An Elevated Metrorail as a Source of Orthopedic Injuries and Death at a Level-I Trauma Center

Abstract

Background:

Elevated Metrorail systems differ from conventional trains by their slower speeds and collisions with pedestrians predominantly occurring at accessible stations or platforms. Here, the orthopedic implications of pedestrians struck by a Metrorail are evaluated, as were the correlations of substance abuse and psychiatric history on injury and death.

Methods:

Retrospective cohort study at a single Level-1 trauma center of patients requiring admission with orthopedic injuries following Metrorail impact from 1/2004-2/2017. Demographics, substance abuse, psychiatric history, intentionality, LOS, follow-up, fracture characteristics, and management were studied.

Results:

33 patients sustained 104 total orthopedic injuries requiring admission; nine sustained 15 traumatic amputations. There were at least 37 open fractures, with some incomplete data in deceased (5) and amputation (9) patients. Suicide attempts were completed at 35.7% and were associated with a documented psychiatric illness and prior psychiatric evaluation. Spine injuries were associated with increased traumatic brain injuries, rib fractures, and open pelvic ring injuries, yet fewer humerus fractures. Open fractures were significantly predictive of death. 14 patients (42.4%) required ICU admission, and 26 (78.8%) patients required orthopaedic surgery (mean 1.3 ± 1.4 operations).

Conclusions:

Metrorail systems are unique sources of orthopaedic injuries requiring high rates of critical care and surgical intervention. Patients sustain multiple injuries, many with amputations. With this mechanism, there is a high rate of open fractures and suicide. Trauma centers should emphasize an extensive evaluation of orthopaedic injuries in this patient setting.

Level of Evidence: II

Introduction

There are currently 108 rail transport systems operating across the United States, and of these, 29 are classified as Metrorail commuter rail systems.¹ In 2016, over 8,000 injuries occurred in American railway systems, resulting in 786 deaths.² In the city of our study, there are 25 miles of elevated railway used by approximately 33,000 people per month.³ This system differs from traditional train systems in that a Metrorail typically maintains slower speeds, avoids motor vehicle interaction, and only interacts with pedestrians at stations/ platforms. While much literature has been published regarding the dangers of vehicles or cyclists being struck by conventional trains,^4,5 there is almost no literature regarding the orthopaedic ramifications of pedestrians struck by an elevated Metrorail system.

This study is the largest single-center review of Metrorail injuries to date and is the first attempt at risk factor correlation with analysis of age, gender, substance abuse and psychiatric illness. Due to its central location in a major city, our Level-I trauma center was well-equipped to study Metrorail injuries. Riding a Metrorail is low-cost, easily accessible, and a frequently utilized mode of travel; however, Metrorail stations have minimal, if any, safety barriers in place. Pedestrians often fail to adhere to critical safety guidelines requiring safe distances from the Metrorail tracks. Substance abuse, psychiatric comorbidities, and suicidal ideation increase the risk of intentional and unintentional injury at such locations.⁶

Currently there is limited data demonstrating frequency of suicide in regards to elevated Metrorails⁷ as well as concomitant factors associated with being struck. The primary objective was to describe patient characteristics, detail intentionality, and analyze the anatomic distribution of injury with a focus on orthopedic fractures. Secondary objectives include risk factor identification and patient outcomes. Such information may allow for possible safety measures and management strategies to prevent or mitigate consequences from these injuries.

Materials and Methods

Study Design

After receiving Institutional Review Board (IRB) approval, a retrospective cohort study was performed at a single Level-I trauma center of patients requiring inpatient admission for significant orthopedic injuries after being struck by a Metrorail from 1/2004-2/2017.

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. For this type of study formal consent is not required.

Participants

“Orthopedic injuries” were defined as joint capsule violations, joint dislocations, tendon lacerations, and all fractures except those of the face/skull and ribs. Patients who presented to the emergency room but were then discharged without requiring admission were excluded. Patients who expired prior to arrival or shortly thereafter were included, albeit with missing data due to incomplete fracture workup and medical history.

Charts were reviewed for age, gender, psychiatric illness, substance abuse, hospital length of stay (LOS), intensive care unit (ICU) duration, follow-up time, orthopedic injuries, traumatic brain injuries (TBI), rib fractures, hospital course, and outcomes. Orthopedic injury information was based on radiologist reports, clinical notes, operative notes and discharge summaries. We also documented prior formal psychiatric evaluations at our institution based on chart review. Drug abuse was defined as a documented history of using illicit drugs, and alcohol use defined as those who reported daily intake.

