Outcome of Delayed Surgery in Treatment of Acute Traumatic Central Cord Syndrome without Fracture or Instability: A Retrospective Study

Background: The optimal surgical timing for acute traumatic central cord syndrome (ATCCS) without fracture or instability has not been established. The purpose of this study is to explore the outcome of delayed surgery in treatment of ATCCS and to investigate potential factors associated with postoperative neurological improvement. Methods: Patients who underwent delayed surgery for ATCCS with at least 2 year follow up time were retrospectively reviewed. Parameters including age, gender, traumatic mechanism, interval to operation, surgical procedures, and complications were collected. Magnetic resonance imaging was performed to determine levels of spinal cord signal change and concomitant pathology. American Spinal Injury Association (ASIA) classication and Japanese Orthopedic Association (JOA) scores were evaluated and compared at admission and the 2 year follow-up visit for neurologic function assessment. Correlations of neurological improvement and age, traumatic mechanism, interval to operation, surgical procedures, concomitant pathology, and preoperative neurological function were investigated by Spearman ’ s correlation test. Results: A total of 39 patients (M:F=28:11, mean age 52.2±10.4 yrs) were enrolled into this study. 21 cases were caused by falls followed by 16 by motor-vehicle accidents and 2 by sports. 19 patients presented with preexisting cervical disc herniation (CDH) (cid:0) 12 with spinal canal stenosis (SCS), 5 with OPLL, and 3 with a combined pathology of CDH and CS. 14 samples received ACDF procedure, 8 obtained ACCF, and 17 underwent posterior unilateral open-door laminoplasty. The mean interval from trauma to surgery was 20.8±3.7 days. All cases except three (ASIA B) showed improvement of ASIA grades with a mean improvement of 1.1±0.5 grades at 2-year follow-up. JOA scores signicantly improved from 6.3±3.1 points at admission to 11.4±3.9 points at 2-year follow up. No difference of neurological improvement was found between different procedures groups. No correlation was showed between neurological improvement and age, concomitant pathology, traumatic mechanism, interval to operation, surgical procedures, or preoperative neurological function. Conclusions:


Background
Since rst described by Schneider et al 1,2 in 1954, acute traumatic central cord syndrome (ATCCS), characterized by more impairment of the upper extremities, bladder dysfunction and varying degrees of sensory loss below the traumatic level, has become the most common incomplete spinal cord injury. With the advancement of anesthetic and surgical technique, the prevailing therapy has transformed from conservative treatment to surgical management [2][3][4] . For ATCCS secondary to spinal column fracture or instability, early decompression and stabilization has been universally approved for satisfactory neurologic recovery, decreased possibility of secondary injury, low rate of complications, and shorter length of hospitalization 5,6 . However, for patients suffering CCS without fracture or instability, up to now, controversy still exists on the optimal surgical timing 7 . Although delayed surgery performed on clinical plateau stage has been proposed, the rationality and necessity are still controversial for mixed results of previous studies 6, 8,9 . So, it is of signi cance to explore the prognosis of patients with ATCCS without fracture or instability treated by delayed operation and to investigate factors related to neurological recovery.
The purpose of this article is to investigate the outcome of delayed surgery in treatment of ATCCS without fracture or instability and to investigate potential factors associated with neurological improvement thus to provide a reference in determining appropriate surgical timing for ATCCS.

Patient Population
We obtained the ethic approval of this study from our hospital's Ethics Board. Patients who underwent delayed surgery on time greater than 2 weeks after trauma for ATCCS without fracture or instability were retrospectively reviewed. All patients were followed up for more than 2 years.

Neurological assessment
All cases except three showed improvement of ASIA grades at the 2 year follow-up visit (P < 0.05). Of the 14 patients classi ed into grade B at admission, 8 improved to grade C, 3 to grade D, and 3 had no change at the 2 year follow-up point. Among the 16 cases with grade C at admission, 12 improved to grade D and 4 to grade E. In the 9 patients evaluated as grade D at admission, all recovered to grade E at last follow up. The mean improvement of ASIA grades was 1.1±0.5 (range 0-2) grades and there was no patient with deterioration of ASIA grades ( Table 2). Improvement of JOA scores were found in all patients and the mean JOA scores signi cantly improved from 6.3±3.1 (range 2-12) points at admission to 11.4±3.9 (range 4-17) points at the 2 year follow up visit with an average improvement of 5.1±1.9 (range 1-10) (P < 0.05) ( Table 3). No difference of ASIA grades and JOA scores was found between different procedure groups (Table 3). Spearman ' s correlation test showed no signi cant correlations of improvement in ASIA grades and JOA scores with age, traumatic mechanism, concomitant pathology, interval to operation, surgical procedures, or preoperative neurological evaluation scale (Table 4).

