A National Study of Emergency Thoracotomy for Trauma.

INTRODUCTION: The role of resuscitative thoracotomy in the emergency department (ED) for patients that have suffered severe thoracoabdominal trauma has been the subject of much debate. Most studies that characterize emergency thoracotomy are from urban, academic, and trauma centers. We sought to describe patient and hospital characteristics of a nationally representative sample of ED thoracotomy (EDT).
METHODS: The health-care cost and utilization project 2013 National ED Sample (NEDS) and the 2013 National Inpatient Sample (NIS) maintained by the agency for health-care research and quality were used to generate a nationally representative estimate of resuscitative thoracotomies performed in the ED. We obtained patient demographics and clinical characteristics and compared the descriptive statistics of the two datasets.
RESULTS: The NEDS dataset identified 124 unsuccessful EDTs, whereas the NIS dataset identified 77 admissions for thoracotomy. When weighted to create a national estimate, these represent 952 emergency thoracotomies performed in the US in 2013. Most were male (82.5% and 88.2% in NEDS and NIS, respectively). In addition, 32.9% and 36.4% in NEDS and NIS, respectively, were between the ages of 20 and 29. The majority of thoracotomies were performed at metropolitan teaching hospitals (64.2% and 75.3%, NEDS and NIS, respectively). The mean total ED charges for patients who had an unsuccessful thoracotomy were $32,664 and the mean total inpatient charges were $141,215.
CONCLUSION: Nearly 1000 thoracotomies are performed annually on the day of presentation to U. S. hospitals. Although emergency thoracotomy for trauma is an infrequently performed procedure, it almost always occurs at an urban, high volume, and level I or level II trauma centers. Copyright:

INTRODUCTION

Emergency thoracotomy is a resuscitative intervention for patients who have suffered severe thoracoabdominal trauma. Considered a salvage therapy, emergency department thoracotomy (EDT) has been the subject of much debate including which patients should receive the procedure and which centers should perform it.[1] The risks associated with EDT are not benign and include health-care provider exposure to blood-borne pathogens.[2] Studies have shown varying rates of overall survival from the procedure, ranging from 5% to 25%.[34567] A meta-analysis of 24 studies of EDT with a total of 4620 cases included demonstrated an overall survival of 7.4%, with survival from penetrating trauma the highest at 8.8%.[8]

The American College of Surgeons Committee on Trauma recommends EDT be performed in traumatic arrest only with witnessed signs of life and short transport times in penetrating trauma and in blunt trauma patients only with witnessed arrest at a trauma center.[9] However, nationally between 30% and 50% of patients with major injuries are hospitalized in nontrauma centers.[101112] Most studies that characterize emergency thoracotomy are from urban, academic, and trauma centers. While such centers have trauma teams at the ready, often community and rural emergency departments (ED) are operating under very different constraints in terms of resources and personnel which may lead to different outcomes. The incidence of EDT in a nationally generalizable sample of all hospitals, both urban and rural, is not known. As there is limited nationally representative data, we sought to describe the patient and hospital characteristics of emergency thoracotomy for trauma.

METHODS

We performed a secondary analysis of the Health-care Cost and Utilization Project (HCUP) 2013 National ED Sample (NEDS) and the 2013 National Inpatient Sample (NIS) conducted by the Agency for Healthcare Research and Quality. NEDS is a nationally representative estimate of all ED admissions in the United States. NIS is the largest publicly available all-payer inpatient health-care database in the United States and contains data from more than 7 million hospital stays each year. The Colorado Multiple Institutional Review Board approved this analysis as “not human subjects research.”

We used both the NEDS and NIS sample design to generate a nationally representative estimate of resuscitative thoracotomies performed. To obtain nationwide estimates, discharge-level weights were calculated for each visit by applying a formula that uses the American Hospital Association Annual Survey Database as the standard. These weights were used to account for the NIS and NEDS sampling schemes and allow for the calculation of descriptive statistics for patient and hospital characteristics.[13]

We analyzed demographic data (age and sex), month of admission, whether the admission occurred on a weekend, region the hospital was located in, teaching status of the hospital, urban or rural location, and total charges. In the NEDS sample, we also analyzed trauma center status and ED annual volume. We searched for the thoracotomy current procedural terminology codes 32100, 32110, 32160 in any of the 15 captured procedure fields. To investigate only those visits in which a thoracotomy was performed for resuscitative purposes, we limited visits to those with an injury diagnosis. We restricted our results to deceased ED patients in the NEDS sample to account for a limitation of this database, which is unable to determine the location of a performed procedure for a particular visit, to ensure only thoracotomies that occurred in the ED were captured. In the NIS sample, we limited our query to EDTs performed on the day of admission to capture resuscitative thoracotomy performed in the setting of trauma and not secondary to other disease processes. It was not possible to determine if the resuscitative thoracotomy was performed in the ED or operating room in NIS.

