The management of colon injuries has undergone many radical changes in the last few decades, resulting in a dramatic reduction of colon-related mortality from about 60% during World War I to about 40% during World War II to about 10% during the Vietnam War and to lower than 3% in the last decade. However, the colon-related morbidity remains unacceptably high and in most prospective studies the abdominal sepsis rate is about 20% (Table 33-1).1–6 In patients with destructive colon injuries, high Penetrating Abdominal Trauma Index (PATI), or multiple blood transfusions the incidence of intra-abdominal sepsis has been reported to be as high as 27%.7,8
Table 33-1 Incidence of Abdominal Septic Complications in Colon Injuries (Prospective Studies) |Favorite Table|Download (.pdf)
Table 33-1 Incidence of Abdominal Septic Complications in Colon Injuries (Prospective Studies)
|Author||Number of Patients||Abdominal Sepsis (%)|
|George et al.1||102||33|
|Chappuis et al.2||56||20|
|Demetriades et al.3||100||16|
|Ivatury et al.4||252||17|
|Gonzalez et al.5||114||24|
|Demetriades et al.6||297||24|
In the United States, the vast majority of colon injuries are due to penetrating trauma, usually firearms. In abdominal gunshot wounds the colon is the second most commonly injured organ after the small bowel and it is involved in about 27% of cases undergoing laparotomy.8,9 In anterior abdominal stab wounds the colon is the third most commonly injured organ after the liver and small bowel and is found in about 18% of patients undergoing laparotomy. In posterior stab wounds the colon is the most commonly injured organ and is diagnosed in about 20% of patients undergoing laparotomy.10 The transverse colon is the most commonly injured segment after gunshot wounds and the left colon the most commonly injured segment after stab wounds.
Stab wounds or low-velocity civilian gunshot wounds usually cause limited damage and most of them are amenable to debridement and primary repair (Fig. 33-1). High-velocity penetrating injuries, such as in war-related trauma, cause major tissue damage and almost always require colon resection (Fig. 33-2).
Low-velocity gunshot wounds cause local damage to the colon.
High-velocity destructive injury to the colon.
Blunt trauma to the colon is uncommon and occurs in about 0.5% of all major blunt trauma admissions or in 10.6% of patients undergoing laparotomy.11,12 Most of these injuries are superficial and only 3% of patients undergoing laparotomy have full-thickness colon perforations.11,13 Traffic trauma is the most common cause of blunt colon injury. The usual mechanism is rapid deceleration that may cause mesenteric tears and ischemic necrosis of the colon (Fig. 33-3). Another possible mechanism is the transient formation of a closed loop and blowout perforation. Seatbelt use increases the risk of hollow viscus perforations. The presence of a seatbelt mark sign should increase the index of suspicion for hollow viscus injury. In rare cases a colonic wall hematoma or contusion may result in delayed perforation several days after the injury. The left colon is the most commonly injured segment followed by the right colon and the transverse colon.11
High-speed motor vehicle injury with mesocolon avulsion and necrosis of the colon.
In blast injuries such as in war or terror-related explosions, hollow viscera are more susceptible to injury than solid organs (Fig. 33-4). Often there is no evidence of major external abdominal trauma. The blast wave is more likely to cause colon rupture than any other intra-abdominal organ.14
Blast injury to the colon. (Courtesy of Captain P. Rhee.)
The diagnosis of colon injury is almost always made intraoperatively. In patients with a penetrating abdominal trauma, selected for nonoperative management, the diagnosis is based on CT scan evaluation and serial clinical examinations. A rectal examination may show blood in the stool, especially in cases with distal colon or rectal injuries. The sensitivity and specificity of the intravenous contrast CT scan are about 90% and 96%, respectively.15 Other investigations, such as ultrasound, diagnostic peritoneal lavage, or laparoscopy, have little or no role in the evaluation of suspected colon injuries.
