Trauma in Obstetrical Patients.

INTRODUCTION: Pregnant trauma patients are an underdescribed cohort in the medical literature. Noting injury patterns and contributors to mortality may lead to improved care.
METHODS: Female patients between 14 and 49 years of age were identified among entries in the 2017 National Trauma Data Bank. Data points were compared using Chi-square test, Fisher's exact test, Student's t-test, Mann-Whitney rank-sum, or multiple logistic regression as appropriate. P < 0.05 was used to determine the findings of significance.
RESULTS: There were 569 pregnant trauma patients identified, which was 0.54% of the 105,507 women identified. Overall, mortality was low among all women and not different between groups (1.2% for pregnant women vs. 2.2% for nonpregnant, P = 0.12). Pregnant women with head injuries had a higher mortality rate than pregnant women without (4.2% vs. 0.47%, P < 0.01). Head injuries (Abbreviated Injury Severity Score [AIS] head >1) were associated with an increased risk for mortality (odds ratio: 3.33, 95% confidence interval: 3.0-3.7, P < 0.01).
CONCLUSION: There was no increase in mortality for trauma patients who are pregnant when controlling for covariates. Factors such as head injuries, the need for blood, and comorbid diseases appear to have a more significant contribution to mortality. We also report the prevalence of head, cervical spine, and extremity injuries in pregnant trauma patients. Multidisciplinary simulation, jointly crafted protocols, and expanding training in regional anesthesia may be the next steps to improving care for pregnant trauma patients. Copyright:

INTRODUCTION

Current recommendations suggest treating pregnant trauma patients similarly to nonpregnant trauma patients with a few additional concerns.[12] One unique recommendation is: “Pregnant trauma patients with… uterine tenderness, significant abdominal pain, vaginal bleeding, sustained contractions, rupture of the membranes, atypical or abnormal fetal heart rate pattern, high risk mechanism of injury, or serum fibrinogen < 200 mg/dL should be admitted for observation for 24 h.”[[1] A few reports have characterized pregnant trauma patients, but gaps in knowledge remain.[345678] Many recommendations are based on consensus opinions. We analyzed the 2017 National Trauma Data Bank (NTDB) to further describe pregnant trauma patients and to provide data to support or refute current recommendations. We assessed for a difference in outcomes for pregnant trauma patients. We also attempted to assess if additional care for pregnant trauma patients significantly increased the burden on the medical system. Further, we endeavored to describe the injury patterns sustained and procedures undergone including the incidence of emergent post cardiac arrest cesarean sections.

METHODS

This retrospective study was performed after institutional approval and the manuscript was referenced against the STROBE checklist. Female patients between 14 and 49 years of age were identified among entries in the NTDB for the year 2017. The age range examined was based on years of fertility and to align with criteria utilized in previous studies.[78] The NTDB defined pregnancy as having reached at least 20 weeks post conceptual age.[9]

Data collected included pregnancy status, age at time of injury, ethnicity, insurance status and type, length of hospital stay, length of intensive care unit (ICU) stay, number of days of mechanical ventilation, International Classification of Diseases version 10 (ICD-10) diagnosis and procedure codes, Injury Severity Score (ISS), Abbreviated Injury Severity Score (AIS) of all body regions, hospital mortality, comorbid diseases (i.e. diabetes, hypertension, and history of smoking), results of alcohol and drug screens, hospital complications (i.e. thrombotic events, embolic events, respiratory distress syndrome, unplanned intubation, and pneumonia), abuse report status, and mechanism of injury. Specific diagnoses were identified using AIS diagnosis codes as well as ICD-10 diagnosis codes. If there was a discrepancy in diagnosis for a patient between the AIS diagnosis code and the ICD-10 diagnosis code (i.e. disease or injury present for a patient with one coding system but not the other), the disease was assumed to be present. Procedures were identified using ICD-10 procedure codes. Only patients with full information for gender, age, ISS, and discharge status were included in the analysis.

Data were prepared and analyzed using Microsoft Excel (Microsoft, Redmond, WA) and SigmaPlot 13 (Systat, San Jose, CA). Nonparametric variables were summarized as the median and interquartile range (IQR) and were compared with the Mann–Whitney U-test. Categorical variables were reported as the number and percentage of patients and compared using the Chi-square test. Data points were compared using Chi-square, Fisher's exact test, Student's t-test, Mann–Whitney rank-sum, or multiple logistic regression as appropriate. P < 0.05 was used to determine the findings of significance.

