To the Editor:
We routinely review all articles that mention Ceclor (cefaclor, Eli Lilly and Co, Indianapolis, Indiana). Thus, we have recently reviewed an article that appeared in the November 1991 issue oí Pediatric Annals entitled "Antibiotic Treatment Failures in Acute Otitis Media" by Christopher J. Harrison, MD, and Thomas Belhorn, MD, PhD.1 We were concerned about misleading statements in that article related to our product with regard to: twice-daily dosing of Ceclor for otitis media, serum sickness-like reactions, and in vitro efficacy-zone diameters.
First, Ceclor has been approved by the Food and Drug Administration for use in a twice-daily dosage regimen for otitis media. Ample clinical data to substantiate this regimen can be obtained from Eli Lilly and Co. Additionally, a study published in the Pediatric Infectious Disease Journal provides evidence of continued clinical efficacy.2
Second, we would like to respond to the serum sickness-like reaction seen infrequently with Ceclor. In clinical trials, a serum sickness-like reaction was reported in 2 of 8346 (0.024%) patients receiving Ceclor.3 The incidence rate obtained by Levine4 was somewhat higher - 5 of 1017 (0.5%) in children 14 years of age or younger receiving multiple courses of Ceclor in a study performed during two winter seasons. In contrast, the data in the clinical trial database were collected over a 10-year period and involved subjects of all ages. Hence, the Levine study may describe an upper limit for expected incidence in a select population of all Ceclor recipients.
During the past decade, the efficacy and safety of Ceclor have been well-documented in more than 100 million courses of therapy. When a serum sicknesslike reaction does occur, full recovery can be expected. This reaction is not life-threatening and has not been associated with deaths or serious sequelae. The symptoms usually are not severe, but may infrequently require hospitalization. The median hospital stay is 2 or 3 days, with a range of 1 to 7 days.
The occurrence of serum sickness-like reactions associated with Ceclor has remained stable over the past several years. The authors' comment "The increasing recognition of serum sickness. . ." should not be equated with an increasing occurrence of this infrequent reaction.
Finally, Harrison and Belhorn call into question the efficacy of Ceclor against Moraxelfo catanhaUs; this was not referenced and is totally unfounded. The apparent increase in resistance of Haemophilus influenzae, noticed in some laboratories, was coincident with the adoption of new National Committee for Clinical Laboratory Standards (NCCLS) guidelines. These guidelines have recently been reviewed by NCCLS and revised recommendations are forthcoming. Additionally, Harrison and Belhorn inappropriately try to correlate these in vitro changes directly with diminished clinical efficacy. Ceclor remains a safe and effective antibiotic in the treatment of otitis media.
Jennifer L. Stotka, MD
Associate Clinical Research Physician
EU Lilly and Co
1. Harrison CJ, Belhorn TH. Antibiotic treatment failures in acute otitis media. Pea Annals. 1991;20:600-608.
2. Hendrickse WA, Kusmiesz H, Shelton S, et al. Five vs ten days of therapy for acute otitis media. Pedia» infect Dis]. 1988;7:14-23.
3. Platt R, Dreis MW, Kennedy DL, et al. Serum sickness-like reactions to amoxicillin, cefaclor, cephalexin, and trimethoprim/sulfamethoxazole. ] infect Dis. 1988;1 58:474476.
4. Levine LR. Quantitative comparison of adverse reactions to cefaclor vs amoxicillin in a surveillance study. Pediatr infect Dis. 1985;4:358-361.
The authors were asked to respond to this letter.
We read with interest the comments of Dr Stotka concerning our statements in our article, "Antibiotic Treatment Failures in Acute Otitis Media," which appeared in the November 1991 issue of Pediatric Annah. Dr Stotka took exception to three points.
First, Dr Stotka mentions our lack of enthusiasm for twice-daily dosing of cefaclor for otitis media and in vitro susceptibility of pathogens to cefaclor. She quotes "ample clinical data," all of which appears to be Lilly's in-house data and which to our knowledge dates from more than 5 years ago. The recent peerreviewed study evidencing its efficacy involved a relatively old population with histories of little previous difficulty with otitis media. Given a disease with a high degree of spontaneous remission where at least a third of the isolates are Streptococcus pneumoniae (spontaneous resolution would respond to placebo and S pneumoniae to penicillin) and the fact that the data are not recent (5 years old), we are not sufficiently comforted by these data to use cefaclor to treat today's pathogens in the young infant with recalcitrant otitis media.
Our concern with twice-daily dosing relates to two factors, one being the in vitro susceptibility of gramnegative otitis pathogens. The other is the rapid excretion and short elimination half-life of cefaclor. Because Dr Stotka quotes Lilly's data on file without providing data for perusal, we feel we must present our data although some have yet to be published in peer-reviewed forums. We have recently isolated nontypable Haemophilus influenzae and Moraxella catanhaUs from specimens from clinical studies undertaken in Bardstown, Kentucky; Cincinnati, Ohio; and Omaha, Nebraska with MIC90 values that are higher than those previously seen. These MIC90 values exceed concentrations likely to be achieved in middle ear fluid even though they remain at or below the National Committee for Clinical Laboratory Standards (NCCLS) approved breakpoints for susceptibility.
