INTRODUCTION

Trauma is one of the leading causes of morbidity and mortality in the United States (Drury et al., 2021; Mondello et al., 2014; Prin & Li, 2016). In 2019 alone, unintentional injuries were the third leading cause of mortality behind heart disease and cancer. These injuries contributed to 6.1% of all deaths (Centers for Disease Control and Prevention, 2019). Additionally, for patients younger than 65 years, trauma has been reported to be the most significant cause of years of life lost (DiMaggio et al., 2017).

Due to the high prevalence and severity of trauma, it is imperative to identify risk factors that contribute to morbidity and mortality among critically ill trauma patients. One such risk factor is an unhealthy body mass index (BMI) above and below the normal range (Aune et al., 2016). BMI is a calculation of body weight and height used by health care professionals to estimate body fat and categorize individuals as either underweight, normal weight, overweight, or obese (Weir & Jan, 2020). Interestingly, the medical literature has reported that having a BMI that correlates with being underweight, overweight, or obese is positively associated with mortality. Several researchers even describe the association between BMI and mortality as a J-shaped curve, with the highest rate of mortality being among patients with obese and underweight BMI and the lowest rate being among individuals with normal weight BMI (Aune et al., 2016; Kelly et al., 2010; Klatsky et al., 2017). Similarly for morbidity, adiposity associated with obesity can initiate toxic metabolic and inflammatory cascades that can eventually drive the onset of conditions, such as cardiovascular disease, type 2 diabetes (Aune et al., 2016; Dixon, 2010), cancers, stroke, respiratory problems, sleep apnea, hypertension (Centers for Disease Control and Prevention, 2020), and organ failure (Nelson et al., 2012). Based on these findings, which suggest that clinical outcomes may be influenced by BMI, we hypothesized that obese and underweight states, in comparison to normal weight BMI, may be associated with a higher risk for morbidity and mortality after sustaining trauma. Therefore, using the American College of Surgeons Trauma Quality Improvement Program (ACS TQIP) database, we studied the association between BMI and morbidity and mortality among all patients who sustained either blunt or penetrating trauma.

Objective

This study aimed to analyze the association between BMI and morbidity and mortality after trauma.

KEY POINTS

* Trauma patients with overweight and obese body mass indexes, in comparison to normal weight body mass index, were independently associated with a lower risk of mortality.

* Patients with underweight, overweight, obese, severely obese, and morbidly obese body mass indexes, in comparison to normal weight body mass index, were independently associated with a higher odds of developing a complication, irrespective of preexisting comorbidities, injury severity score, and mechanism of injury.

* To prevent inhospital complications, vigilant care should be maintained for both underweight and obese patients after trauma.

METHODS

After gaining approval from our Institutional Review Board, we acquired the 2016 ACS TQIP database for this study. As per the TQIP Participant Use File (PUF) User Manual, this database includes all patients 16 years and older admitted to a Level I or II trauma center in 2016 for blunt or penetrating trauma. Additionally, patients with an abbreviated index score (AIS) 05/08 injury with severity of greater than or equal to 3 in body regions 1-8 were included. However, patients with either severe burns specifically disclosed in the TQIP PUF User Manual, advanced directives to withhold life-sustaining interventions, an emergency department discharge disposition to "home without services," "home with services," "transfer to another hospital," "other," or "left against medical advice," and/or a combination of vital signs that are specifically disclosed in the TQIP PUF User Manual were excluded.

After collating all patients from the ACS TQIP database, we stratified them into six BMI groups: underweight (UW, BMI <18.50 kg/m²), normal weight (NW, BMI 18.50-24.99 kg/m²), overweight (OW, BMI 25.00-29.99 kg/m²), obese (OB, BMI 30.0-34.99 kg/m²), severely obese (SO, BMI 35.00-39.99 kg/m²), and morbidly obese (MO, BMI >=40.00). These align with the current classifications from the National Institutes of Health and the Centers for Disease Control and Prevention (National Institutes of Health, n.d.; Weir & Jan, 2020).

In univariate analysis, we compared the NW control group to the five BMI groups (UW, OW, OB, SO, and MO) for differences in demographic factors, preexisting comorbidities, injury severity score (ISS), length of stay (LOS), mechanism of injury (MOI), posttrauma complications, and mortality. Not all of the comorbidities and complications that were analyzed are shown in our results. However, a complete list of these parameters can be referenced in Supplemental Digital Content Appendix 1 (available at: http://links.lww.com/JTN/A45).

