Slides Shared Resource

1. Epidemiology/Health Services/Psychosocial Research ORIGINAL ARTICLE Fluoxetine for Depression in Diabetes A randomized double-blind placebo-controlled trial with those…
of 6
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  • 1. Epidemiology/Health Services/Psychosocial Research ORIGINAL ARTICLE Fluoxetine for Depression in Diabetes A randomized double-blind placebo-controlled trial with those receiving placebo, but the drug PATRICK J. LUSTMAN, PHD LINDA S. GRIFFITH, MSW KENNETH E. FREEDLAND, PHD RAY E. CLOUSE, MD had significant adverse effects on glycemic control. Path analysis, controlling for oppos- ing effects, showed that improvement in depression had a clinically significant bene- fit on glycemic control: depression remission OBJECTIVE — Depression is prevalent in patients with diabetes. It is associated with poor was associated with a 0.8–1.2% reduction in glycemic control and is linked to an increased risk for diabetic complications. In this study, we glycated hemoglobin over the 8-week study assessed the efficacy of fluoxetine for depression in patients with diabetes. period (20,21). Hyperglycemia has not been reported RESEARCH DESIGN AND METHODS — Sixty patients with diabetes (type 1, n = 26; in patients treated with newer classes of type 2, n = 34) and major depressive disorder entered an 8-week randomized placebo-con- antidepressant agents such as the selective trolled double-blind trial. Patients were given daily doses of fluoxetine (up to 40 mg/day). The serotonin reuptake inhibitors (SSRIs) Beck Depression Inventory (BDI) and Hamilton Rating Scale for Depression (HAMD) were (22,23). The efficacy of fluoxetine used to measure the severity of depression and to determine the percentage of patients who hydrochloride, the first SSRI available in achieved substantial improvement or complete remission. GHb levels were obtained to mon- the U.S., for treating depression in healthy itor glycemic control. patients has been established in a number RESULTS — Reduction in depression symptoms was significantly greater in patients treated of controlled clinical trials (24–26), but its with fluoxetine compared with those receiving placebo (BDI, 14.0 vs. 8.8, P = 0.03; HAMD, usefulness in diabetic patients has been 10.7 vs. 5.2, P = 0.01). The percentage of patients achieving a significant improvement in unknown. Tollefson et al. (27) found that depression per the BDI was also higher in the fluoxetine group (66.7 vs. 37.0%, P = 0.03). Addi- fluoxetine was less effective in patients over tionally, trends toward a greater rate of depression remission (48.1 vs. 25.9%, P = 0.09 per the age 60, which might, the study suggested, HAMD) and greater reduction in GHb ( 0.40 vs. 0.07%, P = 0.13) were observed in the flu- partially result from more comorbid med- oxetine group. ical illness in this age group. The efficacy of depression treatment may be limited by CONCLUSIONS — Fluoxetine effectively reduces the severity of depression in diabetic lifestyle restrictions, pain, impairment, and patients. Our study demonstrated that after only 8 weeks, this treatment also produced a trend toward better glycemic control. disability—realities that often accompany advancing diabetes (20,21). This study was Diabetes Care 23:618–623, 2000 designed to determine the antidepressant efficacy of fluoxetine in diabetic patients with major depressive disorder. A sec- these associations can be altered by the ondary aim was to study the effects of treat- ajor depressive disorder is present M successful treatment of depression. ment and depression improvement on in 15–20% of patients with type 1 In general, little is known about the effi- glycemic control. or type 2 diabetes (1) and has cacy of antidepressant pharmacotherapy in implications that exceed its recognized diabetic patients. Nortriptyline hydrochlo- adverse effects on daily functioning and RESEARCH DESIGN AND ride, a secondary amine tricyclic antidepres- quality of life (2–4). Depression has been METHODS sant, is the only agent previously tested in a associated with poor compliance with the placebo-controlled trial with diabetic diabetes regimen (5,6), poor glycemic con- Patients patients (20). Reduction in depression trol (7–15), and an increased risk for A study to determine the usefulness of flu- symptoms was significantly greater in micro- and macrovascular complications oxetine for depression in diabetic patients patients treated with nortriptyline compared (16–19). It is not known, however, whether was reviewed and approved by the Human Studies Committee of Washington Univer- sity School of Medicine. The study was publicized within the Washington Univer- From the Department of Psychiatry (P.J.L., K.E.F L.S.G., R.E.C.) and the Department of Internal Medicine ., (R.E.C.), Washington University School of Medicine; and the Department of Veterans Affairs Medical Cen- sity Medical Center community and ter (P.J.L.), St. Louis, Missouri. through various advertisements in the St. Address correspondence and reprint requests to Patrick J. Lustman, PhD, Department