Do antidepressant medications relieve chronic low back pain?

Journal of Family Practice, Dec, 1993 by Judith A. Turner, Mary C. Denny

Results

We found only six controlled or comparison studies that evaluated the efficacy of antidepressant medication for low back pain. The studies and their sample sizes, medications and dosages, length of trial, and results reported are listed in the Table. Three studies[31-33] were randomized, placebo-controlled trials. One study[34] used a cross-over design with randomly assigned treatment-placebo sequences. Two uncontrolled studies by Ward and colleagues[35,36] compared two antidepressants, but it was not stated whether subjects were randomly assigned to treatments. Ward's 1986 study[35] was an extension of the 1984 study by Ward et al,[36] and included data from subjects in the 1984 study. The six identified studies varied greatly in terms of subject characteristics (eg, depressive symptoms, level of pain intensity and disability, duration of pain), antidepressant medication(s) studied, dosage, study design, and measures.

Is antidepressant medication superior to placebo in decreasing pain? Only three studies attempted to determine whether antidepressant medication was superior to placebo in decreasing pain. Of the three, two found no difference. Jenkins el al[33] reported similar improvements in patient ratings of pain on a visual analogue scale and in clinician ratings of patients' pain level in imipramine and placebo groups. However, no statistical analyses were reported. Goodkin et al[32] found no difference between trazodone and a placebo on visual analogue scale ratings of pain intensity. The third study, by Alcoff et al,[31] reported a borderline (P = .058) statistically significant difference between imipramine and placebo on patient ratings of back pain severity.

Research by Ward and colleagues[36] also addressed the efficacy of antidepressants for pain relief after "eliminating placebo responders" following a 2-week placebo washout period. This study compared doxepin and desipramine to examine whether a more sedating antidepressant (doxepin) might prove more useful for chronic low back pain. Patients in both antidepressant groups reported significantly reduced levels of pain severity (mean for all patients decreased from 5.8 to 3.0 on a scale of 1 to 10, P [less than] .005) and frequency (from pain experienced 90% of the time to pain experienced 52% of the time, P [less than] .01). Pain severity, but not frequency, decreased significantly more in the doxepin than in the desipramine group (data and statistics were not reported for this finding). The lack of a placebo group makes it impossible to assess how much of the change in reported pain is attributable to specific effects of the antidepressants and how much represents improvement associated with the natural course of the pain or the nonspecific effects of being in a pain treatment trial.

If antidepressants do exert more than a placebo influence on pain, there should be some kind of a relation between blood levels of medication and pain symptoms. (Of course, this may not be a linear relation; medication may nonly need to be avove a certain threshold level to affect pain, or there may be a therapeutic window.) Only two studies reported data bearing on this relation; one found no linear relation between imipramine blood levels and patient symptoms,[31] and the other found that increased trazodone blood level was correlated with increased pain as measured by a visual analogue scale.[32]