Corticosteroid Treatment for Idiopathic Facial Nerve Paralysis: A Meta-analysis

The Laryngoscope Lippincott Williams & Wilkins, Inc., Philadelphia © 2000 The American Laryngological, Rhinological and Otological Society, Inc.

Corticosteroid Treatment for Idiopathic Facial Nerve Paralysis: A Meta-analysis Mitchell Jay Ramsey, CPT, MC, USA; Rebecca DerSimonian, ScD; Michael R. Holtel, CDR, MC, USN; Lawrence P.A. Burgess, COL, MC, USA

Objective: A meta-analysis was designed to evaluate facial recovery in patients with complete idiopathic facial nerve paralysis (IFNP) by comparing outcomes of those treated with corticosteroid therapy with outcomes of those treated with placebo or no treatment. Study Design: Meta-analysis of prospective trials evaluating corticosteroid therapy for idiopathic facial nerve paralysis. Methods: A protocol was followed outlining methods for trial selection, data extraction, and statistical analysis. A MEDLINE search of the English language literature was performed to identify clinical trials evaluating steroid treatment of IFNP. Three independent observers used an eightpoint analysis to determine inclusion criteria. Data analysis was limited to individuals with clinically complete IFNP. The endpoints measured were clinically complete or incomplete facial motor recovery. Effect magnitude and significance were evaluated by calculating the rate difference and Fisher’s Exact Test P value. Pooled analysis was performed with a random effects model. Results: Forty-seven trials were identified. Of those, 27 were prospective and 20 retrospective. Three prospective trials met the inclusion criteria. Tests of heterogeneity indicate the trial with the smallest sample size (RD ⴝ ⴚ0.19; 95% CI, ⴚ0.58 – 0.20), to be an outlier. It was excluded from the final analysis. Analyses of data from the remaining two studies indicate corticosteroid treatment improves complete facial motor recovery for individuals with complete IFNP. Rate difference demonstrates a 17% (99% CI, 0.01– 0.32) improvement in clinically complete recovery for the treatment group based on the random effects model. Conclusions: Corticosteroid treatment provides a clinically and statistically

Presented at the Meeting of the Western Section of the American Laryngological, Rhinological and Otological Society, Denver, Colorado, January 9, 1999. From the Tripler Army Medical Center (M.J.R. M.R.H., L.P.A.B.), Honolulu, Hawaii, and the National Institutes of Health (R.D.), Bethesda, Maryland. The views expressed in this article are those of the authors and do not reflect the official policy or position of the Department of the Army, Department of the Navy, Department of Defense, or the US Government. Editor’s Note: This Manuscript was accepted for publication December 10, 1999. Send Correspondence to Mitchell Jay Ramsey, CPT, MC, USA, Tripler Army Medical Center, 1 Jarrett White Road, Division of Otolaryngology-HNS, MCHK-DSH, Honolulu, HI 96859, U.S.A.

Laryngoscope 110: March 2000

significant improvement in recovery of function in complete IFNP. Key Words: Meta, analysis, corticosteroids, idiopathic facial nerve paralysis, Bell’s palsy. Laryngoscope, 110:335–341, 2000

INTRODUCTION Bell’s palsy, also known as idiopathic facial nerve paralysis (IFNP), is the most common cause of facial paralysis. The incidence is estimated to range from 20 to 30 of every 100,000 individuals.1 Bell’s palsy results in facial motor dysfunction, the degree of which ranges from minor weakness to complete paralysis. The clinical findings depend on the amount of neural injury. Although many grading schemes have been proposed to quantify the extent of facial weakness, which is an indicator of neural injury, none is universal. As a result, many clinicians simply differentiate between incomplete IFNP, which represents paresis of some degree, and complete IFNP, which indicates paralysis of facial musculature. Treatment is aimed at improving recovery of facial function and the prevention of neural degeneration and its associated complications. Of the multiple treatment modalities evaluated over the past three decades, corticosteroid therapy has become the most widely accepted. However, despite the widespread clinical acceptance of corticosteroid therapy for IFNP, efficacy of steroid therapy has not been clearly demonstrated in the literature. Many of the available trials lack outcome stratification by severity of facial motor dysfunction. This is important because incomplete IFNP patients—those with paresis— have a very good prognosis. On the other hand, individuals with complete IFNP—those with paralysis— have the worst prognosis. Seventy percent to 80% of all patients with IFNP will have complete or near complete recovery regardless of whether treatment is given.2 The remaining patients without good recovery have more severe neural injury and typically represent patients with near complete or complete paralysis. Although the rate of complete recovery is unknown in this subgroup, nerve excitability testing has demonstrated that denervation, with associated contracture or synkinesis, occurred in 34% of untreated patients with complete paralysis.3 The important measure that this meta-analysis is designed to determine is whether corticosteroids improve the rate of Ramsey et al.: Idiopathic Facial Nerve Paralysis

