Pre-operative feeding

Poor nutritional status is known to increase operative risks such as infective complications, the development of deep vein thrombosis, impaired wound healing and respiratory failure⁽Reference Visbal, Allen and Miller^6–Reference Law, Wong and Kwog¹⁰⁾. Significant respiratory complications are reported to occur in 18–30% of cases following oesophagectomy and can carry a 50% mortality rate⁽Reference Smithers, Gotley and Martin^11–Reference Rente, Bull and Harpole¹⁴⁾. All attempts should be made to address nutritional issues as soon as possible after diagnosis, ideally with a dietetic review at the initial clinic appointment.

The majority of patients with swallowing difficulties can tolerate oral supplementation. The enteral route should always be used when possible and on occasion this may involve the passage of a nasogastric feeding tube. Pre-operative total parenteral nutrition should be avoided in well-nourished or mildly undernourished patients as it provides either no benefit or increased morbidity⁽Reference Braga, Ljungqvist and Soeters¹⁵⁾.

In a study by Gianotti et al. 305 patients undergoing surgery for gastrointestinal cancer were randomised to either pre-operative nutrition alone, pre- and peri-operative nutrition or no nutritional supplementation⁽Reference Gianotti, Braga and Nespoli¹⁶⁾. A clear benefit was demonstrated in the group of patients receiving pre-operative nutrition compared with those in the no-supplement group with a reduction in infection rates from 30 to 13%. Length of hospital stay was shortened from 14 to 11 d. This study demonstrated no additional benefit to the addition of peri-operative nutrition to those patients receiving pre-operative nutrition.

A large number of patients considered suitable for curative resection will receive pre-operative or neo-adjuvant therapy⁽Reference Halliday, Skipworth and Wall¹⁷⁾. In the United Kingdom most patients receive pre-operative chemotherapy alone based on the Medical Research Council OEO2 trial of Cisplatin and 5-fluorouracil delivered in two cycles⁽¹⁸⁾. This demonstrated a 2-year survival advantage of 9% in the group treated with chemotherapy followed by surgery over the surgery alone group (43% v. 34%). A smaller but significant survival advantage remained at 5-year follow-up (5-year 23% v. 17%). During the time of chemotherapy some patients will have an improvement in their swallowing as a result of tumour response to treatment.

Early post-operative feeding

Published literature on the role of early post-operative feeding in patients undergoing oesophagectomy does not demonstrate clear benefit. A meta-analysis of thirteen randomised controlled trials of early feeding (within 24 h) in patients undergoing intestinal surgery reported a reduced mortality but an increased rate of vomiting⁽Reference Lewis, Andersen and Thomas¹⁹⁾. While vomiting may not be a significant risk factor in colorectal surgery it is a very major concern in oesophageal surgery where the anastomosis lies proximally and vomiting should be avoided at all costs. An anastomotic leak is a catastrophic complication with a high mortality⁽Reference Lerut, Coosemans and Decker²⁰⁾. Gianotti et al. have reported that pre-operative supplementation is as effective as peri-operative nutrition in improving outcome⁽Reference Gianotti, Braga and Nespoli¹⁶⁾. Similarly, a prospective randomised trial of early enteral feeding after upper gastrointestinal surgery with immune-enhancing formula compared with intravenous crystalloid solutions reported no significant differences in minor, major or infectious complications between the two groups⁽Reference Heslin, Latkany and Leung²¹⁾. The authors recommended that early enteral feeding with immune-enhancing formula was not beneficial. Evidence that immediate post-operative feeding may have a detrimental effect has been reported⁽Reference Watters, Kirkpatrick and Norris²²⁾. This study involved patients undergoing either oesophagectomy or pancreatoduodenectomy who received immediate post-operative enteral feeding via jejunostomy or no enteral feeding during the first six post-operative days. Post-operative vital capacity and forced expiratory volume (1 s) were lower in the group of patients receiving feeding. Post-operative mobility was also lower in the feeding group. Early mobilisation is essential to reduce the risk of chest infection and pulmonary embolism following oesophagectomy. Post-operative feed should be administered at a controlled rate with particular attention paid to patient's symptoms of abdominal bloating and nasogastric tube output. While few upper gastrointestinal surgeons would support a prolonged period of ‘nil-by-mouth’ following major surgery many units, including our own, prefer to commence feeding within 48 h of surgery⁽Reference Wakefield, Mansell and Baigrie²³⁾.

