Another Staph aureus vaccine failure

The April 3rd JAMA contains the results of a wonderfully designed and executed study by Fowler and colleagues assessing the benefits of a novel (V710) vaccine targeting S. aureus infection in cardiothoracic surgery. The study was negative (i.e. the vaccine didn't work) and highlights how little we know in 2013 about how humoral immunity works for this important human commensal. But first the study and results...

Among 8031 patients entered into the study, the vaccine was given between 14 and 60 days pre-op to 3981 randomly-selected patients prior to surgery while 3982 received saline placebo. The study was powered to detect a 20% reduction in S. aureus bacteremia or deep surgical site infection through 90-days post-op. Secondary outcomes were any invasive S. aureus infection. Three interim analyses were planned but the data monitoring committee terminated the study after the second interim analysis. Thus, only around ~5100 patients completed the full study with 90-day endpoints. Of note, 67% were male, 24% had diabetes, 18% were S. aureus colonized and 2% were MRSA colonized. Randomization appears excellent.

As I suggested, the results are sobering. The primary outcome (deep SSI or bacteremia) occurred in 2.6% of vaccinated patients and 3.2% of controls, p=0.58. There were more adverse events at 14 days in the vaccinated group (31% vs 22% ) and a greater incidence of multiorgan failure. Depressingly, there was far higher mortality secondary to S. aureus infection in the vaccinated group, 23 vs 4.2 per 100 person-years. The overall comparison between vaccinated and unvaccinated for S. aureus related outcomes is striking and humbling. I've pasted Table 4 from the study below, for your review (click on it to enlarge).

The authors of the study point out that their study isn't the first to find worse outcomes in a vaccinated group and also that MRSA infections were more common in the vaccinated arm of the trial. But I don't buy their suggestion that MRSA was a significant contributor to the poorer S. aureus related outcomes in the vaccinated group. There just wasn't enough MRSA disparity between the vaccinated and control groups. I wonder if part of the answer to higher mortality in the vaccinated group could be explained by this old paper by Heiman Wertheim and colleagues that I discussed several years ago. Wertheim report significantly lower mortality among S. aureus carriers vs. non-carriers. Could the vaccine be upsetting the delicate protective effect of pre-existing S. aureus colonization?

There is an excellent accompanying editorial by Preeti Malani and I urge you to read it. She does a wonderful job putting this study in historical context and pointing a path forward for HAI prevention research. The money quote from her editorial is: "Ultimately, the results of the study by Fowler et al say more about the need to rigorously study infection prevention in general than about S aureus per se." I couldn't agree more.

Image source: Vance Fowler, MD MHS from Duke

Going negative

Those of you that have had to sit through one of my "Business Case" talks know that one of my favorite books to recommend is William Ury's, "The Power of the Positive No." I probably should post more on the subject at some later point, but the take home message of the book is that you should say "no" to unreasonable demands by saying "yes" to yourself and your infection control team.

While flying back from San Diego I had a chance to read through two articles with a negative theme. The first was a negative study by Kyle Popovich and collegues at Stroger Hospital in Chicago published in the May Intensive Care Medicine. The group studied the benefit of daily chlorhexidine baths in a surgical ICU in preventing CLABSI. The results: negative, meaning that there was no benefit of CHG baths with a CLABSI rate between soap-and-water and CHG bathing periods (3.81/1,000 central line days vs. 4.6/1,000 central line days; p = 0.57). What I really liked about this study was that a group so highly linked to the benefits of CHG put the effort into reporting this negative study; this effort should not be underestimated. Both the journal and the authors should be applauded.

A related and I think very important article by Isabelle Boutron was published in JAMA this week. The group looked at how authors report and discuss results of randomized clinical trials with statistically non-significant results. It was interesting to see how they quantified the "spin" and how the Discussion sections didn't often match the results. As a side note, the senior author on the paper, Douglas Altman, has been an author on many great epi and stats papers, just search his name and BMJ in PubMed; he has made and continues to make wonderful contributions to clinical research.