The Immo Traffic Light System as a Decision-Making Tool for Prehospital Spinal Immobilization

Abstract

Background

Spinal injuries are difficult injuries to assess yet can be associated with significant neurological damage. To avoid secondary damage, immobilization is considered state of the art trauma care. The indication for spinal immobilization must be assessed, however, for potential complications as well as its advantages and disadvantages.

Methods

This systematic review addressing the question of the correct indication for spinal immobilization in trauma patients was compiled on the basis of our previously published analysis of possible predictors from the Trauma Registry of the German Society for Trauma Surgery. A Delphi procedure was then used to develop suggestions for action regarding immobilization based on the results of this review.

Results

The search of the literature yielded 576 publications. The 24 publications included in the qualitative analysis report of 2 228 076 patients. A decision tool for spinal immobilization in prehospital trauma care was developed (Immo traffic light system) based on the results of the Delphi procedure. According to this system, severely injured patients with blunt trauma, severe traumatic brain injury, peripheral neurological symptoms, or spinal pain requiring treatment should be immobilized. Patients with a statistically increased risk of spinal injury as a result of the four cardinal features (fall >3m, severe trunk injury, supraclavicular injury, seniority [age >65 years]) should only have their spinal motion restricted after weighing up the pros and cons. Isolated penetrating trunk injuries should not be immobilized.

Conclusion

High-quality studies demonstrating the benefit of prehospital spinal immobilization are still lacking. Decision tools such as the Immo traffic light system can help weigh up the pros and cons of immobilization.

About 1–2 % of all trauma patients suffer injury to the spine, while severe neurological damage occurs in about one fifth of these (1, 2). Spinal injuries are both overestimated and underestimated (3– 6). Spinal immobilization is intended to prevent secondary neurological damage, but it can also have a detrimental effect or create a false sense of security.

The use of prehospital spinal immobilization currently relies on decision-making tools that were originally intended to provide an indication for imaging to confirm the diagnosis, such as the NEXUS criteria or the Canadian C-Spine Rule (7, 8). Randomized controlled trials have failed to demonstrate any benefit of prehospital immobilization (9– 13). This leaves room for individual interpretations, especially with regard to patients in whom the indication for immobilization is less obvious than in the case of the severely injured patient (9– 13).

Furthermore, various disadvantages of immobilization have been reported, for example, increased intracranial pressure from cervical collars (14), positional pain (15– 17), prolonged prehospital times (18), difficult intubation conditions (19), or pressure ulcers (20).

From a practical point of view, immobilization would appear an appropriate measure if there were a risk of aggravation of the injury by movement, for example in the case of unstable fractures.

Aims and objectives

The aim of the present article is to use available studies to develop a practical decision-making tool for spinal immobilization in prehospital trauma care.

Methods

As part of preliminary work by the authors, an analysis was conducted using the Trauma Register of the German Society for Trauma Surgery (TR-DGU) to identify predictors of spinal injury (3, 21) (box).

BOX. Results of the trauma register analysis (3, 21).

• Predictors of a clinically relevant injury to the spine:

  • peripheral motor/neurological deficit

  • fall >3 m height

  • traumatic brain injury

  • seniority (age >65 years)

  • severely injured patient

These were verified in the present review using literature that complies with the PRISMA statement for systematic reviews and the PICO scheme (PICO, population, interventions, comparison, outcome). The review is registered in the PROSPERO systematic review register (ID: CRD42021232806). Details can be found in the eMethods section.

Results

Systematic study selection

The literature search initially yielded 576 publications, of which 24 were included in the study (figure 1).

Figure 1.

PRISMA flow diagram of literature search and study selection

Study characteristics

The included reports involved 2 228 076 patients; the patient number is uncertain for four of the publications. The results are evaluated purely descriptively due to heterogeneous data and endpoints and due to study quality. The literature evaluation revealed low-to-medium quality studies, with a mean quality index score of 14.0 for randomized controlled trials (RCTs) and 11.7 for observational studies (eTable 1 and eTable 2).

eTable 1. Results of the literature review using the Downs & Black checklist 1998.

