Objective To evaluate the effectiveness and safety of different injection sites for collagenase chemonucleolysis for lumbar intervertebral disc protrusion (LIDP). Methods We searched for randomized controlled trials (RCTs) and quasi-RCTs in the following electronic databases: Pubmed (1966 to May 2006), EMBASE(1966 to May 2006), The Cochrane library (Issue 2, 2006), CRD(Center for Reviews and Dissemination),York University, CBM (1978 to May 2006 ), CNKI(1994-2006)and VIP(1989-2006). Quality assessment and data extraction were conducted by two reviewers independently. Disagreement was resolved through discussion. Results Eight studies involving a total of 1035 participants met the inclusion criteria. Meta-analysis was not carried out because of apparent heterogeneity. Four studies made comparisons among intradisc, extradisc, and combined intra- and extra-disc injection. One study (62 participants) showed that intradisc injection was superior to extradisc injection (RR 3.71, 95% CI 1.19 to 11.58, Plt;0.05). Another study (240 participants) showed that intradisc injection was superior to combined intra- and extra-disc injection after 3 months and 6 months of follow-up (RR 0.88, 95% CI 0.80 to 0.98, Plt;0.05). The other two studies showed no significant difference among intradisc, extradisc, and combined intra- and extra-disc injection. Four studies (436 participants in total) showed that amongst three extradisc injections, both anterior epidural space injection and lateral epidural space injection were superior to posterior epidural space injection (Plt;0.05). Although one study indicated that anterior epidural space injection was superior to lateral epidural space injection (RR 1.24, 95% CI 1.03 to 1.51, Plt;0.05), no statistical significance was found between anterior epidural space injection and lateral epidural space injection in two other studies (Pgt;0.05). Conclusions There is insufficient evidence to identify which injection site for collagenase is the most effective for lumbar intervertebral disc protrusion. The included studies showed that both anterior and lateral epidural space injection were superior to posterior epidural space injection. However, these studies are too small and poor quality, and have different diagnostic criteria, follow-up time points, outcome measures and efficacy parameters. Thus, more high quality and large-scale RCTs are needed.