Fibromyalgia is a chronic musculoskeletal disorder

Transcription

1 A Randomized, -controlled, Double-blind, Flexible-dose Study of in the Treatment of Women with Fibromyalgia Lesley M. Arnold, MD, Evelyn V. Hess, MD, James I. Hudson, MD, SM, Jeffrey A. Welge, PhD, Sarah E. Berno, BPh, Paul E. Keck, Jr, MD PURPOSE: To assess the efficacy of fluoxetine in the treatment of patients with fibromyalgia. SUBJECTS AND METHODS: Sixty outpatients (all women, aged years) with fibromyalgia were randomly assigned to receive fluoxetine (10 80 mg/d) or placebo for 12 weeks in a double-blind, parallel-group, flexible-dose study. The primary outcome measures were the Fibromyalgia Impact Questionnaire total score (score range, 0 [no impact] to 80) and pain score (score range, 0 10). Secondary measures included the McGill Pain Questionnaire, change in the number of tender points, and total myalgic score. RESULTS: In the intent-to-treat analysis, women who received fluoxetine (mean [ SD] dose, mg/d) had significant (P 0.005) improvement in the Fibromyalgia Impact Questionnaire total score compared with those who received placebo, with a difference of 12 (95% confidence interval [CI]: 19 to 4). They also had significant (P 0.002) improvement in the Fibromyalgia Impact Questionnaire pain score (difference, 2.2 [95% CI: 3.6 to 0.9]), as well as in the Fibromyalgia Impact Questionnaire fatigue (P 0.05) and depression (P 0.01) scores and the McGill Pain Questionnaire (P 0.01), when compared with subjects who received placebo. Although counts for the number of tender points and total myalgic scores improved more in the fluoxetine group than in the placebo group, these differences were not statistically significant. CONCLUSION: In a 12-week, flexible-dose, placebo-controlled trial, fluoxetine was found to be effective on most outcome measures and generally well tolerated in women with fibromyalgia. Am J Med. 2002;112: by Excerpta Medica, Inc. Fibromyalgia is a chronic musculoskeletal disorder characterized by widespread pain and tenderness, and is often accompanied by fatigue, sleep disturbance, and morning stiffness (1). The prevalence of fibromyalgia is estimated to be 2% in the United States, mostly in women aged 50 years and older (2). The pathophysiology of fibromyalgia is unknown, but central monoaminergic neurotransmission may play a role in its etiology. Two recent meta-analyses of trials of antidepressants have shown tricyclic antidepressants to be moderately effective in the treatment of patients with fibromyalgia (3,4). Although the use of selective serotonin reuptake inhibitors (SSRIs) in fibromyalgia has not been studied extensively, these medications are likely to be better tolerated than the tricyclic antidepressants (5). From the Women s Health Research Program (LMA, SEB), Department of Psychiatry, University of Cincinnati Medical Center, Cincinnati, Ohio; Department of Internal Medicine (EVH), Division of Immunology, University of Cincinnati Medical Center, Cincinnati, Ohio; Biological Psychiatry Laboratory (JIH), McLean Hospital, Belmont, Massachusetts; Department of Psychiatry (JIH), Harvard Medical School; Department of Biostatistics and Epidemiology (JIH), Harvard School of Public Health, Boston, Massachusetts; Biological Psychiatry Program (JAW, PEK), Department of Psychiatry, University of Cincinnati Medical Center, Cincinnati, Ohio. Supported by an investigator-initiated grant from Eli Lilly and Company, Indianapolis, Indiana. Requests for reprints should be addressed to Lesley M. Arnold, MD, University of Cincinnati Medical Center, 231 Albert B. Sabin Way, P.O. Box , Cincinnati, Ohio Manuscript submitted March 19, 2001, and accepted in revised form October 29, The effects of SSRIs in fibromyalgia have been examined in four placebo-controlled trials involving citalopram (6,7) and fluoxetine (8,9). In the studies of citalopram, Nørregaard and associates (6) found no significant differences in efficacy measures between the citalopram and placebo groups; whereas Anderberg et al. (7) reported significant improvement in one of the measures of pain and a significant decrease in depressive symptoms. There were no significant differences between groups in the global assessment of improvement (7). Results from the trials in fluoxetine were similarly conflicting. Wolfe