INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Acute otitis media (AOM) is one of the most frequently diagnosed infectious diseases of childhood, and Streptococcus pneumoniae is responsible for up to 50% of cases of AOM (5, 24). On the other hand, the incidence of penicillin-insensitive S. pneumoniae strains is increasing worldwide (1), and in Spain up to 54% of pneumococci isolated from middle ears (MEs) are penicillin insensitive (6, 16). Although the antimicrobial treatment of AOM is still controversial, antibiotics are capable of producing a rapid microbiological cure and avoiding severe complications, such as meningitis or other infections (4, 14).

AOM induced in gerbils is a well-established model for studying AOM as well as for comparing the efficacies of different antimicrobial regimens (3, 10).

The aim of the present investigation was to compare the clinical and bacteriological efficacies of two antibiotics in a gerbil model of AOM produced by a penicillin-resistant S. pneumoniae strain. This organism is a rather frequent cause of AOM in children, who are usually empirically treated with either amoxicillin or cefuroxime.

Part of this work has been presented at the 8th European Congress of Clinical Microbiology and Infectious Diseases, Lausanne, Switzerland, 25 to 28 May 1997 [6a].)

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacteria. A strain of S. pneumoniae type 23 isolated from a bacteremic patient was used. The MIC and minimum bactericidal concentration (MBC) of benzylpenicillin for the strain were both 2 µg/ml. Bacterial virulence was maintained by passage in mice.

Antibiotics. The antibiotics used for in vitro studies were amoxicillin trihydrate (SmithKline Beecham Pharmaceuticals, Worthing, England) and cefuroxime sodium (Sigma Chemical Co., St. Louis, Mo.). For in vivo (therapeutic) use, the contents of commercial vials (Clamoxyl [SmithKline Beecham Pharmaceuticals, Toledo, Spain] and Curoxima [Glaxo, Madrid, Spain]) were reconstituted in nonpyrogenic sterile distilled water to the desired concentrations.

In vitro studies. MICs and MBCs were determined by a broth microdilution method by previously described methods (19, 23). The median of five separate determinations was used to calculate the MIC and MBC of each antibiotic.

Animals. Eight- to 9-week-old adult female Mongolian gerbils weighing 44 to 54 g each were purchased from B & K Universal G.J.S.L. (Grimston, United Kingdom). They were given free access to food and water and were housed in a protected unit with a slight negative pressure and with a 12-h light and 12-h dark cycle. Two weeks prior to the experiments, the anteroinferior part of the auricle was removed to facilitate access to the ear drum. For invasive procedures the animals were anesthetized with 50 mg of ketamine (Ketolar; Parke-Davis, Barcelona, Spain) per kg of body weight and 13 mg of xylazine (Rompun; Bayer, Leverkusen, Germany) per kg.

Experimental otitis. An overnight culture of the organism was kept in aliquots at 70°C, and on each day of experimentation a freshly thawed aliquot was incubated for 4 h at 37°C in a 5% CO₂ atmosphere in brain heart infusion broth enriched with 5% horse serum. The number of viable bacteria in this culture was determined by the colony counting method. The animals included in the different treatment or control groups were inoculated bilaterally with 20 µl of this bacterial culture (approximately 10⁷ CFU), which was introduced directly into the ME bulla, on each day of the experiment. The tympanic membrane was left intact and swelled without rupture during the inoculation. A normal tympanic aspect and correct inoculation were verified with an operating microscope.

Treatment regimen and efficacy studies. Each antibiotic was tested at two doses (5 and 20 mg/kg/dose) and was administered subcutaneously in 500 µl at 2, 10, and 18 h postinoculation (p.i.). The animals in the control group received apyrogenated sterile distilled water in the same way.

Pharmacokinetic studies. Amoxicillin and cefuroxime were administered to healthy animals by the subcutaneous route at doses of 5 and 20 mg/kg. Groups of six animals were killed with CO₂ and were exsanguined by intracardiac puncture at 0.25, 0.5, 1, and 2 h after drug administration. Amoxicillin and cefuroxime, concentrations were determined by microbiological assay with Micrococcus luteus ATCC 9341 and Bacillus subtilis 1904E, respectively. The concentration of antibiotic in the samples was derived from a standard solution prepared in pooled gerbil serum. The variability of the assay for individual samples was <10%. Pharmacokinetic analysis were performed by routine graphical methods (13).

