Fistuloclysis can successfully replace parenteral feeding in the nutritional support of patients with enterocutaneous fistula

Transcription

1 Original article Fistuloclysis can successfully replace parenteral feeding in the nutritional support of patients with enterocutaneous fistula A. Teubner, K. Morrison, H. R. Ravishankar, I. D. Anderson, N. A. Scott and G. L. Carlson Intestinal Failure Unit, Department of Surgery, Hope Hospital, Salford, UK Correspondence to: Mr G. L. Carlson, Department of Surgery, Hope Hospital, Salford M6 8HD, UK ( Background: Use of total parenteral nutrition (TPN) in patients with acute intestinal failure due to enteric fistulation might be avoided if a simpler means of nutritional support was available. The aim of this study was to determine whether feeding via an intestinal fistula (fistuloclysis) would obviate the need for TPN. Methods: Fistuloclysis was attempted in 12 patients with jejunocutaneous or ileocutaneous fistulas with mucocutaneous continuity. Feeding was achieved by inserting a gastrostomy feeding tube into the intestine distal to the fistula. Infusion of enteral feed was increased in a stepwise manner, without reinfusion of chyme, until predicted nutritional requirements could be met by a combination of fistuloclysis and regular diet, following which TPN was withdrawn. Energy requirements and nutritional status were assessed before starting fistuloclysis and at the time of reconstructive surgery. Results: Fistuloclysis replaced TPN entirely in 11 of 12 patients. Nutritional status was maintained for a median of 155 (range ) days until reconstructive surgery could be safely undertaken in nine patients. Two patients who did not undergo surgery remained nutritionally stable over at least 9 months. TPN had to be recommenced in one patient. There were no complications associated with fistuloclysis. Conclusion: Fistuloclysis appears to provide effective nutritional support in selected patients with enterocutaneous fistula. Presented to a meeting of the European Society of Parenteral Nutrition, Cannes, France, September 2003 Paper accepted 5 January 2004 Published online 22 March 2004 in Wiley InterScience ( DOI: /bjs.4520 Introduction Effective management of patients with severe intraabdominal infection may necessitate repeated laparotomy 1 and, in some cases, leaving the abdomen open (laparostomy) until the infection resolves 2. Management is complicated in up to 25 per cent of such patients by the development of secondary intestinal fistulas that present through an open or dehisced abdominal wound 3.These fistulas are particularly demanding to manage, as lack of a cutaneous surface on which to secure a stoma appliance creates technical difficulties for control of fistula effluent 4. In addition, spontaneous fistula closure does not usually occur because healing of the abdomen by secondary intention inevitably results in mucocutaneous continuity at the fistula site. These complex problems continue to be associated with a significant mortality rate 3 and account for an important component of the workload of specialized intestinal failure units 5. As it is not technically possible to restore intestinal continuity until there has been reepithelialization of the abdominal wall and development of a neoperitoneal cavity, a process that can take up to 6 months 6, the nutritional and metabolic needs of a patient with a high-output fistula during this period are generally met by a prolonged period of parenteral nutrition, which is frequently provided at home. Parenteral nutrition is complex, expensive and highly demanding of expert nursing care if acceptable morbidity and mortality rates are to be achieved. In addition, it may not be possible to train the elderly and those with impaired vision or dexterity to administer home parenteral nutrition, necessitating prolonged hospital admission. Previous studies in adults 7 and neonates 8, however, have indicated that parenteral feeding can be avoided in selected patients after massive enterectomy, with enteral feeding administered via a mucous fistula, mixed with chyme Copyright 2004 British Journal of Surgery Society Ltd British Journal of Surgery 2004; 91:

