Liver Trauma: Experience in 348 Cases [620648]

Liver Trauma: Experience in 348 Cases
Jing-mou Gao, M.D., Ding-yuan Du, M.D., Xing-ji Zhao, M.D., Guo-long Liu, M.D., Jun Yang, M.D.,
Shan-hong Zhao, M.B., Xi Lin, M.B.
Department of Traumatology, Chongqing Emergency Medical Center, 1 Jiank ang Road, 400014 Chongqing, People’s Republic of China
Published Online: May 13, 2003
Abstract. Liver trauma, the main cause of death in patients suffering ab-
dominal injury, remains an unresolved problem, especially in its most se-vere forms. The objective of this study was to probe effective surgical pro –
cedures and improve the outcome for patients with severe hepatic injury. Aretrospective study of 348 patients with hepatic trauma seen in our insti-tution during the past 12 years was carried out. Of these 348 patients, 259(74.4%) underwent surgery. To manage severe liver trauma (American As-sociation for the Surgery of Trauma grade III to grade V), procedures suchas packing of the laceration with omentum, hepatectomy or direct controlof bleeding vessels within the liver substance by means of the Pringle ma-neuver, selective hepatic artery ligation, retrohepatic caval repair wi th to-
tal hepatic vascular occlusion, and perihepatic packing were selected an d
combined based on the specific injury. In the 259 patients treated opera-tively, the survival rate was 86.9% (225/259); and 15 of 40 with retrohepat ic
venous injury (RHVI) were cured with the maximum blood transfusion of60 units. In 42 patients treated by perihepatic packing, the bleeding wasstopped in 20 of 25 (80%) with RHVI and in 14 of 17 (82%) without suchinjury ( p> 0.75). The percentage of failure of nonoperative management
was 17.2% (17/99); and it was 46.7% (14/30) in patients with grade III–Vinjury. Death occurred in 3 (50%) of 6 failures of grade IV–V injury. Theoverall mortality rate was 11.8% (41/348), and 51% of the deaths were dueto exsanguination. The results suggest that severe hepatic injuries, esp e-
cially grade IV–V injuries, usually require surgical intervention; reas on-
able surgical procedures based on classification of liver trauma and com-bined application of techniques can increase the survival rate; andperihepatic packing is effective in dealing with RHVI.
Liver trauma has been the main cause of death in patients with
severe abdominal injuries, with a mortality rate between 10% and15% [1]. Furthermore, severe hepatic trauma remains an unre-solved problem; especially retrohepatic venous injury (RHVI),main hepatic venous injury, and retrohepatic caval injury (RHCI)alone or in combination, continue to be associated with high mor-tality rates ranging from 50%–100% in the literature, and the de-crease in mortality appears to have leveled off in the past two de-cades [2, 3].
We hypothesized that severe hepatic injuries usually require sur-
gical intervention; that reasonable surgical procedures based onclassification of hepatic injury and combined application of surgicaltechniques might increase the survival rate among such patients;and that perihepatic packing is effective for patients with and with-out RHVI. A retrospective study involving 348 patients with he-patic injuries was therefore carried out to investigate these hypoth-eses.
Materials and Methods
Patients with liver trauma treated at the Department of Trauma-
tology, Chongqing Emergency Medical Center (CEMC, Level Itrauma center), China, from January 1988 to January 2000 werestudied. The charts of patients were reviewed for information re-garding sex, age, cause of injuries, duration of time pre-admission,and time taken to get the patient to the operating room. In addi-tion, we recorded shock state on admission, amount of blood trans-fusion, severity of hepatic injuries (including grade of hepatic injuryand Injury Severity Score [ISS] value), associated intraabdominaland extraabdominal injuries, method of diagnosis, therapeutic pro-cedures, and the outcome (uneventful recovery, postoperativecomplications, mortality and cause of the death).
In this series, liver trauma was proved by surgical exploration or
at autopsy, except in those patients who were treated nonopera-tively whose diagnosis was defined by computed tomography (CT)scan or hepatic angiography. The grade of hepatic injury was estab-lished according to the Liver Injury Scale of the American Associa-tion for the Surgery of Trauma (AAST, OIS) [4]; after reference tovarious authors’ opinions [1, 3, 5, 6], grade III to grade V wereregarded as severe liver trauma. In some cases the grade was re-vised after operative exploration, if the initial assessment based onCT scan was observed to be incorrect.