	Connell et al. 2003 (25)	Cooper et al. 1995 (34)	McCoy et al. 2017 (e5)	Domeier et al. 1997 (29)	Flabouris 2001 (30)	Häske et al. 2020 (3)	Hauswald et al. 1998 (39)	Haut et al. 2010 (26)	Oosterwold et al. 2017 (32)	Oteir et al. 2017 (35)	Reich et al. 2016 (e1)	Tian et al. 2009 (31)	Underbrink et al. 2018 (e12)	Vanderlan et al. 2009 (27)
Reporting	9	6	8	8	8	10	8	10	9	9	10	8	8	10
External validity	1	1	2	2	2	3	2	1	2	2	2	2	1	2
Internal validity – confounding	1	2	5	2	3	3	4	3	4	5	3	4	4	1
Internal validity – bias	2	1	2	2	2	2	1	4	3	3	2	3	3	4
Power	0	0	0	0	0	0	0	3	0	0	0	0	0	0
Total score (mean quality index)	13	10	17	14	15	18	15	21	18	19	17	17	16	17

0–10 points are achievable in the subscale “Reporting”, in subscale “External validity” 0–3 points, in subscale “Internal validity – confounding” 0–7 points, in subscale “Internal validity – bias” 0–6 points, and in subscale “Power” 0–1 point.

The mean quality index score for randomized controlled trials (RCTs) should be at least 14.0 points and for non-RCTs 11.7.

In all, 0–27 points are possible for the mean quality index score.

eTable 2. Assessment of the systematic reviews using AMSTAR 2.

	Connor et al. 2013 (17)	Hawkridge et al. 2020 (e38)	Hood & Considine 2015 (10)	McDonald et al. 2016 (40)	Oteir et al. 2015 (18)	Sundstrøm et al. 2014 (2)	Theodore et al. 2013 (e39)	Velopulos et al. 2018 (e24)	Walters et. al. 2013 (e23)
Quality index	low	moderate	moderate	moderate	moderate	low	low	moderate	low
1. PICO	No	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
2. Statement	No	Partial Yes	Partial Yes	Partial Yes	Partial Yes	No	No	Partial Yes	No
3. Study design	Yes	Yes	Yes	N o	Yes	Yes	Yes	Yes	Yes
4. Literature search	N o	N o	N o	Yes	N o	N o	N o	Yes	N o
5. Publications: study selection	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	no
6. Publications: data extraction	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
7. Excluded studies	No	Yes	Yes	Yes	Yes	Yes	N o	Yes	N o
8. Included studies	Yes	Yes	Partial Yes	Yes	Yes	Yes	Yes	Yes	Partial Yes
9.1. Techniques for assessing the risk of bias in RCTs	Partial Yes	Partial Yes	Partial Yes	Partial Yes	Partial Yes	Partial Yes	Partial Yes	Partial Yes	no
9.2. Techniques for assessing the risk of bias in RSI	Yes	Yes	Yes	Yes	Yes	N o	N o	Yes	Partial Yes
10. Funding	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
11.1. Statistical combination of results RCT	N o	N o	N o	N o	N o	N o	N o	N o	N o
11.2. Statistical combination of results NRSI	N o	N o	N o	N o	N o	N o	N o	Yes	Yes
12. Impact of risk of bias	Yes	Yes	Yes	Yes	N o	N o	Yes	Yes	Yes
13. Discussed risk of bias	Yes	Yes	Yes	Yes	Yes	no	no	Yes	no
14. Explanation	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
15. Quantitative synthesis	N o	N o	N o	N o	N o	N o	N o	Yes	N o
16. Conflict of interest	Yes	Yes	N o	Yes	Yes	Yes	Yes	Yes	Yes

The criteria may be answered with Yes, Partial Yes or No if the items are fulfilled, fulfilled at times or not fulfilled, respectively. The total score gives the Quality Index.

