1569
Letters
LETTERS TO THE EDITOR Serum and tissue autoantibodies to colonic epitheium in ulcerative colitis SIR,-I read with interest and disappointment the above article by Snook et al which appeared recently.' The authors confirmed previous reports from us and others regarding the presence of cytotoxic serum antibodies in a proportion of patients with inflammatory bowel disease, especially those with active disease, except that their percentage of positivity was much lower (20%) than all the other published reports (which were about 50% or more) and most frequent in ulcerative colitis.23 One reason for this could be the use ofdifferent target cells - for example in this study the authors used HT-29 colon cancer cells compared with other studies (including ours) where RPMI-4788 colon cancer cells or normal colon epithelial cells were used as targets. Indeed, this is important because one of the target molecules on the colon cells recognised by ulcerative colitis serum antibody in cytotoxicity assay is associated with the Mr 40K protein' which acts as an autoantigen in ulcerative colitis.56 Monoclonal and polyclonal antiMr 40K antibodies block the cytotoxicity of ulcerative colitis sera on RPMI-4788 and DLD- 1 colon cancer cells,47 showing the role of Mr 40K protein in this recognition. While normal colon epithelial cells, RPMI-4788, and DLD- 1 colon cancer cells express Mr 40K protein, several colon cancer cell lines, including HT-29, which was used by Snook et al, did not express detectable amounts of the Mr 40K protein. Therefore, the use of normal colon cells or selection of appropriate colon cancer target cells is very important for the cytotoxicity study. On the basis of the cytotoxicity data using a single cell line - that is HT-29 - it is indeed misleading for Snook et al to conclude that 'lack of association with disease, extent and activity . . . lack of cytotoxic activity all strongly suggest this antibody is merely an
epiphenomenon.' Using 'ulcerative colitis (UC)-colon extracted IgG (CCA-IgG)', these authors reported completely negative results in their 'functional studies' such as 'cytoxicity assay' using HT-29 target cells and 'immunohistochemical staining' of colon tissue. These results are in contrast to a number of our independent reports"' and the results of many other investi-
gators."'2 There are several major problems with the study of Snook et al. Firstly, the authors did not provide any evidence of intact immunoglobulin or immunoreactive Fab' fragments being present in their ulcerative colitis colon extracted materials ('CCA-IgGs'). The presence of IgG reactivity was reported only with ELISA. Fragmented IgG and Fc fragments will react in the ELISA as performed by them, but will be functionally inactive in the assays they have used. Indeed, six of the 'CCA-IgG' preparations prepared by these authors were examined by us for Ig analysis and immunoreactivity to tissue antigen(s). None of the samples had detectable intact IgG when we analysed them several times by SDS-polyacrylamide gel electrophoresis. This was communicated to the authors well before the publication of the article.
Secondly, as mentioned above, HT-29 cells do not seem to be the right target cells for the cytotoxicity study as they do not express any detectable amount of Mr 40K protein. Therefore, the 'negative functional studies' can easily be explained by the fact that the 'CCA-IgG' extracted by the authors are most likely fragmented by proteases, which are plentiful in colon particularly in ulcerative colitis. Such fragmentations are common if the extraction procedure is performed at room temperature. A striking increase of local IgG production in ulcerative colitis,9' with reactivity to colonic epithelial antigen(s)" have been reported by many investigators, besides our several independent studies.5`8 The immunoreactive colon antigen, Mr 40K protein has been analysed by us using two sensitive techniques, by immunotransblot analysis"7 and by immunocytochemistry using monoclonal antibody.8 Recently, Trond Halstensen et al'2 beautifully demonstrated by two colour immunofluorescence technique the presence of colon epithelial bound IgG (IgG, subtype) and also epithelial deposition of activated complement products (C3b and terminal complement complex) in active ulcerative