335

complete facial recovery over that of the natural rate of recovery for complete IFNP patients. If, for example, the natural history for complete recovery of complete IFNP patients is 40%, do steroids provide a 25% treatment effect, (i.e., improve outcomes from 40 –50%)? If so, then corticosteroids could provide a statistically and clinically significant improvement in recovery. Corticosteroid therapy and recently antiviral therapy4 have become the most widely accepted empiric treatments for IFNP. Owing to wide acceptance and utilization of corticosteroids, it is unlikely that another stratified, randomized, controlled, double-blinded study evaluating steroid efficacy will ever be conducted. At present, a large body of data exists within the literature that, if analyzed appropriately, may give us important information concerning the scientific validity of treating IFNP patients with steroids. Meta-analysis is a form of review in which studies are systematically assembled, appraised, and combined by use of a predefined protocol to reduce bias and subjectivity.5 The strengths of a meta-analysis as outlined by Sacks et al.6 are 1) to increase statistical power for primary end points and for subgroups, 2) to resolve uncertainty when reports disagree, 3) to improve estimates of effect size, and 4) to answer questions not posed at the start of individual trials. Several excellent reviews are available discussing the application and characteristics of meta-analysis.5– 8 We have designed this meta-analysis in an attempt to resolve uncertainty about corticosteroid efficacy in the treatment of IFNP. We have limited our evaluation to individuals with complete IFNP because of the poorer prognosis for recovery. The null hypothesis is that no difference exists between outcomes for complete IFNP patients treated with corticosteroids or controls (placebo or no treatment).

MATERIALS AND METHODS

criteria. The inclusion criteria were 1) prospective studies with concurrent controls; 2) IFNP was diagnosed with the exclusion of other causes of facial paralysis; 3) treatment started within 7 days of onset of paralysis; 4) treatment with a total prednisone equivalent dose of 400 mg or greater; 5) exclusion of multiple interventions other than corticosteroids; 6) outcome data for complete IFNP patients and controls; 7) follow-up of at least 4 months, or to complete recovery; and 8) unilateral, nonrecurrent IFNP. Abstracts and unpublished data, as well as clinical trials with duplicate data were excluded. Data extraction was attempted for each trial regardless of inclusion criteria. Outcome measures for treatment and control populations were determined for subjects with complete IFNP. Outcomes were defined as complete and incomplete facial motor recovery. Complete paralysis was defined as a House-Brackmann grade 6, or otherwise defined by the trial as total or complete paralysis. Cure was defined as complete or total recovery of facial paralysis or House-Brackmann grade 1. Data from excluded trials were identified when available. This was done to provide the most complete presentation of all available data regarding corticosteroid efficacy for patients with complete IFNP. Although the data from excluded trials cannot be directly compared with the pooled results from the included trials, the data may provide a frame of reference for interpreting the overall results of the meta-analysis. Only data from trials comparing corticosteroids with placebo or no treatment are presented.

Quality Analysis A quality index was calculated for trials meeting inclusion criteria by subjecting them to a quality assessment evaluation based on criteria adapted from Chalmer’s method for assessing randomized controlled trials.9 The trials were scored independently by three of the co-authors (M.R., M.H., L.B.). The scores were averaged and a final quality assessment score was obtained. A quality index ranging from zero to one was calculated for each study based on the percentage of total points earned. This index was used as a general assessment of the quality of the trial design.

Literature Search

Statistical Methods of Data Analysis

Most articles were identified by literature searches of the MEDLINE and PreMEDLINE databases covering 1966 through January 1998 and limited to English-language articles. Searches were performed independently by the principal investigator and a medical reference librarian. MeSH and free-text terms included the following: “facial paralysis,” “Bell’s Palsy,” and “idiopathic facial nerve paralysis,” combined with “drug therapy” and “therapy.” The results were then combined with “corticosteroids,” “adrenal cortex hormones,” “synthetic glucocorticoids,” “prednisolone,” “prednisone,” and “steroids.” The term “therapeutic use” was applied in some cases to the drug terms. Additional articles were found through scanning Current Contents and bibliographies of relevant articles and book chapters. Articles considered for inclusion were any trial evaluating corticosteroid therapy for treatment of IFNP patients.