RESULTS

Pregnant versus nonpregnant patients

There were 569 pregnant patients between 14 and 49 years of age identified (0.54% of the 105,507 women with complete datasets). A higher percentage of pregnant patients were white than nonpregnant patients (48.2% vs. 34.3%, P < 0.01) [Table 1]. Pregnant patients were younger (median age: 26 years, IQR: 22–31 vs. 30, IQR: 22–39, P < 0.01. ISS was lower among pregnant women (2, IQR: 2–5 vs. 5, IQR: 4–10, P < 0.01). Pregnant women had lower rates of head injuries as well (25.0% vs. 30.8%, P < 0.01). The transfusion rate was similar between groups. More pregnant patients had an abuse report filed related to their injuries (2.8% vs. 1.6%, P = 0.04). Pregnant patients had lower rates of smoking (8.8% vs. 23.8%, P < 0.01), hypertension (1.9% vs. 7.8%, P < 0.01), and positive alcohol screens (6.2% vs. 15.2%, P < 0.01). Overall, mortality was low among all women and not different between groups (1.2% for pregnant women vs. 2.2% for nonpregnant, P = 0.12). The rate of deaths in pregnant women in the emergency department was not different from the rate among nonpregnant women (0.5% for pregnant patients vs. 0.8%, P = 0.71). The inhospital mortality rate was lower among pregnant patients (0.7% vs. 2.1%, P = 0.02).

Table 1

Demographics of pregnant compared to nonpregnant patients

Measured variable	Pregnant patients (%)	Nonpregnant patients (%)	P
White race	48.2	34.3	<0.01
Privately insured	39.4	46.5	<0.01
Age (years)	26 (22-31)	30 (22-39)	<0.01
Abuse report filed	2.8	1.6	0.04
Head injury (AIS >1)	25.0	30.8	<0.01
ISS	2 (2-5)	5 (4-10)	<0.01
Mortality rate	1.2	2.2	0.11
Transfusion rate	4.3	4.4	0.97
Positive alcohol screen	6.2	15.2	<0.01
Hypertension	1.9	7.8	<0.01
Smoking	8.8	23.8	<0.01
Diabetes	3.0	4.0	0.29

Data are presented as percentage or median (IQR). Data were analyzed using Chi-square or Mann-Whitney rank-sum as appropriate. ISS: Injury Severity Score, AIS: Abbreviated Injury Severity Score, IQR: Interquartile range

Pregnant patient injury patterns

Head injury (HI), defined as AIS head >1, was identified in 142 pregnant patients (25% of all pregnant patients) [Table 2]. There were 102 pregnant patients identified with a concussion (72% of pregnant head trauma patients), 55 with a closed skull fracture (38.7% of pregnant head trauma patients), and 3 with an open skull fracture (2.9%). Five had basilar skull fractures. Pregnant patients with a HI were more severely injured than those without (median ISS: 6, IQR: 4–14 vs. 2, 1–5, P < 0.01). Pregnant HI patients also exhibited a higher cervical spine injury (CSI) (42.0% vs. 18.2%, P < 0.01) and mortality rates (4.2% vs. 0.47%, P < 0.01) compared to pregnant women without a HI. A HI was associated with increased mortality (odds ratio [OR] 3.33, 95% confidence interval [CI]: 3.0–3.7, P < 0.01), whereas Glasgow Coma Scale (GCS) was not [Table 3]. Among all female patients with a GCS ≤13, both pregnant and nonpregnant, only 61.8% had an identified HI (AIS head ≥2).

Table 2

Injury patterns among pregnant patients

Body region	Incidence rate (%)
Head	25.0
Face	27.2
Neck	28.1
Chest	29.2
Abdomen	46.0
Spine	36.0
Upper extremity	43.4
Lower extremity	58.7
Soft tissue without specific location	10.7

Data presented are percentage of pregnant patients with respective type of injury

Table 3

Odds for mortality

Independent variable	OR	P
Pregnant	0.71 (0.30-1.72)	0.49
ISS	1.09 (1.09-1.10)	<0.01
AIS head >1	3.33 (3.0-3.70)	<0.01
White race	0.86 (0.78-0.96)	0.01
Hypertension	0.59 (0.47-0.75)	<0.01
Smoking	0.36 (0.31-0.42)	<0.01
Diabetes	0.92 (0.69-1.23)	0.59
Positive alcohol screen	0.48 (0.43-0.53)	<0.01
Abuse report filed	1.15 (0.82-1.63)	0.42
Transfusion in first 24 h	4.73 (4.19-5.29)	<0.01
Private insurance	0.52 (0.47-0.57)	<0.01
Penetrating mechanism	4.32 (3.62-5.17)	<0.01
Age (per year increase)	1.02 (1.01-1.02)	<0.01