colitis. All these studies indicate the presence of autoantibodies against epithelial cellular protein(s) in ulcerative colitis which might play an important role in the pathogenesis. While we share the 'disappointment' of the authors for their 'uniform failure in their efforts of extraction of functionally active tissue autoantibodies' in ulcerative colitis, their conclusion of 'no evidence of extractable, epithelial-bound autoantibody' is unfounded. KIRON M DAS Division of Gastroenterology and Hepatology, Department of Medicine, University of Medicine and Dentistry of New3jersey, Robert Wood Johnson Medical School, New Brunswick, NJ 08903-0019, USA 1 Snook JA, Lowes JR, Wu KC, Priddle JD, Jewell DP. Serum and tissue autoantibodies to colonic epithelium in ulcerative colitis. Gut 1991; 32:
163-6. 2 Nagai T, Das KM. Demonstration of an assay for specific cytolytic antibody in sera from patients with ulcerative colitis. Gastroenterology 1981; 80: 1507-12. 3 Auer 10, Grosch L, Hardorfer C, Roder A. Ulcerative colitis specific cytotoxic IgG-autoantibodies against colonic epithelial cancer cells. Gut 1988; 29: 1639-47. 4 Biancone L, Ebert EC, Das KM. The Mr K40,000 colonic protein is associated with antibody dependent cell-mediated cytolysis (ADCC) by ulcerative colitis (UC) sera. Gastroenterology 1990; 98(5): A159. 5 Takahasi F, Das KM. Isolation and characterization of a colonic autoantigen specifically recognized by colon tissue-bound IgG from idiopathic ulcerative colitis. J Clin Invest 1985; 76: 311-8. 6 Takahasi F, Shah H, Wise L, Das KM. Circulating antibodies against human colonic extract enriched with a 40 kDa protein in patients with ulcerative colitis. Gut 1990; 31(9): 1016-20. 7 Das KM, Sakamaki S, Vecchi M. Ulcerative colitis: specific antibodies against a colonic epithelial Mr 40,000 protein. Immunological Investigations 1989; 18: 459-72. 8 Das KM, Vecchi M, Sakamaki S, Diamond B. The production and characterization of monoclonal antibodies to a human colonic antigen associated with ulcerative colitis: Cellular local-
ization of the antigen using the monoclonal antibody. J Immunol 1987; 139: 77-84. 9 Kett K, Rogmem TO, Brandtzaeg P. Mucosal subclass distribution of IgG producing cells is different in ulcerative colitis and Crohn's disease of the colon. Gastroenterology 1987; 93: 919-24. 10 Scott MG, Nagm MH, Macke K, Nash GS, Bertovich MJ, MacDermott RP. Spontaneous secretion of IgG subclasses by intestinal mononuclear cells: differences between ulcerative
colitis, Crohn's Disease, and controls. Clin Exp Immunol 1986; 66: 209-15. 11 Hibi T, Aiso M, Ishikawa M, Watanabe M, Yoshida T, Kobayashi K, Asakura H, Tsuru S, Tsuchiya M. Circulating antibodies to the surface antigens on colonic epithelial cells in ulcerative colitis. Clin Exp Immunol 1983; 54: 163-8. 12 Halstensen TS, Mollnes TE, Garred P, Fausa 0, Brandtzaeg P. Epithelial deposition of imrnunoglobulin G, and activated complement (C3b and terminal complement complex) in ulcerative colitis. Gastroenterology 1990; 98: 1264-71.
Reply SIR,-We are grateful to Dr Das for his comments, and especially for the information that HT-29 cells do not express the Mr 40K protein. We were also unable to show this
protein by immunohistochemical staining using cytospin preparations, but did not use more sensitive methods. We were, however, somewhat surprised by his comments about our negative findings for CCA-IgG being due to fragmentation of IgG. Dr Das rightly informed us that he was unable to detect CCA-IgG in our samples nor could he detect intact IgG. We therefore rechecked not only the frozen aliquots of the samples but also reconstituted freeze dried material which had been sent to Dr Das. The IgG content in the original aliquots and the freeze dried samples differed by less than 10%, and intact IgG was present. Furthermore, we sent the samples to an independent laboratory within the John Radcliffe Hospital which confirmed our findings. These results were communicated to Dr Das who was also sent a copy of the manuscript before it was submitted to Gut. We conclude that CCA-IgG could not be detected in our patients, although we accept that the use of other targets may have shown cytotoxicity. J A SNOOK D P JEWELL Gastroenterology Unit, The Radcliffe Infirmary, Oxford OX2 6HE
Correspondence to: Dr Snook.