Individual studies were analyzed using rate difference (RD) with 95% CI as the measure of effect magnitude. Significance was determined at ␣ ⫽ 0.05 level. This was calculated using a Fisher’s exact test (two tailed). Drug efficacy is indicated when the RD is positive; however, if the RD 95% CI contains zero, the results are not statistically significant. Homogeneity was measured using the Q statistic.7 The Q statistic evaluates the null hypothesis that all study effects are similar, i.e., homogeneous. A small Q statistic with a large P value implies homogeneity. A large Q statistic with a small P value implies heterogeneity. Pooled results were obtained using a random effects model with 99% CI (DerSimonian and Laird). Data from several excluded trials were analyzed in a similar fashion. RD was used as the measure of effect magnitude and significance was determined with Fisher’s exact test. Only trials with treatment and control (placebo or no treatment) were evaluated. Studies comparing multiple medical therapies were not evaluated. Results from excluded trials did not undergo pooled analysis.

Trial Selection and Data Extraction Identified trials underwent independent evaluation by three of the co-authors (M.R., M.H., L.B.) following a prospectively defined protocol. For each trial a standardized form was completed by recording the following; inclusion criteria, treatment rendered, epidemiological characteristics, treatment and control numbers, and, if available, outcome data. All of the publications were completely reviewed. Inclusion was determined if a trial met all eight inclusion

Laryngoscope 110: March 2000

336

RESULTS Trial Selection Forty-seven trials were identified. Twenty trials were excluded because they were retrospective.10 –29 Of the reRamsey et al.: Idiopathic Facial Nerve Paralysis

maining 27 prospective trials only 3 met the inclusion criteria. The remaining prospective trials were excluded for a variety of reasons (Table I). Eighteen studies were excluded for lack of outcome data. Nine studies had multiple medical therapy or surgical therapy. Nine trials used corticosteroid doses that did not meet the inclusion criteria. The three trials meeting the inclusion criteria are listed in Table II.

presentation of the data. Of the total possible points, 60% account for study protocol characteristics, 30% for statistical analysis, and 10% for presentation of results. The total score ranges from zero to one. The quality assessment scores for each study ranged from 33% to 73%, with a mean of 59%. Individual quality scores are listed in Table II.

Outcomes Quality Assessment Each trial was evaluated by three of the co-authors (M.R., M.H., L.B.) using Chalmer’s method for assessing a randomized controlled trial.9 The quality of a trial is based on a variety of study design characteristics as well as the

Outcomes of individual trials. A total of 230 complete IFNP patients were identified from three trials. Complete recovery rates ranged from 27% to 100% for treatment groups and 24% to 84% for control groups. Analyses of individual trials indicate that in all but one

TABLE I. Prospective Studies: Reasons for Exclusion.

Failure to Exclude Other Etiologies

Trial

Abraham et al.37 (1987) Adour and Swanson3 (1971) Adour et al.38 (1972) Adour et al.4 (1996)* Al-Husaini et al.39 (1986)† Brown32 (1982) De-Diego et al.40 (1998)‡ Devriese41 (1977) Ekstrand and Glitterstam42 (1979) Hyden et al.43 (1993) Kawai et al.44 (1988) Kinishi et al.45 (1991)㛳 May et al.46 (1985) Racic et al.47 (1997)¶ Stennert48 (1979) Stennert49 (1981)# Tani50 (1988)** Taverner51 (1954) Taverner et al.52 (1966) Taverner et al.53 (1967) Taverner et al.54 (1971) Watanabe55 (1996)‡‡ Williamson et al.56 (1996) Wolf57 (1978)§§

Treatment Started After 7 Days

Steroid Dose Did Not Meet Protocol Criteria

Multiple Medical Treatment or Surgical Therapies

✕

Less than 4-Month Follow-up

✕

✕

✕ ✕

Inability to Extract Complete IFNP Outcomes

✕

✕ ✕

✕ ✕

✕

✕

✕

✕ ✕ ✕

✕§

✕ ✕ ✕

✕

✕

✕ ✕ ✕†† ✕

✕ ✕ ✕

✕ ✕ ✕ ✕ ✕

✕ ✕ ✕ ✕ ✕ ✕ ✕

No studies were excluded because of individuals with bilateral or recurrent IFNP patients. ✕ marks indicate that particular study had those exclusion criteria. * Steroid versus steroid and acyclovir. † No control group, treatment with prednisone or adrenocorticotrophic hormone. ‡ Prednisone versus acyclovir. § Adrenocorticotrophic hormone was used. 㛳 Low-dose oral steroid versus intravenous high-dose steroid, dextran, and pentoxifylline. ¶ Hyperbaric oxygen versus prednisone. # No controls, treatment with intravenous steroid, dextran, and pentoxifylline. ** Compared historic controls treated with oral steroid to patients treated with intravenous steroid, dextran, and pentoxifylline. †† Adrenocorticotrophic hormone was used. ‡‡ No controls, treatment with intravenous steroid, Hespander, Mannitol. §§ Outcomes from this study could not be directly determined. IFNP ⫽ idiopathic facial nerve paralysis.

