The value of serum uric acid as a prognostic biomarker in amyotrophic lateral sclerosis: Evidence from a meta-analysis.

OBJECTIVE To determine the value of uric acid (UA) as a prognostic biomarker for amyotrophic lateral sclerosis (ALS) using a meta-analysis of hazard ratio-based studies. METHODS We included data from Tokushima University (47 patients with ALS) and three previous studies (1835 patients with ALS) with a hazard ratio (HR) identified by a systematic computational search. A total of four studies and 1882 patients were enrolled in the pooled analysis. We pooled HRs of death or tracheostomy, which were estimated by a Cox proportional hazard model, using a random-effects model. Heterogeneity was assessed by Q statistic, and a p value < 0.1 was considered significant heterogeneity. Furthermore, sensitivity analysis was performed to assess the effect of each single study and the robustness of the summary effect. We evaluated publication bias by visual assessment of the funnel plot and Egger's test, and adjusted the bias using a trim-and-fill method. RESULTS This meta-analysis revealed that UA could be a prognostic factor for ALS (all, HR = 0.87, p < 0.001; men, HR = 0.83, p < 0.001; women, HR = 0.76, p < 0.001). The included studies were homogeneous (all, p = 0.43; men, p = 0.9; women, p = 0.49). Sensitivity analysis confirmed the robustness of these summary effects. Publication bias was detected, which was adjusted for by a trim-and-fill method. The adjusted results showed significant summary effects (all, HR = 0.88, p = 0.002; men, HR = 0.83, p < 0.001; women, HR = 0.77, p < 0.001). CONCLUSION The present meta-analysis suggests that the serum UA level could be a prognostic biomarker in patients with ALS. Sensitivity analyses and the trim-and-fill method supported the robustness of these results.