Context Low back pain (LBP) contributes to considerable disability and lost wages in the United States. Commonly used opioid and nonopioid analgesic drugs produce adverse effects and are of limited long-term benefit in the management of this patient population. Objective To compare the effectiveness of a novel nonpharmacologic pain therapy, percutaneous electrical nerve stimulation (PENS), with transcutaneous electrical nerve stimulation (TENS) and flexion-extension exercise therapies in patients with long-term LBP. Design A randomized, single-blinded, sham-controlled, crossover study from March 1997 to December 1997. Setting An ambulatory pain management center at a university medical center. Patients Twenty-nine men and 31 women with LBP secondary to degenerative disk disease. Interventions Four therapeutic modalities (sham-PENS, PENS, TENS, and exercise therapies) were each administered for a period of 30 minutes 3 times a week for 3 weeks. Main Outcome Measures Pretreatment and posttreatment visual analog scale (VAS) scores for pain, physical activity, and quality of sleep; daily analgesic medication usage; a global patient assessment questionnaire; and Health Status Survey Short Form (SF-36). Results PENS was significantly more effective in decreasing VAS pain scores after each treatment than sham-PENS, TENS, and exercise therapies (after-treatment mean ± SD VAS for pain, 3.4 ± 1.4 cm, 5.5 ± 1.9 cm, 5.6 ± 1.9 cm, and 6.4 ± 1.9 cm, respectively). The average ± SD daily oral intake of nonopioid analgesics (2.6 ± 1.4 pills per day) was decreased to 1.3 ± 1.0 pills per day with PENS (P<.008) compared with 2.5 ± 1.1, 2.2 ± 1.0, and 2.6 ± 1.2 pills per day with sham-PENS, TENS, and exercise, respectively. Compared with the other 3 modalities, 91% of the patients reported that PENS was the most effective in decreasing their LBP. The PENS therapy was also significantly more effective in improving physical activity, quality of sleep, and sense of well-being (P<.05 for each). The SF-36 survey confirmed that PENS improved posttreatment function more than sham-PENS, TENS, and exercise. Conclusions In this sham-controlled study, PENS was more effective than TENS or exercise therapy in providing short-term pain relief and improved physical function in patients with long- term LBP.