Amoxicillin speeds resolution of acute sinus infections, but imperceptibly? (RCT, JAMA)

Amoxicillin may hasten the recovery from acute sinusitis (sinus infections), with more patients feeling just a tiny bit better after a week of amoxicillin compared to placebo, according to a clinical trial published in the February 15 JAMA. After 10 days, those taking placebo felt as well as those taking antibiotics.

You may have heard something else in the widely reported results of this trial (CNN: "Sinus infection? Antibiotics won't help"; CBS News: "Sinus infections not cured with antibiotics"). Authors downplayed the positive finding at 7 days, saying the difference was likely too small for patients to notice. The authors' bias here is defensible, since reporting the findings "straight up" could have exacerbated the problem they are not so-secretly trying to defeat by conducting this study: high rates of antibiotic prescription by primary care doctors for sinus infections.

What They Did: At 10 primary care physician offices in St. Louis, MO, Jane Garbutt et al enrolled 166 adults with bona-fide acute rhinosinusitis according to CDC criteria (maxillary pain, purulent nasal secretions, symptoms > 7 days and < 28 days, or getting worse after initial improvement) rated on a validated scale as moderate-to-severe. They randomized them to take either placebo or amoxicillin 500 mg t.i.d. for 10 days.

Results: There were no differences between groups in symptom scores at day 3 using SNOT-16 scores (yes, that's the real name of the tool). At day 7, however, 74% of patients taking amoxicillin reported symptom improvement, compared to 56% of those taking placebo (p=0.02). The score improvement was quite small: 0.2 points on the 3 point SNOT-16 scale, with 0.5 points being considered clinically significant. At day 10, there was again no difference between groups, suggesting those taking placebo got better on their own and "caught up" in the next few days to those taking antibiotics.

Garbutt et al parsed their results carefully, concluding only that "a 10-day course of amoxicillin compared with placebo did not reduce symptoms at day 3 of treatment." They pooh-poohed the finding of improvement with amoxicillin at day 7, saying it was too small to be meaningful to patients.

The news is bound to buoy the spirits of legions of primary care physicians who are secretly fed up and worn down by patients' insistence on being prescribed antibiotics for what are usually (~80%) self-limited infections. One in five antibiotic prescriptions in the U.S. are now written for sinusitis, and the CDC has warned that penicillin resistance is on the rise, believed partly due to such overuse of antibiotics. First author Garbutt tried to throw those hassled docs a bone, saying in an interview:

"We hope this study provides scientific evidence that doctors can use with patients to explain that an antibiotic is not likely to help an acute sinus infection."

Clinical Takeaway: It would be great if we could stop antibiotic overuse. But who defines "overuse?" Several studies and a meta-analysis have shown a faster resolution of sinusitis with antibiotics (although other studies have not). And can we really extrapolate this study's lack of proven benefit to be true for all antibiotics, as Garbutt's quote seems to assume? If I'm a patient reading about this, all it says to me is that amoxicillin doesn't cure sinus infections. And that patient might ask, What, you want me to be the 20% who gets worse, Doc? Can't you give me a better antibiotic?

This is a single study showing a lack of benefit of a single antibiotic. It's a public relations victory for antibiotic stewards (whose mission I support), and should help primary docs make their case for not writing antibiotic prescriptions they feel are unneeded. But the authors and the press went a bit beyond the data (and beyond CDC's guidelines) by suggesting that all antibiotics should be withheld from all people with acute sinusitis.

Garbutt JM et al. Amoxicillin for Acute Rhinosinusitis. A Randomized Controlled Trial. JAMA 2012;307:685-692.