DOI 10.1212/WNL.0b013e318275978c
; Published online before print November 7, 2012;Neurology
Gary S. Gronseth and Remia Paduga
American Academy of Neurology
palsy : Report of the Guideline Development Subcommittee of the
Evidence-based guideline update: Steroids and antivirals for Bell
November 7, 2012This information is current as of
http://www.neurology.org/content/early/2012/11/07/WNL.0b013e318275978c
located on the World Wide Web at:
The online version of this article, along with updated information and services, is
rights reserved. Print ISSN: 0028-3878. Online ISSN: 1526-632X.
Allsince 1951, it is now a weekly with 48 issues per year. Copyright © 2012 by AAN Enterprises, Inc.
® is the official journal of the American Academy of Neurology. Published continuouslyNeurology
www.medlive.cn
SPECIAL ARTICLE
Gary S. Gronseth, MD,
FAAN
Remia Paduga, MD
Correspondence & reprint
requests to American Academy of
Neurology:
guidelines@aan.com
Supplemental data at
www.neurology.org
Supplemental Data
Podcast
Evidence-based guideline update: Steroids
and antivirals for Bell palsy
Report of the Guideline Development Subcommittee of the
American Academy of Neurology
ABSTRACT
Objective: To review evidence published since the 2001 American Academy of Neurology (AAN) prac-
tice parameter regarding the effectiveness, safety, and tolerability of steroids and antiviral agents for
Bell palsy.
Methods: We searched Medline and the Cochrane Database of Controlled Clinical Trials for studies
published since January 2000 that compared facial functional outcomes in patients with Bell palsy
receiving steroids/antivirals with patients not receiving thesemedications.We graded each study (Class
I–IV) using the AAN therapeutic classification of evidence scheme. We compared the proportion of
patients recovering facial function in the treated group with the proportion of patients recovering facial
function in the control group.
Results: Nine studies published since June2000on patientswith Bell palsy receiving steroids/antiviral
agents were identified. Two of these studies were rated Class I because of high methodologic quality.
Conclusions and Recommendations: For patients with new-onset Bell palsy, steroids are highly
likely to be effective and should be offered to increase the probability of recovery of facial nerve
function (2 Class I studies, Level A) (risk difference 12.8%–15%). For patients with new-onset
Bell palsy, antiviral agents in combination with steroids do not increase the probability of facial
functional recovery by.7%. Because of the possibility of a modest increase in recovery, patients
might be offered antivirals (in addition to steroids) (Level C). Patients offered antivirals should be
counseled that a benefit from antivirals has not been established, and, if there is a benefit, it is
likely that it is modest at best. Neurology� 2012;79:1–5
GLOSSARY
AAN5American Academy of Neurology;AE5 adverse event;CI5 confidence interval;NNT5 number needed to treat;RD5 risk
difference.
Bell palsy is an acute, peripheral facial paresis of unknown cause.1 Usually the diagnosis is established
without difficulty.2 Up to 30% of patients with Bell palsy fail to recover facial function completely.3
The disease is common, with an annual incidence of 20 per 100,000. Thus, thousands of patients
with Bell palsy are left with permanent, potentially disfiguring facial weakness each year.
In 2001, the Quality Standards Subcommittee of the American Academy of Neurology (AAN)
published an evidence-based practice guideline for the treatment of Bell palsy.4 The 2001 guideline
concluded that steroids were probably effective and antivirals (acyclovir) possibly effective in increas-
ing the probability of complete facial functional recovery in patients with Bell palsy.
This update, developed by the AAN Guideline Development Subcommittee (see appendices e-1
and e-2 on the Neurology® Web site at www.neurology.org), systematically reviews studies published
since June 2000 that are considered relevant to this question: For patients with new-onset Bell palsy,
does treatment with steroids or antiviral agents (acyclovir, famciclovir, valacyclovir) improve facial
functional recovery?
From the Department of Neurology, University of Kansas Medical Center, Kansas City.
Approved by the Guideline Development Subcommittee on January 21, 2012; by the Practice Committee on May 14, 2012; and by the AAN Board of
Directors on August 21, 2012.
Study funding: This guideline was developed with financial support from the American Academy of Neurology. None of the authors received reimbursement,
honoraria, or stipends for their participation in development of this guideline.
Go to Neurology.org for full disclosures. Disclosures deemed relevant by the authors, if any, are provided at the end of this article.