Next, we performed multivariate logistic regression analyses to identify risk factors for developing a complication after trauma. All significant variables from our univariate analyses were included as independent covariates in the multivariate logistic regression analyses. Each consisted of patients from the NW control group and one of the non-NW BMI groups (UW, OW, OB, SO, and MO). Developing a complication was the dependent variable, and it consisted of the sum of complications listed in Supplemental Digital Content Appendix 1 (available at: http://links.lww.com/JTN/A45). These analyses were replicated using the same independent variables to identify risk factors for mortality, our second dependent variable, after trauma. Of note, odds ratios (ORs) for covariates other than BMI (comorbidities, MOI, demographic factors, etc.) were not shown in our data table, as we primarily sought to emphasize the independent association between BMI and morbidity and mortality.

Statistical Analysis

Microsoft Excel was used for data stratification and statistical analysis was conducted using Minitab 17. [chi]² and Fischer's exact tests were used to analyze categorical variables, and we utilized Student's t test to analyze continuous variables in univariate analysis. Of note, in univariate analysis, weighted averages were not used and all groups were equal. For both univariate and multivariate analyses, the significance level was p < .05.

RESULTS

In total, 257,726 patients were included in our study population. There were 10,967 (4.26%) UW patients, 98,690 (38.29%) NW patients, 82,491 (32.00%) OW patients, 39,510 (15.33%) OB patients, 15,201 (5.90%) SO patients, and 10,867 (4.22%) MO patients. Overall, 61.5% of the patients were male, and the M (SD) age, in years, was 54.06 (21.74). The UW population was the oldest and was significantly older than the NW cohort (M [SD], 59.80 [23.80] years vs. 53.20 [23.70] years, respectively; p < .001). Additionally, the UW cohort had the lowest average ISS, which was significantly lower than that of the NW cohort (M [SD], 13.82 [8.83] vs. 15.10 [8.21], respectively; p < .001). In contrast, patients with SO and MO BMI had the highest ISS, which was significantly greater than that of NW patients (M [SD], 15.55 [8.31] vs. 15.10 [8.21], respectively; p < .001) and (M [SD], 15.55 [8.37] vs. 15.10 [8.21], respectively; p < .001). Corresponding with these findings, high-energy MOI, such as motor vehicle trauma, was more prevalent among SO and MO patients. In contrast, low-energy MOI, such as falls, was more prevalent among UW patients (data not shown). Furthermore, UW patients had the lowest rate of penetrating trauma, whereas NW patients had the highest rate (4.39% vs. 8.46%, p < .001, Table 1).

Table 1. Patient Characteristics, Comorbidities, and Complications

Patients with a BMI above and below NW BMI had more preexisting comorbidities than NW patients (Table 1); 82.01% of the UW cohort had at least one comorbidity, which was significantly more than that of the NW cohort (p < .001, Table 1). Likewise, MO patients had significantly more comorbidities than the NW population (78.79% vs. 73.88%, p < .001, Table 1). More specifically, diabetes mellitus and hypertension were significantly higher in the OW and obese cohorts than in the NW group.

Analysis of complications showed that patients with OW, OB, SO, and MO BMI developed significantly more posttrauma complications than NW patients (Table 1). The incidence of complications among NW patients was 18.19%, which increased to 23.08% among MO patients (p < .001, Table 1). Specific complications, such as acute kidney injury, deep vein thrombosis, decubitus ulcer, and admission to the intensive care unit, were higher among OW, OB, SO, and MO patients (p < .05, Table 1). Interestingly, UW patients did not develop more complications than NW patients despite having significantly more comorbidities (Table 1). Corresponding with these findings, no differences in LOS were observed between NW and UW patients (M [SD], 7.38 [8.83] vs. 7.56 [9.46], respectively; p = .051). However, OW, OB, SO, and MO patients had significantly longer hospital stays (p < .001). The longest hospitalization period was observed among MO patients, which on average was about 2 days longer than that of NW patients (M [SD], 10.20 [12.20] vs. 7.56 [9.46], respectively; p < .001).

MO patients had the highest mortality rate, followed by UW and SO patients (p < .05). Interestingly, the rates of mortality among OB and OW patients were similar to that of NW patients, and there was no statistical difference (Figure 1).

Figure 1. (A) Figure compares the rate of mortality in each body mass index (BMI) cohort to the NW group. (B)

Multivariate Logistic Regression for Developing a Complication and Mortality

Irrespective of preexisting comorbidities, gender, ISS, high-energy MOI, or age, UW, OW, OB, SO, and MO BMI were independent risk factors for developing at least one complication after trauma. More specifically, MO and SO BMI were associated with 36.2% and 19.6% higher risk of developing a complication than NW patients [MO (OR 1.362; 95% confidence interval [CI], 1.293, 1.435; p < .001)] and [SO (OR 1.196; 95% CI 1.143, 1.250; P < .001)].

