Hearing loss in adults surviving pneumococcal meningitis is associated with otitis and pneumococcal serotype


Corresponding author: D. van de Beek, Department of Neurology, Center of Infection and Immunity Amsterdam (CINIMA), Academic Medical Centre, University of Amsterdam, PO Box 22660, 1100DD Amsterdam, the Netherlands
E-mail: d.vandebeek@amc.uva.nl


Clin Microbiol Infect 2012; 18: 849–855


We assessed the incidence of hearing loss and its relationship with clinical characteristics and pneumococcal serotypes in adults surviving pneumococcal meningitis. We analysed hearing loss in 531 adults surviving pneumococcal meningitis included in two prospective nationwide cohort studies performed from April 1998 through to October 2002 and March 2006 through to January 2009. Hearing loss was evaluated on admission and discharge for all patients. Severe hearing loss was assessed by pure tone average on audiology and corrected for age, or by the combination of hearing loss on discharge and a score on the Glasgow Outcome Scale below 5, which could not be explained by other neurological sequelae. A total of 531 episodes of pneumococcal meningitis with non-lethal outcome were included. Predisposing conditions for pneumococcal meningitis were present in the majority of patients (64%), most commonly otitis (36%). Hearing loss was present at discharge in 116 patients (22%) and was classified as mild in 53% and severe in 47%. Hearing loss was related to otitis (odds ratio [OR], 2.58; 95% confidence interval [CI], 1.66–4.02; p < 0.001) and inversely related to serotype 23F infection (OR, 0.36; 95% CI, 0.13–0.98; p = 0.025), but not with parameters of disease severity or indicators of cerebrospinal fluid inflammation severity. Meningitis due to pneumococcal serotype 3 was associated with the highest rate of hearing loss. Hearing loss frequently complicates pneumococcal meningitis. Risk factors for hearing loss were infection with pneumococcal serotype 23F and otitis, but not disease severity. Otitis and resulting perilympathic inflammation contribute to meningitis-associated hearing loss.


Bacterial meningitis is a severe and life-threatening infectious disease [1]. Streptococcus pneumoniae is the most severe cause of bacterial meningitis and currently accounts for ∼70% of all cases of community-acquired bacterial meningitis.[2] Hearing loss commonly complicates the clinical course of pneumococcal meningitis and is an important cause of disability, with reported rates of 7% to 36% among survivors of pneumococcal meningitis.[1,3–6]

Few studies have identified clinical risk factors for meningitis-associated hearing loss.[3–12] The most commonly described risk factors for hearing loss were age, a low level of consciousness, infection with S. pneumoniae, and cerebrospinal fluid (CSF) and serum parameters of inflammation. [3–12] A recent retrospective cohort study showed that advanced age, female sex, presence of co-morbidity and pneumococcal serotype were associated with an increased risk of hearing loss.[4] The influence of serotype on hearing loss has not been described previously and may influence vaccination policy and development.

In the current study we assessed the incidence of hearing loss in two prospective nationwide studies on pneumococcal meningitis and studied the role of pneumococcal serotypes in the development of hearing loss.


We pooled data from patients with pneumococcal meningitis from two prospective nationwide cohort studies with similar design.[3,13] In these studies adults (defined as patients older than 16 years of age) were included who had bacterial meningitis defined by positive cerebrospinal fluid (CSF) culture, and were listed in the database of the Netherlands Reference Laboratory for Bacterial Meningitis from April 1998 through to October 2002, and from March 2006 to January 2009. This laboratory receives CSF isolates from approximately 85% of all patients with bacterial meningitis in the Netherlands (population, 16.2 million).[3,13] Daily updates were provided of hospitals where patients with bacterial meningitis had been admitted in the preceding 2–6 days. Physicians were informed about the study by telephone. Patients or their legal representatives received written information concerning the study and were asked to give written informed consent for participation. Case record forms were used to collect data. Patients with negative CSF cultures or hospital-associated meningitis were excluded. Patients with an altered immune status due to the use of immunosuppressive drugs or splenectomy, diabetes mellitus or alcoholism were considered immunocompromised, as were patients infected with HIV.

