Cycling-related cranio-spinal injuries admitted to a Major Trauma Centre in the cycling capital of the UK

Abstract

Background

The increased popularity of cycling is leading to an anticipated increase in cycling-related traffic accidents and a need to better understand the demographics and epidemiology of craniospinal injuries in this vulnerable road user group. This study aims to systematically investigate and characterise cycling-related head and spine injuries seen in the Major Trauma Centre for the Eastern region, which has the highest cycling rates in the UK.

Methods

We performed a retrospective cohort study comparing the incidence, patterns, and severity of head and spine injuries in pedal cyclists presenting to the Major Trauma Centre in Cambridge between January 2012 and December 2020. Comparisons of injury patterns, characteristics, and associations were made according to mechanism of injury, helmet use, patient age and gender.

Results

A total of 851 patients were admitted after being involved in cycling-related collisions over the study period, with 454 (53%) sustaining head or spine injuries. The majority of victims (80%) were male and in mid-adulthood (median age 46 years). Head injuries were more common than spine injuries, with the most common head injuries being intracranial bleeds (29%), followed by skull fractures (12%), and cerebral contusions (10%). The most common spine injuries were cervical segment fractures, particularly C6 (9%), C7 (9%), and C2 (8%). Motorised collisions had a higher prevalence of spine fractures at each segment (p < 0.001) and were associated with a higher proportion of multi-vertebral fractures (p < 0.001). These collisions were also associated with impaired consciousness at the scene and more severe systemic injuries, including a lower Glasgow coma scale (R = −0.23, p < 0.001), higher injury severity score (R = 0.24, p < 0.001), and longer length of stay (R = 0.21, p < 0.001). Helmet use data showed that lack of head protection was associated with more severe injuries and poorer outcomes.

Conclusion

As cycling rates continue to increase, healthcare providers may expect to see an increase in bicycle-related injuries in their practice. The insights gained from this study can inform the treatment of these injuries while highlighting the need for future initiatives aimed at increasing road safety and accident prevention.

Highlights

• Study of 851 cycling-related trauma patients in Cambridge, UK, shows high rates of head & spine injuries.

• Motorised collisions were associated with more severe injuries and impaired consciousness at the scene.

• The lack of helmet use was linked to more severe head injuries and impaired consciousness, but not to a longer hospital stay.

• Rising cycling rates may lead to increased incidence of these injuries in clinical practice.

• Our findings may be relevant for clinicians treating cycling-related traumatic injuries to head and spine.

Keywords:

• Cycling
• trauma
• injury prevention
• road traffic accidents
• road traffic collision
• traffic safety
• TBI
• traumatic brain injury
• skull fracture
• spine fracture
• cerebral haemorrhage

1. Introduction

Cycling is an increasingly popular form of transportation, with the average miles cycled in the UK more than doubling in the last 20 years.¹ In response, the British government recently announced a £2 billion strategy to further double cycled distance by 2025, which includes initiatives such as vouchers for cycle repair, plans for safer cycling infrastructure, and a cycle to work scheme.² This strategy aims to combat obesity, reduce carbon emissions, decrease COVID-19 transmission on public transportation, and reduce traffic volumes.

These initiatives suggest a likely increase in the number of cyclists on British roads and data suggests that an increase in cycling may also lead to an increase in cycling injuries. Between 2004 and 2020, while pedal cycle traffic in the UK grew by 96%, fatalities increased by 5% and serious injuries rose by 26%.³ In addition, a study from the Irish National Spinal Unit showed that spinal injuries involving cyclists increased by 200% from 2010 to 2013.⁴ In the UK, 8% of road traffic collisions (RTCs) in 2019 involved cyclists, and this group had the second highest accident rate per miles travelled, second only to motorcyclists.

These data highlight the vulnerability of cyclists on the road and their susceptibility to head and spine injuries. A study of 11,772 trauma cases in Canada found that 69.9% of cyclists sustained head injuries and 41.1% sustained spine injuries.⁵ A review by Thompson, Rivara, and Thompson (1999) found that head injuries accounted for one-third of emergency department visits, two-thirds of hospital admissions, and three-quarters of deaths among cyclists.⁶ Benefits of helmets in preventing head injuries have been suggested in both adult and paediatric populations.^7,8

As the use of pedal cycles increases, there is a risk of a corresponding rise in traffic-related accidents and head and spine injuries. Therefore, it is important to understand the epidemiology of cycle-related head and spine trauma, the mechanisms of injury, and the resulting patterns of injury, which have implications for patient management, risk mitigation strategies and public policy. Previous studies on cycling injuries have often focused on the effect of helmet use on patient outcomes,^9–12 differences between pedal cycles, motorcycles, and e-cycles,^13–15 or described head^16,17 and spine injuries^4,18,19 separately. This study aims to comprehensively analyse cycling-related accidents with a focus on injuries routinely seen by a neurosurgical unit.

