Depression and Generalized Anxiety Disorder: Co-Occurrence and Longitudinal Patterns in Elderly Patients

Depression and Generalized Anxiety Disorder Co-Occurrence and Longitudinal Patterns in Elderly Patients Robert A. Schoevers, M.D., D.J.H. Deeg, Ph.D. W. van Tilburg, M.D., Ph.D., A.T.F. Beekman, M.D., Ph.D.

Objective: The authors sought to establish the natural course and risk-profile of depression, generalized anxiety disorder (GAD), and depression with co-existing GAD in later life. Methods: A total of 2,173 community-living elderly persons were interviewed at baseline, and at a 3-year follow-up. The course of “pure” depression, “pure” GAD, and depression with coexisting GAD was studied in 258 subjects with baseline psychopathology. Authors assessed bivariate and multivariate relationships between risk factors and course types. The risk-profile for onset of pure depression, pure GAD, and the mixed condition at follow-up was studied in 1,915 subjects without baseline psychopathology. Results: Remission rate at follow-up was 41% for subjects with depression-only, 48% for pure GAD, and significantly lower (27%) for depression with coexisting GAD. A pattern of temporal sequencing was established, with anxiety often progressing to depression or depression with GAD. Onset of pure depression and depression with co-existing GAD was predicted by loss events, ill health, and functional disability. Onset of pure GAD, and, more strongly, that of depression with coexisting GAD, was associated with longstanding, possibly genetic vulnerability. Conclusions: In comparison with either depression-only or anxiety-only, the co-occurrence of these represents more severe and more chronic psychopathology, associated with longstanding vulnerability. In elderly persons, GAD often progresses to depression or to the mixed condition. These findings mostly favor a dimensional, rather than a categorical, classification of anxiety and depression. (Am J Geriatr Psychiatry 2005; 13:31–39)

C

o-occurrence of depression and generalized anxiety disorder (GAD) is a consistent finding across different age-groups.1–3 In adults, comorbidity is as-

sociated with a poorer prognosis,1,4 greater functional disability,5 and higher service utilization than seen in persons with a single disorder.6 In elderly persons,

Received March 22, 2004; revised June 15, July 13, September 2, 2004; accepted September 8, 2004. From the Department of Psychiatry, VU University Medical Center, Amsterdam, The Netherlands (RAS,WVT,ATFB), Mentrum Mental Health Care, Amsterdam (RAS), and the Institute for Research on Extramural Medicine, VU University Medical Center, Amsterdam (DJHD). Send correspondence and reprint requests to Robert Schoevers, M.D., Mentrum GGZ, Amsterdam, 2e Constantijn Huijgensstraat 37, 1054 AG Amsterdam, The Netherlands. e-mail: robert. [email protected] 䉷 2005 American Association for Geriatric Psychiatry

Am J Geriatr Psychiatry 13:1, January 2005

31

Co-Occurrence of Depression and GAD much less is known about the course and consequences of depression with coexisting GAD.7 Available data, however, suggest that, also in elderly patients, comorbidity of depression and GAD may appear in a group with unfavorable characteristics in terms of severity and prognosis of psychopathology.8 Cross-sectional findings from the Amsterdam Study of the Elderly (AMSTEL) showed that the combined condition differs from both single-disorder conditions in the sense that comorbidity increases with higher severity levels of both depression and GAD.9 Given the high prevalence and potential adverse consequences of this type of comorbidity in elderly patients, there is an urgent need for longitudinal research in this domain.10 Apart from its clinical implications, the frequent co-occurrence of depression and anxiety has also fuelled the discussion concerning the validity of current categorical classifications of mental disorders, with critics suggesting that a dimensional approach is more appropriate.11–13 GAD, in particular, may be on a continuum with depression.4 It has also been suggested that, although the same genetic factors appear to predispose to both generalized anxiety and depression, specific environmental factors may be differentially associated with the development of either anxiety or depression.14,15 Studies on the association of risk factors with each of these disorders in elderly patients are, however, sparse and showed contradictory results in cross-sectional analyses.2,8,9 This may be a result of differences in study designs. It may also be due to the fact that a cross-sectional design is unsuited to differentiating between factors that co-occur with or result from psychopathology and factors that actually antedate and may be etiologically related to the occurrence of a condition. Thirdly, anxiety, depression, and the mixed condition may also represent subsequent stages of disorder. Although temporal sequencing has been described in younger adults, with anxiety often occurring before the onset of depression, very little is known about longitudinal patterns in elderly patients.3,16 The current study describes the natural course and risk-profile of “pure” depression, “pure” GAD, and depression with co-existing GAD in later life. It seeks to establish patterns of temporal sequencing, and investigates whether comorbidity, in fact, represents a more severe and more chronic condition than single disorders. It also explores questions on the specificity

32

of risk factors for onset of these disorders, in light of the debate on dimensional versus categorical classification of depression and anxiety. The study was performed in a large cohort of community-living elderly patients, by use of a prospective design, with a 3-year follow-up interval.

