Keywords otitis media - meningitis - adult - imaging - computed tomography
Introduction
Intracranial complications secondary to otitis media (OM) arise when the infectious/inflammatory
process extends beyond the temporal bone to the adjacent organs, such as the dura.
With the advent of antibiotic and surgical therapies over the last century, the incidence
of intratemporal/intracranial complications from OM has substantially decreased.[1 ]
[2 ] Otogenic meningitis (OgM) is the most commonly reported intracranial complication
of OM in adults, and its incidence rate has been reported to range from 0.32 to 0.42%
of all adult OM cases.[3 ]
[4 ]
[5 ]
[6 ]
[7 ] Otogenic meningitis is typically diagnosed in adults presenting with a recent history
of OM, with signs and symptoms suggestive of meningitis. After obtaining a cerebrospinal
fluid (CSF) sample for chemistry and bacteriology via lumbar puncture (LP), the mainstay
of OgM treatment consists of parenteral antibiotics and corticosteroids. Myringotomy
can also be of a therapeutic value, because it aims to drain the middle ear cavity
and obtain cultures. This is in contrast with the common practice in the preantibiotic
era, which included cortical/radical mastoidectomy.[8 ]
Unlike the pediatric population, who had seen a major intervention aimed to decrease
the burden of pneumococcal infections with the implementation of pneumococcal conjugate
vaccines (PCVs), adult patients, who are not the usual target population for these
vaccines, may benefit from the herd immunity effect. A recent study from Connecticut, USA, evaluated the proportion rates of invasive
pneumococcal disease (IPD) cases, including meningitis, among adults aged > 40 years
old. The study showed that in areas with a higher proportion of children that did
not complete the recommended 4-dose PCV7 series, there was a higher incidence of overall
adult IPD cases, including meningitis, caused by PCV7 serotypes.[9 ] A contemporary Dutch audit also showed that following the sequential PCV7 and PCV10
immunization of the pediatric population in the Netherlands, the overall IPD incidence
proportionally decreased also in men aged between 20 and 39 years old, who, likewise,
were not the target population of these vaccines.[10 ] Yet, the data on the changing epidemiology of adult OgM is partial, since most studies
pool meningitis cases with other IPDs, or do not differentiate OgM from all-cause
meningitis cases, and conclusions are hard to withdraw. In this single-center report,
we sought to report on our experience in a subgroup of adult patients with OgM.
Patients and Methods
Ethical Considerations
The present study was approved by the local Institutional Review Board (protocol number:
WOMC-0229–16).
Objectives
The primary objective was to study the incidence of OgM from all-cause adult meningitis
cases, and the secondary outcome was to examine the impact of the indirect effect
of the implementation of PCVs in the local National Immunization Program (NIP) on
adult OgM cases.
Study Design and Population
We identified retrospectively all adult patients (>18 years old) who presented to
the emergency department and were later hospitalized between January 1, 2005 and December
31, 2014 with acute meningitis and had International Classification of Diseases-9
(ICD-9) discharge codes of acute meningitis (320.X, bacterial meningitis; 321.X, meningitis
due to other organisms; and 322.X, meningitis of unspecified cause). Adults with previous
ear/mastoid disease (i.e., cholesteatoma) or surgery were excluded, as well as patients
who developed acute meningitis secondary to neurosurgical procedures or if they had
any known immunodeficiencies.
Setting
The study was performed in a secondary healthcare center located in central Israel.
This hospital is one of the 9 public hospitals in the greater Tel Aviv area, which
provide healthcare to 3.85 million people (2016 data), and patients have access to
all of them when needed. Our local community has some 300,000 adults, and there were
no appreciable differences in the population size during the study years. On-call
neurology and otolaryngology consultations are available 24/7. Computerized tomography
(CT) scans are available 24/7, and magnetic resonance imaging (MRI) studies are performed
within several hours.
Otogenic Meningitis
The diagnosis of meningitis was based on clinical findings (nuchal rigidity, photophobia,
papilledema, nystagmus, gait instability), in the context of systemic signs and symptoms
(fever, headache, nausea/vomiting, confusion, seizures). Laboratory tests for CSF
samples were compatible with meningitis findings: elevated opening pressure (> 25cm
H2 O); elevated white blood cell (WBC) count, > 100 cells/µL; low glucose level (< 40%
of serum glucose) and elevated protein level (> 50 mg/dL). All CSF samples were also
sent for Gram staining and standard microbiological cultures.
