Objective: To evaluate the efficacy and safety of antidepressant augmentation of antipsychotics in schizophrenia. Methods: Systematic literature search (PubMed/MEDLINE/PsycINFO/Cochrane Library) from database inception until 10/10/2017 for randomized, double-blind, efficacy-focused trials comparing adjunctive antidepressants vs. placebo in schizophrenia. Results: In a random-effects meta-analysis (studies = 42, n = 1934, duration = 10.1 ± 8.1 weeks), antidepressant augmentation outperformed placebo regarding total symptom reduction [standardized mean difference (SMD) = −0.37, 95% confidence interval (CI) = −0.57 to −0.17, P < 0.001], driven by negative (SMD = −0.25, 95% CI = −0.44–0.06, P = 0.010), but not positive (P = 0.190) or general (P = 0.089) symptom reduction. Superiority regarding negative symptoms was confirmed in studies augmenting first-generation antipsychotics (FGAs) (SMD = −0.42, 95% CI = −0.77, −0.07, P = 0.019), but not second-generation antipsychotics (P = 0.144). Uniquely, superiority in total symptom reduction by NaSSAs (SMD = −0.71, 95% CI = −1.21, −0.20, P = 0.006) was not driven by negative (P = 0.438), but by positive symptom reduction (SMD = −0.43, 95% CI = −0.77, −0.09, P = 0.012). Antidepressants did not improve depressive symptoms more than placebo (P = 0.185). Except for more dry mouth [risk ratio (RR) = 1.57, 95% CI = 1.04–2.36, P = 0.03], antidepressant augmentation was not associated with more adverse events or all-cause/specific-cause discontinuation. Conclusions: For schizophrenia patients on stable antipsychotic treatment, adjunctive antidepressants are effective for total and particularly negative symptom reduction. However, effects are small-to-medium, differ across antidepressants, and negative symptom improvement seems restricted to the augmentation of FGAs.