Original Authors: Brian C. Lund, Pharm.D., BCPP, Paul J. Perry,
Ph.D.
Latest Revisers: Mitchell Barnett, Pharm.D., M.S., Paul J. Perry,
Ph.D.
Creation Date: April 2001
Last Revision Date: February 2005
Peer Review Status: Internally Peer Reviewed
Introduction
The topic of late-life of depression is extraordinarily broad. There are numerous important areas including pathophysiology, presentation, diagnosis, treatment efficacy and safety. The purpose of this chapter is to provide an introduction to the breadth of clinically important issues facing mental health providers in the setting of depression in late-life.
Aging and the Treatment of Major Depression
Taylor et al (1995) used data from Iowa Medicaid program to estimate medical care utilization and expenditures associated with depression among Iowa Medicaid recipients aged 65 and younger. Taylor used computerized claims from a 10% random sample of continuously enrolled non-institutionalized non-elderly patients. A total of 357 depressed patients with ICD-9 codes of major depression for either single episode, recurrent episode or MDD non-specified were contrasted to a population of 9,484 non-depressed Iowa Medicaid patients. Medical care utilization expenditure when stratified by age (18-24, 25-34, 35-44,45-54, and 55-64 years) was significantly greater for all age groups of depressed patients versus the controls. Additionally, the medical care expenditures increased with age. This is expected since individuals develop more medical conditions over time, in addition, the medical conditions tend to become more severe. However medical care costs for depressed 55-64 year olds increased disproportionately greater compared to their non-depressed counterparts. Based on the trend in the curve you’d expect this cost to be around $8,000. Instead, Taylor observed costs to be $13,543 versus $4,418 in the controls. These results suggest an interaction between the cost of treating major depressive disorder and age. As patients age, taking care of major depressive disorder becomes increasingly complex, probably as it interacts with medical illnesses.
Prevalence of Depression in the Elderly
Liebowitz et al (1997) reported a 1-2% rate of major depressive disorder among community dwelling elderly populations. Additionally, a 2% dysthymia rate and a 13-27% estimate of minor depression (subsyndromal depression) were noted. Placement in a nursing home causes dramatic change in the prevalence rates. Minor depression rates can reach 50%, while new cases of major depression after a year in the nursing home may reach 13%. More recently, Steffens et al (2000) reported an estimate of major depression in the elderly of 4.4% in females and 2.7% in males. It is unclear if this is an actual increase or if this reflects an increased ability to detect major depressive disorder in this population.
Risk Factors
The risk factors for late-life depression include female gender, unmarried status, especially widowed, having stressful life events, medical comorbidities and lack of a social support network. In general these risk factors are similar to early life major depressive disorder (Liebowitz et al 1997).
Consequences of Unrecognized or Inadequately Treated Late-life Major Depressive Disorder
Perhaps the most significant outcome related to major depressive disorder in late life is suicide, as evidenced by the fact that older Americans comprise about 13% of the US population, but account for 18% of all suicides and as a group, are the most successful suicide completers (Bruce et al 2004). Further, after the age of 65 there is a rapid increase in the suicide rates reported among males. This rate doubles from the ages of 65 to 85. Furthermore, this is a global phenomenon. Pearson and Conwell (1995) reported suicide data from 37 international sites. In 36 of 37 sites, the greater than 74 year-old age groups had the highest suicide rates among males. In women the relationship with age is less clear. In only 25 of the 37 sites, the 75-84 year old group had the highest suicide rate for women. In the US, the highest rate in women was between those 45-55 years old, suggesting some sort of relationship between menopause and suicide.
This begs the question, why do suicide rates in males increase so dramatically over the age of 65 years? The conventional thought is that the rates of suicide increase because as patients get older, they become more medically ill, they experience a greater degree of functional impairment, and their quality of life is impaired. It is these alleged effects that produce an increase in suicide rates to nearly 60/100,000 in males ≥ 80 years old (MMWR 1996). However, studies examining the relationship between functional impairment and medical illness and suicide provided some further insights. The North American Follow Back Study followed 9,181 individuals aged 65 or older (Grabbe et al 1997). The investigators contrasted disease states and assessments of activities of daily living, etc between patients who died from suicide versus individuals dying from natural causes. It was thought that if there was a strong relationship between the degree of functional impairment or medical illness and suicide, the patients committing suicide would be more likely to have major medical illnesses or limitations in activities and daily living. However, most of the conditions had similar rates. Further, patients committing suicide were actually statistically less likely to have a history of stroke or have diabetes or limitations in activities of daily living compared to the natural caused decedents.
Conwell et al (2000) investigated this relationship in 42 suicide completers and 196 matched controls. Among suicide completers, there were greater rates of depression, a greater percentage of patients seeing a psychiatrist in the past year, a greater history of antidepressant use, and more significant medical illnesses including worse physical health. However, a sub-analysis among all the depressed patients yielded different findings, specifically among the suicide completers, approximately 85% were depressed. Comparison of the make-up of the depressed controls to the depressed suicide completers found no significant differences in physical health or overall functioning. Thus when controlling for depression, the relationship between medical illness and functional impairment in suicide becomes less clear. It appears the most critical issue in treating depressed patients with suicidal depression is to reverse the depression with antidepressants as quickly as possible. Of the suicide completers, only 29% were being treated with an antidepressant at the time of suicide, despite the fact that more than 70% of suicide completers saw a primary medical doctor in the month prior to suicide. Suicide completers are not patients outside of the medical system but are being seen on a regular basis by a primary medical doctor and their depressive symptoms or suicidal ideation are going unchecked or unobserved. Once again it has been demonstrated that recognition and treatment of depression is the most critical factor in the prevention of suicide.
Additionally, there appears to be conceptual barriers in the treatment of depression and suicidal ideation in the geriatric population among health care professionals as demonstrated by Uncapher and Arean (2000) in a survey of suicidal ideation among primary medical doctors. In the survey, primary care investigators administered a clinical vignette of a patient with clear signs of depression and suicidal ideation. They had two forms of the vignette – one in which the patient age was described as 38 years old and a second with a patient who was identified as 78 years old. Other than the age difference, there were no other differences in the vignettes. The vignettes were randomly distributed to primary care physicians in conjunction with a survey that asked questions about assessment and clinical treatment for the “patient.” The physicians who reviewed the 78 year old patient case were much more likely to describe the suicidal ideation in the patient as being rational or normal. They were less likely to recommend treatment for the patient.