NRSI, non-randomized studies of interventions; PICO, population, intervention, comparison, outcome; RCT, randomized controlled trial

Results of the literature review

The results of the review are presented in line with the key points highlighted in the previous analysis of the TR-DGU, and considerations for prehospital implementation are then added. Data used to calculate predictive values were reported in only 7 of the 24 publications.

Isolated penetrating trauma

The analysis of the TR-DGU shows a high prevalence for spinal injuries AIS3+ associated with blunt trauma (odds ratio [OR] = 4.066 for the whole spine and OR = 3.640 for the cervical spine [CSp]), but not after penetrating trauma (3, 21). An unadjusted retrospective analysis of more than 30 000 patients reported that only 0.034% (n = 12) of conscious patients with penetrating trauma also had spinal injuries associated with significant neurological symptoms (25). An analysis of 45 284 patients with penetrating trauma showed an unadjusted mortality in the immobilized group of 14.7% versus 7.2% (p <0.001) in the non-immobilized group (positive predictive value [ppV]: 14.7%, 95% confidence interval: [13.1; 16.3]; negative predictive value [npV]: 92.8% [92.6; 93.0]) (26). Another study demonstrated that more immobilized than non-immobilized patients with penetrating cervical trauma died (OR = 2.77 [1.18; 6.46], p = 0.02) (27).

Obviously severely injured patients

The probability of spinal trauma increases with injury severity and the number of body regions involved (28). Analysis of the TR-DGU reveals systolic blood pressure of 90 mmHg or lower as a significant predictor of a spinal injury (21). This is interpreted as a surrogate indicator of severe injury.

(Peripheral) motor/neurological deficits

Peripheral motor/neurological deficits are the strongest predictors of spinal trauma in the TR-DGU analysis (OR = 3.171 for the whole spine and OR = 7.462 for the cervical spine, p <0.001) and must therefore be considered a warning sign (3). Domeier et al. found neurological symptoms in 14.5 % of patients with spinal injuries (29), and, according to a retrospective analysis, patients with injury to the spine present neurological symptoms more often (42% versus 17%, p = 0.035) (30).

Traumatic brain injury

According to the TR-DGU analysis, spinal injury presents in 11.9% of patients with seemingly isolated traumatic brain injury (TBI) and in as many as 40.7% of those with TBI associated with multiple injuries (3). Schinkel et al. confirmed this for patients with severe TBI (28) and Tian et al. for comatose patients with TBI (ppV: 11.6% [7.9; 16.4]; npV: 96.0% [93.1; 97.9], unadjusted) (31).

Spinal pain

Studies on pain along the lines of the present research question are hard to find. In general, pain should be considered as a warning sign of possible injury (32, 29).

Kinematics

Various studies have indicated that accident kinematics may provide evidence of spinal injury (8, 33). Cooper et al. reported falls as the most common cause of injury to the spine (unadjusted OR, ppV: 6.0% [4.9; 7.3]; npV: 96.7% [95.9; 97.4]) (34). Oteir and colleagues also share this assessment (adjusted ppV: 0.3% [0.2; 0.3]; npV: 99.8% [99.8; 99.9]) (35). A prospective cohort study of 6500 patients found that mechanism of injury did not help prediction of the resulting injuries (36). The TR-DGU analysis showed that only a fall from a height of more than three meters was a risk factor for spinal injury (OR = 2.243; p <0.001) (21). It should be noted that less severely injured patients are not included in the TR-DGU (33).

Severe associated injury

The TR-DGU analysis showed that spinal injuries are associated with thoracic and abdominal injuries. The incidence of thoracolumbar fractures in patients with and without severe associated injuries (OR = 1.9, [1.4; 2.6]; p <0.001) also confirms this (34). Schinkel et al. found significantly more thoracic and lumbar spine injuries in connection with corresponding injuries to the chest and abdomen (28).

