Cytokine Analysis of Middle Ear Effusions During Acute Otitis Media: Significant Reduction in Tumor Necrosis Factor Alpha Concentrations Correlates With Bacterial Eradication

Acute otitis media (AOM) is a common disease in children. The middle ear is capable of generating an immune response similar to that seen in other parts of the respiratory tract. The inflammatory reaction in this site is complex and multifactorial, involving infectious agents, leukocytes, macrophages and multiple cytokines.1 Juhn et al.2 confirmed the presence of significant concentrations of interleukin (IL)-1 and tumor necrosis factor alpha (TNF-alpha) in middle ear effusion (MEF) during AOM. However, the role of cytokines in the pathogenesis, natural history and mechanism of recovery from AOM is unknown. This, in part, is caused by the absence of MEF samples drawn at several times after onset of disease. We utilized MEF samples obtained immediately before antibiotic administration and during the course of treatment, for detection of the proinflammatory cytokine, TNF-alpha, and of the T helper 1 type cytokines, interferon-gamma (IFN-gamma) and IL-2.3 Our aim was to extend previous knowledge on cytokine characterization in AOM that was based on a single tap of the middle ear. Patient and methods. This study was conducted between October, 1995, and May, 1997. Children ages 3 months to 3 years with AOM were enrolled in different antibiotic studies at Soroka University Medical Center; these studies were approved by the local and national ethics committees. Informed consent was obtained at enrollment from parents or guardians. The diagnosis of AOM was based on symptoms of fever, irritability or ear ache, signs of inflammation (red or yellow color or bulging) of the tympanic membrane and the presence of MEF as documented by tympanocentesis. All enrolled patients had an acute illness for ≤7 days, they had no spontaneous perforation and no tympanostomy tubes were present. The patients received after enrollment different antibiotic treatments for AOM, including amoxicillin, cefaclor, azithromycin and intramuscular ceftriaxone, and were followed at least to Day 4 or 5, at which time a second tympanocentesis and MEF culture were performed. In three children both ears were tapped during the same episode of AOM. Study conduct. All patients had tympanocentesis performed at enrollment by an otolaryngologist. Antisepsis of the ear canal before tympanocentesis was done with 70% alcohol instilled for 1 min. After removal of the alcohol by suction and with the use of a 20-gauge spinal needle attached to a 1.0-ml sterilized tuberculin syringe, the anteroinferior portion of the intact tympanic membrane was punctured. The fluid (0.1 to 0.2 ml) was immediately aspirated into the sterile syringe by suction, and part of it was applied onto a sterile swab and sent for bacteriologic culture in transport medium (MW173 Amies medium; Transwab; Medical Wire and Equipment, Potley, UK) for processing within 12 h at the Clinical Microbiology Laboratory. The unused MEF was frozen at −70°C for cytokine determination. After tympanocentesis, patients were randomized to receive various antibiotic treatments (amoxicillin and cefaclor for 10 days, azithromycin and im ceftriaxone for 3 days). Bacteriology. The identification methods used for all pathogens recovered have been described elsewhere.4 Cytokine assays. Preparation: The processing consisted of weighing the sample, dilution with phosphate-buffered saline to 1 ml, centrifugation to remove cellular elements and aliquoting of the supernatant. The processed aliquots were frozen at −70°C for enzyme-linked immunosorbent assays (ELISA). On the day of the assay specimens were vortexed and centrifuged at 1000 rpm for 10 min. Supernatants were aliquoted (20 μl/aliquot) and assayed or were frozen at −70°C until assayed. Before the assay each aliquot was diluted 1/10 with phosphate-buffered saline and vortexed. Cytokine measurements: ELISA assays for TNF-alpha, IFN-gamma and IL-2 were performed by ELISA according to manufacturer's protocol (PharMingen, San Diego, CA). Forty-eight-well ELISA plates were coated with 2 μg/ml of capture monoclonal antibody (mAb) and incubated overnight at 4°C. On the next day known standards and duplicate test samples were added and incubated again overnight at 4°C. After the binding of TNF-alpha, IFN-gamma and IL-2 to the immobilized mAb, a second biotinylated detecting mAb (1 μg/ml) was added for 45 min, thus forming a "sandwich." Peroxidase-conjugated streptavidin (2.5 μg/ml) was added for 60 min to form an avidin-biotin complex with the biotinylated mAb. For color development substrate solution (ABTS) with 0.03% H2O2 was added. Plates were read at 405 nm with a microplate spectrophotometer. Readouts used for analysis were taken when 1000 pg/ml standard reached 0.5 to 0.8 OD. Standard curves were generated using known concentrations of recombinant TNF-alpha, IFN-gamma and IL-2 in a series of eight dilutions from 2000 pg/ml to 15 pg/ml. Because concentration in each sample corresponds to the OD readout of the sample, its value was derived from standard curves by regression analysis. Final concentrations were calculated by multiplying the given values by the dilution factor, the results are reported as log10 pg/ml ± SD. Each experiment was repeated at least twice. Statistical analysis. The statistical analysis was performed on log transformed variables using the paired t test when the results from Day 4 to 5 were compared with results from Day 1 or by use of the Student t test when independent samples were compared. Two-tailed analysis was used throughout with significance defined as P ≤ 0.05. Results.Table 1 summarizes the clinical characteristics of our study group. Twenty paired MEF samples obtained from 17 patients were analyzed on Day 1 and Day 4 or 5 (40 samples overall). Nineteen pathogens were isolated from 15 of 20 (75%) MEF samples obtained on Day 1: 14 nontypable Haemophilus influenzae; 4 Streptococcus pneumoniae; and 1 Streptococcus