The association between hearing loss and meningitis

Transcription

1 Time course of hearing loss in an animal model of pneumococcal meningitis BRADLEY W. KESSER, MD, GEORGE T. HASHISAKI, MD, JONATHAN H. SPINDEL, PhD, ROGER A. RUTH, PhD, and W. MICHAEL SCHELD, MD, Charlottesville, Virginia A leading cause of morbidity from bacterial meningitis is an irreversible, usually profound sensorineural hearing loss, with an incidence as high as 30% in some studies. Bacterial meningitis remains the most common cause of acquired postnatal sensorineural deafness. Although several clinical studies have examined the long-term outcome of hearing in meningitis, few studies have examined the time course of hearing loss during the acute course of the disease. We have developed an animal model of meningogenic hearing loss in the rat and have plotted the time course of that hearing loss. Serial auditory brain stem responses (ABRs) were measured in rats inoculated in the cisterna magna (subarachnoid space) with Streptococcus pneumoniae (10 5 to 10 7 colony-forming units). All rats injected developed meningitis as evidenced by increased cerebrospinal fluid (CSF) white cell counts and positive CSF cultures. Serial ABR measurements taken 6, 12, 15, 18, 21, and 24 hours after inoculation demonstrated significant threshold shifts and eventual loss of the ABR waveform as compared with measurements in control rats injected with sterile culture medium. Hearing loss began approximately 12 to 15 hours after inoculation and progressed to complete loss by 24 hours (17 of 18 animals). No correlation was found between the magnitude of hearing loss and CSF white cell count or bacterial titer. Temporal bone histology of rats with meningitis shows a dense inflammatory cell infiltrate throughout the subarachnoid space. Labyrinthine inflammatory cells were confined to the scala tympani. From the Departments of Otolaryngology (Drs Kesser, Hashisaki, Spindel, and Ruth) and Internal Medicine (Dr Scheld), University of Virginia Medical Center. Supported by the Research Fund of the American Otological Society, Fellowship Training Grant, Presented at the Annual Meeting of the American Academy of Otolaryngology, San Francisco, CA, September 7-10, Reprint requests: George T. Hashisaki, MD, Department of Otolaryngology-, University of Virginia Medical Center, Box 430, Charlottesville, VA Copyright 1999 by the American Academy of Otolaryngology Foundation, Inc /99/$ /1/92772 The cochlear aqueduct is the proposed route of infection from the meninges to the labyrinth (scala tympani). Endolymphatic hydrops was also noted throughout the cochlea. These experiments both establish a reproducible animal model of meningogenic hearing loss and support the hypothesis that this hearing loss is progressive rather than abrupt in onset and is related to the duration of untreated infection. CSF inflammatory cells appear to enter the cochlea through the cochlear aqueduct. This reliable animal model will enable future studies directed toward further understanding the pathogenesis and pathophysiology of this hearing loss. (Otolaryngol Head Neck Surg 1999;120: ) The association between hearing loss and meningitis has been well documented The hearing loss that occurs as a sequela of bacterial meningitis in most patients is a bilateral, severe-to-profound, irreversible sensorineural loss. 1,2,4,5,8-12 The mechanism of this hearing loss, however, remains unclear. In addition, why some children and adults sustain significant hearing impairment whereas others escape this severe debility is unknown. Meningogenic hearing loss remains the most common cause of acquired postnatal childhood sensorineural deafness. 1,2,4,15 It is estimated that 90% of children with an acquired profound sensorineural hearing loss (SNHL) lost their hearing from bacterial meningitis. 15 Neurologic sequelae of bacterial meningitis include hearing loss (the most common), mental retardation, epilepsy, ataxia, blindness, and subtle cognitive deficits. 7 In addition, the aftereffects of meningitis on the labyrinth (ie, labyrinthine ossification) make placement of an implantable hearing device very difficult. 16,17 Interestingly, the incidence of hearing loss varies with the etiologic organism: pneumococcal (Streptococcus pneumoniae) meningitis causes hearing loss in as much as 30% of children, whereas meningitis caused by Haemophilus influenza or Neisseria meningitidis results in an incidence of only 10% to 20%. 3,9,14 Why S pneumoniae causes a proportionally greater incidence of hearing loss is unknown. The pneumococcus is currently the second most common organism to cause bac- 628

