The U.S. Public Health Service trials of isoniazid treatment of latent tuberculous infection provide useful 10-year follow-up information for the untreated, control groups (82). A variety of studies were conducted, of which those with household contacts and inmates of mental hospitals are probably the most relevant to health care workers. Overall, tuberculous infection marked by a PPD reaction of ≥10 millimeters occurred in 2.9 percent of infected household contacts and 1.2 percent of infected mental hospital inmates. Approximately one-third of cases developed during the first year of observation. Attack rates were higher in adults (essentially at the overall levels cited above) than children, and at least in the household contact group, the adult rate did not change through age 55.
A small study of a shipboard outbreak in the Dutch navy was cited by Ferrebee in her review (82). Tuberculosis developed in 12 of 128 seamen (9.4 percent) not given isoniazid in a trial of the effectiveness of this therapy, a figure much higher than that reported in any other investigation. This study, although small, is of interest because it reflects results in employed individuals. Shipboard exposures have been found to be more intense and have higher attack rates than those in other situations, and the population was probably skewed toward the young-adult age group that has the highest risk.
In an editorial dealing with the use of isoniazid for the treatment of latent tuberculous infections in young adults, George Comstock and Phyllis Edwards used published and unpublished data from both BCG trials and isoniazid chemotherapy trials to estimate the lifetime risk of tuberculosis among tuberculin skin test reactors (93). They noted that the risk declined with passing years. Lumping together their estimates of lifetime tuberculosis risks for tuberculin-positive black and white males and females, their estimates were approximately 3.5 to 4.5 percent at age 25, 3.1 to 3.6 percent at age 35, 2.6 to 3.0 percent at age 45, 2.0 to 2.5 percent at age 55, and 1.2 to 1.6 percent at age 65.
HIV infection, even before the onset of frank AIDS, increases the risk of tuberculosis in infected individuals. In fact, tuberculosis is one of the major intercurrent infections of dually infected persons, and tuberculosis often occurs at a time when immune function is relatively well preserved.
Setting aside some excellent studies done in Africa as perhaps not applicable to American health care workers, the 2-year prospective study of Selwyn and colleagues conducted in a New York City methadone clinic in 1985–1987 provided what has been widely quoted and generally ac-