The stimulation of the colon after eating is within most individuals' experience and an exaggerated response is frequently a problem in patients with functional diarrhoea or irritable bowel syndrome. Early studies showed that fatty meals stimulated colonic pressure waves more than carbohydrate meals, whereas amino acid meals tended, if anything, to inhibit motility. Although these initial studies recorded only from the rectum and distal sigmoid colon we now know that the whole colon responds, though the response is less pronounced and of shorter duration in the proximal compared with the distal colon. As well as stimulating phasic activity, eating also increases colonic tone both proximally and distally.1

Modern, solid state manometric systems, which use strain gauges mounted on tubes as in the study by Rao et al (see page 205), have a much smaller diameter and are more comfortable for the patients than the older water perfused manometry catheters and permit prolonged ambulatory studies. Twenty hour recordings have shown a noticeable circadian rhythm, with pronounced activation on wakening and after eating. These recordings also make it possible to get an accurate estimate of the frequency of infrequent forceful propulsive waves, known as high amplitude propagated contractions (HAPCs). These are characterised by high amplitude (>100 mm Hg) and prolonged duration (>12 seconds), with a relatively rapid rate of propagation (0.8–1.3 cm/s).2 They are more likely to occur soon after wakening (33%) or after meals (50%) and some are followed by a desire to defecate.

Most colonic pressure waves are in fact non-propagated, isolated contractions which are responsible for mixing and facilitating absorption, shifting luminal contents only short distances to and fro.

Rao et al's study has many advances compared with earlier studies, including the use of solid state technology which allowed prolonged studies to be performed without the infusion of water which, with multichannel recording, has often been substantial (60 ml/hour or 1.4 litres/24 hours). After a tap water enema the probes were placed colonoscopically, with the tip positioned 60 cm from the anus. The colon was then allowed 10 hours to re-establish its normal function before the first meal was consumed and 27 hours before the second meal. The position of the probe in the descending colon was confirmed fluoroscopically at the end of each meal. The two meals tested contained 4.18 MJ (1060 kcal) of which 60% was provided as either fat or carbohydrate. Unlike some previous meals they were equally palatable and appetising. The analysis was done by investigators blind to which meal had been received, an important point as a substantial amount of the analysis was performed manually and could otherwise be subject to investigator bias. The authors provide clear criteria in the text to identify the various pressure wave patterns. These were non-propagated pressure waves, by far the commonest type, propagated pressure waves to be distinguished from specialised propagated pressure waves which the authors defined objectively as pressure waves with an amplitude and a duration above the 95% confidence interval for non-specialised pressure waves (>103 mm Hg and >13 seconds respectively). They also defined simultaneous pressure waves and much less frequent retrograde pressure waves.

The key findings were that the fatty meal increased the number of pressure waves for three hours, whereas the carbohydrate response was shorter lived, returning to baseline by two hours. The fatty meal also increased propagated pressure waves one to three hours post- prandially, whereas simultaneous pressure waves were increased in the first hour only. Lesser responses were seen with the carbohydrate meal, which increased the propagated pressure waves but did not increase the simultaneous pressure waves. Importantly, retrograde waves also increased after the fatty meal but not after the carbohydrate meal. Eleven of the 18 subjects showed the specialised propagated pressure waves in the first hour after the fatty meal, whereas only nine showed these after the carbohydrate meal.

Many authors have observed this enhanced effect of fat but so far the cause has eluded explanation. Fatty meals tend to empty more slowly, and are associated with a larger increase in visceral blood flow and greater release of cholecystokinin (CCK), neurotensin and peptide YY. However, recent studies with CCK-A receptor antagonists appear to indicate that CCK is not an important component.3 The colonic response to feeding has both a cerebral and gastric component, being elicited both by sham feeding and gastric distension with a balloon.4 These immediate effects, however, appear to be transient and the more prolonged effects probably originate from the stimulation of enteroendocrine cells and possibly nerves further down the gut.

The overall effect of this stimulation in normal subjects is usually mixing with very little net movement, though this does vary with region.5 More detailed examination reveals to and fro mixing movements within the ascending6 and transverse colon, especially after a fatty meal. Much less movement is seen in the descending colon where the contents are normally more solid. These patterns change in disease states, thus in diarrhoeal states with more fluid colonic contents, isotope injected into the splenic flexure shows exaggerated backward and forward movements7 whereas in constipation movement is much reduced.8 Recent studies have suggested that although there is a decrease in mixing activity and tone, the number of propagated contractions increases in both functional diarrhoea9 and ulcerative colitis.10 The functional outcome evidently depends on the balance between propulsive propagated contractions, mixing non-propagated contractions, and retropulsive retrograde pressure waves.

We are only just beginning to understand the pharmacology of HAPCs but they cannot be induced simply by cholinergic stimulation. HAPCs seen to involve a local reflex which starts with stimulation of mucosal afferents. Luminally active contact laxatives such as bisacodyl and senna induce HAPCs, as do irritant substances such as long chain fatty acids and glycerol, an action which can be blocked by lignocaine and atropine. Distension of the colon alone does not induce this motor pattern.2 5-Hydroxytryptamine (5-HT) may play a key role. 5-HT₄ receptor agonists such as prucalapride can induce HAPCs11 whereas 5-HT₄ antagonists impair the gastrocolonic response. Understanding how to control the colonic response to food may well be valuable therapeutically in treating both diarrhoea and constipation.