Friday, September 28, 2012

"Say Anything", Regardless of the Data

As we've learned from prior publications, the conclusions section of the abstract is the ideal location to "spin" your article to generate news releases.  This article, from JAMA Neurology, compares thrombolysis to endovascular intervention for acute carotid artery occlusions and states "Intravenous thrombolysis should be administered as first-line treatment in patients with early cervical ICA occlusion."

That's a pretty strong statement, without qualifiers.  And, it means it received press coverage from MedPage Today, the ACEP News network, etc.

And, they base that statement on retrospective review of a cohort of 21 patients, 13 of whom received thrombolysis and 8 of whom received endovascular intervention.  The tPA patients did better, done and done, OR for early neurologic recovery 77 (95% CI 3 to 500).  But, then, Table 2 is a mini-systematic review of prior studies – and it turns out the rate of neurologic recovery is more like 40-50% with endovascular treatment, not the 1 in 8 they found in their retrospective cohort.  Indeed, the authors go on to state in the article "These findings are in contrast to the results of previous studies", and have an entirely reasonable, non-conclusive discussion of their findings in context of the other daa.

But, if you want news coverage, say something "interesting" in your abstract.

"Stroke From Acute Cervical Internal Carotid Artery Occlusion"
http://www.ncbi.nlm.nih.gov/pubmed/23007611

Wednesday, September 26, 2012

Acetaminophen and Asthma

If this article strikes your fancy – then you'll never look at acetaminophen the same way again.


Published in Pediatrics, this is a bit of a commentary summarizing epidemiological data in both children and adults related to the association between acetaminophen (paracetamol) use and asthma.  Specifically, that there is one, based on the studies he reviews, including:
• A prospective childhood asthma study of 520,000 subjects suggesting a dose-response effect between acetaminophen and asthma in children, up to an increased OR for wheezing of 3.25 for children taking acetaminophen at least once a month.
• A meta-analysis of six pediatric studies with a pooled increased OR for wheezing of 1.95 related to acetaminophen use.
• A meta-analysis of six adult studies with up to an increased OR for asthma of 2.87 for adults taking acetaminophen weekly.

...and several others.  The author does not suggest any specific mechanism through which acetaminophen increases airway reactivity, but he has changed his practice to reduce acetaminophen usage to the minimum.  I can't say I disagree with his hypothesis, and there does appear to be a preponderance of mounting evidence, although I wouldn't say this is an area where I am intimately familiar with the literature.


"The Association of Acetaminophen and Asthma Prevalence and Severity"
www.ncbi.nlm.nih.gov/pubmed/22065272

Monday, September 24, 2012

Don't Believe The Data

This NEJM study published a couple days ago addresses the effect of funding and methodological rigor on physicians' confidence in the results.  It's a prospective, mailed and online survey of board-certified Internal Medicine physicians, in which three studies of low, medium, and high rigor were presented with three different funding sources: none, NIH award, or industry funding.

Thankfully, physicians were less confident and less likely to prescribe the study drug for studies that were of low methodological quality and were funded by industry.  Or, so I think.  The study authors – and the accompanying editorial – take issue with the harshness with which physicians judge industry funded trials.  They feel that, if a study is of high methodological quality, the funding source should not be relevant, and we should "Believe the Data".  Considering how easy it is to exert favorable effects on study outcomes otherwise invisible to ClincalTrials.gov and the "data", I don't think it is safe or responsible to be less skeptical of industry-funded trials.

Entertainingly, their study probably doesn't even meet their definition of high rigor, considering the 50% response rate and small sample size....

"A Randomized Study of How Physicians Interpret Research Funding Disclosures"
www.ncbi.nlm.nih.gov/pubmed/22992075