Residual symptoms in older patients treated for major depression

INTERNATIONAL JOURNAL OF GERIATRIC PSYCHIATRY

Int J Geriatr Psychiatry 2005; 20: 1196–1202.

Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/gps.1418

Residual symptoms in older patients treatedfor major depression

Celia F. Hybels*, David C. Steffens, Douglas R. McQuoid and K. Ranga Rama Krishnan

Department of Psychiatry and Behavioral Sciences, Center for the Study of Aging, Duke University Medical Center,NC, USA

SUMMARY

Objective The purpose of this study was to identify residual symptoms in a sample of older adults treated for majordepression and compare individual symptoms present at baseline with those at three months by remission status.Methods The sample was comprised of 229 patients with DSM-IV major depression who were participants in the NIMHMental Health Clinical Research Center at Duke University. Symptoms were measured using the Montgomery-AsbergDepression Rating Scale (MADRS).Results At three months, 86 patients (37.6%) had remitted, or had a MADRS score less than or equal to 9. In the remittedgroup, the most frequently reported symptoms at three months were inner tension and lassitude. Among nonremitters, themost frequently reported symptoms were reported and apparent sadness, as well as lassitude and inner tension. In the sampleas a whole, the symptoms most likely to be present at baseline but not three months were pessimistic and suicidal thoughts,while the most frequently reported emergent symptoms were reduced appetite and inner tension. Patients were much morelikely to no longer have a particular symptom than to acquire a new symptom. Overall, the symptoms present at three monthswere not severe in either group.Conclusions In older adults treated for major depression, residual symptoms at three months may include emergent symp-toms as well as persistent symptoms, and are likely to include symptoms of anxiety as well as sadness. These findings haveclinical implications for the treatment of late-life depression. Copyright # 2005 John Wiley & Sons, Ltd.

key words— major depression; residual symptoms; elderly

INTRODUCTION

Residual symptoms in adults treated for major depres-sion are common. In clinical samples across all agegroups, as many as one-third of patients treated formajor depression have partial remission with residualsymptoms (Paykel et al., 1995; Cornwall and Scott,1997). Patients with residual symptoms after remis-sion are more likely to relapse or relapse earlier thanthose without (Faravelli et al., 1986; Paykel et al.,1995; Judd et al., 1998; Van Londen et al., 1998;Kanai et al., 2003).

Residual symptoms among those with a history ofmajor depression are associated with lifetimeincreased use of medical and psychiatric outpatientservices, psychiatric inpatient services, emergencyservice use, disability benefits, thoughts of complet-ing suicide and attempted suicide (Judd et al., 1997)and with current functional impairment (Judd et al.,1996). Kennedy and Paykel (2004) recently examinedthe long-term outcome of patients treated for majordepression and found patients with residual symp-toms continued to have more symptoms over fol-low-up, although not at full criteria for majordepression, and showed greater impairment in socialfunctioning. There were no differences regarding thenumber of recurrences, readmission, chronic epi-sodes, or clinical global outcome. Among primarycare patients with depression, partial remission withresidual symptoms was associated with residual dis-ability (Ormel et al., 1993).

Received 1 February 2005Copyright # 2005 John Wiley & Sons, Ltd. Accepted 20 June 2005

*Correspondence to: C. F. Hybels, Department of Psychiatry andBehavioral Sciences, Center for the Study of Aging, DukeUniversity Medical Center, Box 3003, Durham NC 27710, USA.Tel: (919) 660-7546. Fax: (919) 684-8569.E-mail: [email protected]

Contract/grant sponsor: NIMH; contract/grant numbers: K01MH066380, K24 MH70027, R01 MH54846.

Studies of residual symptoms in older adults areless common. Brodaty et al. (1993) reported residualsymptoms in 38% of elderly depressives at 1 year and20% at 2–4 years. Gasto et al. (2003) reported 82.3%of older patients treated for severe major depressionin remission at nine months showed residual symp-toms, and that medical burden, chronic stress, andsubjective social support predicted the severity ofthese symptoms. Chronicity and persistence ofdepressive symptoms are also associated with declin-ing health and disability in older adults (Kennedyet al., 1991). An earlier analysis of the data used herefound residual symptoms predicted relapse (Steffenset al., 2003).

