Read Core Topics in General & Emergency Surgery: Companion to Specialist Surgical Practice Online
Authors: Simon Paterson-Brown MBBS MPhil MS FRCS
Core Topics in General & Emergency Surgery: Companion to Specialist Surgical Practice (52 page)
Although upper gastrointestinal and small-bowel bleeding has been covered in
Chapters 5
and
7
, there remains the problem of patients who bleed from the colon. In approximately 20% of cases, the colon and rectum are the source of acute gastrointestinal haemorrhage. The most common aetiologies of this haemorrhage include diverticulosis, angiodysplasia, ischaemic colitis, inflammatory bowel disease, neoplasms, aorto-enteric fistulas and anorectal diseases such as haemorrhoids.
82
The general principles of management of many of these conditions have already been discussed, but each patient needs to be managed individually. Where possible, management of colonic bleeding is non-operative and usually stops spontaneously, permitting investigation to proceed at a leisurely pace. In some patients, however, the bleeding either does not spontaneously stop or is profuse and urgent investigation, followed occasionally by emergency surgery, is required. In these patients, the source of bleeding is often difficult to identify.
Diagnostic tests utilised will vary according to institutional availability, patient characteristics, and the providers' experience and expertise. These investigations are used to localise the site of haemorrhage, cause it to stop or confirm its cessation. Options include radionuclide imaging, colonoscopy, CT angiography and selective angiography.
Radionuclide imaging with [
99m
Tc]pertechnetate-tagged red cells detects the slowest bleeding rates (0.1–0.5 mL/min) but cannot reliably localise the site of haemorrhage.
83
The timing of the radionucleotide blush is important. An immediately positive blush (within the first 2 minutes of scanning) is highly predictive of a positive angiogram (60%) and the need for surgery (24%), while a negative scan is highly predictive of a negative angiogram (93%) and the need for surgery falls to 7%.
84
A negative nuclear scan provides objective evidence that the patient is not actively bleeding and may be evaluated by colonoscopy.
Many authors believe that colonoscopy has the highest efficacy and should be the first study in patients with major bleeding that appears self-limited.
85
Whether colonoscopy should be undertaken as an ‘emergency’ depends on the patient's stability. In patients with hypotension and ongoing haemorrhage, it is difficult to safely cleanse the bowel with lavage solutions, and the continued bleeding limits intraluminal visualisation and the ability to utilise therapeutic options. In general, these patients require prompt attention with angiography or surgery.
In stable patients with self-limited haemorrhage, colonoscopy is the preferred diagnostic study. The need for bowel preparation is controversial. Colonoscopy without preparation can be described as ‘emergent’ while administering a mechanical preparation and then performing colonoscopy within 24 hours of presentation is best termed ‘urgent’ colonoscopy. The rapid time preferred for mechanical cleansing usually mandates a lavage method.
Proponents of ‘emergency colonoscopy’ have demonstrated high caecal intubation rates (95%) and a diagnostic accuracy of 72–86%.
86
However, many of the reported series described atypical aetiologies for ‘massive haemorrhage’, including ischaemic colitis, inflammatory bowel disease and cancer. The usual rate of bleeding in these conditions is more amenable to urgent colonoscopy (within 24 hours) than to emergent colonoscopy. Higher bleeding rates are more common with diverticular or angiodysplastic sources.
The benefit of colonoscopy depends on its ability to provide a definitive localisation of ongoing active bleeding and the potential for therapy. Many landmarks for colonoscopy may be obscured during haemorrhage. Even when pathology is identified, establishing the presence of stigmata of recent haemorrhage is more difficult than in the upper gastrointestinal tract. In addition, colonoscopy is associated with complications such as perforation. Inconclusive findings owing to technically unsatisfactory results are frequent and failure to achieve a firm diagnosis may reflect stricter diagnostic criteria rather than inferior diagnostic skill.
Because of the inability to appreciate all intraluminal landmarks and accurately locate the segment that is bleeding, once the enodoscopist highlights a source of bleeding, the region of the intestine is best marked with a tattoo of India ink. If the haemorrhage continues and the patient fails medical management, a tattoo greatly assists the surgeon in localising the site of bleeding. The endoscopist has many therapeutic options to control the bleeding, including thermal agents (heater probes and bipolar coagulation), monopolar or bipolar diathermy, topical or intramucosal epinephrine injection and endoscopically applied clips.