Lastly, Bruce et al (2004) conducted the Prevention of Suicide in Primary Care Elderly: Collaborative Trial (PROSPECT), a randomized study designed to examine the combined components of a protocol algorithm recommending citalopram use, and the utilization of depression care managers. The study encompassed 20 primary care clinics (10 assigned to intervention, 10 assigned to usual care) in the Eastern US. Clinic patients were screened for depression via telephone, and subsequently followed for two years (May 1999- August 2001). Rates of suicidal ideation in patients seen at the intervention clinics declined faster than seen at patients utilizing the usual care clinics starting at 4 months and continuing throughout the two years of the study. Among patients reporting suicidal ideation, resolution was faster among intervention patients; further, intervention patients had a more favorable course of depression in both degree and speed of symptom reduction. The study highlights the positive effects of interventions designed to screen and treat depression in the elderly, and the effectiveness of SSRIs and a care management approach in treating depression in the elderly.
In summary, suicide rates increase with age after age 65, particularly in males, and appear to be related primarily to depression. Since the majority of these patients have contact with the health care system in close proximity to the suicide event, improvement in physicians’ ability to assess and treat major depressive disorder in late-life is in order. Rutz et al (1989) demonstrated that more intensive efforts in screening for and treating major depressive disorder in late-life impact the suicide rate. The investigators used the Swedish island of Gotland as their study site. The intervention consisted of two-day educational program focussing on the diagnosis and treatment of depressive disorders including late-life depression. The intervention was targeted at primary care physicians. It was initiated in 1983 and repeated again in 1984. The primary care physicians' participation rate was approximately 90%. Prior to the initiation of the suicide intervention program in 1983 the suicide rates in mainland Sweden versus the island of Gotland were nearly identical. However, beginning in 1983 when the intervention was initiated and continuing on through 1984 to 1985, the suicide rates in Gotland decreased from 25/100,000 to 7/100,000. During the same period of time, the suicide rate on mainland Sweden remained unchanged at 25 to 26/100,000. These data strongly suggest that the intervention program had a strong impact on suicide rates among this population. However it is important to note that a follow-up in 1988 demonstrated a slight increase in Gotland suicide rates. Thus continued attention and continued focus and education to primary physicians are required to maintain the effect.
A meta-analysis of antidepressant efficacy studies in late-life depression demonstrated that the different classes of antidepressants are equally effective and more effective than placebo (SSRI 58%, TCA 63%, and placebo 27%) in the treatment of depression (Mittman et al 1997). A later randomized, placebo controlled trial (Allard et al 2004) found rates of efficacy and side effects for venlafaxine to be statistically equal to citalopram in geriatric outpatients with MMD.
Generally these results are difficult to interpret because of the strict exclusion criteria used in clinical trials of antidepressants. The trials generally exclude patients with significant medical illness, which is an important issue to when extrapolating these studies for use in treating patients, since the prevalence of major medical disorders, especially significant mental illness disorders, is common in late-life. However a recent randomized, double blind placebo controlled trial (Sheikh et al 2004) of 752 patients found sertraline to be efficacious in reducing symptomatology, regardless of the presence of comorbid medical illnesses, and overall superior to placebo.
The majority of clinical trials in late-life depression focus on individuals who have a life-long history of major depressive disorder. This has important implications in the treatment of late-life depression in the clinical setting. It seems reasonable that a patient aged 65 or 75 who presents with a major depressive disorder with a life-long history of major depressive disorder may be different from the same aged patient who presents with major depressive disorder for the first time. This difference in presentation may have implications for treatment. If there is a differing underlying pathophysiology involved in these two processes, it may be unreasonable to assume that antidepressants are going to work equally well in late-onset patients.
Major Depressive Disorder and Cardiovascular Disease
There have been numerous studies examining the relationship between major depression and cardiovascular disease. A majority (51%) show positive associations between depressive symptoms and mortality, however 23% are unequivocal, while the remainder (26%) show different rates for men and women (Gump et al 2004). Frasure-Smith et al (1993) studied 222 consecutively admitted patients with acute myocardial infarction (MI). The age in this population was approximately 60 years old (range of 24-88 years). Individuals were interviewed and screened for major depressive disorder within 10 days of their MI presentation. Major depressive disorder criteria were met by 35 of the 222 (16%) of the patients, which suggested an interaction between major depressive disorder and cardiovascular disease. Further, patients with a diagnosis of major depressive disorder were 3.5 times more likely to be dead at 6 months than non-depressed patients; however it is unknown from the study design whether major depression put patients at risk for an MI, or whether the MI lead to the development of depression. The relationship was maintained after controlling for underlying cardiovascular disease severity; eliminating the chance that depression was showing up simply as a marker for severity of cardiovascular illness. As expected, the diagnosis of congestive heart failure was also identified as a risk factor. Interestingly, a diagnosis of major depression was as important a risk factor as congestive heart failure. In addition, there were important implications for the treatment of depression in the post-MI setting, as only 3 of the 35 patients identified with major depressive disorder were treated. Therefore, it is unknown whether more aggressive treatment of major depressive disorder post-MI would remove major depression as a risk factor for mortality at 6 months.