Laryngoscope 110: March 2000

Ramsey et al.: Idiopathic Facial Nerve Paralysis

337

TABLE II. Characteristics of Included Studies. Trial

Quality index score Total subjects No. treatment cures/no. of treated (%) No. control cures/no. of control (%) Fisher’s Exact test P value Rate difference (95% CI) Total prednisone equivalent dose

May31 (1975)

0.71 24 3/11 (27) 6/13 (46) 0.42 ⫺0.19 (⫺0.58–.20) 410 mg

study the data imply a beneficial treatment effect. Data from the trial by Shafshak et al.30 indicate a statistically significant treatment effect (RD ⫽ 0.18; 95% CI, 3–33; P ⫽ .02). Data from Austin’s trial indicate a marginally significant (P ⫽ .12) treatment effect (RD ⫽ 0.14; 95% CI, ⫺0.04 – 0.32). Results of the RD favor treatment in all but the study by May et al.31 Data from this trial demonstrate that corticosteroids resulted in poorer facial recovery than placebo. Data from all three trials are presented in Table II. Analysis of the data collected from each co-investigator (M.R., M.H., L.B.) indicate 100% correlation of individually identified complete IFNP outcome data. Data from all three trials are presented in Figure 1 and Table II. Tests of heterogeneity. All three trials meeting inclusion criteria were analyzed to determine their homogeneity. This analysis provides a statistical measure of their similarity. Trials that are significantly different, or heterogeneous, may not provide accurate results when combined. Analysis of the three trials demonstrates significant heterogeneity (P ⫽ .06). The study by May et al.,31 which had a small sample size and showed a worse outcome with corticosteroids, accounted for the largest amount of the observed heterogeneity. Excluding the data from the May study resulted in minimal heterogeneity (P ⫽ .20). Pooled results. Results of pooled analysis from the two trials included in the final meta-analysis indicate that patients with complete IFNP treated with corticosteroids had complete recovery 17% more on average than patients who received placebo or no treatment (99% CI, 0.01– 0.32;

Austin et al.36 (1993)

Shafshak et al.30 (1994)

0.72 0.32 46 160 20/20 (100) 39/93 (42) 22/26 (85) 16/67 (24) 0.12 0.02 0.14 (⫺0.04–0.32) 0.18 (0.03–0.33) 405 mg (est.) 425 mg avg.

P ⫽ .005). Analysis of the Shafshak30 trial and the Austin trial36 indicated minimal heterogeneity (P ⫽ .20). Results from pooled data are displayed in Table III. A sensitivity analysis was performed to examine the effect of including May’s study, which was identified as an outlier and excluded from the pooled analysis. Data analysis from all three trials meeting inclusion criteria trials indicates that patients with complete IFNP treated with corticosteroids had complete recovery on average 12% more often than those receiving placebo or no treatment (99% CI, ⫺0.07– 0.31). Statistical significance is not demonstrated; Fisher’s exact P value is .11. In light of the statistical heterogeneity (P ⫽ .06) of the observed effects in the three studies meeting inclusion criteria, the overall pooled estimate of 12% does not reflect the true effect of corticosteroid treatment. May’s study, which had a small sample size, accounted for the largest amount of the observed heterogeneity and was considered separately because of its outlying results. Results from excluded trials. Outcome data were collected from one prospective and five retrospective studies during the initial evaluation of trials. RDs with 95% CI were calculated as the effect measure of magnitude and significance was determined using Fisher’s exact test to calculate a P value. The results show the trend in these studies is uniformly in favor of the treatment group, with complete recovery rates ranging from 49% to 97% for the treatment group and 23% to 64% for the control group. Data from all studies, except Brown’s,32 demonstrate statistically significant positive treatment effects. Significant variation exists in trial size and design. The outcomes, characteristics, and treatment regimens are presented in Table IV. These data were not included in the metaanalysis, nor were they pooled in any other manner because of failure to meet one or more of the inclusion criteria.

DISCUSSION

Fig. 1. Rate difference for the three trials selected in the metaanalysis as well as the rate difference from the pooled results with all trials included and with the outlier removed. Rate differences between (0 –1) indicate a positive treatment effect. Rate differences between (⫺1– 0) indicate a negative effect. The larger the rate difference, the larger the treatment effect.