Much less is known about which symptoms aremost likely to be residual. Identification of suchsymptoms could allow treatment to be targetedtoward specific symptoms to potentially reduce theadverse outcomes associated with residual symptoms.Paykel et al. (1995) found in a younger sample thatresidual symptoms were generally mild and typicalof depression, such as depressed mood, guilt, hope-lessness, impairment of work and activities, anorexia,early insomnia, and anxiety. Mean symptom scoresdifferentiated those with and without residual symp-toms with the exception of symptoms typical ofsevere depression, which were not prevalent in eithergroup. Opdyke et al. (1996/1997) studied the effect ofcontinuation treatment on residual symptoms in olderpatients with depression over four months and foundvariability was most apparent in depressed mood,apathy, psychological and somatic anxiety, feelingsof guilt, anergia, insomnia, and loss of libido. Othersymptoms including psychomotor disturbance, agita-tion, hypochondriasis, loss of appetite, loss of weight,suicidality, and loss of insight showed evidence ofclearer resolution. Gasto et al. (2003) recentlyreported depressed mood, low interest in work andactivities, insomnia, and somatic anxiety were themost frequent symptoms in elderly remitters at ninemonths.

These studies leave several unanswered ques-tions. It is not clear whether these residual symp-toms in older patients in remission are persistentor emergent symptoms, knowledge of which couldguide treatment. Less is known whether residualsymptoms in patients in remission differ from thosesymptoms in patients still depressed, and whetherthe differences are in the presence of the symptomsthemselves or in the severity of the symptoms. Thepurpose of this study was to identify and describeresidual symptoms in a group of older patients trea-ted for major depression and compare symptoms

among those patients in remission to those patientsnot in remission at three months. We also sought toidentify those symptoms present both at baselineand three months, those reported at three monthsbut not at baseline, and those successfully resolved,and whether these symptom changes differed byremission status.

METHOD

Sample

Patients were inpatients and outpatients 60 years ofage or older who met DSM-IV (American PsychiatricAssociation, 1994) criteria for major depression andwere participants in the NIMH Mental Health ClinicalResearch Center (MHCRC) for the Study of Depres-sion in Later Life at Duke University. Patients wereexcluded if any of these conditions were met: (1)any comorbid major psychiatric illness; (2) activealcohol or drug abuse or dependence; (3) any primaryneurologic illness; and (4) metal in the body whichprecluded magnetic resonance imaging (MRI) of thebrain. Patients were recruited through clinician refer-ral from psychiatry and primary care clinics at Duke.The sample included both new (incident) and recur-rent (prevalent) cases.

The study is ongoing and now in its tenth year, andutilizes a prospective cohort design with patientsundergoing naturalistic treatment. Specifically,patients were treated according to a guideline basedalgorithm, the Duke Somatic Treatment Algorithmfor Geriatric Depression (STAGED) approach(Steffens et al., 2002). Treatment was determinedaccording to history of antidepressant use and severityof depression at time of enrollment. Patients wereclassified as follows: (1) no previous medication trialsand not currently on medications; (2) previous posi-tive medication trial and not currently on medica-tions; (3) single failed trial of medication; (4) failedat least two trials; and (5) multiple treatment failures.Patients classified in stages 3–5, 58% of the sample,were taking medications at enrollment. Medicationsprescribed during the study included selective seroto-nin reuptake inhibitors (SSRIs), tricyclic antidepres-sants (TCAs), monoamine oxidase inhibitors(MAOIs), Lithium, and other antidepressants. Somepatients also received electroconvulsive therapy(ECT) or psychotherapy. The study was described tothe patients at enrollment, and all patients providedwritten informed consent. The research protocol wasapproved by the Duke University Institutional ReviewBoard (IRB) and is reviewed annually.

residual symptoms in older patients treated for major depression 1197

Copyright # 2005 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2005; 20: 1196–1202.

Measures

For this analysis, we compared the symptom endorse-ment and score totals using the Montgomery-AsbergDepression Rating Scale (MADRS) (Montgomeryand Asberg, 1979) at baseline and three monthspost enrollment. The MADRS rates each of tendepressive symptoms on a six-point scale. RawMADRS scores were used in the analyses, but forsome comparisons of symptom reporting, we classi-fied the symptoms as present (code 1–6) vs absent(code 0). Remission was defined as a MADRS scoreof 9 or less (Hawley et al., 2002) at 3 months. Wechose 3 months post enrollment as the examinationpoint to minimize risk of remission and subsequentrelapse prior to examination.

