The postoperative care of patients with gastric and small bowel injuries is usually relatively straightforward. Complications when they do arise are more often related to associated injuries or to delays in the operative management of the stomach and bowel injuries. Antibiotics are limited to a 24-hour course, usually of a single agent such as cefoxitin or ampicillin–sulbactam.79 However, appropriate dosing may be problematic in patients undergoing massive volume resuscitation with crystalloid and blood products. Some advocate the use of antifungal postoperatively in gastric perforations as fungal infections are more common after stomach perorations than small or large bowel perforations.
The advisability of routine nasogastric decompression following procedures involving an intestinal anastomosis is still controversial, despite prospective randomized controlled trials finding no advantage to this practice. A meta-analysis of selective versus routine nasogastric decompression after elective laparotomy was conducted by Cheatham et al on 3964 patients from 26 published trials.80 Routine nasogastric decompression was not supported by this meta-analysis of the literature. However, these studies did not involve trauma patients. The potential impact of other clinically important variables including the presence of multiple associated injuries, hemorrhagic shock, and postresuscitation bowel edema as well as an impaired sensorium from head injuries or drugs and alcohol may make nasogastric decompression the more prudent choice. It is our practice to continue nasogastric decompression, which was initiated during the initial resuscitation, until ileus resolves. It is also essential to have a properly functioning nasogastric tube in place to decompress the proximal GI tract following a damage control laparotomy and planned reestablishment of gastrointestinal tract continuity. In patients in whom a jejunostomy feeding tube is placed at laparotomy it is also useful to decompress the stomach and monitor gastric outputs as jejunal enteral feeds are initiated. Jejunal feeding may increase gastric output significantly which can lead to pulmonary aspiration in these patients.81
In uncomplicated cases involving stomach or small bowel injuries there is no evidence to support routine nutritional support of patients who were well nourished pre-injury. On the other hand, in critically ill or injured patients it is prudent to start nutritional support early before hypermetabolism or sepsis intervenes.82 Available clinical evidence suggests that moderately to severely injured patients (ISS > 16 < 25) should have enteral feedings started between 24 and 48 hours postinjury. Those with more severe injuries are more likely to have intolerance to enteral feedings.
There is convincing evidence in the literature that patients with blunt and penetrating injuries sustain fewer septic complications when fed enterally as opposed to parenterally.83 However, total parenteral nutrition should be started by day 7 in severely injured patients who do not tolerate enteral feeding or fail to tolerate at least 50% of their goal rate of enteral feedings. Parenteral nutrition has not been found to be of benefit in trauma patients and is associated with worse outcome. With the decreased use of parenteral nutrition over the years, the complications and hospital resources have been decreased without negatively affecting outcome.84,85,86 It is safest to start enteral feedings at the end of active shock resuscitation. It appears that starting enteral feedings up to 36 hours postinjury and at a “trophic infusion rate” (15 mL/h) for up to 4 days is effective in decreasing pneumonia rates without untoward effects on the ICU course of patients with severe blunt abdominal trauma.87 There does not appear to be a clear advantage to postpyloric enteral feeding versus gastric feeding in trauma patients. Thus, with few exceptions (severe closed head injury, or severe pancreatic duodenal injury) surgical feeding jejunostomy (with the exception of a needle catheter jejunostomy) should not be routinely performed at the initial laparotomy. Further, bowel edema may make this simple procedure a far greater challenge than necessary. If patients do not tolerate intragastric feeds, it is far simpler to place a nasojejunal tube by hand, or endoscopically at the time of initial laparotomy or postoperatively in the ICU.
Complications directly related to gastric and small injuries include intra-abdominal septic complications and anastomotic disruption. The most important etiologic factor relating stomach or small bowel injuries to intra-abdominal abscess formation is delayed recognition and surgical treatment.16,17 Intra-abdominal septic complications most often presents as an intra-abdominal abscess. Anastomotic failures may present as peritonitis and/or the development of an external fistula. Infectious complications following gastric injury are most common following blunt trauma especially if there is an associated colon injury. In these patients, ongoing fever and leukocytosis should mandate diagnostic imaging of both the chest and abdomen to look for foci of infection to drain.