of Psychiatry, Wash- Louis, Missouri, metropolitan area. Patients ington University School of Medicine, 4940 Children’s Pl., St. Louis, MO 63110. E-mail: lustmanp@ with type 1 or type 2 diabetes who were 21–65 years of age were eligible to partici- Received for publication 13 September 1999 and accepted in revised form 25 January 2000. Abbreviations: ANCOVA, analysis of covariance; BDI, Beck Depression Inventory; DIS, National Insti- pate, provided they were able to give tute of Mental Health Diagnostic Interview Schedule; HAMD, Hamilton Rating Scale for Depression; SSRI, informed consent and answer questions selective serotonin reuptake inhibitor. and fill out research forms on their own. A table elsewhere in this issue shows conventional and Système International (SI) units and conversion Patients were required to meet diagnostic factors for many substances. 618 DIABETES CARE, VOLUME 23, NUMBER 5, MAY 2000
  • 2. Lustman and Associates criteria for major depression and to have a GHb is a function of the glucose concen- referred to his or her primary care physi- score of 14 on either the Beck Depression tration to which the erythrocytes are cian for follow-up, and a letter was sent to Inventory (BDI) or the Hamilton Rating exposed. The half-life of the erythrocytes is the physician summarizing the patient’s Scale for Depression (HAMD). Patients 120 days; HbA1c therefore reflects the progress during study treatment. Placebo- were excluded from participation in the average level of glycemic control during treated patients who remained depressed study if they reported active suicidal that period (40,41). Whole-blood HbA1c (BDI score 10) were given open-label ideation or a history of attempted suicide; concentrations were determined using treatment with the active agent (fluoxetine) had a history of bipolar disorder or any cation exchange column-based high-per- for 1 additional month. The fluoxetine- psychotic disorder; had a current alcohol or formance liquid chromatography on the treated patients who remained depressed at other substance abuse disorder; or were Bio-Rad Variant analyzer (Bio-Rad Labora- study’s end were referred for further treat- currently taking psychoactive medications. tories, Hercules, CA) (42). The range of ment to either their primary care physician Patients for whom fluoxetine therapy was GHb levels for normal nondiabetic persons (for antidepressant medication) or a psy- contraindicated were also excluded. This in the Barnes-Jewish Hospital outpatient chotherapist. group included patients who were preg- laboratory is 4.0–6.0%. The presence of nant or lactating, had a history of convul- complications of diabetes (neuropathy, Statistical analyses sions or seizure disorder, had clinically retinopathy, and nephropathy) was deter- Differences in the demographic and clinical significant hepatic dysfunction (liver func- mined by the physician-investigator characteristics of subjects receiving fluoxe- tion tests 3 times normal), and renal (R.E.C.) on the basis of review of each tine or placebo were determined using insufficiency (serum creatinine 3 mg/dl). patient’s medical history, current symp- Fisher’s exact test for categorical data and toms, physical examination findings, and the Student’s t test for continuous data. The Assessment of depression objective test data (which were obtained results of an intention-to-treat analysis of Presence of major Axis I clinical syndromes through review of clinical records). the depression outcomes are provided for was determined with the National Institute the purpose of comparison to other studies of Mental Health Diagnostic Interview Study design (43,44). The primary analyses of study out- Schedule (DIS) (28) and diagnosed accord- Patients who gave informed consent and comes focused on the “completer” sample. ing to the criteria of the Diagnostic and Sta- met inclusion criteria were randomly Analyses of covariance (ANCOVAs) were tistical Manual of Mental Disorders (29). assigned to receive fluoxetine or placebo used to determine the effects of treatment The reliability and validity of the DIS in administered in identical-appearing opaque on depression (per the BDI and the HAMD) psychiatric and epidemiological studies capsules. A computer algorithm deter- and glycemic control (per GHb), with mean have been extensively reported (30). There mined the randomization pattern. The baseline scores on these measures used as the covariates. The 2 and Fisher’s exact is also evidence for the sensitivity and util- 8-week treatment period began with the ity of this diagnostic technique in patients baseline visit and included subsequent vis- tests were used to examine differences in the with diabetes, in whom some of the its at the end of treatment weeks 3, 5, and percentages of patients in each study group somatic symptoms of major depression 8. Fluoxetine dosing began at 20 mg a day achieving significant improvement in (such as fatigue, weakness, and sleep and in the morning and could be increased to a depression (posttreatment BDI and HAMD appetite