If the hepatic injury was defined by CT scan with stable hemody-
namics, treatment was confined to nonoperative measures (includ-ing selective hepatic arterial angiography and embolization). If thepatient’s status worsened, surgical intervention was initiated.Emergent laparotomy was indicated by unstable hemodynamics orby a positive abdominal tap, with aspiration of fresh blood soonafter admission or after an initial negative tap. During the opera-tion, one or more surgical procedures were employed.
When perihepatic packing for severe hepatic injuries associated
with RHVI was required, we stressed the need to give special at-tention to the hepatic wound, as well as to the area of RHVI. Pre-
Correspondence to: Jing-mou Gao, M.D., e-mail: lieuky@sina.com
WOR LD
Journal of
SURGERY
© 2003 by the Socie ´te´
Internationale de ChirurgieWorld J. Surg. 27, 703–708, 2003
DOI: 10.1007/s00268-003-6573-z

cise packing is accomplished as follows: (1) Absorbable gelatin
sponges soaked in vasoconstrictor and coagulant are used to fill inthe injured area, especially around the site of the RHVI; (2) theomentum is either partially removed from the left segment of thetransverse colon, or cut in a “ Z” figure providing its blood flow can
be maintained, down to the diaphragm to cover the damaged liveror the resected surface evenly. It is important that omentum alsoclose the area of an RHVI; (3) on top of the omentum, gauze padsare opened flat and spread in layers 2–3 deep to compress bleedinginjuries of the liver and to support the omentum against retrohe-patic cava. The pads are removed 2–3 days after the operation,while the omentum already attached to the injured tissue andhealed to avoid recurrent hemorrhage. To ensure even pressurebetween the liver and the diaphragm, the drainage tubes are placedat a distance from the injured area. Vasoconstrictor and coagulantcan be administered topically through the drainage tube if slightoozing of blood occurs postoperatively.
All survivors of grade III–V injuries treated by operation were
followed for at least 6 months after discharge. The efficacy of treat-ment and the factors related to the deaths were analyzed and Stu-dent’s
/H92732tests were used for statistical analysis. A value of p< 0.05
was considered statistically significant.
Results
From January 1988 to January 2000, 348 patients with liver trauma
were treated in the Department of Traumatology, ChongqingEmergency Medical Center. Among these 348, there were 286males and 62 females with ages ranging from 7 to 73 years (median28). Injury was defined as blunt trauma in 240 patients and pen-etrating trauma in 108 patients. The AAST Liver Injury Scale re-sults were as follows: grade I in 38 patients; grade II, 92 patients;grade III, 105 patients; grade IV, 68 patients; and grade V, 45 pa-tients, including 40 with RHVI, 16 of which were RHCI. Associatedintraabdominal or extraabdominal injuries, or both, in 251 (72%)patients are shown in Table 1.
Shock (systolic blood pressure /H1134980 mmHg) presented in 184
patients (53%) at admission. The median ISS of all 348 patients was27 (range, 4–75). Duration of pre-entrance was from 25 minutes to96 hours (median 2.5 hours).
In 208 patients, liver injury was confirmed only at urgent lapa-
rotomy. In 133 patients with stable hemodynamics, the diagnosis ofhepatic injury was obtained with computed tomography (CT) scanor B-mode echography followed by surgical exploration (51), CTscan followed by hepatic angiography without operation (3), andCT scan without operation (79). In the remaining 7 patients, all ofwhom died before reaching the operating room, the diagnosis wasestablished by autopsy.
Ninety-nine patients with stable hemodynamics at admission
were treated nonoperatively in the first instance. Of these 99 cases,82 were cured including grade I–II (66), grade III (8), grade IV (6),and grade V (2); and among them 3 underwent selective hepaticartery angiography and embolization. The remaining 17 patientsrequired a subsequent laparotomy within 8–24 hours because he-modynamic status worsened. During the operation, 6 with gradeIII–V injury died of excessive blood loss because of the delay. Thepercentage of failure of nonoperative management was 17.2% (17/99); and it was 46.7% (14/30) in grade III–V injuries. In injuriesbelow grade III, 6 of 11 failures were due to other than liver; incontrast, all 6 failures of grade IV–V were attributed to liver hem-orrhage.
