440 Gut 2000;46:440–442

ing and subsequent regular surveillance 15 Orme SM, McNally RJ, Cartwright RA, et al. above that of the normal population cannot Mortality and cancer incidence in acromegaly: a retrospective cohort study. United Kingdom LETTERS TO be supported by the evidence currently avail- Acromegaly Study Group. J Clin Endocrinol able. Metab 1998;83:2730–4. Gut: first published as 10.1136/gut.46.3.440d on 1 March 2000. Downloaded from THE EDITOR A G RENEHAN STO’DWYER Reply Department of Surgery, Christie Hospital NHS Trust, Wilmslow Road, Withington, EDITOR,—Our conclusion that acromegaly Colorectal neoplasia in acromegaly: the Manchester M20 4BX, UK may be a high risk condition for the develop- ment of colorectal cancer is based not only on reported increased prevalence is S M SHALET our own data (we have now discovered 10 overestimated Department of Endocrinology, Christie Hospital NHS Trust patients with cancer from approximately 210 patients who have had a colonoscopy), but EDITOR,—We read with interest a recent also on those of several other studies. In the paper by Jenkins et al (Gut 1999;44:585– 1 Jenkins PJ, Fairclough PD, Richards T, et al. prospective studies by Archambeaud- 587). However, we are concerned with their Acromegaly, colonic polyps and carcinoma. Mouveroux and colleagues1 and Ituarte ,2 assertion that acromegaly is a high risk Clin Endocrinol (Oxf) 1997;47:17–22. et al 1a Atkin WS, Morson BC, Cuzick J. Long-term cancers were discovered in 12.5% and 20% condition for colorectal neoplasia, and their risk of colorectal cancer after excision of respectively of patients with acromegaly. Ret- recommended advice on colonoscopic rectosigmoid adenomas. N Engl J Med 1992; rospective epidemiological surveys may have screening and surveillance. Jenkins and col- 326:700–2. 2 Correa P, Strong JP, Reif A, . The epidemi- failed to show an increased risk of cancer leagues had found that 33 (26%) of 129 et al ology of colorectal polyps: prevalence in New because of diVerences in methodology—for patients (updated to 155), treated for ac- Orleans and international comparisons. Cancer example, in one study case ascertainment romegaly at St Bartholomew’s Hospital, had 1977;39:2258–64. 3 Eide TJ, Stalsberg H. Polyps of the large depended upon death certificate entries and at least one adenoma and six (5%) had intestine in Northern Norway. 1978; : 1 Cancer 42 cancer registrations which may have been adenocarcinomas. 2839–48. incomplete.3 Furthermore, these studies did 4 Williams AR, Balasooriya BA, Day DW. Polyps We feel that the choice of controls in this not discuss the relevance of the age of the study was inappropriate because although and cancer of the large bowel: a necropsy study in Liverpool. Gut 1982;23:835–42. patients. The mean and range of age during there is no ideal control population, the 5 Blatt LJ. Polyps of the colon and rectum: follow up were not stated,3 and in another authors used comparative data on the inci- incidence in northern Norway. Dis Colon study the mean age of the patients at diagno- dence of adenomatous polyps from only two Rectum 1961;4:277–82. 6 Arminski TC, McLean DW. Incidence and dis- sis of acromegaly was 52 years old and that at cohorts: a published study of left sided tribution of adenomatous polyps of the colon 4 1a follow up was only 61 years old. Our results adenomas, and colonoscopic records of all and rectum based on 1000 autopsy examina- clearly show that colorectal cancer is a late patients without acromegaly that had been tions. Dis Colon Rectum 1964;7:249–61. 7 Stemmermann GN, Yatani R. Diverticulosis complication of acromegaly, as the mean age examined by one of the authors. Matched for and polyps of the large intestine. A necropsy of our aVected patients was 67 years old. age (and side), the relative risk of adenomas study of Hawaii Japanese. Cancer 1973;31: The situation for adenomas is less clear and was higher in patients with acromegaly when 1260–70. 8 Rickert RR, Auerbach O, Garfinkel L, et al. we agree that there is a lack of proper control compared with data from the first study, but Adenomatous lesions of the large bowel: an groups. However, many of the prevalence fig- not when compared with data from the autopsy survey. Cancer 1979;43:1847–57. ures given by Renehan and colleagues were second. 