The liver is effectively an intrathoracic organ (Fig. 1) 1201. It is protected from external forces, except for those severe enough to break ribs or to split the diaphragm (Fig. 2) 1202. Twenty-five per cent of blunt injuries and 5 per cent of penetrating injuries are fatal. If kinetic forces have rent the liver nonetheless, other organs are likely to have been be injured, and a Budd-Chiari syndrome may result from herniation of the liver through the diaphragm.

Little's cartoon (Fig. 3) 1203 shows that in 120 cases of liver injury the brain was injured 15 times (10 fatally), the spleen 21 times (6 fatally), and bones 45 times (14 fatally). It is important to deal immediately with remote injuries that have the potential to become lethal, even if at the moment they don't seem dangerous. Otherwise they'll be forgotten.

Only wide exposure permits examination and treatment of the liver and of other organs, such as the right colon and the pancreas. Elevate the patient's left arm on an ‘ether screen’ or other atraumatic means of support to facilitate exposure when a sternotomy or a thoracoabdominal incision is demanded (Fig. 4) 1204. Figure 5 1205 illustrates an abdomen that was explored for bleeding through a low-transverse incision without avail. When that incision was converted to a proper incision (the vertical component) the haemorrhage was stopped easily. Lyse the coronary and the triangular ligaments to mobilize the liver in all directions for inspection and repair (Fig. 6) 1206.

Although until recently all liver injuries were handled as if they were incipient catastrophes, CT scans show that in many of these injuries blood loss is less than 500 ml; for these stable and relatively minor injuries observation alone is satisfactory. Indeed, 60 per cent of blunt injury and 90 per cent of penetrating injury can be managed conservatively provided that CT examination ascertains that the hepatic flexure of the colon and the right kidney are not damaged. To the contrary, unstable patients do not benefit from the use of CT scans because the interval from arrival in the emergency department to the time of surgery is too long for safety. Unstable patients need an urgent laparotomy.

One must be aggressive if the patient deteriorates or if the magnitude of bleeding increases. Control intravascular losses by infusing volume-expanding solutions and blood products through a 14-G needle in a major vein of each upper extremity or through a subclavian vein cannula. Do not use lower extremity veins for this purpose because the inferior vena cava might have to be cross clamped, blocking blood flow cephalad. Although one might imagine that controlling haemorrhage in a fragmented liver would be a formidable task, even if the liver is semipulped manual pressure to the liver against the diaphragm generally stops haemorrhage because bleeding is usually from hepatic veins, and their pressure is low. Two dozen Mikulicz pads for tamponade are a further aid, however.

Although one cannot pump heat into a patient, who normally generates the heat of a 60-W light bulb, he or she can be kept from losing heat by having the room temperature above 26°C from the start. If this temperature, or a higher one, can be maintained, and if the patient's skull is covered with an internally reflecting aluminium cap, the core temperature will probably stay within a physiological range, allowing normal enzymic processes to work, particularly those associated with coagulation of blood. The Bair Hugger® , recently introduced, facilitates thermoregulation in patients at risk of hypothermia because it permits a controlled flow of warm air to stream over body surfaces enveloped in plastic manifolds and designed specifically for warming truncal and axial organs.

When bleeding persists and its site of origin is undiscoverable, think first of a tear in the inferior vena cava, especially if blood drips from the diaphragm or emerges from behind the liver and is present in the lesser sac. Use a disposable 18-F Frazier aspirator to clear the field of blood. It is without price in every operation associated with the likelihood of major blood loss.

In principle, try to get the liver on to the abdominal wall, just as in mobilizing a spleen for splenectomy. When a rent in the vena cava or in a proximate vein declares itself, apply local pressure until instruments and help arrive. The most important instruments are Judd–Allis vascular forceps (clamps). Apply five or more Judd–Allis clamps to the lips of the rent to set the stage for undersewing them or using them to approximate the posterior vein wall through an open anterior wall.

