The management of potential spine injuries is an important component of pre-hospital care. Sections on spinal assessment and treatment figure prominently in international trauma guidelines (American College of Surgeons Committee on Trauma, 2008; Campbell, 2011; Pons and McSwain, 2016), and it has been estimated that 5 million patients are transported with spinal immobilisation in the United States each year (Orledge and Pepe, 1998). Similar figures are not available for other countries, though rough estimates might be proportional to reported rates of traumatic spinal cord injury. This annual incidence varies widely by country, and is reported to be approximately 7.3 per million in the UK; 40 per million in the United States; 3.6 per million in Canada; and 15 per million in Australia (Jazayeri et al, 2015). A number of position papers and treatment guidelines currently advocate for a shift in treatment practices from uniform immobilisation to more flexible spinal motion restriction or spinal precautions (Connor et al, 2013; Morrissey et al, 2014; White IV et al, 2014). In this context of evolving standards, some of the basic assumptions about the harms and benefits of pre-hospital care are being reconsidered.
The sole rationale for spinal precautions of any kind in the pre-hospital setting is to prevent an unstable column injury from causing or worsening a neurologic deficit arising from damage to the spinal cord (Theodore et al, 2013b). This concept is intuitively clear enough to be labelled ‘good sense’ (Theodore et al, 2013b) and ‘logical’ (Connor et al, 2013), yet recent commentary notes that the practice of taking spinal precautions has been widely adopted without clear answers to basic questions (White et al, 2014). For example, what is the rate of neurologic deterioration among patients with spinal injuries? How many would suffer deterioration if not immobilised or restricted in some way? These questions are difficult to answer partly because the term used to describe deterioration – secondary spinal injury – is often ambiguous. Resolving this ambiguity will have a direct impact on future research and pre-hospital care.
In general, secondary injury refers to ‘the destructive and self-propagating biological changes in cells and tissues that lead to their dysfunction or death over hours to weeks after the initial insult’ (Borgens and Liu-Snyder, 2012). Specific uses of the term in areas such as secondary traumatic brain injury reflect this definition (Goldberg et al, 2015). Secondary spinal injury, however, uniquely encompasses two meanings: some literature uses the term to describe physiologic processes after an initial insult (Tator and Fehlings, 1991; Wilson et al, 2013; Anwar et al, 2016), whereas other research uses it in the sense of additional traumatic injury (Kwan et al, 2009; Connor et al, 2013; Theodore et al, 2013b). Some authors explicitly acknowledge both (Todd et al, 2015), but it is rarely written about in pre-hospital or emergency department literature (Nickson, 2012). More often, studies that investigate the course of patients with spinal injuries use the term secondary injury inconsistently, sometimes conflating both meanings and obscuring an already cloudy picture. This lack of distinction, however, has not always been the case: a survey of frequently cited papers reveals both the evolution of the use of secondary spine injury over time and the importance of re-establishing clear definitions.
Discussion
Early studies on the management of spinal trauma paid close attention to how they defined secondary injury. In a case series from 1966, Geisler et al aimed specifically to determine the frequency of additional traumatic injury within a group of patients with spinal injuries. They found that 3% of their patients suffered motor or sensory deficits after the original injury and, in their judgment, specifically because of additional movement. In 1975, Botterell et al advanced two distinct terms to differentiate between types of secondary injury: ‘progressive’ and ‘sequential’ (p. 373). In their scheme, ‘progressive’ refers to deterioration arising from the clinical course of the original injury, and ‘sequential’ signifies additional traumatic insult.
