Injury 54 (2023) 469–480

Contents lists available at ScienceDirect

Injury
journal homepage: www.elsevier.com/locate/injury

Review

Obesity and outcomes in trauma - a systematic review and

meta-analysis
Paul M Cromwell a,∗, Ian S Reynolds a, Helen M Heneghan a, Simon M Glasgow b
a
Department of Surgery, St. Vincent’s University Hospital, Dublin, Ireland
b
Centre for Trauma Sciences, Blizard Institute of Cell and Molecular Biology, Queen Mary University of London, London, United Kingdom

a r t i c l e i n f o a b s t r a c t

Article history: Background: The physiological abnormalities relating to obesity and metabolic syndrome can contribute
Accepted 23 October 2022 to worse outcomes following trauma especially in class 2 and 3 obesity. The aim of this systematic review
was to determine whether patients with a higher class of obesity who suffer traumatic injury have a
Keywords: higher risk of worse outcomes including in-hospital mortality than normal-weight patients.
Trauma Methods: A systematic search of MEDLINE, EMBASE, CENTRAL, Web of Science and CINAHL was per-
Obesity formed for studies that reported a comparison of in-hospital obesity-related outcomes against normal-
Body mass index weight individuals aged 15 years and older following trauma. Single or multiple injuries from either blunt
WHO obesity classes and/or penetrating trauma were included. Burn-related injuries, isolated head injury and studies focusing
Trauma surgery
on orthopaedic related perioperative complications were excluded.
Results: The search yielded 7405 articles; 26 were included in this systematic review. 945,511 patients
had a BMI>30. A random-effects meta-analysis was performed for analysis of all four outcomes. Patients
with class 3 obesity (BMI>40) have significantly higher odds of in-hospital mortality than normal-BMI in-
dividuals following blunt and penetrating trauma (OR, 1.75; 95% CI, 1.39-2.19, p=<0.0 0 0 01), significantly
longer hospital LOS (SMD, 0.23; 95% CI, 0.21-0.25; p<0.0 0 0 01) and significantly longer ICU LOS (SMD,
0.19; 95% CI, 0.12-0.26; p<0.0 0 01). In contrast, studies that examined blunt and penetrating trauma and
classified obesity with a threshold of BMI>30 found no significant difference in the odds of in-hospital
mortality (OR, 0.94; 95% CI, 0.86-1.02, p=0.13).
Conclusions: There is a higher risk of in-hospital mortality in patients living with class 3 obesity following
trauma when compared with individuals with normal BMI. The management of patients with obesity is
complex and trauma systems should develop specific weight related pathways to manage and anticipate
the complications that arise in these patients.
Systematic review registration number
PROSPERO registration: CRD42021234482
Level of Evidence: Level 3
© 2022 Elsevier Ltd. All rights reserved.

Background from comorbidities that may affect their physiological response to

trauma, including diabetes, cardiovascular and respiratory disease
The worldwide prevalence of obesity has tripled in the last [3–5]. The impact this population has on trauma systems is sig-
three decades. More than 1.9 billion adults aged over 18 years nificant and only likely to increase as the global obesity epidemic
are overweight, of these, over 650 million are classified as obese worsens [6].
[1]. Obesity is a chronic complex disease quantitatively defined as At present, there remains clinical equipoise regarding the effect
those with a body mass index (BMI) of greater than 30kg/m2 [2]. of obesity on patient outcomes following trauma. The last system-
These individuals have a 48% higher annual risk of injury com- atic review of the subject in 2013 identified significantly higher
pared to normal-BMI individuals and are more likely to suffer mortality and intensive care unit (ICU) length of stay (LOS) in
obese versus non-obese patients [7]. Despite this, several studies
suggest an ‘obesity paradox’, attributing a protective influence of
∗
Corresponding author. excess adipose tissue through a so-called ‘cushion effect’ leading to
E-mail addresses: paulcromwell@rcsi.ie (P.M. Cromwell), reynoli@tcd.ie (I.S. higher survival rates among obese compared to non-obese patients
Reynolds), Helen.heneghan@svhg.ie (H.M. Heneghan), s.glasgow@nhs.net (S.M. Glas- [8–11]. Although BMI classes were introduced by the World Health
gow).

https://doi.org/10.1016/j.injury.2022.10.026
0020-1383/© 2022 Elsevier Ltd. All rights reserved.
P.M. Cromwell, I.S. Reynolds, H.M. Heneghan et al. Injury 54 (2023) 469–480

Table 1 the following criteria. Many adult trauma centres admit adoles-
World Health Organisation nutritional status. Body Mass Index
cents and analyse those individuals as part of their studies on their
(BMI) calculated by a patients weight in kilograms divided by
the square of the persons height in metres1 . datasets. Studies were required to be published in English from
2010 onwards. The preceding ten years were chosen to identify
BMI (kg/m2 ) Nutritional Status
studies that were reporting outcomes in patients treated within a
Below 18.5 Underweight modern trauma system. Studies of patients admitted to the hospi-
18.5-24.9 Normal BMI tal with blunt or penetrating trauma with isolated or multiple in-
25-29.9 Pre-obesity
juries were included. All articles were required to report a compar-
30-34.9 Obesity class 1
35-39.9 Obesity class 2 ison with either non-obese individuals or normal-BMI individuals.
Above 40 Obesity class 3 Only randomised control trials, prospective observational and ret-
1
Adapted from World Health Organisation. Obesity and Over-
rospective observational studies on humans carried out on a civil-
weight [1]. ian population were included. Studies from the previously pub-
lished meta-analysis were included, but only if they met the in-
clusion and exclusion criteria for our study [12]. Isolated confer-
Organisation (WHO) over two decades ago, many older studies ence abstracts, case reports and case studies were excluded, as
have not used this method of stratification to investigate outcomes were studies that did not report BMI, BMI derivable data or at least
(Table 1)(12). More recent studies have utilised this to show signif- one of the outcome measures of interest. Studies analysing burn-
icant effects in outcomes when considering higher classes of obe- related injuries or isolated head injury were excluded, as were
sity [13–15]. studies looking at specific orthopaedic related perioperative com-
These additional data and more granular analysis permitted plications.
by the WHO obesity class system suggests the need for an up-
dated systematic review to further analyse the impact of obe-
sity in trauma. We hypothesise that the “obesity paradox” is in- Outcomes and data analysis
correct for class 2 and class 3 obesity following traumatic in-
jury and that these groups may have been under-represented in The primary study outcome measure was all-cause in-hospital
previous studies. The aim of this systematic review and meta- mortality. Secondary outcome measures included duration of me-
analysis is therefore to critically analyse the available literature to chanical ventilation in ICU and overall ICU and hospital LOS. Addi-
determine whether obese patients of differing BMI classes have tional data including reported complications (acute kidney injury,
higher in-hospital mortality following traumatic injury compared organ failure, acute respiratory distress syndrome and lower respi-
to those of a normal BMI (18.5-24.9). The secondary aims of the ratory tract infection), study characteristics, ISS score and mech-
study are to evaluate whether class of obesity is associated with anism of injury were also collected to allow for further sub-
prolonged mechanical ventilation in ICU and longer hospital and analysis of the data as required. All collected data were organ-
ICU LOS. ised for synthesis and analysis into the BMI categorisation used
by the individual study. Studies that categorised obesity as simply
Methods BMI>30 were analysed separately from those that used the WHO
BMI classification (Table 1). The studies were further separated by
A systematic review of the available literature was performed mechanism of injury to compare outcomes in an attempt to off-
using the preferred reporting items for systematic review and set heterogeneity between results. The mechanism of injury cate-
meta-analysis (PRISMA 2020) statement and meta-analyses of ob- gories selected included blunt and penetrating, blunt alone, pen-
servational studies in epidemiology (MOOSE) checklist [16,17]. The etrating alone, thoracic blunt and penetrating and thoracic blunt
study protocol was registered on the PROSPERO database in March alone.
2021 (CRD42021234482) [18]. The meta-analysis was performed using RevMan 5 review man-
agement software (Version 5.4, Cochrane Collaboration, 2020) [19].
A random-effects model was applied to all analyses as differ-
Search strategy ences in inclusion and exclusion criteria, type of trauma, lack of
BMI data in clinical databases and variations in the practice of
The primary search for published literature was performed in trauma care can result in significant clinical heterogeneity between
MEDLINE, EMBASE, CENTRAL and CINAHL. Unpublished work and the studies. Standardised Mean Difference (SMD) was utilised as
other sources of grey literature were identified using Web of Sci- the effects measure for appropriate continuous variables. Statisti-
ence and the search was further expanded through its citation cal heterogeneity was assessed using the I2 statistic. An I2 value
search function and interrogation of the reference lists from core between 50-75% indicated substantial heterogeneity and a value
articles. above 75% indicated considerable heterogeneity [20]. Statistical
The literature search was performed using the MeSH headings significance was accepted at a p-value <0.05 unless otherwise
‘Obesity’, ‘Body Mass Index’, ‘Wounds and Injuries’ and ‘Accidents’ stated.
and the following keyword combinations: (‘obesity’ OR ‘obese’) The measure of effect was reported as odds ratios and 95% con-
AND (‘Body Mass Index’ OR ‘BMI’) AND (‘Trauma’ OR ‘Injur∗ ’). Two fidence intervals. Two people independently performed a risk of
researchers performed the search independently in January 2021. bias assessment using a risk of bias tool (MINORS) [21]. The low
Citations were exported to Mendeley Desktop v1.19.8 (Elsevier, risk of bias retrospective study that compares groups can achieve a
London UK). Duplicate papers were removed before the remaining maximum score of 24. The MINORS risk of bias tool was also used
titles and abstracts were screened for eligibility. Data was captured to perform sensitivity analysis to assess how the studies with the
from all included studies using paper-based data collection forms. highest risk of bias altered the combined effect. Funnel plots were
used where appropriate to display publication bias. The Newcas-
Inclusion and exclusion criteria tle OTTAWA Scale was used to assess the quality of the included
non-randomised studies. Central tendency for continuous data was
We included all clinical studies that reported obesity-related measured using medians and inter-quartile range (IQR) to account
outcomes following trauma in patients >/-15 years if they met for the non-parametric distribution of data common to this patient