A recently published study (Gump et al 2004) utilizing data from The Multiple Risk Factor Intervention Trial (MRFIT) found depressive symptoms associated with an increase in certain types of mortality. The MRFIT study, consisted of 12,866 men between the ages of 35 and 57, enrolled between 1973 and 1975 at 22 centers in 18 US cities. Men were eligible for study inclusion if they were without evidence of coronary heart disease (CHD) at screening, but had above average risk of CHD because of factors including high cholesterol levels, smoking, or hypertension. The MRFIT study offered benefits over previous CHD depression studies, namely 1) MRFIT offered a much longer study time frame (7 years), with an 18 year follow-up, 2) the study provided a more reliable assessment of risk factors, e.g., smoking, cardiac disease, by including 6 yearly assessments throughout the trial, 3) there was nearly 100% ascertainment of mortality, and 4) the study utilized the Center for Epidemiologic Studies Depression Scale (CES-D) score, in which 11,216 men completed. Subjects were subsequently divided in quintiles based on their depression scores. The median length of follow-up for the 11,216 men was 18.4 years. During this time, 1,788 cardiovascular (CVD) and 1,909 non-cardiovascular deaths occurred. Using a linear trend analysis, the unadjusted hazards ratio was higher for each quintile of depression (HR=1.05; 95% CI=1.03-1.07) for all cause mortality, as well as CVD mortality (HR=1.06; 95% CI=1.02-1.10). Specifically, stroke was associated with a significantly higher risk for men with higher CES-D scores (HR=1.17; 95% CI=1.01-1.07). After the addition of covariates, higher depression quintiles were still associated with an increased risk of all cause mortality (HR=1.03; 95% CI=1.01-1.06); CVD mortality (HR=1.05; 95% CI=1.01-1.08); and stroke mortality (HR=1.20; 95% CI=1.07-1.34), but not non-CVD mortality (HR=1.02; 95% CI=0.98-1.05). Further, the study found the magnitude of the effect of depressive symptoms (quintile rank) on stroke mortality to be larger than that of smoking or cholesterol, but less than age and hypertension.
A study looking at depression and CVD exclusively in postmenopausal women (Wassertheil-Smoller 2004) examined women enrolled in the Women’s Health Initiative (WHI) study, a cohort of 93,676 multiethnic women aged 50-79 enrolled in 40 centers from 1993 to December of 1998. Utilizing the CES-D scale, the authors found a high overall rate of depression (15.8%). Further, depression was significantly related to risk factors for CVD, as well as CVD. Utilizing a Cox proportional hazards model to determine the risk of death among women without prior CVD disease while controlling for demographic factors, life style (smoking, exercise, etc), and comorbid factors, the authors found an increased risk associated with depression for both all cause (RR=1.32; 95% CI=1,16-1.52) and CVD (RR=1.50; 95% CI=1,10-2.03) mortality. Also noteworthy, was the observation that of the (20,130) women classified as depressed, only 16.4% were taking an antidepressant medication.
A meta-analysis (Rugulies 2002) of 11 studies (n=730-7,894, mean follow-up=3-37 years) found that depression predicted the development of CHD Relative Risk (RR=1.64; 95% CI=1.63-4.43). A second meta-analysis of 20 studies (n=52-2,432, mean follow-up=3 months-10 years) found that depressive symptoms increased the risk of mortality in CHD patients. Specifically, the risk of CHD depressed patients dying within 2 years of assessment was over twice as high (OR=2.24; 95% CI=1.37-3.60). Further, the presence of depression became a stronger risk factor after two years out (OR=2.61; 95% CI=1.53-4.47). Taken together, these results suggest that depression places healthy patients at risk for CHD, and CHD patients at risk for increased mortality.
The aforementioned Frasure-Smith study provides an illustration of some of the different hypotheses and relationships between major depression and illness. The first hypothesis is of course the null hypothesis that there is no relationship. There are certainly disease states where this is true. However, among cardiovascular patients and patients with other diseases including cancer, etc., there appears to be a relationship. The question becomes what is the nature of the relationship? One hypothesis is that depression occurs as a reaction to being medically ill. This may be as either an emotional or physical reaction to the medical illness. The Frasure-Smith data potentially support this hypothesis in that having a myocardial infarction placed patients at risk for becoming depressed.
A second hypothesis is the opposite hypothesis that depression causes the medical illness. More commonly it is thought that depression does not necessarily cause medical illness but rather exacerbates or worsens the prognosis in medical illness. This effect was demonstrated in the prior study of myocardial infarction in that patients with a diagnosis of depression had higher rates of mortality at 6 months following the MI. Additionally, Cole and Bellavance (1997) conducted a meta-analysis of eight studies of 265 depressed inpatients ≥ 60 years old. They concluded that elderly medical depressed inpatients have a very poor prognosis. The recovery rate among these patients is low and the mortality rate high.
A third hypothesis is that depression occurs through the same biologic mechanisms as the underlying disease. Vascular depression is an example. This hypothesis originates from the observation of the high frequency of major depressive disorder associated with hypertension, diabetes, coronary artery disease, stroke, etc. This hypothesis revolves around the impression that the vascular lesions occurring in the peripheral system that are associated with medical illnesses such as hypertension, stroke, etc., also occur within the brain. These vascular lesions lead to the production of depressive symptoms. The hypothesis has several important implications for medical treatment. First, it is unknown whether aggressive screening and treatment with for medical illnesses with antihypertensive agents, cholesterol lowering agents or antiplatelet drugs will also reduce the risk or prevent vascular depression. Secondly, if depressive symptoms can stem from central vascular lesions, a mechanism that probably differs from patients with life-long depressions, can we expect traditional antidepressant treatment to produce equivalent clinical response for vascular depression?
An example of this important treatment implication is in post-stroke depression (major depression or minor depression and a baseline HAMD score ³ 12) (Robinson et al 2000). In this randomized controlled trial, the investigators treated 104 depressed (DSM-IV) and non-depressed acute stroke patients with either nortriptyline up to 100 mg per day, fluoxetine 40 mg per day or placebo. Non-depressed patients were included in this study to determine whether administering an antidepressant to non-depressed post-stroke patients would improve outcomes and recovery. Among the nortriptyline treated depressed patients, the HAMD scores were reduced from 22.5 at baseline to a mean of 9.0 at endpoint. In contrast, patients treated with fluoxetine went from a baseline of 20.4 to an endpoint of 18.5, suggesting very little therapeutic impact from fluoxetine. The placebo patients decreased from a baseline of 17.5 to an endpoint of 12.2. Therefore it was concluded that nortriptyline was more effective than either fluoxetine or placebo in the treatment of post-stroke depression. The results of this study reinforce the important implication that patients with late-onset depression, in this case post-stroke, may have a differential response to traditional antidepressants.