Data presented as OR (95% CI). ISS: Injury Severity Score, AIS: Abbreviated Injury Severity Score, OR: Odds ratios, CI: Confidence interval

CSI occurred in 137 pregnant trauma patients (24.1% of all pregnant trauma patients). Pregnant patients with a CSI were more severely injured than those without a CSI (median ISS: 4, IQR: 2–10 vs. 2, 1–5, P < 0.01). The most common CSI was a ligamentous injury (93 patients, 68% of CSI) while cervical fractures (13 patients, 9.5%) and subluxations (13 patients, 9.5%) were less common. The remaining CSIs were unspecified. Of pregnant patients with a CSI, 58 (42.3%) had a concomitant HI and 35 (25.5%) suffered a more significant HI (AIS > 2).

Bony facial trauma occurred in twenty pregnant patients (3.5% of all pregnant patients) with six (30%) sustaining concurrent mandibular fractures. There were no isolated mandibular fractures. Five patients (0.9%) had dental fractures and ten (1.8%) had oropharyngeal soft-tissue injuries.

Thoracic trauma (TT) was identified (AIS thorax ≥1) in 166 pregnant patients (29.2% of pregnant trauma patients) while 44 patients (7.7%) sustained more significant TT (AIS ≥2). Patients with an AIS thorax ≥2 trended toward a higher mortality rate (4.65% vs. 0.95%, P = 0.16). The most common thoracic injuries were chest wall injuries (76 patients, 45.8% of all TT), with almost half of all chest wall injuries involving sternal or rib fracture (44 patients, 26.5%). There were 16 patients with a pneumothorax (9.6% of TT), 4 (2.4%) with a hemothorax, 10 (6.0%) with a pulmonary contusion, 2 (1.2%) with a pulmonary laceration, and 1 (0.6%) with a diaphragmatic laceration.

Procedures and surgery for pregnant trauma patients

There were 354 pregnant trauma patients who required surgery or a procedural intervention during their hospitalization (62.3% of all pregnant patients) [Table 4]. No women underwent post cardiac arrest cesarean section. The most common procedures were peripheral orthopedic procedures (involving a bone and/or joint) or soft-tissue procedures. The next most common operation type was open abdominal or pelvic surgery, which included 10 uterine surgeries (2.8% of patients requiring an operation), 8 splenectomies (2.3%), 2 hepatic surgeries (0.6%), and 13 other open surgeries of the gastrointestinal tract without further specification (3.7%). Intrathoracic surgery was performed on 14 pregnant trauma patients during admission (4% of those requiring an operation). There were an additional seven (2%) patients who underwent chest tube placement without surgery.

Table 4

Percentage of pregnant patients undergoing procedures

Body region of procedure	Patients requiring procedure during admission* (n=354)	Patients requiring procedure hospital day 1** (n=155)
Oral	2.3	2.6
Tracheostomy	0.9	0
Intrathoracic surgery	2.3	0.7
Intracranial	1.4	3.2
Face or scalp	2.6	2.6
Blood vessels	5.1	5.8
All intra-abdominal or pelvic	14.2	12.9
Uterus	2.8	0.7
Spleen	2.3	1.3
Liver	0.6	1.2
Gastrointestinal	3.7	3.9
Soft tissue (not face or scalp)	14.5	18.7
Peripheral orthopedic surgery	21.8	18.7
Hip or femur surgery	7.9	5.2
Pelvic bone surgery	1.1	0
Evacuation of fetal content	6.2	14.2

*Data presented as percentage of all pregnant patients who needed an operation during hospital admission, **Data presented as percentage of all patients who needed an operation on hospital day 1

There were 155 pregnant patients who required surgery or a procedure on hospital day 1 (27.2% of all pregnant patients and 43.8% of those requiring an operation or procedure). The most common were intra-abdominal (20 patients, 12.9% of patients undergoing procedures on hospital day 1), soft tissue (29 patients, 18.7%), and peripheral orthopedic surgery (29 patients, 18.7%). Only 51 patients were transferred directly from the emergency department to the operating room.

There were 92 pregnant HI patients who required a procedure (64.8% of all pregnant head trauma patients). Of this group, 40 required a procedure on hospital day 1 (28.1% of all pregnant head trauma and 43.5% of all pregnant HI who required a procedure). There were five patients with a HI who underwent intracranial surgery within the first hospital day. An additional five had a ventriculostomy placed on hospital day 1 without any other intracranial procedure. Among patients with HI, five also underwent surgery of the oropharynx, eight underwent surgery of the face or scalp, and five underwent surgery on their cervical spine.