2 Otolaryngology Volume 120 Number 5 KESSER et al 629 Fig 1. Distribution of stimulus thresholds as a function of time in rats with pneumococcal meningitis. White bar, median; brackets, 99th percentile range; gray shaded area, 25% to 75% of measured thresholds; black shaded area, 95% CI for median; solid bar, outlier. terial meningitis in children, behind H influenza. With the introduction of the H influenza type B vaccination, fewer cases of invasive disease (meningitis, epiglottitis, sinusitis) caused by H influenza are likely to occur. The pneumococcus will therefore likely replace H influenza as the leading cause of bacterial meningitis in children. 3 The time course for hearing loss during the acute course of bacterial meningitis has not been well characterized. Studies conflict in documenting the onset of hearing loss during meningitis. Patients seek treatment at various stages of disease, and it is difficult to know the true onset of disease and its relation to the onset of hearing loss. Although ABRs have been performed in patients at the time of diagnosis with follow-up ABRs weeks to months after recovery, no human study has documented the time course of hearing loss in meningitis by using serial ABRs during the acute course of the disease. It is unclear whether the hearing loss associated with meningitis is abrupt in onset or progressive, or whether it occurs early or late in the clinical course. Factors that potentially predispose to hearing loss are also unknown. Brookhouser et al 8 pointed out a few general principles culled from the literature: 1. Onset of hearing loss associated with bacterial meningitis most likely occurs early in the course of the illness, perhaps during the bacteremia phase, before other signs and symptoms are evident. Keane et al 10 observed the presence of hearing loss shortly after the onset of meningitis, based on behavioral observations and play audiometric testing. On the other hand, Vienny et al 13 could not show a significant correlation between duration of symptoms before treatment and occurrence of deafness. 2. Patients with normal ABRs after the first few days of hospitalization with meningitis are unlikely to have later SNHL. Late onset SNHL is uncommon. 3. Some patients demonstrating ABR consistent with hearing loss early in the course of meningitis may have normal ABRs at discharge or in follow-up testing (some of this improvement corresponds to resolution of a concomitant conductive hearing loss). Alternatively, complete absence of the ABR predicts irreversible sensorineural deafness Most but not all patients who showed improvement of auditory thresholds over time initially demonstrated a mild-to-moderate SNHL, rather than a severe-to-profound loss. Establishing the time course of hearing loss during meningitis would therefore be helpful in patient and family counseling, in defining the role of adjunctive therapeutic measures such as steroids (which are still controversial in bacterial meningitis), and in exploring the pathophysiology of meningogenic hearing loss. How infection spreads from the meninges to the inner ear has also been a subject of debate. Temporal