Urgent laparotomy was performed in 242 patients soon after ad-
mission (166 within 1 hour of arrival). The indication for urgentintervention was hemodynamic instability, a positive abdominal tapwith aspiration of fresh blood (177), or signs of peritoneal irritationwith other injuries requiring laparotomy (65). A further 17 patientsinitially treated nonoperatively later underwent laparotomy. Of the259 patients treated operatively, 195 had grade III–V injury (97grade III, 55 grade IV, and 43 grade V), and 64 had grade I–IIinjury. In the latter group laparotomy was indicated for treatmentof associated injuries while simple procedures such as suturing orfibrin glue were used to treat the liver trauma. Time from admis-sion to arrival in the operating room ranged from 10 minutes to 24hours (median 45 minutes). Perioperative blood transfusionranged from 2 to 60 units (median 12 units). In 259 patients, 331times of surgical procedures were used. There were 14 anatomichepatic resections including right hemihepatectomy (9), left hemi-hepatectomy (1), left lateral segment resection (1), and segmentec-tomy (3); and 25 debridemental resections were done. The mainoperative methods and the data of hepatectomy are shown in Table2 and 3, respectively.
Of the 225 survivors who underwent laparotomy, 51 had no de-
tectable blood pressure on admission with a maximum blood trans-fusion of 60 units, and 28 had ISS /H1135050 with a maximum ISS of 66.
In 40 patients with RHVI, 6 of 16 (37.5%) with RHCI, and 9 of 24(37.5%) with main hepatic venous injury survived ( p> 1.00,
/H92732=
0). In 42 patients treated by perihepatic packing, bleeding ceased in20 of 25 (80%) with RHVI and in 14 of 17 (82%) without suchinjury ( p> 0.75,
/H92732= 0.044). However, death occurred from extra-
abdominal injuries in 5 patients with RHVI and in 4 without it.Selective hepatic artery ligation in combination with other proce-dures succeeded in controlling the bleeding in all 28 patients but 7of 22 with RHVI; 8 of 21 survivors had a transient rise in SGPT levelTable 1. Associated injuries in 251 patients.
Associated injuries No. of patients
Head 112
Facial 19Chest 131Abdomen (other than liver) 188Bone and joint (spinal, pelvic, and limb) 149Soft tissue (skin, muscle, nerve, and vessel) 164Table 2. Surgical procedures and outcome in 259 patients.
Main surgical procedureNo. ofpatientsNo. ofsurvivors
Hepatorrhaphy 125 124Packing with omentum 43 38Repair and selective hepatic artery ligation 6 6Hepatectomy or intrahepatic hemostasis 36 30Repair and perihepatic packing 11 6Perihepatic packing and selective hepatic artery
ligation22 15
RHCI repair with hemihepatectomy and
perihepatic packing74
RHCI repair with hepatotomy and perihepatic
packing20
RHCI repair with hepatectomy by total hepatic
vascular occlusion72
RHCI: retrohepatic caval injury.704 World J. Surg. Vol. 27, No. 6, June 2003

within a week after this procedure. In all 225 survivors, postopera-
tive complications included blood oozing in 12 and hemobilia in 13,all of which were treated conservatively (2 of the latter treated byangioembolization); subphrenic abscess in 9, all of whom recoveredafter surgical drainage; abdominal compartment syndrome in 4,which cleared once the packing was removed; bile leak in 15, all ofwhom healed spontaneously 3–47 days after operation. The mor-tality rate of the operative management group was 13.1% (34/259).Comparison of groups treated by initial nonoperation and urgentoperation in grade III–V is shown in Table 4.
In this study, the overall mortality rate was 11.8% (41 of 348 pa-
tients) with a mean ISS of 48.6. The grade I and grade II death ratewas 0; grade III, 3.8% (4/105); grade IV, 13.2% (9/68); l grade V,62.2% (28/45). Seven deaths during resuscitation were all grade IVinjury; one was due to head trauma, 6 delayed and exsanguinatedbefore entrance. Of 34 deaths among patients undergoing laparot-omy, 11 were intraoperative (1 head trauma, 10 exsanguination),and 23 were postoperative (10 head trauma, 2 chest trauma, 1 se-vere pelvic fracture, 5 multiple organ dysfunction syndrome, and 5continuous liver bleeding). Overall, liver hemorrhage was the causeof death in 21 of the 41 patients who died (51%), and in 15 of the 34patients who died after undergoing laparotomy (44%). The relativecauses of the 41 deaths are shown in Table 5.
The term of hospitalization ranged from 1 to 122 days (median
26.5 days). The median in the initially nonoperative managementgroup was 25 days, and that in the urgent laparotomy group was27.5 days. Follow-up involved 161 patients with severe liver traumatreated by operation who were followed for 6 months after dis-charge. One patient died of hepatic cancer 8 months later; in an-other patient, “spontaneous gallbladder cut-off” (cholecystic atro-phy due to devascularization) occurred after proper hepatic arteryligation and was proved by cholecystography. All of the remainingpatients were healthy with normal function of liver and kidney.