METHODS Sampling and Non-Response The Amsterdam Study of the Elderly (AMSTEL) is a longitudinal study of a large and representative sample of community-dwelling elderly persons; the study investigates mental health problems, medical diagnoses, and demographic characteristics. The sampling and data collection procedures have been described elsewhere.17,18 In short, the population base for AMSTEL included all non-institutionalized individuals in the 65–84-year-old age bracket who lived in the city of Amsterdam. The profile of the study sample corresponded to the non-institutionalized Amsterdam population in terms of age and gender. An age-stratified sample was drawn. All 4,051 subjects (71.5%) who responded and gave their informed consent were interviewed at baseline. Non-response in the younger-old (⬍age 75) was associated with health problems and poorer performance on cognitive tests. In the older-old, we found no correlates of non-response.19 Of the 4,051 subjects who initially responded, 2,244 (55.4%) were re-interviewed 3 years later (median: 38 months). Nonresponse consisted of 656 subjects (16.2%) who had died, 662 persons (16.3%) who refused further cooperation, 282 (7.0%) who were too ill or cognitively impaired to respond, and 207 (5.1%) who were not available for interview for other reasons. Subjects with an organic diagnosis at baseline (GMS-AGECAT case levels for organicity ⬎2; N⳱71) were excluded, yielding a total sample of 2,173 subjects. The study sample for the investigation of course types of psychopathology consisted of subjects with baseline GMS-AGECAT case levels of depression, anxiety, or both, who attended follow-up assessment. The sample for these assessments thus consisted of 258 subjects. The analyses of the association between risk factors

Am J Geriatr Psychiatry 13:1, January 2005

Schoevers et al. and incidence of depression, anxiety, and depression with coexisting GAD at follow-up were performed in a sample consisting of all responders who did not have depression, GAD, or dementia at baseline (GMS-AGECAT levels ⬍3 for depression, anxiety, and organicity; N⳱1,915).

and “psychotic” depression. Psychotic depression was classified as more severe than neurotic depression. In subjects with any of these three types of psychopathology at baseline, remission was defined as no longer having any of these diagnoses at followup.

Measures

Risk factors. Educational status was dichotomized as lower (primary school or less) and higher (more than primary school) education. Marital status was assessed with the appropriate questions in the GMS-AGECAT. Social support was ascertained by the question: “Do you get help from your children, neighbors, or other acquaintances?” A personal history was ascertained by the relevant CAMDEX question, asking subjects whether they had ever experienced depression and/or anxiety of such severity that treatment had been sought. Age at first onset was recorded. Also translated from the CAMDEX questionnaire was presence of a family history of psychiatric illnesses. The presence of chronic diseases was assessed with the pertinent CAMDEX questions on cardiovascular diseases, cancer, lung disease, diabetes, Parkinson disease, arthritis, and epilepsy. A dichotomous variable indicated whether or not subjects had any chronic illnesses. Cognitive status was assessed by MMSE score. Subjects were considered to have functional disability if their ADL or IADL scores were 2 or more points below the maximum score on the respective scales (ADL, 12 and IADL, 16 points). This means that subjects needed help to perform at least two of the tasks in the respective domains. Differences that had occurred between baseline and follow-up were also noted. Stressors thus defined were personal life events (partner loss), changes in physical health (emergence of one or more chronic diseases), a substantial decrease in ADL or IADL functioning (2 scale points or more), and cognitive decline (incident organic syndrome; GMS-AGECAT organic Level 3 or more at follow-up).

We conducted a 1-hour interview consisting of the Dutch translation of the Mini-Mental State Exam,20 the Geriatric Mental State Examination (GMS) items related to organic, affective, and generalized anxiety syndromes,21 the Activities of Daily Living (ADL) scale,22 the Instrumental Activities of Daily Living (IADL) scale,23 and the CAMDEX interview.24 The interview was administered during home visits by lay interviewers who were specially trained by use of video sessions and were regularly supervised. Psychiatric syndromes. Diagnoses of depression, organic “caseness,” and generalized anxiety were reached by use of the GMS-AGECAT system.21,25 AGECAT has proven its reliability for epidemiological work.25 The Dutch language version has been validated.26 In the analyses, syndrome levels were used for psychopathology, because the diagnostic hierarchy in the GMS-AGECAT would otherwise bias the results. Diagnostic Levels 1 and 2 are classified as subcases; Levels 3–5 have been proven valid to detect cases requiring clinical attention in the community.27–29 Depression, generalized anxiety, and organic caseness were defined as a GMS-AGECAT at Level 3 or higher. Subjects meeting GMS criteria for depression and not meeting GMS criteria for anxiety were classified as Pure Depression (DEP). Subjects with GMS generalized anxiety disorder without depression caseness were classified as Pure GAD. Subjects meeting GMS criteria for both depression and anxiety were labeled Depression With Co-existing GAD (DEP/GAD). GMS-AGECAT case levels also represent levels of increasing severity of the disorder, with Case Level 5 indicating the most severe end of the spectrum. To determine the prognostic value of different severity levels at baseline, a dichotomous variable was constructed for severity of psychopathology. GAD Case Levels 4 and 5 were classified as higher severity, and Level 3 as less-severe illness. Depressed subjects were subdivided into “neurotic”