3 630 KESSER et al Otolaryngology May 1999 Fig 2. Change in hearing over time in rats with pneumococcal meningitis. This curve and Fig 3 represent the best fit predicted time course based on data points at 0, 12, 15, 18, 21, and 24 hours. Fig 3. Comparison of hearing over time between control animals injected with sterile culture medium and experimental animals injected with S pneumoniae (P < ). bone histology in human beings with meningitis has implicated a route of infection from the subarachnoid space into the cochlear aqueduct and scala tympani. 18,19 Another proposed route of infection involves white cells and bacteria pushing into the internal auditory canal (IAC) along the cochlear nerve and into the modiolus. 18 Finally, entry into the cochlea through the cerebrovascular system has also been postulated. 20 The mechanism of damage to the inner ear has not been well characterized either. One other animal model of meningogenic hearing loss has previously been documented in the rabbit. 21 This model has also been used to study the time course of hearing loss in meningitis. We report a second animal model of meningogenic hearing loss, characterize the time course of that hearing loss, and display temporal bone histology to document the potential routes of infection from the meninges to the inner ear. METHODS AND MATERIAL Preparation of Inoculum S pneumoniae type III are initially plated on sheep blood agar and grown for 24 hours in a 37 C incubator. From the plate, a high-inoculum suspension of bacteria is created in a

4 Otolaryngology Volume 120 Number 5 KESSER et al 631 Fig 4. Scatter plot of hearing thresholds at 24 hours versus CSF white cell count 24 hours after injection with S pneumoniae. 10-mL solution of premixed maltose culture medium; 50 µl of this high-inoculum suspension is then pipetted into an additional 10 ml of maltose culture medium and grown for 8 to 12 hours to midlogarithmic phase. This suspension yields an inoculum of 10 5 to 10 7 colony-forming units (CFU)/µL; 50 µl of this suspension are then injected into the subarachnoid space of the animal. Animal Preparation and Inoculation This protocol was reviewed and approved by the Institutional Animal Care and Utilization Committee of the University of Virginia Medical Center. All experiments were performed at the University of Virginia, and strict adherence to NIH guidelines on the use of laboratory animals was maintained. Female Wistar rats (175 to 195 g) were anesthetized with a ketamine/xylazine mixture intramuscularly. Rats were anesthetized before each stereotactic manipulation and before each ABR recording to ensure minimal artifact from muscle electrical activity. After the animal was anesthetized, a 25-g butterfly infusion needle was stereotactically guided into the subarachnoid space at the level of the cisterna magna. External landmarks that direct the needle were easily palpable. Entrance into the subarachnoid space was marked by brisk capillary return of cerebrospinal fluid (CSF) into the infusion set. The capillary tubing was clamped off, and 50µL of S pneumoniae inoculum was injected. Control animals received an inoculum of sterile culture medium. Hearing Assessment After bacteria were injected intracisternally, the ABR recordings were collected with a click, broadband stimulus to provide information on changing auditory function across the time course of bacterial infection. Click-evoked ABRs were collected with a 50 µsec acoustic click. A closed-canal stimulus system was used with a tube transducer (Etymotic, ER-2) and probe microphone system (Etymotic, ER-7C). Tubes for both devices were coupled to the ear with an infant ear speculum and ear mold impression material. The probe microphone was used to provide true in-the-canal stimulus amplitude and frequency spectrum calibration. With the animal anesthetized, ABR electrodes were placed with the standard vertex (+), external auditory canal ( ), and hind quarters (ground) configuration. These signals were preamplified with a low-noise preamplifier (Stanford Research, SR560) gain setting of 20,000 and bandpass filtered (3 to 3000 Hz). An inhouse designed postamplifier and digital display gain provided additional amplification, which yielded a total ABR amplification of 800,000. ABR averaging was accomplished with a computerized signal averager and oscilloscope system (RC Electronics, ISC-16). With an initial stimulus intensity of 50 db pespl, an ABR series was collected for each animal preparation over a range of decreasing stimulus intensities until threshold was found. Threshold was determined as the minimum sound intensity (db pespl) that would elicit a repeatable ABR waveform. All animals were examined for the presence of middle ear fluid that would confound the SNHL associated with meningitis.

