The aging of the US population and the high prevalence of depression among the elderly attest to its public health significance.1 The prevalence of depression is 8% to 16% in community-dwelling older persons, and the prevalence of mixed anxiety-depression is 1.8%.2 Depression complicates medical illnesses and their management, and it increases health care use, disability, and mortality.3 This article focuses on the recent research data on diagnosis, etiopathogenesis, treatment, and prevention in unipolar, bipolar, psychotic, and subsyndromal depression.
Differential diagnosis and treatment of depressive spectrum disorders
Depressive symptoms in the elderly are not adequately captured by the criteria of major depression as outlined in DSM-III through DSM-IV-TR.4 There is evidence from epidemiological, longitudinal, neuroimaging, cognitive, and genetic studies that supports the idea of a continuum of depressive disorders, ranging from the very mild subthreshold to major unipolar and bipolar depression.5 Minor depression affects up to 50% of residents in long-term—care facilities and up to 25% in the primary care setting, and it is associated with considerable discomfort, disability, and risk of morbidity, as well as excessive use of non—mental health services.6,7 The term “subsyndromal depressive spectrum disorder” describes a group of states operationally defined by the presence of 2 or more concurrent symptoms of depression for most or all of the time, for at least 2 weeks, associated with evidence of social dysfunction, in patients who do not meet criteria for a diagnosis of minor depression, major depression, and or dysthymic disorder. (Table 1)8,9
Bipolar affective disorder is fairly common in elderly persons, with a prevalence of 0.1% to 0.43% in the United States. However, between 10% and 20% of geriatric patients with mood disorders have bipolar disorder, as do 5% of those admitted to geropsychiatric inpatient units.10 While late-onset bipolar disorder is relatively rare, recurrence of remitted disease frequently occurs in late life.11 Subramaniam and colleagues12 identified higher cerebrovascular risk in elderly patients with late-onset bipolar disorder compared with patients who had the early-onset disorder. These patients often present a tremendous treatment challenge because of substantial medical comorbidity and age-related variations in response to therapy. Unfortunately, the management of geriatric bipolar disorder has been relatively neglected compared with the condition in the younger population.10,11,13
Lithium, divalproex sodium, carbamazepine, lamotrigine, atypical antipsychotics, and antidepressants have been found to be beneficial in the treatment of elderly patients with bipolar disorder. In an open trial, flexible dose aripiprazole with a mean daily dose of 10.2 mg was effective as an adjunct medication.14 The NIMH-funded Geriatric Bipolar (Geri-BD) trial is ongoing and will provide additional information about the use of lithium and valproic acid with and without risperidone in older adults with bipolar mania. Major depression with psychotic features occurs in 15% of community sample populations and in 25% of inpatient mixed-age sample populations. Among geriatric inpatients, the rate may reach 45%.15 Unfortunately, the appropriate diagnosis of psychotic depression is missed in about 30% of emergency department admissions.16 Despite recommended antidepressant/antipsychotic combination treatment, the use of antipsychotics in combination therapy for psychotic depression remains low (5%).17
Alexopoulos and colleagues18 and Krishnan and colleagues19 proposed the concept of vascular depression on the premise that cerebrovascular disease, documented on structural neuroimaging (CT and MRI) studies, may be causally related to geriatric depressive syndromes. Subcortical ishemic depression affects clinical presentation, long-term outcomes, and response to antidepressant therapy.20 The Figure delineates the pathways to geriatric depression.
In a study of patients with DSM-III-R criteria for depression, silent cerebral infarctions were reported in 65% of 20 patients.21 Recent observations suggest that in elderly persons, smaller brain volumes and white matter hyperintensities may provide complementary, albeit autonomous pathways to late-life major depressive disorder.22,23 Taylor and colleagues24 observed that greater anterior white matter lesion volumes were associated with higher diffusivity and lower anisotropy in the white matter of the dorsolateral prefrontal cortex and with higher diffusivity of the internal capsule and white matter lateral to the anterior cingulate cortex. The neuropathological correlates of small strokes or signal hyperintensities on MRI or CT are diverse and represent ischemic vascular change.25
Although there are no specific treatments for vascular depression, general vascular prevention approaches using lipid-lowering drugs, fish oil, or aspirin are applicable. The recent study by Jorge and colleagues26 examined the efficacy of repetitive transcranial magnetic stimulation (rTMS) for treatment of vascular depression in 92 patients who were randomly assigned to receive active or sham rTMS of the left dorsolateral prefrontal cortex. Response rates to rTMS were negatively correlated with age and lower frontal gray matter volumes. This approach applies to subcortical ischemic depression, a diagnosis corresponding to the “vascular depression” hypothesis.