Bleeding complications after gastric or small bowel trauma are rare, but may present as bleeding into the peritoneal cavity or into the bowel lumen. Bleeding from the short gastric vessels or from a torn splenic capsule is a common iatrogenic source of bleeding in this area. Bleeding from the mesentery or lesser curvature of the stomach may not be apparent intraoperatively in the hypotensive patient. This may only become clinically apparent when the patient normalizes blood pressure; continued bleeding postoperatively then manifests as hypotension and a falling hematocrit.88 Suture line bleeding although rare, can be troublesome and may manifest as bloody nasogastric secretions. Endoscopic hemostatic techniques may be carefully employed in this setting, particularly if the bleeding is from the stomach as it is more challenging to accomplish small bowel endoscopy.
Anastomotic leak following repair of gastric and small bowel injury can lead to significant morbidity and mortality. The definition of anastomotic leak is variable in both emergency and elective gastrointestinal reviews on the topic.89 Anastomotic failure may present as a contained leak, diffuse peritonitis, or a gastrocutaneous or enterocutaneous fistula (ECF). Risk factors for breakdown of intestinal repair include resection and anastomosis rather than repair, massive perioperative blood and fluid administration, associated pancreatic injuries, and the development of the abdominal compartment syndrome. In patients with enteric injuries managed with an open abdomen, failure to obtain fascial closure after post-injury day 5 has been found to result in an increased leak rate.77,80 Additional factors include evidence for ongoing hypoperfusion and the use of vasopressors during the initial resuscitation and the early post-injury ICU management.90
CT is the best diagnostic imaging study to identify anastomotic leaks that are not clinically obvious. Therapeutic options include medical care only (NPO, TPN, perhaps octreotide, somatostatin, and NG suction) if there is a tiny radiographically evident but clinically insignificant leak. Reoperation is necessary with primary repair and drainage for a small leak discovered early postoperatively if there is minimal peritonitis in the otherwise stable patient. Percutaneous drainage is useful for a symptomatic leak presenting in a delayed fashion as an intra-abdominal abscess. If there is complete disruption of an anastomosis with widespread peritoneal contamination, it is advisable to consider a proximal diverting enterostomy.
Anastomotic leaks diagnosed in the immediate postoperative period may be surgically approached, as the relevant tissue planes are amenable to surgical dissection. After 10 or 14 days, the inflammatory process makes dissection of bowel extremely difficult. In this case proximal diversion and/or controlled external drainage of the leak may be safer.
An ECF is a dreaded complication following trauma laparotomy and may be the result of an anastomotic leak, missed injury, delayed perforation or complications from an open abdomen following a damage control laparotomy. An ECF developing with an open abdomen is referred to as an enteroatmospheric fistula (EAF) and prior to massive transfusion protocols and the resulting diminished occurrence of the damage control laparotomy, was the most common type of ECF encountered by trauma surgeons.91,92 An EAF may result from anastomotic breakdown or de novo from exposed bowel in the open abdomen. Factors associated with the development of an EAF include deserialization and iatrogenic injury to the bowel in an open abdomen and dense granulation tissue with adhesion of bowel loops to the fascial rim or adjacent bowel loops.92 Excessive force on the bowel from coughing or even movement by the patient in this setting can lead to shearing of the bowel and bowel disruption. It is advisable to protect the bowel to avoid this complication. Measures include keeping omentum and/or nonadherent dressing materials over exposed bowel (such as Owens Gauze), gentle dressing changes by experienced caregivers, and aggressive attempts to obtain abdominal fascial or “skin only” closure as “early as possible.” Vacuum packs, vacuum-assisted wound management, and progressive closure with abdominal retention sutures are useful techniques to facilitate “early” fascial closure. Absorbable mesh materials and human acellular dermal materials (HADM) or other non-cross-linked biological materials may also be used to facilitate early abdominal closure.76 However, prolonged use of vacuum-assisted closure (VAC) of abdominal wounds and absorbable mesh materials may also contribute to the development of an EAF.92
A single institution review of the development of an ECF in the era of open abdomen management was published by Fischer et al.93 The overall incidence of ECF following trauma laparotomy was 1.9%. Patients with open abdomen had a higher ECF incidence (8% vs 0.5%) and a lower rate of spontaneous closure (37% vs 43%). The development of an ECF initiates the requirement for a prolonged ICU and hospital length of stay as well as the need for a team of dedicated and experienced nurses, wound care therapists, and surgeons.
