|Home | About | Journals | Submit | Contact Us | Français|
Recognition of the significance of anxiety disorders in older adults is growing. The revision of the Diagnostic and Statistical Manual of Mental Disorders (DSM) provides a timely opportunity to consider potential improvements to diagnostic criteria for psychiatric disorders for use with older people. The authors of this paper comprise the Advisory Committee to the DSM5 Lifespan Disorders Work Group, the purpose of which was to generate informative responses from individuals with clinical and research expertise in the field of late-life anxiety disorders.
This paper reviews the unique features of anxiety in later life and synthesizes the work of the Advisory Committee.
Suggestions are offered for refining our understanding of the effects of aging on anxiety and other disorders (e.g., mood disorders) and changes to the DSM5 criteria and text that could facilitate more accurate recognition and diagnosis of anxiety disorders in older adults. Several of the recommendations are not limited to the study of anxiety but rather are applicable across the broader field of geriatric mental health.
DSM5 should provide guidelines for the thorough assessment of avoidance, excessiveness, and comorbid conditions (e.g., depression, medical illness, cognitive impairment) in anxious older adults.
It is estimated that by 2050, there will be two billion older adults globally (Kalache et al., 2005) with a corresponding increase in the number of older adults who suffer from anxiety disorders (Blazer, 2003; Schutzer and Graves, 2004). Recognition of anxiety in this age group has grown in recent years. Epidemiological studies indicate prevalence of approximately 10% in community-dwelling older adults (Beekman et al., 1998; Trollor et al., 2007; Gum et al., 2009; Byers et al., 2010), with higher rates in medical settings (Kim et al., 2001), and variability across geographic regions (Copeland et al., 1987) and ethnic groups (Diefenbach et al., 2009). The revision of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-TR; American Psychiatric Association, 2000) provides a timely opportunity to consider potential improvements to the diagnostic criteria for psychiatric disorders in older adults. The authors of this paper were invited to contribute to that process as members of the Advisory Committee to the DSM5 Lifespan Disorders Work Group. We provide recommendations based on recent research, suggestions for changes to the DSM-IV-TR (2000) criteria, and recommendations for conducting field trials that could improve the recognition and diagnosis of anxiety disorders in older adults.
The presentation of anxiety may be different in older versus younger adults (Wolitzky-Taylor et al., 2010). Developmental, neural, and cognitive processes all contribute to heterogeneity in the presentation of anxiety in later life, leading to challenges in identifying anxiety disorders in older adults. For example, avoidance and the presence of “excessive” anxiety, hall-marks of DSM criteria for anxiety disorders, may be challenging to detect in older people. Consequently, when anxiety disorders do appear in older individuals, they may not be recognized (Krasucki et al., 1999; Bryant, 2010) and are thus underdiagnosed and inadequately treated (Harman et al., 2002; Jeste et al., 2005). Thus, we present recommendations to improve the criteria and accompanying text in DSM5, with the ultimate goal of improving the recognition and treatment of anxiety disorders in older adults.
Heterogeneity increases as people age (Zarit and Zarit, 2006), and the optimal diagnostic scheme should alert clinicians to the fact that presentations of anxiety are also likely to differ with age (Bryant, 2010). Older adults may regard and report symptoms of anxiety differently than younger adults. In DSM-IV-TR (2000), diagnostic issues related to age are sometimes addressed under the headings, “Specific Culture, Age, and Gender Issues.” However, older age is only addressed in this category for one disorder (obsessive-compulsive disorder), whereas issues related to childhood are discussed for five anxiety diagnoses. Under the “Prevalence” heading in DSM-IV-TR, neither childhood nor older adulthood are addressed except for as age relates to specific phobias, and under the “Course” heading, discussion of older adulthood is absent (whereas childhood is represented by age-of-onset discussions for six diagnoses). An explicit discussion of older age in the DSM5 could (1) include a description of common sources of anxiety and worry in older adults; (2) emphasize the need to probe more actively for evidence of disproportionate functional consequences in the older population; (3) describe prevalence data; and (4) describe how symptoms and diagnostic status may be influenced by older adulthood.
One commonly hypothesized age-based difference is the tendency for older individuals to underreport psychological symptoms of anxiety and to overendorse somatic symptoms (Heimberg et al., 2004). However, these differences have not yet been subjected to rigorous empirical study (Flint, 2005; Wolitzky-Taylor et al., 2010), and some evidence suggests that anxious older adults report increased anxiety symptoms in the absence of increased somatic distress (Wetherell et al., 2010). Differences of this sort may be driven by the stigma associated with disclosure of psychiatric symptoms and a general lack of knowledge of mental health problems, in addition to actual phenomenological differences (Montorio et al., 2003; Wetherell et al., 2009). DSM5 could recommend obtaining reports from collateral informants (e.g., friends, family) when possible, which could improve the accuracy of this and other sensitive assessment domains.