Medical and cognitive comorbidities
There has been an exponential growth in studies in the interface between primary care and psychiatry in the past 10 years.27,28 Studies in the medically ill usually report the negative impact of depression on the rates and speed of recovery as well as the overall outcome of the medical illness (coexisting depression produces excess disability and increases the cost of care).29,30 Depression can be successfully managed in medical settings with depression care strategies. Evidence-based practice can be imported into primary care settings to accelerate recovery and reduce the burden of depressive symptoms, suicidal ideation, and caregiver distress.1
Late-life depression is associated with cognitive deficits, but there is some debate over the profile and significance of the deficits. Performance in episodic memory, verbal fluency, information processing, executive function, and visuospatial ability have been reported to be lower in patients who are depressed than in nondepressed volunteers.31,32 The level of executive dysfunction at baseline has been shown to predict outcomes in terms of increased vulnerability to relapse and resistance to treatment.33 After treatment, some of the deficits are resolved, but the level of performance on executive and memory tasks often does not reach the premorbid level, which suggests that the depression and the residual cognitive deficits are expressions of an underlying and related brain dysfunction. Patients with minor depression show functional and cognitive deficits similar to those demonstrated by patients with major depression.34
Several studies have shown that late-life depression and even depressive symptoms are risk factors for mild cognitive impairment and Alzheimer disease.35-37 For example, semantic clustering—an executive strategy for learning a list of words—mediated executive dysfunction in late-life depression.34 These results support the conceptualization of major and nonmajor depression as a gradual continuum that is associated with increasing cognitive deficits but with the degree of difference dependent on the specific domain of cognitive ability.
Etiology and pathogenesis
Depression is a truly multifactorial disorder.38 Given the importance of genetic and environmental risk factors, aging processes, neurodegenerative and cerebrovascular disease processes, and medical comorbidity, the integration of basic and clinical neuroscience research approaches is critical for understanding the variability in the course of the illness, as well as the development of prevention and intervention strategies that are most effective.39
The results of association studies in behavioral genetics have been inconsistent. However, genetic studies are adding to our understanding of pathophysiology, risk factors, treatment response, and adverse-effect profile. Findings from a study by Taylor and colleagues24 indicate that Val66Met polymorphism of the brain-derived neurotrophic factor gene is associated with cognitive and neuroimaging changes. In a large study of older adults with depression and controls, the depressed subjects were more likely to be Met66 allele carriers than were the comparison subjects (38.8% vs 24.4%). Met66 allele carriers have almost double the odds of having geriatric depression than do Val66 allele homozygotes. Steffens and colleagues40 recently reported that older age, lower Mini-Mental State Examination score, and apolipoprotein E e4 allele were each correlated with gray matter lesion volume in patients with depression. Results from the study by Whyte and colleagues41 showed that platelet activation is increased in elderly depressed patients, especially those with the 5-HTTLPR l/l genotype.
In addition, data from a clinical study of geriatric depression by Lavretsky and colleagues42 showed that dopamine transporter genotype, but not serotonin transporter genotype, was associated with executive dysfunction and preferential response to methylphenidate. Lenze and colleagues43 reported that the 5HTR1A (-1019) G allele is associated with increased depressive symptoms after hip fracture, which in turn accounts for poorer functional recovery. Lotrich and colleagues44 reported an interaction between early paroxetine concentration and the 5-HTTLPR genotype on symptomatic improvement over 12 weeks.
Investigations using structural neuroimaging (CT or MRI) continue to document the importance of cerebrovascular disease. Teodorczuk and colleagues45 demonstrated in a pan-European multicenter study of 639 older adults that white matter changes seen on MRI predate and are associated with the development of depressive symptoms. Sheline and colleagues46 recently reported on the differential significance of the localization of white matter hyperintensities in geriatric depression. In a comparison of patients with depression and controls, greater hyperintensities were seen in the following white matter tracts: the superior longitudinal fasciculus, frontooccipital fasciculus, uncinate fasciculus, extreme capsule, and in-ferior longitudinal fasciculus. In depressed patients, but not in controls, volumes of 3 of these regions correlated with executive function; whole brain white matter hyperintensities correlated with executive function; whole brain white matter correlated with episodic memory, processing speed, and executive function; and whole brain gray matter correlated with processing speed.