Am J Geriatr Psychiatry 13:1, January 2005

Data Analysis Characteristics at baseline and follow-up were described, and response rate according to baseline characteristics was calculated using bivariate and multivariate statistics. In the baseline case sample, possible patterns of temporal sequencing across both assessments of DEP,

33

Co-Occurrence of Depression and GAD GAD, and DEP/GAD were calculated in cross-tabulation with chi-square analysis. The probability of different course types was tested by dichotomizing outcome as “remission versus persistence,” and “developing DEP/GAD versus not” and drawing up 2 ⳯ 2 cross-tabulations. In this way, we compared the courses of baseline GAD and DEP, and the courses of subjects who developed DEP/GAD. In the non-case baseline sample, associations of independent variables with the incidence of either DEP, GAD, and DEP/GAD at follow-up were assessed, calculating odds ratios (ORs), using the subjects without depression and anxiety at follow-up as the reference category. When the 95% confidence interval (CI) did not include 1, the association was regarded to be statistically significant. Multivariate logistic-regression analysis was used to achieve independent predictive ability for each diagnostic category using the full model of potential risk factors. Differences between strata were considered statistically significant according to the criterion that the confidence intervals of a risk factor mutually exclude the point estimates in one or both of the other diagnostic categories.30 Backward stepwise analysis was subsequently used to reach the most parsimonious model for each diagnostic category.

RESULTS Non-response at the follow-up assessment was predicted by higher age, male gender, lower education level, chronic disease, ADL and IADL impairment, and organic caseness.18 After subjects with dementia at baseline and those who died between measurements were excluded from the sample, only lower education, not having a marital partner, and baseline cognitive deficits (MMSE ⬍26) remained as significant predictors of non-response. Importantly, controlling for other factors in multivariate logistic-regression analysis, neither DEP (relative risk [RR]: 0.87; 95% CI: 0.68–1.12), GAD (RR: 0.97; 95% CI: 0.50– 1.88), nor DEP/GAD (RR: 0.85; 95% CI: 0.48–1.48) were significantly associated with the risk of attrition (overall likelihood v2[16] test for full model: 85.29; p ⬍0.001). Table 1 shows the sample characteristics and the associations with GAD, DEP, and DEP/GAD at follow-up.

34

Table 2 shows that the 25 subjects with GAD had the highest rate of remission (48%); three subjects (12%) remained in the same category, but six subjects (24%) developed DEP. DEP/GAD emerged in four persons (16%). Pure depression. Of the 199 subjects with DEP at baseline, 82 (41.2%) no longer had DEP or GAD at follow-up; 86 subjects (43.2%) remained in the same category at both assessments. Only 4 subjects (2%) of those with DEP at baseline developed GAD, whereas 27 (13.6%) developed DEP/GAD. Mixed anxiety-depression. Among the 34 subjects with DEP/GAD at baseline, remission rates were low (27%). A relatively high proportion of subjects remained within the same diagnostic category (44%); DEP appeared in 27%; only one person developed GAD (3%). Cross-tabulation of all course types proved statistically significant (Pearson v2[6]⳱ 28.14; p ⬍0.001). Dichotomizing course types showed that there was no statistically significant difference in the probability of remission between DEP and GAD. Subjects with DEP/GAD showed significantly lower remission rates than subjects with DEP- or GAD-only (Pearson v2[1]⳱2.96; one-tailed p⳱0.043). Comparing baseline categories on the outcome of DEP/GAD or “not DEP/GAD” did not show a statistically significant difference between GAD and DEP, but showed that subjects with baseline DEP/GAD were more likely to keep the combined disorder than change into either of the pure forms (Pearson v2[1]⳱18.47; p ⬍0.001). Associations of Risk Factors With Course of Illness Stable versus worsened. Bivariate analyses showed that in subjects with either GAD- or DEPonly, developing DEP/GAD was associated with the occurrence of chronic diseases between assessments (RR: 2.68; 95% CI: 1.32–5.42; overall v2[1]⳱8.16; p⳱0.0043; data not shown). Severity of DEP and/or GAD was not associated with course type, nor was the presence of a personal history. In the multivariate model (backward stepwise logistic-regression modeling) the association with new chronic illness remained (RR: 2.90; 95% CI: 1.43–5.89), and mixed anxiety-depression was also inversely associated with an

Am J Geriatr Psychiatry 13:1, January 2005

Schoevers et al. age above 74 years (RR: 0.45; 95% CI: 0.22–0.95; overall v2[2] at last step⳱12.8; p⳱0.0017; data not shown). Stable versus remitted. None of the risk factors showed significant differences between strata of subjects with a single disorder at baseline who had remitted at follow-up. In subjects with depression coexisting with generalized anxiety disorder at baseline, partial or complete remission was not associated with any of the risk factors, either. Predictors of Incidence of Pure GAD, Pure DEP, and Mixed Anxiety/Depression at Follow-Up In the sample without depression, anxiety, or dementia at baseline (N⳱1,915), 24 subjects (1.3%) had TABLE 1.