disturbances) may be manifesta- maximum of 40 mg daily in the morning scores 50% of the pretreatment scores tions of diabetes (31,32). dependent upon side effects and clinical [45]) and depression remission (posttreat- The severity of the current depression response. Study outcomes were assessed at ment BDI score 9 (35), HAMD score 7 symptoms was measured by the BDI (33) baseline and at the conclusion of the 8-week [46]). Continuous variables are reported as and the HAMD (34). The BDI is both a reli- treatment period. At each evaluation, the means ± SD unless otherwise stated. able and valid measure of the severity of assessments of depression and glycemic RESULTS depression (35). It requires a self-rating control were made and scored indepen- from 0 to 3 on each of 21 items; a cumula- dently of one another. The study personnel tive total from the addition of individual who monitored patient progress were not Demographic and clinical symptom scores is recorded. Depression involved in treatment, and assessors were characteristics presents similarly on the BDI in diabetic blinded to treatment assignments. Three A total of 65 patients gave informed consent and psychiatric patients (32). The 17-item patients were permitted to continue to participate and were evaluated to assess HAMD is a clinician-rated measure of the amitriptyline for neuropathic pain (25–50 their eligibility. Five (7.7%) of these patients severity of depression and has been the mg/day); otherwise, patients were not were excluded from participation, and 60 primary outcome measure in the majority allowed to take other psychoactive med- (92.3%) met all inclusion criteria and were of studies of antidepressant pharmacother- ications during protocol treatment. randomly assigned to 1 of the 2 study apy. There is evidence that the HAMD is a Study personnel did not provide dia- groups. Of the 5 excluded patients, 1 (20%) valid (36–38) and modestly reliable (39) betes treatment, and patients continued to had an exclusionary psychiatric condition, measure of depression severity. see their diabetologists as scheduled or as and 4 (80%) were unwilling to take medica- needed during the trial. At each study visit, tion. Of the 60 patients who were randomly Assessment of diabetes patients were asked to describe any major assigned to treatment, 54 (90%) completed GHb levels were measured to assess changes in diabetes treatment, and alter- the 8 weeks of treatment and 6 (10%) dis- glycemic control. The GHb level correlates ations were recorded. At the end of the continued participation prematurely. Of the with the average blood glucose concentra- 8-week trial, any patient who was classified 6 who did not complete the study, 3 (50%) tion over time because the proportion of as not depressed (BDI score 9) was were in the fluoxetine group and 3 (50%) DIABETES CARE, VOLUME 23, NUMBER 5, MAY 2000 619
  • 3. Fluoxetine in diabetes posttreatment scores on the BDI and the Table 1—Selected characteristics of the study sample HAMD, using pretreatment values as the covariate. The results of these comparisons Characteristic Fluoxetine Placebo are shown in Fig. 1. Mean posttreatment scores on the BDI and the HAMD were n 27 27 significantly lower in the fluoxetine group Age (years) 45 ± 13.0 47.7 ± 11.5 compared with the placebo group (BDI, 9.6 Race ± 8.5 vs. 13.6 ± 10.7, P = 0.03; HAMD, 9.4 Caucasian 25 (92.6) 24 (88.9) ± 9.1 vs. 14.3 ± 10.6, P = 0.01). Reduction Nonwhite 2 (7.4) 3 (11.1) in depression symptoms was significantly Female sex 22 (81.5) 16 (59.3) greater in patients treated with fluoxetine Marital status compared with those receiving placebo, Married 19 (70.4) 17 (63.0) whether measured with the BDI ( 14.0 vs. Not married 8 (29.6) 10 (37.0) 8.8, P = 0.03) or the HAMD ( 10.7 vs. Level of education (years) 14.3 ± 2.0 13.4 ± 2.2 5.2, P = 0.01). Type of diabetes Depression outcomes were also stud- Type 1 13 (48.2) 11 (40.7) ied by comparing in the fluoxetine and Type 2 14 (51.8) 16 (59.3) placebo groups 1) the percentages of Insulin treatment 20 (74.1) 21 (77.8) patients who achieved a clinically signifi- Duration of diabetes (years) 12.2 ± 7.1 14.1 ± 12.2 cant improvement in depression (a GHb level (%) 8.4 ± 1.7 8.6 ± 1.6 decrease 50% in scores on the BDI and Weight (lb) 191.2 ± 67.2 207.6 ± 60.1 the HAMD) and 2) the number of patients Complications of diabetes who achieved depression remission (post- Neuropathy 17 (63.0) 17 (63.0) treatment BDI score 9, posttreatment Nephropathy 2 (7.4) 2 (7.4) HAMD score 7). The categorical out- Retinopathy 9 (33.3) 7 (25.9) comes are summarized in Table 2. Previous episodes of major depression 9.4 ± 8.4 7.3 ± 7.4 The percentage of patients with signifi- BDI score 23.6 ± 8.2 22.4 ± 9.1 cant clinical improvement was significantly HAMD score 20.1 ± 5.6 19.5 ± 6.9 greater in the fluoxetine group than in the Data are n, means ± SD, or n (%). placebo group when improvement was measured with the BDI (66.7% [18/27] vs. were in the placebo group. Of the 3 non- type of diabetes, duration of diabetes, pre- 37.0% [10/27], P = 0.03). The