Long-Term Impact of Residual Symptoms in Treatment-Resistant Depression

Transcription

1 Long-Term Impact of Residual Symptoms in Treatment-Resistant Depression Abebaw Fekadu, MD, PhD, MRCPsych 1 ; Sarah C Wooderson, PhD 2 ; Lena J Rane, MD, MRCPsych 3 ; Kalypso Markopoulou, MD, MRCPsych 4 ; Lucia Poon, RMN 5 ; Anthony J Cleare, PhD, MRCPsych 6 Objective: Although commonly encountered, little work has defi ned the longitudinal course of treatment-resistant depression (TRD) and the infl uence of residual posttreatment symptoms on longer-term outcome. The aim of our study was to assess the impact of posttreatment clinical states on longer-term outcome. Method: Patients (n = 118) with TRD received specialist inpatient treatment and were followed-up for a median of 3 years. Longitudinal outcome dichotomized into good and poor outcome was used as the primary outcome and functional measures were used as secondary outcomes. Results: Among 118 treated patients, 40 (34%) entered clinical remission, 36 (31%) entered partial remission, and 42 (37%) remained in episode at discharge. At follow-up, 35% had longitudinally defi ned poor outcome. Posttreatment clinical status was the main predictor of both poor and good outcome. Nearly 50% of patients achieved postdischarge recovery, and subsequently had longer-term outcome, comparable with patients discharged in remission. Patients who remained in episode posttreatment were more symptomatically and functionally impaired. Conclusion: Posttreatment clinical states are a useful guide to clinicians for projecting the longer-term outcome of patients with TRD. The persistence of residual or syndromal symptoms predicts a poorer longer-term outcome, whereas treatment to remission is associated with better outcomes. Can J Psychiatry. 2011;56(9): Clinical Implications Residual symptoms are important predictors of poorer longer-term outcome. Treatment to full remission is a realistic goal in treatment-resistant illness and should be the goal of treatment, even for TRD. Limitations Our study was based on a clinical sample. The specialist nature of the treatment setting limits generalizability. The follow-up duration was variable. Key Words: treatment-refractory depression, treatment-resistant depression, follow-up study, course, outcome 549

2 Soon after the introduction of the first ADs, it was recognized that a significant proportion of patients treated with ADs fail to show satisfactory response. 1 6 Now, 60 years later, after the deployment of numerous alternative ADs, TRD remains a major public health challenge. It is common 7 10 and is associated with high levels of comorbidity, disability, and cost TRD remains underresearched: its etiology, pathophysiology, and course are poorly understood. Unlike in less chronic and refractory forms of depression, where remission has been advocated as a potentially measurable and clinically meaningful treatment end point, 8,16 18 no explicit treatment end point 17 has been suggested in TRD. The extent to which residual symptoms confer a poorer outcome in TRD is also unclear. In the context of chronic and refractory illness, residual symptoms might be expected, to some extent, and be less concerning than in more acute illnesses. Indeed, given the poor long-term prognosis in TRD, 19 it could be argued that a more realistic aim of treatment in TRD could be amelioration of symptoms and symptom management, rather than symptom remission or cure. Conversely, it may also be that TRD is essentially an extension of ordinary depressive disorder, and that residual symptoms will confer an adverse prognosis in the same way. If that is so, the explicit aim of treatment in TRD should also be for a full remission of symptoms. Our study assessed the use of posttreatment clinical states as potential treatment end points in TRD. We hypothesized that patients discharged still in episode, or only in partial remission rather than full remission, would have poorer longer-term outcomes than patients discharged in remission, in terms of clinical status, functioning, and quality of life outcomes measured, both longitudinally and cross-sectionally. Method Participants The cohort consisted of patients with TRD who had received intensive inpatient treatment in a specialist unit for the treatment of complex mood disorders in the United Abbreviations AD GAF HDRS ICD IQR LIFE PSR antidepressant Global Assessment of Functioning Hamilton Depression Rating Scale International Classifi cation of Diseases interquartile range Longitudinal Interval Follow-up Evaluation Psychiatric Status Rating QIDS-C Quick Inventory of Depressive Symptomatology Clinician Rating TRD treatment-resistant depression Kingdom. Patients had received individualized treatment packages that consisted of pharmacotherapy (mostly using medication combinations), individual and couple psychological therapy, and other physical therapies as indicated. 