Assessment scales

In a recent systematic review of anxiety assessment scales used with adults aged 65 years and older in the published literature, Therrien and HunsleyReference Therrien and Hunsley¹² identified a total of 91 instruments. Of these, they concluded that twelve were in common use, but in only three was there sufficient psychometric evidence to support their use in this age group: the Beck Anxiety Inventory (BAIReference Kabacoff, Segal, Hersen and Van Hasselt¹³), the Penn State Worry Questionnaire (PSWQReference Beck, Stanley and Zebb¹⁴), and the Geriatric Mental Status Examination (GMSEReference Copeland, Kelleher, Kellett, Gourlay, Gurland, Fleiss and Sharpe¹⁵). In recent years, a few instruments to identify and measure anxiety have been validated or developed specifically for use with older adults. These include the ten-item Short Anxiety Screening Test (SAST) in medical in-patients and out-patients,Reference Sinoff, Ore, Zlotogorsky and Tamir¹⁶ the four-item FEAR in older adults attending primary care,Reference Krasucki, Ryan, Ertan, Howard, Lindesay and Mann¹⁷ the Worry Scale (WS),Reference Wisocki, Handen and Morse¹⁸ and the RAID (rating anxiety in dementia) in patients with dementia.Reference Shankar, Walker, Frost and Orrell¹⁹ Another recent development with promising psychometric properties is the self-report Geriatric Anxiety Inventory (GAIReference Pachana, Byrne, Siddle, Koloski, Harley and Arnold²⁰); this appears to perform well in both cognitively intact and cognitively impaired subjects, and in community, home care, and residential care settings.Reference Boddice, Pachana and Byrne^21, Reference Diefenbach, Tolin, Meunier and Gilliam²² A short 4-item version of the GAI has also been validated, for use in epidemiological surveys and clinical screening.Reference Byrne and Pachana²³

The prevalence, incidence, course and outlook of anxiety in old age

Bryant et al.Reference Bryant, Jackson and Ames²⁴ have reviewed studies of the prevalence of anxiety in older adults for the period 1980–2007. They performed a systematic database search for articles that reported data on anxiety disorders or symptoms in both community and clinical populations, in subjects aged 60 years and older. A total of 49 articles from a wide range of North American and European epidemiological studies met their criteria, but methodological differences meant that formal meta-analysis was not possible.

There was a wide variation in the prevalence estimates of anxiety and its disorders in these older adult community populations. A few studies reported the prevalence of anxiety symptoms, which ranged from 15% (two or more symptoms on the Hopkins Sympton Checklist (HSC)Reference Mehta, Simonsick, Penninx, Schulz, Rubin, Satterfield and Yaffe²⁵) to 52.3% (one or more symptoms on Geriatric Mental State (GMS)Reference Schaub and Linden²⁶). So far as anxiety disorders were concerned, overall prevalence rates ranged from 1.2% (GMS-AGECATReference Copeland, Gurland, Dewey, Kelleher, Smith and Davidson²⁷) to 14% (DSM-IVReference Ritchie, Artero and Beluche²⁸). For specific disorders, the prevalence of phobic disorders ranged from 1.4% (GMS-AGECATReference Copeland, Gurland, Dewey, Kelleher, Smith and Davidson²⁷) to 25.6% (Phobic Disorder Screen (PDS)Reference Lindesay and Banerjee²⁹), with specific phobia being the most common phobic disorder, and social phobia the least common. The prevalence of generalized anxiety disorder (GAD) ranged from 1.3% (Composite International Diagnostic Interview (CIDI)Reference Hunt, Issakidis and Andrews³⁰) to 7.1% (HSCReference Uhlenhuth, Balter, Mellinger, Cisin and Clinthorne³¹); GAD was the commonest anxiety disorder reported in these studies. Prevalence data on the less common anxiety disorders in old age are much more limited. The Longitudinal Aging Study in Amsterdam has reported 6-month prevalence rates of 0.9% for case-level post-traumatic stress disorder (PTSD) and 13.1% for sub-threshold PTSD in their sample.Reference van Zelst, de Beurs, Beekman, Deeg and van Dyck³² Because of the episodic nature of panic attacks and panic disorder, they are difficult to capture reliably in cross-sectional surveys, so it is not surprising that the studies that have reported on this disorder have found very low rates (1% or less).