Post-operative nutrition should be via the enteral route whenever possible to maintain gut integrity and function. A combination of enteral and supplementary parenteral feeding may be required in some cases⁽Reference Braga, Ljungqvist and Soeters¹⁵⁾. The cost of total parenteral nutrition is reported to be three to ten times that of enteral nutrition and requires close biochemical monitoring⁽Reference Wakefield, Mansell and Baigrie^23–Reference Baigrie, Devitt and Watkin²⁵⁾. Enteral access can be either by jejunostomy or nasoduodenal tube. A randomised trial comparing feeding jejunostomy with nasoduodenal feeding in patients undergoing oesophagectomy reported both methods were effective in providing enteral feeding⁽Reference Han-Geurts, Hop and Verhoef²⁶⁾. Tube displacement occurred in 23% of patients with a nasoduodenal tube compared with 6% with a jejunostomy. Thirty percent of patients in the jejunostomy group and 21% in the nasoduodenal tube group required feeding beyond 14 d. Seven patients with anastomotic leaks were discharged with enteral feeding.

Long-term nutritional needs

The main indications for nutritional support in patients undergoing oesophagectomy are for patients who develop a post-operative complication or those with prolonged poor nutritional intake. The need for prolonged periods of feeding following complications of oesophagectomy have been reported in several series⁽Reference Sica, Sujendran and Wheeler^27–Reference Han-Geurts, Verhoef and Tilanus²⁹⁾. In a UK study of 262 consecutive patients undergoing oesophagectomy 63% of patients required enteral nutrition for 10 d or more⁽Reference Sica, Sujendran and Wheeler²⁷⁾. Feeding was started on the first post-operative day via a jejunostomy. Nineteen percent required nutritional support for more than 20 d. In a similar-sized study from India, patients undergoing oesophagectomy received a mean of 16 d on jejunostomy feeding⁽Reference Gupta²⁸⁾. Almost two-thirds of the patients developed an anastomotic leak, and half of the patients with post-operative complications required enteral feeding beyond 30 d. No serious complications relating to the jejunostomy were reported in the 204 patients and home feeding was established in those who felt well enough for discharge⁽Reference Gupta²⁸⁾. Han-Geurts et al. reported an anastomotic leak rate of approximately 9% in 150 patients undergoing oesophagectomy⁽Reference Han-Geurts, Verhoef and Tilanus²⁹⁾. The median duration of enteral support via jejunostomy was 11 d with a range of 2–126 d. In patients who require long-term nutritional support a feeding jejunostomy is considered more comfortable than a nasoduodenal tube and has a much lower displacement rate⁽Reference Sica, Sujendran and Wheeler²⁷⁾.

Feeding jejunostomies

Several varieties of tube have been described for use as a surgical jejunostomy. The use of a Foley catheter or t-tube is advocated by some and these have the advantage of a wide bore, allowing administration of medications in addition to feed and a potential to reduce the risk of tube blockage. Complications have been reported in relation to intestinal obstruction as a result of over-distension of the catheter balloon⁽Reference Chester and Turnbull³⁰⁾. It is practice in some units to perforate the balloon at the time of insertion to prevent this occurrence. The larger diameter does potentially increase the leak rate and there are now several purposely designed jejunostomy kits available. A feeding jejunostomy can be placed in the vast majority of patients undergoing oesophagectomy⁽Reference Gupta²⁸⁾. Reported complication rates for feeding jejunostomy inserted during elective surgery vary greatly from 1·1 to 45%⁽Reference Han-Geurts, Verhoef and Tilanus^29, Reference Han-Geurts, Verhoef and Tilanus^31–Reference Date, Clements and Gilliland³⁵⁾. The majority of complications reported in modern series are of a minor nature. These include occlusion, catheter displacement and local cellulitis at the site of insertion⁽Reference Gupta²⁸⁾. Serious complications include leakage into the peritoneal cavity resulting in peritonitis, volvulus at the point of fixation to the anterior abdominal wall, aspiration pneumonia, necrotising fasciitis or jejunal necrosis at the site of catheter insertion, septicaemia and pneumatosis intestinalis⁽Reference Wakefield, Mansell and Baigrie²³⁾.