Mechanism of injuries among pregnant patients

There were 369 pregnant trauma patients who were involved in a motor vehicle crash (MVC) (64.8% of pregnant trauma patients), 331 (89.7%) of whom were restrained by a seat belt [Table 5]. In 186 cases, an airbag was deployed (front, side, or both) (50.4% of MVCs). There were 33 patients injured in an MVC that involved intrusion into the vehicle cockpit (8.9%) and 11 patients (3.0%) were ejected from the vehicle. There were 17 pregnant patients involved in a fall of >20 feet or 3 times the patient's height (3% of all trauma patients). There were 16 patients who sustained penetrating injuries (2.8%) and 17 who sustained a burn (3%). The remaining causes were indeterminate.

Table 5

Mechanism of injury for pregnant patients versus nonpregnant patients

Mechanism	Pregnant patients	Nonpregnant patients	P
MVC	70.7	62.2	<0.01
Motorcycle	0	9.5	<0.01
Fall	3.7	5.7	0.75
Burn	3.7	3.2	0.86
Penetrating	3.0	2.7	0.73

Data presented as percentage of patients in each group. MVC: Motor vehicle crash

Hospitalizations

There was no difference in rates of hospitalization of pregnant and nonpregnant women (85.6% vs. 83.8% respectively, P = 0.29). However, pregnant women had a greater likelihood of being admitted to a hospital (OR: 1.68, 95% confidence interval [CI]: 1.31–2.16, P < 0.001) when controlling for age (OR: 1.02 per year, 95% CI: 1.01–1.20, P < 0.001) and ISS (OR: 1.14, 95% CI: 1.14–1.14, P < 0.001). A higher percentage of pregnant women were admitted for observation of < 24 h (6.35% vs. 4.32%, P = 0.03). There was a significantly lower rate of ICU admission for pregnant women (8.8% vs. 16.7%, P < 0.01). There were 32 pregnant patients who required mechanical ventilation (5.6%) with a median length of 2 days (IQR: 2–6).

DISCUSSION

We found no increase in mortality for trauma patients who are pregnant when controlling for covariates. In contrast, Deshpande et al. described higher odds for mortality among pregnant patients.[7] Our model may be more accurate for several reasons. First, Deshpande et al. utilized GCS, whereas we included an independent variable for HI (AIS head >1). In our model, there was a significantly higher risk for mortality associated with a HI, whereas a decreased GCS was not associated with a change. In our dataset, there were a significant number of patients with a depressed GCS without an identified HI (possibly intoxicated patients), which may be why we did not find an association between mortality and GCS. AIS head was a better variable for mortality risk stratification. Second, Deshpande et al. did not include the transfusion of blood in their model, whereas we demonstrated that the need for a transfusion in the first 24 h of admission was associated with mortality. Third, we included comorbid conditions such as diabetes, hypertension, and smoking to better control for the differences among patients. Our findings suggest that overall injury patterns, presence of a HI, the need for a transfusion, age, race, socioeconomic status, and comorbid conditions are associated with changes in mortality for female trauma patients, while pregnancy is not. Our findings are similar to other reports in a broader sample of trauma patients. Race, socioeconomic status, diabetes, and metabolic syndrome have been described as risk factors for mortality.[1011] HIs, along with injury severity and transfusions, have also been associated with increased mortality.[1213] While the previous seven variables are associated with mortality in our model, pregnancy was not; therefore, the data do not support the need to treat pregnant trauma patients differently, with the notable exception of assessment for fetal or uterine injury.

Greco et al. recommended transporting pregnant trauma patients with a viable fetus to: “a maternity facility… when injuries are neither life-nor limb-threatening… [However, when] the injury is major, the patient should be transferred or transported to the trauma unit or emergency room…”[1] We found no difference in outcomes for pregnant trauma patients based on trauma level designation. Previous reports also found no difference. Distelhorst et al. found no maternal, infant, or fetal mortality difference at trauma-designated hospitals compared to hospitals without a trauma designation.[3] In addition, Distelhorst et al. found similar outcomes at a level 1 or 2 trauma hospitals compared to hospitals with a lower trauma designation.[4] We reaffirm the findings of both previous studies and, consistent with the recommendation by Greco et al., find no reason to recommend diverting pregnant trauma patients to higher designation trauma centers.

