Digestive system diseases

Clinical record A previously well 4-month-old girl presented to an Australian Capital Territory hospital emergency department in August 2009 with a 2-day history of fever, vomiting and bloody diarrhoea with rectal prolapse. On arrival, she was alert, active and afebrile but had profuse diarrhoea and tachycardia. She was admitted, and stool cultures were requested. Intussusception was considered unlikely, and oral rehydration via a nasogastric tube was commenced. Over the following 48 hours, diarrhoea persisted, with some vomiting and low-grade fevers. Stool culture showed Salmonella species. After 4 days, the child was afebrile, had good oral intake and was discharged from hospital, although diarrhoea continued. No other family members were ill. Salmonellosis is a notifiable infection in all Australian states and territories, and interviews are sought with all affected patients in the ACT to locate sources of infection. Following receipt of the laboratory notification, an environmental health officer interviewed the mother of the infant. Exclusive breastfeeding meant that food-related illness was unlikely. The infant’s siblings were reported to handle kangaroo excreta and to examine the infant for new teeth. This was nominated as a possible source and mode of transmission. The household pet cat and dog were healthy. No other environmental exposures, including recent travel, were nominated. Subsequently, serotyping characterised the salmonella as Salmonella enterica serotype Rubislaw (S. Rubislaw), the first recorded instance of infection with this serotype in the ACT. S. Rubislaw has been isolated from various non-human sources in Australia: native animals and birds, feral goats, kangaroos and water supplies from the northern regions of Australia, and also captive snakes and lizards, including bearded dragons in South Australia in 2008 and New South Wales in 2009 (National Enteric Pathogens Surveillance Scheme [NEPSS], September 2009, unpublished data). The infant’s mother was re-interviewed and revealed that the family kept a 4-year-old pet eastern bearded dragon (Pogona barbata) (Figure). The mother reported that she cared for the lizard, fed it and cleaned its terrarium. The reptile was not permitted to roam the house but was removed from the terrarium on occasion to be held by the mother. There was no report of direct contact between the infant and lizard, and only rare contact with other children. The family was advised about the high Salmonella carriage in reptiles and the risk to young children. Environmental samples were collected from the terrarium (lizard faeces, drinking water, bark, swabs of the terrarium environment) and the household vacuum cleaner filter. S. Rubislaw was isolated from all these environmental samples. After being advised of these results, the family of the infant felt an unacceptable risk remained and the lizard was euthanased. Salmonellosis is frequently classified as a foodborne disease because contaminated food, mainly of animal origin, is the predominant mode of transmission.1 In the United States, “exotic pets” such as reptiles have become popular and reptile-associated salmonellosis has emerged as a public health problem.2-5 In Australia, although accurate figures on pet reptile ownership and reptile-associated disease are difficult to obtain, the risk to human health must be considered. International evidence suggests reptile-associated Salmonella infections are more likely to be associated with invasive disease, more commonly lead to hospitalisation and more frequently involve infants and younger children than do other Salmonella infections.4 The particular susceptibility of children to disease transmission from reptiles kept in the family home may be due to greater exposure to the contaminated reptile and greater host vulnerability.5 Much of the available evidence regarding risk in children comes from case–control studies conducted in the US.2,5 There have also been numerous reports of sporadic cases of invasive disease and hospitalisation of children and infants with salmonellosis linked to reptiles.3 A fatal case of S. Rubislaw meningitis in a 3-week-old baby was linked to a pet water dragon in England.6 It has been estimated that 90% of reptiles are colonised with Salmonella, which is shed intermittently in their faeces.7 Such high rates of colonisation suggest Salmonella are commensal organisms of the gastrointestinal tracts of these animals.4 There are over 2500 Salmonella serotypes, with most (including S. Rubislaw) belonging to S. enterica subspecies I.8 Serotypes of S. enterica subspecies I cause most Salmonella infections in humans and other warm-blooded animals. By contrast, over half of the S. enterica serotypes reported from reptiles in Australia belong to subspecies IIIa, IIIb and IV, which are rare in humans (NEPSS, unpublished data). Salmonella carried by reptiles can be transmitted either through handling a reptile, or indirectly by contact with an object contaminated by a reptile or its faeces.5 The droppings of free-roaming reptiles may result in widespread contamination of the home environment. The hands of younger children are readily contaminated, while infants may be indirectly infected via their parents or siblings. Ultimately, bacteria are transferred to the mouth and ingested. Hand washing may not eliminate the risk of infection. As the entire surface of a reptile may be contaminated, so may be the clothes and skin (beyond the hands) of the handler.9 Between nine and 19 human cases of S. Rubislaw were reported annually in Australia from 2000 to 2009 (NEPSS, unpublished data), predominantly from the northern tropical regions. Two-thirds of patients were aged less than 5 years (and most often less than 1 year), suggesting the importance of contact with environmental sources. We suspect the Australian public may not be familiar with reptile-associated salmonellosis. In the US, public health authorities have advocated the prohibition of the sale or gifting of reptiles without written point-of-sale education being provided by pet-store owners.3 Australian states and territories vary in their regulation of reptile ownership. NSW pet stores are not permitted to buy or sell reptiles or even to have them on their premises, whereas commercial reptile dealing occurs in Victoria, SA, the Northern Territory and ACT.10 In the ACT, eastern bearded dragons are classified as Category A reptiles (licence exempt), under the Nature Conservation Act 1980, and can be kept by individuals without prior experience in keeping such animals. Within the ACT there is no compulsion for pet stores, breeders or reptile keepers to inform prospective reptile owners or members of the public of possible risks to human health associated with reptile ownership or contact. We recommend that owners and prospective owners of reptiles be better informed of the risks to human health, particularly the threat to children. Lessons from practice Some Salmonella serotypes that occupy specific environmental niches infect humans through environmental sources rather than foods. Clinicians managing a young child with salmonellosis should ask parents or guardians about contact with household pets, including lizards, snakes, turtles and fish. Investigating individual cases of unusual infections can shed light on novel and evolving public health risks. Reptile-associated salmonellosis may become a more significant issue for medical practitioners and public health authorities in Australia. Parents and guardians should be advised about the risks (including invasive salmonellosis) that pet reptiles may pose to infants and young children.

Cameron R M Moffatt BBusHealthAdmin, MPH, MAppEpid · Antony R Lafferty MB ChB, FRACP · Sara Khan MB BS · Radomir Krsteski BApplSc, DipApplSc · Mary Valcanis BSc, MPH, MASM · Joan Powling BAgrSc · Mark Veitch MB BS, MAppEpid, FAFPHM

Objective: To evaluate the assessment and treatment outcomes of a prison hepatitis service.Design and setting: A retrospective, observational cohort study of prison inmates who attended hepatitis clinics from 1996 to 2005 at correctional centres in New South Wales.Patients: Inmates who attended the clinics, including a nested case–control series of patients who received antiviral treatment and age- and sex-matched patients who did not receive treatment.Main outcome measures: Demographic and clinical characteristics of patients who attended the service; correlates of selection for antiviral treatment; and clinical and virological outcomes of treatment.Results: Of the 1043 inmates who attended the clinics, 851 were men (82%) and 994 (95%) were referred for HCV infection; the mean age for this group was 33 years (range, 18–74 years). In the case–control series (185 treated and 186 untreated patients), selection for treatment was not biased by culturally and linguistically diverse background, current methadone treatment or psychiatric status. In the treated group, 76 of 138 genotyped patients had a genotype that is predictive of favourable treatment response, and a small minority of those with available liver biopsy results had established cirrhosis (7/119 patients). Of treated patients for whom complete follow-up data were available, 55% achieved sustained virological response and 100% adhered to therapy. In addition, treatment episodes were not especially complicated.Conclusion: Although the prison population has high rates of injecting drug use and poor mental health, imprisonment offers an opportunity for assessment and treatment of chronic HCV infection.

Leng Boonwaat RN, MPH · Paul S Haber BSc, MD, FRACP · Michael H Levy MB BS, MPH, FAFPHM · Andrew R Lloyd MB BS, MD, FRACP

A 79-year-old woman was admitted with shortness of breath that had worsened over the previous week. Two months earlier, she had developed dysphagia with nocturnal regurgitation, cough and occasional stridor. Physical examination showed no abnormalities apart from an oxygen saturation level of 85% on room air. The patient had a history of achalasia, which had been treated 5 years earlier with pneumatic dilation. A chest computed tomography scan showed a severely dilated and tortuous oesophagus with retained food (Figure, A) causing compression of the trachea (Figure, B). The woman was successfully treated with laparoscopic Heller myotomy and discharged.