470
P.M. Cromwell, I.S. Reynolds, H.M. Heneghan et al. Injury 54 (2023) 469–480

Fig. 1. PRISMA 2020 flow chart

population and the outcome measures considered. No ethical ap- trauma group, 322,036 in the blunt trauma only group, 849 in the
proval was required for this study. penetrating trauma only group and 30,036 in the studies reporting
on thoracic trauma only. The stated outcomes of each study are
Results provided in supplementary material 1.
Only seven studies used the WHO obesity classes to report their
Literature search and study selection data, fifteen used a dichotomous cut-off of BMI>30 with the re-
mainder using differing methods of BMI stratification (Table 2).
The literature search yielded 7405 articles, of which 63 un- Twelve studies included ISS grading, five of which adjusted for
derwent full text review. From these, twenty-six studies met ISS severity in their analysis. The quality of the included stud-
the criteria for inclusion and were included in the final analy- ies ranged from 4-9 using the Newcastle Ottawa assessment scale
sis (Fig. 1). Eleven studies reported on both blunt and penetrat- (supplementary material 2) [46]. The included studies achieved the
ing trauma [4,22–31], nine studies reported on blunt trauma alone lowest score in the comparability category as many did not ac-
[5,14,32–40], three studies reported on penetrating trauma [41– count for head trauma in their design or for patient co-morbidities.
43] and three studies reported on isolated thoracic blunt and pen- Most studies had at least a moderate risk of bias, the results from
etrating trauma [13,44,45]. A consensus was agreed for the fi- MINORS risk of bias assessment is shown in supplementary mate-
nal number of studies for inclusion without the need for a third rial 3 with results ranging from 12 to 17 [21]. As all studies were
reviewer. comparative in nature the maximum score a study could achieve
is 24.
Study characteristics
Study outcomes
A total of 2,186,107 patients were included in the 26 studies
of which nearly half (945,511) had a BMI>30 and just under 10% In hospital mortality
(144,046) had a BMI>40. Of the 26 studies, 17 were carried out All 26 studies reported the primary outcome measure mortality
in the United States, 5 in Germany, and there was 1 study each as an outcome. An overview of each outcome that used obesity
from France, Taiwan, Turkey and the Netherlands (Table 2). Over- classes is shown in Table 3 and the remainder in supplementary
all, 1,833,186 patients were reported in the blunt and penetrating material 4.

471
 P.M. Cromwell, I.S. Reynolds, H.M. Heneghan et al.
Table 2
Characteristics of included studies.
Years Total BMI BMI BMI obesity Exclusion Head trauma Death in ED
Study reported Country Design Trauma type number (n) >30 (n) >40 (n) categories used Inclusion criteria criteria excluded excluded ISS grade
Farhat 2020 2013-2015 Germany Retrospective, Blunt and 988988 417481 89433 18.5-24.9; Age >18y Direct Renal No Unclear All
Multi centre, Penetrating 25-29.9; Trauma; (Adjusted
Database 30-34.9; Previous ISS >25 in
35-39.9; >40 nephrectomy multivariate
analysis)
Rock 2019 2006-2014 United Retrospective, Blunt and 406214 372695 NA >30 All trauma Unclear No Unclear All
States Multi centre, Penetrating
Database
DeJong 2016 2008-2012 France Retrospective, Blunt and 910 119 NA >30 All trauma; ICU Died in ED, No Yes All
Single centre Penetrating admissions (Adjusted
ISS >25 in
multivariate
analysis)
Chuang 2009-2013 Taiwan Retrospective, Blunt and 5391 880 NA >30 All trauma Unclear No Unclear All
2016 Single centre, Penetrating
Trauma
registry
Osborne 2008-2012 United Retrospective, Blunt and 2196 656 NA >30 (some All trauma Age <18; Burns, No No Stratified
2014 States single centre, Penetrating analysis < 18.5, Pregnancy
Trauma 18-24.9, 25-29.9,
registry 30-39.9, >40)
Mica 2013 1996-2008 United Retrospective, Blunt and 628 49 NA <25, 25-29.9, ISS >16; Age >16; Unclear No No ISS > 16
472