Herein, we describe the epidemiological characteristics of head and spine injuries sustained in bicycle-related collisions among all patients admitted over a 9-year period to a Major Trauma Centre (MTC) in Cambridge, a region with the highest rates of cycling in the UK.¹ In Cambridge, 44.8% of people cycle at least once a week, compared to 35.9% in the second highest region.¹

2. Materials and methods

This is a retrospective study of head (any traumatic brain injury, intracranial bleed, skull fracture) and spine (vertebral fractures and cord injuries) injuries sustained by cyclists admitted to the major trauma service at Cambridge University Hospitals over a nine-year period (January 2012 to December 2020). The study used the Trauma Audit and Research Network database (TARN), which includes information on age, gender, injury severity score (ISS), Glasgow Coma Scale (GCS) at the scene, and mechanism of injury (MOI). Additional details were obtained from electronic medical records, including helmet and alcohol use, length of stay, and specific injuries sustained and their locations. Exclusion criteria were adapted from previous research on cycling-related injuries using the TARN database.²⁰

Patients were divided into two groups based on MOI: (1) collision with motorised vehicles (car, heavy goods vehicle, motorcycle), and (2) self-accident (such as collision with non-motorised vehicles or objects, falls from a cycle). The study compared injury patterns, characteristics, and associations between these groups and according to patient age and gender. The primary outcomes studied were GCS at the scene and associated traumatic brain injury (TBI) severity. Secondary outcomes included ISS and hospital length of stay (LOS).

Statistical analysis was conducted using MATLAB.²¹ Appropriate statistical tests were chosen based on the type and distribution of the data. For associations between categorical variables, Fisher’s exact test was used. For continuous variables, normality was assessed using the Anderson-Darling test. If the data were normally distributed, a two-sample Student’s t-test was used to compare means of the populations. For categorical data the non-parametric Mann-Whitney U test was used to compare medians of the populations. Correlations were analysed using Spearman’s Rank Correlation Coefficient. Results with a p-value less than 0.05 were considered statistically significant.

3. Results

3.1. Trends in incidence of cycling-related injuries

Over the 9 years of study period, there were 851 patients admitted with cycling-related injuries. The annual median number of patients admitted with any cycling-related injuries was 88 (interquartile range: 78.5–96). Of those, 29 (25.25–38.25) patients presented with head injuries, while 19 (12.75–23) patients presented with spinal injuries (Figure 1).

Figure 1. Bar plot showing the yearly incidence of cycling-related injuries over the study period. Dotted lines indicate trends identified using linear regression.

Further analysis revealed a weak upward trend in the incidence of all cycling-related injuries, with a 25% increase observed over the study period (R² = 0.32). The trend in the incidence of head injuries was characterized as very weak, with a 53% increase over the study period (R² = 0.08), while the trend in the incidence of spinal injuries was weak, with an 82% increase observed (R² = 0.19) (Figure 1).

The analysis of the monthly distribution of patients admitted with cycling-related injuries revealed that the highest number of admissions occurred in July (median 10, IQR 9–13), followed by August (median 9, IQR 6.75–11), June (median 8, IQR 6.75–12), and May (median 8, IQR 4.75–9.75). The lowest number of admissions was observed in February (median 3, IQR 2–5.5) (Figure 2).

Figure 2. Boxplot (median, interquartile range, minimum and maximum) showing the monthly distribution of hospital admissions for cycling-related injuries.

3.2. Patient demographics

Of the 851 patients involved in cycling-related collisions, 166 (19.5%) were female and 685 (80.5%) were male, with a male-to-female ratio of 4.13. A total of 454 (53.4%) patients sustained head or spine injuries. Patients with cranio-spinal injuries were predominantly male (370, 81.5%) and in mid-adulthood (43.7 ± 19.4 years) (Figure 3). The most commonly injured age group was 50–60 years old (Figure 3).

Figure 3. Histogram showing the demographic profile of hospitalized patients with cycling-related injuries.

3.3. Mechanism of injury

There were 301 cases (35.4%) due to collisions with motorised vehicles, including 236 men (78.4%) and 65 women (21.6%) with no significant gender difference (p = 0.278). 207 patients (68.8%) sustained cranio-spinal injuries in motorised collisions, compared to 550 non-motorised collisions (64.6%) resulting in 247 craniospinal injuries (44.9%) (p < 0.001). Patients involved in motorised collisions were younger (41.3 ± 18.0 years) than those injured in other accidents (46.0 ± 19.0) (p < 0.001). There were 8 deaths (0.9%), with all patients succumbing to cranial injuries, and all but one being involved in a motorised collision.

3.4. Head injuries

3.4.1. Skull fractures

One hundred and two patients (12.0%) sustained skull fractures, with 128 separate fractures among all patients. The average age in this group was 43.6 years and males represented 78.4% of this group. There were no differences in age (p = 0.439) or gender (p = 0.594) with regards to these injuries. Cyclists involved in collisions with motorised vehicles had a higher prevalence of skull fractures than any other accident type (p < 0.001).

There were 70 fractures of the skull calvarium, 58 fractures of the skull base, and the most commonly fractured bone was the temporal bone (32.8%). Collisions with motorised vehicles more often resulted in calvarial (p = 0.004), base of skull (p < 0.001), and occipital bone (p < 0.001) fractures (Table 1). Fracture distributions are shown in Figure 4. Wearing a helmet was associated with a lower prevalence of all but occipital skull fractures (p = 0.390) (Table 1).

Figure 4. Skull fractures by mechanism of injury and helmet use. (A) Overall patterns; (B) Mechanism of injury; (c) Helmet use.

Table 1. Number of injuries by fractured bone, mechanism of injury and helmet use.

Bone	Motorised	Other	Total	P-value (FE test)	Helmet	No helmet	P-value (FE test)
Calvarial	36	34	70	0.004	7	40	<0.001
Basilar	33	25	58	<0.001	5	30	<0.001
Any	55	47	102	<0.001	12	70	<0.001
Frontal	9	12	21	0.493	2	11	0.023
Parietal	12	10	22	0.070	0	16	<0.001
Temporal	22	22	44	0.051	6	31	<0.001
Occipital	16	6	22	<0.001	4	8	0.390
Sphenoid	8	5	13	0.075	0	10	0.002

Other than in the case of ‘any’ skull fracture, individuals with fractures in more than one region have been included multiple times. P-values from Fisher exact (FE) test.