Major Depressive Disorder and Dementia
The relationship between depression and dementia in late-life is complex. Several studies have demonstrated a history of depression to be an independent risk factor for dementia, particularly of the Alzheimer type (DAT) (Reding et al 1985, Jorm et al 1991, Buntinx et al 1996, Steffens et al 1997, Wetherell et al 1999). It has been proposed that physiologic changes induced by or associated with depression may predispose an individual to the subsequent development of dementia (Lupien et al 1994). However, the majority of case-control studies have noted a temporal relationship between the age of onset of depressive illness and the subsequent development of dementia (Jorm et al 1991, Wetherell et al 1999). That is, depression tends to be a significant risk factor only when it occurs within 10 years of the onset of dementia. Other studies have found that late-onset depression (> 45-50 years) is more predictive of dementia than early-onset depression (Alexopoulos et al 1993a, Buntinx et al 1993, Steffens et al 1997). These findings support an alternative hypothesis that late-life depression may occur as a prodrome of dementia.
Additional studies have found that elderly patients who demonstrate cognitive impairment while depressed are more likely to develop an irreversible dementing illness. Follow-up data on a series of 44 elderly patients (mean age 76.5 years) with depressive pseudodementia has been reported (Kral and Emery 1989). All patients were successfully treated for depression with cognitive symptoms returning to premorbid levels. With continued follow-up every 6 months over a period of 4-18 years (average of 8 years), 39 of the 44 patients (89%) developed DAT. In another series, 183 consecutive geriatric patients (³ 60 years) with depression were evaluated for dementia (Alexopoulos et al 1993a). Patients were considered to have late-onset depression if their index episode occurred at 50 years of age or older. The percentage of late-onset depression cases was greater in patients with concurrent dementia, 83%, versus 66% in patients with depression alone. Late-onset depression was also predictive of DAT, independent of apolipoprotein e4 allele status (Steffens et al 1997). Based on these studies, it appears that late-onset depression is predictive of the development of subsequent dementia. The risk for developing DAT for patients with late-onset major depressive disorder with no e4 alleles present had the same three-fold risk ratio as no major depressive disorder with 2 e4 alleles. It is possible that the depression occurs as a result of the same pathophysiologic process that later produces dementia.
Depressive symptoms also commonly occur after the onset of clinically evident dementia. Although the DSM-IV considers depression as a subtype of dementing illnesses, many clinicians and investigators believe that is possible to diagnose major depressive disorder (MDD) and dysthymia in these patients (Katz 1998). The diagnosis of an independent depressive disorder becomes difficult in the demented patient. Many symptoms used to diagnose depression, including impaired concentration and cognition, irritable mood, psychomotor retardation, and loss of interest in daily activities, are also observed in a dementing illness (American Psychiatric Association, 1994). The loss of motivation and initiative commonly observed in patients with dementia is often termed apathy. While apathy has traditionally been considered a component of depression, there is data to suggest that apathy and depression may occur as separate entities in dementia populations (Marin et al 1993, Levy et al 1998). The apathy spectrum includes decreases in interest, motivation, spontaneity, affection, enthusiasm, and emotion (Marin 1991, Levy et al 1998). In contrast, the affective spectrum includes symptoms such as depressed mood, sadness, worthlessness, hopelessness, and suicidal ideation (Levy et al 1998). It has been proposed that affective symptoms should be emphasized, and motivational symptoms minimized, when establishing a diagnosis of major depression in a demented patient (Forsell et al 1993, Levy 1998, Katz 1998).
The relationship between apathy and depressive symptoms has been contrasted between patients with DAT and patients with MDD (Marin et al 1994). In the depression group, the mean apathy score was 39.8 on the Apathy Evaluation Scale (AES) and the mean depression score was 20.1 on the Hamilton rating scale for depression (HAMD). However, the mean apathy and depression scores in the DAT group were 45.2 and 5.4, respectively. These data support the observation that significant apathy symptoms may occur as a component of MDD. Furthermore, a high level of apathy can be observed in patients with DAT (similar to that observed with MDD), in spite of low levels of depressive symptoms. When evaluating such studies it is important to consider that DAT has a progressively deteriorating course with a wide range of symptom severity. The symptomatology of a newly diagnosed patient with mild DAT will differ drastically from that of a patient with severe DAT. Many of the depression assessment studies in patients with dementia do not give a clinical indication of the symptom severity of the study population. One study looking at disturbances in mood (depression) versus motivation (apathy) stratified patients by dementia severity (Forsell et al 1993). The severity categories included non-demented, questionable, mild, moderate, and severe dementia. Interestingly, the mean depression scores steadily increased from the non-demented to mild severity groups, but dropped off dramatically in patients with moderate and severe dementia. Conversely, mean apathy scores were low in non-demented and questionable patients, but rose rapidly in the mild group with a gradual increase in moderate and severe dementia groups. Therefore, disturbances in affect may be more common and more apparent in early stages of dementia, but apathy predominates in later stages.
The separation of depressive symptoms into components of affect versus apathy has important implications in the treatment of dementia. In 1997, the American Association for Geriatric Psychiatry, the Alzheimer's Association, and the American Geriatrics Society jointly published a consensus statement regarding the diagnosis and treatment of Alzheimer disease and related disorders (Small et al 1997). The authors suggested that all patients with DAT and depressive symptoms be considered for pharmacotherapy, even if they fail to meet the criteria for a depressive syndrome. This approach was also endorsed in the American Psychiatric Association guidelines for Alzheimer's and other dementias of late-life (APA 1997). While endorsing the use of antidepressants in dementia patients, the authors also recognized the lack of clinical trial support for this recommendation (APA 1997). A few small placebo-controlled antidepressant trials have been performed in dementia populations (Passeri et al 1987, Reifler et al 1989, Nyth et al 1990, Nyth et al 1992, Fuchs et al 1993, Petracca et al 1996, Roth et al 1996, Lyketsos et al 2000). There are two studies assessing the efficacy of the tricyclic antidepressants, imipramine (Reifler et al 1989) and clomipramine (Petracca 1996). Both of these studies were essentially negative and found the TCA to be equivalent to placebo in the treatment of depression in DAT. However, since both drugs are highly anticholinergic thereby increasing confusion, memory difficulty and orientation problems the results were not surprising. Two studies utilized the SSRI, citalopram in the treatment of major depressive disorder in DAT and vascular dementia (DVT) (Nyth and Gottfries 1990 Nyth 1992). Unfortunately both of these studies involved a heterogeneous population of depressed and non-depressed patients. Thus their findings on the outcome measures are difficult to interpret. However the conclusion in both studies was that citalopram provided some benefit in the treatment of depression in the DAT but not the DVT patients.
