To investigate the prognostic values of clinical factors 72 h within traumatic cervical spinal cord injury (TCSCI). Data were extracted from the medical materials of 57 TCSCI cases. AIS was used as the outcome measure and divided into dichotomous variables by two methods, i.e. “complete(AIS = A)/incomplete(AIS ≠ A) SCI” and “motor complete(AIS = A or B)/incomplete(AIS ≠ A and B) SCI”. Relationships between evaluated factors and outcomes were investigated by univariate and multivariate methods. MRI Cord transection (MCT) cases, most significantly related to complete SCIs by univariate analysis (P = 0.006), all showed complete SCIs when discharged, which makes it unsuitable for logistic regression. With MCT cases removed, univariate analysis was conducted again, then logistic regression. At last, only C5 spine injury (P = 0.024, OR = 0.241) was related to complete SCI. Cases with compression flexion injury mechanism (CFIM), most significantly related to motor complete SCIs by univariate analysis (P = 0.001), was also unsuitable for logistic regression for the same reason. At last, C3 spine injury (P = 0.033, OR = 0.068) and high energy injury (P = 0.033, OR = 14.763) were related to motor complete SCIs with CFIM cases removed. The results show that MCT and C5 spine injury are good predictors for complete/incomplete SCIs. CFIM, C3 spine injury and high energy injury are good predictors for motor complete/incomplete SCIs.

No abstract