Evidence for using somatostatin analogues in the treatment of enterocutaneous fistula.

Enterocutaneous fistulas cause considerable mortality and morbidity following gastrointestinal surgery. Most occur after surgery for malignant, diverticular or inflammatory bowel disease; fewer patients with these conditions develop fistulas spontaneously. Care is often long and costly1, although spontaneous closure can occur with non-operative management in 60–75 per of patients2. The goals of management for enterocutaneous fistula are therefore to promote conditions for spontaneous fistula closure or, if this is not possible, to restore physiological stability in order to optimize the patient for definitive surgery. Although somatostatin analogues are used widely as part of this management plan, the evidence for their use is questionable. Once sepsis has been treated, nutritional status restored and distal obstruction excluded, the management of enterocutaneous fistula is centred on reduction of fistula output. Output can be modified by slowing gastrointestinal transit with loperamide, optimization of luminal absorption with oral rehydration salts and elemental or lowresidue feeds, and reduction of enteric secretions with proton pump inhibitors and somatostatin or its analogues. Although there is heterogeneity between studies, it is generally agreed that production of at least 500 ml/day represents a high-output fistula1,3. The premise of reducing fistula output is based on the observation that low-output fistulas are associated with greater likelihood of spontaneous closure. A series of 110 conservatively managed fistulas reported an odds ratio for spontaneous closure of 0·17 (95 per cent confidence interval 0·07 to 0·40; P < 0·001) in the context of high output3. Somatostatin is an endogenous hormone that reduces enteric secretion, increases water and electrolyte absorption, and inhibits pancreatic exocrine secretion. It is postulated that somatostatin could expedite spontaneous fistula closure both by directly reducing fluid output and in some situations by reducing the destructive enzymatic content of fistula effluent. Such treatment may reduce the length of hospital stay or decrease the number of patients who eventually require operative fistulamanagement, ultimately reducing treatment costs. The short half-life of somatostatin, however, necessitates continuous intravenous infusion for prolonged periods, making the use of longer-acting somatostatin analogues (octreotide, lanreotide) more appealing. Octreotide (Sandostatin; Novartis, Basle, Switzerland) and lanreotide (Somatuline; Ipsen, Paris, France) are synthetic octapeptides with greater potency and a longer half-life than somatostatin. They have poor affinity for somatostatin receptors (SSTRs), binding mostly to SSTR-2 and SSTR-54,5. Human gastric and colonic tissues contain predominantly SSTR-1 or SSTR46, but the receptor subtypes in human small bowel are not known. Downregulation of SSTR expression may limit treatment efficacy with prolonged administration7, and intermittent dosing schedules may impair the clinical effect of octreotide owing to a rebound increase in secretions between doses5,8. The novel somatostatin analogue pasireotide (SOM230; Novartis), which acts on multiple SSTRs (SSTR-1, -2, -3 and -5), has not been evaluated in the treatment of enterocutaneous fistula9. Although low-output fistulas have a greater likelihood of spontaneous closure, it is unclear whether conversion of a high-output fistula into one with low output actually increases the probability of spontaneous closure3. Fistula output is itself intimately related to the underlying anatomy of the fistula track. Small bowel defects give rise to large volumes of caustic effluent compared with colonic defects, a factor that may explain the poorer healing of high-output fistulas regardless of the absolute volume of effluent produced. Complex fistulas may arise from multiple levels of the gastrointestinal tract, often associated with intra-abdominal abscess formation, and are less likely to close spontaneously than simple fistulous tracks10. Poiseuille’s law may also explain other anatomical factors likely to influence fistula healing. The physics of resistance to flow would suggest that small defects with long tracks are more likely to heal than short, wide defects or where there is complete obliteration of bowel continuity or continued distal obstruction. Long-standing fistulas with epithelialized tracks will not heal and comorbid conditions may impair the ability of the fistula track to close