Background: Gout is an inflammatory arthritis that is characterised by the deposition of monosodium urate crystals in synovial fluid and other tissues. The natural history of articular gout is generally characterised by three periods: asymptomatic hyperuricaemia, episodes of acute gout and chronic gouty arthritis. Non-steroidal anti-inflammatory drugs (NSAIDs) including selective cyclo-oxygenase-2 (COX-2) inhibitors (COXIBs) are commonly used to treat acute gout. Published guidelines recommend their use to treat acute attacks, using maximum recommended doses for a short time. Objectives: To assess the benefit and safety of NSAIDs (including COXIBs) for acute gout. Search methods: We searched the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE and EMBASE for studies to 7 October 2013, the 2010 and 2011 ACR and EULAR abstracts and performed a handsearch of reference lists of articles. We searched the World Health Organization (WHO) trial register and ClinicalTrials.gov. We applied no date or language restrictions. Selection criteria: We considered all published randomised controlled trials (RCTs) and quasi-randomised controlled clinical trials that compared NSAIDs with placebo or another therapy (including non-pharmacological therapies) for acute gout. Major outcomes were pain (proportion with 50% or more reduction in pain or mean pain when the dichotomous outcome was unavailable), inflammation (e.g. measured by joint swelling/erythema/tenderness), function of target joint, participant's global assessment of treatment success, health-related quality of life, withdrawals due to adverse events and total adverse events. Data collection and analysis: Two review authors independently selected the studies for inclusion, extracted the data, performed a risk of bias assessment and assessed the quality of the evidence using the GRADE approach. Main results: We included 23 trials (2200 participants). One trial (30 participants) of low-quality evidence compared an NSAID (tenoxicam) with placebo. More participants taking NSAIDs reported at least a 50% reduction in pain after 24 hours (11/15 participants) compared with those taking placebo (4/15 participants) (risk ratio (RR) 2.75, 95% confidence interval (CI) 1.13 to 6.72). There was no difference in the proportion of participants with at least 50% improvement in joint swelling after 24 hours (5/15 participants taking NSAIDs versus 2/15 participants taking placebo; RR 2.50, 95% CI 0.57 to 10.93). The trial did not measure function, participant global assessment of treatment success and health-related quality of life. There were no adverse events reported with the use of tenoxicam; two adverse events (nausea and polypuria) were reported in the placebo group. No between-group differences in outcomes were observed after four days. Moderate-quality evidence based upon four trials (974 participants) indicated that NSAIDs and COXIBs produced similar benefits in terms of pain, swelling and global improvement, but COXIBs were associated with fewer adverse events. Pain reduction was 1.9 points on a 0- to 10-point scale with COXIBs (0 was no pain) while pain reduction with NSAIDs was 0.03 points lower or better (mean difference (MD) -0.03, 95% CI -0.19 to 0.13). Joint swelling in the COXIB group was 1.64 points on a 0- to 3-point scale (0 is no swelling) and 0.13 points higher with NSAIDs (MD 0.13, 95% CI -0.08 to 0.34). Function was not reported. Participant-reported global assessment was 1.56 points on a 0- to 4-point scale with COXIBs (0 was the best score) and was 0.04 points higher with NSAIDs (MD 0.04, 95% CI -0.12 to 0.20). Health-related quality of life assessed using the 36-item Short Form showed no evidence of a statistically significant between-group difference (MD 0.49, 95% CI -1.61 to 2.60 for the physical component). There were significantly fewer withdrawals due to adverse events in participants treated with COXIBs (3%) compared with NSAIDs (8%) (RR 2.39, 95% CI 1.34 to 4.28). There was a significantly lower number of total adverse events in participants treated with COXIBs (38%) compared with NSAIDs (60%) (RR 1.56, 95% CI 1.30 to 1.86). There was moderate-quality evidence based on two trials (210 participants) that oral glucocorticoids did not differ in pain reduction, function or adverse events when compared with NSAIDs. Pain reduction was 9.5 on a 0- to 100-point scale with glucocorticoids, pain reduction with NSAIDs was 1.74 higher or worse (MD 1.74, 95% CI -1.44 to 4.92). The trials did not assess inflammation. Function measured as walking disability was 17.4 points on a 0- to 100-point scale with glucocorticoids, function with NSAIDs was 0.1 lower or better (MD -0.10, 95% CI -4.72 to 4.52). The trials did not measure participant-reported global assessment and health-related quality of life. There were no withdrawals due to adverse events. There was no evidence of a difference in total number of adverse events with glucocorticoids (31%) versus NSAIDs (49%) (RR 1.58, 95% CI 0.76 to 3.28). Authors' conclusions: Limited evidence supported the use of NSAIDs in the treatment of acute gout. One placebo-controlled trial provided evidence of benefit at 24 hours and little or no harm. We downgraded the evidence due to potential selection and reporting biases, and imprecision. While these data were insufficient to draw firm conclusions, they did not conflict with clinical guideline recommendations based upon evidence from observational studies, other inflammatory arthritis and expert consensus, which support the use of NSAIDs in acute gout. Moderate-quality evidence suggested that selective COX-2 inhibitors and non-selective NSAIDs are probably equally beneficial although COX-2 inhibitors are likely to be associated with significantly fewer total and gastrointestinal adverse events. We downgraded the evidence due to an unclear risk of selection and reporting biases. Moderate-quality evidence indicated that systemic glucocorticoids and NSAIDs were also equally beneficial in terms of pain relief. There were no withdrawals due to adverse events and total adverse events were similar between groups. We downgraded the evidence due to unclear risk of selection and reporting bias. There was low-quality evidence that there was no difference in function. We downgraded the quality due to unclear risk of selection bias and imprecision.