In vitro activity of amikacin against Mycobacterium avium

In a recent article by L. Heifets and P. Lindholm-Levy (Antimicrob. Agents Chemother. 33:1298-1301, 1989), the authors contend that they could not confirm our conclusions about the in vitro activity of amikacin against the Mycobacterium avium complex (C. B. Inderlied, L. S. Young, and J. K. Yamada, Antimicrob. Agents Chemother. 31:16971701, 1987) and found our observations on the clinical efficacy of amikacin paradoxical (C. B. Inderlied, P. T. Kolonoski, M. Wu, and L. S. Young, Antimicrob. Agents Chemother. 33:176-180, 1989). We disagree with these statements. Heifets and Lindholm-Levy reported amikacin MICs for 50 and 90% of isolates tested of 2.65 and 6.75 ,ug/ml, respectively. These values are quite similar to our values of 4.8 and 13.4 ,ug/ml using a similar method but a more exacting endpoint (99.9% inhibition). When amikacin was tested in an animal model of disseminated M. avium infection, we found it to be effective, and the addition of ciprofloxacin and/or imipenem did not improve the efficacy of amikacin. Heifets and Lindholm-Levy interpreted our results as suggesting evidence for a synergistic interaction between the drugs. We disagree and continue to interpret our results as evidence for the efficacy of amikacin alone. Criteria for the definition of susceptibility of M. avium are neither clearly established nor generally accepted by the community of scientists studying the M. avium complex. Drug concentrations that can be achieved in serum provide only one criterion, and even this criterion must be interpreted in light of the nature of the infection and drug pharmacokinetics. The rather arbitrary interpretive criteria used by Heifets and Lindholm-Levy need to be assessed in terms of clinical efficacy and until substantiated should be regarded as conjecture and potentially misleading.