The View from the Office
EAR INFECTIONS IN CHILDREN: TO TREAT OR NOT TO TREAT?
A brief history of ear infections (otitis media) is in order before going on to intelligently review the current state of the art. Once upon a time that a few of us still alive well remember, there were no antibiotics, and recurrent ear infections in children were a serious problem. Untreated, they could lead to mastoiditis, meningitis, and even death. When antibiotics came along, the attitude was "Of course, we'll treat. This is a bad disease." And treatment worked. Mastoiditis became a rare problem. Meningitis became even more rare, especially after new vaccines. Ultimately, ear infections in children became a very benign and simple problem. For the last 50 years or so, there have been hundreds of studies comparing every new antibiotic made by a pharmaceutical company to the old traditional ones trying to show that the new one was better in some way--either it cured more cases or had fewer side effects. The general lesson we have learned from these studies is that the choice of antibiotic doesn't make much difference, so you just use the least expensive one.
Then, a funny thing happened in the 1970s. Since otitis media was becoming a minor illness with just a few days of fussiness and moderate discomfort which would in the majority of cases go away even if it was not treated, European investigators started asking the question or whether you really need antibiotics at all for this newly less serious infection. The Netherlands pioneered this effort, and there it became standard policy NOT to treat ear infections with antibiotics at the first visit. At the first visit doctors would just recommend tylenol and analgesic drops (like Auralgan) in the ears. Then parents were given instructions that if symptoms were not much better by 3 days, they should return to get antibiotics. It turns out that this policy worked very well. The overall cure rates were the same as in the US, where antibiotics were being used routinely, and there were no increases in complications like mastoiditis.
It took a while for this approach to drift over to the US, but in the 80s and early 90s there were a number of studies that compared antibiotic treatment of otitis media to the outcomes in children who were not treated with antibiotics. What these studies showed consistently is that there was a small, statistically significant, short-term benefit in the children who were treated with antibiotics. The precise benefit was one day less of symptoms (shorter duration of illness) in 1 out of 7 children treated. This was shocking, of course, to physicians who were trained to treat ear infections. Thus, despite the very clear literature, only a minority of physicians actually stopped prescribing antibiotics; old habits change hard. In my practice I have always preferred to share the evidence with the patients (in this case, the parents) and let them make their own decision. I neither urged nor discouraged antibiotic use. Instead I asked the parents if they would want to use antibiotics for their child for the sake of possibly (a 1 in 7 chance) having one day less of symptoms at the risk of a drug reaction, most commonly a rash, upset stomach, or diarrhea. In my experience about 75% of parents chose to use the antibiotic.
Since the 1990s there have been very few studies of ear infections in children and absolutely no change in the bottom line--antibiotics confer a 1 in 7 chance of 1 day less of symptoms. That is why it is a noteworthy event that 2 recent studies have just been published in the New England Journal of Medicine on the treatment of ear infections in children--one in children less than 2 years of age in Pittsburgh PA, and the other in children aged 6 to 35 months in Finland. They both reach the same conclusion--that the benefits of antibiotic treatment are greater than has previously been estimated.
Let's take a look at those two studies and see how convincing this conclusion is.
In the under-2 years of age study, 291 children aged 6 to 23 months were diagnosed by experts using stringent criteria for the diagnosis of ear infection, and when an ear infection was diagnosed one group received a potent antibiotic (Augmentin) and the other group received a placebo. Among the children who received Augmentin (amoxicillin-calvulanate), 35% had initial resolution of symptoms by day 2, 61% by day 4, and 80% by day 7. This compared to rates of, 28% resolution by day 2, 54% resolution by day 4, and 74% resolution by day 7. These small differences were NOT statistically significant. For sustained relief of symptoms, the rates were 20% vs 14% at 2 days, 41% vs 36% at 4 days, and 67% vs 53% at 6 days. This result achieved borderline statistical significance (P value=0.04). Mean symptoms scores over the first 7 days were lower for the children treated with antibiotics than for those who received placebo (P=0.02). The rate of clinical failure--defined as the persistence of signs of acute infection on otoscopic examination--was also lower among the children treated with antibiotics than among those who received placebo: 4% vs 23% at or before the visit on day 4 or 5 (P < 0.001) and 16% vs 51% at or before the visit on day 10 to 12 (P < 0.001).
