|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2003 American Society for Clinical Oncology Fluoxetine Versus Placebo in Advanced Cancer Outpatients: A Double-Blinded Trial of the Hoosier Oncology Group
From the Department of Palliative Care and Rehabilitation, University of Texas M.D. Anderson Cancer Center, Houston, TX; Division of Hematology/Oncology, Indiana University; Walther Cancer Institute, and Division of Biostatistics, Indiana University School of Medicine, Indianapolis; Department of Psychology, Indiana State University, Terre Haute, IN; Department of Palliative Care, University of Kentucky, Lexington, KY; Department of Biostatistics and Bioinformatics, Duke University, Durham, NC; and Division of Medical Oncology, Washington University, St Louis, MO. Address reprint requests to Michael J. Fisch, MD, MPH, University of Texas M.D. Anderson Cancer Center, Box 008, Room P12.2911, 1515 Holcombe Blvd, Houston, TX 77030-4009; email: mfisch{at}mdanderson.org.
Purpose: To determine whether fluoxetine improves overall quality of life (QOL) in advanced cancer patients with symptoms of depression revealed by a simple survey. Patients and Methods: One hundred sixty-three patients with an advanced solid tumor and expected survival between 3 and 24 months were randomly assigned in a double-blinded fashion to receive either fluoxetine (20 mg daily) or placebo for 12 weeks. Patients were screened for at least minimal depressive symptoms and assessed every 3 to 6 weeks for QOL and depression. Patients with recent exposure to antidepressants were excluded. Results: The groups were comparable at baseline in terms of age, sex, disease distribution, performance status, and level of depressive symptoms. One hundred twenty-nine patients (79%) completed at least one follow-up assessment. Analysis using generalized estimating equation modeling revealed that patients treated with fluoxetine exhibited a significant improvement in QOL as shown by the Functional Assessment of Cancer Therapy–General, compared with patients given placebo (P = .01). Specifically, the level of depressive symptoms expressed was lower in patients treated with fluoxetine (P = .0005), and the subgroup of patients showing higher levels of depressive symptoms on the two-question screening survey were the most likely to benefit from treatment. Conclusion: In this mix of patients with advanced cancer who had symptoms of depression as determined by a two-question bedside survey, use of fluoxetine was well tolerated, overall QOL was improved, and depressive symptoms were reduced.
THE CARE of patients with advanced cancer is becoming increasingly challenging because of the growing numbers of patients living with cancer and the increasing expectations of patients and their families for effective palliative care. According to a recent report from the Institute of Medicine, "A major problem in palliative care is the underrecognition, underdiagnosis, and thus undertreatment of patients with significant distress ranging from existential anguish to anxiety and depression."1 Significant depressive symptoms occur in roughly 25% to 35% of cancer patients.2,3 However, there are several inherent difficulties in diagnosing depression in this population. The most obvious problem is that sadness and grief are normal responses to the changes associated with the diagnosis of cancer and at transitional points in the disease. In addition, the physical signs of depression may be caused by the malignancy or by medications commonly used for cancer patients. Furthermore, patients and their family caregivers often do not recognize or accept the diagnosis of depression. A symptom-based approach is commonly used for the management of phenomena such as pain and nausea, as well as psychologic distress in cancer patients. This involves ascertaining whether the symptom is present, elucidating further details, and deciding on the proper management approach for the patient. In addition, there are no universally accepted criteria for diagnosing depression in the terminally ill.4 Multiple depression questionnaires have been developed for the primary-care population, with instruments ranging in length from one to 30 items and ranging in administration times from less than 1 to 5 minutes.5 In the largest survey of depressive symptoms in ambulatory cancer patients, the 20-item Zung Self-Rating Depression Scale (ZSDS) was administered to more than 1,000 outpatients.6 An 11-item version of the ZSDS, the Brief Zung Self-Rating Depression Scale (BZSDS), has been found to have acceptable levels of criterion validity.7 However, several studies have demonstrated the utility of one- or two-item questionnaires in both primary care and palliative care