Oesophageal complications in epidermolysis bullosa SIR,-Drs Walton and Bennett have written a good overview of oesophagocutaneous diseases (Gut 1991; 32: 694-7). Such diseases are rare and information on their appropriate management is therefore limited. We have an interest in epidermolysis bullosa (EB) and have recently reviewed 258 cases covering all major forms of EB to determine the prevalence of oesophageal lesions, among other features.' Some comments made by Walton and Bennett need to be qualified. Firstly, oesophageal lesions needing dilatation can occur in dominant dystrophic EB. Some 20% of our 57 patients with dominant dystrophic EB had dysphagia, anid oesophageal dilatation was needed in about half of these.' Dysphagia was most common in recessive dystrophic EB (72% of 36 patients), but oesophageal dilatation was needed in a similar proportion. Secondly, cervical strictures are not 'easily dilated'. We try to perform dilatation under general anaesthetic at the same time that the patient is having another operative procedure (such as division of acquired syndactyly), and this requires expert anaesthetic care
to avoid trauma to the orofacial tissues, apart from negotiating with a guide wire a stricture that may start 18 cm from the teeth. We are
1570
Letters. Book reviews
currently using balloon dilatation as opposed to bougies, in the hope that shearing stress on the fragile mucosa is reduced,2 but repeat dilatation after either technique seems to be equally common. Thirdly, we feel that oesophageal replacement should be avoided, owing to the particular problems of surgical and anaesthetic management.3 Strictures do not always recur after dilatation and none of our 258 patients has needed oesophageal surgery. Lastly, it is worth noting that an 'inverse' form of dystrophic EB exists, probably with recessive inheritance, in which oropharyngeal and oesophageal involvement may be severe, but skin lesions relatively mild.4 The management of such difficult problems requires a multidisciplinary approach, with cooperation between dermatologists, gastroenterologists, ENT surgeons, plastic surgeons, anaesthetists, radiologists, and dietitians. S P L TRAVIS Gastroenterology Unit, Radcliffe Infirmary, Oxford OX2 6HE
RAJ EADY, The Institute of Dermatology and StJohn's Dermatology Centre, St Thomas's Hospital, LondonSE) 7EH R P H THOMPSON Gastroenterology Laboratory, The Rayne Institute, St Thomas's Hospital, London SEI 7EH
1 Travis SPL, Turnbull AJ, Schofield OM, Fitzgerald O'Connor A, Mayou B, Eady RAJ, et al. Gastrointestinal complications of epidermolysis bullosa. Gastroenterology 1991; 100: A2 1. 2 Feurle GE, Weidauer H, Baldauf G, SchulteBraucks T, Anton-Lamprecht I. Management of esophageal stenosis in recessive dystrophic epidermolysis bullosa. Gastroenterology 1984; 87:
1376-80. 3 Touloukian RJ, Schonholz SM, Gryboski JD, et al. Perioperative considerations in esophageal replacement for epidermolysis bullosa: report of two cases successfully treated with colon interposition. AmJ Gastroenterology 1988; 83: 857-61. 4 Pearson RW, Paller AS. Dermolytic (dystrophic) epidermolysis bullosa inversa. Arch Dermnatol 1988; 124:544-7.
Limitations of faecal chymotrypsin as a screening test
SIR,-We read with interest the paper by Riedel et al (Gut 1991; 32: 321-4) on the levels of chymotrypsin in the stools of South African patients with chronic pancreatitis and apparently healthy people. There are some points that we would like to make. Firstly, we wonder if the faecal chymotrypsin test should be regarded as a screening test for chronic pancreatitis. This is in part due to the characteristics of the disease: chronic pancreatitis is rare enough not to make a diagnosis in the asymptomatic phase practical; moreover, we have no proof that early treatment can modify its course. On the other hand, the faecal chymotrypsin assay itself is not a good screening test, since it is consistently abnormal only in patients with advanced exocrine impairment. We, for example, found normal levels in only four of 31 patients with exocrine insufficiency, but in 13 of 44 of the patients with chronic pancreatitis.' In our opinion, therefore, the test
should be considered as a test for pancreatic insufficiency only. This is not simply a semantic problem. It implies (i) that the control
subjects should be patients with chronic pancreatitis without exocrine impairment, rather than asymptomatic subjects from the general population; (ii) that we should choose the cut off point that best distinguishes between
patients with and without insufficiency (malabsorption or severely impaired secretinpancreozymin test) rather than between asymptomatic control subjects and patients with pancreatitis. We wonder, therefore, what was the specificity of the test for pancreatic insufficiency considering this lower cut off point. We found the relation between faecal chymotrypsin and stool pH interesting. The correlation was found only in the control subjects. If we have correctly interpreted the data, the chronic pancreatitis patients were studied while on a hospital diet, whereas all the controls were outpatients. The large fibre intake in controls might have induced a large faecal bulk with possible diluting effects on faecal chymotrypsin. All the references in the paper arguing against a diluting effect by faecal bulk on faecal chymotrypsin dealt with the effects of fat malabsorption in Western patients with chronic pancreatitis on a low fibre diet,23 which represents a different condition. Moreover, different time periods are likely to have elapsed between the bowel movements and stool pH measurement in controls (one morning) and in chronic pancreatitis patients. When we measured stool pH immediately and one and six hours after the bowel movement, a progressive reduction in pH was found in five of six chronic pancreatitis patients (from mean (SEM) 7 03 (0-18) to 6-85 (0 11) to 6-7 (0-12)). We wonder, therefore, if both the low pH and chymotrypsin values in some controls may represent hydrolysis by intestinal bacteria for the longer time elapsed in a warm and sunny country. Again, this would argue against the use of the assay as a screening test in the general population, but not in inpatients. G CAVALLINI Istituto di Clinica Medica, Policlinico L BENINI Divisione di Riab Gastroenterologica, COC di Valeggio sM. Universiti di Verona, Italy
1 Cavallini G, Benini L, Brocco G, et al. The fecal chymotrypsin photometric assay in the evaluation of exocrine pancreatic capacity. Comparison with other direct and indirect pancreatic function tests. Pancreas 1989; 4: 300-4. 2 Ammann RW, Akovbiantz A, Haecki W, Largiader F, Schmid M. Diagnostic value of the fecal chymotrypsin test in pancreatic insufficiency, particularly chronic pancreatitis: correlation with the pancreozymin-secretin test, fecal fat excretion and final clinical diagnosis. Digestion 1981; 21: 281-9. 3 Bode C, Bode JC. Usefulness of a simple photometric determination of chymotrypsin activity in stools - results of a multicentre study. Clin Biochem 1986; 19: 333-7.