5 632 KESSER et al Otolaryngology May 1999 Fig 5. Scatter plot of hearing thresholds at 24 hours versus CSF bacterial titer 24 hours after injection with S pneumoniae. ABR recordings were taken every 3 or 6 hours (depending on the experiment) starting 12 hours after injection (our pilot studies have shown that rats do not lose hearing in the first 12 hours of their disease). At 24 hours after injection, a final ABR was performed. To document central nervous system infection, CSF was drawn off at 24 hours and analyzed for white cell count (with a standard hemocytometer system) and bacterial titer. Serial CSF bacterial dilutions were grown on blood agar plates to determine the bacterial concentration in the CSF. Tissue Preparation Animals were euthanized after the final ABR, 24 hours after the injection of bacteria. They were decapitated, immersion-fixed in formalin, and prepared for histology. The heads were decalcified in a 10% formic acid solution and embedded in paraffin. Serial sections were cut at a thickness of 8 µm. Tissue sections were mounted on glass slides and stained with hematoxylin and eosin. In addition, some animals were transcardially perfused with saline solution and then 4% paraformaldehyde. The temporal bones were removed and immersed in fresh fixative. They were then processed and sectioned according to the above protocol. RESULTS All animals injected in the cisterna magna with 10 5 to 10 7 CFU S pneumoniae developed meningitis during the ensuing 24 hours, as evidenced by increased CSF white cell count (CSF leukocytosis) and positive CSF bacterial cultures. Spinal fluid was analyzed before bacteria injection in several random animals to ensure preexperimental sterility. These samples grew no bacteria and had no white blood cells. CSF white cell counts at 24 hours after inoculation varied widely, but all rats injected with bacteria displayed CSF leukocytosis (mean CSF white cell count = 29,600 cells/µl; SD = 28.6). Bacteria were also present in spinal fluid after 24 hours (mean CSF bacterial titer = CFU/mL; SD = 37.4); sufficient spinal fluid for serial bacterial dilution could not be obtained from every animal. Animals that develop meningitis demonstrate significant hearing losses. This hearing loss starts approximately 12 to 15 hours after injection and rapidly progresses to severe-to-profound, bilateral SNHL 24 hours after injection. Hearing loss is documented by a threshold shift in the ABR. As the animal progresses further into its disease, hearing thresholds shift such that increasing intensities of sound are required to generate an ABR response. Severity of hearing loss is clearly related to duration of infection. By 24 hours, intensities of 100 to 120 db pespl are required to elicit electrical activity in the brain stem; some rats exhibit no response at these intensity levels. Figure 1 shows the distribution of stimulus thresholds as a function of time during the course of meningitis. Figure 2 shows the change in threshold as a function of time. In each case, there was a very significant linear increase in hearing loss with

6 Otolaryngology Volume 120 Number 5 KESSER et al 633 Fig 7. Low-power midmodiolar section of infected rat that developed complete hearing loss. Fig 6. Low-power photomicrograph of rat brain stem showing cellular inflammatory infiltrate (PMNs) throughout subarachnoid space around cochlear nucleus. increase in time (all P < ). All rats were examined before and after the experiment for evidence of middle ear pathology; no rat in the experiment had middle ear fluid or infection. In contrast, CSF samples taken from control animals 24 hours after injection did not demonstrate evidence of meningeal inflammation (no white cells or positive CSF cultures). Again serial ABR recordings were measured and thresholds were recorded. None of the control rats lost hearing, as evidenced by thresholds that remained constant throughout the course of the 24-hour experiment (Fig 3). χ 2 analysis demonstrates a strong difference between hearing loss for the experimental as compared with the control animals (χ 2 = 21, df = 3, P < ). The hearing loss increased linearly throughout the 24-hour evaluation period (slope = ). An attempt was made to correlate the degree of meningeal inflammation with hearing loss by plotting CSF white cell count and bacterial titer as a function of change in hearing threshold. We hypothesized that higher bacterial titers in the CSF and higher CSF white cell counts would lead to greater degrees of hearing loss, but no correlation was found (Figs 4 and 5). Fig 8. Higher power midmodiolar view. Note cellular infiltrate in scala tympani, predominantly acellular infiltrate in scala vestibuli, and endolymphatic hydrops. Temporal bone histology was reviewed after both immersion and perfusion fixation techniques. Light microscopic analysis demonstrated a dense cellular infiltrate (predominantly polymorphonuclear leukocytes [PMNs]) throughout the meninges, consistent

