Objective: We performed a systematic review of randomized controlled trials (RCTs) to assess the efficacy of interleukin-2 (IL-2) for the treatment of patientswith unresectable or metastatic renal cell carcinoma (RCC).Methods: We searched the literature to identify RCTs or meta-analyses of RCTscomparing treatment regimens with IL-2 to those without. Outcomes of interestincluded overall or progression-free survival, response rate, toxicity andquality of life.Results: We identified 36 RCTs, and 6 met the eligibility criteria (1098 patients).We studied IL-2 alone and in combination with other agents, including interferon-alpha (IFN-a), 5-fluorouracil (5-FU), and 13-cis-retinoic acid or tamoxifen.No trials comparing high-dose IL-2 to non-IL-2 regimens were identified.A meta-analysis of 1-year mortality data from the 6 trials did not show a differencebetween IL-2-based regimens and non-IL-2 controls. Two of the 6trials detected statistically significant longer survival with IL-2 combinedwith IFN-a and 5-FU. Of the 4 trials that assessed progression-free survival, 3 reported significantly longer progression-free intervals with IL-2-based regimens.Five trials reported response rates; pooling the rates from these trialsgave an overall weighted response rate of 13.3% (range 9%–39%) and 5.3%(range 0%–20%) for IL-2-containing regimens and non-IL-2 regimens, respectively.IL-2-based regimens were more toxic than were non-IL-2 controls;the most frequently reported grade 3–4 toxicities were hypotension (range6%–68%), fever (2%–56%), nausea or vomiting or both (6%–34%), diarrhea(1%–28%) and cardiac toxicity (11%–25%). None of the trials reported healthrelatedquality-of-life data.Conclusion: Non-high-dose IL-2 containing regimens do not provide superiortreatment efficacy over non-IL-2-based regimens, with added toxicity, and thereforeshould not be used as standard treatment for patients with unresectableor metastatic RCC. High-dose IL-2 should only be used by experienced physiciansin the context of a clinical trial or investigative setting.