2 626 A. Teubner, K. Morrison, H. R. Ravishankar, I. D. Anderson, N. A. Scott and G. L. Carlson collected from the proximal stoma. This technique has failed to gain popularity, possibly owing to its complexity and aesthetic considerations. It is unclear whether chyme admixture is essential to provide satisfactory enteral feeding in patients with discontinuity of the intestinal tract. In addition, previous studies of the feasibility of enteral nutrition in patients with a proximal stoma after major small bowel resection specifically excluded patients with intestinal fistulas. This study reports the results of an attempt to adapt these techniques for the nutritional support of patients with small intestinal fistulation within an open abdomen, and shows that enteral feeding via an intestinal fistula (fistuloclysis) can successfully replace parenteral nutrition, without the need for reinfusion of chyme. Patients and methods Inclusion criteria Between 1 November 1997 and 31 December 2002, fistuloclysis was attempted in all patients with a small intestinal fistula and mucocutaneous continuity within a partially or completely dehisced abdominal wound referred to the intestinal failure unit at Hope Hospital. Patients were selected for fistuloclysis if they were haemodynamically stable, free from evidence of active infection, had a mean daily fistula output in excess of 500 ml, and contrast radiology via the most distal fistula opening confirmed a minimum of 75 cm of intact small intestine available for absorption. All patients had either undergone previous surgery with formation of a laparostomy for the control of abdominal sepsis or had developed postoperative intestinal fistula within a totally or partially dehisced abdominal wound, and had been receiving total parenteral nutrition (TPN) for varying intervals before the commencement of fistuloclysis. fistula output over 3 consecutive days immediately before initiation of fistuloclysis with that in the 3 consecutive days after fistuloclysis had been successfully established (see below). Oral energy and nitrogen intake were measured using 24-h dietary recall and hospital food record charts, and dietary energy and nitrogen intakes were calculated using standard food composition tables 11. No corrections were made for stool or insensible losses of nitrogen and energy. Successful fistuloclysis was defined as the ability to maintain or increase bodyweight and normal serum biochemistry without the need to resume TPN or parenteral fluid therapy until definitive reconstructive surgery could be undertaken. Technique of fistuloclysis The total length of small intestine above and below each fistula was determined using contrast radiology and by findings at reconstructive surgery where relevant. After confirmation of the integrity and length of the small intestine beyond the most distal fistula opening, the fistula was intubated with a MIC-G TM balloon retention gastrostomy tube (Vygon, Cirencester, UK). The catheter was advanced to a depth of 5 10 cm in the lumen of the distal intestine under radiological control (Figs 1 and 2) and 5 ml water was placed in the catheter balloon. The fistula was enclosed within a stoma appliance linked to a Hollister universal catheter access port (Hollister, Wokingham, UK), which allowed collection and measurement of proximal enteric content as enteral feed was infused via the fistula (Fig. 3). Enteral feeding was commenced at an initial rate of 30 ml/h, with a standard polymeric feed (Fresubin, 1 0 kcal/ml, 300 mosm/kg, 15 0 per cent medium-chain triglycerides; Fresenius-Kabi, Warrington, UK) and the rate of infusion increased by up to 20 ml/h each day, Assessment A single dietician weighed the patients and calculated the body mass index on commencement of fistuloclysis, at weekly intervals during hospital admission, and at intervals of between 2 and 12 weeks after discharge from hospital, depending on clinical need. Energy and nitrogen requirements for each individual were assessed using the Schofield equation 9 and Elia normogram 10 respectively. Serum albumin, urea and electrolyte concentrations were measured weekly. Total feed energy and nitrogen, volume administered orally, enterally and parenterally, and fistula output were recorded daily. The effect of fistuloclysis on fistula output was calculated by comparing the mean Fig. 1 Fistuloclysis catheter placed in a loop of small intestine fistulating through a laparostomy wound

3 Fistuloclysis for nutritional support 627 Fig. 2 Small bowel contrast study via fistuloclysis catheter, showing healthy small intestine distal to the fistula before the start of fistuloclysis Fig. 3 Stoma appliance with cone system, allowing simultaneous infusion of feed and collection of proximal fistula effluent in all cases with oral omeprazole 40 mg twice daily, loperamide 4 mg four times daily and codeine phosphate mg four times daily. Failure to tolerate polymeric feed, or to be able to advance to target feeding rates, resulted in alteration to the feeding regimen. The regimen was changed initially to a semielemental feed (Perative,1 3 kcal/ml, 380 mosm/kg, 42 0 per cent lipid as medium-chain triglyceride; Abbott Nutrition, Maidenhead, UK), and the rate of feeding increased again as described above. If feed intolerance persisted, the