Management

Per institutional protocol, all patients with open fractures received a 3-day course of intravenous (IV) antibiotics (2g ceftriaxone, daily) initiated immediately upon presentation. Usually a patient with this mechanism of injury (MOI) received a full body computed tomography (CT) scan if hemodynamically stable at presentation. If unstable, management varied depending on the degree of instability, with some patients receiving bedside radiographs before urgent surgical management with general trauma surgery. Open fractures were irrigated with 2-4 liters of normal saline in the trauma bay and then immobilized, and large wounds loosely approximated with nylon sutures prior to formal operative debridement.

Data analysis

Statistical analysis was performed using SAS University Edition. Correlations were determined using either analysis-of-variance (ANOVA) or Chi-square/Fisher’s Exact Test, for parametric and non-parametric data, respectively. Means and frequencies were analyzed and p-values of <0.05 were considered significant.

Results

Orthopedic injuries

A total of 33 patients with 104 total orthopedic injuries met our inclusion criteria. The cohort included 25 males (75.8%) and 8 females (24.2%) with a mean age of 37.7 ± 13.8 years (range 19-79 years) (Table 1). Nine patients (4 male, 5 female) did not survive their hospital stay (mean LOS of 8.7 ± 8.6 days, median 1 day). Of these, one acutely expired on arrival and the orthopedic injuries were not accounted for; this patient was excluded from further statistical analysis. Of the 24 surviving patients, the mean hospital LOS was 24.7 ± 25.0 days (median 14 days). The entire cohort had a mean of 1.2 ± 1.3 trips to the operating room for orthopedic issues, and 4 (12.1%) patients had operations performed for spinal injuries.

Table 1.

Patient Demographics & Hospital Course Stratified by Surviving & Death (At Arrival / While In-Patient)

Demographics	Totala	Aliveb	Deadc	p-value
Total (with completed Hx)	33	24 (72.7%)	9 (27.3%)	-
Mean age ± SD	37.7 ± 13.8	38.1 ± 13.9	37.3 ± 14.3	0.894
Males	25 (75.8%)	21 (87.5%)	4 (44.4%)	0.009
Females	8 (24.2%)	3 (12.5%)	5 (55.6%)	0.009
White	18 (54.5%)	16 (66.7%)	2 (22.2%)	0.022
Black	11 (33.3%)	6 (25.0%)	5 (55.6%)	0.103
Hispanic	4 (12.1%)	2 (8.3%)	2 (22.2%)	0.291
Psychiatric History	17 (51.5%)	14 (58.3%)	3 (42.9%)	0.512
Drug/EtOH abuse	19 (57.6%)	17 (70.8%)	2 (28.6%)	0.072
Intentional Injury	14 (42.4%)	9 (37.5%)	5 (55.6%)	0.141
Hospital Course
Mean open fractures ± SD	1.2 ± 1.3	0.75 ± 0.99	2.4 ± 1.5	0.019
Mean LOS ± SD	19.5 ± 23.3	24.7 ± 25.0	8.7 ± 8.9	0.010
Mean Ortho OR trips ± SD	1.2 ± 1.3	1.3 ± 1.4	0.9 ± 1.1	0.339
Spine OR trips	4 (12.1%)	4 (36.4%)	0 (0.0%)	0.038
Mean follow-up days ± SD	-	538.5 ± 1129.1	-	-
Mean days to death ± SD	-	5.4 ± 8.7	-	-

^aPercentages listed refer to percent of total n=33.

^bPercentages listed refer to percent of surviving patient n=24.

^CPercentages listed refer to percent of non-surviving patient n=9.

*2 deceased patients were excluded from statistical analysis of psychiatric and drug/EtOH abuse due to incomplete history, as were 2 surviving patients and 1 dead patient being excluded from statistical analysis of open fractures due to incomplete assessment. All significant findings are listed in bold. (Hx) history; (Ortho) Orthopedic Surgery (LOS) Length of Stay; (EtOH) Alcohol; (OR) Operating Room; (SD) Standard Deviation

There were 37 (35.2%) upper extremity (9 radius, 5 scapula, 4 clavicle, 2 hand, 6 ulna, and 9 humerus), 33 (31.4%) lower extremity (9 femur, 9 tibia, 11 fibula, and 3 foot), 15 (14.3%) vertebral, 6 (5.7%) pelvic ring, and 5 (4.8%) acetabular injuries (Figure 1). Of the 33 patients, 20 (60.6%) patients sustained open fractures, with 9 (27.3%) patients sustaining more than one (Table 2). There were 37 open fractures in total (11 (29.7%) upper extremity, 22 (59.5%) lower extremity, and 4 (10.8%) pelvic ring). 9 (27.3%) patients sustained a total of 15 traumatic amputations (7 below knees, 2 above knees, 5 humeral, and 1 hand).