Supraclavicular injuries

Data analysis of the TR-DGU failed to classify maxillofacial injuries to be predictive of cervical spine injuries (OR = 0.183 for facial injuries and OR = 0.876 for skull injuries), but here it is important to bear in mind the selective patient population of the TR-DGU (3). Other studies show that maxillofacial injuries may double the risk of cervical spine injury (e1, e2). An increased risk of cervical spine injury has been reported for geriatric patients in particular after simply tripping and falling (e3, e4).

Age

Patients older than 65 years are more likely to suffer injury to the cervical spine even after low-energy trauma. Analysis of the TR-DGU confirmed this risk constellation with an OR of 1.344 ([1.236; 1.461], p <0.001) (3). McCoy et al. even showed a 3.27-fold increased risk in patients aged 65 years and over (relative risk [RR] = 3.27 [1.66; 6.45]; ppV: 12.6% [9.1; 16.8]; npV: 95.4% [92.9; 97.3], unadjusted) (e5). Reduced mobility and degenerative changes are considered to be causes (e6, e7).

Suggestion for recommended action

The Immo traffic light system was developed as a suggestion for use by the emergeny medical services for prehospital spinal immobilization and is based on the results of the Delphi procedure (figure 2). The Immo traffic light system divides trauma patients into three categories, with the highest priority given to the outcome of the clinical examination. In addition, a few risk factors are also taken into consideration following analysis of the trauma register (3).

Figure 2.

The Immo traffic light system for responsive adult patients. The ABC approach to the unstable patient has priority over immobilization. With “yellow” patients or patients with more than one of the 4 cardinal feature points, there should be a sound justification why, after appropriate risk assessment, immobilization was not performed but instead only restriction of movement. The subjective parameter “pain” requires assessment and interpretation by a qualified member of the medical staff, as does the parameter “obviously severely injured”.

TBI, traumatic brain injury; GCS, Glasgow Coma Scale; NRS, numeric rating scale

Discussion

The aim of the present article was to develop a feasible decision-making tool for justifying spinal immobilization in prehospital trauma care. This should not be interpreted as being the same as reaching a diagnosis, which usually requires cross-sectional imaging (e8). Schweigkofler et al. demonstrated that spinal injuries in the severely injured patient are assumed to present less frequently than their true incidence (31% versus 34%) (e8). Spinal immobilization was advocated by the emergency services in order to prevent neurological damage. Given the lack of studies demonstrating efficacy of prehospital spinal immobilization and the numerous publications demonstrating potential disadvantages of immobilization, there has been an increasingly restrictive approach when discussing spinal immobilization (11, e9, e10). The following suggested categorization is aimed at facilitating risk assessment:

Category Red: Category red includes patients in whom examination findings indicate the presence of spinal injury and who should therefore receive full spinal immobilization. Various guidelines support this view (37, 11). Apart from a peripheral or central neurological deficit, associated spinal pain as an abnormal finding must be regarded as a red flag. Spinal immobilization would appear indicated at the latest when the injured patient requires pharmacological pain relief. The rescue team must therefore be trained to interpret pain as a warning sign and provide treatment (29, 32, 38).

Category Yellow: Patients are assigned to category yellow when the spinal examination is unremarkable (no pain, no limitation of motion, etc.). However, patients in this category also present risk factors for spinal injury.

According to Hauswald et al., there was less neurologic disability in patients with blunt spinal injuries who were not immobilized than in those who were (OR adjusted = 2.03 [1.03; 3.99]; p = 0.04; ppV: 21.0% [16.8%; 25.8%]; npV: 89.2% [82.2%; 94.1%]) (39). Neurological deterioration in the non-immobilized patient with spinal injuries was not reported (40). Other authors recommend the use of the NEXUS criteria (e11). A cautious immobilization strategy in patients aged 60 years and over did not retrospectively affect the incidence of neurological deficit (6.5%, n = 8 versus 5.3, n = 6; p = 0.69) (e12). Two other literature reviews found no benefit in spinal immobilization; there is also reported evidence that prehospital immobilization does not appear to be beneficial (13, e13).

