Article type
Year
Abstract
Background: In recent meta-analyses proinflammatory cytokines, such as interleukin-6 (IL-6), have been associated with major depression in humans. We wanted to know if the prior accumulated evidence on animal experimentation could lead and support these findings.
Objectives: To perform a rapid review on the association of IL-6 and depression models in animal research, differentiating between models without direct inflammatory implications (chronic stress and related models) and models with direct inflammatory implications (use of substances that trigger an immune response).
Methods: A PubMed search was run with appropriate terms. Effect sizes (standardized mean differences, SMD) were estimated directly from the data and/or statistics reported in the included studies. We used a random-effects model to pool the effect estimates and estimate prediction intervals. A sensitivity analysis was performed and funnel-plots drawn to assess the presence of small study bias.
Results: Fifteen trials on the alteration of IL-6 levels in models without direct inflammatory implications were included. Overall IL-6 increased in depression (random-effects model SMD = 0.57; 95% CI 0.34 to 0.80; prediction interval = 0.32 to 0.82; I2 0%). Sensitivity analysis showed no trial exerted a significant influence on the pooled estimate, and a funnel plot gave no indication of small studies bias. Twelve trials on the alterations of IL-6 levels in models with direct inflammatory implications were included. Pooled results showed an increase of IL-6 in depression (random-effects model SMD = 1.72; 95% CI 0.89 to 2.55). However results were highly heterogeneous (I2 79.3%), giving a large prediction interval (-0.79 to 4.23), and a funnel plot with indications of small studies bias.
Conclusions: Disparate results were obtained with two different models of animal depression, and the possibility of publication bias was suggested in one. Similar problems could be presented in other areas of translational research, and must be deal with appropriately before jumping directly from the bench to the human.
Objectives: To perform a rapid review on the association of IL-6 and depression models in animal research, differentiating between models without direct inflammatory implications (chronic stress and related models) and models with direct inflammatory implications (use of substances that trigger an immune response).
Methods: A PubMed search was run with appropriate terms. Effect sizes (standardized mean differences, SMD) were estimated directly from the data and/or statistics reported in the included studies. We used a random-effects model to pool the effect estimates and estimate prediction intervals. A sensitivity analysis was performed and funnel-plots drawn to assess the presence of small study bias.
Results: Fifteen trials on the alteration of IL-6 levels in models without direct inflammatory implications were included. Overall IL-6 increased in depression (random-effects model SMD = 0.57; 95% CI 0.34 to 0.80; prediction interval = 0.32 to 0.82; I2 0%). Sensitivity analysis showed no trial exerted a significant influence on the pooled estimate, and a funnel plot gave no indication of small studies bias. Twelve trials on the alterations of IL-6 levels in models with direct inflammatory implications were included. Pooled results showed an increase of IL-6 in depression (random-effects model SMD = 1.72; 95% CI 0.89 to 2.55). However results were highly heterogeneous (I2 79.3%), giving a large prediction interval (-0.79 to 4.23), and a funnel plot with indications of small studies bias.
Conclusions: Disparate results were obtained with two different models of animal depression, and the possibility of publication bias was suggested in one. Similar problems could be presented in other areas of translational research, and must be deal with appropriately before jumping directly from the bench to the human.