By comparison, the prevalence studies carried out in clinical settings, such as hospitals, nursing homes and primary care facilities are limited, but the results are again very variable, with anxiety symptoms ranging from 15% (geriatric hospital, GMS-AGECATReference Ames, Flynn and Harrigan³³) to 56% (general hospital, GMS-AGECATReference Ames and Tuckwell³⁴), and anxiety disorders from 1% (general hospital, GMS-AGECATReference Ames and Tuckwell³⁴) to 24% (primary care, CIDI-SFReference Tolin, Robinson, Gaztambide and Blank³⁵).

More recent surveys continue to report prevalence rates for anxiety disorders in older adults within the range identified by Bryant et al.Reference Bryant, Jackson and Ames²⁴ The US National Co-morbidity Survey – Replication (NCS-R) found that the 12-month prevalence of DSM-IV anxiety disorder in a representative sample of those aged 55 years and older was 12.0%Reference Byers, Yaffe, Covinsky, Friedman and Bruce³⁶; specific phobia was the commonest specific anxiety disorder identified (7.0%). The prevalence rate of anxiety disorder declined with increasing age in this sample. In the National Australian Survey of Mental Health and Well-Being,Reference Goncalves, Pachana and Byrne³⁷ the 12-month prevalence of DSM-IV GAD was 2.8%.

Some of the differences in the reported prevalence rates may reflect genuine differences between populations, but much is due to conceptual and methodological differences between the studies.Reference Bryant, Jackson and Ames²⁴ A key factor is variation in the level of severity that is used to define a case. Studies using GMS-AGECAT demonstrate this; typically, they report quite high rates for sub-cases but very low rates for cases, indicating that its threshold is high. Similarly, the prevalence of DSM-IV anxiety rises considerably if all of the sub-threshold manifestations of anxiety in assessments of older adults are considered.Reference Grenier, Préville, Boyer, O'Connor, Béland, Potvin, Hudon and Brassard³⁸ Another important reason for the difference in prevalence between studies is the use of different diagnostic rules. In particular, if hierarchical rules are applied, so that cases of anxiety associated with other disorders such as depression and dementia are not counted (e.g. GMS-AGECAT, DIS-DSM-III), this will result in far fewer cases of anxiety disorder being identified. For example, in a study where different diagnostic criteria for phobic disorders were applied to the same population,Reference Lindesay and Banerjee²⁹ a hierarchical system (GMS-AGECAT) yielded a prevalence of 3.6%, as against 25.6% when a non-hierarchical system (PDS) was used. Hierarchical exclusion is particularly problematic in older adult populations, where there is extensive co-morbidity between depression and anxiety, particularly GAD.Reference Lindesay, Briggs and Murphy^39–Reference Jeste, Hays and Steffens⁴¹ Indeed, the extent of this co-morbidity has led some to question the utility of depression and GAD as distinct diagnoses, and propose instead a dimensional construct for these disorders.Reference Tyrer^42–Reference Haslam⁴⁴ The prevalence of Common Mental Disorder as defined by the Revised Clinical Interview Schedule (CIS-R) in the UK National Survey of Psychiatric Morbidity was 9% in males and 14% in females over the age of 65 years.Reference Evans, Singleton and Meltzer⁴⁵

