Keywords Otitis externa - Osteomyelitis - Temporal bone - Pseudomonas aeruginosa - Skull base
Introduction
Osteomyelitis of the temporal bone (OTB), also referred to as otitis externa maligna,
is a rare but serious disease with an increasing incidence over the past decades [1 ]
[2 ]. The diagnosis is primarily based on radiological, clinical and microbiological
findings. A diabetic metabolic state or immunosuppression are considered to be decisive
risk factors [3 ]
[4 ]
[5 ]. While the first symptoms of OTB, such as chronic otalgia and otorrhea, may be rather
unspecific, the short- and long-term consequences can be severely threatening. Some
of the most feared complications are cranial nerve palsies and inflammatory involvement
of intracranial structures [6 ]
[7 ]
[8 ]
[9 ]. In addition to the high risk of possible irreversible morbidity, mortality rates
are reported to reach 20–50% in complicated cases [10 ]
[11 ]
[12 ]. Due to the wide range of unspecific clinical symptoms and the absence of standardized
diagnostic criteria [13 ]
[14 ]
[15 ], it is challenging to make an exact diagnosis in an early stage of the disease.
However, a rapid diagnosis is important for the timely initiation of therapy and thus
the clinical outcome. The identification of the underlying pathogen is essential to
initiate an effective antibiotic treatment [16 ]
[17 ]
[18 ]
[19 ]
[20 ]
[21 ]
[22 ].
In one of the most fundamental studies, which defined the central characteristics
of OTB back in 1987, Pseudomonas aeruginosa was described to be not only the main pathogen but its presence was even one of the
key diagnostic criteria of OTB [23 ]. Henceforth, P. aeruginosa has remained the most frequently detected pathogen in this disease [6 ]
[12 ]
[24 ]
[25 ]
[26 ]. However, a current study reported an increasingly heterogeneous spectrum of microorganisms.
In particular, a progressively dominant role of methicillin-resistant S. aureus (MRSA) has been observed, especially in conjunction with a history of ear surgery
[27 ].
In this study, which covered a period of 10 years, we aimed to evaluate whether changes
in the pathogen spectrum could also be observed in our patient group.
Material and Methods
Patients
All patients diagnosed with OTB at our department from January 2010 to April 2020
were included in this retrospective study (n=39). The diagnostic criteria were: Clinical
findings of otitis externa, lack of response to previous local and systemic therapy
and computed tomography signs of an invasive inflammatory disease of the temporal
bone originating from the external auditory canal. The patient data (age, sex, symptoms,
potential risk factors) and the provided treatment were taken from archived and/or
electronic medical records. A follow-up of the patients’ clinical course and outcome
was conducted until May 2024 (follow-up period, 19.8 ± 15.8 months; range, 0–58 months).
We performed a classification of the outcome in three categories based on the documented
anamnestic data during the follow-up period: (i) improvement of symptoms, (ii) persistence
of symptoms and (iii) death.
This study was approved by the local Research Ethics Committee (file numbers 20–673
and 20–673_1).
Microbiological diagnostics, radiological imaging and laboratory diagnostics
Microbiological diagnostics were performed for all OTB patients (n=39). Samples were
taken from the affected ear pre- or intraoperatively. Routine diagnostic microbiological
and laboratory procedures were performed under strict quality-assured conditions according
to DIN EN ISO 15189 standards (certificate number D-ML-13102–01–00).
Clinical specimens were streaked onto various solid culture media: Columbia blood
agar, chocolate agar, MacConkey agar, Schaedler agar and Sabouraud dextrose agar.
Then the samples were inoculated into thioglycolate bouillon for an enrichment culture
(agar and bouillon from Oxoid, Wesel, Germany). In the next step, the agar plates
were incubated under aerobic culture conditions at 36±1 °C for at least 48h and in
case of Sabouraud dextrose agar for 7 days. Schaedler agar and thioglycolate bouillon
were incubated anaerobically at 36±1 °C for 7 days. Media were inspected daily for
microbial growth. The detected microorganisms were identified by the VITEK MS system
and analyzed for their antimicrobial resistance by using the VITEK 2 system (BioMérieux).
All patients (n=39) underwent routine blood tests (blood count, electrolytes, inflammation
parameters). All patients (n=39) underwent computed tomography (CT) of the temporal
bone as part of the radiological diagnostics. In n=17 cases (43.6%) an additional
magnetic resonance imaging (MRI) of the skull was performed.
