(87). Tuberculosis developed in 0.4 percent (111/28,478) of individuals initially tuberculin positive with reactions of 10 millimeters or greater. Tuberculosis developed in 0.03 percent (272/794,721) of those with initial tuberculin reactions smaller than 10 millimeters. The morbidity for men 10 percent underweight was three times that for men 10 percent overweight.

Horwitz, Wilbek, and Erickson followed more than 626,000 persons aged 15 to 44 years in Denmark who were tuberculin tested in 1950–1952 for 12 years (88). Their results were age stratified and are shown in Table C-13. It is clear that the risk was greatest in the postpubertal years and decreased later in adulthood.

Useful data can be obtained from follow-up of the tuberculin-positive individuals excluded from BCG vaccination trials. In general, this information is excellent because the tuberculin skin testing was usually done by skilled, specifically trained individuals, the studies were done prospectively, and great effort was put into data management. For the purposes of this review, many of them have the disadvantage of having been conducted in children with follow-up through the adolescent and postpubertal years. Studies done in high-tuberculosis-prevalence, developing countries have not been considered here, as they are unlikely to represent the risks to American health care workers.

Sol Roy Rosenthal and coworkers gave BCG vaccine to student nurses in Chicago during the 1940s and early 1950s (19). Among initially tuberculin-positive students followed for 12 years, tuberculosis developed in 0.7 percent (3/420).

The British Medical Research Council conducted a trial of BCG vaccination among more than 58,000 schoolchildren in 1950–1952 and reported results of a 15-year follow-up of 54,239 of them (89). Including as tuberculin positive both those who reacted to 3 tuberculin units of PPD and those