In the supplement to Contemporary Pediatrics, "Management of conjunctivitis: Diagnosis and treatment of bacterial disease" (November 2000), Dr. Francis Gigliotti states that the 73% incidence of otitis media in children with conjunctivitis reported in my study (Bodor FF: Pediatrics 1982;69:695) is out of line with other studies, and that no more than a quarter to a third of children with bacterial conjunctivitis have concurrent otitis media. By "other studies" he means the study by his own group (Gigliotti F et al: J Pediatr 1981;98:531) and the one by Harrison and colleagues (Harrison CJ et al: Pediatr Infect Dis J 1987;6:536). In making this conclusion he apparently overlooks the different methods used in selecting subjects for the three studies.
In the study by Gigliotti et al, 90% of subjects were from a six-person pediatric practice group; 10% were walk-in patients. Patients who had received systemic or topical antibiotics during the previous week were excluded. Patients with otitis media whose conjunctivitis had cleared spontaneously in the previous week also were not included. Furthermore, the report does not indicate whether subjects with conjunctivitis alone who had received topical therapy were followed up for the possible development of otitis media within the following 10 days.
In the study by Harrison et al, patients also were recruited from a pediatric group practice. Of 130 patients, 47 presented with conjunctivitis-otitis media syndrome and 83 with conjunctivitis alone. Of the 83 with conjunctivitis alone, 42 were randomized to placebo and 41 were treated with amoxicillin for 10 days plus topical bacitracin-polymyxin ointment for seven days. Eleven of the placebo group and two of the amoxicillin group developed otitis media. We might predict, therefore, that, if neither group had received treatment, approximately 22 of the 83 patients with conjunctivitis only would have developed otitis media. If these 22 are added to the 47 initially diagnosed with conjunctivitis-otitis media syndrome, this would make 69 patients (53%) who had conjunctivitis associated with otitis media. If Harrison had also included patients with otitis media who had had bacterial conjunctivitis (treated or not) in the week before the study, his figures would have been closer to those in my study.
In my study, the great majority of subjects came from my private practice. In addition to patients who initially presented with both components of the syndrome, all patients who had purulent conjunctivitis alone and developed otitis media within 10 days, as well as patients with a history of purulent conjunctivitis in the week before the study, were included.
For example, a mother of three children called my office, asking for a prescription for one of her children who had "pink eye." The receptionist advised her to bring the child to the office for examination of a possible eye and ear infection. She came in with all three children, and stated that the other two had had red eyes with yellow discharge in the previous week; that their condition cleared without treatment; but that they still all had a "bad cold." I was able to examine the three children; all had otitis media. Haemophilus influenzae type B was recovered from cultures of the conjunctivae of all three.
Based on my studies (Pediatrics 1982;69:695; Pediatrics 1985;76:26; Pediatr Infect Dis J 1989;8:287), I firmly believe that every child with bacterial conjunctivitis must have otoscopy at the time of initial examination, as well as follow-up otoscopy within seven to 10 days if a systemic ß-lactamase-resistant antibiotic is not used. If, for any reason, this is not feasible on the day the parent calls, an ophthalmic topical antibiotic may be used until otoscopy can be performed the following day. However, if appropriate systemic antibiotics are used for the treatment of conjunctivitis-otitis media syndrome caused by ß-lactamase-positive (resistant to ampicillin) H influenzae, then a topical antibiotic is unnecessary (Bodor FF: Pediatr Infect Dis J 1989;8:287). Because of a high incidence of ampicillin-resistant H influenzae in many areas of the United States, amoxicillin should not be used for the treatment of conjunctivitis-otitis syndrome unless culture of the conjunctivae shows that the microorganism is susceptible to ampicillin.
Last, I emphasize that H influenzae conjunctivitis is highly contagious; if a sibling or other child in the same family develops otitis media at approximately the same time, then that child should receive the same antibiotic coverage as the child with both components of the syndrome.
The author replies: Dr. Bodor disagrees with my statement that approximately a quarter to a third of patients who present with conjunctivitis have concurrent otitis media. In support of his stance, he reviews his published experience and reanalyzes published observations of Harrison et al.
I have written and spoken extensively on this topic. Each time, in preparation for my undertaking, I review the literature. Having read Dr. Bodor's paper many times, it was never clear why so many of his patients had concurrent conjunctivitis and otitis. I believed there must have been ascertainment bias in his patient population to account for his extraordinarily high (73%) incidence of concurrent conjunctivitis and otitis not confirmed by other investigators. Dr. Bodor's letter now clarifies this issue. The patients he gives as examples do not really meet the definition of concurrent conjunctivitis and otitis. His report and reanalysis of the paper by Harrison would be more accurately characterized as a description of the natural history of conjunctivitis and otitis. In contrast, my colleagues and I and Harrison described the incidence of concurrent conjunctivitis and otitis at presentation. I stand by my statement that a quarter to a third of patients who present to a physician with conjunctivitis have concurrent otitis.
With regard to more substantive issues, I agree completely with Dr. Bodor's observations. For example, in reviewing our experience and that of Dr. Bodor, I wrote, "Thus it is important to examine the tympanic membranes of every child presenting with conjunctivitis" (Contemporary Pediatrics 1986;3:75). Likewise, his observations on treatment are consistent with our earlier placebo-controlled trial of antibiotic management of conjunctivitis (Gigliotti F et al: J Pediatr 1984;104:623).
Office-based research has contributed significantly to our knowledge of pediatric infectious diseases. Clinicians like Dr. Bodor should be commended for their participation in office-based research. I would make a plea, however, that, for this type of data to be interpreted correctly, authors should be precise about how they report their materials, methods, patient populations, and results.
Readers' Forum. Contemporary Pediatrics 2001;6:133.