Statistics
The correlation of the detection of microbial pathogens with various patient parameters
was evaluated using the Wilcoxon-Mann-Whitney U-test. The Kolmogorov-Smirnov test
with Lilliefors correction was utilized as a pre-test to check the normal distribution
assumption. Moreover, contingency table tests, Chi-square tests or exact Fisher tests
as well as correlation tests (according to Pearson or Spearman) were also used for
data analysis. The influence of individual quantitative and dichotomous patient characteristics
on the occurrence of different events (death, pathogen detection or other events)
was tested by univariate logistic regression.
The data are presented as mean and standard deviation. In principle, statistical tests
were performed bilaterally and with a significance level of alpha=5%. SPSS 22.0 (SPSS
Inc., Chicago, IL, U.S.A.) was used to statistically analyse the data. The graphs
were created with GraphPadPrism 10.2.3. (GraphPad Software, Boston, MA, U.S.A.).
Results
Patient characteristics, radiological findings, treatment and clinical course
The patient group with OTB identified during the study period (n=39) included 32 male
(82.1%) and 7 female (17.9%) patients with an average age of 73.2 (±14.1) years (range,
44–100 years). Three predominant symptoms were reported: otalgia in 29 patients (74.4%),
otorrhea in 24 (61.5%) and hearing loss in 21 patients (53.8%). Facial nerve palsy
occurred in 12 cases (30.8%). There was no statistical correlation between clinical
findings and the detected pathogen spectrum (p>0.05).
In a considerable number of patients (n=33, 84.6%) predisposing factors were identified.
Among the most common were diabetes mellitus (n= 26, 66.7%) and arterial hypertension
(n=23, 59.0%).
CT imaging revealed a washed-out bone texture of the temporal bone in 38 patients
(97.4%), obstructions of the auditory canal, mastoid and middle ear in 31 patients
(79.5%) and lysis and destruction of the temporal bone in 30 patients (76.9%). MRI
morphologically, 14 of the examined cases (82.4%) had a soft tissue affection and
13 cases (76.5%) each had a T1 signal reduction and T2 signal increase ([Table 1 ]). However, there was no statistical correlation between the radiological and microbiological
findings (p>0.05).
Table 1 Radiological findings.
Patients (n)
Ratio (%)
Total
39
100.0
Received Imaging
39
100.0
17
43.6
Localization of the inflammatory process
37
94.9
13
33.3
5
12.8
CT criteria
Obstructions
31
79.5
2
5.1
2
5.1
3
7.7
1
2.6
Bone texture alterations
38
97.4
8
20.5
3
7.7
Lyses/destructions
30
76.9
Complications (abscess)
5
12.8
MRI criteria (as a proportion of the examinations performed)
13
76.5
13
76.5
12
70.6
14
82.4
All 39 patients received intravenous antibiotic therapy. In 36 patients (92.3%) the
OF was surgically treated and in one case (2.6%) a CT-guided biopsy was performed
for histological and microbiological diagnosis. In 12 cases (30.8%) revision surgery
had to be performed and 3 patients (7.7%) were operated three times. There was no
statistical correlation between the revision rate and the detected pathogen spectrum
(p>0.05).
The follow-up period was up to one year for 29 patients (74.4%), up to three years
for 6 patients (15.4%) and up to five years for 4 patients (10.3%).
During the follow-up period after therapy 10 patients (25.6%) reported an improvement
in their symptoms, half of these patients (n=5) experienced a complete remission of
their symptoms. 23 patients (59.0%) complained of persisting symptoms of OTB. 4 patients
(10.3%) deceased. There was no statistical correlation between the presence of cranial
nerve palsies and a protracted course of the disease or the death of patients (p>0.05).
Data such as clinical symptoms, pre-existing illnesses, treatment and clinical course
are summarized in [Table 2 ] and [Table 3 ].
Table 2 Patient characteristics.
Patients (n)
Portion (%)
a Standard deviationb Chronic obstructive pulmonary disease
Total
39
100.0
Sex
32
82.1
7
17.9
Age , average (± SDa )
73.2 (±14.1)
Symptoms and clinical findings
29
74.4
24
61.5
21
53.8
12
30.8
2
5.1
1
2.6
1
2.6
1
2.6
1
2.6
Potential risk factors
26
66.7
2
5.1
24
61.5
23
59.0
9
23.1
5
12.8
4
10.3
4
10.3
3
7.7
3
7.7
2
5.1
2
5.1
1
2.6
1
2.6
1
2.6
1
2.6
1
2.6
1
2.6
1
2.6
1
2.6
Table 3 Treatment and clinical course.