There are three phases in the management of an ECF.92 Phase 1 is the recognition of the fistula and patient stabilization. Initial clinical priorities include fluid and electrolyte imbalance, control of sepsis, nutrition, and wound care. The patient may present with enteric content coming from the wound or indolent sepsis and a leak eventually identified by imaging studies. External loss of intestinal fluids rich in electrolytes, minerals, and protein leads to fluid and electrolyte imbalances as well as eventual malnutrition.
Identification of the fistula site and/or measurement of the electrolyte composition of the fistula effluent are sometimes helpful for fluid replacement (Table 31-4). However, in most cases, normal saline with 10–20 mEq of potassium per liter is a suitable fluid to use for the initial intravenous fluid replacement. Patients with ECF may also develop significant calcium, magnesium, and phosphate deficits that should be corrected.
TABLE 31-4Composition and Volume of Gastrointestinal Secretions ||Download (.pdf) TABLE 31-4 Composition and Volume of Gastrointestinal Secretions
|Type ||Volume (mL/day) ||Na (mEq/L) ||K (mEq/L) ||Cl (mEq/L) ||HCO3 (mEq/L) |
Fistulas are classified as high output (>500 mL/d), moderate output (200–500 mL/d), or low output (<200 mL/d). This may be important to classify fistulas in this manner as it may allow anticipation of the method for nutritional support, and may be useful in predicting the likelihood of spontaneous closure and mortality.92
Control of sepsis may include image-guided or surgical drainage of intra-abdominal abscesses identified by CT. Empiric antibiotic should be started in septic patients and modified after relevant culture data are obtained. The presence of an ECF without clinical signs of sepsis does not warrant antibiotic therapy.
The provision of adequate nutritional support is critical in the stabilization phase. Total parental nutrition has long been recognized to be an important factor in the management of ECF. These patients are usually hypercatabolic and generally require 25–32 cal/kg/d in total calories with a calorie to nitrogen ratio of 150:1 and a protein intake of at least 1.5 g/kg/d. Attempts at enteral nutritional support may aggravate fluid and electrolyte imbalances in the early phase of ECF management.
Adjuncts to control fistula drainage include nasogastric drainage, and acid suppression with H2-receptor antagonists or protein pump inhibitors. Fistula output can be reduced with somatostatin and octreotide. These agents reduce GI secretions and prolong transit times, thereby simplifying management of fistula output. However, no evidence exists that these agents hasten spontaneous closure rates. Administration of somatostatin and its analog octreotide has been shown to have an inconsistent effect on fistula output and time to fistula closure. Furthermore, the use of these agents does not increase the rate of nonoperative closure of fistulas.94 If used, fistula output should be monitored before and after a trial with the use of these agents to determine efficacy.
Protecting the integrity of the skin surrounding the fistulas will improve the quality of the surrounding tissues and decrease infectious complications. Low-output fistulas are usually managed with conventional measures. High-output fistulas or fistula(s) in the patient with an open abdomen may benefit by the use of the VAC system. These may be applied over the entire wound with the fistula or as a VAC dressing with openings for stoma pouches of the fistula openings.95,96 The main benefit of the use of the VAC system for ECF appears to be improved wound care before definitive surgery.