developed GAD at follow-up; 250 (13.1%) had DEP, and, in 49 persons (3.6%), DEP/GAD had developed. Table 3 shows the multinomial logistic-regression models for each diagnostic category. Only having a personal history of depression or anxiety was significantly associated with the onset of GAD. Because the numbers were relatively small, these associations were also tested in bivariate analyses and yielded no other statistically significant associations with GAD at follow-up. The occurrence of DEP was associated with having chronic diseases at baseline and baseline IADL disability, with loss of spouse and with a decrease in functional abilities (IADL) between assessments. DEP/GAD was predicted by having a personal history and by a decrease in IADL functioning.

Sample Characteristics Follow-Up

Risk Factors Age ⬎74 years at baseline Sex: female Education: ⱕprimary school Marital status: unmarried/ divorced/widowed Social support: help from relatives/friends Personal history of depression Family history of psychiatric disorder Baseline chronic diseases Baseline ADL disability Baseline IADL disability Low MMSE (⬍26) Changes from baseline: life events Partner died New ADL disability New IADL disability New chronic diseases

Baseline

No Depression or Anxiety (Nⴔ1,695)

GAD-Only (Nⴔ32)

DEP-Only (Nⴔ351)

DEP/GAD (Nⴔ95)

v2[df]

1,066 (49.1) 1,372 (63.1) 803 (37.0)

808 (47.7) 1,033 (60.9) 612 (36.1)

13 (40.6) 24 (75.0) 16 (50.0)

202 (57.5) 244 (69.5) 137 (39.0)

43 (45.3) 71 (74.7) 38 (40.0)

12.9[3]** 17.1[3]** 3.9[3] (p⳱0.27)

1,100 (50.6)

848 (50.1)

13 (40.6)

192 (54.7)

47 (49.5)

3.8[3] (p⳱0.28)

373 (17.2)

273 (16.1)

4 (12.5)

78 (22.2)

18 (18.9)

8.4[3]*

327 (15.0)

207 (12.2)

8 (25.0)

72 (20.5)

40 (42.1)

75.7[3]***

258 (11.9) 1,073 (49.4) 103 (4.7) 393 (18.1) 217 (10.0)

185 (10.9) 792 (46.7) 61 (3.6) 255 (15.0) 158 (9.3)

3 (9.4) 18 (56.3) 3 (9.4) 7 (21.9) 6 (18.8)

50 (14.2) 205 (58.4) 27 (7.7) 104 (29.6) 39 (11.1)

20 (21.1) 58 (61.1) 12 (12.6) 27 (28.4) 14 (14.7)

11.2[3]* 22.0[3]*** 26.3[3]*** 49.3[3]*** 6.5[3] (p⳱0.09)

155 (7.1) 128 (5.9) 448 (20.6) 422 (19.4)

97 (5.7) 86 (5.1) 308 (18.2) 300 (17.7)

1 (3.1) 2 (6.3) 10 (31.3) 8 (25.0)

43 (12.3) 29 (8.3) 98 (27.9) 82 (23.4)

14 (14.7) 11 (11.6) 32 (33.7) 32 (33.7)

28.0[3]*** 11.1[3]* 29.8[3]*** 19.7[3]***

Note: Values are N (%), unless otherwise indicated. GAD: generalized anxiety disorder; DEP: depression; DEP/GAD: depression with anxiety; MMSE: Mini-Mental State Exam; ADL: activities of daily living; IADL: instrumental activities of daily living. Baseline sample excluded subjects with dementia (dichotomized variables), N⳱2,173 with psychopathology at follow-up. Chi-square tests with two sided p-value: *p ⬍0.05; **p ⬍0.01; ***p ⬍0.001.

TABLE 2.

Stability of Diagnostic Categories Across Assessments Follow-Up Diagnosis

Baseline Diagnosis GAD-only (N⳱25) DEP-only (N⳱199) DEP/GAD (N⳱34)

No DEP/GAD

GAD-Only

DEP-Only

DEP/GAD

12 (48.0) 82 (41.2) 9 (26.5)

3 (12.0) 4 (2.0) 1 (2.9)

6 (24.0) 86 (43.2) 9 (26.5)

4 (16.0) 27 (13.6) 15 (44.1)

Note: Values are N (%), unless otherwise indicated. GAD: generalized anxiety disorder; DEP: depression; DEP/GAD: depression with anxiety.

