CONTROLLED TRIAL OF FAECAL OCCULT BLOOD TESTING IN THE DETECTION OF COLORECTAL CANCER

Saturday 2 July

CONTROLLED TRIAL OF FAECAL OCCULT BLOOD TESTING IN THE DETECTION OF COLORECTAL CANCER P. A. FARRANDS* T. W. BALFOUR

J. D. HARDCASTLE

Department of Surgery, University Hospital, Nottingham NG7 2UH

J. CHAMBERLAIN Royal Marsden Hospital, Sutton, Surrey SM2 5PT S. S. AMAR

University Hospital, Nottingham NG7 2UH M. G. SHELDON

Department of Community Health, University of Nottingham, Queens Medical Centre, Nottingham NG7 2UH 20 525 patients from general practitioners’ lists were randomly allocated into test and control groups. The 10 253 test subjects were invited to perform haemoccult faecal occult blood testing over 3 days. 3613 (36·8%) of the 9807 who received their invitations completed the test. Compliance was improved by direct invitation from the general practitioner and by prior health education by letter or interview. 77 people (2·1%) had a positive test result, and 50% of these on investigation had

Summary

neoplastic disease—12 had invasive carcinomas (9 Dukes’ stage A, 2 stage B, 1 stage C) and 27 had 40 adenomas (12 over 2 cm, 2 of which contained areas of severe dysplasia). In the year following the screening test 1 carcinoma (stage C) has presented in the group which accepted the test, and 10 carcinomas (4 stage B, 4 stage C, 2 stage D) have presented in the control group. This respresents a 3·6 times greater detection rate per 1000 persons in the test group than in the control group. Only 8 adenomas have presented in the control and non-responding groups. Fibreoptic sigmoidoscopy identified the 10 carcinomas within its range and 39 of the 40 adenomas. Double-contrast barium enema identified only 9 of the 12 carcinomas and 24 (62%) of the 40 adenomas. All 3 carcinomas not identified by barium enema were polypoid Dukes’ stage-A lesions.

Introduction CARCINOMA of the large bowel is now the second in the Western world, 16 500 individuals of the disease in England and Wales each year.There dymg has been little change in the survival rate over the past 20 years, 24% of patients registered in 1959 surviving 5 years, compared with 30% in 1975.z2

commonest cancer

*Present address:

Department of Surgery, University ot Southampton.

1983

There is evidence that the early diagnosis of rectal carcinoma by means of sigmoidoscopy leads to a greater yield of localised tumours and an improved 5-year surviva1.3,4 At the present time the only practical method of population screening for colorectal cancer is by the detection of occult blood in the faeces.5 A guaiac-impregnated filter-paper test (’Haemoccult’) has been widely used in the USA and has been included in a national screening service in the Federal German Republic since 1977. Pilot studies in the United Kingdom have shown that faecal occult blood can be tested with haemoccult in general

practice.6-8 Whether the early diagnosis of colorectal cancer leads to improved survival can be answered only by the results of prospective controlled studies. One such study is being undertaken in the USA,9 but the study and control groups have been selected from health-conscious volunteers, many of whom are undertaking regular health checks. It is likely, therefore, that the results obtained from this work will not apply to the population as a whole. Preliminary results show that 65% of tumours detected through screening are localised to the bowel wall," but no information relating to the control group is yet available. This present study was undertaken to compare the yield of neoplasia in a test group offered haemoccult screening with that in a control group, both groups having been randomly allocated from the same population. The test and control groups have been followed up for one year after the screening test, and the pathological stage of tumours in both groups has been recorded. Factors affecting the acceptance of the haemoccult test have been investigated. Methods

patients aged 45-74 were identified in nine general in practices the Nottingham area. The practices provided a crosssection of rural and urban populations and referred all their patients to the two Nottingham hospitals. Patients with known bowel disease or a previous history of cancer and those otherwise designated unsuitable by the general practitioner were excluded from the study. Families within each practice were randomly allocated to test and control groups, so that husband and wife, where both were eligible, were included in the same group. 10 253 people were allocated to the test group and 10 272 to the control group. Control subjects were allocated dates of entry comparable to the test group’s dates of invitation. The 10 253 test subjects were sent instructions that the haemoccult test should be used on 3 consecutive days, a sheet of instructions, and a stamped addressed envelope in which to return the completed test. Subjects were asked not to eat red meat or to take vitamin-C preparations before collection of the faecal samples. Initially individuals were randomly allocated, within each practice, to receive either an invitation letter from their own general 20 525

