The concept of minimally invasive lumbar disc surgery comprises reduced muscle injury. The aim of this study was to evaluate creatine phosphokinase (CPK) in serum and the cross-sectional area (CSA) of the multifidus muscle on magnetic resonance imaging as indicators of muscle injury. We present the results of a double-blind randomized trial on patients with lumbar disc herniation, in which tubular discectomy and conventional microdiscectomy were compared. In 216 patients, CPK was measured before surgery and at day 1 after surgery. In 140 patients, the CSA of the multifidus muscle was measured at the affected disc level before surgery and at 1 year after surgery. The ratios (i.e. post surgery/pre surgery) of CPK and CSA were used as outcome measures. The multifidus atrophy was classified into three grades ranging from 0 (normal) to 3 (severe atrophy), and the difference between post and pre surgery was used as an outcome. Patients' low-back pain scores on the visual analogue scale (VAS) were documented before surgery and at various moments during follow-up. Tubular discectomy compared with conventional microdiscectomy resulted in a nonsignificant difference in CPK ratio, although the CSA ratio was significantly lower in tubular discectomy. At 1 year, there was no difference in atrophy grade between both groups nor in the percentage of patients showing an increased atrophy grade (14% tubular vs. 18% conventional). The postoperative low-back pain scores on the VAS improved in both groups, although the 1-year between-group mean difference of improvement was 3.5 mm (95% CI; 1.4-5.7 mm) in favour of conventional microdiscectomy. In conclusion, tubular discectomy compared with conventional microdiscectomy did not result in reduced muscle injury. Postoperative evaluation of CPK and the multifidus muscle showed similar results in both groups, although patients who underwent tubular discectomy reported more low-back pain during the first year after surgery.