States singe centre Penetrating >30 Admission within 24

hours
Hoffman 20 05-20 08 Germany Retrospective, Blunt and 5766 760 NA <20, 20-24.9, Alive on arrival; Burns; Dead on No Yes ISS > 9
2012 Multi centre, Penetrating 25-29.9, >30 >16y; ISS >9 arrival
database
Dvorak 2020 2014-2015 United Retrospective, Blunt and 415807 118267 22739 18.5-24.9; >18y; All trauma Outliers No No ISS > 9 / <9
States Multi centre, Penetrating 25-29.9; anaysed for
database 30-34.9; mortality
35-39.9; >40
Treto 2017 2012-2015 United Retrospective, Blunt and 6049 1451 207 19-24, 25-29.9, Age >16; All trauma Burns; Died in No Yes ISS <16,
States Single centre Penetrating 30-39.9, >40 ED 16-25, >25
Mica 2012 1996-2002 United Retrospective, Blunt and 651 50 NA <25, 25-29.9, ISS >16; Age >16 Unclear No No ISS >16
States Single centre Penetrating >30
Andruszkow 2005-2011 Germany Retrospective, Blunt and 586 82 NA >20, 20-24.9, ISS >16; Age >16; Unclear No No ISS >16
2013 Single centre Penetrating 25-29.9, >30 ICU within 24 hours
Hwabejire unclear United Retrospective, Blunt 1976 581 110 18.5-24.9; Haemorrhagic Traumatic brain Yes Unclear All
2015 States Multi centre 25-29.9; shock + blunt injury (AIS >4)
database 30-34.9; trauma, SBP <90 or GCS motor <
35-39.9; >40 within 60 minutes, 3, age <18,
no brain injury, anticipated
blood transfusion survival <24
within 12 hours, hours, immuno-
intact cervical spinal suppression,
cord CHF, CKD, liver

Injury 54 (2023) 469–480

cirrhosis
(continued on next page)
 P.M. Cromwell, I.S. Reynolds, H.M. Heneghan et al.
Table 2 (continued)

Study Years Country Design Trauma type Total BMI BMI BMI obesity Inclusion criteria Exclusion Head trauma Death in ED ISS grade
reported number (n) >30 (n) >40 (n) categories used criteria excluded excluded

Choi 2020 2016 Unitied Retrospective, Blunt 28475 8207 1481 18.5-24.9; Aged 18-64 Unclear No No All
States Multi centre 25-29.9; (adjusted)
database 30-34.9;
35-39.9; >40
Ditillo 2014 2007-2010 Unitied Retrospective, Blunt 32780 NA 16390 BMI >40 >18y; Blunt injury; Dead in ED; Yes No All
States Multi centre analysed BMI >40 and BMI Burns; isolated
database 18.5-25; head trauma
(AIS >3);
Winfield 20 01-20 08 Unitied Retrospective, Blunt 455 130 29 18-24.9, 25-29.9, Blunt; AIS >2; Head Dead in ED; GCS Yes Yes ISS 1-15,
2010 States Multi centre 30-39.9, >40 injury AIS <2 <8; Spinal cord 16-25, >25
database injury
Durgun 2013-2014 Turkey Prospective Blunt 1398 294 54 <25, 25-29.9, Age >15; All blunt Firearm; Sharp; No No All
2016 single centre 30-34.9, >35 trauma Simple cuts
Barry 2019 2014-2016 Unitied Retrospective, Blunt 2411 919 NA BMI >30 Age 15-70; Blunt Unclear No No ISS >15
States Mutli centre MVC or fall (one
database outcome)
Fu 2020 2013-2015 Unitied Retrospective, Blunt 48043 12756 1795 18.5-24.9; Blunt Abdominal Unstable; No Yes ISS <16 (in
States Multi centre 25-29.9; trauma; Stable Penetrating multivariate
database 30-34.9; patient; All ages; ISS trauma analysis)
35-39.9; >40 < 16
2011-2015 Unities Retrospective, Blunt 2452 801 148 18.5-24.9; Motor Vehicle Unclear No No All
473

Brandenberger States Single centre 25-29.9; Collision; ICU (adjusted)

2020 30-39.9; >40 admission; Intubated
Joseph 2017 2007-2010 Unitied Retrospective, Blunt 204046 NA 10260 BMI >40 Blunt Motor Vehicle Age <16; Yes Yes ISS <16,
States Multi centre analysed Collision Isolated head >16
database injury; Dead on
arrival
Hsiao 2019 unclear (4 Unitied Retrospective, Penetrating 100 25 NA BMI >30 >18y; abdominal Unclear No No All
year period) States Single centre stab wound
Bloom 2015 20 0 0-2012 Unitied Retrospective, Penetrating 249 48 NA 18.5-29.9, >15y, abdominal Unclear No No All
States Single centre 30-34.9, >35 stab wound
Nash 2019 2010-2013 Unitied Retrospective, Penetrating 500 132 NA <25, 25-29.9, penetrating Pregnancy, no Yes No All
States Single centre >30 thoraco-abdominal BMI data,
injury (gunshot and intracranial
stab wounds) injury,
Schieren 20 04-20 09 Germany Retrospective, Thoracic 828 276 NA 18-24.9, 25-29.9, AIS thorax > 3 Unclear No No All
2018 Multi centre Blunt and >30
database Penetrating
Cone 2020 2013-2014 Retrospective, Thoracic 28820 8809 1400 18.5-24.9; Chest AIS >3, <5 <20y; >89y; No No ALL
Netherlands Multi centre Blunt 25-29.9; BMI <10; BMI (Patients
database 30-34.9; >55; AIS >2 in included
35-39.9; >40 other areas were based
on AIS)
Eckhardt 2010-2015 Germany Retrospective, Thoracic 388 43 NA <20, 20-24.9, Chest trauma, Age <18; Dead No No ISS >9
2020 Single centre Blunt and 25-29.9, >30 on arrival

Injury 54 (2023) 469–480

Penetrating
P.M. Cromwell, I.S. Reynolds, H.M. Heneghan et al. Injury 54 (2023) 469–480

Table 3
Statistical analysis of outcomes in studies that used WHO classes of obesity.

Mortality Hospital LOS ICU LOS Mechanical Ventilation

Study Obesity Class Yes No Yes No Yes No Yes No

∗
Farhat 2020 Blunt and Penetrating Class 1 X X X X
∗
Class 2 X X X X
Class 3 X∗ X X X
Dvorak 2020 Blunt and Penetrating Class 1 X∗ X∗ X∗ X∗
Class 2 X∗ X∗ X∗ X∗
Class 3 X∗ X∗ X∗ X∗
Hwabejire 2015 Blunt Class 1 X∗ X X X
Class 2 X∗ X X X
Class 3 X X X X
Choi 2020 Blunt Class 1 X∗
Class 2 X∗
Class 3 X∗ X X
Ditillo 2014 Blunt Class 1
Class 2
Class 3 X X X
Fu 2020 Blunt Class 1 X∗
Class 2 X∗
Class 3 X∗
Brandenberger 2019 Blunt Class 1 X∗
Class 2 X∗
Class 3 X∗ X∗
Cone 2019 Thoracic Blunt Class 1 X∗ X X
Class 2 X∗ X X
Class 3 X∗ X X
∗
Indicates that multi-variate statistical analysis was utilised

