and 5/286 (1.7%, 95% CI: 0.2, 3.3) in Belle Glade (p=0.2 in comparison with 1986). Independent predictors of HIV infection in both 1986 and 1998-2000 were having a history of sexually transmitted disease, number of sex partners, and exchanging money or drugs for sex. A history of having sex with men was a risk factor among men in 1986 but not in 1998-2000; residence in specific neighborhoods was a risk factor in 1998-2000 but not in 1986. The authors conclude that heterosexually acquired HIV infection did not spread throughout the community between 1986 and 1998 but persisted at a low level in discrete neighborhoods. Interventions targeting HIV-endemic neighborhoods will be needed to further reduce HIV prevalence in this area.

The Glades Health Survey, a population-based survey of tuberculosis and HIV infection, provides a model for building community±research partnerships with local health departments in ethnically diverse communities. The survey was initiated without broad community participation; a year and a half of organizing established community leadership of the project. Essential factors in the success of the project included a shared objective, direct confrontation of fears about research, inclusion of all socioeconomic and racial/ethnic groups, and community participation in performing the research. These activities led to establishment of a communitybased organization that received funding for HIV counseling and testing and HIV prevention case management.

No abstract