Nicola Mumoli MD

To the Editor: Optimism about our growing ability to recognise coeliac disease (CD) is tempered by a worrying trend towards misuse of genetic testing for CD risk, as illustrated by this salutary case. A 42-year-old woman with nausea and lethargy attended her doctor, concerned that she may have CD, because her niece had the condition, and she knew it was “genetic”. A laboratory representative had told her doctor that a test for the HLA-DQ gene was the best test for CD; this test was performed, and detected genotype susceptibility for CD (HLA-DQ2 and HLA-DQ8 alleles). The patient was told she had CD and was referred to a dietitian to commence a gluten-free diet. She contacted the Coeliac Society of Australia and learned that this was an unusual way of diagnosing CD. This led to her being referred to a gastroenterologist. Further testing showed normal IgA and IgG transglutaminase antibody (TGA) levels (8 enzyme-linked immunosorbent assay [ELISA] units [reference range, 0–20 ELISA units]), and normal iron, folate and IgA levels. Mild nausea and diagnostic ambiguity prompted gastroscopy and duodenal biopsy. Findings of these tests were normal, showing no intraepithelial lymphocytosis. CD is common, affecting up to 1% of Australians,1 and causing a spectrum of complaints from classic malabsorption to more subtle problems (eg, osteoporosis and iron deficiency). It is clarified by sensitive serological detection methods (TGA levels).2 This gluten-triggered autoimmune disease represents one of the best characterised models of gene–environment interaction, in that an “at-risk” genome is needed to provide disease predisposition.3 However, CD genotyping (looking for HLA-DQ2 and HLA-DQ8 alleles) is increasingly requested under the misconception that a “genetic” test must be a more reliable detector of disease than other tools. The HLA risk genotype is necessary but not sufficient for development of CD: 20% of Australians share this risk factor, but only 5% of these express CD. Serological testing (TGA) provides approximately 90% sensitivity and specificity for CD in the appropriate clinical setting, but the diagnostic “gold standard” still involves biopsy confirmation of gluten-responsive tissue changes to justify the life-long social and financial costs of gluten avoidance. CD genotyping should be considered when: gluten challenge is not possible or acceptable; serological–histological discrepancies exist; or endoscopic biopsy is difficult or potentially non-diagnostic (eg, certain paediatric cases, or patients being treated with anticoagulants or immunotherapies [eg, prednisolone] that may normalise biopsy appearances). Our group is seeing more misuse of CD genotyping, with some practitioners using a positive result to confirm CD. As a centre of expertise in CD, we recommend that CD genotyping be confined to its negative predictive role: negative CD genotyping indicates a substantially reduced likelihood of CD (< 1%).4

Neil J Porter · Huy A Tran · Glenn E M Reeves

Objective: To examine the prevalence of perianal Crohn’s disease (PCD) and the eligibility of PCD patients to access anti-tumour necrosis factor-alpha (anti-TNFα) treatment under current Australian Pharmaceutical Benefits Scheme (PBS) guidelines.Design, setting and participants: A retrospective study of patients with Crohn’s disease (CD) and PCD attending four large adult inflammatory bowel disease (IBD) centres in Australia between January 2004 and May 2008. Patients for whom anti-TNFα therapy was clinically indicated were assessed to determine whether they satisfied PBS criteria for subsidised medication.Main outcome measures: Prevalence of CD and PCD in patients attending different IBD centres; eligibility of PCD patients for PBS-subsidised anti-TNFα medication.Results: Data were available on 3589 patients, representing about 6% of all patients with IBD in Australia. Of the 1815 patients with CD, 310 (17%) had PCD. Anti-TNFα therapy was deemed clinically indicated for 166 patients with PCD (54%), of whom 49 (30%) did not qualify for PBS-funded therapy.Conclusion: Thirty per cent of patients with clinically significant PCD currently do not have access to PBS-subsidised optimal medical treatment. We believe that PBS criteria should be extended to include this subgroup of IBD patients.

Daniel C Burger BSc, MB BS(Hons) · Ian C Lawrance MB BS(Hons), FRACP, PhD · Peter A Bampton MB BS, MD, FRACP · Ruth Prosser RN, BNurs · Anthony Croft BSc(Hons) · Kristen Gilshenan BMaths(Hons), BInfoTech · Graham L Radford-Smith MRCP, FRACP, DPhil · Timothy H Florin BSc(Hons), MSc, FRACP

McKittrick–Wheelock syndrome is a rare but recognised complication of hypersecretory rectosigmoid villous adenoma. Fluid and electrolyte imbalances require close monitoring because of large-volume losses of water, sodium and potassium. We report an unusual presentation of the syndrome associated with the development of acute pseudo-obstruction of the colon, presumably due to electrolyte dysfunction and acute renal failure. (MJA 2010; 192: 225-227) Clinical recordA 68-year-old, white, previously fit and healthy man was transferred to our facility with oliguric acute renal failure. The patient reported 72 hours of lethargy, myalgia, generalised weakness and cramps, following 3 weeks of watery diarrhoea, with pink-stained mucous discharge. He reported 10 motions per day, and no abdominal pain or vomiting. Some bloating had been noted in the preceding 24 hours. On examination, the patient had a heart rate of 85 beats/min, blood pressure of 115/65 mmHg, a respiratory rate of 20 breaths/min, and a temperature of 36.2°C. Oxygen saturation was 95% on room air. He was clinically dehydrated, with reduced tissue turgor and dry mucous membranes. Normal breath sounds were audible. His abdomen was noted to be grossly distended and non-tender. No masses or organomegaly were evident. Digital examination demonstrated an enlarged rectum with no palpable mass. Initial laboratory investigations (Box 1) showed hyponatraemia, hypokalaemia and elevated urea and creatinine levels. Haemoconcentration was evident. Results of investigations for intrarenal causes of renal failure were negative. Analysis of arterial blood gases showed a transient respiratory alkalosis and net mild metabolic acidosis. There were multiple causative factors, suggesting the presence of an underlying metabolic alkalosis. Renal tract ultrasound showed no obstructing lesion, with normal kidney size and morphology. Plain imaging (Box 2) and computed tomography (CT) of the abdomen and pelvis with intravenous and oral contrast demonstrated distension of the small bowel up to 4 cm in diameter, with a caecal diameter of 10 cm. Fluid material filled the sigmoid colon and rectum. A non-obstructing, exophytic mass arising from the lateral wall of the rectum, measuring 5 × 9 × 7 cm, was noted. A presumptive diagnosis of colonic pseudo-obstruction due to fluid and electrolyte imbalance was made. The patient was aggressively rehydrated and administered N-acetyl cysteine. He was admitted to the intensive care unit immediately after his CT scans for continuous haemofiltration, vasopressor support and total parenteral nutrition (TPN). A rectal tube was inserted, which drained at a rate of 4 L/day. Haemofiltration and noradrenaline were ceased on Day 2 and TPN on Day 3. Renal function and serum biochemistry normalised by Day 4. Colonoscopy revealed an exophytic, pedunculated tumour about 10 cm long, located 10 cm from the anal sphincters. The large bowel was successfully decompressed. A biopsy of the lesion demonstrated tubulovillous adenoma. The rectal tube was removed on Day 5, and the patient, having opened his bowels, was transferred to the ward. Stool cultures were negative for viral or bacterial pathogens. Loperamide and codeine phosphate were introduced, with effect. Given the patient’s recent renal impairment, indomethacin was not commenced. The patient was discharged home on Day 10. Staging magnetic resonance imaging of the pelvis performed on the day of discharge clearly demonstrated a large, sessile tumour arising from the left rectal wall (Box 3). Appearances were thought to be consistent with a T2 lesion. After 3 weeks of outpatient convalescence, the patient returned for elective low anterior resection, which revealed a large, exophytic lesion within the rectum (Box 4). Histopathological examination of the resected specimen (Box 5) confirmed hypersecretory tubulovillous adenoma with low-grade dysplasia. DiscussionMcKittrick–Wheelock syndrome was described in 1954 and is a rare complication of villous adenoma.1 It is typified by large-volume secretory diarrhoea, prerenal acute renal failure, and severe electrolyte dysfunction (primarily hyponatraemia, hypochloraemia, hypokalaemia and metabolic acidosis). The causative lesion is usually in the rectosigmoid and is normally over 4 cm in diameter.2 One series of 18 patients described tumours ranging between 7 cm and 18 cm.3 Roughly 2% of patients with rectosigmoid villous adenoma will develop hypersecretory complications.4 Cellular composition of non-secretory and secretory villous adenomas differs markedly. Light microscopy of non-secretory villous adenoma reveals relatively few goblet cells within the tumour epithelium, while much of the epithelial architecture of secretory adenomas is composed of these mucin-secreting structures. It has been postulated that the large surface area of the lesion participating in mucin production is a contributing factor to the volume of diarrhoea. The distal location of these lesions means there is a minimal area of normal colonic mucosa remaining to allow fluid absorption.5 The electrolyte composition of mucin secreted by abnormal cells within the causative lesion is of interest. Whereas normal bowel absorbs sodium and water and secretes potassium, segments of intestine affected by villous adenoma have been found to secrete water, sodium and potassium. Absorptive capacity was largely unchanged. In both cases, net movement of water was directly related to net movement of sodium. There was no relationship between net movement of water and potassium loss.6 The cause of this abnormal secretory function has been postulated to be secretagogue-mediated. Rectal effluent from a patient with villous adenoma of the rectum demonstrated prostaglandin E2 (PGE2) levels three to six times higher than normal.7 Tissue from villous adenoma and carcinoma synthesises more PGE2 than normal colonic mucosa. Mucosa adjacent to adenomatous polyps has been found to be unaffected by this, but carcinoma-associated mucosa was demonstrated to synthesise larger amounts of PGE2.8 These secretagogues are active at sites containing the prostaglandin synthetic pathway. Non-reversible cyclooxygenase-inhibiting agents have been used to reduce PGE2 production, and consequently loss of sodium and water through the rectum.9 Cyclic nucleotides have also been implicated.10 It has previously been suggested that elevated PGE2 levels may contribute to the adenoma–carcinoma sequence,11 explaining the beneficial role of non-steroidal anti-inflammatory drugs in preventing colorectal cancer as well as controlling diarrhoea. The presence of colonic pseudo-obstruction added another level of complexity to our case.12 Given the pathophysiology of hypersecretory villous adenoma and the predisposing factors for pseudo-obstruction, this association is not unexpected. Less than 5% of patients will develop colonic pseudo-obstruction idiopathically.13 In an analysis of 378 patients with pseudo-obstruction, 16 cases (4.23%) were precipitated by acute renal failure, and a further 15 (3.97%) by electrolyte dysfunction.14 In another retrospective analysis of 48 cases, 83% of patients were found to have some degree of electrolyte dysfunction.15 The mechanism by which homeostatic derangement causes pseudo-obstruction remains poorly understood. 1 Relevant admission laboratory results* Result Reference range Serum biochemistry Sodium (mmol/L) 120 135–145 Potassium (mmol/L) 3.2 3.5–5.0 Chloride (mmol/L) 86 99–107 Bicarbonate (mmol/L) 18 24–32 Urea (mmol/L) 54.6 4.0–9.0 Creatinine (μmol/L) 580 60–105 Glucose (mmol/L) 7.5 4.0–7.0 Osmolality (mmol/kg) 331 275–300 Corrected calcium (mmol/L) 2.14 2.23–2.50 Magnesium (mmol/L) 1.96 0.7–1.1 Phosphate (mmol/L) 5.15 0.6–1.3 Urine biochemistry Osmolality (mmol/kg) 383 50–1200 Sodium (mmol/L) 8 Variable Protein (g/L) 1.42 0.01–0.14 Full blood examination Haemoglobin (g/L) 191 122–170 White cell count (× 109/L) 15.8 4.6–10.5 Platelets (× 1012/L) 326 150–400 Haematocrit (%) 0.53 0.36–0.49 Mean cell volume (fL) 83 80–97 Neutrophils (× 109/L) 13.4 1.9–8.0 Arterial blood gas pH 7.47 7.35–7.45 PaCO2 (mmHg) 25 35–45 PaO2 (mmHg) 126 75–99 Base excess − 4 − 3 to 3 Other Albumin (g/L) 47 35–52 * Some initial volume resuscitation with normal saline had taken place prior to investigations. 2 Erect (A) and supine (B) abdominal x-rays taken on admission Note the uniform dilation of the large bowel, with multiple air–fluid levels (arrows). 3 Sagittal (A) and coronal (B) magnetic resonance images of the causative lesion Note the large mass occupying the mid to distal rectum, arising from the left rectal wall (arrows). The protruding mass extends almost to the anal verge. These images show no evidence of involvement of perirectal structures. The darkened signal on T2 images is caused by large amounts of mucin within the lesion. H = head. F = feet. A = anterior. P = posterior. R = right. L = left. 4 Macroscopic photograph of rectal villous adenoma found at operation The causative lesion was 12 cm in diameter, with elevation of 5 cm. 5 Low-power (A) and high-power (B) microscopic photographs of the causative lesion These images demonstrate features consistent with villous adenoma with low-grade dysplasia. Of particular interest is the large volume of gelatinous mucin produced by the lesion, as well as the high population of goblet cells evident along the epithelial border of the lesion (arrows).

