Cefixime compared with amoxicillin for treatment of acute otitis media

PEDIATRIC PHARMACOLOGY AND THERAPEUTICS

Cefixime compared with amoxicillin for treatment of acute otitis media C a n d i c e E, Johnson, MD, PhD, Susan A, Carlin, MD, Dennis M, Super, MD, MPH, James M, Rehmus, MD, David G, Roberts, MD, Norman C, Christopher, MD, Judy K, Whitweli, RN, a n d Paul A, Shurin, MD* From the Department of Pediatrics, Case Western Reserve School of Medicine, MetroHealth Medical Center, Cleveland, Ohio Cefixime was compared with amoxicillin for treatment of acute otitis media in a randomized trial. Results of tympanocentesis on day 3 to 5 of therapy were used as the major outcome. Total daily doses were 8 mg/kg of cefixime and 40 m g / k g of amoxicillin. One hundred twenty-six patients were randomly assigned to receive treatment; 64 cultures grew pathogens. Pathogens were eradicated from the middle ear after 3 to 5 days of therapy in 27 (79.4%) of 34 children given amoxicillin and 26 (86.7%) of 30 children given cefixime (p = 0.47). When Strept o c o c c u s p n e u m o n i a e cases were analyzed, bacteriologic cure occurred in 14 (93.3%) of 15 children given amoxicillin and 12 (75%) of 16 given cefixime (p = 0.333). When cases of Haemophilus influenzae infection were analyzed, significantly more cures occurred with cefixime (10/10, 100%) than amoxicillin (8/13, 62%) (p = 0.046). Pathogens associated with failure of amoxicillin therapy were 14. influenzae (five cases, two ~-lactamase-positive), $. p n e u m o n i a e (one case), and Moraxella catarrhalis (one case, ~-lactamase-positive). The four failures with cefixime therapy were all in patients infected with S. pneumoniae. Rates of rash, diarrhea, and vomiting were the same in both groups and did not necessitate stopping therapy. We conclude the following: (I) Cefixime and amoxicillin were equivalent in overall clinical and bacteriologic efficacy for otitis media. (2) Cefixime was more efficacious than amoxicillin in treating/4. influenzae otitis media and should be preferred when 14. influenzae is the suspected etiologic agent. (3) Side effects of both drugs were mild and equivalent. (J PEDIATR1991;119:117-22) Because of the increasing frequency of ~3-1actamase-producing strains of Haemophilus influenzae 1 and Moraxella (previously Branhamella) catarrhalis 2 in respiratory tract infections, there have been many recent studies of new cephalosporins that are/3-1actamase stable. Although inSupported by Lederle Laboratories, Pearl River, N.Y. Presented at the Ambulatory Pediatrics Society, Anaheim, California, May 1990 (abstract published in Am J Dis Child 1990;144:421 ). Submitted for publication Dec. 13, 1990; accepted Feb. 14, 1991. Reprint requests: Candice E. Johnson, MD, PhD, Department of Pediatrics, MetroHealth Medical Center, 3395 Scranton Rd., Cleveland, OH 44109. *Now at College of Physicians and Surgeons, Columbia University, New York, N.Y. 9/25/28733

creasing rates of resistance of H. influenzae to amoxicillin in vitro have been documented in nationwid e surveys, 1 there is little published evidence of an increased clinical or bacteriologic failure rate for amoxicillin in otitis media. We compared amoxieillin with cefixime, a new/3-1actamasestable cephalosporin, using the results of a second tymMIC

Minimum inhibitory concentration

[

panocentesis on day 3 to 5 of therapy as the major outcome; this tympanocentesis has been referred to by Howie et al. ~ as the "in vivo sensitivity test in otitis media," METHODS Patient selection. Children from 8 weeks to 13 years of age attending the MetroHealth Medical Center Pediatric 117

118

Johnson et al.