Per our protocol, any patient with suspected meningitis undergoes neurologic and ophthalmologic
evaluation, and is always sent for a head computed tomography (CT) before LP is performed.
Routine otolaryngologic evaluation for each adult meningitis is not performed.
The diagnosis of OgM was made in patients presenting with concurrent meningitis and
OM, diagnosed on otoscopy and/or ear swab culture, but also ad hoc, after reviewing
imaging studies, mostly CT scans, which had shown the presence of opacification in
the tympanomastoid cavity. In these ad hoc cases, patients were subsequently referred
for otomicroscopy in the otolaryngology service, which reinforced the diagnosis of
OM and were managed accordingly.
Medical charts of eligible patients were reviewed and verified for the accuracy of
OgM episodes independently by 2 authors (Shemesh S., Tamir S. O.). Data collected
and analyzed included age, gender, history of current disease, history of previous
OM episodes, past medical and surgical history, laterality, pre-hospital antibiotic
treatment, OgM symptoms and signs, ear and CSF culture results, antibiotic treatment
during hospitalization, timing of imaging studies, findings on CT/MRI studies, OgM-related
sequelae surgical interventions during hospitalization (myringotomy, mastoidectomy
or neurosurgical drainage). We did not collect data on viral studies because they
were not consistently performed in all adult OgM cases.
Because meningitis is an IPD, it is requested by the Israeli Ministry of Health to
send all Streptococcus pneumoniae isolates from CSF cultures for serotyping in its central laboratory for national
surveillance purposes. These data are not sent to the sending hospitals because it
has no additive, practical therapeutic value in the management of meningitis patients
in real time. Because this information is irrelevant for decision-making purposes
(unlike bacterial antibiotic sensitivity tests), we did not collect this data.
Pneumococcal Conjugated Vaccines Periods
In brief, PCV7 was introduced in the Israeli NIP in children in July 2009, which was
shortly replaced later in November 2010 by PCV13, which offered a broader coverage
of 6 additional serotypes. Pneumococcal conjugated vaccines are routinely administered
to infants only in their first 2 years of life. By June 2011 and December 2012, ∼
80% and ∼ 90%, respectively, of Israeli children between 7 and 11 months old received
≥ 2 PCV7 and/or PCV13 doses; by June 2014, ∼ 95% received ≥ 2 PCV13 doses. By June
2011 and December 2012, 36 and 87%, respectively, of children between 24 and 35 months
old, received ≥ 3 PCV7/PCV13 doses; by June 2014, 91% received ≥ 3 PCV13 doses.[11 ]
The period between 2005 and 2009 was considered as the “pre-PCV years”, when the general
population could be considered as naïve, vaccine-free, and the period between 2010
and 2014 was considered as the “post-PCV years”, when the pediatric population has
been gradually vaccinated. Adults are not requested to be immunized with PCVs, but
may indirectly be affected by this intervention in a nonmeasurable way. Therefore,
we hypothesized that the clinical presentation of OgM may have changed in the post-PCV
years.
Statistical Analysis
Data were recorded on a Microsoft Excel spreadsheet (Microsoft Corporation, Redmond,
WA, USA). The unit of analysis was an OgM episode. Statistical analysis was performed
using IBM SPSS Statistics for Windows, Version 21.0 (IBM Corp., Armonk, NY, USA).
Events were compared using the chi-squared test or the Fisher exact text, appropriately.
All of the tests were two-sided and were considered significant at p < 0.05.
Results
Demographics
Overall, there were 45 cases of all-cause adult meningitis. Of them, 26 (58%) cases
were classified as OgM, and were the focus population of our study. Of them, 20 were
diagnosed by concurrent otomicroscopy and diagnostic imaging modality (CT scan) upon
admission, and 6 additional cases were initially diagnosed as having normal otoscopy
or did not undergo otoscopic examination before the CT scan had been performed. A
total of 10 (38%) of the study patients were male. None of the patients had received
any PCV dose. Demographic and clinical data are presented in [Table 1 ]. In 3 (12%) cases, a history of recent OM (< 14 days prior to admission) was documented.
A preadmission antibiotic treatment was documented in 5 (19%) patients. The mean hospitalization
duration was 20 days (range: 3–81 days).