Angiography aids as a diagnostic and therapeutic option in the treatment of intestinal haemorrhage. Acute, major haemorrhage with ongoing bleeding requires emergency angiography, while patients with an early blush during nuclear scintigraphy may benefit from therapeutic angiography. Angiograms may also define a potential source for haemorrhage in occult and recurrent gastrointestinal haemorrhage. In order to appreciate an angiographic blush of contrast, the study requires a bleeding rate of at least 1–1.5 mL/min. Appropriate patient selection increases yields and avoids excessive use of angiograms. Superselective catheterisation of branch arteries and multiple injections of contrast may be required to examine the entire territory. It is important to take late films in the venous phase as they may demonstrate abnormalities that may not be seen in the arterial phase. Major complications of diagnostic arteriography include arterial thrombosis, embolisation and renal failure caused by the contrast material.
The angiographic blush may suggest a specific aetiology, but this finding lacks the accuracy of colonoscopy. Highly accurate localisation also provides for focused therapy. The haemorrhagic site may receive highly selective, intra-arterial vasopressin infusion. This medication produces potent arterial contraction that may reduce or halt the haemorrhage. Infusion rates of vasopressin begin at concentrations of 0.2 U/min and may progress to 0.4 U/min. The systemic effects and cardiac impact of vasopressin may limit maximising the dosage. Vasopressin infusion controls bleeding in as many as 91% of patients, but unfortunately bleeding may recur in up to 50% of patients once the vasopressin is reduced. However, this has been reported as allowing surgery to be performed electively in 57% of patients.
87
Angiographic technology also allows for arterial embolisation to control haemorrhage. Arterial embolisation of larger vessels produces intestinal ischaemia or infarction in approximately 20% of patients. Much safer superselective mesenteric angiography using microcatheters allows embolisation of the intestinal vasa recta or vessels as small as 1 mm.
The high re-bleeding rate with vasopressin has led to the suggestion that other, more definitive therapeutic manoeuvres should be undertaken during angiography such as superselective embolisation, which is effective in controlling colonic haemorrhage and is associated with a low rate of post-embolisation colonic ischaemia. In a study of 27 patients, all were initially controlled with arterial embolisation although six re-bled, five of whom needed surgical intervention.
88
Successful embolisation therapy provides immediate arrest of the bleeding. Embolisation agents include gelfoam pledgets, coils and polyvinyl alcohol particles. Success has been better in diverticular bleeds, which are usually fed by one vessel, than with angiodysplasias, which often have multiple feeding vessels.
Advanced CT angiography using thinly sliced, fast image acquisition combined with three-dimensional software packages has revolutionised the imaging of the vascular tree.
89
Vessels smaller than ‘named’ vessels can be visualised, and use of CT angiography has been reported in chronic conditions such as mesenteric ischaemia and inflammatory bowel disease. Case reports and animal modelling suggest feasibility for gastrointestinal haemorrhage. The sensitivity and specificity of CT angiography in patients with gastrointestinal haemorrhage are unknown and require further research.
Patients require surgical intervention if they continue to bleed profusely and medical, endoscopic and angiographic interventions have failed. Most sources of bleeding spontaneously resolve or are controlled with the therapeutic interventions described previously. Surgical therapy for intestinal bleeding is required infrequently and associated with significant mortality. Patients that are haemodynamically unresponsive to the initial resuscitation obviously require emergency surgery. In other patients the site of haemorrhage may be localised, yet the available therapeutic interventions fail to control the bleeding. Patient mortality increases with transfusion requirements. It has been reported that there is a reduced mortality (7%) for patients requiring less than 10 units of blood, while the mortality increases to 27% for patients receiving in excess of 10 units.
90
Therefore, patients with ongoing haemorrhage who require more than 6–7units of blood during the resuscitation should undergo surgical intervention.
Surgeons should tailor their operative approach based on the preoperative diagnostic evaluations. Surgery starts with a thorough examination of the entire intestine through a large midline open laparotomy incision. The first objective is location of intraluminal blood with the hope of segmentally isolating a possible source of bleeding.