Morbid obesity (Class 3) BMI counterparts [33]. Farhat et al (2020) similarly found no dif-
Nine of ten studies reported a significantly higher in-hospital ference in risk to patients with class 2 obesity (OR 1.02, 95% CI
mortality in class 3 obesity following trauma (Table 3) [4,5,13– 0.97-108, p = 0.49) but found that all classes of obesity had a step-
15,33,34,36,37,47]. Six studies were amenable to meta-analysis wise increased risk of haemodialysis after trauma (class 2: OR1.89
[5,33,34,36,37,47]. Those who suffered blunt trauma had higher CI 1.66-2.15). A meta-analysis was not possible in this weight cate-
odds of in-hospital mortality when compared with normal-BMI gory due to the heterogeneity and lack of data. No study reported
individuals (OR, 1.75; 95% CI, 1.39-2.19, p =<0.0 0 0 01) (Fig. 2a). a protective effect of having class 2 obesity.
All six studies scored between 15-17 in the MINORS assessment.
Ditillo et al 2014 had the highest risk of bias scoring a 15. When Obesity (Class 1)
this study was removed from the pooled analysis the results sug- Six studies examined class 1 obesity against normal-BMI in-
gest that patients with class 3 obesity were almost twice as likely dividuals. All of these studies reported no significant difference
to die than normal weight individuals (OR, 1.92; CI 1.48-2.49, P in mortality in this weight category (Table 3) [4,5,13–15,33,35]. A
< 0 0 0 0 01). Similarly, studies that reported outcomes in patients meta-analysis was not performed due to heterogeneity of trauma
following blunt and penetrating injury reported that morbidly type.
obese patients had a significantly higher in-hospital mortality than
normal-BMI individuals [4,15]. Meta-analysis was not possible in
this group due to a lack of reported data [28]. There was signifi- Non-class based obesity analysis
cantly higher mortality in patients with class 3 obesity following Fifteen studies used a cut-off point of BMI>30 to analyse obe-
thoracic injury in the single study that reported outcomes in that sity (Table 2). Six of eight studies that combined blunt and pen-
weight category [13]. Across all the weight classes, meta-regression etrating trauma reported no difference in mortality when com-
to estimate the combined effect of the primary outcome was not pared with the BMI<30 group [22,24–26,29,31]. A random-effects
possible as fewer than ten studies with low heterogeneity were meta-analysis showed no significant difference in odds of mor-
available for analysis. A funnel pot was generated to investigate tality between groups (OR, 0.94; 95% CI, 0.86-1.02, p = 0.13)
for bias (supplementary material 5). Some studies were large (Fig. 2b) [4,22,24–27,29–31]. Four of these studies scored between
thereby increasing precision but the small number of studies in 12-14 on MINORS assessment thereby having a higher risk of bias
the analysis made it difficult to determine bias. [4,22,27,31]. Sensitivity analysis was performed removing all four
studies. There was no change in the odds of mortality between
groups (OR, 1.01; 95% CI, 0.65-1.55, p = 0.38). Three studies that
Severe obesity (Class 2) examined blunt trauma only in this category were amenable to
Seven studies examined the odds of mortality of patients with meta-analysis [5,39,40]. There was a significantly increased odds of
BMI 35-40 versus normal-BMI (Table 2) [4,13–15,33,39,47]. Five of mortality in this group on random-effects meta-analysis (OR, 2.28;
the seven studies reported significantly higher mortality in class 95% CI, 1.09-4.76; p=0.03) (Fig. 2c). When Durgun et al 2016 was
2 obesity [13–15,39,47]. Hwabejire et al (2015), a retrospective excluded for sensitivity analysis due to having a MINORS score of
study investigating patients admitted with haemorrhagic shock af- 13 there remained a significantly higher odds of mortality in the
ter blunt trauma, found no significant difference in the odds of BMI > 30 group than the BMI < 30 (OR, 1.70; 95% CI, 1.02-2.85,
mortality for class 2 obesity. However they did report that patients p = 0.04). Four studies reported on outcomes following penetrat-
with class 2 obesity were more likely to experience ARDS (OR 1.89, ing trauma and all showed no significant difference in mortality
95% CI 1.08-3.30, P = .025) and abdominal compartment syndrome [23,41–43]. A meta-analysis was not possible in this group due to
(OR 1.66, 95% CI 1.08 - 2.57, P = .022) compared to their normal- a lack of available data. Two studies reported outcomes following

474
P.M. Cromwell, I.S. Reynolds, H.M. Heneghan et al. Injury 54 (2023) 469–480

Fig. 2. Forrest plots of the primary outcome, in-hospital mortality. 2a: Blunt trauma only BMI > 40 (class 3) versus normal weight, 2b: Blunt and penetrating, BMI > 30
versus BMI < 30, 2c: Blunt trauma only BMI > 30 versus BMI < 30.

thoracic injury, there was no significant difference in mortality be- Mean Difference (SMD), 0.23; 95% CI, 0.21-0.25; p <0.0 0 0 01)
tween groups [44,45]. (Fig. 3a). When the study with the highest risk of bias was re-
moved [14] there was no difference in the result of the pooled ef-
Hospital length of stay fect of the two remaining studies (SMD 0.25; 95% CI, 0.16-0.33; p
Twenty studies reported hospital LOS as an outcome. Thirteen < 0.0 0 0 01). There was insufficient data in the remaining studies to
studies used BMI>30 and compared it against BMI<30 (Supple- perform further analysis.
mentary material 4). Seven used WHO classes and compared them
against normal-BMI patients (Table 3). Obesity class 1-2
Four studies reported on hospital LOS outcomes in the class 2
Morbid obesity (Class 3) obesity group [4,13,15,33]. Farhat et al (2020) (median days, 4.0 vs
All seven studies utilising the WHO obesity class system re- 3.0, p < 0.001), Cone et al (2020) (6.5 vs 5.4 days p <0.01) and Dvo-
ported a significantly longer hospital LOS in BMI>40 compared rak et al (2020) (IRR 1.117 (CI 1.105–1.129) all report significantly
with normal-BMI individuals [4,5,13–15,33,37]. This included two longer LOS when compared with normal-BMI patients [4,13,15]. Al-
blunt and penetrating trauma studies, three blunt trauma only and though Hwabejire et al (2015) reported no significant difference in
one blunt thoracic trauma study. The three blunt trauma stud- LOS in class 2 obesity they did report longer LOS in class 3 (29 +/-
ies were suitable for random-effects meta-analysis [14,33,37]. Pa- 27 vs 21 +/- 22, p = .007) and class 1 ((26 +/- 30 vs 21 +/- 22,
tients with class 3 obesity were found to have a significantly longer p = .020) [33]. Meta-analyses was not possible due to heterogene-
hospital LOS compared to normal BMI individuals (Standardised ity in types of trauma reported.

475
P.M. Cromwell, I.S. Reynolds, H.M. Heneghan et al. Injury 54 (2023) 469–480

Fig. 3. Forrest plots of secondary outcomes, Hospital length of stay, ICU length of stay and mechanical ventilation. 3a Hospital LOS, blunt trauma only, BMI>40 (class 3)
versus normal weight. 3b Hospital LOS, blunt and penetrating trauma, BMI>30 versus BMI<30. 3c ICU length of stay, blunt trauma class 3 versus normal weight, 3d ICU
length of stay, blunt and penetrating trauma, BMI>30 versus BMI <30, 3e duration of mechanical ventilation, blunt and penetrating trauma, BMI>30 versus BMI <30.