Outcome was graded according to the Glasgow Outcome Scale. A score of 1 on this scale indicates death; a score of 2 a vegetative state (the patient is unable to interact with the environment); a score of 3 severe disability (the patient is unable to live independently but can follow commands); a score of 4 moderate disability (the patient is capable of living independently but unable to return to work or school); and a score of 5 mild or no disability (the patient is able to return to work or school). A favourable outcome was defined as a score of 5, and an unfavourable outcome as a score of 1–4. The Glasgow Outcome Scale is a well-validated instrument with good interobserver agreement.[14] At discharge, all surviving patients underwent a neurological examination performed by a neurologist.

Hearing loss was evaluated on admission, during admission and on discharge for all patients. Audiological follow-up was performed at the discretion of the treating physician. Audiograms performed within 1 year of admission were retrieved in patients with a history of hearing loss on admission, during admission or at discharge. Audiograms were re-evaluated by an otolaryngologist (EH) and a neurologist (SH). Pure tone audiometry at 500, 1000, 2000 and 4000 Hz was used to calculate the pure tone average (PTA). The PTA was corrected for any coexistent conductive hearing loss by subtracting the air-bone gap. Finally, an age- and sex-matched dataset was used to correct the PTA values.[15] By correcting for conductive hearing loss, age and sex, we obtained the corrected pure tone average (cPTA). We categorized patients into four categories: patients with a cPTA of <10 decibels (dB) uni- or bilaterally (no hearing loss), patients with a cPTA of 10–30 dB uni- or bilaterally (mild hearing loss), patients with a cPTA of 31–70 dB uni- or bilaterally (moderate hearing loss) and patients with a cPTA >70dB (severe hearing loss).[4] We analysed patients with any hearing loss and those with moderate to severe hearing loss, which was defined as by cPTA >30 Patients with the combination of hearing loss at discharge and a score on the Glasgow Outcome Scale below 5, which was not caused by other neurological sequelae, were also categorized as having moderate to severe hearing loss. Clinical hearing loss in the case record form was not specified as uni- or bilateral hearing loss.

The Mann–Whitney U-test was used to identify differences between groups with respect to continuous variables, and dichotomous variables were compared by use of the chi-squared test. All statistical tests were two-tailed, and a p value of <0.05 was regarded as significant. Analyses were undertaken with PASW software, version 18 (SPSS, Armonk, NY, USA). We used logistic regression analysis to assess the association between potential prognostic factors and the probability of any or severe hearing loss.


A total of 709 episodes of pneumococcal meningitis in 697 patients were included in the two cohort studies; 352 in 1998–2002 and 357 in 2006–2009. One hundred and seventy-eight episodes were fatal (25%), so 531 episodes of adult pneumococcal meningitis were available for evaluation in this study.

Predisposing conditions for pneumococcal meningitis were present in 342 patients (64%), and consisted of otitis in 195 patients (36%; Table 1). On presentation, classic symptoms and signs of bacterial meningitis were present in a large proportion of patients. Headache occurred in 87% of episodes, neck stiffness in 81%, fever in 86%, and a change in mental status (defined by a Glasgow Coma Scale score below 14) in 81%. In 15% of episodes, the patients were comatose on admission, and in 9% of episodes, a hemiparesis was present on presentation. A lumbar puncture was performed in all patients, and the median white blood cell (WBC) count was 3004 (interquartile range [IQR], 893–8121).

Table 1.   Baseline characteristics of 531 patients with pneumococcal meningitis
  1. Numbers are number/number evaluated or median (interquartile range [IQR]).

  2. GCS, Glasgow coma scale; CSF, cerebrospinal fluid; ESR, erythrocyte sedimentation rate.

  3. aCSF white cell count was available for 498 episodes.

  4. bCSF protein was available for 492 episodes.

  5. cCSF/blood glucose ratio was available for 486 episodes.

  6. dCSF pressure was available for 172 episodes.

  7. eESR was available for 369 episodes.