There is now one well-designed study involving DAT patients with major depression (Lyketsos et al 2000). This 12-week study involved 22 outpatients with DAT and MDD. Inclusion criteria required a mini mental score (MMSE) of ³10 for participation in this study. (A score of 10 is typically used as a cut-off for severe dementia.) The mean baseline MMSE score was 15 to 16. These scores suggest that the population had moderate to possibly mild DAT. The study design involved randomized assignment to sertraline to 150 mg/day or placebo in a double-blind fashion. The mean peak dose of sertraline was 80 mg/d. The primary efficacy measure was a global rating determined by two psychiatrists. They rated the patient as either not improved, partially improved or fully improved. A clinical response was defined as either partial improvement or full improvement. Using this definition, 9 of the 12 patients assigned to sertraline were considered responders, 3 of these 4 responders and 6 partial responders in contrast to the placebo group in which 2 patients were considered responders, 1 a full responder and 1 a partial responder (p < 0.05). Investigators also used other assessments for depression including the Cornell scale for depression in dementia and the Hamilton depression scale. There was a significant improvement rate between groups on the Cornell scale but not the Hamilton scale in the study. Overall the study suggests that the SSRIs, specifically sertraline, are probably effective in the treatment of depression in the presence of mild to moderate DAT. However, replication of the study is necessary.
To date, only one controlled study has looked at treating depressive symptoms in late stage Alzheimer’s disease (Magai et al 2000). This study involved 31 female patients in late stage Alzheimer’s disease and involved mostly minor depression. Only 10% of their population met criteria for definite major depressive disorder. This is not a surprising finding among this population and probably reflects that their population expressed mostly apathy symptoms and minimal affective symptoms. Investigators used several outcome measures including the Cornell scale for depression and the knit brow face response. This response is somewhat controversial is presumed to reflect a level of emotional and physical distress on the part of the patient. The study involved randomized double-blind treatment with either sertraline 100 mg or placebo. The Cornell scale, which was the primary efficacy measure, improved significantly over time in both groups equally. The authors suggested this might have been expectation on the part of the raters that patients would improve or a reflection of the general waxing/waning nature of the symptoms. In addition the investigators considered the possibility that improvement was due to the increased attention and interaction the patients experienced because of participating in a clinical trial. While none of the outcome measures demonstrated a significant response or significant drug effect, there was a trend in the knit brow face response that favored drugs. However the interpretation of this finding is difficult. Even if there is a true drug response on this measure it’s unclear what is being measured and what is changing in the patient. In this study, the correlation between Cornell scale scores and the knit brow face response was only 0.2. This suggests that whatever the knit brow face response is measuring, it may not be depression. While it seems clear in the study that sertraline did not improve depressive symptoms or improve depression in this population, it does open the door for further research with SSRIs in this population. Perhaps the effect demonstrated on the knit brow face response is related to either decreased anxiety or other symptoms and there may be some benefits to administering the SSRI antidepressants in this population outside of depression.
Overall, it is difficult to compare these studies due to differences in type and severity of dementia, characterization of depressive symptoms, choice of antidepressant, and definition of response. In general, it appears that antidepressants can improve depressive symptoms in early-stage demented patients (APA 1997). Due to unfavorable adverse effect profiles, tricyclic antidepressants (TCAs) and monoamine oxidase inhibitors (MAOIs) should be considered second line agents in dementia populations. If depressive symptoms do improve, the duration of therapy is unknown. Inferring from geriatric depression in non-demented populations, some authors argue for the routine use of maintenance treatment. However, it could be argued that once the dementia becomes severe a trial off the antidepressant would be warranted. The APA guidelines also addressed the distinction between depression and apathy (APA 1997). Stimulants such as methylphenidate and other dopaminergic agents (e.g. bupropion, amantadine, and bromocriptine) may have a role in the treatment of apathetic symptoms, but clinical trials in patients with DAT and other dementia disorders are lacking (Marin et al 1995).
In summary, the occurrence of depressive symptoms in patients with dementia presents a challenge to clinicians. Based on treatment guidelines and limited clinical studies, such patients should be considered for antidepressant therapy even if they do not meet criteria for a depressive syndrome (Small et al 1997, APA 1997). In evaluating patients for pharmacotherapy, the severity of dementia and the spectrum of depressive symptoms must be considered. Current studies suggest that persistent affective symptoms (e.g. depressed mood), particularly in the early stages of dementia, are responsive to antidepressant therapy. However, apathetic symptoms (e.g. amotivation), particularly in moderate to severe dementia, may not be responsive to antidepressant therapy. In either case, if antidepressant therapy is to be initiated, the depressive symptoms that are being targeted must be carefully documented and monitored. Furthermore, the clinician must be committed to stop or change antidepressant therapy if the target symptoms do not improve. If successful, the duration of antidepressant therapy is unknown. Apathy symptoms probably do not respond to antidepressants as well as affective symptoms, but may respond to stimulants. Further research is needed to define the role of antidepressants and stimulants in the treatment of depressive symptoms in patients with dementia.
