In pre-antibiotic times, acute otitis media often resulted in serious infectious complications such as mastoiditis, meningitis, and sepsis. Experts often cite the statistic that it accounted for 30 percent of pediatric hospitalizations at Bellevue Hospital in the 1930s. (J Pediatr 1939;14:730.) With this historical perspective, it seems a little counterintuitive that a strategy of watchful waiting (initially withholding antibiotics for uncomplicated cases) is not only widely practiced but endorsed by the American Academy of Pediatrics, the American Academy of Family Physicians, and the American Academy of Otolaryngology (Pediatrics 2004;113:1451), especially in this era of superbugs and extra-strength antibiotics.
A recent meta-analysis showed present day otitis media left untreated usually resolves on its own with suppurative complications occurring in only 0.12 percent of patients when antibiotics are withheld. (Laryngoscope 2003;113:1645.) Even more recently, a trial of ED patients reported that a wait-and-see approach resulted in significantly less antibiotic use without apparent increase in bad outcomes. (JAMA 2006;296:1235.) Perhaps widespread use of childhood vaccinations has had a role in taming the savage pediatric acute otitis media scourge of the 1930s. (Pediatr Infect Dis J 2009;28:e271.) Critics of this body of literature would be correct in highlighting the numerous methodological limitations of these studies. But watchful-waiting advocates would likely point out that if withholding antibiotics for routine acute otitis media was indeed harmful, we'd be up to our ears with complications by now.
Treatment of Acute Otitis Media in Children Under 2 Years of Age
Hoberman A, et al
N Engl J Med
A Placebo-Controlled Trial of Antimicrobial Treatment for Acute Otitis Media
Tahtinen PA, et al
N Engl J Med
These two articles published this year in the same issue of the New England Journal of Medicine have the potential of turning the “watching ears” strategy on its collective head. The University of Pittsburgh trial enrolled 291 children between 6 months and 23 months. (N Engl J Med 2011;364:105.) Children were randomized to 10 days of amoxicillin-clavulanate vs. placebo. The authors reported a non--statistically significant trend toward more rapid recovery and a statistically significant reduction in both seven-day symptoms (p=0.02) and overall treatment failure (p<0.001).
The second trial from Turku University Hospital in Finland enrolled 319 children between 6 months and 3 years of age. (N Engl J Med 2011;364:116.) Children were randomized to seven days of amoxicillin-clavulanate vs. placebo. The authors of this trial reported significantly less treatment failure in the antibiotic group (44.9% vs. 18.6%; p<0.001).
An editorial accompanying the articles concluded that these two studies settled the seemingly controversial issue once and for all, and that children with otitis media recover more quickly when given appropriate antibiotics. Looking deeper into these trials, however, might lead to a different conclusion.
In the Pittsburgh trial, a statistically significant reduction in parent--reported symptoms was found in the antibiotic group at seven days. The absolute difference between the groups, however, was only six percent (80% vs. 74%; p=0.02). Additionally, only 58 percent of eligible parents consented for their children to participate. One shouldn't overlook the possible impact on the study's significance that the 42 percent of nonparticipants might have had, especially when the difference in the measured endpoint is so small. If one considers that these same parents who chose not to participate might have intrinsically been prone to grade their children's symptoms differently in some way (not too much of a stretch), the impact on the study results could be far greater.
Another factor worth considering is that “treatment failure” in this trial was defined independent of symptoms and based solely on the persistence of otoscopic findings. Many of these “treatment failures” actually improved clinically, bringing into question how many patients truly failed therapy. It should be mentioned that two authors in the Pittsburgh trial disclosed honoraria and grant support from the manufacturer of the chosen study antibiotic. Finally, and perhaps most importantly, one must weigh the potential benefits of therapy with the incurred side effects. The incidence of diarrhea in the antibiotic group was more than three times that of the placebo group (24% vs. 7%).
In the Finland study, diagnostic criteria for acute otitis media were extremely well-defined and included three overall criteria and at least three out of five tympanic membrane abnormalities. These criteria are much more stringent than typically used in a busy ED, and are arguably much more specific for bacterial acute otitis media. The potential to show a benefit from antibiotics using these criteria is much higher. In the real world, acute otitis media is largely overdiagnosed, and often based on a combination of less-than-ideal factors: cursory otoscopy, no insufflator available, uncooperative children, impacted cerumen, and no luxury of performing a time-intensive research-protocol exam. These much less certain diagnoses are much less likely to benefit from antibiotics. Finally, if you don't buy the argument that true acute otitis media in the ED is a challenging diagnosis to make (or alternatively if you are indeed the best otoscopist in your practice, city, or state), it is notable to point out that acute otitis media resolved in two-thirds of the Finland trial placebo group without any rescue therapy whatsoever!
Both articles were also right to point out the potential risks of promoting antibiotic resistance by unnecessarily prescribing antimicrobials.
These two articles are certain to make waves in the medical community and headlines in the lay press. But perhaps we shouldn't let them cloud our judgment or cause a rush to change practice. To me, the management of the acute otitis media is still rather clear. As before, reserve upfront antibiotics for acute otitis media when patients are very ill or when the clinical suspicion is extremely high, and when strict criteria are used to make the diagnosis. More importantly, for uncomplicated acute otitis media or equivocal diagnoses, continue your literature-based, time-validated, academy-endorsed strategy of watchful waiting, and let these renewed pleas for extra-strength antibio-tics figuratively fall on deaf ears..
Comments about this article? Write to EMN at firstname.lastname@example.org.