9 Vatn MH, Stalsberg H. The prevalence of obtained from necropsy studies and therefore In an attempt to estimate more appropri- polyps of the large intestine in Oslo: an autopsy study. Cancer 1982;49:819–25. cannot provide a valid comparison because ately the prevalence of adenomas in the nor- 10 Jass JR, Young PJ, Robinson EM. Predictors of the resected bowel was thoroughly washed on mal population, we have carried out a presence, multiplicity, size and dysplasia of up to three occasions, repeatedly examined http://gut.bmj.com/ comprehensive review of the literature on colorectal adenomas. A necropsy study in New under magnification in optimal lighting, and adenoma prevalence per decade of life from Zealand. Gut 1992;33:1508–14. 11 DiSario JA, Foutch PG, Mai HD, et al. lesions as small as l mm were classified as which two groups of studies emerged. The Prevalence and malignant potential of colorec- adenomas. This gives an increased prevalence first group comprised six necropsy studies tal polyps in asymptomatic, average-risk men. of the disease compared with incidences of (n=2914), and the second comprised three Am J Gastroenterol 1991;86:941–5. 12 Lieberman DA, Smith FW. Screening for colon neoplasia revealed by colonoscopic screen- screening colonoscopy studies of asympto- malignancy with colonoscopy. Am J Gastroen- ing. Furthermore, these studies were of matic average risk volunteers (n=720, table terol 1991;86:946–51. populations with very diVerent demographic, 2–13 1). With the exception of those patients 13 Rex DK, Lehman GA, Ulbright TM, et al. socioeconomic and dietary influences, which

Colonic neoplasia in asymptomatic persons on September 30, 2021 by guest. Protected copyright. between 50–59 years old, the prevalence rates with negative fecal occult blood tests: influence are factors known to influence the incidence of adenomas in patients with acromegaly are of age, gender, and family history. Am J Gastro- and prevalence of colonic adenomas. By con- remarkably similar to those for individuals in enterol 1993;88:825–31. trast, our control groups were taken from screening colonoscopy studies, and less than 14 Ron E, Gridley G, Hrubec Z, et al. Acromegaly and gastrointestinal cancer. Cancer 1991;68: similar populations to the patients with those from necropsy studies. We find no evi- 1673–7. acromegaly, and in one group, the colono- dence that patients with acromegaly are at increased risk of developing adenomas. For colorectal cancer, Jenkins and Table 1 Prevalence (%) of adenomatous polyps by decade of life colleagues1 used comparative data from a regional cancer registry and estimated in- Age group (years) creased relative risks of 13–90. These figures n <40 40–49 50–59 60–69 70+ Total are exaggerated in comparison with studies using standardised age and sex adjusted Autopsy studies2–4* population data. Ron and colleagues14 re- Blatt5 446 0 21 34 34 44 39 ported 13 colonic cancers in 1041 male Arminski and McLean6 1000 20 19 27 35 46 33 7 veteran acromegalics (standardised incidence Stemmermann and Yatani 202 43 – 72 59 63 62 Rickert and colleagues8 518 – 17 35 56 58 47 ratio (SIR) 3.1; 95% CI 1.7 to 5.1) and, using Vatn and Stalsberg9 445 – 6 23 31 46 33 uniform methods of ascertainment of cases Jass and colleagues10 303 3 11 33 30 36 24 and comparison groups, Orme and 15 Colonoscopy studies in screened populations colleagues found 12 cases of colonic cancer DiSario and colleagues11 119 – – 21 45 53 41 in a larger study of 1362 patients with Lieberman and Smith12 105 – – 28 41 58 41 acromegaly (SIR 1.68; p=0.06). Rex and colleagues13 496 – – 20 33† 31‡ 26 There are approximately 1500 patients Weighted averages with acromegaly in the United Kingdom, and Necropsy studies 2914 9 15 31 40 48 37 it would seem sensible for strategies for large Screening studies 720 – – 20 38 37 31 bowel screening to be evidence based. The v 1 data given above suggest that the reported Acromegalics (Jenkins and colleagues ) 129 8 12 29 36 39 26 increased prevalence of colorectal neoplasia *Three other autopsy studies (Correa and colleagues2; Eide and colleagues3; Williams and colleagues4) were in patients with acromegaly is overestimated considered but not included as the age bands in these studies did not correspond with those used by Jenkins and thus, the recommendations given by et al.1 these authors for early colonoscopic screen- †Patients aged 60–64; ‡patients aged 65–75. Letters, Book reviews, Notes 441 scopies were performed by the same operator. stump appeared mildly inflamed. Biopsy and/or measurement of the portocaval pres- A comparison between these groups showed samples of the rectal stump were consistent sure gradient, together