Predictors of fluoxetine remission for hospitalized patients with major depressive disorder


  • Ching-Hua Lin MD, PhD,

    1. Kai-Suan Psychiatric Hospital
    2. Graduate Institute of Medicine
    Search for more papers by this author
  • Hsien-Yuan Lane MD, PhD,

    1. Institute of Clinical Medical Science, China Medical University
    2. Departments of Psychiatry, China Medical University Hospital, Taichung, Taiwan
    Search for more papers by this author
  • Cheng-Chung Chen MD, PhD,

    1. Kai-Suan Psychiatric Hospital
    Search for more papers by this author
  • Suh-Hang Hank Juo MD, PhD,

    1. Department of Medical Genetics
    2. Center of Excellence for Environmental Medicine, Kaohsiung Medical University
    3. Department of Medical Research, Kaohsiung Medical University Hospital, Kaohsiung
    Search for more papers by this author
  • Cheng-Fang Yen MD, PhD

    Corresponding author
    1. Department of Psychiatry, Faculty of Medicine, College of Medicine
    2. Department of Psychiatry, Kaohsiung Medical University Hospital
      Cheng-Fang Yen, MD, PhD, Department of Psychiatry, Kaohsiung Medical University Hospital, 100 Tzyou 1st Rd, Kaohsiung 807, Taiwan. Email:
    Search for more papers by this author

Cheng-Fang Yen, MD, PhD, Department of Psychiatry, Kaohsiung Medical University Hospital, 100 Tzyou 1st Rd, Kaohsiung 807, Taiwan. Email:


Aim:  The goal of treating major depressive disorder is to achieve remission. This prospective study aimed to identify predictors of remission in a cohort of depressive inpatients who received fluoxetine.

Methods:  A total of 131 newly hospitalized patients with major depressive disorder received a fixed dose of 20 mg/day (the recommended dose from the literature) of fluoxetine for 6 weeks. Symptom severity was assessed using the 17-item Hamilton Depression Rating Scale at weeks 0, 1, 2, 3, 4 and 6. Remission was defined as a score of ≤7 on the 17-item Hamilton Depression Rating Scale after 6 weeks of treatment. We compared the remitters and non-remitters in terms of baseline variables. The Short-Form-36 pain interference item was used to assess pain. It was classified as high (score ≥ 3) or low (score < 3).

Results:  A total of 31 (27.7%) of 112 completers remitted after 6 weeks of treatment. The remitters and non-remitters did not differ in baseline variables, except pain interference, baseline depression severity, and depression improvement at week 1.

Conclusion:  These findings obtained from newly hospitalized major depression patients support the previous notion that pain interference, depression severity, and early improvement can be the predictors for remission. Patients with high pain interference, a greater depression severity or a less early improvement are likely to require aggressive treatment early. These data require confirmation and extension to outpatients and other antidepressants.

MOST CONTEMPORARY GUIDELINES and experts have recommended that achieving remission (i.e. minimal or no depressive symptoms) should be viewed as the goal in the acute treatment of depression.1–7 This recommendation is based on studies consistently demonstrating that depressed patients who have improved with treatment but failed to achieve remission experience more psychosocial impairment and have higher likelihood of recurrence of full depressive symptoms.8–10

However, attaining complete remission remains a major clinical challenge. In defining remission using the 17-item Hamilton Depression Rating Scale (HAMD-17)11 of ≤7, only 25–35% of patients in clinical trials experience remission,3,12,13 so that the majority of patients with major depressive disorder are actually expected to fail to achieve remission with a single trial of monotherapy with antidepressants.14

In the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study,13 adequate doses of citalopram were given for 14 weeks. The remission rate (by HAMD-17 score ≤ 7) was only 28%. The result demonstrates that even with adequate dosages and duration, about 70% of patients fail to achieve remission.

Numerous studies have been conducted to identify predictors of treatment outcome,13,15,16 such as REM latency,17 occupancy of brain serotonin transporter by SSRI,18 a lower baseline concentration of homovanillic acid in plasma,19 and early improvement.20,21 However, few predictors have yet been substantiated.

