G Model

ANL-2359; No. of Pages 6

Auris Nasus Larynx xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

Auris Nasus Larynx

journal homepage: www.elsevier.com/locate/anl

High-dose corticosteroids improve the prognosis of Bell’s palsy

compared with low-dose corticosteroids: A propensity score
analysis
Takashi Fujiwara *, Yasuharu Haku, Takuya Miyazaki, Atsuhiro Yoshida, Shin-ich
Sato, Hisanobu Tamaki
Department of Otolaryngology Head and Neck Surgery, Kurashiki Central Hospital, Miwa 1-1-1, Kurashiki, Okayama 710-8602, Japan

A R T I C L E I N F O A B S T R A C T

Article history: Objective: The aim of this study was to evaluate the effectiveness of high-dose corticosteroid
Received 10 August 2017 (120 mg prednisolone equivalent daily) in Bell’s palsy compared with low-dose corticosteroid
Accepted 13 September 2017 (60 mg PSL equivalent).
Available online xxx
Methods: A single-center retrospective observational study was performed. We included adult
Bell’s palsy patients who were treated within 7 days after disease onset. We compared high- and
Keywords:
low-dose corticosteroid for the non-recovery rate at 6 months after disease onset using inverse
Bell’s palsy
Facial palsy
probability-weighted propensity score analysis (IPW-PS).
Prognosis Results: A total of 368 Bell’s palsy patients (281 in the high-dose and 87 in the low-dose group)
Steroids were included. The non-recovery rate without IPW-PS was 13.8% in the low-dose and 8.2% in the
Propensity score high-dose group. After IPW-PS adjustment, the non-recovery rate was 13.1% in the low-dose and
7.8% in the high-dose group (difference = 5.28%, 95% confidence interval [CI] 12.7% to
2.1%, p = 0.040). High-dose corticosteroid decreased the non-recovery rate in severe Bell’s palsy
patients with a Yanagihara score of 0–10 (difference = 16.1%, 95% CI 38.5% to 6.2%,
p = 0.012), but did not decrease in moderate Bell’s palsy patients with a Yanagihara score of 12–18
(difference = 2.0%, 95% CI 11.0% to 7.0%, p = 0.591). Subgroup analysis revealed that the
efficacy of high-dose corticosteroids was higher when patients were treated within 3 days after
disease onset, but not when patients were treated at 4 days or later after disease onset.
Conclusions: Physicians would be better to treat severe Bell’s palsy patients with high-dose
corticosteroids when the patients are treated within 3 days after disease onset.
© 2017 Elsevier B.V.. All rights reserved.

1. Introduction palsy is favorable, and approximately 70% of patients achieve

Bell’s palsy, defined as an acute facial nerve paralysis of
unknown origin, is the most common cause of peripheral facial
palsy. Epidemiologic studies have reported an annual incidence
of 20–30 cases per 100,000 persons [1]. The prognosis of Bell’s
* Corresponding author.
E-mail address: t.fujiwarabi@gmail.com (T. Fujiwara).
complete recovery of facial movement without treatment [2]. In
addition, approximately 90% of Bell’s palsy patients achieve
complete recovery of facial movement using a combination of
corticosteroids and antiviral agent therapy [3,4].
The etiology of Bell’s palsy remains unclear, but reactivation
of latent herpes simplex type 1 infection in the geniculate
ganglion is considered a major cause [5]. The reactivation of
herpes simplex type 1 introduces edematous changes in the

http://dx.doi.org/10.1016/j.anl.2017.09.008
0385-8146/© 2017 Elsevier B.V.. All rights reserved.

