Secondary Logo

Initial Severity Effects on Residual Symptoms in Response and Remission: A STAR*D Study During and After Failed Citalopram Treatment

Madhoo, Manisha MD*; Levine, Stephen Z. PhD

Journal of Clinical Psychopharmacology: August 2015 - Volume 35 - Issue 4 - p 450–453
doi: 10.1097/JCP.0000000000000354
Brief Reports

The effects of initial severity on the time to and course of residual symptoms based on response or remission periods, and during and after failed response to citalopram in major depressive disorder are unknown. STAR*D data during and after failed citalopram treatment were reanalyzed to examine the effect of initial severity on the time to and course of residual symptoms using the Quick Inventory of Depressive Symptomatology–Self-Report (QIDS-SR). During and after failed citalopram treatment, Cox regression and Generalized Estimating Equation models were computed to examine mild and moderate residual symptoms during (1) response based on at least a 50% QIDS-SR reduction, as well as (2) remission based on a QIDS-SR score below 6. Generally, initial severity significantly (P < 0.05) increased the time to and course of residual symptoms at the time of response and remission. The course of select mild and moderate residual symptoms was significantly (P < 0.05) more likely to persist in the presence of initial severity during response than remission (eg, energy) across treatment levels. It is concluded that initial severity is a predictor of the time to and course of residual symptoms. The presence of residual symptoms is more likely during response than remission, thereby directing their definition as a treatment target.

From the *Shire Development LLC, Wayne, PA; and †Department of Community Mental Health, University of Haifa, Haifa, Israel.

Received November 25, 2014; accepted after revision April 27, 2015.

Reprints: Stephen Z. Levine, PhD, Department of Community Mental Health, University of Haifa, Haifa, 31905, Israel (e-mail:

This article was part of a research grant from Shire Development LLC to Dr Levine at the University of Haifa.

Data used in the preparation of this article were obtained from the limited access data sets distributed from the NIH-supported “Sequenced Treatment Alternatives to Relieve Depression” (STAR*D). STAR*D focused on nonpsychotic MDD in adults seen in outpatient settings. The primary purpose of this research study was to determine which treatments work best if the first treatment with medication does not produce an acceptable response. The study was supported by NIMH Contract # N01MH90003 to the University of Texas Southwestern Medical Center. The identifier is NCT00021528. The views in this article may or may not reflect the views of the STAR*D investigators and Shire.

Both authors are responsible for the study concept and design, interpretation of the data analysis, drafting the manuscript, and critical revision; Levine, data acquisition and statistical analysis.

This is an open-access article distributed under the terms of the Creative Commons Attribution-Non Commercial-No Derivatives License 4.0 (CCBY-NC-ND), where it is permissible to download and share the work provided it is properly cited. The work cannot be changed in any way or used commercially.

Residual symptoms of major depressive disorder (MDD) are associated with worse outcomes, such as relapse1 and so represent a key treatment target.2 Theoretically, residual symptoms of MDD are defined as symptoms that persist despite a meaningful clinical response to an adequate course of treatment.3 Empirical studies have examined the presence of residual symptoms at response (ie, a 50% symptom reduction) or remission (ie, symptom-free).1,2,4 Response and remission, however, differ5 and do not necessarily coincide.6 Moreover, research reports that the most common residual symptom among responders without remission is insomnia (94.6%),4 yet among remitters is weight gain (71.3%).1 Currently, the course of residual symptoms after response and remission is unexamined. Debate also exists as to whether initial symptom severity predicts symptom change,7,8 yet the role of initial residual symptom severity is unknown in response and remission. Collectively, these gaps in our current knowledge may adversely affect the understanding and treatment of residual symptoms.

The current study aimed to examine the association between initial severity and the time to and course of residual symptoms in response and remission among STAR*D study participants during and after failed citalopram treatment.

Back to Top | Article Outline


The current study is based on the STAR*D study data. The STAR*D study rationale has been described in depth elsewhere, as have the primary reports and project reviews.9–12 Briefly, STAR*D was designed to identify the most effective of several MDD treatments for nonpsychotic outpatients with an unsatisfactory clinical response to initial citalopram treatment. Ethical approval was received by 14 regional center institutional review boards as well as the National Coordinating Center and the Data Coordinating Center.

Participants were recruited from 18 primary and 23 specialty care settings across the United States. Eligible participants were female and male outpatients (aged 18–75 years) with nonpsychotic MDD for whom outpatient treatment with an antidepressant was considered safe and appropriate by the treating clinician. STAR*D exclusion criteria included major psychiatric disorders, a history of nonresponse or intolerance, medical conditions contraindicating protocol treatments (eg, seizures), and current substance abuse or dependence.9 These selection criteria aimed to obtain a representative sample of persons with MDD treated in everyday practice.