8340

2

or a similar letter from the department of community health in the University of Nottingham. The effect of increasing knowledge of colorectal cancer by a preliminary letter or personal interview, and the effect of taking ispaghula husk (’Fybogel’) during specimen collection were subsequently studied in a similar randomised manner. Returned haemoccult occult-blood slides were tested in the department of surgery, without rehydration, by the addition of two drops of stabilised hydrogen peroxide. A blue discoloration occurring at 30 s was recorded as positive. Subjects with a positive haemoccult test had a full history taken, and a clinical examination, including digital rectal examination, proctoscopy, rigid sigmoidoscopy and 60 cm fibreoptic sigmoidoscopy, was carried out. All patients subsequently had a double-contrast barium enema (administered by S. S. A.). The reasons for these investigations were carefully explained to the patients, who gave their consent. Tumours were staged according to Dukes’ system, modified to include a stage D when histologically proven liver metastases were present. Colorectal neoplasia developing within 1 year of the test in both test and control groups was determined by checking the hospital pathology registers, the Trent Cancer Registry, and general practitioners’ lists.

practitioner

Results The distribution of patients in test and control groups is shown in the accompanying figure. 19 patients saw their general practitioner because of recent rectal bleeding. 1 carcinoma of the sigmoid colon (stage B) and 2 tubular adenomas (I - 0- 1 -5 cm) were detected. Examination of the 77 haemoccult-positive patients revealed neoplastic disease in 39. 12 persons were found to have infiltrating carcinoma, and 27 persons had 1 or more adenomas. 9 (75%) of the patients with carcinoma and 22 (81 -5%) of those with adenoma had no symptoms. 1 carcinoma was identified by digital rectal examination. Rigid sigmoidoscopy to 20 cm was achieved in all 39 patients with neoplasia and identified 4 of the 12 carcinomas (33%) and 10 of the 40 adenomas (25%). Flexible fibreoptic sigmoidoscopy identified 10 carcinomas (83%) and 39 (97-5%) of the adenomas. Full insertion of the instrument was achieved in 36 patients. Double-contrast barium enema identified only 9 (75%) of the 12 carcinomas and 24 (62 ° 5%) of the 40 adenomas. 3 stage-A tumours were missed-2 in the sigmoid colon and 1 in

the transverse colon. 16 adenomas were missed in the sigmoid colon. Yield of Carcinomas 10 of the 12 patients with infiltrating carcinomas underwent colectomy, and 2 were treated with polypectomy. In 2 patients the carcinoma was situated in the rectum, in 4 in the sigmoid, in 4 in the rectosigmoid, and in 2 in the transverse colon. 9 (75%) of the carcinomas were stage-A lesions; 2 (16 - 7%) were stage-B, and 1 stage-C (8 - 3%). All patients underwent a curative resection. 8 of the carcinomas were polypoid growths, 5 of which were pedunculated and 3 sessile. 9 of the carcinomas were moderately differentiated lesions, and 3 were well differentiated. 4 of the polypoid carcinomas were associated with adenomatous tissue (2 tubulovillous and 2 tubular). Yield of Adenomas

40 adenomas, 5 mm

or more

in

size,

were

identified in 27

patients. 35 (87%)

adenomas were below 40 cm on endoscopy; 4 in the descending colon and 1 was in the transverse colon. Their sizes ranged from 5 to 50 mm, the majority being 10-20 mm. 12 were over 20 mm, and 4 were over 30 mm. Histological examination showed that 24 were tubular, 14 tubulovillous, and 2 villous adenomas. Dysplasia was present in 22 adenomas. It was severe in a 1-55 cm tubulovillous adenoma and a 2 - 5 cm tubular adenoma. 11adenomas had foci of moderate dysplasia, 3 arising in tubular adenomas, 6 in tubulovillous adenomas, and 2 in villous adenomas. Mild dysplasia was present in 9 adenomas. 25 patients were treated with colonoscopy and polypectomy; in 2 patients laparotomy and colotomy was were

necessary.

