The incidence of regression after the non-surgical treatment of symptomatic lumbar disc herniation: a systematic review and meta-analysis

BACKGROUND: Although the regression of symptomatic lumbar disc herniation (SLDH) has been widely reported, little data exist regarding the generalized incidence of regression (IR). We aimed to review the varying IRs and to synthesize the pooled IR of non-surgically-treated SLDH. METHODS: Four electronic databases were searched for relevant studies pertaining to the regression of SLDH after non-surgical treatment and for potential studies that may have reported morphological changes in lumbar disc herniation in the follow-up results of SLDH patients treated non-surgically. The main outcome was the regression of SLDH. A random effects model was used to determine the pooled IR of SLDH. RESULTS: We identified 13,672 articles, 38 of which were eligible for analysis. Our analysis included 2219 non-surgically treated SLDH patients, 1425 of whom presented regression. The pooled IR was 63% (95% CI 0.49–0.77). In subgroup analyses, studies that quantitatively measured the regression of SLDH yielded statistically higher pooled IRs than those that used qualitative methods. The pooled IRs gradually increased in randomized controlled trials and prospective and retrospective studies. The pooled IR varied from 62 to 66% after the sequential omission of any single study. Meta-regression showed that study types, herniation levels and regression measurements caused heterogeneity. CONCLUSIONS: We report an overall IR of 63% among non-surgically treated SLDH patients, thus providing clinical decision makers with quantitative evidence of IR. Based on our systematic review, we suggest a follow-up timeline with time points 4 and 10.5 months after onset when deciding whether to perform surgery for SLDH.