Depressive patients often have a complex set of overlapping symptoms, including emotional and physical complaints. Physical complaints typically include medically unexplained pain.22,23 Pain is not a diagnostic symptom of major depressive disorder according to the DSM-IV;24 however, evidence from a growing database suggests that chronic pain and depression appear to share common biological pathways and neurotransmitters.25–27 It is reported that as many as 76% of patients with depression suffer from painful physical complaints.28 The high co-occurrence of depression and pain has led to the speculation that pain symptoms are a core component of depression.29 Some reports suggest that pain may contribute to a reduced improvement from depression treatment.30–35 Fava et al.36 studied antidepressant effects on pain and remission in 495 depressive patients. A decrease in overall pain scores was associated with an increase in the probability of achieving remission of depressive symptoms. These results suggest that pain may act as a barrier to remission in patients with major depressive disorder.

In this study conducted in the depression-care ward in a major psychiatric center in Taiwan, we enrolled newly admitted patients with major depressive disorder, aiming to detect the predictors of remitters from baseline demographic and clinical features (including pain), to help clinicians formulate treatment decisions and shorten the time spent on ineffective treatments.



The study was approved by Kai-Suan Psychiatric Hospital's institutional review board and conducted in accordance with Good Clinical Practice procedures and the current revision of the Declaration of Helsinki (Project number: KSPH-2007–16).

Patients were recruited from Kai-Suan Psychiatric Hospital, a major psychiatric center in Taiwan, between May 2007 and February 2010. All patients newly hospitalized for acute treatment of major depressive disorder were screened and evaluated by experienced psychiatrists. The Structured Clinical Interview for DSM-IV37 was used to ensure the accuracy of the diagnosis. Han Chinese patients in Taiwan were enrolled in this study if they: (i) were physically healthy and had all laboratory parameters within normal limits; (ii) were aged 18–70 years old; (iii) satisfied DSM-IV24 criteria for major depressive disorder; (iv) had an HAMD-17 ≥18, and a Clinical Global Impression of Severity (CGI-S)38≥4 at baseline; (v) had no DSM-IV diagnosis of substance abuse or dependence (including alcohol) within the past 6 months; and (vi) gave written informed consent to participate in the study after a full explanation of the study's aims and procedures. Patients excluded from this study were: (i) those with a history of serious adverse reaction to fluoxetine, or a history of epilepsy or organic mental disorders; (ii) those with psychotic depression, or bipolar I or II disorder; (iii) those with schizophrenia or any other psychotic disorder; (iv) those with severe cognitive impairment; (v) female subjects who were pregnant, or at risk of pregnancy or lactation; (vi) those patients initiating or stopping formal psychotherapy within 6 weeks prior to enrollment; (vii) those with treatment-resistant depression (defined as lack of response to two or more adequate courses of antidepressant treatment); or (viii) those who had a history of poor response to fluoxetine (20 mg/day for ≥4 weeks) or previously received electroconvulsive therapy.


After a washout period of at least 72 h, patients received open-labeled fluoxetine treatment at a fixed dose of 20 mg daily39 for 6 weeks. Benzodiazepine (≤4 mg lorazepam equivalent) was allowed at bedtime as needed for insomnia. No other psychotropic agents were used. Drug adherence was monitored and ensured by psychiatric nurses.

Symptom severity was assessed at baseline, and again at weeks 1, 2, 3, 4 and 6 by trained, senior psychiatrists using HAMD-17. The intraclass correlation coefficient (ICC) of reliability was 0.95 between the raters. To maintain high interrater reliability and prevent rater drift, raters met at least once a month for training and reliability re-testing. The research psychiatrists who conducted the clinical ratings did not know the detailed study design, or the remitter versus non-remitter status of patients as defined during the study.