Patients (n)
Ratio (%)
a includes primary and revision surgeryb includes changes in initial antibiotic treatmentc during follow-up period
Total
39
100.0
Treatment
Surgical interventiona
36
92.3
21
53.8
21
53.8
11
28.2
11
28.2
Antibiotic treatmentb
39
100
18
46.2
14
35.9
12
30.8
8
20.5
4
10.3
Clinical coursec
10
25.6
23
59.0
4
10.3
Microbiological and laboratory findings
The microbiological findings could be divided into five pathogen-related groups: gram-negative
rods (n=26, 66.7%), gram-positive cocci (n=16, 41.0%), fungi (n=10, 25.6%), gram-positive
rods (n=6, 15.4%) and anaerobes (n=5, 12.8%). A patient age over 78 years was significantly
correlated with the presence of an infection with gram-negative rods (p<0.05, [Fig. 1 ]). The bacteria most frequently detected from cultures of smears were P. aeruginosa (gram-negative rod) in 21 patients (53.8%) and S. aureus (gram-positive coccus) in 6 patients (15.4%), followed by Enterococcus faecalis (gram-positive coccus) and Corynebacterium spp. (gram-positive rod) in 3 patients (7.7%) each. In addition, fungi were found
in 10 patients (25.6%). Here, the most frequent types were: Candida parapsilosis (n=4, 10.3%) and Candida albicans (n= 3, 7.7%) ([Table 4 ]). In the samples of 3 patients (7.7%) no microbial growth was detected. In all other
cases (n=36) an antibiogram and antimycogram were performed. There was no association
between the detected microbial spectrum including P. aeruginosa and a history of previous ear surgery (n=9; p>0.05, [Fig. 2 ]).
Table 4 Microbiological spectrum.
Patients (n)
Portion (%)
a Gram-negative rodsb Gram-positive coccic Gram-positive rodsd Anaerobes
Bacterial pathogens
21
53.8
6
15.4
3
7.7
3
7.7
3
7.7
2
5.1
2
5.1
2
5.1
1
2.6
1
2.6
1
2.6
1
2.6
1
2.6
1
2.6
1
2.6
1
2.6
1
2.6
1
2.6
1
2.6
Fungal pathogens
4
10.3
3
7.7
1
2.6
1
2.6
1
2.6
Fig. 1 Association between the detection of gram-negative rods and patient age (≤ 78 years,
19 patients; >78 years, 20 patients). *p<0.05.
Fig. 2 Detection of Pseudomonas aeruginosa and previous ear surgery (9 patients with previous
ear surgery; 30 patients without previous ear surgery). ns, not significant.
With regard to P. aeruginosa , the most frequently detected microorganism, all isolates were susceptible to piperacillin/tazobactam,
ceftazidime, cefepime, imipenem and meropenem. Two isolates (9.5%) were found to be
resistant against fosfomycin. There was no significant correlation between the detection
of P. aeruginosa and a fatal outcome (p > 0.05).
The laboratory blood tests showed that leukocytosis was present in only 8 patients
(20.5%) with an average of 10.5 (±2.7) leukocytes/nl (reference value 4–10 leukocytes/nl).
No correlation between the detected pathogens and leukocytosis was detected (p > 0.05).
However, an elevated concentration of the C-reactive protein (CRP) was detected in
35 patients (89.7%) with an average value of 5.1 (±3.6) mg/dl (reference value 0.0–0.5
mg/dl). The detection of P. aeruginosa was significantly associated with an increased CRP value (p<0.05, [Fig. 3 ]).
Fig. 3 Association between the CRP value (reference value 0.0–0.5 mg/dl) and Pseudomonas
aeruginosa detection (Pseudomonas-positive, 21 patients; Pseudomonas-negative, 18
patients). *p<0.05.
Discussion
OTB is a chronic progressive disease that, if left untreated, can spread through the
petrous bone and base of the skull, thus causing life-threatening complications [10 ]
[12 ]
[28 ]
[29 ]. Effective and targeted antimicrobial therapy is essential for the management of
this disease. Therefore successful identification of the causal pathogen is key [16 ]
[17 ]
[21 ]
[22 ]. In recent decades P. aeruginosa has been known to be the main pathogen of OTB [12 ]
[23 ]
[30 ]
[31 ]. Chen et al. [27 ] compared the spectrum of pathogens in patients with OTB in Taiwan during the periods
1990–2001 and 2002–2011. Although it should be noted that the patients in these two
groups were heterogeneous in terms of a number of characteristics (e.g. proportion
of diabetic patients), data revealed a change in the spectrum of pathogens with an
increasing prevalence of MRSA, especially in patients with previous ear-related surgeries
[27 ].