7 634 KESSER et al Otolaryngology May 1999 Fig 9. Organ of Corti in an infected animal. Cells seen in scala media are red blood cells. with CSF leukocytosis. PMNs were also seen in the subarachnoid space around the cochlear nucleus (Fig 6). In the cochlea, inflammatory cells were present throughout the scala tympani. In some specimens an acellular inflammatory infiltrate was also noted in the scala vestibuli and less frequently in the scala media (Figs 7 and 8). Very few inflammatory cells were noted in the scala vestibuli; no inflammatory cells were seen in the scala media. In general, the scala media was free of infiltrate; however, a bulging of Reissner s membrane was noted throughout all levels of the cochlea (Figs 7 and 8). The cochlear aqueduct could not be identified. Tissue preparation artifact did not allow a detailed examination of the organ of Corti, but the outer hair cells in one specimen appeared to be intact (Fig 9). In addition, inflammatory cells were also noted along the medial course of the cochlear nerve from the brain stem (Fig 10). Very few inflammatory cells were noted laterally in the IAC or in the modiolus. The cochlear, vestibular, and facial nerves all appeared normal. Fig 10. Vestibular and cochlear nerves entering the IAC. PMNs are confined to the medial portion of the canal and out in the subarachnoid space of the brain stem. The lateral IAC appears free of inflammation. DISCUSSION This set of experiments establishes a reliable animal model for the study of meningogenic hearing loss. Animals given pneumococcal meningitis experience a period of approximately 12 hours in which they have little or no hearing loss. As the disease continues past 12 hours, hearing loss becomes rapid and progressive, resulting in almost no brain stem electrical activity to the standard auditory click stimulus at 24 hours in all but 1 animal (17 of 18). These results indicate that meningogenic hearing loss is not immediate in onset it has a latency period. Once started, however, it is progressive in its time course. Severity of hearing loss during the 24 hours does appear to be related to duration of (untreated) infection. Hearing loss does not correlate with degree of meningeal inflammation (CSF leukocytosis or bacterial titer). Because these animals were untreated, it is impossible to know at what time point the hearing loss is reversible or even capable of being stabilized. Future studies will attempt to document treatment effects. In addition, by 24 hours the animals are quite moribund, and the cause of the hearing loss could potentially result from global brain stem dysfunction rather than a specific injury to the cochlea. A similar animal model of meningogenic hearing loss in the rabbit supports these findings; rabbits inoculated with S pneumoniae had an initial latency period of 12 hours followed by a rapid progression to deafness at an average of 36 hours. 21 That study documented a strong positive correlation between incidence and severity of hearing loss with duration of untreated meningitis. The 2 animal models of meningogenic hearing loss support the hypothesis that hearing loss is an unavoidable sequela of untreated meningitis and that all animals will lose hearing if left untreated and measured for a long enough period of time. In fact, a small subset of animals in both studies did not have a profound hearing deficit during the course of the experiment. A few animals seemed to be somewhat resistant to hearing loss. Changes in the anatomy of the cochlear aqueduct and/or its patency (see below) could account for such findings. Differences in the animals immune responses to the infection may have resulted in labyrinthine suppuration late in the course of the disease. Finally, these resistant animals may have developed hearing loss if recorded

8 Otolaryngology Volume 120 Number 5 KESSER et al 635 for an extended period of time. When 1 animal in a study by Bhatt et al 21 was measured at 60 hours, significant hearing loss did develop. A few animals infected in our early experiments also did not develop profound hearing loss; when these animals were allowed to progress to 40 hours, significant hearing loss was measured farther along in the course of meningitis. Clinically, why some patients lose hearing and others do not remains a mystery. In fact, the retrospective clinical data seem to indicate that meningogenic hearing loss is idiosyncratic and unpredictable. Some studies have shown a correlation between duration of symptoms and morbidity, but others have not supported this correlation. The human studies that have attempted to characterize the time course of hearing loss in meningitis have used auditory evoked potentials. Patients who had hearing loss from meningitis in one study tended to have some hearing dysfunction early in the course of their disease (within 48 hours of presentation). 