Lachlan F Miles MB BS(Hons) · Christopher J Wakeman MB ChB, MMedSci, FRACS · K Chip Farmer MB BS(Hons), FRACS, FCSSANZ

To the Editor: Hellard and Wang1 are correct in emphasising the importance of the general practitioner in the management of hepatitis C virus (HCV) infection. As the authors note, HCV infection is a considerable source of morbidity and mortality in the community, and the infection may cause a substantial burden of illness in the future if it is not appropriately managed. The GP plays a pivotal role in managing HCV infection, being the first and most likely point of contact for patients. However, Hellard and Wang fail to note that the GP’s most useful role is to inform patients that “alcohol abstinence is strongly recommended before and during antiviral therapy”.2 The well recognised role of alcohol in disease progression is emphasised in the position papers of both the American Gastroenterological Association and the United States National Institutes of Health.2,3 From a public health perspective, it is difficult to think of a more cost-effective approach to the management of such a public health issue.

Anne E Duggan · John M Duggan

In reply: Duggan and Duggan are correct to highlight the well recognised role of alcohol consumption in progression of hepatitis C virus (HCV) infection. Alcohol consumption has been found to increase viral load and accelerate hepatic fibrosis in HCV infection.1,2 While studies have reported that a history of alcohol consumption adversely affects treatment outcomes (with some reporting a dose–response relationship),3,4 treatment success has also been reported among patients who continue to consume moderate amounts of alcohol during treatment.5 Although there are biologically plausible mechanisms through which alcohol consumption might negatively affect treatment, low rates of treatment success among drinkers may also be related to lack of adherence to treatment regimen in this population.1 To date, no study has specifically measured the effect of alcohol consumption during treatment while adequately controlling for the effects of compliance, disease progression and baseline viral load. Until studies are undertaken that measure the direct effect of alcohol consumption on treatment success, while adjusting for compliance, it seems reasonable to advise patients to decrease their level of alcohol consumption before and during HCV treatment. However, given that some patients have successfully completed treatment without abstaining from alcohol consumption, this should not be an automatic exclusion criterion.

Margaret E Hellard · Yung-Hsuan J Wang · Rachel Sacks-Davis

To the Editor: Over-the-counter herbal supplements to promote weight loss have become increasingly popular. Several of these products contain potentially hepatotoxic substances. We present the first reported Australian case of acute hepatotoxicity associated with the weight-loss product Hydroxycut Hardcore (Iovate Health Sciences, Blasdell, NY, USA). Hydroxycut contains various ingredients, including extracts of the herbs Garcinia cambogia and Camelia sinesis (green tea root), and the chromium salt, chromium polynicotinate. A recent review cites these three ingredients as possible causes of Hydroxycut hepatotoxicity.1 A 23-year-old, previously well, construction worker presented to his doctor with a 2-month history of lethargy and jaundice. Test results confirmed serum liver enzyme derangement, and the patient was advised to stop taking the weight-loss supplement. Results of repeat testing a week later showed worsening liver enzyme levels and he was referred to our hospital. On arrival, the patient denied any symptoms except moderate lethargy and icteric sclera. He was usually well with no relevant medical or surgical history. He weighed 83 kg. He denied drinking alcohol, using prescription or non-prescription drugs, or receiving a blood transfusion. He had had unprotected sexual intercourse 2–3 months earlier. A previously obtained tattoo was being extended by a reputable tattoo parlour. He had been using Hydroxycut Hardcore daily for 10 weeks (obtained from his local outlet of a global nutritional products retailer) in an attempt to lose weight and tone muscle. He claimed to have taken the dosage recommended by the manufacturer (six capsules daily); he was taking no other supplements. Physical examination revealed mild jaundice without other features of chronic liver disease or portal hypertension. However, in addition to abnormal blood levels of liver enzymes, bilirubin and iron, results of the patient’s liver biopsy showed severe acute hepatitis (Box). The abnormal test results were consistent with acute drug toxicity. The patient improved without specific treatment and was discharged 8 days after presentation, with near-normal serum liver enzyme levels. He was well on follow-up at 4 weeks. Two reports from the United States link Hydroxycut Hardcore to acute liver injury in otherwise well young males.2,3 The American Food and Drug Administration in May 2009 advised consumers to stop using the product, based on 23 reports linking it to serious injury, including one case of liver failure leading to death.4 In May 2009, Australia’s Therapeutic Goods Administration (TGA) issued a warning to consumers about the product, although noting that no adverse events had so far been reported in Australia.5 In view of this first reported Australian case of Hydroxycut hepatotoxicity, we advise medical practitioners and consumers in this country to be wary of the product, and call on the TGA to re-examine its continued availability. Investigation results for a 23-year-old man with liver dysfunction after using Hydroxycut Hardcore Investigation Result (reference range) Blood tests Aspartate aminotransferase (U/L) 1182 (12–36) Alanine aminotransferase (U/L) 2950 (5–40) Alkaline phosphatase (U/L) 121 (50–140) Bilirubin (µmol/L) 113 (3–18) Prothrombin time (seconds) 13 (11–15) Iron (µmol/L) 68 (11–30) Ferritin (µg/L) 1897 (30–400) Iron saturation (%) 99 (16–50) Paracetamol Not detected Hepatitis A, B, C Negative HIV Negative Autoantibodies Not detected Epstein–Barr virus, cytomegalovirus, toxoplasma, leptospira, coxiella Negative Haemochromatosis genetic testing No abnormality Other tests Abdominal ultrasound No abnormality Percutaneous liver biopsy Severe acute lobular hepatitis with areas of bridging necrosis; no bridging fibrosis or cirrhosis were seen Hepatic iron index 1.1 (< 2.0)

N Nudrat Rashid · Jason Grant

Charles Darwin visited New Zealand in December 1835, and Australia from January until March 1836, on the return portion of his voyage around the world in HMS Beagle. Despite the shortness of these visits, he retained an interest in these countries throughout his life, maintaining correspondence and receiving many biological specimens. His experiences in these places influenced his thinking on evolution, particularly on the evolution of man. Aspects of his health recorded during this part of the voyage support a new hypothesis for the diagnosis of the illness that Darwin endured for most of his life.