  8. fC-reactive protein was available for 389 episodes.

  9. gThrombocyte count was available for 501 episodes.

Age (years)59 (45–68)
Male sex242/531 (46%)
Duration of symptoms >24 h251/509 (49%)
Predisposing conditions342/531 (64%)
 Otitis195/522 (37%)
 Sinusitis90/509 (18%)
 Pneumonia66/523 (13%)
 Immunocompromise111/530 (21%)
Symptoms and signs on admission
 Headache417/479 (87%)
 Neck stiffness420/519 (81%)
 Fever385/447 (86%)
 Diastolic blood pressure <60 mmHg39/523 (7%)
GCS on admission11 (9–13)
 <14 (indicating change in mental status)428/531 (81%)
 <8 (indicating coma)80/531 (15%)
Focal neurological abnormalities on admission
 Hemiparesis47/498 (9%)
 Cranial nerve palsy66/498 (13%)
 Hearing loss26/454 (6%)
Indices of CSF inflammation
 White cell count (cells/mm3)a3004 (893–8121)
 CSF protein (g/L)b3.9 (2.4–5.9)
 CSF/blood glucose ratioc0.04 (0.01–0.24)
 CSF pressure (cm H2O)d40 (29–50)
Blood tests
 ESR, mm/he42 (22–72)
 C-reactive protein, mg/Lf185 (92–203)
 Thrombocyte count, platelets/mm3g207 (163–262)

Hearing loss was noted during 187 of 531 episodes (35%): on or during admission only in 71 patients, and at discharge in 116 patients (Table 2). Audiological examination was performed in 82 of 187 patients with hearing loss (44%). According to our predefined criteria, hearing loss was present in 73 audiograms and was classified as mild in 30 episodes (37%), moderate in 25 (30%) and severe in 18 episodes (22%). For the other 105 episodes complicated by hearing loss in which audiology was not performed, hearing loss was still present on discharge in 43 (41%) of these patients. In 12 (28%) of these episodes without audiograms, hearing loss was classified as severe hearing loss as it caused unfavourable outcome. Combining the clinical and audiology data we concluded that any hearing loss was present on discharge in 116 episodes (22%). The degree of hearing loss was classified as moderate to severe hearing loss in 55 episodes (47%). Clinical features of patients with audiological examination and with clinical hearing loss were similar, except for diastolic hypotension, which was more common on presentation in patients receiving audiological follow-up (17% vs. 2%, p = 0.03). Hearing loss in audiograms was most pronounced in the highest frequency (4000 Hz).

Table 2.   Characteristics of 531 patients with pneumococcal meningitis according to occurrence of hearing loss
CharacteristicAny hearing lossNo hearing lossp value
  1. Numbers are number/number evaluated or median (IQR).

  2. GCS, Glasgow coma scale; CSF, cerebrospinal fluid.

  3. aCSF white cell count was available for 498 episodes.

  4. bCSF protein was available for 492 episodes.

  5. cCSF/blood glucose ratio was available for 486 episodes.

No. of episodes116415 
Age (years)56 ± 1457 ± 150.49
Male sex49/116 (42%)193/415 (47%)0.41
Duration of symptoms >24 h53/111 (48%)198/398 (50%)0.71
Predisposing conditions
 Otitis64/116 (55%)131/406 (32%)<0.001
 Sinusitis24/109 (22%)66/400 (17%)0.18
 Pneumonia15/114 (13%)51/409 (12%)0.84
 Immunocompromise22/116 (19%)89/414 (21%)0.55
Symptoms and signs on admission
 Headache96/108 (89%)321/371 (87%)0.52
 Neck stiffness91/114 (80%)329/405 (81%)0.73
 Diastolic blood pressure <60 mmHg13/115 (11%)26/408 (6%)0.08
 Hemiparesis12/104 (12%)35/394 (9%)0.41
 Score on GCS
  <14 (indicating change in mental status)92/115 (80%)336/415 (81%)0.81
  <8 (indicating coma)13/115 (11%)67/415 (16%)0.20
Indices of CSF inflammation
 White cell count (cells/mm3)a3370 (762–8692)2933 (902–8000)0.40
 CSF protein (g/L)b4.1 (2.8–5.8)3.8 (2.3–5.9)0.27
 CSF/blood glucose ratioc0.04 (0.00–0.21)0.05 (0.01–0.24)0.57

