Trius Therapeutics (TSRX) will soon be reporting data from its second phase III trial (ESTABLILSH-2) and the recently announced publication of the data from its first phase III trial (ESTABLISH-1) in the Journal of the American Medical Association (JAMA) that provides incremental evidence that should further derisk the upcoming results. While this might not seem like an important milestone (as we already know the data), it is actually a nice incremental positive that should not be ignored. Data published in peer-reviewed formats go through a rigorous process that ensures the quality of the article. For instance, JAMA only publishes 9% of the articles that it receives and is one of the highest ranked journals in the field of medicine (ranked 2nd according to journal-rankings.com and 2nd as well if using the ISI impact factor). In general, then, the publication represents, at the very least, additional confirmation as to the quality and robustness of the results from ESTABLISH-1 but also provides some additional details about the trial results.
Perhaps the most important additional information is the robustness effect of tedizolid across varying ways of analyzing the data. As we already know, the trial showed that tedizolid is non-inferior to linezolid with 79.5% of tedizolid patients responding to therapy versus 79.4% of linezolid treated patients. Given these similar rates of success, however, are the reasons for treatment failure the same? Generally, there are some interesting differences. For instance, of the 68 non-responders to tedizolid 41 (60% of tedizolid non-responders) were because of missing data and only 27 (40% of tedizolid non-responders) are confirmed failures. In contrast, of the 69 linezolid non-responders 34 (50% of linezolid non-responders) were due to missing data and 35 (50% of linezolid non-responders) were due to clear failure. While this is not statistically significant given the sample size, it is interesting that only 8.1% of tedizolid patients were clear failures to treatment versus 10.4% for linezolid.
The article also has a series of additional tests to examine how tedizolid and linezolid stack up under differing end points. Of the 5 endpoints they examined tedizolid was better in 3 and linezolid better in 2. Of course, that is not necessarily meaningful because in all of the endpoints there was never a statistically significant difference. In addition, the authors found that early clinical response was highly correlated with end of treatment response for both drugs. In particular, only 2.4% (7 patients) of the tedizolid early clinical responders became failures at the follow up and this is statistically indistinguishable from linezolid where only 2% (6 patients) of early clinical responders became failures. Overall, the article just confirms what we already knew and that is tedizolid is as effective in treating acute bacterial skin and skin structure infections (ABSSI) as linezolid and this is true across different measures and ways of defining success.
The article also went into more details as to the safety profiles of the two drugs. In general, there were no major differences in adverse events that would indicate tedizolid is less tolerable than linezolid. In fact, the only large difference was in terms of gastrointestinal disorder, where 16.3% of the tedizolid group had an AE versus 25.4% of the linezolid group. Give the shorter course of treatment this is not necessarily a surprising outcome. Perhaps the only AE advantage for linezolid was in terms of abnormal treatment-emergent alanine aminotransferase elevations, where this occurred in 4.1% of the tedizolid group but only 3.5% of the linezolid group. This difference is not statistically significant and none of these patients discontinued because of it. In fact, Prokocimer et al (2013:565) noted that "no apparent pattern suggesting liver dysfunction or toxicity emerged." In general, however, the side effect profiles are similar and likely favor tedizolid as being the more tolerable of the two drugs.
JAMA also published an editorial on the trial in the issue (Doran and Boucher 2013), which came to similar conclusions. Doran and Boucher (2013: 608) concluded that tedizolid "is a new oral antibiotic that appears efficacious using a short course and may have a better safety profile than linezolid." What is interesting is that they focused less on the gastrointestinal issues and more on a benefit in terms of platelet counts, where in the tedizolid group only 2.3% saw substantially low platelet counts versus 4.9% in the linezolid group. They point out that more data are necessarily to establish whether this difference is clinically significant but argue it is important to further examine as low platelet counts are a treatment limiting side effect for linezolid.
So what do these articles mean for Trius? I think the quality of the data and the results are another piece of evidence that the upcoming phase III trial results have little clinical risk. The articles also highlight what is going to be the differentiating factor for tedizolid, which is safety and, to a lesser extent, ease of use. Trius is trading at a market capitalization of just over $200M and has a lead compound with an efficacy on par with linezolid and a safety profile that appears cleaner (although the data are still early in this regard). Linezolid (marketed by PFE) sold $1.2B in 2011 and likely over $1.3B in 2012. While it is unlikely that tedizolid will reach those sales heights, it is a significant market opportunity and one that the current value of Trius seems to be discounting too heavily.