Management of liver injuries: Predictors for the need of operation and damage control surgery

Injury, Int. J. Care Injured xxx (2014) xxx–xxx

G Model

JINJ-5645; No. of Pages 5

Management of liver injuries: Predictors for the need of operation anddamage control surgery

Supparerk Prichayudh *, Chayatat Sirinawin, Suvit Sriussadaporn, Rattaplee Pak-art,Kritaya Kritayakirana, Pasurachate Samorn, Sukanya Sriussadaporn

Chulalongkorn University, Surgery, Rama4 rd, Bangkok, Thailand

A R T I C L E I N F O

Article history:

Received 9 December 2013

Received in revised form 29 January 2014

Accepted 5 February 2014

Keywords:

Liver injury

Non-operative management

Damage control surgery

A B S T R A C T

Management of liver injuries: Predictors for the need of operation and damage control surgery,

Introduction: The advancement in the management of liver injuries, including the use of non-operative

management (NOM), damage control surgery (DCS) and angiographic embolisation (AE); has resulted, in

improvement of outcomes. The aim of this study is to analyse the outcome of liver injury patients in our

institution and to identify predictors for the need of operative management (OM) and DCS.

Patients and methods: We retrospectively reviewed 218 patients with liver injury admitted to King,

Chulalongkorn Memorial Hospital from May 2002 to May 2011. Data collection included demographic,

data, emergency department parameters, detail of liver injuries, and outcome in terms of mortality rate

(MR). Stepwise logistic regression was performed to identify mutually independent predictors for the

need of OM and DCS.

Results: Two hundred and eighteen patients with liver injury were identified (156 blunt and 62

penetrating). One hundred fifty-four patients (70.6%) underwent OM due to hemodynamic instability,

(96), peritonitis (24), and other indications (34). DCS (perihepatic packing and temporary abdominal,

closure) was utilised in 45 patients. NOM was attempted in 64 patients (29.4%), 6 of these, subsequently

required laparotomy (success rate 90.6%). Angiography was performed in 47 patients, (14 in NOM, 33 in

OM) and 40 patients received AE (10 in NOM, 30 in OM). Overall MR was 17.4%, the, MR was significantly

higher in OM than in NOM (24 vs. 1.6%; p < 0.001, OR 19.92). The mutually independent predictors for

the need of operation were low Glasgow Coma Score (GCS), penetrating mechanism, tachycardia, and

hypotension; while the independent predictors for DCS were high grade (>4) liver injury, tachycardia,

and blunt mechanism.

Conclusions: Overall MR of liver injury patients was 17.4%. NOM carried a low MR and should be,

attempted in the absence of hemodynamic instability and peritonitis. Patients with low GCS, penetrating

injury, tachycardia, and hypotension were more likely to require operation. DCS should be considered

while operating on patients with high grade liver injury, tachycardia, and blunt mechanism.

� 2014 Elsevier Ltd. All rights reserved.

Contents lists available at ScienceDirect

Injury

jo ur n al ho m epag e: ww w.els evier . c om / lo cat e/ in ju r y

Introduction

During the past 40 years, the management of liver injuries hasevolved significantly due to the development of angiographicembolisation (AE) and non-operative management (NOM) [1–6].This has resulted in less operative interventions required inmanaging liver injuries, especially blunt liver injury patients in

* Corresponding author. Tel.: +66 2256 4117; fax: +66 2256 4194.

E-mail address: [email protected] (S. Prichayudh).

Please cite this article in press as: Prichayudh S, et al. Management ocontrol surgery. Injury (2014), http://dx.doi.org/10.1016/j.injury.201

http://dx.doi.org/10.1016/j.injury.2014.02.013

0020–1383/� 2014 Elsevier Ltd. All rights reserved.

which NOM could be performed in 50–80% of the patients andcarries successful rate of 88–98% [1–7]. Furthermore, the accep-tance of damage control surgery (DCS) techniques has alsoimproved the outcomes of the patients with complex liver injuryrequiring operation [8–11]. Recently, the reported mortality ratesof liver injury patients ranged from 5 to 52% [1,3–7,11–14], thehigher mortality rates are from the series with more severe liverinjuries and higher operative rates [11–14]. The objectives of thisstudy are to review the outcome of liver injury management in theera of AE and DCS, and identify the predictors for the need ofoperation and DCS.

f liver injuries: Predictors for the need of operation and damage4.02.013


mailto:[email protected]


http://www.sciencedirect.com/science/journal/00201383

www.elsevier.com/locate/injury


Table 1Demographic data and emergency department (ED) parameters.

