Spinal cord injury (SCI) is a devastating disease that has a global impact on individuals and society. The number of SCI cases in 2016 was 27 million worldwide, which was predominantly due to falls and road traffic collisions [1]. Alarmingly, the number of new SCI cases in most countries has risen over the last few decades [2]. Patients often experience multiple sequalae to injury, such as loss of sensory and motor function below the site of spinal injury, and in the long term develop complex conditions, including chronic pain [3,4]. As mortality is rare and SCI often occurs at a young age, it can cost up to $5.4 million USD in lifetime care per patient [5]. To date, there are no effective treatments available for SCI and the understanding of the pathophysiology of SCI remains limited. Therefore, animal research models, mainly in the form of rodents have been developed and used [6,7].

There are many rodent models of SCI, ranging from hemisection injury in relation to gunshot and stab wounds, to contusion and compression injuries in close imitation of injuries from road traffic collisions and falls [6,7]. The most clinically relevant SCI model is contusion injury induced by force or spinal cord displacement [8]. Many of these preclinical studies have given rise to promising therapies that have transitioned into the clinical trials stage, but sadly their success has halted there. To date, there have been over 1,100 clinical trials since 1986 aiming to improve outcomes for SCI patients [5]. Several studies have focused on drug therapies in the form of neuroprotection [9], neuroregeneration [10], and cell-based therapy [11]. Although the pre-clinical trials have demonstrated success and reproducibility in various SCI models, all have struggled to reproduce these results in large human trials.

Clinicians and researchers have suggested that the failure of SCI clinical trials is due to a variety of reasons within the clinical trials such as small sample sizes, heterogeneity in SCI severity (i.e. ASIA A vs ASIA D), and variations in spinal level studied (e.g. cervical vs lumbar) [9]. However, one factor that has not been fully addressed is the suitability and simplicity of animal SCI models for the clinical setting. The creation of traumatic SCI animal models has been designed to represent human patients as much as possible [12]. For example, in contusion injury, a transient force is applied onto the spinal cord. Not only is the mechanism of injury similar between rats and humans, but also the majority of the pathophysiological responses such as neurodegeneration, neuroinflammation and cyst formation are also similar [13]. Interestingly, this does not hold true for mice, who have several neuropathological differences after SCI when compared to rats. Mice have limited cyst formation, reduced glial scar formation and diminished blood-spinal cord barrier disruption after SCI [14]. Other sources of error may result from the frequent use of young, female adults that are healthy, unintoxicated and injured at the thoracic, rather than cervical level. This does not represent the average human SCI and may make the results clinically irrelevant.

Many preclinical animal studies focus on SCI in young adult rodents. This neglects to account for the increasing number of older SCI patients, often owing to fall-induced tetraplegic injury [15]. It should not be assumed that young adults respond similarly to older adults after SCI. One rodent study focusing on SCI-induced chronic central pain syndrome found that spontaneous locomotor recovery, completion of behavioural test training and development of neuropathic behaviour differed in the young adult. Recovery in the younger rodent (2 months rat age, equivalent to approximately 15 years human age) progressed faster than middle aged rats (12 months rat age, equivalent to approximately 33 years human age) [16-18]. Interestingly, it was shown that aged rats (15 months age, equivalent to approximately 41 years human age) with SCI had a higher vulnerability to mortality after SCI surgery and cell transplantation interventions [17]. However, in SCI clinical trials, the age of participants often ranges from 18 to 65 years old, and, in some studies, the maximum age reaches 80 years old [5]. This large age range in clinical trials could introduce heterogeneity in the cohort of SCI patients, with peaks of SCI incidence occurring at 16-30 years and again at over 65 years. A potential solution is to limit the age range in the clinical trials, but this would decrease the sample size and lengthen the duration of the study.

Although there are more male SCI patients, with a ratio of approximately 3.6 : 1 (males : females) [15], much of the pre-clinical rodent research focuses on females. This is mostly due to their relative practical benefits, including manual bladder emptying following SCI. Additionally, their smaller size and less aggressive nature facilitates ease of handling and group housing [17]. However, it has recently been shown that recovery of motor function and preservation of grey and white matter after SCI was greater in females than males [19]. It was hypothesised that oestrogen may be a contributing neuroprotective factor, but a study with young adult male and ovariectomized female rats with SCI disproved this [20]. The effects of age and sex in pre-clinical SCI studies have been recently reviewed in detail [21]. Therefore, it is important to consider the inclusion of males in a program of preclinical studies for the development of a therapy.

Approximately 25% of SCI incidents involve patients that have consumed alcohol. Of these, 51% had a persistent drinking problem and/or had driven under the influence of alcohol [22]. Furthermore, one or more illicit drugs, such as marijuana, cocaine, and amphetamines, were found in the system of 44% of SCI patients [22]. This is unsurprising given alcohol and illicit drugs can impair vision, balance, reaction time, and judgment. They additionally alter behavioural responses, leading to aggression and neglect, all of which dramatically increases the risk of SCI [23,24]. Interestingly, 45% of SCI patients after injury onset exhibit some form of alcohol dependency, which is a far larger proportion than the 13% affected in the general population. Notably, this percentage amongst SCI patients decreases 17 months after injury [25].