Table 1
Demographic and clinical presentation of otogenic meningitis (n = 26)
Variance
Value
Male
10 (38%)
Mean age, years old (SD)
62.4 (13)
Previous otologic history
3 (11%)
Preadmission antibiotic treatment
5 (19%)
Otitis media (on otoscopy)
21 (81%)
Otogenic meningitis diagnosed after CT
6 (23%)
Myringotomy performed
6 (23%)
Abbreviations: CT, computed tomography; SD, standard deviation.
Otogenic Meningitis Presentation and Management
Upon admission, at least 2 typical signs/symptoms of meningitis were documented in
25 (96%) patients, and included headache (23 patients; 88%), nuchal rigidity (n = 23; 88%), body temperature > 38.1°c (n = 21; 81%), confusion (n = 21; 81%), nausea or vomiting (n = 19; 73%), photophobia (n = 7; 27%) and papilledema (n = 2; 7%). Otogenic signs/symptoms were not universal: 18 (69%) patients reported
otalgia, and 5 (19%) reported otorrhea prior to hospital admission.
Intravenous antibiotic treatment was administered in all OgM cases. Ceftriaxone was
the most frequent treatment, prescribed in 24 (92%) patients, followed by vancomycin
(n = 14; 54%) and ampicillin (n = 11; 42%). In 18 (69%) patients, the antibiotic treatment was switched during the
hospitalization course, due to bacterial culture results and antibiotic resistance
studies reported during the course of the hospitalization.
Bacteriology
Cultures were collected and sent for bacterial studies from LP samples (24 patients,
92%), myringotomy (n = 18; 69%), spontaneous otorrhea (n = 1; 3%) and from the mastoid cavity during surgery (n = 1; 3%). Of the 28 analyzed cultures, 22 (79%) grew bacteria. Of the positive cultures,
S. pneumoniae grew in 20 (91%) cases, Listeria monocytogenes in 1 (5%) case, and Haemophilus influenzae in 1 (5%) case. There were no positive cultures for other typical meningitis/OM pathogens,
such as Neisseria meningiditis or Moraxella catarrhalis . There were no cases of mixed bacterial growth. One patient presented with 2 different
OgM episodes, with a complete resolution between them. His 1st CSF culture grew S. pneumoniae , and the 2nd and 3rd cultures obtained during the subsequent hospitalization 3 months later also grew
S. pneumoniae again in both CSF and mastoidectomy cultures, but had different sensitivities to
antibiotic agents from the S. pneumoniae isolated in the 1st episode. He had no OM episodes in between these episodes.
All of the the 26 (100%) OgM patients had CSF cultures obtained during LP. Of them,
20 (77%) had also ear cultures, obtained during myringotomy, swabbing of the external
ear and mastoidectomy. Of them, 8 (40%) had bacterial growth. In these patients, ear
cultures were obtained via myringotomy in 6 (75%) patients, on the 2nd day following admission, on average (range 0–8 days); culture from spontaneous otorrhea
in 1 (13%) patient, and culture from mastoidectomy cavity in 1 (13%) patient. [Table 2 ] shows the concordance between the CSF and the ear cultures in those patients. Of
the 6 patients who underwent myringotomy, 3 (50%) patients had identical CSF and ear
culture results (which both grew S. pneumoniae ). The other 3 (50%) patients who underwent myringotomy had Staphylococcus coagulase negative (which was interpreted as contaminant) or no growth in their ear
cultures. Because the patients had already been treated with a broad-spectrum antibiotic
treatment that covered these pathogens beforehand, there was no need for a change
in the antibiotic policy.
Table 2
Comparison between cerebrospinal fluid and middle ear cultures (n = 8)
Patient
Ear Culture Source
Day of Culture
CSF Culture Result
Ear Culture Result
1
Myringotomy
1
Streptococcus pneumoniae
Streptococcus pneumoniae
2
Myringotomy
0
Streptococcus pneumoniae
Staphylococcus coagulase negative
3
Myringotomy
0
Streptococcus pneumoniae
No growth
4
Myringotomy
3
No Growth
Staphylococcus coagulase negative
5
Myringotomy
8
Streptococcus pneumoniae
No growth
6
Spontaneous otorrhea
1
Streptococcus pneumoniae
Streptococcus pneumoniae
7
Mastoidectomy
3 months
Streptococcus pneumoniae
Streptococcus pneumoniae
8
Myringotomy
1
Streptococcus pneumoniae
No growth
During the pre-PCV years, there were 12 (46%) cases, compared with 14 (54%) cases
during the post-PCV years. Since there were no appreciable changes in the population
size in our area, this reflects a stable OgM admission rate during the whole study
period. Of the 12 pre-PCV cases, 9 (75%) had S. pneumoniae grown in their CSF cultures, while of the 14 post-PCV cases, 11 (78%) were positive
for S. pneumoniae in their CSF cultures (p = 0.59). Since OM is not considered to be IPD and serotyping from ear cultures is
not routinely performed in clinical settings like ours, we could not show if there
was a serotype shift phenomenon from vaccine to non-vaccine strains.
