RESUMEN
Abstract Objectives: The benefit of corticosteroids following facial nerve neurorrhaphy in the setting of complete transection is questionable. This systematic review and meta-analysis aimed to evaluate corticosteroid efficacy on facial nerve regeneration and functional recovery after complete disruption and neurorrhaphy. Methods: Randomized controlled trials on both human and animal models from Ovid MEDLINE and Ovid EMBASE studying corticosteroid efficacy in complete facial nerve disruption followed by neurorrhaphy were included. Data were extracted and pooled for meta-analysis. The outcomes were evaluated from electrophysiology, histology, and functional recovery. However, no randomized controlled trial in human was performed. Possibly, performing human trials with histopathology may not be feasible in clinical setting. Results: Six animal studies (248 participants) met inclusion criteria. Electrophysiologic outcomes revealed no differences in latency (Standardized Mean Difference (SMD) = −1.97, 95% CI −7.38 to 3.44, p = 0.47) and amplitude (SMD = 0.37, 95% CI −0.44 to 1.18, p = 0.37) between systemic corticosteroids and controls. When analysis compared topical corticosteroid and control, the results provided no differences in latency (Mean Difference (MD)=0.10, 95% CI −0.04 to 0.24, p = 0.16) and amplitude (SMD = 0.01, 95% CI −0.08 to 0.10, p = 0.81). In histologic outcomes, the results showed no differences in axon diameter (MD = 0.13, 95% CI −0.15 to 0.41, p = 0.37) between systemic corticosteroid and control; however, the result in myelin thickness (MD = 0.06, 95% CI 0.04 to 0.08, p < 0.05) favored control group. When comparing systemic corticosteroid with control in eye blinking, the results favored control (MD= 1.33, 95% CI 0.60 to 2.06, p = 0.0004). Conclusions: This evidence did not show potential benefits of systemic or topical corticosteroid deliveries after facial nerve neurorrhaphy in complete transection when evaluating electrophysiologic, histologic, and functional recovery outcomes in animal models.