Hormones and Late-Life Depression
Among women, considerable effort has been expended to determine if a relationship exists between menopause and depression. However, it has been concluded that there really is insufficient evidence to maintain that menopause causes major depressive disorder (Nicol-Smith 1996). Despite this conclusion, there does appear to be a relationship between menopause and the presence of depressive symptoms or minor depression. This is also reflected in assessing the impact of hormone replacement on depressive symptoms. In general, patients receiving hormone replacement therapy (HRT) have more lower depression scores and fewer and more mild depressive symptoms. However it is well known in these types of studies the women who receive HRT differ from women not receiving these agents. Therefore it’s unclear whether the effect on depression scores is a true effect of HRT (Zweifel and O’Brien 1997). Also, it has been found in several studies that estrogen use, as primary antidepressant treatment in women with major depression is generally ineffective. This has important implications for treatment. In perimenopausal or recently postmenopausal woman who present only with depressive symptoms, the evidence suggests that the use of HRT may help depressive symptoms. However if the patient meets criteria for a major depressive episode, there is no evidence to suggest that placing the individual on HRT will be effective, instead the patient should be treated with a traditional antidepressant. It has also been suggested in the literature that HRT modulates or effects response to antidepressants. In a retrospective analysis of data from a fluoxetine versus placebo study in women over the age of 60 with a minimum Hamilton depression score of 16, the overall treatment response rate did not differ between fluoxetine (38%) versus placebo (28%) (Schneider et al 1997). However, a response was observed after separating women into those ERT users, and non-users. Specifically, fluoxetine displayed a significantly greater response rate than placebo (42% vs. 17%) among those receiving ERT, but not for those who did not receive ERT (36% vs. 29%). The authors hypothesized that ERT augmented the fluoxetine response; however this conclusion is not supported by the raw data. Analyzing the distribution of response rates between estrogen user and non-users, the most reasonable conclusion from this study was that ERT blunted the placebo response.
A lack of relationship between ERT and antidepressant response is supported in the follow-up study by Amsterdam et al (1999). The effectiveness of fluoxetine 20 mg/d for 8 weeks was contrasted in depressed women receiving ERT to those not receiving ERT. The minimum age of the women in the studies in this retrospective analysis was 45 years. They found no benefit for ERT. In fact, the trend was that patients receiving ERT had lower response rates in general (62% vs 70%, HAMD change ≥ 50%) remission rates (41% vs 56%, HAMD ≤ 7). In men, there have been few studies investigating the treatment of major depressive disorder with hormones. One randomized study in younger patients administered DHEA at 90 mg/day or placebo for 6 weeks in depressed patients with a mean age of 44. In this study, 5 of 11 DHEA patients versus none of the placebo patients responded (Wolkowitz et al 1999). This is statistically significant difference suggesting at least some efficacy for the use of hormones in treating major depressive disorder in males.
Safety Issues in the Treatment of Late Life Depression?
The impact of medications in elderly populations cannot be over-emphasized. Many elders are taking numerous medications, in fact a recent study of chronically ill elderly VA patients found patients taking a mean of 12.5 ±4.8 prescription and OTC meds (Kaboli 2004). It has been estimated that in patients greater than 70 years of age, as many as 1/6 of all hospitalizations in this population are drug-induced. Therefore it is important to reinforce that geriatric psychopharmacology cannot be considered in a vacuum. Furthermore, antidepressants cannot be administered indiscriminately in this population. Elderly patients have significant pharmacokinetics and pharmacodynamics differences that worsen with age. The most notable, of course, is reduced renal or hepatic clearance. In addition, elderly patients are likely more susceptible to adverse effects from a pharmacodynamic standpoint. The occurrence of polypharmacy is the norm rather than the exception. This increases the risk of clinically significant drug interactions. The tricyclic antidepressants caused a number of adverse effects that are significant in elderly populations to the point where these agents are considered generally contraindicated for any reason in elderly populations (Beers et al 1997). In addition, tricyclics also have important adverse effects on cardiovascular systems and in general should be avoided in patients with cardiovascular disease.
The most worrisome effect is the proarrhythmic effect of these agents. There is evidence that administration of tricyclic antidepressants to patients with cardiovascular disease increases their risk for sudden cardiac death (Moir et al 1971). In general the SSRIs are much safer in cardiovascular disease than TCAs. One potential exception that is somewhat surprising is the effect of SSRIs on fall risk. As TCAs are known to cause orthostatic hypertension, it’s not surprising that these agents have been associated with an increased risk of fall. However inspection of comparative studies in fall risk between SSRIs and tricyclics suggest these two antidepressant classes exhibit similar risks of fall (Sleeper et al 2000).
The effect of SSRIs on clotting is another important issue in late-life depression treatment. The SSRIs, specifically fluoxetine, has been demonstrated to modify markers for platelet stickiness including platelet factor 4 and beta-thromboglobulin. These platelet stickiness markers are increased in patients with major depressive disorder and also in patients with cardiovascular disease, especially after an AMI. However there are two treatment studies comparing nortriptyline versus paroxetine that have demonstrated that paroxetine significantly decreases these markers of platelet stickiness toward normal levels. In contrast, the nortriptyline treated patients had no change in their platelet stickiness markers even with resolution of depression (Musselman et al 1996, Pollack et al 2000). It appears that the SSRIs may have some additional benefits in treating major depression in cardiovascular disease mediated by an improvement in these platelet stickiness markers. However this effect is a double-edged sword in that any medication that can reduce platelet stickiness can also increase the risk for significant bleeding. The association between SSRIs and significant bleeding events was analyzed in a case control design (de Abajo 1999). The investigators found that SSRIs (fluoxetine, fluvoxamine, paroxetine, citalopram, sertraline, trazodone, clomipramine) were significantly associated with the risk of clinically significant bleeding and reported a relative risk of 3.0 (95% CI 2.1-4.4). In contrast, antidepressants devoid of serotonergic activity (nortriptyline, desipramine, protriptyline, trimipramine maprotiline, and amoxapine) were observed to have a relative risk of 0.8 (95% CI 0.2-2.4), which suggests they are not a risk factor for significant bleeding. Also of interest, antidepressants with moderate levels of serotonergic activity, including amitriptyline, doxepin, and imipramine, showed a moderate risk of 1.4 (95% CI 1.1-1.9). This dose response relationship indicates a relationship between the degree of serotonergic activity and the risk for bleeding. This finding is not surprising since the platelet stickiness factors are mediated and controlled through a serotonergic mechanism. An additional important finding in this study was the relationship with SSRIs and NSAIDs. The relative risk for NSAID use in this study is 3.7. What is interesting is patients receiving both an NSAID and an SSRI had a relative risk of a clinically significant bleed of 15.6. These data suggest a potential interaction between SSRIs and NSAIDs that can significantly increase bleeding risks. This stands to reason from a theoretical standpoint, as these 2 classes of medications have different mechanisms of action related to increased bleeding risk.
