An increased incidence of carcinoma of the small bowel and colon has been described in patients with Crohn's disease. Tumours arising in the rectum and anus are reported less often. Between 1940 and 1992, of some 2500 patients with Crohn's disease seen at this hospital, 15 are known to have developed carcinoma of the lower gastrointestinal tract. Malignancy occurred in the colon in two patients, in the upper two thirds of rectum in one, in the lower third of rectum in seven, and in the anus in five. The 12 patients with carcinoma arising in the anus or lower rectum had longstanding severe anorectal Crohn's disease, which included a stricture in four, fistula in four, proctitis in one, abscess in two, and enlarged anal skin tags in one. The development of malignancy in patients with Crohn's disease may apply particularly to those with chronic complicated anorectal disease.

Persistent, refractory diarrhoea continues to be an important clinical problem. The mechanisms involved are associated with reduced intestinal absorption and increased intestinal secretion. Reduced intestinal absorption can result from small intestinal resection or from disorders in which there is damage to the small intestine. Motility disorders may also impair absorptive function. The rationale for using octreotide in refractory diarrhoea, intestinal motility disorders, and fistulae relates to its ability to promote intestinal absorption and inhibit gastric, pancreatic, and intestinal secretion. Several clinical studies in patients with short bowel syndrome have reported a reduction of intestinal output in patients taking octreotide compared with controls. Additionally, a number of studies have shown that octreotide improves secretory diarrhoea resulting from neuroendocrine tumours, intestinal infections in AIDS patients, and intestinal graft v host disease. Octreotide may be of use in patients suffering from intestinal motility disorders such as those associated with systemic sclerosis. Octreotide may also be of value in promoting closure of gastrointestinal and pancreatic fistulae.

Bleeding from oesophageal varices has a high death rate. Injection sclerotherapy is the most appropriate treatment but facilities for this are not always available. Balloon tamponade and vasoactive therapy may be used as stop gap measures. Somatostatin and octreotide are therapeutic candidates for the treatment of variceal bleeding and there are several trials that have compared somatostatin and octreotide with other treatments for this condition. The results of these trials are summarised and discussed. A meta analysis of the group of trials of placebo or H2 antagonists v somatostatin or octreotide showed a significant advantage of somatostatin or octreotide in terms of efficacy, but no difference in mortality. The trials discussed seem to show that somatostatin and octreotide are at least as effective as other treatments, with the benefit of fewer adverse effects, and thus represent the best vasoactive agents. Additionally, they may have a role as adjuvant treatment to emergency sclerotherapy for active bleeders and this must be further investigated.

The effects of octreotide on six normal subjects and five patients with scleroderma were investigated. Changes in intestinal motility and in plasma motilin were examined after a single injection of octreotide. Octreotide stimulated intense intestinal motor activity in normal subjects. Motility patterns in the scleroderma patients were chaotic and non-propagative, but, after octreotide was given, became well coordinated, aborally directed, and nearly as intense as in normal volunteers. Clinical responses and changes in breath hydrogen were also evaluated in the five scleroderma patients who had further treatment with octreotide at a dose of 50 micrograms/day subcutaneously for three weeks. A reduction in symptoms of abdominal pain, nausea, vomiting, and bloating was seen. Additionally, there was an improvement in bacterial overgrowth as objectively measured by breath hydrogen testing. The effects of octreotide (100 micrograms/day subcutaneously) on the perception of rectal distension were investigated in a double blind, placebo controlled study in healthy volunteers. Octreotide was shown to reduce the perception of rectal distension without affecting motor pathways or local rectal reflexes. This enhanced tolerance to volume distension seems to result from inhibition of sensory afferent pathways as shown by electroencephalographic studies showing diminished evoked spinal and cortical potentials after octreotide. In irritable bowel syndrome patients with rectal urgency, octreotide reduces rectal pressures and perception after rectal distension to near normal values.

Intestinal mucosal injury that results from local ischaemia can be detected by early increases in gut permeability, followed by later morphological, histological, and biochemical abnormalities. Local adaptive mechanisms (for example increased oxygen extraction) can cope with reductions in blood flow of up to 50%, as may occur during episodes of septic shock or cardiac tamponade. Why then does hypoxic injury develop? The peculiar vascular anatomy of the villi allows for oxygen short circuiting to occur at their base, when blood flow is low. Although overall oxygen extraction efficiency may be high, regional hypoxia at the villus tip may, paradoxically occur. The severity of reperfusion injury depends on the duration of preceding hypoxia. Free radical generation through the hypoxanthine xanthine oxidase system is important in mediating cellular damage. In addition, luminal aggressive factors (for example, pancreatic proteases) may cause mucosal damage, as suggested by earlier studies. More recent studies in pigs suggests that pancreatic duct ligation merely delays, but does not prevent development of gut reperfusion injury. Enteral nutrition should benefit patients with the ischaemic intestine because in comparison with total parental nutrition, it stimulates regional blood flow, and attenuates mucosal injury. There are no randomised trials to verify this, but use of tonometry to monitor local ischaemia may help resolve the issue.