This attention to different types of secondary injury was not widely followed, and specific terminology such as Botterell's has not been adopted. When estimating rates of neurologic deterioration, subsequent studies focused on additional traumatic or ‘sequential’ causes. In 1983, Podolsky et al asserted that 25% of neurologic deficits result from ‘improper handling during the transport and evaluation stage’ (p. 461). This rate of additional traumatic injury remains widely cited (Conrad et al, 2013; Dixon et al, 2014; Hong et al, 2014). In 1988, Toscano described a case series of patients with spinal injuries admitted to a specialty treatment center, and calculated the rate of neurologic deterioration due to improper handling in this setting to be 26%. This figure also appears frequently as the upper limit of the rate of secondary injury (Mazolewski and Manix, 1994; Flabouris, 2001; Consortium for Spinal Cord Medicine, 2008; Shafer and Naunheim, 2009; Schouten et al, 2012; Benjamin and Lessman, 2013; Engsberg et al, 2013; Theodore et al, 2013a; Theodore et al, 2013b; Foster et al, 2014; Sundstrom et al, 2014).
The repeated citations of similar figures as an outer range illustrate both the influence of these studies and how their assumptions have been carried forward. These assumptions deserve scrutiny. The paper by Podolsky et al (1983) cites Cloward (1980):
A survey of a large series of patients with fatal injuries treated at the Edinburgh Royal Infirmary shows that 25% of fatal complications were related to the period between the accident and arrival of the victim in the emergency room (p. 15).
This statement, without a reference, does not support the oft-cited conclusion that up to 25% of spine injuries deteriorate from improper handling. In contrast, the Toscano (1988) paper provides more detail. However, among other methodological choices that affect our interpretation of his results, Toscano (1988) includes two patients who met the contemporary immobilisation standard in the group that he judges to have deteriorated from improper handling. Although he acknowledges the difference between deterioration due to additional trauma as opposed to physiological factors, he nevertheless ascribes all deterioration to movement, even in these cases where there is no evidence of it.
The practice of lumping together of all cases of deterioration under the heading of additional traumatic injury appears to have become conventional. In 2006, for example, Sundheim and Cruz performed a back-of-the-envelope calculation to try to derive a number-need-to-treat (NNT) for spinal immobilisation. They referenced rates of secondary injury provided by a number of studies, including Geisler et al (1966) and Toscano (1988) as their lower and upper ranges, respectively. In attempting to quantify the benefit of immobilisation, the authors followed the practice of assuming that all neurologic deterioration resulted from movement.
Although there appears to be a general tendency to ascribe cases of neurologic deterioration to additional traumatic events, studies that calculate a rate of secondary injury can be parsed in the manner of the Toscano (1988) report. A recent structured review aims to identify and examine cases in which patients suffered neurologic damage from movement (Oto et al, 2015). This review finds that few cases show a clear association between motion and deterioration – and none of these occurred in the pre-hospital setting. Although the patients in this series vary in many respects, the clinical course of their neurologic deterioration does not support the presumption that deficits follow movement.
It is clear that direct links between events and deterioration are virtually impossible to establish after the fact. However, simply assuming that all deterioration follows movement or mishandling – as has often been the case – undoubtedly overstates the real incidence of additional traumatic spine injury. By contrast, restricting that category only to cases that have a confirmed traumatic event potentially understates it. It is likely not possible to derive an accurate rate from the available literature. More important than any single figure, however, is the recognition that motor and sensory deficits as a result of spinal injuries can change for both mechanical and physiological reasons. Current pre-hospital literature tends to emphasise the former. How does this mindset affect options for future research and treatment?
The evolving standard of care in pre-hospital spinal immobilisation is opening up new study possibilities that were unimaginable a few years ago. As services consider moving away from applying a uniform treatment to every patient who cannot be cleared in the field, it will become probable, at the very least, to compare outcomes among patients in different jurisdictions treated under different protocols. One goal of future research will undoubtedly be to investigate which treatment provides the best possible patient care while minimising the rate of neurologic deterioration after the original event. In answering this question, any ambiguity around the cause of observed deterioration will be an important, if not catastrophic, confounder. Although certainty around causes might not always be possible, our language must be precise enough to reflect the fact that different types of injury exist. Whether this comes in the form of explicit terminology such as ‘progressive’ and ‘sequential’, or some other way of signaling the distinction between mechanical and physiological complications, future research must find a way to recognise and address different types of harm.