The Journal of Pediatrics July 1991

T o b l e I. Patient population in the two treatment groups

Mean age (mo) Sex (%) Male Female Race (%) White Black Ears affected (%) Unilateral Bilateral Fever (%) Present Absent Previous episodes of otitis media (%) >3 --<3

Amoxicillin (n = 62)

Cefixime (n = 64)

19.9 _+ 3.5

16.3 _+ 2.2

61 39

56 44

74 26

72 28

37 63

41 59

28 72

4l 59

26 74

27 73

Clinic were eligible for the study if they had symptomatic otitis media. The committee on investigation in humans approved the study, written informed consent was obtained from a parent, and assent was obtained from children older than 7 years of age. Acute otitis media was defined as a middle ear effusion confirmed by two observers using pneumatic otoscopy, with a history of either fever, earache, or irritability of recent onset. Patients were excluded for (1) a suspected allergy to penicillins or cephalosporins, (2) an underlying serious chronic disease or immunodeficiency, (3) the presence or history of pressure-equalization tubes or eraniofaciaI defects, (4) treatment with another antibiotic during the preceding 3 days, or (5) having a non-Englishspeaking parent. Stratification was performed by age (--<2 years and >2 years), race (white, nonwhite), number of episodes of otitis media (-<3 or >3), and laterality (one versus both ears involved). Children were then randomly assigned to receive eefixime at 8 mg/kg per day or amoxicillin at 40 mg/kg per day. The first 19 children to receive eeflxime were given two divided daily doses, and the last 45 children were given one daily dose. This change in dosing was made because other trials showed equivalent efficacy of once- and twice-daily dosing (data on file at Lederle Laboratories). The investigators and study nurses were unaware of drug group assignment, and medication was dispensed by a pharmacist. Bottles were labeled, but parents were instructed not to inform the investigators of the name of the medication or the frequency of dosing. Bacterlologie methods. All children had external auditory canal cultures before tympanocentesis to identify contaminants. Sterilization of the ear canal was not performed. Staphylococcus epidermidis, diphtheroids, and Pseudo-

monas aeruginosa were classified as contaminants. Tympanic aspiration was performed using a 20-gauge spinal needle and Senturia trap (Storz), and both ears were aspirated in cases of bilateral disease. Specimens were cultured immediately onto sheep blood and chocolate agar plates. They were also inoculated into modified Anderson broth, containing Difco Brain Heart Infusion and 1 ml of deftbrinated horse blood per liter, and 10 mg/L nicotinamideadenine dinucleotide (Sigma Diagnostics, St. Louis, Mo.). Mean inhibitory concentrations were determined by microtiter panels (Gibco Laboratories, Gaithersburg, Md.) and by Kirby-Bauer disk sensitivities. ~-Lactamase production was tested by the chromogenic cephalosporinase assay (Cefinase, Becton Dickinson Microbiology Systems, Cockeysville, Md.). 4 If both left and right tympanocenteses yielded the same organism, only one of each pair was included in analyses of MIC. Follow-up visits. Children returned at 3 to 5 days, 10 days, and 4 weeks after enrollment. Tympanocentesis was repeated on day 3 to 5 if a pathogen grew from the initial middle ear fluid. Those cases in which the pathogen was not eradicated by day 3 to 5 were considered bacteriologic failures, and a different antibiotic, selected by susceptibility testing, was prescribed. Clinical failure was defined as persistent fever, pain, irritability, or otorrhea on day 3 to 5, and these children were also given a different antibiotic. In no child with clinical improvement and bacteriologic success was the treatment regimen discontinued because of in vitro test results that showed resistance to the drug being administered. Adverse reactions were recorded at each visit. Side effects were defined as those beginning or worsening after drug therapy began and included all events within the 10 days of therapy. Diarrhea was defined as watery stools; mild diarrhea was defined as 1 to 2 stools per day, moderate as 3 to 6 stools per day, and severe as greater than 7 stools per day. Serum for hematologic, renal, and hepatic studies was obtained on days 1 and 10, and tests with abnormal results were repeated. Middle ear effusion was defined as markedly decreased tympanic membrane mobility or the presence of an air-fluid level. Two independent observers examined each ear, and tympanometry was performed at each visit. If the observers disagreed, tympanometry was used to define the presence of effusion. Tympanograms were interpreted as previouslY described. 5 Recurrence of otitis media was defined as new middle car effusion that appeared during follow-up or new symptoms of earache or fever in a child with persistent effusion. Clinical success was defined as absence of fever, otorrhea, earache, and irritability. Compliance monitoring. Compliance was evaluated by obtaining a urine or blood sample at the day-3 to day-5 visit. Cefixime assays were performed on an Escheriehia eoli American Type Culture Collection disk diffusion plate sen-

Volume 119 Number 1, Part 1

Cefixime therapy f o r acute otitis media

119

T a b l e II. Middle ear isolates from initial tympanocentesis according to treatment group Amoxicillin (n = 62)

S. pneumoniae H, influenzae ~-Lactarnase-positive 13-Lactamase-negative M. catarrhalis t3-Lactamase-positive fl-Lactamase-negative Group A streptococcus E. coli Sterile/nonpathogens*

Cefixime (n = 64) Total

Pure culture

Mixed culture

No.

