Background:

Conventional paired meta-analyses have shown inconsistent results regarding the safety and efficacy of different interventions.

Objective:

To perform a network meta-analysis (NMA) and systematic review based on randomized controlled trials (RCTs) evaluating the efficacies of different interventions for lumbar spinal stenosis (LSS).

Methods:

We searched PubMed, Embase, Cochrane Library, Web of Science, and major scientific websites from inception to October 10, 2019, for randomized controlled trials comparing the nine most commonly used interventions for LSS. The main outcomes were disability and pain intensity. The PROSPERO number was CRD42020154247.

Results:

First, laminotomy was better in improving patients’ short- and long-term dysfunction (probability 49% and 25%, respectively). Second, decompression, decompression plus fusion, endoscopic decompression, interspinous process spacer device implantation, laminectomy, laminotomy and minimally invasive decompression were significantly more efficacious in relieving pain than non-surgical interventions (mean difference in the short-term -21.82, -22.00, -16.68, -17.47, -17.75, -17.61 and -18.86; in the long-term -37.14, -34.04, -34.07, -39.79, -36.14, -32.75 and -39.14, respectively). Third, endoscopic decompression had a lower complication rate (probability 51%). In addition, laminotomy had a lower reoperation rate (probability 45%). Fourth, decompression plus fusion resulted in more blood loss than any other surgical intervention (probability 96%). Finally, endoscopic decompression had the shortest hospitalization time (probability 96%).

Conclusions:

There were no significant differences among the different interventions in improving patient function. Surgical interventions were associated with better pain relief but a higher incidence of complications. Decompression plus fusion is not necessary for patients. In addition, endoscopic decompression as a novel and less invasive surgical approach may be a good choice for LSS patients.