semielemental feed was replaced by elemental feed (Emsogen, kcal/ml, 539 mosm/kg, 83 0 per cent lipid as medium-chain triglyceride; SHS International, Liverpool, UK) and the rate of feeding increased as detailed above. TPN was continued while fistuloclysis was being established, but ceased when the optimal infusion rate of 90 ml/h had been achieved for at least 5 days without adverse symptoms necessitating a reduction in the rate of feeding. If fistuloclysis with elemental feeding was poorly tolerated, the procedure was abandoned and TPN continued or reinstated. Cost comparison between fistuloclysis and parenteral nutrition Comparison of nutrition-related care costs was made using the current costs of parenteral ( 60 per day) and enteral ( 15 per day) feeding solutions in this institution. Cost differences were calculated solely on the basis of feeding solutions, without taking into account the costs of infusion apparatus or nursing time. The differential cost for fistuloclysis and parenteral nutrition during inpatient care, from the time at which parenteral nutrition was discontinued until time of discharge, was used to calculate the difference in expenditure on nutrients, per patientday of treatment and for the group as whole. Differential costs of fistuloclysis and parenteral nutrition after hospital discharge were not calculated as funding and expenditure data for patients in primary care were neither available nor pertinent to the acute hospital budget. depending on the patient s ability to tolerate feeding, until a final target feeding rate of 90 ml/h was achieved. All patients were fed by continuous infusion, for h overnight, but were encouraged to eat a low-residue diet during the day. To minimize output from the proximal limb of small bowel fistulas, oral fluid intake was restricted to a maximum of 1500 ml/day. Symptoms of feed intolerance (diarrhoea and abdominal pain) were recorded daily. Proximal fistula output and diarrhoea associated with fistuloclysis were controlled Results Patients Between 1 November 1997 and 31 December 2002, 40 patients with small bowel fistulas were admitted to the intestinal failure unit. A total of 12 patients (seven men and five women), of median age 54 (range 20 79) years, met the entry criteria for the study. All patients had small intestinal enterocutaneous fistulas with direct mucocutaneous continuity (Fig. 1). Nine patients had

4 628 A. Teubner, K. Morrison, H. R. Ravishankar, I. D. Anderson, N. A. Scott and G. L. Carlson intestinal fistulas within laparostomy wounds and three had fistulas within partially or completely dehisced abdominal wounds. The cause of intestinal failure in all patients was postoperative small bowel fistulation. Patients had developed enterocutaneous fistulas with mucocutaneous continuity following surgery for Crohn s disease (three), colonic diverticular disease (two), appendicitis (one) and gastric carcinoma (one). Five patients had developed fistulas as a result of surgery for incarcerated hernia (one), small bowel obstruction (one), abdominal trauma (one), carcinoma of the rectum (one) and hysterectomy (one). In nine patients the small bowel distal to the fistula was in continuity with the colon. Radiological studies before commencement of fistuloclysis revealed a median of 105 (range ) cm of small intestine from the duodenojejunal junction to the most proximal fistula opening and 120 (range ) cm of small intestine beyond the most distal fistula opening. Nutritional assessment and requirements At the time of entry to the study all patients had been receiving parenteral nutrition for median of 7 (range 5 7) nights per week and had been doing so for 4 0 (range ) weeks. All patients were considered dependent on TPN for either maintenance of satisfactory fluid balance (two) or nutritional status (ten) before attempted fistuloclysis. Median estimated oral energy intake was 550 (range ) kcal/day and nitrogen intake 3 0 (range ) g/day. Predicted energy and nitrogen requirements, and bodyweight, body mass index and serum albumin concentration, on commencement of fistuloclysis are shown in Table 1. Table 1 Nutritional status, requirements and nutritional support administered before fistuloclysis and at time of readmission for surgery Before fistuloclysis (n = 12) At time of surgery (n = 9)* TPN energy (kcal/day) 1800 ( ) TPN nitrogen (g/day) 9 4 ( ) Energy requirement 26 5 ( ) 25 7 ( ) (kcal per kg per day) Nitrogen requirement (g/day) 13 5 ( ) 13 2 ( ) Weight (kg) 70 (46 106) 78 (53 104) Body mass index (kg/m 2 ) 25 5 ( ) 25 0 ( ) Albumin (g/dl) 33 5 ( ) 40 (28 45) Fistuloclysis energy (kcal) 1300 ( ) Fistuloclysis nitrogen (g) 9 (5 15) Values are median (range). TPN, total parenteral nutrition. *Two patients were not readmitted