13

4

5 8 5 1 1 1 21

Total

%

Pure culture

Mixed culture

No.

%

17

27.4

15

3

18

28.0

1 5

6 13

9.7 21.0

0 8

1 2

1 10

1.6 16.1

3 0 1 0 NA

8 1 2 1 21

12.9 1.6 3.2 1.6 33.8

5 1 0 0 31

1 0 0 0 NA

6 1 0 0 31

9.4 1.6 0 0 48.4

NA, Not applicable. *S. epidermidis, diphtheroids,and P. aeruginosa wereconsideredcontaminants.

T a b l e III. Number of isolates with the minimum inhibitory concentrations noted for cefixime and amoxicillin among isolates cultured from the middle ear before treatment No. of isolates in MIC (#g/L) Pathogen

Antibiotic

MIC <0.5

MIC t

MIC 2

MIC 4

MIC >8

S. pneumoniae

Cefixime Amoxicillin Cefixime Amoxicillin Cefixime Amoxicillin

35 35 30 12 14 3

0 0 0 9 2 0

0 0 0 2 0 2

0 0 0 0 0 0

0 0 0 7 0 11

H. influenzae M. catarrhalis

p"

NS <0.001 <0.001

NS, Not significant. *Wilcoxonrankedsumstest.

sitive to 0.017 #g/ml, and amoxicillin was assayed with a Bacillus subtilis disk diffusion method sensitive to 0.156 ~g/ml. Statistical analysis. Because the bacteriologic success rates were similar for twice-daily versus once-daily administration of cefixime (91.7% vs 83.3%; p = 0.58), the two groups were combined for all analyses. All outcomes were analyzed by the Fisher Exact Test (two-tailed). The MICs were compared by the Wilcoxon matched-pairs signed-rank test. RESULTS Study population. One hundred twenty-six children were enrolled from October 1986 to June 1989. The demographic characteristics of the two groups were comparable at entry (Table I). Fever was analyzed only if documented in our clinic and was therefore less commonly reported than in other studies. The mean age of the children at entry was 18 months, and 85% were younger than 2 years of age. One

fourth of the children had had three or more episodes of otitis media. Bacterial etiology. Pathogenic organisms were isolated from 74 (58.7%) of the children (Table II). Of the 30 isolates of H. influenzae, 7 (23.3%) were t3-1actamase positive, as were 14 (87.5%) of 16 strains of M. catarrhalis. No penicillin-resistant pneumococci were isolated. All 35 isolates of Streptococcus pneurnoniae were susceptible to both antibiotics (Table III). All 30 isolates of 14. influenzae were susceptible to cefixime, but 7 strains were resistant to amoxicillin because of/3-1actamase production. For M. catarrhalis, all 16 strains were susceptible to cefixime, but 11 of 16 were resistant to amoxicillin. The MICs of amoxicillin for M. catarrhalis strains that produced/3-1actamase were 0.5 #g/L (1 strain), 2 #g/L (2 strains), 8 #g/L (2 strains), and 16 t~g/L (9 strains). The MICs for both H. influenzae and M. catarrhalis were significantly lower (p <0.001) for cefixime than for amoxicillin. Bacteriologic outcome. Only patients whose cultures of

120

Johnson et al.

The Journal of Pediatrics July 1991

Table IV. Eradication of bacterial pathogens from middle ear on days 3 to 5 of therapy Amoxicillin n*

All patients with bacterial isolates S. pneumoniae H. influenzae /3-Lactamase positive /3-Lactamase negative TOTAL

Cefixime %

n'

%

p

86.7 75.0

0.469 0.333

27/34 14/15

79.4 93.3

26/30 12/16

1/3 7/10 8/13

33.3 70.0 61.5

1/ 1 9/9 10/'10

M. catarrhalis ~-Lactamase positive /3-Lactamase negative TOTAL

6/7 1_~ 7/8

85.7 100 87.5

Group A Streptococcus pyogenes

1/ 1

100

100 100 100

--0.046

5,/6 ill 6/7

83.3 100 85.7

--1,00

0/0

0

--

*Numberof patientswith eradicationof bacterial pathogensfrom middleear/numberof patientstreated.

