Statistical Fragility of Steroid–Antiviral Therapy for Bell’s Palsy: A Systematic Review of RCTs

CLINICAL QUESTION: How statistically robust is the evidence supporting combination corticosteroid–antiviral therapy compared with corticosteroid monotherapy for Bell’s palsy?

BOTTOM LINE: Evidence supporting the addition of antivirals to corticosteroids in Bell’s palsy is statistically fragile. Across randomized controlled trials (RCTs), only a few patient outcome reversals would nullify reported significance, suggesting current conclusions about combination therapy lack robustness.

BACKGROUND: Bell’s palsy is an acute, idiopathic facial paralysis that affects 15-40 per 100,000 people annually. Corticosteroids remain the established therapy, improving rates of full recovery. Antivirals such as acyclovir and valacyclovir are often added to target presumed viral etiologies, but RCTs have produced mixed results. Given small sample sizes and reliance on P values, this study evaluated the “statistical fragility” of published evidence comparing corticosteroid monotherapy with combination therapy.

STUDY DESIGN: Systematic review and fragility analysis of 11 RCTs published between 1976 and 2024, including a total of 2,944 participants. Studies compared corticosteroid-only versus corticosteroid–antiviral therapy for Bell’s palsy. Each dichotomous outcome was reassessed using fragility index (FI) and fragility quotient (FQ). These are the metrics that quantify how many outcome reversals would change statistical significance, adjusted for trial size. Risk of bias was evaluated using the Cochrane RoB 2 tool.

SETTING: International RCTs conducted across Europe, Asia, and North America and published in major medical and otolaryngology journals.

SYNOPSIS: Across 57 dichotomous outcomes, only seven (12%) favored combination therapy with statistical significance (P < .05). The mean FI across all outcomes was 5.56, indicating that changing outcomes for just over five patients could overturn significance. The mean FQ was 0.031 (3.1%), meaning 3% of study participants could shift results. Among significant outcomes, the mean FI was even lower (2.7), emphasizing the fragility of positive findings. Nonsignificant outcomes had a reverse FI of 5.96, reflecting similar instability. High risk of bias, variable outcome measures (e.g., House–Brackmann grading), and small sample sizes further undermined statistical strength. Compared with robust cardiovascular trials (median FI ≈13–22), Bell’s palsy studies demonstrated markedly weaker reliability. The authors concluded that corticosteroids remain the standard of care, while the routine addition of antivirals should be approached cautiously until larger, high-quality RCTs with standardized endpoints confirm a consistent benefit.

CITATION: Sharma S, et al. Statistical fragility of steroid–antiviral therapy for Bell’s palsy: a systematic review of randomized controlled trials. Otolaryngol Head Neck Surg. 2025. 173:1069-1078. doi: 10.1002/ohn.1363.

COMMENT: This article investigates the statistical strength of RCTs supporting the use of combination (steroid plus antiviral) therapy versus steroid treatment alone for patients with acute Bell’s palsy. The authors demonstrate evidence that the addition of antivirals is weak. More randomized trials investigating the best treatment for acute Bell’s palsy need to be performed, especially since up to 30% of patients do not make a complete recovery. Matthew Q. Miller, MD