13 A subset of patients with mild-to-moderate abnormalities on the initial ABR had a reversal of their hearing deficit and had normal ABR recordings weeks to months later, 13 but most patients found to have an initial SNHL maintained stable hearing deficits over time. 8 Establishing a time course for hearing loss in meningitis in human beings remains elusive because of the difficulty in documenting duration of symptoms and onset of infection. Two studies failed to show a significant correlation between duration of symptoms before initiation of treatment and development of SNHL. 2,13 These investigators argued that hearing loss theoretically occurs early in the course of meningitis. A critical period during the first week of meningitis has been proposed in which changes are seen on the ABR, but these changes are reversible. These initial hearing deficits on the ABR may correspond to a reversible serous labyrinthitis, which may eventually suppurate during the course of meningitis and result in permanent hearing loss. 13 These findings seem to correlate with the latency period documented in our set of experiments. Certainly the time course for the rat will be different from that of human beings, but some general principles seem to be applicable. Meningogenic hearing loss appears to be gradual in onset and progressive if left untreated, with the severity of hearing loss related to the duration of infection. We hypothesized that animals with higher CSF white cell counts and higher bacteria titers in their spinal fluid would develop a more rapid or greater degree of hearing loss. In fact, there was no correlation between magnitude of hearing loss and measured CSF parameters. This finding is not supported by clinical studies in which correlations have been found between CSF white cell count/glucose level and hearing loss. 1,5 Other clinical studies, however, have not shown a correlation between CSF cytochemistry (white count and glucose level) with hearing loss. 9,12 Although studies in the rabbit model showed definite increases in CSF white count and protein level with time, no analysis was made to correlate degree of hearing loss or rapidity of onset of hearing loss with CSF cytochemistry. At this time, it is impossible to predict, based on clinical factors or CSF cytochemistry, which patients will have hearing loss as a sequela of meningitis. Although we did not find a correlation between CSF leukocytosis/bacterial titer and hearing loss, our animal model may nevertheless help identify factors that predispose or that have a negative impact on hearing during the course of meningitis. Current recommendations are for all children who have recovered from bacterial meningitis to undergo hearing evaluation. Temporal bone histology in our study supports the hypothesis that the hearing loss resulting from meningitis is caused by a secondary suppurative labyrinthitis caused by the invasion of bacteria into the perilymphatic space from the meninges. Animals in our study developed suppurative labyrinthitis as evidenced by a cellular inflammatory infiltrate confined most predominantly to the scala tympani. Acellular infiltrates were seen in the scala vestibuli (there were occasional cells in the scala vestibuli) and, rarely, the scala media (no inflammatory cells were seen in the scala media). Suppurative labyrinthitis after meningitis has been reported for the rabbit 21,22 and infant rat models, 23 as well as in postmortem human temporal bone studies. 18,24 The cochlear aqueduct has been implicated as a pathway of infection from the meninges to the labyrinth based on both human postmortem temporal bone studies and animal models of experimental meningitis with secondary suppurative labyrinthitis 18-23,25 ; our results support this hypothesis as well the densest areas of inflammation and PMNs were found in the scala tympani. We did not identify the cochlear aqueduct itself in our specimens. We also did not see a clear gradient of inflammation progressing from base to apex in our cochlear specimens, as would be predicted by the hypothesis and as was noted in the rabbit model. 22 The IAC has also been proposed as a second potential route of infection from the meninges to the inner ear. 18 Although we did see inflammatory cells medially around the porus, there did not appear to be leukocytes laterally in the cribriform area of the canal. A similar finding was also noted in the rabbit model. 21 The hematogenous route of infection cannot be completely ruled out. Infant rats inoculated intraperi-

9 636 KESSER et al Otolaryngology May 1999 toneally developed meningitis and labyrinthitis, with the greatest degree of meningeal and labyrinthine inflammation seen 48 hours after inoculation. 