In support of these findings, Yuan and colleagues47 looked at the first episode of late-life depression and found that fractional anisotropy in white matter was lower in patients than in controls at the right superior frontal gyrus, left inferior frontal gyrus, left middle temporal gyrus, right inferior parietal lobule, right middle occipital gyrus, left lingual gyrus, right putamen, and right caudate. These results suggest that the white matter integrity of the whole brain was disrupted in first-episode remitted geriatric depression.
Alexopoulos and colleagues48 used diffusion tensor imaging to test the hypothesis that older adults with depression who do not attain remission have microstructural white matter abnormalities in corticostriatolimbic networks, which are implicated in geriatric depression. Subjects who did not achieve remission had lower fractional anisotropy in multiple frontal limbic brain areas, including the rostral and dorsal anterior cingulate, dorsolateral prefrontal cortex, genu of the corpus callosum, white matter adjacent to the hippocampus, multiple posterior cingulate cortex regions, and insular white matter, relative to those who achieved remission. In addition, lower fractional anisotropy was detected in the neostriatum and midbrain as well as select temporal and parietal regions.
There have been attempts to use structural neuroimaging to identify correlates of depressive spectrum disorders. Lavretsky and colleagues49 reported that depression was associated with a reduction in gray matter volumes in the orbitofrontal cortex, but a highly prevalent (in the elderly) syndrome of apathy was associated with the reduction in the gray matter volume in right anterior cingulate. Results from another study demonstrate that the lacunar volume in the deep white matter was associated with symptoms of apathy, anergia, anhedonia, and depression in older persons with and without cognitive impairment.50 The findings indicate that patients with minor depression present with specific neuroanatomical abnormalities that are comparable with the major depression group but significantly different from controls.51 Neuroanatomical abnormalities may represent one aspect of a broader neurobiological diathesis to late-life mood disorders.
Functional neuroimaging correlates have been identified for late-life depression and include widespread reductions in glucose metabolism and cerebral blood flow on positron emission tomography (PET) and single photon emission computed tomography (SPECT) scans. Sheline and colleagues52 used PET scans and [18F]altanserin, a ligand with high affinity for the 5-HT2A receptor, to examine the relationship between 5-HT2A receptor density and geriatric depression. Depressed subjects had less hippocampal 5-HT2A receptor binding than controls. Depressed subjects not previously treated for depression had less hippocampal 5-HT2A receptor binding than previously treated subjects. It may be that earlier medication treatment provides a compensatory up-regulation of the 5-HT2A receptor.
Meltzer and colleagues53 explored the role of the serotonin type 1A (5-HT1A) receptor as a regulator of treatment response, particularly the 5-HT1A autoreceptor in the dorsal raphe nucleus. They used [11C]WAY 100635 and PET scans in the dorsal raphe nucleus and prefrontal cortex and observed significantly diminished [11C]WAY 100635 binding in depressed patients compared with controls that was correlated with pretreatment Hamilton Depression Rating Scores in the depressed cohort.
Postmortem tissue can help clarify the pathways to depression, especially if the postmortem data can be integrated with antemortem neuroimaging findings. Autopsies in 6 persons with a clinical diagnosis of late-life depression, in fact, had dementia; the brains were neuropathologically consistent with dementia, primarily of the Alzheimer type.54
In general, the pharmacological treatment of nonpsychotic major depressive disorder in elderly persons is only partially successful: only about half of older adults with depression improve with initial antidepressant monotherapy.55 Many factors may predict a more difficult-to-treat depression, including coexisting anxiety, low self-esteem, poor sleep, and a high coexisting medical burden.56 Being aware of these and other predictors of a difficult-to-treat depression gives the clinician more reasonable expectations about a patient’s likely treatment outcome. Getting well and staying well is the goal of patients; thus, clinicians should treat to remission of depression, not merely to response.56
Maintenance treatment with the same regimen that has been successful in relieving the depression strongly improves the patient’s chances of remaining depression-free. In a recent trial of maintenance treatments, patients 70 years or older with major depression who had responded to initial treatment with paroxetine and psychotherapy were less likely to have recurrent depression if they received 2 years of maintenance therapy with paroxetine.57 Monthly maintenance psychotherapy did not prevent recurrent depression. This is unexpected given earlier reports of improved response to the combined treatments, and thus, needs to be replicated.
Randomized trials have compared different strategies for management of nonmajor depression, mostly in primary care settings. In a recent meta-analysis, Pinquart and colleagues58 integrated the results of 89 controlled studies of treatments for acute major depression (37 studies) and other depressive disorders (52 studies conducted with mixed diagnostic groups, including patients with major depression, minor depression, and dysthymia). Psychosocial treatments had larger effect sizes for less severe forms of depression than pharmacological treatments.
