Author
Subjects
Treatment
Brain areas
Results
Vakili et al. [100]
38 patients (17 males, mean age 38.5 ± 10.0 years)
20 healthy controls (9 males, mean age 40.3 ± 10.4 years)
Fluoxetine, 20 mg/day
Prospective, 8 weeks
Hippocampus
Larger right hippocampal volume in female responders vs. nonresponders
Salloway et al. (2002) [86]
59 patients under sertraline (mean age 69.22 ± 5.63 years)
111 patients under citalopram (mean age 79.43 ± 4.13 years)
Sertraline, citalopram
Prospective, 8 weeks
Open label
Subcortical hyperintensities
Larger level of subcortical hyperintensities in citalopram group vs. sertraline group, strong correlation with age
No correlation between subcortical hyperintensities and treatment response
Hsieh et al. [44]
60 patients (24 males, mean age 68.57 ± 6.43 years)
Multiple treatment
(SSRIs, bupropion, venlafaxine, nefazodone, mirtazapine, nortriptyline)
Prospective, 12 weeks
Hippocampus
Smaller right hippocampal volume in nonremitting patients vs. remitting patients
Pizzagalli et al. [81]
20 melancholic patients (7 males, mean age 36.57 ± 12.9 years)
18 non-melancholic patients (8 males, mean age 33.17 ± 8.8 years)
18 healthy controls (8 males, mean age 38.1 ± 13.6 years)
Nortriptyline up to 150 ng/ml
Longitudinal, 6 months
Automated analysis
No volumetric differences between groups
Frodl et al. [28]
30 patients (12 males, mean age 48,4 ± 13,9 years)
30 healthy controls (12 males, mean age 45,7 ± 12,9 years)
Multiple drugs
(fluvoxamine, paroxetine, sertraline, citalopram, venlafaxine, mirtazapine, amitriptyline, doxepin, trimipramine, and reboxetine)
Longitudinal, 1 year f-up
Amygdala,
hippocampus
Smaller bilateral hippocampal volumes in nonremitting patients vs. remitted patients, both at baseline and at f-up
Vythilingam et al. [102]
38 patients (15 males, mean age 41 ± 11 years)
33 healthy controls (12 males, mean age 34 ± 10 years)
Multiple drugs
Longitudinal: remission (22 patients, 20 fluoxetine, 1 venlafaxine, 1 sertraline)
(subanalysis with 20 patients: fluoxetine)
Longitudinal, 7 ± 3 months
Hippocampus and temporal lobe
No treatment effect
Lavretsky et al. [59]
11 treatment exposure patients (4 males, mean age 67 ± 6.1 years)
30 drug-naive patients
(5 males, mean age 71.7 ± 7.8 years)
41 healthy controls (21 males, mean age 72.2 ± 7.3 years)
Multiple drugs
Cross-sectional
Automated analysis
Larger OFC GM volumes in treated depressed patients compared to drug-naive depressed patients, but smaller than those in normal controls
Papakostas et al. [77]
50 patients (33 males, mean age 41.2 ± 10.2 years)
Fluoxetine, 20 mg/day
Prospective, 8 weeks
Subcortical and periventricular WM hyperintensities
Greater severity of hyperintensities in nonresponders vs. responders
Chen et al. [19]
17 patients (5 males, mean age 44.06 ± 8.36 years)
Fluoxetine, 20 mg/day
Prospective, 8 weeks
Automated analysis
Larger GM volume in anterior cingulate cortex, insula, and right temporoparietal cortex in responders vs. nonresponders
Colla et al. [22]
24 patients (9 males, mean age 54.5 ± 11.9 years)
14 healthy controls (6 males, mean age 53.8 ± 17.7 years)
Paroxetine, up to 40 mg/day
Amitriptyline, up to 150 mg/day
Prospective, 4 weeks
Hippocampus
No treatment effect
Frodl et al. [28]a
38 patients (13 males, mean age 46.1 ± 11.3 years)
30 healthy controls (11 males, mean age 43.6 ± 11.3 years)
Multiple drugs
(fluvoxamine, paroxetine, sertraline, citalopram, venlafaxine, mirtazapine, amitriptyline, doxepin, trimipramine, and reboxetine)
Longitudinal, 3 years f-up
Automated analysis
F-up:
Smaller decline in left hippocampus, left anterior cingulum, left dorsomedial prefrontal cortex, and bilateral dorsolateral prefrontal cortex in remitting patients vs. nonremitting
MacQueen et al. [64]
14 remitted patients (8 males, mean age 30.5 ± 9.5 years)
32 nonremitted patients (15 males, mean age 27.6 ± 10.5 years)
Multiple drugs
(citalopram, venlafaxine, bupropion, mirtazapine, sertraline, and fluvoxamine)
Prospective, 8 weeks
Hippocampus
Larger pretreatment hippocampal body/tail volume in remitted patients vs. nonremitted patients