Reply SIR,-We welcome the opportunity to reply to the letter from Drs Cavallini and Benini. Certainly, faecal chymotrypsin is a test to assess exocrine pancreatic function. The diagnostic value of the test for screening for chronic pancreatitis by showing exocrine pancreatic insufficiency has been suggested by others' and was not the aim of our study. It has been clearly shown that cases of mild or early disease are often missed.2 Most patients with chronic pancreatitis presenting to our unit have advanced disease with exocrine insufficiency. In these patients faecal chymotrypsin determination is a valuable screening test provided faecal pH is taken into consideration. The suggestion that control subjects should be patients with chronic pancreatitis but without insufficiency is of interest but was not the
objective of our study. We emphasise that the question posed was the effect of faecal pH on faecal chymotrypsin. It is correct that we found a correlation between faecal chymotrypsin and stool pH only in control subjects. Patients with chronic pancreatitis and control subjects were not on a hospital diet. They were studied as outpatients on free living diets. It should be noted that the dietary fibre consumption of urban and rural blacks has decreased. In fact in urban blacks dietary fibre consumption is much lower than in Western populations.3 This makes it unlikely that faecal bulk was an important factor. There was no time lapse between bowel movement and stool pH measurement in control subjects and patients. In all subjects stool was put on ice and pH was measured within six hours. Samples were then deep frozen and faecal chymotrypsin was measured within 10 days. Faecal chymotrypsin activity has been shown to be very stable over several days at room temperature. L RIEDEL I SEGAL
Gastroenterology Unit,
Baragwanath Hospital and University of the Witvatersrand, PO Bertsham, 2013, South Afirica 1 Ammann RW, Akovbiantz A, Haecki W, Largiader F, Schmid M. Diagnostic value of the fecal chymotrypsin test in pancreatic insufficiency, particularly chronic pancreatitis: correlation with the pancreozymin-secretin test, fecal fat excretion and final clinical diagnosis. Digestion 1981; 21: 281-9. 2 Deurr HK, Forell MM, Bode JC. Fecal chymotrypsin: a study on its diagnostic value by comparison with the cholecystokinin test. Digestion 1978; 17: 404-9. 3 Segal I, Walker ARP. Low fat intake with falling fibre intake commensurate with rarity of noninfective bowel disease in blacks in Soweto, Johannesburg, South Africa. Nutr Cancer 1986; 8:185-91.
BOOK REVIEWS Progress in hepatic, biliary and pancreatic surgery. By J S Najarian and J P Delaney. (Pp 172; illustrated; £72.) London: Year Book Medical Publishers, 1991. The basis of this annual book, with over 60 contributors, is the continuation course in surgery at the University of Minnesota. Its purpose is to provide an update on the specific but fairly broad areas of general surgery. The course from which this book was generated, dealt with the liver, biliary tract, and pancreas. Despite the title, several of the chapters are more medical in orientation but the book does not lose value from that. For example, the chapter by Dame Sheila Sherlock on 'Viral hepatitis and the surgeon' is a model ofits kind. The authors have brought together a galaxy of surgical talent, including Henri Bismuth, Martin Adson, Babs Moossa, Seymour Schwartz, Frank Moody, and Ben Eiseman. Books such as this are essentially ephemeral, but with the rapid pace of development in this field, even encyclopaedic tomes are usually of only temporary interest. The particular value of this book is that all the contributions are brief, written attractively, and well illustrated.