20 Diagnosis and level of treatment resistance were established within the unit before discharge. Only patients who had failed to respond to at least one adequate dose of ADs were included in the study. Adequacy of treatment was defined according to the Maudsley Prescribing Guideline. 21 Clinical and demographic details are given in Table 1. Outcomes Primary Outcome. A dichotomous summary outcome (good, compared with poor, outcome) was generated based on the available longitudinal data from the LIFE chart. 22 We considered that spending less time in remission than the median percentage time of remission of the whole follow-up cohort, and spending more time in episode than the median percentage time spent in episode by the whole cohort, would constitute a poor outcome. Conversely, we assumed that spending more time in remission than the cohort median and less time in episode than the cohort median would constitute a good outcome. As TRD is a longitudinal condition, summaries based on the longitudinal course were deemed more appropriate as primary outcomes than crosssectional follow-up measures. Secondary Outcomes. The main secondary outcome was the clinical status of the cohort at end of follow-up; that is, whether patients were in full remission, in episode, or significantly symptomatic without being in episode (subthreshold state). Being in full remission was rated when virtually no depressive symptoms were exhibited, irrespective of the state of comorbid symptoms or disorders, such as anxiety and obsessive compulsive disorders. An episode was defined by the occurrence of depressive symptoms sufficient to meet criteria for a depressive episode according to the ICD-10 criteria. 23 Additional secondary outcomes comprised longitudinal symptomatic state and cross-sectional severity of illness, determined using dimensional symptom scales, and functional and quality of life outcomes. The study was approved by the local ethics committee, and participants gave informed consent for participation. Assessment Instruments Clinical Outcome. The LIFE chart was the main assessment tool used to establish longitudinal clinical outcome. This is a well-established follow-up evaluation scale that allows the weekly or monthly symptomatic state of a patient to be rated retrospectively at follow-up intervals of 6 months or longer. 22 Symptoms are rated on a 6- or 7-point scale (the latter a United Kingdom modification), 24,25 called the PSR, and linked with symptoms in the ICD In this study, the 7-point scale was used to record the monthly clinical status of patients for the interval between discharge and followup. The PSR scores range from an asymptomatic state 550 La Revue canadienne de psychiatrie, vol 56, no 9, septembre 2011

3 Long-Term Impact of Residual Symptoms in Treatment-Resistant Depression Table 1 Baseline sociodemographic and clinical characteristics of cohort Discharge status Characteristic Remission n = 40 % Partial remission n = 36 % Episode n = 42 % P a Sex Male b Female Marital status Single b Married Postmarital Employment Employed b Unemployed Retired Mean (SD) Mean (SD) Mean (SD) Age at index admission, years 51.3 (12.9) 47.1 (12.7) 44.9 (11.1) 0.06 Education, years 13.5 (3.6) 13.1 (3.0) 12.7 (3.3) 0.37 Duration of illness, years 22.2 (15.9) 18.5 (13) 16.0 (12.7) 0.17 Duration of admission, weeks 29.3 (21.1) 28.1 (19.6) 23.3 (19.5) 0.37 Duration of follow-up, months 34.2 (18.4) 41.5 (23.3) 41.8 (23.7) 0.22 Median (IQR) Median (IQR) Median (IQR) Duration of index episode, years 2 (3.5) 3 (8) 4 (6) 0.08 c a All tests of signifi cance based on 1-way ANOVA except when specifi ed b Derived from Pearson chi-square test c Derived from Kruskal Wallis test (score of 1) to a severe episode with psychosis or severe impairment (score of 7). The cross-sectional symptomatic severity at the end of the follow-up was ascertained using the 17-item HDRS 26 and the QIDS-C, and self-rated QIDS versions. 27 The HDRS is the most widely used instrument in the study of treatment response in depression, 28 while the QIDS has been recently developed as an alternative with good psychometric properties. 