Kronmüller et al. [56]
57 patients (24 males, mean age 43.54 ± 12.82 years)
30 healthy controls
Multiple drugs
Prospective, 2 years
Hippocampus
Smaller bilateral hippocampal volume in male relapsing patients vs. healthy controls
Costafreda et al. [24]
37 patients (9 males, mean age 43.2 ± 8.8 years)
37 healthy controls (9 males, mean age 42.8 ± 6.7 years)
Fluoxetine, 20 mg/day
Prospective, 8 weeks
Automated analysis
Larger GM volume in right rostral anterior cingulate cortex, left posterior cingulate cortex, left middle frontal gyrus, and right occipital cortex in remitted patients vs. nonremitted
Gunning et al. [35]a
22 remitted patients (mean age 71.0 ± 5.6 years)
19 nonremitted patients (mean age 70.0 ± 6.3, years)
Escitalopram, 10 mg/day
Prospective, 12 weeks
Anterior cingulate
Smaller dorsal and rostral anterior cingulate GM volumes in nonremitters vs. remitters
Li et al. [61]
25 nonremitted (5 males, mean age 46.5 ± 10.4 years)
19 remitted (6 males, mean age 42.6 ± 13.0 years)
Multiple drugs
(SSRIs, SNRIs, or bupropion)
Prospective, 6 weeks
Automated analysis
Smaller left DLPFC in nonremitters compared to remitters
Lorenzetti et al. [63]
27 remitted patients (9 males, mean age 35,07 ± 9,96 years)
29 currently ill patients (7 males, mean age 32,52 ± 8,28 years)
31 healthy controls (10 males, mean age 34,68 ± 9,86 years)
Multiple drugs
Cross-sectional
Amygdala
Larger left amygdala volumes in remitted patients vs. healthy controls and currently ill patients
Sheline et al. [90]a
217 patients (96 males, mean age 68.4 ± 7.2 years)
Sertraline, up to 200 mg/day
Prospective, 12 weeks
WM hyperintensities
Lower signal hyperintensities in remitters vs. nonremitters
Gunning-Dixon et al. [36]a
22 remitted patients (9 males, 69.61 ± 4.71 years)
20 nonremitted patients (8 males, 71.18 ± 6.95 years)
25 healthy controls (9 males, 70.68 ± 5.82 years)
Escitalopram, 10 mg/day
Prospective, 12 weeks
Signal hyperintensities
Greater signal hyperintensities in nonremitters vs. healthy controls and remitters
Lai and Hsu [58]
15 patients (5 males, mean age 35.87 ± 9.59, years)
15 healthy controls (4 males, mean age 34.30 ± 9.87 years)
Duloxetine, 60 mg/day
Longitudinal, 6 weeks
Automated analysis
F-up:
Larger GM volumes over left inferior frontal cortex, right occipital fusiform gyrus, and right cerebellum VIIIa regions in remitted patients compared to baseline
Sneed et al. [94]
28 nonremitted patients (17 males, mean age 66.5 ± 7.9 years)
10 remitted patients (5 males, mean age 64.7 ± 6.5 years)
Sertraline, up to 200 mg/day
Nortriptyline, 1 mg/kg/day
Prospective, 12 weeks
WM hyperintensities
Greater hyperintensities in nonremitted patients vs. remitted patients
Sheline et al. [91]a
168 patients (mean age 67.96 ± 7.49 years)
50 healthy controls (mean age 73.00 ± 5.31 years)
Sertraline, up to 200 mg/day
Prospective, 12 weeks
Hippocampus, amygdala, parahippocampus, and caudate, anterior cingulate gyrus, frontal pole, superior frontal gyrus, orbital frontal gyrus, and middle frontal gyrus
Smaller hippocampal volume and frontal cortical thickness in nonremitters vs. remitters
Smith et al. [93]
13 patients (4 males, mean age 35.23 ± 9.04 years)
10 healthy controls (1 males, mean age 35.67 ± 12.3 years)
Sertraline, up to 200 mg/day
Longitudinal, 12 weeks
Automated analysis
F-up:
Enlargement in DLPFC in patients compared to baseline
Ribeiz et al. [83]
30 patients (7 males, mean age 70.73 ± 6.59 years)
22 patients (5 males, mean age 70.41 ± 7.58)
Not specified antidepressant treatment
Prospective, 24 weeks
Automated analysis
Larger left lateral OFC in remitted patients vs. nonremitted
Kong et al. [54]
28 drug-naive patients (11 males, mean age 34.42 ± 8.24 years)
28 healthy controls (14 males, mean age 32.07 ± 9.27 years)
24 treated patients acquired at f-up (10 males, mean age 36.12 ± 5.73 years)
Fluoxetine, 10 – 40 mg/day
Longitudinal, 8 weeks
Automated analysis
Larger left middle frontal gyrus and right OFC in treated patients at f-up vs. healthy controls
No differences between drug-naive and treated patients
