BACKGROUND: Arthroscopic knee surgery remains a common treatment for symptomatic knee osteoarthritis, including for degenerative meniscal tears, despite guidelines strongly recommending against its use. This Cochrane Review is an update of a non-Cochrane systematic review published in 2017. OBJECTIVES: To assess the benefits and harms of arthroscopic surgery, including debridement, partial menisectomy or both, compared with placebo surgery or non-surgical treatment in people with degenerative knee disease (osteoarthritis, degenerative meniscal tears, or both). SEARCH METHODS: We searched the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, Embase, and two trials registers up to 16 April 2021, unrestricted by language. SELECTION CRITERIA: We included randomised controlled trials (RCTs), or trials using quasi-randomised methods of participant allocation, comparing arthroscopic surgery with placebo surgery or non-surgical interventions (e.g. exercise, injections, non-arthroscopic lavage/irrigation, drug therapy, and supplements and complementary therapies) in people with symptomatic degenerative knee disease (osteoarthritis or degenerative meniscal tears or both). Major outcomes were pain, function, participant-reported treatment success, knee-specific quality of life, serious adverse events, total adverse events and knee surgery (replacement or osteotomy). DATA COLLECTION AND ANALYSIS: Two review authors independently selected studies for inclusion, extracted data, and assessed risk of bias and the certainty of evidence using GRADE. The primary comparison was arthroscopic surgery compared to placebo surgery for outcomes that measured benefits of surgery, but we combined data from all control groups to assess harms and knee surgery (replacement or osteotomy). MAIN RESULTS: Sixteen trials (2105 participants) met our inclusion criteria. The average age of participants ranged from 46 to 65 years, and 56% of participants were women. Four trials (380 participants) compared arthroscopic surgery to placebo surgery. For the remaining trials, arthroscopic surgery was compared to exercise (eight trials, 1371 participants), a single intra-articular glucocorticoid injection (one trial, 120 participants), non-arthroscopic lavage (one trial, 34 participants), non-steroidal anti-inflammatory drugs (one trial, 80 participants) and weekly hyaluronic acid injections for five weeks (one trial, 120 participants). The majority of trials without a placebo control were susceptible to bias: in particular, selection (56%), performance (75%), detection (75%), attrition (44%) and selective reporting (75%) biases. The placebo-controlled trials were less susceptible to bias and none were at risk of performance or detection bias. Here we limit reporting to the main comparison, arthroscopic surgery versus placebo surgery. High-certainty evidence indicates arthroscopic surgery leads to little or no difference in pain or function at three months after surgery, moderate-certainty evidence indicates there is probably little or no improvement in knee-specific quality of life three months after surgery, and low-certainty evidence indicates arthroscopic surgery may lead to little or no difference in participant-reported success at up to five years, compared with placebo surgery. Mean post-operative pain in the placebo group was 40.1 points on a 0 to 100 scale (where lower score indicates less pain) compared to 35.5 points in the arthroscopic surgery group, a difference of 4.6 points better (95% confidence interval (CI) 0.02 better to 9 better; I2 = 0%; 4 trials, 309 participants). Mean post-operative function in the placebo group was 75.9 points on a 0 to 100 rating scale (where higher score indicates better function) compared to 76 points in the arthroscopic surgery group, a difference of 0.1 points better (95% CI 3.2 worse to 3.4 better; I2 = 0%; 3 trials, 302 participants). Mean post-operative knee-specific health-related quality of life in the placebo group was 69.7 points on a 0 to 100 rating scale (where higher score indicates better quality of life) compared with 75.3 points in the arthroscopic surgery group, a difference of 5.6 points better (95% CI 0.36 better to 10.68 better; I2 = 0%; 2 trials, 188 participants). We downgraded this evidence to moderate certainty as the 95% confidence interval does not rule in or rule out a clinically important change. After surgery, 74 out of 100 people reported treatment success with placebo and 82 out of 100 people reported treatment success with arthroscopic surgery at up to five years (risk ratio (RR) 1.11, 95% CI 0.66 to 1.86; I2 = 53%; 3 trials, 189 participants). We downgraded this evidence to low certainty due to serious indirectness (diversity in definition and timing of outcome measurement) and serious imprecision (small number of events). We are less certain if the risk of serious or total adverse events increased with arthroscopic surgery compared to placebo or non-surgical interventions. Serious adverse events were reported in 6 out of 100 people in the control groups and 8 out of 100 people in the arthroscopy groups from eight trials (RR 1.35, 95% CI 0.64 to 2.83; I2 = 47%; 8 trials, 1206 participants). Fifteen out of 100 people reported adverse events with control interventions, and 17 out of 100 people with surgery at up to five years (RR 1.15, 95% CI 0.78 to 1.70; I2 = 48%; 9 trials, 1326 participants). The certainty of the evidence was low, downgraded twice due to serious imprecision (small number of events) and possible reporting bias (incomplete reporting of outcome across studies). Serious adverse events included death, pulmonary embolism, acute myocardial infarction, deep vein thrombosis and deep infection. Subsequent knee surgery (replacement or high tibial osteotomy) was reported in 2 out of 100 people in the control groups and 4 out of 100 people in the arthroscopy surgery groups at up to five years in four trials (RR 2.63, 95% CI 0.94 to 7.34; I2 = 11%; 4 trials, 864 participants). The certainty of the evidence was low, downgraded twice due to the small number of events. AUTHORS' CONCLUSIONS: Arthroscopic surgery provides little or no clinically important benefit in pain or function, probably does not provide clinically important benefits in knee-specific quality of life, and may not improve treatment success compared with a placebo procedure. It may lead to little or no difference, or a slight increase, in serious and total adverse events compared to control, but the evidence is of low certainty. Whether or not arthroscopic surgery results in slightly more subsequent knee surgery (replacement or osteotomy) compared to control remains unresolved.