Am J Geriatr Psychiatry 13:1, January 2005

35

Co-Occurrence of Depression and GAD There were three significant differences between depression-only and depression with co-existing GAD; the mixed category shows significantly higher ORs for age (65–74 versus 75–84 years), a personal history of depression and/or anxiety disorder, and decreases in both ADL and IADL functioning. The most parsimonious models (Table 4) showed similar results for GAD and DEP, with a personal history just failing to reach statistical significance in DEP. Having a personal history was more strongly associated with the occurrence of DEP/GAD. DEP/GAD was also predicted by baseline ADL and newly occurring functional disabilities (ADL and IADL) and loss of spouse. Risk ratios of baseline and newly occurring functional limitations were generally higher in DEP/GAD. Loss of spouse showed similar associations with the occurrence of both categories.

DISCUSSION The aim of the current study was to describe the natural course of depression-only, GAD-only, and depression with co-existing GAD in a large and representative sample of community-living elderly persons and to investigate whether different risk profiles were associated with occurrence and prognosis of each of these types of psychopathology. TABLE 3.

Assessment of the natural course of the three diagnostic categories showed that, first of all, the prognosis of either category in elderly patients is poor, with remission rates under 50% after a 3-year interval. When comparing categories, GAD was the least stable category, with only 12% of anxious subjects remaining within that category at follow-up. A significantly larger number of subjects with baseline GAD developed either DEP or DEP/GAD at follow-up. DEP and DEP/GAD appeared to be much more stable diagnostic categories, and these subjects were highly unlikely to develop GAD at follow-up. Although, because of relatively small numbers, not all of these transitions could be statistically tested, these data clearly suggest a fixed pattern of temporal sequencing. GAD in elderly subjects either remits or progresses into depression or mixed anxiety-depression, whereas subjects with DEP or DEP/GAD are highly unlikely to develop GAD-only at follow-up. Although this finding is in line with those for younger adults,31 such sequential patterns have not been systematically described in this age-group before. Studying the course of DEP/GAD showed that, with only 27% remission at follow-up, the prognosis of this mixed category is significantly worse than that of either of the single conditions. This confirms findings of earlier, mostly cross-sectional research show-

Associations of Risk Factors With Incidence of Generalized Anxiety Disorder, Depression, and Depression With CoExisting Generalized Anxiety Disorder Follow-Up Diagnosis

Risk Factors Age ⬎74 years Female Only basic education Not / no longer married Social support Personal history of depression/anxiety Family history of mental illnesses Chronic illnesses at baseline ADL disability at baseline IADL disability at baseline MMSE ⬍26 Loss of spouse New medical illness Recent decrease in ADL functioning Recent decrease in IADL functioning

GAD-Only (Nⴔ24)

DEP-Only (Nⴔ250)

DEP/GAD (Nⴔ49)

1.33 (0.53–3.32) 2.62 (0.95–7.22) 1.16 (0.49–2.74) 0.56 (0.22–1.42) 0.53 (0.14–1.93) 2.79 (1.06–7.35) 0.96 (0.27–3.39) 1.37 (0.60–3.14) 0.78 (0.00–6.29) 1.31 (0.44–3.92) 2.02 (0.63–6.44) — 1.71 (0.66–4.44) 0.60 (0.00–4.79) 1.56 (0.60–4.09)

1.27 (0.94–1.72) 1.03 (0.75–1.42) 0.96 (0.72–1.29) 1.19 (0.85–1.66) 0.83 (0.56–1.21) 1.42 (0.96–2.09) 1.39 (0.92–2.09) 1.47 (1.11–1.95) 1.25 (0.65–2.39) 1.80 (1.26–2.57) 0.67 (0.40–1.12) 3.36 (2.09–5.38) 1.31 (0.94–1.83) 1.02 (0.57–1.81) 1.43 (1.02–1.99)

0.61 (0.32–1.18) 1.32 (0.68–2.57) 1.27 (0.68–2.34) 0.78 (0.39–1.57) 0.58 (0.25–1.38) 4.27 (2.25–8.13) 0.89 (0.35–2.22) 1.06 (0.58–1.95) 2.62 (0.90–7.61) 1.50 (0.71–3.17) 1.45 (0.62–3.40) 2.54 (0.96–6.68) 1.74 (0.90–3.35) 2.41 (0.99–5.86) 3.04 (1.58–5.86)

Note: N⳱1,915; multinomial logistic-regression model: odds ratios (95% confidence intervals [CIs]); italics: significant difference between strata according to the criterion of mutually-exclusive CIs. Overall-likelihood chi-square test: 135.5; df: 45; p ⬍0.001. GAD: generalized anxiety disorder; DEP: depression; DEP/GAD: depression with anxiety; MMSE: Mini-Mental State Exam; ADL: activities of daily living; IADL: instrumental activities of daily living.