for surgery. All patients experienced abdominal discomfort and/or diarrhoea within 24 h of initiating fistuloclysis. Target rates of feeding using polymeric feed were achieved in three patients whereas feed intolerance (abdominal pain and/or diarrhoea) precluded progression to a full rate of enteral feeding in nine. Four of these patients were able to tolerate target rates of feeding after introduction of a semielemental feed. Five patients were also unable to tolerate the semielemental feed, but four of these were able to tolerate target rates of feeding with elemental feed. Fistuloclysis successfully replaced TPN in 11 of the 12 patients, by a median of 28 (4 68) days after commencement. Total daily energy and nitrogen administered to these patients is shown in Table 1. Fistuloclysis was unsuccessful in one patient. This patient had been at home for 287 days before recommencing parenteral nutrition 4 nights per week (1800 kcal plus 9 0 g nitrogen per day) owing to weight loss. Data on fistula output were available for only 11 of the 12 patients. Median proximal fistula output before fistuloclysis was 1360 (range ) ml/day. There was no significant effect of fistuloclysis on proximal fistula output. In the 11 patients in whom fistuloclysis was successfully established, the median fistula output was 1170 (range ) ml/day. In six patients fistuloclysis was associated with an increase in proximal fistula output by ml/day, and in four patients the output decreased by ml/day. Nine of the 11 patients successfully underwent reconstructive surgery a median of 155 (range ) days after commencing fistuloclysis. Two patients underwent surgery during the same hospital admission, 19 and 31 days respectively after fistuloclysis had been established. Seven of the nine patients received fistuloclysis at home and were subsequently admitted for reconstructive surgery a median of 210 (range ) days after discharge. The weight, body mass index and serum albumin concentration in these nine patients at the time reconstructive surgery are shown in Table 1. One patient died from ischaemic heart disease at home, having been considered unfit for reconstructive surgery, and another did not undergo reconstructive surgery because of medical co-morbidity. Both patients received fistuloclysis as the sole form of nutritional support at home, for 9 and 18 months respectively, without developing nutritional impairment (data not shown). The total cost of enteral feeds, for all patients, from the day the TPN ceased to the day of discharge from hospital was Assuming a typical cost of TPN of 60

5 Fistuloclysis for nutritional support 629 per patient per day the cost of providing parenteral nutrition for this group of patients for the duration of hospital admission would have been Discussion Provision of safe and effective nutritional and metabolic support for patients with high-output small intestinal fistulas has contributed significantly to the improvement in prognosis over the past four decades 12, and parenteral nutrition remains the principal route by which nutritional support is administered in these circumstances. The practical difficulties of managing patients with an open abdomen (laparostomy) associated with, or complicated by, the development of a small intestinal fistula has led to the development in the UK of specialized intestinal failure units. Such units are run by a multidisciplinary team capable of providing the medical, nursing, dietetic and psychological care required for the rehabilitation of these patients 5. Although the ultimate goal of surgical treatment is to reconstruct the intestinal tract, inability to re-enter the abdomen until a neoperitoneal cavity has been established usually requires a period of approximately 6 months 6, during which time there is a continued need for nutritional and metabolic support. Provided sepsis has been controlled and the patient is metabolically and nutritionally stable, this treatment is usually administered as home parenteral nutrition 6, obviating the need for many months of hospital admission. The ability to provide prolonged parenteral nutrition and to train patients to administer home parenteral nutrition with minimal morbidity 13,14 is a key element in the delivery of this care and remains a principal reason for referral of patients to specialized units 15. Although it would be inappropriate to cannulate the track of an enterocutaneous fistula if spontaneous closure was a possibility, the finding of intestinal mucosa on the skin surface or embedded within the granulation tissue of a wound healing by secondary intention signifies that spontaneous closure will not occur. Although successful maintenance of nutritional and metabolic status has been reported after