1 year Follow-up of

Among

the 3536

Test and Control

patients

with

a

Groups negative

haemoccult

test

1 patient presented a year later with a stage-C carcinoma of the ascending colon. No adenomas have presented in this group. 11 of the 6143 subjects who either refused or failed to reply subsequently presented with neoplastic disease-8 with carcinomas (5 stage B [3palliative resections with local tumour remaining] and 3 stage D [liver metastases]). 3

result,

subjects presented with tubulovillous adenomas (1-2 cm). Among the 10 272 control subjects, 10 presented with carcinoma of the colon and 5 with adenomas during the first year of the study. Of the carcinomas 4 were stage B (2 with local tumour remaining), 4 were stage C, and 2 stage D (liver metastases). 4 of the adenomas measured 1 - 5-3’ 0 cm and 1 villous adenoma 7 cm x 6 cm. A comparison of the yield of neoplasia in the screened group, the non-responders, and the control group is shown in the accompanying table. The detection rate of carcinoma at first screening (3-6/1000 persons screened) is 3-6times greater than the annual incidence in the control group (1’ 0/1000 persons) and the adenoma detection rate is 16 times ’

greater. Factors

Distribution of subjects:

test

and control groups.

Affecting Compliance

1673 of 4766 men (35-1%) and 1940 of 5052 women (38-4%) who received the test accepted (p<0-01). The response rates within 5-year age groups were similar (38-40%) up to age 65; above this age the compliance was less, especially in those over 70 (27%).

3 YIELD OF COLORECTAL NEOPLASMS IN ONE YEAR FROM ENTRY TO THE STUDY

stage-A lesions, regional nodes cancers

the

majority having already spread

to

the

liver.l1,12 The pathological staging of detected in the control group is similar; no stage-A or

carcinomas have presented; 2 of the 4 patients presenting with stage-B lesions had local tumour remaining after operation; 4 patients had lymphatic involvement (stage C), and 2 had liver metastases at the time of operation. Similar results were found in the 8 patients presenting with carcinoma who failed to do the test (non-responders): 5 had stage-B carcinomas and 3 had liver metastases at the time of

operation.

Among any group of people with preclinical

cancer there case-mix of differing durations of the preclinical phase, and the length of this phase may vary considerably, *The 19 patients who, as a result of being sent the test, consulted their general some cancers progressing rapidly whereas others may be very practitioner because of rectal bleeding are included in this group; 1 carcinoma slow-growing. Screening at any single point in time is likely to and 2 adenomas were diagnosed. pick up a disproportionate number of slow-growing lesions, because prevalence is a function of duration and the slowly Effect of varying the source of the invitation.-1271 progressive cancers have a longer duration. 13 A large number invitations from the patients’ own practitioner and 1260 from of well-differentiated tumours might therefore have been the department of community health, University of expected among those detected at screening. In fact only 3 of Nottingham, were randomly distributed in two general the 12 carcinomas were well differentiated, the rest being practices. 455 (38’3%) accepted the GP invitation, compared moderatly differentiated, suggesting that screening is able to with only 303 (25 ° 8°70) who accepted the invitation from the pick up aggressive tumours at an early stage. department of community health (p<0-01). Identification of the 27 subjects with one or more adenomas Effect of preliminary education by letter or interview.-442 by means of faecal occult blood testing is also important, for subjects were sent a letter informing them about colorectal there is increasing evidence that these lesions are cancer and the purpose of the test 2 weeks before being sent premalignant, 14 especially those measuring over 2 cm, villous the test invitation, 472 received only the invitation letter from or tubulovillous types, and those showing mucosal the GP. 194 (46’7%) patients receiving the educational letter12 of the adenomas detected in this study were dysplasia." accepted, compared with 171(38’07o) of those sent only the over 2 cm, and 4 were over 3 cm. 16 were villous or invitation letter (p<0’02). 756 subjects randomly selected tubulovillous, and 2 contained areas of severe dysplasia. from eight general practices were seen at interview, during The false-negative rate of the haemoccult test may be as which colorectal cancer and the use of the test were discussed. high as 30% in known colorectal cancers. 16,17 However, little The overall compliance in this group was 51 -6% (p<0’ 00 1). information on the false-negative rate in population screening Addition of a stool-bulk ing agent. -3020 patients were asked is available. Only 1 patient with a negative test result to take fybogel for 3 days covering the period of the occult presented in the following year with colorectal cancer. blood test, and 3025 were offered the haemoccult test alone. The carcinoma detection rate in the present study, 935(32-5%) patients asked to take fybogel accepted the occult 3’6/1000 persons screened, is higher than that previously blood test, compared with 1102 (37’8%) asked only to do the reported in studies in the United Kingdom in which only test (p<0-01). 14 (1 -5%) of the 835 subjects performing the rigid sigmoidoscopy was used. This can possibly be haemoccult test with fybogel had a positive test, compared attributed to the routine use of the flexible fibreoptic with 21(1-9%) of the 1102 haemoccult-only group. sigmoidoscope in the investigation of haemoccult-positive persons. The yield of this examination is greater than that of Discussion rigid sigmoidoscopy,18 and double-contrast barium enema had been shown to be unreliable in the detection of polypoid of occult blood means Population screening by testing, lesions in the sigmoid colon, particularly in the presence of in several can be countries, although widely practised only diverticular disease.’9 In the present study 3 of 12 carcinomas if it be shown that the of the can justified mortality group and 16 of 40 adenomas were missed by this examination. for a screened colorectal cancer is lower than similar control It is encouraging that the screening test was more readily group. The present study indicates that faecal occult blood accepted by patients under the age of 65 (38-40%) than by those over 70 (27%), since early diagnosis in this group is screening can detect neoplasia at a pathological stage more favourable than that found in patients with symptoms. Of the likely to be most rewarding. It is known that people refusing occult blood testing are less 10 272 control subjects, 10 have presented in the first year with colorectal carcinoma and 5 with adenomas. Of the 3613 health-conscious, know less about cancer (especially test subjects, 12 were found to have carcinoma (3-6/1000 colorectal cancer), and are less well informed about medical matters than people accepting the occult blood test (Farrands persons screened-a detection rate 3,6times that of the annual incidence), and 27 were found to have one or more PA, Chamberlain J, Moss S, Hardcastle JD, unpublished). It adenomas (7-4/1000 persons screened-a detection rate 16 was encouraging, therefore, that more patients accepted the test when the invitation was preceded by an educational letter times higher than that of the annual incidence). Not only was the neoplastic yield greater in the screened group, but 9 (75%) (47%) or an interview (52%). of the carcinomas identified were Dukes’ stage A lesions and 2 This study has shown that the pathological stage of Dukes’ stage B, and only 1 patient had lymphatic tumours detected in the screened group is more favourable involvement. In most regional hospitals only 5-10% of than that of tumours presenting in the control and noncolorectal cancers presenting symptomatically are Dukes’ responding groups at one year, but a larger study is necessary