The second phase involves anatomic definition of the fistula. CT and/or fistulogram help define the anatomic details and any associated pathology that guide further interventions. Nutrition is continued by the parenteral route with attempts at the use of the enteral route. It is likely that at least 4 ft of functioning small bowel between the ligament of Treitz and the fistula is necessary for significant absorption of even low-residue formulas. However, recent reports have advocated that the provision of at least some of the caloric requirement should be by the enteral route. This may be helpful for “trophic” effects of enteral feedings on the intestine and, if tolerated, allow easier management in the outpatient setting. In patients with diversion of the proximal small bowel as an ostomy, reinfusion of the succus entericus into the distal GI tract may be helpful as well.97
Spontaneous closure of ECFs in patients provided adequate nutritional support and free of sepsis usually occurs within 4–6 weeks. Unfortunately spontaneous closure occurs only in about 30% of trauma patients.93 Definitive surgery (phase 3) in a patient with a persistent ECF is usually delayed 4–6 months following the initial operation. Failure to obtain spontaneous closure should not be a primary factor in the timing of operative intervention. But rather nutritional and wound status, as well as the overall clinical condition, of the patient should be optimal before embarking on an often long and difficult surgical procedure. The procedure should include complete lysis of adhesions to eliminate the possibility of distal obstruction as a contributing factor for the failure of spontaneous closure. Options include resection and reanastomoses of the involved bowel segments or over sewing or wedge resection of the fistula. Recurrence of an ECF is related to the method of surgical closure. In a study by Lynch et al, oversewing or wedge resection of the ECF was associated with 36% recurrence rate, while resection with reanastomoses had a 16% recurrence rate.98 After resection of the fistula and reestablishment of GI continuity, attention is then directed to closure of the abdominal wound. Closure with autologous tissue, often requiring unilateral or bilateral component separation, is optimal (Fig. 31-9).99,100
An illustration of the component separation technique as first described by Ramirez et al. A generous components separation will allow for expansion of the abdominal wall to all for coverage of a gap of up to 20 cm. (Reproduced with permission from Shestak KC, Edington HJ, Johnson RR. The separation of anatomic components technique for the reconstruction of massive midline abdominal wall defects: anatomy, surgical technique, applications, and limitations revisited. Plast Reconstr Surg. 2000 Feb; 105(2):731–738; quiz 739.)
The use of HADM or a non-cross-linked biological material is a second option if closure with autologous tissue is not practical (Fig. 31-10).101 Absorbable (vicryl) mesh closure and acceptance of a later incisional hernia is another viable option and may be the better part of valor in those with multiple fistulae and loss of abdominal domain. Later reconstruction once intestinal continuity has been established and the patient has tolerated a period of enteral nutrition is not unreasonable. In these patients the fistulae can be resected or closed, and the overlying skin simply reclosed or resected in the case of an overlying skin graft if viability is questioned. If overlying skin graft is resected, skin flaps can be developed and a “skin only” closure accomplished with the remaining abdominal skin. The use of other prosthetic materials, including cross-linked biological materials, is ill-advised due to concerns for the breakdown of intestinal anastomotic repair or the development of a de novo intestinal fistula.102,103
An enterocutaneous fistula that fails to close spontaneously should be managed operatively when the nutritional and wound statuses are optional (A). Closure with non-cross-linked biological materials may be used when autologous tissue is not available (B).
Small bowel obstruction (SBO) is a well-known complication following abdominal operation. Patients with nontherapeutic laparotomies for trauma had a 2.4% incidence of SBO in a report by Renz and Feliciano.104 The rate is higher if operative repair is required and may be up to 7.4% in patients with penetrating abdominal trauma and 10.8% in patients with small or large bowel injuries.105 However, trauma laparotomy does not appear to have added risk versus that reported following elective colorectal and general surgery on the development of early SBO and need for operative management.
CT imaging has superior sensitivity and specificity, compared with plain radiographs for making the diagnosis of mechanical SBO.106 The identification of a transition zone and “small bowel feces sign” on CT make the diagnosis of mechanical SBO more certain.106,107 However, the presence of radiographic transition zones does not increase the likelihood of need for operative intervention.108 CT findings suggestive of bowel ischemia include decreased bowel wall enhancement, mucal thickening, congestion of mesenteric veins, or ischemia. However, these findings could not discriminate between patients with strangulated and those with nonstrangulated SBO in a report by Rocha et al.106
The absence of these findings may be helpful in deciding on a course of conservative management for at least 10–14 days following initial laparotomy. Various clinicoradiological scores have been proposed to predict the risk for strangulated SBO.108,109 Fortunately, early postoperative SBO often resolves spontaneously. Thus, it may be initially treated expectantly with only a small risk of bowel strangulation.110
Resection of significant amounts of intestine may lead to problems with malabsorption. In general, jejunal resections are better tolerated than ileal resections. Removal of significant portions of the jejunum may lead to lactose intolerance; however, this is usually self-limited. Resection of the distal ileum often leads to vitamin B12 deficiency as well as bile salt deficiencies, and subsequent fat malabsorption. Ileal resection also removes the “ileal breaking mechanism” that may cause decreased transit time throughout the gut. This may result in profuse diarrhea and significant fluid and electrolyte imbalances. The ileocecal valve also has an important role as it acts to decrease the volume of stool by slowing intestinal transit time.