Like the primary STAR*D study, we examined initial citalopram treatment (level 1) among 2876 persons eligible for analysis who consented to the study, and had HAM-D scores over 14.10 Failed responders to citalopram treatment were eligible for randomization to level 2 treatments, with 4 switch options (sertraline, bupropion, venlafaxine, and cognitive therapy) and 3 citalopram augment options (bupropion, buspirone, and cognitive therapy). The level 2 data consisted of 1439 trial participants.9 The protocol recommended assessments at 2, 4, 6, 9, and 12 weeks (with an optional 14-week visit if needed). We examined data to week 12 at each level, because week 14 was by design not completed by all trial participants. Inclusion of these participants would induce unequal follow-up times and alter the symptom presentation rate.

The Quick Inventory of Depressive Symptomatology–Self-Report (QIDS-SR) was used to examine the residual symptoms of depression. The QIDS-SR consists of 16 items with 4 response options. Examined symptoms of MDD based on the STAR-D protocol were sleep disturbance (early, middle, and late insomnia plus hypersomnia); psychomotor disturbance (agitation and retardation); appetite (increase or decrease) and weight disturbance (increase or decrease), depressed mood, decreased interest, decreased energy, worthlessness/guilt, concentration/decision making, and suicidal ideation.

Back to Top | Article Outline


All analyses were computed at levels 1 and 2 of STAR*D. First, all QIDS-SR items were categorized into “at least mild” (equal to or over 1), and “at least moderate” (equal to or over 2) residual symptom levels, like prior research.1 Second, we computed the percent of patients with a mild or moderate baseline symptom that was present at the first week of response or remission as well. Third, we examined the association between the presence or absence of each baseline symptom and the time to and course of each residual symptom in response and remission. Periods examined were response based on a 50% QIDS-SR reduction, and remission based on a QIDS-SR score below 6. To examine the time taken for each residual symptom to be present in response and remission, Cox regression models were computed. Cox regression models provide a hazard ratio (HR), with associated 2-sided 95% confidence intervals (CIs), and significance levels that were set at P < 0.05 (2-sided). The HR values are the time to (ie, risk of) each residual symptom at the first time in response or remission. To examine the course of residual symptoms, Generalized Estimating Equations (GEEs) assuming a binomial distribution were computed.13 This provides odds ratios (OR), with associated 2-sided 95% Wald-type CI values. An OR and CI above 1 means that a mild/moderate symptom at baseline is also present in response or remission. The GEE modeling is robust because it requires no assumptions about the data distribution, capitalizes on maximal data, and adjusts for the within-individual correlation at different visits. To compare response and remission differences in the time and course of each symptom, each model was examined for non-overlapping CI values, implying a significant difference between response and remission. All Cox and GEE models were computed following standard statistical guidelines.13 Baseline mild or moderate severity was set as a factor and adjusted for sex, age, years of education, race (white vs other), and age at onset, that are associated with remission.12 Significance levels were set at P < 0.05 (2-sided).

Back to Top | Article Outline


Description of Residual Symptoms at First Remission and Response

The percentage of mild and moderate residual symptoms at the first postbaseline remission and response visit for each participant was computed (1) during citalopram treatment (level 1; Table 1), and (2) following failed response to citalopram (level 2; Table 2). Across treatment levels, the most frequent mild and moderate residual symptom was mid-nocturnal insomnia. Differences between residual symptom percentages in response and remission were not marked.





Back to Top | Article Outline

Effects of Initial Severity on Time to Residual Symptoms in Remission and Response

On the basis of HR values from Cox regression models, baseline severity consistently increased the likelihood that a residual symptom would be present in response or remission (Tables 1 and 2). This result replicated across residual symptoms at level 1 and level 2. As the HR-associated CI values overlapped, the baseline prediction of the time to each residual symptom in response compared to in remission did not significantly differ.

Back to Top | Article Outline

Course of Residual Symptoms

To examine the course of residual symptoms based on baseline severity, GEE models were computed and adjusted for confounders. Generally, initial severity was a significant (P < 0.05) predictor of the subsequent presence of residual symptoms. Non-overlapping CI values were examined to identify significant differences in the likelihood of each residual symptom during response and remission. At level 1, mild residual symptoms of suicidal ideation, involvement, energy/fatigability, and psychomotor slowing were significantly more likely to be present in response than remission when baseline severity was present. Moderate residual symptoms of appetite, concentration, outlook, suicidal ideation, energy, and slowing were significantly more likely to be present in response than remission (Table 1). At level 2, mild residual symptoms of concentration and energy were significantly more likely in response than remission. Also, the likelihood of moderate residual symptoms of energy, and slowing were significantly more likely in response than remission if the symptom was present at baseline (Table 2).