3.4.2. Intracranial haemorrhages

Two hundred and forty-six patients (28.9%) developed intracranial bleeds, with the average age of this group being 44.7 years and the majority being male (79.3%). There were no significant effects of age (p = 0.657) or gender (p = 0.568) on the number of intracranial bleeds. Collisions with motorised vehicles were significantly associated with a higher number of intracranial bleeds (p = 0.006) compared to non-motorised accidents.

The most common intracranial bleed was the acute subdural hematoma (aSDH; 36.6%), followed by traumatic subarachnoid haemorrhage (tSAH; 34.69%), and extradural hematoma (EDH; 18.4%). Bleeds in cerebral ventricles were the least common (3.1%). Accidents involving motorised vehicles were associated with a higher prevalence of intracranial bleeds overall (p < 0.001), and especially tSAH (p < 0.001), but not with other types of bleeds (Table 2). Helmet use was associated with a lower prevalence of all types of bleeds (p < 0.001) except for intraventricular haemorrhage (p = 1.000).

Table 2. Frequency of head and spine injuries and P-values from t-tests for the relation between age and injuries and from Fisher exact (FE) tests for the relation between gender, accident type, or helmet use and injuries.

Injury	Total (N = 851)	Motorised (N = 351)	Other (N = 550)	Age (t-test)	Gender (FE test)	Accident type (FE test)	Helmet (FE test)
Head injury	329 (38.7%)	144 (41.0%)	18 (3.3%)	0.439	0.859	0.028	<0.001^c
Contusion	85 (10.0%)	42 (12.0%)	43 (7.8%)	0.177	0.886	0.006	<0.001^c
DAI	10 (1.8%)	8 (2.3%)	2 (0.4%)	0.083	0.696	0.005	1.000
HBI	7 (0.8%)	4 (1.1%)	3 (0.6%)	0.577	0.332	0.254	1.000
Pneumocephalus	7 (0.8%)	5 (1.4%)	2 (0.4%)	0.819	0.139	0.104	0.220
Skull fracture	102 (12.0%)	55 (15.7%)	47 (8.5%)	0.439	0.594	<0.001	<0.001^c
Calvarial	70 (8.2%)	36 (10.3%)	34 (6.2%)	0.283	0.528	0.030	<0.001^c
Basilar	58 (6.8%)	33 (9.4%)	25 (4.6%)	0.111	1.000	0.005	<0.001^c
Intracranial bleed	246 (28.9%)	105 (29.9%)	141 (25.6%)	0.657	0.568	0.006	<0.001^c
aSDH	117 (13.8%)	47 (13.4%)	70 (12.7%)	0.030^b	0.209	0.253	<0.001^c
tSAH	111 (13.0%)	58 (16.5%)	53 (9.6%)	0.016^b	0.441	<0.001	<0.001^c
EDH	59 (6.9%)	25 (7.1%)	34 (6.2%)	<0.001^a	1.000	0.260	<0.001^c
ICH	23 (2.7%)	11 (3.1%)	12 (2.2%)	0.235	1.000	0.268	<0.001^c
IVH	10 (1.8%)	5 (1.4%)	5 (0.9%)	0.167	0.029	0.337	1.000
Spinal fracture	176 (20.7%)	103 (29.3%)	83 (15.1%)	0.510	0.210	<0.001	0.025^d

^a Population with injury is younger, ^bpopulation with injury is older. Significant difference with regard to gender means higher prevalence of injury in males in all cases. Significant difference with regard to accident type means higher prevalence of injury in motorised collisions in all cases. ^cLower prevalence of injury in helmeted group, ^dhigher prevalence of injury in helmeted group.

DAI: diffuse axonal injury; HBI: hypoxic brain injury; aSDH: acute subdural haematoma; tSAH: traumatic subarachnoid haematoma; EDH: extradural haematoma; ICH: intracerebral haemorrhage; IVH: intraventricular haemorrhage.

3.4.3. Other brain injuries

One hundred and nine cyclists (12.8%) suffered from brain injuries other than bleeds or fractures, including contusions, diffuse axonal injury (DAI), or hypoxic brain injury (HBI) (henceforth referred to as ‘Other TBI’). As seen in Table 3, the average age of this group was 45.83 years and male patients represented 79.8% of this group. There was no age (p = 0.342) or gender (p = 0.897) predilection with regards to this type of injury.

Table 3. GCS at the scene by the mechanism of injury with respect to different types of injury.

	Motorised	Other	Overall	P-value (MW U)
Skull fracture	15 (13.25–15)	15 (15–15)	15 (14–15)	0.041
Intracranial bleed	12 (6–15)	15 (13.75–15)	14 (9–15)	<0.001
Other TBI	10.5 (6–14)	14 (7–15)	12 (6–15)	0.006
Spinal injury	15 (13.25–15)	15 (15–15)	15 (14–15)	0.041
Overall	15 (12–15)	15 (15–15)	15 (15–15)	<0.001

Reported as median (interquartile range). P-values from Mann-Whitney U (MW U) test.