Laryngoscope 110: March 2000

338

Bell’s palsy, otherwise known as idiopathic facial nerve paralysis, is a common problem affecting the facial nerve. The most readily apparent symptom of this condition is hemifacial paresis or paralysis. Patients afflicted with IFNP have variable recovery largely dependent on the degree of motor dysfunction, which reflects the amount of neural degeneration. The literature clearly shows that virtually all patients with clinically incomplete Ramsey et al.: Idiopathic Facial Nerve Paralysis

TABLE III. Pooled Statistical Data From the Random Effects Model. No. of Trials

No. of Patients

All 3 trials

3

230

All except (May) outlier

2

206

Trials

OR* (99% CI)

RD (99% CI)

2.01 (0.36–11.21) P ⫽ .31 3.27 (0.76–14.10) P ⫽ .04†

0.12 (⫺0.07–0.31) P ⫽ .11 0.17 (0.01–0.32) P ⫽ .005†

Test of Heterogeneity P Value

P ⫽ .06

P ⫽ .20

Magnitude effect is represented by rate difference (RD) and odds ratio (OR). The Q statistic is a measure of variability between studies, and it helps determine if heterogeneity is present. The Q statistic evaluates the null hypothesis that all studies are similar, i.e. homogeneous. A small Q statistic with a large P value implies homogeneity. A large Q statistic with a small P value implies heterogeneity. * In the calculation of OR for Austin’s study, 0.5 added to each cell. † Results are the same for both fixed and random effects models.

paralysis that remains incomplete have excellent recovery of facial function independent of treatment. Individuals with complete paralysis comprise the worst prognostic group and do not have the same recovery. Many often suffer from persistent weakness or the complications of neural degeneration. Because the large majority of patients with incomplete paralysis will have good recovery regardless of any therapy given, the more pertinent question to address is what effect corticosteroid therapy will have on patients with complete IFNP. Several papers have reviewed the available literature to clarify the overall effectiveness of steroid therapy,33–35 but the controversy still remains. Reasons why the data are inconclusive range from poor study design to inadequate trial size. Perhaps most importantly, many of the earlier studies did not stratify outcomes based on severity of facial weakness. The lack of stratification limits the interpretation of the data as previously demonstrated.33 Meta-analysis was used to provide a systematic, unbiased method of synthesizing data from related studies.

Our null hypothesis states that corticosteroid treatment is no better than placebo or no therapy for complete IFNP patients. Three trials met all of the inclusion criteria. One study was identified as an outlier because the inclusion of its data into the meta-analysis resulted in effect heterogeneity. Because pooled data from heterogeneous studies can yield erroneous conclusions about treatment effect, we did not include the data from this outlier study in the overall effect estimation. We believe that the pooled results, with outlier data removed, represents the most accurate estimate of the benefit of corticosteroid treatment. Corticosteroids improve the overall incidence of complete recovery by 17%. This is statistically significant (95% CI, 0.01– 0.32; P ⫽ .005). The clinical significance of the treatment effect is illustrated by comparing the absolute improvement, which is the overall increase in the cure rate with the relative improvement, which is the change produced in that proportion of the population not expected to recover. The data from this study indicate the natural rate of complete recovery without treatment, for patients with

TABLE IV. Additional Data From Excluded Trials. Trial

Treatment

Adour et al.12 Brown et al.32 Devi et al.14 Jarvis20 Saito et al.27 Sellars28

Prednisone, dose unstated Cortisone, 400 mg Prednisone, est. 400–600 mg Prednisolone, 520 mg Multiple Multiple

Reason for Exclusion

Treatment Cures/ Total Treated (%)

Control Cures/Total Controls (%)

P Value

Rate Difference with 95% CI

1,4

23/29 (79%)

10/25 (40%)

P ⫽ .005

0.39 (0.15–0.63)

4 1,3

20/41 (49%) 45/61 (74%)

4/10 (40%) 6/26 (23%)

P ⫽ .731 P ⬍ .001

0.09 (⫺0.25–0.43) 0.51 (0.31–0.71)

1

38/43 (88%)

42/66 (64%)

P ⫽ .004

0.24 (0.09–0.39)

1,5 1

35/36 (97%) 163/195 (84%)

14/24 (58%) 67/105 (64%)

P ⬍ .001 P ⬍ .001

0.39 (0.18–0.60) 0.20 (0.09–0.31)

Reasons for exclusion: 1. Retrospective or unstated and assumed to be retrospective. 2. Other causes not excluded. 3. Treatment started ⬎7 days from onset, or unknown. 4. Inadequate steroid treatment or unknown dosage. 5. Multiple medical or surgical therapy. 6. Follow-up ⬍4 months or unknown. 7. Inability to determine outcomes. 8. Bilateral or recurrent palsy.