settings.8,9 We have piloted the use of a two-item questionnaire that assesses the cardinal symptoms of depressed mood and anhedonia in advanced cancer outpatients and have found it easy to administer and acceptable to patients.10 For patients who are treated for the syndrome of major depression, 50% to 60% respond to initial therapy with antidepressants, psychotherapy, or both.11 Compared with placebo, patients with subsyndromal depression or dysthymia also benefit from treatment with an antidepressant or psychotherapy with similar response rates.12 Fluoxetine is a familiar antidepressant to many oncologists, and it is the first and most widely studied selective serotonin-reuptake inhibitor. Uncontrolled trials published in the late 1980s and 1990s indicated that antidepressants might also be helpful in selected cancer patients.13–18 Only six published, randomized, placebo-controlled trials have compared an antidepressant drug (a tricyclic antidepressant or serotonin-reuptake inhibitor) with placebo for the treatment of depression in patients with cancer.18–23 The trend in these data shows a modest benefit of an antidepressant compared with placebo, but there is concern about the generalizability of the data because of patient dropout and the relative preponderance of women with breast or gynecologic malignancies in these studies. It is clear that patients with depressive disorders benefit from treatment, but it is also evident that there are major barriers to diagnosing depressive disorders in outpatients with advanced cancer. The purpose of this study was to explore the efficacy and feasibility of treating outpatients with advanced cancer with an antidepressant on the basis of the presence of at least minimal depressive symptoms. A randomized, placebo-control design was needed to obtain an accurate assessment of efficacy. The primary objective was to compare the change in quality of life (QOL) of these patients; the secondary objective was to compare the change in depressive symptoms.
Patients Between July 1998 and October 2000, 163 adult ambulatory patients with advanced, incurable malignancy were enrolled from one of 15 sites of the Hoosier Oncology Group (three academic centers and 12 community sites). Patients with an expected survival between 3 and 24 months were eligible for study participation. Expected survival was judged by clinicians on the basis of the histology, stage, and comorbid status of the patient and any other prognostic data typically used by the individual clinician (physician or nurse). Clinician-predicted survival was 3 to 12 months for 59% of the patients. During the study period, all participating Hoosier Oncology Group clinicians made available to each outpatient a two-item screening survey to assess depressed mood and anhedonia. These data were used for clinical purposes and were not limited to patients who were possible candidates for this study. The questions were as follows: (1) "During the past month, have you often been bothered by feeling down, depressed, or hopeless?" and (2) "During the past month, have you often been bothered by having little interest or pleasure in doing things?" To be eligible for this study, each patient had to have a score of 2 or greater on this Two-Question Screening Survey (TQSS). Each of the items (depressed mood and anhedonia) in the survey has five possible responses that were assigned values of 0 to 4 as follows: 0, not at all; 1, a little bit; 2, somewhat; 3, quite a bit; and 4, very much. The score on the survey was the sum of the two questions. As it pertains to this trial, the TQSS was not being used as a casefinding instrument; rather, it was being used to eliminate from consideration enrollment of patients who have no problems at all with depressed mood or anhedonia. This was an ethical issue raised by study reviewers because these asymptomatic patients were believed to have a very low possibility of benefiting from treatment with an antidepressant. Patients could be receiving any anticancer therapy or only the best supportive care. Exclusion criteria included the following: serious suicidal risk or psychotic behaviors; inability to swallow oral medications; regular use of antidepressants or psychotropic drugs (other than phenothiazine-type antiemetics or benzodiazepines) within 6 weeks of the baseline study evaluation; uncontrolled brain or leptomeningeal disease; bilirubin more than 2.0 mg/dL; current use of a monoamine oxidase inhibitor; enrollment onto another clinical trial with QOL as the primary end point; recent or active substance abuse; and major depression diagnosed by a psychiatrist in the past 6 months. Informed consent was obtained for each patient with appropriate institutional review.
