A meta-analysis and modelling of the duration of symptoms in Bell’s palsy in the Scottish and Scandinavian trials

Abstract Background Idiopathic peripheral facial paralysis (Bell’s palsy) is commonly treated with corticosteroids and sometimes with antiviral therapy in combination. The benefits of corticosteroids for long term complete recovery are well established but the effect of treatment and different treatment combinations on duration of disease are not well discussed in the literature and so is the focus of this investigation. Methods Two randomised controlled studies that were broadly similar in their objectives (being primarily the chances of recovery by end of study), design (different treatment combinations that were very similar in both studies), instruments (measurement of severity), delivery (a visit regime over approximately one year) and in their consideration of outputs (proportion recovered) additionally provided sufficient data for a secondary investigation of duration of symptoms. Probability models for duration were compared and the best-fitting model was selected. Results A significant reduction in median time to recovery after treatment with prednisolone was identified, compared with treatment without prednisolone (54 days vs. 95 days: significance probability SP < 0.0001). Antivirals had no observable effect on time to recovery whether considered alone (101 days vs. 90 days; SP = 0.571) or in combination with prednisolone (53 days vs. 54 days; SP = 0.750). The fitted model was used to provide estimated quantiles and estimates of recovery rates at key intervals. Conclusion These results amplify those previously identified when exploring achievement of final recovery but provide an important additional information about the rate of recovery to patients concerned about their immediate prognosis, and to aid patients and practitioners in their discussion of treatment options.