Vancomycin vs TMP/SMX for MRSA – Not the Final Word

…the etymology of vancomycin reflects marketing hype and irrational exuberance…

Vancomycin got a bad reputation since the early 90ies when it became clear that it was slow to provide clinical improvement and microbiologic cure in MRSA patients.  It was associated with prolonged bacteremia even when combined with rifampin [1].  And for MSSA, anti-staph penicillins like nafcillin were always the better choice.

Then, in a series of trials comparing vancomycin to linezolid (LZD), the oxazolidinone showed superior outcomes.  Well, because these studies were registration trials which individually only showed non-inferiority with a trend towards superiority, the label of LZD is what it is and vancomycin still can claim non-inferiority[2].  However, with vancomycin’s poor penetration into pulmonary tissues and consistently lower clinical response rates in several large well-controlled trials, most pulmonologists and ID specialists would choose to treat S.aureus pneumonia with LZD today.

To add insult to injury, a heightened awareness of ‘red neck syndrome’, or ‘red man syndrome’ and the need for therapeutic drug monitoring (TDM) made vancomycin even less popular.

There was always interest in using TMP/SMX (trimethoprim/sulfamethoxazole) for serious MRSA infections [3].  Susceptibility data suggest that most MRSA in the US remain susceptible.  So, it seems only logical to use this combination before resorting to the newer but much more expensive antibiotics like ceftaroline/Teflaro, dalbavancin/Dalvance, oritavancin/Orbactiv or daptomycin/Cubicin.  However, comparative clinical data for the two old drugs were not available.

So, the recently published clinical trial by Paul, a side-by-side comparison of vancomycin and TMP/SMX in patients with severe MRSA infections, fills an important gap.[4].

The authors report on 252 patients enrolled at 4 centers in Israel between 2007 and 2014.  Half of the patients had post-surgical infections, and 1/3 of cases had bacteremia.  This was a randomized but open-label trial in which treatment success or failure was determined on Day 7 by a blinded adjudication committee.

The study finding was surprising.  Despite a large 15% non-inferiority margin, TMP/SMX was inferior to vancomycin, in particular in the subgroup of bacteremic patients.  The authors felt that TMP/SMX could not be recommended for severe MRSA infections.

Several caveats are in order before accepting their conclusions.  This was what the authors describe as a ‘practical’ study.  In many aspects it did not meet the quality markers of the linezolid trials mentioned above.  As an open trial which allowed concomitant anti-staphylococcal medicines up to 48 hrs prior to enrolment, it is certainly not registration-quality even by 2007 standards.  A 15% NI margin may seem generous, but FDA would have insisted on a double-blinded comparison, two-sided alpha and an adjustment for the 2 interim analyses performed a la Pocock or more likely, the O’Brien Fleming alpha-spending method. 

It is hard to do a practical study, but even harder to do a well-controlled study

There are other methodologic concerns which detract from the value of this study.  The effect of concomitant surgeries (in approx. 75% of patients) during the initial 7 days of treatment is hard to ignore: Were these wound debridements?  Were topical antibiotics administered?  And some very basic questions remain: What was the MIC distribution of organisms?  Why were there so many polymicrobial infections (37%)?

30-Day mortality is an efficacy and a safety endpoint.  The fact that no difference in survival between the 2 groups at this nor at the 7-day efficacy time point was observed should be noted.  Hospitalization was of equal length in both treatment arms.  Makes you wonder whether the 2 antibiotics are really all that different.

My take-away: Admittedly, it is hard to do a practical study, but it is even harder to do a well-controlled study.  Because the trial fails to meet these higher standards, it is not the last word on TMP/SMX.  If the authors had compared it to LZD, presumably there would have been clear difference in efficacy.

Then again, without sponsors to finance a more definitive study of these two off-patent drugs, the Paul study may be the last word for a long time.


[1] D Levine Ann Int Med. 1991; 115:674
[2] R Wunderink CID 2012; 54:621
[3] N Markowitz AAC. 1983; 23:450
[4] M Paul BMJ 2015;350:h2219 doi: 10.1136/bmj.h2219

Print Friendly, PDF & Email

Leave a Reply

Your email address will not be published.