Please cite this article in press as: Fujiwara T, et al. High-dose corticosteroids improve the prognosis of Bell’s palsy compared with low-dose
corticosteroids: A propensity score analysis. Auris Nasus Larynx (2017), http://dx.doi.org/10.1016/j.anl.2017.09.008
G Model
ANL-2359; No. of Pages 6
2 T. Fujiwara et al. / Auris Nasus Larynx xxx (2017) xxx–xxx
facial nerve [6], causing compression of the nerve in the facial
canal. Corticosteroids are anti-inflammatory agents, reducing
edema and inflammation of the facial nerve in the acute
presentation of Bell’s palsy. A recent Cochrane systematic
review reported that systemic corticosteroid improves the
prognosis of Bell’s palsy [3]. However, the initial dosage of
corticosteroid included in this systematic review varied from
50 to 1000 mg (prednisolone [PSL] equivalent) daily [3].
Several clinical practice guidelines recommend the use of
steroids during the acute phase of Bell’s palsy [7–9], but the
optimal dosage of corticosteroids remains unclear. Therefore,
various corticosteroid dosing regimens were recommended in
the clinical practice guidelines [7–9]. An initial PSL dose of
50 or 60 mg daily is commonly used, and the American
Academy of Otolaryngology-Head and Neck Surgery proposed
two regimens: PSL at 50 mg daily for 10 days and PSL at 60 mg
daily for 5 days with a 5-day taper [8]. The Japan Society of
Facial Nerve Research also proposed PSL at 60 mg daily with a
10-day taper for moderate-to-severe Bell’s palsy patients
[7]. However, the Japan Society of Facial Nerve Research
proposed treatment options of PSL at 120–200 mg daily with a
10-day taper for severe Bell’s palsy patients.
PSL at 120–200 mg daily with a 10-day taper is an optional
treatment for Bell’s palsy in Japan, but a randomized controlled
trial or well-designed comparative study is not yet available.
Additionally, the adverse effects of corticosteroids increase in a
dose-dependent manner and it is important to identify
subgroups of patients who may benefit from high-dose
corticosteroids. In this study, we evaluated the effectiveness
of high-dose corticosteroid (120 mg PSL equivalent) compared
with low-dose corticosteroid (60 mg PSL equivalent), and
identified patients who may benefit from high-dose corticoste-
roids based on a propensity score analysis.
2. Methods
2.1. Study design and setting
We conducted a retrospective observational study at
Kurashiki Central Hospital from October 2009 to September
2016. The institutional review board of Kurashiki Central
Hospital approved this study. We included Bell’s palsy patients
aged 18 years or older who first visited the hospital during the
study period. Patients were identified using the following
International Statistical Classification of Diseases and Related
Health Problems 10 (ICD-10) codes: B022, G510, G519, S045,
and T812.
We included patients who were diagnosed with Bell’s
palsy and prescribed corticosteroid within 1 week after
disease onset. Patients whose facial palsy was caused by Hunt
syndrome or other causes (e.g., trauma, iatrogenic, or parotid
tumors) were excluded, as were patients who were prescribed
a step-up dose of corticosteroid (e.g., initially prescribed
1.0 mg/kg/day prednisolone and later prescribed 2.0 mg/kg/
day prednisolone after a few days because of worsening
facial palsy). Patients who visited our hospital for a second
opinion and were followed by another hospital; who died due