Procedures
Longitudinal assessments were performed at baseline and every 3 to 6 weeks thereafter and included the measurement of QOL and depression. The visit interval varied among patients and often depended on the schedule for anticancer therapy. Patients were assessed for 12 weeks, and complete assessment involved three to five sessions of data collection (depending on the individual patient’s visit intervals). After the 12-week study period, the patients were given the option to continue the study drug (blinded) for up to 9 more months. The flow of study participants is summarized in Fig 1
Measures QOL was measured using a reliable and valid 28-item self-report instrument called the Functional Assessment of Cancer Therapy–General (FACT-G, version 3).24 Spiritual well-being was measured using a related instrument, the Functional Assessment of Chronic Illness Therapy–Spiritual, with 12 items25 (data for the latter will be reported separately). Depressive symptoms were measured using the 11-item BZSDS.6 Higher scores represent better QOL, better spiritual well-being, and more depressive symptoms (worse health) in these assessment tools, respectively. These assessment tools were chosen because they can easily be used in outpatient cancer-care settings and because of their face validity.
Statistical Analysis An additional analysis planned a priori involved a comparison of the best-change score between the patient groups. The best-change score was defined as the difference between the baseline FACT-G score and the average of the best consecutive scores. The use of summary measures, such as the best change score, has been described as a relevant and statistically valid way to evaluate longitudinal data,27 and it has been used to analyze QOL data in clinical cancer research.28 We defined a best-change score of 6 points or greater as showing a response. This 6-point change was chosen as the response threshold because a 6-point change in the FACT-G is roughly equivalent to a one-level change in the Eastern Cooperative Oncology Group performance status and because, in similar patient populations, it represents an effect size of 0.4 to 0.5, which is generally considered a small-to-moderate effect.24 For example, a patient with a baseline FACT-G score of 90 who had a follow-up score of 100 and then was lost to follow-up would be considered a nonresponder with a best-change score of 5 points. If this same patient had one more follow-up with a score of 94, the best-change score of 7 points would have been sufficient for the patient to be considered a responder.
Sample Size Calculations
The baseline questionnaires were completed by 159 patients (98%), and at least one follow-up assessment was completed by 129 patients (81%). The baseline characteristics of the 30 patients who did not complete at least one follow-up are listed in Table 1  . Patients who completed at least one follow-up assessment were assessable for the primary outcome variable of change in QOL over time.
Patients in each study arm were comparable at baseline with respect to age, sex, performance status, symptom status regarding pain and depression, disease distribution, and current treatment with chemotherapy. These patient characteristics are listed in Table 2 . A family history of depression in first-degree relatives and current or prior psychologic counseling (individual or group) since their diagnosis was more frequent in patients randomly assigned to receive placebo. The mean interval between visits (in weeks) was similar for patients in each treatment arm. The patients’ cancer treatments are listed in Table 3. There were significantly more fluoxetine patients currently receiving radiation therapy, and more placebo patients had undergone surgery.
Patient compliance to the study drug was not assessed with pill counts but by patient self-report at follow-up. Six patients discontinued the study drug (3.7%). Four patients dropped out of the fluoxetine arm (two because of daily headaches and two because of nausea or vomiting), and two patients dropped out of the placebo arm (one because of nonspecific side effects and one because of nausea or vomiting). Fifteen patients had unexpected hospitalizations during the study; nine in the fluoxetine arm and six in the placebo arm (P = .59 by two-sided Fisher’s exact test). To assess a common side effect of fluoxetine, clinicians were asked to report at each follow-up the number of times the patient vomited in the past week. There was no difference in the frequency of emesis for the second, third, and fourth visits, but the frequency was significantly greater in patients in the fluoxetine arm at the study completion visit; nine (33%) of 27 patients in the fluoxetine arm reported one or more episodes of emesis at the study completion visit, compared with two (4.6%) of 43 patients receiving placebo (P = .01 by the two-sided Wilcoxon rank sum test).
