Background: Indications and techniques of rehabilitation differ widely across types of lumbar surgery, including timing (before or after surgery) and prescriptions (surgeons but also medical or paramedical professionals).Objectives: This project aimed to build consensual recommendations for practice in this context.Methods: The SOFMER methodology was used to establish recommendations for physical medicine and rehabilitation: a steering committee defined the types of lumbar surgery involved and developed the main questions to be addressed; a scientific committee performed a literature review for grading evidence and proposed the first version of recommendations, which were discussed during a dedicated session at the national Physical and Rehabilitation Medicine congress; then an e-Delphi method with cross-professional experts was used to finalise recommendations and reach a multidisciplinary consensus.Results: The main questions developed were the value of rehabilitation before and after surgery, timing and type of rehabilitation, benefit of supervision and instrumental rehabilitation, value of patient education, and complementary interventions concerning rehabilitation for discectomy, fusion, and disc prosthesis (excluding decompression for spinal stenosis). The literature review identified 60 articles, but for several of the questions, no article in the literature addressed the issue. The multidisciplinary scientific committee analysed the literature and addressed the questions to propose the first version of a set of 23 recommendations. The congress session failed to answer all questions or to reach consensus for all items. After a three-step e-Delphi, 20 recommendations were retained, for which consensus among experts was reached. The recommendations are applicable only to patients without a neurological lesion.Conclusions: These recommendations provide important and consensual knowledge to assist clinicians in decision-making for rehabilitation in lumbar surgery. Despite many of the recommendations relying exclusively on expert opinion rather than published evidence, this approach is an important advance to improve concordance among healthcare professionals.