to other diseases within 6 months after facial palsy; who were
Please cite this article in press as: Fujiwara T, et al. High-dose corticosteroids improve the prognosis of Bell’s palsy compared with low-dose
corticosteroids: A propensity score analysis. Auris Nasus Larynx (2017), http://dx.doi.org/10.1016/j.anl.2017.09.008
on bad terms with Kurashiki Central Hospital because of
violence or failure to pay medical bills; who had slight facial
palsy (Grade 1–2 on House–Brackmann facial grading
system, Yanagihara facial nerve grading system score of
32–40 [10,11]); who had sequelae, such as facial spasm,
synkinesis, or contracture before disease onset; or who could
not be followed for more than 6 months or until complete
recovery were excluded. The aim of our study was to
compare high-dose corticosteroid (120 mg PSL equivalent)
with low-dose corticosteroid (60 mg PSL equivalent); thus,
we excluded Bell’s palsy patients who were treated with an
initial PSL dose of 30 mg daily or less.
2.2. Treatment of Bell’s palsy in our hospital
During the study period, Bell’s palsy was treated using the
following strategy: 30 mg (0.5 mg/kg) PSL equivalent daily for
3 days with a 6-day taper, 60 mg (1.0 mg/kg) PSL equivalent
for 3 days with a 6-day taper, or 120 mg (2.0 mg/kg) PSL
equivalent for 3 days with a 6-day taper. Corticosteroids were
administered orally at an initial dose of 0.5 or 1.0 mg/kg, and
intravenously at an initial dose of 2.0 mg/kg. In most cases, one
of these three steroid regimens was used, but physicians
modified the dose of steroids for individual patients. Whether
the Bell’s palsy patients were prescribed antiviral agents was
not standardized for protocol at our hospital. The physicians
advised patients to massage their faces, but did not prescribe
rehabilitation by a speech-language therapist.
2.3. Outcome measures
The primary outcome was non-recovery at 6 months after
onset. Recovery was defined as an improvement in the
Yanagihara facial nerve grading system score to 36 or more
without sequelae, according to the facial paralysis guidelines of
the Japan Society of Facial Nerve Research.
Data on patient age, gender, date of Bell’s palsy onset, side
of palsy, symptoms of zoster sine herpete (ZSH) (auricular
redness, intense pain, or taste disturbance), concurrent medical
diseases (diabetes mellitus or hypertension), Yanagihara facial
nerve grading system score (at first hospital visit and the worst
score), and time from the onset to the start of treatment were
collected. The initial dose and the use of antiviral drugs were
also determined. Different corticosteroid agents were used and
the dosage of corticosteroid was converted to the PSL
equivalent based on anti-inflammatory potency (e.g., prednis-
olone:betamethasone = 4:25). The dose of PSL was divided
into low and high doses. When the initial dose of PSL
equivalent was less than 100 mg daily, it was defined as a low
dose; initial doses of 100 mg PSL equivalent daily or more
were defined as high doses. For subgroup analysis, Yanagihara
facial nerve grading system score were converted to House–
Brackmann facial nerve grading system, as reported previous-
ly [11].
In our hospital, Bell’s palsy patients are treated in an
outpatient setting. During the acute phase of Bell’s palsy,
patients on low-dose corticosteroids visit the hospital once a
week, while those on high-dose corticosteroids visit every day.
G Model
ANL-2359; No. of Pages 6