Laryngoscope 110: March 2000

Ramsey et al.: Idiopathic Facial Nerve Paralysis

339

complete IFNP, is approximately 40% (control outcomes averaged from Austin and Shafshak alone, and from all trials in Table II and Table IV). The absolute increase in complete recovery of 17% corresponds to a 28% relative improvement for the remaining 60% of the subjects who would not have recovered without any treatment. That is, of every 100 subjects treated with steroids, about 40 would have recovered regardless of the treatment, of the remaining 60 who would have not recovered without steroids, about 17 (28%) additional subjects will recover with corticosteroid therapy. Certain factors should be considered when interpreting this meta-analysis. First, despite the identification of multiple trials, only three met the inclusion criteria. Of those three, one trial was excluded based on the significant heterogeneity observed when compared with the other studies. Although the results from two studies are pooled, the data are weighted so that they reflect the results of Shafshak’s30 study to a larger degree than Austin’s.36 Despite these considerations, the information gained is statistically significant and clearly supports the use of corticosteroids in treatment of complete IFNP patients. A separate analysis of individual studies was performed on data from excluded trials. This was not a metaanalysis. This was done to provide the reader with all of the available data possible. It is not presented to support or refute the evidence determined by the meta-analysis. Although the data from the excluded studies appear to strongly support the efficacy of corticosteroid treatment, there are significant differences in the design and quality of the studies and the results of each study should be individually interpreted after complete review. It is interesting that in all but one study, our analysis shows statistically significant evidence of improved cure rates for complete IFNP patients treated with steroids. The data from Brown’s study did not demonstrate significance. The total dose of corticosteroid in that study was equivalent to 80 mg of prednisone. This is much less than any other study. The results of this meta-analysis have important implications for physicians caring for patients with IFNP and for physicians planning future research. This meta-analysis clearly shows a statistically significant benefit for complete IFNP patients treated with a total prednisone dose equivalent of 400 mg or more, started within 7 days of onset of paresis. Unfortunately, approximately 20% of patients with incomplete IFNP will progress to complete paralysis within five days after onset. For this reason, our therapeutic protocol is to treat all patients regardless of severity with prednisone and famciclovir as early as possible unless contraindicated. Prednisone is dosed at 80 mg a day for 5 days then tapered dependent on whether the paresis progresses to paralysis or remains stable. If the palsy remains stable, we either stop steroids without tapering or rapidly taper the dose over the next 4 days for a total dose of 530 mg. If the patient presents with paralysis or progresses to complete paralysis, the dose is tapered over the next 7 days for a total dose of 680 mg. Based on Adour’s4 study, we also prescribe an antiviral medication, famciclovir 500 mg three times daily for 10 days. It is discontinued after 5 days if the palsy remains incomplete. Famciclovir is recommended over acyLaryngoscope 110: March 2000

340

clovir because of easier dosage schedule and fewer gastrointestinal side effects. Eye protection with ophthalmic ointment and saline drops is also prescribed. The meta-analysis provided no information on whether treatment started early versus late within the first 7 days had improved outcome. However, corticosteroids should be initiated as soon as possible after onset of palsy. Shafshak’s30 study evaluating patients with HouseBrackmann grades of 5 and 6 demonstrated statistically significant improvement in outcome for patients treated with prednisone within 24 hours of onset. No difference existed between patients who started treatment after 24 hours and the control group. Brown32 also demonstrated improved outcome when complete IFNP patients received treatment within the first 24 hours of onset. In his study, 41 patients with complete IFNP were treated with steroids. Eighteen of 20 who recovered completely were treated within 24 hours. Of the nine patients with complete IFNP who had poor recovery, all received treatment on the third day after onset. Thus despite the low dose given, these data show a dramatic improvement in outcome for those treated early.

CONCLUSION Results of this meta-analysis clearly demonstrate better outcomes for individuals with complete IFNP who are treated with steroids. Patients with complete IFNP should be treated with a minimum of 400 mg of prednisone, or an equivalent dose of corticosteroid, started within 7 days of onset of facial weakness. Corticosteroid treatment improves the incidence of complete facial recovery by 17%. This is clinically and statistically significant. Recovery of patients with IFNP is largely dependent on severity of facial motor dysfunction, which reflects neural injury. Approximately 60% of individuals with complete IFNP will not be expected to have complete facial recovery without treatment. All patients should be treated initially and as close to onset as possible regardless of severity of palsy. A significant number of patients presenting with incomplete IFNP may progress to complete paralysis, prompting treatment of all IFNP patients initially.

BIBLIOGRAPHY 1. Hauser WA, Karnes WE, Annis J, Kurland LT. Incidence and prognosis of Bell’s palsy in the population of Rochester, Minnesota. Mayo Clin Proc 1971;46:258 –264. 2. Peitersen E. The natural history of Bell’s palsy. Am J Otol 1982;4:107–111. 3. Adour KK, Swanson PJ Jr. Facial paralysis in 403 consecutive patients: emphasis on treatment response in patients with Bell’s palsy. Trans Am Acad Ophthalmol Otolaryngol 1971;75:1284 –1301. 4. Adour KK, Ruboyianes JM, Von Doersten PG, et al. Bell’s palsy treatment with acyclovir and prednisone compared with prednisone alone: a double-blind, randomized, controlled trial. Ann Otol Rhinol Laryngol 1996;105: 371–378. 5. Rosenfeld RM. How to systematically review the medical literature. Otolaryngol Head Neck Surg 1996;115:53– 63. 6. Sacks HS, Berrier J, Reitman D, Ancona-Berk VA, Chalmers TC. Meta-analyses of randomized controlled trials. N Engl J Med 1987;316:450 – 455.