The average age and hospitalization duration was similar in both groups. There were
significantly more patients with a medical history of diabetes mellitus and hypertension
in the post-PCV years (25%, pre-PCV years versus 71%, post-PCV years). A total of
5 (36%) patients in the post-PCV years had a bilateral disease, while none of the
pre-PCV patients had a bilateral disease. Due to these small numbers, we could not
perform the statistical analysis.
Complications varied from neurological sequelae, such as partial to complete paresis
and alteration in consciousness level, permanent sensoneural hearing loss, to cardiac
arrest and death. In the pre-PCV years, there were 4 (33%) patients with OgM sequelae,
compared with 6 (43%) in the post-PCV patients (p = 0.39). The overall mortality rate was 15%: 3 (25%) patients in the pre-PCV group
eventually died from OgM complications, and 1 (7%) patient in the post-PCV died of
neurological sequalae (p = 0.12).
Imaging Studies
All of the patients underwent CT scans, and 2 patients also underwent MRI studies
for suspected intracranial complications (both were in the post-PCV years). A total
of 4 patients (15%) had impacted cerumen blocking the external ear canal, plausibly
leading to the delayed diagnosis of OgM ([Fig. 1A ]).
Fig. 1 Representative computed tomography (CT) Scans of Adult Patients with Ad hoc Otogenic Meningitis. (A ) Axial CT scan showing an impacted cerumen (white arrow), obscuring the right tympanic
membrane. (B ) Axial CT scan showing bilateral opacifications of the tympanomastoid cavities (white
star) as well as the middle ear cleft and ethmoidal sinusitis (white arrow). (C ) Coronal CT scan showing opacification of the right middle ear cavity and ipsilateral
dehiscence of the tegmen tympani (white arrow). (D ) Axial CT scan showing opacification of the left petrous apex of the temporal bone
(white arrow), compatible with petrous apicitis.
As for findings in imaging studies, 6 (50%) patients in the pre-PCV years had tympanomastoid
opacifications alone, while the other patients had concomitant rhinosinusitis findings
([Fig. 1B ]). Osteomyelitis of the temporal bone was observed in 6 (43%) patients. Tegmen dehiscence
was the most common pathology ([Fig. 1C ]). In the post-PCV years, there were 7 (50%) patients with isolated OgM, while the
others had concomitant rhinosinusitis findings in their imaging studies. Osteomyelitis
was observed in 5 (42%) patients, with, again, tegmen dehiscence as the common pathology,
and one case of petrous apicitis ([Fig. 1D ]). We did not find any comorbidity that had significant correlation with either pathology
observed in imaging studies.
Discussion
We report that 58% of all-cause adult meningitis in our cohort could be classified
as originating from an otogenic source. The overall outcome of adult OgM patients
treated with intravenous antibiotic therapy was good, and only a few required surgical
interventions, which were minor in most cases (only myringotomy). There were no considerable
differences in the pre- and post-PCV years in the epidemiology and clinical presentation
of adult OgM. The present retrospective study did not find any change in the proportion
of OgM growing S. penumoniae on ear or CSF cultures.
It is widely accepted that OM is the most common cause for pediatric meningitis, due
to the proximity of the tympanomastoid cavity to the dura, as well as the presence
of inner ear malformations and the relative ease for the hematogenous spread of the
infectious process from the middle ear cleft.[12 ] Connections between the hematopoietic bone marrow of the temporal bone and the middle
ear in autopsies of infants with meningitis and OM have also been suggested as a potential
route for the infection.[13 ] However, OM as a cause for adult OgM is a less accepted notion, and in our view, under-appreciated. Ear-related signs
and symptoms were noticeable in only 70% of our patients (a few patient families were
interviewed and stated this fact only after a “cause directed” interview, because
the patient could not be interviewed due to his neurological status), and OM as a
cause for meningitis was absent in 20% of the cases. We showed an incidence rate of
56% of OgM from all-cause meningitis, which is higher than previously published pre-PCV
era reports: 15% (13/87) in Bristol, United Kingdom,[7 ] 4% (8/201) in Oporto, Portugal,[14 ] or 30% in the Danish National Registry.[15 ] Unlike these studies, which also included infants and children, our population consisted
of verbal adult patients who reported their symptoms and could be questioned for their
complaints.