The most prevalent type of parenchymal brain injury was contusions (78.0%), followed by diffuse axonal injury (DAI; 9.2%) and hypoxic brain injury (HBI; 6.4%), with pneumocephalus being equally prevalent (6.4%). Accidents with motorised vehicles had a higher prevalence of other brain injuries overall (p < 0.001), and specifically contusions (p = 0.006) and DAI (p = 0.005) as seen in Table 2. The helmeted patient group had significantly lower prevalence of contusions (p < 0.001) but not other types of brain injuries (Table 2). The majority of parenchymal injuries were localized to the frontal (39.7%) and temporal (31.0%) lobes.

3.5. Spine injuries

One hundred seventy-six patients (20.7%) had spinal injuries, with the mean age of this group being 45.0 years and it consisting predominantly of men (83.9%). There was no association between age (p = 0.510) or gender (p = 0.210) and spinal fractures. Most spinal fractures occurred in collisions with motorised vehicles (55.4%), which had a significantly higher prevalence of spinal fractures compared to other accidents (p < 0.001). Fifty-one patients with spinal injuries also had cranial injuries.

There were 339 vertebral fractures in this study, with the C6 level being the most frequently fractured (9.4%), followed by C7 (9.1%) and C2 (7.7%) (Figure 5A). In motorised collisions, fractures mostly affected the C5, C7, and L1 vertebrae (8.4% each), while in other accidents, the C6 (12.4%), C7 (10.2%), and C2 (9.5%) were most often fractured (Figure 2A).

Figure 5. Distribution of spinal fractures in hospitalized patients with cycling-related injuries by localisation and mechanism of injury. (A) Fracture localization by vertebral level; (B) Fracture localization by spinal segment; (C) Multivertebral fractures.

Overall, the cervical segment accounted for the majority of fractures (41.9%), followed by the thoracic segment (38.3%), with the lumbar segment being the least commonly fractured (19.8%). Motorised collisions were associated with a higher prevalence of all spinal fractures (p < 0.001) and this was also true for cervical (p = 0.008), thoracic (p < 0.001), and lumbar (p < 0.001) fractures (Figure 5B).

Most patients (51.1%) sustained a single vertebral fracture (Figure 5C). However, those involved in collisions with motorised vehicles had a higher prevalence of multivertebral fractures (53.4%) compared to those involved in other accidents (43.4%) (p < 0.001). No effect of helmet use on spinal fractures was seen in the cervical (p = 0.866), thoracic (p = 0.078), or lumbar (p = 0.124) segments, although it was weakly associated with a higher prevalence of spinal fractures overall (p = 0.025) (Table 2).

Twenty-one patients had spinal cord injuries, with 7 being ASIA A, 3 ASIA B, and 2 ASIA D. 9 patients had an unspecified neurological deficit, which was associated with central cord syndrome in 3 cases and cord contusion in 2 cases. All spinal cord injuries were associated with vertebral fractures. There was no significant difference in the incidence of spinal cord injuries between motorised collisions (11 cases) and non-motorised accidents (10 cases) (p = 0.258).

3.6. Relationships between injuries

During the initial analysis, combinations of intracranial bleeds, skull fractures, spinal fractures, and traumatic brain injuries were observed (Table 4). Significant associations were found between different types of head injuries (skull fractures, intracranial bleeds, and other TBI) (p < 0.001). Spinal fractures were also associated with concurrent intracranial bleeds (p = 0.035) but not with other brain injuries (p = 1.000) or skull fractures (p = 0.125).

Table 4. Coexistence of injuries, expressed as P-values from Fisher exact (FE)-tests.

	Skull fracture	TBI	Spinal injury
Intracranial bleed	<0.001	<0.001	0.035
Skull fracture		<0.001	0.125
Other TBI			1.000

TBI: traumatic brain injury.

3.7. Helmet use

Helmet use data was available for 455 patients (53.5%). Of these, 215 (47.3%) wore helmets and 240 (52.8%) did not. The group wearing helmets was older (median age 50 years, IQR 35–59) than the non-helmeted group (median age 39 years, IQR 23–56.5) (p < 0.001). There was no significant difference in gender with respect to helmet use (p = 0.463).

3.8. Accident velocity and helmet effectiveness

In our study, we divided the accidents with available helmet data into two groups based on self-reported velocity: high-velocity (above 12 mph or 19.3 kph) and low-velocity (below 12 mph or 19.3 kph) based on the European standard for helmets for pedal cyclists (EN1078). The high-velocity group consisted of 158 accidents (34.7%), while the low-velocity group consisted of 297 accidents (65.3%). There were no significant differences in age or gender distribution between the two groups (p = 0.309 and p = 0.368, respectively).

Under low-velocity impact conditions, we found that helmet use was protective of skull fractures of all types, reducing the risk by 76.6%. It also lowered the risk of intracranial bleeds overall (RRR = 0.49) and specific types of intracranial bleeds such as aSDH (RRR = 0.41), tSAH (RRR = 0.60), and EDH (RRR = 0.67), as well as cerebral contusions (RRR = 0.65) and head injury overall (RRR = 0.48) compared to the non-helmeted group (Table 5).

Table 5. Bicycle helmet efficacy with respect to the type of injury and accident velocity (high velocity – >12 mph, low velocity – < = 12 mph).