Total

(n = 218)

NOM

(n = 64)

OM

(n = 154)

p-value

Sex

Male 186 (85.3%) 50 (78.1%) 136 (89.6%) 0.084

Female 32 (15.7%) 14 (21.9%) 18 (10.4%)

Age

Mean (SD) 30.4 (12.4) 31.6 (14.0) 25.9 (11.7) 0.610

Mechanism

Blunt 156 (71.6%) 60 (93.7%) 96 (62.3%) <0.001

Penetrating 62 (28.4%) 4 (5.3%) 58 (37.7%)

ED parameters (Mean (SD))

SBP 110.8 (33.6) 121.6 (21.7) 106.3 (36.6) 0.002

PR 97.8 (26.5) 90.4 (19.3) 100.9 (28.5) 0.007

GCS 12.7 (4.0) 14.7 (0.7) 11.8 (4.4) <0.001

ISS 25.9 (14.2) 20.8 (9.0) 28.1 (15.4) 0.001

RTS 6.93 (1.74) 7.76 (0.26) 6.58 (1.97) <0.001

TRISS 85.1 (27.6) 97.1 (3.1) 80.1 (31.5) <0.001

Hematocrit 33.7 (10.2) 36.8 (5.2) 32.4 (11.4) 0.004

BE �7.18 (7.5) �6.9 (6.6) �7.2 (7.6) 0.188

ED: emergency department, SBP: systolic blood pressure, PR: pulse rate, GCS:

Glasgow Coma Scale, ISS: Injury Severity Score, RTS: Revised Trauma Score, TRISS:

Trauma and Injury Severity Score, BE: base excess, SD: standard deviation. The italic

value in table is meant to highlight the significant values (p < 0.05).

218 Liver injury

patients

Unstable* (96 )Stable (122)

OR

-Generalize d

periton itis (24)

- SW wit h

evisce ration, GI

blee ding, hem aturia

(20)

-GSW through the

abdo men** (2)

CT scan (76)

Injuries requiring

surgery?***

No (64)

Yes (12)

NOM att empted

(64)Fail ed (6)Succes sful

(58)

No obv ious

indic ations for laparotomy

Fig. 1. Management of liver injury patients. SW: stab wound, GI bleeding:

gastrointestinal bleeding, GSW: gunshot wound, CT: computed tomography, NOM:

non-operative management, OR: operating room. *Unstable was defined as

persistent hypotension (systolic blood pressure <90 mmHg) not responding to

the initial fluid resuscitation. **All gunshot wounds through the abdomen other

than an isolated GSW to the right upper quadrant. ***Injuries requiring surgery

included pancreatic and/or hollow viscus injuries identified on CT scan.

S. Prichayudh et al. / Injury, Int. J. Care Injured xxx (2014) xxx–xxx2

G Model


Patients and methods

A retrospective study was performed on liver injury patients atKing Chulalongkorn Memorial Hospital; a 1300-bed universityhospital and a level 1 trauma center in Bangkok, Thailand, fromMay 2002 to May 2011. The study was approved by ourinstitutional review board. The management of abdominal injurypatients in our institution depends mainly on the patient’shemodynamic status, the presence of generalised peritonitis, andthe mechanism of injury. The patients with (1) persistent hypoten-sion (systolic blood pressure <90 mmHg not responding to theinitial fluid resuscitation) and/or (2) generalised peritonitis wouldimmediately undergo operative management (OM). In penetratingtrauma, the additional indications for exploratory laparotomy were(1) stab wound (SW) with evisceration, gastrointestinal (GI)bleeding and hematuria; and (2) all gunshot wounds (GSWs)through the abdomen other than an isolated GSW to the right upperquadrant. SW patients without indications for operation wouldundergo serial physical examination. Stable patients with anisolated GSW or SW to the right upper quadrant and stable blunttrauma patients would be considered candidates for NOM andwould be sent for computed tomography scan (CT) of the abdomen.The patients with positive CT findings for pancreatic and/or hollowviscus injuries would also undergo OM, while all other patients withliver injury identifying on CT would be managed non-operatively.The diagnosis of liver injury in the present study was made based oneither intra-operative findings or CT findings. The severity of liverinjuries was graded by the American Association for the Surgery ofTrauma-Organ Injury Scale [15].