Back to Top | Article Outline


The current study examines the time to and course of mild and moderate residual symptoms during response and remission in nonpsychotic MDD to citalopram treatment, and after failed response to citalopram based on the STAR*D project data.

The most frequent residual symptom at first remission and response as well as across treatment levels was mid-nocturnal insomnia corresponding with prior research.4 The time to residual symptoms did not differ between response and remission. The course of select residual symptoms (eg, energy) was, however, significantly more likely to be present during response than remission, if the symptom was present at baseline. This is consistent with the notion that response and remission differ,5,6 and conceptualizations of residual symptoms as occurring during response.4 This may reflect the different criteria applied to define each concept. Response is defined as the percentage reduction from baseline to end point, but does not provide information on how symptomatic the participant is at end point. Remission may refer to the absence of significant symptoms, but does not provide information regarding the participant’s symptom improvement.

Generally, baseline severity was a significant predictor of the time to and course of residual symptoms. As the definition of residual symptoms includes a severity and not a change cutoff value, the higher the initial severity, the higher the likelihood that the symptoms will be above the severity cutoff at end point. The result that initial severity is associated with the time to and course of residual symptoms resembles the initial value hypothesis. Thus, the presence of at least mild or moderate symptom severity at baseline seems to be a study design consideration for future clinical trials of residual symptoms, and a treatment target.

Back to Top | Article Outline

Limitations and Conclusions

Although STAR*D provides a sufficient duration to examine residual symptoms, the visit schedule was not ideally designed for this purpose. Periods of residual symptoms could have been missed. For instance, if a participant assessed at weeks 2 and 4 responded at week 3 only, his or her response would be omitted. The study design, which allowed patients to switch to other medications while remaining in the study, may have contributed to the response rate on citalopram (ie, the first medication). Also, comparison of the onset and course of residual symptoms between this and other studies requires caution as studies are designed differently (eg, follow-up times, times between visits).

The current study results suggest that in up to 24 weeks of follow-up during and after failed treatment with citalopram, select key residual symptoms (eg, insomnia) seem to be difficult to target in MDD. Residual symptoms are slightly more prevalent in response than remission and are predicted by initial symptom severity.

Back to Top | Article Outline


Madhoo is an employee and stock holder in Shire Development LLC. Levine has received research support, and/or consultancy fees and/or travel support from Shire, F. Hoffmann-La Roche and Eli Lilly.

Back to Top | Article Outline


1. Nierenberg AA, Husain MM, Trivedi MH, et al. Residual symptoms after remission of major depressive disorder with citalopram and risk of relapse: a STAR*D report. Psychol Med. 2010; 40: 41–50.
2. Madhoo M, Keefe RS, Roth RM, et al. Lisdexamfetamine dimesylate augmentation in adults with persistent executive dysfunction after partial or full remission of major depressive disorder. Neuropsychopharmacology. 2014; 39: 1388–1398.
3. Fava M, Ball S, Nelson JC, et al. Clinical relevance of fatigue as a residual symptom in major depressive disorder. Depress Anxiety. 2014; 31: 250–257.
4. McClintock SM, Husain MM, Wisniewski SR, et al. Residual symptoms in depressed outpatients who respond by 50% but do not remit to antidepressant medication. J Clin Psychopharmacol. 2011; 31: 180–186.
5. Jakubovski E, Bloch MH. Prognostic subgroups for citalopram response in the STAR*D trial. J Clin Psychiatry. 2014; 75: 738–747.
6. 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): 5–9.
7. Fournier JC, DeRubeis RJ, Hollon SD, et al. Antidepressant drug effects and depression severity: a patient-level meta-analysis. JAMA. 2010; 303: 47–53.
8. Kirsch I, Deacon BJ, Huedo-Medina TB, et al. Initial severity and antidepressant benefits: a meta-analysis of data submitted to the Food and Drug Administration. PLoS Med. 2008; 5: e45.
9. Rush AJ, Fava M, Wisniewski SR, et al. Sequenced treatment alternatives to relieve depression (STAR*D): rationale and design. Control Clin Trials. 2004; 25: 119–142.
10. 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: 28–40.
11. Warden D, Rush AJ, Trivedi MH, et al. The STAR*D Project results: a comprehensive review of findings. Curr Psychiatry Rep. 2007; 9: 449–459.
12. Howland RH. Sequenced Treatment Alternatives to Relieve Depression (STAR*D). Part 2: Study outcomes. J Psychosoc Nurs Ment Health Serv. 2008; 46: 21–24.
13. Hothorn T, Everitt BS. A Handbook of Statistical Analyses Using R. Boca-Raton, FL: Chapman & Hall; 2014.

citalopram; residual symptoms; major depressive disorder; course

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.