A clinical study involving over 1000 patients demonstrated that imaging studies, and thus presumably also immobilization of the cervical spine, is only rarely indicated in the conscious trauma patient without supraclavicular complaints (e14). No clear benefit was demonstrated for the use of cervical collars either (2, 10, e15).

On the other hand, it has been reported in various case series that unstable spinal injuries can occur even with initially unremarkable neurological findings and can develop a neurological deficit if left untreated (e16, e17). In a review of expert medicolegal reports involving 59 patients with spinal injury, 27 patients were found to have developed neurological deterioration over time; 23 of these were attributed to absent/inadequate immobilization (e18).

Completely dispensing with spinal immobilization measures despite the presence of risk factors would therefore not appear appropriate even if the examination findings are unremarkable, especially as involuntary movements during transport cannot be ruled out.

The data analysis of the TR-DGU identified four independent risk factors for spinal injury that were included in the Immo traffic light system as four cardinal features:

• Seniority (age >65 years)

• Fall >3 m

• Supraclavicular injuries

• Severe associated thoracoabdominal injury.

All four predictors taken alone already indicate an—additively increasing—risk of spinal injury. The specified age limit of 65 years should not be regarded as an absolute value; but rather, the risk increases with age as degenerative changes also increase.

Although only falls from more than three meters emerged as a predictor of spinal injury in the TR-DGU analysis, discovery of energetically comparable kinematics during the physical examination should lead to increased awareness of spinal injury.

As the severity and number of torso injuries increase, so does the risk for spinal trauma. The NEXUS criteria understand “distracting injuries” as inadequate pain perception with respect to the spine; this has already been critically questioned or refuted (e19, e11). The authors regard torso injury to be a risk factor for spinal injury regardless of pain intensity. The benefits and risks of immobilization should be weighed against each other in the conscious patient with unremarkable examination findings (e13, e20– e22).

Category Green: Category green includes patients with clinically unremarkable investigation findings of the spine in the absence of other risk factors, as well as those who have sustained an isolated penetrating injury to the torso. Several review articles favor prompt surgical intervention rather than (cervical spine) immobilization (17, 18, 27, e13, e23, e24).

Providing immobilization

The ABC approach to the unstable patient has priority over immobilization in the sense of “treat first, what kills first”. This a basic principle in emergency medicine and requires no further explanation. As with all medical measures, potential disadvantages of immobilization should also be taken into account for any differentiated justification.

Suggested action for Category Red: Immobilization

The authors recommend full spinal immobilization for this category. Published results on the use of a vacuum mattress or spine board are contradictory, and the quality of the studies is variable (e15, e25– e27). The aim is complete immobilization of the entire spine. Several studies suggest a possible increase in intracranial pressure associated with the use of a rigid cervical collar in patients with higher grade traumatic brain injury (e28– e31). Therefore, alternative methods of immobilization of the cervical spine, such as head blocks or a manual technique, should be considered, especially in the presence of signs and symptoms of intracranial pressure (pupillary dilation, extension synergisms, extension response to painful stimulus, progressive clouding of consciousness). The decisive factor is ultimately the successful immobilization of the cervical spine, not the procedure itself (e14, e19, e25, e26).

Trauma patients with significantly impaired consciousness or disorientation without adequate torso control should receive spinal immobilization.

Suggested action for Category Yellow: Restrict movement

Patients in this category require restriction of movement of the spinal segment at risk in order to prevent major involuntary spinal movements by maintaining torso control (e31, e32).

