Fluoroquinolones: are they essential to treat multidrug-resistant tuberculosis?

To the Editors:

We read with interest the excellent letter by Holtz and Cegielski 1 contributing to the current discussion on extensively drug-resistant (XDR)-tuberculosis (TB). Several publications have already demonstrated that resistance to fluoroquinolones (FQ) is independently associated with poor outcome and/or that the possibility of including FQ in regimens improves treatment outcomes of multidrug-resistant (MDR)-TB cases 2–4. This happened before the (recent) description of XDR-TB 1, 5. We do not know how many of the patients with MDR-TB strains were, in fact, infected with XDR Mycobacterium tuberculosis.

We wanted to establish the role of the different XDR-defining components (e.g. isoniazid and rifampicin, FQ and injectable second-line drugs) in determining poor treatment outcomes.

Our group has shown for the first time that XDR-TB cases in Italy and Germany have a five-fold increase in the risk of death (relative risk (RR) 5.45; 95% confidence interval (CI) 1.95–15.27; p<0.01), require longer hospitalisation than MDR-TB cases (241.2±177.0 versus 99.1±85.9 days; p<0.001), have a longer treatment duration (30.3±29.4 versus 15.0±23.8 months; p<0.05) and, for the few cases who converted, need a longer time to smear/culture conversion (p<0.01) 6. The findings of a second study, which included additional cases from Estonia and the Russian Federation, demonstrated that XDR-TB cases had an RR of 1.58 to achieve death or failure compared with MDR-TB cases resistant to all first-line drugs (95% CI 1.14–2.20; p<0.05) and an RR of 2.61 (95% CI 1.45–4.69; p<0.001) compared with MDR-TB cases in which susceptibility to at least one first-line drug still existed 7. These data support the observation that the loss of first-line drugs different from rifampicin and isoniazid has a role in worsening prognosis of MDR-TB cases.

In order to better understand the role of FQ in determining poor treatment outcomes in MDR-TB cases, we re-analysed data from the four-country study 7 to assess whether there is any difference in death or mortality in MDR-TB cases resistant or susceptible to FQ. The overall sample included 425 MDR-TB cases (361 MDR, 64 XDR). A total of 87 (20%) were resistant to FQ, 23 (26%) being MDR and 64 (74%) XDR. Although the proportion of MDR-TB cases resistant to FQ was similar in the three countries reporting FQ resistance (i.e. 18, 24 and 24% in Italy, Germany and Estonia, respectively), the proportion of XDR-TB cases among FQ-resistant cases was largely different (50, 27 and 88% in Italy, Germany and Estonia, respectively). FQ-resistant MDR-TB cases yielded a higher proportion of deaths than non-FQ-resistant cases (20 versus 12%; p = 0.020), as well as a higher proportion of treatment failures (19 versus 9%; p = 0.038; table 1⇓).

At the multiple regression analysis, the presence of XDR-TB is the only independent risk factor for both death (odds ratio (OR) 2.07; 95% CI 1.05–4.05; p<0.034) and failure (OR 2.37; 95% CI 1.14–4.89; p<0.02). The findings of our analysis suggest that FQ contribute to increase the risk of death and failure, being a key XDR-defining variable.

In conclusion, apart from linezolid, fluoroquinolones represent the only “new” class of active drugs currently available to treat drug-resistant tuberculosis. They are effective and relatively well tolerated. Furthermore, fluoroquinolones have the potential to allow a reduction in the (still long) short-course chemotherapy regimens. Unfortunately, rapid selection of drug resistance mutants to fluoroquinolones is a well-known phenomenon. Prevention of development of further drug resistance is imperative until new drugs become available in the treatment arena.

Support statement

The study received financial support from the University Hospitals of Leicester NHS Trust.

Statement of interest

None declared.