Analysis procedures

All analyses were run using SAS analysis software(SAS Institute, 1998). As directed by our researchquestions, statistical analyses used were eitherdescriptive univariate analyses or bivariate compari-sons. All tests were two-tailed, and � was set at0.01 to adjust for multiple tests.

RESULTS

A total of 283 patients had been enrolled at the start ofthis analysis. Three patients had not been in the studyfor three months, leaving 280 eligible patients. A totalof 14 patients dropped out prior to three months, and34 patients were missing 3-month MADRS scores butremained in the study. Finally, three patients weremissing one or more MADRS items, leaving an ana-lysis sample of 229 patients. The sample was predo-minantly female (69%) and white (88%), with a meanage of 69.8 years (std¼ 7.6). The patients on averagehad 13.3 years of education (std¼ 3.2; range 0–17).

A total of 86 patients, 37.6% of our sample, hadremitted at three months. Those patients who hadremitted did not differ from those who had notremitted in terms of demographic characteristics(age, race, sex) or in health status. Specially, to assesscomorbid medical illness, we compared the percen-tage of patients in the two groups reporting twelvechronic conditions, and group differences were notsignificant for any of the conditions. A total of 25%of the nonremitters compared to 19% of the remittersreported heart disease ( p¼ 0.24), 40% of the nonre-mitters compared to 47% of the remitters reportedhypertension ( p¼ 0.35), and 6% of the nonremitterscompared to 10% of the remitters reported a historyof stroke (p¼ 0.09).

The two groups did not differ in clinical variables atbaseline, including age of onset and number of life-time depression spells, as well as mean MADRS scoreat baseline (27.8 in the remitted group compared to28.3 in the nonremitted group, t-test with 227 df,p¼ 0.67). While both groups showed a lowerMADRS score at three months compared with base-line, subjects who had remitted, by definition, hadlower scores than those who had not remitted(mean¼ 3.7, std¼ 2.9 and mean 18.4, std¼ 6.9respectively; t-test with 206 df; p< 0.0001). Thetwo groups did not differ in the types of antidepres-sant medications received during the three months.A total of 33.7% of those who had remitted receivedSSRIs, compared to 31.5% of those who had not.Similarly, the proportion receiving TCAs was 2.3%in the remitted group and 3.5% in the nonremittedgroup, receiving other antidepressants 24.4% in theremitted group and 25.2% in the nonremitted group,and receiving combination medications 22.1% in theremitted group and 32.9% in the nonremitted group.While none of these differences was statistically sig-nificant, a higher percentage of patients (93%) in thenonremitted group received any antidepressant medi-cation than in the remitted group (83%) (p¼ 0.01).Eleven percent of the patients did not receive any anti-depressant medication in the first three months.

At baseline, the most frequently endorsed symp-toms were reported sadness, apparent sadness, andinability to feel (loss of interest/anhedonia). At threemonths, the three most reported symptoms werelassitude, inner tension, and reported sadness(see Table 1). In the remitted group, the most fre-quently reported symptoms were inner tension andlassitude, while in the nonremitters, the most reportedsymptoms were reported sadness, apparent sadness aswell as lassitude and inner tension. While the propor-tion of remitted patients reporting sadness at 3 monthswas over 30%, these results suggest the residualsymptoms in the remitted group may be somewhatdifferent than those observed in the nonremittedgroup, with the remitters having less apparent andreported sadness. The high percentage of patients inboth groups endorsing lassitude and inner tensionsuggests although symptoms relating to depressedmood may be less prevalent when the depression istreated, symptoms of anxiety and difficulties initiatingactivities are common.

We compared the odds of having a particular symp-tom in the group that had not remitted compared tothe remitted group. As shown in Table 1, the oddsof reporting each symptom are higher in the nonre-mitted group, although the estimates are imprecise

1198 c. f. hybels ET AL.


as noted by the width of the confidence intervals.These results suggest overall there is not a symptomwhich has a greater probability of being reported bysubjects who had remitted compared to those whohad not remitted.

We examined the severity of the residual symptomsamong those in both groups still reporting the symp-tom using the symptom score as a continuous vari-able. Overall, the symptoms, if present at threemonths, were not severe. In the remitted group, theMADRS symptom reported with the highest meanscore at three months was reduced appetite(mean¼ 2.00, std¼ 0.58), while among those patientswho had not remitted, the MADRS symptom with thehighest mean score was reported sadness(mean¼ 2.60, std¼ 1.05).