and associates (8) found no significant difference between fluoxetine and placebo on any efficacy measure; however, this study was limited by a high rate of withdrawal in the placebo group, a trial duration of 6 weeks, and a modest dose (20 mg/d) of fluoxetine. By contrast, Goldenberg and colleagues (9) found that both fluoxetine 20 mg/d and amitriptyline 25 mg/d had significantly greater efficacy than did placebo in a crossover trial, and were more effective when administered together. However, interpretation of the findings might have been limited by unblinding when subjects crossed over to treatments with different side effects; carryover effects when subjects switched from active to placebo arms; and an inadequate washout period for fluoxetine, which requires at least a month for elimination (10). Finally, the efficacy of treatment with the combination of fluoxetine and amitriptyline may have been due, in part, to a pharmacokinetic interaction between the two agents. Since fluoxetine 2002 by Excerpta Medica, Inc /02/$ see front matter 191 All rights reserved. PII S (01)

2 typically increases tricyclic levels substantially (11,12), it may have increased the dose of amitriptyline to greater than 25 mg/d. Based on these limited data regarding the efficacy of SSRIs in the management of fibromyalgia, we conducted a 12-week, parallel-group, placebo-controlled study to assess the efficacy of fluoxetine in 60 outpatients with fibromyalgia. SUBJECTS AND METHODS Subjects were recruited from local rheumatology clinical outpatient practices and by advertisements for a medication trial for fibromyalgia. They were eligible for inclusion in the study if they were at least 18 years of age and met the American College of Rheumatology 1990 criteria for fibromyalgia (1). The screening protocol included an interview for demographic information; medical, psychiatric, and family histories; the Structured Clinical Interview (13) to identify any current or past Axis I psychiatric disorders as defined by the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) (14); and the 17-item Hamilton Depression Rating Scale (15). Patients were excluded if they had evidence of traumatic injury, inflammatory rheumatic disease, or infectious or endocrine-related arthropathy; clinically unstable medical illness; a history of seizure, head trauma, or stroke; a lifetime history of hypomania, mania, psychosis, or dementia; alcohol or substance dependence during the past 6 months; a substantial risk of suicide; any current Axis I diagnosis; a score of 10 or greater on the 17-item Hamilton Depression Rating Scale; received monoamine oxidase inhibitors, tricyclics, lithium, SSRIs, or other antidepressants within 2 weeks before randomization; received investigational medications within 3 months before randomization; or previously received fluoxetine for fibromyalgia. The Institutional Review Board at the University of Cincinnati, Ohio, approved the protocol, and all subjects provided written informed consent. Study subjects were randomly assigned to treatment with fluoxetine or placebo for 12 weeks. All medications were in identical capsules (20 mg of fluoxetine or placebo). Subjects began the double-blind treatment at a starting dose of 20 mg, administered daily for the first 2 weeks. If this dose was not tolerated, it was decreased to 20 mg every other day. After 2 weeks, the dose could be titrated in 10- to 20-mg increments every 2 weeks to a maximum of 80 mg/d. Adjustments within the range of 1 capsule every other day to 4 capsules per day were made at the discretion of the investigator and until a patient improved or intolerance occurred. Subjects completed the Fibromyalgia Impact Questionnaire (16), a self-report instrument composed of 19 items. The first 10 items comprise a physical functioning scale; each item is rated on a 4-point Likert-type scale. On items 11 and 12, subjects indicated the number of days that they felt well or missed work because of fibromyalgia symptoms. Items 13 through 19 are 10-cm, visual analog scales along which subjects rated the difficulty in performing their job responsibilities, pain, fatigue, morning tiredness, stiffness, anxiety, and depression. All subscores with the exception of the two work-related scores were summed to yield the total score of