Current issues of ACP Journal Club are published in Annals of Internal Medicine


Meta-analysis: Isoniazid prophylaxis was beneficial for asymptomatic young adults with positive tuberculin skin test reactions and no risk factors for drug-related hepatitis

ACP J Club. 1991 May-June;114:74. doi:10.7326/ACPJC-1991-114-3-074

Source Citation

Colice GL. Decision analysis, public health policy, and isoniazid chemoprophylaxis for young adult tuberculin skin reactors. Arch Intern Med. 1990 Dec;150:2517-22.



To review and re-analyze decision analysis studies that examine whether asymptomatic young adults with positive tuberculin skin test reactions should receive isoniazid chemoprophylaxis.

Data sources

Studies published in English in or after 1966 were identified using MEDLINE, using the key words isoniazid, isoniazid therapeutic use, tuberculosis, tuberculosis chemoprophylaxis, isoniazid hepatitis, isoniazid adverse effects, decision analysis, decision support techniques, tuberculosis, reactivation tuberculosis, tuberculosis prevention and control, and tuberculosis drug therapy. Bibliographies from identified articles were searched by hand.

Study selection

Decision analysis studies of tuberculosis chemoprophylaxis were selected if they included patients between 20 and 34 years of age who had a normal chest roentgenogram and no predisposing conditions for tuberculosis activation. A second set of studies for estimating the rate of isoniazid toxicity was identified using these criteria: use of isoniazid (300 mg/d for 12 months), monthly monitoring for hepatitis, and categorization of isoniazid related hepatitis cases and death by patient age. A third set of studies for estimating the tuberculosis case-fatality rate was identified using these criteria: documentation of the number of tuberculosis cases in a defined population during the late 1970s and 1980s (a period during which optimal chemoprophylaxis was available) and indication of the number of deaths related to and not related to tuberculosis.

Data extraction

Data were pooled to calculate case-fatality rates for isoniazid-related hepatitis and tuberculosis. Life expectancy was recalculated using one of the decision analysis studies.

Main results

The 3 decision analysis studies reached conflicting conclusions: 1 endorsed isoniazid use, 1 recommended no isoniazid use, and 1 found the benefits of isoniazid chemoprophylaxis too marginal to make a recommendation. The 3 studies of isoniazid toxicity yielded a pooled rate of probable isoniazid-related hepatitis of 1.1% and a case-fatality rate of 0% (95% CI 0% to 3.5%). The 2 studies estimating the tuberculosis case-fatality rate gave a pooled rate of 7.9% (CI 6.7% to 9.1%). The confidence intervals from these rates were used to recalculate life expectancy in the decision analysis study that had found marginal benefits. It was estimated that isoniazid therapy gave a 30-year-old white man a mean of 0.7 to 9.2 extra days of life. Re-analysis of all 3 decision analysis studies favored isoniazid chemoprophylaxis if the isoniazid-related hepatitis case-fatality rate was < 1%, and the tuberculosis case-fatality rate was ≥ 6.7%.


Isoniazid chemopreventive therapy is beneficial in asymptomatic young adults who had positive tuberculin skin test reactions, given observed case-fatality rates for isoniazid-related hepatitis (< 1%) and tuberculosis (≥ 6.7%).

Source of funding: Not stated.

Address for article reprint: Dr. G.L. Colice, Chief, Pulmonary Medicine, VA Medical and Regional Office Center, White River Junction, VT 05001, USA.


Risk-benefit analyses conducted by the American Thoracic Society (ATS) (1) and the Centers for Disease Control (CDC) (2) led to the recommendation that tuberculin-positive persons < 35 years of age who had no other risk factors for tuberculosis activation should receive chemoprophylaxis. However, as Colice reports, 2 of 3 decision analysis studies failed to reach the same conclusion, so the policy has remained controversial. In reviewing these studies, Colice demonstrated that their conclusions were critically influenced by the single probability estimate used for the case-fatality rate of isoniazid-related hepatitis. Colice re-analyzed the decision tree from one of the studies, substituting revised estimates for the case-fatality rates of isoniazid-related hepatitis and tuberculosis, estimates that were based on data pooled from additional studies. The re-analysis showed that prophylaxis had a small, beneficial effect on life expectancy.

Colice's review strengthens the case for adopting the ATS-CDC recommendations. However, some caveats must be stated. First, Colice derived the case-fatality rate for tuberculosis from 2 British studies done in the 1970s, one of which included only patients > 55 years of age, which may have caused the risk for death from tuberculosis in younger patients to be overestimated. Second, the low case-fatality rate for hepatitis was based on three studies that excluded patients with conditions associated with an increased risk for drug-related hepatitis and that followed patients with monthly clinical and serologic evaluations for hepatitis. These studies show the importance of good patient selection and adequate supervision during isoniazid therapy in keeping the risk:benefit ratio low. Finally, none of the decision analyses included an estimate of the societal benefit of prophylaxis through the prevention of secondary cases among contacts.

Philippa Easterbrook, MD, MPH
Johns Hopkins HospitalBaltimore, Maryland, USA.


1. American Thoracic Society. Preventive therapy of tuberculous infection. Am Rev Respir Dis. 1974;110:371-4.

2. Preventive therapy of tuberculous infection. MMWR. 1975;24:71-8.