Ramsey et al.: Idiopathic Facial Nerve Paralysis

7. DerSimonian R, Laird N. Meta-analysis in clinical trials. Controlled Clin Trials 1986;7:177–188. 8. Berlin JA, Laird NM, Sacks HS, Chalmers TC. A comparison of statistical methods for combining event rates from clinical trials. Stat Med 1989;8:141–151. 9. Chalmers TC, Smith H, Blackburn B, et al. A method for assessing the quality of a randomized control trial. Controlled Clin Trials 1981;2:31– 49. 10. Adour KK, Bell DN, Wingerd J. Bell palsy. Dilemma of diabetes mellitus. Arch Otolaryngol 1974;99:114 –117. 11. Adour KK, Wingerd J. Idiopathic facial paralysis (Bell’s palsy): factors affecting severity and outcome in 446 patients. Neurology 1974;24:1112–1116. 12. Adour KK, Byl FM, Hilsinger RL Jr, Kahn ZM, Sheldon MI. The true nature of Bell’s palsy: analysis of 1,000 consecutive patients. Laryngoscope 1978;88:787– 801. 13. Bryant, FL. Intratympanic injection of steroid for treatment of facial paralysis. Laryngoscope 1973;83:700 –706. 14. Devi S, Challenor Y, Duarte N, Lovelace RE. Prognostic value of minimal excitability of facial nerve in Bell’s palsy. J Neurol Neurosurg Psychiatry 1978;41:649 – 652. 15. Devriese PP. Bell’s palsy in children. Acta Otorhinolaryngol Belg 1984;38:261–267. 16. Devriese PP, Schumacher T, Scheide A, de Jongh RH, Houtkooper JM. Incidence, prognosis and recovery of Bell’s palsy. A survey of about 1000 patients (1974 –1983). Clin Otolaryngol 1990;15:15–27. 17. Hilsinger RL Jr, Adour KK, Doty HE. Idiopathic facial paralysis, pregnancy, and the menstrual cycle. Ann Otol Rhinol Laryngol 1975;84:433– 442. 18. Hughes GB. Practical management of Bell’s palsy. Otolaryngol Head Neck Surg 1990;102:658 – 663. 19. Ing DK. Bell’s palsy—a retrospective study. Med J Malaya 1974;28:248 –250. 20. Jarvis JF. A review of 250 cases of Bell’s palsy. S Afr Med J 1974;48:593–596. 21. Katusic SK, Beard CM, Wiederholt WC, Bergstralh EJ, Kurland LT. Incidence, clinical features, and prognosis in Bell’s palsy, Rochester, Minnesota, 1968 –1982. Ann Neurol 1986;20:622– 627. 22. McCafferty GJ. Bell’s palsy—a clinical survey. J Otolaryngol Soc Aust 1972;3:410 – 413. 23. McGovern FH, Estevez J. Observations on 160 patients with peripheral facial paralysis seen in private practice. Laryngoscope 1978;88:1427–1438. 24. Prescott CA. Idiopathic facial nerve palsy (the effect of treatment with steroids). J Laryngol Otol 1988;102:403– 407. 25. Prescott CA. Idiopathic facial nerve palsy in children and the effect of treatment with steroids. Int J Pediatr Otorhinolaryngol 1987;13:257–264. 26. Ramamurthi, Ramamurthi B. Prognosis in Bell’s palsy—influence of early treatment. J Indian Med Assoc 1974;62:281–282. 27. Saito O, Aoyagi M, Tojima H, Koike Y. Diagnosis and treatment for Bell’s palsy associated with diabetes mellitus. Acta Otolaryngol Suppl Stockh 1994;511:153–155. 28. Sellars SL. Idiopathic facial nerve palsy. S Afr Med J 1983; 64:379 –381. 29. Thomas MH. Treatment of Bell’s palsy with cortisone and other measures. Neurology 1955;5:882– 886. 30. Shafshak TS, Essa AY, Bakey FA. The possible contributing factors for the success of steroid therapy in Bell’s palsy: a clinical and electrophysiological study. J Laryngol Otol 1994;108:940 –943. 31. May M, Wette R, Hardin WB Jr, Sullivan J. The use of steroids in Bell’s palsy: a prospective controlled study. Laryngoscope 1976;86:1111–1122. 32. Brown JS. Bell’s palsy: a 5 year review of 174 consecutive cases: an attempted double blind study. Laryngoscope 1982;92:1369 –1373. 33. Burgess LP, Yim DW, Lepore, ML. Bell’s palsy: the steroid controversy revisited. Laryngoscope 1984;94:1472–1476. 34. Huizing EH, Mechelse K, Staal A. Treatment of Bell’s palsy.