fibromyalgia impact, which ranges from 0 (no impact) to 80 (maximum impact). Subjects also completed the McGill Pain Questionnaire, in which subjects chose words to describe their subjective pain experience (17). The McGill Pain Questionnaire consists of the sum of the ranked values associated with adjectives describing the severity of pain. Finally, subjects were examined for the number of positive tender points and total myalgic score. To obtain the total myalgic score, a Fischer dolorimeter (18) with a rubber disc of 1 cm 2 was applied at a 90 vertical angle to 18 tender point sites, and pressure was increased at a rate of 1 kg/s. Subjects were instructed to indicate verbally when they first felt discomfort. The pressure was then stopped, and the measurement was recorded. The individual scores were summed to give the total myalgic score. Subjects were evaluated by a rheumatologist (EVH) who confirmed the diagnosis of fibromyalgia based on a review of medical records, history, and results of the physical examination. If subjects were taking a medication that was not permitted, they tapered the dose and discontinued the medication. Subjects were allowed to continue acetaminophen or nonsteroidal anti-inflammatory medications on their usual schedule. Subjects returned after 2, 4, 6, and 8 weeks for assessment of medication dose, any adverse events, and completion of the Fibromyalgia Impact Questionnaire and McGill Pain Questionnaire. Subjects returned after 12 weeks of treatment or, if possible, at their final visit to repeat the above assessments and to complete the dolorimetry examination and tender point count. After the first 10 subjects entered the study, we changed the technique of measuring baseline and endpoint tender point counts by digital pressure to the use of a dolorimeter to achieve a more consistent application of 4 kg of pressure. Therefore, the tender point counts for the first 10 subjects are excluded in the data analysis. The primary outcome measures were the Fibromyalgia Impact Questionnaire total score and pain score. Because pain is the predominant symptom of fibromyalgia, a secondary outcome measure for pain, the McGill Pain Questionnaire pain rating index (19,20), was used to assess other subjective qualities of the pain experience in fibromyalgia. Other secondary outcome measures included the change in the number of tender points and total myalgic score. 192 February 15, 2002 THE AMERICAN JOURNAL OF MEDICINE Volume 112

3 Statistical Analysis Pretreatment comparisons between groups were made using the Fisher exact test for categorical variables and two-sample t tests for continuous variables. The primary time point for efficacy analyses was the time of the last postbaseline measurement obtained for each subject during the study. For each outcome measure in each subject, we computed the change from baseline to the last observation (last observation value minus baseline value). This analysis uses the last known state of all subjects, including those who failed to complete the study (assuming that at least one postbaseline measurement was available), in accordance with an intent-to-treat analysis. The mean ( SD) change from baseline to last observation between groups was tested for difference from zero with a twosided t test at To measure the sensitivity of this unadjusted analysis to possible residual confounding (i.e., covariate imbalance not eliminated by randomization), we used analyses of covariance to compare mean differences after adjusting for these variables. RESULTS Sixty subjects were enrolled in the study from March 1998 to June 2000; 30 were assigned to each treatment group. All subjects were women, and at baseline were similar in age, race, education, marital status, employment status, duration of fibromyalgia, and history of a major depressive disorder, although the difference between the groups in marital status approached statistical significance (P 0.06; Table 1). Their baseline values for the outcome measures were also comparable (Table 2). Twenty-three subjects withdrew during the course of the study; 12 from the placebo group and 11 from the fluoxetine group. The reasons for withdrawal were an adverse effect or event (n 12), including nervousness, headache, insomnia, nausea, gastrointestinal reflux, increased