John A Hayman MB BS, MD, FRCPA

A 70-year-old man developed melaena after taking aspirin and clopidogrel for 2 months to treat ischaemic heart disease. Other medications the patient was taking included metformin and metoprolol. A lipoma, the endoscopic appearance of which resembled a finger (Figure), was the only endoscopic abnormality. It was resected endoscopically, with the aid of an endoloop to cut off its blood supply before removal with a snare. The diagnosis was histologically confirmed but ulceration was not found, possibly because one section of the lesion was not retrieved. Lipomas are soft and typically appear as an isolated bulge of smooth mucosa, often with a yellow hue. Such lesions bleed rarely.

Hayley Clifford · Millie Lui · Roger Lee · Andrew Thomson

General practitioners hold the key to expanding access to treatment Hepatitis C virus (HCV) is an important cause of morbidity and mortality in Australia. More than 200 000 people are estimated to be living with chronic HCV infection, with over 80% of these infections resulting from unsafe injecting drug use. About 10 000 new infections occur annually, although incidence is thought to be declining.1 Following primary HCV infection, persistent viraemia and chronic hepatitis occurs in 50%–80% of patients; after 20 years approximately 7% develop cirrhosis, and a small proportion of these patients develop hepatocellular carcinoma.1 Treatment of HCV infection has advanced over the past 10 years, leading to improved outcomes; the most effective current treatment is pegylated interferon combined with ribavirin. The aim of treatment is viral eradication, and treatment is deemed successful if a patient has a sustained virological response (SVR).2 The subsequent reduction in liver disease progression in patients who obtain an SVR3 suggests that the burden of advanced liver disease could be reduced if more patients received treatment. Treatment uptake has increased over the past 10 years, but has remained low. The removal of restrictions to prescribing for treatment — including the requirement for patients to have abnormal alanine aminotransferase levels or liver biopsy results (restrictions that were removed in 2005 and 2006, respectively) — has increased access,4 but the total number of individuals being treated for HCV infection in Australia remains low, at around 3500 individuals per year.5 Estimates vary, but recent modelling suggests that at least 6000, and closer to 10 000, people with chronic HCV infection need to be treated annually to reduce the burden of advanced liver disease in the future.1 While we acknowledge that not all people want or can have treatment for HCV infection, the number could be increased. HCV treatment can only be prescribed by certain medical practitioners or specialists, or at liver clinics, and cannot be prescribed by most general practitioners. Opportunities for treatment in tertiary hospitals and opioid pharmacotherapy clinic settings need to be expanded,6 but equally important is increasing GPs’ capacity to manage and treat patients living with HCV. GPs are usually the initial point of contact for patients with or at risk of HCV. It is imperative that GPs provide clear, accurate and up-to-date advice on HCV risk, prognosis and management. Several surveys in the past 10 years have reported that most GPs want further education about treatment, interpretation of test results, pre- and post-test counselling, and referral information.7,8 Many GPs’ knowledge of HCV is limited. A study undertaken in the period 2005–2006 found that, although GPs were aware of which patients are at risk of HCV (injecting drug users in particular), many underestimated the large number of Australians infected with HCV.9 A 2002 study reported that 39% of surveyed GPs mistakenly believed that positive results from HCV serological testing, as opposed to positive results from HCV RNA testing, differentiated current and resolved infection.7 An anti-HCV antibody test only provides information on whether a patient has ever been exposed to HCV — a positive result does not necessarily indicate an ongoing infection. An HCV RNA test is required to determine whether a patient has an ongoing infection or has spontaneously cleared their infection. GPs’ awareness of HCV treatment was also limited. The 2005–2006 study showed that only 42% of surveyed GPs were aware of the effectiveness of current HCV treatment, and only 28% were aware of the eligibility criteria for access to subsidised treatment.9 The 2002 study showed that 52% of GPs were not aware that pegylated interferon–ribavirin combination therapy was the most effective HCV treatment.7 Other studies have shown that fewer than 52% of people living with HCV in Australia had ever been referred to a specialist liver clinic10,11 despite the general acceptance that this should occur for the vast majority of patients with HCV. In addition, many GPs have reported being uncertain about when to refer patients to hepatitis specialists.9 Ongoing education is vital if GPs are to remain up to date on the management of HCV infection. A range of options are required to meet the needs of different GPs and their patients. One option is training GPs to fully manage their own patients, including training to become accredited prescribers of pegylated interferon and ribavirin under the federal government’s Section 100 (s100) Highly Specialised Drugs Program. Currently, the availability of such training is limited and varies between states and territories. A shared care model is a second option. Some GPs could be actively involved in the management and follow-up of patients without being an s100 prescriber. This model could suit GPs who manage only a few patients with HCV infection, as well as nurse practitioners and GPs in rural and regional areas. A specific training program could be developed to provide GPs and nurse practitioners with the necessary knowledge and skills, particularly for managing the side effects of HCV treatment. A third option should be provided to potentially the largest group of GPs — those who have patients at risk of or infected with HCV, but who do not want to be directly involved in HCV management. Through ongoing and regularly updated training programs, these GPs need to stay up to date on who is at risk of HCV infection, what tests should be ordered to diagnose and monitor HCV infection, and when patients should be referred to a specialist. In addition, all GPs need to have a broad understanding of what HCV treatments are available, that current and recent injecting drug users are eligible for HCV treatment, and that a liver biopsy is no longer required for a patient to have access to treatment (Box). Encouraging GPs to undertake HCV training in a setting of competing education priorities is a major challenge. Perhaps the first step should be to highlight that HCV occurs in about 1% of the population and that treatment options and outcomes for their patients have improved considerably during the past 5 years and are likely to continue to do so. Hepatitis C virus (HCV): what general practitioners should know HCV is common — more than 200 000 Australians have ongoing infection. Approximately 25% of people infected with HCV spontaneously clear their infection.12 An anti-HCV antibody test only detects whether a patient has ever been exposed to HCV — it does not detect ongoing infection. An HCV RNA test is required to determine whether a patient has an ongoing infection or has spontaneously cleared their infection. HCV can be successfully treated with pegylated interferon combined with ribavirin. The most common HCV genotypes in Australia are genotype 1 and genotype 3.1 Genotype 1 HCV infection usually requires 48 weeks of treatment, and genotype 3 usually requires 24 weeks of treatment; the chance of successfully clearing the virus with such treatment is approximately 45% and 75%, respectively.13 Treatment is deemed successful if a patient has a sustained virological response — defined as having a negative HCV RNA test result 24 weeks after completion of treatment. After successful treatment, patients will be HCV RNA negative but will remain anti-HCV antibody-positive in the vast majority of cases. Subsidised treatment is available to people older than 18 years who are anti-HCV antibody-positive, have detectable serum HCV RNA levels, have compensated liver disease, and have not had prior treatment with pegylated interferon or interferon alfa.

Margaret E Hellard FRACP, PhD, FAFPHM · Yung-Hsuan J Wang MB BS, FRACGP, MAppEpid

Objective: To describe the patterns of screening for hepatitis C virus (HCV) infection in methadone-maintained pregnant women and their infants.Design, setting and patients: Retrospective review of medical records from one rural and two metropolitan hospitals in New South Wales for pregnant women on methadone maintenance treatment and infants born to these women between 1 January 2000 and 31 December 2006, as well as records for pregnant women who were not on methadone treatment.Main outcome measures: Rates of anti-HCV antibody and HCV RNA testing for pregnant women and their infants, and ages at which infants attended follow-up appointments.Results: Of 295 pregnant women on methadone maintenance treatment, 288 were tested for anti-HCV antibodies (98%), compared with 1995 of 9987 women who were not on methadone treatment (20%) (P < 0.001). Seropositive results were obtained for 243 women in the methadone group (84%) and 54 in the non-methadone group (3%) (P < 0.001), of whom 44 (18%) and 17 (31%), respectively, were subsequently tested for HCV RNA (P = 0.03). HCV RNA test results were positive for 31 (70%) and 10 (59%) seropositive women in the methadone and non-methadone groups, respectively (P = 0.39). Of infants of HCV-seropositive methadone-maintained mothers, 27% of those for whom we had follow-up attendance data received HCV screening, and one of these infants tested positive for anti-HCV antibodies and HCV RNA.Conclusions: Screening for HCV infection in the high-risk population of pregnant women on methadone maintenance treatment and their infants is inadequate. This could lead to significant underdetection of active HCV infection in this high-risk population, and their infants. Current screening guidelines may therefore need to be revised.