The need to insert an atriocaval shunt is rare (Fig. 7) 1207. Occasions for its insertion are so infrequent that few surgeons have an opportunity to use it regularly. In our hospital a single insertion was successful in the 1960s; no further use of it has been recognized or undertaken. Even in Houston, Texas, only 31 patients were treated with an atriocaval shunt over a period of nearly 10 years. At San Francisco General Hospital, where the shunt was devised, shunts were inserted in only 27 patients over a span of about 20 years. Survival of 12 patients gave a mortality rate of 55 per cent.

A decrease in end tidal Pco&sub2; is the most specific sign of air embolism, except for a windmill murmur in the heart. Depending upon the patient's size, his position on the operating table, and the degree of cardiovascular compensation, an air embolism of 20 ml, or less, can kill. Although the commonly invoked remedy is to turn the patient left side down to permit air to rise into the right heart, this action is seldom a specific because air is certainly already in the cardiac ventricles. Having a defibrillator at ready is more important, because attempts at removal of air from the right heart by aspiration through a jugular or subclavian central line are less successful than might be anticipated and because air often enters the cannula during fruitless efforts at removing intra-atrial froth.

This is the great dichotomy between surgeons and anaesthesiologists. The surgeon would like the blood to clot on demand. The anaesthesiologist is imbued with the need to deliver a plethora of lactated Ringer's solution to support physiological processes. An initial appropriate compromise is for the anaesthesiologist to administer not more than 2 litres of Ringer's solution unless massive volume replacement has to be undertaken with only Ringer's solution at hand. Otherwise, the formula should be to transfuse both 10 units of platelets and 5 units of packed cells for every 5 units of blood given after the first five. Once haemodilution has occurred there is no use trying to correct the situation at the operating table. Pack the abdomen and restore homeostasis in an intensive care unit.

Although Pringle's manoeuvre (compression of the hepatic vascular triad) is sometimes touted as a sovereign remedy for bleeding of all sorts in hepatic trauma, even Pringle's attempts to make compression work were virtually unavailing, and matters are often little better today. Notwithstanding, compress the liver against the diaphragm as described above and use Pringle's manoeuvre for the fastest and the most effective ways of dealing with fragmented liver and torrential bleeding from an unknown source. The problem is that accessory and replaced right and left hepatic arteries are common; the ability to compress aberrant blood vessels outside of the portal triad is a matter of luck. Occlusion of the infradiaphragmatic vena cava is sometimes helpful, but impeded venous return to the heart is often more than the patient can stand. Although a normal liver can withstand an hour of arterial ischaemia during an elective resection to remove a neoplasm, the tolerance of a burst and bleeding liver is far less.

Given that major tears of the parahepatic vena cava and of the right hepatic vein are usually fatal, devote attention to local haemostasis in the parts of the liver under your control: namely, the shattered hepatic parenchyma and the major and minor intrahepatic veins liable to bleed or to suck air, or both. Drain the liver bed with suction catheters above and below. Although drainage is sometimes frowned upon by the experts, experts won't be reading this chapter.

Place bolsters of Teflon felt approximately 4 cm long and 1 cm wide over and under regions of hepatic bleeding. Coapt the Teflon felt with no. 1 chromic catgut insofar as is possible, avoiding nearby structures in the portal triad (Fig. 8) 1208. Felt does not extrude, even in the presence of sepsis. If a ragged surface of liver presents itself, try converting it to a fishmouth contour to make closure easier, with or without Teflon bolsters.

When minor, persistent bleeding thwarts efforts to clear the field of blood, spray the sites with thrombin solution; then place a Hemopad® of crystalline collagen on the raw surface, followed by a rubber dam of appropriate size. When the dam is eventually removed, the thrombin and collagen layer remain. Although thrombin glue may work the same way, it is not approved for use in America because of the risk of hepatitis C contamination.