Letters
LETTERS TO THE EDITOR Serum and tissue autoantibodies to colonic epitheium in ulcerative colitis SIR,-I read with interest and disappointment the above article by Snook et al which appeared recently.' The authors confirmed previous reports from us and others regarding the presence of cytotoxic serum antibodies in a proportion of patients with inflammatory bowel disease, especially those with active disease, except that their percentage of positivity was much lower (20%) than all the other published reports (which were about 50% or more) and most frequent in ulcerative colitis.23 One reason for this could be the use ofdifferent target cells - for example in this study the authors used HT-29 colon cancer cells compared with other studies (including ours) where RPMI-4788 colon cancer cells or normal colon epithelial cells were used as targets. Indeed, this is important because one of the target molecules on the colon cells recognised by ulcerative colitis serum antibody in cytotoxicity assay is associated with the Mr 40K protein' which acts as an autoantigen in ulcerative colitis.56 Monoclonal and polyclonal antiMr 40K antibodies block the cytotoxicity of ulcerative colitis sera on RPMI-4788 and DLD- 1 colon cancer cells,47 showing the role of Mr 40K protein in this recognition. While normal colon epithelial cells, RPMI-4788, and DLD- 1 colon cancer cells express Mr 40K protein, several colon cancer cell lines, including HT-29, which was used by Snook et al, did not express detectable amounts of the Mr 40K protein. Therefore, the use of normal colon cells or selection of appropriate colon cancer target cells is very important for the cytotoxicity study. On the basis of the cytotoxicity data using a single cell line - that is HT-29 - it is indeed misleading for Snook et al to conclude that 'lack of association with disease, extent and activity . . . lack of cytotoxic activity all strongly suggest this antibody is merely an
epiphenomenon.' Using 'ulcerative colitis (UC)-colon extracted IgG (CCA-IgG)', these authors reported completely negative results in their 'functional studies' such as 'cytoxicity assay' using HT-29 target cells and 'immunohistochemical staining' of colon tissue. These results are in contrast to a number of our independent reports"' and the results of many other investi-
gators."'2 There are several major problems with the study of Snook et al. Firstly, the authors did not provide any evidence of intact immunoglobulin or immunoreactive Fab' fragments being present in their ulcerative colitis colon extracted materials ('CCA-IgGs'). The presence of IgG reactivity was reported only with ELISA. Fragmented IgG and Fc fragments will react in the ELISA as performed by them, but will be functionally inactive in the assays they have used. Indeed, six of the 'CCA-IgG' preparations prepared by these authors were examined by us for Ig analysis and immunoreactivity to tissue antigen(s). None of the samples had detectable intact IgG when we analysed them several times by SDS-polyacrylamide gel electrophoresis. This was communicated to the authors well before the publication of the article.
Secondly, as mentioned above, HT-29 cells do not seem to be the right target cells for the cytotoxicity study as they do not express any detectable amount of Mr 40K protein. Therefore, the 'negative functional studies' can easily be explained by the fact that the 'CCA-IgG' extracted by the authors are most likely fragmented by proteases, which are plentiful in colon particularly in ulcerative colitis. Such fragmentations are common if the extraction procedure is performed at room temperature. A striking increase of local IgG production in ulcerative colitis,9' with reactivity to colonic epithelial antigen(s)" have been reported by many investigators, besides our several independent studies.5`8 The immunoreactive colon antigen, Mr 40K protein has been analysed by us using two sensitive techniques, by immunotransblot analysis"7 and by immunocytochemistry using monoclonal antibody.8 Recently, Trond Halstensen et al'2 beautifully demonstrated by two colour immunofluorescence technique the presence of colon epithelial bound IgG (IgG, subtype) and also epithelial deposition of activated complement products (C3b and terminal complement complex) in active ulcerative