The preoperative diagnosis of colon injury following blunt trauma can be difficult, especially in unevaluable patients, due to severe associated head or spinal cord injuries. The diagnosis is often suspected by the presence of free gas, unexplained free peritoneal fluid, or thickened colonic wall on the routine abdominal CT scan (Fig. 33-5). Luminal contrast extravasation is an infrequent finding and its absence does not rule out an injury. In some cases the diagnosis may be delayed by many days with catastrophic consequences.
CT scan of a victim involved in a high-speed traffic injury. Note the thicken ascending colon wall (circle). The patient had rupture of the cecum.
Intraoperatively in penetrating trauma, every paracolic hematoma should be explored and the underlying colon should be evaluated carefully. Failure to adhere to this important surgical principle is a serious error with medical and legal implications. In blunt trauma there is no need for routine exploration of paracolic hematomas, unless there is a strong suspicion for an underlying perforation.
The American Association for the Surgery of Trauma (AAST) developed a grading system for colon injuries that is useful in predicting complications and comparing therapeutic interventions. The AAST Colon Injury Scale is shown in Table 33-2.16
Table 33-2 AAST Colon Injury Scale |Favorite Table|Download (.pdf)
Table 33-2 AAST Colon Injury Scale
|I||(a) Contusion or hematoma without devascularization|
|(b) Partial thickness laceration|
|II||Laceration ≤50% of circumference|
|III||Laceration >50% of circumference|
|IV||Transection of the colon|
|V||Transection of the colon with segmental tissue loss|
The first guidelines regarding the management of colon injuries were published by the US Surgeon General and mandated colostomy for all colon wounds. This unusual directive was initiated because of the very high mortality associated with colorectal injuries, in excess of 50%,17,18 during the early years of World War II. Although these guidelines were not based on any scientific evidence, they were credited for the improved outcomes in the last years of the war. However, during this period many other major advances such as faster evacuation from the battlefield, improved resuscitation, and introduction of penicillin and sulfadiazine could all have contributed to the reduction of mortality. The policy of mandatory colostomy remained the unchallenged standard of care until late 1970s. Stone and Fabian reported the first major scientific challenge of this policy in 1979.19 A prospective randomized study, which excluded patients with hypotension, multiple associated injuries, destructive colon injuries, and delayed operations, concluded that primary repair was associated with fewer complications than colostomy. The exclusion criteria were perceived as risk factors for anastomotic leak and were absolute indications for diversion.
The validity of the “standard” contraindications for primary repair or resection and anastomosis was challenged in subsequent studies. New prospective randomized studies with no exclusion criteria demonstrated the safety of primary repair, at least in nondestructive colon injuries. An alternative to primary repair or colostomy was exteriorized repair, which was introduced in the 1970s. This technique included suturing and exteriorization of the colon. If the repair remained intact in the next 4–5 days, the colon was returned to the peritoneal cavity. If the repair broke down, it was converted to a loop colostomy.20,21 The enthusiasm for this approach waned in the 1980s due to the overwhelming evidence of the superiority of primary repair.
In the 1990s and 2000s primary repair gained widespread acceptance and the role of colostomy was challenged, even in cases with perceived risk factors.
Management of Nondestructive Colon Injuries
Nondestructive injuries include those involving <50% of the bowel wall and without devascularization. There is now enough class I evidence supporting primary repair in all nondestructive colon injuries irrespective of risk factors. Chappuis et al.2 in a randomized study of 56 patients with no exclusionary criteria concluded that primary repair should be considered in all colon injuries irrespective of risk factors. In a subsequent study in 1995, Sasaki et al.22 randomized 71 patients with colon injuries to either primary repair or diversion, without any exclusionary criteria. The overall complication rate was 19% in the primary repair group and 36% in the diversion group. In addition, the complication rate for colostomy closure was 7%. The study concluded that primary repair should be performed in all civilian penetrating colon injuries irrespective of any associated risk factors.
In another prospective randomized study in 1996, Gonzalez et al.5 randomized 109 patients with penetrating colon injuries to primary repair on diversion. The sepsis-related complication rate was 20% in the primary repair group and 25% in the diversion group. The authors continued their study and the series increased to 176 patients. They concluded again that all civilian penetrating colon injuries should be primarily repaired.