If no source appears obvious after the exploration phase, the surgeon may consider intestinal enteroscopy. The enteroscope or colonoscope exposes the luminal surface and transilluminates the intestinal wall. Transillumination may identify vascular anomalies, small ulcers or tumours. Endoscopic access to the intestine may require a transoral, transgastric, transcolonic or transanal approach, or a combination of these. Intraoperative endoscopy is, however, a technically difficult endeavour. A team approach with two surgeons or the availability of an experienced endoscopist is important in order to identify the elusive lesions causing the haemorrhage.
Once the haemorrhage site is identified, an appropriate segmental resection can be performed. If no source of bleeding is confirmed, but appears to arise from the colon, the surgeon should perform a subtotal or total colectomy. Stable patients will tolerate a primary ileosigmoid or ileorectal anastomosis, while unstable patients are best served with an end ileostomy with closure of the rectal stump or a sigmoid mucous fistula.
Critical issues with operative management include delaying surgery until the haemorrhage reaches a critical point beyond 10 units of blood, with an associated mortality rate between 10% and 35%, and recurrence bleeding rates of 10% due to imprecise localisation of the bleeding.
91
Recurrence rates are higher (e.g. 20%) if a limited right or left colectomy is performed without precise localisation of the haemorrhage.
92
A total colectomy offers the same mortality but with a lower chance of recurrent or persistent bleeding. Therefore, a total colectomy is the preferred option if preoperative localisation is not possible, or a definite site of bleeding cannot be found at laparotomy.
Key points
There has been a recent swing towards greater emphasis on preoperative investigations, especially CT, and daytime operating by specialist consultant surgeons for most urgent surgery. In cases of colonic emergencies where there is faecal contamination of the peritoneal cavity, emergency surgery may still be life-saving.
Ischaemia
Mild cases are treated with bowel rest and optimisation of cardiovascular status. Severe cases usually require resection.
Colonic obstruction
Volvulus.
Emergency decompression using a proctoscope, a long soft tube such as a chest drain, or a colonoscope should, in most cases, be followed by elective definitive surgery.
Acute colonic pseudo-obstruction.
Some patients can be managed conservatively, employing regular review and frequent plain abdominal radiographs to check caecal diameter. As caecal diameter increases, decompression is required. Many patients will respond to a trial of neostigmine. If this treatment is not successful, the colon can be decompressed with the colonoscope. Failure may necessitate caecostomy or resection. If perforation or necrosis is suspected, a full laparotomy is necessary.
Malignant disease.
Right hemicolectomy with primary anastomosis is the treatment of choice for most patients with right-sided or transverse colonic obstruction. A subtotal colectomy may be the preferred option for patients who have a lesion at the splenic flexure. Most other patients who have an obstructing carcinoma in the left colon will be best treated by segmental colectomy with primary anastomosis where possible. Patients who are thought to be particularly at risk from lengthy emergency operations are best served by either segmental resection or colostomy.
Inflammation and/or infection
Diverticulits.
CT is now the investigation of choice for those patients presenting as emergencies. Abscesses can then be drained percutaneously and perforations requiring surgery diagnosed early. One-stage surgery is ideal; however, in left-sided disease this is not appropriate for the unstable patient or in those with gross faecal contamination.
Typhlitis.
There should be a high index of suspicion for this condition in the neutropenic patient. The majority respond to conservative measures including antibiotics and complete bowel rest, but perforation should be actively excluded along the course of the disease, preferably with CT.
Perforation
Stercoral perforation.
This rare condition may be difficult to diagnose preoperatively. Awareness of the possibility of postoperative re-perforation, especially where limited resections have been performed, may be the key to a successful outcome.
Anastomotic leak.
Postoperative assessment of the abdomen is complex, with many signs being attributable to the surgery itself. In equivocal cases, further investigations should be instigated where there is uncertainty. Active observation should be employed by senior surgical staff. In some instances, transfer to a specialist colorectal unit needs to be considered after initial resuscitation and assessment.
Colonoscopic damage.
Signs and symptoms of perforation are not always obvious at the time of colonoscopy. A high index of suspicion, especially following interventional procedures, may allow conservative management or primary repair if surgery is performed early.
Haemorrhage
The majority of colonic bleeding settles spontaneously, allowing subsequent investigation to take place ‘electively’. However, ongoing bleeding should be investigated by CT angiography and then selective angiography, with or without embolisation. Surgery may be required and unless an obvious cause has been identified, subtotal colectomy with stoma formation is probably the safest procedure.