	High velocity			Low velocity
	RR (CI)	ARR	RRR	RR (CI)	ARR	RRR
Head injury	0.73 (0.49–1.07)	Ns	Ns	0.51 (0.37–0.71)	24.5%	48.2%
Contusion	0.53 (0.24–1.19)	Ns	Ns	0.35 (0.15–0.78)	33.0%	64.9%
DAI	0.61 (0.11–3.40)	Ns	Ns	0.61 (0.11–3.40)	Ns	Ns
HBI	1.23 (0.17–8.84)	Ns	Ns	0.98 (0.24–3.95)	Ns	Ns
Pneumocephalus	0.61 (0.11–3.40)	Ns	Ns	0.39 (0.03–4.70)	Ns	Ns
Skull fracture	0.53 (0.28–1.01)	Ns	Ns	0.23 (0.10–0.53)	38.9%	76.6%
Calvarial	0.51 (0.23–1.14)	Ns	Ns	0.17 (0.04–0.64)	42.1%	82.9%
Basilar	0.17 (0.02–1.16)	Ns	Ns	0.24 (0.06–0.90)	38.3%	75.4%
Sphenoid	0.22 (0.01–3.17)	Ns	Ns	0.17 (0.01–2.51)	Ns	Ns
Frontal	0.82 (0.26–2.59)	Ns	Ns	0.13 (0.00–1.91)	Ns	Ns
Parietal	0.11 (0.00–1.76)	Ns	Ns	0.15 (0.01–2.17)	Ns	Ns
Temporal	0.58 (0.24–1.39)	Ns	Ns	0.19 (0.05–0.73)	40.8%	80.3%
Occipital	0.54 (0.15–1.88)	Ns	Ns	1.31 (0.58–2.94)	Ns	Ns
Intracranial bleed	0.49 (0.28–0.84)	20.5%	50.6%	0.50 (0.35–0.73)	25.0%	49.2%
aSDH	0.40 (0.21–0.75)	30.3%	59.7%	0.59 (0.37–0.94)	20.6%	40.5%
tSAH	0.51 (0.24–1.06)	Ns	Ns	0.40 (0.21–0.75)	30.3%	59.7%
EDH	0.43 (0.15–1.23)	Ns	Ns	0.32 (0.11–0.92)	34.2%	67.2%
ICH	0.22 (0.01–3.17)	Ns	Ns	0.17 (0.02–1.16)	Ns	Ns
IVH	1.23 (0.45–3.34)	Ns	Ns	0.98 (0.24–3.95)	Ns	Ns
Spinal fracture	1.42 (1.13–1.78)	21.1%*	−42.5%*	1.00 (0.74–1.36)	Ns	Ns

RR: risk ratio; CI: confidence interval; ARR: absolute risk reduction; RRR: relative risk reduction; Ns: non-significant result; DAI: diffuse axonal injury; HBI: hypoxic brain injury; aSDH: acute subdural haematoma; tSAH: traumatic subarachnoid haematoma; EDH: extradural haematoma; ICH: intracerebral haemorrhage; IVH: intraventricular haemorrhage.

* Increased risk of injury when wearing helmet.

In the high-velocity accident group, helmet use was associated with a lower risk of intracranial bleeds (RRR = 0.51), specifically aSDH (RRR = 0.60). However, no other significant protective effects on other types of head injuries were observed in this group. Interestingly, helmet use was associated with an apparent increase in the rates of spinal fractures (RRR = 1.42) in the high-velocity group (Table 5).

3.9. Injury severity and patient outcomes

3.9.1. Injury severity score

Overall, the median ISS was 9 (IQR 9–22). Motorised accident group had a higher ISS (17, 9–25.25) than cyclists involved in other accidents (9, 8.25–18) (p < 0.001). Higher ISS was correlated with head injuries overall (r(327) = 0.50, p < 0.0001), and particularly with intracranial bleeds (r(244) = 0.43, p < 0.001) including aSDH (r(115) = 0.30, p < 0.001) and tSAH (r(109) = 0.27, p < 0.001), contusions (r(83) = 0.26, p < 0.001) and skull fractures (r(100) = 0.25, p < 0.001) including calvarial (r(68) = 0.20, p < 0.001) and basilar (r(56) = 0.19, p < 0.001) fractures. Patients wearing helmets had lower ISS (9, 9–20) than those not using head protection (16, 9–25) (p = 0.010).

3.9.2. Glasgow coma scale at the scene

The median GCS for all injuries overall was 15 (15–15). Parenchymal brain injury group had the lowest GCS (12, 6–15) and spinal injury group had the highest (15, 14–15) (Table 3). Cyclists involved in collisions with motorised vehicles, as a group overall, had significantly lower GCS on scene (15, 12–15) than those involved in other accidents (15, 15–15) (p < 0.001, Mann Whitney U test). As seen in Table 3, this trend was consistent across all types of injury. A Pearson correlation coefficient was computed to assess the linear relationship between GCS and different types of craniospinal injuries. There was a moderate negative correlation between GCS and any type of head injury (r(327) = −0.44, r < 0.001) and intracranial bleeds (r(244) = −0.35, p < 0.001) and specifically tSAH (r(109) = −0.36, p < 0.001). There was a weak correlation between GCS and SDH (r(115) = −0.24, p < 0.001), HBI (r(5) = −0.29, p < 0.001), DAI (r(8) = −0.28, p < 0.001), contusions (r(83) = −0.22, p < 0.001) and skull fractures (r(100) = −0.21, p < 0.001) including basilar (r(56) = −0.24, p < 0.001) but not calvarial (r(68) = −0.08, p < 0.001) fractures. Use of bicycle helmets was associated with higher GCS at scene (15, 15–15) compared to no head protection gear (15, 14–15) (p < 0.001).