The longitudinal data regarding QOL are listed in Table 4
There was a statistically significant improvement in the total FACT-G scores in the fluoxetine group, compared with the placebo group, using the GEE method of regression (P = .05). This difference remained highly significant after multivariate adjustments were made for age, sex, the two-question screening score, performance status, family history of depression, and current cancer therapy (P = .01). The fluoxetine group also improved significantly, compared with placebo, on the depression scale (P = .0005). The full GEE model characteristics for depression and QOL are listed in Table 7 . In both the QOL and depression model, female sex was associated with significantly better outcomes (P = .004 for QOL and P = .04 for depression). Moreover, the TQSS was a significant predictor of overall QOL and depression (P = .001 and P = .01, respectively), having a stronger association with both outcomes (QOL and depression) than did performance status. Patients with no family history of depression had less depression (P = .02). Two subscales of the FACT-G, social well-being and emotional well-being, showed statistically significant improvement with fluoxetine on the basis of the GEE model, but there were no significant differences in the mean best-change scores for any of the subscales.
The TQSS score was less than 4 in 32% of the patients. In this subgroup, the GEE model showed no treatment effect on QOL (P = .19) or depression (P = .35). However, the treatment effect was more pronounced in the subset of patients whose TQSS score was 4 or greater, with significant improvement in QOL (P = .05) and depression (P = .0008). This subgroup analysis on the basis of the TQSS was unplanned and is considered exploratory. The conclusions were the same when the subset analysis was performed using the baseline BZSDS data dichotomized by those patients scoring in the normal range (0 to 21; 64 patients) versus patients with mild or moderate depression at baseline (95 patients).
The survival of the treatment groups is illustrated using the Kaplan-Meier method in Fig 2
This study is the first to explore an oncologist-driven, symptom-management solution to the prevalent problem of depressive symptoms in advanced cancer patients. To our knowledge, it is also the largest study examining the placebo response rate associated with the FACT-G instrument for QOL assessment of outpatients with advanced cancer. Our findings show that simple symptom screening followed by antidepressant treatment is both feasible and associated with improvement in overall QOL and depressive symptoms. The overall benefit to QOL in the fluoxetine arm, although statistically significant, is of uncertain clinical significance because the proportion of patients with a 6-point difference in their best-change score, indicating clinically important improvement, did not differ between the treatment arms. Determining the clinical significance of changes in self-reported health measures measured longitudinally is difficult, and this is an area of active research and debate.29 The best-change score represented an attempt to provide a very conservative summary of change of longitudinal data that was both meaningful and simple to calculate. By taking the difference between the baseline measure and the average of the best consecutive measures, this method gives greater weight to any sustained improvements and discounts the treatment effect when there are missing data. If there had been a significant difference between the treatment arms with this best-change score method, the clinical significance of the statistical findings would have been convincing, but it is clearly more difficult to interpret a no-difference result. Therefore, more research is needed to validate this method. In addition to the controversy surrounding the definition of clinically significant changes, there is also controversy regarding the appropriate use and interpretation of QOL end points in studies in which the intervention is expected to affect a specific kind of clinical problem. Somerfield et al30 have argued persuasively that a more reasonable expectation of an intervention is that it would be effective at the more proximal target problem (in this case depression) rather than a more distal, global outcome such as overall QOL. This was indeed borne out in our study, in which there was a more convincing proximal effect of the antidepressant (decrease in the depressive symptoms) and a statistically significant but less convincing distal effect. It seems that the depressive symptoms did not abate as the result of adaptation and regression to the mean but, instead, as a result