T. Fujiwara et al. / Auris Nasus Larynx xxx (2017) xxx–xxx 3

To avoid bias due to this difference in visit frequency, we 3.2. Comparison of non-recovery rate between low- and high-
assessed adverse effects in patients who were treated in an dose corticosteroid
inpatient setting due to the presence of diabetes mellitus. New
prescriptions to counter adverse effects were noted (proton The non-recovery rates of patients receiving low- and high-
pump inhibitors [PPIs], histamine H2-receptor blockers dose corticosteroid are shown in Table 2. Prior to IPW-PS
[H2Bs], sleeping pills, laxatives or stool softeners, or adjustment, the non-recovery rate was higher in the low-dose
medications for hiccups). group, but the non-recovery rate was higher in the high-dose
group in Bell’s palsy patients with a Yanagihara score of 12–18.
Table 2 also shows the IPW-adjusted non-recovery rates with
2.4. Statistical analysis low- and high-dose corticosteroid, and the non-recovery rate
was higher in the low-dose group in Bell’s palsy patients with a
Data are presented as frequencies and percentages for
Yanagihara score of 12–18 after IPW-PS adjustment. High-
categorical variables and continuous variables are expressed as
dose corticosteroid significantly decreased the non-recovery
mean and standard deviation. The non-recovery rate was
rate, and the non-recovery rate was 13.1% in the high-dose
compared using the unpaired t-test. In the clinical setting,
group and 7.8% in the low-dose group. A significant difference
patients with severe facial palsy tend to receive high-dose
was observed in Bell’s palsy patients with a Yanagihara score of
corticosteroid. Therefore, baseline characteristics should be
0–10, but not in those with a Yanagihara score of 12–18.
accounted for when estimating the effectiveness of high-dose
Subgroup analysis was performed, as shown in
corticosteroid. We calculated the average treatment effects of
Fig. 2. Overall, the non-recovery rate was lower in the high-
high-dose corticosteroid using inverse probability-weighted
dose group, although no significant difference was observed in
propensity score analysis [12,13]. Based on inverse probabili-
patients because of the small sample size. In subgroup analysis
ty-weighted propensity score analysis, we performed logistic
of age, diabetes mellitus, and symptoms of ZSH, high-dose
regression to estimate the probability of being treated with
corticosteroids tended to decrease non-recovery rate. However,
high-dose corticosteroid according to baseline characteristics.
the non-recovery rate was similar between high- and low-dose
We included the following factors in the logistic regression:
group in patients with an initial Yanagihara score of 12–18, with
age, sex, time from disease onset to start of treatment, side of
start of treatment at 4 day or later after disease onset, and with
palsy, initial Yanagihara facial grading system score, diabetes
hypertension.
mellitus, hypertension, and use of antiviral medication. We
Table 3 shows the numbers of patients who had new
then calculated the average treatment effects of high-dose
prescriptions to counter adverse effects (subdivided into high-
prednisolone using inverse probability-weighted propensity
and low-dose corticosteroid groups). Patients on high-dose
score (IPW-PS) analysis. Number of patients who had new
corticosteroids tended to use sleeping pills, laxatives or stool
prescriptions for adverse effects was compared using odds
softeners, or medications for hiccups. PPIs or H2Bs were
ratio. Data were analyzed using Stata for Mac software (ver.
prescribed, before corticosteroid administration, to 52 (100%)
14.0; Stata Corp., College Station, TX, USA).
patients on high-dose corticosteroids and 16 (84.2%) patients
on low-dose corticosteroids.

3. Results 4. Discussion
3.1. Patient characteristics We evaluated the correlation between corticosteroid dose
and prognosis in Bell’s palsy. IPW-PS analysis revealed that a
During the study period, 1226 patients were identified, of
high-dose of corticosteroid improved prognosis in Bell’s palsy.
whom 368 met the inclusion and exclusion criteria (Fig. 1). Of
The effectiveness of high-dose corticosteroid was greater in
these 368 patients, 87 were treated with low-dose corticosteroid
patients with a Yanagihara score of 0–10. High-dose
and 281 were treated with high-dose corticosteroid. The
corticosteroid tended to decrease the non-recovery rate in
baseline characteristics are shown in Table 1. Among the
patients with a Yanagihara score of 12–18, but no significant
included 368 patients, the overall non-recovery rate was 9.5%. difference was observed based on IPW-PS analysis. Thus,
The non-recovery rates of patients receiving low- and high-dose
60 mg PSL equivalent daily would be sufficient for Bell’s palsy
corticosteroid were 13.8% and 8.2%, respectively.
with a Yanagihara score of 12–18, and the treatment strategy in
Facial nerve decompression surgery is an optional treatment
the clinical practice guidelines of the Japan Society of Facial
for Bell’s palsy patients with a Yanagihara score of 0–8 and
Nerve Research is adequate.
electroneurography value of 10% or less [14]. Of the
Corticosteroids are clearly effective for the treatment of
368 patients included in this study, 31 met these criteria but
Bell’s palsy, and most clinical practice guidelines recommend
none had undergone facial nerve decompression surgery.
the use of corticosteroids during the acute phase of Bell’s palsy
However, during the study period, two of the patients [7–9]. In 1972, Adour et al. proposed 1 mg/kg per day of
underwent such surgery: one visited our hospital at 3 weeks
corticosteroid, tapered from day 5 to 10, as a treatment regimen
after disease onset and the other was lost to follow-up within
for Bell’s palsy. A recovery rate of 87.6% [15] was reported for
6 months after disease onset. These patients were excluded from
this regimen, and an initial prednisolone dosage of 50–60 mg
the analysis.
daily is now commonly used in Bell’s palsy patients. In 1979,