Figure 1.

Anatomic distribution of the 104 orthopedic injuries sustained from being struck by a Metrorail.

Table 2.

Association of Spine Injury (Bone or Ligamentous) with Other Orthopedic Injuries

Injury	Total	Spine Injury	No Spine Injury	p-value
n (patients)	32*	15	17
TBI (patients)	19	12 (80.0%)	7 (41.2%)	0.024
Rib fracture (patients)	12	9 (60.0%)	3 (17.7%)	0.015
Open fractures	37	13 (35.1%)	24 (64.9%)	0.237
Upper Extremity (Injuries)
Scapula	5	3 (20.0%)	2 (11.8%)	0.544
Clavicle	4	1 (6.7%)	3 (17.7%)	0.354
Humerus	9	1 (6.7%)	8 (47.1%)	0.011
Bilateral humerus	1	0 (0.0%)	1 (5.9%)	0.332
Open humerus	6	0 (0.0%)	6 (35.3%)	0.010
Open elbow joint	2	1 (6.7%)	1 (5.9%)	0.930
Radius	9	4 (26.7%)	5 (29.4%)	0.868
Open radius	2	1 (6.7%)	1 (5.9%)	0.930
Ulna	6	3 (20.0%)	3 (17.7%)	0.871
Open ulna	3	2 (13.3%)	1 (5.9%)	0.498
Hand	2	0 (0.0%)	2 (11.8%)	0.163
Pelvis (Injuries)
Pelvic ring	6	5 (33.3%)	1 (5.9%)	0.063
Open pelvic ring	4	4 (26.7%)	0 (0.0%)	0.042
Acetabulum	5	4 (26.7%)	1 (5.9%)	0.131
Lower Extremity (Injuries)
Femur	9	3 (20.0%)	6 (35.3%)	0.260
Bilateral femur	1	0 (0.0%)	1 (5.9%)	0.332
Patella	0	0 (0.0%)	0 (0.0%)	-
Open knee joint	1	0 (0.0%)	1 (5.9%)	0.332
Tibia	9	2 (13.3%)	7 (41.2%)	0.119
Bilateral tibia	4	1 (6.7%)	3 (17.7%)	0.354
Open tibia	11	3 (20.0%)	8 (47.1%)	0.273
Fibula	11	4 (26.7%)	7 (41.2%)	0.167
Bilateral fibula	3	0 (0.0%)	3 (17.7%)	0.083
Open fibula	10	2 (13.3%)	8 (47.1%)	0.130
Foot	4	4 (26.7%)	0 (0.0%)	0.104
Bilateral foot	2	2 (13.3%)	0 (0.0%)	0.334
Open foot	1	1 (6.7%)	0 (0.0%)	0.334

*1 patient was excluded from statistical analysis due to unknown presence or absence of spine injury. All significant findings are listed in bold; (TBI) Traumatic Brain Injury

Spine injuries

In reviewing each patient’s injury profile, of particular interest was the distribution of spinal injuries and their association with other orthopaedic injuries. Four surviving patients were taken to the operating room for spinal injuries (p=0.038). Patients sustaining spinal injuries had a significantly higher incidence of comorbid TBI (p=0.024), >1 rib fracture (p=0.015), and open pelvic ring fractures (p=0.042), as well as a significantly lower incidence of humerus fractures (p=0.011) than those patients who did not. Furthermore, there was no significant association between spinal injury and gender or intentionality.

Survivors vs Deceased

Comparing the 24 (72.7%) surviving patients to the 9 (27.3%) who died either at presentation or as an inpatient, surviving patients had a mean age of 38.1 ± 13.9 years (range of 22-79 years), compared to non-surviving patients with a mean age of 37.3 ± 14.3 years (range 19-64 years; p=0.894) (Table 1). Gender was a predictor of survival with 21/25 (84.0%) males surviving, compared to just 3/8 (37.5%) females (p=0.009). Surviving patients had a mean hospital stay of 24.7 ± 25.0 days, and those who died in-hospital had a mean hospital stay of 8.7 ± 8.6 (p=0.010), with an average of 5.4 ± 8.7 days to death. While surviving patients sustained an average of 0.75 ± 0.99 open fractures, compared to the mean of 2.4 ± 1.5 open fractures in deceased patients (p=0.019), a full evaluation of deceased patients may have been limited at presentation.

