Follow-up of tuberculosis patients and link with HIV infection in a Frenchdistrict, Gironde, 1995-1996

Objective - The authors had for aim to describe the outcome of tuberculosis patients enrolled by the Reseau Tuberculose Gironde (RTG) and identify the prognostic factors of success and survival after treatment, with an emphas is on HIV co-infection. Patients and methods - Data on tuberculosis, collected from 1995 to 1996 th rough an active surveillance system in the district of Gironde, allowed to enroll patients, then follow them up after treatment. Data was described ac cording to the Kaplan-Meier survival method and the prognostic factors of s urvival were studied with the Cox regression model. A logistic regression w as used to study the factors associated with treatment success. Results - 296 patients were enrolled, including 29 seropositive (HIV+). Out come data was available after treatment for 279 patients (94%): 73% of them were declared bacteriologically cured, 2.5% were treatment failures, 14.5% had died by the end of treatment, 10% were reported lost to follow-up. The death risk was statistically associated to age (Hazard Ratio HR = 1.40; 95 % Confidence interval [CI = 1.03-1.90] for a 10 year increase), and to the absence of tuberculosis treatment, as observed for 26 patients (HR - 25.5 [ 8.1-80.1]). The death risk was also higher among HIV+ patients (HR = 4.9 [1 .2-19.4]) compared to HIV- patients. Absence of standard tuberculosis treat ment and old age remained the only risk factors for treatment failure in th e logistic regression (HR = 27.9 [6.2-125.3] and HR = 1.15 [1.01-1.32] for a 10 year increase, respectively). Conclusions - Tuberculosis remains a serious disease in France, even in the short-term, when considering the current diagnostic and case management co nditions. (C) 2000 Editions scientifiques et medicales Elsevier SAS.