Please cite this article in press as: Fujiwara T, et al. High-dose corticosteroids improve the prognosis of Bell’s palsy compared with low-dose
corticosteroids: A propensity score analysis. Auris Nasus Larynx (2017), http://dx.doi.org/10.1016/j.anl.2017.09.008
G Model
ANL-2359; No. of Pages 6

4 T. Fujiwara et al. / Auris Nasus Larynx xxx (2017) xxx–xxx

Fig. 1. Flow diagram of patients’ selection. Retrospective chart review of the 1226 potentially eligible patients was conducted, and 368 patients were included in
analysis. The reasons for exclusion were shown in this flow diagram.

Stennert proposed high-dose cortisone (initial dose equivalent
to 200–250 mg PSL daily) in combination with dextran plus
pentoxifylline regimens. A total of 73 patients, including 31
(56%) complete paralysis patients, underwent this regimen, and
70 (96%) patients showed good recovery [16]. The prevalence
of adverse effects of prednisolone depends of the dosage, and
the regimen proposed by Stennert is not commonly used. At this
time, 120–200 mg of prednisolone is used to treat only severe
Bell’s palsy patients in Japan. Some studies compared regimens
according to the corticosteroid dose [17,18], but these studies
were non-randomized controlled trials and the baseline
characteristics were not adjusted. Therefore, the studies could
not avoid patient selection bias, and the effectiveness of high-
dose corticosteroid was unclear. In this study, we used
propensity score analysis to adjust for baseline characteristics,
and a causal relationship between corticosteroid dose and Bell’s
palsy prognosis was observed.
Subgroup analysis revealed that the efficacy of high-dose
corticosteroid could be observed regardless of age, diabetes
mellitus, and symptoms of ZSH. However, the efficacy was not

Table 1
Baseline characteristics.

Low-dose corticosteroid High-dose corticosteroid

(n = 87) (n = 281)
Characteristics
Age (year) 55.8  18.5 53.0  17.0
Gender (male/female) 42/45 140/141
Facial palsy
Side of palsy (right/left) 48/39 142/139
Yanagihara score at first visit 15.3  7.2 13.7  6.3
Yanagihara score at first visit
0–10 26 (29.9%) 99 (35.2%)
12–18 31 (35.6%) 120 (42.7%)
20– 30 (34.5%) 62 (22.1%)
Yanagihara score at worst
0–10 31 (35.6%) 127 (45.2%)
12–18 30 (34.5%) 105 (37.4%)
20– 26 (29.9%) 49 (17.4%)
Concurrent medical disease
Hypertension 21 (24.1%) 66 (23.5%)
Diabetes mellitus 19 (21.8%) 52 (18.5%)
Treatment
Time from onset to treatment (day) 3.01  1.48 3.61  1.61
Initial dose (PSL equivalent) (mg) 57.7  10.6 126.5  24.0
Antivirals prescription 53 (60.9%) 179 (63.7%)
Non-recovery at 6 month 12 (13.8%) 23 (8.2%)

Please cite this article in press as: Fujiwara T, et al. High-dose corticosteroids improve the prognosis of Bell’s palsy compared with low-dose
corticosteroids: A propensity score analysis. Auris Nasus Larynx (2017), http://dx.doi.org/10.1016/j.anl.2017.09.008
G Model
ANL-2359; No. of Pages 6

T. Fujiwara et al. / Auris Nasus Larynx xxx (2017) xxx–xxx 5

Please set column gutter less than 3cm.
Table 2
Non-recovery rate of patients receiving low- and high-dose corticosteroid.