Anthony J W Liu,* MB BS, FRACP, MPH · Ethan I An,* BMedSc, MB BS(Hons) · Henry G Murray MB ChB, MRCOG, FRACOG · Emma Tetstall BSc(Hons), MB BS(Hons) · Marcel J Leroi FRACP, FRCPA, MMed(ClinEpi) · Ralph K H Nanan Dr med Habil (Germany), FRACP

To the Editor: The recent withdrawal of faulty faecal occult blood testing (FOBT) kits by the government highlights the flaws in the National Bowel Cancer Screening Program (NBCSP), with 475 000 kits needing to be replaced. In their recent editorial,1 Ee and Olynyk attempted to persuade us that the NBCSP, which has now temporarily suspended the issuing of new invitations to participate, is both rational and appropriate for an affluent country such as Australia. The facts about bowel cancer mortality are not disputed — it is second only to lung cancer, equating to about 12 Australian deaths per day from a potentially preventable cancer. There are only two serious contenders for bowel cancer screening: FOBT and colonoscopy. Annual FOBT has the potential to reduce mortality by 15% (realistic) to 33% (optimistic).2 Colonoscopy is both diagnostic and preventive, with a predicted reduction in mortality of 76% (realistic) to 90% (optimistic).3 Recent United States guidelines advocate colonoscopy at 10-yearly intervals as the test of choice.3 Cost and potential mortality are two major issues raised by proponents of FOBT versus colonoscopy. However, Australian data using all costs, including infrastructure, colonoscopy, surgery, pathology, anaesthesia, chemotherapy and disinfection, have shown that 10-yearly colonoscopy is of equal cost–benefit to annual FOBT.4 A recent publication from Germany highlights some important facts.5 In a study of 269 144 colonoscopies, the completion rate (ie, reaching the caecum) was high, at 97.3%. The perforation rate was low (0.02%), as was the incidence of bleeding (0.16%), though this was predictably greater in those requiring polypectomy (0.8%). There were no deaths attributable to perforation or bleeding. Colonoscopy is therefore a rational option for screening, with the potential for prevention — a feature not shared by FOBT. The NBCSP’s letter of invitation and FOBT kit are sent only to individuals aged 50, 55 and 65 years. It ignores other age groups and gives no educational advice about alternative strategies such as colonoscopy. Ee and Olynyk1 clearly believe that centrally organised government screening programs are to be preferred, and they disapprove of colonoscopy screening “driven by patients”. Recent case series show colonoscopy to be safe and to have a higher completion rate than older series.5 Current initiatives in Australian colonoscopy training programs should improve safety and completion rates further. For those with no active interest in their health, centralised, directive programs are likely to be the most effective. For the informed, however, discussion about lifestyle changes, screening programs and healthy dietary alternatives should be their right.

Terry D Bolin · Melvyn G Korman · Alistair E Cowen

Australia can benefit from lessons learned in the epidemic of C. difficile infection in Europe and North America It is 30 years since Clostridium difficile was shown to be the cause of pseudomembranous colitis and many cases of antibiotic-associated diarrhoea in humans. In the interim, C. difficile has risen from relative obscurity to become a major hospital pathogen. Two factors were particularly important in its emergence during the 1980s. First, increased and inappropriate use of some broad-spectrum antibiotics, particularly cephalosporins, predisposed more patients to infection with C. difficile. Second, contamination of the hospital environment with C. difficile spores was, and remains, a significant problem, as the spore is likely to be the infective particle. The epidemiology of C. difficile infection continues to evolve, and developments overseas in the past decade threaten not only parts of Australia’s vast agricultural sector but also the country’s health care system. Since 2002, rates of C. difficile infection have escalated, with outbreaks of severe infection in North America and Europe caused by an epidemic strain — polymerase chain reaction (PCR) ribotype 027 (also known as North American pulsed-field type 1 [NAP1]). This strain is characterised by the production of greater amounts of toxins A and B and an additional, binary toxin, as well as resistance to fluoroquinolone antimicrobials.1 When this editorial was submitted for publication in January 2009, there was no evidence that this epidemic strain was present in Australia. However, C. difficile PCR ribotype 027 has now been isolated for the first time in Australia, as reported in this issue of the Journal2 (Riley et al). Although the patient most probably acquired the organism while travelling in North America, this case illustrates the ease with which it could be introduced into Australia. Thought to be driving the epidemic in humans in North America and Europe are the overuse of fluoroquinolones and fluoroquinolone resistance, but the ageing population and improved case ascertainment may also be contributing to the dramatic increase in cases. Other factors may also be important, such as the increase in prescription of proton-pump inhibitors, which coincided with the emergence of epidemic C. difficile.3 Several recent observations from overseas have broad relevance for Australia. First, there has been an apparent increase in community-acquired C. difficile infection in the absence of classic risk factors such as antibiotic exposure, leading to suggestions that all patients with community-acquired diarrhoea should be tested for C. difficile.4 Assertions that community-acquired C. difficile infection is a new disease4 are not correct — it has been recognised in Australia for over 15 years but is underdiagnosed.5 Therefore, it is difficult to determine whether this increase is a true increase or rather reflects better case ascertainment. Nonetheless, the suggestion that C. difficile infection should be considered more than just a hospital problem is valid, and general practitioners need to be aware of this change in epidemiology. The prevalence of binary toxin-producing C. difficile in human disease is also increasing, and there is an association between binary toxin-producing isolates and community acquisition.6 Second, it is speculated that C. difficile is part of a zoonosis, and that transmission of infection via spores may be foodborne.7 There is compelling evidence for the former, but none for the latter. C. difficile is known to colonise many animals.8 Indeed, as in humans, it probably colonises the gastrointestinal tracts of most infant animals until weaning. There was alarm at a report that 20% of a small sample (n = 60) of retail beef in Canada contained C. difficile.9 Equally disturbing are reports that many pig herds in the United States are infected with C. difficile. The overall prevalence of C. difficile in piglets from 10 herds in North Carolina was 48%, and ranged from 0 to 97% across the herds. Mortality for piglets with C. difficile infection is 15%, and animals that survive are 10% underweight when they go to market.10 Most animal isolates of C. difficile produce binary toxin, and both pigs and cattle harbour PCR ribotype 078 — a strain that, like ribotype 027, produces increased amounts of toxins A and B, in addition to binary toxin. In the Netherlands, the prevalence of human C. difficile infection with ribotype 078 strains has increased since 2005; these infections were in a younger population and more frequently community-acquired than infections with ribotype 027 strains. In the eastern Netherlands, where more than 90% of the country’s pig farms are located, over 20% of human isolates are now ribotype 078, and human and pig strains of C. difficile are highly genetically related.11 In Australia, little is known about the prevalence of C. difficile in pigs. A small study in 2007 found C. difficile in 10 of 37 samples (27%) from piglets with diarrhoea, but none of the isolates were ribotype 078 (unpublished data). Why is C. difficile infection increasing in pigs in Europe, and what are the implications for Australia? The use of antimicrobials for growth promotion was banned from 2006 in Europe, and even earlier in Denmark, starting in 1995. However, since 2000, the use of therapeutic antimicrobials in production animals has increased in Europe in general, and specifically in Denmark, a big producer of pork. Of real concern is evidence of greater use of cephalosporins in animals. While the number of pigs in Denmark increased by 50% in the past 15 years, the amount of penicillinase-susceptible penicillins used increased by 400%, and cephalosporins by 1000%. Most of this increase was in piglets and sows.12 Although the total amount of cephalosporins used remains small, this is a worrying trend. If the situation is similar in the Netherlands, and anecdotal evidence suggests that it is, then this may be analogous to the situation in humans in the 1980s when there was a dramatic increase in C. difficile in many hospitals, driven by cephalosporin use.13 The overlap between the location of pig farms in the Netherlands and the occurrence of human ribotype 078 infections suggests a common source.11 This is likely to be the environment. The Netherlands has one of the highest population densities in the world. If infection rates in pig farms in the Netherlands are as high as those in the US,10 then it is likely that a large proportion of the Dutch population comes into contact with C. difficile spores every day. Individuals are at risk of infection if they are taking antimicrobials or any other medication that perturbs the gut flora. The good news for Australia is that, with our very low population density, a similar risk to humans is unlikely to develop. However, this is no reason for complacency. Every effort should be made to stop epidemic C. difficile from becoming established in our production animals. Unfortunately, the mere perception of C. difficile infection as a foodborne disease will damage the industry. Even before the first isolation of C. difficile PCR ribotype 027 in a patient in Australia, health care practitioners were becoming justifiably concerned. A proposal for C. difficile to be made notifiable in all states and territories of Australia was approved at the Australian Health Ministers’ Advisory Council meeting in November 2008. Australia’s conservative policies on fluoroquinolone use in humans and animals may offer some protection. However, if cephalosporin use is driving C. difficile infection in animals overseas, then additional efforts to target cephalosporin use in veterinary medicine may be needed in Australia. The solution to these problems continues to lie in surveillance for the emergence of virulent strains of C. difficile, promotion of judicious use of antimicrobials in both human and veterinary medicine, and environmental cleanliness, the latter perhaps easier said than done outside health care facilities.