Non-class based obesity analysis (BMI>30) [25,26]. Sensitivity analysis also suggested a significantly longer
Eight studies reported on hospital LOS following blunt and pen- length of stay in the obese group (MD, 1.44; 95% CI, 0.65-2.23;
etrating trauma [22,24–27,29–31]. In the five studies with avail- p = 0.0 0 03). Both studies that reported on outcomes following
able data, there was a significantly higher risk of longer hospital blunt trauma in this weight category reported a significantly longer
LOS in the obese group on random effects meta-analysis (MD, 1.76; LOS in the obesity groups [23,40]. In the three penetrating trauma
95% CI, 1.01-2.51; p <0.0 0 0 01) (Fig. 3b) [22,25–27,30,31]. Only two only studies, only one reported a significantly longer LOS in the
studies scored a MINORS risk of bias 15 or above from this group obese group[41]. Two studies that examined thoracic blunt and

476
P.M. Cromwell, I.S. Reynolds, H.M. Heneghan et al. Injury 54 (2023) 469–480

penetrating trauma reported a significantly longer hospital LOS in (SMD, 0.10; 95% CI, -0.23-0.43; p=0.56). Two of these studies
BMI>30 [5,44]. scored less than 15, therefore no further sensitivity analysis was
performed. One of two studies examining blunt trauma reported
ICU length of stay significantly longer mechanical ventilation in the obesity group
Twenty-one studies reported data on ICU LOS. Six studies used [23,40]. Schieren (2018) reports that obese patients who suffered
WHO classes (Table 3). Eleven studies used BMI>30 for their obe- thoracic trauma had a significantly longer duration of mechanical
sity analysis and 2 studies used miscellaneous categories (supple- ventilation than non-obese [44].
mentary material 4).

Obesity classes 1-3 Discussion

A random-effects meta-analysis of all three studies reporting on
ICU LOS following blunt trauma showed a significantly longer ICU This study provides a contemporary analysis of the impact of
LOS in the morbid obesity group (SMD, 0.19; 95% CI, 0.12 – 0.26; p obesity on outcomes in traumatic injury. Whilst the last system-
<0.0 0 0 01) [14,33,37] (Fig. 3c). When sensitivity analysis was per- atic review of this nature published a decade ago primarily de-
formed and the study with the highest risk of bias was removed scribed outcomes using a dichotomous definition of obesity, this
[14] there remained a significantly longer ICU LOS in the morbid is the first systematic review and meta-analysis of its kind to
obesity group (SMD, 0.24; 95% CI, 0.11-0.38; p=0.0 0 04). One of use the WHO obesity class system to interpret outcomes follow-
the two studies reporting blunt and penetrating trauma reported a ing trauma [48,49]. In doing so, our study shows that those indi-
longer ICU LOS in morbid obesity [4,15]. Morbidly obese patients viduals with class 3 obesity appear to be at significantly greater
also had a significantly longer ICU LOS following blunt thoracic risk for worse outcomes following blunt trauma. Those living with
trauma (5.6 versus 4.4 days, p <0.001) [13]. The four studies that class 3 obesity carry a much higher risk of death, require sig-
reported on obesity class 1 and class 2 mirrored the results found nificantly longer hospital and ICU LOS and are associated with
in the morbid obesity group (Table 3) [4,13,15,33]. Three studies more prolonged mechanical ventilation compared to those of a
report a significantly longer ICU LOS in both class 1 and class 2 normal BMI. This remained true after sensitivity analysis remov-
obesity [13,15,33] however Farhat et al (2020) reports no signifi- ing the studies with the highest risk of bias. The data suggests
cant difference for either class 1 (3.0 days vs 3.0 days, p = 0.28) an increased risk for class 2 obesity also. Furthermore, there was
or class 2 (3.0 days vs 3.0 days, p = 0.38) versus normal weight no protective effect reported in patients with class 2 or class
individuals [4]. 3 obesity.
Nine out of the ten studies that analysed class 3 obesity com-
Non-class based obesity analysis pared to normal-BMI individuals reported significantly higher odds
Four of the seven studies reporting blunt and penetrating of in-hospital mortality. This finding is in contrast to studies that
trauma showed a significantly longer ICU LOS in the BMI>30 define obesity as a dichotomous BMI>30 versus BMI<30 variable,
category[24–27,29–31]. A pooled estimate showed a significantly in which only a minority of studies report a significantly higher
longer LOS in the obesity group (MD, 2.23; 95% CI, 1.33-3.14; rate of in-hospital mortality within the obese cohort. Hwabejire
p<0.0 0 0 01) (Fig. 3d). The pooled estimates were similar when two et al (2015) demonstrate this concern in their patient cohort, ini-
of the studies with the highest risk of bias were excluded (MD, tially dividing cases dichotomously and finding no significant dif-
2.11; 95% CI 1.16-3.06; p <0.0 0 01) [14,30]. One of two studies re- ference in mortality before stratifying BMI by class and identifying
ported a significantly longer ICU LOS following blunt trauma, and morbidly obese patients as having twice the likelihood of death
one study reported no significant difference following penetrating compared to normal-BMI patients [33].
trauma [23,40]. Schieren (2018) reported that obese patients had Better characterising the association between obesity and mor-
significantly longer ICU LOS following thoracic injury [44]. tality following trauma is challenging; however, as suggested in
previous modelling work by both Dvorak et al and Choi and col-
Mechanical ventilation leagues in 2019 and 2020 respectively, the relationship appears
Duration of mechanical ventilation was reported in 12 of the non-linear [14,15]. Both suggest that, as seen in this study, higher
included studies. Four studies used the WHO obesity classes levels of obesity such as those in class 2 and 3, have a higher risk
(Table 3), seven studies used a BMI>30 cut-off for their analysis of in-hospital mortality than those of normal-BMI. In contrast, class
with the remainder reported miscellaneous weight categories (sup- 1 obesity appears to have no increased risk of death post trauma
plementary material 4). compared to normal weight individuals. Only one of seven studies
analysed in this systematic review reported a significantly higher
Obesity classes 1-3 in-hospital mortality in class 1 obesity and analysing class 2 and 3
One of two studies reporting blunt and penetrating trauma in obese patients together with an often larger, lower risk cohort of
class 3 obesity reported a statistically significant longer duration individuals, risks concealing poorer outcomes amongst more obese
of mechanical ventilation [4,15]. In this study, class 1 patients had patient cohorts [39].
a 22% increase in ventilator days (IRR 1.217; 95% CI 1.171-1.263), The effect of isolated penetrating injury on survival amongst
class 2 had a 30% increase (IRR 1.295; 95% CI 1.228-1.366) and class this patient population remains more uncertain. Hsiao et al in 2019
3 had a 54% increase (IRR 1.536; 95% CI 1.450 – 1.627) [15]. Three showed a BMI of greater than 30 to be an independent risk factor
studies examining blunt trauma reported significantly longer me- for GI resection following isolated abdominal stab injury [41]. This
chanical ventilation in class 3 obesity [5,33,36]. Metanalysis was is in contrast to Nash and colleagues, who in their published study
not possible due to heterogeneity. of the same year, reported no significant difference in either the
rate of operative intervention nor post-operative adverse outcomes
Non-class based obesity analysis when comparing obese and non-obese patients suffering penetrat-
All four studies on blunt and penetrating trauma report a sig- ing injury [43]. Whilst previous work in 2015 by Bloom et al sug-
nificantly longer duration of mechanical ventilation in the BMI>30 gested there to in fact be a protective influence from obesity in
category when compared with BMI<30 [26,29–31]. Three stud- penetrating trauma, identifying a lower incidence of visceral injury
ies were amenable to meta-analysis, albeit with significant het- amongst obese compared to non-obese patients following isolated
erogeneity, but revealed no significant difference between groups abdominal stab injuries [42].