Statistical analysis

All analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA). Using NEDS and NIS data files, we obtained patient demographics and clinical characteristics with basic descriptive statistics. Our analyses were adjusted for sample weighting to provide nationally representative estimates.

RESULTS

We identified 124 ED visits that had an unsuccessful emergency thoracotomy performed from, along with 77 NIS visits that were identified where a thoracotomy was performed in the setting of injury. When weighted to create a national estimate, these represent 567 unsuccessful emergency thoracotomies in NEDS and 385 emergency thoracotomies NIS for a total of 952 emergency thoracotomies performed in the US in 2013. The characteristics of these visits are shown in Table 1. Most were male (82.5% and 88.2% in NEDS and NIS, respectively). Patients aged 20–29 accounted for 32.9% and 36.4% in NEDS and NIS, respectively. The majority of thoracotomies were performed at metropolitan teaching hospitals (64.2% in NEDS and 75.3% in NIS). Seventy percent of thoracotomies occurred in EDs with >60,000 visits per year. The mean total ED charges for patients who had an unsuccessful thoracotomy were $32,664 and the mean total inpatient charges were $141,215.

Table 1

Characteristics of patient visits with emergency thoracotomy

Variables	NEDS - died in ED		NIS - inpatient sample
	n	Weighted % (95% CI)	n	Weighted % (95% CI)
Demographic characteristics
Overall	124	-	77	-
Age years, mean (95% CI)	-	36.5 (32.6-40.5)		36.7 (33.4-40.1)
Male sex	103	82.5 (75.8-89.3)	67	88.2 (82.5-93.8)
Weekend admission	53	42.2 (29.9-54.5)	28	36.4 (26.9-45.8)
Month of arrival
December-February	23	17.5 (8.9-25.9)	26	33.8 (26.2-41.4)
March-May	17	14.2 (7.0-21.4)	18	23.4 (15.1-31.6)
June-August	26	21.0 (11.1-30.8)	18	23.4 (15.1-31.6)
September-November	18	13.0 (4.4-21.7)	15	19.5 (12.0-27.0)
Missing	40	34.3 (8.1-60.5)	-	-
Hospital characteristics
Region
Northeast	<10	6.0 (3.8-8.1)	11	14.3 (12.8-15.8)
Midwest	17	13.9 (1.6-26.3)	12	15.6 (14.0-17.2)
South	52	43.8 (24.7-62.8)	27	35.1 (30.1-40.0)
West	48	36.3 (21.6-51.1)	27	35.1 (29.1-41.0)
Trauma designation
Trauma center Level I	64	55.4 (38.8-71.9)	-	-
Trauma center Level II	17	12.6 (6.2-19.1)	-	-
Trauma center Level I or II	29	23.4 (11.3-35.4)	-	-
Level III/nontrauma center	14	8.6 (5.5-11.7)	-	-
Teaching and urban-rural status
Metropolitan nonteaching	46	35.0 (20.8-49.1)	16	20.8 (18.6-23.0)
Metropolitan teaching	77	64.2 (49.9-78.6)	58	75.3 (72.7-77.9)
Micropolitan/rural	<10	0.8 (0.5-1.1)	<10	3.9 (3.5-4.3)
Annual ED visit volume, mean (95%CI)	-	72,840 (62,478-83,201)	-	-
<40,000	<10	7.2 (0.0-15.6)	-	-
40,000-59,999	25	22.5 (3.3-41.7)	-	-
60,000-79,000	57	46.9 (18.7-75.1)	-	-
≥80,000	33	23.4 (3.9-43.0)	-	-
Outcomes
ED disposition
Died in ED	124	100.0	-	-
Died in hospital	-	-	22	28.9 (20.7-37.2)
Home health or routine	-	-	37	48.7 (38.5-58.8)
Transfer other	-	-	10	13.2 (7.8-18.6)
Transfer to short term hosp	-	-	<10	9.2 (3.6-14.8)
Injury severity score, mean (95% CI)*	-	17.2 (13.0-21.3)	-	-
Total charges, mean (95% CI)	-	$32,664 (22,469-42,858)	-	$141,215 (83,243-199,185)
Mechanism of injury**
Cut	13	10.2 (4.8-15.6)	-	-
Firearm	58	46.4 (32.0-60.8)	-	-
MVT	31	25.4 (9.6-41.1)	-	-
Fall	<10	1.4 (0.0-3.1)	-	-
Intent assault	49	38.4 (24.1-52.6)	-	-
Intent self-harm	<10	2.8 (0.0-6.1)	-	-
Intent unintentional	50	41.1 (22.5-59.7)	-	-

