Background: Specific anti-inflammatory agents may yield antidepressant effects; however, no study has gathered the evidence on all anti-inflammatory drugs including a detailed assessment of side effects and risk of bias. Objectives: To systematically review all randomized clinical trials (RCTs) that studied antidepressive treatment and side effects of pharmacological antiinflammatory intervention. Methods: We searched CENTRAL, PubMed, EMBASE, Psychinfo, Clinicaltrials.gov, and review articles for trials published prior to January 1, 2018. We included RCTs investigating anti-inflammatory intervention in adults with depressive symptoms or major depressive disorder (MDD). Two independent reviewers extracted data. Pooled standard mean differences (SMD) including 95%-confidence intervals (95%-CI) were calculated. Outcomes included depression scores after treatment including remission and response and side effects. Results: A total of 41 RCTs (N=10,344) were included, whereof 13 investigated NSAIDs (N=4,362), 11 cytokine-inhibitors (N=3,533), 9 statins (N=2,118), 3 minocycline (N=158), 3 pioglitazone (N=114), and 2 glucocorticoids (N=59). Overall, anti-inflammatory agents improved antidepressant treatment effects compared to placebo by a SMD of-0.46 (95%-CI=-0.62 to-0.30; I2=89%; N=9,459), which was present as add-on in patients with MDD (SMD=-0.56; 95%-CI=-0.84 to-0.27; I2=66%; N=634) and as monotherapy against depressive symptoms (SMD=-0.41; 95%-CI=-0.60 to-0.22; I2=93%, N=8,825). Better antidepressant effects were observed for NSAIDs as add-on (SMD=-0.82; 95%-CI=-1.17 to-0.46; I2=0%; N=132) and monotherapy (SMD=-0.29; 95%-CI=-0.51 to-0.06; I2=84%; N=4,082), cytokineinhibitors as monotherapy (SMD=-0.65; 95%-CI=-1.04 to-0.26; I2=95%; N=3,285), statins as add-on (SMD=-0.73; 95%-CI=-1.05 to-0.42; I2=0%; N=164), minocycline as monotherapy (SMD=-1.06; 95%-CI=-1.68 to-0.44; N=46), and glucocorticoids as add-on (SMD=-0.90; 95%-CI=-1.44 to-0.36; I2=0%; N=59). All studies were associated with high risk of bias. Only 19 trials reported on adverse events, and we found no increased risks for pain/muscle aching, gastrointestinal or cardiovascular events but a trend towards an increased risk for infections. Conclusion: The majority of anti-inflammatory agents improved antidepressant treatment effects. Future large, high-quality RCTs need to include longer follow-up, identify optimal doses and subgroups of patients that can benefit from anti-inflammatory intervention, possibly guided by neurovegetative symptoms, somatic comorbidity and pro-inflammatory markers.