Altered Colorectal Compliance and Anorectal Physiology in Upper and Lower Motor Neurone Spinal Injury May Explain Bowel Symptom Pattern.

OBJECTIVES: Supraconal spinal cord injury (SCI) and lower motor neurone spinal cord injury (LMN-SCI) cause bowel dysfunction; colorectal compliance may further define its pathophysiology. The aim of this study was to investigate rectal (RC) and sigmoid (SC) compliance and anorectal physiology parameters, in these subjects. METHODS: Twenty-four SCI subjects with gut symptoms (14 RC, 10 SC) and 13 LMN-SCI subjects (9 RC, 4 SC) were compared with 20 spinal intact controls (10 RC, 10 SC). Staircase distensions were performed using a barostat. Anorectal manometry, including rectoanal inhibitory reflex (RAIR) measurement, was performed in all. Data presented as mean±standard error (SCI/LMN-SCI vs. controls). RESULTS: SCI subjects had a higher RC (17.0±1.9 vs. 10.7±0.5 ml/mm Hg, P<0.05) and SC (8.5±0.6 vs. 5.2±0.5 ml/mm Hg, P=0.002). LMN-SCI subjects had a lower RC (7.3±0.7 ml/mm Hg, P=0.0021) while SC was unchanged (8.3±2.2 ml/mm Hg, P>0.05). Anal resting pressure was decreased in SCI (55±5 vs. 79±7 cmH2O, P=0.0102). Anal squeeze pressure was decreased in LMN-SCI (76±13 vs. 154±21 cmH2O, P=0.0158). In SCI and LMN-SCI, the amplitude reduction of the RAIR was greater (62±4% and 70±6% vs. 44±3%, P=0.0007). CONCLUSIONS: Colorectal compliance abnormalities may explain gut symptoms: increased RC and SC contributing to constipation in SCI, reduced rectal compliance contributing to fecal incontinence (FI) in LMN-SCI. Reduced resting anal pressure in SCI and reduced anal squeeze pressure in LMN-SCI along with a greater RAIR amplitude reduction may be factors in FI. These co-existing abnormalities may explain symptom overlap, and represent future therapeutic targets to ameliorate neurogenic bowel dysfunction.

Delayed gastric emptying in Parkinson's disease.

Gastrointestinal symptoms are evident in all stages of Parkinson's disease (PD). Most of the gastrointestinal abnormalities associated with PD are attributable to impaired motility. At the level of the stomach, this results in delayed gastric emptying. The etiology of delayed gastric emptying in PD is probably multifactorial but is at least partly related to Lewy pathology in the enteric nervous system and discrete brainstem nuclei. Delayed gastric emptying occurs in both early and advanced PD but is underdetected in routine clinical practice. Recognition of delayed gastric emptying is important because it can cause an array of upper gastrointestinal symptoms, but additionally it has important implications for the absorption and action of levodopa. Delayed gastric emptying contributes significantly to response fluctuations seen in people on long-term l-dopa therapy. Neurohormonal aspects of the brain-gut axis are pertinent to discussions regarding the pathophysiology of delayed gastric emptying in PD and are also hypothesized to contribute to the pathogenesis of PD itself. Ghrelin is a gastric-derived hormone with potential as a therapeutic agent for delayed gastric emptying and also as a novel neuroprotective agent in PD. Recent findings relating to ghrelin in the context of PD and gastric emptying are considered. This article highlights the pathological abnormalities that may account for delayed gastric emptying in PD. It also considers the wider relevance of abnormal gastric pathology to our current understanding of the etiology of PD.

Consensus best practice pathway of the UK scleroderma study group: gastrointestinal manifestations of systemic sclerosis.

Systemic sclerosis is an autoimmune connective tissue disorder, which can be progressive with multisystem involvement. Guidance on the management of complications is based on a limited data set and practice amongst clinicians can vary. The UK Scleroderma study group set up several working groups to agree some consensus pathways for the management of specific complications. Approximately nine out of ten patients with systemic sclerosis will have involvement of the gastrointestinal system and in this review article we explore the management of these complications in a symptom-based approach. The algorithms are a useful tool for clinicians, which we hope, will be a point of reference and highlight the need for further research in these areas.

Fecal incontinence in systemic sclerosis is secondary to neuropathy.

