Traumatic cervical spinal cord injuries are one of the most life restraining injuries with a significant clinical and socioeconomic impact if survived [18]. In our study of severely injured patients, we could show that mortality prediction with the RISC II score for patients with a cervical spinal cord injury AIS 6 is highly overestimated with 81.4% as compared to the observed mortality rate of 63.9%. In contrast, for patients with a maximum AIS of 5 for cervical spinal cord injuries, the prediction (18.5%) is correct with an 18.5% observed mortality rate. There could be several reasons for this discrepancy such as the parameters which are included in the RISC II score, particularly the worst injury, age, motor function and physiological parameters which are represented by the ISS, GCS, CPR and blood pressure in the data of the TraumaRegister DGU®. We assume that these parameters, which all affect the outcome of spinal cord injuries and are all included in RISC II Score, could be a reason for the significant difference between the estimated and the real survival rate in this group of patients with a severe spinal cord injury.
The annual incidence of SCI´s ranges around the world. In the developed countries like the US it varies from 40 to 50 per million population, whereas in Europe the range estimate is reported to be between 13 and 19 per million population [19, 3, 20, 4-6, 21]. For traumatic cervical spine injuries, the incidence was previously reported to be 16,5 per 100,000 hospital admissions in a Norwegian population [22]. Passias et.al. [23] showed that in 2017 the US population had an incident rate of 5.0% for traumatic cervical spine injuries. These previous findings are consistent with our overall incidence rate of 5.6% for all traumatic spine injuries in our data set.
The prevalent cause for spinal injuries in most studies are motor vehicle accidents (MVA) or falls. Passias et.al. [23] reported numbers for MVA and falls with 29.3% and 23.7%, which is relatively consistent with our study for MVA (35.5%) and falls (28.9%). The small variation seems to be due to our small sample size of 612 patients. Jackson et al. [24] showed a decrease for spinal cord injuries in the elderly in a study on the US Model Spinal Cord System, but other authors have shown that with increasing age there is an increasing risk for spinal cord injuries, especially for cervical injuries [25, 26]. Elderly patients often have a narrow spinal canal and a stiff spinal column, sometimes associated with Bechterews Disease, which could lead to cervical spinal injuries even with falls lower than 2metres [27]. In our population, 54% of all patients with cervical spinal cord injuries with an AIS of 6 were 60 years or older, whereas in the overall population only 36.5% were 60 years or older.
Guidelines for trauma care suggest that transferring patients with a cervical injury to a level 1 trauma centre could benefit them. In the study by Varma et al. [28] they could show, that 62% of all spinal cord injury patients were transferred to a level 1 centre. In our study, patients were transferred in 78.1% of all cases. However, as there are different health care systems and small sample sizes, these findings are difficult to compare. In the US, 84% of all citizens have access to level I or II trauma centre within one hour, although over 46 million residents, mostly in rural areas, do not have one-hour access to level 1 or 2 trauma care [29]. In another country, Canada, it is estimated that about 22.5% of all residents do not have access to level I or II trauma care within one hour [30]. Due to the existence of the TraumaNetzwerk DGU®, there is a trauma network with the potential of a level I or II trauma care within a 30 km radius. We assumed that this infrastructure could be a reason for the higher survival rate of patients with cervical spinal cord injuries in Germany.
Varma et al. [28] also observed that injuries of the cervical spine causes death at the accident site more often than other multiple injuries with cardiovascular instability. For patients with cervical spinal cord injuries, cardiovascular dysfunctions are initially the most life-threatening events, especially hypotension and bradycardia with resulting in an on-site CPR. Guly et. al. [31] showed an incidence rate of 19% for hypotension and spinal shock of all patient with cervical spinal cord injury. According to Hagen et.al. [32] these findings result from injuries to the autonomic nervous system following a spinal cord injury. The injuries to the autonomic nervous system may also cause bradycardia which could be another cause for onsite CPR [31, 33]. Even the results are difficult to compare, in our group of patients the systolic blood pressure on scene was 63mmHG for the non-survivors and 99mmHG for the survivors. Regarding onsite CPR, in our group we observed that 78% of all non-survivors had a CPR, in contrast to 38% of all survivors. In contrast to our findings, a study from Lockey et.al. [34] showed that for 909 patients with traumatic cardiac arrest, the long-term survival rate with discharge from hospital was 7.5%. However, in this study they only had 6 (8.8%) cervical spine injuries out of 68 survivors. Lockey considered hypoxia as the main cause of traumatic cardiac arrest besides hypovolemia. In our study, we could only generate onsite oxygen saturation for 200 patients. Here we have to admit that based on the register style and the missing information we cannot draw conclusions about hypoxia in our study. In contrast to the American College of Surgeons and the National Association of EMS Physicians [35] we believe that based on our findings traumatic cardiac arrest from a cervical spinal cord injury should be resuscitated no matter of pulselessness or apneic events.
