Tuberculosis in prisons

Abstract

Sir—In their report on tuberculosis treatment in prisons, Rudi Coninx and colleagues (March 20, p 969) highlight the low rates of treatment success (54%) obtained in a situation where resistance to antituberculosis drugs is high, despite use of the directly observed (short-course) treatment strategy (DOTS). This finding leads the researchers to question the strategy, particularly the first-line regimens. What is surprising in their results is the low efficacy of treatment in individuals with susceptible bacilli. Results of drug susceptibility testing before treatment are available for only 101 (28%) of the 357 participants who completed treatment in prison. Of the 101 patients, 37 were failures (sputumpositive at the end of treatment), of whom three were not resistant and seven were resistant to a single drug before treatment. The treatment failure rate in Coninx’s study is 12% in susceptible patients and 28% in monoresistant patients (table). These results conflict with those reported by Mitchison and Nunn in a review of clinical trials that aimed to analyse the effect of primary resistance on treatment results. They showed that if the strain was susceptible to all antituberculosis drugs or monoresistant to isoniazid or streptomycin, the failure rate of treatment regimens that contain rifampicin for 6 months was less than 1% (table). Coninx and colleagues argue that the three failures with strains initially sensitive to all drugs may have been reinfected during treatment by a multiresistant strain. The same would need to be true for the seven monoresistant individuals. Even in a situation where the prevalence of HIV-1 is high, which does not appear to be the case here, this hypothesis does not seem reasonable. If the first-line regimens were truly so ineffective with susceptible or monoresistant strains, any antituberculosis treatment strategy would be vain. The third-line regimens, which are much less effective and more toxic, can in no way represent an alternative. These results question the data presented by Coninx and colleagues. Their report is more an analysis of routine data within a programme than a prospective study; there may have been bias in the recruitment of patients or errors in recording results. Furthermore, the prison context creates behaviour patterns that are unusual. As indicated by the researchers, patients with tuberculosis are better off than the other prisoners, if only from the point of view of diet. As a result, it is not in their interests to be cured. Although they are theoretically under DOTS, do prisoners really swallow the drugs? Do the sputum samples at the end of treatment belong to the right patients? Whatever the reasons for these failures, changing the first-line regimen itself will not provide better results. If the results for susceptible and monoresistant strains are doubtful, what could one say of the others? The difficulty posed by cases of multidrug-resistant tuberculosis in exUSSR prisons is serious. This issue requires particular attention, starting with a better understanding of what causes these multiresistant cases. The question is an important one, and it is not groundless criticism of DOTS, which has proved successful e l s e w h e r e , , 5 that will help find a solution to the problem.