In a univariate analysis, the presence of otitis on presentation was associated with hearing loss at discharge (odds ratio [OR], 2.58; 95% confidence interval [CI], 1.66–4.02; p < 0.001; Table 2). Of the 195 patients with otitis, 64 developed hearing loss (33%). The association with otitis was present in patients with audiologically confirmed hearing loss (p < 0.001) and in patients with clinical hearing loss (p = 0.003). Clinical severity, as reflected in scores on the Glasgow Coma Scale, and parameters of CSF inflammation, were not predictive for the development of hearing loss. Patients with severe hearing loss had lower CSF/blood glucose ratios, as compared with those without severe hearing loss (ratio of 0.03 [IQR, 0.00–0.08] vs. 0.05 [IQR 0.01–0.25]; p = 0.02), but all other clinical and laboratory characteristics were similar between groups. In a multivariate analysis with possible predictors of hearing loss (sinusitis, otitis, diastolic hypotension and coma on admission), otitis on admission remained the only predictor of hearing loss (p < 0.001). Treatment data for otitis were not available for a sufficient number of patients to assess the influence of, for example, mastoidectomy on hearing loss on discharge.

Serotype analysis was performed for 504 of 531 pneumococcal isolates (95%; Table 3). The most common capsular serotypes were 3, 7F, 23F, 14, 6B and 19F, together accounting for 47% of isolates. The coverage of the 7-, 10- and 13-valent vaccines of these isolates would be 40%, 53% and 68%, respectively. Compared with the reference serotype 3 (the most common serotype), the risk of hearing loss was lower for all other serotypes. Any hearing loss during clinical course occurred less frequently in patients infected with S. pneumoniae serotype 14 (OR, 0.28; 95% CI, 0.09–0.80; p = 0.007) and serotype 23F (OR, 0.36; 95% CI, 0.13–0.98; p = 0.025), as compared with serotype 3. Infection with the S. pneumoniae serotype 23F was also associated with less hearing loss at discharge (OR, 0.32; 95% CI, 0.08–1.19; p = 0.05) and showed a trend towards less severe hearing loss at discharge (OR, 0.19; 95% CI, 0.01–1.71; p = 0.10), but confidence intervals were wide because of small numbers. S. pneumoniae serotype 14 was not related to hearing loss at discharge.

Table 3.   Distribution of pneumococcal serotypes in 504 episodes of pneumococcal meningitis according to occurrence of hearing loss
SerotypeFrequency (%)Hearing loss during clinical course (%)Odds ratio (95% CI)pHearing loss at discharge (%)Odds ratio (95% CI)p
  1. aIncluded in seven valent pneumococcal conjugate vaccine.

14a38 (8)8 (30)0.28 (0.09–0.80)0.0076 (16)0.52 (0.15–1.70)0.23
19Fa28 (6)10 (36)0.58 (0.20–1.66)0.268 (29)1.11 (0.35–3.51)0.85
22F23 (5)7 (30)0.46 (0.14–1.46)0.145 (22)0.77 (0.20–2.83)0.66
23Fa39 (8)10 (26)0.36 (0.13–0.98)0.0254 (10)0.32 (0.08–1.19)0.05
349 (10)24 (49)1.00 (reference)13 (27)1.00 (reference)
4a22 (4)9 (41)0.72 (0.23–2.24)0.536 (27)1.04 (0.29–3.67)0.94
6Ba33 (7)12 (36)0.60 (0.22–1.61)0.268 (24)0.89 (0.28–2.74)0.82
7F48 (10)18 (38)0.63 (0.26–1.52)0.259 (19)0.64 (0.22–1.85)0.84
831 (6)12 (39)0.66 (0.24–1.81)0.378 (26)0.96 (0.31–3.00)0.94
9Va27 (5)11 (41)0.72 (0.25–0.52)0.498 (30)1.17 (0.36–3.72)0.77
Other serotypes166 (33)58 (35)38 (22)
Total504 (100)179 (36)113 (22)