These diagnostic problems mean that it is difficult to know to what extent the prevalence rates of anxiety disorders change with increasing age. Those epidemiological studies that have examined prevalence rates across the adult age range report rather different patterns. In the US Epidemiological Catchment Area (ECA) study, the prevalence of anxiety disorders generally declined with increasing age, although this difference was less pronounced for some categories, such as phobic disorder.Reference Robins and Regier⁴⁶ In the UK National Survey of Psychiatric Morbidity the prevalence of GAD peaked in the 50s and declined thereafter.Reference Jenkins, Lewis, Bebbington, Brugha, Farrell, Gill and Meltzer⁴⁷ Across the 1993–2007 period covered by the three UK National Psychiatric Morbidity Surveys, there has been no meaningful or consistent variation of common mental disorder or symptoms of worry across cohorts.Reference Spiers, Bebbington, McManus, Brugha, Jenkins and Meltzer⁴⁸ Few studies report data on the very oldest age groups; there is a suggestion that anxiety rates may decline in extreme old age,Reference Shaub and Linden^49, Reference Gum, King-Kallimanis and Kohn⁵⁰ but not all surveys support this.Reference Beekman, Bremmer, Deeg, van Balkom, Smit, de Beurs, van Dyck and van Tilburg⁵¹ It is of course important to bear in mind that most of our information on this subject comes from cross-sectional studies, and must be interpreted with care. It may not simply be the case that our vulnerability to developing anxiety declines as we grow old; this pattern may be the result of an excess mortality associated with anxiety (healthy survivor effects), or of different age cohorts having different vulnerabilities to mental illness.Reference Klerman⁵²

There is little evidence available regarding the incidence rates of anxiety in older age groups. In studies that have looked retrospectively at age of onset, such as the ECA, most of the cases of GAD had their onset prior to the age of 65 years;Reference Blazer, Hughes, George, Robins and Regier⁵³ however, a more recent study of US adults aged 55 years and older found that only 33.7% of those with current GAD reported an onset prior to the age of 50.Reference Chou⁵⁴ In a secondary analysis of data from the Australian National Survey of Mental Health and Well-Being, Goncalves and ByrneReference Goncalves and Byrne⁵⁵ found that earlier age of onset of GAD (<26 years) was associated with childhood physical abuse and worse clinical outcomes, suggesting that this may be an indicator of increased vulnerability.

Prospective studies of the course and outcome of anxiety disorders in late life are sparse, but data from the Aging Study Amsterdam indicate that they continue to be chronic conditions, with 6-year remission rates of only 31%.Reference Schuurmans, Comijs, Beekman, de Beurs, Deeg, Emmelkamp and van Dyck⁵⁶ This may not be true of all anxiety disorders, however, as one recent study found that social phobia showed substantial improvement and recovery over 5-year follow-up in an older adult general population sample.Reference Karlsson, Sigström, Waern, Ostling, Gustafson and Skoog⁵⁷ The association between depression and increased mortality risk in late life is well established, but that between anxiety and mortality is less clear. In one 10-year follow-up study of an older adult community sample, no excess mortality was found in association with either generalized anxiety or mixed anxiety-depression;Reference Holwerda, Schoevers, Dekker, Deeg, Jonker and Beekman⁵⁸ in another study population of adults of all ages (the Norwegian HUNT study), anxiety co-morbid with depression in fact reduced mortality risk compared with depression alone, and there was a U-shaped relationship between anxiety symptoms and mortality.Reference Mykletun, Bjerkeset, Overland, Prince, Dewey and Stewart⁵⁹ It may be that specific groups are at risk; in one study, anxiety symptoms were found to be an independent predictor of all-cause mortality in older adult black individuals, but not whites.Reference Brenes, Kritchevsky, Mehta, Yaffe, Simonsick, Ayonayon, Rosano, Rubin, Satterfield and Penninx⁶⁰