appetite, decreased appetite or weight loss, fatigue, sedation, anorgasmia, hot feelings, weakness, decreased coordination, decreased concentration, migraine, worsening pain, herniated lumber disc, fall with muscle strain, ear pain, or diagnosis of breast cancer; lack of efficacy (n 5); loss to follow-up (n 4); or for other reasons that were not related to medication (n 2). A significantly greater proportion of placebo-treated subjects discontinued treatment because of lack of efficacy (17% [n 5] of placebo-treated subjects vs. 0% of fluoxetine-treated subjects, P 0.05). Endpoint data for the intent-to-treat analysis of the Fibromyalgia Impact Questionnaire scores and McGill Pain Questionnaire scores were not available for 9 of the subjects who withdrew (5 from the fluoxetine group and 4 from the placebo group). Because of subject withdrawals and exclusion from the analysis of some tender point Table 1. Baseline Characteristics of 60 Women with Fibromyalgia Receiving or * Characteristics Mean SD or Number (%) Age (years) Duration of fibromyalgia (years) White 27 (90) 29 (97) Education High school 5 (17) 7 (23) Some college or higher 25 (83) 23 (77) Married 23 (77) 15 (50) Employment status Working 20 (67) 22 (73) On disability for fibromyalgia 1 (3) 1 (3) History of major depression 17 (57) 20 (67) * There were no statistically significant differences between the treatment groups. measures, endpoint intent-to-treat data on tender point counts were not available for 24 subjects (12 from each treatment group), and total myalgic scores were not avail- Table 2. Baseline Measurements of 60 Subjects with Fibromyalgia Receiving or * Measure (score range) Mean SD Fibromyalgia Impact Questionnaire Total score (0 80) Subscores Physical impairment (0 9.99) Days felt good ( ) Work missed (0 10) Work impairment (0 10) Pain (0 10) Fatigue (0 10) Feeling tired upon awakening (0 10) Stiffness (0 10) Anxiety (0 10) Depression (0 10) Tender points (0 18) Myalgic score McGill Pain Questionnaire (0 78) * There were no statistically significant differences between the treatment groups. Recoded so that a higher number indicates greater impairment. (n 20), placebo (n 25); subjects could delete this item, and it is not included in the total score. (n 24), placebo (n 26); subjects could delete this item, and it is not included in the total score. (n 26), placebo (n 24). (n 29). February 15, 2002 THE AMERICAN JOURNAL OF MEDICINE Volume

4 Table 3. Outcome Measures for Subjects with Fibromyalgia Receiving or : Intent-to-Treat Analysis Measure (score range) Mean ( SD) Change From Baseline to Endpoint (n 25) (n 26) Between-Group Difference (95% Confidence Interval) P Value Fibromyalgia Impact Questionnaire Total score (0 80) ( 19.4 to 3.6) Subscores Physical Impairment (0 9.99) ( 1.9 to 0.6) 0.28 Days felt good ( ) ( 3.6 to 0.2) 0.08 Work missed* (0 10) ( 1.0 to 0.9) 0.88 Work impairment (0 10) ( 3.4 to 1.0) 0.27 Pain (0 10) ( 3.6 to 0.9) Fatigue (0 10) ( 3.0 to 0.0) 0.05 Feeling tired upon awakening (0 10) ( 2.5 to 0.4) 0.15 Stiffness (0 10) ( 2.9 to 0.1) 0.07 Anxiety (0 10) ( 2.5 to 0.5) 0.19 Depression (0 10) ( 3.5 to 0.5) 0.01 Tender points (0 18) ( 3.7 to 0.7) 0.17 Myalgic score ( 4 to 13.8) 0.27 McGill Pain Questionnaire (0 78) ( 15.9 to 2.3) 0.01 * Not included in total score; fluoxetine (n 17), placebo (n 19). Not included in total score; fluoxetine (n 18), placebo (n 21). (n 18), placebo (n 18). (n 20), placebo (n 23). able for 17 subjects (10 from the fluoxetine group and 7 from the placebo group). The mean ( SD) dose at endpoint for patients treated with fluoxetine was mg/d for those who completed 12 weeks of treatment and mg/d for all fluoxetine-treated subjects. The Fibromyalgia Impact Questionnaire total score improved significantly in the fluoxetine group compared with the placebo group in the intent-to-treat analysis (P 0.005; Table 3, Figure 1) and in the subjects who completed treatment (P 0.03; Table 4). The pain score also decreased significantly in the fluoxetine group compared with the placebo group in the intent-to-treat analysis (P 0.002; Table 3, Figure 2) and in subjects who completed treatment (P 0.01; Table 4). Although threshold severity scores were not established for the Fibromyalgia Impact Questionnaire, a 25% improvement in scores corresponded to our