colitis. All these studies indicate the presence of autoantibodies against epithelial cellular protein(s) in ulcerative colitis which might play an important role in the pathogenesis. While we share the 'disappointment' of the authors for their 'uniform failure in their efforts of extraction of functionally active tissue autoantibodies' in ulcerative colitis, their conclusion of 'no evidence of extractable, epithelial-bound autoantibody' is unfounded. KIRON M DAS Division of Gastroenterology and Hepatology, Department of Medicine, University of Medicine and Dentistry of New3jersey, Robert Wood Johnson Medical School, New Brunswick, NJ 08903-0019, USA 1 Snook JA, Lowes JR, Wu KC, Priddle JD, Jewell DP. Serum and tissue autoantibodies to colonic epithelium in ulcerative colitis. Gut 1991; 32:
163-6. 2 Nagai T, Das KM. Demonstration of an assay for specific cytolytic antibody in sera from patients with ulcerative colitis. Gastroenterology 1981; 80: 1507-12. 3 Auer 10, Grosch L, Hardorfer C, Roder A. Ulcerative colitis specific cytotoxic IgG-autoantibodies against colonic epithelial cancer cells. Gut 1988; 29: 1639-47. 4 Biancone L, Ebert EC, Das KM. The Mr K40,000 colonic protein is associated with antibody dependent cell-mediated cytolysis (ADCC) by ulcerative colitis (UC) sera. Gastroenterology 1990; 98(5): A159. 5 Takahasi F, Das KM. Isolation and characterization of a colonic autoantigen specifically recognized by colon tissue-bound IgG from idiopathic ulcerative colitis. J Clin Invest 1985; 76: 311-8. 6 Takahasi F, Shah H, Wise L, Das KM. Circulating antibodies against human colonic extract enriched with a 40 kDa protein in patients with ulcerative colitis. Gut 1990; 31(9): 1016-20. 7 Das KM, Sakamaki S, Vecchi M. Ulcerative colitis: specific antibodies against a colonic epithelial Mr 40,000 protein. Immunological Investigations 1989; 18: 459-72. 8 Das KM, Vecchi M, Sakamaki S, Diamond B. The production and characterization of monoclonal antibodies to a human colonic antigen associated with ulcerative colitis: Cellular local-
ization of the antigen using the monoclonal antibody. J Immunol 1987; 139: 77-84. 9 Kett K, Rogmem TO, Brandtzaeg P. Mucosal subclass distribution of IgG producing cells is different in ulcerative colitis and Crohn's disease of the colon. Gastroenterology 1987; 93: 919-24. 10 Scott MG, Nagm MH, Macke K, Nash GS, Bertovich MJ, MacDermott RP. Spontaneous secretion of IgG subclasses by intestinal mononuclear cells: differences between ulcerative
colitis, Crohn's Disease, and controls. Clin Exp Immunol 1986; 66: 209-15. 11 Hibi T, Aiso M, Ishikawa M, Watanabe M, Yoshida T, Kobayashi K, Asakura H, Tsuru S, Tsuchiya M. Circulating antibodies to the surface antigens on colonic epithelial cells in ulcerative colitis. Clin Exp Immunol 1983; 54: 163-8. 12 Halstensen TS, Mollnes TE, Garred P, Fausa 0, Brandtzaeg P. Epithelial deposition of imrnunoglobulin G, and activated complement (C3b and terminal complement complex) in ulcerative colitis. Gastroenterology 1990; 98: 1264-71.
Reply SIR,-We are grateful to Dr Das for his comments, and especially for the information that HT-29 cells do not express the Mr 40K protein. We were also unable to show this
protein by immunohistochemical staining using cytospin preparations, but did not use more sensitive methods. We were, however, somewhat surprised by his comments about our negative findings for CCA-IgG being due to fragmentation of IgG. Dr Das rightly informed us that he was unable to detect CCA-IgG in our samples nor could he detect intact IgG. We therefore rechecked not only the frozen aliquots of the samples but also reconstituted freeze dried material which had been sent to Dr Das. The IgG content in the original aliquots and the freeze dried samples differed by less than 10%, and intact IgG was present. Furthermore, we sent the samples to an independent laboratory within the John Radcliffe Hospital which confirmed our findings. These results were communicated to Dr Das who was also sent a copy of the manuscript before it was submitted to Gut. We conclude that CCA-IgG could not be detected in our patients, although we accept that the use of other targets may have shown cytotoxicity. J A SNOOK D P JEWELL Gastroenterology Unit, The Radcliffe Infirmary, Oxford OX2 6HE
Correspondence to: Dr Snook.
