Optimizing Second-Line Therapy for Drug-Resistant Tuberculosis: the Additive Value of Sequencing for Multiple Resistance Loci

LETTER

Molecular probe-based assays for common resistance mutations to isoniazid and rifampin (multidrug-resistant tuberculosis [MDR-TB]) are of considerable clinical value (1, 2), and yet once MDR-TB has been identified, the more vexing challenge arises of determining a regimen of several medications to which the Mycobacterium tuberculosis isolate is susceptible. Numerous reports describe increasing resistance to second-line medications among MDR-TB isolates (4). Thus, we read with great enthusiasm the study by Campbell et al. (3), which describes the experience at the Centers for Disease Control and Prevention (CDC) with sequencing multiple additional resistance loci, particularly for those drugs that one would quickly consider for MDR-TB therapy: pyrazinamide (pncA), ethambutol (embB), the fluoroquinolones (gyrA), kanamycin (rrs and eis), amikacin (rrs), and capreomycin (rrs and tlyA). We have utilized this CDC sequencing service on several occasions and describe its value through an illustrative case.

A 32-year-old previously healthy woman, originally from Belarus, presented with cough and a right upper-lobe cavity on chest radiograph. She did not have a history of TB or prior contacts with drug-resistant TB. Sputum smear microscopy was positive for acid-fast bacilli, and she was started on first-line TB treatment with isoniazid, rifampin, ethambutol, and pyrazinamide at a local hospital (Fig. 1). By day 14 of treatment, M. tuberculosis was confirmed, and thereafter the isolate had conventional first-line drug susceptibility testing (DST), which revealed MDR-TB. Amikacin, cycloserine, and moxifloxacin were empirically added to the regimen of pyrazinamide and ethambutol, and the isolate was sent to the CDC. Within 3 days, sequencing results were received and revealed the A1401G rrs mutation but no mutations in gyrA. With these preliminary data, the fluoroquinolone was continued, the injectable drug was held, and the regimen was expanded to include linezolid and ultimately para-aminosalicylic acid. These decisions were later supported by conventional DST, but in contrast to the 3-day turnaround from CDC, it took up to day 101 of treatment before all desired conventional DST was complete. To date, the patient has tolerated the regimen without side effect or sign of disease recurrence and will have completed 18 months of therapy in June 2011. We cannot understate the value of speed offered by these molecular results for the patient, health department, and clinician.

Fig. 1.

Timeline of second-line regimen construction for patient with complex drug resistance.

While we applaud the progress to date, further work remains. The rich sequence data described by Campbell et al. on a large number of diverse M. tuberculosis isolates reveal a nagging rate of conventional resistance to certain medications such as capreomycin and ethambutol despite wild-type sequence. Furthermore, the rate of pncA mutation in pyrazinamide-susceptible isolates was not clinically insignificant. These data caution against unbridled enthusiasm for molecular DST, particularly for probe-based methods, which cannot provide such broad sequence information (5). The work of Campbell et al. helps the clinician a great deal, but the need for confirmatory conventional DST remains in most settings, and faster methods for determining susceptibility to para-aminosalicylic acid, cycloserine, ethionamide, linezolid, and the carbapenems must be pursued.