Obesity May Not be Protective in Abdominal Stab Wounds.

CONTEXT: Current protocols for the management of abdominal stab wounds were established based on retrospective data from prior decades. Few have investigated whether higher body mass index (BMI) affects outcomes after these injuries. AIM: The aim was to determine the effects of obesity on outcomes in abdominal stab wound patients. SETTING AND
DESIGN: This was a retrospective cohort study at a Level I university-associated trauma center in the United States.
MATERIALS AND METHODS: We reviewed medical records of 100 adult patients admitted to our trauma center with abdominal stab wounds. Demographics, types of internal organ injury, gastrointestinal (GI) resection and repair, mortality, length of hospital stay (LOS), units of blood transfused within 24 h of admission, need and indications for exploratory laparotomy, surgical site infections (SSI), and need for re-operation were compared between obese and nonobese patients. STATISTICAL ANALYSIS: Categorical and continuous outcome variables were compared between the two groups using Chi-squared and independent-samples t-tests, respectively. BMI was evaluated as a predictor of outcomes using univariate and multivariate logistic regression.
RESULTS: Records of 100 adult abdominal stab wound patients were reviewed. Twenty-five patients were obese. The obese group was older (38.76 vs. 31.23, P = 0.018). Rates of therapeutic laparotomy were similar between obese and nonobese patients (20 [80.00%] vs. 64 [85.33%]). Obesity was associated with longer LOS (9.6 vs. 6.5, P = 0.026). In the multivariate analysis, increasing BMI was an independent predictor of need for GI resection (odds ratio: 1.10 [1.02-1.18], P = 0.018). One patient from the obese group died.
CONCLUSIONS: Obese patients with abdominal stab wounds have longer LOS than nonobese patients. Increasing BMI was an independent predictor of need for GI resection.

INTRODUCTION

In managing patients with abdominal stab wounds, most of the institutions follow protocols based largely on retrospective data from prior decades, which do not consider patient characteristics such as obesity. One-third of the United States population are now obese, and two-thirds are obese or overweight. Protocols derived from data of prior decades may no longer be applicable or cost-effective for all patients.

Trauma surgeons have speculated whether obesity may be protective in trauma due to increased soft tissue protecting the abdominal cavity. In blunt trauma, obesity is an independent risk factor for mortality, and obese patients have higher rates of multiorgan failure, longer hospital stays, longer periods of mechanical ventilation, pulmonary complications, and higher rates of other in-hospital complications.[12345] Little data exist on the effect of obesity in penetrating trauma, though one review found that increased body mass index (BMI) is associated with lower incidence of severe injury and need for operation.[6]

With the goal of updating care guidelines, we conducted a retrospective chart review on nonballistic penetrating injuries in obese patients. We assessed the relationship between BMI and outcome measures including length of hospital stay (LOS), internal organ injury, need for operation, surgical complications, and mortality. We hypothesized that obesity and higher BMI would be associated with decreased need for operation but increased rate of complications such as surgical site infections.

MATERIALS AND METHODS

Following IRB approval, a retrospective chart review was conducted at our Level I university-associated trauma center. Records of patients admitted with penetrating torso wounds over a 4-year period were identified from a database of trauma patients. To increase the uniformity of the study population, only patients aged 18 years and older were included. The group was further narrowed to patients who had received stab wounds only (rather than gunshot wounds) and who had a BMI recorded in the chart.

We then reviewed individual medical records to select patients that were found to have isolated anterior abdominal stab wounds, as all patients with this injury pattern were admitted and managed per trauma division protocol based on prior evidence published from our institution.[7] This was to eliminate a possible source of selection bias if one group was admitted rather than discharged from the emergency department at a disproportionately high rate. The anterior abdomen was defined as the area bounded by the costal margins superiorly, the inguinal creases and pubis inferiorly, and between the anterior axillary lines laterally. According to institutional protocol, patients with hemodynamic instability (systolic blood pressure ≤90 mmHg), evisceration, or peritonitis underwent immediate exploratory laparotomy. Patients without these findings were admitted for serial abdominal examinations.