NOM protocol in our institution consisted of continuousmonitoring of vital signs, serial physical examination, and serialblood tests for hematocrit and amylase every 6 h. The indicationsfor angiography with possible AE in our institution are (1) contrastextravasation in the injured liver demonstrated by CT scan, and (2)the suspicion of persistent bleeding from hepatic artery in patientsundergoing either NOM or OM [11]. DCS was defined as anabbreviated operative intervention performed in liver injurypatients including perihepatic packing and temporary abdominalclosure. The decision to perform DCS in liver injury patients wasmade when other simpler maneuvers failed to control the bleedingand when the patient’s physiology became deranged (i.e., whenhypothermia, acidosis, and coagulopathy occurred).

Data collection included demographics data, emergencydepartment (ED) parameters (vital signs, trauma scores, hemato-crit, and base deficit), detail of liver injuries, type of management,and outcomes in terms of blood component transfusion, complica-tions, Intensive Care Unit (ICU) days, ventilator days, length of stayand mortality. Statistical analysis was done by the Window SPSSprogram version 17.0 with the statistical significance was set atp < 0.05. Univariate analysis was performed with the chi-squaredtest for comparison of categorical variables, and the Student’s t testfor comparison of continuous variables. The non-parametric(Mann–Whitney U) test was used for comparison of non-parametric variables (blood transfusion, ICU days, ventilator days,and length of stay). Multivariable analysis for the mutuallyindependent predictors for the need of OM and DCS was doneby a stepwise logistic regression.

Results

There were 218 patients with liver injury identified during May2002 to May 2011 (156 blunt and 62 penetrating). One hundredeighty-six patients were male and 32 were female, the mean agewas 30.4 years. Sixty-four patients were candidates for NOM(29.4%, NOM group). In the OM group, there were 154 patients(70.6%) operated on initially due to hemodynamic instability (96);


peritonitis (24); stab wound with evisceration, GI bleeding, andhematuria (20); gunshot wound through the abdomen (2); andpositive CT scan for hollow viscus and/or pancreatic injuries (12).The demographic data and the ED parameters were shown inTable 1. The management of liver injuries was summarised inFig. 1.

The diagnoses of liver injury were made intraoperatively in 142patients undergoing immediate laparotomy and by CT scan in 76patients. The details of liver injury were shown in Table 2. Theassociated intraabdominal injuries were also shown in Table 2.Two patients in the NOM group were suspected to have rightdiaphragmatic injuries from CT finding (diaphragmatic irregulari-ty) and were managed conservatively.

Angiography was performed in 47 patients (14 in NOM group,33 in OM group) and 40 patients received AE (10 in NOM group, 30in OM group). The indications for angiography were contrast



Table 2The details of liver injuries.

Total (n = 218) NOM (n = 64) OM (n = 154) p-value

Gradea

I 18 (8.3%) 5 (7.8%) 13 (8.4%) 0.654

II 84 (38.5%) 19 (29.8%) 65 (42.2%)

III 60 (27.5%) 28 (43.8%) 32 (20.8%)

IV 32 (14.7%) 10 (15.6%) 22 (14.3%)

V 24 (11.0%) 2 (3.1%) 22 (14.3%)

Anatomy

Left 54 (24.8%) 12 (18.8%) 42 (27.3%) 0.249

Right 138 (63.3%) 46 (71.9%) 92 (59.7%) 0.124

Combined 26 (11.9%) 6 (9.3%) 20 (13.0%) 0.603

Juxtahepatic IVC involvement 13 (6.0%) 0 (0) 13 (8.4%) 0.168

Associated Intraabominal injuries Diaphragm 31 (14.2%) Diaphragm 2 (3.1%) Diaphragm 29 (18.8%) 0.005

Kidney 20 (9.2%) Kidney 6 (9.4%) Kidney 14 (9.1%) 0.841

Spleen 22 (10.1%) Spleen 6 (9.4%) Spleen 22 (14.3%) 0.446

Pancreas 12 (5.5%) Pancreas 12 (7.8%) 0.027

Stomach 8 (3.7%) Stomach 8 (5.2%) 0.117

Duodenum 9 (4.1%) Duodenum 9 (5.8%) 0.080

Small bowel 6 (2.7%) Small bowel 6 (3.9%) 0.241

Colon 18 (8.2%) Colon 18 (11.7%) 0.003

Gall bladder 6 (2.8%) Gall bladder 6 (3.9%) 0.241

Urinary bladder 1 (0.5%) Urinary bladder 1 (0.6%) 1

Abdominal vascular 7 (3.2%) Abdominal vascular 7 (4.5%) 0.168

IVC: Inferior Vena Cava, N/A: non applicable. The italic value in table is meant to highlight the significant values (p < 0.05).a Graded by the American Association for the Surgery of Trauma-Organ Injury Scale.