OBJECTIVES: Systemic sclerosis (SSc) is a chronic multi-system autoimmune disorder with gastrointestinal tract (GIT) involvement in up to 90% of patients and anorectal involvement occurs in up to 50% of patients. The pathogenesis of gastrointestinal abnormalities may be both myogenic and neurogenic. We aimed to identify which anorectal physiological abnormalities correlate with clinical symptoms and thus understand the pathophysiology of anorectal involvement in SSc. METHODS: In total, 44 SSc patients (24 symptomatic (Sx) (fecal incontinence) and 20 asymptomatic (ASx)) and 20 incontinent controls (ICs) were studied. Patients underwent anorectal manometry, rectal mucosal blood flow (RMBF), rectal compliance (barostat), and rectoanal inhibitory reflex assessment (RAIR). RESULTS: Anal squeeze pressure was lower in the IC group compared with both the ASx and Sx groups (IC: 46.95 (30-63.9)) vs. ASx: 104.6 (81-128.3) vs. (Sx: 121.4 (101.3-141.6); P < 0.05). Resting pressure was lower in the IC group. RMBF and rectal compliance did not differ between groups. Anal, but not rectal, sensory threshold, was significantly attenuated in Sx patients (Sx: 10.4 (8.8-11.4) vs. ASx: 6.7 (5.7-7.7) vs. IC: 8.5 (6.5-10.4); P < 0.05). There was a positive correlation between anal sensory thresholds and incontinence score in SSc patients (r = 0.54; P < 0.05). RAIR was absent in 11/24 Sx patients but only in 2/20 ASx and in 1/20 IC patients. CONCLUSIONS: Fecal incontinence in SSc is related to neuropathy as suggested by absent RAIR and higher anal sensory threshold and is related less so to sphincter atrophy and rectal fibrosis.

Internal anal sphincter atrophy in patients with systemic sclerosis.

OBJECTIVES: SSc is a connective tissue, multisystem disorder of unknown aetiology. The gastrointestinal tract (GIT) is affected in up to 90% of patients. The exact pathophysiology of GIT involvement is not known, but it is related to both neurogenic and myogenic abnormalities as well as possible vascular and ischaemic changes. Thinning of the internal anal sphincter (IAS) has been demonstrated in SSc with faecal incontinence. We aimed to investigate anal sphincter structure in patients with SSc. METHODS: Forty-four SSc patients [24 symptomatic (Sx) and 20 asymptomatic (ASx)] and 20 incontinent controls (ICs) were studied. Patients underwent anorectal manometry and endoanal US. RESULTS: In the ICs, external anal sphincter defects were more common, but the IAS was less atrophic, evident by both atrophy scores and IAS thickness. There was no significant difference in atrophy scores [Sx: 2 (1.5-3) vs ASx: 2 (1-2)] or IAS thickness [Sx: 1.85 (1.5-2.3) vs ASx: 1.8 (1.7-2.25)] between the Sx and ASx SSc patients. CONCLUSION: Patients with SSc (both Sx and ASx) have thin and atrophic IAS, suggesting that IAS atrophy develops even in ASx patients and this may be amenable to treatment with sacral neuromodulation.

Anal canal vector volume manometry.

BACKGROUND: Anal manometry is routinely used in the assessment of the anal sphincters in patients with fecal incontinence or suspected sphincter injury. Such physiological information is complementary to the anatomical assessment provided by anal endosonography. The evolution of 3-dimensional anal endosonography provides more diagnostically useful information in complex cases. Vector volume manometry has been developed to give a 3-dimensional view of the anal sphincters. OBJECTIVE: We reviewed the published literature on this technique, with the intention of deriving a system of standardization based on the published literature and to summarize the derivation and physiological meaning of the parameters measurable by vector volume studies, as well. DATA SOURCES: We undertook a MEDLINE search using the terms "vector volume" or "vector manometry" and "anal canal." We also reviewed further publications found from references cited in the original articles identified from the above search. STUDY SELECTION: Only English language articles of studies performed on humans were reviewed. INTERVENTION: Anal canal vector volume manometry was the intervention. RESULTS: With the development of automated puller systems and associated software, parameters such as total vector volume, maximum pressure, mean pressure, anal canal symmetry, anal canal length, and the length of the high-pressure zone can be readily calculated. LIMITATIONS: There are conflicting studies related to the clinical value of both anal manometry and vector volume manometry, in part, because of the lack of standardization of equipment and technique. CONCLUSIONS: The vector volume parameters have been shown to correlate with both imaging results and incontinence scores with automated puller systems. The clinical utility of vector volume manometry would be improved further by the standardization of equipment and technique. The main clinical utility may lie in the treatment selection and preoperative assessment of patients awaiting surgery for anal pathology that has yet to be evaluated.