Initial antibiotic treatment consisted of penicillin or amoxicillin monotherapy in 41% of patients, third-generation cephalosporin monotherapy in 22% of patients, and a combination of penicillin or amoxicillin and a third-generation cephalosporin in 27% of patients, and different regimens were used in 10% of patients. Use of aminoglycosides was reported in 25 episodes and hearing loss at discharge was found in three of those patients (12%). Aminoglycosides are associated with ototoxicity and hearing loss; however, there was no increased risk of hearing loss in patients receiving aminoglycosides in our study.[16]

Between the first and second cohort studies, after the publication of a clinical trial and meta-analysis,[17,18] adjunctive dexamethasone was introduced as routine treatment in the Netherlands.[13] In our patient cohorts, the standard regimen of dexamethasone, 10 mg four times daily, started with or before the first dose of antibiotics, was administered in 240 of 531 episodes (45%). The proportion of patients with hearing loss treated with or without standard regimen dexamethasone was similar (48 of 240 [20%] vs. 68 of 291 [23%]; p = 0.35). The proportion of patients with severe hearing loss treated with or without standard regimen dexamethasone was also similar (20 of 240 [8.3%] vs. 35 of 291 [12%]; p = 0.16).


The current study shows that hearing loss frequently complicates pneumococcal meningitis. In our adult population, hearing loss was present in about one in four patients surviving their disease. Previous studies have reported rates of hearing loss varying from 7 to 36%. A recent retrospective Danish study based on a nationwide registration during a 5-year period (1999–2003) reported a rate of 54%. The majority of these patients had hearing loss on audiological examination only, without complaints. Our study was performed nationwide and, therefore, we were able to study a representative sample of adults surviving pneumococcal meningitis. Our prospective approach allowed us to collect comprehensive data on signs and symptoms, clinical course and outcome.

We identified otitis as the main risk factor for hearing loss (OR, 2.58; CI, 1.66–4.02; p < 0.001). This is consistent with findings of a previous prospective cohort study that reported a similar rate of hearing loss among patients with pneumococcal meningitis and otitis (33%).[7] Other studies have found a relation between markers of CSF inflammation and hearing loss.[4] Our finding that hearing loss is related to otitis, but not to disease severity or levels of CSF inflammation, suggests a role for otitis and resulting perilympathic inflammation in the pathogenesis of meningitis-associated hearing loss, consistent with the aggravated high-frequency loss in our patients. Research in animals and patients with bacterial meningitis showed that the main site of the lesion in meningitis-associated hearing is the cochlea.[5] Predominant morphological correlates of acute meningitis-associated hearing loss are damage to the blood-labyrinth barrier, hair cells and spiral ganglion.[5] However, the route of entry of bacteria into the cochlea has been subject to debate. Bacteria can reach the cochlea via the bloodstream, the vestibulocochlear nerve, or through the perilymphatic duct. Klein et al. [5] concluded that the most probable route of infection is spread of bacteria from the subarachnoid space through the perilymphatic duct to the perilymphatic space of the cochlea. However, in otitis media toxic substances (bacterial products and inflammatory mediators) may pass from the middle ear passage through the round window membrane, causing perilymphatic inflammation, starting within the cochlear basal turn.[19] This localization is consistent with the prevalence of sensorineural hearing loss at higher frequencies, as found in patients with post-meningitis hearing loss.[5] Subsequently, these toxic substances may cause permanent cochlear damage, resulting in hearing loss.[19] The association of hearing loss with otitis instead of severity of CSF inflammation would further be supported if the side of otitis corresponded with the side of hearing loss. The side of otitis, mastoiditis and hearing loss was not systematically registered in our study, which is a limitation of our results and leaves room for both versions of the pathophysiology of hearing loss in pneumococcal meningitis.