© 2012 American Academy of Neurology 1
ª 2012 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
Published Ahead of Print on November 7, 2012 as 10.1212/WNL.0b013e318275978c
www.medlive.cn
DESCRIPTION OF THE ANALYTIC PROCESS We
searched Medline for articles published from June
2000 through January 2012 using the term “Bell’s palsy”
and the sensitive, therapeutic clinical filter5 (see appendix
e-3 for the specific search strategy employed). The
Cochrane Database of Systematic Reviews and Con-
trolled Clinical Trials was also searched. A secondary
search of the references of selected articles and review
articles (including Cochrane systematic reviews) was per-
formed to identify studies missed by our search strategy.
The titles and abstracts of the identified citations were
reviewed for relevance to the clinical question. The full
text of potentially relevant articles was retrieved and
included in the analysis if these studies determined facial
functional outcomes after at least 3 months of follow-up
in at least 20 patients with new-onset Bell palsy. We
included only controlled trials with prospective data col-
lection comparing outcomes in patients treated with ste-
roids or antiviral agents with patients not treated with
these medications. Both authors independently
reviewed articles and completed data abstraction. Dis-
crepancies were resolved through discussion.
Facial functional recovery was defined as “good” or
“complete” using the same criteria used in the 2001
practice guideline. The quantitative measure of the treat-
ment effect employed was the difference in the propor-
tion of patients attaining complete or good facial recovery
in the treatment group relative to the comparative group
(i.e., the risk difference [RD]). In studies using theHouse
and Brackmann6 facial function scoring system, we con-
sidered an outcome of grade I or II a good recovery.
When comparing the proportion of patients recovering
complete facial function, we considered an outcome of
grade I a complete recovery. The measure of statistical
precision used was the 95% confidence intervals (CIs) of
the RD, and an RD $10% was considered clinically
meaningful. The frequency and severity of the adverse
events (AEs) from the treatments employed were also
abstracted.
Studies were rated for their risk of bias using the AAN
4-tiered classification of evidence scheme for therapeutic
studies (appendix e-4). Studies from the original guideline
were re-rated using the updated classification of evidence
scheme. The strength of practice recommendations was
linked to the strength of evidence (appendix e-5).
For the purpose of formulating conclusions and rec-
ommendations, we used the term “steroids” regardless of
the specific type, dose, and route of steroids used in the
reviewed studies. Likewise, we used the term “antiviral
agents” regardless of the specific type of agent used in the
reviewed studies.
ANALYSIS OF EVIDENCE Our search strategy iden-
tified 340 citations. We reviewed the full text of 38
potentially relevant articles. Nine articles7–15 fulfilled
the inclusion criteria.
For patients with new-onset Bell palsy, does treatment with
steroids improve facial functional recovery? The search
strategy identified 3 articles8,11,15 published since the ini-
tial review comparing outcomes in patients with Bell
palsy treated with steroids with those not treated with
steroids. Table 1 summarizes these studies and includes
2 of the studies reviewed in the original guideline that
attained a grading of Class II or better.16,17 Only Class I
and Class II studies are discussed further.
Table 1 Design characteristics and outcomes in Class I and Class II controlled studies of patients with Bell palsy treated with antiviral agents
and steroids relative to patients treated with steroids alone
Author and
year
Cohort
size Age, y
Steroid dose
duration Rx
Severity,
%a
Duration,
db
Follow-
up, mo
Completion
rate, %c Blind Class
NH
%d
RD good
recovery (CI)
RD complete
recovery (CI)
Engström8
2008
422 Median 39
(IQR 23–54)
Prednisolonee 60 mg
daily 3 5, taper
Med HB 4
IQR 3–5
3 12 99 Yes I 56 — 15% (8%–
21%)
Sullivan11
2007
551 Mean 44
(16.4 SD)
Prednisolone 25 mg BID Mean HB
3.6 6 1.3
3 9 90 Yes I 82 — 12.8%
(7.2%–
18.6%)
Lagalla15
2002
58 Range
15–84
Prednisone 1 g IV 3 3 d
then 0.5 g IV 3 3 d
24 3 12 100 Yes II 75 7% (214%
to 27%)
—
May16
1976
51 53% .30 Prednisone 410
mg 10 d
47 2 6 100 Yes IIf 81 20.75%
(218% to
22.5%)
—
Taverner17
1954
26 Mean 40
(range
12–76)
Hydrocortisone 1 g 8 d 23 9 NS 100 Yes IIf 67 5.25%
(227% to
55%)
—
Abbreviations: CI 5 95% confidence interval; HB 5 House Brackmann score; IQR 5 interquartile range; NH 5 natural history; NS 5 not stated; RD 5 risk
difference (positive values results favoring steroids).
a Percentage of patients with complete palsy.
bMaximum duration of palsy before steroids started.
c Percentage of subjects followed to study completion.
d Percentage of patients not treated with steroids who attained a good outcome.
e Prednisolone and prednisone are dose-equivalent steroids.
f Downgraded by one Class from the rating in the original practice parameter because of no description of allocation concealment.