The mean baseline score and percent change inseverity from the baseline score for each symptomis shown in Table 2, using each symptom score as acontinuous variable. Specifically, we computed the

percent change in symptom score for each individualand present here mean percent change for each symp-tom across all individuals. Overall, the symptoms thatshowed the most change in severity were apparentsadness, inability to feel, and reported sadness.Among the patients who had remitted, the symptomsthat showed the greatest percent change were appar-ent and reported sadness, while the symptom thatshowed the smallest change was suicidal thoughts,most likely because this symptom was not severe atbaseline. Among those patients who had not remitted,the symptoms that showed the greatest change wereapparent sadness, reduced sleep, and inability to feel,while the symptom that showed the smallest percentchange was inner tension. For all symptoms, changein severity was greater in the group that had remittedby three months.

Overall, the proportion of patients reporting emer-gent symptoms was small. A total of 99 patients hadall MADRS symptoms at baseline. Of the remaining

Table 1. Number and percent of patients with individual MADRS symptoms at three months by remission status (n¼ 229)

Total Remitted Not remitted Test p-value OR (95% CI)(n¼ 229) MADRS � 9 MADRS > 9 statistic*

(n¼ 86) (n¼ 143)

Apparent sadness 160 (69.9%) 22 (25.6%) 138 (96.5%) 128.3 p< 0.0001 80.3(29.1, 221.6)Reported sadness 166 (72.5%) 27 (31.4%) 139 (97.2%) p< 0.0001 75.9 (25.4, 226.6)Inner tension 169 (73.8%) 36 (41.9%) 133 (93.0%) 72.7 p< 0.0001 18.5 (8.5, 40.0)Reduced sleep 127 (55.5%) 26 (30.2%) 101 (70.6%) 35.5 p< 0.0001 5.5 (3.1, 10.0)Reduced appetite 66 (28.8%) 7 (8.1%) 59 (41.3%) 28.7 p< 0.0001 7.9 (3.4, 18.4)Inability to concentrate 143 (62.5%) 26 (30.2%) 117 (81.8%) 60.9 p< 0.0001 10.4 (5.6, 19.4)Lassitude 171 (74.7%) 36 (41.9%) 135 (94.4%) 78.4 p< 0.0001 23.4 (10.2, 53.9)Inability to feel 146 (63.8%) 20 (23.3%) 126 (88.1%) 97.8 p< 0.0001 24.5 (12.0, 49.8)Pessimistic thoughts 119 (52.0%) 12 (14.0%) 107 (74.8%) 79.7 p< 0.0001 18.3 (8.9, 37.6)Suicidal thoughts 69 (30.1%) 5 (5.8%) 64 (44.8%) 38.7 p< 0.0001 13.1 (5.0, 34.3)

*Chi-square with 1 df used to compare proportions; Fisher’s Exact test was used for reported sadness due to small cell counts.

Table 2. Mean symptom score at baseline and mean percent change in item scores for individual symptoms from baseline to three monthsby remission status (n¼ 229)

Mean score Mean percent Mean percent Mean percent Test statistic df p-valueat baseline change total change in change in group (t-test)

(std) sample remitted group not remitted(n¼ 229) (n¼ 86) (n¼ 143)

Apparent sadness 3.4 (1.1) 39.7% 66.2% 23.8% �13.47 227 p< 0.0001Reported sadness 3.5 (1.1) 36.2% 66.4% 18.1% �14.81 218 p< 0.0001Inner tension 2.9 (1.2) 22.1% 54.8% 2.4% �8.48 221 p< 0.0001Reduced sleep 3.2 (1.3) 35.6% 58.8% 21.6% �6.74 218 p< 0.0001Reduced appetite 1.9 (1.8) 20.9% 47.1% 5.2% �5.49 212 p< 0.0001Inability to concentrate 2.9 (1.1) 29.5% 59.0% 11.7% �6.79 221 p< 0.0001Lassitude 3.3 (1.1) 31.1% 57.8% 15.1% �10.02 227 p< 0.0001Inability to feel 3.1 (1.1) 37.8% 64.8% 21.6% �10.45 206 p< 0.0001Pessimistic thoughts 2.4 (1.3) 33.3% 58.5% 18.2% �7.28 204 p< 0.0001Suicidal thoughts 1.5 (1.3) 25.7% 40.9% 16.6% �3.76 227 p¼ 0.0002