will be

,

a

4

determine whether altered. to

’

mortality

rates

in the

long-term

V’e thank Eaton Laboratories for supplying the haemoccult faecal occult blood tests and Reckitt and Colman for supplying the fybogel This work was supported by a grant from the Department of Health and Social Security. P. A. F. was supported by a Cancer Research Campaign Surgical Research

Fellowship. Correspondence should

be addressed

to

J. D. H.

REFERENCES 1.

Office of Population Censuses and Surveys 1980 mortality statistics, series DH2, no. 7 London HM

Stationery Office,

2. Office of Population Censuses and

1982.

Surveys.

Cancer statistics—survival,

series

MB1,

9 London HM Stationery Office, 1982. 3. Sherlock P, Winawer SJ The role of early diagnosis in controlling large bowel cancer an overview. Cancer 1977; 40: 2609-15 4. Gilbertsen VA, Nelms JM. The prevention of invasive cancer of the rectum. Cancer no.

1978; 41: 1137-39. 5. Hardcastle 6 7 8. 9.

10

JD. Screening for colorectal cancer. In: Wright R, ed. Recent advances in gastrointestinal pathology. Philadelphia. WB Saunders, 1980 311-29 Hardcastle JD, Balfour TW, Amar SS Screening for symptomless colorectal cancer by testing for occult blood in general practice Lancet 1980; i: 791-93. Million R, Howarth J, Turnberg E, Turner LA Faecal occult blood testing for colorectal cancer in general practice. Practitioner 1982; 226: 659-63 Farrands PA, Griffiths RL, Britton DC A practical solution to the diagnosis and treatment of colorectal cancer? Lancet 1981, i: 1231-32. Gilbertsen VA, Church TR, Crewe FJ, Mandel JS, McHugh RB, Schuman LM, Williams SE. The design of a study to assess occult blood screening for colon cancer J Chron Dis 1980; 33: 107-14. Gilbertsen VA, McHugh RB, Schuman LM, Williams SE. The earlier detection of colorectal cancer. A preliminary report on the results of the occult blood study.