Remitters and non-remitters were compared in terms of sex, marital status, family history of depressive disorders, melancholic features, anxious depression, pain interference, age, age at onset, educational level, number of previous episodes, baseline CGI-S score, baseline HAMD-17 score, and HAMD-17 score improvement (%) at week 1. Each subject's family history was taken to indicate whether first-degree relatives had a history of depressive disorders. Anxious depression was defined as major depressive disorder with high levels of anxiety symptoms, as reflected in an HAMD-17 anxiety/somatization factor score ≥ 7.40 The anxiety/somatization factor includes six items from the HAMD-17: the items for psychic anxiety, somatic anxiety, gastrointestinal somatic symptoms, general somatic symptoms, hypochondriasis, and insight. Age at onset was regarded as age at the first major depressive episode.

We assessed pain in this population using a measure of the level to which pain interfered with daily activities. Pain interference over the past 1 month was measured by the pain interference item of Medical Outcomes Study Short-Form-36 (SF-36).33,35 Pain interference was assessed with the question, ‘During the past 4 weeks, how much did pain interfere with your normal work, including both work outside the home and housework?’ It is rated on a five-point scale ranging from 1 (not at all) to 5 (extremely). A higher score means more pain. The SF-36 pain interference item was classified as high (score ≥ 3) or low (score < 3). This classification has been used previously.33,41–43

Side-effects were evaluated at each visit by the Udvalg for Kliniske Undersogelser (UKU) Side Effect Rating Scale,44 with scores ranging from 0 (none) to 3 (severe). A score of 1, 2 or 3 on any UKU item that first occurred or worsened during treatment indicated ‘cases’ of adverse events. UKU was administered at baseline and at weeks 1, 2, 3, 4 and 6. Adverse events were conceptualized as a dichotomous variable, either ‘present’ or ‘not present’.

Remission of symptoms was chosen as the outcome measure. Remission was defined as a score of ≤7 on the 17-item HAMD-17 after the 6-week fluoxetine treatment.45 Thus, remission status was a dichotomous variable, operationally defined as ‘remission’ or ‘non-remission’. HAMD-17 scores can range from 0 to 52, with higher scores indicating more severe depression.

Statistical analyses

Pearson's χ2-test or Fisher's exact test were used to compare categorical variables; Mann–Whitney U-test was used for continuous variables. All tests were two-tailed, and significance was defined as an alpha of less than 0.05. All data were processed by spss version 17.0 for Windows (spss Inc., Chicago, IL, USA).


A total of 131 acutely ill inpatients with major depressive disorder were enrolled. A total of 112 (85.5%) of the 131 patients completed the 6-week trial of fluoxetine (completers). Twenty-four (21.4%) of the completers were male and 88 (78.6%) were female. The mean age was 42.9 ± 11.0 years. The remaining 19 did not complete the trial due to: lack of efficacy (n = 3), premature discharge (n = 14) or withdrawal of consent (n = 2). No patient withdrew from the study due to adverse events. The dropout patients (n = 19) and the completers (n = 112) were comparable for sex (P = 0.14), age (P = 0.32), age at onset (P = 0.06), number of previous episodes (P = 0.67), baseline CGI-S scores (P = 0.85), and baseline HAMD-17 scores (P = 0.19) (data not shown in Table).

Of the 112 completers, 27.7% (n = 31) of the subjects were classified as remitters after the 6-week treatment. Remitters and non-remitters did not differ in sex, marital status, family history of depressive disorders, melancholic features, anxious depression, age, age at onset, and number of previous episodes. Compared to remitters, non-remitters were likely to have high pain interference, higher baseline CGI-S scores, higher baseline HAMD-17 scores, and lower HAMD-17 score improvement (%) at week 1 (Table 1).

Table 1.  Clinical characteristics comparing remitters and non-remitters following 6 weeks of treatment
VariablesRemitters (n = 31)Non-remitters (n = 81)Analysis
Sex: Female2477.46479.00.85
Marital status: Married or cohabiting1548.44251.90.74
Family history of depressive disorders825.82227.20.89
Melancholic features2890.36479.00.27
Anxious depression2477.47187.70.18
Pain interference: High1445.25669.10.019*
  • *

    Statistically significant.