Statistical analysis. The number of ears with a positive count divided by the total number of ears was calculated to give the percentages of ears with positive counts for each group of animals. The percentage of treated animals with positive counts was compared with that for animals in the untreated control group, and the percentages between the groups were compared by Fisher's exact test corrected for by the Bonferroni method, due to the multiple comparisons that were performed. The overall type I error was 5%, adjusted for the 10 pairwise treatment group comparisons. Bacterial counts in untreated and treated animals were expressed as the arithmetic mean log₁₀ CFU per 20 µl of washing fluid; culture-negative samples were included in the calculation of means, assuming a value at the detection limit, and the results were analyzed by Kruskal-Wallis and Mann-Whitney tests, and the Bonferroni correction ( = 0.005) was used. These tests were also used to analyze the differences in the volume of fluid recovered from the ME as well as the relative weight loss.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

In vitro studies. The respective median MICs and MBCs for the test strain were 1 and 1 µg/ml and 4 and 4 µg/ml for amoxicillin and cefuroxime, respectively.

Therapeutic efficacy in experimental otitis. After inoculation, unilateral or bilateral purulent otitis media was detected in 92.3% of untreated animals at day 3, with 88.5% of the ME specimens being culture positive. The mean bacterial concentration was 2.58 ± 1.67 log₁₀ CFU per 20 µl of washing fluid. Intra- and extracellular pneumococci were seen in Gram-stained ME samples. From the first day of the experimental study the animals started to lose weight, with the loss being more marked on day 2 and then with a slow recovery afterward. Lethargy and otorrhea were observed in most animals on day 1, and the otoscopic examination, performed on day 3, showed marked inflammation with retrotympanic exudate. By day 7, 62.5% of the untreated animals had cleared the organism from their ME samples, and their behavior was normal, with almost complete weight recovery. Table 1 presents the comparative therapeutic results for control and treated animals. Both doses of amoxicillin and the high dose of cefuroxime significantly reduced the number of culture-positive ME specimens (P = 0.0001 for amoxicillin at 5 or 20 mg/kg and P = 0.0009 for cefuroxime at 20 mg/kg) compared with the results obtained for untreated animals. When the results obtained with different antibiotics and doses were compared, only amoxicillin at 5 mg/kg was found to be more efficacious than cefuroxime at 5 mg/kg (P = 0.0012). As far as the mean concentration of bacteria in the ME is concerned, amoxicillin (at both doses) and cefuroxime (at 5 mg/kg) significantly reduced the number of organisms compared with the numbers in the MEs of the untreated controls. No differences between antibiotics and doses were observed.

View larger version (30K): [in this window] [in a new window]   FIG. 1.   Inflammatory response on day 3 (evaluated by the presence of polymorphonuclear leukocytes [PMN]) and number of recovered microorganisms (in log10 CFU/20 µl) from washing fluid. Shaded bars, percent polymorphonuclear cells; open bars, bacterial counts (log10 CFU/20 µl). AMX, amoxicillin; CXM, cefuroxime; 5 and 20, doses of 5 and 20 mg/kg, respectively.

Pharmacokinetic and pharmacodynamic data. Table 2 presents the results of the pharmacokinetic and pharmacodynamic analyses in relation to the inoculated organism. A bacteriological eradication rate of 75% was obtained only when the concentration at 15 min/MIC ratio was >10, when the area under the concentration-time curve/MIC ratio was >10, or when the levels in serum exceeded the MIC for the pathogen for >15% of the dosing interval, which were obtained only with amoxicillin.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

AOM is usually a self-limited infection with a favorable evolution even in the absence of antimicrobial treatment (4, 26). For these reasons, it is very difficult for clinical trials to show the superiority of one regimen over another, by which bacteriological efficacy should be always considered. On the other hand, animal models of AOM can provide both clinical and bacteriological parameters that can be used to compare the efficacies between antibiotics. Amoxicillin and cefuroxime are both commonly used for the treatment of respiratory tract infections, including AOM, usually with favorable clinical results (17, 22). However, the emergence of penicillin-resistant S. pneumoniae strains (1, 6, 16) has created a great deal of concern about the efficacies of such antibiotics because these strains usually show reduced susceptibilities to both antibiotics. We selected for the animal model one S. pneumoniae strain of a serotype that is usually involved in AOM in humans (16, 20) and that was also penicillin resistant, which is the case for up to 54% of pneumococcal strains involved in this pathology (6, 8). It is also common that the cefuroxime MICs for such strains are higher than those of amoxicillin (12). In addition, this strain was able to induce AOM in gerbils, as described by other investigators (3, 10).