Subcapsular haematomas If they do not expand or burst, let them heal themselves as the blood and plasma are absorbed.Stab wounds not bleeding when they are examined in the operating room require only observation initially. Do not drain them or stir them up.

Partial major amputations Do not try to mate them to their former locations. Because it contains growth factors that probably facilitate haemostasis and healing, pack omentum into major crevices in the liver and tie it in place with non-strangulating heavy chromic catgut.

The burst liver Try to preserve an artery to any segment or segments provided that the portal vein blow flow is normal and that collateral blood vessels are intact. Otherwise pack the liver with gauze pads beneath, above, and around itself. Wrapping and compressing the liver with polyglactin mesh is a potential solution, but not one commonly applied. Arteriography after homeostasis is gained is a reasonable option in some instances to identify post-traumatic false hepatic artery aneurysms (Fig. 9) 1209.

Deal with false aneurysms in the radiology suite, not in the operating room. ( Figure 10 1210 shows the result of haemobilia, made clear after an injection of radiographic contrast medium, i.e., blood filling the biliary system.)

Therapeutic embolization with thrombin and Gianturco–Wallace coils may be appropriate. If injury demands occlusion of a major branch of the portal vein, use Doppler duplex ultrasonography to be certain that the portal vein, carrying 80 per cent of oxygen to the liver, is patent and functional.

Practically, the vena cava is part of the liver, even if it is not anatomically recognized as being so. Thus, the final aspect of recognizing and treating hepatic trauma is again to be alert to associated vena caval injury.

Remember that gauze packing of an injured liver is a practically foolproof way even for surgeons of little experience in hepatic trauma to deal with a potentially lethal situation. Ignore the smell of Pseudomonas and Proteus species. They seldom require specific treatment. When packs are removed after 12 or 24 h, following restoration of homeostasis, if the surgeon does not consider his experience adequate to deal with a complex problem, the patient can generally be transported to another hospital easily and with no hazard.

In extremis, consider the possibility of liver transplantation.

Burch JM, Feliciano DV, Mattox KL. The atriocaval shunt. Facts and fiction. Ann Surg, 1988; 207: 555–68.

Ciresi KF, Lim RC Jr. Hepatic vein and retrohepatic vena caval injury. World J Surg, 1990; 14: 472–7.

Dobson DE, et al. 1-Butyryl-glycerol: a novel angiogenesis factor secreted by differentiating adipocytes. Cell, 1990; 61: 223–30.

Feliciano DV. Surgery for liver trauma. Surg Clin N Am, 1989; 69: 273–84.

Hollands MJ, Little JM. Non-operative management of blunt liver injuries. Br J Surg, 1991; 78: 968–72.

Hollands MJ, Little JM. The role of hepatic resection in the management of blunt liver trauma. World J Surg, 1990; 14: 478–82.

John TG, et al. Liver trauma: a 10-year experience. Br J Surg, 1992; 79: 1352–6.

Krige JEF, Bornman PC, Terblanche J. Therapeutic perihepatic packing in complex liver trauma. Br J Surg, 1992; 79: 43–6.

Pachter HL, et al. Significant trends in the treatment of hepatic trauma. Experience with 411 injuries. Ann Surg, 1992; 215: 492–502.

Reed RL, et al. Continuing evolution in the approach to severe liver trauma. Ann Surg, 1992; 216: 524–38.

Schweizer W, et al. Management of traumatic liver injuries. Br J Surg, 1993; 80: 86–8.

Sharp KW, Locicero RJ. Abdominal packing for surgically uncontrollable haemorrhage. Ann Surg, 1992; 215: 467–75.

Sheldon GF, Rutledge R. Hepatic trauma. Adv Surg, 1989; 22: 179–94.

Stevens SL, Maull KI, Enderson BL. Total hepatic mesh wrap for hemostasis. Surg Gynecol Obstet, 1992; 175: 181–2.