There may also be age-related differences in those symptoms that differentiate clinical from subthreshold cases. For instance, Generalized Anxiety Disorder (GAD) among older adults is associated with less perceived control over worry as compared with controls (Montorio et al., 2003). Uncontrollability distinguishes older (Diefenbach et al., 2003; Wetherell et al., 2003) but not younger adults meeting criteria for GAD from subsyndromal cases (Andrews and Hobbs, 2010). Because older adults with GAD may more readily endorse “uncontrollability” than their younger counterparts, we recommend retaining this symptom in diagnostic criteria for use with older people.
Detection of impairment in work or relationships may be complicated by changes in social roles. This may enable avoidance of certain safe but anxiety-provoking situations (Bryant, 2010). Table 1 provides some age-specific examples of this type of maladaptive avoidance, which can be subtle and difficult to detect in older individuals.
The utility of categorical diagnosis based on specifically defined operational criteria may be limited, an issue perhaps best highlighted by the high comorbidity of anxiety and depression in later life (King-Kallimanis et al., 2009). Older adults might not be aware of the distinction between these mood states, leading to potential problems in assessment (Diefenbach et al., 2003). Alternatively, many older people experience coexistent anxiety and depression (Cassidy et al., 2005; Lenze et al., 2005) but may not necessarily meet criteria for a specific anxiety or depressive disorder as defined by DSM-IV-TR (2000), leading to diagnostic ambiguity. Additionally, this syndrome cannot be diagnosed if the individual has a history of another mood or anxiety disorder (e.g., Major Depressive Disorder, GAD). This criterion may function as an obstacle to the detection of this particular syndrome in older people.
This anxiety-depression syndrome is associated with a negative prognosis, including elevated symptom severity, diagnostic “shifting” from one disorder to another as opposed to symptom remission, suicidal ideation, increased autonomic symptoms, exacerbation of disability in the presence of medical conditions, and lower social functioning. However, very little is known about optimal treatment strategies (Cassidy et al., 2005; Lenze et al., 2005). There is also the suggestion that the pattern of co-occurrence of anxiety and depression varies according to the setting; prevalence of depression may be higher in nursing homes than in the community or in primary care, with stand-alone anxiety disorders virtually absent (Smalbrugge et al., 2005). Yet, in the community, the authors found that about 75% of the DSM-IV anxiety disorders were not comorbid with a DSM-IV depressive disorder. Longitudinal studies suggest that affective symptoms are often fluid and dynamic (Schoevers et al., 2005; Teachman, 2006), leading to a seemingly “blurred” boundary between categories (Beck et al., 2003). Thus, depending on when a person comes to clinical attention, he or she may receive an anxiety or depressive diagnosis, even though neither diagnosis alone adequately reflects the true nature of the problem. It has also been suggested that depression can be masked by anxiety, highlighting the need for careful assessment (King-Kallimanis et al., 2009; Grenier et al., 2010).
To increase sensitivity in detecting mixed anxiety-depression in the older people, DSM5 could include information on age-related differences in prevalence, use of longitudinal timelines, and additional probes (e.g., probing suicidal ideation even when depression accompanying anxiety is subthreshold; specific probes for detecting fluctuation of symptoms over time).
One barrier in diagnosing anxiety disorders is the fact that older adults may not readily endorse clinical criteria of impairment and distress due to unfamiliar terminology (Grenier et al., 2010), particularly including the failure to characterize the anxiety as “excessive.” One relevant study found that the prevalence of social phobia nearly doubled when ignoring the criterion of whether fear was considered to be excessive (Karlsson et al., 2009). If an older patient denies excessive anxiety or worry but indicates comparable cognitive activity (e.g., thinking too much, having concerns) or avoidance, the patient’s own descriptive words can be used in place of “excessive” during the remainder of the interview (Mohlman et al., in press). Questioning about how often the symptom occurs “compared to other people of your age” can be used, and also a greater emphasis on consequences and affective reactions to those consequences. Table 2 provides suggestions for conducting a thorough analysis of the excessive nature of worry. We also recommend that more examples of the functional domains of older adults (e.g., visiting grandchildren, volunteering, attending church services) be included consistently across diagnoses.
Anxiety disorders are prevalent in the context of medical illness; thus, accurate detection of anxiety disorders in this subgroup is particularly important. For example, prevalence of anxiety disorders in chronic obstructive pulmonary disease ranges from 10% to 19% in patients with stable disease, 9% to 58% in individuals recovering from a recent acute exacerbation, and 50% to 75% in those with severe disease (Chung et al., 2006; Todaro et al., 2007). Anxiety in the context of medical illness can have serious consequences by increasing risk of hospitalization (Maurer et al., 2008) and leading to poorer rehabilitation outcomes (Yohannes et al., 2000).