Temporary sacral nerve stimulation alters rectal sensory function: a physiological study.

BACKGROUND: The indications for sacral nerve stimulation are increasing, but the mechanism remains poorly understood. OBJECTIVE: This study aimed to examine the effect of sacral nerve stimulation on rectal compliance and rectal sensory function. DESIGN: This was a prospective study. SETTINGS: This study took place at a university teaching hospital. PATIENTS: Twenty-three consecutive consenting patients (22 female; median age, 49 y) undergoing temporary sacral nerve stimulation for fecal incontinence were prospectively studied. Clinical response was assessed by the use of bowel diaries and Wexner scores. MAIN OUTCOME MEASURES: Anal manometry, rectal compliance, volume and pressure thresholds to rectal distension (barostat), and rectal Doppler mucosal blood flow were measured before and at the end of stimulation. RESULTS: Sixteen patients (70%) had a favorable clinical response. Median anal squeeze pressures increased with stimulation from 40 (range, 6-156) cmH2O to 64 (range, 16-243) cmH2O. Median rectal compliance did not significantly change with stimulation (prestimulation: 11.5 (range, 7.9-21.8) mL/mmHg, poststimulation: 12.4 (range, 6.2-22) mL/mmHg, P = .941). Rectal wall pressures associated with urge (baseline: 15.4 (range, 11-26.7) mmHg, poststimulation: 19 (range, 11.1-42.7) mmHg, P = .054) and maximal tolerated thresholds (baseline: 21.6 (8.5-31.9) mmHg, poststimulation: 27.1 (14.3-43.3) mmHg, P = .023) significantly increased after stimulation. Rectal Doppler mucosal blood flow did not significantly change with stimulation (baseline: 125.8 (69.9-346.8), poststimulation: 112.4 (50.2-404.1), P = .735). Changes in anal resting pressure and rectal wall pressures with stimulation were evident only in responders; however, changes in anal squeeze pressures were evident in both responders and nonresponders. LIMITATIONS: The study reports results following short-term stimulation in a small but homogenous group of patients. A larger long-term study will follow. CONCLUSION: Temporary sacral nerve stimulation does not change rectal compliance, but is associated with significant changes to the pressure thresholds of rectal distension. This, together with the observation that outcome is not related to sphincter integrity, supports the hypothesis of an afferent-mediated mechanism of action.

Assessment of gastrointestinal symptoms in patients with systemic sclerosis in a UK tertiary referral centre.

OBJECTIVES: The gastrointestinal tract (GIT) is affected in up to 90% of patients with SSc to a variable extent and severity. We aimed to establish the frequency and range of gastrointestinal (GI) symptoms in SSc patients at the Royal Free Hospital, a tertiary referral centre. METHODS: A 52-item, previously validated, questionnaire capturing SSc-related gut dysfunction was given to consecutive patients with SSc attending the rheumatology outpatient department. The questionnaire assesses the 'frequency' of five categories of symptoms and their 'impact' on social functioning and emotional well-being. Patients' notes were reviewed to establish disease subtype, autoantibody profile and other internal organ involvement. RESULTS: We collected 402 completed questionnaires (357 females; mean age 55). Sixty-nine per cent of patients had lcSSc and 30% dcSSc with mean disease duration of 11 years. Mean questionnaire scores showed that patients have a wide range of GI symptoms. Ninety-four per cent of patients reported upper and 79% lower GI symptoms, 3% of patients reported no symptoms and 10% reported daily symptoms. There was no association between disease subtype or autoantibody profile and GI symptoms. There was a positive correlation between diarrhoea scores (high scores = best health) and pulmonary fibrosis (r = 0.134, P = 0.0068). No other association between GI symptoms and other internal organ involvement was found. CONCLUSIONS: GI symptoms, both upper and lower, are common in patients with SSc. Patients should be asked specifically about GI symptoms as they may be under-reported and therefore under-treated. GI focused questionnaire is an effective way to assess gut symptoms and adjust treatment.

Twitter debate: controversies in functional gastrointestinal disorders.

The new 'Controversies In…' series for the Frontline Gastroenterology Twitter debates addressed the difficult area of functional gastrointestinal disorders, facilitated by the former editor-in-chief Anton Emmanuel. Key topics discussed included distinguishing functional dyspepsia from genuine gastroparesis, when we should investigate for bile acid malabsorption, the current treatments for constipation-predominant irritable bowel syndrome and, importantly, how to manage consultations with complex patients presenting with functional bowel disease. The debate generated over a million impressions on twitter and this article aims to summarise the key educational points from the event.