with imaging of the a significant increase in the relative risk of with active colitis and a putative diagnosis of portal venous system in order to exclude por- adenomas in patients with acromegaly, al- diversion colitis was made. Treatment with tal or splenic vein thrombosis. The Gut: first published as 10.1136/gut.46.3.440d on 1 March 2000. Downloaded from though we accept that this risk is not as high steroid enemas was eVective. Six months later importance of this is illustrated by the series as that for cancer, and this raises intriguing the patient developed bleeding and mucus of patients with GAVE after bone marrow questions about the causes of colorectal can- discharge through the colostomy for the first transplantation, cited by Spahr et al.2 In this cer in patients with acromegaly. It is possible time. Stool cultures including analysis for series, all patients with available liver histol- that the adenoma–carcinoma sequence in Clostridium diYcile were negative. Colonos- ogy were found to have hepatic veno- patients with acromegaly diVers from that in copy through the colostomy to the limit of the occlusive disease, a well recognised cause of the non-acromegalic population, or that the examination revealed severe, active colitis. portal hypertension. A further link between cancers arise de novo without an adenoma- The biopsy samples on this occasion were GAVE and autoimmune disorders may be tous stage. consistent with ulcerative colitis with chronic explained by extrahepatic or non-cirrhotic We fully concur that strategies for large inflammation in the lamina propria, cryptitis portal hypertension in some of those 13 bowel screening should be evidence based. and crypt abscess formation. The patient patients. The possible association of GAVE Our initial screening recommendations are responded well to treatment with oral with chronic renal failure cited by Spahr et al based on current data and will be modified corticosteroids and mesalazine. presumably refers to an early series of according to our continuing prospective We have postulated previously that the patients with diVuse haemorrhagic gastric studies. Subsequent strategies might take into development of ulcerative colitis may be initi- lesions in the absence of any overt liver account not only the age of the patients but ated by inflammation of another aetiology at disease; however, these lesions would prob- also the activity of their acromegaly, because an anatomically discontinuous site of the ably not be classified as GAVE by current cri- preliminary data suggest that those patients bowel. This description of an additional case teria, therefore this association is 14 in whom the disease is more active (elevated would support the hypothesis that diversion questionable. serum IGF-I) are more likely to develop colitis might be one such trigger for ulcerative Thus, the pathogenesis of this interesting adenomas. Until more multicentre studies colitis. The mechanisms underlying this disorder remains uncertain but the strong involving larger numbers of patients undergo- remain speculative but may involve the association of portal hypertension (whether ing careful and total colonoscopy allow the recruitment of activated lymphocytes from overt or covert) with the majority of cases of risks to be better quantified, it seems prudent the diverted colon to the phenotypically simi- GAVE means that this is likely to be a key for patients to undergo colonoscopic screen- lar vascular endothelium of the instream contributory factor. ing every five years, or every three years if an colon. N C FISHER adenoma is found. Specialist Registrar in Gastroenterology, AGLIM Manor Hospital, P JENKINS WLIM Walsall, P FAIRCLOUGH Department of Gastroenterology, M BESSER West Midlands WS2 9PS, UK Epsom General Hospital, Departments of Endocrinology and Gastroenterology, Epsom, Surrey KT19 7EG, UK St Bartholomew’s Hospital, West Smithfield, 1 Gostout CJ, Viggiano TR, Ahlquist DA, et al. London EC1A 7BE, UK The clinical and endoscopic spectrum of the Gastric antral vascular ectasia and its watermelon stomach. J Clin Gastroenterol relation with portal hypertension 1992;15:256–63. 2 Tobin RW, Hackman RC, Kimmey MB, et al. 1 Archambeaud-Mouveroux F, GeVray I, Teissier Bleeding from antral vascular ectasia in mar- MP, et al. Prevalence of colorectal carcinoma EDITOR,—Spahr and colleagues recently pub- row transplant recipients. Gastrointest Endosc and polyps in acromegaly [abstract]. 