Yuen et al. [105]
16 patients with apathy (7 males, mean age 71.6 ± 5 years)
29 patients without apathy (10 males, 69 ± 5.8 years)
49 healthy controls (16 males, mean age 70.6 ± 6.4 years)
Escitalopram 10 mg
Prospective, 12 weeks
Cingulate cortex
Larger left posterior subgenual cingulate volume correlated with apathy amelioration
Taylor et al. [96]
11 patients (5 males, mean age 64.6 ± 4.4 years)
Citalopram fluoxetine, sertraline, venlafaxine
Prospective 3 and 6 months
WM hyperintensities
Greater WM hyperintensities at cingulum bundle associated with worse outcome
Jung et al. [48]
26 nonresponders (7 males, mean age 40.8 ± 12.7 years)
24 responders (7 males, mean age 43.0 ± 10.1 years)
29 healthy controls (8 males, mean age 43.6 ± 13.4 years)
Bupropion, duloxetine, escitalopram, venlafaxine, paroxetine
Prospective, 8 weeks
Automated analysis
Smaller right superior frontal gyrus in nonremitters
Larger lingual gyrus in remitters
Korgaonkar et al. [55]
40 nonresponders (males 19, mean age 37.1 ± 14.8 years)
34 responders (18 males, mean age 28.2 ± 7.4 years)
Escitalopram, sertraline, venlafaxine
Prospective, 8 weeks
Automated analysis
Smaller left middle frontal gyrus and larger right angular gyrus in nonremitters
Klauser et al. [52]
29 patients currently ill (7 males, mean age 33.09 ± 8.25 years)
27 remitted patients (9 males, mean age 35.02 ± 9.72 years)
Multiple treatments (SSRIs, SNRIs, TCAs, NsSSAs, MAOIs, NRIs)
Cross-sectional
Automated analysis
No treatment effect
Fu et al. [30]
32 patients (19 males, mean age 42.2 ± 11.2 years)
25 healthy controls (12 males, mean age 38.8 ± 9.9 years)
Duloxetine 60–120 mg
Longitudinal, 8 weeks
MR scan 0, 1, 8, 12 weeks
Amygdala, anterior cingulate, hippocampus
Larger hippocampal volumes among remitters
Phillips et al. [80]
14 remitted patients (3 males, mean age 44.7 ± 10.5 years)
12 remitted (5 males, mean age 47.5 ± 10.6 years)
Multiple treatments
Longitudinal, 6 months from remission to 12 months of nonremission
Hippocampus, rostral middle frontal gyrus, orbitofrontal cortex, rostral and caudal anterior cingulate cortices, and inferior temporal gyrus
Larger hippocampal volume and cortical thickness in the rostral middle frontal gyrus, orbitofrontal cortex, and inferior temporal gyrus in remitters vs. nonremitters
Even though the neurobiology of depression has not yet completely understood, brain structural alterations have been reported in frontotemporal, hippocampal, and striatal regions [1]. Clinical depression includes different domains that involve cognitive, emotional, and neurovegetative symptoms. Neurobiological hypotheses on the etiopathology of this heterogeneous syndrome involve stress response mechanisms, immunomodulatory impairment, and neurochemical alterations. Thus, brain area alteration studies have been focused on those circuits, which are implicated in stress-related response or on those neurotransmitters such as serotonin and norepinephrine, which have been selected as target biochemicals for antidepressant treatment. Evidence in this field is still inconclusive and brain imaging offers a valid and widely available instrument to investigate the neurobiological implications of the disease and the mechanism of action of effective therapies [9]. Also, imaging results can differentiate groups of patients who respond to antidepressant therapies from those who do not. Indeed, rates of treatment response to antidepressants reach about 30 %, even after multiple trials. Enrollment of patients with treatment resistance due to not yet recognized cause might limit the power of drug trials to demonstrate efficacy of new compounds; furthermore, time spent on inefficacious treatments causes a delay in the consideration of more efficacious approaches for refractory patients. For these reasons, ability to predict treatment response is a challenge for the progression of research on novel treatments for depression. In this field, imaging features can offer an objective and replicable predictor of treatment efficacy or response.