impression of clinically meaningful improvement. A 25% improvement in the total score was observed in 8 (32%) of the 25 fluoxetine-treated subjects and 4 (15%) of the 26 subjects who received placebo (P 0.19), whereas a 25% improvement in the pain score was observed in 14 (56%) of the fluoxetine-treated subjects and 4 (15%) of the subjects who received placebo (P 0.003). The other Fibromyalgia Impact Questionnaire scores that improved significantly in the fluoxetine-treated subjects compared with the subjects who received placebo Figure 1. Intent-to-treat analysis of individual change in Fibromyalgia Impact Questionnaire total scores for fluoxetine and placebo treatment groups. The corresponding means with 95% confidence intervals are to the right of the data. The asterisk (*) indicates P versus placebo. 194 February 15, 2002 THE AMERICAN JOURNAL OF MEDICINE Volume 112

5 Table 4. Outcome Measures for Subjects with Fibromyalgia Receiving or : Completer Analysis Measure (score range) Mean ( SD) Change From Baseline to Endpoint (n 19) (n 18) Between-Group Difference (95% Confidence Interval) P Value Fibromyalgia Impact Questionnaire Total score (0 80) ( 21 to 1) 0.03 Subscores Physical Impairment (0 9.99) ( 1.3 to 1.3) 0.99 Days felt good ( ) ( 4.4 to 0.7) 0.14 Work missed* (0 10) ( 1.1 to 0.2) 0.18 Work impairment (0 10) ( 2.0 to 2.0) 0.98 Pain (0 10) ( 3.8 to 0.5) 0.01 Fatigue (0 10) ( 3.8 to 0.2) 0.03 Feeling tired upon awakening (0 10) ( 3.2 to 0.6) 0.18 Stiffness (0 10) ( 2.9 to 0.8) 0.26 Anxiety (0 10) ( 2.7 to 1.2) 0.43 Depression (0 10) ( 3.7 to 0.4) 0.02 Tender points (0 18) ( 4 to 1) 0.22 Myalgic score ( 4.6 to 16.2) 0.27 McGill Pain Questionnaire (0 78) ( 13.3 to 1.1) 0.02 * Not included in total score; fluoxetine (n 13), placebo (n 13). Not included in total score; fluoxetine (n 14), placebo (n 13). (n 16), placebo (n 15). (n 18), placebo (n 18). included the depression score, which improved in the intent-to-treat analysis (P 0.01) and the completer analysis (P 0.02), and the fatigue score, which improved in the intent-to-treat analysis (P 0.05) and the completer analysis (P 0.03). Since the difference in marital status between the treatment groups at baseline approached significance, marital status was considered a covariate in a secondary analysis of covariance for the primary outcomes. There was no significant interaction between marital status and treatment, and the magnitude and significance of the overall treatment effect were not sensitive to the inclusion of marital status. We also assessed whether the effect of fluoxetine on pain was influenced by a history of major depressive disorder or by baseline level of depression. However, analysis of covariance revealed no significant interaction between treatment and either a history of major depressive disorder or by baseline level of depression, nor were the magnitude or significance of the treatment effects sensitive to the inclusion of these two covariates in the model. Finally, we assessed the possibility that the effect of fluoxetine on pain was mediated through improvements in depression. Indeed, the effect of fluoxetine on the pain score was still significant (P 0.02) after adjustment for change in the depression score. The secondary pain measure, the McGill Pain Questionnaire, improved significantly in the fluoxetine group compared with the placebo group in the intent-to-treat analysis (P 0.01; Table 3) and the completer analysis (P 0.02; Table 4). Although the counts for tender points decreased and the myalgic scores increased (representing less tenderness) further in the fluoxetine group than in the placebo group, the differences were not significant (Tables 3 and 4). The most common adverse events reported by the fluoxetine-treated subjects were headache, insomnia, sedation, and nausea. There were no significant differences between the treatment groups in the incidence of these events. DISCUSSION In this flexible-dose, placebo-controlled, 12-week study of fluoxetine in fibromyalgia, we found fluoxetine treatment significantly superior to placebo in improving the Fibromyalgia Impact Questionnaire total score and the Fibromyalgia Impact Questionnaire measures of pain, fatigue, and depression. In addition, patients who received fluoxetine had a significantly decreased severity of pain as measured by the McGill Pain Questionnaire. Counts for tender points and myalgic scores improved in the fluoxetine group, but not significantly more than in the placebo group. was also well tolerated. In the previous placebo-controlled trials of fluoxetine administered at 20 mg/d, one found no therapeutic effect (8), while the other reported a modest effect (9). Given the beneficial effects of higher doses of fluoxetine in our February 15, 2002 THE AMERICAN JOURNAL OF MEDICINE Volume