477
P.M. Cromwell, I.S. Reynolds, H.M. Heneghan et al. Injury 54 (2023) 469–480

A high BMI can impact all stages of the patient journey from Conclusions and recommendations
the initial response to injury and resuscitation through to definitive
management and rehabilitation. Adipose tissue is a complex or- The obesity paradox is not an accurate representation of the
gan, the physiology of which, even outside of traumatic injury, re- complexities involved in caring for the injured obese patient. There
mains incompletely understood. Morbid obesity has been shown to is a higher risk of in-hospital mortality in patients living with class
be associated with secretion of pro-inflammatory mediators such 3 obesity following trauma, this is likely true for class 2 also. Fur-
as IL-1, IL-6 and TNF-α from hypertrophied adipocytes leading to thermore, all obese patients have a significantly longer hospital
a continuous state of low grade inflammation [50]. Similarly, An- stay, ICU stay and a trend towards longer duration of mechanical
giotensin II, a potent vasoconstrictor, likely contributes to obesity- ventilation. Therefore, trauma systems should develop weight class
related hypertension but is also involved in the activation of path- related trauma pathways to manage and anticipate the complica-
ways of oxidative stress [51]. Also, major trauma itself provokes tions that arise when managing these complex patients.
a dysfunctional inflammatory mediator release that can result in Prospective, multi-centred studies are needed and future stud-
multi-organ dysfunction syndrome (MODS) [52]. ies should stratify and define obesity using the WHO obesity
Further challenges, both physiological and mechanical also classes (Table 1) and not a BMI > 30 dichotomy. This will allow
present opportunities to influence outcomes amongst this popu- for more precise analysis of the obesity-related complications of
lation. It is well recognised for example that obese patients more trauma management as obesity has a non-linear relationship with
commonly suffer from concomitant cardiovascular and pulmonary mortality.
conditions, present with more difficult airways and vascular access
and generate greater logistical challenges during extrication, trans- Author contributions
portation and investigation of their injuries compared to normal-
BMI individuals. This can result in poorer overall outcomes [53,54]. Study design – PC
Such factors may in part explain the longer hospital LOS observed Literature search – PC, IR
in this study amongst obese individuals. Data Collection – PC, IR
Similarly, the longer ICU admissions and more prolonged peri- Data analysis – PC
ods of mechanical ventilation found in class 3 obese patients in Data interpretation – PC
this review are likely a product of factors inherent to the obese Writing – PC, SG
population. Higher rates of multi-organ failure, acute respiratory Critical revision – PC, IR, SG, HH
distress syndrome, sepsis, AKI and dialysis observed in general
amongst obese individuals are likely to translate from non-trauma Financial support
ICU admissions to trauma-based populations. Hence the finding
during this review of factors such as prolonged ventilation being None
significantly increased even on dichotomous analysis of obesity is
unsurprising and widely supported in previous reports in this area Declaration of Competing Interest
[4,24,26,27,30,31,33,37,48].
Early recognition of this at-risk population and anticipating the All authors declare no support from any organisation for the
challenges that present across the paradigm of trauma care is key submitted work; no financial relationships with any organisations
to improving future outcomes. This involves, for example antic- that might have an interest in the submitted work in the previous
ipating the monitoring and investigation difficulties experienced three years; no other relationships or activities that could appear
with these patients, the nutritional challenges they present with to have influenced the submitted work.
and ensuring consideration of their varied pharmacokinetics. Struc-
tured, obesity-specific protocols are just one initiative which may Supplementary materials
be utilised to address such issues [55]. These include measures
such as weight-based dosing of LMWH with appropriate pneu- Supplementary material associated with this article can be
matic compression to reduce the significantly higher rates of ve- found, in the online version, at doi:10.1016/j.injury.2022.10.026.
nous thromboembolic events associated with obesity; the use of
patient-appropriate bariatric beds for ease of transfer and patient References
repositioning; aggressive pulmonary toileting with adjunctive res-
piratory physiotherapy; and the early implementation of nutri- [1] World Health Organisation. Obesity and overweight [Internet]. WHO Int.
2021 [cited 2021 Jul 4]. Available from: https://www.who.int/news-room/
tional regimens using recognised guidelines such as ASPEN [9,56– fact-sheets/detail/obesity- and- overweight
59]. [2] Obesity Guidelines for Clinicians. 2019 obesity algorithm E-BookMain [In-
ternet]. [cited 2021 Jun 19]. Available from: https://obesitymedicine.org/
obesity-algorithm/
Limitations [3] Tracy BM, Wilson JM, Staley C, Frias B, Schenker ML, Gelbard RB.
Metabolic syndrome: major risk factor for morbidity and mortality in
We recognise a number of limitations remain in relation to severely injured trauma patients. J Am Coll Surg 2020;230(1):145–50
[Internet]Available from. https://www.embase.com/search/results?subaction=
this work. All but one of the studies included in this system- viewrecord&id=L2003779842&from=export
atic review were retrospective cohort studies, pooling data into [4] Farhat A, Grigorian A, Nguyen NT, Smith B, Williams BJ, Schubl SD, et al. Obese
a combined effect should be therefore interpreted with a degree trauma patients have increased need for dialysis. Eur J trauma Emerg Surg Off
Publ Eur Trauma Soc 2020;46(6):1327–34 Dec.
of caution. Unfortunately, meta-regression was not possible due
[5] Winfield RD, Delano MJ, Dixon DJ, Schierding WS, Cendan JC, Lottenberg L,
to the lack of studies with standardised data and whilst attempts et al. Differences in outcome between obese and nonobese patients following
were made to reduce heterogeneity by grouping studies into dif- severe blunt trauma are not consistent with an early inflammatory genomic
response. Crit Care Med 2010;38(1):51–8 Jan.
ferent trauma types and obesity categories, again given the na-
[6] Finkelstein EA, Chen H, Prabhu M, Trogdon JG, Corso PS. The relation-
ture of clinical trauma management the data should be interpreted ship between obesity and injuries among U.S. adults. Am J Health Promot
with some caution. Finally, as highlighted above, the lack of iso- 2007;21(5):460–8.
lated penetrating data prevented interpretation of the effects of [7] Liu T, Chen JJ, Bai XJ, Zheng GS, Gao W. The effect of obesity on outcomes in
trauma patients: a meta-analysis. Injury 2013 Sep 1;44(9):1145–52.
obesity on this population and further work is needed in this [8] Cone JT, Benjamin ER, Alfson DB, Demetriades D. Isolated severe blunt trau-
area. matic brain injury: effect of obesity on outcomes. J Neurosurg 2020:1–8 Jun.