36

Am J Geriatr Psychiatry 13:1, January 2005

Schoevers et al. ing that, also in elderly patients, higher levels of comorbidity indicate more severe psychopathology.8,9 These findings may have important implications. Recent data suggest that DEP/GAD can effectively be treated in elderly patients.32 Early intervention in depressed subjects with comorbid anxiety may significantly reduce chronicity and all of its negative consequences, both for the individual patient and for the use of health services.33 Studying whether specific risk factors were associated with either persistence or remission of baseline psychopathology, or with the development of DEP/ GAD, yielded a limited number of statistically significant associations. The finding that the progression to mixed anxiety-depression in subjects with GAD or DEP was predicted by comorbid medical illness supports earlier research showing that medical comorbidity carries a higher risk for chronicity and worsening of affective disorder.34 The finding that the progression from a single disorder to depression with co-existing GAD was also associated with age below 74 fits the concept of a decrease of this type of vulnerability at a greater age, with the most vulnerable subjects selectively leaving the population at an earlier stage.35–37 Apart from studying prognosis, we also attempted to study the risk profile for onset of any of the three subtypes of psychopathology at follow-up in subjects without baseline psychopathology. GAD was only predicted by a personal history of depression and/or anxiety, but this finding may have been due to relatively low numbers. Onset of both DEP and DEP/ GAD was associated with functional limitations, chronic illnesses, and loss events. This is in line with findings from earlier studies on depression incidence38 that did not make a distinction between depression with or without coexisting anxiety. According to the criterion that the confidence limits should mutually exclude the point estimates of the relative TABLE 4.

risk in the opposite stratum, we found three risk factors that showed statistically significant differences between DEP and DEP/GAD. Again, the youngerold were more likely to have the mixed condition, in line with the idea of more severe pathology. Decreases in functional abilities also showed significantly stronger associations with DEP/GAD than with GAD. The strongest predictor of DEP/GAD was the presence of longstanding, and possibly genetically-defined, vulnerability, exemplified by a personal history of depression and/or anxiety. The current study is an improvement over existing studies on depression and generalized anxiety in elderly persons, in the sense that it uses a prospective design in a community sample, using established measurement instruments and a comprehensive set of risk factors associated with late-life psychiatric disorders. The principal limitation is the fact that symptoms have been assessed two times, over a 3-year interval. Subjects may have seen both relapse and remittance during the study interval, and not all of this variation is necessarily accounted for in the data. Likewise, episodes of both depression and anxiety with a shorter duration are generally underrepresented in studies with a limited number of assessments. This would suggest that, overall, the actual course of depression and GAD in the population may be somewhat less gloomy than this study shows. Still, the level of chronicity is comparable to what is generally found in studies on late-life depression,34 even when they were able to use more frequent measurements.39 A second limitation is the fact that this study is a naturalistic follow-up in the community and did not control for ongoing treatment that may have affected the results. Comparable samples of community-living elderly persons have demonstrated that case-finding is generally poor, with the majority of subjects with de-

Multivariate Logistic Regression

GAD-Only (Nⴔ34) Personal history 2.58 (1.01–6.58)

DEP-Only (Nⴔ250)

DEP/GAD (Nⴔ49)

Personal history: 1.44 (0.99–2.11) Chronic diseases (baseline): 1.45 (1.10–1.91) IADL disability (baseline): 1.78 (1.28–2.48) Loss of spouse: 2.93 (1.93–4.47) Recent IADL decrease: 1.53 (1.12–2.10)

Personal history: 4.48 (2.38–8.38) ADL disability (baseline): 2.84 (1.04–7.77) Loss of spouse: 2.90 (1.22–6.93) Recent ADL decrease: 2.39 (1.01–5.65) Recent IADL decrease: 2.91 (1.56–5.44)

Note: Odds ratios (confidence intervals) in parsimonious model with p[in] ⬍0.05, p[out] ⬍0.10; backstep procedure. GAD: generalized anxiety disorder; DEP: depression; DEP/GAD: depression with anxiety.

Am J Geriatr Psychiatry 13:1, January 2005

37

Co-Occurrence of Depression and GAD pression and/or anxiety not receiving any treatment at all for their symptoms.40,41 Our study is no exception to this. It is however fair to conclude that, if treatment had actually affected prognosis in some subjects, this would have led to an underestimation of the level of chronicity and other adverse consequences in our study. Third, although the total number of subjects included in this study was relatively large, and the prevalence of generalized anxiety (4.7%) was well within the range suggested by earlier findings,2,7,42 stratification of the sample sometimes resulted in low numbers of subjects in the different cells. Especially when analyzing GAD-only, this may have limited the possibility of finding statistically significant differences with the risk-profile of the other categories of psychopathology. A conclusion on specificity of risk factors for GAD can therefore only be tentative. Still, both the temporal patterns and the differences between depression and depression with co-existing GAD were rather pronounced and statistically significant, also, in more detailed analyses. It should be noted, however, that reduction of the alpha level, for example, by using Bonferroni correction because of multiple comparisons, was not done in this study. This would be warranted if a conclusion were based on a single outcome within a larger set of calculations. However, when assessing multiple risk factors and searching for risk-profiles with an a-priori hypothesis, this method may actually mask valid results, especially when numbers are limited. A fourth limitation is that selective attrition to the follow-up assessment of the study has occurred among the very frail participants. Subjects with cognitive impairment and those with lower educational level were more likely to be nonresponders at followup. This a normal pattern in epidemiological studies. If this has affected the results, it again would lead to