Overall, collective review of all published prospective randomized studies identified 160 patients with primary repair and 143 patients treated with diversion. The abdominal sepsis complication rate was 13.1% and 21.7%, respectively (Table 33-3). In addition, numerous prospective observational studies (class II evidence) supported routine primary repair in nondestructive injuries.1,3,4,23 In conclusion, there are sufficient class I and II data to support routine primary repair of all nondestructive colon injuries, irrespective of risk factors.
Table 33-3 Primary Repair versus Diversion: Prospective Randomized Studies with No Exclusion Criteria |Favorite Table|Download (.pdf)
Table 33-3 Primary Repair versus Diversion: Prospective Randomized Studies with No Exclusion Criteria
|Study||No. of Patients||Abdominal Septic Complications||No. of Patients||Abdominal Complications|
|Chappuis et al.2||28||4 (14.3%)||28||5 (17.9%)|
|Sasaki22||43||1 (2.3%)||28||8 (28.6%)|
|Gonzalez23||89||16 (18%)||87||18 (21%)|
|Total||160||21 (13.1%)||143||31 (21.7%)|
Despite the available scientific evidence, many surgeons still consider colostomy as the safest procedure in high-risk colon injuries. In a survey of 317 Canadian surgeons in 1996, 75% of them chose colostomy in low-velocity gunshot wounds to the colon.24 In another survey in 1998, of 342 American trauma surgeons, members of the AAST, a colostomy was the procedure of choice in 3% of injuries with minimal spillage, in 43% of injuries with gross spillage, in 18% of injuries involving >50% of the colon wall, and in 33% of cases with colon transection.25 It is obvious that old habits still play a significant role in modern surgical practice.
Management of Destructive Colon Injuries
Destructive colon injuries include those with loss of more than 50% of the bowel wall circumference or with devascularization (Fig. 33-2) and require a segmental colonic resection. Destructive injuries were traditionally managed with diversion because of the perceived high risk for intra-abdominal sepsis. Small prospective studies in the 1990s suggested that primary anastomosis may be safe. Collectively, these studies included only 36 patients with colon resection and anastomosis. The incidence of anastomotic leak was 2.5% and no deaths occurred. These studies concluded that primary anastomosis is the procedure of choice irrespective of the presence of any risk factors for abdominal complications.2,22,23 However, another prospective observational study with 25 patients treated by resection and anastomosis and 2 patients treated by resection and colostomy reported two fatal anastomotic leaks (8%).7 The study concluded that some high-risk patients (PATI >25 or ≥6 U of blood transfusions or delayed operation) with destructive colon injuries might benefit from diversion. The study included only 2 patients with diversion, making any comparison with the primary anastomosis group impossible.
There are two retrospective studies, which included only destructive colon injuries requiring resection. In an analysis of 43 patients who were managed by resection and anastomosis Stewart et al.26 reported an overall anastomotic leak rate of 14%. However, in the subgroup of patients with blood transfusion >6 U the leak rate was 33%. The study suggested that diversion should be considered in patients receiving massive blood transfusions or in the presence of underlying medical illness. In another retrospective study of 140 patients with destructive colon injuries requiring resection Murray et al.27 reported similar intra-abdominal sepsis rates with primary anastomosis or diversion. Univariate analysis identified Abdominal Trauma Index ≥25 or hypotension in the emergency room to be associated with increased risk of anastomotic leak. The study suggested that a diversion procedure should be considered in these high-risk subgroups of patients.
In summary, the available prospective randomized studies, which include only a small number of cases, recommend resection with anastomosis irrespective of risk factors. Two larger retrospective studies suggest that diversion should be considered in selected patients with PATI ≥25, multiple blood transfusions, or associated medical illness.26,27 The guidelines of the Eastern Association for the Surgery of Trauma (EAST) published in 199828 recommended that a diversion procedure should be considered in patients with shock, significant associated injuries, peritonitis, or underlying disease. However, these guidelines could not be supported by the literature and were based exclusively on class III evidence. There were only 40 patients in class I studies with resection and anastomosis and the anastomotic leak rate was 2.5% and without mortality. There were only 12 patients in class II studies who underwent anastomosis and the leak rate was 8.3% without mortality. In class III retrospective studies there were 303 patients with anastomosis with a leak rate of 5.2% and 3 deaths (1%) due to the leak.
