The Benefits of Evaluating Close Associates of Child Tuberculin Reactors from a High Prevalence Group PAUL M. SULLAM, MD, GARY SLUTKIN, MD, AND PHILIP C. HOPEWELL, MD Abstract: To determine the efficacy of evaluating persons (as- sociates) in close contact to children with significant tuberculin reactions, we prospectively evaluated 831 associates of 297 children younger than eight years who had significant (.10 mm) tuberculin reactions. Eighty-seven per cent of the index reactors were foreign- born, as were 84 per cent of the associates. All associates were evaluated by tuberculin skin testing; chest roentgenograms and sputum cultures were obtained if indicated. Four hundred sixty-one (55 per cent) of the associates had significant tuberculin reactions, and 15 had current tuberculosis. However, only three of these cases Introduction Programs designed to reduce the incidence of tubercu- losis within the United States have concentrated on the identification and treatment of current cases, and on preven- tive therapy with isoniazid in selected groups. One proposed method for detecting previously unrecognized cases of tu- berculosis is the investigation of close contacts (associates) of children found to have significant tuberculin reactions.' Evaluation of programs utilizing this technique during the 1950s and early 1960s indicated a yield of two to six cases of tuberculosis for every 100 associates investigated.2" With the subsequent decline in the prevalence of tuberculous infection and disease, however, many communities aban- doned the investigation of associates as a case-finding meth- od, because it was judged to be inefficient.9-2 In recent years, some US cities have undergone a significant demographic change, due to a large influx of foreign-born persons. With these new arrivals, the incidence of current tuberculosis and the prevalence of tuberculin reactors have increased,'3 suggesting that the investigation of associates may once again be a useful means of identifying previously undiscovered cases of tuberculosis. The population served by the Division of Tuberculosis Control, San Francisco Department of Public Health (DPH), includes a large number of recent immigrants from countries where tuberculosis is endemic. In view of the high incidence of tuberculosis within its population, the DPH has attempted to identify current cases by several strategies. One method has been the investigation of associates of all children less than eight years old, who have a significant tuberculin reaction. Nevertheless, the value of this case-finding tech- nique in populations such as ours has not been determined. Therefore, we evaluated the effectiveness of our associate investigation program in identifying new cases of current tuberculosis. From the Department of Public Health, City and County of San Francisco; The Medical Service, San Francisco General Hospital and Medical Center; and the Department of Medicine, University of California, San Francisco. Address reprint requests to Philip C. Hopewell, MD, The Chest Service, SKI, San Francisco General Hospital, 1001 Potrero Avenue, San Francisco, CA 94110. This paper, submitted to the Journal July 26, 1985, was revised and accepted for publication January 21, 1986. C 1986 American Journal of Public Health 0090-0036/86$1.50 were newly discovered (total case rate: 1.81/100, new case rate: 0.36/100). Two of the three new cases were detected in the associates of children younger than three years of age. In addition, 338 candidates for isoniazid (INH) preventive therapy were found. We conclude that although the yield of new cases was low, the evaluation of associates did provide a convenient, high yield method of identifying candidates for INH preventive therapy. Moreover, it served as a useful mechanism for monitoring the adequacy of other case-finding activities. (Am J Public Health 1986; 76:1109-11 11.) Methods The San Francisco DPH performs tuberculin testing as part of a mandatory health evaluation for all new school enrollees, and as part of routine pediatric care in well-child clinics. These populations contain a large proportion of Asians (43 per cent) and Latinos (18 per cent) with a high prevalence of tuberculous infection (34 per cent and 20 per cent, respectively). Children with significant tuberculin reactions (-10 mm of induration) to five tuberculin units of purified protein derivative applied by the Mantoux method were referred to the Division of Tuberculosis Control for further evaluation. These index reactors were assessed for current tuberculosis by interview and chest radi- ography; sputum specimens were collected for mycobacterial stains and cultures if indicated. Household members and other close associates of the child reactors were identified and evaluated by tuberculin skin testing. Associates with significant tuberculin reactions had chest radiographs taken. Three sputum specimens were col- lected for mycobacterial stains and cultures from those associ- ates with any radiographic abnormality consistent with possible tuberculosis. Associates were classified as having current tu- berculosis if: 1) Mycobacterium tuberculosis was isolated from sputum cultures; 2) abnormalities on chest roentgenographs improved during three months of empiric therapy for tubercu- losis with isoniazid, rifampin, and ethambutol; or 3) they were receiving treatment for tuberculosis that had been diagnosed previously. Associates were classified as having previous tu- berculosis if they had negative sputum cultures and chest radiographic abnormalities that had remained stable for one year, of if they showed no radiographic response to three months of empiric therapy with the above regimen. Based on the results of these evaluations, the prevalence of current tuberculosis was calculated for all associates. Since previous studies indicated that associates of very young reac- tors may have a high prevalence of tuberculosis,2' 3 we also determined the prevalence of current disease in associates to children younger than three years of age. The prevalence of tuberculosis among these associates was compared with that among associates to children three years of age or older by means of the Fisher exact test. Results From May 1, 1982 through December 31, 1983, 297 children younger than eight years of age were referred to the Division of Tuberculosis Control because of a significant AJPH September 1986, Vol. 76, No. 9 1109