Extensively drug-resistant tuberculosis: is its definition correct?

To the Editors:

Two articles by Migliori and co-workers 1, 2, which recently appeared in the European Respiratory Journal (ERJ), may support the idea that the current definition of extremely drug-resistant tuberculosis (XDR-TB) 3, 4 is not the most adequate.

In recent years, XDR-TB has become a major concern as it leads to incurable TB in a significant proportion of patients 4, 5. XDR-TB was first defined in March 2006 as multidrug-resistant TB (MDR-TB; resistance to isoniazid and rifampicin) plus resistance to at least three of the six second-line anti-TB drug groups (fluoroquinolones, aminoglycosides, polypeptides, thioamides, cycloserine and para-aminosalicylic acid (PAS)) 6. However, this definition permits the possibility of susceptibility to fluoroquinolones and aminoglycosides (kanamycin, amikacin) and/or polypeptides (capreomycin) in some XDR-TB patients, meaning much higher success rates could be achieved if such drugs were used. Amongst second-line drugs, only the fluoroquinolones and injectables (aminoglycosides and polypeptides) have bactericidal activity and could be considered very effective. Fluoroquinolones and injectables therefore seem to represent the same as isoniazid and rifampicin amongst the first-line drugs. Acknowledgement of the fact that the success of treatment with second-line drugs depends on the use of fluoroquinolones and injectables (aminoglycosides and polypeptides), in addition to the fact that susceptibility testing to these drugs produces more reliable and reproducible results, prompted a modification of the definition of XDR-TB. Currently, XDR-TB is defined as MDR-TB plus resistance to fluoroquinolones and to at least to one of the second-line injectables (kanamycin, amikacin and capreomycin) 3, 4.

Although two recent studies have shown that the current definition of XDR-TB is predictive of a poorer clinical outcome than MDR-TB 1, 7, this definition may still be inappropriate even though it is clearly better than the first. This is particularly so because, in special cases, it allows the use of some first-line drugs and/or one of the injectables (kanamycin, amikacin or capreomycin).

The current definition permits the possibility of susceptibility to ethambutol and/or pyrazinamide, which, although rare, could be present in any given case. In a recent study, Migliori et al. 1 presented the unfavourable outcome of MDR-TB cases resistant to all first-line drugs compared with other MDR-TB cases in which there was susceptibility to ethambutol, pyrazinamide or streptomycin. Table 1⇓ shows that the role of pyrazinamide was avery important in three studies and that a success rate of >90% was achieved with pyrazinamide plus ethionamide and cycloserine 8–10, 15.

The current XDR-TB definition also allows the use of one of the injectables (kanamycin, amikacin or capreomycin) along with ethionamide, cycloserine and PAS, thereby reaching a possible cure rate of >80% 15, as can be observed in four of the studies presented in table 1⇑ 11–14. The possibility of success in an XDR-TB patient using an injectable and all of the second-line drugs could therefore be very close to that achieved in patients with MDR-TB and without XDR-TB (for example, in MDR-TB patients with susceptibility to all of the injectables) 15. In their more recent study, Migliori et al. 2 evidenced the favourable outcome of XDR-TB patients with susceptibility to capreomycin, the least frequently used injectable in the world, and, for this reason, susceptibility is possible in many patients with resistance to kanamicyin and/or amikacyn.

For all of these reasons, the most accurate definition of extensively drug-resistant tuberculosis would be cases with resistance to all first-line drugs (not only those defining multidrug-resistant tuberculosis) and to fluoroquinolones and all of the injectables (not just to one), the two most potent second-line drugs groups. Usually, these patients have a possible treatment success rate of <50% and clearly stand apart from exclusively multidrug-resistant tuberculosis patients. If the current extensively drug-resistant tuberculosis definition is maintained, we will shortly be in need of a new definition for those cases, thus constituting a new level of difficulty in antituberculosis treatment for which the best classification would be XXDR (extensively extensively drug-resistant tuberculosis).

Statement of interest

None declared.