Table 3Outcomes of liver injury patients.

Total

(n = 218)

NOM

(n = 64)

OM

(n = 154)

p-value

Blood transfusion (Median (IQR))

PRC 4 (0–12) 1 (0–4) 6 (2–19) <0.001

FFP 2 (0–12) 0 (0–2) 4 (0–20) <0.001

Platelets 0 (0–10) 0 (0) 0 (0–15) <0.001

Activated factor VII 7 (3.2%) 0 (0%) 7 (4.5%) 0.168

Angiography 47 (21.6%) 14 (21.9%) 33 (21.4%) 0.920

Embolisation 40 (18.3%) 10 (15.6%) 30 (19.5%) 0.632

Complications

Liver relateda 34 (15.6%) 6 (9.3%) 28 (18.2%) 0.153

Other 38 (17.4%) 7 (10.9%) 31 (20.1%) 0.152

ICU days (Median (IQR)) 0 (0–4) 0 (0) 1 (0–5) <0.001

Ventilator days (Median (IQR)) 0 (0–4) 0 (0) 2 (0–5) <0.001

LOS (Median (IQR)) 10 (6–21) 11 (8–16) 9.5 (4–27) 0.935

Mortality 38 (17.4%) 1 (1.6%) 37 (24.0%) <0.001

Causes of death

Exsanguinations 23 (10.6%) 0 (0%) 23 (14.9%)

MOF 6 (2.8%) 0 (0%) 6 (3.9%)

Cardiac failure 2 (0.9%) 1(1.6%) 1 (0.6%)

Severe head injury 7 (3.2%) 0 (0%) 7 (4.5%)

PRC: Packed Red blood Cells, FFP: Fresh Frozen Plasma, ICU: Intensive Care Unit,

LOS: Length of stay, IQR: inter-quartile range, SD: standard deviation. The italic

value in table is meant to highlight the significant values (p < 0.05).a Forty liver related complications occurring in 34 patients comprised false

aneurysm formation/hemobilia (7), bile leak/intraabdominal collection (14),

abdominal compartment syndrome (9), liver necrosis (8), and liver failure (2).

S. Prichayudh et al. / Injury, Int. J. Care Injured xxx (2014) xxx–xxx 3

G Model


extravasation in 14, persistent bleeding in patients undergoingNOM in 4, and persistent bleeding in patients undergoing OM and/or DCS in 29. We performed angiography with possible embolisa-tion prior to packing removal in 24 out of 45 patients undergoingperihepatic packing (DCS), 21 patients in this group received AE.

Of the 64 patients in NOM group, 58 (90.6%) underwentsuccessful NOM, while the other 6 patients (9.4%) requiredsubsequent laparotomy due to delayed hemorrhage (4) anddevelopment of peritonitis on serial examination (2). Among the4 patients with delayed hemorrhage, 3 had active bleeding fromliver injuries and 1 had active bleeding from a kidney injury. Twopatients who developed peritonitis on serial examination receivednon therapeutic laparotomy (no hollow viscus/pancreatic injuriesfound). Eleven out of 12 patients with severe liver injuries (grade 4and 5) in the NOM group were successfully managed non-operatively. NOM was also carried out successfully in 4 patientssustaining penetrating injuries (2 SWs and 2 GSWs to the rightupper quadrant). As mentioned above, we used AE as an adjunct toNOM in 10 patients. Only one patient died in the NOM group due tosudden cardiac arrest (mortality rate 1.6%).

In the OM group, the patients had significantly higherpenetrating mechanism, pulse rate, Injury Severity Score (ISS);and significantly lower systolic blood pressure, Glasgow ComaScore (GCS), Revised Trauma Score (RTS), Trauma and InjurySeverity Score (TRISS), hematocrit than in the NOM group (Table 1).As shown in Table 3, the OM group patients also requires moreblood product transfusions; had longer ICU days, and ventilatordays than the NOM group. Seven patients in the OM group receivedactivated factor VII, 4 survived and 3 died of multiple organ failure(2) and severe head injury (1). Three out of 24 patients operated onfor generalised peritonitis had negative laparotomy. The overallmortality rate was 17.4%, the mortality rate was significantlyhigher in the OM group than in the NOM group (24% vs. 1.6%,p < 0.001, OR 19.92, 95% CI 2.7–148.7). Two patients in the OMgroup undergoing ED thoracotomy with subsequent laparotomydied from exsanguinations within the first 24 h. The causes ofdeath were also demonstrated in Table 3.

