Saturday, February 28, 2009


I spent a half day a week this month in preop clinic. While there, I took the time to go back and read the POISE trial that much ado was made about last year. POISE was a randomized placebo controlled, I believe double blinded study of extended release metoprolol in patients undergoing non-cardiac surgery. They had >8000 patients. MANY studies have shown the benefit of perioperative beta blockade in reducing perioperative MI, the internist's most feared perioperative complication.

POISE verified this finding, with reduction in nonfatal MI, but they discovered that more people died in the metoprolol group. Oops. What did they die of? Stroke. There was a 0.5% increase in stroke in the metoprolol group (statistically significant).

What the hell happened?

Well, if you look at their methods it is instantly obvious what went wrong. The treatment arm, whow as beta blocker naive, received a whopping 100 mg of extended release metoprolol 2-4 hours before their surgery. They chased it with another dose 6 hours postop, or earlier if they ad a HR > 80 or SBP >100. WOW! That's a lot of metoprolol for somebody who has never seen the drug before. 200 mg total of extended release metoprolol a day was then continued for patients that were able to tolerate it without severe bradycardia or hypotension. Gee, I wonder where the strokes came from? Any IM intern could have seen this coming.

Two take-home points here. Peri-operative betablockade doesn't need and shouldn't have the mega doses used in this trial. Also, if you have time before surgery, why not establish the drug for a week or two beforehand?

ACC/AHA 2007 guidelines for perioperative cardiovascular eval for non-cardiac surgery

Wednesday, February 4, 2009

Cough, Cough

Well, sometimes it seems that nearly every patient you see in an outpatient clinic at this time of the year has a URI. I find myself constantly giving instructions that may or may not be helpful: "Use a cough suppressant with dextromethorphan", "Use a nasal spray with oxymetazoline either during the day so you can breathe and sound normal at work" (the way I do it) or "use a nasal spray with oxymetazoline at night so you can breathe and sleep". My personal favorite: "Drink plenty of water." These conversations are even more common now that cold medicine is essentially off the market for kids. So I embarked on a literature search to figure out if any of this stuff is really helpful. I searched for:

"upper respiratory infection" "supportive care" efficacy

and came up with the article I refer to here. I had heard of this study before, but I had never read it.

The authors at Pennsylvania State randomly assigned 105 patients to receive either buckwheat honey, dextromethorphan thickened and artificially flavored like honey, or nothing. It is not clear if the "nothing" syringe contained a placebo, but I suspect it did. The study population was kids 2-18 with a diagnosis of URI. They asked parents about the patient's coughing frequency, severity, bothersome-ness, affect on patient sleep, and affect on parent sleep. This was surveyed on presentation and diagnosis with the URI and again after treatment. In paired comparisons, honey was superior to "nothing" for cough frequency and combined symptom score. It was also "marginally significant" for child sleep and bothersome nature of cough. It was not significant for cough severity or parents sleep. No difference was detected between dextromethorphan and honey.

Interesting stuff. Maybe I should be comfortable recommending honey for cough. They go on to hypothesize about how this might actually work. There are theories about sweetness leading to reflex salivation, secretion of airway mucous, and a "demulcent effect" in the pharynx and larynx. There is also a theory about these secretions improving mucociliary clearance in the airway via an expectorant mechanism.