Discussion
In recent years, nonoperative therapy of liver injuries has become
an acceptable approach to management of hemodynamically stablepatients without associated injury requiring laparotomy, and theindication is extended [6–10]. However, Parks and some other sur-geons recommend that nonoperative management should be initi-ated only for injuries below grade III in patients with stable hemo-dynamics; grade III to grade V injuries usually require surgicalintervention [1]. Strong et al. have an active attitude and recom-mend anatomic resection for severe liver trauma [5]. We agree withthe viewpoints of Parks and Strong; in our opinion, vital signs ofpatients (pulse, blood pressure, etc.) are critical in deciding wheth-er nonoperative management can be applied. In other words, sur-gical intervention is mainly indicated by unstable hemodynamics.Our department, in which two hepatobiliary surgeons are availablefor liver trauma, is a specialized unit for polytrauma. Many patientswith mild liver trauma are admitted to other departments of thishospital such as Department of General Surgery. In the presentstudy from Department of Traumatology, therefore, the patientswith severe liver trauma accounted for 63% (218/348). This per-centage may be higher than other reports. Surgical intervention inthe patients with grade I or grade II injury all was due to severeinjury to other organs requiring laparotomy; in the patients treatednonoperatively who survived, 6 had grade IV injuries and 2 hadgrade V injuried. It has been shown that if the patient is hemody-namically stable, nonoperative management can be used even inTable 3. Comparison of anatomic and debridement hepatectomy.
ResectionNo. of
patients GradeNo. ofsurvivors Complications
Anatomic 14 V with RHCI 6 Blood oozing 1;
subphrenicabscess 1; ACS 1
Debridement 25 V 12, IV 9,
III 419 Blood oozing 11;
hemobilia 6; bileleak 10; subphenicabscess 6; ACS 3
ACS: abdominal compartment syndrome.Table 4. Comparison of groups treated by initial nonoperation and
operation in grade III–V injuries.
ManagementPatientsand gradeDied ofliverbleeding Complications
Initial nonoperation Total 30 Total 6* Total 25 (times)**
III 16 3 Failure
a8 (3 other
injury); blood oozing 1;hemobilia 3; subphrenicabscess 1
IV 10 2 Failure 4; blood oozing 1;
hemobilia 2; bile leak 1
V 4 1 Failure 2; hemobilia 2
Urgent operation Total 181 Total 9 Total 46 (times)
III 89 0 Blood oozing 2; hemobilia
6; bile leak 3;subphrenic abscess 1
IV 51 0 Blood oozing 5; hemobilia
3; bile leak 9;subphrenic abscess 5;ACS 3
V 41 9 Blood oozing 3; hemobilia
1; bile leak 2;subphrenic abscess 2;ACS 1
*Compared with another,
/H92732= 6.67, p< 0.01.
**Compared with another, no significance of statistics in various com-
plications but hemobilia higher ( /H92732= 11.7, p< 0.01); count out failure.
aFailure of nonoperative management.
Table 5. Factors related to mortality.
FactorType ofinjuryNo. ofpatients Deaths
/H92732p
AAST grade I–II 130 0
III–V 218 41 27.71 < 0.005
Injured mechanism Penetrating 108 4
Blunt 240 37 9.83 < 0.005
RHVI No 308 16
Yes 40 25 106.41 < 0.005
Associated injuries No 97 5
Yes 251 36 5.68 < 0.025
Shock No 164 6
Yes 184 35 19.69 < 0.005
Blood transfusion < 2000 228 6
(ml) /H113502000 120 35 53.26 < 0.005
Coagulopathy No 313 19
Yes 35 22 92.28 < 0.005
AAST: American Association for the Surgery of Trauma; RHVI: ret-
rohepatic venous injury.705 Gao et al.: Liver Trauma

certain cases with grade IV or grade V injury, such as huge or mul-
tiple intrahepatic hematomas. Nevertheless, in our series the per-centage of failure of nonoperative management was as high as46.7% (14/30) in grade III–V injuries. Furthermore, Table 4 showsthat the mortality from liver bleeding in the group treated nonop-eratively is significantly higher than in the operative group ( p<
0.01). Of 6 patients with grade IV and V injury ultimately treated bylaparotomy, 3 (50%) died of hepatic excessive hemorrhage due tothe delay. In contrast, none of the 140 patients with grade III and IVinjury who underwent urgent laparotomy died of liver bleeding. Inthat group, all deaths resulted from associated injuries. Compari-son of the various complications that occurred in the two groupsshowed no statistically significant differences, except for an in-creased incidence of hemobilia in the nonoperative group. Failureof nonoperative management is difficult to compare with operativetreatment, but it appears to have increased blood loss and de-creased tolerance to subsequent operation. The results reveal thatoperative therapy is far more valuable in grade III–V injuries, es-pecially grade IV–V, than in grade I–II.
