When is primary anastomosis safe in the colon?

Author Information

B. Ostrow, MD, FRCS(C), Adjunct Lecturer, Office of International Surgery, University of Toronto, Canada

Reviewed July 2005 and updated August 2007

 

Introduction

The repair by suture or resection of diseased or injured colon is one of the most important skills in general surgery. Untreated or treated improperly these conditions cause significant morbidity in terms of intra-abdominal infection, fistulae, or death from generalized peritonitis. Today in Western medicine the two most common conditions requiring emergency colonic resection are obstructing colon cancers and complicated diverticulitis. In Africa, the equivalent conditions are gangrenous sigmoid volvulus and penetrating injuries resulting in colonic perforation. The deleterious consequences of primary anastomosis of the obstructed left colon with a tenuous blood supply, filled with pathogenic organisms and/ or of the perforated colon with various degrees of peritoneal contamination fostered a surgical paradigm which precluded primary anastomosis of the left colon unless preoperative mechanical bowel preparation and antibiotic prophylaxis had been undertaken.

This was certainly the paradigm I was taught and which I practised throughout 1980-2000 in Canada. As Primary Surgery, Non-Trauma, an exemplary text for International surgeons from which I plan to quote extensively in these Reviews, states: “You cannot safely anastomose large gut to large gut when it is obstructed and instead you have to let its contents escape through an ostomy. When large gut is to be anastomosed, it has to be carefully prepared first with enemas and antibiotics…“(1). A similar attitude held for colonic injuries. The definitive International Committee of the Red Cross 1990 edition of Surgery for Victims of War states: “The general rule is that a damaged colon should be exteriorized or defunctioned by proximal colostomy” (1). Over the last decade the tenets underlying this paradigm have been seriously questioned and the whole approach to emergency colonic surgery has been revised. This Review will briefly consider the historical development of the paradigm and then detail its decline through experience with colon injuries, the morbidity of colostomies themselves, alternatives in acute colonic disease such as obstruction, diverticulitis and sigmoid volvulus and finally the fundamental questioning of the value of mechanical bowel preparation.

History – Development of the Paradigm

The history of the development of treatments for perforated, gangrenous or obstructed colons reflects the fear surgeons had for the dangers of anastomotic dehiscence. The late 19th century saw the development of standardized techniques for “two-layer” bowel anastomosis (3) and at the same time non-anastomotic techniques for the acutely diseased colon – the colostomy. The PaulMikulicz operation exteriorized the diseased colon as a double-barreled colostomy. Mayo in 1907 recommended temporary transverse colostomy for obstructing diverticulitis. In the 20th century treatment for perforated diverticulitis evolved towards the Hartmann procedure, proposed for colon cancer by Henri Hartmann in 1921, with resection of the perforated segment and proximal colostomy. This procedure showed an improvement in mortality over proximal colostomy, suture and drainage (4).
For obstructing left-sided carcinomas a traditional three-staged process was initially proposed which included initial transverse colostomy, followed by resection-anastomosis and fi nally closure of the colostomy. Because of the recognized delay and difficulties of three operations, alternatives such as sub-total colectomy as well as techniques for on-table bowel cleansing and intra-colonic stents were developed. Lesions of the right and transverse colon could be managed by forms of right hemicolectomy and ileo-colic anastomosis. But the taboo of anastomosing unprepared left colon without proximal diverting colostomy persisted.
Ogilvie’s WWII experience established exteriorization of the damaged bowel and fecal diversion as the primary modalities of therapy for wartime colon injuries. Mortality dropped from 50% to 10% and this approach was initially applied to civilian penetrating colonic injury. The treatment for extra-peritoneal rectal injuries included colostomy, irrigation of the rectal stump and wide perineal drainage.
While some form of fecal diversion was considered mandatory for emergency left-sided colon lesions, an elaborate preparation was undertaken for elective colon surgery. Mechanical bowel preparations (MBP) to remove the fecal contents were developed. These progressed from prolonged fasting, fl uids only and cathartics to rapid polyethylene glycol solution and sodium phosphate administration. At the same time multiple different regimes of peri-operative antibiotics have been found to be useful in reducing post-operative septic complications and mortality (5). The antibiotics need to be active against enteric aerobic and anaerobic organisms and should be limited to the operative period. A single dose is adequate for operations less than 2 hours in duration. Gentamycin and metronidazole are a good combination. Parenteral and oral antibiotics may have a cumulative effect (6). Of all the traditional recommendations concerning colon surgery, only those relating to peri-operative antibiotics have withstood scientific scrutiny and are uniformly practiced.
While much of surgical practice is a learnt behaviour arising from the apprenticeship system of surgical education, the second half of the 20th century has seen the rise of evidence-based medicine and the prospective, randomized clinical trial (RCT). These types of studies have begun to alter surgical practice.