The second study randomized 319 children aged 6 to 35 months in Finland who were diagnosed by strict criteria as having ear infections to either placebo or the same antibiotic as above. Treatment "failure" occurred in 18.5% of the children who received the antibiotic vs 44.9% of the children were received placebo (P < 0.001). At 3 days after diagnosis, the failure rate was 13.7% in the antibiotic group vs. 25.3% in the placebo group. Overall, the antibiotic reduce the progression to treatment failure by 62%. Side-effects were much more common in the antibiotic group, and 47.8% of the treated children experienced diarrhea. The authors of this study comment, "Future studies should identify patients who may derive the greatest benefit, in order to minimize unnecessary antimicrobial treatment and the development of bacterial resistance."
An editorialist reviewing these two studies concludes "Is acute otitis media a treatable disease? The investigators in Pittsburgh and Turku, Finland have provided the best data yet to answer the question, and the answer is yes; more young children with a certain diagnosis of acute otitis media recover more quickly when they are treated with an appropriate antimicrobial agent."
Fig 1: An acute ear infection: a very bulging tympanic membrane.
To all this I have to say, "Not so fast!" First, the authors used, as is appropriate for a formal study, a "strict" definition of an ear infection. In the Pittsburgh study, the criteria for a diagnosis of ear infection were: (1) onset within 48 hours preceding the office visit; (2) a symptom assessment score of at least 3 on a scale of 0-14; the symptoms evaluated on a 3 point scale (none, a little, a lot) were: tugging on ear, crying, irritability, difficulty sleeping, diminished activity, diminished appetite, and fever. This is, in fact, admirably rigorous. The problem, however, is that out in the real world of general practice, primary care physicians do not do all this. Their history is very brief, no symptom scores are assessed, the ears are looked at very quickly, and lots of cases are diagnosed as ear infections that have no actual bulging of the tympanic membrane. What this means is that the group of patients being treated in ordinary offices is significantly different, probably with milder disease, than the patients included in the study. Thus we do not know if the conclusion from the formal study should apply to children with less severe illness.
Clinical "failure" was strictly defined as either a lack of clinical improvement (and this is what parents really care about) or worsening of signs on otoscopic examination (which parents don't care about). My perspective as a practicing physician is that, if the child is not having any symptoms, I don't care what the ear drum looks like. I would never extend or repeat an antibiotic just because the ear drum look abnormal, whether moreso or not. What matters is how the patient is doing.
Finally, both studies arbitrarily chose to use an advanced antibiotic (Augmentin). This is a powerful (probably more powerful than one needs) and effective antibiotic, but it causes a very high rate of gastrointestinal side-effects. In one studied it caused diarrhea in almost half of the children treated. This would mean that treatment is like trading one problem (an ear infection) for another (diarrhea), and it is hard to say which is worse. I can tell you which is messier. Certainly there is a need for additional studies, if we are going to making a recommendation to change general practice, to determine which antibiotic will cause the least side-effects while retaining most of the effectiveness. There are many good candidates.
Thus in my opinion these studies don't change anything. It still comes down to a close call whether young children with mild ear infections should be treated. This decision should be made by the parents, not by the doctor. The doctor's role is continue to give parents the benefits of the latest information and help them to make a reasonable decision.
Another strategy that has been well-defined as safe and effective in multiple studies is for the physician not to treat at the time of the first visit, but to give a prescription for an antibiotic to use if the symptoms become worse or do not resolve in 2-3 days. This was called a "delayed prescription" for an antibiotic. Only about 50% of parents given a "delayed prescription" actually get them filled, which means that many, many children get better all by themselves. This strategy significantly reduces the use of antibiotics and thus will lead to less emergence of antibiotic resistance among common bacteria. I use this strategy a lot.
References: Hoberman A et al. (Pittsburgh) Treatment of acute otitis media in children under 2 years of age. NEJM 2011; 364:105-15. Tahtinen PA et al. A placebo-controlled trial of antimicrobial treatment for acute otitis media. NEJM 2011; 364: 116-26. Klein JO. Is acute otitis media a treatable disease? (editorial) NEJM 364: 168-9.