In this study, selective hepatic arterial angiography with emboli-
zation was used in 5 cases, including nonoperative therapy (3), and
management of postoperative complications (2); satisfactory he-mostasis was achieved in every case. This technique is very benefi-cial in the diagnosis and control of expanding hematoma or signifi-cant hemobilia, and is an important feature of both nonoperativeand postoperative management.
To deal with severe liver trauma, temporary control of hemor-
rhage and thorough exposure of the injured liver are the precondi-tions for success. Occlusion of the hepatic pedicle by the Pringlemaneuver can provide a relatively avascular field if RHVI is notpresent. As a general rule, occlusion of the portal triad in normo-thermia is safe for at least 60 minutes without any temporary inter-ruption; this was maintained up to 85 minutes in the report of Hu-guet and co-workers [11]; and a survivor of our group withstood thisprocedure for 110 minutes without any postoperative signs of liverdysfunction. However, we prefer to keep the time of occlusion asshort as possible, because the tolerance of the liver to hypoxia de-creases in hemorrhagic shock. Data cited by Huguet and othersrefer to elective cases only. To expose the liver fully, the ligamen-tous attachments (falciform, right triangular, and coronary) shouldbe divided to allow rotation of the liver to the incision as the ma-jority of severe liver trauma involves right lobe.
Various therapeutic modalities have been suggested as useful
approaches for management of liver trauma during the past 20years. Among these procedures, hepatectomy is frequently used forsevere liver trauma. The benefit of anatomic resection has beenproved by Strong and his associates in their excellent practice; theyhave pointed out that, “When severe parenchymal disruption ispresent, anatomic resection achieves the dual roles of removing thesource of bleeding and the site of necrosis. It leaves a smooth re-sected surface, with viable liver and a low propensity for septic com-plications” [5]. In the present study, all anatomic resections wereused in grade V injuries with RHCI, and the survival rate is notideal. However, Table 3 shows that complications such as postop-erative blood oozing, hemobilia, bile leak, and subphrenic abscessin the survivors undergoing anatomic resection appeared lowerthan in debridement group. The mortality rate of severe livertrauma with RHVI is still high. Sometimes, death is associated withselection of inappropriate surgical procedures. Strong et al. havepointed out that, “The surgical procedure embarked on must beappropriate for the individual injury” [5]. In the present study, we
have emphasized that different methods of operation should be se-lected by AAST(OIS) grade of liver trauma, and in many cases,more than two procedures should be used. Faulty proceduresshould be avoided. For example, if the laceration of liver is deep,hepatorrhaphy may lead t o a a life-threatening upper gastrointes-
tinal bleed that can be catastrophic. In other words, posttraumatichemobilia can develop due to intrahepatic hematoma with bili-vascular fistulae. In this series, one patient sustaining hepatorrha-phy developed multiple intrahepatic hematomas and abruptly lostblood—over 1500 ml within 4 hours postoperatively. Fortunately,she was treated successfully by angioembolization of selected he-patic arteries. In our series, the efficacy of hepatic artery ligationwas confirmed when hemorrhage could be stopped by that maneu-ver. When the bleeding was mainly from retrohepatic veins or theportal vein, however, hepatic artery ligation alone will fail to con-trol the bleeding unless adjunct procedures such as packing areadded. In a word, patients with severe liver trauma will die of con-tinuous bleeding postoperatively, if haemostatic procedures are in-
adequate. In this study, the mortality rate associated with grade IVinjuries was relatively low compared with other reported series; thisresult might be attributable to reasonable selection of the surgicalprocedures.
Many high-risk trauma patients with abdominal injuries are af-
flicted with severe damage to the liver, frequently complicated byRHVI. A significant number of such patients tend to develop con-sumption coagulopathy. Under these conditions, complex surgicalprocedures such as hepatic bypass and total hepatic vascular occlu-sion are virtually impossible to perform with success [12]. Althoughhepatotomy or hemihepatectomy may occasionally be performedwith or without bypass or total hepatic vascular occlusion to exposeand repair RHVI directly, perihepatic packing is a well acceptedtechnique for severe liver trauma with or without RHVI when rou-tine procedures cannot control the bleeding. Selective hepatic ar-tery ligation can be used as an adjunct to this procedure. Perihe-patic packing, which was abandoned after World War II, was putforward once again, as uncontrollable bleeding from severe livertrauma remains an unresolved problem [13–17]. Beal and othersrecommended perihepatic packing as a definitive hemostatic pro-cedure. They first selected it for the treatment of severe livertrauma with or without RHVI; their experience proved that peri-hepatic pressures can stanch low-pressure venous hemorrhage [13,14]. Our practice supported Beal’s viewpoint; as for the hemostaticefficacy of packing, however, there wasn’t a statistically significantdifference ( p> 0.75) between subgroups with and without RHVI.