Revision of the paradigm — colonic injuries

It was in the field of colon trauma that exceptions to the rule prohibiting primary colon repair initially began to be advocated. Beginning first with right-sided injuries; subsequently extending to left sided lesions; those that could be repaired without resection; then to resection with proximal colostomy and finally the most recent recommendations accepting resection and primary anastomosis without fecal diversion as appropriate surgery — a new paradigm has emerged for the treatment of colon injuries. The initial caveats to this most radical treatment included limited time from injury to repair, minimal contamination and few associated injuries. Whereas Cornwell (2) questioned the use of primary repair for high risk penetrating trauma patients, by 2001, the same group (8) having undertaken a multicenter prospective study, concluded that all colon injuries could be managed without fecal diversion without influencing morbidity or mortality. Gonzalez (2) drew similar conclusions. However, Miller (10) still raised a cautionary note for high risk patients, those with high trauma scores, destructive colon injuries, need for multiple blood transfusions, etc. In a meta-analysis, (which pools results from multiple RCTs), Singer (11), clearly favoured primary repair without diversion for all risk categories. Kamwendo (12) from South Africa showed that increasing time after injury did not preclude primary repair. Esraghi (13) reflected the changing opinions of American surgeons on this question. Demetriades (14) sums up current thinking. Note should be made that the need remains for colostomy in rectal injuries, because of the particular anatomy of this problem. These changes are reflected in the 1998 edition of Surgery for Victims of War which tentatively states that fecal diversion for colonic injuries “has been the rule in war surgery, but the ICRCS working experience has shown that it is seldom necessary”(15).

Revision of the paradigm — the morbidity of colostomy

While colostomy itself in the proper setting may be a valuable procedure, the need for and risks of additional surgical procedures increase the morbidity to the patient. Khoury (16) reviewed the results of colostomy closure in 46 patients with a complication rate of 24%. Riesener (17) reviewed 548 closures since 1972. He advocated MBP and peri-operative antibiotics. However their best complication rate was still 27%. In Mealy (18) only 50% of patients had their stomas closed. Berne (19) reviewed colostomy closures after trauma and also found high morbidities, although this was not the case after rectal injury. Chandramouli (20) reviewed the experience of colostomies and their closure in children.

Revision of the paradigm –diverticulitis, obstruction and sigmoid volvulus

With the recognition of the increased morbidity of subsequent surgical procedures have come alternatives such as on-table lavage (21) and intra-colonic stenting. However, in the 1990s more surgeons began to perform resection and primary anastomosis in acute non-traumatic colon conditions. Trillo  (22) reported 43 cases without mortality. Goozen (23) reported 45 patients with acute obstructive or perforated diverticulitis treated with resection and primary anastomosis. There were three deaths from anastomotic leaks-all in obstructed patients. They cautioned its use in these cases. Salem (24) reviewed 98 published studies since 1957 and called for a RCT comparing resection and colostomy with primary anastomosis in acute surgical diverticulitis. Whetstone (25) giving guidelines for management of diverticulitis insisted that the Hartmann procedure is the appropriate operation when emergency surgery is indicated.
Various forms of right-hemicolectomy and ileo-colic anastomosis have been accepted for some time as the surgical procedure of choice in acute right-sided obstructions. For left-sided obstructions some authors advocate subtotal colectomy and ileo-colic anastomosis. Remst (26) is an example. Edino (27) carried out on table lavage in 32 consecutive patients without mortality. Harris (28) reviewed the use of endoluminal stents obviating the need for emergency surgery. DeSalvo (29) carried out a literature review and concluded that there existed inadequate studies comparing primary anastomosis and colostomy in left-sided obstructions and recommended further RCTs. Clearly no definitive conclusion can be drawn concerning the safety of primary anastomosis without colostomy for acute left-sided obstructions.
The non-Western experience with sigmoid volvulus also favours divergence from the paradigm. Kuzu (30) reported 102 patients with sigmoid volvulus requiring emergency surgery. Although the operative procedure was chosen by the surgeon about half the patients had resection and primary anastomosis and their infectious complications and death rates were similar to those with the Hartmann pro