Intent, psychiatric illness, and substance abuse

A secondary parameter of interest was the intent behind injury (intentional vs accidental) and any associations with patient psychiatric history, comorbid substance abuse, and mortality (Table 3). Nineteen of the 33 patients (57.6%) had a history of substance abuse and 17/19 (89.5%; p=0.072) patients survived. No significant associations were found between intent and substance abuse or mortality. Furthermore, 17/33 (51.5%) had a known history of psychiatric illness and 14/17 (82.3%; p=0.512) patients survived. Out of the 14 patients sustaining intentional injuries, 11 (78.6%) also had a history of psychiatric disorder (p=0.014), and 11/14 (78.6%) had a prior psychiatric evaluation at our institution (p<0.001). Interestingly, those suffering intentional injuries with a prior psychiatric evaluation were injured an average of 24 months (range 1 day-7.8 years) from their last evaluation. Of the 31 patients with known intentionality, injuries were believed to be intentional in 14 (45.2%) patients, with 5/14 (35.7%; p=0.141) achieving suicide.

Table 3.

Intention Behind Injury Versus History of Psychiatric Disorder, Psychiatric Evaluation, Substance Abuse, and Death

	Total	Intentional	Accidental	p-value
n	31*	14*	17*	-
Documented psychiatric Hx	17 (54.8%)	11 (78.6%)	6 (35.3%)	0.014
Prior documented psychiatric evaluation	13 (41.9%)	11 (78.6%)	2 (11.8%)	<0.001
Days from last psychiatric	-	730.2 ± 902.5	19.0 ± 15.6	0.026
evaluation to injury (mean ± SD)
Drugs/EtOH Abuse Hx	18 (58.1%)	10 (71.4%)	8 (47.1%)	0.098
Alive	24 (77.4%)	9 (64.2%)	15 (88.2%)	0.138
Dead	7 (22.6%)	5 (35.7%)	2 (11.8%)	0.141

*2 patients were excluded from statistical analysis due to unknown intention behind injury; All significant findings are listed in bold; (EtOH) Alcohol; (Hx) History (SD) Standard Deviation; 730.2 days is approximately 24 months.

Discussion

The available literature regarding train-related injury has focused on traditional railways. Our study specifically evaluated an elevated Metrorail system, allowing us to uniquely exclude any incidences involving motor vehicles. Train-related trauma results in extremely high hospitalization costs. Documented case series have shown a monetary rate averaging $18,698 in one study and $61,159 in another,^8,9 with some hospitalizations approaching $165,000 per patient.⁸ In 1998 Goldberg et al.⁹ documented significant deficits in reimbursement compared to cost of hospitalization in train trauma, with an estimated cost to society of $300 million annually.

The majority of the patients involved were male—a fact correlating with multiple studies describing male gender as a possible risk factor in traditional traumatic train injuries.^8,10-13 While males were the primary victims; they were also significantly more likely to survive than females. This study demonstrated a need for orthopedic intervention in 78.8% of cases, especially given that 60.6% of patients sustained ≥1 open fracture, highlighting the unique, high-energy mechanism of injury. Open fractures were significantly associated with death, and the presence of ≥1 open fracture correlated with an increased risk of fatal injuries. Traumatic amputations occurred in 27.3% of cases, with 15 amputations in 9 patients, reflecting their prevalence in train-pedestrian accidents^9,13 Moore¹⁴ described a series of accidents in which there were 10 traumatic amputations in 8 victims. Blazar et al.⁸ described a series of 17 train-child accidents in which there were 20 traumatic amputations in 17 victims.

Our investigation demonstrated an approximately equal distribution of upper (35.2%) and lower (31.4%) extremity fractures. Prior studies document a higher incidence of lower extremity injuries. Singer reported 19 lower and 2 upper extremity amputations, with the foot being the most commonly injured body segment, with 52% of patients sustaining direct fractures and/or amputations.¹⁵ Agalar also documented a high rate of lower extremity amputation in traditional train-pedestrian trauma.¹³ In contrast to traditional, ground-level train tracks, Metrorail stations and tracks are usually elevated, with minimal access to track level, aside from intentionally or unintentionally falling onto them. Traditional tracks provide easier means of access, which may explain the higher incidence of lower extremity injuries, as both train and pedestrian are at ground level. Overall, the amputation rate is extremely high (38% to 61%),^8,15,16 causing high morbidity in survivors.