Presentation of an existing anxiety disorder can be changed by the presence of some medical problems, which can produce symptoms that mimic anxiety disorders (e.g., diabetes, thyroid disease; Thomas, 1995). Certain medical conditions are associated with panic attacks or other anxiety symptoms, most notably angina, tachycardia, obesity, stomach ulcer, and arthritis (Chou, 2010); thus, the presence of these conditions might lead to a suspicion that the anxiety is primarily due to the medical condition. The presence of a medical illness, however, does not preclude existence of co-occurring anxiety that can exacerbate the patient’s functional decline. For instance, panic attacks in an individual with chronic obstructive pulmonary disease could ensue from hypoxic episodes, yet still contribute to functional problems (e.g., agoraphobia). In the absence of careful questioning, anxiety symptoms can be obscured by polypharmacy or mimicked by medications used to treat medical conditions (e.g., corticosteroids; Smith et al., 1995).
The DSM currently excludes cases in which anxiety disorders may be “due to the direct physiological effects of a general medical condition,” and therefore may impose barriers to diagnosis and treatment. This is particularly problematic when patients are seen by non-mental health providers with less training in recognizing psychiatric disorders in patients. New-onset anxiety in the context of an acute physical disorder can be regarded as a physiological consequence of the physical disorder or its medications. Either chronic anxiety that clearly predates an acute physical disorder or acute anxiety developing in a person with chronic but stable medical illness suggests that the anxiety is not a direct physiological consequence of the medical condition (see Table 3). Although some have argued that sudden versus insidious onset is a useful distinction, there can be precipitating life events that bring about sudden onset of GAD (e.g., a medical test suggesting a serious health problem and subsequent worry about its many implications). Thus, this distinction, although important, is not sufficient on its own. Further research is needed to establish whether, for example, anxiety in the context of a new medication is characterized by a predominance of physiological over cognitive features of anxiety.
Furthermore, if all medical conditions were acute (as most are in younger adults), there would be a purpose in distinguishing anxiety as secondary to a medical condition (in which case the anxiety would be expected to resolve when the medical condition is treated) from stand-alone anxiety (in which case the anxiety needs to be treated separately). If the medical condition is chronic (as most are in older adults), then classifying anxiety as secondary to medical illness will probably result in the anxiety being less than optimally treated. In other words, the current exclusion criterion of “Psychiatric condition due to a general medical condition” is in question. However, this is not consistent with the non-etiologic philosophy underlying DSM diagnoses and often creates an artificial dichotomy between mental and physical conditions.
Finally, the relationship between anxiety and medical illness is likely to be bidirectional. In prospective studies, GAD was associated with higher rates of cardiovascular events after controlling for cardiac disease severity, medications, major depression, and other possible confounds (Martens et al., 2010). Panic disorder, posttraumatic stress disorder (PTSD), and anxiety disorder not otherwise specified also have been associated with increased risk of myocardial infarct independent of depression (Chen et al., 2009; Scherrer et al., 2010). Based on this ongoing debate, field trials should examine the implications of the current “due to General Medical Condition” exclusion on prevalence and presentation of disorders, especially in older adults. We recommend that the field trials consider the potential alternatives displayed in Table 4.
Defining anxiety in the presence of dementia is a complicated challenge unique to the field of geriatrics, and the best evidence suggests a bidirectional relationship between anxiety and cognitive performance (Beaudreau and O’Hara, 2009; Yaffe et al., 2010). Although DSM-IV-TR includes vascular dementia subtype “With Depressed Mood,” there is no analogous subtype for anxiety, which is commonly comorbid (Ballard et al., 2000). Dementia-specific criteria have been proposed (Starkstein et al., 2007) with GAD or an agitated state reported as the most common manifestation (Mintzer and Brawman-Mintzer, 1996). The literature also consistently reports that anxiety arising in the context of dementia is associated with depression (Ferretti et al., 2001).
The course of anxiety in dementia is complex. The temporal sequence of the onset of anxiety symptoms and cognitive decline is of value in identifying clinically significant anxiety. If the anxiety symptoms occurred some time before the cognitive decline, then this suggests a separate anxiety disorder. If anxiety symptoms and cognitive difficulties emerge concurrently, the anxiety may be part of the disease course that gives rise to the cognitive decline, or could represent fears or worry related to increasing cognitive difficulties. Anxiety symptoms can be a prodromal manifestation of dementia, so new-onset anxiety associated with mild cognitive impairment may be a “red flag” of a developing disease process. If patients with dementia have insight into their mental decline, then anxiety is likely to be present (Seignourel et al., 2008). However, as neural areas that govern insight are increasingly affected by the disease, anxiety is likely to reduce, which further clouds the diagnostic process. One heuristic is that certain types of anxiety that are triggered or associated with specific events or situations seem less likely to be caused by cognitive decline (e.g., PTSD, specific phobia). In fact, recent evidence suggests that a prior history of PTSD may increase the likelihood of developing dementia in old age (Yaffe et al., 2010), although intelligence may account for risk of PTSD and dementia separately (Pitman, 2010).