High-resolution anal manometry: Repeatability, validation, and comparison with conventional manometry.

BACKGROUND: Accurate measurement of anal sphincter function is potentially of value in defining treatment of common pelvic floor disorders. The aim of this study was to establish repeatability and validate high-resolution anorectal manometry (HRAM) by comparison to conventional manometry (CM). Arising from this work would be definitive normal range data. METHODS: Eighty healthy volunteers (40 female) underwent a test-retest repeatability study. A 16-channel water-perfused HRAM catheter was compared to an 8-channel conventional catheter using a station pull-through technique. KEY RESULTS: High-resolution anorectal manometry had similar precision to conventional manometry when measuring resting pressure (intraclass correlation coefficient [ICC] 0.73 vs 0.68, HRAM vs CM) and squeeze increment (ICC 0.90 vs 0.94, HRAM vs CM). HRAM measured resting pressures 10% lower than CM and squeeze pressure 27% higher than CM. CONCLUSIONS AND INFERENCES: High-resolution anorectal manometry is a valid technique with comparable precision to CM. HRAM measurements differ considerably to CM, and a new set of normal values must be used.

Joint Hypermobility Syndrome Affects Response to a Low Fermentable Oligosaccharide, Disaccharide, Monosaccharide and Polyol Diet in Irritable Bowel Syndrome Patients: A Retrospective Study.

BACKGROUND: The low fermentable oligosaccharide, disaccharide, monosaccharide and polyol (FODMAP) diet causes significant clinical improvement in patients with irritable bowel syndrome (IBS). Joint hypermobility syndrome (JHS), defined as musculoskeletal symptoms in a hypermobile individual in the absence of systemic rheumatological disease, may be associated with functional gastrointestinal symptoms, including IBS. The aim of this study is to examine whether JHS can affect the response to the low FODMAP diet in patients with IBS. METHODS: In this retrospective study, we included patients with IBS according to Rome III criteria who had followed a low FODMAP diet. Symptoms scores were measured before and after the low FODMAP diet. RESULTS: A total of 165 patients (130 females, age 44 ± 14 years) were included. Diarrhea predominant IBS (IBS-D) was present in 40.6% of our patients while JHS was present in 21.2%. The score for abdominal pain was higher for JHS compared to non-JHS prior to intervention (P = 0.011). Symptoms improved in both groups of patients after a low FODMAP diet (P < 0.0001). The largest effects were shown with significant decreases of the average score and bloating. When broken down by JHS and IBS type, a low FODMAP diet significantly improved pain, bloating, diarrhea, constipation, and the average score with the largest effect in JHS/constipation predominant IBS (IBS-C), JHS/mixed IBS and unclassified IBS (IBS-M), JHS/IBS-D, non-JHS/IBS-C and JHS/IBS-M, respectively. CONCLUSIONS: Our study suggests that a low FODMAP diet has a greater effect on IBS symptoms in JHS than non-JHS patients.

The patient burden of opioid-induced constipation: New insights from a large, multinational survey in five European countries.

BACKGROUND: Despite its high prevalence, opioid-induced constipation (OIC) remains under-recognised and undertreated, and its true impact on wellbeing and quality of life (QoL) may be underestimated. METHODS: A quantitative, questionnaire-based international survey was conducted. RESULTS: Weak-opioid users appeared as bothered by constipation as strong-opioid users (38% vs 40%, respectively; p = 0.40), despite it causing less-severe physical symptoms and impact on QoL. Strong-opioid users meeting Rome IV OIC criteria appeared to experience greater symptomatic and biopsychosocial burden from constipation than those not satisfying these criteria. Almost one-fifth of respondents were dissatisfied with their current constipation treatment and around one-third found balancing the need for adequate pain relief with constipation side effects challenging. Consequently, more than half failed to adhere to their prescribed treatment regimens, or resorted to suboptimal strategies, e.g. 40% reduced their opioid intake, to relieve constipation. Almost 60% of healthcare professionals did not adequately counsel patients about constipation as a common side effect of opioid use. CONCLUSIONS: Findings suggest that both weak- and strong-opioid users suffer comparable bother and decreased QoL, Rome IV criteria can identify patients with more-severe OIC, but may underdiagnose patients showing fewer symptoms, and increased education is needed to manage patients' expectations and enable improved OIC self-management.

Efficacy and mechanism of sub-sensory sacral (optimised) neuromodulation in adults with faecal incontinence: study protocol for a randomised controlled trial.