Patient charts were then reviewed in detail for patient age, sex, and race. Our primary outcome variables were mortality, need for exploratory laparotomy, and rate of therapeutic laparotomy. The secondary outcomes were indications for exploratory laparotomy if it was performed, occurrences of visceral and solid organ injuries, LOS, mean units of blood transfused within 24 h of admission, surgical site infection (SSI) occurrence, need for gastrointestinal (GI) resection and repair, and need for reoperation. A therapeutic laparotomy was defined as a laparotomy in which one of the following interventions was performed: control of hemorrhage from the mesentery or a solid organ by repair or resection; repair or resection of any part of the stomach, small bowel, or colon; diaphragm repair; repair, shunting, or grafting of a vascular injury; or repair of traumatic hernia.

Patients were classified as obese (BMI ≥30) and nonobese (BMI <30) in accordance with universally used classifications established by the National Heart, Lung, and Blood Institute.[8]

Statistical analysis

The sample size was chosen based on available resources, and no a priori test of power was performed. Descriptive statistics were performed on all variables using Stata Statistical Software: Release 14 (StataCorp., 2015, College Station, TX, USA). Categorical outcome variables were compared between the two groups using Chi-squared tests. In cases where cell frequencies were <5, Fisher's exact test was used rather than the Chi-squared test. Continuous outcome variables were compared using independent-samples t-tests. Then, BMI was evaluated as a predictor for adverse outcomes using univariate logistic regression. Multivariate logistic regression was also used to determine whether BMI was an independent predictor of these outcomes.

RESULTS

Five hundred and sixteen adult patients were admitted with traumatic penetrating torso wounds. Of these, 382 had a BMI recorded at the time of admission. Two hundred and sixty-one of these suffered from stab wounds only; the remainder had firearm injuries. One hundred patients were found to have isolated anterior abdominal stab wounds on review of their medical records. Twenty-five patients were obese and 75 were nonobese. The obese group was older (38.76 vs. 31.23, P = 0.018) and had a higher BMI (35.54 vs. 23.85, P = 0.000). There was no difference in sex or race characteristics between the two groups [Table 1].

Table 1

Demographics of abdominal stab wound patients

Variable	Observations	Overall	BMI <30	BMI ≥30	P
n	100	100	75	25
Age	100	33.110 (13.910)	31.227 (13.618)	38.760 (13.492)	0.018*
BMI	100	26.768 (6.582)	23.845 (3.587)	35.535 (5.681)	0.000*
Male sex (%)	100	88 (88.000)	67 (89.333)	21 (84.000)	0.477
White race (%)	100	45 (45.000)	30 (40.000)	15 (60.000)	0.082

For numeric data, the mean of each variable is shown, and the standard deviation is in parentheses. For categorical data, the number of patients in each category is shown, and the percentage of patients in each group falling into that category is in parentheses. *Result statistically significant with P<0.05. BMI: Body mass index

Twenty (80.00%) obese patients and 64 (85.33%) nonobese patients received an exploratory laparotomy (P). Of these 84 patients who received an exploratory laparotomy, the obese and nonobese groups had similar indications for laparotomy: hemodynamic instability (5 [25.00%] vs. 9 [14.06%]), peritonitis (5 [25.00%] vs. 15 [21.88%], P = 0.77), evisceration (6 [30.00%] vs. 19 [29.69%]), concerning imaging or diagnostic peritoneal lavage (DPL) finding (2 [10.00%] vs. 11 [17.19%]), and concerning examination (2 [10.00%] vs. 8 [12.50%]) [Table 2]. Seventeen (85.00%) obese patients and 58 (90.63%) nonobese patients had a therapeutic laparotomy (P 0.478). There were no statistically significant differences in the types of intra-abdominal, visceral, and solid organ injuries or need for GI repair or resection between nonobese and obese patients [Table 2].

