converted their tuberculin reactions while students in the classes of 1930 to 1953 (83). Among 1,480 such students, 1,353 were alive at the time of follow-up in 1953; none of the decedents had tuberculosis at the time of death. In total, there were 39,205 person-years of follow-up. There were 92 cases of clinical tuberculosis in these individuals, or a rate of 6.2 percent (92/1,480). It is possible that this rate is low because of the short period of follow-up of recent reactors. Myers and colleagues used tuberculin skin testing techniques with high doses of old tuberculin. While Minnesota is a geographic region where nonspecific reactivity is rare, it is also possible, although not likely, that the number of reactors was overstated, again leading to underestimation of the risk.
In a parallel study covering the same time period, Myers and colleagues surveyed the graduates of three Minnesota nursing schools (84). Follow-up information was obtained on 2,880 of 3,192 graduates (90.2%). Of nursing students who either were tuberculin reactors on nursing school entry or who became tuberculin positive during their nursing studies, 5.2 percent (33/637) developed clinical tuberculosis. As with the earlier studies, the caveats about possible underestimation of the risk apply. In both the medical and nursing student populations, the risk was greater for students who converted in school than for those who were infected prior to entry. This observation is consonant with others that suggest that the primary school years may be years of infection with a relatively low risk of subsequent disease in comparison with the risk from infection during the years of young adulthood.
A longitudinal study of tuberculin reactors in Britain from 1933 to 1944 (Prophit study) has been recently reanalyzed by Sepkowitz (85). Nearly 1,500 medical students and more than 3,000 nurses were followed for a decade. Among medical students, the rates were 1.0/100 person-years for females and 0.6/100 person-years for males. Nurses were classified as high and low exposure depending on their current work status. In the high-exposure group, the tuberculosis incidence was 1.5/100 person-years, and in the low exposure group it was 0.7/100 person-years. The higher attack rate among high-exposure nurses was thought by Sepkowitz to indicate exogenous reinfection as a source of some of the disease. As in the studies of Myers and colleagues, old tuberculin was used for skin testing.
In her postal survey of physicians who graduated from California medical schools prior to 1975, Barrett-Connor found that 5.0 percent (100/ 1,988) of doctors who had been tuberculin-positive on medical school entry or who converted their tuberculin tests later developed tuberculosis (86).
Lydia B.Edwards and colleagues obtained follow-up information as of the end of 1969 for 823,199 (85 percent of those tested) naval recruits tuberculin tested with 5 tuberculin units of PPD between 1958 and 1967