infusion of enteral feed (and chyme) into defunctioned intestinal loops after small bowel resection for severe abdominal sepsis 7,8,16, these studies specifically excluded patients with intestinal fistulas and the potential for enteral feeding via an intestinal fistula (fistuloclysis) has not been explored previously. The findings of the present study suggest that fistuloclysis can satisfactorily maintain nutritional and metabolic status in a group of patients who would otherwise have required TPN. In addition, delivery of enteral feed into the small intestine appeared to be effective, even though no attempt was made to collect or reinfuse chyme. Although this seems surprising, the feeding solutions selected for fistuloclysis were chosen because of their high content of medium-chain triglycerides, which may be absorbed in the small intestine (and possibly in the colon), independent of biliary and pancreatic secretions 17. In addition, although enteral feeds were selected with the aim of minimizing osmolality, and thus diarrhoea 18, the elemental and semielemental feeds appeared better tolerated than the polymeric feed. The small number of patients in the present study precluded statistical analysis, but there was no apparent relationship between the length of small intestine and colon available for feed absorption and the type of feed tolerated. This suggests that factors other than the mechanics of feed absorption determine enteral feed tolerance in these patients. Medium- and long-chain triglycerides in enteral feeds have been shown differentially to activate gut hormone release and gut motility 19. This variability in the pattern of hormone release and/or motility of the remaining intestine may explain the lack of correlation between the residual length of intestine and feed tolerance. Although 100 cm or less of small intestine was available for absorption in six patients, this did not seem to preclude effective enteral nutrition, possibly because the majority of these patients also had intact colon. Previous studies have shown that short bowel syndrome can be avoided with even shorter lengths of small intestine, provided that the colon is intact and in continuity 20. Nevertheless, the ability of two patients with only 100 cm of small intestine (and no colon) distal to the fistula to maintain satisfactory nutritional and metabolic status over several months with fistuloclysis alone is an important observation. This is particularly so as a quarter of the patients had originally undergone colectomy for inflammatory bowel disease, known to be a major factor in the epidemiology of short bowel syndrome following later complicated laparotomy 21. Enteral feeding may have considerable advantages over parenteral feeding in critically ill surgical patients such as those in the present study. In particular, enteral, as opposed to parenteral, nutrition has been suggested to improve intestinal barrier function 22, reduce the rate of infectious complications in critically ill patients 23 and to maintain immune function 24. From a practical standpoint, surgery to restore intestinal continuity is technically difficult in patients in whom the small intestine distal to a fistula has been deprived of enteral nutrition for several months, at least in part because of marked atrophy of the small intestine. This results in disparity between the bowel ends

6 630 A. Teubner, K. Morrison, H. R. Ravishankar, I. D. Anderson, N. A. Scott and G. L. Carlson and relatively poor quality of intestinal tissues for anastomosis. Although no markers of intestinal atrophy were measured in the present study, the surgeons noted that the small intestine that had been used for nutrient delivery in all patients who underwent reconstructive surgery after a period of fistuloclysis was of much greater calibre and thickness and held sutures more readily than usual (unpublished observations). It is assumed that the delivery of enteral feed had prevented or at least limited the intestinal atrophy associated with prolonged parenteral nutrition. Although the aim of this study was to evaluate the practicality and efficacy of fistuloclysis, rather than to perform a detailed cost benefit comparison with parenteral nutrition, avoidance of prolonged parenteral nutrition in the most complex group of patients with acute intestinal failure would be expected to be associated with significant cost savings. Taking only the cost of the nutritional support into account, and using local costs for parenteral and enteral nutrient mixtures, fistuloclysis provided nutritional support for at least 45 less per patient-day than parenteral nutrition. The potential savings would have been significantly greater if the cost of the sterile infusion apparatus and