Risk factors for anxiety in old age

The factors that affect the risk of developing an episode of anxiety may be classed as pre-morbid vulnerability factors, those that trigger the onset of a particular episode (destabilization), and those that determine the duration of the episode (restitution).Reference Goldberg and Huxley² While exposure to specific risk factors changes across the lifespan, with some (e.g. disability) being more prevalent in old age, ageing per se does not appear to be a risk factor for anxiety,Reference Beekman, Bremmer, Deeg, van Balkom, Smit, de Beurs, van Dyck and van Tilburg⁵¹ and may in fact be protective.Reference Acierno, Ruggiers, Kilpatrick, Resnick and Galea^61, Reference Heun, Papassotiropoulos and Ptok⁶² The evidence relating to factors associated with anxiety and depression in old age has been reviewed by Vink et al.Reference Vink, Aartsen and Schovers⁶³ They reviewed studies published between 1995 and 2005. Over this period, studies of risk factors for depression outnumbered those of anxiety. Once again, methodological differences between the studies precluded any formal meta-analysis. The various risk factors were classified as either: biological, psychological or social. The principal biological risk factors associated with anxiety disorders in cross-sectional studies were: number of chronic health problems, poor self-perceived health, and functional limitation.Reference Schoevers, Beekman, Deeg, Jonker and van Tilburg^64, Reference Beekman, de Beurs, van Balkom, Deeg, van Dyck and van Tilburg⁶⁵ Evidence from the few longitudinal studies was limited, with only poor self-perceived health and sensory deprivations emerging as predictors of anxiety symptoms.Reference van Zelst, de Beurs, Beekman, Deeg and van Dyck^32, Reference de Beurs, Beekman, Geerlings, Deeg, Van Dyck and Van Tilburg⁶⁶ The psychological risk factors for anxiety identified in both cross-sectional and longitudinal studies were: external locus of control and neuroticism, poor coping strategies, and psychopathology.Reference van Zelst, de Beurs, Beekman, Deeg and van Dyck^32, Reference Schoevers, Beekman, Deeg, Jonker and van Tilburg^64, Reference Beekman, de Beurs, van Balkom, Deeg, van Dyck and van Tilburg⁶⁵ Not many of the included studies had examined social risk factors for anxiety in late life in any detail. Social demographic associations such as female gender and lower educational level are reported,Reference Beekman, Bremmer, Deeg, van Balkom, Smit, de Beurs, van Dyck and van Tilburg^51, Reference Heun, Papassotiropoulos and Ptok^62, Reference Schoevers, Beekman, Deeg, Jonker and van Tilburg⁶⁴ and the size and quality of social network has been identified in some studies.Reference van Zelst, de Beurs, Beekman, Deeg and van Dyck^32, Reference Beekman, Bremmer, Deeg, van Balkom, Smit, de Beurs, van Dyck and van Tilburg^51, Reference Forsell⁶⁷ In comparative studies examining both anxiety and depression, infrequent social contacts, childlessness, adverse life events and low income were associated with anxiety only, whereas small social network size and being unmarried were associated with depression. Regarding life events, depression seems to be associated more with losses and anxiety with threats and trauma. Overall, however, the risk profile for anxiety and depression and for symptoms and disorders appear very similar, supporting the dimensional rather than categorical construct of these conditions.Reference Vink, Aartsen and Schovers⁶³

There is little evidence on the extent to which prevalence rates of anxiety disorders differ in different racial and ethnic groups in old age. In the US, elderly Puerto Ricans may express more mixed anxiety-depression than elderly African-Americans.Reference Diefenbach, Disch, Robison, Baez and Coman⁶⁸ In elderly African-Americans, PTSD is the anxiety disorder with the highest 12-month prevalence rate, in marked contrast to the results from studies of nationally representative samples.Reference Ford, Bullard, Taylor, Toler, Neighbors and Jackson⁶⁹

More recent studies of risk factors associated with late-life anxiety support the general conclusions of Vink et al.Reference Vink, Aartsen and Schovers⁶³ In the cross-sectional National Australian Survey of Mental Health and Well-Being,Reference Goncalves, Pachana and Byrne³⁷ GAD was associated with functional limitations, psychiatric co-morbidity and increased medication intake.