3.9.3. Hospital length of stay

Worse outcomes, as measured by the severity of TBI, were observed with respect to helmet use, whereby patient group without helmets sustained less mild (79.2% vs. 90.7%) and more moderate (5.8% vs. 1.4%) and severe (13.3% vs. 7.4%) TBI (Figure 6A). Similar trends were observed in accidents involving motorised vehicles, where 17.7% of patients sustained severe TBI compared to 5.1% in the non-motorised accident group. There were also more cases of moderate TBI (8.1% vs. 3.1%), and conversely, less cases with mild TBI (74.8% vs. 91.8%) (Figure 6B).

Figure 6. Severity of traumatic brain injury (TB) in relation to mechanism of injury and helmet use. (A) Mechanism of Injury; (B) Helmet Use. Mild (GCS 13–15), moderate (GCS 9–12), severe (GCS 3–8).

The median LOS across all injury groups was 9 days (interquartile range, 4–22.25). The longest hospital stay was associated with parenchymal brain injuries (16, 9–35), followed by intracranial bleeds (11, 4–23.50) and skull fractures (10.50, 4–22) (Table 6). Involvement in motorised collision (16, 6–27.75) was associated with longer LOS than other accidents (6, 4–14) (p < 0.001). This trend was significant with regards to skull fractures (p = 0.035), intracranial bleeds (p = 0.005), and spinal injury (p = 0.046), but not parenchymal brain injuries (p = 0.780) (Table 6). LOS was correlated with head injury of any type (r(327) = 0.50, p < 0.001), and less strongly with intracranial bleeds (r(244) = 0.43, p < 0.001) including aSDH (r(115) = 0.30, p < 0.001) and tSAH (r(109) = 0.27, p < 0.001), contusions (r(83) = 0.26, p < 0.001) and skull fractures (r(100) = 0.25, p < 0.001). There was little to no effect of the localisation of skull fractures to calvarium (r(68) = 0.08, p < 0.001) or base of skull (r(56) = 0.16, p < 0.001). There was no difference in LOS with respect to helmet use (6,4–15) compared to no head protection (8,4–23) (p = 0.126, Mann Whitney U test). There was no correlation between patient age and outcomes, as measured by GCS (r(849) = 0.04, p = 0.286), LOS (r(849) = −0.02, p = 0.605), or ISS (r(849) = 0.03, p = 0.542).

Table 6. Hospital length of stay, measured in days, by type and mechanism of injury.

	Motorised	Other	Overall	P-value (MW U)
Skull fracture	18 (5.75–24.75)	6 (4–14.5)	10.50 (4–22)	0.035
Intracranial bleed	16 (7.50–28)	6 (4–15)	11 (4–23.50)	0.005
Other TBI	15 (8–33)	17 (9–35)	16.50 (8.50–34)	0.780
Spinal injury	13 (6–27)	7 (5–17)	9 (5–24)	0.046
Overall	16 (6–27.75)	6 (4–14)	9 (4–22.25)	<0.001

Reported as median (interquartile range). P-values from Mann-Whitney U (MW U) test.

3.10. Intoxication status

Forty-two patients were intoxicated at the time of accident (4.9%) and all were male. Median age was 44.5 years (IQR 20–57) and there was no difference in age between intoxicated and sober patient groups (p = 0.454). Thirty-three patients sustained cranial injuries and 9 did not, which was statistically significant (p < 0.001). Alcohol use was associated with poorer helmet use adherence (p = 0.004) but not with the mechanism of injury (p = 0.510).

4. Discussion

The popularity of cycling has increased in recent years and is expected to continue, with many governments promoting the activity. However, as the number of cycling-related traffic accidents is likely to rise, it is important to understand the demographics and epidemiology of cranio-spinal injuries (including skull and vertebral fractures, intracranial bleeds, and brain injury) in this vulnerable group of road users. In this study, we analysed injury patterns among 851 patients admitted to our Major Trauma Centre over a period of 9 years.

4.1. Demographics

In accordance with previous research,^5,13,14,16 our study found that most head and spine injuries occurred in middle-aged adults aged 41–60 years (mean ± standard deviation: 43.7 ± 19.4 years). Previous studies¹⁶ have suggested that older age is associated with a lower prevalence of epidural hematomas and a higher prevalence of subarachnoid and subdural haemorrhages. This may be due to brain atrophy,²² which increases the distance between the arachnoid and dura mater and thus increases the risk of tears and subdural haemorrhages in older individuals. Conversely, the tighter adherence of the dura mater to the skull in the elderly²³ may explain the higher prevalence of epidural hematomas in younger populations. Our study also found that older age was associated with subarachnoid haemorrhages,²⁴ in line with previous reports. However, we did not find a significant correlation between patient age and injury severity or outcomes, which differs from some previous studies^16,25,26 that have found a relationship between older age and poorer outcomes. This discrepancy could be due to the heterogeneous nature of the population in our study and the positive skew of the age strata, as well as the low prevalence of paediatric injuries.