80th lished a case series which described the poor 1996;44:223–9. Annual meeting of The Endocrine Society; response of a haemorrhage from gastric antral 3 Webb LJ, Sherlock S. The aetiology, presenta- http://gut.bmj.com/ 1998 June 2427; New Orleans. Bethesda, MD: tion and natural history of extrahepatic portal The Endocrine Society, 1998:504. vascular ectasia (GAVE) to portal decom- venous obstruction. QJMed1979;48:627–39. 2 Ituarte EM, Petrini J, Hershman JM. Acrome- pression by insertion of a transjugular intra- 4 Cunningham JT. Gastric telangiectasias in galy and colon cancer. Ann Intern Med hepatic portosystemic shunt (TIPS) (Gut chronic haemodialysis patients: a report of six 1984;101:627–8. cases. Gastroenterology 1981;81:1131–3. 3 Orme SM, McNally RJ, Cartwright RA, et al. 1999;44:739–742). However, the authors’ Mortality and cancer incidence in acromegaly: claim that this indicates the absence of a rela- a retrospective cohort study. United Kingdom tion between GAVE and portal hypertension Reply Acromegaly Study Group. J Clin Endocrinol is seriously flawed. The failure of this condi- Metab 1998;83:2730–4.

tion to respond to portal decompression can- on September 30, 2021 by guest. Protected copyright. 4 Ron E, Gridley G, Hrubec Z, et al. Acromegaly EDITOR,—We appreciate Dr Fisher’s com- and gastrointestinal cancer. Cancer 1991;68: not exclude a primary role for portal 1673–7. ments on our recent paper, in which we pro- hypertension in the pathogenesis of the vided evidence against the role of portal disorder. Furthermore, this study would have hypertension in the pathogenesis of GAVE. Diversion colitis: a trigger for ulcerative been more informative if aortoportography— Firstly, lowering or normalisation of portal that is, superior mesenteric and splenic colitis in the instream colon pressure was not followed by improvement in angiography with venous phase imaging, had either the endoscopic findings or the rate of EDITOR,—We published in 1999, to our been performed in order to determine the transfusions needed for recurrent bleeding. knowledge, the first description of diversion pattern of portosystemic shunting before and In this case series, one patient had to be colitis appearing to trigger instream ulcera- after TIPS insertion. The authors have not transfused repeatedly for five years despite a tive colitis (Gut 1999;44:279–282). We have addressed the question of whether portosys- patent surgical end to side portocaval shunt recently encountered a fourth patient with temic shunts supply the vascular lesions of (portocaval gradient 2 mm Hg). Further- the same rare condition. GAVE; it would have been helpful if they had more, the degree of residual portal hyper- A 66 year old non-smoking white man pre- performed angiography in at least some tension was not correlated with clinical and sented in 1997 with large bowel obstruction. patients after the TIPS procedure to confirm endoscopic evolution in patients treated by His mother had died of colorectal carcinoma that blood flow had been restored to the liver TIPS. Interestingly, the only patient that at the age of 71 years. There was no family from all portal vein tributaries, as it cannot be responded to TIPS still had an increased history of inflammatory bowel disease. An assumed that normalisation of portal vein gradient after the treatment (14 mm Hg); in instant barium enema revealed obstruction at pressure will completely ablate all preformed this patient, the favourable outcome of GAVE the level of the mid-sigmoid colon with portosystemic shunts. Finally, it was not was parallelled by a noticeable improvement radiological features suggestive of sigmoid stated whether splenic vein thrombosis was in liver function. carcinoma. The patient underwent a Hart- excluded in all patients. If present, this could One study has suggested that arteriopor- mann’s procedure with colostomy. The resec- have caused ongoing segmental portal hyper- tography may have a diagnostic value in tion specimen showed a Dukes’s B carcinoma tension which could not have been expected GAVE.1 However, typical findings were of the sigmoid colon with no evidence of coli- to respond to TIPS insertion. shown on the arterial phase (hypervasculari- tis. The patient then received adjuvant It is evident from previous case series, sation of the antrum and early arteriovenous chemotherapy consisting of 12 weeks of infu- which suggested that GAVE might occur shunting), and none of our patients had a sional 5-fluorouracil. Eighteen months later