Aids to be considered for this purpose are a stretcher combined with a cervical collar (CSp), spine board with head blocks, or a vacuum mattress, if necessary, in combination with head blocks or cervical collar—these are to be used as determined by the affected body region. Manual restriction of cervical spine mobility is also a possibility. A cervical collar does not immobilize completely (e33– e37), but the restriction of movement is merely intended to stabilize against involuntary movement while torso control is maintained, which a cervical collar for the cervical spine undoubtedly provides (e32– e36). The above-mentioned limitations in the use of the cervical collar in patients with severe traumatic brain injury do not apply to this category since these patients are neurologically unremarkable.

Limitations

Apart from the register analysis, the Immo traffic light system is also based on systematic literature searches of studies of no high methodological quality. This carries with it the risk of unknown confounders which must be taken into account in the evaluation. Validation of the Immo traffic light system for everyday care is yet to be carried out.

Supplementary Material

eMethods section

As part of preliminary work, the authors conducted an analysis of the Trauma Register of the German Society for Trauma Surgery (TR-DGU), which was used to identify predictors of spinal injury (3, 21).

The Trauma Register of the German Society for Trauma Surgery

The TR-DGU was founded in 1993. Since then, over 450 000 treatment histories have been documented. Participating departments are primarily located in Germany, but departments from other European and non-European countries are increasingly contributing to the register. The aim of this multicenter database is to gather pseudonymized and standardized documentation of severely injured patients. Data acquisition is conducted prospectively in four consecutive phases:

TR-DGU inclusion criteria are either admission to hospital via the resuscitation room with subsequent need for intensive care or arrival at the hospital with vital signs, but death before admission to the intensive care unit. The basic population is defined as patients with a maximum abbreviated injury scale severity score (MAIS) of three or more and patients with an MAIS of two who either died or were in the ICU (22). Currently, around 80% meet this criterion for the basic population, of which 54% had an injury severity score (ISS) of 16 or more (22).

This analysis produced independent predictors of injury to the spine (box). A Delphi procedure using systematically selected literature was then applied to develop a feasible decision-making tool.

Systematic review

These predictors were verified in the present review using literature that complies with the PRISMA statement for systematic reviews and the PICO (population, interventions, comparison, outcome) scheme. The review is registered in the PROSPERO systematic review registry (ID: CRD42021232806).

Search

A systematic literature search for articles in English or German was performed using the electronic databases PubMed and Web of Science with the following search terms and filters: (spine OR spinal*) AND (immobilization OR stabilization) AND (trauma OR injur*) AND (prehospital OR pre-hospital OR out-of-hospital OR emerg*); filters: clinical study, clinical trial, meta-analysis, observational study, randomized controlled trial, review, systematic review). A 10-year period from February 2011 thru February 2021 was searched. This period was chosen because the introduction of certified training courses changed health care strategies while safety features and equipment in the vehicles were also improved.

The bibliographies of the retrieved publications as well as Google Scholar and the SpringerLink Library were also searched to find additional publications.

Inclusion criteria

Published studies recommending indications for spinal immobilization were included.

Study selection and evaluation

After exclusion of duplicates, all titles and abstracts were independently reviewed by two authors, and a decision was made on whether to obtain full-text access according to the inclusion criteria. Full texts were assessed for their relevance and included where appropriate. There was the option to call in an additional author if there were any discrepancies.

Randomized controlled trials (RCTs) and observational studies were evaluated for their quality using the Downs and Black checklist (23).

Systematic reviews were assessed for randomized and non-randomized trials using AMSTAR 2 (24). The quality index comprised the following criteria: “critically low”, “low”, “moderate” and “high” (eTable 1 and eTable 2).

Data analysis

Different study designs were included to reflect the heterogeneous nature of the data, and their results were described qualitatively. If event frequencies were reported in the included sources, an additional calculation of predictive values with 95% exact confidence intervals was performed for the unadjusted four-field tables using the Clopper and Pearson method.

• Prehospital

• Resuscitation room and subsequent surgery

• Intensive care unit

• Discharge