BACKGROUND: Faecal incontinence (FI) is a substantial health problem with a prevalence of approximately 8% in community-dwelling populations. Sacral neuromodulation (SNM) is considered the first-line surgical treatment option in adults with FI in whom conservative therapies have failed. The clinical efficacy of SNM has never been rigorously determined in a trial setting and the underlying mechanism of action remains unclear. METHODS/DESIGN: The design encompasses a multicentre, randomised, double-blind crossover trial and cohort follow-up study. Ninety participants will be randomised to one of two groups (SNM/SHAM or SHAM/SNM) in an allocation ratio of 1:1. The main inclusion criteria will be adults aged 18-75 years meeting Rome III and ICI definitions of FI, who have failed non-surgical treatments to the UK standard, who have a minimum of eight FI episodes in a 4-week screening period, and who are clinically suitable for SNM. The primary objective is to estimate the clinical efficacy of sub-sensory SNM vs. SHAM at 32 weeks based on the primary outcome of frequency of FI episodes using a 4-week paper diary, using mixed Poisson regression analysis on the intention-to-treat principle. The study is powered (0.9) to detect a 30% reduction in frequency of FI episodes between sub-sensory SNM and SHAM stimulation over a 32-week crossover period. Secondary objectives include: measurement of established and new clinical outcomes after 1 year of therapy using new (2017 published) optimised therapy (with standardised SNM-lead placement); validation of new electronic outcome measures (events) and a device to record them, and identification of potential biological effects of SNM on underlying anorectal afferent neuronal pathophysiology (hypothesis: SNM leads to increased frequency of perceived transient anal sphincter relaxations; improved conscious sensation of defaecatory urge and cortical/subcortical changes in afferent responses to anorectal electrical stimulation (main techniques: high-resolution anorectal manometry and magnetoencephalography). DISCUSSION: This trial will determine clinical effect size for sub-sensory chronic electrical stimulation of the sacral innervation. It will provide experimental evidence of modifiable afferent neurophysiology that may aid future patient selection as well as a basic understanding of the pathophysiology of FI. TRIAL REGISTRATION: International Standard Randomised Controlled Trial Number: ISRCTN98760715 . Registered on 15 September 2017.

Gastrointestinal symptoms related to autonomic dysfunction following spinal cord injury.

The impact of spinal cord injury on an individual's gastrointestinal tract function is often poorly understood by the general public and also by those involved with persons with spinal cord injury. This chapter reviews the anatomy, physiology and function of the gastrointestinal tract, with particular emphasis on neurological control mechanisms. In turn, it relates the effect that spinal cord injury has on the neurological control of the gastrointestinal tract. The symptoms that are encountered by patients in the acute phase following injury, and by individuals in the months/years after injury, with particular reference to the effect of altered autonomic nervous system control of the gastrointestinal tract, are discussed. Together with a following summary of current bowel management regimens and techniques, this chapter aims to provide an overall view of the effect that autonomic dysfunction due to spinal cord injury has on gastrointestinal function.

Patterns of constipation in urogynecology: clinical importance and pathophysiologic insights.

OBJECTIVE: We have analyzed the prevalence and patterns of constipation in women with urinary symptoms and/or genital prolapse. STUDY DESIGN: Seven hundred and eighty-six consecutive urogynecologic patients underwent a questionnaire and structured clinical assessment. Comparison between constipated and nonconstipated women was made. Fisher exact test, Wilcoxon rank sum test, and logistic regression were used for statistical analysis (P < .05 for significance). RESULTS: Thirty-two percent of women were constipated (172 difficult stool passage, 13 reduced stool frequency, 64 both). A genital prolapse > or = 2 degree Half Way System (HWS) was present in 44% of women. A posterior colpocele was more frequent in constipated women (35% vs 19%; P < .0001), resulting in a risk factor for constipation (OR 2.31; 95% CI 1.63-3.27). By contrast, higher degrees of anterior colpocele appeared to protect against constipation (OR 0.80; 95% CI 0.66-0.96). No differences in prevalence of constipation were observed for urinary symptoms or urodynamic diagnosis. CONCLUSION: Bowel dysfunction correlates exclusively with posterior aspects of the pelvic floor support.

The effect of different macronutrient infusions on appetite, ghrelin and peptide YY in parenterally fed patients.