fluoxetine completers in the fluoxetine group, 1 was treatment GHb level, number of previous and placebo groups were not statistically discontinued because of medication side episodes of depression, and pretreatment different when improvement was measured effects, and the remaining 2 withdrew with- scores on the BDI and the HAMD. with the HAMD (59.3% [16/27] vs. 40.7% out explanation. Of the 3 noncompleters in [11/27], P = 0.17). A posttreatment BDI the placebo group, 1 was discontinued Effect of treatment on depression score 9 was observed in 59.3% (16/27) of because of a cardiac event, and 2 withdrew The effect of treatment on depression the fluoxetine group compared with 40.7% without explanation. There were no signifi- symptoms was studied using ANCOVA of (11/27) of the placebo-treated patients (P = cant differences between noncompleters and completers on any of the measured demo- graphic and clinical characteristics: age, race, sex, marital status, education, monthly income, type of diabetes, duration of dia- betes, mode of diabetes treatment, preva- lence of diabetes complications (neuropathy, nephropathy, retinopathy), or pretreatment GHb and depression levels. There was no evidence of differential attrition. There were no significant differences between the fluox- etine and placebo groups on the report of any of 14 potential side effects. Only 2 side effects (nausea and appetite loss) were reported by 10% of the subjects taking fluoxetine. Selected demographic, depression, and diabetes characteristics of the 54 patients who completed treatment are shown in Figure 1—Mean scores on depression measures before and after treatment with fluoxetine or placebo. Table 1. No statistically significant differ- Posttreatment severity of depression was significantly lower in the fluoxetine group compared to the ences were seen between the study groups placebo group when measured by the BDI (9.6 ± 8.5 vs. 13.6 ± 10.7, P = 0.03) or the HAMD (9.4 ± in age, race, sex, education, marital status, 9.1 vs. 14.3 ± 10.6, P = 0.01). 620 DIABETES CARE, VOLUME 23, NUMBER 5, MAY 2000
  • 4. Lustman and Associates were more likely to be younger (P = 0.07) Table 2—Response to fluoxetine and placebo in the study sample and to have type 1 diabetes (P = 0.10). Fluoxetine (%) Placebo (%) Difference (%) P CONCLUSIONS — Our findings indi- cate that fluoxetine is superior to placebo in Intention-to-treat sample (n = 30/group) the management of major depression in Improved patients with comorbid diabetes. The med- BDI 60.0 33.3 26.7 0.04 ication was well tolerated, and the rate of HAMD 53.3 36.7 16.6 0.19 study noncompletion was low (10%) and Remitted unrelated to fluoxetine treatment. Change BDI 53.3 36.7 16.6 0.17 in the severity of depression symptoms, HAMD 43.3 23.3 20.0 0.09 whether measured by the BDI or the Completer sample (n = 27/group) HAMD, was significantly greater in patients Improved treated with fluoxetine than it was in the BDI 60.0 33.3 26.7 0.04 placebo-treated control subjects. The per- HAMD 53.3 36.7 16.6 0.19 centage of patients whose depression Remitted significantly improved was also higher in BDI 59.3 40.7 18.6 0.17 the fluoxetine group. Although the per- HAMD 48.1 25.9 22.2 0.09 centage of patients achieving full syndro- Unweighted average 21.1 mal remission did not reach statistical Data are %. Remitted = posttreatment BDI score 9, posttreatment HAMD score 7; improved = posttreatment significance over the 8 weeks of the study, score on the measure is 50% of the baseline score. a trend in this direction was observed. Response rates might have been higher if 0.17). A posttreatment HAMD 7 was the course of the trial, improvement in the treatment interval had been longer, as is observed in 48.1% (13/27) of the fluoxetine mean GHb level was greater in the fluoxe- customary in clinical settings. group compared with 25.9% (7/27) of the tine group than in the placebo group, but In patients who completed the trial, the placebo group (P = 0.09). the difference was not statistically significant average rate of response [(BDI improved An intention-to-treat analysis was also ( 0.40 vs. 0.07%, P = 0.13). This statis- BDI remitted HAMD improved performed. In this conservative analysis, tical trend could not be attributed to differ- HAMD remitted)/4] was 58.4% in the flu- noncompleters of the study were treated as ential effects of treatment on weight or on oxetine group compared with 36.1% in if they did not achieve depression remission the severity of depression symptoms. the placebo group, a difference of 22.3%. or significant improvement. The percentage ANCOVA was used to test the effect of treat- This result compares favorably with the of patients with significant clinical improve- ment (fluoxetine vs. placebo) on posttreat- average difference in response, 18.4%, ment (a decrease of 50% in scores on the ment weight, with weight at ba
  • BET Theory

    Jul 23, 2017
    We Need Your Support
    Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

    Thanks to everyone for your continued support.

    No, Thanks