Despite the risk of bleeding, the high rates depression in patients with ischemic heart disease have historically been very problematic for clinicians, as there were no antidepressants found to be safe and effective in this population. Glassman et al (2002) conducted a randomized, double-blind, placebo-controlled trial of sertraline in 369 patients with major depressive disorder. Patients had been hospitalized due to a myocardial infarction (74%) or unstable angina (26%). There was a 2-week placebo run-in period, the result of which was that the sertraline-treated patients received drug an average of 1 month after the acute cardiac incident. Another limitation was that patients with other life-threatening medical conditions were excluded. Regardless of these limitations, this trial was important. Patients received flexible-dose sertraline (50-200 mg) or placebo for 24 weeks. The primary outcome was left ventricular ejection fraction. With regard to this measure, sertraline showed no benefit. However, while the HAM-D showed no significant differences between the groups, response rates on the CGI-Improvement scale favored sertraline in the total sample (67% vs. 53%). The CGI-Improvement response rates also favored sertraline treated patients in a subset with at least one prior episode of depression (72% vs. 51%) and in patients with more severe MDD (78% vs. 45%). Thus, sertraline appears to be a relatively safe and effective medication for treating MDD in the face of ischemic heart disease. It is unclear exactly how early sertraline could be initiated after an MI, but initiation 1-month post-MI is probably safe and may even be conservative. One final concern in the treatment of these patients is that many cardiac medications are metabolized by cytochrome P450 enzymes. The potential drug interactions should be assessed carefully when deciding whether to initiate antidepressant treatment in patients with unstable heart disease.
REFERENCES
Alexopoulos GS, Meyers BS, Young RC (1993a). Geriatric depression: age of onset and dementia. Biol Psychiatry 34:141-5.
Alexopoulos GS, Meyers BS, Young RC et al (1993b). The course of geriatric depression with "reversible dementia": a controlled study. Am J Psychiatry 150:1693-9.
Alexopoulos GS, Meyers BS, Young RC et al (1997). 'Vascular depression' hypothesis. Arch Gen Psychiatry 54:915-22.
American Psychiatric Association (1994). Diagnostic and Statistical Manual of Mental Disorders, Fourth edition. Washington, DC. American Psychiatric Association
American Psychiatry Association (1997). Practice guideline for the treatment of patients with Alzheimer's disease and other dementias of late life. Am J Psychiatry 154(May suppl):1-39.
Amsterdam J, Garcia-Espana F, Fawcett J et al (1999). Fluoxetine efficacy in menopausal women with and without estrogen replacement. J Affective Disord 55:11-7.
Anonymous (1996). Suicide among older persons - United States, 1980-1992. MMWR 45:3-6.
Barth J, Schumacher M, Herrmann-Liungen C (2004). Depression as a risk factor for mortality in patients with coronary heart disease: a meta-analysis. Psychosom Med 66:802-813.
Beers MH (1997). Explicit criteria for determining potentially inappropriate medication use by the elderly: an update. Arch Int Med 157:1531-6.
Gump BB, Matthews KA, Eberly LE et al (2004). Depressive symptoms and mortality in men results from the multiple risk factor intervention trial. Stroke 36:98-102.
Buntinx F, Kester A, Bergers J, Knottnerus JA (1996). Is depression in elderly people followed by dementia? A retrospective cohort study based in general practice. Age Aging 25:231-3.
Cole MG, Bellavance F (1997). Depression in elderly medical inpatients: a meta-analysis of outcomes. CMAJ 157:1055-60.
Conwell Y, Lyness JM, Duberstein P et al (2000). Completed suicide among older patients in primary care practices: acontrolled study. J Am Geriatrics Soc 48:23-9.
DeAbajo FJ, Garcia Rodriguez LA, MonteroD (1999). Association between selective serotonin reuptake inhibitors and upper gastrointestinal bleeding: population based case-control study. BMJ 319:1106-9.
Forsell Y, Jorm AF, Fratiglioni L et al (1993). Application of DSM-III-R criteria for major depressive episode to elderly subjects with and without dementia. Am J Psychiatry 150:1199-1202.
Frasure-Smith N, Lesperance F, Talajic M (1993). Depression following myocardial infarction: impact on 6-month survival. JAMA 270:1819-25.
Fuchs A, Hehnke U, Erhart C et al (1993). Video rating analysis of effect of maprotiline in patients with dementia and depression. Pharmacopsychiat 26:37-41.
Glassman AH, O’Connor CM, Califf RM, et al (2002). Sertraline treatment of major depression in patients with acute MI or unstable angina. JAMA 288:701-9.
Grabbe L, Demi A, Camann MA et al (1997). The health status of elderly persons in the last year of life: a comparison of deaths by suicide, injury, and natural causes. Am J Public Health 87:434-7
Jorm AF, Van Duijn CM, Chandra V et al (1991). Psychiatric history and related exposures as risk factors for Alzheimer's disease: a collaborative re-analysis of case-control studies. Int J Epidemiol 20(suppl 2):S43-7.
Katz IR (1998). Diagnosis and treatment of depression in patients with Alzheimer's disease and other dementias. J Clin Psychiatry 59(suppl 9):38-44.
Kral VA, Emery OB (1989). Long-term follow-up of depressive pseudodementia of the aged. Can J Psychiatry 34:445-6.
Lebowitz BD. Pearson JL, Schneider LS. et al (1997). Diagnosis and treatment of depression in late life. Consensus statement update. JAMA 278:1186-90.
Levy ML, Cummings JL, Fairbanks LA et al (1998). Apathy is not depression. J Neuropsychiatry Clin Neurosci 10:314-9.
Lupien S, Lecours AR, Lussier I, et al (1994). Basal cortisol levels and cognitive deficits in human aging. J Neurosci 14:2893-903.
