QUESTION: This umbrella review and guidelines aimed to provide evidence to support the rational choice of selected adjunctive therapies for schizophrenia.

STUDY SELECTION AND ANALYSIS: Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) and World Federation of Societies of Biological Psychiatry (WFSBP)-grading recommendations, 63 randomised control trials (RCTs) (of which 4219 unique participants have completed the RCTs) and 29 meta-analyses were analysed.

FINDINGS: Provisional recommendations (WFSBP-grade 1) could be made for two molecules in augmentation to antipsychotics: (1) N-acetyl-cysteine (NAC, 1200-3600 mg/day, for >12 consecutive weeks) in improving negative symptoms, general psychopathology (positive and negative syndrome scale for schizophrenia (PANSS) general psychopathology factor (G)-G subscale), with the RCTs with the longer duration showing the most robust findings; (2) polyunsaturated fatty acids (3000 mg/day of eicosapentaenoic acid, for >12 weeks) in improving general psychopathology. Weaker recommendations (ie, WFSBP-grade 2) could be drawn for sarcosine (2 g/day) and minocycline (200-300 mg/day) for improving negative symptoms in chronic schizophrenia (not early schizophrenia), and NAC for improving positive symptoms and cognition. Weak recommendations are not ready for clinical practice. There is provisional evidence that oestrogens and raloxifene are effective in some patients, but further research is needed to determine their benefit/risk ratio.

CONCLUSIONS: The results of this umbrella review should be interpreted with caution as the number of RCTs included in the meta-analyses was generally small and the effect sizes were weak or medium. For NAC, two RCTs with low risk of bias have provided conflicting results and the WFSBP-grade recommendation included also the results of meta-analyses. These drugs could be provisionally prescribed for patients for whom no other treatments have been effective, but they should be discontinued if they prove ineffective.