The short bowel syndrome may result from traumatic injury to the intestine and/or its blood supply. In a series of 196 adult patients evaluated at a single institution over 23 years, 8% of short bowel syndrome cases were secondary to traumatic injury.111 Eighty percent of trauma-related short bowel syndrome was due to mesenteric injuries. Clinical manifestations include malabsorption, diarrhea, steatorrhea, fluid and electrolyte disturbances, and malnutrition. Late complications include cholelithiasis and kidney (oxalate) stones. Although resection of up to 100 cm of ileum causes diarrhea, short bowel syndrome is fully manifest when the remaining jejunum and ileum is less than 200 cm in length. Plasma citrulline may be a useful biomarker to index small bowel enterocyte mass.112
Physiologic adaptation of patients with short bowel syndrome follows three phases.113 The acute phase occurs during the immediate postoperative weeks and may last 1–3 months. This phase is marked by poor absorption of almost all macronutrients and micronutrients. Ostomies, if present, may have outputs exceeding 5 L/d during the first few days. Aggressive intravenous fluid and electrolyte replacement is necessary to prevent life-threatening dehydration and electrolyte imbalances. Gastric hypersecretion is frequent in this phase and may be treated with proton pump inhibitors. Loperamide, codeine, or diphenoxylate, or even tincture of opium, is used to slow gastric and intestinal transit to control diarrhea. Careful monitoring of fluid and electrolytes, particularly potassium, magnesium, and calcium, is critical. Fluid needs can be monitored by urinary and fistula output, as well as by urinary sodium and osmolarity. The primary route of nutrition is parenteral; however, enteral feeding should be slowly initiated later in this phase.113
Adaptation occurs during the subacute phase and enhances bowel absorption. This adaptive process in humans occurs over a period of months to 1–2 years and is due to dilation of the remaining bowel coupled with improved cell transport function and prolonged intestinal transit time. Intestinal adaptation may be mediated by growth factors and nutrients including human growth hormone (HGH), insulin-like growth factor, epidermal growth factor, transforming growth factor a, and glucagen-like peptide 2 (GLP-2).114 Nutrients including glutamine and fatty acids may also act as growth factors in intestinal adaptation. The maintenance phase is characterized by achievement of maximal absorptive capacity. The goal of this phase is to achieve nutritional and metabolic homeostasis by primarily oral feeding.
Enteral feeding to provide intraluminal nutrients to maintain gut mass is necessary for the adaptation response to occur. Thus, enteral nutrition remains the primary therapy in maximizing luminal nutrient absorption in the intestinal remnant. The use of growth hormone, glutamine, and nutrients to facilitate bowel adaptation has been shown to help lessen TPN-dependence in adults with short bowel syndrome.115,116,117 The addition of GLP-2 may have a synergistic effect with HGH, glutamine, and optimal dietary management.
A number of surgical techniques have been used in an attempt to slow transit time and/or increase functional bowel length. Most of these techniques have met with limited success and risk further bowel loss.118 Restoring the colonic remnant to the GI tract may be helpful as the colon takes on an absorptive function by deriving energy from short-chain fatty acids and prolonging transit time, particularly if the ileocecal valve is intact. However, at least 3 ft of small intestine is required to prevent diarrhea and perianal complications. Surgical lengthening with the Bianchi procedure or serial transverse enteroplasty (STEP) may improve efficacy of enteral nutrition and reverse complications of TPN. The STEP is technically easier to perform than the Bianchi procedure. Both have been shown to improve absorption of nutrients by increasing the function of the remnant small bowel (Fig. 31-11).118
The STEP procedure used for lengthening segments of dilated small intestine to promote increased transit time and absorption of nutrients in short gut.115 (Reproduced with permission from Kim H, Fauza D, Garza J, et al. Serial transverse enteroplasty (STEP): a novel bowel lengthening procedure. J Pediatr Surg. 2003;38(3):425–429. Copyright © Elsevier.)
Intestinal transplantation is reserved as a last alternative for patients unable to compensate and adapt following intestinal resection. It may also be offered to patients who require TPN to maintain body mass and have limited or no remaining venous access for parenteral nutrition, or parenteral nutrition–related liver disease. Trauma patients seem to have equivalent long-term survival rates as compared with nontrauma patients following intestinal transplantation. A multidisciplinary approach to the treatment of intestinal failure due to short bowel syndrome appears to be the best outcome for these patients.119