The management of enterocutaneous fistulae and how to avoid them

The management of intestinal fistulae is best carried out by taking a multidisciplinary team (MDT) approach.¹ An MDT approach ensures that the patient is managed holistically and has access to doctors from various specialties (including gastroenterologists and intestinal failure physicians, surgeons, radiologists, intensivists, biochemists and psychiatrists), experienced nurses, pharmacists, dieticians, stoma therapists, psychologists, occupational therapists and physiotherapists. Involving an MDT can generate conversations, views and strategies that facilitate optimum patient care. Indeed, engaging expertise from different disciplines helps to prevent blind spots in treatment plans and facilitates timely decision making, whilst ensuring that patients are aware that there is a structured course of action. Each patient will not necessarily require all of the professionals mentioned above to be involved in their care, but it is an important strategic aim of any service seeking to care for these patients to have ready access to such professionals. Where some of these professionals are not available, it requires roles within the team to be extended to compensate for the deficiency. This situation may mean that a patient misses out on key knowledge and education, attitude and empathy or timely decisions that could have a profound impact on morbidity and mortality.

A careful, methodical, systematic approach to management is what keeps a patient’s care pathway moving forward. Being systematic also prevents high-risk strategies (i.e. trying out the latest ‘flavour of the month’) from being pursued and ensures that the entire MDT has a safe, cohesive and structured approach. The optimal strategy for managing gastrointestinal fistulae  that has stood the test of time is the ‘Sepsis-Nutrition-Anatomy-Plan’ or ‘SNAP’ approach (figure 1).¹

In some cases, a fistula may spontaneously heal without having to resort to surgery and it is important to manage patient expectations regarding the possibility of spontaneous fistula closure. In those patients who are nutritionally well with optimal medical therapy, it is our experience that of those fistula that could close spontaneously more than 90% will have done so by 6 weeks. So, if at 6 weeks the fistula has not closed, it is highly unlikely to close and will therefore require surgical intervention. There are several factors that make spontaneous fistula closure less likely, including: spontaneous development of a fistula; a fistula that developed more than 6 weeks ago; a high-output fistula; intercurrent sepsis; and the presence of a distal intestinal obstruction. Whilst these are the most common and important factors, it is clear that other comorbidities, the presence of steroids or an ageing population may prevent spontaneous healing. It is also important to examine the patient’s abdomen—if intestinal mucosa can be seen on the surface of the abdomen then the fistula is less likely to close. 

There is rarely any benefit in prolonging the recommendation for nil by mouth beyond 4–6 weeks simply in the hope that this will help the fistula close spontaneously. The role of preventing oral intake should be to determine the natural baseline volume of effluent produced via the fistula (e.g. for a 48-hour period while the patient is receiving parenteral support), which can then be used to help determine the impact of reintroducing food and drink. In the presence of a fistula that’s feeding an intra-abdominal cavity and/or collection, reducing enteric contents containing food can also help control infection. There is a risk of psychological harm from prolonged lack of eating so the reason to recommend a nil-by-mouth approach should be carefully discussed with the patient, with an outline of the expected timeframe given. 

A series of questions already exists for the assessment of patients with short bowel syndrome and the principles of dietary intake and fluid management applied to these patients should also be applied to patients with enterocutaneous fistulae. In the absence of control and due to the loss of absorptive capacity within the gut, the natural societal urge to drink more when thirsty will actually dehydrate further and deplete the patient of key electrolytes, such as sodium, potassium and magnesium. Water absorption should, therefore, be optimised by ensuring patients are consuming oral glucose saline solutions that have a sodium concentration >90 mmol (e.g. double-strength diarolyte). To minimise enteric secretions (thus water loss), slow gut transit time and maximise absorption, medications, including proton pump inhibitors (PPI), loperamide and codeine (used with caution in elderly patients), should also be optimised. 

Enterocutaneous/atmospheric effluent can burn the skin and cause leakage from stoma appliances, worsening the integrity of the skin. Checking the pH of the fistula effluent can ensure that effective dosing of PPIs occurs, reducing acid injury to the skin, and may have the added benefit of reducing the effluent output as well. 

Octreotide is a somatostatin analogue that reduces pancreatic secretions and has been proposed to be used in the management of gastrointestinal fistulae because it reduces gastrointestinal secretions. A metaanalysis of nine randomised controlled trials comparing ocreotide with placebo for the management of enterocutaneous fistulae found in favour of using octreotide, with an increased closure rate and shorter closure time.² However, there was heterogeneity in the trial designs and/or outcomes measured and frequently only a small number of patients were included. Several of the studies evaluated also included pancreatic or biliary fistulae, which are more likely to respond to octreotide therapy than enterocutaneous fistulae. Furthermore, it is painful for patients to have the required three subcutaneous injections of octreotide per day and longer acting somatostatin analogues are more expensive. If trialled in individual patients, then the likely efficacy of these medications needs to be measured against the factors that make spontaneous fistula closure less likely (see mistake 3) and for a maximum of 72 hours. 

Prior to reconstructive surgery, consideration should be given to distal enteral tube feeding (fistulocylsis) for at least 2 weeks, but optimally for 6 weeks. Enteral tube feeding down the most distal limb of an enterocutaneous fistula promotes rejuvenation of the gut mucosal surface so that it can start absorbing nutrients. In addition, distal feeding can reverse any atrophy of the gut segment, facilitating surgical reanastomosis and potentially reducing the risk of an anastomotic leak or refistulation. 

Many intestinal fistulae that occur in patients who have Crohn’s disease are the result of a surgical complication, rather than being caused by the Crohn’s disease itself. The efficacy of anti-TNF agents for closing enterocutaneous fistula is limited—long-term closure occurs in less than one-fifth of cases and an intra-abdominal abscess forms in nearly one-third of those treated with anti-TNF agents.³ In addition, (as per mistake 3), it is important to recognise the factors that make fistula closure less likely when deciding whether or not to commence biological agents to manage an enterocutaneous fistula. 

The optimum timing of surgery is vital to ensure that each patient undergoes as few further procedures as possible (as outlined in mistake 2).  In particular, if the fistula is an iatrogenic complication, there can be a desire to reoperate and resolve the problem as soon as possible; however, operating before sepsis is drained, nutrition optimised and the anatomy clearly defined exposes the patient to an increased risk of an unsuccessful outcome. Experience also suggests that  post-surgical intraperitoneal adhesions soften with time, facilitating reoperation during what is likely to be a complex procedure. Finally, even after a patient is stabilised, a period of waiting, often with a patient at home on home parenteral nutrition, can improve the chances of a successful and definitive reconstruction. The optimum advocated time in the literature varies, with some recommending as early as 6 weeks⁴ whereas most recommendations are for at least 6 months.^{5, 6}