  • Pearson's χ2-test.

  • Fisher's exact test.

  • §

    Mann–Whitney U-test.

  • CGI-S, Clinical Global Impression of Severity; HAMD-17, 17-item Hamilton Depression Rating Scale.

Age at onset42.113.238.511.10.11§
Educational level (years)11.53.310.93.60.58
Number of previous episodes2.§
Baseline CGI-S score6.00.856.40.60.027*§
Baseline HAMD-17 score28.77.532.76.00.005*§
HAMD-17 score improvement (%) at week 144.118.627.420.0<0.001*

Adverse events occurring at an incidence of 20% or higher in any treatment group are shown in Table 2. Non-remitters had significantly higher rates of adverse events and were more likely to experience asthenia/increased fatigability, failing memory, tension, orthostatic dizziness and palpitation. No severe side-effects (score = 3) in any of the UKU items were noted in any of our patients.

Table 2.  Adverse events occurring in at least 20% of patients in either group
Adverse eventsRemitters (n = 31)Non-remitters (n = 81)Analysis
  • *

    Statistically significant.

  • Fisher's exact test.

  • Pearson's χ2-test.

At least one adverse event2580.68098.8<0.001*
Concentration difficulties412.92227.20.14
Asthenia/increased fatigability26.52125,90.03*
Failing memory13.21721.00.02*
Accommodation disturbances39.72024.70.11
Reduced salivation825.83340.70.14
Orthostatic dizziness825.83846.90.04*


The main finding of this study was that patients with high pain interference, higher baseline scores of depressive symptom severity (baseline HAMD-17 or CGI-S scores), and less HAMD-17 score improvement (%) at week 1 were less likely to remit during the 6-week fluoxetine treatment. It provides the foundation that these patients are likely to require aggressive treatment to increase remission rates.

In the current study, high pain interference predicted a poor antidepressant remission. Our finding is in accordance with those from a variety of studies, in which pain indicates difficult-to-treat depression.30,31,33,34,36 If we used Pearson's correlation to determine the correlation between baseline HAMD-17 score and baseline SF-36 pain interference score (a continuous variable), the results would reveal that the baseline HAMD-17 score was correlated with the baseline SF-36 pain interference score (r = 0.24, P = 0.009). This result is in accordance with a previous report that depression and pain have a reciprocal correlation in that each heightens the severity of the other.46 Seventy (62.5%) of the subjects reported high pain interference in their normal activities at work or at home in the current study; this means that pain is a common problem in depressed inpatients. For depression treatment, the high rate of co-occurrence and the adverse effects of pain suggest it no longer makes sense to treat depression without considering pain. With respect to pain, this may involve antidepressants with analgesic effects (e.g. tricyclic antidepressants, or serotonin-norepinephrine reuptake inhibitors).33

A number of investigators have examined predictors of remission to acute treatment of major depressive disorder. The consistent finding is that greater baseline severity of depressive symptoms is a significant predictor of non-remission.10,13,47–50 Patients with severe depression tend to have a lower probability of remission and a greater risk of subsequent recurrent episodes.51 For severe depression, the STAR*D study has also emphasized the limitations of current acute treatments and the frequent need for multiple treatment trials to reach remission of symptoms.13

Like baseline severity of depressive symptoms to predict remission, our finding is consistent with a couple of reports that have suggested that early improvement can be used as a predictor of antidepressant remission.20,21,52

If variables (i.e. pain interference, baseline CGI-S score, baseline HAMD-17 score, and HAMD-17 score improvement (%) at week 1) that discriminated between the groups (P < 0.05) are used as potential predictors in a forward stepwise logistic regression model to determine the strong predictor(s) for fluoxetine remission, baseline HAMD-17 score (odds ratio [ratio of odds of remission vs non-remission] = 0.90, 95% confidence intervals [CI] = 0.83–0.96, P = 0.003) and HAMD-17 score improvement (%) at week 1 (odds ratio = 1.05, 95%CI = 1.02–1.07, P < 0.001) are the best predictors. Clinicians could adjust their treatment plan for those patients with greater baseline depressive severity or less early improvement. For example, an increased dose of a current antidepressant or augmentation with another psychotropic agent have been suggested.52

In our study, non-remitters had significantly higher rates of adverse events. The possible reason is that there may be an overlap between adverse events to fluoxetine and symptoms of depression. Patients with more complaints of adverse events were positively related to more severity of depressed mood,53 and had greater difficulty achieving remission.

