It has become increasingly clear that depressed patients quite often suffer from some degree of physical pain, which may adversely impact treatment outcome. Indeed, research suggests that as the pain severity of patients increases, the odds of achieving a successful antidepressant response decreases (Bair MJ et al., Psychosom Med 2004;66:17-22). Given that pain is associated with poor treatment response, a clear need exists for treatments that can effectively target both depression and physical pain.

Enter Cymbalta (duloxetine), Eli Lilly’s popular antidepressant. Most clinicians are probably familiar with Lilly’s “Depression Hurts” marketing campaign, which highlights the link between depression and physical pain. Cymbalta has been marketed as a treatment that can alleviate both typical depressive symptoms and physical pain. Indeed, Cymbalta has FDA indications for the treatment of diabetic neuropathic pain and fibromyalgia, which suggest that it might be a good choice for depressed patients who are also in physical pain. Does the scientific literature support such a notion?

Several clinical trials have compared Cymbalta to a placebo in treating depression. In some of these studies, physical pain was also assessed. A meta analysis including data from 714 patients taking duloxetine and 562 patients taking placebo from five such trials found that the advantage of Cymbalta over placebo was not quite statistically significant (p = .057). Several measures were used in the trials to measure pain. Overall, the effect size was 0.115, which statisticians generally consider as indicating a very small effect (Spielmans GI., Psychother Psychosom 2008;77:12-16). A recent placebo-controlled trial involving 327 patients who presented with at least moderate pain found a statistically significant difference: on the Brief Pain Inventory (BPI), each individual item and the total score showed a statistically significant advantage for Cymbalta (Brecht S et al., J Clin Psychiatry 2007;68:1707-1716). However, in a letter to the editor, a group of researchers argued that while the benefits of Cymbalta over placebo might have been statistically significant, they were not clinically significant. They noted that Cymbalta showed less than a one point advantage over placebo on the BPI 24-hour average pain item, which was the primary outcome measure (Griffith JM et al., J Clin Psychiatry 2008;69:1022).

Additionally, three Lilly-sponsored Cymbalta trials included pain measures but did not fully report outcomes; these trials only reported that some results were statistically significant but provided no numbers on any of the pain outcome measures. One trial found an advantage over placebo for over half of the pain measures used (Detke MJ et al., Eur Neuropsychopharmacol 2004;14:457-470) while another study found an advantage for Cymbalta on only one of several pain outcome measures (Perahia et al., Eur Psychiatry 2006;21:367-378). A recent study in elderly patients found that that Cymbalta offered a statistically significant advantage over placebo on two of six pain measures (Raskin J et al., Am J Psychiatry 2007;164:900-909). When examining all available data on Cymbalta as an analgesic for depressed patients, it appears to offer, at best, a small benefit over placebo. While many depressed patients on Cymbalta experience a substantial decrease in pain, it appears that a placebo offers many of the same analgesic benefits.

How does Cymbalta compare to other antidepressants? Five trials have compared Cymbalta to Paxil (paroxetine) in treating pain in depression. However, two of these Lilly-sponsored trials did not report how pain outcomes compared between the two drugs, which may indicate that Cymbalta did not outperform Paxil in these trials. The Spielmans meta-analysis found no difference between Cymbalta and Paxil in two trials (p = .76). An additional trial had similar results (Lee P. et al.; Psychiatry Clin Neurosci 2007;61:295-307).

While a handful of studies have compared Cymbalta to Lexapro (escitalopram) and Effexor (venlafaxine), pain measures were not a focus of these investigations. For depression, the studies showed mixed results, with two trials showing statistically significant but small advantages for Lexapro on some measures (Khan A et al., Clin Drug Investig 2007;27:481-492; Wade A et al., Curr Med Res Opin 2007;23:1605-1614) and another finding no difference between the drugs (Pigott TA et al., Curr Med Res Opin 2007;23:1303- 1318). A trial comparing Cymbalta to Effexor found that the drugs produced similar antidepressant effects (Perahia DGS et al., J Psychiatr Res 2008;42:22-34). At this point, there is no convincing clinical trial data suggesting that Cymbalta relieves the pain accompanying depression more than any other antidepressant.

While Cymbalta appears to offer nothing special for pain accompanying depression, it is still possible that it is more robust than other antidepressants for other pain conditions, such as diabetic neuropathy and fibromyalgia. Unfortunately, there is no research directly comparing Cymbalta to other antidepressants for either of these pain syndromes. As another article in this issue shows, both Cymbalta and tricyclics have demonstrated efficacy for fibromyalgia, while the evidence for SSRIs is mixed. In addition, tricyclics have demonstrated efficacy for pain relief in diabetic neuropathy (Wong M et al., BMJ 2007;335:87).

To summarize, Cymbalta appears little better than a placebo in alleviating pain in depression and is no more effective than Paxil, and probably by extension, other SSRIs, in treating such symptoms.

TCPR Verdict:

Cymbalta: No advantage over SSRIs or Effexor for painful depression