When one uses the recommended 8 µ?^p?? to define "susceptible," organisms are classified as susceptible based on achievable serum concentrations. Because serum concentrations cannot be expected to occur in middle ear fluid, this susceptibility is irrelevant when treating otitis media. The concentrations of beta-lactam antibiotics and cephalosporins in middle ear fluid would not be expected to reach more than half of that seen in serum. Therefore, with expected average cefaclor serum concentrations of 8 ^g/mL, one might expect ^4 µg/mL to reach the middle ear. Thus, if we used 2 to 4 µg/mL as the breakpoint (based on putative middle ear concentrations), many of these same isolates would not be considered susceptible. Therefore, our concern over increased resistance is not related to revised NCCLS guidelines, but to changes in the MIC90 in the most common gram-negative pathogens causing respiratory infections from several geographic locations in relation to achievable concentrations at the site of infection.
Sophisticated clinicians realize that sufficient drug is required at the site of infection to reach twice and preferably four times the MIC of the pathogen, thus providing the best chance to achieve a good clinical outcome. As an example, consider that cefaclor exceeds 2 ^gJmL in serum for less than 3 hours after each dose, and the probability is that only half of this peak concentration reaches the ear. Two doses daily yield serum concentrations of 2*2 µg/mL for at most 6 hours and three doses for at most 9 hours.
This is important because 2 to 4 µg/mL is a conservative estimate of the MIC90 for the gramnegative pathogens causing otitis media in 1991 and 1992. Using the identical methodology and formulation for medium in 1990 to 1992 as used in 1984 to 1985, we observed increases in the MIC90 for both beta-lactamase nonproducing (BL-) and betalactamase producing (BL+ ) H influenzae from those in 1985 (Table).
In addition, the relative frequency of betalactamase production has increased from ^45% to >90% for M catarrhalis. These two findings have decreased our enthusiasm for cefaclor, which remains an expensive antibiotic. The extra expense should hopefully include extra potency. However, the MIC90 for H influenzae and M catarrhalis in relation to achievable drug concentrations and the elimination half-life for cefaclor are less favorable than for extended spectrum cephalosporins, amoxicillin plus clavulanate or even Lilly's new experimental drug. Therefore, we feel that if one intends to prescribe an extended spectrum antimicrobial for otitis media, it is best to choose one with a reasonably long half-life (at least 12 hours/day with serum concentrations at least twice the MIC90) and a truly extended spectrum within the middle ear fluid. Cefaclor does not meet these criteria for M catanhaUs and H influenzae even with three-times-a-day dosing, while routinely recommended dosing of Cefixime, cefuroxime axetil, and amoxicillin/clavulanate (and several soon-to-bereleased compounds) appear to.
The second point Dr Stotka questioned concerned serum sickness-like illnesses due to cefaclor. We did not state that the rate of these reactions was increasing but that clinicians were recognizing these reactions more often. We did not state that the rate is high. Even so, the reaction rate with cefaclor is 100 times higher than with other antibiotics used for otitis media.1 It is unclear what provoked Dr Stotka's undue sensitivity to our straightforward statement in this regard.
The last point that Dr Stotka took exception to regarded concern over efficacy against M catanhaUs. Given the increase in the proportion of isolates of M catanhaUs that produce beta lactamase and the MIC90 of these isolates (4 µ^p??) and the putative middle ear concentration of cefaclor (2 to 4 igltaiX it is difficult to be enthusiastic about the chances that cefaclor will exceed the MIC90 of M catanhaUs at the site of infection in the middle ear with any regularity, whereas the same data for extended spectrum drugs now or soon to be marketed (such as Cefixime, cefuroxime axetil, and amoxicillin/clavulanate) appear adequate.
In summary, if the breakpoint of 2 to 4 µ-g/mL is used (analogous to putative middle ear concentrations of cefaclor), then die nearly 95% susceptibility of H influenzae and M catanhaUs to cefaclor is reduced to <25% of BL + H influenzae (N = 21 ), <50% of BL - H influenzae (N = 39), and approximately 50% of BL+ M catanhaUs (N = 26) using our data from isolates obtained in pediatric clinical studies from three divergent geographical areas between 1990 and 1992. If these gram-negative pathogens make up 50% to 60% of middle ear isolates (as is the case in our patients), then one would expect that one fourth to one third of otitis patients could fail therapy based on pharmacokinetic data and on in vitro susceptibility patterns taking into account middle ear antibiotic concentrations. In the late 1970s and early 1980s, cefaclor had major advantages over the drugs available at the time. More potent antibiotics have since been released and cost little if any more than cefaclor. These can be used with more confidence in the 1990s, especially when BL + organisms may be involved.
We feel it is important to share some of our data in response to Dr Stotka's concerns in order to remove any disquiet that may have arisen from them.
Christopher J. Harrison, MD
Department of Pediatrics and Medical Microbiology
Creighton University and University of Nebraska
1, Piatt R1 Dreis MW, Kennedy DL, et al. Serum sickness-like reactions to amoxicillin, cefaclor, cephalexin, and trimethoprim-sulfamethoxazole. J infect Dis. 1988:158:474477.
We would like to thank the following people, who were involved in the work we used to generate the above table: Stephen A. Chartrand, MD; Jay R. Pollack, PhD; Shirley Reising, PhD; James A. Hedrick, MD; Stanley Block, MD; and Warren Webb, MD.