Depression

As noted above, the symptoms of anxiety and depression are dimensional constructs that commonly occur together, particularly in the majority of less severely affected individuals seen in primary care settings; co-morbidity between these conditions at the case level is also common at all ages. There is extensive co-morbidity between GAD and depression in old age in cross-sectional surveys,Reference Schoevers, Beekman, Deeg, Jonker and van Tilburg⁶⁴ but evidence regarding the temporal nature of their relationship in this age group is limited. The findings of Lenz et al.Reference Lenze, Mulsant, Mohlman, Shear, Dew, Schulz, Miller, Tracey and Reynolds⁷⁰ suggest that GAD precedes depression and persists as a chronic disorder while the co-morbid depression may remit and relapse over time. Co-morbidity is associated with greater GAD severity,Reference Hopko, Bourland, Stanley, Beck, Novy, Averill and Swann⁷¹ greater suicidal ideation,Reference Bartels, Coakley, Oxman, Constantino, Oslin, Chen, Zubritsky, Cheal, Durai, Gallo, Llorente and Sanchez⁷² and impaired response to treatment.Reference Flint and Rifat⁷³ Co-morbidity also occurs between depression and other anxiety disorders in old age, notably panic disorder, agoraphobia, and social phobia.Reference Corna, Cairney, Herrmann, Veldhuizen, McCabe and Streiner^74–Reference Cairney, Corna, Veldhuizen, Hermann and Streiner⁷⁶ Bipolar disorder is co-morbid with anxiety disorders at all ages, including late life.Reference Sajatovic, Blow and Ignacio⁷⁷ As with younger adults, anxiety disorders in old age, particularly social phobia, also show significant co-morbidity with alcohol abuse and dependence.Reference Chou⁷⁸ Axis II personality disorders have also been found to be more prevalent in older adults with GAD.Reference Mackenzie, Reynolds, Chou, Pagura and Sareen⁷⁹

Cognitive impairment

Since both anxiety and cognitive impairment are relatively common in the older adult population, it is not surprising that they often occur together.Reference Forsell, Palmer and Fratglioni⁸⁰ Over and above this chance association, however, there is evidence to suggest a causal relationship in a proportion of cases. For example, baseline anxiety symptoms or disorders have been shown to be predictive of cognitive decline in prospective follow-up studies.Reference Sinoff and Werner^81, Reference Potvin, Forget, Grenier, Preville and Hudon⁸² Studies of US veterans have found that those with PTSD are more likely to develop dementia than those without.Reference Yaffe, Vittinghoff, Lindquist, Barnes, Covinsky, Neylan, Kluse and Marmar⁸³ At all ages, anxiety can interfere with effective cognitive functioning via poor concentration, autonomic arousal, sleep impairment and intrusive worries, and in some cases individuals present with concerns about cognitive impairment. Subjective forgetfulness is associated primarily with anxiety and depression,Reference Cooper, Bebbington, Lindesay, Meltzer, McManus, Jenkins and Livingston⁸⁴ so it is important that these conditions are screened for in patients referred to memory clinics for the assessment of mild cognitive impairments. It is also important (although difficult) to determine the sequence of events; anxiety that develops after individuals become aware of their cognitive decline may be a response to the decline and its implications.

In people with dementia, the epidemiological evidence suggests that there is an association with anxiety symptoms, but not anxiety disorders.Reference Schaub and Linden^85, Reference Hwang, Masterman, Ortiz, Fairbanks and Cummings⁸⁶ However, the reported prevalence of anxiety in this population ranges widely, reflecting differences in methodology, and difficulties in diagnosing anxiety in the presence of dementia.Reference Wolitzky-Taylor, Castriotta, Lenze, Stanley and Craske^11, Reference Seignourel, Kunik, Snow, Wilson and Stanley⁸⁷ It may be that anxiety can manifest in different ways as dementia progresses, for example as agitation,Reference Twelftree and Qazi⁸⁸ which would lead to it being under-diagnosed in studies using the standard classifications. Rates of anxiety may be higher in vascular dementia than in Alzheimer's disease, but the evidence is limited.Reference Seignourel, Kunik, Snow, Wilson and Stanley⁸⁷