Thomas V Riley MAppEpid, PhD, FRCPath

Objective: To characterise epidemiological, clinical and laboratory features of children in New South Wales with chronic hepatitis B (HBV) or C (HCV) infections.Design and setting: Retrospective record review of epidemiological, clinical, laboratory, liver biopsy and treatment data for children (aged < 18 years) referred to tertiary referral paediatric and refugee clinics in NSW with chronic HBV or HCV during 2000–2007; and comparison with NSW Health notification data for the same period.Main outcome measures: Numbers and characteristics of referred children with HBV and HCV, and notifications to NSW Health.Results: During 2000–2007, 79 children with chronic HBV and 29 with HCV infection were referred to specialist clinics, while 930 children with HBV and 777 with HCV infection were reported to NSW Health. Most of the referred children with HBV were born overseas, while most with HCV were born in Australia to mothers with a history of intravenous drug use. Of the 79 HBV-infected children, 56 were e-antigen positive. Most HCV-infected children (23/29) had alanine aminotransferase levels ≤ 2 times the upper limit of normal, and more than half of those who had genotype determined had type 2 or 3. Fibrosis was evident in liver biopsies performed for both HBV and HCV.Conclusions: Although advanced liver disease was uncommon in children referred with HBV or HCV infection, a large number of infected children in NSW were not referred for specialist medical care, indicating that opportunities to intervene early in the natural history of these infections, particularly HCV, are being missed.

Scott Nightingale BMed(Hons), MClinEpid, FRACP · Michael O Stormon MB BS, FRACP · Andrew S Day MD, FRACP · Murray T Webber BMed, FRACP · Kate A Ward BPhty(Hons), MPH · Edward V O’Loughlin MD, FRACP

To the Editor: We read with interest Talley’s excellent and informative editorial about the risks associated with proton-pump inhibitors (PPIs).1 Other possible serious side effects of PPIs that need to be taken into account are potential drug interactions with aspirin and clopidogrel. Aspirin is a weak acid that crosses the mucosa in its lipid state. The suppression of acid production reduces the lipophilic nature of this drug and, theoretically, might reduce its absorption and bioavailability.2 On the other hand, clopidogrel is a prodrug that is converted in the liver to an active metabolite. This bioactivation is mediated by hepatic cytochrome P450 isoenzymes,3 with cytochrome P450 2C19 (CYP2C19) playing a particularly important role. There is evidence suggesting that some PPIs (omeprazole, lansoprazole and rabeprazole) can inhibit CYP2C19, which would alter the effectiveness of clopidogrel and potentially lead to an increased risk of adverse cardiovascular outcomes. In a recent Canadian case–control study among patients prescribed clopidogrel after acute myocardial infarction, current use of PPIs was associated with an increased risk of reinfarction (adjusted odds ratio, 1.27; 95% CI, 1.03–1.57).4 The risk was limited to patients currently taking a PPI (the authors did not find any association with more distant exposure to PPIs), and did not extend to pantoprazole, a drug that does not interfere with the conversion of clopidogrel to its active form.

Francisco J Fernández-Fernández · Gonzalo Pía · Pascual Sesma

To the Editor: In his recent editorial, Talley discusses a range of risks of proton-pump inhibitors (PPIs).1 Another rare but serious side effect of PPIs of which every doctor should be aware is hyponatraemia. Eleven cases of hyponatraemia caused by PPIs have been published.2,3 Consistent features were the rapid onset of hyponatraemia within days of commencement of the PPI therapy, the severity of hyponatraemia often being associated with confusion or delirium, and rapid recovery after cessation of the PPI medication. Test results in each case were consistent with inappropriate release of antidiuretic hormone. One case occurred 5 days after a patient changed from lansoprazole to esomeprazole.4 Hyoponatraemia needs to be considered whenever there is clinical deterioration, even after brief exposure to a PPI.

Adam P Morton

In reply: Proton-pump inhibitors (PPIs) are often coprescribed for patients taking aspirin and clopidogrel to reduce gastrointestinal bleeding. There are emerging data that omeprazole diminishes the therapeutic effect of clopidogrel because the active enzyme in the liver, cytochrome P450 2C19 (CYP2C19), metabolises omeprazole and activates clopidogrel.1,2 In a large cohort study of 8205 patients with acute coronary syndrome and taking clopidogrel, 64% were also taking a PPI (60% omeprazole); 21% of those who were taking clopidogrel but no PPI died or were rehospitalised for acute coronary syndrome, versus 30% of those taking clopidogrel as well as a PPI.3 Notably, not all the PPIs have the same metabolic pathway. For example, omeprazole and esomeprazole are principally metabolised by CYP2C19 in contrast to lansoprazole, which is metabolised by cytochrome P450 3A4 (CYP3A4), and pantoprazole, which is metabolised by CYP2C19 O-demethylation then rapid sulfate conjugation. Thus, the negative interaction with clopidogrel may not apply to all PPIs (and pantoprazole may be the drug of choice if a PPI is required, as cytochrome P450 interactions are least likely).1 However, until more data are accumulated, all PPIs should probably be avoided where possible in patients who have been prescribed clopidogrel, unless there is no alternative. It is correct that hyponatraemia has, rarely, been reported in patients taking PPIs. However, this knowledge is based solely on case report data, and therefore the level of evidence for cause and effect is relatively weak.

Nicholas J Talley · Aneta Dimoska · Kevin Gan

An 81-year-old man presented with a 1-day history of right iliac fossa pain that radiated to the left iliac fossa. Vital signs were normal. Examination revealed abdominal distension with peritonism. An abdominal x-ray showed small bowel obstruction, and a computed tomography scan suggested a small bowel volvulus. Laparotomy revealed a small bowel obstruction due to a pendulous appendix (20 cm in length) that had wrapped around a loop of bowel, causing obstruction and ischaemia (Figure). Histopathological examination confirmed the presence of acute appendicitis with ischaemic small bowel. The patient recovered uneventfully. The average length of the adult appendix is 7.5 cm in men and 6.7 cm in women (SD, 2.1 cm; range, 2–26 cm).1 Small bowel obstruction secondary to “appendiceal tourniquet” is extremely rare, with only 12 cases previously reported.2,3

Richard H Riley · Richard J Martin

An appropriate strategy is to test patients with symptoms and signs attributable to coeliac disease In recent years, knowledge about coeliac disease (CD) has improved significantly, and we now have a better understanding of the diagnosis and pathogenesis of the disease. In this issue of the Journal Chin and colleagues report the largest community study of CD undertaken in Australia.1 The study found a prevalence of CD of between 0.56% and 0.96%,1 similar to prevalences found in European population studies. A recent population-based serological study (without biopsy) in the United States found a prevalence of 0.95%.2 The screening strategy used by Chin and colleagues involved IgA and IgG anti-tissue transglutaminase (anti-tTG) antibody assays, HLA-DQ2 and HLA-DQ8 haplotyping, and, where appropriate, gastroscopy and duodenal biopsy.1 We believe that the data do not justify population-based screening for CD in Australia, but the diagnostic approach developed by Chin et al is appropriate for diagnosing patients in whom CD is suspected. Genetic and environmental factors are important in the genesis of the disease. The role of the HLA-DQ2 and -DQ8 haplotypes is yet to be fully understood. One or other haplotype is present in nearly all CD patients, yet they are also present in 20%–40% of subjects without CD. The “gold standard” for diagnosis has long been histological findings from a duodenal biopsy, despite the discordance between clinical symptoms and these findings. CD is caused by an immune response to antigens in gluten proteins. Wheat, barley and rye, but not corn or rice, contain gluten proteins. Ingested gluten is partially digested to peptides, some of which are absorbed into the lamina propria of the small intestine. The enzyme, tissue transglutaminase, deamidates certain glutamine residues in the peptides to glutamic acid. In patients with CD, this results in enhanced binding to HLA-DQ2 or -DQ8 cells. Within the lamina propria, T cell recognition of the amino acid sequences presented by the DQ2 or DQ8 cells produces a series of inflammatory changes, with the release of cytokines and the activation of lymphocytes. This leads to the characteristic histological findings in the proximal duodenum — lymphocytic infiltration, villous damage or loss, and crypt hyperplasia.3 Increases in transglutaminases in the lamina propria produce several effects: they catalyse the crosslinking of gluten peptide residues to lysine residues in various local proteins, including in transglutaminase itself, creating an immunogenic gluten–peptide–protein complex. This becomes an autoantigen, producing the characteristic IgA/IgG anti-tTG antibody of CD, which is best detected by an enzyme-linked immunosorbent assay. The mainstay of diagnosis of CD remains the anti-tTG assay, followed by duodenal biopsy, with the patient on a gluten-containing diet (equivalent to four slices of wheat bread a day) for several weeks. It is important to recognise that false-negative test results do occur, an underappreciated cause being IgA deficiency, which occurs in one in 40 people and can be identified using a combined IgG/IgA anti-tTG assay. Recent evidence suggests that the IgA/IgG anti-tTG antibody test has superior performance (sensitivity of about 90%, specificity of about 80%) to the anti-endomysial antibody (EMA) test, is less expensive, and overcomes the potential problem of IgA deficiency.4 If doubt persists, HLA haplotyping is useful, and a negative test for both DQ2 and DQ8 effectively excludes the diagnosis.5 Several factors should raise suspicion of the presence of CD and prompt testing: first-degree relatives of CD patients; adult type 1 diabetes mellitus; osteoporosis; ill-defined abdominal symptoms (eg, irritable bowel syndrome); iron or other nutritional deficiency; unexplained transaminase elevations; or Down syndrome. In patients in whom testing points to CD, the benefits of a gluten-free diet should be considered. Given that a gluten-free diet is arduous, we need to be assured that the risk–benefit ratio is appropriate. The benefits in those with florid symptoms are life-transforming (“Doctor, I never knew what it was like to be well before the diet”), but for the symptomless patient the gains in quality or quantity of life are less clear.2 Recent studies show a mild increase in risk of death from malignancy, which appears to diminish with time from diagnosis. Future research needs to quantify the benefits, if any, of lifelong treatment of people with asymptomatic and currently undiagnosed CD. What, then, is the role of screening for CD? To be justified, population screening must be effective, acceptable and cost-effective, and these conditions are not met for CD.6 Focused screening is preferable, as shown by a number of primary care studies. A study from nine surgeries in central England screened 1000 sequential patients with an anti-EMA test and found 30 patients in whom CD was confirmed by small-bowel biopsy. Half of these 30 patients presented with anaemia.7 Another study from five general practices in Yorkshire reported positive serological results in 12 of 1200 subjects, a prevalence of 1%. In patients with irritable bowel syndrome, the figure was 3.3%; in those with iron-deficiency anaemia, 4.7%; and in those complaining of fatigue, it was 3.3%.8 In a more recent US study in primary care, all individuals with symptoms or conditions known to be associated with CD were tested: 30 of 976 patients had positive results of an anti-tTG test and 22 (2.25%) were diagnosed with CD. These studies support our recommendation of a low threshold for serological testing of patients with coeliac-associated conditions. In summary, CD is now recognised as a common disorder, with a population prevalence of nearly one in 100 people. In most people, its manifestations are mild or non-existent. In others, it is associated with symptoms such as fatigue, iron-deficiency anaemia, and vague abdominal complaints. Rarely, CD is associated with more significant problems, such as strange neurological symptoms, hyper- or hypothyroidism and hepatic dysfunction. The most appropriate strategy for general practitioners is to have a high level of suspicion when patients present with symptoms and signs that may be attributable to CD and to test these patients.