Lyketsos CG, Sheppard JME, Steele CS et al (2000). Randomized, placebo-controlled, double-blind clinical trail of sertraline in the treatment of depression complicating Alzheimer's disease: initial results from the depression in Alzheimer's disease study. Am J Psychiatry 157:1686-9.
Magai C, Kennedy G, Cohen CI et al (2000). A controlled clinical trial of sertraline in the treatment of depression in nursing home patients with late-stage Alzheimer's disease. Am J Geriatr Psychiatry 8:66-74.
Marin RS (1991). Apathy: a neuropsychiatric syndrome. J Neuropsychiatry Clin Neurosci 3:243-54.
Marin RS, Firinciogullari S, Biedrzycki RC (1993). The sources of convergence between measures of apathy and depression. J Affect Disord 28:117-24.
Marin RS, Firinciogullari S, Biedrzycki RC (1994). Group differences in the relationship between apathy and depression. J Nerv Ment Dis 182:235-9.
Marin RS, Fogel BS, Hawkins J et al (1995). Apathy: a treatable syndrome. J Neuropsychiatry Clin Neurosci 7:23-30.
Mittmann N. Herrmann N. Einarson TR et al (1997). The efficacy, safety and tolerability of antidepressants in late life depression: a meta-analysis. J Affective Disord 46:191-217
Moir DC, Cornwell WB, Dingwall-Fordyce I et al (1972). Cardiotoxicity of amitriptyline. Lancet 2:561-4
Musselman DL, Tomer A, Manatunga AK et al (1996). Exaggerated platelet reactivity in major depression. Am J Psychiatry 153:1313-7.
Nicol-Smith L (1996). Causality, menopause, and depression: a critical review of the literature. BMJ 313:1516.
Nyth AL, Gottfries CG (1990). The clinical efficacy of citalopram in treatment of emotional disturbances in dementia disorders. Br J Psychiatry 157:894-901.
Nyth AL, Gottfries CG, Lyby K et al (1992). A controlled multicenter clinical study of citalopram and placebo in elderly depressed patients with and without concomitant dementia. Acta Psychiatr Scand 86:138-45.
Passeri M, Cucinotta D, DeMello M, et al (1987). Comparison of minaprine and placebo in the treatment of Alzheimer's disease and multi-infarct dementia. Int J Geriatr Psychiatry 2:97-103.
Pearson JL, Conwell Y (1995). Suicide in late life: challenges and opportunities for research. Int Psychogeriatr 7:131
Petracca G, Teson A, Chemerinski E et al (1996). A double-blind placebo-controlled study of clomipramine in depressed patients with Alzheimer's disease. J Neuropsychiatry Clin Neurosci 8:270-5.
Pollack BG, Laghrissi-Thode F, Wagner WR (2000). Evaluation of platelet activation in depressed patients with ischemic heart disease after paroxetine or nortriptyline treatment. J Clin Psychopharmacol 20:137-40.
Reding M, Haycox J, Blass J (1985). Depression in patients referred to a dementia clinic: a three year prospective study. Arch Neurol 42:894-6.
Reifler BV, Teri L, Raskind M et al (1989). Double-blind trial of imipramine in Alzheimer's disease patients with and without depression. Am J Psychiatry 146:45-9.
Robinson RG, Schultz SK, Castillo C et al (2000). Nortriptyline versus fluoxetine in the treatment of depression and in short-term recovery after stroke: a placebo-controlled, double-blind study. Am J Psychiatry 157:351-9.
Roth M, Mountjoy CQ, Amrein R et al (1996). Moclobemide in elderly patients with cognitive decline and depression: an international double-blind, placebo-controlled trial. Br J Psychiatry 168:149-57.
Rutz W, von Knorring L, Walinder J (1989). Frequency of suicide on Gotland after systematic postgraduate education of general practitioners. Acta Psychiatr Scandinavica.80:151-4.
Rugulies R (2002). Depression as a predictor for coronary heart disease: a review and meta-analysis. Am J Prev Med 23:51-61.
Schneider LS, Small GW, Hamilton SH et al (1997). Estrogen replacement and response to fluoxetine in a multi-center geriatric depression trial. Am J Geriatr Psychiatry 5:97-106.
Sheikh JI, Cassidy EL, Doraiswamy PM, et al (2004). Efficacy, safety, and tolerability of sertraline in patients with late-life depression and comorbid medical illness. J Am Geriatr Soc 52:86-92.
Sleeper R, Bond CA, Rojas-Fernandez C (2000). Psychotropic drugs and falls: new evidence pertaining to serotonin reuptake inhibitors. Pharmacotherapy 20:308-17.
Small GW, Rabins PV, Barry PP et al (1997). Diagnosis and treatment of Alzheimer disease and related disorders. JAMA 278:1363-71.
Steffens DC, Plassman BL, Helms MJ, et al (1997). A twin study of late-onset depression and apolipoprotein E4 as risk factors for Alzheimer's disease. Biol Psychiatry 41:851-6.
Steffens DC, Skoog I, Norton MC et al (2000). Prevalence of depression and its treatment in an elderly population: the Cache County study. Arch Gen Psychiatry 57:601-7.
Taylor T, Perry, PJ, Rohland B et al (1995). Medical care utilization and expenditures associated with depression in a Medicaid population. Presented to the Iowa Medicaid Drug Utilization Review Commission, April.
Uncapher H, Arean PA (2000). Physicians are less willing to treat suicidal ideation in older patients. J Am Geriatrics Soc. 48:188-92.
Wassertheil-Smoller S, Shumaker S, Ockene J, et al (2004). Depression and cardiovascular sequelae in postmenopausal women. Arch Intern Med 164:289-298.
Wetherell JL, Gatz M, Johansson B, Pedersen NL (1999). History of depression and other psychiatric illness as risk factors for Alzheimer disease in a twin sample. Alz Dis Assoc Dis 1:47-52.
Wolkowitz OM, Reus VI, Keebler A et al (1999). Double-blind treatment of major depression with dehydroepiandrosterone. Am J Psychiatry 156:646-9.
Zweifel JE, O'Brien WH (1997). A meta-analysis of the effect of hormone replacement therapy upon depressed mood Psychoneuroendocrinol 22:189-212.