DCS (perihepatic packing and temporary abdominal closure)was utilised in 45 patients in the OM group. Thirteen patients died(exsanguinations in 8, multi organ failure in 3, and severe head


injury in 2) before packing removal and abdominal closure couldbe achieved, while 32 patients survived to receive packing removaland abdominal closure. Abdominal closure techniques used in thepresent study included absorbable mesh closure (plan ventralhernia) in 26, delayed primary fascial closure in 1, and componentseparation in 5, with the mean closure time of 4.2 days. Fivepatients suffered multiple organ failure and subsequently de-ceased after receiving abdominal closure, resulting in the mortalityrate of 40% in the DCS patients.



Table 4Mutually independent predictors of the need for operation.

Step Variable Adjusted OR (95% CI) p-value

1 GCS < 13 42.46 (5.7–314.2) <0.001

2 Penetrating mechanism 9.06 (3.1–26.2) <0.001

3 PR � 100 2.95 (1.6–5.74) <0.001

4 SBP � 90 4.72 (1.8–12.5) 0.022

Stepwise regression: variables considered for the model were sex (p = 0.073),

mechanism, systolic blood pressure (SBP), pulse rate (PR), GCS, ISS (p = 0.747), RTS

(p = 0.944), TRISS (p = 0.606), and hematocrit (p = 0.459).

The non-significant p-values represent the p-value if each variable was added in

turn to the model shown.

Table 6Mutually independent predictors of the need for DCS.

Step Variable Adjusted OR (95% CI) p

1 Grade � 4 6.92 (3.2– 15.1) <0.001

2 PR � 100 3.92 (17–9.2) 0.014

3 Blunt mechanism 6.38 (2.5–16.3) 0.037

Stepwise regression: variables considered for the model were mechanism, systolic

blood pressure (p = 0.985), pulse rate (PR), GCS (p = 0.701), ISS (p = 0.249), RTS

(p = 0.524), TRISS (p = 0.832), and grade of liver injuries.

The non-significant p-values represent the p-value if each variable was added in

turn to the model shown.

S. Prichayudh et al. / Injury, Int. J. Care Injured xxx (2014) xxx–xxx4

G Model


All 13 Juxtahepatic Inferior Vena Cava (IVC) injuries (Grade 5vascular liver injuries) were managed operatively with a mortalityrate of 61.5% (8 died). Of the 5 patients who survived, 2 underwentonly perihepatic packing, while 3 patients underwent combinedlaparotomy and median sternotomy to expose and repair thejuxtahepatic IVC with a successful outcome. Atriocaval shunt wasutilised in a 15-year-old male sustaining motorcycle accident withjuxtahepatic IVC injury who had undergone 3 previous operationsand 1 AE in the first 24 h that failed to definitively control thebleeding. After a period of resuscitation in the ICU, the physiologicderangements were corrected and he eventually received atriocavalshunt on the 4th operation to facilitate control and repair of thebleeding IVC. The patient survived and was discharged on postoperative day 84.

The multivariable analysis demonstrated that the mutuallyindependent predictors for the need of operative managementwere GCS < 13 (p < 0.001, OR 42.46), penetrating mechanism(p < 0.001, OR 9.06), tachycardia (p < 0.001, OR 2.95) andhypotension (p = 0.022, OR 4.72) (Table 4). The univariable analysisshowed that the patients who needed DCS had more bluntmechanism, lower systolic blood pressure, higher pulse rate,higher ISS, lower TRISS, and higher numbers of severe liver injuries(grade � 4) than those who did not undergo DCS (Table 5). Themutually independent predictors for DCS demonstrated by amultivariable analysis were high grade (�4) liver injury (p < 0.001,OR 6.92), tachycardia (p = 0.014, OR 3.92), and blunt mechanism(p = 0.037, OR 6.38) as shown in Table 6.

Table 5Univariate analysis of factors associated with the need of Damage Control Surgery

(DCS).

OM group

(n = 154)

No DCS

(n = 109)

DCS

(n = 45)

p-value

Sex

Male 136 (88.3%) 98 (89.9%) 38 (84.4%) 0.494

Female 18 (11.7%) 11 (10.1%) 7 (15.6%)

Age (Mean (SD)) 29.9 (11.8) 29.4 (11.5) 31.1 (12.5) 0.689

Mechanism

Blunt 96 (62.3%) 57 (52.1%) 39 (86.7%) <0.001

Penetrating 58 (37.7%) 52 (47.9%) 6 (13.3%)

ED parameters

(Mean (SD))