Most of the cyclists in this study were male (80.0%), a trend observed in other studies.^14,16 The reason for this is not clear, but it may be due to higher uptake of cycling among men or increased risk-taking behaviour in men, among other potential factors.²⁷ This matches the wider gender imbalance in head injuries in middle age, from any mechanism. However, there was no significant difference in the gender distribution of injuries among those involved in motorised vehicle accidents compared to other types of accidents. The mortality rate in this study (0.95%) was similar to that reported by Spoerri et al.¹⁴ (1.2%), and the patients who died had similar profiles of injuries and mechanisms of injury as those seen in other studies.¹⁶

4.2. Mechanism of injury

In our study, we analysed 851 cycling-related accidents, of which 351 (41.0%) involved motorised vehicles. This percentage has ranged from 23.1%¹⁴ to 51.2%¹⁶ in other studies. We found that patients involved in motorised collisions were younger (mean age 41.3 years) than those injured in other types of accidents (mean age 45.9 years) (p < 0.001), a finding similar to the age difference observed by Depreitere et al.¹⁶ (34.5 vs 45.9 years). Motorised collisions were associated with a higher prevalence of head injuries overall, and specifically contusions, diffuse axonal injury, skull fractures, and intracranial bleeds including subarachnoid haemorrhages. This was also true for spinal fractures.

These types of collisions often involve high impact due to higher vehicle speeds and momentum, resulting in more severe injuries. Additionally, the biodynamics of falls from cycles differ from those induced by vehicular impact.²⁸ Perhaps due to the high prevalence of serious cranio-spinal injuries, motorised collisions were slightly correlated (Pearson’s r between 0.21 and 0.24) with lower Glasgow Coma Scale scores, longer lengths of stay, and higher Injury Severity Scores. Similar associations between the involvement of motorised vehicles and worse outcomes have been reported previously.²⁹

4.3. Injury patterns and associations

The prevalence of head and spine injuries, and the relationships between the different subtypes of injuries (Table 4) were similar to those observed in previous reports.^{13,14,19,30–32}

4.3.1. Head injuries

There is limited research on the localization of head and spine injuries in cyclists. In our study, we found that temporal skull fractures were twice as common as frontal, parietal, or occipital fractures, which all had similar prevalence. Motorised collisions were associated with a higher prevalence of skull fractures overall, but this association was significant only for occipital fractures. It is worth noting that patients with fractures involving multiple bones were counted multiple times in our analysis. These findings may be explained by more serious injuries and multiple skull areas affected by accidents involving motorised vehicles, or by different mechanisms of impact leading to falls on the back of the head, possibly due to loss of consciousness due to acceleration-deceleration injuries.

Previous research by Depreitere et al.¹⁶ found that head injuries most commonly affected the frontal (55.2%) and temporal (34.4%) regions, which is consistent with our findings (40% frontal, 31% temporal). This pattern may be due to the vector of impact, which is unlikely to cause falls that affect the occipital and parietal lobes directly. Additionally, the thickness of brain parenchyma may absorb some of the force based on its viscoelastic properties,³³ protecting deeper areas of the brain. However, these results may be biased as we did not distinguish between coup and contrecoup brain injuries and information on the direction of impact was not collected. Reconstruction of impact forces from radiological findings is also unreliable.¹⁶

The overall and relative prevalence of intracranial bleeds in our study is consistent with previous research. Studies by Roberts et al.⁵ and Baschera et al.¹³ found relatively low rates of epidural hematomas (8.5% and 6.0%, respectively), similar rates of subdural hematomas (26.0% and 18.0%, respectively), and higher rates of subarachnoid haemorrhages (21.7% and 21.2%, respectively). These figures differ from and are lower than those reported by Depreitere et al.,¹⁶ possibly because their study only analysed patients who underwent neurosurgical procedures, implying more severe injuries in their studied population. In our study, the highest proportion of intracranial bleeds was subarachnoid haemorrhages (13.0%), followed by subdural hematomas (13.8%) and epidural hematomas (6.9%). The rates of DAI in our study were lower than in previous studies, possibly due to the lack of routine MRI scanning for all patients with brain injury at our centre during the study period.

4.3.1. Spine injuries

In our study, the most commonly fractured segment of the spine was the cervical region, followed by the thoracic and lumbar regions. However, this differed significantly depending on the involvement of a motorised vehicle, which carried a higher risk of lumbar and thoracic spine injury. In a study of electric bicycle injuries, Wu et al.¹⁹ found that the most commonly fractured vertebral level was L1, followed by C7 and T12 fractures. Our results showed that while a similar pattern of low cervical and high lumbar fractures was true for pedal cyclists involved in motorised collisions, most cyclists overall sustained both low and high cervical fractures, which is consistent with other studies.^30,34,35 This may be because electric cycle riders may not be representative of pedal cyclists, as suggested by studies comparing electric cycle, pedal cycle, and motorcycle-related injuries.^13,14 There was a high prevalence of multivertebral fractures, especially in the motorised accident group, which may be due to the higher impact force and speeds associated with these accidents. Spinal fractures were associated with intracranial bleeds, but not with other types of injuries, a finding supported by previous research.³⁶

A third of patients with spinal fractures also suffered from head injuries, and more than 10% of patients with spinal fractures also suffered from spinal cord injuries. These figures are consistent with previous research showing that high cervical fractures, which were the most common injuries analysed, were less frequently associated with cord injuries compared to lower segment fractures, which were underrepresented in our cohort.³⁷ Interestingly, spinal injuries were not correlated with worse outcomes and in fact had the shortest length of stay and highest Glasgow Coma Scale scores at the scene of the accident among all injury groups.

4.4. Outcomes

Outcomes after cycling-related injuries,^{6,31,32,38,39} including traumatic brain injuries,^{10,16,17,40,41} have been reported in previous research, suggesting that older age is associated with higher severity of injuries and poorer outcomes, and that accidents involving cyclists colliding with motorised vehicles carry a risk of more severe injuries and worse outcomes. Our study found similar results.