All perihepatic packing incorporated other definitive proceduressuch as hepatectomy, selective hepatic artery ligation, or even re-pair of RHVI; we never used it alone. When the packs were re-moved, therefore, no additional surgical management of livertrauma was generally needed, and significant bleeding rarely re-curred because the omentum had already attached to the injuredtissue and healed. Well executed perihepatic packing should alsocare for the area of RHVI. When the Pringle maneuver fails tocontrol bleeding from the retrohepatic area, then RHVI should beconsidered. In this case, it is unwise to turn the liver over for inspec-tion of the site of RHVI; such handling will lead to life-threateningexsanguination. The correct approach is to push back and up (spineand diaphragm) the liver to stop the hemorrhage temporarily. Thisshould be followed by prompt removal of nonviable hepatic paren-chyma and careful perihepatic packing. Only when hemorrhage
706 World J. Surg. Vol. 27, No. 6, June 2003

cannot be controlled with packs, are complex procedures such as
total vascular occlusion or bypass required. Based on our experi-ence, packing (laparotomic pads) is removed 2–3 days after opera-tion. This timing appears to be reasonable, corresponding with therecommendation of both Caruso and Abikhaled and their co-workers with this timing, neither recurrent hemorrhage caused byremoval of the packing nor complications from prolonged packingdevelop easily [15, 16]. The most common complication associatedwith perihepatic packing is subphrenic abscess. Special attentionshould be given to prevention or early detection and drainage.Other complications have been mentioned by some surgeons, in-cluding renal failure from excessive compression of the inferiorvena cava due to packing or cardiopulmonary dysfunction becauseof the rise of diaphragm [18]. In addition, the abdominal compart-ment syndrome has been reported by many surgeons in recentyears, and this can lead to similar and other complications [19, 20].In our opinion, these complications could be avoided or decreasedby proper techniques. In our group of patients treated with perihe-patic packing, 4 patients developed abdominal compartment syn-drome. In every case the complications were relieved by immediateremoval of the packing.
In 7 of our patients with RHCI repaired under total hepatic vas-
cular occlusion, only 2 survived. We did this procedure with clamp-ing of the portal triad and the supra- and infrahepatic vena cavawithout clamping aorta or performing venovenous bypass. None ofthese patients developed cardiac arrest at the time of clamping.Some authors believe that clamping the vena cava can lead to car-diac arrest because of the sudden decrease in cardiac preload; theysuggest clamping the aorta simultaneously to increase cardiac post-load [21, 22]. We found that there was not a great change in theblood flow returning to the right atrium before and after clampingthe vena cava because of prior exsanguination from an injured ret-rohepatic cava in these moribund patients; therefore, the risk ofcardiac arrest was relatively low. On the other hand, declamping ofthe aorta still can lead to lethal risk from the abrupt washout [2, 22].
Perihepatic drainage with both rubber tube and cigarette used in
grade IV and grade V injury is helpful in decreasing the incidenceof subphrenic abscess and in observing oozing of blood and bile.Additional cigarettes, which increase the early efficacy of drainage,are removed 48 hours after the operation; and the tube is main-tained as closed suction drainage at least a week if bile leakage doesnot occur or the drainage is prolonged. Closed drains are more ben-eficial than open ones because they provide an accurate record ofthe amount of fluid evacuated and they decrease septic complica-tions. In addition, vasoconstrictor and coagulant can be adminis-tered topically through the drain tube, if oozing does occur.
Conclusions
Our hypotheses are demonstrated by the results of this study. Se-
vere hepatic injuries, especially grade IV and grade V, usually re-quire surgical intervention. Reasonable surgical procedures basedon classification of liver trauma and enforcing the combined use ofmultiple techniques can improve the survival rate of such patients.Perihepatic packing is an available option when circumstances aresuch that some complex surgical procedures cannot be carried outfor treating RHVI; when packing is used, special attention shouldbe given to the hepatic wound as well the area of RHVI.