We sought to characterize the increase in the burden of hospitalization due to the current recommendation for 24-h observation in patients with possible uterine or fetal injuries. We found pregnant trauma patients were more likely to be admitted for hospitalization after controlling for injury severity. However, they were less severely injured and less likely to be admitted to the ICU. More pregnant women were admitted for a 24-h period of observation than nonpregnant patients in our dataset, which reflects the recommendations from expert guidelines.[1] The result of these recommendations for observation is a very low increased rate of admission (only about a 2% increase) in an already small population. A recent retrospective report by Sert et al. described a nearly 20% rate of obstetric complications in the first 24 h after trauma for pregnant patients.[14] Owattanapanich et al. described an 11% rate of obstetrical complications following MVCs in another retrospective review.[15] Based on our data and the two previous reports, there would be little cost-saving or other societal gains from trying to reduce the number of pregnant patients admitted for observation. The current recommendations for 24-h observation of select patients should remain unchanged as it causes a minimal increase in the burden of care and may detect the occurrence of obstetrical complications during this timeframe.

We attempted to further characterize the injury patterns suffered by pregnant trauma patients to help prepare clinicians for what they may encounter in practice. Previously, Battaloglu et al. described a higher rate of thoracic injuries in pregnant compared to nonpregnant female trauma patients but did not relate injury patterns to mortality.[8] Kissinger et al. reported the incidence of pelvic fractures (7.5%) and HI (9.7%) but did not further characterize injury patterns.[6] Our report adds to the previous studies and is the first to quantify the incidence of CSI among pregnant patients. Nearly one-quarter of all pregnant trauma patients had an identified CSI. Importantly, roughly two-thirds of pregnant trauma patients with a CSI underwent a procedure on hospital day 1, likely requiring anesthesia and airway management. Parturients are recognized to be at an increased risk for difficult or failed airway management and aspiration.[16171819] Cervical spine precautions such as in-line stabilization can impair intubation.[20] Consequently, these patients have two risk factors for difficulty with intubation or other airway-related complications. Multidisciplinary simulation events and mutually created protocols involving trauma surgeons, obstetricians, neonatologists, emergency medicine, and anesthesia providers may help prepare teams.

We elucidated the types and distribution of procedures pregnant trauma patients undergo. We failed to identify any cesarean sections performed, pre or post morbid. The most common procedures involved the patient's extremities. Of particular concern are the procedures these women underwent urgently or emergently on the day of hospital admission. Of the 155 who underwent a procedure on day 1, only 51 were transferred directly to the operating room from the emergency department for care. Some of the remaining 104 procedures were potentially performed in the emergency department. The majority of the intracranial or thoracoabdominal procedures likely occurred in the operating room. There is no way to confirm the location of the other procedures, but it is reasonable to assume there are a significant number of patients who required some form of anesthesia in the emergency department. Anesthetizing parturients can be challenging, especially due to airway concerns and the physiologic changes of pregnancy. General anesthesia (GA) with a rapid sequence intubation (RSI) is recommended.[2122] With such a significant number of procedures occurring outside of the operating room, many procedures may not involve GA with an RSI. Regional anesthesia (RA) may be an option in these cases. In other trauma populations, RA has been shown to be safe and effective, with a wide array of additional benefits.[232425] A regional technique can serve as the sole anesthetic, potentially avoiding GA and significant opioid use.[2627] RA catheters may be used beyond the operating room for postoperative analgesia, thus extending the benefit of an opioid-sparing technique for the pregnant trauma patient. RA may be beneficial given that pregnant patients undergoing GA for nonobstetric surgeries had a higher risk of preterm delivery and lower birth weight.[2829] In many cases, especially with procedures occurring outside of the operating room, anesthesia (analgesia and/or sedation) may be provided by an emergency room physician or other nonanesthesiologists. Training involved clinicians in the core concepts of local anesthetics and regional anesthesia may enhance care for pregnant trauma patients.

Our report has several limitations. The data are retrospective and observational. Our logistic regression can only show associations rather than causality. We had a modest sample size of 569 pregnant patients with limited mortality.[30] Despite these limitations, pregnancy does not seem to confer a higher risk for mortality among trauma patients. Current recommendations regarding similar treatment of trauma regardless of pregnancy status should remain unchanged. Multidisciplinary simulation, jointly crafted protocols, and expanding training in and use of regional anesthesia may be the next steps to improving care for pregnant trauma patients and should be further investigated.

Research quality and ethics statement

This retrospective data analysis was exempted from the Institutional Review Board / Ethics Committee approval (IRB Number: 357097-1). The authors followed applicable EQUATOR Network (http://www.equator-network.org/) guidelines during the conduct of this research project.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.