It can be difficult to achieve remission in individuals with late-life depression (LLD) and they often require aggressive treatment. This challenge is in part due to age-related vascular changes that are common in LLD. Successful antidepressant treatment involve changes across affective, cognitive, and default mode networks. We hypothesize that in LLD, vascular disease adversely affects response to antidepressants by disrupting connectivity of these networks. The primary goal of this project is to characterize how focal vascular damage affects regional connectivity and response to antidepressants. Based on past work and pilot data, we a priori focus on the cingulum bundle and uncinate fasciculus. These key fiber bundles connect frontal, temporal, and cingulate regions involved in cognition and affective responses. Our central hypothesis is that ischemic damage to the cingulum bundle and uncinate fasciculus contributes to structural and functional connectivity deficits of those tracts. This results in a disconnection effect that alters the function of connected regions. In turn, this increases the risk of a poor response to antidepressants.

Our approach is to enroll up to 130 adults over age 60 years with a diagnosis of Major Depressive Disorder. Subjects will complete clinical evaluation, cognitive testing, and MRI/functional MRI (fMRI) sessions, including an fMRI emotional oddball task that includes attentional and affective components. Participants will be stratified by cerebral lesion severity and randomized in a 2:1 ratio to a double-blinded 8-week trial of escitalopram or matching placebo. Those who do not remit will transition to an 8-week trial of open-label bupropion, an antidepressant with a different mechanism of action. This will allow us to determine if different and distinct circuit deficits affect response to antidepressants with different mechanisms of action while also accounting for the placebo response.

Individuals with late-life depression (LLD) experience high levels of disability, mortality, and poor responses to antidepressants. The MRI hallmark of 'vascular depression' in this population is significant ischemic white matter lesion (WML) severity, a finding associated with poor antidepressant outcomes. Despite observations of diffuse white matter disease in LLD, we propose in our "disconnection hypothesis" that focal damage to fiber tracts negatively affects the function of connected regions, resultantly contributing to cognitive deficits and clinical symptoms such as depression severity and negativity bias. Focal WMLs may also reduce the likelihood of an antidepressant response when the specific antidepressant used acts on neurotransmitter systems projecting through the damaged fiber tract. This implies that the effect of tract damage on clinical response may differ between drugs with different mechanism of action.

We propose that the cingulum bundle (CB) and uncinate fasciculus (UF) are key tracts in LLD as they are components of cognitive, affective and default mode networks and contain monoamine neurotransmitter projections. Supporting our model, past work implicates CB and UF deficits in LLD and our new pilot data associates tract damage with poor antidepressant response.

In a cohort of up to 130 depressed elders we will test our central hypotheses: a) focal CB and UF WMLs disrupt connectivity and function of connected regions and contribute to cognitive deficits and affective symptoms; and b) antidepressants acting on neurotransmitter systems projecting through these tracts will be less effective.

Overall Study Design: After obtaining informed consent, we will assess for eligibility. Individuals who meet eligibility criteria will complete neuropsychological testing and a one-hour magnetic resonance imaging (MRI). Subjects will then be randomized in a 2:1 ratio to receive either blinded escitalopram or identical placebo. After 8 weeks of study drug, they will be assessed for remission. Those whose depression has remitted will end their study participation. Those who remain symptomatic will be transitioned to open-label bupropion for 8 weeks, after which their participation will end. We will work with participants on plans for clinical care after the study and will offer two additional visits to facilitate that transition.

Specific Aim 1: To characterize the effect of cingulum bundle (CB) and uncinate fasciculus (UF) WMLs on tract connectivity, function of connected regions, and the cognitive and affective presentation of LLD.

Hypothesis 1: Greater CB WML volume is associated with a) reduced resting-state connectivity of frontal, temporal, and cingulate regions and b) deficits in attention, memory and processing speed.

Hypothesis 2: Greater UF WML volume is associated with a) reduced resting state functional connectivity between frontocingulate and medial temporal regions, and b) greater depression severity and negativity bias.

Exploratory Hypothesis: Greater CB WML volume is associated with failure of anterior and posterior default mode network nodes to deactivate during attentional components of the fMRI task. Greater UF WML volume is associated with greater medial temporal reactivity during the functional MRI task.

Specific Aim 2: To determine whether deficits in tract structural / functional connectivity predict nonremission to antidepressant treatments and if these relationships vary based on antidepressant mechanism of action.

Hypothesis 3: Nonremission to escitalopram will be predicted by: a) greater WML volume in the CB and UF, and b) reduced resting functional connectivity between structures connected by the CB and UF. Greater WML severity and reduced functional connectivity in these tracts will not significantly predict response to placebo.

Hypothesis 4: Nonremission to bupropion will be predicted by a) greater WML volume in the UF but not CB, and b) reduced resting functional connectivity deficits between structures connected by the UF but not CB.

Exploratory Aims: Expl. Aim 1) To determine if specific cognitive measures may serve as markers of focal tract WML damage. Expl. Aim 2) To use whole-brain multimodal imaging approaches to examine how connectivity differences in other brain regions may also predict nonremission to antidepressants.

Inclusion Criteria:

1. Age 60 years or older.
2. Current diagnosis of major depressive disorder (DSM-IV-TR), single episode, recurrent or chronic, without psychotic features, as detected by Mini-International Neuropsychiatric Inventory (MINI) and clinical exam.
3. Minimum MADRS score ≥ 15.
4. Mini-Mental State Exam ≥ 24.
5. Fluent in English.

Exclusion Criteria:

1. Current or past diagnoses of other Axis I psychiatric disorders, except for generalized anxiety disorder (GAD) symptoms occurring during a depressive episode
2. History of alcohol or drug dependence or abuse in the last three years
3. History of developmental disorder or Intelligence Quotient score < 70
4. Presence of acute suicidality
5. Acute grief (< 1 month)
6. Current or past psychosis
7. Primary neurological disorder, including but not limited to dementia, stroke, brain tumors, epilepsy, Parkinson's disease, or demyelinating diseases
8. MRI contraindications
9. Any physical or intellectual disability adversely affecting ability to complete assessments
10. Electroconvulsive therapy in last 6 months
11. Use of antidepressant medications or other psychotropic medications in the last 4 weeks (or the last 6 weeks for fluoxetine). Occasional use of benzodiazepines or non-benzodiazepine sedatives (such as zolpidem, eszopiclone, or zaleplon) during this period is allowable.
12. A failed therapeutic trial of escitalopram in the current depressive episode (defined as at least 6 weeks of treatment at a daily dose of 10mg or higher)
13. Known allergy or hypersensitivity to escitalopram or bupropion
14. Current or planned psychotherapy

Participants randomized in 2:1 drug:placebo ratio.

Adverse event data were collected during participants' duration in the study following randomization. This lasted for 8 weeks for the randomized, blinded escitalopram or placebo phase, and an additional 8 weeks for individuals progressing into the subsequent open-label bupropion phase.

Single-site study; some missing data for self-report other outcome data