dressings (required for parenteral nutrition but not fistuloclysis) and the additional nursing time required for parenteral nutrition had been taken into account. Finally, the training period for fistuloclysis (less than 2 weeks in all seven patients who were discharged from hospital and received fistuloclysis at home) was significantly shorter than that required for home parenteral nutrition (usually 4 6 weeks) 25. Fistuloclysis may therefore reduce the length of hospital stay for selected patients. Although fistuloclysis cannot replace parenteral nutrition in the management of all patients with intestinal fistula, it may provide an effective, safe and inexpensive alternative to parenteral nutrition for those in whom secondary intestinal fistulation occurs within an open abdomen. References 1 Anderson ID, Fearon KC, Grant IS. Laparotomy for abdominal sepsis in the critically ill. Br J Surg 1996; 83: Mughal MM, Bancewicz J, Irving MH. Laparostomy : a technique for the management of intractable abdominal sepsis. Br J Surg 1986; 73: Bosscha K, Hulstaert PF, Visser MR, van Vroonhoven TJ, van der Werken C. Open management of the abdomen and planned reoperations in severe bacterial peritonitis. Eur J Surg 2000; 166: Hughes S, Myers A, Carlson G. Care of intestinal stoma and enterocutaneous fistula. In Intestinal Failure, Nightingale JMD (ed.). Greenwich Medical Media: London, 2001; Scott NA, Leinhardt DJ, O Hanrahan T, Finnegan S, Shaffer JL, Irving MH. Spectrum of intestinal failure in a specialised unit. Lancet 1991; 337: Scripcariu V, Carlson G, Bancewicz J, Irving MH, Scott NA. Reconstructive abdominal operations after laparostomy and multiple repeat laparotomies for severe intra-abdominal infection. Br J Surg 1994; 81: Levy E, Frileux P, Sandrucci S, Ollivier JM, Masini JP, Cosnes J et al. Continuous enteral nutrition during the early adaptive stage of the short bowel syndrome. Br J Surg 1988; 75: Al-Harbi K, Walton JM, Gardner V, Chessell L, Fitzgerald PG. Mucous fistula refeeding in neonates with the short bowel syndrome. JPediatrSurg1999; 34: Schofield, WN. Predicting basal metabolic rate. New standards and review of previous work. Human Nutr Clin Nutr 1985; 39(Suppl 1): Elia M. Artifiicial nutritional support. Med Int 1990; 82: McCance RA, Holland B, Widdowson EM. McCance and Widdowson s the Composition of Foods (5th rev). Royal Society of Chemistry: Cambridge, Meguid MM, Campos AC. Nutritional management of patients with gastrointestinal fistulas. Surg Clin North Am 1996; 76: Williams N, Carlson GL, Scott NA, Irving MH. Incidence and management of catheter-related sepsis in patients receiving home parenteral nutrition. Br J Surg 1994; 81: Williams N, Scott NA, Irving MH. Catheter-related morbidity in patients on home parenteral nutrition: implications for small bowel transplantation. Ann R Coll Surg Engl 1994; 76: Carlson GL. Surgical causes and management. In Intestinal Failure, Nightingale JMD (ed.). Greenwich Medical Media: London, 2001; Calicis B, Parc Y, Caplin S, Frileux P, Dehni N, Ollivier J-M et al. Treatment of postoperative peritonitis of small bowel origin with continuous enteral nutrition and succus entericus reinfusion. Arch Surg 2002; 137: Jeppesen PB, Mortensen PB. The influence of a preserved colon on the absorption of medium chain fat in patients with small bowel resection. Gut 1998; 43: Keohane PP, Attrill H, Love M, Frost P, Silk DB. Relation between osmolality of diet and gastrointestinal side effects in enteral nutrition. BMJ (Clin Res Ed) 1984; 288: Vu MK, Verkijk M, Muller ES, Biemond I, Lamers CB, Masclee AA. Medium chain triglycerides activate distal but not proximal gut hormones. Clin Nutr 1999; 18: Messing B, Lemann M, Landais P, Gouttebel MC, Gerard-Boncompain M, Saudin F et al. Prognosis of patients with nonmalignant chronic intestinal failure receiving long-term home parenteral nutrition. Gastroenterology 1995; 108: Agwunobi AO, Carlson GL, Anderson ID, Irving MH,

7 Fistuloclysis for nutritional support 631 Scott NA. Mechanisms of intestinal failure in Crohn s disease. Dis Colon Rectum 2001; 44: Kudsk KA. Current aspects of mucosal immunology and its influence by nutrition. Am J Surg 2002; 183: Kudsk KA, Croce MA, Fabian TC, Minard G, Tolley EA, Poret HA et al. Enteralversus parenteral feeding. Effects on septic morbidity after blunt and penetrating abdominal trauma. Ann Surg 1992; 215: Shou J, Lappin J, Daly JM. Impairment of pulmonary macrophage function with total parenteral nutrition. Ann Surg 1994; 219: Morley JP, Myers A, O Toole J, Malik RA, Scott NA. Home parenteral nutrition (HPN) training. Does a dexterity score relate to the suitability and length of time required for a patient to become competent? Clin Nutr 1999; 18(Suppl 1): 17 (Abstract).