To determine if visually-guided arthroscopic irrigation is an effective therapeutic intervention in patients with early knee osteoarthritis.Ninety patients with knee osteoarthritis were randomized in a double-blind fashion to receive either arthroscopic irrigation with 3000 ml of saline (treatment group) or the minimal amount of irrigation (250 ml) required to perform arthroscopy (placebo group). The primary outcome variable was aggregate WOMAC score.The study did not demonstrate an effect of irrigation on arthritis severity as measured by aggregate WOMAC scores, the primary outcome variable; the mean change in aggregate WOMAC score at 12 months was 15.5 (95% CI 7.7, 23.4) for the full irrigation group compared to 8.9 (95% CI 4.9, 13.0) for the minimal irrigation group (P=0.10). Full irrigation did have a statistically significant effect on patients' self-reported pain as measured by the WOMAC pain subscale and by a visual analog scale (VAS) (the secondary outcome variables). Mean change in WOMAC pain scores decreased by 4.2 (95% CI -0.9, 9.4) for the full irrigation group compared with a mean decrease of 2.3 (95% CI -0.1, 4.7) in the minimal irrigation group (P=0.04). Mean VAS pain scores decreased by 1.47 (95% CI -1.2, 4.1) in the full irrigation group compared to a mean decrease of 0.12 (95% CI 0.0, 0.3) in the minimal irrigation group (P=0.02). A hypothesis-generating post-hoc analysis of the effect of positively birefrigent intraarticular crystals showed that patients with and without intraarticular crystals had statistically significant improvements in pain assessments and aggregate WOMAC scores at 12 months; patients with crystals had statistically greater improvements in pain.Visually-guided arthroscopic irrigation may be a useful therapeutic option for relief of pain in a subset of patients with knee OA, particularly in those who have occult intraarticular crystals.