Table V. Clinical outcomes according to drug treatment in children with bacterial otitis media Amoxicillin

Symptomatic improvement at 3 to 5 days (No. improved/No, examined) Middle ear effusion at 10 days (N0. with effusion/No, examined) Patients Ears Middle ear effusion at 4 to 5 wk (No. with effusion/No, examined) Patients Ears Recurrence of otitis media (No. recurrences/Nol examined)

Cefixime

n

%

n

%

p

30/34

88.2

27/30

90.0

NS

13/24 18/48

54.2 37.5

16/21 24/42

76.2 57.1

NS NS

10/20 16/40 3/20

50.0 40.0 15.0

15/20 25/40 3/20

75.0 62.5 15.0

NS NS NS

NS, Not significant.

middle ear fluid grew a pathogenic organism and who returned on day 3 to 5 for additional tympanocentesis were analyzed. One hundred ten children returned on day 3 to 5, including 64 (86.5%) of 74 children whose cultures were growing pathogens. Overall bacteriologic success rates were 79.4% for amoxicillin and 86.7% for cefixime (p = 0.469, two-tailed; Table IV), The 95% confidence limits were 65.8% to 93.0% for amoxicillin and 74.6% to 98.8% for cefixime. Serum or urine samples for compliance were obtained from 62 of these 64 children on the day-3 to day-5 visit, and all had detectable antibiotic. Both children from whom urine and serum were not obtained had bacteriologic and clinical successes. When all children seen on day 3 to 5 were analyzed, 105 (97%) of 108 had detectable antibiotic. There were four treatment failures with cefixime in S. pneumoniae otitis media; one of these patients was also in-

fected with M. catarrhalis. Only one child treated with amoxicillin failed to clear S. pneumoniae; however, the difference in success rates was not statistically significant (p = 0.333). There were five treatment failures with amoxicillin in patients with H. influenzae otitis media; two of these organisms produced/3-1actamase. All 10 children with 1t. influenzae otitis media treated with cefixime were cured (p = 0.046). One child in each group infected with M. catarrhalis had treatment failure. Clinical outcome. Of 57 children who were clinically improved, 6 (10.5%) had documented bacteriologic failures. In contrast, 5 of 7 children (71.4%) with clinical failure also had bacteriologic failure. Of these 7 children who were clinically unimproved, 2 had no growth on cultures after a second tympanocentesis, 3 had H. influenzae growing on cultures, and 2 had S. pneumoniae growing on cultures. Of the 64 children who underwent repeat tympanocente-

Volume 119 Number 1, Part 1

sis, 45 (70%) returned on day 10 and 40 (62.5%) returned at the 4- to 5-week visit. When the two treatment groups were compared for symptomatic improvement, presence of ear effusion, and rates of recurrent otitis media, no significant differences were noted (Table V); the trend toward fewer effusions in the amoxicillin group at both time points was not statistically significant. The most common side effect noted was diarrhea, present in 29.4% of patients given amoxicillin and 33.9% of patients treated with cefixime (Table VI). In no case did the severity of any side effect require stopping the study medication. Four children had transient increases in alanine aminotransferase values during treatment, but all four children had normal aspartate aminotransferase levels and alkaline phosphatase levels. Two of these children were receiving amoxicillin and two, cefixime. In one child, alkaline phosphatase lew:ls were five times above normal at presentation and fell to the normal range by day 10 of therapy. DISCUSSION Cefixime is a newly licensed oral cephalosporin that has been formuiiated to be active against all common respiratory and urinary pathogens. 6 It is resistant to both plasmid and chromosomal/3-1actamases, and it may be given only once daily. Three previously published studies have evaluated cefixime for treatment of otitis media in children.79 McLinn7 reported a multicenter study of cefixime compared with amoxicillin and found equivalent clinical cure rates (93% vs 94%). However, 34 of the 120 randomly selected children were excluded from analysis because the infecting organism was resistant to either drug in vitro. The causative agent is rarely known in clinical practice, so a true efficacy assessment cannot be done. Howie and Owen 8 performed a comparison of cefixime and amoxicillin in 140 children, wJith a second tympanocentesis as the end point. Although overall cure rates were similar, bacteriologic failure occurred in 7 of 19 pneumococcal cases. A larger series of cefixime cases is being prepared for publication, with a better success rate (V. Howie: personal communication, March 19911). The third study of cefixime for treatment of otitis media, by Kenna et al., 9 compared it with cefaclor and used clinical end points; no difference in cure or recurrence rate was observed. We failed to demonstrate a difference in overall efficacy between cefixime and amoxicillin. Because the difference in bacteriologic efficacy between the two drugs was only 7%, we would have needed to enroll more than 740 subjects to confirm a superior result with cefixime with an 80% power. Documentation of a significant difference in clinical outcome would have required an even larger number of subjects. We were able to demonstrate a significantlyhigher success rate for cefixime than for amoxicillin in the treat-