Non-recovery rate Mean difference (95% CI) P value

Low dose High dose
Analysis without adjustment
All 12/87 13.8% 23/281 8.2% 5.6% ( 12.7 to 1.5%) 0.120
Yanagihara score 0–10a 8/26 30.8% 16/99 16.2% 14.6% ( 37.7 to 2.5%) 0.094
Yanagihara score 12–18a 3/31 9.7% 17/120 14.2% 3.8% ( 6.1 to 13.8%) 0.446
Yanagihara score 20–a 1/30 3.3% 0/62 0.0% 3.3% ( 7.9 to 1.2%) 0.152
House–Brackmann grade III–IVb 2/46 4.3% 2/112 1.8% 2.6% ( 8.0 to 2.9%) 0.355
House–Brackmann grade V–VIb 10/41 24.4% 21/148 14.2% 12.0% ( 24.1 to 0.2%) 0.053
Analysis after inverse probability-weighted propensity score adjustment
All 13.1% 7.8% 5.28% ( 12.7 to 2.1%) 0.040
Yanagihara score 0–10a 32.1% 16.0% 16.1% ( 38.5 to 6.2%) 0.012
Yanagihara score 12–18a 7.7% 5.7% 2.0% ( 11.0 to 7.0%) 0.591
House–Brackmann grade III–IVb 3.8% 1.8% 2.0% ( 7.4 to 3.5%) 0.254
House–Brackmann grade V–VIb 20.5% 12.1% 8.3% ( 21.0 to 4.3%) 0.053
a
Initial score of Yanagihara facial nerve grading system.
b
Initial grade of House–Brackmann facial nerve grading system.

observed in patients who started treatment at 4 days after
disease onset and in patients with hypertension. The efficacy of
corticosteroids could be increased when Bell’s palsy patients
are treated during the acute phase. Although the window period
of corticosteroid in Bell’s palsy remains unclear, a subgroup
analysis of one randomized controlled trial reported benefits of
corticosteroid use only if administered within 48 h of onset
[19]. The results of subgroup analysis in our study suggested
that we would be better to avoid 120 mg PSL equivalent daily
for patients who started treatment at 4 days or later after disease
onset. In subgroup analysis, the efficacy of high-dose
prednisolone was not observed in patients with hypertension.
Both hypertension and diabetes mellitus are believed to affect
the prognosis of Bell’s palsy [20], but the dose response of
prednisolone was observed only in patients with diabetes
mellitus. It remains unclear whether the results were statisti-
cally significant or based on a specific mechanism, and further
validation studies are required.
This study has some limitations. First, 39 patients were
excluded because they were lost to follow-up. This is likely to
have occurred because the natural history of Bell’s palsy is
favorable, and treatment is limited to the acute phase. Less
severe facial palsy patients tend to be lost to follow-up [4]. In
this study, 9.6% (39/407) of patients were lost to follow-up;
inclusion of those patients in the analyses would have decreased
the overall non-recovery rate. Second, we did not collect data on
body mass index (BMI), hematologic parameters (e.g.,
neutrophil-to-lymphocyte ratio), or stapedial muscle reflex,
and some studies have suggested associations between these
factors and the prognosis of Bell’s palsy [21–23]. This
information was not collected because we did not measure
BMI, perform blood tests, or measure stapedial muscle reflex in
all patients in our hospital. Third, the mean initial dose of
prednisolone in the high-dose group was 120 mg PSL
equivalent daily. We did not evaluate the initial PSL dose of
200 mg daily for the prognosis of Bell’s palsy. Fourth, we could

Fig. 2. Non-recovery rates of the high- and low-dose corticosteroid subgroups. The non-recovery rate was calculated using inverse probability-weighted propensity
score analysis. To calculate the propensity scores, age, sex, time from disease onset to the start of treatment, side of palsy, initial Yanagihara facial grading system
score (Y-score), and status regarding diabetes mellitus (DM), hypertension (HTN), and use of antiviral medications were used. Subgroup analysis of following factors
were conducted; initial Y-score, age (<65 versus 65 years), time from disease onset to to the start of treatment (72 versus >72 h), status of DM, status of HTN, and
symptom of zoster sine herpete (ZSH). In subgroup analyses, we calculated the propensity scores as described above, except for the factor under analysis (e.g., in the
subgroup analysis of age, we calculated propensity scores for all factors except age).