Physical illness

The co-morbidity of anxiety and physical illness raises important issues for detection and diagnosis. Physical ill-health and disability are relatively common in old age, so a degree of co-occurrence with common psychiatric disorders such as depression and anxiety would be expected by chance. However, the rates of co-morbidity observed between anxiety and certain medical conditions, such as cardiac and respiratory disorders, are such that there must be some causality; for example, studies report that about one-third of older cardiac patients and half of chronic obstructive pulmonary disease (COPD) patients also have an anxiety disorder.Reference Torado, Shen and Raffa^89, Reference Vogele and von Leupoldt⁹⁰ Other physical disorders associated with anxiety symptoms and disorders in older patients include postural disturbance and vestibular disease,Reference Gagnon, Flint, Naglie and Devins⁹¹ Parkinson's disease,Reference Pontone, Williams, Anderson, Chase, Goldstein, Grill, Hirsch, Lehmann, Little, Margolis, Rabins, Weiss and Marsh⁹² and self-reported limitations of mobility.Reference Mehta, Yaffe, Brenes, Newman, Shorr, Simonsick, Ayonayon, Rubin and Covinsky⁹³ The precise nature of this causal relationship is complex; for example, individuals may become anxious when they experience the symptoms of a physical illness, their anxiety symptoms may be interpreted as being physically caused, leading to further anxiety, and so on in a vicious circle. Alternatively, the symptoms of some physical disorders (or the side-effects of their treatments) may resemble the symptoms of anxiety. On the one hand, this state of affairs can lead to over-investigation and misdiagnosis of physical disorder in those who are primarily anxious; on the other, important symptoms of physical illness may be dismissed as being psychological in origin – a difficult dilemma for the clinician, and one that can only be addressed by careful and unprejudiced assessment of the individual case.Reference Norhus, Hynninen, Pachana, Laidlaw and Knight⁹⁴

Non-pharmacological

Compared with the number of studies of CBT for common mental disorders in younger adults, the evidence base supporting its use in older adult patients is still small. In a recent Cochrane review of psychotherapeutic treatments for older depressed adults,Reference Wilson, Mottram and Vassilas⁹⁸ only nine studies were amenable to meta-analysis, and the conclusions were limited accordingly. In a systematic review of CBT for late-life anxiety disorders, Hendriks et al.Reference Hendriks, Oude Voshaar, Keijsers, Hoogduin and van Balkom⁹⁹ examined studies that compared CBT with a waiting-list control and an active control condition in patients aged 60 years and older. They also identified nine randomized controlled comparisons involving a total of 297 patients that met their inclusion criteria, and meta-analysis showed that CBT was significantly superior at controlling symptoms of anxiety, worry and depression than either a waiting-list or active control. In a more broadly based meta-analysis, Thorp et al.Reference Thorp, Ayers, Nuevo, Stoddard, Sorrell and Wetherell¹⁰⁰ reviewed 19 studies of behavioural treatments for anxiety symptoms in older adults, and concluded that overall these were more effective than active controls, and that the effect size was comparable to those seen for behavioural treatments in younger adults and for pharmacotherapy in older adults. Although direct comparison was difficult, their analysis suggested that CBT (alone or in combination with relaxation training) did not offer anything beyond relaxation training alone. In another review, Ayers et al.Reference Ayers, Sorrell, Thorp and Wetherall¹⁰¹ looked at 17 studies that met their criteria for evidence-based treatment for anxiety disorders in older adults. They concluded that there was good evidence to support the use of CBT and relaxation training, and some evidence for supportive therapy and cognitive therapy. The patients in these studies were suffering from GAD or mixed anxiety and depression, and the authors note the need for studies in other conditions, such as phobic disorders and post-traumatic stress disorder.

