112 / 264
who died from intracranial complications (1 =
Klebsiella
pneumoniae
, 1 =
P aeruginosa
; 1 = negative culture).
Based on culture sensitivity, 52.6% (n = 10) received
culture-specific therapy. Those with negative cultures
received empirical intravenous ceftazidime and oral fluoro-
quinolone. The mean (SD) duration of antibiotic therapy
was 42.2 (15.3) days. Outpatient therapy was used in 63.2%
(n = 12) to reduce inpatient stay, reducing the mean (SD)
inpatient antibiotic duration to 24.3 (20.1) days.
There was no statistically significant difference in out-
come between those who received culture-specific therapy
and those who received ceftazidime and fluoroquinolone
empirically (70.0% of the directed therapy group had dis-
ease that resolved vs 55.6% in the empirical therapy group;
P
= .650).
Discussion
Malignant otitis externa had associated mortality rates of
50% when it was first characterized in 1968.
1
The develop-
ment of anti-pseudomonal antibiotics has since reduced
mortality significantly. Studies have reviewed factors such
as clinical presentation, laboratory and imaging findings,
and microbiology in an attempt to identify prognostic fac-
tors. However, conclusive prognostic factors have yet to be
identified, and there is still no consensus on the optimal
choice of antibiotics and duration of therapy.
Demographic Factors
The impact of age on prognosis is disputed. Franco-Vidal et
al
5
reviewed 46 patients and found that age did not affect
outcome. On the contrary, Soudry et al
4
studied 57 patients,
and their results suggested a lower life expectancy for those
aged 70 years and older. However, this was also attributed
to additional risks contributed by diabetes and other
atherosclerosis-related comorbidities frequently present in
these patients. Our experience suggests that age is not an
accurate prognostic factor.
The link with diabetes is well established,
1,6
with recent
studies reporting the prevalence of diabetes in MOE to be
between 65% and 95%.
4,5,7
Our results concurred, with
94.7% (n = 18) having diabetes in our study. Diabetes-
related microangiopathy is thought to predispose to MOE as
the impaired local circulation within the EAC diminishes
the ability of immunological cells to respond to invasion by
P aeruginosa.
8,9
Glycemic control is linked to the severity
of microangiopathy in the retina and renal circulations.
Hence, we postulated that poorer glycemic control would
imply more severe microangiopathy within the EAC, and
this could predict severe disease. However, analysis of
HbA1c levels failed to show relation to outcome. This con-
curs with earlier studies that also failed to show any relation
between degree of glucose intolerance and outcome.
6,9
A
possible explanation could be that as most patients are
elderly with longstanding diabetes, microangiopathy within
the EAC is already end stage, therefore making little differ-
ence in the local immune response.
Presentation
Initial symptoms are indistinguishable from simple otitis
externa. It is usually only after multiple failed treatment
attempts that MOE is suspected. The mean diagnostic delay
was significantly long at 6.79 weeks, similar to the literature,
which has been reported to be between 1 and 7 months.
10,11
Similar to other studies,
4,12
our results suggest that a delay in
the commencement of intravenous antibiotics does not
adversely affect outcome. However, this delay leads to pro-
longed suffering, and physicians should always maintain a
high index of suspicion in susceptible patients.
Soudry et al
4
and Franco-Vidal et al
5
found that those
with facial nerve palsy at presentation had a poorer out-
come. This trend was noted in our series, but the number of
patients with facial nerve involvement was small, and the
findings did not reach statistical significance (50.0% poor
outcome in the group with cranial nerve involvement vs
33.3% poor outcome in the group without;
P
= .603).
Hematological Parameters
Leukocytosis, a traditional marker of inflammation, was not
prominent, and its use in monitoring activity is limited. In
contrast, the usefulness of CRP and ESR as markers for
activity is apparent, concurring with previous studies.
3,6
Our
results showed that the absolute ESR and CRP levels could
not be relied upon alone to predict outcome, but the useful-
ness of these markers became evident when serial readings
showed that trends were seen to correlate closely with disease
activity. Patients with disease that resolved after 6 weeks of
intravenous therapy showed a 21.71% reduction in mean
ESR values compared with the group with persistent disease,
in which ESR values remained unchanged (
Figure 1
). A
similar downward trend was also seen in CRP levels. By
trending inflammatory marker levels, clinicians are able to
decide more confidently which patients are suitable for anti-
biotic cessation or outpatient treatment, especially when one
does not have ready access to radionuclide scans.
Imaging Findings
Radionuclide bone and white cell–tagged scans have proven
valuable in MOE. Bone scans are sensitive because the
radionuclide tracer (technetium-99m methylene diphospho-
nate) accumulates at areas of osteoblastic activity. However,
osteoblastic activity persists long after the infective process,
hence limiting the usefulness for charting progress or con-
firming resolution. White cell–tagged scans using gallium
citrate (Ga67) have been more useful in monitoring disease
as the tracer is incorporated directly into granulocytes at
sites of infection. These scans are undoubtedly valuable;
however, they may not be available in all institutions deal-
ing with MOE. In our experience, we have used anatomic
imaging modalities such as CT and MRI to assess our
patients.
Computed tomography is ideal for assessing bony invol-
vement. In theory, the anatomical extent of disease should
be useful for prognosis. Peleg et al
13
reported that severe
Otolaryngology–Head and Neck Surgery 148(6)
91