*The injury severity score was only available in the NEDS dataset, **The mechanism of injury was only available in the NEDS dataset. NEDS, National emergency department sample, NIS: National inpatient sample, ED: Emergency department; CI: Confidence interval, MVT: Motor vehicle trauma. Less than 10 observations not reported as specific number of observations per HCUP requirements. HCUP: Healthcare cost and utilization project

Among the visits for which an unsuccessful EDT was performed, the top diagnoses were crushing injury or internal injury (42.6%), cardiac arrest and ventricular fibrillation (23.9%), and open wounds of head, neck, and trunk (20.4%). Penetrating trauma accounted for 56.6% of cases in the NEDS population. In the NIS dataset, patients who survived to hospital admission had the most common diagnoses of crushing injury or internal injury (49.4%); open wounds of head, neck, and trunk (13.0%); and intracranial injury (11.7%).

DISCUSSION

We conducted an analysis of a nationally representative sample of hospital and patient characteristics for visits where an EDT was performed, using two large, national datasets. Consistent with prior studies,[25] the majority of patients who underwent EDT were male, and approximately 30% were aged 20–29. Penetrating trauma was the indication for emergency thoracotomy in 56.6% of NEDS patients. In almost one-third of visits, thoracotomy was performed for motor vehicle trauma, consistent with a recent report by DuBose, et al.[5] Similarly, western Europe has also seen a greater proportion of EDTs performed for blunt trauma.[14] We did not have the ability to identify if EDT was performed for nontraumatic injuries, as EDT is becoming expanded to other conditions.[15]

Most thoracotomies were performed at high volume, Level I trauma centers, predominantly in the south and west of the US and in metropolitan settings. Fewer than 5% of emergency thoracotomies occurred in nonurban and nonteaching facilities. Several studies have reported that higher institutional volume is associated with improved patient outcomes, with respect to high impact, low frequency procedures.[16171819] A recent study in a single state by Dumas et al.[7] found that patients presenting to centers that had a higher mean annual volume of EDTs performed had significantly higher odds of survival. Unfortunately, we are restricted by the HCUP policies on reporting data with <10 observations and are unable to report the survival among this population.

Our study demonstrates that one-third EDTs are performed at urban nonteaching compared to almost two-thirds at urban teaching institutions. Although urban institutions overall perform a higher volume of EDTs, evidence has not shown that teaching status has a significant effect on patient outcomes from EDT.[7]

In the cases that survived to hospital admission, obtained from the NIS dataset, almost half (48.7%) survived to discharge home from the hospital. This finding is somewhat higher than has been reported previously in the literature.[8] We were not able to discern the proportion of patients that did or did not have signs of life before EDT, as well as other features of the clinical presentation to better understand this finding. More qualitative work will need to be performed to characterize the decision-making involved in performing EDTs from our data.

Our study helps to provide some insight into the locations in which EDTs occur, including a small percentage that occur in rural locations. This is supported by the fact that 22% of observations for EDT were transferred to other facilities. It might be interesting to further study the characteristics of rural locations as well as those patients who are transferred to other facilities in an effort to understand the features of their injuries as well as their clinical course.

Limitations

Our study has several limitations. Given the manner in which the data are collected, the HCUP is subject to the limitations of administrative data, with possible errors in data collection and coding. We primarily used limits such as the presence of injury in our analysis of the NIS dataset to identify the patients who underwent a thoracotomy in the ED. While we believe we have excluded the majority of admissions, for which a thoracotomy was performed outside of the ED, it was not possible to prevent this completely as some thoracotomies for trauma may occurred in the operating room on the day of admission rather than in the ED. Finally, while it is possible that newer techniques in trauma resuscitation, namely, resuscitative endovascular balloon occlusion of the aorta (REBOA), may have potentially decreased the incidence of thoracotomy, the study was conducted in 2013 before the release of more modern REBOA catheters which makes a significant effect less likely.

CONCLUSION

We estimate nearly 1000 thoracotomies are performed annually on the day of admission to US hospitals. Although emergency thoracotomy for trauma is an infrequently performed procedure, it almost always occurs at an urban, high volume, level I or level II trauma center. Given limited data on the efficacy in blunt trauma patients, a surprising proportion were after motor vehicle trauma.

Research quality and ethics statement

The authors of this manuscript declare that this scientific work complies with reporting quality, formatting, and reproducibility guidelines set forth by the EQUATOR Network. The authors also attest that this clinical investigation was determined to require the Institutional Review Board review and has been approved. We also certify that we have not plagiarized the contents in this submission and have done a plagiarism check.

Financial support and sponsorship

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Conflicts of interest

There are no conflicts of interest.