In order to address these limitations, the AAST sponsored a prospective multicenter study to evaluate the safety of primary anastomosis or diversion and identify independent risk factors for colon-related complications in patients with penetrating destructive colon injuries.6 The study included 297 patients with penetrating colon injuries requiring resection who survived at least 72 hours. Rectal injuries were excluded. The overall colon-related mortality was 1.3% (4 deaths) and all deaths occurred in the diversion group (P = .01). The most common abdominal complication was an intra-abdominal abscess (19% of patients) followed by fascia dehiscence (9%). The incidence of anastomotic leaks was 6.6% and no death occurred in these cases. Multivariate analysis identified severe fecal contamination, ≥4 U of blood transfusions within the first 24 hours, and inappropriate antibiotic prophylaxis as independent risk factors for abdominal complications. In the presence of all these three risk factors the incidence of abdominal complications was about 60%, in the presence of two factors the complication rate was 34%, in the presence of only one factor the rate was about 20%, and with no risk factors it was 13%. The method of colon management (anastomosis or diversion), delay of operation >6 hours, shock at admission, site of colon injury, PATI >25, ISS >20, or associated intra-abdominal injuries were not found to be independent risk factors. In a second analysis, the study compared colon-related outcomes in high-risk groups (hypotension at admission, blood transfusions> 6 U, delay of operation >6 hours, severe peritoneal contamination, or PATI >25) after primary anastomosis or colostomy. These risk factors have been suggested by many surgeons as indications for diversion. The colon-related mortality in this high-risk group was 4.5% (4 of 88 patients) in the colostomy group and no deaths in the 121 patients who underwent primary anastomosis (P = .03). The adjusted relative risk of abdominal septic complications was similar with the two operative procedures, in both the low- and high-risk patients (Table 33-4). There was a trend toward longer ICU and hospital stay in the colostomy group. The study concluded that “In view of these findings and the fact that colon diversion is associated with worse quality of life and requires an additional operation for closure, colon injuries requiring resection should be managed by primary repair, irrespective of risk factors.”6
Table 33-4 AAST Study of Destructive Colon Injuries: Comparison of Abdominal Complications between Primary Anastomosis and Diversion in High- and Low-Risk Patients6 |Favorite Table|Download (.pdf)
Table 33-4 AAST Study of Destructive Colon Injuries: Comparison of Abdominal Complications between Primary Anastomosis and Diversion in High- and Low-Risk Patients6
|Abdominal Complications (%)|
|Patient Population (N = 297)||Primary Anastomosis||Diversion||Adjusted Relative Risk (95% CI)||P-value|
|All patients||22||27||0.81 (0.55–1.41)||0.69|
|Low-risk patientsch33fn1||13||8||1.26 (0.21–8.39)||0.82|
|High-risk patientsch33fn1||28||30||0.90 (0.53–1.40)||0.67|
The optimal management of destructive colon injuries in patients undergoing damage control procedures is not clear, and the literature on this issue is scanty. It has been suggested that anastomosis may be safe because follow-up reexploration identifies any anastomotic problems and fecal diversion can be performed at this stage. There are some theoretical disadvantages of having a colostomy, because it is an open source of fecal material, near an open abdomen. In addition, the subsequent closure of the colostomy, especially end colostomy, might be a major technical challenge because of the hostile intra-abdominal environment. For those patients managed with stapling off of the injured colon as part of damage control, there is class III evidence that delayed primary anastomosis may be a safe option.29,30 This issue requires further investigation in larger and better controlled studies.
Risk Factors for Abdominal Complications after Colon Injuries
The incidence of abdominal complications after colon injuries is very high, with a sepsis rate higher than 20% (Table 33-1). Various conditions have been suggested as possible risk factors for colon-related complications but most of them failed scientific scrutiny.