Table 2

Outcomes in abdominal stab wound patients

Variable	Observations	Overall	BMI <30	BMI ≥30	P
Systolic blood pressure in ED	99	123.283 (23.853)	124.160 (23.807)	120.542 (24.299)	0.521
Injury severity score	100	8.850 (4.865)	8.867 (5.076)	8.800 (4.262)	0.953
Number of intraoperative transfusions	99	0.071 (0.435)	0.093 (0.498)	0.000 (0.000)	0.362
Number of OR visits	100	0.970 (0.540)	0.987 (0.507)	0.920 (0.640)	0.596
GI injury (%)	100	68 (68.000)	50 (66.667)	18 (72.000)	0.621
GU injury (%)	100	2 (2.000)	2 (2.667)	0 (0.000)	1.000†
Vascular injury (%)	100	6 (6.000)	5 (6.667)	1 (4.000)	1.000†
Diaphragm injury (%)	100	6 (6.000)	5 (6.667)	1 (4.000)	1.000†
Solid organ injury (%)	100	26 (26.000)	19 (25.333)	7 (28.000)	0.792
Spleen injury	100	8 (8.000)	6 (8.000)	2 (8.000)	1.000†
Liver injury	100	17 (17.000)	12 (16.000)	5 (20.000)	0.759†
Kidney injury	100	1 (1.000)	1 (1.333)	0 (0.000)	1.000†
Traumatic hernia repair (%)	100	43 (43.000)	34 (45.333)	9 (36.000)	0.414
Cholecystectomy (%)	100	3 (3.000)	1 (1.333)	2 (8.000)	0.153†
Mesenteric, omental appendage repair (%)	100	12 (12.000)	10 (13.333)	2 (8.000)	0.725†
DPL (%)	100	12 (12.000)	12 (16.000)	0 (0.000)	0.034†
Positive DPL (%)	100	6 (6.000)	6 (8.000)	0 (0.000)	0.332†
GI resection (%)	100	1 (1.000)	0 (0.000)	1 (4.000)	0.250†
GI repair (%)	100	17 (17.000)	11 (14.667)	6 (24.000)	0.282
Exploratory laparotomy (%)	100	84 (84.000)	64 (85.333)	20 (80.000)	0.529
Indications for exploratory laparotomy (%)
Hemodynamic instability	84	14 (16.667)	9 (14.062)	5 (25.000)	0.305†
Peritonitis	84	19 (22.619)	14 (21.875)	5 (25.000)	0.766†
Evisceration	84	25 (29.762)	19 (29.688)	6 (30.000)	0.979
Imaging/DPL finding	84	13 (15.476)	11 (17.188)	2 (10.000)	0.724†
Concerning examination	84	10 (11.905)	8 (12.500)	2 (10.000)	1.000†
Therapeutic exploratory laparotomy	84	75 (89.286)	58 (90.625)	17 (85.000)	0.478

For numeric data, the mean of each variable is shown, and the standard deviation is in parentheses. For categorical data, the number of patients in each category is shown, and the percentage of patients in each group falling into that category is in parentheses. Result statistically significant with P<0.05, †Evaluated using Fisher’s exact test given small frequency values. BMI: Body mass index, GI: Gastrointestinal, DPL: Diagnostic peritoneal lavage, ED: Emergency department, GU: Genitourinary, OR: Odds ratio

Obesity was associated with longer LOS (9.6 days vs. 6.5 days, P 0.026) [Table 3]. Units of blood transfused in the first 24 h, SSI, and need for reoperation were similar in the two groups. One obese patient, and no nonobese patients, died (P = 0.250). Twelve (16.00%) of nonobese patients and no obese patients received a DPL (P = 0.034). Higher BMI was predicted decreased likelihood of receiving a DPL (odds ratio [OR]: 0.815 [0.710–0.979]), increased likelihood of GI resection (OR: 1.10 [1.02–1.18], P = 0.018), and longer LOS (Coefficient 0.243 [0.067–0.419], P = 0.007). In a multivariate regression model that included BMI, age, race, and sex, BMI was an independent predictor of higher rates of GI resection (OR: 1.09 [1.01–1.19], P = 0.020).

Table 3

Outcomes in abdominal stab wound patients

Variable	Observations	Overall	BMI <30	BMI ≥30	P
Blood 1st 24 h (units)	100	0.790 (3.003)	0.960 (3.431)	0.280 (0.737)	0.329
LOS	100	7.300 (5.984)	6.533 (3.761)	9.600 (9.836)	0.026
Death (%)	100	1 (1.000)	0 (0.000)	1 (4.000)	0.250†
SSI (%)	100	7 (7.000)	6 (8.000)	1 (4.000)	0.677†
Reoperation (%)	100	12 (12.000)	9 (12.000)	3 (12.000)	1.000†

For numeric data, the mean of each group is shown with the standard deviation in parentheses. For categorical data, the number of patients in each group is shown, with the percentage out of the total number of patients in each group (overall, nonobese, or obese). Result statistically significant with P<0.05, †Evaluated using Fisher’s exact test given small frequency values. SSI: Surgical site infection, BMI: Body mass index, LOS: Length of hospital stay

CONCLUSION

In our review of 100 patients with stab wounds to the abdomen, obesity was not shown to be protective by our primary outcome measures of mortality, need for exploratory laparotomy, and rate of therapeutic laparotomy. Of our secondary outcome variables, obesity was associated with longer hospital stays. Higher BMI was predictive of higher GI resection rates and longer LOS. Otherwise, outcomes were similar among the two groups.