Oesophageal complications in epidermolysis bullosa SIR,-Drs Walton and Bennett have written a good overview of oesophagocutaneous diseases (Gut 1991; 32: 694-7). Such diseases are rare and information on their appropriate management is therefore limited. We have an interest in epidermolysis bullosa (EB) and have recently reviewed 258 cases covering all major forms of EB to determine the prevalence of oesophageal lesions, among other features.' Some comments made by Walton and Bennett need to be qualified. Firstly, oesophageal lesions needing dilatation can occur in dominant dystrophic EB. Some 20% of our 57 patients with dominant dystrophic EB had dysphagia, anid oesophageal dilatation was needed in about half of these.' Dysphagia was most common in recessive dystrophic EB (72% of 36 patients), but oesophageal dilatation was needed in a similar proportion. Secondly, cervical strictures are not 'easily dilated'. We try to perform dilatation under general anaesthetic at the same time that the patient is having another operative procedure (such as division of acquired syndactyly), and this requires expert anaesthetic care
to avoid trauma to the orofacial tissues, apart from negotiating with a guide wire a stricture that may start 18 cm from the teeth. We are
1570
Letters. Book reviews
currently using balloon dilatation as opposed to bougies, in the hope that shearing stress on the fragile mucosa is reduced,2 but repeat dilatation after either technique seems to be equally common. Thirdly, we feel that oesophageal replacement should be avoided, owing to the particular problems of surgical and anaesthetic management.3 Strictures do not always recur after dilatation and none of our 258 patients has needed oesophageal surgery. Lastly, it is worth noting that an 'inverse' form of dystrophic EB exists, probably with recessive inheritance, in which oropharyngeal and oesophageal involvement may be severe, but skin lesions relatively mild.4 The management of such difficult problems requires a multidisciplinary approach, with cooperation between dermatologists, gastroenterologists, ENT surgeons, plastic surgeons, anaesthetists, radiologists, and dietitians. S P L TRAVIS Gastroenterology Unit, Radcliffe Infirmary, Oxford OX2 6HE
RAJ EADY, The Institute of Dermatology and StJohn's Dermatology Centre, St Thomas's Hospital, LondonSE) 7EH R P H THOMPSON Gastroenterology Laboratory, The Rayne Institute, St Thomas's Hospital, London SEI 7EH
1 Travis SPL, Turnbull AJ, Schofield OM, Fitzgerald O'Connor A, Mayou B, Eady RAJ, et al. Gastrointestinal complications of epidermolysis bullosa. Gastroenterology 1991; 100: A2 1. 2 Feurle GE, Weidauer H, Baldauf G, SchulteBraucks T, Anton-Lamprecht I. Management of esophageal stenosis in recessive dystrophic epidermolysis bullosa. Gastroenterology 1984; 87:
1376-80. 3 Touloukian RJ, Schonholz SM, Gryboski JD, et al. Perioperative considerations in esophageal replacement for epidermolysis bullosa: report of two cases successfully treated with colon interposition. AmJ Gastroenterology 1988; 83: 857-61. 4 Pearson RW, Paller AS. Dermolytic (dystrophic) epidermolysis bullosa inversa. Arch Dermnatol 1988; 124:544-7.
Limitations of faecal chymotrypsin as a screening test
SIR,-We read with interest the paper by Riedel et al (Gut 1991; 32: 321-4) on the levels of chymotrypsin in the stools of South African patients with chronic pancreatitis and apparently healthy people. There are some points that we would like to make. Firstly, we wonder if the faecal chymotrypsin test should be regarded as a screening test for chronic pancreatitis. This is in part due to the characteristics of the disease: chronic pancreatitis is rare enough not to make a diagnosis in the asymptomatic phase practical; moreover, we have no proof that early treatment can modify its course. On the other hand, the faecal chymotrypsin assay itself is not a good screening test, since it is consistently abnormal only in patients with advanced exocrine impairment. We, for example, found normal levels in only four of 31 patients with exocrine insufficiency, but in 13 of 44 of the patients with chronic pancreatitis.' In our opinion, therefore, the test
should be considered as a test for pancreatic insufficiency only. This is not simply a semantic problem. It implies (i) that the control
subjects should be patients with chronic pancreatitis without exocrine impairment, rather than asymptomatic subjects from the general population; (ii) that we should choose the cut off point that best distinguishes between
patients with and without insufficiency (malabsorption or severely impaired secretinpancreozymin test) rather than between asymptomatic control subjects and patients with pancreatitis. We wonder, therefore, what was the specificity of the test for pancreatic insufficiency considering this lower cut off point. We found the relation between faecal chymotrypsin and stool pH interesting. The correlation was found only in the control subjects. If we have correctly interpreted the data, the chronic pancreatitis patients were studied while on a hospital diet, whereas all the controls were outpatients. The large fibre intake in controls might have induced a large faecal bulk with possible diluting effects on faecal chymotrypsin. All the references in the paper arguing against a diluting effect by faecal bulk on faecal chymotrypsin dealt with the effects of fat malabsorption in Western patients with chronic pancreatitis on a low fibre diet,23 which represents a different condition. Moreover, different time periods are likely to have elapsed between the bowel movements and stool pH measurement in controls (one morning) and in chronic pancreatitis patients. When we measured stool pH immediately and one and six hours after the bowel movement, a progressive reduction in pH was found in five of six chronic pancreatitis patients (from mean (SEM) 7 03 (0-18) to 6-85 (0 11) to 6-7 (0-12)). We wonder, therefore, if both the low pH and chymotrypsin values in some controls may represent hydrolysis by intestinal bacteria for the longer time elapsed in a warm and sunny country. Again, this would argue against the use of the assay as a screening test in the general population, but not in inpatients. G CAVALLINI Istituto di Clinica Medica, Policlinico L BENINI Divisione di Riab Gastroenterologica, COC di Valeggio sM. Universiti di Verona, Italy
1 Cavallini G, Benini L, Brocco G, et al. The fecal chymotrypsin photometric assay in the evaluation of exocrine pancreatic capacity. Comparison with other direct and indirect pancreatic function tests. Pancreas 1989; 4: 300-4. 2 Ammann RW, Akovbiantz A, Haecki W, Largiader F, Schmid M. Diagnostic value of the fecal chymotrypsin test in pancreatic insufficiency, particularly chronic pancreatitis: correlation with the pancreozymin-secretin test, fecal fat excretion and final clinical diagnosis. Digestion 1981; 21: 281-9. 3 Bode C, Bode JC. Usefulness of a simple photometric determination of chymotrypsin activity in stools - results of a multicentre study. Clin Biochem 1986; 19: 333-7.