The spectacular success of parenteral nutrition in supporting patients during small intestinal adaptation after massive resection, tends to obscure the prolonged periods often needed for such adaptation to take place. After neonatal small intestinal resection for example, it may take more than five years before adaptation is complete. There is therefore a strong argument for examining ways in which adaptation can be facilitated, in particular, by the addition of novel substrates to enteral feeds. Pectin is completely fermented by colonic bacteria to short chain fatty acids. In the rat, addition of pectin to enteral feeds led to a more rapid adaptive response in both the small and large intestine after massive small intestinal resection, although faecal nitrogen losses were increased. In a similar rat model, the provision of 40% of non-protein energy as short chain triglycerides facilitated the adaptive response in the jejunum, colon, and pancreas. The importance of glutamine as a metabolic substrate for the small intestine makes it another potential candidate and some, but not all animal studies, have suggested a therapeutic effect: increasing the glutamine content of feeds to 25% of total amino acids produced enhanced jejunal and ileal hyperplasia, even on a hypocaloric feed, and an improved overall weight gain. Studies in humans are very limited, but such promising results in the experimental animal suggest that this is probably a fruitful area for further study.

At present, there is limited evidence for the role of enteral nutrition as a primary therapy in cancer patients. Cachexia commonly occurs in patients with advanced cancer. A consensus view from a large number of studies suggests that cachexia cannot be fully reversed by vigorous enteral nutritional support. A review is included of the available data on the effects of enteral nutritional support on the common indices of nutritional state and on the final outcome of patients receiving enteral nutrition in conjunction with radiotherapy or chemotherapy, or both. The 'nutritional' effects are probably limited because the duration of the nutritional support in most studies consists of a few weeks while malnutrition in the cancer patients often occurs over many months.

The feeding of a protein hydrolysate based 'elemental' diet supplemented with added glutamine did not provide superior protection to the small intestine of dogs subjected to therapeutic pelvic irradiation. Comparison of diets with and without the added glutamine showed significant protection of the intestine from radiation injury. Both histological examination and electron microscopy showed lack of tissue injury with both diets. The activity of the free radical generating enzymes, scavengers, and antioxidants were similar in the intestinal mucosa of dogs fed either diet. After radiation, however, the activity of xanthine oxidase, superoxide dismutase, and glutathione peroxidase were significantly (p < 0.002) higher in the intestine of dogs fed elemental diet without the added glutamine. If the activities of these enzymes are important in the protection of the intestine from radiation injury, then the addition of extra glutamine may provide no benefit.

The developments in enteral feeding for Crohn's disease in the past decade are critically reviewed. The advent of amino acid based chemically defined elemental diets signalled the end of 'total bowel rest' in the management of these patients. Subsequently, controlled clinical trials showed that elemental diets were as effective as corticosteroids in inducing clinical remission in patients with acute exacerbations of Crohn's disease. The later use of peptide based elemental diets, in Crohn's disease produced somewhat conflicting results. The initial uncontrolled studies suggest that polymeric whole protein diets might also be effective in the management of acute exacerbations of the disease, casting in turn doubts concerning the role of dietary antigens in the pathogenesis of Crohn's disease. Results of controlled studies comparing the use of elemental and polymeric diets as primary therapy in Crohn's disease have, however, also produced conflicting results. The results of one recent controlled trial in which the use of polymeric diet was compared with that of corticosteroids does, however, suggest that these diets may have a primary therapeutic effect in Crohn's disease. An analysis of the composition of some of the enteral diets used in different trials suggest that the effectiveness of enteral diets in treating active Crohn's disease might relate more to their fat than nitrogen composition. A hypothesis is proposed that the effectiveness of enteral nutrition in the primary therapy of acute exacerbations of Crohn's disease occurs because the successful diets used contain insufficient precursors for arachidonate derived eicosanoid synthesis.

The normal gut is adapted to intermittent feeding with complex macromolecular substrates of low sodium content. The high permeability of the upper small intestine to sodium, together with sodium rich saliva and pancreaticobiliary secretions results in large sodium fluxes into the lumen. These substantial sodium influxes are matched by equally large effluxes from the ileum and proximal colon, which are comparatively impermeable to sodium and capable of active sodium absorption. Resection of these distal, sodium absorbing regions of the intestine, lead to problems with sodium depletion. Controlled transit of chyme is essential to permit time for optimum digestion and absorption and a range of feedback control mechanisms exist. Partially digested nutrients, both in the duodenum and ileum, exert inhibitory feedback to delay delivery of further nutrients and here again surgery may compromise these reflexes. Brush border hydrolase values are strongly influenced by luminal nutrient concentrations, being impaired by malnutrition and total parenteral nutrition, but restored by enteral feeding. Viscous fibre slows absorption and may delay transit through mechanisms that are as yet uncertain. Whether and how novel substrates activate normal control mechanisms will be important factors determining their effectiveness and patient acceptability.

The informational aspects of nucleic acid synthesis have attracted much more attention than the quantitative significance of DNA, rRNA, tRNA, and nucleotide synthesis. Animal and human studies suggest that in energetic terms, 5-10% of the energy used in synthesising tissue protein is expended in manufacturing an appropriate amount of synthetic machinery, that is the ribosome and tRNA. The two sources for synthesis of nucleotides are salvage of nucleotides released by intracellular degradation or derived from the diet, and nucleotides synthesised de novo from amino acids (for example, glutamine) and sugars (glucose). The comparative importance of these two processes is not well defined, but rRNA production requires a high de novo input in cell types with the capacity for rapid division (for example, lymphocytes). The gut is unusual in requiring a ready arterial supply of nucleotides synthesised by hepatic de novo pathways. Animal studies show that an exogenous supply of nucleotides (salvage) can improve liver regrowth, immune responsiveness to a microbial challenge, and gut morphology in diarrhoea models. Humans adapt to dietary nucleotide intake by downregulating de novo pathways. All total parental nutrition regimens, and most enteral regimens lack nucleotides, which may predispose to an inadequate supply of preformed nucleotides to gut and immune cells in the critically ill, artificially fed patient. Unfortunately, there are no clinical studies that answer this point at present.