27 Functional Outcome. We administered the GAF, 29 a widely used dimensional scale that allows rating of impairment in psychological, social, and occupational functioning. The GAF scores vary from 0 to 100. Defi nition of Main Clinical States. The main clinical states were determined using the LIFE chart. PSR scores of 1 to 2 were equated with remission, scores of 3 to 4 with partial remission (subthereshold), and scores of 5 to 7 with episode. The PSR was linked with an ICD-10 symptom checklist 30 so that symptomatic state could be matched with the ICD-10 episode status. The term partial remission used at discharge was replaced during follow-up by the term subthreshold because of the difficulty of deciding whether a patient is remitting from an episode or actually in a relapse process. Using the monthly PSR ratings, the percentage of time patients spent in each prespecified clinical state (remission, subthreshold, and episode) was computed, this, in turn, allowing specification of the occurrence of the primary outcome (a poor outcome, compared with a good outcome). Relapse was assessed in patients who entered partial or full remission either at the end of treatment or during the follow-up period. Relapse was defined as the return of an ICD-10 depressive episode (that is, a PSR of 5, 6, or 7) lasting for at least 1 month, after an episode-free period of 1 month or longer. Data Management. SPSS Version 15 (SPSS Inc, Chicago, IL) was used for data entry and analysis. As per the hypothesis, data were stratified according to discharge status for the main analysis. Three main analytic strategies were followed: for simple comparisons of 2 independent groups, independent samples t test or Mann Whitney U test were used; for comparison of multiple groups, ANOVA or Kruskal Wallis tests were used. Pearson chi-square tests were used for categorical data; for the analysis of the main summary outcomes 551

4 Figure 1 Time spent in various clinical states as a factor of clinical state at discharge from unit Time spent, % Time in remission Time in subthreshold Time in episode Remission Subthreshold Episode Clinical status at discharge (good and poor outcome) we used the logistic regression model to quantify the independent effect of discharge status on these outcomes. As our hypothesis was focused on discharge status, different potential confounding variables were progressively added into the initial model in which only discharge status was fitted. Discharge status was either entered as continuous variables based on the PSR scores (that is, scores of 1 to 7) or as categories using dummy variables, depending on the analysis. Finally, we were also interested in patients who remained in episode at the end of treatment, but who later attained partial or full remission during follow-up. We compared this subgroup with patients who remained in a persistent depressive episode, and those who had already reached partial or full remission by the end of treatment. Results Baseline Sociodemographic and Clinical Characteristics Among the 150 patients with TRD who had received treatment and were approached for our study, 118 (78.7%) were successfully followed-up. Reasons for nonparticipation in the study were as follows: 7 refused to take part, 8 were not traced because of change of address, 13 were deceased, and 4 were excluded because of incomplete data. There was no significant difference in any of the baseline characteristics, including illness severity, between participants and nonparticipants, except for duration of inpatient treatment. This was significantly longer for participants (median 24.0 weeks, IQR = 26 weeks) than nonparticipants (median 9.5 weeks, IQR = 21) (P = 0.007). Among the 118 patients, 40 (34%) were discharged in clinical remission, 36 (31%) in partial remission, and 42 (37%) in episode. Most patients were women (75%), married or cohabiting (56%), unemployed (68%), or retired (14%). The median duration of the index depressive episode was 3.0 years (IQR = 7) while the mean (SD) total duration of illness was 18.9 years (14.1). Patients were followed-up for a median of 37.5 (IQR = 37.2) months, with a minimum of 8 months and a maximum of 84 months. Five patients were followed-up for just under a year (1 for 8 months, 2 for 10 months, and another 2 for 11 months). Removing these 5 from the data has only a marginal impact on duration of follow-up. Clinical Outcome Primary Outcome: Longitudinal Course. During followup, the mean (SD) percentage of time patients spent in remission, subthreshold state, and episode was 36% (39.3%), 39% (39.2%), and 25% (33.8%), respectively. The percentage of time spent in different clinical states was significantly different among the groups stratified according to discharge status as shown in Figure 1. Figure 1 depicts the mean percentage of time spent in various clinical states. Thus patients discharged in remission spent significantly more follow-up time in remission (mean 79%, SD 24.8%, compared with patients discharged in partial remission (mean 19%, SD 26.2%) and in episode (mean 16%, SD 28.2%). Similarly, patients discharged in episode spent most of their follow-up time in episode, while patients 552 La Revue canadienne de psychiatrie, vol 56, no 9, septembre 2011