SBP 106.9 (36.1) 110.4 (34.6) 98 (38.5) 0.038

PR 100.5 (28.3) 97.2 (28.0) 108.9 (27.6) 0.027

GCS 11.8 (4.4) 12.2 (4.3) 10.8 (4.6) 0.080

ISS 27.8 (15.3) 24.7 (14.4) 35.6 (14.9) <0.001

RTS 6.6 (2.0) 6.8 (1.9) 6.1 (2.1) 0.099

TRISS 80.1 (3.2) 84.3 (28.8) 71.9 (35.0) 0.037

Hematocrit 32.5 (11.3) 33.2 (11.5) 30.9 (10.6) 0.290

BE �7.2 (7.7) �4.8 (5.7) �9.3 (8.7) 0.201

Grade I 4 44 (28.6%) 18 (16.5%) 26 (57.8%) <0.001

ED: emergency department, SBP: systolic blood pressure, PR: pulse rate, ISS: Injury

Severity Score, RTS: Revised Trauma Score, TRISS: Trauma and Injury Severity Score,

BE: base excess. The italic value in table meant is to highlight the significant values

(p < 0.05).


Discussion

The increase in evidences has resulted in the increased use ofNOM in liver injuries, especially in blunt trauma which yieldedhigh success rates (88–98%) [1–7,16]. The application of NOM hasalso been expanded to penetrating abdominal trauma, in whichNOM could be performed in 33% of penetrating liver injury patientswith a success rate of 86% in one series [17]. CT scan of theabdomen is crucial when considering NOM in liver injury patients,since it helps grade the liver injuries and identify associatedintraabdominal organ injuries [1–7,17,18]. In the present study,the patients were considered candidates for NOM when they werehemodynamically stable and had no peritonitis nor otherindications for surgery. With both penetrating (28.4%) and blunt(71.6%) mechanisms in our series, NOM could be carried out in29.4% of the liver injury patients (38.5% in blunt and 6.3%% inpenetrating trauma). The low NOM rate in the present study maybe explained by a higher proportion of the hemodynamicallyunstable patients in our series (44% overall and 44% in blunt injury)since significant proportion of the patients (36%) had been referredfrom other hospitals to our institution due to the high severity ofinjuries. The severe liver injury (grade 4 and 5) per se is notconsidered a contraindication for NOM as demonstrated in thepresent study (success rate of NOM was 91.7% in severe liver injurypatients) and in the previous reports.[1–7,16]. The success rate ofNOM in this study was 90.6%, which is comparable with otherstudies.

The authors used angiography as an adjunct to both NOM(22%) and OM (21%) of liver injury patients who had contrastextravasation detected on CT or had signs of ongoing hemor-rhage. The AE was performed in 40 out of 47 patients undergoingangiography, resulting in the therapeutic angiography rate of85%, which is corresponding to the previously reported rates inthe recent studies (50–75%) [19,20]. Angiography was utilisedmore frequently after DCS (53%) and AE was done in 21 out of 24patients, leading to the higher therapeutic angiography rate(88%) in the DCS group. Hence, AE should be considered as animportant adjunct to DCS in the management of liver injurypatients.

Several studies have proposed the factors associated with theneed for operation and failure of NOM including increasing age,male sex, hypotension, low GCS, high ISS, low TRISS, low initialplatelet count, high fluid requirement, and concomitant abdominaltrauma [6,7,16,21]. In the present study, the authors identified lowGCS, penetrating mechanism, tachycardia, and hypotension asmutually independent predictors for the need of operation. Theseparameters are in agreement with the previous studies.

Since the introduction of Damage Control in the early 1980s, it hasbeen generally accepted that patients with severe injury andphysiologic derangements(i.e., hypothermia, acidosis, coagulopathy)are likely to benefit from DCS [8,10,22,23]. As mentioned before, thedecision to perform DCS in our institution was made intraoperativelywhen other simpler maneuvers failed to control the bleeding andwhen the patient’s physiology became deranged. In the present



S. Prichayudh et al. / Injury, Int. J. Care Injured xxx (2014) xxx–xxx 5

G Model


study, the mutually independent predictors for DCS were high grade(�4) liver injury, tachycardia, and blunt mechanism. Although ourblunt liver injury patients were less likely to require operation, whenthey did they tended to need more DCS (41% of blunt injury patients)than in those with penetrating injury (10%). This may be explained byhigher grade of liver injury in blunt injury patients (mean liver injurygrade of 3.1) than in penetrating injury patients (mean liver injurygrade of 2.3) in the OM group. Nevertheless, the lack of theintraoperative physiologic parameters (i.e., coagulogram, blood pH,and body temperature) and the retrospective nature of the presentstudy making our analysis arguable and a prospective study shouldbe performed on this subject, particularly in the liver injury patients,to clarify the issue.

