We investigated the effects of continuous lumbar traction in patients with lumbar disc herniation on clinical findings, and size of the herniated disc measured by computed tomography (CT). In this prospective, randomized, controlled study, 46 patients with lumbar disc herniation were included, and randomized into two groups as the traction group (24 patients), and the control group (22 patients). The traction group was given a physical therapy program and continuous lumbar traction. The control group was given the same physical therapy program without traction, for the same duration of time. Data for the clinical symptoms and signs were collected before and after the treatment together with calculation of a herniation index, from the CT images that showed the size of the herniated disc material. In the traction group, most of the clinical findings significantly improved with treatment. Size of the herniated disc material in CT decreased significantly only in the traction group. In the traction group the herniation index decreased from 276.6 +/- 129.6 to 212.5 +/- 84.3 with treatment (p < 0.01). In the control group, pretreatment value was 293.4 +/- 112.1, and it decreased to 285.4 +/- 115.4 after the treatment (p > 0.05). Patients with greater herniations tended to respond better to traction. In conclusion, lumbar traction is both effective in improving symptoms and clinical findings in patients with lumbar disc herniation and also in decreasing the size of the herniated disc material as measured by CT.