Re´sume ´. Les le ´sions traumatiques du foie sont la cause principale de
mortalite ´ chez le traumatise ´ de l’abdomen. Parmi celles-ci, les le ´sionsgraves (grades III a ` V selon la classification AAST) restent un proble `me
majeur, non comple `tement re ´solu. Le but de cette e ´tude a e ´te´ d’analyser
diffe´rents proce ´de´s chirurgicaux et d’essayer d’ame ´liorer l’e ´volution des
patients victimes de traumatisme he ´patique se ´ve`re. Dans cette e ´tude
re´trospective de 348 patients victimes de traumatismes dans les 12
dernie `res anne ´es, 259 patients (74.4%) ont e ´te´ ope´re´s. Une large gamme de
proce ´de´ sae ´te´ utilise ´e pour traiter ces le ´sions graves, comprenant
l’omentoplastie, inse ´re´e dans les plaies profondes, l’he ´patectomie ou le
contro ˆle direct des vaisseaux saignant sous clampage du pe ´dicule
he´patique (manoeuvre de Pringle), la ligature de l’arte `re he ´patique, la
re´paration de la veine cave re ´trohe ´patique sous exclusion he ´patique totale
ou un packing pe ´rihe´patique, parfois combine ´s, selon les cas. Chez les 259
patients traite ´s, le taux de survie a e ´te´ de 86.9% (225/259); 15 des 40
patients avec une le ´sion veineuse re ´trohe ´patique (LVRH) ont gue ´ri, parfois
ne´cessitant jusqu’a ` 60 unite ´s de sang. Chez 42 victimes traite ´es par packing
pe´rihe´patique, on a pu arre ˆter l’he ´morragie chez 20 des 25 (80%) avec
LVRH et chez 14 des 17 (82%) sans LVRH ( p> 0.75). Le pourcentage
d’e´chec en cas de traitement non-ope ´ratoire a e ´te´ de 17.2% (17/99); 46.7%
(14/30) en cas de le ´sions grades III–V. Trois des six e ´checs de traitement
non-ope ´ratoire (le ´sons de grade IV–V) sont de ´ce´de´s. La mortalite ´ globale a
e´te´ de 11.8% (41/348) et 51% des de ´ce`s ont e ´te´ en rapport avec
l’exsanguination. Ces re ´sultats sugge `rent que les le ´sions he ´patiques
graves, surtout des grades IV–V, ne ´cessitent souvent une intervention
chirurgicale; la combinaison des proce ´de´s, utilise ´s en accord avec la
classification des le ´sions, pourrait ame ´liorer le taux de survie: la technique
de packing pe ´rihe´patique est efficace dans le traitement des LVRH.
Resumen. En los traumatizados de abdomen, la mortalidad viene dada,
principalmente, por la gravedad de las lesiones hepa ´ticas, tanto ma ´s
cuanto que el tratamiento de los traumatismos graves del h ı´gado sigue
siendo un problema irresuelto. El objetivo del trabajo fue comprobar laeficacia de diversos procedimientos quiru ´rgicos en pacientes con graves
traumatismos hepa ´ticos. Se efectuo ´ un estudio retrospectivo de 12 an ˜os,
recopila ´ndose 348 pacientes con traumatismos graves de h ı´gado. 259
(74.4%) fueron tratados quiru ´rgicamente. En el tratamiento de lesiones
hepa´ticas graves (AAST grados III–V) se emplearon diversos procederes
quiru´rgicos solos o combinados, tales como: taponamiento de la
dislaceracio ´n hepa ´tica con epiplon mayor, hepatectom ı´as, control directo
de los vasos sangrantes (tras maniobra de Pringle), ligadura selectiva dearteria hepa ´tica, reparacio ´n de la cava retrohepa ´tica tras oclusio ´n vascular
hepa´tica total y, taponamientos perihepa ´ticos. La tasa de supervivencia fue
del 86.9% (225/259) y 15 de 40 lesiones venosas retrohepa ´ticas (RHVI)
sobrevivieron sin necesitar transfusiones de sangre superiores a las 60unidades. En 20 de 25 (80%) de las lesiones RHVI as ı´ como en 14 de 17
(82%) sin este tipo de lesiones, el taponamiento perihepa ´tico fue capaz de
controlar la hemorragia ( p< 0.75). En los pacientes no tratados
quiru´rgicamente la tasa de fracasos terape ´uticos fue del 17.2% (17/99) y
para las lesiones de los grados II I a V del 46.7% (14/30). 3 de 6 fracasos en
pacientes con lesiones I V y V fallecieron (50%). La mortalidad total fue del
11.8% (41/348) y 51% de las muertes se debieron a exanguinacio ´n. Nuestros
resultados demuestran que los traumatismos hepa ´ticos graves, sobre todo
los grados IV y V, requieren tratamiento quiru ´rgico, ya que segu ´n la
gravedad del trauma pueden combinarse diferentes procedimientosquiru´rgicos, incrementando la supervivencia de estos pacientes. En el
tratamiento de las lesiones RHVI el taponamiento (packing) perihepa ´tico
es muy efectivo, constituyendo el me ´todo de eleccio ´n.