5 Long-Term Impact of Residual Symptoms in Treatment-Resistant Depression Figure 2 Clinical state at end of follow-up as a factor of discharge clinical characteristics Status at final follow-up, % Remission Subthreshold Episode 10 0 Remission Partial remission Episode End of treatment clinical status Figure 3 Functional and clinical status at follow-up (expressed as a mean score) as a function of discharge clinical status Mean score Remission Subthreshold Episode GAF QIDS-C QIDS-SR HDRS Table 2 Clinical and functional outcomes as a factor of discharge clinical status Discharge status, mean (SD) Remission Partial remission Episode P a QIDS-C 5.4 (4.3) 6.7 (5.0) 11.8 (6.4) <0.001 HDRS (5.4) 8.7 (6.6) 15.3 (9.1) <0.001 Beck Anxiety Inventory 17.9 (14.1) 17.3 (13.3) 26.7 (16.8) 0.01 GAF 76.3 (10.2) 70.5 (13.0) 58.3 (18.8) <0.001 a All tests of signifi cance based on 1-way ANOVA 553

6 Table 3 Present state rating scores at discharge as predictors of longitudinal outcome dichotomized into poor and good outcome, in univariate and adjusted models Poor outcome a Good outcome b Model fi t Model fit Variable OR (95% CI) P χ 2 (df) P OR (95% CI) P χ 2 (df) P Unadjusted PSR score c 2.50 ( ) < (5) ( ) < (5) 0.12 Adjusted for Sex 2.50 ( ) < (6) ( ) < (7) 0.23 Education 2.43 ( ) < (7) ( ) < (8) 0.67 Employment 2.46 ( ) < (7) ( ) < (8) 0.25 Marital status 2.50 ( ) < (8) ( ) < (7) 0.19 Age (at admission) 3.08 ( ) < (8) ( ) < (8) 0.11 Life events and (or) social support d 2.46 ( ) < (8) ( ) < (7) 0.28 Duration of admission 2.57 ( ) < (8) ( ) < (8) 0.11 Duration of follow-up 2.59 ( ) < (8) ( ) < (8) 0.15 Fully adjusted PSR score 3.11 ( ) < (8) ( ) < (8) 0.77 a Poor outcome = spending in full remission for less than the median duration of follow-up and spending in episode for longer than the median duration of follow-up b Good outcome = spending in full remission at least for the median duration of follow-up and spending in episode for less than the median duration of follow-up c PSR scored in reverse for the model on good outcome; that is, score of 7 (psychotic depression) replaced by a score of 1 (asymptomatic remission) d Adjusted for life events in model for poor outcome and for social support in model for good outcome discharged in partial remission spent most of their follow-up time in a subthreshold state (Figure 1). Cumulatively, 81% of participants achieved either remission or partial remission at least once; however, the proportion achieving full remission at any given follow-up time was never above 40%. Cross-Sectional Outcome. In cross-sectional assessment, the PSR scores mirrored the longitudinal findings. For the month of the final follow-up assessment: most patients discharged in episode were found to be in episode; most patients discharged in partial remission were in subthreshold state; and most patients discharged in remission were in remission (Figure 2). The assessments on other scales (HDRS and QIDS-C) also confirmed these differential outcomes (Table 2 and Figure 3). Relapse. Although there was no significant difference between the various discharge groups, the proportion relapsing (that is, showing recurrence of a full episode for at least 1 month) was lowest for patients who were discharged in full remission (48%; n = 19). Relapse rates were higher for patients discharged in partial remission (64%; n = 23). Among those discharged still in episode but who entered remission during follow-up, 53% (n = 10) subsequently relapsed. Patients who had been discharged in remission but who then relapsed experienced a shorter relapse episode (median 3.0; IQR = 4 months) than patients who had been discharged in partial remission (median 4.0; IQR = 8) and those who had only entered remission postdischarge (median 5.5; IQR = 13.5). The time to relapse was not statistically significantly different between the groups (log rank χ 2 = 0.436, df = 2, P = 1.66), although patients discharged in remission remained well for a numerically longer period before relapse (median 18.5 months, IQR = 25.5), compared with patients discharged in partial remission (median 14.0 months, IQR = 31.5) and patients discharged in episode but who attained remission postdischarge (median 12.0 months, IQR = 23). Functional Outcome. Details are presented in Table 2 and Figure 3. The mean GAF score overall was 68.2 (SD 16.3) and significantly higher for patients discharged in remission, compared with patients discharged in partial remission (P = 0.04) and those discharged in episode (P < 0.001). Patients discharged in partial remission also had better functioning than those discharged in episode (P = 0.002). 554 La Revue canadienne de psychiatrie, vol 56, no 9, septembre 2011