Patients in the current study received the same fixed dose, 20 mg daily, of fluoxetine treatment. Earlier fixed-dose studies39,54 have demonstrated that 20 mg of fluoxetine daily is the optimal dose for most patients, and is associated with fewer and less severe side-effects than higher doses. A meta-analysis study by Beasley et al.55 also found that fluoxetine therapy at 20 mg daily is a critical factor for adequate therapy and has good treatment tolerance. However, the rate and quality of response to fluoxetine are highly individualized. Certainly, a proportion of patients may benefit from higher doses.

Several strengths of this study could be addressed. First, the subjects were inpatients requiring acute treatment. Hospitalized patients constituted only a small proportion of the patients in the previous studies.56 As inpatients, the subjects were carefully monitored for symptom assessment, the development of side-effects, and medical adherence. They also had similar environments. Second, inpatients reflect greater severity of depression than outpatients. It has been reported that patients with more severe depression get fewer placebo effects.57 Third, HAMD-17 was originally developed for inpatients.11 It might be sufficiently, or even more, sensitive in detecting changes in depressive symptoms with greater severity.50,58 Finally, unlike other studies that have used last-observation-carried-forward (LOCF) analysis to account for missing data, we analyzed only the trial completers. LOCF analysis, assuming that a subject's severity rating at the time of dropout would be the same as his or her rating at the end of the trial, could add a negative bias to the results across time.59

The limitations of this study included a short-term, open-labeled treatment design, the use of a single antidepressant agent (fluoxetine), and having been conducted in only one psychiatric center. The ideal duration for an antidepressant trial is not well characterized. The traditional period of 6 weeks may not be most appropriate given that a subset of patients go on to achieve remission after 6 weeks. However, Stassen et al.60 suggest that extending the clinical trial beyond 6 weeks certainly identifies only about 5% of late improvement but does not produce more new information. Although this was an open-labeled study, its goal was early identification of remitters rather that demonstrating treatment efficacy. Additionally, an open trial resembles the clinical practice setting,61 in which patients and clinicians both know the medication and both expect the outcomes. Consequently, the results may be more generalizable to clinical settings. Furthermore, the subjects were inpatients who had been hospitalized due to severe symptoms, severe functional impairment, or suicidal tendencies. Thus, inclusion of a placebo group gave rise to ethical concerns. Moreover, it was also unclear if the conclusions could be generalized to other antidepressants or outpatients.

In addition, our findings demonstrated that baseline pain was a strong predictor of fluoxetine non-remission. But pain was solely based on self-report measures, that is, the SF-36 pain interference score. As pain is mainly a subjective perception and experience, self-reports of pain are commonly considered accurate.62 Studies that have examined the concordance between subjective reports of pain and external measures of pain have found good agreement between the two.63

Further studies, preferably involving other antidepressants, larger inpatient or outpatient groups from multi-centers, and a treatment period of longer than 6 weeks, are needed to better determine the predictive value of baseline depressive symptoms and pain for ultimate treatment remission.


This study was funded by the Kai-Suan Psychiatric Hospital in 2007–2009, Food and Drug Administration, Department of Health, Taiwan (DOH100-FDA-43002-L) and National Science Council, Taiwan (NSC-97–2314-B-039–006-MY3, NSC-99–2627-B-039–001, NSC-100-2627-B-039-001), the National Health Research Institutes, Taiwan (NHRI-EX-100-9904NI), the Department of Health, Taiwan (DOH99-TD-I-111-TM001), and the Taiwan Department of Health Clinical Trial and Research Center of Excellence (DOH100-TD-B-111–004).