The proportions of patients with mild (85.2%), moderate (4.8%), and severe (9.3%) traumatic brain injuries were in line with other studies,^13,14 although we observed a higher proportion of severe injuries. This may be due to the operating model of the regional trauma network and the inherent selection bias whereby less severely injured patients do not present to this centre and are treated in local hospitals. The mean Injury Severity Score of 11.37 in our study cohort, corresponding to mild injuries overall, was lower than the 17 reported by Roberts et al.,⁵ although their study focused on severe injuries from the outset. Higher Injury Severity Scores were strongly correlated with head injuries (r(327) = 0.50, p < 0.001), and moderately correlated with intracranial haemorrhages (r(244) = 0.43, p < 0.001). The median length of stay for the cohort overall was 9 days, and 6 days for patients without head or spine injuries. Similar lengths of stay for cycling-related injuries of 5–6 days have been reported in previous studies.^5,14 Patients with head injuries had a longer median length of stay (10.5–16.5 days) than those with spinal injuries (9 days). This may be due to the treatment and rehabilitation of intracranial neurological injuries requiring longer support in the neurocritical care unit or other hospital departments. It is worth noting that our study only included patients with a length of stay of at least 3 days and we do not have information on the proportion of the length of stay spent in the intensive care unit or the proportion of patients needing intensive care unit admission. Previous research suggests that approximately a quarter of severely injured cyclists will require intensive care unit admission.⁵

4.5. Use of head protection

The effect of helmets use on patient injuries and outcomes and the benefits of helmets use in preventing head injuries in both adult and paediatric populations have been previously studied.^7–12 In our study, helmet data was available for over half of all patients. Our data suggests that lack of head protection was associated with more severe injuries and more seriously impaired consciousness at the scene, but not with a longer hospital stay.

Helmet wearers were older in our study cohort. Although it is known that children are more likely to wear helmets, this group was underrepresented in our cohort. Conversely, young adults and those in mid-adulthood have been shown to use head protection less frequently.⁴² These patient groups made up the majority of our studied demographic, which may explain our findings. There was no effect of gender on helmet wearing.

We found that the non-helmeted group had a higher proportion of moderate and severe traumatic brain injuries compared to patients with head protection, which corresponded to a higher prevalence of cerebral contusions and almost all types of skull fractures, excluding occipital fractures. This could be explained by helmet design being less protective of the posterior skull and the higher impact energy required for occipital fractures due to higher bone density and thickness of the bone table. This mechanism of injury would correspond to high-velocity collisions,⁴³ which is consistent with our finding that almost three-quarters of occipital fractures were sustained in collisions with motorised vehicles. The lack of statistically significant differences with respect to other types of traumatic brain injury can be explained by low statistical power due to the low overall prevalence of these injuries (at most 10 cases).

Helmets appeared to be more effective under low impact velocity conditions. They were protective of all types of head injury, including skull fractures (calvarial, basilar, and especially temporal), bleeds (aSDH, tSAH, EDH) and contusions. In high-velocity accidents, helmets were associated with reduced risk of intracranial bleeds (including aSDH) only.

Previous studies have suggested that helmets may increase risk of spine fractures due to increased torque at the neck during accidents.⁴⁴ Our study found an increased risk of spinal fractures in helmeted patients involved in high-velocity accidents compared to non-helmeted patients. This may be due to increased risk-taking behaviour among helmet wearers^45,46 or selection bias where helmets protect the head but not the rest of the body during high-velocity accidents with motorised vehicles. Consistent with this explanation, our data shows that motorised collisions were associated with a higher injury severity and increased prevalence of spinal fractures. However, our data does not reflect increased risk of spinal fractures in low-velocity accidents, consistent with recent studies that found no increased risk of spinal injury from helmets.^47,48

4.6. Limitations

There are several limitations to this study. Cambridge has the highest cycling rates in the UK¹ and the cycling infrastructure is well-developed, which may contribute to increased safety for cyclists. Because of how common cyclists are on the roads, a herd protection effect may be at play whereby awareness of cyclist presence and driving patterns increases their visibility for other traffic participants.

Cambridge University Hospital (CUH) is the designated level 1 MTC for the East of Anglia region, serving population of around 6.1 million.⁴⁹ With this large catchment population, the risk of selection bias is reduced. However, patients with less severe injuries treated by other hospitals within the trauma network will not be reported, nor will patients not requiring hospitalisation at all.

Although widely used, self-reported cycling speeds may not accurately determine crash severity and helmet effectiveness. A study by Thompson et al.^35,50 found that self-reported speeds were accurate for children and recreational cyclists, but further research is needed to assess accuracy in other populations and the impact of cognitive biases and external influences.

With retrospective studies, the temporal aspects of cycling-related trauma and its management may be difficult to assess. This study design is also prone to recall bias and may suffer from incomplete historical data (as was the case for helmet use in this study). Finally, the epidemiology and patterns of injury observed in the cohort managed within Cambridge MTC might not be representative of other regions in the country, especially those with much lower cycling rates.

4. Conclusions

In conclusion, this study characteried the head and spine injuries sustained by 851 cycling accident patients presenting to a major trauma centre over a 9-year period. The findings suggest that cyclists are a vulnerable road user group with a high risk of head and spine injuries, and that accidents involving motorised vehicles are associated with poorer outcomes. As cycling rates continue to increase and efforts to popularize it as a sport and means of transportation continue, healthcare providers may expect to see an increase in bicycle-related injuries in their practice. The insights gained from this study can inform the prevention and treatment of these injuries.

Disclosure statement

No potential conflict of interest was reported by the author(s).