7 Long-Term Impact of Residual Symptoms in Treatment-Resistant Depression Postdischarge Recovery. Patients who recovered symptomatically, attaining either partial or full remission at any point during follow-up (n = 19) after being discharged in episode (n = 42), had comparably good clinical and functional outcome at follow-up as patients discharged already symptomatically improved (in partial or full remission). This difference is more apparent when the postdischarge recovery group are compared with those with persistent illness: the mean GAF score for pateints with persistent illness was 45.6 (SD 13.0), while for patients who recovered postdischarge it was 71.7 (SD 14.1). This difference was highly significant (mean difference 26.1; 95% CI 34.9 to 17.3; P < 0.001). Similarly there were significant differences between the 2 postdischarge groups clinical illness severity as measured using the HDRS and the QIDS. Compared with patients discharged in partial or full remission, patients discharged in episode but subsequently attained postdischarge recovery showed no statistically significant difference in any of the outcome domains. Prediction of Outcome. Discharge PSR score consistently predicted outcomes (good, compared with poor, outcome) as depicted in Table 3. For poor outcome defined as longer time in episode and shorter time in remission a unit increase in the discharge PSR score was associated with a 3-fold increase in the likelihood of poor outcome (AOR 3.11; 95% CI 1.91 to 5.07) (Table 3). Good outcome was predicted similarly, with minimal change with adjustment for confounders: a unit decrease in the discharge PSR was associated with a near 3-fold increase in the chance of good outcome (AOR 2.74; 95% CI 1.74 to 4.37). Neither life events nor social support confounded prediction of good outcome. Discussion The first main finding to emerge from our study is that even in TRD, treating to full remission offers patients the prospect of a better longer-term outcome in several domains. In the longitudinal clinical domain, patients treated to remission spent significantly less time in episode and more time in remission. In terms of cross-sectional symptom severity, patients discharged in remission had significantly less severe illness at the end of the follow-up period. This was also shown in the functional domain, in which patients discharged in remission had significantly better functional outcomes, both self-rated and measured by clinicians. The second main finding is that, while overall, patients discharged still in episode have the worst longer-term outcome, even among patients discharged in episode, a degree of recovery can be attained in nearly 50% of cases. Such postdischarge recovery was associated with comparable clinical and functional outcome with patients who attained predischarge partial or full remission. A third finding is that, as expected, the worst functional outcomes are found in patients who remain in a persistent depressive episode, with no periods of full or partial recovery (20%; n = 23). Follow-up outcome measures were worse in this persistently unwell group than patients who showed a full or partial response to specialist treatment, and worse than for patients who showed postdischarge recovery. At least 3 main conclusions may be drawn from these findings for clinicians treating patients with TRD. First, the PSR discharge clinical states are a useful guide for projecting the longer-term outcome of the disorder, and thus may be of clinical use. Second, even among patients who are discharged still in episode after treatment in a specialist unit, a reasonable proportion (almost one-half) continue to recover and, following recovery, attain better longer-term outcome trajectories, comparable with patients discharged in remission and partial remission. Finally, full remission is a potentially achievable goal in TRD: the proportion in remission at the time of follow-up was about 70%, 50%, and 