130 patients, 39 (30%) reported one or more emergentsymptoms. Across the entire sample, the symptomsmost frequently reported at three months but not atbaseline were reduced appetite and inner tension. Atotal of 15 patients (6.6%) reported reduced appetiteas an emergent symptom and nine patients (3.9%)reported inner tension. The symptoms most likely tobe present at baseline but not at 3 months were pessi-mistic and suicidal thoughts. A total of 102 patients(44.3%) no longer reported pessimistic thoughts and107 (46.7%) reported that suicidal thoughts hadresolved. Finally, the symptoms most frequentlyreported at both assessments were lassitude, reportedsadness, inner tension, and apparent sadness. Wecomputed a McNemar chi-square test with 1 df foreach symptom for the two measurements, with thenull hypothesis being that an individual’s responsefor each symptom would be the same at the two pointsin time (Stokes et al., 1995) and were able to reject thenull hypothesis for each symptom (p< 0.0001) for thesample as a whole, indicating significant symptomchange for all symptoms.

DISCUSSION

This study reports several new findings. At 3months, approximately one-third of our olderpatients had remitted, a proportion consistent withstudies of younger adults (Paykel et al., 1995;Cornwall and Scott, 1997), supporting researchindicating the recovery rate in older patients is simi-lar to that in younger patients (Alexopoulos et al.,1996). In the remitted group, the most frequentlyreported symptoms at 3 months were inner tensionand lassitude, while in the non-remitted group, themost endorsed symptoms were reported and appar-ent sadness, although lassitude and inner tensionwere also frequently reported. At 3 months, residualsymptoms in both groups included symptoms notreported at baseline as well as persistent symptoms,but overall, patients were more likely to no longerhave a particular symptom than to acquire a newsymptom. Residual symptoms at 3 months, if pre-sent, were not severe. The symptoms that showedthe greatest change in severity from baseline wereapparent and reported sadness and inability to feel.Less change in severity was due, in part, to lessseverity at baseline.

This finding that inner tension was a frequentlyreported residual symptom is consistent with earlierfindings reported by Blazer et al. (1989). Specifically,they found among patients hospitalized for depres-sion, 32.1% of the patients reported anxiety 1–2 years

following hospitalization. A total of 68.5% of thosewith no anxiety at follow-up reported they had fullyrecovered from their index episode, compared to40.0% of those with early morning anxiety and only15.4% of those who reported both morning and eve-ning anxiety (Blazer et al., 1989). Our finding thatanxiety may be reduced but still persist has potentialtreatment implications. In a recent report from ourstudy, patients with generalized anxiety symptomshad a longer time to remission (Steffens andMcQuoid, 2005). In a similar study, among elderlypatients with major depression, anxiety at the timeof the index episode did not predict outcome, butthose patients who remained anxious at the time ofremission of the index episode had a significantlyshorter time to relapse/recurrence (Flint and Rifat,1997). These findings also confirm that more researchis needed in the area of comorbid anxiety and depres-sion and how anxiety affects the course of late lifedepression (Devanand, 2002), and that co-occurenceof depression and anxiety disorders are common inthis population (Beekman et al., 2000). Treatmentaimed at lassistude may also be helpful, and there issuggestion that treating fatique as a residual symptomwill be beneficial (Menza et al., 2003).

Gasto et al. (2003) found depressed mood was themost frequent symptom in elderly remitters at 9months, reported by 38.4% of their sample. Whilenot the most common symptom in our sample ofremitters, 31.4% reported sadness at 3 months, a pro-portion somewhat similar to their results.

One of the surprising findings from our study wasthat 30.1% of our sample reported suicidal thoughtsdespite being in treatment for depression for at leastthree months. There were only eight patients (3.5%)who had emergent suicidal thoughts at three months,so the majority of patients with this symptom had per-sistent thoughts. Suicidal thoughts at baseline, if pre-sent, were not severe (mean at baseline 1.6 for thosewho had remitted and 1.5 for those who had not), inagreement with clinical observation that the majorityof these were ‘passive’ thoughts, noted but not requir-ing special treatment. Both groups showed a lowermean score for this symptom at three months. Changein suicidal thoughts was not associated with type ofantidepressant in our sample.

