Thirteen studies comprising four RCTs (n=253) and nine observational studies (n=325) were included in the review. Overall, study quality was low. All of the RCTs and most of the observational studies had limitations in their methodology, details of which were reported in the review. Sample sizes ranged from 19 to 102.
All of the studies with the exception of one observational study (n=55) reported a beneficial effect with fusion. An overall statistically significant benefit for clinical outcome was reported for fusion in comparison with decompression alone (relative risk was 1.40; 95% CI: 1.04 to 1.89, p<0.05; two RCTs and five comparative observational studies). However, this finding was associated with significant statistical heterogeneity (I2=59.3%), which was mainly due to the inclusion of one observational study.
Sensitivity analyses suggested that RCTs reported a larger effect size and did not show significant levels of heterogeneity. Instrumented spinal fusion was associated with a significantly increased probability of solid fusion (relative risk was 1.37; 95% CI: 1.07 to 1.75, p<0.05; two RCTs and three comparative observational studies) in comparison with non-instrumented fusion; no significant improvement in clinical outcome was reported. Both analyses were associated with significant levels of heterogeneity (I2=69.9% and 59%) that were shown in further analyses to be due to the inclusion of particular studies. Again RCTs were reported as showing greater effect sizes.
Details of further analyses were reported in the review.