Perioperative nutrition

The papers in the perioperative nutrition section this year nearly all have a pancreatic flavor. Last year Thorrell et al. [1] reviewed the short-term changes in insulin secretion and resistance that accompany surgical stress. They pointed out that, although hyperglycemia is an important marker of this state, insulin has a profound effect on all anabolic processes, not only in carbohydrate but also in protein and fat metabolism. Wolfe [2] has shown, for example, that the negative nitrogen balance after burn injury persists for many weeks, and after the acute stage relates more to diminished protein synthesis than to increased protein catabolism. This is mediated by reduced substrate availability due to lower than normal levels of amino acids in the intracellular pool. This in turn is caused by impaired uptake of amino acids across the cell membrane. Wolfe showed further that amino acid uptake could be enhanced by high doses of insulin, and that this results in normalization of protein synthesis. The complications associated with hyperglycemia, with its increased risks of nosocomial infection, were reviewed by Khondhiar et al. [3] in the same issue. These observations concerning short-term changes in insulin sensitivity are extended in the present issue by Wright (pp. 5-10) and by Coulston (pp. 11-15), who discuss not only the transient chronic insulin resistance of stress, but also the effect of feeding in chronic insulin resistance syndromes such as type 2 diabetes mellitus. It could be argued simplistically that diabetic persons should be given the same nutritional formulation as nondiabetic persons, and that the hyperglycemia should be controlled using sliding scale insulin dosage. The thoughtful paper by Wright suggests that, although this approach may be appropriate in short-term parenteral nutrition, it may not be ideal for patients requiring enteral nutritional support, particularly when this is long term. He points out that these considerations are of considerable numeric importance, because a quarter or more of critically ill patients are likely to have carbohydrate intolerance, or frank diabetes, either pre-existing or secondary to the stress of illness. The prevalence of diabetes in the population is around 5%, with type 2 diabetes accounting for the majority of the cases. This situation is worsening steadily as the prevalence of obesity in the population rises. In the UK, for example, it has risen from 6 to 20% over the past 20 years. When these patients undergo surgery, they have in effect an acute insulin resistance superimposed on that of the chronic condition. Therefore there is reason to question whether current feeding formulas are optimal. Fashion has swung over the past 50 years from carbohydrate restriction in the 1950s and 1960s to regimens with 50-55% slowly absorbed carbohydrate. As both Wright and Coulston point out in this issue, however, diets containing as much as 50% of total energy as fat (30% as monounsaturated fatty acids) and only 30-40% as carbohydrate give an even better lipoprotein profile and glycemic control. It may be difficult to persuade diabetic persons to take such a diet orally in the normal course of events, but there seems to be good ground for adopting such a formulation for enteral feeding, particularly where this is likely to be prolonged. Coulston, in her review of enteral feeding, points out that most standard enteral formulas contain 25-40% total calories from fat, with n-6 fatty acids as the major source. These may have potentially adverse effects on immune function and enhance the production of inflammatory cytokines. For this reason she argues for diabetic diets not only high in monounsaturated, but also in n-3, fatty acids. Whether such an approach might be relevant to the nondiabetic person with stress-induced insulin resistance is an interesting question. As Coulston points out, this is an area that is ripe for clinical research. The subject of insulin sensitivity and insulin resistance continues to be an area of active interest in nutrition support. At a congress of the European Society of Parenteral and Enteral Nutrition (ESPEN) held in Stockholm on September 4-8, 1999, new research was presented regarding insulin at both the laboratory and clinical levels. Griesen et al. [4] presented evidence that acute pain decreases insulin sensitivity, apparently by affecting nonoxidative glucose metabolism. Myrenfors et al. [5] described a combination of microdialysis and isotopic glucose to determine insulin-stimulated glucose uptake in muscle and in adipose tissue of normal humans. Soop et al. [6] explored the role of muscle in postoperative insulin resistance. Using muscle biopsies from patients, they reported that the condition is not associated with skeletal muscle fiber atrophy or loss of force-generating capacity of single muscle fibers. Nutritional support is only one aspect of the total care of a surgical patient, and can not compensate for less than ideal management in other respects. In a symposium on perioperative care at the recent ESPEN meeting, authors from Denmark and from Sweden reviewed their experience with a package of care. This involved giving a carbohydrate-containing drink 2 h preoperatively, with consequent reductions in insulin resistance, negative nitrogen balance and duration of hospital stay. The use of nasogastric tubes was largely avoided in small intestinal and distal bowel surgery. This treatment allows oral intake to begin on the first postoperative day when aggressive mobilization is undertaken. Low thoracic epidural analgesia was provided for 4 days postoperatively, which provided pain-free mobilization and avoided the constipating effect of extensive opiate use. In control patients, quadriceps mass and strength diminished by 