Two collaborative trials have addressed the efficacy of antidepressant treatment in patients with depressive syndromes treated in primary care settings. The Prevention of Suicide in the Primary Care Elderly Collaborative Trial (PROSPECT) study, which was supported by the NIMH, and theImproving Mood: Promoting Access to Collaborative Treatment for Late Life Depression (IMPACT) study, which was supported by the Hartford Foundation, evaluated the effectiveness of current models by nurse health specialists.59,60 Results showed that the PROSPECT intervention substantially reduced suicidal ideation at rates comparable to those seen in earlier efficacy studies and in specialty mental health clinical settings.61 The IMPACT intervention was effective in improving symptomatic and functional outcomes for older primary care patients with major depression or dysthymic disorder.60 The studies on collaborative treatments are important and the components of collaborative care can be transported into clinical practice emphasizing early diagnosis and treatment of depression in primary care settings.
Esther, a 72-year-old, presents with complaints of insomnia, weight loss, and sadness over the loss of her husband 2 years ago. She is socially isolated, lives alone, and rarely leaves her house. Esther is able to cook simple meals. She still drives short distances, but on several occasions she has forgotten her appointments.
Her medical problems include hypertension, coronary artery disease, hyperlipidemia, and stress incontinence. When her primary care physician prescribed 10 mg of nortriptyline at night to treat the depression, incontinence, insomnia, and poor appetite, she became agitated, confused, and disoriented. She was taken to the emergency department where delirium was diagnosed. The patient was admitted to the geriatric psychiatry unit.
A CT scan of her head showed some age-related generalized atrophy and old microvascular disease in the deep white matter and periventricular but no acute stroke. The ECG showed QTC-interval prolongation but no acute ischemic changes.
Esther was given 0.5 mg of haloperidol for agitation and belligerence, and nortriptyline was discontinued. Over the next 5 days, her confusion and agitation subsided, but she showed mild extrapyramidal signs caused by the haloperidol, which was discontinued by day 5.
Neuropsychological testing revealed deficits in executive functioning and psychomotor slowing; however, dementia was ruled out. Esther was given 15 mg of mirtazapine at night: she became more animated, less withdrawn, participated in group activities, and enjoyed the company of other patients. Her appetite and sleep improved as well. There were no changes from baseline on subsequent ECGs, and Esther was discharged with recommendations for admission to a day treatment program and individual psychotherapy or grief counseling with structured activities to increase social stimulation, as well as recommendations for local resources and support groups.
The antidepressant selection that was ultimately successful for Esther considered her vulnerability to anticholinergics and the central and cardiac adverse effects of nortriptyline that led to the patient’s confusion and hospitalization. The selection of mirtazapine addressed her symptoms of depression, as well as incontinence, insomnia, and poor appetite without exposing Esther to the detrimental anticholinergic adverse effects.
Despite recent advances in our understanding of pathophysiology of geriatric depression and the development of treatment approaches similar to that in younger adults, the available treatments are only modestly effective. Our current understanding of the pathophysiology of geriatric depression complicated by medical, vascular, and neurodegenerative processes should lead to the development of more personalized treatment approaches based on individual risk factors. For example, treatment of depression should be different for those who are depressed due to complicated bereavement than for those who become depressed at the beginning of Alzheimer disease. So far, recommended treatments have not differed and have followed the drug development paradigm for younger adults: no consideration has been given to differences in pathophysiology in aging adults, which poses challenges to future research.
General considerations when choosing a drug regimen for older adults should include age-related changes in pharmacokinetics and pharmacodynamics of the drugs and in the structure and function of the brain and other organs that lead older adults to be more sensitive to both therapeutic and toxic effects of psychotropic medications. Such changes necessitate lower doses and slower dose titration. Although the effective dose may be similar to that for younger adults, the initial dose of psychotropic medications for older adults should be one-third to one-half of that for younger adults and titrated slowly. The groups that may be particularly vulnerable to the adverse effects and drug-drug interactions include frail older adults with multiple medical problems or those with cognitive impairment, dementia, or other neurodegenerative disorders or who have had a stroke. Lower doses of drugs and augmentation strategies with other drugs and supplements should be considered (eg, cognitive enhancers and neuroprotective agents in patients with cognitive impairment; or psychostimulants or dopaminergic agents in patients with apathy or Parkinson disease). Table 2 (Table restricted. Please see print version for content.) lists the recommended dosages of antidepressants for older adults.