Cefixime therapy for acute otitis media

1 21

Table VI, Side effects of antibiotic therapy within 10 days of therapy in all children returning at the day-3 to day-5 visit Amoxicillin (n = 51) No."

Generalized rash Diaper rash Stool changes Loose stools Mild diarrhea Moderate diarrhea Vomiting Tinnitus

%

Cefixime (n = 59) No.'

%

p

1 4

2.0 7.8

3 10

5.1 16.9

NS NS

3 12 3 2 1

5.9 23.5 5.9 3.9 2.0

3 19 1 5 0

5.1 32.2 1.7 8.5 0

NS NS NS NS NS

*Numberof patients.

NS, Not significant.

ment ofH. influenzae otitis media (100% cure vs 62%). The apparent trend toward more failures of cefixime against pneumococci was not statistically significant; 146 subjects with pneumococcal otitis media would have been required to show significance with a power of 80%. The 79% cure rate for amoxicillin in this study is comparable to the 83% rate reported by Howie et al. 3, 10, 11 from 1972 to 1977. Clinical success rates for amoxicillin or ampicillin in the treatment of otitis media reported from 1980 to 1987 range from 83% to 96%. 12-15These findings do not support the hypothesis that amoxicillin treatment failure has greatly increased. However, the rate of/3-1actamase-producing H. influenzae increased nationwide1 from 12.1% to 15.6% between 1984 and 1986. At our hospital the rate increased from 8.1% in the period 1979 to 1984 to 23% in this study (p = 0.046). Although there has been an increase in the incidence of M. catarrhalis infections as a cause of otitis media, 2 no great increase in amoxicillin treatment failure has occurred at our hospital. Only two failures to cure M. catarrhalis were found in this study; one failure occurred with amoxicillin and one with cefixime in a child who was also infected with S. pneumoniae. This finding is consistent with the hypothesis that despite production of/3-1actamase, M. catarrhalis is a less virulent pathogen than S. pneumoniae or H. influenzae,16 and it can be eradicated more easily from the middle ear. Review articles that combine rates of ~3-1actamase production in H. influenzae and M. catarrhalis and assume treatment failure with amoxicillin are therefore misleading. The choice of an antibiotic for otitis media is influenced by the cost of the drug and by the cost of a second office visit for treatment failure.17, 18Amoxicillin is less expensive than the newer cephalosporins such as cefixime. Streptococcus pneumoniae remains the most common cause of otitis me-

12 2

Johnson et al.

dia, and amoxicitlin appears to remain the preferred drug when the pathogen is unknown. Amoxicillin may be used as the first-choice drug, but in two situations cefixime would be preferable. The first is the otitis-conjunctivitis syndrome, which is almost always caused by H. influenzae, w The second role for cefixime is for children whose condition fails to improve during treatment with amoxicillin, because such children are probably infected with H. influenzae. In summary, cefixime and amoxicillin gave equivalent results in bacteriologic and clinical efficacy, side effects, and I-month recurrence rates in children with acute otitis media. Cefixime was significantly more effective in those cases caused by H. influenzae. Cefixime is preferred when H. influenzae is the proven or suspected etiologic agent, as in amoxicillin treatment failures. The most important advantage of cefixime over other currently available drugs is its once-daily dosing, which may enhance compliance and therefore treatment success. 2~ We thank Doris Fulton and Chang Kim for laboratory assistance, Deborah Fattlar and D. Murdell-Panek for nursing support, and the nurse practitioners and pediatric house staff (particularly Jean Reitz and Virginia Turezyk) for their referral of patients. REFERENCES

