Editorial

Drug-resistant tuberculosis: issues in epidemiology and challenges for public health

WHO estimated that of 9.4 million cases of tuberculosis (TB) worldwide in 2008, 440,000 (3.6%) had multidrug-resistant (MDR)-TB. Childhood TB is estimated at 10-15% of the total burden, but little is known about the burden of MDR-TB in children. Children in close contact with MDR-TB cases are likely to become infected with the same resistant strains and are vulnerable to develop disease. Although MDR-TB is a microbiological diagnosis, children should be treated empirically according to the drug susceptibility result of the likely source case, as often cultures cannot be obtained from the child. MDR-TB treatment in children is guided by the same principles, using the same second-line drugs as in adults, with careful monitoring for adverse effects. Co-infection with HIV poses particular challenges and requires early initiation of antiretroviral therapy. Preventive therapy for high-risk MDR-TB contacts is necessary, but no consensus guidance exists on how best to manage these cases. Pragmatic and effective Infection control measures are essential to limit the spread of MDR-TB.

The aim of this study was to enhance tuberculosis (TB) treatment outcome monitoring by linking diverse surveillance systems and estimating treatment outcomes including relapse.

Tuberculosis treatment was surveyed in the Lazio region (Italy) from 1999 to 2001; a six-year follow-up of notified cases was undertaken to detect relapses. The results were analyzed as a population-based case–control study comparing each unsuccessful outcome and relapse with eligible controls.

Of the 974 patients who entered the survey, 805 (82.6%) had complete treatment evaluations; 398 (49.4%) had a successful outcome, 401 (49.8%) had an unsuccessful outcome, and six developed chronic TB. Death was associated with age >64 years (OR 5.9; 95% CI 3.1–11.2), male gender (OR 2.1; 95% CI 1.0–4.4), and using second-line drugs (OR 2.3; 95% CI 1.0–5.4). Treatment failure was associated with previous treatment (OR 3.0; 95% CI 1.4–6.7) and being male, being foreign born (OR 6.6; 95% CI 2.1–21.2), receiving second-line drugs (OR 7.4; 95% CI 1.8–29.5), and receiving modified therapy (OR 5.1; 95% CI 1.7–14.9). Relapses after successful outcomes were detected in 5.5%, for which the strongest predictor was having extrapulmonary lesions (OR 22.8; 95% CI 1.8–287.3).

Linking our survey data to other surveillance systems improved the mortality estimates and detected a high rate of relapse. Having received previous treatment and being a foreigner were independent determinants of treatment failure, suggesting that both acquired and primary drug resistance affect TB patients in Lazio.

To investigate the anti-tuberculosis drug resistance pattern of pulmonary tuberculosis isolates in southern Taiwan, we performed a hospital-based surveillance at a southern Taiwan medical centre from 1996 to 2000. The combined drug resistance rates to at least one of four first-line agents (isoniazid, rifampicin, ethambutol, streptomycin) was 52.4%, and to both isoniazid and rifampicin was 11.4%, indicating high resistance rates compared with those reported in the World Health Organization (WHO)/International Union Against Tuberculosis and Lung Disease (IUATLD) global project and in northern Taiwan. The resistance rates to two second-line drugs, cycloserine, and kanamycin, were 75.7 and 23.7%, respectively. A significant decreasing trend in resistance rates to all tested drugs except streptomycin was observed during the 5-year period. The resistance rates in 1996 and 2000 were 43.1 and 16.4% for isoniazid, 23.4 and 9.5% for rifampicin, 23.4 and 12.1% for ethambutol, 92.7 and 50.9% for pyrazinamide. The combined drug resistance rate may not be the most accurate tool as it includes previously treated cases that may inflate the resistance rate and cases without a history of treatment. However, the observation of trends in the susceptibility of pulmonary tuberculosis with the increasing percentages of tuberculosis patients receiving the complete treatment course and the decreasing percentages of cases lost to follow-up in Kaohsiung after the institution of new governmental regulations for case management in 1997, suggest that such intervention programs are useful.

After decades of decline in developed countries, there was a resurgence of tuberculosis in the mid-1980s accompanied by increased recognition that this infectious disease has long remained a major public health problem at the global level. New methods from molecular biology, in particular DNA ‘fingerprinting’ (of Mycobacterium tuberculosis), made it clear that current transmission and recent infection (in contrast to reactivation of earlier, latent infection) were much more significant than previously believed. Studies of tuberculosis outbreaks using these new tools pointed to complex networks through which infection was spreading and highlighted the need for new approaches to outbreak investigation and disease control. In the study reported here a new approach — combining methods from molecular biology, epidemiology and network analysis — was used to examine an outbreak of tuberculosis in Houston, Texas. Initial investigation using conventional strategies revealed few contacts among 37 patients with identical (six-band) DNA (IS6110-based) fingerprints but subsequent research uncovered over 40 places (including many gay bars) to which patients in this outbreak could be linked. Network methods were used to reconstruct an outbreak network and to quantify the relative importance (here, ‘betweenness’ centrality) of different actors (persons and places) playing a role in the outbreak. The multidisciplinary work provides the basis for a new approach to outbreak investigation and disease control.