Laryngoscope 110: March 2000

35. 36.

37. 38. 39. 40.

41. 42.

43.

44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57.

An analysis of the available studies. Acta Otolaryngol Stockh 1981;92:115–121. Stankiewicz JA. Steroids and idiopathic facial paralysis. Otolaryngol Head Neck Surg 1983;91:672– 677. Austin JR, Peskind SP, Austin SG, Rice DH. Idiopathic facial nerve paralysis: a randomized double blind controlled study of placebo versus prednisone. Laryngoscope 1993; 103:1326 –1333. Abraham IL, Oosting J, Hart AA. Bell’s palsy: factors affecting the prognosis in 200 patients with reference to hypertension and diabetes mellitus. Clin Otolaryngol 1987;12:349 –355. Adour KK, Wingerd J, Bell DN, Manning JJ, Hurley JP. Prednisone treatment for idiopathic facial paralysis (Bell’s palsy). N Engl J Med 1972;287:1268 –1272. Al-Husaini A, Jamal GA, Hilmi AM, Mathur VS. Steroid therapy in Bell’s palsy. Int J Clin Pharmacol Ther Toxicol 1986;24:430 – 432. De-Diego JI, Prim MP, De-Sarria MJ, Madero R, Gavilan J. Idiopathic facial paralysis: a randomized, prospective, and controlled study using single-dose prednisone versus acyclovir three times daily. Laryngoscope 1998;108: 573–575. Devriese PP. Prednisone in idiopathic facial paralysis (Bell’s palsy). ORL J Otorhinolaryngol Relat Spec 1977;39: 257–271. Ekstrand T, Glitterstam K. Bell’s palsy— beneficial effect of treatment with adrenocorticotrophic hormone (ACTH) in patients with a poor prognosis. Acta Otolaryngol Stockh 1979;87:9 –15. Hyden D, Roberg M, Forsberg P, et al. Acute “idiopathic” peripheral facial palsy: clinical, serological, and cerebrospinal fluid findings and effects of corticosteroids. Am J Otolaryngol 1993;14:179 –186. Kawai M, Inamura H, Koike Y. On the conservative treatment of Bell’s palsy. Acta Otolaryngol Suppl (Stockh) 1988; 446:106 –110. Kinishi M, Amatsu M, Hosomi H. Conservative treatment of Bell’s palsy with steroids and dextran-pentoxiphylline combined therapy. Eur Arch Otorhinolaryngol 1991;248:147–149. May M, Klein SR, Taylor FH. Idiopathic (Bell’s) facial palsy: natural history defies steroid or surgical treatment. Laryngoscope 1985;95:406 – 409. Racic G, Denoble PJ, Sprem N, Bojic L, Bota B. Hyperbaric oxygen as a therapy of Bell’s palsy. Undersea Hyperb Med 1997;24:35–38. Stennert E. Bell’s palsy—a new concept of treatment. Arch Otorhinolaryngol 1979;225:265–268. Stennert E. Pathomechanisms in cell metabolism: a key to treatment of Bell’s palsy. Ann Otol Rhinol Laryngol 1981; 90:577–583. Tani M, Kinishi M, Takahara T, Hosomi H, Amatsu M. Medical treatment of Bell’s palsy. Oral vs. intravenous administration. Acta Otolaryngol Suppl Stockh 1988;446:114 –118. Taverner D. Cortisone treatment of Bell’s palsy. Lancet 1954; 2:1052–1054. Taverner D, Fearnley ME, Kemble F, Miles DW, Peiris OA. Prevention of denervation in Bell’s palsy. BMJ 1966;5484: 391–393. Taverner D, Kemble F, Cohen SB. Prognosis and treatment of idiopathic facial (Bell’s) palsy. BMJ 1967;4:581–582. Taverner D, Cohen SB, Hutchinson BC. Comparison of corticotrophin and prednisolone in treatment of idiopathic facial paralysis (Bell’s palsy). BMJ 1971;4:20 –22. Watanabe S, Kenmochi M, Kinoshita H, Kato I. Effects of administration of high dose hydrocortisone on Bell’s palsy. Acta Otolaryngol Suppl (Stockh) 1996;522:108 –110. Williamson IG, Whelan TR. The clinical problem of Bell’s palsy: is treatment with steroids effective? Br J Gen Pract 1996;46:743–747. Wolf SM, Wagner JH, Davidson S, Forsythe A. Treatment of. Bell palsy with prednisone: a prospective, randomized study. Neurology 1978;28:158 –161.

Ramsey et al.: Idiopathic Facial Nerve Paralysis

341