Left versus right colon injuries: For many years there was an anecdotal but widespread belief that left colon injuries are associated with a higher risk of anastomotic leaks and septic complications following repair or colocolostomy than right colon injuries. This perception was based mainly on the anatomical differences between the two sides of the colon. This led to the practice of liberal primary repair in the right colon and colostomy in the left colon. However, no clinical or experimental study has ever demonstrated any healing differences between the two sides of the colon or any evidence that the two anatomical sides should be treated differently. Experimental work in baboons, which have anatomy and bacteriology very similar to those of humans, showed no difference of the healing properties between the right and left colon.31 The healing was evaluated clinically (anastomotic leak or abscess), biochemically (hydroxyproline concentrations), and mechanically (breaking strength of the anastomosis), in both normovolemic and hypovolemic conditions.31,32 However, there is strong evidence that ileocolostomy is associated with significantly fewer leaks than colocolostomy and it should be the procedure of choice in cases with right hemicolectomy.31,32 Good blood supply is the cornerstone of successful colon healing and this should be taken into account when repairing injuries in the watershed region of the splenic flexure.
- Associated abdominal injuries: Earlier retrospective studies suggested that multiple or severe associated intra-abdominal injuries (PATI > 25) are associated with a high incidence of anastomotic leaks and therefore a colostomy should be performed, especially in patients with destructive colon injuries. However, class I and II studies have shown that although multiple associated intra-abdominal injuries are significant risk factors for intra-abdominal sepsis, the method of colon management does not affect the incidence of abdominal sepsis.3,5–7,33 Some studies have even suggested that the presence of a colostomy in these high-risk patients may independently contribute to abdominal sepsis.33 There is class II evidence that the presence of pancreatic or urine leaks is associated with increased risk of anastomotic failure.31,32
Shock: There is now sufficient class I and II evidence that preoperative or intraoperative shock is neither an independent risk factor for abdominal sepsis nor a contraindication for primary colon repair or anastomosis.3,5,6 The duration and severity of hypotension might be important factors not taken into account in these studies. This is an area that needs further investigation.
Blood transfusions: Multiple blood transfusions (≥4 U of blood within the first 24 hours) have been shown to be a major independent risk factor for abdominal septic complications.6,33 In a large prospective AAST study of 297 patients with penetrating destructive colon injuries multiple blood transfusions were the most important independent factor for abdominal sepsis. However, the method of colon management, primary anastomosis or colostomy, did not influence the complication rate in this group of patients.6 It is possible that massive transfusion might be an important risk factor for anastomotic failure and this possibility requires further investigation.
Fecal contamination: Severe fecal spillage is a major independent risk factor for abdominal sepsis.1,6,11,27,33,34 This finding led some authors to suggest that this condition should be a contraindication for primary repair or anastomosis.1,11,34,35 However, all prospective randomized studies and recent large prospective observational studies have shown that the method of colon management does not influence the septic complication rate.2,5,6
Time from injury to operation: Although delays in the operative management of colon perforations increase the risk of septic complications, the length of delay over which the complication rate increases is not clear. Some studies suggest that this critical delay is 6 hours, while others extend it to 12 hours.7,27,36 It seems that the degree of contamination is much more important than the operative delay and the time delay in itself should not be used as an absolute criterion for primary repair or diversion.
Retained missiles: There is no evidence that retained bullets, which passed through the colon, are associated with increased risk of local sepsis. Removal of the missiles does not reduce the risk of infection. In a study of 84 patients with gunshot wounds of the colon, the bullet remained in the body in 40 and was removed in 44. The incidence of local septic complications was 5% and 7%, respectively.37
Closure of the skin wound: Closure of the skin incision after colonic injuries, especially in the pressure of fecal spillage, is associated with a high incidence of wound infection that is often complicated by necrotizing fasciitis or fascia dehiscence.38 In these cases the skin should be left open and delayed closure should be performed a few days later.