Varma et.al.[28] mentioned severe preexisting comorbidities are a main predictor for early death after trauma. In the TR-DGU only the ISS is reported, including all trauma related injuries, whereas the register does not report preexisting comorbidities. In case of our study we note that all patients with an AIS 6 injury are automatically set to an ISS of 75 due to the calculation method of the ISS [15]. Trauma patients with an ISS of 75 are often described as the most severe injuries with the lowest possible survival rate among trauma researchers. In several studies[36, 37] researchers excluded this type of patients. But Peng et. al.[38] showed that among severely injured patients with an ISS of 75, 48.6% of all patients survived. We could confirm this in our study, as for patients with a spinal cord injury AIS of 6 saw a survival rate of 37.2%.
In our population the median stay on ICU for AIS 6 survivors was 20 days and in hospital stay was 44 days as mentioned in Table 2. Costa et.al. [39] laid out 2016 ventilator associated pneumonia is highly accompanied by physician staff and nurse work environment. As this is only one of many problems in the treatment of spinal cord injuries [40] with better equipment and highly trained staff the outcome gets better over time. We deem the implementation of spinal cord centers in Germany could also improve the survival and outcome of AIS 6 injuries.
The ISS has become the most cited and used trauma score in the last decades not only for trauma surgeons but also for researchers. But we have to note that the score disregards multiple injuries in the same body region and may underestimates head injuries [41, 42]. Paffrath et al.[43] asked whether the ISS based approach to severely injured patients is sufficient. They pointed out, that this approach on a purely anatomical background includes a major number of patients who are not at major risk to die. Here based on the register information of the TR-DGU Lefering et al. published the Revised Injury Severity Classification, version I and II [44, 17]. In the latest version of this trauma score 15 items, including anatomical and physiological parameter, were used out of the documentary of the TR-DGU to predict the mortality rate of each patient. Even with missing items in the register, the score is possible to calculate. As mentioned in the presentation of the RISC II, there are injuries which are overestimated with increasing risk of death. Here we could show that even the RISC II Score performs correctly for patients with a maximum spinal cord injury of AIS 5, but the group of cervical spinal cord injuries with an AIS 6 is highly overpredicted. Therefor in a revision of the RISC II there should be a reconsideration of the role of these AIS 6 injuries.
Furthermore, the ISS relies on the AIS codebook, which is repeatedly changed and updated. However, the revision in 2005 (update 2008) showed only major revisions for pelvic fractures, extremities and head trauma with almost no changes in spine, neck, abdomen and external injuries [45, 46]. Therefore, there were no changes that effected our study. For the years 2002 – 2008 the AUC performed a new coding with the 2008 codebook for all cases entered the TR-DGU, therefore no changes in count of cSCI are scarcely to be expected. Although there is a new revision in 2015, the AIS codebook of 2008 is still in use for the TR-DGU.
As well there were changes in coding in the revision 2005/08 there were also changes in the characterization of AIS 6 injuries. The wording changed from unsurvivable to not treatable. With this wording these injuries are better described.
We have to emphasize, that there are no injury scores or prediction models which weight all the different AIS 6 injuries of all body regions. Due to the small sample size, these injuries show no influence on the general quality of the prediction models.
As we have demonstrated above, we believe that neither changing the AIS codebook for cervical spinal cord injuries nor changing all mortality prediction models would be a solution, as these prediction models perform well in the majority of cases with exception of AIS 6 injuries. Therefor in a revision of the RISC II there should be a reconsideration of the role of the AIS 6 injuries.
Nonetheless in all clinical cases with an cSCI AIS 6 we speak for the treatment of these patients, because as we could show, there are more to survive than expected.