Multidrug resistant tuberculosis in France 1992-4: two case-control studies

Subjects, methods, and results

Annual prevalence of multidrug resistance was calculated by dividing the number of cases of multidrug resistant tuberculosis—patients who had at least one isolate resistant to isoniazid and rifampicin in the calendar year—by the total number of cases with tuberculosis confirmed by culture that the laboratories reported. Multidrug resistant tuberculosis was defined as secondary in patients who had been treated for 1 month or more before the first known multidrug resistant isolate, and as primary in all other cases. DNA fingerprinting was performed on multidrug resistant strains sampled in 1993 and 1994.2 Factors associated with multidrug resistant tuberculosis were analysed by comparing cases of multidrug resistant tuberculosis reported by the laboratories with cases that were notified for the same period by 69 (of 100) French districts where HIV infection was consistently monitored. We compared primary and secondary cases of multidrug resistant tuberculosis in two case-control studies with all notified new cases and all notified cases with a history of previous tuberculosis respectively. We performed multivariate analysis by logistic regression.

In 1992, 48 out of 8521 cases of tuberculosis confirmed by culture were multidrug resistant (0.6% (95% confidence interval 0.4% to 0.7%)); in 1993, 40 out of 8539 (0.5% (0.3% to 0.6%)); and in 1994, 58 out of 7752 (0.7% ( 0.5% to 0.9%) (P=0.10 for trend). Prevalence did not vary significantly between the 22 administrative regions.

The 146 cases occurred in 125 patients, of whom 116 (93%) had pulmonary tuberculosis (70 had a positive sputum smear test). Of 122 patients with information on previous treatment, 31 had primary and 91 secondary multidrug resistant tuberculosis. Overall, 91 out of 122 (74%) of the patients were men and 58 out of 122 (49%) were born outside Europe, without significant difference between primary and secondary cases. Primary cases were significantly younger than secondary cases (median age 35 years v 40 years, P=0.02) and were more likely to be infected with HIV (35% (11/31) v 13% (12/91), P<0.01).

We analysed DNA fingerprints for 66 of the 88 patients whose cases were reported on in 1993 and 1994. Only two patients had identical fingerprints. One was a French citizen resident in New York City who tested positive for HIV and returned to France after multidrug resistant tuberculosis was diagnosed. During his stay in hospital, where he had respiratory symptoms and a positive sputum smear test, he came in contact with the other patient, who was also HIV positive and developed multidrug resistant tuberculosis 2 months later. The strain was the “W” strain first recognised in several outbreaks in New York City.3

The only factor associated with primary multidrug resistant tuberculosis in multivariate analysis was infection with HIV (table). Secondary multidrug resistant tuberculosis was independently associated with young age and non-European origin but not with HIV infection (table). These results were unchanged when analysis was restricted to patients with known HIV status.

Comment

Our results do not show an epidemic of multidrug resistant tuberculosis in France. However, although HIV infection was not associated with secondary multidrug resistant tuberculosis, it was an independent risk factor for primary multidrug resistant tuberculosis. The increased risk of primary multidrug resistant tuberculosis in people infected with HIV has recently been shown by the nosocomial outbreaks reported in London and Madrid. 4 5 To assess failures in tuberculosis control, the prevalence of multidrug resistance should be monitored throughout Europe.