23 These experiments used infant rats, a different strain of bacteria, and the intraperitoneal route of inoculation; these differences could account for the longer period of time to the development of labyrinthitis. The route of infection, however, must be intravascular because bacteremia set up by the intraperitoneal injection resulted in meningitis and labyrinthitis. The hematogenous route of infection would predict an equal distribution of inflammatory cells throughout the cochlear chambers, perhaps more cells in the scala media. Although our studies cannot exclude this possibility, it is less likely given the far greater proportion of cells in the scala tympani and lack of cells in the scala media. Finally, we cannot exclude the cochlear nucleus and cochlear nerve as potential sites of injury. Although morphologically intact, the functional status of these structures must be called into question as the ABR results suggest. Histologically, the cochlear nerve and nucleus are surrounded by the inflammatory cellular infiltrate and are potential targets of cellular injury. We are currently measuring otoacoustic emissions during meningitis to differentiate whether the hearing loss is sensory or neural. One unexpected finding was the presence of endolymphatic hydrops. The appearance of hydrops could be artifactual from tissue fixation and/or perfusion, but it has also been reported in the rabbit model. 21 A recent study examining ultrastructural changes in the organ of Corti during experimental pneumococcal meningitis showed disruption of the reticular lamina, the tight junctions between apical hair cells, and supporting cells that provide the ionic and electrical barrier between the scala media and scala tympani. 26 A breach in the reticular lamina could disrupt the endocochlear potential and cause ionic changes resulting in osmotic fluid shifts between the cochlear chambers. Alternatively, obstruction of the endolymphatic duct from hemorrhagic, fibrinous, or infectious material could cause a hydropic state. Further studies are needed to measure the endocochlear potential during meningitis or possibly decompress the endolymphatic sac before inoculation. Experiments are currently under way to attempt to determine whether the hearing loss associated with bacterial meningitis is caused by the infecting organism itself or by the host inflammatory response an autoimmune phenomenon. Although we did not attempt to identify or locate bacterial organisms in the cochlea, other studies have shown bacteria in the perilymphatic spaces during meningitis. 22,23 Tarlow et al 27 perfused the scala tympani directly with bacterial endotoxin. They found a consistent drop in the size of the compound action potential and cochlear microphonic elicited by stimulation with a 1-msec 10-kHz tone pip. Histologically, there was generalized swelling of the tectorial membrane and damage to hair bundles of inner and outer hair cells, mostly to the stereocilia. This study documents both morphologic and functional damage to the cochlear from administration of an inflammatory molecule as opposed to a live organism. Is meningogenic hearing loss caused by the infecting organism itself, the endotoxin molecule, or the resultant inflammatory cascade? Future studies using our animal model seek to answer these questions. The role of steroids in bacterial meningitis remains controversial; studies aimed at defining the role of the host inflammatory response may offer insight into the use of steroids or other anti-inflammatory compounds to prevent and/or reverse the morbidity associated with bacterial meningitis. We thank Karen Pieper, MS, for her statistical assistance. REFERENCES 1. Nadol J. Hearing loss as a sequela of meningitis. Laryngoscope 1978;88: Dodge P, Hallowell D, Feigin R, et al. Prospective evaluation of hearing impairment as a sequela of acute bacterial meningitis. N Engl J Med 1984;311: Fortnum HM. Hearing impairment after bacterial meningitis: a review. Arch Dis Child 1992;67: Baldwin R, Sweitzer R, Freind D. Meningitis and sensorineural hearing loss. Laryngoscope 1985;95: Berlow S, Caldarelli D, Matz G, et al. Bacterial meningitis and sensorineural hearing loss: a prospective investigation. Laryngoscope 1980;90: Taylor HG, Mills E, Ciampi A, et al. The sequelae of Haemophilus influenza meningitis in school-age children. N Engl J Med 1990;323: Pomeroy S, Holmes S, Dodge P, et al. Seizures and other neurologic sequelae of bacterial meningitis in children. N Engl J Med 1990;323: Brookhouser P, Auslander M, Meskan M. The pattern and stability of postmeningitic hearing loss in children. Laryngoscope 1988;98: Ozdamar O, Kraus N, Stein L. Auditory brainstem responses in infants recovering from bacterial meningitis: audiologic evaluation. Arch Otolaryngol 1983;109: Keane WM, Potsic WP, Rowe LD, et al. Meningitis and hearing loss in children. Arch Otolaryngol 1979;105: Guiscafre H, Martinez MC, Benitez-Diaz L, et al. Reversible hearing loss after meningitis: prospective assessment using auditory evoked responses. Ann Otol Rhinol Laryngol 1984;93: Kotagal S, Rosenberg C, Rudd D, et al. Auditory evoked potentials in bacterial meningitis. Arch Neurol 1981;38: Vienny H, Despland PA, Lutschg J, et al. Early diagnosis and evolution of deafness in childhood bacterial meningitis: a study using brainstem auditory evoked potentials. Pediatrics 1984;73: Kaplan SL, Catlin FI, Weaver T, et al. Onset of hearing loss in children with bacterial meningitis. Pediatrics 1984;73: Davis A, Wood S. The epidemiology of childhood hearing impairment: factors relevant to planning of services. Br J Audiol 1992;26:77-90.