This research has several limitations. Remissionwas defined by MADRS score only and not bythe clinician’s assessment. In addition, the MADRSwas given at baseline and at 3 months, so we can-not determine if the symptoms were continuousor intermittent. This was a naturalistic treatmentstudy in which study psychiatrists followed a



treatment algorithm allowing for use of a variety ofantidepressants alone or in combination. A morerigid treatment protocol would have allowed us todiscount medication-specific variables such as treat-ment intensity and compliance as a source of varia-bility in outcome. We may also not have accountedfor medication side effects that mimic residualsymptoms. For example, lassitude in both groupsmight be SSRI induced apathy. We did not have suf-ficient information about the use of nonpsychiatricmedications or nonpharmacologic treatments,including concurrent psychotherapy, which couldaffect symptom reporting. Our sample size wassmall and therefore our ability to detect differenceswas somewhat limited. Our sample is predominantlywhite, well educated and cognitively intact and ourpatients were treated in an academic setting so theseresults may not be applicable to all older adults withmajor depression.

This is an ongoing study that will allow us toexamine these trends in symptom endorsement overthe course of the study. Patients are clinically eval-uated over time rather than relying only on self-report of symptoms. This is a not a naturalistic studyof untreated patients or a highly controlled clinicaltrial, but rather a clinical study employing a treat-ment algorithm used in a geriatric affective disor-ders clinic.

The identification of residual symptoms amongolder patients treated for depression has clinicalimplications in that treatment can focus on specificresidual symptoms. For example, in a study ofdepressed outpatients with a mean age of 43.7 years,patients reported a mean of 2.7 residual symptoms atthe completion of successful treatment. The mostcommon residual symptoms were generalized anxi-ety, somatic anxiety, and irritability. Patients ran-domly assigned to cognitive behavioral treatmenttargeted to residual symptoms had a lower rate ofrelapse at 2 years compared to a group clinically man-aged (Fava et al., 1994). In a similar study, patientswith residual symptoms of major depression wererandomized to clinical management alone or clinicalmanagement plus cognitive therapy. The addition ofCT produced modest improvements in social and psy-chological functioning (Scott et al., 2000).

Overall, these results confirm earlier findings thatsymptom scores alone may be insufficient in definingremission if residual symptoms are present, andthat the goal of antidepressant therapy should be thelowest degree of depressive symptomatology (Tayloret al., 2004). In addition, our findings point to theneed to monitor individual residual symptoms in

patients in remission, and the need for new medica-tions or more effective use of existing medicationsto treat comorbid anxiety in older adults recoveringfrom an episode of major depression.

ACKNOWLEDGEMENTS

This research was supported by NIMH grants K01MH066380, K24 MH70027, and R01 MH54846.The authors would like to thank Dr. Carl Pieper forhis statistical consultation. These results were pre-sented, in part, at the annual meeting of the AmericanAssociation of Geriatric Psychiatry, February 21-24,2004, Baltimore, Maryland.

REFERENCES

Alexopoulos GS, Meyers BS, Young RC, et al. 1996. Recovery ingeriatric depression. Arch Gen Psychiatry 53: 305–312.

American Psychiatric Association. 1994. DSM-IV: Diagnostic andStatistical Manual of Mental Disorders, 4th edn. American Psy-chiatric Association: Washington, DC.

Beekman ATF, de Beurs E, van Balkom AJLM, et al. 2000. Anxietyand depression in later life: Co-occurence and communality ofrisk factors. Am J Psychiatry 157: 89–95.

Blazer D, Hughes DC, Fowler N. 1989. Anxiety as an outcomesymptom of depression in elderly and middle-aged adults. Int JGeriatr Psychiatry 4: 273–278.

Brodaty H, Harris L, Peters K, et al. 1993. Prognosis of depressionin the elderly. Br J Psychiatry 163: 589–596.

Cornwall PL, Scott J. 1997. Partial remission in depressive disor-ders. Acta Psychiatr Scand 95: 265–271.

Devanand DP. 2002. Comorbid psychiatric disorders in late lifedepression. Biol Psychiatry 51: 236–242.

Faravelli C, Ambonetti A, Pallanti S, Pazzagli A. 1986. Depressiverelapses and incomplete recovery from index episode. Am J Psy-chiatry 143: 888–891.

