Lack of evidence of isoniazid efficacy for the treatment of MDR/XDR-TB in the presence of the katG 315T mutation

To the Editor:

We are grateful for the response by E. Heldal to our correspondence. Similar to other reports [1] from the World Health Organization (WHO) Euro-Region we found that a high proportion (88%) of Mycobacterium tuberculosis strains from patients with multidrug-resistant (MDR)/extensively drug-resistant (XDR) tuberculosis (TB) in the Republic of Moldova had a katG 315T mutation, which is strongly suggestive that high-dose isoniazid should not be part of a standardised treatment regimen for patients with MDR/XDR-TB in this country.

We respectfully disagree with the conclusion made by E. Heldal, as related to the presence of the katG 315T mutation, that “…a majority of strains with low or moderate-level resistance can be effectively treated with isoniazid at normal or high doses…” as this statement is not sufficiently supported by the scientific literature.

In 2011 the Clinical and Laboratory Standards Institute (CLSI) defined a breakpoint for isoniazid at 0.2 mg·L⁻¹ for Middlebrook 7H10 media [2]. CLSI also set 1 mg·L⁻¹ as an additional breakpoint in order to define low-level resistance and identify patients that may possibly profit from high-dose isoniazid treatment. Full range minimum inhibitory concentration (MIC) testing in 7H10 showed an MIC distribution from 0.032 to 0.125 mg·L⁻¹ for wild-type isolates (n=79) [3]. In a recent cohort of patients with MDR/XDR-TB from Germany and Sweden, all 18 strains of M. tuberculosis with a mutation in katG 315T had an MIC for isoniazid of >3 mg·L⁻¹ in liquid culture (MGIT) (J. Heyckendorf and co-workers, unpublished results). In 52 additional isolates with katG 315T mutations without other known resistance mechanisms for isoniazid, MICs were found to be between 4 and 32 mg·L⁻¹ on Middlebrook 7H10 media (T. Schön and co-workers, unpublished results). In 27 of these 52 katG 315T mutated isolates the MICs for isoniazid were 8 mg·L⁻¹. Thus, most isolates with a katG 315T mutation showed MICs at least 30 times higher than current breakpoints and wild-type isolates and not a single strain out of 68 tested M. tuberculosis strains with isoniazid resistance due to a mutation in katG 315 had a MIC <3 mg·L⁻¹ by two different methods.

In the correspondence by Otto-Knapp et al. [4], referred to by E. Heldal as a study which indicates that strains with katG 315T mutations have highly variable MICs, the authors defined intermediate level drug resistance to isoniazid by an MIC of >1 and <5 mg·L⁻¹ using Middlebrook 7H10 agar. The concentrations tested and the presence of other isoniazid resistance mechanisms in these isolates is not mentioned in further detail. We are unaware of an international consensus or standards on the definition of intermediate drug resistance to isoniazid within this MIC range. Furthermore, the suggestion in the same report [4] that isoniazid or any anti-tuberculosis drug is effective just if maximum serum concentrations are above the in vitro generated MIC is, to our knowledge, not supported by clinical data, nor is it consistent with current concepts of establishing clinical breakpoints [5]. The pharmacokinetic and pharmacodynamic (PK/PD) target for isoniazid needed to obtain a clinical cure has yet to be established. When considering treatment of isolates with resistance mutations and elevated MICs against isoniazid, pharmacokinetic variability should be taken into account as it strongly influences PK/PD ratios [6]. In contrast to other pathogens, clinical breakpoints including an intermediate range are not established for tuberculosis and should be based on MIC distributions, PK/PD-analysis and clinical outcome data [5].

In patients with isoniazid-monoresistant TB it appears that different levels of isoniazid resistance do not substantially affect TB treatment outcomes. In one study with 134 patients with culture-confirmed isoniazid-monoresistant TB, treatment success rates were similar (81.8% versus 86.7%) between low- and high-concentration isoniazid-resistant TB [7]. In another study with 395 patients with culture-confirmed isoniazid-monoresistant pulmonary TB, the treatment success rates were also similar in patients with high-level and low-level isoniazid-resistant TB (82.2% versus 83.4%) and among those taking anti-tuberculosis treatment with and without isoniazid (83.1% versus 83.0%) [8]. Of note, a cut-off of >1 mg·L⁻¹ was used to define high-level isoniazid resistance in both of these trials and drug concentrations were not measured. In patients with isoniazid-resistant TB, bacillary mutations in katG 315T, but not in the inhA promotor, have been associated with unfavourable treatment outcomes [9].

The TBNET/RESIST-TB consensus statement highlights that evidence is lacking as to whether adding isoniazid to an MDR/XDR-TB regimen improves treatment outcomes when the genetic basis of the isoniazid resistance is a mutation in katG 315T [10]. Without sufficient evidence, high-dose isoniazid should not be given to patients with MDR/XDR-TB carrying strains with katG 315 mutations since the risk of toxicity may outweigh the potential benefit. Individualised therapy based on comprehensive drug susceptibility testing (DST) appears to be a more promising technique to improve therapy outcomes in MDR/XDR-TB. Early bactericidal activity studies with high-dose isoniazid in MDR/XDR-TB and clinical studies evaluating treatment outcomes in MDR/XDR-TB in patients randomised to receive high-dose isoniazid or a placebo in addition to a DST-guided regimen are needed to address the possible benefits of high-dose isoniazid therapy adequately.

Disclosures