References
1. Parks RW, Chrysos E, Diamond T. Management of liver trauma. Br. J.
Surg. 1999;86:1121–1135
2. Buechter KJ, Gomez GA, Zeppa R. A new technique for exposure of
injuries at the confluence of the retrohepatic veins and the retrohepaticvena cava. J. Trauma 1990;30:328–331
3. Menegaux F, Langlois P, Chigot JP. Severe blunt trauma of the liver:
study of mortality factors. J. Trauma 1993;35:865–869
4. Moore EE, Cogbill TH, Jurkovich GJ, et al. Organ injury scaling:
spleen and liver (1994 revision). J. Trauma 1995;38:323–324
5. Strong RW, Lynch SV, Wall DR, et al. Anatomic resection for severe
liver trauma. Surgery 1998;123:251–257
6. Boone DC, Federle M, Billiar TR, et al. Evolution of management of707 Gao et al.: Liver Trauma

major hepatic trauma: identification of patterns of injury. J. Trauma
1995;39:344–350
7. Brasel KJ, Delisle CM, Olson CJ, et al. Trends in the management of
hepatic injury. Am. J. Surg. 1997;174:674–677
8. Pachter HL, Knudson MM, Esrig B, et al. Status of nonoperative man-
agement of blunt hepatic injuries in 1995: a multicenter experience with404 patients. J. Trauma 1996;40:31–38
9. Sherman HF, Savage BA, Jones LM, et al. Nonoperative management
of blunt hepatic injuries: safe at any grade? J. Trauma 1994;37:616–621
10. Meredith JW, Young JS, Bowling J, et al. Nonoperative management of
blunt hepatic trauma: the exception or the rule? J. Trauma 1994;36:529–535
11. Huguet C, Gavelli A, Chieco PA, et al. Liver ischemia for hepatic re-
section: where is the limit? Surgery 1992;111:251–259
12. Chen RJ, Fang JF, Lin BC, et al. Surgical management of juxtahepatic
venous injuries in blunt hepatic trauma. J. Trauma 1995;38:886–890
13. Beal SL. Fatal hepatic hemorrhage: an unresolved problem in the man-
agement of complex liver injuries. J. Trauma 1990;30:163–169
14. Stylianos S. Abdominal packing for severe hemorrhage. J. Pediatr.
Surg. 1998;33:339–342
15. Caruso DM, Battistella FD, Owings JT, et al. Perihepatic packing of
major liver injuries: complications and mortality. Arch. Surg. 1999;134 :
958–96216. Abikhaled JA, Granchi TS, Wall MJ, et al. Prolonged abdominal pack-
ing for trauma is associated with increased morbidity and mortality.Am. Surg. 1997;63:1109–1112
17. Mayberry JC. Bedside open abdominal surgery. Utility and wound
management. Crit. Care Clin. 2000;16:151–172
18. Meldrum DR, Moore FA, Moore EE, et al. Cardiopulmonary hazards
of perihepatic packing for major liver injuries. Am. J. Surg. 1995;170:537–542
19. Burch JM, Moore EE, Moore FA. The abdominal compartment syn-
drome. Surg. Clin. N. Am. 1996;76:833–842
20. Ivatury RR, Porter JM, Simon RJ, et al. Intra-abdominal hypertension
after life-threatening penetrating abdominal trauma: prophylaxis, inc i-
dence, and clinical relevance to gastric mucosal pH and abdominalcompartment syndrome. J. Trauma 1998;44:1016–1023
21. Rogers FB, Reese J, Shackford SR, et al. The use of venovenous bypass
and total vascular isolation of the liver in the surgical management ofjuxtahepatic venous injuries in blunt hepatic trauma. J. Trauma 1997;43:530–533
22. Baumgartner F, Scudamore C, Nair C, et al. Venovenous bypass for
major hepatic and caval trauma. J. Trauma 1995;39:671–673708 World J. Surg. Vol. 27, No. 6, June 2003