Current issues of ACP Journal Club are published in Annals of Internal Medicine


Amitriptyline or cyclobenzaprine provided only short-term benefit for fibromyalgia

ACP J Club. 1994 July-Aug;121:10. doi:10.7326/ACPJC-1994-121-1-010

Source Citation

Carette S, Bell MJ, Reynolds WJ, et al. Comparison of amitriptyline, cyclobenzaprine, and placebo in the treatment of fibromyalgia. A randomized, double-blind clinical trial. Arthritis Rheum. 1994 Jan;37:32-40.



To compare the efficacy of amitriptyline (AT), cyclobenzaprine (CB), and placebo in the treatment of fibromyalgia and to identify predictors of response.


Randomized, double-blind, placebo-controlled, 6-month trial.


9 university centers and 2 private practices in Canada.


208 patients (mean age 45 y, 94% women) who met the 1990 American College of Rheumatology criteria for fibromyalgia; scored ≥ 4 on at least 1 of 2 self-administered visual analog scales (VAS) evaluating pain and global assessment of fibromyalgia symptoms; and had normal erythrocyte sedimentation rates, creatine kinase levels, and thyroid-stimulating hormone levels. Exclusion criteria were inflammatory rheumatic disease; untreated endocrine, neurologic, infectious, or osseous disorders; previous treatment with AT or CB; or a history of glaucoma, urinary retention, myocardial infarction, arrhythmia, or congestive heart failure. Follow-up was 88%.


84 patients were assigned to receive placebo CB plus AT, 10 mg nightly during week 1, 25 mg nightly for weeks 2 through 12, and 50 mg nightly for the 12 weeks thereafter. 82 patients received placebo AT plus CB, 10 mg nightly during week 1, 20 mg daily for weeks 2 through 12, and 10 mg in the morning plus 20 mg at bedtime for the next 12 weeks. 42 patients received placebo tablets of the active drugs.

Main outcome measures

50% improvement in at least 4 of the following criteria: pain, sleep, fatigue, patient and physician global assessment, or 1-kg increase in mean total myalgia score obtained by dolorimeter to measure point tenderness.

Main results

After 1-month improvement was shown in 21% of patients receiving AT, 12% receiving CB, and none of those receiving placebo (AT vs placebo, P = 0.002; CB vs placebo, P = 0.02). At 3 and 6 months, improvement continued but the differences among the 3 groups did not reach statistical significance. Patients receiving AT or CB had greater improvement in all VAS compared with placebo (P < 0.05). No differences existed between the 2 active treatment groups for any variables. Myalgia scores did not improve. Predictors of a favorable response to both AT and CB at 1 month were a normal profile on the Minnesota Multiphasic Personality Inventory at baseline (odds ratio [OR] 3.3, 95% CI 1.2 to 9.0) and higher educational level (OR 2.4, CI 1.0 to 5.9). Many adverse effects were noted in all groups (98% in the cyclobenzaprine group, 95% in the amitriptyline group, and 62% in the placebo group) (no significant differences among groups). Withdrawal from the study because of adverse effects was also high although no statistically significant differences were shown among the groups (8% for amitriptyline, 16% for cyclobenzaprine, and 5% for placebo).


Amitriptyline and cyclobenzaprine elicited favorable responses over placebo in a small percentage of patients with fibromyalgia after 1 month, but by 6 months improvement in the active drug groups was no different from that in the placebo group.

Sources of funding: Canadian Arthritis Society and Merck Frosst Canada.

For article reprint: Dr. S. Carette, Toronto Hospital, Western Division, 399 Bathurst Street, FP-1224, Toronto, Ontario M5T 2S8, Canada. FAX 416-603-4348.


The study by Carette and colleagues is one of several that provides evidence that CB and AT, which are very closely related (only 1 double bond distinguishes them), may be useful in the management of fibromyalgia. This study was longer in duration than those done previously, and the improvement in the 2 active-drug groups did not remain significantly different from that seen in the placebo group over the longer term.

Two important differences between clinical practice and this clinical research must be borne in mind. First, in administering tricyclic drugs, the experienced clinician titrates the dose to maximize clinical effect. In this study, every patient was ultimately maintained on 50 mg/d of AT. Those of us who prescribe a lot of AT find that some patients achieve acceptable results on as little as 10 mg/d, whereas others do not improve until the dose is raised to 75, 100, or even 150 mg/d.

Second, in day-to-day practice, patients not given any drug would ordinarily not enjoy the extensive attention lavished on these research participants (who spent many hours being examined and answered hundreds of questions) and might, instead, feel underserved and neglected.

Combining the information gained from this and other studies with anecdotal evidence, it would appear that some patients with fibromyalgia can be expected to improve slowly when provided with either a large measure of special attention and hope or with moderate doses of tricyclics. CB and AT may be equally efficacious and are about equally well tolerated. Dry mouth and weight gain are common side effects. Dizziness and somnolence were the adverse reactions which most commonly led to withdrawal from the study. When low doses fail, higher doses are recommended. The largest clinical experience has been with AT.

Mary E. Moore, PhD, MD
Albert Einstein Medical CenterPhiladelphia, Pennsylvania, USA