Most of the investigations surrounding body composition and trauma outcomes have focused on the systemic ramifications of obesity, including immunosuppression, respiratory compromise, and decreased cardiopulmonary reserve, rather than the anatomic impact of obesity on immediate injury patterns. Rates of massive transfusion and multiple organ failure have been shown to be higher in injured obese individuals.[910] In a prospective study of over 1000 critically ill trauma patients, obese patients had a markedly increased risk of blood, urinary tract, and respiratory infections, intensive care unit admissions, and death after controlling for age and ISS score.[10]

However, in each of these studies, blunt trauma patients comprised around 70% of the study population. The specific relationship between obesity and penetrating trauma has garnered far less attention. Yet, understanding the role of body composition in penetrating injury patterns is important. projectiles or instruments generally must enter the abdominal cavity to cause significant injury. Urban Level I trauma centers in the United States see hundreds of patients with abdominal gunshot and stab wounds annually. Determining whether body composition affects rates of peritoneal violation and injury patterns will guide future management and resource allocation.

Bloom et al. recently performed the first dedicated review of obesity and abdominal stab wounds.[6] In 249 patients stratified into four BMI groups, patients in higher BMI groups had fewer visceral injuries and injuries that required therapeutic operation. They concluded that thin patients lacking both subcutaneous abdominal wall and intraperitoneal adipose tissue should be observed vigilantly for intra-abdominal injury, whereas obese patients may more safely be managed by observation alone. They do not recommend changes to existing protocols. Our results showed a lack of outcomes differences that preclude safely suggesting alterations to current practice guidelines.

Our trauma division has followed the same guidelines for the evaluation and treatment of patients with anterior abdominal stab wounds for several years. These guidelines were derived from data gathered at our institution in 2004 and from the Western Trauma Association (WTA) published in 2011.[711] At our center, Tsikitis et al. in 2004 examined 77 patients with anterior abdominal stab wounds. Those with hemodynamic instability, evisceration, or peritonitis were taken for immediate laparotomy, whereas 30 patients who did not require immediate operation were admitted for 24 h and received serial physical examinations and laboratory monitoring every 8 h. These 30 patients were discharged without event, and none required readmission.

In 2011, the WTA validated their algorithm for the management of abdominal stab wounds in a multicenter study. Patients were taken for immediate laparotomy for the same indications described above. They also concluded that serial clinical examination alone was safe for abdominal stab wound patients who did not require an immediate operation. Interestingly, the authors suggest a computed tomography scan in their algorithm for patients with a BMI above 30 but do not cite references to support the practice. Eighty-seven percent of patients that underwent immediate laparotomy had a therapeutic operation, comparable to results in our series.

We are unwilling to change our practice patterns based on ours or Bloom's data and continue to support the guidelines based on the two above studies. A thorough cost analysis of a larger number of patients would be required to increase the power of our study to detect smaller clinical differences and make changes to current protocols.

The most important limitation of this study was the small sample size, which precluded further stratification of BMI groups and detection of subtle outcome differences. In addition, as a retrospective study, the data are subject to selection bias. Despite institutional protocol, obese individuals with small stab wounds may more frequently have been discharged from the emergency department at the discretion of an attending trauma surgeon, while we reviewed only admitted patients. We also acknowledge that BMI may not accurately represent the body composition and distribution of soft tissues, which may be more accurately measured by abdominal wall thickness. Increased depth of subcutaneous fat may decrease the incidence of abdominal wounds in blunt trauma, while intra-abdominal fat may not. Obtaining measurements of abdominal wall thickness from imaging or physical examination may be worthwhile in the future to further clarify whether increased abdominal wall thickness is, in fact, protective in abdominal stab wounds.

In summary, obesity does not appear to be protective for abdominal stab wound patients in our series. We suggest continued adherence to established and published guidelines that stress serial clinical examinations for those patients that do not require immediate laparotomy. Future work with larger sample sizes through multicenter collaboration, and more accurate measures of body composition, is required.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.