Meta-analysis of molecular imaging of serotonin transporters in major depression

Abstract

The success of serotonin-selective reuptake inhibitors has lent support to the monoamine theory of major depressive disorder (MDD). This issue has been addressed in a number of molecular imaging studies by positron emission tomography or single-photon emission computed tomography of serotonin reuptake sites (5-HTT) in the brain of patients with MDD, with strikingly disparate conclusions. Our meta-analysis of the 18 such studies, totaling 364 MDD patients free from significant comorbidities or medication and 372 control subjects, revealed reductions in midbrain 5-HTT (Hedges' g=-0.49; 95% CI: (-0.84, -0.14)) and amygdala (Hedges' g=-0.50; 95% CI: (-0.78, -0.22)), which no individual study possessed sufficient power to detect. Only small effect sizes were found in other regions with high binding (thalamus: g=-0.24, striatum: g=-0.32, and brainstem g=-0.22), and no difference in the frontal or cingulate cortex. Age emerged as an important moderator of 5-HTT availability in MDD, with more severe reductions in striatal 5-HTT evident with greater age of the study populations (P<0.01). There was a strong relationship between severity of depression and 5-HTT reductions in the amygdala (P=0.01). Thus, molecular imaging findings indeed reveal widespread reductions of ∼10% in 5-HTT availability in MDD, which may predict altered spatial-temporal dynamics of serotonergic neurotransmission.

Figure 1

Reductions of 5-HTT sites in midbrain and thalamus of major depressive disorder (MDD) patients. (A) A forest plot of reports on 5-HTT binding in the midbrain summarizing to an effect size of −0.49 s.d. (B) The corresponding funnel plot displays study precision as a function of effect estimate. Its slightly asymmetric appearance in this instance suggests the lack of one or two data points on the lower right, which implies some publication bias favoring small studies reporting significant decreases of 5-HTT in MDD. (C) A forest plot for imaging studies appraising thalamic 5-HTT, showing a slight reduction in depressed patients. Note the deviation of data published by Cannon et al, the 95% confidence intervals (CI) of which are separated by a substantial gap from the 95% CI of the summary effect of all studies. (D) The corresponding funnel plot appears symmetrical except for one outlier at the right bottom. Δ5-HTT, mean difference in serotonin reuptake transporter level in units of s.d.

Figure 2

Reductions in striatal 5-HTT with major depressive disorder (MDD) are moderated by age. (A) A forest plot for studies exploring the striatal 5-HTT in depressed patients, revealing a notable reduction of 5-HTT. (B) Partial regression plot showing the relationship between mean difference in 5-HTT (Δ5-HTT) and mean age corrected for standardized between-group age differences (Δage) in the current set of data. Here, residuals of effect size against age difference are plotted with residuals of mean age. (C) The definition of the meta-regression model links the mean age of study population to the changes in 5-HTT in the striatum, accounting for Δage. Uniquely in the striatum, age emerged as a factor explaining a significant part of between-study heterogeneity in effect sizes. The black line in the plot shows the predicted effect size at a given age with Δage fixed at zero. Corresponding 95% CI for predictions are indicated by the dashed lines.

Figure 3

Meta-analysis revealed significant reductions of 5-HTT in amygdala, but no effects in cortical regions. (A) A forest plot of molecular imaging studies of the amygdala in depressed patients, showing a significant reduction in 5-HTT levels. Forest plots for (B) frontal cortex and (C) cingulate cortex show no conclusive effect of major depression on 5-HTT availability. Δ5-HTT, mean difference in serotonin reuptake transporter level in units of s.d.