Fava GA, Grandi S, Zielezny M, Canestrari R, Morphy MA.1994. Cognitive behavioral treatment of residual symptoms inprimary major depressive disorder. Am J Psychiatry 151:1295–1299.

Flint AJ, Rifat SL. 1997. Two-year outcome of elderly patients withanxious depression. Psychiatry Res 66: 23–31.

Gasto C, Nevarro V, Catalan R, Portella MJ, Marcos T. 2003.Residual symptoms in elderly major depression remitters. ActaPsychiatr Scand 108: 15–19.

Hawley CJ, Gale TM, Sivakumaran T and the HertfordshireNeuroscience Research Group. 2002. Defining remission bycut off score on the MADRS: selecting the optimal value. J AffectDisord 72: 177–184.

Judd LL, Akiskal HS, Maser JD, et al. 1998. Major depressive dis-order: a prospective study of residual subthreshold depressivesymptoms as predictor of rapid relapse. J Affect Disord 50:97–108.

Judd LL, Akiskal HS, Paulus MP. 1997. The role and clinical sig-nificance of subsydromal depressive symptoms (SSD) in unipo-lar major depressive disorder. J Affect Disord 45: 5–18.

Judd LL, Paulus MP, Wells KB, Rapaport MH. 1996. Socioeco-nomic burden of subsyndromal depressive symptoms and majordepression in a sample of the general population. Am J Psychia-try 153: 1411–1417.



Kanai T, Takeuchi H, Furukawa TA, et al. 2003. Time to recurrenceafter recovery from major depressive episodes and its predictors.Psychol Med 33: 839–845.

Kennedy GJ, Kelman HR, Thomas C. 1991. Persistence and remis-sion of depressive symptoms in late life. Am J Psychiatry 148:174–178.

Kennedy N, Paykel ES. 2004. Residual symptoms at remissionfrom depression: impact on long-term outcome. J Affect Disord80: 135–144.

Menza M, Marin H, Opper RS. 2003. Residual symptoms in depres-sion: can treatment be symptom-specific? J Clin Psychiatry 64:516–523.

Montgomery SA, Asberg M. 1979. A new depression scaledesigned to be sensitive to change. Br J Psychiatry 134: 382–389.

Opdyke KS, Reynolds III CF, Frank E, et al. 1996/1997. Effect ofcontinuation treatment on residual symptoms in late-life depres-sion: how well is ‘well’? Depress Anxiety 4: 312–319.

Ormel J, Oldehinkel T, Brilman E, vanden Brink W. 1993. Outcomeof depression and anxiety in primary care. Arch Gen Psychiatry50: 759–766.

Paykel ES, Ramana R, Cooper Z, Hayhurst H, Kerr J, Barocka A.1995. Residual symptoms after partial remission: an importantoutcome in depression. Psychol Med 25: 1171–1180.

SAS Institute. 1998. Statistical Analysis System, Version 8. SASInstitute: Cary, NC.

Scott J, Teasdale JD, Paykel ES, et al. 2000. Effects of cognitivetherapy on psychological symptoms and social functioning inresidual depression. Br J Psychiatry 177: 440–446.

Steffens DC, McQuoid DR. 2005. Impact of symptoms of general-ized anxiety disorder on the course of late-life depression. Am JGeriatr Psychiatry 13: 40–47.

Steffens DC, McQuoid DR, Krishnan KR. 2002. The Duke SomaticTreatment Algorithm for Geriatric Depression (STAGED)approach. Psychopharmacol Bull 36: 58–68.

Steffens DC, McQuoid DR, Krishnan KRR. 2003. Partial responseas a predictor of outcome in geriatric depression. Am J GeriatrPsychiatry 11: 340–348.

Stokes ME, Davis CS, Koch GG. 1995. Categorical Data AnalysisUsing the SAS System. SAS Institute Inc: Cary, NC.

Taylor WD, McQuoid DR, Steffens DC, Krishnan KRR. 2004.Is there a definition of remission in late life depression thatpredicts later relapse? Neuropsychopharmacology 29:2272–2277.

Van Londen L, Molenaar RPG, Goekoop JG, ZwindermanAH, Rooijmans HGM. 1998. Three- to 5-year prospectivefollow-up of outcome in major depression. Psychol Med 28:731–735.



Documents

Residual symptoms in older patients treated for major depression