Personal series

Fifty consecutive patients undergoing oesophagectomy at the Royal Infirmary of Edinburgh under the care of the author were identified from the unit database. Notes were available for forty-eight patients. There were nine females and thirty-nine males. Thirty-eight patients had adenocarcinoma, eight squamous cell carcinoma, one high-grade dysplasia and one neuroendocrine tumour. The average age was 62 years (range 43–82). Twenty-seven patients underwent oesophageal resection using a right thoracic approach and twenty-one a thoracoscopically assisted oesophagectomy⁽Reference Akiyama, Tsurumaru and Udagawa^36, Reference Smithers, Gotley and McEwan³⁷⁾. Thirty-five percent (17/48) had received pre-operative chemotherapy. A feeding jejunostomy using a FREKA® surgical jejunostomy set was inserted in all cases and feeding started at 48 h post-operatively. Supplemental feeding was continued until a dietary review determined that adequate nutrition could be taken orally.

The median length of jejunostomy tube insertion in patients with complications was 11·5 weeks (range 4–32 weeks). In patients without complications the median duration was 8 weeks (1–32 weeks).

The unit policy is to discharge patients home with a jejunostomy tube in situ and all patients are taught to flush the tube with water daily. Nine patients (19%) were discharged home on feeding. The main reason was poor oral intake in five patients. Two patients with anastomotic leaks, one patient felt to be at high risk of aspiration and one patient with a temporary recurrent laryngeal nerve palsy were also discharged on home feeding. Four patients required feeding to be restarted at clinic follow-up, all due to poor oral intake. Recorded weights pre-operatively, at 4 weeks and 6 months post-operatively were available for thirty-five out of forty-eight patients (Fig. 1). At 6 months the average weight loss was 8·4 kg with only 8% regaining their pre-operative weight.

Fig. 1. Weight change in thirty-five patients undergoing oesophagectomy at 4 weeks and 6 months post-operatively.

A significant reduction in weight at 6 months post-operatively might be expected in those patients developing a significant post-operative complication. A total of twenty-four complications occurred in twenty-one patients (21/48 (44%)). These included ten patients with pneumonia, five anastomotic leaks, one post-operative haemorrhage, one chyle leak, three patients with atrial fibrillation, one prolonged ileus, one recurrent laryngeal palsy with resolution, one aspiration and one pulmonary embolism. There were no mortalities. Weights were recorded at 6 months post-operatively in seventeen out of twenty-one patients. Large weight losses were recorded in patients irrespective of the development of post-operative complications (Table 1).

Table 1. Weight loss at 6 months following oesophagectomy in thirty-seven patients in relation to the development of post-operative complications and recurrent disease

All forty-eight patients were beyond 24 months from operation when reviewed. Fourteen patients who developed recurrent disease within 24 months were considered as early recurrence. Three patients survived less than 6 months from surgery. The average weight loss in patients with recurrent disease developing within 24 months at 6 months post-oesophagectomy was similar to the group of patients with no evidence of recurrent disease (Table 1).

Of patients 12·5% (6/48) had minor tube complications. Four fell out at 9 weeks, 12 weeks, 6 months and 7 months without complication. A tube pulled out at 7 d caused no problems. A superficial wound infection at 4 weeks was treated with removal and a short course of oral antibiotics. No tube required replacement or reoperation.