of a real effect from the antidepressant. In addition, for those who might speculate that use of an antidepressant might cause the average patient worse problems than the depression itself because of polypharmacy, drug interactions, or antidepressant side effects, there are reassuring data that the overall QOL was improved in the treatment arm. Just as overall QOL might be regarded as a distal outcome for an antidepressant trial, survival would be even more distal. Survival was measured in this trial as a descriptive end point to better understand this patient population and its generalizability to other groups of patients. As expected, there was no difference in survival between the treatment arms. Since United States Food and Drug Administration approval of fluoxetine in 1987, there has been a national trend toward greater use of psychotropic drugs and greater involvement of physicians in the management of depression in outpatients.31 Thus, our data may broaden the comfort zone of oncologists for prescribing antidepressants for some patients. A particular factor favoring the use of antidepressants in patients with advanced cancer is that access to behavioral health professionals and psychologic interventions is not always feasible for very sick patients. It is also important that the TQSS proved to be quite easy to adapt to clinical practice as an initial assessment of psychologic distress. The exploratory analysis revealed that the subset of patients with mild or moderate depressive symptoms (on the basis of the baseline TQSS or the baseline BZSDS) benefited the most from the fluoxetine intervention. Therefore, the TQSS, as an example of simple bedside screening, is worthy of additional research as a predictor of response to antidepressant therapy in this population. Selection bias is one of several important limitations to this study. These 163 patients were enrolled over a period of 30 months and represent a small fraction of the total eligible patients. The pace of accrual was affected by the physicians’ lack of familiarity with placebo-controlled symptom research, and there was a substantial learning curve in this regard. More important, collecting questionnaire data using existing office staff is cumbersome and was not always feasible on a day-to-day basis. This group’s willingness to pursue this research is a credit to oncologists and their drive to learn more about their patients’ psychologic distress. Another important limitation was the substantial missing data. This is an inherent problem in advanced cancer outpatient research, and it was no worse in this study than in other similar studies. Nevertheless, nonrandom reasons for missing data introduce possible bias into this study. Although new methods for the analysis of studies with missing data such as this study are being explored, more important innovations for preventing missing data also are being pursued. Better funding for symptom research projects and use of telephone assessments may lessen the missing data problem in future research. In summary, this study demonstrated that simple two-question screening for depressive symptoms followed by treatment with a once-a-day pill proved effective in identifying and treating depressive symptoms and, hence, improving QOL in patients with advanced cancer.
We thank the nurses and physicians of the Hoosier Oncology Group and study design consultants Ian Tannock, PhD, and Jimmie Holland, MD, and we acknowledge the generosity and courage of the patients who agreed to participate in this study.
Supported in part by the Mary Margaret Walther Program for Cancer Care Research, Indianapolis, IN. Fluoxetine, placebo, and the study notebooks were provided by the Eli Lilly Company, Indianapolis, IN. Presented in part at the Thirty-Seventh Annual Meeting of the American Society of Clinical Oncology, San Francisco, CA, May 12–15, 2001.
1. Foley KM, Gelband H (eds): Improving Palliative Care for Cancer: Summary and Recommendations. Washington, DC, National Academy Press, 2001 2. Massie MJ, Popkin MK: Depressive disorders, in Holland JC (ed): Psycho-Oncology. New York, NY, Oxford University Press, 1998, 518–540 3. Wilson KG, Chochinov HM, de Faye B, et al: Diagnosis and management of depression in palliative care, in Chochinov HM, Breitbart W (eds): Handbook of Psychiatry in Palliative Care. Oxford, United Kingdom, Oxford University Press, 2000, 25–49 4. Lloyd-Williams, M: Screening for depression in palliative care. Am J Hosp Pall Care 19:7–8, 2002
5. Williams JW Jr, Noel PH, Cordes JA, et al: Is this patient clinically depressed? J Am Med Assoc 287:1160–1170, 2002