In contrast to these findings, our results revealed a persistent dominance of P. aeruginosa. We detected P. aeruginosa in more than half of the patients (n=21, 53.8%), whereas S. aureus was found in only 6 patients (15.4%). Notably, MRSA was not detected among our study
population. Moreover, there was no correlation between a history of previous ear-related
surgery and the observed microbial spectrum (p>0.05). The discrepancy in the frequency
of detected MRSA cases might be biased by the different prevalences of MRSA in Germany
and Taiwan: recent data show a MRSA prevalence in Taiwan between 17 and 27% [32 ], whereas in Germany it is reported to be significantly lower at approximately 8.5%
[33 ].
The increasing development of antimicrobial resistance represents a threat to the
effectiveness of an early treatment of the assumed pathogen. Nearly 10% of the P. aeruginosa isolates detected in our patient cohort turned out to be resistant against fosfomycin.
P. aeruginosa is known to have intrinsic resistance against a variety of antibiotics and an outstanding
capacity to acquire resistance mechanisms during treatment [34 ]
[35 ]. These findings suggest a progressive development of resistance and thus an increasing
complexity of OTB therapy.
The diagnostic criteria for OTB are inconsistent in the literature, which complicates
timely diagnosis. Clinical findings of otitis externa, which showed a lack of response
to previous local and systemic therapy, as well as radiological signs of invasive
inflammation of the temporal bone starting from the auditory canal are listed as core
findings of the disease [14 ]
[15 ]. In our study, a CT scan of the temporal bone was performed in all cases at the
time of diagnosis, which showed the extent of the bony inflammatory process. In some
cases, an additional MRI of the skull was performed to identify soft tissue affections.
The routine use of MRI in the diagnosis of OTB could facilitate the diagnosis in the
future and improve the assessment of the inflammatory spread of the disease [6 ].
In our cohort a high proportion of patients (92.3%) underwent surgical treatment of
the inflammatory process. There is a lack of consensus in the literature regarding
the necessity of a surgical approach in the treatment of OTB. While many authors support
debridement [14 ]
[15 ], other authors limit the surgical procedure only for the purpose of a biopsy for
histological and microbiological examination of the tissue removed [13 ].
Our results confirm previous findings that OTB is a disease of elderly patients [17 ]
[21 ]
[27 ]
[36 ]. Interestingly, we observed a significant correlation between patient age and the
presence of gram-negative rods – mainly P. aeruginosa . Moreover, our data show that an elevation in the CRP value is significantly correlated
with the presence of a P. aeruginosa infection. Thus, in elderly OTB patients – the wide majority of the affected – with
high CRP levels a specific pathogen spectrum can be anticipated and consequently a
calculated antibiotic therapy covering P. aeruginosa can be initiated even before microbiological results are available. To our knowledge,
the significant correlation between an elevated CRP value and detection of P. aeruginosa has not yet been described in the literature. In contrast to CRP levels, leukocyte
count does not appear to be a predictive factor for the underlying pathogen.
In our study, no correlation between the detected spectrum of microorganisms and mortality
was found. Nonetheless, our follow-up data underline that OTB is a serious disease
with high morbidity rates and partial symptom persistence (59% in our patient cohort)
even after extensive therapy confirming data from previous studies [10 ]
[16 ]
[21 ].
The limitations of this study relate to its retrospective and monocentric character
and the rather small study population which is due to the rarity of the disease. Moreover,
the lethal outcomes in our patient series could not be causally attributed to OTB.
Therefore, no disease-related mortality rate can be derived from our cohort. Furthermore,
the long-term effects of OTB on quality of life and secondary complications were not
systematically recorded. These parameters could be included as primary endpoints of
future studies.
Conclusion
Our results do not confirm a shift in the microbial spectrum of OTB toward MRSA. P. aeruginosa appears to remain the major pathogen of this devastating disease. Thus, we propose
that especially in a typical OTB patient cohort of elderly patients an empiric, calculated
antibiotic therapy covering P. aeruginosa should be initiated. Owing to the increasing prevalence of antibiotic resistances
the choice of an adequate and effective therapy can become increasingly difficult.
Effective monitoring and surveillance systems will be necessary to evaluate the future
development of resistance in order to establish the most effective empiric antibiotic
treatment.