Our experimental model showed that for animals treated with either amoxicillin or cefuroxime, the number of cultures of ME fluids that were positive was reduced and there was also a reduction in the number of microorganisms in treated animals on day 3 compared with the numbers in untreated animals. However, only amoxicillin at either dose was able to significantly reduce both the number of culture-positive samples and the bacterial concentration in ME samples versus those for untreated animals.

Comparison of efficacy between antibiotics showed that at the low dose amoxicillin was significantly more active than cefuroxime. The increase in the dose did not increase therapeutic efficacy. The beneficial effects of the antibiotics and the superiority of amoxicillin were also confirmed by the ponderal control of the animals, with those treated with amoxicillin losing the least body weight.

The lower efficacy of cefuroxime shown in our experimental model correlates with the recently published clinical and bacteriological failures of the drug in treating AOM caused by penicillin-resistant pneumococcal strains. Gehanno et al. (11) reported 93.5% clinical efficacy when cefuroxime axetil was used for the treatment of AOM caused by pneumococci, with MICs of this antibiotic being 0.5 µg/ml, while its efficacy fell to 76.3% for strains for which MICs were 1 µg/ml. Bacteriological failure rates of up to 50% were shown by Dagan et al. (8) for those pneumococcal strains for which cefuroxime MICs were 1 µg/l, while a bacteriological failure rate of only 9% was found for those strains for which cefuroxime MICs were 0.5 µg/ml (8).

With our animal model of AOM we also showed that after antimicrobial treatment most animals had higher amounts of ME liquid compared with the amounts in untreated control animals. This could be due, at least in part, to the fact that many untreated animals suffered tympanic perforation, which caused leaks of ME liquid, while in most animals treated with antibiotics, the membrane remained intact. The higher antibiotic doses were related to the presence of more polymorphonuclear cells and to higher volumes of ME liquid compared with those in animals treated with lower doses. This might be explained by the fact that beta-lactam antibiotics increase the level of inflammation in experimental pneumococcal infections (otitis and meningitis) because cell wall debris can induce such an inflammatory response (15, 25). This phenomenon could be related to the frequent complication observed after an AOM episode, namely, otitis media with effusion, which merits further investigation.

Pharmacokinetic parameters alone were not predictive of efficacy since similar concentrations in serum at 15 min and half-lives were obtained with the low doses of both antibiotics. On the other hand, the pharmacodynamic parameters were more predictive of efficacy, with amoxicillin being very efficacious and with the percentage of the dose interval during which the levels of amoxicillin in serum exceeded the MIC for the pathogen being at least two times the percentage obtained with cefuroxime. Nevertheless, these percentages for amoxicillin were lower than those previously suggested for bacteriological efficacy (7), which could be explained by the pharmacokinetics of antimicrobial agents in ME resulting in lower peak concentrations but higher trough concentrations than those in serum (7) and/or longer half-lives in the ME (2).

Although S. pneumoniae was considered cefuroxime sensitive when the MIC for the strain was 8 µg/ml (18), clinical experience has demonstrated that failure may occur when the MICs for the strains are 1 µg/ml, particularly in infections in which the antibiotic has difficulties in penetration, such as to the ME (8, 11). Our data also show that a pneumococcal strain for which the cefuroxime MIC was 4 µg/ml cannot be eradicated when cefuroxime is used for the treatment of AOM, even when it is administered at a high dose. While we were carrying out this investigation we became aware of the most recent proposition of the National Committee for Clinical Laboratory Standards changing the cefuroxime susceptibility breakpoint to 0.5 µg/ml (19); according to that breakpoint our strain should be considered nonsensitive, which would explain both the clinical and bacteriological failures. On the other hand, although the susceptibility of our strain to amoxicillin should be considered intermediate, the use of higher doses in animal models (2, 23) and for the treatment of pneumococcal pneumonia in humans (21) is usually associated with a favorable outcome.