Experimental data have suggested that in neuropathic pain, tricyclic antidepressants may work solely through a β2-agonist action. The aim of this study was to test if the β2-agonist terbutaline relieves painful polyneuropathy. The study was a randomized, double-blind, placebo- and active-controlled, 3-way, cross-over trial among patients with painful polyneuropathy. The treatment periods were of 5 weeks’ duration and were preceded by 1 week for washout and 1 week for baseline observations. The patients received terbutaline (5 to 15 mg), imipramine (30 to 150 mg), or placebo in random order. Drug doses depended on age and metabolizer status. Change in total pain recorded from ratings in diaries (NRS 0-10) was the primary outcome and change in rating of specific pain symptoms (NRS 0-10), patient global impression of change and sleep disturbance were secondary outcomes. Forty-seven patients were randomized. Median score for total pain changed from NRS 6.4 to 6.1 from baseline to week 5 on terbutaline with an average effect during the treatment period as compared to placebo of 0.13 (95% CI -0.12;0.38, p = 0.32). Median score for total pain on imipramine changed from NRS 6.6 to 4.8 with an average effect as compared to placebo of -1.17 (95% CI -1.42; -0.92, p < 0.001). Secondary outcomes were also unaltered by terbutaline but improved by imipramine. The β2-agonist terbutaline has no effect in painful polyneuropathy. β2 -agonism seems not to be an important mechanism of action of TCAs in neuropathic pain.