More recent studies support the general conclusions of these reviews. In a randomized controlled trial (RCT) of CBT for GAD in older adults, Bradford et al.Reference Bradford, Cully, Rhoades, Kunik, Kraus-Schuman, Wilson and Stanley¹⁰² found this to be superior to enhanced usual care, and that the level of early response at 1 month predicted the longer-term outcome at 1-year post-treatment follow-up. In a study comparing paroxetine and CBT in the treatment of panic disorder in 49 patients aged 60 and older, the effect of both approaches was similar, and superior to the waiting-list control.Reference Hendriks, Keijsers, Kampman, Oude Voshaar, Verbraak, Broekman and Hoogduin¹⁰³

The setting in which treatment is provided may also be important. In a RCT of CBT for GAD in older adults in primary care, Stanley et al.Reference Stanley, Wilson, Novy, Rhoades, Wagener, Greisinger, Cully and Kunik¹⁰⁴ found that, compared with enhanced usual care, CBT was associated with improvements in worry, depression and general mental health, but not in the overall severity of the GAD.

When evaluating this evidence, it is important to bear in mind that to date, most studies have involved relatively small numbers of young-old, well-educated females, and it is not clear to what extent the results are generalizable to other groups within the older adult population.Reference Mohlman¹⁰⁵ Moreover, CBT in this age group appears to be associated with high attrition rates,Reference Mohlman¹⁰⁵ so retention strategies may be important to optimize the treatment effect. Long-term follow-up studies are few, so it remains unclear to what extent the effects of interventions such as CBT are sustained in this age group. One post-trial follow-up suggests that the treatment effect of CBT for anxiety disorders may not be as long-lasting as that of a course of pharmacotherapy with sertraline;Reference Schuurmans, Comijs, Emmelkamp, Weijnen, van den Hout and van Dyck¹⁰⁶ the factors predicting less good outcome with CBT in this study were poor perceived health and neuroticism.

The role of CBT in the management of anxiety associated with physical illness or disability, and cognitive impairment, is not as yet clear. In patients with anxiety associated with physical illness, it is not clear to what extent CBT has an effect superior to that of simply providing tailored health information.Reference Kunik, Veazey, Cully, Souchek, Graham, Hopko, Carter, Sharafkhaneh, Goepfert, Wray and Stanley¹⁰⁷ With regard to cognitive impairment, baseline errors in the Mini-Mental State Examination (MMSE) orientation subscale appear to be associated with reduced response to CBT for GAD.Reference Caudle, Senior, Wetherell, Rhoades, Beck, Kunik, Snow, Wilson and Stanley¹⁰⁸ It may be that the goals and techniques of the treatment need to be modified in these patient groups. There have to date been no trials of CBT for anxiety symptoms and disorders specifically in the context of dementia, but a report of two case-studies suggest that a modified approach involving caregivers may be effective.Reference Kraus, Seignourel, Balasubramanyam, Snow, Wilson, Kunik, Schulz and Stanley¹⁰⁹ Research is also needed into the benefits of combined psychological and pharmacological interventions, compared with either on its own.Reference Goncalves and Byrne¹¹⁰

Pharmacological

Psychotropic drugs continue to be the mainstay of the management of anxiety symptoms in older adult patients, despite the paucity of evidence of their effectiveness in this population. What evidence we do have suggests that antidepressant drugs are an effective treatment for GAD in late life, both in the presence and absence of co-morbid depression. In a comparative meta-analysis of behavioural and pharmacological interventions, Pinquart et al.Reference Pinquart and Duberstein¹¹¹ concluded that pharmacotherapy was associated with greater improvement, but that this difference disappeared if non-specific changes in the control groups were controlled for. In their systematic review, Goncalves and ByrneReference Goncalves and Byrne¹¹⁰ found that the pooled treatment effects of pharmacological trials were similar to those of psychotherapeutic trials, with both superior to controls. An SSRI is nowadays the first treatment of choice;⁹⁵ one small trial in older adult patients with GAD has found citalopram to be anxiolytic,Reference Lenze, Mulsant, Shear, Dew, Miller, Pollock, Houck, Tracey and Reynolds¹¹² and escitalopram has also been shown to be effective in studies that have included older subjects.Reference Goodman and Bose¹¹³ Mixed-age studies also support the use of paroxetine and trazodone,Reference Flint¹¹⁴ and there is some evidence that the SNRI venlafaxine is effective in older patients.Reference Katz, Reynolds, Alexopoulos and Hackett¹¹⁵ According to a recent systematic review of trials in GAD in general adult samples,Reference Baldwin, Woods, Lawson and Taylor¹¹⁶ fluoxetine was ranked first for response and remission, and sertraline for tolerability. In a sub-analysis of drugs licensed for this indication in the UK (duloxetine, escitalopram, fluoxetine, paroxetine, sertraline, pregabalin, tiagabine, lorazepam, venlafaxine), duloxetine ranked first for response, escitalopram for remission, and pregabalin for tolerability. There is no systematic evidence to support the use of antidepressants in other anxiety disorders in this age group, but extrapolation from studies in younger adults suggests that an SSRI would also be the drug of choice for the treatment of panic disorder.Reference Flint and Gagnon¹¹⁷ Tricyclic antidepressants are effective against symptoms of anxiety, but, with the exception of lofepramine,Reference Dorman¹¹⁸ the use of these drugs is limited by the risks they pose to physically ill patients, and their toxicity in overdose. They are not supported by the NICE guideline, except as part of a specialist Step 4 intervention.