30% of patients discharged in remission, partial remission, and episode, respectively. Our study also confirms that full remission may be the best goal of intervention in TRD. Patients discharged in full remission had a clear advantage over those discharged in partial remission in most domains of clinical and functional outcome (Table 2 and Figure 3). Although not reaching statistical significance, a much higher proportion of patients discharged in partial remission (64%) went on to relapse, compared with patients discharged in full remission (48%). There was also a trend for delayed relapse and shorter duration of subsequent episodes in relapse in patients discharged in full remission. Given the small sample size in each subgroup, our study was not adequately powered to detect significant difference in these relapse rates, despite the reasonably large size of the difference. At least in the short-term, studies on the outcome of depression have demonstrated increased risk of relapse among patients treated only to partial remission Full remission was also associated with better functioning at follow-up. Although we anticipate little disagreement that full remission has to be the goal of treatment, even in TRD, it is legitimate to ask whether this is actually achievable. In a limited way, our study suggests that a degree of recovery is achievable, even among people who failed to show expected recovery following intensive treatment in a specialist unit. By the end of the follow-up (at a median of 3 years), 81 people (69%) had achieved either remission or partial remission at some point in time. Conversely, the proportion exhibiting full remission at any one follow-up point was under 40%. However, our data suggests that even when it is difficult to achieve the desired goal of full remission, recovering to a subthreshold level is crucial in improving the longer-term prospect of patients in terms of symptoms and function. Thus failing short of the desired goal of full remission, the second best treatment target appears to be interrupting the persistence of depressive episode, even if significant residual symptomatology remains. Functional recovery is an important aspect of the recovery process. It has been suggested as an important end point in trials of AD treatments. 17,34 Functioning is a complex process 555

8 affected not only by illness factors but also other individual and environmental factors. Although clearly linked with clinical outcome among this cohort, a functional scale that takes into account the complex nature of functional outcome is necessary before it can be recommended as a primary treatment end point. Our study was conducted in a tertiary specialist centre where patients with more severe illness and complex treatment needs are seen. This may introduce selection bias and may limit generalizability. However, our findings are in line with the recommendations and findings of research into depression in general. Studies in depression have described the impact of residual symptoms on outcome 17,35 and have proposed full remission as the optimal target for the treatment of depression. 8,17,35 The variable and the long follow-up interval may also be another source of limitation. The LIFE chart is designed, with its provisions for anchoring dates and events to the course of the illness, for longer-term interval assessments. We have also attempted to account for the differential impact of the variable length of follow-up either by adjusting for its effect in the analysis or by looking at different follow-up interval gaps separately. To further limit recall bias, we have attempted to use all sources of information, including informants and records. All assessments of functioning were conducted at the end of follow-up and thus it was not possible to determine the longitudinal course of functional impairment. Another limitation might also be that we do not account for any reduction in severity of episode. Reduction in episode severity (for example, from severe to moderate or mild depression) might also be a valuable outcome, although the study had insufficient power to examine this. Finally, we acknowledge that this is a naturalistic study and that other uncontrolled factors may be affecting patient outcome. To our knowledge, this is the first systematic study to look at a well-characterized cohort of patients with TRD to determine the longer-term outcome as a function of baseline, end of treatment clinical status. We conclude that full remission is a desired goal of treatment, even in TRD, and one that is achievable in a substantial