6 Figure 2. Intent-to-treat analysis of individual change in Fibromyalgia Impact Questionnaire pain scores for fluoxetine and placebo treatment groups. The corresponding means with 95% confidence intervals are to the right of the data. The asterisk (*) indicates P versus placebo. study, some patients who fail to respond to a 20-mg/d dose may improve with higher doses. Our finding of no change in tender points is consistent with meta-analyses of antidepressant treatment studies of fibromyalgia, in which measures of pain, sleep, and global well-being improved moderately, but tender points improved only minimally (3,4), and similar to the crossover study of fluoxetine and amitriptyline in which the tender point score did not improve with antidepressant treatment despite improvement in pain, global well-being, and function scores (9). There are several possible explanations for the lack of improvement in tender point measures. First, it may be that there was insufficient power to detect a difference between the groups. Second, pain and tender points may respond differently to treatment. Indeed, some studies have found that tender point measures and generalized pain are only modestly associated (20) and may represent different aspects of pain in fibromyalgia (21). Third, current methods for measuring tender points may be inadequate to detect significant change in clinical trials; recent attempts to standardize tender point examinations may help (22). Finally, although there is evidence that tender point evaluations have test-retest stability when repeated after 1 week (23), further studies are needed to determine if the reliability of tender point assessments remains stable for the longer periods required in a clinical trial. The mechanism of action of fluoxetine in fibromyalgia remains unknown., an SSRI, may correct an abnormality of serotonin neurotransmission. It is also unknown whether the effect of fluoxetine on different measures is caused by common effects, or by effects on certain measures that in turn improve other measures. Our study demonstrates that higher doses of fluoxetine may be needed to achieve a therapeutic response in some patients. Although subjects with major depressive disorder at the time of the study were excluded, most had a history of major depressive disorder. The high comorbidity of a history of major depressive disorder with fibromyalgia and the positive response of fibromyalgia to fluoxetine support the hypothesis that depression and fibromyalgia share pathophysiological similarities (24,25). Several limitations of this study should be considered. First, the Fibromyalgia Impact Questionnaire is limited by single-item measures of symptoms associated with fibromyalgia, such as depression and nonrestorative sleep. A validated measure that reliably and thoroughly evaluates the full range of symptom domains in fibromyalgia is needed. Second, because the duration of treatment was 12 weeks, the results may not be generalized to longer treatment periods. Since fibromyalgia is a chronic condition that usually requires treatment for longer than 3 months, future studies should assess the long-term efficacy of fluoxetine. Third, patients with all current and several forms of lifetime psychopathology were excluded. Thus, the results may not be generalized to those with certain forms of comorbid psychopathology. However, major depressive disorder is one of the most common comorbid conditions in patients with fibromyalgia, and exclusion of patients with current depression might lead to a more conservative estimate of the effects of fluoxetine. Fourth, the group was relatively small, and our study lacked power to detect potentially relevant differences between groups, particularly on tender point measures. Finally, although the study did not exclude men, only women were enrolled, reflecting the higher prevalence of fibromyalgia in women (2). The results of this study may not therefore be generalized to men. In summary, in a 12-week, randomized, placebo-controlled, flexible-dose trial, fluoxetine was found to be effective on most outcome measures and generally well tolerated in women with fibromyalgia. REFERENCES 1. Wolfe F, Smythe HA, Yunus MB, et al. The American College of Rheumatology 1990 criteria for the classification of fibromyalgia. Report of the Multicenter Criteria Committee. Arthritis Rheum. 1990;33: February 15, 2002 THE AMERICAN JOURNAL OF MEDICINE Volume 112