The problems associated with benzodiazepine use at all ages are well-known,Reference Madhusoodanan and Bogunovic¹¹⁹ and their continued widespread use is presumably a testament to the lack of available alternatives. Despite these problems, however, there is still a place for these drugs in the management of transient, short-term anxiety symptoms in a crisis.⁹⁵ The current wisdom is that drugs with short half-lives and no active metabolites, such as oxazepam, are to be preferred, but there is increasing evidence that these may also be problematic. For example, short-acting benzodiazepines have been found to be associated with hip fracture, and with increased rates of disability in mobility and activities of daily living.Reference Wagner, Zhang, Soumerai, Walker, Gurwitz, Glynn and Ross-Degnan^120, Reference Gray, LaCroix, Hanlon, Penninx, Blough, Leveille, Artz, Guralnik and Buchner¹²¹ Is not clear why this should be; perhaps they are more likely to be prescribed to high-risk patients.

There is evidence that buspirone (an azapirone 5-HT_1A partial agonist) is a safe and effective anxiolytic in all age groups,Reference Chessick, Allen, Thase, Batista Miralha da Cunha, Kapczinski, de Lima and dos Santos Souza¹²² including older patients.Reference Robinson, Napoliello and Schenk^123, Reference Böhm, Robinson, Gammans, Shrotriya, Alms, Leroy and Placchi¹²⁴ Unlike other anxiolytics and like antidepressants, it takes about 2 weeks to become effective, so would be indicated for the management of severe, chronic GAD rather than acute anxiety states. However, longer-term studies of its effectiveness are still needed.Reference Chessick, Allen, Thase, Batista Miralha da Cunha, Kapczinski, de Lima and dos Santos Souza¹²² It may be that non-pharmacological factors may be important determinants of response to buspirone in older adult in-patients; those with high levels of both physical co-morbidity and social contacts showed greater response than those who scored low on these factors.Reference Majercsik and Haller¹²⁵ Pregabalin is another novel compound that has been found to be safe and effective in the treatment of GAD, both in younger adults and in older out-patients with the disorder.Reference Montgomery, Chatamra, Pauer, Whalen and Baldinetti¹²⁶ Like antidepressant drugs and buspirone, the onset of efficacy of pregabalin is at about 2 weeks after initiating treatment.

The use of typical antipsychotic drugs in the management of anxiety is not recommended, because of the risk of adverse side-effects. The role of atypical antipsychotics is less clear; recent systematic reviews of trials involving adults of all ages find some benefit associated with quetiapine in GAD.Reference LaLonde and Van Lieshout^127, Reference Maher, Maglione, Bagley, Suttorp, Hu, Ewing, Wang, Timmer, Sultzer and Shekelle¹²⁸ In older patients, however, adverse events associated with use of atypical antipsychotics included increased risk of death, stroke, extrapyramidal symptoms and urinary tract infections.