In this series, 19% of patients sustained spine injuries. Those with spine injuries were significantly more likely to sustain TBI, rib fractures and open pelvic fractures, suggesting that these patients sustain higher-energy injury, more direct body-centered impact loads, and a poorer prognosis compared to those without. Interestingly, in a 10-year review of pediatric train-related injuries, no intentional injuries were recorded and no spinal injuries documented,⁸ suggesting that spine trauma in train-pedestrian accidents may be related to suicide attempts, intentionality, and mature musculoskeletal systems.

While many prior studies well document pediatric injuries while playing or walking on train tracks, ours included no children. Singer published a series of 211 patients where 8% of the patients were children. Children injured in that population were mainly toddlers who wandered onto train tracks in rural areas. The 10-year review by Blazar et al. focused specifically on pediatric train trauma, detailing 17 cases where children were injured while boarding moving trains, walking on tracks, and running alongside moving trains.⁸ It is likely that pediatric injuries are much less common in our Metrorail-related trauma due to the fact that the tracks are not at ground level and inaccessible to the public. The location of the tracks precludes many mechanisms of accidental trauma described in the pediatric population, providing a mode of protection from such injury.

Both psychiatric disorders and substance abuse are risk factors for Metrorail injuries. The rate of suicide attempts in the current study was 45%, much higher than that in published train-pedestrian studies, with rates ranging from 17-22%.^13,17 Due to the lack of information and relatively small population in each of these studies, it is unclear whether suicide rates are truly higher in Metrorail versus train accidents, warranting more study. Of note, in our study intentionality itself was not predictive of death, since several of our deceased patients had no intentionality documented.

Alcohol intoxication has previously been studied as a possible risk factor for train-pedestrian accidents. Nixon et al. demonstrated alcohol as one of the highest risk factors for injury, second only to being a railway employee.¹⁷ Our study demonstrated alcohol influence in a staggering 58.1% of cases; although this number relies on appropriate documentation and patient disclosure of such information, this number is similar to those in previous studies. In a study of tram-pedestrian injury by Hedelin et al., 60% of cases involved alcohol intoxication.¹⁰ Similarly, in his study of railroad-related deaths, Pelletier found that 78% of cases occurred while under the influence of alcohol.¹² Metrorail-related injury is thus similar to train and tram-related injury, where alcohol is an independent variable associated with a higher risk of injury. Our study did not demonstrate any significant relation between alcohol and intentionality. Out of the 18 cases with documented alcohol use, 10 cases were intentional and 8 were unintentional (p=0.098).

Psychiatric history was evaluated, to our knowledge, for the first time with respect to railway-related accidents. 55% of our cases had a documented psychiatric history. The presence of a known psychiatric condition correlated with a higher risk of intentionality (p=0.014), and those with a previous inpatient psychiatric evaluation demonstrated an even more significant correlation (p<0.001). Thus, when involved in a Metrorail-pedestrian accident, psychiatric patients are more likely to have acted intentionally. Interestingly, we observed that the time since a patient’s previous psychiatric follow-up appointment had a significant impact on intentionality. Of those with a documented history of psychiatric treatment at our facilities, the amount of time from the last known evaluation to the injury was 730.2 days (2 years) in the intentional group compared to just 19 days in the unintentional group (p<0.001). This difference may suggest the importance of follow-up in patients with psychiatric disease, as loss to follow-up and noncompliance may increase the likelihood of suicide.

One of the weaknesses of our study is its retrospective nature, and we relied on documentation available to understand the patient’s injury and circumstances surrounding the event. We also lacked a complete evaluation on total fracture numbers with regards to amputated limbs, and no radiographs were obtained on patients who died shortly after presentation. When investigating the time since last psychiatric evaluation, we were only able to view those occurring at our institution, and any evaluation at an outside practice was unknown to our team, possibly skewing the time to injury. We also frequently relied on the patient’s disclosure regarding sensitive information such as psychiatric illness, intentionality, and substance abuse. Furthermore, our study focused only on admitted patients, yet theoretically some patients with Metrorail injuries may not require orthopedic surgery consultation. Finally, we did not analyze the outcome of wounds treated as outpatients.

In conclusion, this study of train-pedestrian accidents demonstrates an extensive injury profile and associations with mental illness and substance abuse. Unfortunately, both accidental and intentional patients sustain a high rate of open fractures, traumatic amputations and subsequent orthopedic surgery. Open fractures were significantly associated with death. With the increases in Metrorail prevalence, manufacturers, policymakers, and healthcare employees should work towards improving technologies and usage rules to protect bystanders and reduce the burden of pedestrian injuries on healthcare systems.