Latent TB Infection: When to Treat

January 1, 2003

Q:Should I prescribe isoniazid prophylaxis for an elderly patient withmultiple health problems who recently converted from negative topositive tuberculin status?

Q:Should I prescribe isoniazid prophylaxis for an elderly patient with multiple health problems who recently converted from negative to positive tuberculin status?

A:The treatment of tuberculosis (TB) continues to be a challenge in the United States because of rising incidence, the aging population, and the growing numbers of immigrants from countries where TB is endemic. From 1993 to 1998, the TB case rate was 32.9 per 100,000 among foreign-born persons compared with 5.8 per 100,000 among American-born persons.1 Within a few years, most US cases are expected to occur among foreign-born persons.

TB in the elderly. Most cases of TB in elderly persons represent reactivation of latent disease. Before the modern chemotherapy era, which was inaugurated in the early 1950s with the introduction of isoniazid, a large proportion of the American population was considered to be infected because of a positive reaction to tuberculin. Although most reactors with inactive disease have since died, a significant number of infected elderly persons are still alive. As immunologic defense mechanisms against the tubercle bacillus wane, clinical TB may return with subtle or dramatic manifestations.

Treatment of new infection. In contrast, recent conversion (ie, within 2 years) from confirmed negative to positive tuberculin status indicates new infection. Progressive primary or disseminated TB is thus likely to develop, particularly in elderly patients. Even in the absence of comorbid disease, treatment is indicated in this setting. Some refer to such treatment as “preventive therapy,” but that is a misnomer. Although the incidence of adverse effects associated with isoniazid increases with age, the risk of clinically significant drug-related hepatitis is smaller than the risk of progressive TB. The conventional wisdom is that isoniazid monotherapy should be effective in this setting because in cases of recent conversion, the number of invading and multiplying organisms-and thus the probability of spontaneous mutation- is small. My view is that because the risk of isoniazid-resistant disease with single-drug therapy is not zero, a second drug is required for the first 3 months of treatment. I use ethambutol, which, in low doses (ie, 15 mg/kg/d) is very well tolerated. Most actively multiplying organisms are killed within the first 3 months of 2-drug therapy. After 3 months, isoniazid is continued alone, for a total of 9 months. This regimen should be curative.

Treating drug-resistant disease. The above regimen is effective against isoniazid-sensitive organisms. However, if the culprit organisms are resistant, treatment becomes more complex. The incidence not only of isoniazid resistance but of multiple-drug resistance is increasing in areas of the United States with large numbers of immigrants, particularly those from Eastern Europe, Asia, and Africa. If isoniazid resistance is suspected on epidemiologic grounds or because of a patient’s known contact with drug-resistant organisms, treatment should be guided-if possible-by knowledge of drug susceptibilities of the contact person.

Testing and treatment. The intradermal Mantoux test involves the injection of 10 units of purified protein derivative, standard (PPD-S). A positive tuberculin reaction is defined as 10 mm of induration for immunocompetent persons and 5 mm for those who are immunosuppressed. (Some use 3 mm or any degree of induration as the cutoff for immunosuppressed persons.) The reaction should occur at 48 to 72 hours. A conversion of the PPD from negative to positive within 2 years represents active infection and is an indication for treatment. Today there is a concerted effort to identify and treat so-called latent TB- that is, all tuberculin reactors.2 The problem is a positive PPD response in an adult who has no known exposure and for whom the time of infection cannot be determined. Although a positive PPD response means that infection has occurred in the past, it does not mean that the organisms that are sequestered somewhere in the body-often in lymph nodes-are actively multiplying. Only actively multiplying organisms can be killed or inhibited with antituberculous antimicrobials. Only 5% to 10% of persons with latent TB will progress to clinical disease. If the policy of treating all persons with latent TB is implemented, many reactors who do not have active disease will be treated and thus will be placed at risk for significant adverse reactions (including death).3,4 The only randomized clinical trial that showed the benefit of giving isoniazid to a large population was conducted in Alaska in the 1960s.5 The reason that fewer cases of clinical TB developed in the treated group was that active TB was rampant in Alaska in that era. Thus, many persons with occult but active disease were actually treated. Today, AIDS patients are sometimes treated for latent TB. This practice has not been found beneficial in the United States, although it has been effective in Haiti, where TB is much more prevalent. Antituberculous drugs for latent disease should not be given to all persons with a positive tuberculin reaction, but they may be considered in isolated high-risk populations. Treat all patients who have active TB-that is, those who manifest a timed conversion of PPD from negative to positive-with 3 or 4 antituberculous drugs chosen on the basis of drug susceptibility testing. In addition, treat close contacts with at least 2 of the same drugs for 6 months. Do not treat immunocompetent reactors older than 35 years for whom the timing of conversion is unknown unless they have additional risk factors for active TB, such as concomitant disease or a history of immunosuppressive-drug use.




Talbot EA, Moore M, McCray E, Binkin NJ. Tuberculosis among foreign-born persons in the United States,1993-1998.


. 2000;284:2894-2900.


Geiter L.

Ending Neglect: The Elimination of Tuberculosis in the United States

. Washington, DC: NationalAcademy Press; 2000.


Moulding TS, Redeker AG, Kanel GC. Twenty isoniazid-associated deaths in one state.

Am Rev Respir Dis



ATS/CDC. Fatal and severe liver injuries associated with rifampin and pyrazinamide for latent tuberculosisinfection, and revisions in American Thoracic Society/CDC recommendations-United States, 2001.




Ferebee SH. Controlled chemoprophylaxis trials in tuberculosis: a general review.

Bibl Tuberc

. 1970;26:28-106.

Related Content:

Geriatrics | HIV AIDS | Infection