proportion. However, partial remission with ongoing residual symptoms still confers an improved long-term symptomatic and functional outcome, and should represent the minimum achievable target for treatment in TRD. Acknowledgements We are grateful to Mr Girmay Medhin for his statistical advice. Dr Cleare has, within the last 5 years, received honoraria for consulting and (or) speaking from Eli Lilly, Bristol-Myers Squibb, UCB Pharma, and Cyberonics, and unrestricted research grant support from GlaxoSmithKline. This research was supported by the National Institute for Health Research Biomedical Research Centre at the South London and Maudsley National Health Service Foundation Trust and the Institute of Psychiatry (King s College London). References 1. Barker PA, Ashcroft GW, Binns JK. Imipramine in chronic depression. J Ment Sci. 1960;106: Greenblatt M, Grosser GH, Wechsler H. A comparative study of selected antidepressant medications and EST. Am J Psychiatry. 1962;119: Harris JA, Robin AA. A controlled trial of phenelzine in depressive reactions. J Ment Sci. 1960;106: Holdway V. Trial of imipramine. J Ment Sci. 1960;106: Kuhn R. The treatment of depressive states with G (imipramine hydrochloride). Am J Psychiatry. 1958;115: Thiery M. Clinical trial of the treatment of depressive illness: report to the Medical Research Council by its Clinical Psychiatry Committee. Br Med J. 1965;1: Greden JF. The burden of disease for treatment-resistant depression. J Clin Psychiatry. 2001;62(Suppl 16): Rush AJ, Trivedi MH, Wisniewski SR, et al. Acute and longer-term outcomes in depressed outpatients requiring one or several treatment steps: a STAR*D report. Am J Psychiatry. 2006;163: Souery D, Papakostas GI, Trivedi MH. Treatment-resistant depression. J Clin Psychiatry. 2006;67(Suppl 6): Trivedi MH, Rush AJ, Wisniewski SR, et al. Evaluation of outcomes with citalopram for depression using measurement-based care in STAR*D: implications for clinical practice. Am J Psychiatry. 2006;163: Fava M. Diagnosis and definition of treatment-resistant depression. Biol Psychiatry. 2003;53: Keller MB. Issues in treatment-resistant depression. J Clin Psychiatry. 2005;66(Suppl 8): Kornstein SG, Schneider RK. Clinical features of treatment-resistant depression. J Clin Psychiatry. 2001;62(Suppl 16): Nelsen MR, Dunner DL. Clinical and differential diagnostic aspects of treatment-resistant depression. J Psychiat Res. 1995;29: Sackeim HA. The definition and meaning of treatment-resistant depression. J Clin Psychiatry. 2001;62(Suppl 16): Frank E, Prien RF, Jarrett RB, et al. Conceptualization and rationale for consensus definitions of terms in major depressive disorder. Remission, recovery, relapse, and recurrence. Arch Gen Psychiatry. 1991;48: Keller MB. Past, present, and future directions for defining optimal treatment outcome in depression: remission and beyond. JAMA. 2003;289: Nierenberg AA, DeCecco LM. Definitions of antidepressant treatment response, remission, nonresponse, partial response, and other relevant outcomes: a focus on treatment-resistant depression. J Clin Psychiatry. 2001;62(Suppl 16): Fekadu A, Wooderson SC, Markopoulo K, et al. What happens to patients with treatment-resistant depression? A systematic review of medium to long term outcome studies. J Affect Disord. 2009;116: Wooderson SC, Juruena MF, Fekadu A, et al. Prospective evaluation of specialist inpatient treatment for refractory affective disorders. J Affect Disord. 2011;131(1-3): Taylor D, Paton C, Kerwin R. The South London and Maudsley Prescribing Guidelines. 9th ed. London (GB): Informa Healthcare; Keller MB, Lavori PW, Friedman B, et al. The longitudinal interval follow-up evaluation a comprehensive method for assessing outcome in prospective longitudinal studies. Arch Gen Psychiatry. 1987;44: World Health Organization. The ICD-10 Classification of Mental and Behavioural Disorders. Clinical descriptions and diagnostic guidelines. 10th ed. Geneva (CH): WHO; Kennedy N, Abbott R, Paykel ES. Remission and recurrence of depression in the maintenance era: long-term outcome in a Cambridge cohort. Psychol Med. 2003;33: Kennedy N, Paykel ES. Residual symptoms at remission from depression: impact on long-term outcome. J Affect Disord. 2004;80: La Revue canadienne de psychiatrie, vol 56, no 9, septembre 2011