an underestimation of the actual situation in the community with regard to prevalence and prognosis of the disorders under study. The current study provides arguments that may be relevant to the debate on a dimensional-versus-categorical interpretation of GAD and DEP. The pattern of temporal sequencing, with GAD often progressing to DEP or DEP/GAD, can be taken as an indication favoring a communal concept of anxiety and depression. Studying the specificity of risk factors for onset of either category, reflective of the “same genes, different environments” hypothesis put forward by Kendler et al.,14,15,43 however, did not show that the risk-profile for DEP differed from that for GAD. Although this may in part be explained by smaller numbers of GAD cases, our data do not provide support for the idea that anxiety and depression may be associated with different risk-profiles. Further study is however needed to confirm this. The finding that a longstanding, possibly genetic vulnerability is most strongly associated with DEP/ GAD does appear to be congruent with the findings of Kendler et al.15 in twin studies. They suggested a strong and identical contribution to both GAD and DEP by genetic factors. Given that previous research showed that comorbidity indicates higher severity levels of both DEP and GAD, it is in line with this view that these associations were more prominent in subjects with the mixed condition. Overall, our data show that, in elderly patients, both DEP and GAD often have a chronic course, in which GAD often progresses to DEP or DEP/GAD. Comorbidity is associated with longstanding vulnerability, ill health, and functional decline, and has a gloomy prognosis. These findings may be of considerable relevance for both our theoretical understanding of these commonly-occurring disorders in later life, and for preventive purposes.

References 1. Regier DA, Rae DS, Narrow WE, et al: Prevalence of anxiety disorders and their comorbidity with mood and addictive disorders. Br J Psychiatry 1998; (suppl 34):24–28 2. Beekman AT, de Beurs E, van Balkom AJ, et al: Anxiety and depression in later life: co-occurrence and communality of risk factors. Am J Psychiatry 2000; 157:89–95 3. Kessler RC, Nelson CB, McGonagle KA, et al: Comorbidity of DSM-III-R major depressive disorder in the general population: results from the U.S. National Comorbidity Survey. Br J Psychiatry 1996; (suppl 30):17–30 4. Liebowitz MR, Hollander E, Schneier F, et al: Anxiety and de-

38

pression: discrete diagnostic entities? J Clin Psychopharmacol 1990; 10(suppl 3):61S–66S 5. Bijl RV, Ravelli A: Current and residual functional disability associated with psychopathology: findings from The Netherlands Mental Health Survey and Incidence Study (NEMESIS). Psychol Med 2000; 30:657–668 6. Kessler RC, McGonagle KA, Zhao S, et al: Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States: results from The National Comorbidity Survey. Arch Gen Psychiatry 1994; 51:8–19 7. Flint AJ: Epidemiology and comorbidity of anxiety disorders in the elderly. Am J Psychiatry 1994; 151:640–649

Am J Geriatr Psychiatry 13:1, January 2005

Schoevers et al. 8. Lenze EJ, Mulsant BH, Shear MK, et al: Comorbidity of depression and anxiety disorders in later life. Depress Anxiety 2001; 14:86– 93 9. Schoevers RA, Beekman AT, Deeg DJ, et al: Comorbidity and risk patterns of depression, generalised anxiety disorder and mixed anxiety-depression in later life: results from the AMSTEL Study. Int J Geriatr Psychiatry 2003; 18:994–1001 10. Charney DS, Reynolds CF III, Lewis L, et al: Depression and Bipolar Support Alliance Consensus Statement on Unmet Needs in Diagnosis and Treatment of Mood Disorders in Late Life. Arch Gen Psychiatry 2003; 60:664–672 11. Tyrer P: Comorbidity or consanguinity. Br J Psychiatry 1996; 168:669–671 12. Zinbarg RE, Barlow DH, Liebowitz M, et al: The DSM-IV field trial for mixed anxiety-depression. Am J Psychiatry 1994; 151:1153– 1162 13. Goldberg D, Huxley PJ: Common Mental Disorders: A Bio-Social Model. New York, Routledge, 1992 14. Kendler KS, Heath AC, Martin NG, et al: Symptoms of anxiety and symptoms of depression: same genes, different environments? Arch Gen Psychiatry 1987; 44:451–457 15. Kendler KS: Major depression and generalised anxiety disorder: same genes, (partly) different environments, revisited. Br J Psychiatry 1996; (suppl 30):68–75 16. Merikangas KR, Angst J, Eaton W, et al: Comorbidity and boundaries of affective disorders with anxiety disorders and substance misuse: results of an international task force. Br J Psychiatry 1996; (suppl 30):58–67 17. Launer LJ, Dinkgreve MA, Jonker C, et al: Are age and education independent correlates of the Mini-Mental State Exam performance of community-dwelling elderly? J Gerontol 1993; 48:271– 277 18. Schoevers RA, Beekman AT, Deeg DJ, et al: Risk factors for depression in later life: results of a prospective, community-based study (AMSTEL). J Affect Disord 2000; 59:127–137 19. Launer LJ, Wind AW, Deeg DJ: Nonresponse pattern and bias in a community-based, cross-sectional study of cognitive functioning among the elderly. Am J Epidemiol 1994; 139:803–812 20. Folstein MF, Folstein SE, McHugh PR: “Mini-Mental State:” a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975; 12:189–198 21. Copeland JR, Dewey ME, Griffiths-Jones HM: A computerized psychiatric diagnostic system and case nomenclature for elderly subjects: GMS and AGECAT. Psychol Med 1986; 16:89–99 22. Katz S, Ford AB, Moskowitz RW, et al: Studies of illness in the aged: The Index of ADL: a standardized measure of biological and psychosocial functioning. JAMA 1963; 185:914–919 23. Lawton MP, Brody EM: Assessment of older people: self-maintaining and instrumental activities of daily living. Gerontologist 1969; 9:179–186 24. Roth M, Tym E, Mountjoy CQ, et al: CAMDEX: a standardised instrument for the diagnosis of mental disorder in the elderly, with special reference to the early detection of dementia. Br J Psychiatry 1986; 149:698–709 25. Copeland JR, Dewey ME, Henderson AS, et al: The Geriatric Mental State (GMS) used in the community: replication studies of the computerized diagnosis AGECAT. Psychol Med 1988; 18:219–223 26. Hooijer C, Jonker C, Dewey ME: A standardized interview for the

