107 / 264
Two of 3 MRSA-infected patients had resolution of their
infections, averaging 8.5
6
0.7 weeks of definitive antibio-
tic therapy and 12.6
6
0.9 total weeks of antibiotic therapy.
One patient infected with MRSA and
Acinetobacter
died
from an infected central catheter while undergoing treatment
for MOE. Although the treatment duration was longer for
MRSA-infected than for
Pseudomonas
-infected patients,
these differences in duration were not statistically signifi-
cant (
P
= .18 for total and
P
= .10 for definitive therapy).
None of the MRSA-infected patients underwent surgical
intervention beyond debridement of the ear canal. The
MRSA-infected patients were treated with intravenous
vancomycin.
Overall, the 7 patients with documented resolution of
non-
Pseudomonas
infections were treated with antibiotics
for an average of 9.3
6
4.5 weeks of definitive antibiotic
therapy and 10.4
6
4.6 total weeks of antibiotic therapy.
These treatment durations were longer (3.2 and 2.5 weeks,
respectively) than those for the
Pseudomonas
-infected
patients but did not reach significance (
P
= .09 and
P
=
.25). Six of these 7 patients (86%) received intravenous
therapy with nonquinolone antibiotics.
Discussion
Classically, MOE has been thought to be due exclusively to
Pseudomonas
infection. In fact, Cohen and Friedman
10
sug-
gested the presence of
Pseudomonas
on cultures as an obli-
gatory diagnostic criterion for this disease, though noting
that this required further investigation. In 1988, Rubin and
Yu
3
performed a literature review of 260 cases of MOE and
found that virtually all cases (99.2%) were caused by
Pseudomonas
. With increasing frequency, however, non-
pseudomonal cases of MOE are being reported. Fewer than
half (45%) of the patients in our study had cultures that
grew
Pseudomonas
. Similarly, in 2010, Chen et al
7
and
Jacobsen and Antonelli
13
reported relatively low proportions
of pseudomonal MOE, with only 26.9% and 34% of their
respective patients having
Pseudomonas
isolated in cultures.
Commensurate with the decline in
Pseudomonas
isolates
has been a rise of other organisms leading to MOE.
5-7,13
The second most common isolate in our series was
S
aureus
, with 3 isolates being MRSA (15%). To our knowl-
edge, there are few reports documenting MRSA as a causa-
tive organism in cases of MOE.
7,8
With the evolving microbiology of MOE, it is essential
that treatment be tailored to the causative organism(s).
Since its introduction in the late 1980s, oral ciprofloxacin
has commonly been used as a first-line empiric treatment
for MOE.
14
This allowed patients with early-stage disease
to be treated empirically in an outpatient setting. Now, how-
ever, with the increasing frequency of nonpseudomonal
MOE, ciprofloxacin may not always be an effective treat-
ment, as it has poor gram-positive coverage and is ineffec-
tive against MRSA.
Additionally, the incidence of ciprofloxacin-resistant
Pseudomonas
has been rising. Although there was 1 instance
of levofloxacin resistance and not any documented
ciprofloxacin resistance in our
Pseudomonas
isolates, other
investigators have reported an increasing incidence
ciprofloxacin-resistant
Pseudomonas
as a cause of MOE.
14,15
In 2002, Berenholz et al
15
were the first to report MOE
caused by ciprofloxacin-resistant
Pseudomonas
, with 33% of
their
Pseudomonas
isolates being resistant to ciprofloxacin.
The clinical features of MOE caused by
Pseudomonas
and
MRSA were similar in many respects (average age of onset,
signs and symptoms, etc), but some important differences
were noted. Diabetes is a commonly noted comorbidity in
patients with MOE and, along with immunocompromised
state, is thought to be a risk factor for development of the
disease.
3
In our series, all of the
Pseudomonas
-infected
patients had diabetes, whereas only 1 of the 3 MRSA-
infected patients did (
P
= .046). Additionally, only 55% of
the patients with non-
Pseudomonas
infections had diabetes,
which was also significantly less than in the
Pseudomonas
-
infected patients (
P
= .04). These findings illustrate the point
that a diagnosis of MOE must be considered in any patient
with refractory otitis externa, even in those without diabetes.
Furthermore, it suggests that atypical organisms, such as
MRSA, should be suspected in patients without diabetes who
present with MOE.
There were important limitations of this study worth men-
tioning. First, the study was retrospective in nature, and data
were limited to the available records. Additionally, the
patient population was heterogeneous in terms of prior ther-
apy and the manner in which cultures were obtained. It is
possible that prior therapy would eradicate organisms that
would have otherwise have been detected on culture. This
means that potentially pathogenic organisms (in isolation or
as a part of a polymicrobial infection) may have been eradi-
cated or rendered undetectable by culture prior to presenta-
tion to our offices in some instances. Furthermore, ear-swab
culture may not always be reflective of the pathogenic organ-
ism infecting the temporal bone in MOE. Although the spe-
cific role of culture method has not been investigated in
MOE, discordance between swab and bone culture has been
shown for diabetic foot osteomyelitis.
16
For this reason, we
recommend tissue biopsy in cases in which there is poor
or no response to ear-swab culture–directed therapy.
Additionally, the choice of antibiotic was based on the
preference of the treating neurotologist in conjunction
with an infectious disease specialist. Many patients were
treated with intravenous antibiotics, while sensitivities sug-
gested that an oral antibiotic could have been used. Because
MRSA is a rare organism to cause MOE, there are few data
guiding treatment of this offending organism. It is worth
noting that 1 of our MRSA-infected patients was being treated
with oral trimethoprim-sulfamethoxazole at the time of presen-
tation and culture. Although sensitivities indicated that the
organism was susceptible to this antibiotic, the patient’s infec-
tion was not responsive to this treatment. The infection was
resolved with intravenous vancomycin treatment.
In general, treatment durations were planned for 6 weeks
and extended as needed on the basis of clinical or radio-
graphic evidence of persistent or progressive disease. It is
Hobson et al
86