9 Long-Term Impact of Residual Symptoms in Treatment-Resistant Depression 26. Hamilton M. A rating scale for depression. J Neurol Neurosurg. 1960;23: Rush AJ, Trivedi MH, Ibrahim HM, et al. The 16-Item Quick Inventory of Depressive Symptomatology (QIDS), clinician rating (QIDS-C), and self-report (QIDS-SR): a psychometric evaluation in patients with chronic major depression. Biol Psychiatry. 2003a;54: Bagby RM, Ryder AG, Schuller DR, et al. The Hamilton Depression Rating Scale: has the gold standard become a lead weight? Am J Psychiatry. 2004;161: American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4th ed. Washington (DC): APA; World Health Organization. The ICD-10 Classification of Mental and Behavioural Disorders. Diagnostic criteria for research. Geneva (CH): WHO; Pintor L, Gastó C, Navarro V, et al. Relapse of major depression after complete and partial remission during a 2-year follow-up. J Affect Disord. 2003;73: Pintor L, Torres X, Navarro V, et al. Is the type of remission after a major depressive episode an important risk factor to relapses in a 4-year follow up? J Affect Disord. 2004;82: Bech P. Social functioning: should it become an endpoint in trials of antidepressants? CNS Drugs. 2005;19: Paykel ES, Ramana R, Cooper Z, et al. Residual symptoms after partial remission: an important outcome in depression. Psychol Med. 1995;25: Rush AJ, Thase ME, Dube S. Research issues in the study of difficult-to-treat depression. Biol Psychiatry. 2003b;53: Ma nuscript received February 2011, revised, and accepted April Associate Professor, King s College London, Institute of Psychiatry, Section of Neurobiology of Mood Disorders, London, England; Associate Professor, Department of Psychiatry, Faculty of Medicine, Addis Ababa University, Addis Ababa, Ethiopia. 2 Postdoctoral Researcher, King s College London, Institute of Psychiatry, Section of Neurobiology of Mood Disorders, London, England. 3 Clinical Researcher, King s College London, Institute of Psychiatry, Section of Neurobiology of Mood Disorders, London, England. 4 Consultant Psychiatrist, King s College London, Institute of Psychiatry, Section of Neurobiology of Mood Disorders, London, England; Consultant Psychiatrist, Affective Disorders Unit, South London and Maudsley National Health Service Trust, London, England. 5 Research Nurse, Affective Disorders Unit, South London and Maudsley National Health Service Trust, London, England. 6 Professor, King s College London, Institute of Psychiatry, Section of Neurobiology of Mood Disorders, London, England; Professor, Affective Disorders Unit, South London and Maudsley National Health Service Trust, London, England; Professor, National Institute for Health Research Biomedical Research Centre, South London and Maudsley National Health Service Foundation Trust, Institute of Psychiatry, King s College London, London, England. Address for correspondence: Dr A Fekadu, King s College London, Institute of Psychiatry, Department of Psychological Medicine and Psychiatry, Section of Neurobiology of Mood Disorders, PO Box 074, 103 Denmark Hill, London, England SE5 8AZ; Abe.Wassie@kcl.ac.uk or Abe_Fekadu@yahoo.co.uk. Résumé : Effet à long terme des symptômes résiduels de la dépression réfractaire au traitement Objectif : Bien qu on rencontre souvent la dépression réfractaire au traitement (DRT), peu de travaux en ont défi ni le cours longitudinal et l infl uence des symptômes résiduels post-traitement sur le résultat à long terme. Notre étude visait à évaluer l effet des états cliniques post-traitement sur le résultat à long terme. Méthode : Des patients (n = 118) souffrant de DRT hospitalisés ont reçu un traitement spécialisé et ont été suivis pendant 3 ans en moyenne. Le résultat longitudinal dichotomisé en résultat bon et mauvais a été utilisé comme résultat principal et les mesures fonctionnelles ont servi de résultats secondaires. Résultats : Parmi les 118 patients traités, 40 (34 %) sont entrés en rémission clinique, 36 (31 %) sont entrés en rémission partielle, et 42 (37 %) étaient encore dans leur épisode à leur congé. Au suivi, 35 % avaient un mauvais résultat défi ni longitudinalement. L état clinique post-traitement était le principal prédicteur du résultat, bon et mauvais. Près de 50 % des patients ont réussi à se rétablir post-congé, et ont eu subséquemment un résultat à long terme, comparable à celui des patients en rémission au moment de leur congé. Les patients qui sont demeurés dans un épisode post-traitement étaient plus incapacités symptomatiquement et fonctionnellement. Conclusion : Les états cliniques post-traitement sont un guide utile aux cliniciens pour projeter le résultat à long terme des patients souffrant de DRT. La persistance des symptômes résiduels ou syndromaux prédit un résultat à long terme plus mauvais, alors que le traitement vers la rémission est associé à de meilleurs résultats. 557