478
P.M. Cromwell, I.S. Reynolds, H.M. Heneghan et al. Injury 54 (2023) 469–480

[9] Newell MA, Bard MR, Goettler CE, Toschlog EA, Schenarts PJ, Sagraves SG, [32] Webster J, Osborne S. Preoperative bathing or showering with skin antisep-
et al. Body mass index and outcomes in critically injured blunt trauma pa- tics to prevent surgical site infection. Cochrane Database Syst Rev [Internet]
tients: weighing the impact. J Am Coll Surg 2007;204(5):1056–61. [Inter- 2015(2) Available from. doi:10.10 02/14651858.CD0 04985.pub5.
net]May [cited 2021 Jul 4]Available from https://pubmed.ncbi.nlm.nih.gov/ [33] Hwabejire JO, Nembhard CE, Obirieze AC, Oyetunji TA, Tran DD, Fullum TM,
17481540/. et al. Body mass index in blunt trauma patients with hemorrhagic shock: op-
[10] Diaz JJ, Norris PR, Collier BR, Berkes MB, Ozdas A, May AK, et al. Morbid obe- posite ends of the body mass index spectrum portend poor outcome. Am J
sity is not a risk factor for mortality in critically ill trauma patients. J Trauma - Surg 2015;209(4):659–65 Apr.
Inj Infect Crit Care 2009;66(1):226–31. [Internet]Jan [cited 2021 Jul 4]Available [34] Joseph B, Hadeed S, Haider AA, Ditillo M, Joseph A, Pandit V, et al. Obesity and
from https://pubmed.ncbi.nlm.nih.gov/19131831/ . trauma mortality: sizing up the risks in motor vehicle crashes. Obes Res Clin
[11] Alban RF, Lyass S, Margulies DR, Shabot MM. Obesity does not af- Pract [Internet] 2017;11(1):72–8 JanAvailable from. http://proxy.library.rcsi.ie/
fect mortality after trauma. Am Surg 2006;72(10):966–9. [Internet]Oct login?url=https://search.ebscohost.com/login.aspx?direct=true&db=
1 [cited 2021 Jul 5]Available from https://journals.sagepub.com/doi/ cmedm&AN=
abs/10.1177/0 0 0313480607201028?casa_token=IdFKpjaPemwAAAAA% 26996285&site=ehost-live.
3APGzXzNm0SpExIsppa9TCGOXfGYkg286WksfvnNo-Wvc1rgtI7VCD0dorPct6z [35] Fu C-Y, Bajani F, Bokhari M, Tatebe LC, Starr F, Messer T, et al. Obesity is
te309jCDma_mt5pGQ . associated with worse outcomes among abdominal trauma patients undergo-
[12] Obesity: preventing and managing the global epidemic . Report of a WHO con- ing laparotomy: a propensity-matched nationwide cohort study. World J Surg
sultation, 894. Switzerland: World Health Organization technical report series; 2020;44(3):755–63 Mar.
20 0 0. [36] Brandenberger KJ, Culbreth R, Shan S, Gardenhire DS, Warren GL. A retro-
[13] Cone JT, Benjamin ER, Alfson DB, Biswas S, Demetriades D. The effect of spective analysis of associations between BMI and days spent on mechan-
body mass index on outcomes following severe blunt chest trauma. Injury ical ventilation in a level 1 trauma facility. Heart Lung 2020;49(5):605–
2020;51(9):2076–81 Sep. 609.
[14] Choi J, Smiley A, Latifi R, Gogna S, Prabhakaran K, Con J, et al. Body mass [37] Ditillo M, Pandit V, Rhee P, Aziz H, Hadeed S, Bhattacharya B, et al. Morbid
index and mortality in blunt trauma: the right BMI can be protective. Am J obesity predisposes trauma patients to worse outcomes: a National Trauma
Surg 2020 (xxxx). Data Bank analysis. J Trauma Acute Care Surg 2014;76(1):176–9 Jan.
[15] Dvorak JE, Lester ELW, Maluso PJ, Tatebe L, Schlanser V, Kaminsky M, [38] Winfield RD, Delano MJ, Cuenca AG, Cendan JC, Lottenberg L, Efron PA, et al.
et al. The obesity paradox in the trauma patient: normal may not be bet- Obese patients show a depressed cytokine profile following severe blunt in-
ter. World J Surg 2020;44(6):1817–23 [Internet]Available from. doi:10.1007/ jury. Shock 2012;37(3):253–6 Mar.
s00268- 020- 05398- 1. [39] Durgun HM, Dursun R, Zengin Y, Özhasenekler A, Orak M, Üstündağ M, et al.
[16] DF S JA B, SC M I O, GD W D R, et al. Meta-analysis of observational stud- The effect of body mass index on trauma severity and prognosis in trauma
ies in epidemiology: a proposal for reporting. Meta-analysis of observational patients. Ulus Travma Ve Acil Cerrahi Derg = Turk J trauma Emerg Surg TJTES
studies in epidemiology (MOOSE) group. JAMA 20 0 0;283(15):20 08–12. [Inter- 2016;22(5):457–65 Sep.
net]Apr 19 [cited 2021 Jul 6]Available from https://pubmed.ncbi.nlm.nih.gov/ [40] Barry R, Modarresi M, Aguilar R, Sanabria J, Wolbert T, Denning D, et al. The
10789670/ . impact of BMI on adult blunt trauma outcomes. Am Surg 2019;85(12):1354–62
[17] Page M, McKenzie J, Bossuyt P, Boutron I, Hoffman T, Mulrow C, et al. Dec.
The PRISMA 2020statement: an updated guideline for reporting system- [41] Hsiao V, Sim J, Zimmerman A, Stephen A. Obesity may not be protective in
atic reviews. MetaArXiv Prepr 2020:1–36. [Internet](September)Available from abdominal stab wounds. J Emerg Trauma Shock 2019;12(3):168–72.
https://osf.io/preprints/metaarxiv/v7gm2/ . [42] Bloom MB, Ley EJ, Liou DZ, Tran T, Chung R, Melo N, et al. Impact of body mass
[18] Tran A, Yates J, Lau A, Lampron J, Matar M, Tran A, et al. International prospec- index on injury in abdominal stab wounds: implications for management. J
tive register of systematic reviews systematic review and meta-analysis of hy- Surg Res 2015;197(1):162–6 Jul.
potensive vs . conventional resuscitation for traumatic hemorrhage citation re- [43] Nash N, Kimbrough CW, Mackowski M, Benns M V, Smith JW, Harbrecht BG,
view question participants /population intervention. Exposure Comp 2017:1– et al. Is obesity protective in thoracoabdominal penetrating trauma? Am Surg
3. PROSPERO [Internet]Available from https://www.crd.york.ac.uk/prospero/ 2019;85(1):34–8 Jan.
display_record.php?ID=CRD42017070526 . [44] Schieren M, Böhmer AB, Lefering R, Paffrath T, Wappler F, Defosse J.
[19] Cochrane Community. RevMan 5 | cochrane community [Internet]. Impact of body mass index on outcomes after thoracic trauma-A
[cited 2020 Feb 2]. Available from: https://community.cochrane.org/help/ matched-triplet analysis of the TraumaRegister DGU(®). Injury 2019;50(1):96–
tools- and- software/revman- 5 100 Jan.
[20] Higgins J, Thomas J. Cochrane handbook for systematic reviews of interven- [45] Eckhardt T, Horst K, Störmann P, Bläsius F, Hofman M, Herren C, et al. Impact
tions | cochrane training version 6.2 [Internet]. 2021 [cited 2022 Feb 1]. Avail- of chest trauma and overweight on mortality and outcome in severely injured
able from: https://training.cochrane.org/handbook/current patients. J Clin Med 2020;9(9) Aug.
[21] Slim K, Nini E, Forestier D, Kwiatkowski F, Panis Y, Chipponi J. Methodological [46] GA Wells, B Shea, D O’Connell, J Peterson, V Welch, M Losos PT. The Newcastle-
index for non-randomized studies (Minors): development and validation of a Ottawa Scale (NOS) for assessing the quality of non-randomised studies in
new instrument. ANZ J Surg 2003;73(9):712–16. [Internet]Sep 1 [cited 2021 Jul meta-analyses [Internet]. [cited 2021 Jun 21]. Available from: http://www.ohri.
4]Available from https://onlinelibrary.wiley.com/doi/full/10.1046/j.1445-2197. ca/programs/clinical_epidemiology/oxford.asp.
2003.02748.x . [47] Fu C-Y, Bajani F, Butler C, Welsh S, Messer T, Kaminsky M, et al. Morbid obe-
[22] Rock K, Hayward RD, Edhayan E. Obesity and hospital outcomes follow- sity’s silver lining: an armor for hollow viscus in blunt abdominal trauma.
ing traumatic injury: associations in 9 years of patient data from a single World J Surg 2019;43(4):1007–13 Apr.
metropolitan area. Clin Obes 2019;9(2):e12293 Apr. [48] Liu T, Chen J, Bai X, Zheng G, Gao W. The effect of obesity on outcomes in
[23] Osborne Z, Rowitz B, Moore H, Oliphant U, Butler J, Olson M, et al. Obesity in trauma patients: a meta-analysis. Injury 2013;44(9):1145–52 Sep.
trauma: outcomes and disposition trends. Am J Surg 2014;207(3):382–7 Mar. [49] Winfield RD, Bochicchio GV. The critically injured obese patient: a review and
[24] De Jong A, Deras P, Martinez O, Latry P, Jaber S, Capdevila X, et al. Rela- a look ahead. J Am Coll Surg 2013;216(6):1193–206 Jun.
tionship between obesity and massive transfusion needs in trauma patients, [50] Fischer-Posovszky P, Wabitsch M, Hochberg Z. Endocrinology of adipose tis-
and validation of TASH score in obese population: a retrospective study on sue - an update. Horm Metab Res [Internet] 2007;39(5):314–21 May 29
910 trauma patients. PLoS One [Internet]. 2016 Mar 24;11(3):e0152109 Avail- [cited 2022 Feb 11]Available from. http://www.thieme-connect.com/products/
able from. http://proxy.library.rcsi.ie/login?url=https://search.ebscohost.com/ ejournals/html/10.1055/s- 2007- 976539
login.aspx?direct=true&db=cmedm&AN=27010445&site=ehost-live [51] Winfield RD, Southard RE, Turnbull IR, Bochicchio K, Reese S, Freeman BD,
[25] Chuang J-F, Rau C-S, Kuo P-J, Chen Y-C, Hsu S-Y, Hsieh H-Y, et al. Traumatic in- et al. Angiotensin inhibition is associated with preservation of T-cell and
juries among adult obese patients in southern Taiwan: a cross-sectional study monocyte function and decreases multiple organ failure in obese trauma pa-
based on a trauma registry system. BMC Public Health 2016;16:275 Mar. tients. J Am Coll Surg 2015;221(2):486–94 Aug.e4.
[26] Mica L, Keller C, Vomela J, Trentz O, Plecko M, Keel MJ. Obesity and overweight [52] Manson J, Cole E, De’Ath HD, Vulliamy P, Meier U, Pennington D, et al. Early
as a risk factor for pneumonia in polytrauma patients: a retrospective cohort changes within the lymphocyte population are associated with the develop-
study. J Trauma Acute Care Surg 2013;75(4):693–8 Oct. ment of multiple organ dysfunction syndrome in trauma patients. Crit Care
[27] Hoffmann M, Lefering R, Gruber-Rathmann M, Rueger JM, Lehmann W. The [Internet] 2016;20(1):1–10 Jun 7 [cited 2022 Jun 19]Available from. https:
impact of BMI on polytrauma outcome. Injury 2012;43(2):184–8 Feb. //link.springer.com/articles/10.1186/s13054- 016- 1341- 2
[28] Dvorak JE, Lester ELW, Maluso PJ, Tatebe L, Schlanser V, Kaminsky M, et al. [53] De Jong A, Chanques G, Jaber S. Mechanical ventilation in obese ICU patients:
The obesity paradox in the trauma patient: normal may not be better. World J from intubation to extubation [Internet]. Critical Care. BioMed Central Ltd.
Surg 2020;44(6):1817–23 Jun. 2017;21 [cited 2021 Jul 3]. Available from. https://pubmed.ncbi.nlm.nih.gov/
[29] Treto K, Safcsak K, Chesire D, Bhullar IS. Analysis of mortality in traumatically 28320439/
injured patients based on body mass index and mechanism reveals highest [54] Sifri ZC, Kim H, Lavery R, Mohr A, Livingston DH. The impact of obesity
mortality among the underweight in comparison with the ideal weight pa- on the outcome of emergency intubation in trauma patients. J Trauma -
tients. Am Surg 2017;83(4):341–7 Apr. Inj Infect Crit Care [Internet] 2008;65(2):396–400 [cited 2021 Jul 6]Avail-
[30] Mica L, Keel M, Trentz O. The impact of body mass index on the physiology of able from. https://journals.lww.com/jtrauma/Fulltext/20 08/080 0 0/The_Impact_
patients with polytrauma. J Crit Care 2012;27(6):722–6 Dec. of_Obesity_on_the_Outcome_of_Emergency.21.aspx
[31] Andruszkow H, Veh J, Mommsen P, Zeckey C, Hildebrand F, Frink M. Im- [55] Gray S, Dieudonne B. Optimizing care for trauma patients with obesity.
pact of the body mass on complications and outcome in multiple trauma CUREUS. 2018;10(7) Jul.
patients: what does the weight weigh? Mediators Inflamm 2013;2013: [56] Miranda S, Le Besnerais M, Lévesque H, Benhamou Y, Armengol G, Le Cam-
345702. Duchez V, et al. Adjusted value of thromboprophylaxis in hospitalized obese

