Depression and survival of glioma patients: A systematic review and meta-analysis

Introduction: There is currently a lack of a well-formed consensus regarding the effects of depression on the survival of glioma patients. A more thorough understanding of such effects may better highlight the importance of recognizing depressive symptoms in this patient population and guide treatment plans in the future. Objective: The aim of this meta-analysis was to study the effect of depression on glioma patients' survival. Methods: A meta-analysis was conducted according to the PRISMA guidelines. PubMed, Embase, and Cochrane databases were searched for studies that reported depression and survival among glioma patients through 11/06/2016. Both random-effects (RE) and fixed-effect (FE) models were used to compare survival outcomes in glioma patients with and without depression. Results: Out of 619 identified articles, six were selected for the meta-analysis. Using RE model, the various measures for survival outcomes displayed worsened outcomes for both high and low-grade glioma patients with depression compared to those without depression. For binary survival outcomes, the overall pooled risk ratio for survival was 0.70 (95% CI: 0.47, 1.04; 6 studies; I-2 = 54.9%, P-heterogeneity = 0.05) for high grade gliomas (HGG) and 0.28 (95% CI: 0.04, 1.78; I-2 = 0%, P-heterogeneity = 1.00; one study) for low grade gliomas (LGG) was. A sub-group analysis in the HGG group by depression timing (pre- versus post-operative) revealed no differences between depression and survival outcomes (P-interaction = 0.47). For continuous survival outcomes, no statistically significant difference was found among the high and low-grade glioma groups (P-interaction = 0.31). The standardized mean difference (SMD) in survival outcomes was -0.56 months (95%CI: -1.13, 0.02; 4 studies, I-2 = 89.4%, P-heterogeneity < 0.01) for HGG and -1.69 months (95%CI: -3.26, -0.13; one study; I-2 = 0%, P-heterogeneity = 1.00) for LGG. In patients with HGG, the pooled HR of death also showed a borderline significant increased risk of death among depressive patients (HR 1.42, 95% CI: 1.00, 2.01). Results using the FE model were not materially different. Conclusions: Depression was associated with significantly worsened survival regardless of time of diagnosis, especially among patients with high-grade glioma.