For mortality, UW BMI alone was associated with a 19.9% higher risk than NW BMI (OR 1.199; 95% CI, 1.076, 1.337; p = .001). In contrast, OW and OB BMI were associated with 9.6% and 10.5% lower probability of mortality, respectively (p < .001, Table 2).

Table 2. Multivariate Logistic Regression

DISCUSSION

Univariate analysis of the 2016 ACS TQIP database demonstrated that the mortality rate was the lowest among patients with NW BMI, whereas it was the highest among patients with MO BMI, followed by patients with UW BMI. These statistics illustrate a J-shaped curve, which confirms other studies that reported that patients with a high or a low BMI have a higher risk of mortality than NW patients (Antonopoulos & Tousoulis, 2017; Hartrumpf et al., 2017). Correspondingly, it opposes the idea that the relationship between BMI and adverse clinical outcomes is linear. Our multivariate logistic regression confirmed these findings, as UW BMI was associated with a 19.9% higher risk of mortality than NW BMI. This is plausible as UW patients have been associated with comorbid conditions, such as organ dysfunction, heart failure, cancers, malnutrition, and less fat and muscle stores (Hartrumpf et al., 2017). After trauma, these adverse health states can exacerbate these patients' mortality risk. In contrast, multivariate logistic regression demonstrated that OW and OB BMI alone were independently associated with a lower mortality risk than NW BMI. These findings challenge a previous publication from a single institution that reported no difference in the risk of mortality between underweight, normal weight, and obese patients (Drury et al., 2021). This inconsistency is likely due to the fact that BMI is not the most accurate measure of overall health status for OW and OB patients, as it does not distinguish between adiposity and lean mass (Centers for Disease Control and Prevention, 2011; Romero-Corral et al., 2008; Stefan et al., 2013). Patients who are metabolically fit and have more muscle mass may be classified within the same OW or OB BMI category as others who may be metabolically unfit and have more adipose tissue (Romero-Corral et al., 2008; Stefan et al., 2013). These differences in body mass composition and overall fitness affect morbidity and mortality. The risk of mortality among OB and OW patient populations may be lower for metabolically fit patients and higher for unfit patients. Therefore, it is comprehensible as to why OW and OB BMI were associated with a lower risk of mortality in our study and why there are mixed conclusions regarding the effect of OW and OB BMI in the medical literature.

Additionally, OW, OB, SO, and MO patients presented with significantly more preexisting comorbidities and subsequently developed more complications than NW patients. This is understandable as the pro-inflammatory responses associated with obesity can contribute to comorbid conditions, such as cardiovascular disease and diabetes (Ellulu et al., 2017) and subsequently, trauma-related organ dysfunction (Liu et al., 2013). Additionally, because complications were more prevalent among OW, OB, SO, and MO patients than in NW patients, it is likely that the management of these complications contributed to their longer hospital stays (Mondello et al., 2014).

Moreover, the average ISS increased with BMI, which complements our findings that UW patients sustained more low-energy trauma and OW, OB, SO, and MO patients sustained more high-energy trauma. The UW cohort was also the oldest, suggesting that this population may have more frail patients who sustained injuries from falls. In contrast, MO patients were the youngest and sustained more injuries from high-energy mechanisms, such as motor vehicle trauma. In our multivariate logistic regression analyses, high ISS and high energy MOI were risk factors for morbidity and mortality. However, despite the independent effects of these covariates, OW, OB, SO, and MO BMI were still notable risk factors for morbidity.

This study comes with some limitations. Because this is a retrospective study, our analysis was reliant on strict data points from the ACS TQIP database. As a result, there were patients with negative or missing values for numerical data points who had to be excluded from our analysis. More importantly, BMI may not be the most accurate predictor of overall health status because it does not account for the difference between adiposity and muscle mass. Future studies accounting for body mass composition may be warranted to obtain more definitive conclusions regarding obesity and mortality after trauma.

CONCLUSION

Trauma patients with UW, OW, SO, and MO BMI, as opposed to NW BMI, were independently associated with a higher risk of developing complications, irrespective of preexisting comorbidities and ISS. In contrast, both OW and OB BMI were independently associated with a lower probability of posttrauma mortality than NW BMI. The results from our research indicate that a complex relationship exists between OW and OB BMI and clinical outcomes after trauma. We believe that the trauma community would benefit from prospective studies exploring the association between obesity and morbidity and mortality, while controlling for body mass composition.

REFERENCES