Abstract
Purpose: A lack of consensus regarding the radiologic criteria to diagnose spinal non-union limits inferences from clinical research. This systematic review aimed to examine the spectrum of radiologic investigations used to assess lumbar spinal fusion and the definitions of successful spine union used in the spine literature.
Method: We comprehensively searched three electronic databases from 1950 to 2009 (MEDLINE, Embase and the Cochrane Central Register of Controlled Trials) for clinical studies involving posterolateral fusion of the lumbar spine. English-language studies including adult patients and reporting a definition of successful fusion were included. Studies examining the reliability and validity of radiologic investigations were also identified. Key measures included
-
radiologic investigations,
-
definition of successful lumbar fusion and
-
reliability, sensitivity and specificity of the investigations used to assess the spinal fusion.
Results: Among 1165 potentially eligible studies, 91 met our inclusion criteria. Of the studies 78% (n = 71) used plain radiographs to diagnose non-union, 4% (n = 4) used CT scans and 18% (n = 16) used both. Fifty-one studies used both static (xray or CT) and dynamic (flexion-extension xray) images, 35 used only static images and five used only dynamic radiographs. In total, we identified fifty-two different radiographic definitions of successful fusion. More than half of the studies (n = 50, 55%) failed to provide a reference for the definition used. The most common definition of fusion (7 studies) used static radiographs and defined fusion as continuous intertransverse bony bridging with this quality of fusion at all intended levels. Seven studies evaluated reliability of xray criteria but no studies provided complete validation of the definitions. Only 3 studies provided some validation and reliability estimates of thin-slice CT scanning in diagnosing spinal non-union. Significant variability in reliability, sensitivity and specificity exists for all radiologic investigations in the diagnosis of spinal non-union.
Conclusion: The radiologic investigations and definitions of successful posterolateral fusion used in the spine literature vary substantially. Choice of radiologic criteria should be based upon reliability and validity testing. Studies using fusion criteria that have not been shown to be reliable or valid should be interpreted with caution.
Correspondence should be addressed to: COA, 4150 Ste. Catherine St. West Suite 360, Westmount, QC H3Z 2Y5, Canada. Email: meetings@canorth.org
