Pyrazinamide Is a Two-Edged Sword: Do WHO Guidelines Matter?

LETTER

Two recent studies have suggested that pyrazinamide is a two-edged sword in the treatment of multidrug-resistant tuberculosis (MDR-TB). Kempker et al. showed good penetration of pyrazinamide into lung tissue of MDR-TB patients with a significant correlation between the tissue pyrazinamide concentration and necrosis (1). Fofana et al. demonstrated that inappropriately including pyrazinamide for treating MDR-TB could amplify drug resistance (2), possibly by reducing the use of other effective drugs among patients with pyrazinamide-resistant MDR-TB. To harness the unique sterilizing activity of pyrazinamide that might help shorten MDR-TB treatment, rapid drug susceptibility testing (DST) of pyrazinamide is indispensable for categorically distinguishing pyrazinamide-susceptible from pyrazinamide-resistant bacillary strains (3). Besides technical advances, guidelines from authorities matter. While the World Health Organization (WHO) has recommended routinely including pyrazinamide in the MDR-TB treatment regimen (4) unless pyrazinamide resistance is confirmed or pyrazinamide is poorly tolerated, WHO reclassification of pyrazinamide as an additional agent rather than a core drug might inadvertently downplay the importance of pyrazinamide and frustrate the development and implementation of rapid DST of pyrazinamide, thereby potentially promoting the emergence of extensively drug-resistant TB.

With predominant sterilizing activity against nonreplicating persister mycobacteria at acidic pH (5), pyrazinamide works in synergy not only with isoniazid and rifampin in the treatment of drug-susceptible TB (5, 6) but also with second-line TB drugs, notably fluoroquinolones (7), clofazimine (8), and bedaquiline (9). The importance of pyrazinamide is reflected by its inclusion in a number of phase 2 and 3 MDR-TB clinical trials involving one or more novel drugs, including bedaquiline, delamanid, and pretomanid (Table 1). Although novel drugs may also have certain activity against persister mycobacteria, they mainly act synergistically with, but probably cannot replace, pyrazinamide.

TABLE 1.

Registered phase 2 and 3 clinical trials of MDR-TB involving the use of pyrazinamide and novel drugs

Trial (trial registry identifier)	Novel drug(s)
NC-001 (NCT01215851)	Bedaquiline, pretomanid
NC-002 (NCT01498419)	Pretomanid
NC-003 (NCT01691534)	Bedaquiline, pretomanid
NC-005 (NCT02193776)	Pretomanid, bedaquiline
MDR-END (NCT02619994)	Delamanid
NEXT (NCT02454205, PACTR-201409000848428)	Bedaquiline
STAND (NCT02342886)	Pretomanid
DAZZLE (NCT02975570)	Delamanid

Proper use of pyrazinamide in MDR-TB treatment has been confounded by pyrazinamide resistance, which varies widely from 10 to 85% in different geographic settings (3, 10–12), and the technical challenge of phenotypic pyrazinamide DST (3). Even if phenotypic tests were easy, they are not sufficiently quick for timely guidance of the use of pyrazinamide. Rapid molecular tests are needed for determining pyrazinamide susceptibility.

Although pyrazinamide resistance is attributable to more than one molecular mechanism, including discernible mutations in the pncA, rpsA, panD, and clpC1 genes (13, 14), mutations in the pncA gene encoding pyrazinamidase accounts for the majority (72 to 99%; average, 85%) of pyrazinamide resistance in Mycobacterium tuberculosis (15–20), with a good correlation between pncA mutations and lack of pyrazinamidase activity in detecting pyrazinamide resistance. It has been demonstrated that pncA sequencing can be substantially more sensitive and specific than MGIT 960 pyrazinamide DST (21), which is prone to missing low-level pyrazinamide resistance caused by nonsynonymous mutations in pncA (22, 23). Although pncA sequencing may be flawed by false resistance associated with pncA mutations that do not significantly alter pyrazinamidase activity (24–26), as well as false susceptibility related to resistance-associated mutations outside pncA (5, 16, 27), it is less onerous and more affordable than other molecular methods (3). Correlation between mutations and pyrazinamide susceptibility can be optimized by compiling a database of clinical isolates and categorically classifying pncA mutants according to observed predictive values for pyrazinamide resistance (23), with or without supplementary testing for pyrazinamidase activity (28), as exemplified by the establishment of the TB relational sequencing data platform (ReSeqTB) (29, 30).

Reinstating pyrazinamide as a core drug may help harness the two-edged sword for properly treating categorically different MDR-TB subgroups that differ by pyrazinamide susceptibility.