The evidence reviewed here consists of retrospective, unselected individuals, and as such represent a biased sample of traumatic central cord individuals. Generally, a common management strategy is to conservatively manage those individuals who have mild pathology or show spontaneous early neurological improvement, and to consider surgery for those that fail to improve or who worsen. As such, these publications likely represent individuals that failed a period of initial observation. Furthermore, most studies here do not enlist a control group to compare surgical versus nonsurgical intervention, which precludes a robust assessment of the efficacy of decompression.
Aito et al. (2007a) reported their retrospective comparison of surgically and non-surgically treated individuals and is the only study here to use a non-surgical control group. They found no effect of surgery on neurologic outcome. This is consistent with Dvorak et al. (2005) who demonstrated no difference in long-term neurologic outcomes in surgically versus non-surgically managed individuals with traumatic central cord syndrome.
The largest case series for this population is that of Samuel et al. (2015) which followed 1060 surgically treated individuals. Using multivariate analysis, they reported an association between earlier surgery and mortality. Earlier surgery was also associated with minor, but not major, adverse events. Because other associations with mortality included advanced age and comorbidity status, it is possible that medically optimizing and stabilizing these individuals before surgical intervention may reduce minor complications and confer some survival advantage. This study did not assess neurologic outcome. This finding regarding advanced age is consistent with the Aito et al. (2007a) and Chen et al. (2009a) studies.
Other assessments of surgical timing in central cord syndrome reveal mixed results with respect to neurologic recovery. Kepler et al. (2015) found that early (<24 hours) surgery conferred no advantage in terms of AIS motor recovery, but that age was a significant predictor of worse neurological outcome. Anderson et al. (2012a) found no association between surgical timing and AIS motor outcome. In a cohort study Jug et al. (2015) found a significant neurological benefit to those receiving very early surgery (<8 hours) compared to “late” surgery which was performed 8-24 hours post SCI. Guest et al. (2002b) found that early surgery conferred a neurological recovery advantage in those individuals with an acute lesion such as an acute disc herniation or a compressive fracture, but not in those with a chronic process such as spondylosis.
Chen et al. (2009a) followed surgically treated individuals and noted that at 6 months, there were improvements in AIS motor score, but there was no non-surgical comparator group. Among surgically treated individuals, there were no differences in health-related quality of life when assessing early (<4 days) versus late surgery. Similar to previous studies, older age was a negative prognostic factor.
Liu et al. (2017) reported on a series of 11 individuals presenting who underwent surgical treatment for central cord syndrome in a delayed fashion (mean of 90 days after injury). In this cohort, postoperative AIS and quality of life scores showed improvement, but the absence of a nonsurgical comparator group again makes conclusions difficult.
There is level 3 evidence (based on one case control; (Aito et al., 2007a) that surgical treatment of traumatic central cord syndrome does not confer neurologic benefit compared to conservative management.
There is conflicting level 2 and 4 evidence (based on one cohort (Jug et al., 2015) and five case series studies; (Chen et al., 2009a; R. Guest, Craig, Tran, & Middleton, 2015a; Kepler et al., 2015; Y. Liu et al., 2017; Samuel, Bohl, et al., 2015a)that early decompression of central cord lesions have similar neurologic outcomes to late surgery but that the latter may have lower mortality, possibly due to age-and-comorbidity-related factors.