Am J Geriatr Psychiatry 13:1, January 2005

elderly (GMS): reliability studies comparing the Dutch Language Version with the original. Int J Geriatr Psychiatry 1991; 6:71–79 27. Copeland JR, Dewey ME, Wood N, et al: Range of mental illness among the elderly in the community: prevalence in Liverpool using the GMS-AGECAT package. Br J Psychiatry 1987; 150:815– 823 28. Copeland JR, Dewey ME, Griffith-Jones HM: Dementia and depression in elderly persons: AGECAT compared with DSM-III and pervasive illness. Int J Geriatr Psychiatry 1990; 5:47–51 29. Copeland JR, Davidson IA, Dewey ME, et al: Alzheimer’s disease, other dementias, depression, and pseudodementia: prevalence, incidence, and three-year outcome in Liverpool. Br J Psychiatry 1992; 161:230–239 30. Gardner MJ, Altman DG: Confidence intervals rather than p values: estimation rather than hypothesis-testing. Br Med J (Clin. Res. Ed.) 1986; 292(6522):746–750 31. de Graaf R, Bijl RV, Spijker J, et al: Temporal sequencing of lifetime mood disorders in relation to comorbid anxiety and substance use disorders: findings from The Netherlands Mental Health Survey and Incidence Study. Soc Psychiatry Psychiatr Epidemiol 2003; 38:1–11 32. Lenze EJ, Mulsant BH, Dew MA, et al: Good treatment outcomes in late-life depression with comorbid anxiety. J Affect Disord 2003; 77:247–254 33. Katon WJ, Lin E, Russo J, et al: Increased medical costs of a population-based sample of depressed elderly patients. Arch Gen Psychiatry 2003; 60:897–903 34. Cole MG, Bellavance F, Mansour A: Prognosis of depression in elderly community and primary-care populations: a systematic review and meta-analysis. Am J Psychiatry 1999; 156:1182–1189 35. van Ojen R, Hooijer C, Jonker C, et al: Late-life depressive disorder in the community: early onset and the decrease of vulnerability with increasing age. J Affect Disord 1995; 33:159–166 36. Schoevers RA, Geerlings MI, Beekman AT, et al: Association of depression and gender with mortality in old age: results from The Amsterdam Study of the Elderly (AMSTEL). Br J Psychiatry 2000; 177:336–342 37. Penninx BW, Geerlings SW, Deeg DJ, et al: Minor and major depression and the risk of death in older persons. Arch Gen Psychiatry 1999; 56:889–895 38. Cole MG, Dendukuri N: Risk factors for depression among elderly community subjects: a systematic review and meta-analysis. Am J Psychiatry 2003; 160:1147–1156 39. Beekman AT, Geerlings SW, Deeg DJ, et al: The natural history of late-life depression: a six-year, prospective study in the community. Arch Gen Psychiatry 2002; 59:605–611 40. Beekman AT, Copeland JR, Prince MJ: Review of community prevalence of depression in later life. Br J Psychiatry 1999; 174:307–311 41. German PS, Shapiro S, Skinner EA, et al: Detection and management of mental health problems of older patients by primarycare providers. JAMA 1987; 257:489–493 42. Beekman AT, Bremmer MA, Deeg DJ, et al: Anxiety disorders in later life: a report from the Longitudinal Aging Study, Amsterdam. Int J Geriatr Psychiatry 1998; 13:717–726 43. Kendler KS, Neale MC, Kessler RC, et al: Major depression and generalized anxiety disorder: same genes, (partly) different environments? Arch Gen Psychiatry 1992; 49:716–722

39