7 2. Wolfe F, Ross K, Anderson J, et al. The prevalence and characteristics of fibromyalgia in the general population. Arthritis Rheum. 1995;38: Arnold LM, Keck PE Jr, Welge JA. Antidepressant treatment of fibromyalgia. A meta-analysis and review. Psychosomatics. 2000;41: O Malley PG, Balden E, Tomkins G, et al. Treatment of fibromyalgia with antidepressants: A meta-analysis. J Gen Intern Med. 2000; 15: Versiani M, Ontiveros A, Mazzotti G, et al. versus amitriptyline in the treatment of major depression with associated anxiety (anxious depression): a double-blind comparison. Int Clin Psychopharmacol. 1999;14: Nørregaard J, Volkmann H, Danneskiold-Samsø B. A randomized controlled trial of citalopram in the treatment of fibromyalgia. Pain. 1995;61: Anderberg UM, Marteinsdottir I, von Knorring L. Citalopram in patients with fibromyalgia a randomized, double-blind, placebocontrolled study. Eur J Pain. 2000;4: Wolfe F, Cathey MA, Hawley DJ. A double-blind placebo controlled trial of fluoxetine in fibromyalgia. Scand J Rheumatol. 1994; 23: Goldenberg DL, Mayskiy M, Mossey C, et al. A randomized, double-blind crossover trial of fluoxetine and amitriptyline in the treatment of fibromyalgia. Arthritis Rheum. 1996;39: Preskorn S. Targeted pharmacokinetics in depression management: comparative pharmacokinetics of fluoxetine, paroxetine, and sertraline. Int Clin Psychopharmacol. 1994;9: Aranow RB, Hudson JI, Pope HG Jr, et al. Elevated antidepressant plasma levels following addition of fluoxetine. Am J Psychiatry. 1989;146: Schmider J, von Moltke LL, Shader RI, et al. Extrapolating in vitro data on drug metabolism to in vivo pharmacokinetics: evaluation of the pharmacokinetic interaction between amitriptyline and fluoxetine. Drug Metab Rev. 1999;31: First MB, Spitzer RL, Gibbon M, et al. Structured Clinical Interview for the DSM-IV Axis I Disorders Patient Edition (SCID-I/P), Version 2.0). New York: Biometrics Research Department, New York State Psychiatric Institute; American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition. Washington, DC: American Psychiatric Association; Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry. 1960;23: Burckhardt CS, Clark SR, Bennett RM. The Fibromyalgia Impact Questionnaire: development and validation. J Rheumatol. 1991;18: Melzack R. The McGill Pain Questionnaire: major properties and scoring methods. Pain. 1975;1: Fischer AA. Pressure algometry over normal muscles. Standard values, validity and reproducibility of pressure threshold. Pain. 1987; 30: Dwight MM, Arnold LM, O Brien H, et al. An open clinical trial of venlafaxine treatment of fibromyalgia. Psychosomatics. 1998;39: Nicassio PM, Weisman MH, Schuman C, Young CW. The role of generalized pain and pain behavior in tender point scores in fibromyalgia. J Rheumatol. 2000;27: Jacobs JWG, Rasker JJ, van der Heide A, et al. Lack of correlation between the mean tender point score and self-reported pain in fibromyalgia. Arthritis Care Res. 1996;9: Okifuji A, Turk DC, Sinclair JD, et al. A standardized manual tender point survey. I. Development and determination of a threshold point for the identification of positive tender points in fibromyalgia syndrome. J Rheumatol. 1997;24: Jacobs JWG, Geenen R, van der Heide A, et al. Are tender point scores assessed by manual palpation in fibromyalgia reliable? Scand J Rheumatol. 1995;24: Hudson JI, Pope HG. Fibromyalgia and psychopathology: is fibromyalgia a form of affective spectrum disorder? J Rheumatol. 1989; 16(suppl 19):S15 S Hudson JI, Goldenberg DL, Pope HG Jr, et al. Comorbidity of fibromyalgia with medical and psychiatric disorders. Am J Med. 1992;92: February 15, 2002 THE AMERICAN JOURNAL OF MEDICINE Volume