2 Neurology 79 November 27, 2012
ª 2012 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
www.medlive.cn
The 2 Class I studies8,11 randomized patients to ste-
roids vs placebo and described concealed allocation. Both
studies enrolled patients within 3 days of the onset of
facial weakness. Both studies used prednisolone, one at
60 mg/d for 5 days followed by a 5-day taper,8 the other
25 mg BID for 10 days.11 All studies employed masked
outcome assessment and had high rates of follow-up.
The 3 Class II studies15–17 did not describe concealed
allocation. These studies enrolled fewer than 100 patients
each and described complete follow-up and masked out-
come assessment. One of the Class II15 studies used IV
prednisone. The other 2 used oral steroid preparations.
Efficacy. The 2 Class I studies demonstrated a signif-
icant increase in the probability of complete recovery in
patients randomized to steroids (RD favoring steroids
12.8% and 15%), translating to a number needed to
treat (NNT) of 6 to 8. None of the Class II studies dem-
onstrated a significant benefit from steroids. However,
these studies lacked the statistical precision to exclude a
clinically meaningful effect of steroids.
Safety and tolerability. All studies reported AEs from
steroids. In general, these were minor and temporary.
The most common AEs reported were insomnia and
dyspepsia.
Conclusion. For patients with new-onset Bell palsy, it is
highly likely that steroids are effective in increasing the
probability of complete facial functional recovery
(NNT 6–8, 2 Class I studies).
For patients with new-onset Bell palsy, does treatment with
antiviral agents improve facial functional recovery? We
found 8 articles7–14 published since 2000 comparing out-
comes in patients with new-onset Bell palsy treated with
antiviral agents. Five of these studies7,9,12–14 were rated
Class IV because of nonindependent, nonmasked, non-
objective outcome assessment. These studies are not dis-
cussed further. Table 2 summarizes the remaining Class I
and II studies (2 Class I, 1 Class II).8,10,11 The table also
includes the single Class II study18 from the original
guideline.
The Class I studies compared outcomes in patients
randomized to antivirals and placebo. In addition, the
Class I studies compared outcomes in patients on
antivirals plus steroids with patients on steroids alone.
The Class II studies10,18 compared outcomes only of
patients on antivirals plus steroids with patients on
steroids alone.
Valacyclovir was used in one study8 whereas acyclovir
was the antiviral agent employed in the other studies. The
doses are indicated in table 2. The majority of patients
were enrolled within 3 days of onset of facial weakness.
Efficacy. None of the Class I studies demonstrated a
significant improvement with the use of antivirals as
compared with placebo (random-effects Mantel-Haens-
zel pooled RD 4% favoring placebo, 95% CI 23% to
11%). Although a benefit of antivirals was not observed
in comparison with placebo, some authors have sug-
gested antivirals might have an additional benefit
when added to steroids.9
All of the studies reviewed here specifically compared
outcomes in patients on steroids and antivirals with pa-
tients on steroids alone (table 2). No significant benefit
of antivirals added to steroids as compared with steroids
alone was observed in the Class I and II studies. How-
ever, the 95% CIs of the Class I studies indicate that the
studies’ statistical precision was insufficient to exclude a
modest benefit or harm of antivirals added to steroids
(random-effects Mantel-Haenszel pooled RD 0, 95%
CI28% favoring steroids alone to 7% favoring antivirals
plus steroids). Adding the Class II studies to the meta-
analysis fails to importantly increase the precision of the
analysis (pooled RD 4% favoring steroids plus antivirals,
95% CI 24% to 12%).
Safety and tolerability. None of the studies demon-
strated a significant increase in any AE for patients ran-
domized to an antiviral agent.
Table 2 Design characteristics and outcomes in Class I and II controlled studies of patients with Bell palsy treated with antiviral agents and
steroids relative to patients treated with steroids alone
Author and
year
Cohort
size Age, y (range)
Dose duration
Rx
Severity,
%a
Duration
db
Follow-up,
mo
Completion
rate, %c Blind Class
NH
%d
RD complete
recovery (CI)
Engström8
2008
829 Median 39 (IQR
23–54)
VC 3000 mg/day
7 days
Med HB 4
IQR 3-5
3 12 99 Yes I 76 3.4% (24.6–11.3%)
Sullivan11
2007
551 Mean 44
(16.4 SD)
AC 2000 mg/day
10 days
Mean HB
3.6 6 1.3
3 9 90 Yes I 93 23.3% (29.7–2.7%)
Yeo10 2008 91 Mean 41
(17 SD)
AC 1000 mg/d
5 days
23 53%
# 3 d
6 100 Yes II 85 8.1% (25.6–21.6%)
Adour18
1996
99 Mean 43 AC 400 mg 3 5
qd 10 days
20 3 12 83 Yes II 72 15% (20.9–30.8%)
Abbreviations: AC 5 acyclovir; CI 5 95% confidence interval; HB 5 House Brackmann score; IQR 5 interquartile range; RD 5 risk difference (positive
values results favoring antivirals); NH 5 natural history; VC 5 valacyclovir.
a Percentage of patients with complete palsy.
bMaximum duration of palsy before steroids started.
c Percentage of subjects followed to study completion.
d Percentage of patients treated with steroids only who attained a good outcome.