The overall incidence of colon leaks after repair or anastomosis is fairly low. In a collective review of 35 prospective or retrospective studies with 2,964 primary repairs, there were 66 (2.2%) leaks.39 Review of the published prospective studies that included 534 patients with colon repair or resection and anastomosis showed 17 (3.2%) leaks.6,39 Resection and anastomosis is significantly more likely to leak than simple repairs. In a collective review of 362 patients with resection and anastomosis the overall incidence of anastomotic leak was 5.5%.39 In a more recent multicenter prospective study of 197 patients with penetrating colon injuries who underwent resection and primary anastomosis, the leak rate was 6.6%.6
The risk factors for anastomotic leak are not well defined. Colocolostomies have been shown to be at a higher risk of anastomotic leaks than ileocolostomies. Murray et al. reported a leak rate of 4% in 56 patients with ileocolostomies and 13% in 56 colocolostomies.27 A multicenter prospective AAST study reported a leak rate of 4.2% for ileocolostomies and 8.9% for colocolostomies.6 Multiple blood transfusions, severe contamination, and multiple associated injuries were not identified as independent risk factors for anastomotic leak.
The prognosis of colon leaks is usually good and most of the patients can safely be managed nonoperatively with adequate drainage and low-residue diet. However, in some patients the colonic leak may cause severe intra-abdominal sepsis and a proximal diversion procedure may be required. Curran and Borzotta reported no deaths in a collective series of 66 patients with repair leaks.39 However, Murray et al. reported 2 colon-related deaths in a group of 10 patients with anastomotic leak.27 A multicenter AAST study reported no deaths in the 13 patients with anastomotic leaks. The overall mortality due to anastomotic leak-related complications in a collective review of 3,161 trauma patients treated with primary repair or resection and anastomosis was only 0.1%.6,39
In summary, colonic leaks occur more commonly in patients with colocolostomies than in patients with ileocolostomies. The majority of these leaks can safely be managed nonoperatively with adequate drainage and low-residue diet. Reexploration of the abdomen for wide drainage and fecal diversion or resection and reanastomosis should be reserved only for patients with generalized peritonitis or failed percutaneous drainage.
On entering the peritoneal cavity the first step is to control any bleeding. The extent of the colon injury is then assessed by adequate mobilization of the injured segment and careful inspection of the retroperitoneal wall. Any paracolic hematoma due to penetrating trauma should be explored to rule out any underlying perforation. Gentle squeezing of the suspected area may facilitate the diagnosis of any occult injuries by the manifestation of air or colonic content leak. The ureter should always be identified and examined in cases with injuries to the ascending or descending colon. The splenic flexure of the colon is the least accessible segment because of its anatomical location under the left hypochondrium. During its mobilization, caution should be exercised to avoid excessive downward traction of the colon, which may cause avulsion of the splenic capsule and troublesome bleeding (Fig. 33-6). In cases with multiple associated injuries and coagulopathy, many capsular tears may require a splenectomy, which increases the risk of postoperative complications. This iatrogenic complication can be avoided by placing three or four laparotomy pads under the diaphragm, above the spleen. This maneuver provides a good exposure and safe division of the splenocolic ligament.
Excessive downward traction of the splenic flexure of the colon may cause avulsion of the capsule of the spleen and troublesome bleeding.
Adequate debridement of all penetrating wounds, especially gunshot wounds, is critical before any repair is performed. In destructive injuries the resection should ensure well-perfused and not contused edges and the anastomosis should be tension free.
The role of local application of fibrin glue around the anastomosis is not clear. Although some experimental work suggested some protective benefits with fibrin glues,40 others failed to show any benefits.41 No clinical work has been published in this field. Further protection of the anastomosis with omental wrap is a common practice, although randomized studies in nontrauma colon operations failed to show any benefit.42
The method of anastomosis, handsewn or stapled, does not play any significant role in the incidence of anastomotic leaks. In a prospective AAST study of 207 patients with penetrating destructive injuries who underwent resection and anastomosis, 128 cases were managed by handsewn and 79 cases by stapled anastomosis. The incidence of anastomotic leak was 7.8% and 6.3%, respectively.43 Another debated technical issue is the role of one-layer versus two-layer anastomosis. Numerous studies in nontrauma operations have concluded that one-layer anastomosis is as safe as a two-layer anastomosis.44,45