Upon returning from the annual Interscience Conference on Antimicrobial Agents and Chemotherapy, I wanted to share my view of what seemed to emerge as a theme from the many papers published on ear infection treatment.

First, consensus exists on the importance of a bulging tympanic membrane in the diagnosis of acute otitis media (AOM) and its differentiation from otitis media with effusion (OME). Most AOM experts appear to agree that antibiotic treatment is warranted and recommended for children with clear-cut AOM as distinguished by a bulging eardrum.

Watchful observation or placebo treatment should involve less ill, older children who are verbal enough to describe their pain accurately.

Second, in children with a bulging tympanic membrane, the chances for bacterial infection are high, probably around 90%–98%. These children are not the ones with an 80% or greater likelihood of “spontaneous resolution” of their ear infections at the same speed as those who receive antibiotics.

Third, the spontaneous resolution rate of bacterial otitis media depends on when you ask the question: 1–2 days into treatment, on days 3–5, on days 10–14, or on day 28.

Most of the symptomatic benefit of antibiotics occurs during the early days of treatment.

The rates of persistent effusion, measured later, will be lower if appropriate antibiotics are used.

Fourth, the effectiveness of antibiotic therapy should be gauged against the likelihood of resolution that would accompany placebo treatment of otitis media.

Research in this area involves ethical, medicolegal, and practical concerns. The patient populations who enroll in trials in which children could be receiving antibiotic, placebo, or possibly one or two tympanocenteses (ear taps) are almost certainly different from each other and different from what we see in everyday practice.

The situation is dynamic in terms of study populations, investigative sites, causative bacteria, and antibiotic resistance.

Thus, comparisons across studies and across time should be made with extreme caution; frankly, I don't think they should be made at all.

Absent new data to the contrary—and I know of none—among children who have AOM of sufficient severity to receive an ear tap, bacterial eradication by natural host defense occurs 3–5 days after the onset of symptoms in 20% of these patients with Streptococcus pneumoniae, in 50% of those with Haemophilus influenzae, and in 70% with Moraxella catarrhalis. These numbers have been confirmed in multiple studies.

The bacterial profile for AOM in the United States has changed significantly because of the conjugate pneumococcal vaccine (Prevnar).

In vaccinated children, the No. 1 bacterial species is now H. influenzae, and more than half of those organisms make β-lactamase, rendering them resistant to the current first-line antibiotic choice, amoxicillin.

In fact, it appears that about 60% of AOM in Prevnar-vaccinated children involve H. influenzae.

But we can't ignore S. pneumoniae, which makes up about 30% of the total bacterial burden.

Although most of those strains are now penicillin susceptible because of the vaccine, the resistant strains are still around. The ratios may change with time, but S. pneumoniae is more worrisome because of its invasiveness and suppurative complications.

Finally, at least for the moment, we see a clearer distinction emerging among antibiotics as “good, better, and best” for anticipated effectiveness and for tolerability in the current U.S. pathogen mix. (See table below.)

Of course, there are caveats to these distinctions.

Measures of efficacy include bacterial cure (by double tympanocentesis study designs) and pharmacokinetic/pharmacodynamic antibiotic measurements of the key pathogens in middle-ear fluids.

Although I'm tempted to add “tolerability to the pocketbook (monetary cost)” and “annoyance cost of phone calls from the managed care policy police” as measures of tolerability and adherence to the prescribed regimen, the tolerability measurements in the table refer to taste, the number of doses per day, the duration of treatment, and the number of office visits involved.

Not included in the table are cefixime and ceftibuten because these agents are not satisfactorily effective against penicillin-nonsusceptible S. pneumoniae. (Combined with high-dose amoxicillin, however, such a combination would be expected to work very well.)

Also not listed are cefaclor and loracarbef because their efficacy, by current standards, has not been tested and they are anticipated not to be that good, although the tolerability is excellent.

Now the debate will continue around what to do in practice: Do we start with a good antibiotic, then go to a better one? Or start with a better one, and then jump to the best? Or start with the best and go to other “bests”—or to ear taps or tubes?

The best antibiotics in efficacy are not the same as the best in tolerability, so which one takes precedence? The most effective antibiotic won't work if it is not taken, and the most tolerable antibiotic won't work if it is not effective.

Stay tuned for the next chapter.

Good, Better, Best: Antibiotics for AOM