6. Passik SD, Dugan W, McDonald MV, et al: Oncologists’ recognition of depression in their patients with cancer. J Clin Oncol 16:1594–1600, 1998 7. Passik SD, Kirsh K, Donaghy KB, et al: An attempt to employ the Zung Self-Rating Depression Scale as a "lab test" to trigger follow-up in ambulatory oncology clinics: Criterion validity and detection. J Pain Symptom Manage 21:273–281, 2001[CrossRef][Medline] 8. Whooley M, Avins A, Miranda J, et al: Case-finding instruments for depression: Two questions are as good as many. J Gen Intern Med 12:439–445, 1997[CrossRef][Medline] 9. Chochinov H, Wilson K, Enns M, et al: "Are you depressed?" Screening for depression in the terminally ill. Am J Psychiatry 154:674–676, 1997[Abstract] 10. Fisch M, Marx R, Brames M, et al: Use of a two-question screening survey for depression in outpatients with advanced cancer. Proc Am Soc Clin Oncol 18:580a, 1999 (abstr 2241)
11. Whooley MA, Simon GE: Managing depression in medical outpatients. N Engl J Med 343:1942–1950, 2000
12. Williams JW Jr, Barrett J, Oxman T, et al: Treatment of dysthymia and minor depression in primary care: A randomized controlled trial in older adults. J Am Med Assoc 284:1519–1526, 2000 13. Breitbart W: Psycho-oncology: Depression, anxiety, delirium. Semin Oncol 21:754–769, 1994[Medline] 14. Evans DL, McCartney CF, Haggerty JJ Jr, et al: Treatment of depression in cancer patients is associated with better life adaptation: A pilot study. Psychosom Med 50:72–76, 1988 15. Massie M, Holland J: Depression and the cancer patient. J Clin Psychiatry 51:12–17, 1990 (suppl 7)[Medline] 16. Mermelstein H, Lesko L: Depression in patients with cancer. Psychooncology 1:199–215, 1992[CrossRef] 17. Holland J, Romano S, Heiligenstein J, et al: A controlled trial of fluoxetine and desipramine in depressed women with advanced cancer. Psychooncology 7:291–300, 1998[CrossRef][Medline] 18. Razavi D, Allilaire J, Smith M, et al: The effect of fluoxetine on anxiety and depression symptoms in cancer patients. Acta Psychiatr Scand 94:205–210, 1996[Medline]
19. van Heeringen K, Zivkov M: Pharmacological treatment of depression in cancer patients: A placebo-controlled study of mianserin. Br J Psychiatry 169:440–443, 1996 20. Eija K, Tiina T, Pertti NJ: Amitriptyline effectively relieves neuropathic pain following treatment of breast cancer. Pain 64:293–302, 1996[CrossRef][Medline] 21. Purohit DR, Navlakha PL, Modi RS, et al: The role antidepressants in hospitalised cancer patients: A pilot study. J Assoc Physicians India 26:245–248, 1978[Medline] 22. Costa D, Mogos I, Toma T: Efficacy and safety of mianserin in the treatment of depression of women with cancer. Acta Psychiatr Scand 72:85–92, 1985 (suppl)[CrossRef]
23. Musselman DL, Lawson DH, Gumnick JF, et al: Paroxetine for the prevention of depression induced by high-dose interferon alfa. N Engl J Med 344:961–966, 2001
24. Cella DF, Tulsky DS, Gray G, et al: The Functional Assessment of Cancer Therapy Scale: Development and validation of the general measure. J Clin Oncol 11:570–579, 1993 25. Brady MJ, Peterman AH, Fitchett G, et al: A case for including spirituality in quality of life measurement in oncology. Psychooncology 8:417–428, 1999[CrossRef][Medline]
26. Liang KY, Zeger SL: Longitudinal data analysis using generalized linear models. Biometrika 73:13–22, 1986
27. Matthews JN, Altman DG, Campbell MJ, et al: Analysis of serial measurements in medical research. Br Med J 300:230–235, 1990
28. Tannock IF, Osoba D, Stockler MR, et al: Chemotherapy with mitoxantrone plus prednisone or prednisone alone for symptomatic hormone-resistant prostate cancer: A Canadian randomized trial with palliative end points. J Clin Oncol 14:1756–1764, 1996
29. Sloan JA, Cella D, Frost M, et al: Assessing clinical significance in measuring oncology patient quality of life: Introduction to the symposium, content overview, and definition of terms. Mayo Clin Proc 77:367–370, 2002
30. Somerfield M, Jatoi A, Nguyen PL, et al: Hazards of quality of life data for clinical decision making. J Clin Oncol 19:594–595, 2001
31. Olfson M, Marcus SC, Druss B, et al: National trends in the outpatient treatment of depression. J Am Med Assoc 287:203–209, 2002 Submitted August 2, 2002; accepted February 20, 2003.
Related Correspondence
Related Reply
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2003 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
2). Patients were then randomly assigned in a double-blind manner to receive either fluoxetine (20-mg tablets) or an identical placebo tablet. The randomization was performed centrally through a preprinted randomization table, and the study drug was sent by overnight mail directly to the patient. The study drug was self-administered by the patient once daily in the morning. 