Adjunctive dexamethasone therapy was not associated with a lower risk of hearing loss among survivors of pneumococcal meningitis, although a trend was found for reduced risk of disabling hearing loss in patients receiving dexamethasone. Adjunctive anti-inflammatory therapies have been investigated, with the aim of decreasing rates of post-meningitis hearing loss. Systemic adjunctive steroid treatment reduced long-term hearing loss in experimental pneumococcal meningitis.[20] Clinical trials and recent meta-analyses have shown that adjunctive dexamethasone therapy prevents hearing loss in patients with bacterial meningitis.[21,22] We recently showed that dexamethasone therapy has been implemented on a large scale as adjunctive treatment of adults with pneumococcal meningitis in the Netherlands.[13] The prognosis of pneumococcal meningitis on a national level has substantially improved after the introduction of adjunctive dexamethasone therapy.[13] However, as dexamethasone improves outcome and reduces mortality, a larger proportion of patients survive and are at risk of developing hearing loss. Other adjunctive therapies to prevent hearing loss (i.e. antioxidant therapy and neurotrophin-3) have been investigated in animal models and may present promising future treatment options.[23,24]

Our study has several limitations. As audiological examination was performed at the discretion of the treating physician, only 16% of patients received audiological examination. We extended the definition of hearing loss to describe all patients with clinical hearing loss on discharge, even though bedside assessment of hearing loss is of limited value.[4] The limited availability of audiological examinations hinders the interpretation of our results, but reflects clinical practice. Furthermore, in patients with poor clinical outcome, hearing loss may be masked by clinical condition (e.g. decreased level of consciousness). Therefore these patients may have been missed. Finally, treatment data for otitis (mastoidectomy) were not reported, which is a possible confounder for hearing loss on discharge. However, this does not affect the identified association between otitis and hearing loss as a possible treatment effect is included in this analysis.

Our reference pneumococcal serotype, serotype 3, was associated with the highest risk of hearing loss. Serotype 23F was associated with a significantly lower risk of hearing loss, as compared with the serotype 3. The Danish study showed less hearing loss in patients infected with serotype 6B, but also found the highest risk in patients infected with serotype 3.[4] A systematic review and meta-analysis of serotype-specific disease outcomes for patients with pneumonia and meningitis showed that the relative risk of death in patients infected with serotype 3 was increased.[25] Experimental animal studies are needed to further elucidate the role of different pneumococcal serotypes in the risk of meningitis-associated hearing loss.

The high incidence of hearing loss in patients with pneumococcal meningitis warrants consultation of an otolaryngologist in patients with suspected hearing loss.[1] Otitis on presentation is common in pneumococcal meningitis and may require invasive treatment to remove the focus of infection. Hearing loss may not be clinically evident in patients (e.g. those with an altered state of consciousness or patients requiring mechanical ventilation). The threshold for perfoming audiometric evaluation should therefore be low. In patients with hearing loss due to pneumococcal meningitis, obliteration of the cochlear lumen may occur in the weeks after the hearing loss developed. This will further diminish sensorineural hearing and adversely affects options for improving hearing, such as cochlear implantation, even in patients with initial mild hearing loss.[26,27] In these cases, cochlear implantation before obliteration has occurred may prevent a permanent disabling hearing impairment.[28,29] Therefore, otolaryngological evaluation in these patients should be performed in the acute phase of disease.


We are indebted to many physicians in the Netherlands for their cooperation. We thank Floortje Ruijter for her assistance with retrieving audiograms.

Financial Support

This study has been funded by grants from the Netherlands Organization for Health Research and Development (ZonMw; NWO-Veni grant 2006 [916.76.023], NWO-Vidi grant 2010 [917.113.58], both to D.v.d.B) and the Academic Medical Center (AMC Fellowship 2008 to D.v.d.B); M.C.B. is supported by the European Society of Clinical Microbiology and Infectious Diseases and European Federation of Neurologic Societies. The other authors received no financial support.

Potential Conflicts of Interest

All authors: no conflicts.

Transparency Declaration

The authors have no dual or conflicting interest to declare.