pyogenes. Twelve pathogens were isolated from 11 of 20 (55%) MEF samples obtained on Day 4 or 5: 11 nontypable H. influenzae; 1 S. pneumoniae. Persistence or eradication of bacteria on Day 4 or 5 was not related to antibiotic treatment during the 2 weeks preceding the present episode of AOM, number of previous episodes of AOM or body temperature at the time of tympanocentesis.TABLE 1: Clinical characteristics IFN-gamma and IL-2 were not detected in any MEF samples analyzed. Overall the mean TNF-alpha concentrations were lower in the MEF samples obtained on Day 4 or 5 than on Day 1. However, this decrease was not statistically significant (P = 0.07, Table 2). Significantly higher TNF-alpha were measured on Day 1 and on Days 4 or 5 in the MEF samples of patients with culture positive specimens compared with those with culture negative ones (P = 0.011; P = 0.023 on Days 1 and 4 or 5, respectively).TABLE 2: TNF-alpha concentrations* in MEF in relation to bacteriologic results and day of tympanocentesis The TNF-alpha concentrations in 20 paried MEF samples, according to bacteriologic outcome of AOM, disclosed the following. There was a statistically significant reduction (P < 0.005) in TNF-alpha concentrations on Day 4 or 5 (2.39 ± 0.4 pg/ml) compared with Day 1 (3.16 ± 0.17 pg/ml) in the MEF of the patients in whom eradication was achieved. On the contrary no significant reduction was recorded in the MEF TNF-alpha concentrations of patients in whom pathogens persisted on Day 4 or 5 (3.01 ± 0.46 on Day 1 vs. 2.68 ± 0.85 on Day 4 or 5, P = 0.36). In addition TNF-alpha concentrations did not vary between culture-negative MEF samples on Day 1 and those that continued to be negative on Day 4 or 5. Discussion. Our study is the first to document a correlation between bacteriologic outcome and TNF-alpha concentration in MEF of patients with AOM. We report that the mean concentration of TNF-alpha is significantly higher in culture positive than in culture-negative MEF, both before and after initiation of antibiotic therapy. Recently the correlation between bacteriologic outcome and cytokine (interleukin-8) and the inflammatory mediator (leukotriene B4) patterns in MEF of patients with AOM have been reported by Chonmaitree et al.5 Bacteriologic failure was associated with the levels of leukotriene B4 in MEFs collected before initiation of antibiotic, but not with those collected after treatment. Our findings of significant concentrations of TNF-alpha in MEF is in agreement with earlier studies measuring cytokine concentration in MEF. Ophir et al.6 used a biologic assay to document the presence of TNF-alpha in a high percentage of MEF samples obtained from patients with chronic otitis media with effusion (COME). The presence of inflammatory mediators during COME has been shown by Yellon et al.7 to be correlated to TNF-alpha concentrations present in MEF in COME. Other studies analyzing effusion type (purulent vs. mucoid) and inflammatory mediators showed that the degree of middle ear inflammation is reflected by high concentrations of TNF-alpha and IL-1.2 Although we failed to detect IFN-gamma and IL-2 in MEF samples, several studies have detected levels of IFN-gamma and IL-2 in a small number of MEF samples. Yellon et al7 who documented the persistence of IFN-gamma and IL-2 in MEF samples measured the latter in patients with COME. The ongoing chronic inflammatory state associated with COME6 may contribute different cell types and thus different cytokines than those with AOM. Whereas we used ELISA to measure cytokine concentration, Howie et al.,8 using a biologic assay, showed the total IFN activity during bacterial otitis media. The different results from our study and that of Howie et al. are likely to be related to a combination of factors, among them the sensitivity and the specificity of the assay, dilution factor, sample size and type of IFN used. Because cytokine analysis of a single MEF sample collected before treatment was shown to have some predictive value for bacteriologic outcome during AOM5 and because our study also showed that reduction of TNF-alpha is correlated with bacteriologic outcome, we suggest that single or sequential analysis of MEF during AOM might have a clinical role as a predictor of outcome in some patients with AOM. Our findings demonstrate a possibly important role for TNF-alpha in the disease process. TNF-alpha is produced by macrophages and other cells in response to stimuli such as bacterial lipopolysaccharide and viruses. It has numerous activities including promotion of fibroblast proliferation, neutrophil activation, inhibition of growth of vascular endothelial and B cells, stimulation of cytotoxic cell activity and of cartilage and bone resorption.9 Thus similar to other cytokines it has the potential to help clear infection from the middle ear, while at the same time it has the potential to cause tissue damage.6 We hypothesize that monocytes recruited to the middle ear or macrophages in situ secrete large amounts of TNF-alpha and later on, after the initiation of antibiotics, they are unable to produce sufficient concentrations of this cytokine, especially in the absence of bacteria as a result of successful antibiotic therapy. On the other hand it seems that during acute inflammation, regardless of whether a pathogen was recovered, recruitment of lymphocytes to the middle ear with special emphasis on T helper 1 cells, which are known to produce and secrete IFN-gamma and IL-2,10 does not occur. Thus the nature of the cellular infiltrate during the acute inflammation might account for the absence of IFN-gamma and IL-2. The findings of this study suggest a prominent role for the monokine TNF-alpha in the pathogenesis and outcome of acute otitis media in children. Further studies that include a larger number of patients are needed to determine the kinetics of TNF-alpha and other cytokines in MEF samples during acute otitis media. Asher Barzilai, M.D. Benjamin Dekel, M.D. Ron Dagan, M.D. Justen H. Passwell, M.D. Eugene Leibovitz, M.D.