without portal hypertension, that some did coeliac axis arteriogram. On direct portogra- he had a colonoscopy through the rectal not exclude portal hypertension in their phy, it was impossible to show dilated stump and the stoma. The colon proximal to patients.1 A rigorous exclusion of portal mucosal blood vessels in the antrum. In all the colostomy looked normal but the rectal hypertension would require liver biopsy non-responders, bleeding recurred despite a 442 Letters, Book reviews, Notes reversed portal blood flow observed after cancer is receiving due recognition from the that can walk upright, do joined up writing TIPS or surgical shunt. In addition, splenic Government, primary care physicians, man- and has a medical degree. The format is lists vein thrombosis was not observed in any of agers, and the general public. Attention is of key points, many of which are expanded to our patients, obviously because such a finding upon us and our daily work load is ever short paragraphs. Topics are presented alpha- Gut: first published as 10.1136/gut.46.3.440d on 1 March 2000. Downloaded from would contraindicate TIPS, which is not a increasing. How encouraging that at a time betically which leads to a strange contents treatment of segmental portal hypertension. like this the arms of colorectal surgeons from page with Nausea and vomiting coming Therefore, we are still convinced that porto- Penzance to John O’Groats have been or are before Neuroendocrine tumours and after systemic shunting and liver function could soon to be strengthened by the colorectal Menetrier’s disease, but that is a quirk of the both influence liver metabolism of the nurse specialist. This book will be very useful Key Topics series and of little importance. All vasodilating substances that contribute to the to them. the major areas of gastroenterology are pathogenesis of GAVE, whereas portal hyper- Who else will find this book useful? Well covered and although I wouldn’t agree with tension alone has no influence. certainly medical students, trainees in gastro- several of the statements made, these areas G POMIER-LAYRARGUES enterology and surgery, and our general prac- are covered well. The section on small bowel Liver Unit, titioner colleagues. And dare I also suggest transplantation lacks any great substance, Hôpital Suint-Luc, gastroenterologists and colorectal surgeons. however, though it does at least serve to high- 1058, rue Saint-Denis, Not because we’d learn anything of course. light the fact that such work is being carried Montreal, Actually, that’s rubbish because I was full of out. So, who should buy it? Quebec H2X 3J4, Canada. the new facts I learnt over this last weekend Dr Anderson has used up to date refer- and I was educating everyone in theatre this ences throughout, but as he and his collabo- 1 Robertson IR, Tait PN, Jackson JE. Vascular Monday afternoon! I definitely will be using rators point out, this book is not intended to ectasia of the gastric antrum: angiographic this book in preparing for both undergradu- be a complete textbook on gastroenterology. findings. AJR Am J Roentgenol 1996;166:87–9. ate and postgraduate teaching. If you need a reference book, gastroenterolo- gist or general physician alike, this is not for A LEATHER you. Go and look it up on the Internet. If you are a gastroenterology house-oYcer, SHO or BOOK REVIEWS Key Topics in Gastroenterology. Edited registrar consider it, although you should by Anderson SHC, Davies G, Dalton HR. look around first, particularly at the Little Brown series. If you are a non-GI SHO you’re (Pp 286; £21.95.) Oxford: Bios Scientific probably better oV buying a cardiology or Publishers, 1999. ISBN 1859962815. neurology book. The case for buying this 2nd edn. ABC of Colorectal Diseases. book would be stronger if it provided more Edited by Jones DJ. (Pp 109; illustrated; Simon Anderson has put a lot of work into definite clinical paradigms for how to ap- £18.95.) London: BMJ Books, 1999. ISBN this book. He writes clearly, concisely and proach the GI patient. £21 is still a lot of 0-7279-1105-8. well. But, does the world need another hand- money for something that doesn’t provide book on gastroenterology? immediate gratification; however, in these The series of ABC articles in the is What is the role of professional books, large BMJ days when even the is buying books for always enjoyable. This book brings together or small, nowadays? Down at your local Sun schools, if pharmaceutical companies articles that were