Twenty-eight patients with primary fibromyalgia participated in the study. Eight patients received a stellate ganglion blockade with bupivacaine, and 14 days later an intravenous regional sympathetic blockade with guanethidine. The remaining patients served as controls and were randomly allocated to receive either a sham (placebo) injection with physiologic saline superficial to the stellate ganglion (n = 10) or bupivacaine intramuscularly (n = 10). The efficiency of the stellate ganglion blockade was evaluated by measuring skin blood flow (using a laser Doppler flowmeter), skin temperature, and skin conductance responses ('sympathogalvanic reflex'). Trigger and tender points (TePs) were counted, and rest pain in the arm, shoulder and neck evaluated at intervals up to 4 h after the injection. The guanethidine blockade was evaluated 24 h after the injection by counting TePs and by assessment of rest pain in the hand and forearm. The results indicate that a complete sympathetic blockade, produced by a stellate ganglion blockade, markedly reduced the number of TePs and produced a marked decrease in rest pain. The guanethidine blockade reduced the number of TePs, but had no effect on rest pain. The reduction in pain and TePs produced by a sympathetic blockade may be due to an improvement in microcirculation. Sympathetic activity may, in some patients, contribute to the pathogenesis of primary fibromyalgia.