Please cite this article in press as: Fujiwara T, et al. High-dose corticosteroids improve the prognosis of Bell’s palsy compared with low-dose
corticosteroids: A propensity score analysis. Auris Nasus Larynx (2017), http://dx.doi.org/10.1016/j.anl.2017.09.008
G Model
ANL-2359; No. of Pages 6

6 T. Fujiwara et al. / Auris Nasus Larynx xxx (2017) xxx–xxx

Table 3 [4] Fujiwara T, Hato N, Gyo K, Yanagihara N. Prognostic factors of Bell’s

New prescriptions for adverse effecs of patients receiving low- and high-dose palsy: prospective patient collected observational study. Eur Arch
corticosteroid. Otorhinolaryngol 2014;271:1891–5.
[5] Murakami S, Mizobuchi M, Nakashiro Y, Doi T, Hato N, Yanagihara N.
Low- Bell palsy and herpes simplex virus: identification of viral DNA in
dose endoneurial fluid and muscle. Ann Intern Med 1996;124:27–30.
corticosteroid High- [6] Koike Y, Tojima H, Maeyama H, Aoyagi M. Contrast-enhanced MRI of
(n = 19) dose the facial nerve in patients with Bell’s palsy. Eur Arch Otorhinolar-
corticosteroid Odds confidence P value yngol 1994;S346–8.
(n = 52) ratio interval) [7] Baugh RF, Basura GJ, Ishii LE, Schwartz SR, Drumheller CM,
(95% Burkholder R, et al. Clinical practice guideline: Bell’s palsy. Otolar-
yngol Head Neck Surg 2013;149:S1–27.
Sleeping pill 6 (31.6%) 21 (40.4%) 1.468 (0.48–4.48) 0.58
[8] Murakami S. Treatment of peripheral facial paralysis in acute phase.
Laxatives/stool softeners 1 (5.3%) 5 (9.6%) 1.915 (0.21–17.54) 0.97
In: Japan Society of Facial Nerve Research, editor. Guideline for facial
Hiccups medication 2 (10.5%) 15 (28.8%) 3.446 (0.71–16.79) 0.13
paralysis—Bell’s palsy and Hunt syndrome. Tokyo: Kanehara shup-
pan; 2011. p. 55–9.
[9] de Almeida JR, Guyatt GH, Sud S, Dorion J, Hill MD, Kolber MR,
not select non-recovery at 12 months as an endpoint instead of et al. Management of Bell palsy: clinical practice guideline. CMAJ
2014;186:917–22.
non-recovery at 6 months because the number of patients lost to [10] Hato N, Fujiwara T, Gyo K, Yanagihara N. Yanagihara facial nerve
follow-up would increase between 6 and 12 months. Sequelae, grading system as a prognostic tool in Bell’s palsy. Otol Neurotol
such as synkinesis, may occur even at 6 months after disease 2014;35:1669–72.
onset, or beyond; thus, the timing regarding measurement of [11] Berg T, Jonsson L, Engstrom M. Agreement between the Sunnybrook,
non-recovery would affect the results. Fifth, we could not avoid House–Brackmann, and Yanagihara facial nerve grading systems in
Bell’s palsy. Otol Neurotol 2004;25:1020–6.
a bias due to differences in the administration route and where [12] Lunceford JK, Davidian M. Stratification and weighting via the
the administration route impacts on the effectiveness of high- propensity score in estimation of causal treatment effects: a compara-
dose corticosteroids. tive study. Stat Med 2004;23:2937–60.
[13] Curtis LH, Hammill BG, Eisenstein EL, Kramer JM, Anstrom KJ.
