Resistance is useless

Not just a Vogon catchphrase anymore, a new study in Lancet Infectious Diseases has some wondering about the effect of antibiotic resistance in bacteria. I’m actually a little uncomfortable with the way that infectioncontroltoday.com summarized the recently-published article by Lambert et al., titled “Clinical outcomes of healthcare-associated Infections.” Lambert et al. studied almost 120,000 ICU patients over a four-year time span in 10 European countries. Unsurprisingly, they found that patients who acquire a healthcare-associated infection have worse outcomes than those that don’t. However, they go on to say that if the infection acquired is resistant to narrowly-specified antibiotics, the fact of the resistance itself does not make the outcome much worse.

Specifically, the study finds that pneumonia doubles the risk of death and BSIs triple the risk of death, but when the bacteria causing those infections is antibiotic-resistant, there is only a 20% increase in the likelihood of death (results significant just for pneumonia). “Only.”

Now, I understand that 20% is not much when compared with 200%, and that that difference is what generated ICT’s summary title “Resistance Does Not Greatly Impact Mortality in ICU Patients with HAIs,” and that’s a valid viewpoint based on the study’s results. However, imagine someone coming up to you and saying “You have pneumonia; you’re now 200% more likely to die than before. By the way, given that awful news, would you rather be an additional 20% more likely to die or not?” Who would answer “Whatever, no great impact either way”? 20% doesn’t seem like much only because it’s being compared to 200%, but I’d definitely consider a 20% increase in number of deaths a significant impact.

Abstract, with link to full text for those with access: http://www.thelancet.com/journals/laninf/article/PIIS1473-3099%2810%2970258-9/abstract

ICT article: http://www.infectioncontroltoday.com/news/2010/12/resistance-does-not-greatly-impact-mortality-in-icu-patients-with-hais.aspx