25%, whereas this package of care resulted in no reduction in either muscle mass or strength as well as earlier return of gastrointestinal function. The advantages of this approach invite widespread attention. The review by Cooperman et al. (pp. 17-21) in this issue devotes attention to the uncertain role of nutrition in treating malignant disease of the pancreas. It is common clinical wisdom that weight loss and malnutrition have an adverse effect on the outcome of various types of surgical operation. Despite the general acceptance of this idea, a significant number of surgical patients are operated upon with varying evidence of malnutrition. Unfortunately, the incidence of this problem is expected to continue or accelerate due to the increased severity of illness in hospitalized patients and the growing insurance restrictions on both admission and discharge from the hospital. The benefit of nutrition support to treat malnutrition seems intuitively correct, but the application of this therapy remains a subject of controversy. Nowhere is this more evident than in the field of cancer, particularly cancer of the pancreas. Cooperman et al. observe that the influence of nutrition support on outcome can only become evident when there is effective therapy for the primary disease. There are only a limited number of large studies on the influence of nutrition in patients operated upon for cancer of the pancreas. The authors offer three reasons for this situation. First, few institutions are able to accumulate a large number of such patients. Another possible reason pertains to the general misconception that delaying operation to correct malnutrition (or jaundice by placing a stent) will adversely influence outcome. Finally, operative morbidity and mortality for pancreatic cancer have decreased over the past decade without emphasis on perioperative nutrition, implying that nutritional status is of low priority. It is also argued that delay in operation for nutritional rehabilitation might increase patient exposure to antibiotic-resistant organisms in the hospital and increase costs, all for uncertain benefit. Two meta-analyses [7,8] of prospective randomized clinical trials of patients undergoing operation for upper gastrointestinal cancer compared patients with and those without preoperative nutrition support. Four of these studies included patients with pancreatic cancer, in whom no benefit could be shown from nutrition support, either total parenteral nutrition (TPN) or enteral nutrition, with regard to morbidity, mortality or duration of hospital stay. Separate analysis of the patients with pancreatic cancer also showed no benefit. Brennan et al. [9] found no benefit in giving postoperative TPN to patients with pancreatic cancer who were operated upon with an average of 6% weight loss, but major complications were significantly higher in the TPN group. The extensive evidence that perioperative TPN is of little benefit in such patients leaves the question as to why perioperative TPN may predispose to infectious complications. Brennan et al. [9] speculated that the fat in the TPN might cause immunosuppression. Tappy et al. [10] questioned whether carbohydrate might be detrimental, especially when given in amounts that forced lipogenesis, with the respiratory quotient over 1.0. TPN has often been chosen for postoperative nutrition out of worry over prolonged paralytic ileus. Cooperman et al. in this issue note that modern surgical procedures for pancreatic cancer can allow early enteral feeding, but the potential benefits have not been reflected in most studies of outcome. The reported benefits of providing supplementation with glutamine or arginine in critically ill patients have not been studied in cases of pancreatic cancer. Two studies of gastrointestinal cancer patients [11,12] reported that a postoperative enteral immune-enhancing diet, using a proprietary combination of arginine, RNA and n-3 fatty acids, was associated with a significant decrease in the incidence of infection, wound complications and duration of hospital stay. Another study using this commercial formulation in a similar population, however [13], found no benefit. Studies that utilize multiple agents before adequate study of each individual component not only fail to identify which component may be responsible for any benefit shown, but also may obscure evidence of possible benefit when one of the agents is given alone. Cooperman et al. conclude that the routine and indiscriminate use of conventional nutrition support in the hospital for the minimally to moderately malnourished patient cannot be supported. They add, however, that whether nutrition support offers benefit to the severely malnourished patient with pancreatic cancer is unknown, and that the data on immune-enhancing diets are ‘conflicted and inconclusive’. The excellent review by MacFie (pp 23-29), of ethical and legal considerations in providing nutritional support to the perioperative patient, embraces a number of aspects which only recently have been receiving serious attention from the medical community. Underlying many of his considerations remains the importance of the physician carefully appraising the results of treatment and practicing evidence-based medicine. Data from meta-analysis has often been used to assemble sufficiently large numbers of cases for statistical analysis without proper identification of all available data in each clinical trial. The resulting unreliability and bias is receiving public attention in an editorial in the Journal of the American Medical Association [14] followed by front page stories in the lay press. Both the enthusiasts and the critics of nutritional support must unite in the continuing search for when and how to treat the perioperative patient with nutritional support.