published several years ago bookshop, Delia Smith will tell you how to changed their mission from “chicken korma and were well received in the first edition of make an omelette, Charlie Dimmock how to for all by 2000” to something of more tangi- 1993. These articles have been updated and build that water feature; there’s no real alter- ble benefit, perhaps this book would reach a eight additional chapters have been added on native. So, in the hospital who will tell you constipation, diarrhoea, irritable bowel syn- about haemachromatosis? The professional wider audience as it should. http://gut.bmj.com/ drome, inflammatory bowel disease, anal book has always held pride of place as the font J MEENAN cancer, colorectal trauma, tropical colonic of all knowledge. Owning such book gave diseases, and paediatric problems. peace of mind, but this was at a hefty price The editor (DJ Jones) has turned vision and that has always been the problem. How into reality to produce a basic and useful often will people keep paying the ransom for coloproctology text. He has also worked hard the latest edition, or the newest series? It is NOTES as he has written half of the chapters (nine not surprising that cutting edge books on solo and three shared). The authors of the cimetidine still adorn many bookshelves. first chapter on Anatomy and Physiology of Time to mention the “I” word. the Colon, Rectum and Anus are Mr Hill and If anything, the problem currently for on September 30, 2021 by guest. Protected copyright. Professor Irving: the former runs the im- professional books is that knowledge is cheap, Second Annual Gastrointestinal mensely successful M62 Coloproctology plentiful and easily obtained. Case reviews, Course and the latter needs little introduction subject overviews and clinical slides are pain- Cancer Update: A Multidisciplinary and is President of the Association of fully easy to obtain through the Internet. Approach Coloproctology of Great Britain and Ireland. They are free and bang up to date. A 1999 Not surprisingly they combine forces to pro- review on haemachromatosis can be found The Second Annual Gastrointestinal duce three pages with all the basic facts and and printed oV within 15 minutes. Game set Cancer Update conference will be held at 10 superb illustrations. and match to the Internet, Woolworth the Yarrow Hotel and Conference Centre, Chapter 2 takes the reader through exam- vouchers to the poor runner up, books. Not Park City, Utah, USA, on 15–19 March ination (I appreciated time spent on how to quite. Whereas books are concise, tidy, 2000. Further information from: Rosalie perform a rectal examination) and tests from handy, and look good, the Internet is Lammle. Tel: +1 801 581 8664; fax: +1 proctoscopy to colonic transit studies. Chap- disparate and its out-pourings are a mess. 801 581 3647; email: rosalie.lammle@ ters 3–23 then cover all aspects of colorectal Books have editors, the Internet doesn’t. hsc.utah.edu disease before the final chapter on drugs. Today, to be worth buying, a book must Each chapter is concise (3–6 pages) and address either a very specialised audience or a packed with really good photographs and very broad one. A book must gather together European Courses on Laparoscopic illustrations. All credit to the various depart- information that is diYcult to obtain else- Surgery ments of medical illustration involved in this where in a coherent form (e.g. clinical classi- venture! fications and research scoring systems, etc.), The European Courses on Laparoscopic For me the prize chapter is on inflamma- or that is timeless (clinical manifestations). Surgery will be held at the University tory bowel disease: the authors (Mr Scott and The Key Topics series does reflect some of Hospital Saint Pierre, Brussels, Belgium, Professor Thompson) bridge the gap be- these qualities, though weakly. on 4–7 April 2000 and 21–24 November tween a presentation of the vital facts and the The stated church for Key Topics in Gastro- 2000. Further information from: Confer- thorny management issues so eloquently. enterology is senior house oYcers (SHO) and ence Services S.A., Drève des Tumuli, 18, Who should read this book? Let us first registrars preparing for the MRCP and all B-1170 Brussels, Belgium. Tel:+32 2 375 consider the exciting and challenging times doctors wishing to keep up to date—the usual 1648; fax: +32 2 375 3299; email: that we are living in. At long last colorectal publishers’ non-specific audience of anyone [email protected]