The operative management of juxtahepatic IVC injury remainedvery challenging since the inaccessibility of the juxtahepatic IVCmakes it extremely difficult to expose and repair the injury properly.The mortality rates of juxtahepatic IVC injury range from 33% to 81%[3,11,24–26], which correlate well with the present study (mortalityrate 61.5%). In one series, perihepatic packing could be used tosuccessfully control the bleeding in 32% of the blunt juxtahepatic IVCinjury patients, while failure of packing led to direct repair andresulted in a high mortality rate of 55% [26]. Yellin, et al. proposedthe use of combined laparotomy and right thoracotomy withdiaphragm divided down to the vena caval hiatus to obtain control ofjuxtahepatic IVC injury [27]. We found that extending thelaparotomy incision to a median sternotomy and cutting thediaphragm directly downward to the vena caval hiatus also provideda good exposure to control the intrapericardial IVC and repair thejuxtahepatic IVC, as performed in 2 patients in our series.

Since the introduction of the atriocaval shunt by Schrock, et al.in 1968, it has not been widely used despite the theoreticaladvantage of preserving the venous return during the IVC repairbecause of the complexity of the insertion technique and the highmortality rates associated with its use (50–91%) [3,11,25–28]. Webelieve that the insertion of an atriocaval shunt should be reservedonly for relatively stable patients that other modalities have failedto definitively control the bleeding. In the present study, thepatient who underwent atriocaval shunt insertion had undergoneperihepatic packing to temporarily control the bleeding and hadalready been resuscitated to correct physiologic derangementsprior to the procedure. To make the shunt work properly,preparation of the shunt insertion should be performed underan unhurried condition. The bleeding from the injured hepatic veinor IVC should have been effectively controlled with packing untilthe shunt is already inserted. This will allow a safe removal of thepacking and a good opportunity for repairing the injury.

In conclusion, the management of liver injuries requires amultidisciplinary approach, i.e. a combination of NOM, OM bysurgeons, and AE by interventional radiologists. Since the NOMcarries lower mortality rate, it should be attempted in hemody-namically stable patients without other indications for surgery.The patients with low GCS, penetrating mechanism, tachycardiaand hypotension are more likely to undergo surgery; while thosewith high grade liver injury, tachycardia, and blunt mechanism aremore likely to require DCS. When operating on the severe liverinjury patients, the surgeons should be familiar with the conceptand techniques of DCS in order to provide the best possibleoutcome.

Conflict of interest

We hereby certify that there is no conflict of interest in ourstudy.


References

[1] Croce MA, Fabian TC, Menke PG, Waddle-Smith L, Minard G, Kudsk KA, et al.Non-operative management of blunt hepatic trauma is the treatment of choicefor hemodynamically stable patients: results of a prospective trial. Ann Surg1995;221:744–55.

[2] Pachter HL, Knudson MM, Esrig B, Ross S, Hoyt D, Cogbill T, et al. Status ofnonoperative management of blunt hepatic injuries in 1995: a multicenterexperience with 404 patients. J Trauma 1996;40:31–8.

[3] Richardson JD, Franklin GA, Lukan JK, Carrillo EH, Spain DA, Miller FB, et al.Evolution in the management of hepatic trauma: a 25-year perspective. AnnSurg 2000;232:324–30.

[4] Petrowsky H, Raeder S, Zuercher L, Platz A, Simmen HP, Puhan MA, et al. Aquarter century experience in liver trauma: a plea for early computed tomog-raphy and conservative management for all hemodynamically stable patients.World J Surg 2012;36:247–54.

[5] van der Wilden GM, Velmahos GC, Emhoff T, Brancato S, Adams C, Georgakis G,et al. Successful nonoperative management of the most severe blunt liverinjuries: a multicenter study of the research consortium of New Englandcenters for trauma. Arch Surg 2012;147:423–8.

[6] Leppaniemi AK, Mentula PJ, Streng MH, Koivikko MP, Handolin LE. Severehepatic trauma: nonoperative management, definitive repair, or damagecontrol surgery? World J Surg 2011;35:2643–9.

[7] Markogiannakis H, Sanidas E, Michalakis I, Manouras A, Melissas J, Tsiftsis D.Predictive factors of operative or nonoperative management of blunt hepatictrauma. Minerva Chir 2008;63:223–8.

[8] Stone HH, Strom PR, Mullins RJ. Management of the major coagulopathy withonset during laparotomy. Ann Surg 1983;197:532–5.

