Epidemics of HIV and Tuberculosis Foster Drug-Resistance

Abstract & Commentary

Synopsis: It is suggested by the authors of this study that acquired tuberculous drug resistance may emerge regardless of good medical practice.

Source: Bradford WZ, et al. The changing epidemiology of acquired drug-resistant tuberculosis in San Francisco, USA. Lancet 1996;348:928-931.

Ever since drug-resistant tuberculosis reared its ugly head in the early 1990s in New York City, the country’s epidemiologists have watched closely for the localization of drug-resistant disease. Acquired drug resistance (ADR), in these cases of Mycobacterium tuberculosis infection, is defined as strains that are fully susceptible initially and become resistant at any level to ethambutol, isoniazid, or rifampicin. Bradford and colleagues reviewed a group of 3168 cases of tuberculosis reported in San Francisco from 1985 through 1994. Drug susceptibility data was available in 2612 strains, of which 14 showed ADR. Six were rifampicin resistant, and four additional strains were resistant to INH or ethambutol in addition to rifampicin.

Proof that the patients harbored identical strains was demonstrated by DNA fingerprinting in eight of the 14 sensitive/resistant paired isolates tested. The number of ADR cases increased significantly from the 1985-1989 interval to the 1990-1994 interval. Of the 14 patients, eight developed resistance during the initial course of therapy while six grew ADR strains during a relapse. There was no evidence of medical mismanagement. Cases of ADR were more likely to be white, unemployed, and born in the United States. Risk factors detected by multivariate analysis included gastrointestinal symptoms, non-compliance, and AIDS. For the interval of 1990-1994, these data would translate into a rate of one case of ADR per 121 patients, and one in 40 for those with AIDS.


This study suggests that some cases of ADR may emerge regardless of good medical practice. Of course, issues like absorption of anti-tuberculosis drugs and compliance still make close supervision of therapy and symptoms necessary, but with the continuing epidemics of HIV infection and tuberculosis, a small percentage of patients will select resistant strains from an initially susceptible population of infecting organisms. In fact, directly observed therapy (DOT) in this study group was more common in cases than in controls. The authors note that although resistance to rifampicin alone in M. tuberculosis had been unsusual, it was found frequently and exclusively among patients with AIDS.

Why should rifampicin resistance predominate? The rate of rifampicin-resistant mutants usually runs two-fold lower than that for isoniazid or streptomycin resistance. The most common pattern of ADR in non-HIV patients has been INH-resistant strains.

So, what’s going on? Unfortunately, there was too small a number of ADR cases in this study to incriminate or discount as risk factors such variables as combination rifampicin/INH pills, concurrent use of azole antifungal agents, or the use of rifabutin in the treatment of M. avium infection. Interestingly, the authors do speculate that perhaps other factors like zidovudine exposure may contribute to emergence of resistance. It is known that zidovudine selects our isoniazid-resistant M. smegmatis.

The reader may discount such low rates of ADR in this population. Yet, the patients with ADR at the time of isolation of the resistant isolate were all failing. Certainly, more studies like this one need to be done to determine if there are identifiable risk factors to look for when managing patients with advanced HIV infection and tuberculosis.