1. Doern GV, Jorgensen JH, Thornsberry C, et al. National collaborative study of the prevalence of antimicrobial resistance among clinical isolates of Haemophilus influenzae. Antimierob Agents Chemother 1988;32:180-5. 2. Shurin PA, Marchant CD, Kim CH, et al. Emergence of betalactamase-produeing strains of Branhamella catarrhalis as important agents of acute otitis media. Pediatr Infect Dis J 1983;2:34-8. 3. Howie VM, Ditlard R, Lawrence B. In vivo sensitivity test in otitis media: efficacy of antibiotics. Pediatrics 1985;75:8-13. 4. O'Callaghan CH, Morris A, Kirby SM, Shingler AH, Novel method for detection of beta-lactamase using a chromogenic cephalosporin substrate. Antimicrob Agents Chemother 1972;1:283-5. 5. Marchant CD, McMillan PM, Shurin PA, et al. Objective diagnosis of otitis media in early infancy by tympanometry and ipsilateral acoustic reflex thresholds. J PED1ATR 1986;109: 590-5.

The Journal of Pediatrics July 1991

6. Neu HC. In vitro activity of a new broad spectrum, beta-Iactamase-stable oral cephalosporin, cefixime. Pediatr Infect Dis J 1987;6:958-62. 7, McLinn SE. Randomized, open label, mutticenter trial of cefixime compared with amoxicillin for treatment of acute otitis media with effusion. Pediatr Infect Dis J 1987;6:997-1004. 8. Howie VM, Owen MJ. Bacteriologic and clinical efficacy of cefixime compared with amoxicillin in acute otitis media. Pediatr Infect Dis J 1987;6:989-91. 9. Kenna MA, Bluestone CD, Fall P, et al. Ccfixime vs. cefaclor in the treatment of acute otitis media in infants and children. Pediatr Infect Dis J 1987;6:992-6. 10. Howie VM, Ploussard JH, Sloyer J. Comparison of ampicitlin and amoxicillin in the treatment of otitis media in children. J Infect Dis 1974;129(suppl):S181-4. 11. Howie VM, Ploussard JH. Efficacy of fixed combination antibiotics versus separate components in otitis media. Clin Pediatr 1972;11:205-14. 12. Mandel EM, Bluestone CD, Rockette HE, et al. Duration of effusion after antibiotic treatment for acute otitis media: comparison of cefaclor and amoxicitlin. Pediatr Infect Dis J 1982;1:310-6. 13. Giebink GS, Batalden PB, Russ JN, Le CT. Cefaclor versus amoxicillin in treatment of acute otitis media. Am J Dis Child 1984;138:287-92. t4. Shurin PA, Petton SI, Donner A, Finkelstein J, Klein JO. Trimethoprim-sulfamethoxazole compared with ampicillin in the treatment of acute otitis media. J PEDIATR 1980;96:1081-7. 15. Rodriguez W J, Schwartz RH, Sait T, et al. Erythromycinsulfisoxazole vs amoxicillin in the treatment of acute otitis media in children. Am J Dis Child 1985;139:766-70. 16. VanHare GF, Shurin PA, Marchant CD, et al. Acute otitis media caused by Branhamella catarrhalis: biology and therapy. Rev Infect Dis 1987;9:16-27. 17. Weiss JC, Melman ST. Cost effectiveness in the choice of antibiotics for the initial treatment of otitis media in children: a decision analysis approach. Pediatr Infect Dis J 1988;7:23-6. 18. Callahan CW. Cost effectiveness of antibiotic therapy for otitis media in a military pediatric clinic. Pediatr Infect Dis J 1988;7:622-5. 19. Bodor FF. Conjunctivitis-otitis syndrome. Pediatrics 1982;69: 695-8. 20. Eisen SA, Miller DK, Woodward RS, Spitznagel E, Przybeck TR. The effect of prescribed daily dose frequency on patient medication compliance. Arch Intern Med 1990;150:1881-4.