John M Duggan AM, MD, FRACP · Anne E Duggan MHP, PhD, FRACP

Objectives: To determine (i) the prevalence of positive results of anti-tissue transglutaminase (anti-tTG) antibody assays and coeliac disease (CD) in a rural Australian community; and (ii) whether confirmatory testing of a positive assay result with an alternative anti-tTG assay improved the positive predictive value of the test in population screening for CD.Design: Retrospective analysis in December 2004 of stored serum samples taken in 1994–1995 from 3011 subjects in the Busselton Health Study follow-up. Assays for IgA and IgG anti-tTG antibodies were performed, and positive or equivocal samples were retested with a different commercial anti-tTG assay. Available subjects with one or more positive assay results were interviewed, had serum collected for repeat anti-tTG assays and for HLA-DQ2 and HLA-DQ8 haplotyping and, if appropriate, gastroscopy and duodenal biopsy were performed. In unavailable subjects, HLA-DQ2 and -DQ8 haplotyping was performed on stored sera. Total serum IgA levels were assessed in subjects with initially negative assay results.Main outcome measure: Prevalence of anti-tTG positivity and biopsy-proven CD.Results: In 47 of 3011 serum samples (1.56%), at least one anti-tTG assay gave positive results: 31 of the subjects who provided these sera were available for clinical review, and 21 were able to have a gastroscopy. Seventeen subjects (0.56%) were diagnosed with definite CD (14 were confirmed at gastroscopy, and three unavailable subjects had three positive results of anti-tTG assays and an HLA haplotype consistent with CD); in a further 12 unavailable subjects, CD status was considered equivocal, with one or more positive anti-tTG assay results and an HLA haplotype consistent with CD. If these subjects were regarded as having CD, the prevalence of CD would be 0.96%. The positive predictive value when all three anti-tTG assays gave positive results was 94%, but fell to 45.2% with only one positive result.Conclusions: The prevalence of anti-tTG antibodies in this population is 1.56%; the prevalence of CD is at least 0.56%. The utility of a single, positive result of an anti-tTG assay in screening for CD in the community is poor, and repeat and/or collateral assessment with different assays may decrease the need for gastroscopy and distal duodenal biopsy.

Marcus W Chin MB BS, FRACP · Dominic F Mallon MB BS, FRACP, FRCPA · Digby J Cullen MB BS, FRACP · John K Olynyk MB BS, FRACP, MD · Lindsay C Mollison MB BS, MPH, FRACP · Callum B Pearce MD, MRCPI, FRACP

How can we ensure colonoscopy services are available to those who need them most? Bowel (colorectal) cancer is the most common cancer affecting both men and women in Australia, with 13 076 cases diagnosed and 4164 deaths reported in 2005.1 It is the second commonest cause of cancer-related death, behind lung cancer. The incidence of bowel cancer increases exponentially after 50 years of age, with a lifetime risk of about one in 17 among men and one in 26 among women.1 Bowel cancer satisfies most of the World Health Organization criteria for population cancer screening.2 Specifically, it is a common, serious cancer, and its natural history is reasonably well understood. It arises from precursor adenomas, and removal of these prevents cancer development. Importantly, most adenomas and early cancers are asymptomatic. Detection at early stages confers an excellent prognosis, and there are numerous tests for early detection and intervention, with the potential to reduce the incidence, morbidity and mortality of the disease. Cancer screening aims to identify affected individuals who do not suspect they have the disease. This is in contrast to performing diagnostic investigations for symptomatic patients, or targeting individuals with a significant family history of bowel cancer, a history of inflammatory bowel disease, previous adenomatous polyps or previous bowel cancer. However, bowel cancer screening recommendations can be confusing for medical practitioners. Numerous tests of varying performance levels are available, and it is difficult to separate recommendations for the population from those targeted towards individuals. There are also significant differences between the Australian recommendations endorsed by the National Health and Medical Research Council (NHMRC)3 and two recently published American guidelines, from the United States Preventive Services Task Force (USPSTF)4 and from a collaboration of the American Cancer Society, the US Multi-Society Task Force on Colorectal Cancer (representing the American Gastroenterological Association, American Society for Gastrointestinal Endoscopy, American College of Gastroenterology and American College of Physicians) and the American College of Radiology.5 Interestingly, these guidelines differ due to differing interpretations of essentially the same evidence. It is important to note that the newer American guidelines are less relevant to the Australian health care environment, and should not usurp the existing NHMRC recommendations in Australia. The NHMRC guidelines strongly recommend screening from the age of 50 years, by performing a faecal occult blood test (FOBT) at least every second year. The strength of this recommendation arises from three large, population-based, prospective randomised controlled trials demonstrating a mortality reduction of 15%–33%.3-5 No other cancer screening strategy is based on such strong evidence. Flexible sigmoidoscopy performed 5-yearly receives an equivocal recommendation, whereas colonoscopy and computed tomographic (CT) colonography are not recommended because of insufficient evidence. In contrast, the American guidelines recommend presenting information to patients, who then choose from the options, including high-sensitivity FOBT, flexible sigmoidoscopy or colonoscopy.4,5 The American recommendations diverge on CT colonography: the USPSTF does not recommend it on the basis of unknown long-term harm,4 while the collaborative joint guidelines endorse 5-yearly examinations.5 Unlike the US, Australia has moved beyond making passive recommendations to become one of a few countries actively implementing a nationwide population-based bowel cancer screening program — the National Bowel Cancer Screening Program (NBCSP) — although, currently, only individuals aged 50, 55 or 65 years are invited to participate (http://www.cancerscreening.gov.au/internet/screening/publishing.nsf/Content/bowel-about). The NBCSP, which uses a high-sensitivity immunochemical FOBT, began in 2006 after a pilot program (2002–2004) found that bowel cancer screening would be acceptable, feasible and cost-effective.6 The age restriction for invitees is part of a phasing-in process designed to enable resources to cope with increased downstream demand, especially for colonoscopy, which is the recommended investigation for participants with a positive FOBT result. About 7.5% of NBCSP participants have a positive FOBT result, and 5% of these are found to have cancer.7 Although the federal government has made no commitment to the NBCSP beyond June 2011, it is inconceivable that the program will simply be terminated, given the strong evidence basis for its function. The problem facing Australian medical practitioners is what to do in the face of the various disparate guidelines. A simple approach is to recommend participation in the NBCSP for all invitees. Increased awareness of bowel cancer will also cause some individuals outside the eligible ages to enquire about screening. In these cases, the NHMRC recommendations are most appropriate: for medical practitioners wishing to encourage opportunistic screening, an immunochemical FOBT can be recommended for patients aged between 50 and 75 years. Patients with positive FOBT results will then require follow-up colonoscopy. However, these recommendations ignore the growing tendency, often driven by patients, for screening with colonoscopy. In the absence of symptoms, about 500 colonoscopies need to be performed on 50–75-year-olds to identify one cancer,7 with an associated one in 1000 risk of serious complications.8 This contrasts with one cancer found for every 20 colonoscopies in the NBCSP.7 Of course, many patients will have premalignant adenomas that would be found with colonoscopy screening, which in turn creates a substantial requirement for long-term surveillance colonoscopies. Yet the vast majority of patients with adenomas will never develop bowel cancer, even without future intervention.9 Finally, the imperfections of colonoscopy are becoming increasingly recognised, reducing the ratio of its potential benefits to high cost.10,11 Thus, “indiscriminate” use of colonoscopy diverts availability of this expensive resource away from those most in need of it, especially in the public health system. As a step towards improving colonoscopy access, the Western Australian Department of Health will introduce a Colonoscopy Services Model of Care, scheduled for implementation over the next 2 years in the public hospital system, that prescribes appropriate use and referral processes. This will be supported by trained clinical staff and computerised referral systems to improve the quality and triaging of referrals in the public sector. Patients will also be informed of the priority level of their case, and the appropriate waiting time. Other states in Australia should strongly consider implementing similar strategies. Bowel cancer is common, serious and largely preventable. Medical practitioners should be encouraged to refer individuals for bowel cancer screening in compliance with the current Australian NHMRC recommendations and the NBCSP. Such practice will ensure that colonoscopy resources are available to those most in need in our community.