Discussion
As a clinically prevailing spinal cord injury, the most common causes of CCS are falls, motor-vehicle accidents, and diving injuries 2,10 . Base on etiologies and relevant demographic factors, the whole population with CCS can be divided into three subgroups 9 . First are younger patients, less than 50 years, usually with high-energy traumatic spinal column injuries and subsequent spinal fracture or instability.
The second also commonly consists of younger population with an acute central disc herniation. The last is the "classic" central cord injury in elderly patients greater than 50 years, of whom CCS usually occurs after a hyperextension injury and cord compression on the preexisting spondylosis or spinal canal stenosis 2 . In our study, the most popular etiology of CCS was falls followed by motor-vehicle accidents, which was consistent with ndings of previous reports.
Prior studies reported that hyperextension injury was the most common mechanism of CCS 11,12 . Under neck hyperextension situation, the ligamentum avum buckles inward against the posterior aspect of the spinal cord; meanwhile a bulging disc compresses the cord anteriorly 10,13 . In addition, pre-existing pathologies such as CDH, SCS or OPLL have been reported to contribute to the occurrence of CCS, and patients, especially the elderly population, with such lesions might be more inclined to suffer CCS even after a minor injury 2,9 . In the present study, all patients underwent a hyperextension trauma before CCS and most cases had pre-existing CDH or SCS, noting that extra caution should be paid to patients with CDH and SCS after a hyperextension neck injury.
Conservative treatment was previously favored for CCS mainly for the risk of damage to the already injured spinal cord and previously poor prognosis of surgery 7,10,14,15 . However, with more understanding of spinal cord function division and pathophysiological mechanism of CCS, the dominant therapy for CCS has changed gradually 3,4,6 . In clinical practice now, surgery is usually recommended for CCS with spinal fractures or instability if no signi cant contraindication exist 7,16 . But for CCS without fracture or instability, although controversy does not disappear completely, a decompression surgery is also widely proposed 4,12,17 . In our study, while the patients had no fracture or instability, an anterior or posterior decompression operation was still performed to remove the compression and/or widen the spinal canal volume whose bene t for neurological recovery has been proved by various studies 5,7 .
Controversy on the surgical timing for CCS still exists 6 18 conducted a systemic review to explore whether urgent surgical decompression was the optimal treatment for enhancing neurologic recovery in patient with acute CCS without fracture or instability and concluded that early surgical decompression should be considered in patients with profound neurologic de cit (ASIA = C) and persistent spinal cord compression whereas those with less severe de cit (ASIA = D) could be treated with observation followed by surgery at a later date. Besides, there have been various publications reporting no signi cant difference of prognosis between early and delayed surgery for ATCCS 4,5,11,17,20 . In our clinical experience, we prefer to choose a delayed surgery as our priority unless early operation is obligatory or de nite bene ts of early surgery exist. Most patients in our study had neurological improvement after a delayed decompression procedure. Although no change of ASIA grades was found in 3 cases at 2 year follow-up visit, it does not mean the ineffectiveness of delayed surgical procedure considering the improvement of self-reported symptoms and JOA scores. The nonimprovement in ASIA grades might be attributed to the less quantitative evaluation of ASIA grades which might be unable to distinguish a slight neurological change.
The reported time de nition of delayed operation for CCS varied greatly from hours to weeks 17,20,21 . Surgeons favoring delayed surgery usually conduct an operation for CCS at one week after trauma.
However, in our study, the mean interval from injury to surgery was 20.8 (range, 15-29) days, much bigger than that most surgeons adopted in clinics, and there was enough time for spontaneous recovery of general health state and spinal cord function, whose effectiveness in decreasing iatrogenic complications and promoting neurological improvement has been reported previously 6,8,19 . Ventilator dependence, a disturbing morbidity, can occur after surgery to CCS involving or proximal to C3-C5 levels. Earlier operation performed before the arrival of an abundant spontaneous neurological recovery and unexpected surgical stimulus to pre-traumatic spinal cord may contribute to this catastrophic complication 1 . In our study, postoperative was found in 3 cases with ASIA B grade at admission which suggested that extra caution should be paid to prevent ventilator dependence for patients with poor neurological function.
Previous studies suggested that different surgical procedures might provide different neurological improvement for CCS. However, in our study, no difference was found among different surgery groups.
Besides, Spearman ' s correlation test showed there was no signi cant correlation between neurological improvement and age, traumatic mechanism, concomitant pathology, interval to operation, surgical procedures, or preoperative neurological status. The heterogenicity of samples in different studies might contribute to the difference of outcomes.
Our approval of delayed surgery does not mean we resist other surgical timing; in contrary, we applaud any researches on optimal operative timing for CCS. Clinically, we also conduct surgery for patients with CCS on early stage if necessary. But for patients whose general conditions are unstable or cases that are admitted or transformed to medical centers at time exceeding early surgical timing, a delayed operation should be taken into consideration.
Although we tried to objectively explore the outcomes of delayed surgery for CCS without fracture of instability, the inherent limitations of retrospective study in e cacy evaluation should be noted. Besides, the small sample size may compromise the reliability of outcomes in our study. Randomized controlled studies involving large numbers and multi-centers are warranted to further investigate the optimal surgical time for CCS.

Conclusions
Our study preliminarily suggested that delayed surgery was a feasible and effective therapy for ATCCS without fracture or instability although long-term effectiveness and more details still need to be investigated.       Table 4. Correlation analysis of improvement of ASIA grades and JOA scores with some factors.

Abbreviations
Spearman ' s correlation test, p 0.05 means statistically different.