BACKGROUND & AIMS: Patients receiving parenteral nutrition (PN) still feel hungry despite adequate provision of calories intravenously. It is not known whether PN or its constituent macronutrients acutely affect appetite and to what degree this may be mediated by ghrelin and peptide YY (PYY). METHODS: Six medically stable patients (four men) with intestinal failure receiving PN received an isocaloric 200 kcal infusion on three separate occasions following a 12 h fast. The infusions consisted of either carbohydrate (10% dextrose), fat (10% intralipid) or mixed protein/carbohydrate (PN). Changes in ghrelin and peptide YY levels and changes in subjective symptoms of hunger, satiety and nausea during each macronutrient infusion were assessed. RESULTS: None of the three infusions acutely affected subjective symptoms of hunger, satiety and nausea (P>0.05 ANOVA). Ghrelin levels decreased significantly during dextrose [-19.1 (-35.9, -12.4), regression coefficient (95% CI), P<0.001] and parenteral nutrition infusions [-18.2 (-26.8, -9.6), P<0.001]. Lipid infusion had no effect on ghrelin levels but led to a significant decrease in PYY [-0.076 (-0.0123, -0.028), P=0.004]. Dextrose and PN infusion had no significant effect on PYY levels. CONCLUSIONS: Dextrose and PN infusions decrease ghrelin levels. Lipid infusion does not affect ghrelin levels but in contrast to oral nutrients leads to a significant decrease in PYY. Despite these changes, in patients receiving PN, macronutrient infusions do no acutely affect appetite.

Aberrant reflexes and function of the pelvic organs following spinal cord injury in man.

Spinal cord injuries can be devastating on quality of life and lost opportunities due to their impact on bladder, bowel and sexual functions. Supra-sacral spinal lesions can cause incontinence by interrupting those pathways, which normally coordinate the bladder, bowel and sphincters. From a scientific perspective, neural control of the pelvic organs is one of the most intriguing in the body, involving both somatic and autonomic pathways participating in an exquisitely fine integration of lumbo-sacral reflexes. This review details the interactions of somatic and autonomic lumbo-sacral pathways responsible for coordinating the bladder and sphincters, the nature of their aberration post-injury and those aspects of neural control of the pelvic organs that are amenable to neurophysiological examination in man. It will focus in greater detail on how measurement of pelvic floor and sphincter reflexes can be used to assess the modulatory effects of sacral autonomic pathways on sacral somatic reflexes and vice versa including the so called "guarding reflex" and vesical inhibitory reflexes. The effects of volitional modulation of these reflexes will be discussed in relation to people with both complete and incomplete lesions. Finally the possible utility of such neurophysiological measures for complementing the established neurological classification and the assessment of somatic sensory-motor impairment in spinal cord injury will be discussed.

Medical management of diverticular disease.

Diverticular disease is a common disorder affecting 30-55% of the population in the developed world. It has been described as "a deficiency disease of Western civilization", a theory that has been proposed since it was observed that diverticular disease was uncommon in African and other developing countries where dietary fibre consumption was high. At the turn of the century the medical treatment of colonic diverticulosis involved the avoidance of "roughage", which was felt to exacerbate the problem. However, the use of high fibre diets in the prevention and treatment of the symptoms of diverticular disease has now become commonplace, and the development of this practice is discussed in this chapter.

Anorexia nervosa in gastrointestinal practice.

OBJECTIVES: To characterize the demographic, psychosocial and prognostic features of patients with anorexia nervosa (AN) presenting to a gastroenterology service, and to compare them with patients presenting to an eating disorders unit. METHODS: A retrospective study set in two centres providing a local and tertiary service for gastroenterology and eating disorders. The notes of 20 consecutive patients with AN from each centre were compared. Comparison was made with a control group of 20 consecutive patients with slow transit constipation presenting to a gastroenterology service. RESULTS: Patients with AN who presented to a gastroenterology service were significantly older, had often seen a large number of hospital specialists, had a spectrum of gastrointestinal complaints, suffered a substantial delay in being diagnosed, and had undergone a significantly greater number of investigations and hospital admissions than AN patients attending an eating disorders unit. The parents of AN patients presenting to a gastroenterology clinic had a greater burden of physical and psychiatric illness than the parents in either of the other groups, and also tended to have separated when the patients were under the age of 10 years. Adverse prognostic factors among AN patients presenting to a gastroenterology clinic included older age at presentation, long history, unemployment, early parental separation and a body mass index less than 17. CONCLUSIONS: Patients with AN presenting to a gastroenterology service have profound psychosocial morbidity in excess of those presenting to a specialist eating disorders unit. Their diagnosis is often delayed. Early recognition and prompt referral to a specialist eating disorder unit should form the basis of management.