Most patients with sciatica (often caused by disc herniations) are managed conservatively at first. The natural course seems to be favorable. The additional value of many conservative therapies remains controversial. Because a systematic review of the conservative treatment of sciatica is lacking, such a review was carried out. After a standardized literature search and selection procedure, methodology and outcome of randomized controlled trials (RCTs) were assessed by three observers. If possible, trial results were statistically pooled (= metaanalysis). The 19 RCTs that were selected showed greatly varying methodologic quality. Pooled odds ratios were calculated for several treatment types. Neither traction, exercise therapy, nor drug therapy was unequivocally effective. Epidural steroids may be beneficial for subgroups of nerve root compression. The literature suggests possible effectiveness of epidural steroids for sciatica. The authors recommend trials to investigate this treatment option further, especially with regard to patient subgroups (e.g., acute sciatica). The effects of clinical heterogeneity and methodology on review results are illustrated.