479
P.M. Cromwell, I.S. Reynolds, H.M. Heneghan et al. Injury 54 (2023) 469–480

patients: A comparative study of two regimens of enoxaparin: The ITOHENOX [58] Lorenzet R, Napoleone E, Cutrone A, Donati MB. Thrombosis and obesity. Cel-
study. Thromb Res [Internet] 2017;155:1–5 Jul 1 [cited 2021 Jul 3]Available lular bases [Internet] 2021;129:285–9 Thrombosis Research. Thromb Res; 2012
from. https://pubmed.ncbi.nlm.nih.gov/28460259/. [citedJul 3]Available from. https://pubmed.ncbi.nlm.nih.gov/22078462/
[57] Rondina MT, Wheeler M, Rodgers GM, Draper L, Pendleton RC. Weight-based [59] Choban P, Dickerson R, Malone A, Worthington P, Compher C. clinical guide-
dosing of enoxaparin for VTE prophylaxis in morbidly obese, medically-Ill pa- lines: nutrition support of hospitalized adult patients with obesity. J Parenter
tients. Thromb Res [Internet] 2010;125(3):220–3 Mar [cited 2021 Jul 3]Avail- Enter Nutr [Internet] 2013;37(6):714–44 Nov [cited 2021 Jul 3]Available from.
able from. https://pubmed.ncbi.nlm.nih.gov/19272635/ https://pubmed.ncbi.nlm.nih.gov/23976769/

480