Confounding Aminoglycosides?

I’m still stuck on the recent FDA meetings where pathways to regulatory approval for pathogen-specific antibiotics were discussed. The example chosen to illustrate the challenges in identifying a pathway was a fictional drug called X-1 that was entirely specific for Pseudomonas aeruginosa. I suggested a superiority design trial where patients with a high risk for Pseudomonas infection would be enrolled and treated with a combination of a carbapenem, X-1 and, if desired, an aminoglycoside. Controls would include those treated with only the carbapenem plus (or not) aminoglycoside. As I noted, the centers would have to be those where there was a rate of carbapenem-resistance among Pseudomonas but not at such a high level that physicians would feel uncomfortable with empiric carbapenem therapy.  (Globally, the rate for carbapenem resistance in this organism is in the range of 15-20% range.  In this range, most patients are still treated empirically with a carbapenem, usually in combination with an aminoglycoside).   In my design, I suggested that the evaluable population for analysis of superiority would, in fact, be those with carbapenem-resistant infections.

The response I received was a horrified, “But the aminoglycoside is an active drug.  It will confound your results.  You will never know what to attribute to X-1 vs. the aminoglycoside!” Some insisted this would be true even if the aminoglycoside were only administered for the first few days of therapy. I beg to differ (at least for serious infections outside the urinary tract).

There are a number of studies examining the effect of aminoglycosides on Pseudomonas aeruginosa infections. One, from Leibovici et. al. in Israel.  In their prospective observational study, they were unable to demonstrate an advantage of adding an aminoglycoside to B-lactam therapy outside of those patients who were neutropenic and those with bacteremia. But more interesting was that for patients where the aminoglycoside was the only “appropriate” drug based on susceptibility testing, patients had a 40-100% higher mortality compared to appropriate beta-lactam monotherapy.

Another study looked at bacteremia caused by Pseudomonas aeruginosa. In this study from Spain, it appeared as though a higher mortality among those treated “appropriately” was mostly due to those who received an aminoglycoside as the only active drug.

Several other studies confirm the lack of an advantage for combination therapy of beta-lactam plus aminoglycoside compared to beta-lactam alone in treating these infections. Several, like the two I noted here, suggest that there is a treatment disadvantage when, in fact, an aminoglycoside is the only active drug in the regimen.  These studies suggest to me that a few days of an aminoglycoside in the design I suggested would be anything but confounding.