SETTING: The developing world. OBJECTIVE: To compare the strategy of TB vaccination with that of tuberculin skin-testing in conjunction with isoniazid (INH) in preventing tuberculosis in HIV-infected persons. For any clinical scenarios in which immunization would be more effective than INH preventive therapy, to determine the minimum necessary vaccine safety and effectiveness required. DESIGN: Decision analysis. A hypothetical cohort of 10000 HIV-infected persons, 65% of whom were tuberculin positive, living in the developing world, was studied. Probability estimates were based on BCG vaccine for the baseline analysis, and it was assumed that the vaccine cannot protect if given after infection. RESULTS: Under the probability estimates and assumptions of the analysis, tuberculin skin testing/INH preventive therapy would prevent 458 more cases of TB and 45 more deaths due to TB than TB vaccination. One- and two-way sensitivity analyses revealed no thresholds at which TB vaccination would be the preferred strategy. Vaccine safety did not impact the outcome of the analysis. Three-way sensitivity analysis revealed that if the prevalence of anergy were 35% and the risk of progression to active TB among anergic persons 12.2 cases per 100 person-years, a vaccine would have to be at least 87% effective to be preferred over INH preventive therapy. CONCLUSIONS: Under the conditions of the analysis, which did not account for cost or logistics, tuberculin skin testing/INH preventive therapy would be more effective than TB vaccination in preventing TB among HIV-infected persons. The hypothesized TB vaccine would prevent more TB than INH preventive therapy only in areas where the prevalence of anergy and risk of active TB if anergic were high, and vaccine effectiveness exceeded 87%.
SETTING: The developing world. OBJECTIVE: To compare the strategy of TB vaccination with that of tuberculin skin-testing in conjunction with isoniazid (INH) in preventing tuberculosis in HIV-infectedpersons. For any clinical scenarios in which immunization would be more effective than INH preventive therapy, to determine the minimum necessary vaccine safety and effectiveness required. DESIGN: Decision analysis. A hypothetical cohort of 10000 HIV-infectedpersons, 65% of whom were tuberculin positive, living in the developing world, was studied. Probability estimates were based on BCG vaccine for the baseline analysis, and it was assumed that the vaccine cannot protect if given after infection. RESULTS: Under the probability estimates and assumptions of the analysis, tuberculin skin testing/INH preventive therapy would prevent 458 more cases of TB and 45 more deaths due to TB than TB vaccination. One- and two-way sensitivity analyses revealed no thresholds at which TB vaccination would be the preferred strategy. Vaccine safety did not impact the outcome of the analysis. Three-way sensitivity analysis revealed that if the prevalence of anergy were 35% and the risk of progression to active TB among anergic persons 12.2 cases per 100 person-years, a vaccine would have to be at least 87% effective to be preferred over INH preventive therapy. CONCLUSIONS: Under the conditions of the analysis, which did not account for cost or logistics, tuberculin skin testing/INH preventive therapy would be more effective than TB vaccination in preventing TB among HIV-infectedpersons. The hypothesized TB vaccine would prevent more TB than INH preventive therapy only in areas where the prevalence of anergy and risk of active TB if anergic were high, and vaccine effectiveness exceeded 87%.