Symptoms and chronology of anxiety symptoms should be assessed in the presence of cognitive impairment or dementia, including awareness of cognitive impairment and its impact on anxiety. In more severe dementia, assessment of agitation and behavioral disturbance is more appropriate than probing cognitive aspects such as insight. Clinicians should also be aware that concerns about memory problems are common among the older people (Teachman, 2007) and do not necessarily suggest the presence of cognitive impairment.
One example of how DSM fails to capture age-specific variants of anxiety disorders is fear of falling, a common syndrome in community dwelling older adults, with estimates of 7–14% for moderate to severe fear with avoidance of multiple situations and activities (about 3% of community-dwelling older adults avoid leaving their homes or yards due to fear of falling; Arfken et al., 1994). This fear can be excessive relative to objective fall risk; for example, many who report fear have never experienced a fall and do not report any fall risk factors. Thus, an agoraphobia-like syndrome can emerge solely from excessive anxiety about falling (van Haastregt et al., 2008) rather than as a consequence of an actual fall (McCabe et al., 2006). The syndrome can limit mobility and contribute to functional decline and institutionalization (Zijlstra et al., 2007).
Fear of falling does not map well onto the DSM requirement of insight into excessiveness of the fear, which many older patients lack. For example, Gagnon et al. (2005) found that only 1 of 48 subjects with moderate or severe fear of falling considered their fear to be unreasonable, even though the fear frequently resulted in avoidance of activities, in some cases to the point of being housebound (Flint, 2005). Fear of falling also does not map well onto Agoraphobia Without History of Panic Disorder, which is conceptualized as a disorder related to panic attack symptoms. This syndrome is common and impairing and therefore should be encompassed more readily by the DSM.
Hoarding can be a refractory symptom of obsessive-compulsive disorder and subtype of obsessive-compulsive personality disorder that occurs more often and to a more severe degree in older than younger individuals (Teachman, 2007). Hoarding can be sufficiently life-threatening to be classified as elder selfneglect (Frost et al., 2000), and is associated with diagnoses of dementia, debilitating physical conditions, or serious mental illness (Snowdon et al., 2007), a fact that highlights the need for improved assessment strategies.
Epidemiological instruments in mental health may have low sensitivity in older adults because of insight and recall issues. We suggest that the key to successful use of psychiatric measures with older adults is in their administration. The following suggestions to improve sensitivity of assessment instruments are based on clinical observation and findings from empirical research (Mohlman et al., in press).
We recommend adequate representation from a wide range of settings, including
Field trials may opt to include cases or case series if recruiting homogeneous groups of older adults with less prevalent anxiety disorders (e.g., obsessive-compulsive disorder) is not feasible. We also recommend including subsyndromal groups of older adults for the sake of determining appropriate clinical thresholds, which may be more liberal than those applied to younger adults (Grenier et al., 2010).
Anxiety disorders are increasingly heterogeneous in older adults and may be experienced and expressed in ways that are different from their younger counterparts. In particular, it can be challenging to evaluate avoidance and the excessiveness of anxiety in later life—two features considered central to anxiety diagnoses. Older adults are also more likely to experience disorders at a subthreshold level yet still experience functional impairment, increased disability, impaired quality of life, and increased use of health services and medications. It is not surprising, then, that a categorical diagnostic system such as the DSM-IV-TR fails to address the difficulties in diagnosing presentations of anxiety that are phenomenologically and etiologically diverse. We have made recommendations that could largely be encompassed within the existing structure of the DSM. Additionally, we suggest that the text accompanying the criteria include more age-sensitive information and examples to provide clinicians with better guidelines for eliciting relevant information from patients, thus enabling them to make more accurate diagnoses. These changes should lead to better recognition and treatment of anxiety disorders in later life.
This work was partly supported by the VA HSR&D Houston Center of Excellence (HFP90-20).
Conflicts of interest
Dr Flint has received research grants from NIMH and CIHR, honoraria from Pfizer Canada and Lundbeck, and medications have been donated by Eli Lilly and Pfizer to his federally funded research. Dr Lenze has received research funding from Forest laboratories and is a consultant for Fox Learning Systems. Dr Stanley has received funding from NIMH and the South Central MIRECC. Dr Thorp is supported by a VA Career Development Award and has received additional funding from the Department of Defense and the VA. The opinions expressed reflect those of the authors and not necessarily those of the Department of Veterans Affairs or other affiliated institutions.