Neurology 79 November 27, 2012 3
ª 2012 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
www.medlive.cn
Conclusion. For patients with acute-onset Bell palsy, it is
highly likely that antivirals do not moderately (RD.7%)
increase the likelihood of improved facial functional recov-
ery (2 Class I studies). The pooled results of studies with a
low risk of bias lack the statistical precision to exclude a
modest benefit (RD favoring antivirals #7%) or modest
harm (RD favoring steroids alone #8%).
RECOMMENDATIONS For patients with new-onset
Bell palsy, oral steroids should be offered to increase the
probability of recovery of facial nerve function (Level A).
For patients with new-onset Bell palsy, antivirals (in
addition to steroids) might be offered to increase the pro-
bability of recovery of facial function (Level C). Patients
offered antivirals should be counseled that a benefit
from antivirals has not been established, and, if there is
a benefit, it is likely that it is modest at best (RD,7%).
PUTTING THE EVIDENCE INTO A CLINICAL
CONTEXT Although there is strong evidence that ste-
roid use increases the probability of good facial func-
tional recovery in patients with Bell palsy, it does not
necessarily follow that all patients with Bell palsy need
to take steroids. For example, it would be reasonable
for a clinician to opt not to use steroids in a patient with
brittle diabetes mellitus. Other comorbidities potentially
requiring further consideration include morbid obesity,
osteopenia, and a prior history of steroid intolerance.
We found limited evidence of the efficacy of steroids
and antivirals in important Bell palsy subgroups, includ-
ing those with a lower probability of recovery because of
severe palsy at presentation and those with possible zoster
sine herpete. Such studies are particularly important rel-
ative to the efficacy of the addition of antivirals to steroids
given the lack of evidence for moderate efficacy in the
“typical” patient with Bell palsy.
Authors of one Class I study8 performed a preplanned
subgroup analysis on patients with severe palsy at pre-
sentation19 defined by a Sunnybrook Scale score of 0 to
25. This analysis showed no significant difference in
12-month recovery rates between patients treated with
prednisolone alone as compared with patients treated
with prednisolone plus valacyclovir (RD 0.2% favoring
valacyclovir 95% CI, 218% to 17.6%). However, the
analysis lacked the statistical precision to exclude an
important beneficial effect (or harm) from the addition
of valacyclovir. A Class IV study9 observed a significant
improvement in recovery (RD 26.6%) between patients
with severe Bell palsy treated with prednisone alone and
patients with severe Bell palsy treated with prednisone
plus famciclovir (House-Brackmann Scale score of 5 or
6). This study had a high risk of bias because of pseudo-
randomized treatment allocation and unmasked out-
come assessment.
Relative to zoster sine herpete, a Class IV study12
observed no significant difference in recovery after
treatment with prednisolone alone as compared with
treatment with prednisolone plus valacyclovir in a sub-
group of 28 patients with evidence of zoster reactivation
(hazard ratio for recovery 1.6 favoring prednisolone plus
valacyclovir, 95% CI 0.4 to 6.1). The small sample size
and high risk of bias make this observation inconclusive.
These studies in aggregate do not provide strong evi-
dence to identify subgroups of patients that might ben-
efit more or less from treatment.
Because the studies included only patients presenting
early after palsy onset, it is difficult to determine the effect
of steroid or antiviral treatment in patients presenting
later in the course of their illness (e.g., 1 week after the
onset of facial weakness). Likewise, although it seems rea-
sonable to assume that an equivalent dose of alternative
steroids would also be effective, decisions regarding alter-
native steroid dosing regimens necessarily require clini-
cian judgment.
RECOMMENDATIONS FOR FUTURE RESEARCH It
is unlikely that additional research regarding the efficacy
of steroids will change the current estimate of its effect.
Large randomized trials comparing outcomes in patients
with Bell palsy receiving steroids with or without antivi-
rals would help in determining whether the addition of
antivirals to steroid treatment results in a modest benefit.
Such trials should be powered to allow prespecified sub-
grou