Imaging studies on depressive disorder have often focused on limbic system structures, namely, hippocampus and amygdala, since they are involved in functions such as declarative memory and mood regulation, which are impaired in depression. Converging evidence from animal postmortem and clinical examinations reports reduction of volumes in hippocampus in depressive disorder. MR data confirm robustly this observation. Hippocampi have been reported to be reduced in volume in patients with chronic depression [22, 44], nonremitting, first episode, or unmedicated patients [5], with only few contrary results of no volume change in comparison with healthy controls [102]. Lower volumes were reported to progress over time [28, 29]; duration of illness, rather than age, has been hypothesized to determine hippocampal volume reduction. As these results were observed in a heterogeneous population, several hypotheses have been raised to explain lower hippocampal volumes in depression. The connection between stress, hypercortisolemia, and hippocampal function has often been sustained, not only by imaging studies [66]. However, factors such as illness duration, long-term pharmacological treatment, life events (e.g., sexual or childhood abuse [101]), comorbidity with anxiety disorders, somatoform disorders, or alcohol (substance) abuse [95] might have confounded results in this field. The object has been summarized in few meta-analyses since the early 2000s. Campbell and colleagues reported reduced hippocampal volume in patients with depressive disorder [13]. Inconsistencies across studies have been attributed to methodological issues such as slice thickness of images, due to differences in the machinery adopted, or tracing methods, or definition of anatomical boundaries of hippocampus (particularly the limit between hippocampus and amygdala). Indeed, the shape rather than the overall volume of hippocampus might be altered, as effect of fibers’ disruption [12]. Duration of illness has been accounted for one of the major contributors to hippocampal atrophy, which have been observed in schizophrenia as well [12]. Nevertheless, reduction of volume has been confirmed for patients at first depression episode [21]. These data have been further confirmed by Arnone and colleagues [1].
Along with hippocampus, a reduction of volume has been frequently observed for amygdala, also [89]. Reduction of volume has been reported for psychotic, recent-onset, and chronic depression [39]. Recent meta-analyses have been reviewed the opinion of a reduction of volume in the amygdala among patients with depression; they concluded there’s not a volume reduction in amygdala [13] and argued that inconsistency across studies can be related to either technical variables, such as different anatomical boundaries adopted for selecting the area, or slice thickness, or biological variables, such as administration of pharmacological treatment or age or illness duration [1]. Even though structural imaging did not demonstrate a consistent structural alteration in the amygdala among patients with depression, the structure is still considered to have an important implication in the illness. This assumption has been robustly confirmed by functional, rather than structural, imaging studies which reported an altered activation of amygdala during affective tasks and a “retuning” of its activity after successful pharmacological treatment [9].
Cortical alterations have been observed in temporal and frontal lobes in depressive disorder. Prefrontal and orbitofrontal cortex reduction of volume, along with anterior cingulate cortex, have been reported in geriatric as well as recent-onset depression. Specifically, lower gray matter (GM) gyrification index has been observed in prefrontal and anterior cingulate cortex. These alterations have been reported by longitudinal studies to undergo a progressive process, faster than normal aging degeneration [29]. Nevertheless, findings on prefrontal areas are sometimes inconclusive [52], particularly for anterior cingulate [11], though its functional implication in depression has been robustly reported [81]. Heterogeneity of results should take into account differences in illness severity or time course, gender, anatomical variability, or tracing protocols.
Reduction of volume in temporal lobe for severe patients or reduced cortical thickness in recent onset [99] has been reported for some studies even though some results report unaltered volumes [76]. Probably, lateralization (left volume reduction more than right) and duration of illness affect the results [102]. Furthermore, white matter (WM) alterations (hyperintensities) have been observed, mostly in geriatric population [86, 87] but also in first episode, specifically in the corpus callosum. Finally, basal ganglia alterations, such as volume reduction or signal hyperintensities, have been reported, in the caudate, the putamen, globus pallidus, and in the neighboring thalamus.
