Interventions for negative symptoms in schizophrenia: efficacy and clinical interpretability in a meta-analysis of...

Negative symptoms (avolition, anhedonia, asociality, blunted affect, and alogia) are among the most disabling features of schizophrenia spectrum disorders. In the absence of treatment consensus guidelines, this PRISMA-compliant meta-analysis (PROSPERO: CRD42024613967) evaluated efficacy and clinical significance of interventions targeting this dimension. Web of Science/PsycInfo databases were searched from inception to December 2024. Five categories (antipsychotics, other pharmacological agents, brain stimulation, psychosocial, and lifestyle interventions) were analyzed across short/middle/long follow-up times. Categories were divided into 27 subcategories (e.g., 'other pharmacological agents' divided in 14 subcategories including antidepressants, antibiotics, immunomodulators) regardless of follow-up, assessing evidence with GRADE criteria. The primary outcome was the change in negative symptom severity, measured with validated scales (PANSS/SANS/BPRS/CAINS/BNSS) as standardized mean differences (SMD). A clinically meaningful SMD threshold was estimated from the regression between SMD and one-point reductions on the Clinical Global Impression-Severity (CGI-S) scale. This study meta-analyzed 451 trials (n = 42566). The clinically meaningful threshold, obtained from 122 trials reporting CGI-S, was SMD ≥ 0.457. In 214 high-quality studies (n = 19746), 2 category-by-follow-up combinations and 16 subcategories showed significant improvements. Clinically meaningful SMDs for subcategories were antibiotics (0.95; CI: 0.18-1.71; moderate-GRADE), integrated psychosocial interventions (0.93; CI: 0.53-1.33; very-low-GRADE), antidepressants (0.76; CI: 0.33-1.19; moderate-GRADE), physical activity (0.68; CI: 0.39-0.96; very-low-GRADE), transcranial current stimulation (0.52; CI: 0.17-0.86; low-GRADE), and immunomodulators (0.47; CI: 0.26-0.67; high-GRADE), typically as adjuncts to antipsychotics. Heterogeneity was the main limitation. While selected interventions may yield meaningful improvements, more rigorous designs are needed to identify reliable, personalized and scalable treatment options.