Reply SIR,-We welcome the opportunity to reply to the letter from Drs Cavallini and Benini. Certainly, faecal chymotrypsin is a test to assess exocrine pancreatic function. The diagnostic value of the test for screening for chronic pancreatitis by showing exocrine pancreatic insufficiency has been suggested by others' and was not the aim of our study. It has been clearly shown that cases of mild or early disease are often missed.2 Most patients with chronic pancreatitis presenting to our unit have advanced disease with exocrine insufficiency. In these patients faecal chymotrypsin determination is a valuable screening test provided faecal pH is taken into consideration. The suggestion that control subjects should be patients with chronic pancreatitis but without insufficiency is of interest but was not the
objective of our study. We emphasise that the question posed was the effect of faecal pH on faecal chymotrypsin. It is correct that we found a correlation between faecal chymotrypsin and stool pH only in control subjects. Patients with chronic pancreatitis and control subjects were not on a hospital diet. They were studied as outpatients on free living diets. It should be noted that the dietary fibre consumption of urban and rural blacks has decreased. In fact in urban blacks dietary fibre consumption is much lower than in Western populations.3 This makes it unlikely that faecal bulk was an important factor. There was no time lapse between bowel movement and stool pH measurement in control subjects and patients. In all subjects stool was put on ice and pH was measured within six hours. Samples were then deep frozen and faecal chymotrypsin was measured within 10 days. Faecal chymotrypsin activity has been shown to be very stable over several days at room temperature. L RIEDEL I SEGAL
Gastroenterology Unit,
Baragwanath Hospital and University of the Witvatersrand, PO Bertsham, 2013, South Afirica 1 Ammann RW, Akovbiantz A, Haecki W, Largiader F, Schmid M. Diagnostic value of the fecal chymotrypsin test in pancreatic insufficiency, particularly chronic pancreatitis: correlation with the pancreozymin-secretin test, fecal fat excretion and final clinical diagnosis. Digestion 1981; 21: 281-9. 2 Deurr HK, Forell MM, Bode JC. Fecal chymotrypsin: a study on its diagnostic value by comparison with the cholecystokinin test. Digestion 1978; 17: 404-9. 3 Segal I, Walker ARP. Low fat intake with falling fibre intake commensurate with rarity of noninfective bowel disease in blacks in Soweto, Johannesburg, South Africa. Nutr Cancer 1986; 8:185-91.
BOOK REVIEWS Progress in hepatic, biliary and pancreatic surgery. By J S Najarian and J P Delaney. (Pp 172; illustrated; £72.) London: Year Book Medical Publishers, 1991. The basis of this annual book, with over 60 contributors, is the continuation course in surgery at the University of Minnesota. Its purpose is to provide an update on the specific but fairly broad areas of general surgery. The course from which this book was generated, dealt with the liver, biliary tract, and pancreas. Despite the title, several of the chapters are more medical in orientation but the book does not lose value from that. For example, the chapter by Dame Sheila Sherlock on 'Viral hepatitis and the surgeon' is a model ofits kind. The authors have brought together a galaxy of surgical talent, including Henri Bismuth, Martin Adson, Babs Moossa, Seymour Schwartz, Frank Moody, and Ben Eiseman. Books such as this are essentially ephemeral, but with the rapid pace of development in this field, even encyclopaedic tomes are usually of only temporary interest. The particular value of this book is that all the contributions are brief, written attractively, and well illustrated.