Acute alcohol intoxication has been shown to exacerbate injury following trauma in an animal spinal contusion model by altering the biochemical response to injury and potentially worsening the secondary injury [26]. Furthermore, SCI patients with chronic pain may become heavily reliant on opioids, which can result in further misuse [27]. Substance abuse may also worsen pain and pressure ulcers, thereby increasing mortality [28]. Alcohol abuse can interfere with rehabilitation, lengthen hospital stay, and cause or exacerbate mental health disorders such as depression and anxiety [29]. Alcohol and illicit drugs if not declared or identified by SCI clinical trial investigators may interact with treatment and alter the drug treatment’s effect. For example, alcohol has serious side effects in isolation, yet it can additionally enhance the side effects of other medications such as opioids, leading to respiratory depression [30]. Therefore, drug screening and questionnaires may be required in SCI patient recruitment to clinical trials to ensure the therapy under investigation is not compromised or alternatively, investigate in preclinical studies whether SCI treatment is affected by alcohol and/or illicit drugs.

Even though injury to the cervical spinal cord accounts for approximately 50% of clinical SCI cases, over 80% of pre-clinical rodent SCI studies focus on thoracic level injury [31]. Experimenting with thoracic level injury decreases the risk of accidental rodent mortality and reduces the burden of post-operative care as only the hindlimbs are affected [8]. Furthermore, many behavioural tests, such as the universally used Basso, Beattie, and Bresnahan (BBB) locomotor score, predominantly focus on the hindlimbs. Importantly, there are a few rodent studies that focus on cervical level injury [32,33]. Behavioural testing for the forelimb is performed via the Montoya staircase and/or single pellet reaching test. While the training is difficult, it provides a reliable indication of the dexterity and sensorimotor function of the forelimbs and minimises false positive results, hence it has been recommended for use in the stroke field [34]. However, if researchers are interested in cervical injury, it is important to consider various ethical concerns given the greater impact of higher spinal level injuries [7]. As most clinical trials do not discriminate between spinal levels of injuries in patient recruitment, it is important to consider the differences in motor, autonomic and cardiovascular roles of the cervical and thoracic spinal cord. Notably, the thoracic spinal level has comparatively scarce grey matter and the presence of the sympathetic preganglionic neurones within the intermediate lateral horn of the grey matter [35-37]. Therefore, it is important to consider the inclusion of cervical level injury in preclinical animal studies.

Rats in the laboratory setting are often fed on a diet that provides optimum energy and nutrients for healthy living [38]. It is currently not known whether a healthy diet would reduce injury severity and hasten recovery after SCI compared with a high fat, high carbohydrate Western diet. However, dietary therapies including a diet enriched with omega-3 fatty acids, such as docosahexaenoic acid (DHA), have shown to be neuroprotective after spinal hemisection, contusion, and compression injuries [33,39]. By contrast, treatment with omega-6 fatty acids, such as arachidonic acid, can exacerbate hemisection injury [40,41]. Initiating the ketogenic diet (high fat, low carbohydrate), 4 hours post cervical unilateral contusion demonstrated significant improvement in forelimb function and reduced lesion size, suggesting that SCI patients should limit high carbohydrate content [42]. Therefore, more preclinical studies are required to understand how certain diets of high risk SCI patients may influence injury severity and recovery.

Often SCI patients suffer from trauma to other regions due to road traffic collisions and falls. Recent studies have shown that 40-47% of SCI patients additionally have a clinical concomitant traumatic brain injury (TBI) [43,44]. Unsurprisingly, current SCI clinical trials omit patients that have a concurrent TBI since this co-presentation generally requires a longer hospital stay and adversely impacts functional improvement measures [45]. However, if SCI patients with TBI remain excluded from clinical trials, then treatment options for this group will remain sparse. The clinical attitude to this is mirrored in the paucity of preclinical studies, which show limited data. One study has addressed this gap by administering a unilateral cervical contusion alongside an ipsilateral or contralateral unilateral TBI [46]. They demonstrated a complex recovery dependent on the laterality of the SCI and TBI lesions. Therefore, unless more preclinical studies are conducted to understand the pathophysiology of concomitant SCI with TBI, a large proportion of SCI patients will be excluded from SCI clinical trials.

Given that the role of pre-clinical studies is to prepare for the clinical eventuality, it is important that all controllable variables are mimicked as closely as possible to ensure effective therapeutic translation. If this is not achieved, then any significant progress made in pre-clinical trials will be difficult to replicate successfully in SCI patients. To date, many of the factors that could affect the response to spinal cord trauma such as age, sex, drug intake prior and after injury, and head trauma are not included in the majority of pre-clinical SCI models. Unless the pre-clinical animal models represent the clinical setting, there will be more failures in SCI clinical trials.