5.3.1 Features of Treatment Response
Along with structural alterations related to depressive disorder, imaging studies have also investigated structural features of treatment response in terms of effect of single compounds or predictors of treatment response. Only few studies have observed the effect of antidepressant administration over time in a longitudinal design. Increased hippocampal volume and cortical thickness in the rostral middle frontal gyrus, orbitofrontal cortex, and inferior temporal gyrus in remitters and decreased volume or thickness in these regions in nonremitters were observed after 6–12 months of observation after multiple antidepressant treatments [80]. First episode patients with major depression (MD), medication naive, under short-term administration of fluoxetine, showed larger volumes in prefrontal areas – middle frontal gyrus and orbitofrontal cortex – compared to healthy controls, along with treatment response [54]. Sertraline administration determined enlargement of volume of dorsolateral prefrontal cortex (DLPFC) among responsive MD patients, compared to healthy controls [93]. Also, among drug-naive patients with MD, duloxetine treatment determined amelioration of inferofrontal areas shrinkage, observed at baseline, in comparison to healthy controls; enlargement came along with clinical improvement but did not reach the same volume of healthy controls [58]. In a different sample, 12 weeks administration of duloxetine was not associated with structural changes but with increased functional connectivity in prefrontal areas [30]. These data underlie the implication of prefrontal area structural remodeling during response to treatment [59], probably associated with a cortical modulatory effect on emotional neuronal circuits. Indeed, larger prefrontal area volumes and cortical thickness, in middle frontal gyrus, DLPFC, or anterior cingulate, have been reported as predictors of better clinical outcome, in prospective studies [19, 24, 35, 48, 55, 60, 63]: these studies assessed patients only ones and test structural imaging predictors of treatment response after few weeks of treatment. In a sample of geriatric patients, larger anterior cingulate orbitofrontal cortex volume predicted lower levels of apathy after 12 weeks of escitalopram administration [105] or clinical remission [83]. Among geriatric populations, a predictive negative effect of white matter hyperintensities on treatment outcome was observed [36, 78, 90, 94], particularly for cingulate fibers [96].
Several predictive reports have focused on hippocampus, in terms of identification of markers of treatment response, because it is considered as a core area for the pathophysiology of depressive disorder and its volumetric reduction has been correlated to illness severity and poor outcome [73, 100]. Larger hippocampal volume predicted treatment response to antidepressants [30, 64, 91], whereas nonremitting patients had lower volumes at baseline and did not undergo an increase of volume at follow-up, whereas those who respond had larger volume compared to baseline [29]. Hippocampus volumetric response to antidepressant treatment, along with clinical improvement, might be more evident in female patients, suggesting a possible gender effect.
5.4 Summary
The growing recognition of personalized medicine in psychiatry not only by clinicians and researchers but also politicians, patients, and the pharmaceutical industry reflects the need for an improvement of diagnostic processes and therapeutic algorithms. Biomarkers are supposed to improve and expand the diagnostic options and enable patient-tailored treatments. Despite great efforts, imaging biomarkers are still not sensitive nor specific enough to allow the application in clinical practice. However, they offer valid landmarks on the neurobiological substrates of psychiatric diseases. The main imaging findings suggest an altered brain network that encompasses the anterior cingulate, the prefrontal cortex, and the hippocampus in major depression. Few studies have addressed the effects of pharmacological treatment on brain structures. Most of them offer hints of features of treatment response rather than explaining differential neural effects of single compounds. Antidepressants, mainly the SSRIs, might exert neuroprotective effects, which determine a reduction of hippocampal and prefrontal cortex shrinkage, probably through an activation of neuromodulatory factors like BDNF. Altered levels of BDNF belong also to the most replicated findings in the field of blood-based biomarkers. Lower BDNF plasma levels are observed in depressed patients and normalize after successful treatment. Even though the development of blood-based biomarkers in psychiatric disorders is still in its infancy, there are some very promising findings observed within inflammatory pathways and the HPA axis. In both systems, there is accumulating evidence that drugs specifically targeting these systems may be beneficial for the treatment of depression, but only in these patients who have marked alterations detected by the respective biomarkers.
Acknowledgments
PB was partially supported by grants from the Italian Ministry of Health (RF‐2011‐02352308), the BIAL Foundation (Fellowship#262/12), and the IRCCS “E. Medea” (Ricerca Corrente).
Related posts:
Pharmacogenetics of the Efficacy and Side Effects of Antidepressant Drugs
Genomic Studies of Treatment Resistance in Major Depressive Disorder
Pharmacogenetics of Serious Antipsychotic Side Effects
Pharmacogenetics of Mood Stabilizers
Pharmacogenetics of the Efficacy of Antipsychotic Drugs in Schizophrenia
Practical Application of Pharmacogenetics of Antipsychotic, Antidepressant, and Mood-Stabilizing Drugs
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