The association between PCVs and the decrease in pediatric meningitis cases has been
previously described in many countries that implemented PCVs in their NIP.[16 ]
[17 ]
[18 ] In the U.S., the decrease in pneumococcal meningitis was the greatest in the early
post-PCV7 period (introduced in 2000), with an additional small decrease in the post-PCV13
years (introduced in 2010).[2 ] The most pronounced decline in meningitis admission prevalence occurred in children < 1
year old, followed by children aged between 1 and 2 years old. In older children,
the decrease was modest, in accordance with the acute OM prevalence distribution in
the pediatric population.
The introduction of PVCs has positively affected the incidence of IPDs in adults as
well, including all-cause meningitis, as observed in the changes in S. pneumoniae serotypes before and after the implementation of PCVs,[19 ] in concordance with the observed decrease with the national uptake of the vaccines[20 ] and the reduction in overall all-cause and prevalence of pneumococcal meningitis.[11 ] The biggest evidence comes from an epidemiological study that demonstrated a downward
trend in the pneumococcal meningitis hospitalization rates across the U.S. in the
post-PCV13 years, but without a substantial change in the fatality rate.[21 ] Although this report does not specifically address OgM as a subgroup, the authors
showed that in children < 2 years old, who are the target population of PCVs and are
at the peak age for OM, the average annual pneumococcal meningitis admission rate
decreased by 45%, from 2.19 in the late-PCV7 years to 1.20 in the post-PCV13 years
(p = 0.10). In adults aged between 18 and 39 years old, there was also a significant
downward trend, from 0.25 to 0.15 (p = 0.02), similarly to adults aged between 40 and 64 years old, from 0.95 to 0.54
(p = 0.03); and in patients ≥ 65 years old, the admission rate decreased by 38%, from
1.02 to 0.63, though it did not reach statistical significance (p = 0.08) (all rates shown are for pneumococcal meningitis admission cases per 100,000
individuals). The same trend was also observed in nonpneumococcal meningitis admission
rates, but did not reach statistical insignificance. The same decrease has also been
reported in Colombia, where the observed over the expected monthly meningitis rate
decreased in 90.5% in the post-PCV period (2012–2015), when compared with the pre-PCV
period.[22 ] We could not show such a change in the incidence of OgM in the pre- and post-PCV
periods, possibly due to the small sample size. Also, we were unable to show such
a decrease in the S. pneumonia -positive cultures in the post-PCV years.
Our work is based on a meticulous chart review of selected OgM patients, regarding
their ear disease and imaging studies. To our knowledge, this is the first study focusing
on this subgroup of adult meningitis patients in the post-PCV era. We showed that
ad hoc diagnosis is important and relevant for the diagnosis of OgM early during the
disease, due to favorable outcomes. We acknowledge limitations to our work: 1) small
study group, proportional to the size of our single center. This limited us from performing
meaningful statistical analyses. Yet, other reports also include only a few dozens
of patients[4 ]
[7 ]
[12 ]
[23 ]; 2) we do not have data on OgM hospitalization rates in our neighboring hospitals;
3) no serotype data; and 4) due to the retrospective nature of the present study,
we could not get into the details why otoscopy was not performed in all cases or why
the interpretation of the first otoscopy was considered to be normal.
Conclusion
To conclude, OgM is underdiagnosed. Ear-related signs and symptoms in adult patients
with meningitis are not universal, and OM as a cause for meningitis can be easily
missed. Since OM pathogens may be different from meningitis pathogens and the need
to drain the middle ear in OgM cases, we suggest that micro-otoscopy for any adult
OgM should be in the routine work-up for any patient with meningitis. When the reason
for presumptive meningitis is unclear, patients should be evaluated by an otolaryngologist
in addition to imaging and ophthalmologic evaluation. Once OM has been diagnosed on
otoscopy, the otolaryngologist should be involved in the management of these patients.