Hooi C Ee MB BS, FRACP, PhD · John K Olynyk MB BS, FRACP, MD

To the Editor: We read with great interest Drini and colleagues’ recent report of hepatosplenic T-cell lymphoma (HSTCL) associated with inflammatory bowel disease. The occurrence of this rare lymphoma is partly driving a move away from the use of combination thiopurine and anti-tumour necrosis factor alpha therapy. It is important to recognise that risk of HSTCL is not only associated with exposure to thiopurine drugs with or without infliximab. It is associated with immunocompromise in general, and also occurs de novo. Explaining relative risks to patients is never easy and needs to be balanced with the need for treatment in properly selected patients. Contrary to a comment by Drini and colleagues, HSTCL has been reported in association with exposure to adalimumab, another anti-tumour necrosis factor alpha agent.2 Two of these cases occurred in patients with prior exposure to thiopurines and infliximab, and the third occurred in a patient with rheumatoid arthritis who was not exposed to thiopurine or infliximab. There has been a recent spate of reports of HSTCL in patients with inflammatory bowel disease treated with immunomodulator therapy, but most HSTCL appears to occur in patients without exposure to immunosuppressants.3 Also, although cases predominantly occur in young men (under 22 years), HSTCL is not limited to the paediatric age group.3,4 The presentation also occurs in females, including during pregnancy. More generally, in patients with inflammatory bowel disease, treatment with thiopurines appears to increase the risk of lymphoma two- to fourfold.5 Assuming a fourfold increase in risk, this translates into one additional lymphoma per year for every 4357 patients treated with thiopurines at age 20–29 years.5 This should be compared with the lifetime attributable risk of death from any cancer due to a single abdominal computed tomography scan performed at the age of 20, which is in the order of 1 : 2000.6 Currently, it is difficult to completely disentangle the many different factors — including patient age; severity, duration and course of disease;7 type of immunosuppression; and exposure to radiation — when considering the risk of lymphoma in inflammatory bowel disease. However, in many patients with severe disease, this risk appears to be outweighed by the benefits of adequate medical control.

Daniel C Burger · Timothy H J Florin

No drug is completely safe and, while the risks seem small, some side effects can be serious When I went to medical school, the mantra “know thy drugs” was pounded into me repeatedly, and it has served me well. Among the most commonly prescribed drugs in Australia are the proton-pump inhibitors (PPIs). In 2005, one million Australians were dispensed drugs in this class.1 There is no doubt that PPIs are safe relative to most other medications we prescribe, but suppressing gastric acid is not physiological.2 A low level of gastric acid promotes the growth of swallowed and enteric flora in the proximal gut, and these bacteria may be aspirated during episodes of physiological reflux. In this issue of the Journal, Roughead and colleagues assess the risk of pneumonia in Australian veterans taking PPIs (Roughead); they identified a 16% increase in risk, equating to four extra hospitalisations for pneumonia each year for every 1000 people prescribed a PPI.1 A study from Denmark found current use of PPIs was associated with a 1.5-fold (or 50%) increase in the risk of community-acquired pneumonia (95% CI, 1.3–1.7); the attributable proportion (ie, the fraction of pneumonia potentially caused by PPIs) was calculated to be 4%.3 Similarly, a Netherlands study reported an adjusted relative risk for pneumonia among those using PPIs versus those who stopped using the drugs of 1.89 (95% CI, 1.36–2.62).4 There is, therefore, a small increased risk of pneumonia, and this may be further increased in those who have recently begun taking PPIs. However, no preventive strategies are available to reduce this risk, and a causal association has not been established. Indeed, the increase could still be caused by unidentified confounders. For example, many patients in these studies had multiple comorbid conditions that might have given rise to an increased risk of pneumonia in the first place, such as alcoholism, previous stroke, or immune compromise secondary to chronic disease.3,4 A case–control study reported a significantly increased risk of Clostridium difficile infection in those who were exposed to PPIs in the 90 days before the infection, with an odds ratio (OR) of 3.5 (95% CI, 2.3–5.2).5 As expected, previous exposure to antibiotics was also a significant risk factor for C. difficile infection in this study. An association with H2-receptor antagonists, as well as non-steroidal anti-inflammatory drugs (but not aspirin), was also reported.5 An increased risk of other enteric infections has similarly been observed with PPI use (OR, 2.55; 95% CI, 1.53–4.26), although combining data in this meta-analysis was problematic because of significant study heterogeneity.6 Whether other comorbid conditions in patients taking these drugs account for the association remains to be clarified, and any causal link is still speculative. An acid environment is needed for insoluble calcium absorption, but the effects of PPI on dietary calcium absorption are uncertain. Increasing dietary calcium intake and taking calcium citrate as a supplement (which does not require gastric acid for absorption) is not established practice for patients who are prescribed acid-suppressing medications. A large retrospective study from the United Kingdom General Practice Research Database suggested that there was an increased risk of hip fracture in patients taking PPIs, with an adjusted odds ratio of 1.44 (95% CI, 1.30–1.59).7 Importantly, the calculated excess risk was small (about 1263 patients over the age of 50 would need to be treated with PPIs for a year to identify one excess hip fracture). A weaker effect was seen with H2-receptor antagonists. A recent Canadian case–control study in an administrative claims database examined longer-term PPI use and osteoporosis-related fractures; exposure of 7 years or more was modestly associated with an increased fracture risk (OR, 1.92; 95% CI, 1.16–3.18).8 A causal association has not been established, but it may be worth considering increasing calcium intake for prevention and assessing bone mineral density in those requiring a daily PPI for more than 5 years.9 Carpopedal spasm secondary to low serum magnesium and calcium levels that reversed with withdrawal of PPI therapy in two patients has also been reported.10 Iron absorption does not appear to be affected by PPI use, but long-term acid suppression has been linked to malabsorption of vitamin B12, especially in older people.11 Therefore, it may be reasonable to assess vitamin B12 levels annually in older patients requiring maintenance PPI therapy, but this is not routine practice, as the cost versus the benefit is unknown. Patients infected with Helicobacter pylori and who are taking long-term PPI therapy are at increased risk of developing gastric atrophy, but the exact clinical significance remains uncertain.12,13 Gastrin release increases with PPI therapy because of acid suppression, but there is no evidence this leads to neoplastic changes in the stomach. One study reported that 30% of patients taking omeprazole developed gastric atrophy, although this appeared to occur primarily in those concurrently infected with H. pylori.13 However, the development of gastric body intestinal metaplasia is rare, and there is no definite evidence that long-term maintenance PPI therapy in the setting of H. pylori induces dysplasia or gastric cancer. Despite this, some authorities do recommend screening for H. pylori infection if long-term PPI use is contemplated, and offering eradication therapy to those infected, and this is my practice too.14 There appears to be no increased risk of colon cancer in PPI users, which some have speculated could occur secondarily to hypergastrinaemia.15 Finally, an increasing number of cases of acute interstitial nephritis are being reported in association with PPI therapy, and this appears to be an idiosyncratic class effect; sadly, not all patients recover kidney function when they stop taking the drug.16 The bottom line is that no drug is completely safe, and this applies to acid suppression therapy. Fortunately, the risks, if causal, seem small, although preventive strategies are largely unavailable and identifying those at particularly high risk of serious side effects (eg, based on pharmacogenomics to individualise therapy) is not yet an established strategy. However, it is prudent and best practice to warn patients about the potential serious (albeit rare) side effects of PPIs, to prescribe the lowest possible dose of PPI (when indicated) for as short a time as possible, and to consider alternative management options if these are available.

Nicholas J Talley MD, PhD, FRACP