[9] Feliciano DV, Mattox KL, Burch JM, Bitondo CG, Jordan Jr GL. Packing forcontrol of hepatic hemorrhage. J Trauma 1986;26:738–43.

[10] Rotondo MF, Schwab CW, McGonigal MD, Phillips 3rd GR, Fruchterman TM,Kauder DR, et al. ‘‘Damage control’’: an approach for improved survival inexsanguinating penetrating abdominal injury. J Trauma 1993;35:375–82.

[11] Asensio JA, Roldan G, Petrone P, Rojo E, Tillou A, Kuncir E, et al. Operativemanagement and outcomes in 103 AAST-OIS grades IV and V complex hepaticinjuries: trauma surgeons still need to operate, but angioembolization helps. JTrauma 2003;54:647–53.

[12] Sriussadaporn S, Pak-art R, Tharavej C, Sirichindakul B, Chiamananthapong S. Amultidisciplinary approach in the management of hepatic injuries. Injury2002;33:309–15.

[13] Sikhondze WL, Madiba TE, Naidoo NM, Muckart DJ. Predictors of outcome inpatients requiring surgery for liver trauma. Injury 2007;38:65–70.

[14] Lin BC, Fang JF, Chen RJ, Wong YC, Hsu YP. Surgical management and outcomeof blunt major liver injuries: Experience of damage control laparotomy withperihepatic packing in one trauma centre. Injury 2014;45:122–7.

[15] Moore EE, Cogbill TH, Jurkovich GJ, Shackford SR, Malangoni MA, ChampionHR. Organ injury scaling: spleen and liver (1994 revision). J Trauma1995;38:323–4.

[16] Polanco PM, Brown JB, Puyana JC, Billiar TR, Peitzman AB, Sperry JL. Theswinging pendulum: A national perspective of non operative management insevere blunt liver injury. J Trauma Acute Care Surg 2013;75:590–5.

[17] Demetriades D, Hadjizacharia P, Constantinou C, Brown C, Inaba K, Rhee P,et al. Selective nonoperative management of penetrating abdominal solidorgan injuries. Ann Surg 2006;244:620–8.

[18] Velmahos GC, Constantinou C, Tillou A, Brown CV, Salim A, Demetriades D.Abdominal computed tomographic scan for patients with gunshot wounds tothe abdomen selected for nonperative management. J Trauma 2005;59:1155–60.

[19] Johnson JW, Gracias VH, Gupta R, Guillamondegui O, Reilly PM, Shapiro MB,et al. Hepatic angiography in patients undergoing damage control laparotomy.J Trauma 2002;52:1102–6.

[20] Misselbeck TS, Teicher EJ, Cipolle MD, Pasquale MD, Shah KT, Dangleben DA,et al. Hepatic angioembolization in trauma patients: indications and compli-cations. J Trauma 2009;67:769–73.

[21] Duane TM, Como JJ, Bochicchio GV, Scalea TM. Reevaluating the managementand outcomes of severe blunt liver injury. J Trauma 2004;57:494–500.

[22] Rotondo MF, Zonies DH. The damage control sequence and underlying logic.Surg Clin North Am 1997;77:761–77.

[23] Asensio JA, McDuffie L, Petrone P, Roldan G, Forno W, Gambaro E, et al. Reliablevariables in the exsanguinated patient which indicate damage control andpredict outcome. Am J Surg 2001;182:743–51.

[24] Pachter HL, Spencer FC, Hofstetter SR, Coppa GF. Experience with the fingerfracture technique to achieve intra-hepatic hemostasis in 75 patients withsevere injuries of the liver. Ann Surg 1983;197:771–8.

[25] Burch JM, Feliciano DV, Mattox KL. The atriocaval shunt. Facts and fiction. AnnSurg 1988;207:555–68.

[26] Liu PP, Chen CL, Cheng YF, Hsieh PM, Tan BL, Jawan B, et al. Use of a refinedoperative strategy in combination with multidisciplinary approach to manageblunt juxtahepatic venous injuries. J Trauma 2005;59:940–5.

[27] Yellin AE, Chaffee CB, Donovan AJ. Vascular isolation in treatment of juxta-hepatic venous injuries. Arch Surg 1971;102:566–73.

[28] Schrock T, Blaisdell FW, Mathewson Jr C. Management of blunt trauma to theliver and hepatic veins. Arch Surg 1968;96:698–704.


http://refhub.elsevier.com/S0020-1383(14)00084-9/sbref0005





















































































Documents

Management of liver injuries: Predictors for the need of operation and damage control surgery