Background: Although the paramedian approach for epidural blockade is useful in some clinical situation, the parameters which are correlated with the distance from skin to the epidural space has not been established. Methods: We studied in 143 patients having elective continuous epidural blocks for relief of postoperative pain. All blocks were performed using paramedian approach with Tuohy needle in the lumbar (group 1, n=100) and thoracic (group 2, n=45) area. We measured the distance from skin to the epidural space, body weight, height, and the angle between the shaft of the needle and the skin. Data were analyzed by linear regression. The relationships between parameters identified by the F-test with a P value of less than 0.05 were considered statistically significant. Results: The mean distance from skin to the lumbar epidural space was $4.4{pm}0.7$ cm. significant correlation between the body weight and the depth of lumbar epidural space ($gamma$ value : 0.492) was noted with regression equation of depth(cm)=2.293+0.034${ imes}$body weight (kg). Also the significant correlation between the ponderal index (PI) and the depth of lumbar epidural space ($gamma$ value : 0.539) was noted with regression equation of depth(cm)=1.703+0.07${ imes}$PI, The mean distance from skin to the thoracic epidural space was $5.2{pm}0.7cm$ which did not correlated with other anatomic measurements. Conclusion: We found that PI and body weight are the suitable predictors of the depth of the lumbar epidural space, but not the thoracic epidural space.