10 Otolaryngology Volume 120 Number 5 KESSER et al Novak M, Fifer R, Barkmeier J, et al. Labyrinthine ossification after meningitis: its implications for cochlear implantation. Otolaryngol Head Neck Surg 1990;103: Nadol J, Hsu W. Histopathologic correlation of spiral ganglion cell count and new bone formation in the cochlea following meningogenic labyrinthitis and deafness. Ann Otol Rhinol Laryngol 1991;100: Igarashi M, Saito R, Alford B, et al. Temporal bone findings in pneumococcal meningitis. Arch Otolaryngol 1974;99: Igarashi M, Schuknecht H. Pneumococcic otitis media, meningitis, and labyrinthitis: a human temporal bone report. Arch Otolaryngol 1962;76: Rodriguez AF, Kaplan SL, Hawkins EP, et al. Hematogenous pneumococcal meningitis in the infant rat: description of a model. J Infect Dis 1991;164: Bhatt S, Halpin C, Hsu W, et al. Hearing loss and pneumococcal meningitis: an animal model. Laryngoscope 1991;101: Bhatt S, Lauretano A, Cabellos C, et al. Progression of hearing loss in experimental pneumococcal meningitis: correlation with cerebrospinal fluid cytochemistry. J Infect Dis 1993;167: Wiedermann BL, Hawkins EP, Johnson GS, et al. Pathogenesis of labyrinthitis associated with Haemophilus influenzae type B meningitis in rats. J Infect Dis 1986;153: Eavey RD, Gao Y-Z, Schuknecht HF, et al. Otologic features of bacterial meningitis of childhood. J Pediatr 1985;106: Blank A, Davis G, VanDeWater T, et al. Acute Streptococcus pneumoniae meningogenic labyrinthitis: an experimental guinea pig model and literature review. Arch Otolaryngol Head Neck Surg 1994;120: Winter A, Marwick S, Osborne M, et al. Ultrastructural damage to the organ of Corti during acute experimental Escherichia coli and pneumococcal meningitis in guinea pigs. Acta Otolaryngol (Stockh) 1996;116: Tarlow MJ, Comis SD, Osborne MP. Endotoxin induced damage to the cochlea in guinea pigs. Arch Dis Child 1991;66: BOUND VOLUMES AVAILABLE TO SUBSCRIBERS Bound volumes of Otolaryngology are available to subscribers (only) for the 1999 issues from the Publisher, at an individual cost of $ ($ for Canadian, $ for international subscribers) for Vols. 120 (January-June) and 121 (July- December). Shipping charges are included. Each bound volume contains subject and author indexes, and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable blue buckram with the Journal name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact Mosby, Inc., Subscription Services, Westline Industrial Drive, St. Louis, MO , USA; phone, or Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular Journal subscription.