Using inverse probability-weighted estimators in comparative effec-
5. Conclusions tiveness analyses with observational databases. Med Care 2007;45:
S103–7.
[14] Takeda T. Facial nerve decompression surgery for Bell’s palsy and
This study showed that high-dose (120 mg PSL equivalent)
Hunt syndrome. In: Japan Society of Facial Nerve Research, editor.
corticosteroid improved prognosis in Bell’s palsy compared Guideline for facial paralysis—Bell’s palsy and Hunt syndrome.
with low-dose corticosteroid (60 mg PSL equivalent). High Tokyo: Kanehara shuppan; 2011. p. 75–7.
efficacy could be obtained only if patients were treated within [15] Adour KK, Wingerd J, Bell DN, Manning JJ, Hurley JP. Prednisone
3 days after disease onset. treatment for idiopathic facial paralysis (Bell’s palsy). N Engl J Med
1972;287:1268–72.
[16] Stennert E. Bell’s palsy – a new concept of treatment. Arch Otorhi-
Funding nolaryngol 1979;225:265–8.
[17] Inamura H, Tojima H, Saito O, Maeyama H, Takeda K, Aoyagi M, et al.
None. Conservative treatment of idiopathic facial palsy—effects of the
administration of high-dose steroids in Bell’s palsy. Nihon Jibiinkoka
Gakkai Kaiho 1992;95:172–7.
Acknowledgment [18] Kawai M, Inamura H, Koike Y. On the conservative treatment of Bell’s
palsy. Acta Otolaryngol Suppl 1988;446:106–10.
None. [19] Axelsson S, Berg T, Jonsson L, Engström M, Kanerva M, Pitkäranta A,
et al. Prednisolone in Bell’s palsy related to treatment start and age.
Otol Neurotol 2011;32:141–6.
References [20] Kang HM, Jung SY, Byun JY, Park MS, Yeo SG. Steroid plus antiviral
treatment for Bell’s palsy. J Intern Med 2015;277:532–9.
[1] Hauser WA, Karnes WE, Annis J, Kurland LT. Incidence and prognosis [21] Choi SA, Shim HS, Jung JY, Kim HJ, Kim SH, Byun JY, et al.
of Bell’s palsy in the population of Rochester, Minnesota. Mayo Clin Association between recovery from Bell’s palsy and body mass index.
Proc 1971;46:258–64. Clin Otolaryngol 2017;42:687–92.
[2] Peitersen E. Bell’s palsy: the spontaneous course of 2500 peripheral [22] Bucak A, Ulu S, Oruc S, Yucedag F, Tekin MS, Karakaya F, et al.
facial nerve palsies of different etiologies. Acta Otolaryngol Suppl Neutrophil-to-lymphocyte ratio as a novel-potential marker for pre-
2002;549:4–30. dicting prognosis of Bell palsy. Laryngoscope 2014;124:1678–81.
[3] Madhok VB, Gagyor I, Daly F, Somasundara D, Sullivan M, Gammie [23] Ushio M, Kondo K, Takeuchi N, Tojima H, Yamaguchi T, Kaga K.
F, et al. Corticosteroids for Bell’s palsy (idiopathic facial paralysis). Prediction of the prognosis of Bell’s palsy using multivariate analyses.
Cochrane Database Syst Rev 2016;7:CD001942. Otol Neurotol 2008;29:69–72.

Please cite this article in press as: Fujiwara T, et al. High-dose corticosteroids improve the prognosis of Bell’s palsy compared with low-dose
corticosteroids: A propensity score analysis. Auris Nasus Larynx (2017), http://dx.doi.org/10.1016/j.anl.2017.09.008