Cancer 1980; 45: 2899-01. PG, Morris PJ. The survival of patients with colorectal cancer treated in a regional hospital. Br J Surg 1978, 65: 17-20 12 Holliday HW, Hardcastle JD Delay in diagnosis and treatment of symptomatic colorectal cancer. Lancet 1979; i: 309-11. 13. Chamberlain J Screening for early detection of cancer: general principles In: Alderson M, ed. The prevalence of cancer London: Edward Arnold, 1982. 227-58 14. Morson BC. The polyp-cancer sequence in the large bowel. Proc Roy Soc Med 1974; 67: 11 Gill

451-57. 15. Koneshi F, Morson BC Pathology of colorectal adenomas; Pathol 1982, 35: 830-41.

16

a

colonscopy survey J Clin

Macrae FA, St Johns DJB. Relationship between patterns of bleeding and haemoccult sensitivity

in

patients with colorectal

cancer or

adenomas

Gastroenterology 1982;

82: 891-98 17.

18. 19.

Doran J, Hardcastle JD. Patterns of bleeding in colorectal cancer-the effect of aspirin. Implications for faecal occult blood testing. Br J Surg 1982, 69: 711-13. Vellacott KD, Hardcastle JD. An evaluation of flexible fibreoptic sigmoidoscopy. Br MedJ 1981, 283: 1583-86. Vellacott KD, Amar SS, Hardcastle JD Comparison of rigid and flexible fibreoptic sigmoidoscopy with double contrast barium enema Br J Surg 1982, 69: 399-400.

MASKING BY ENZYME INHIBITOR OF RAISED SERUM GLUTAMATE DEHYDROGENASE ACTIVITY IN REYE’S SYNDROME

JEFFREY T.

HOLT DEAN A. ARVAN THEODOR K. MAYER

Clinical Pathology Laboratory Division, University of Rochester School of Medicine and Dentistry, Strong Memorial Hospital, 601 Elmwood Avenue, Box 608, Rochester, New York 14642, USA

glutamate dehydrogenase (GDH) activity was greatly raised (up to 830 times the upper limit of normal) in 16 patients with Reye’s syndrome. The serum activity was masked by an inhibitor, and the rises were observed only after dialysis or sample dilution. Serum GDH values from 38 paediatric patients, including 10 with hyperammonaemia due to other causes, showed no such rise after dialysis. Only 1 of 13 adult patients with liver disease had high GDH activity, but this level was not increased after dialysis. Serum ornithine carbamyl transferase activity was also raised in patients with Reye’s syndrome, but levels were not increased after dialysis. The ratio of dialysed/undialysed GDH activity clearly distinguished all Reye’s patients from controls. The inhibition of a mitochondrial enzyme which regulates ammonia metabolism may contribute to the hyperammonaemia of Reye’s syndrome. Serum GDH levels before and after dialysis would seem to be a useful diagnostic aid in Reye’s disease.

Summary

Introduction

are

BIOCHEMICAL and ultrastructural studies provide evidence of acute mitochondrial insult in Reye’s syndrome. 1-4 Decreased activity of mitochondrial enzymes (glutamate dehydrogenase, ornithine carbamyl transferase, and others) has been reported in liver tissue from patients with this disease5-8 and could be due to decreased synthesis, increased degradation, inactivation, inhibition, or loss from liver cells into serum and other biological fluids. Several reports ascribe the low mitochondrial enzyme activity in the liver to decreased synthesis.9-11Except for a report by Deshmukh et al12 little attention has been given to other possible mechanisms of apparent mitochondrial enzyme depletion. Some enzymes which are primarily cytosolic (eg, aspartate transaminase, alanine transaminase) are increased in sera from patients with Reye’s syndrome. The rises, however, are generally variable and do not correlate well with the severity of disease.9 Such rises may reflect secondary injury to hepatocytes or other cells, so their value in diagnosis is limited. We hypothesised that, if Reye’s syndrome results from a selective or primary injury of mitochondria, serum mitochondrial enzyme levels might provide specific information for diagnosis and management, and might contribute to the understanding of the pathogenesis of the disease. Glutamate dehydrogenase (GDH, EC 1.4.1.3.) is a mitochondrial enzyme that uses nicotinamide adenine dinucleotide (NAD) to reversibly catalyse the oxidative deamination of glutamate, producing ammonia, a-ketoglutarate, and reduced nicotinamide adenine dinucleotide (NADH). The physiological activity of GDH was previously thought to operate largely in the direction of ammonia production.13 Work based on the intramitochondrial concentration of reactants and allosteric modifiers has provided grounds for questioning this concept. In vivo, the reaction may be predominantly towards the production of glutamate (reductive deamination) and thus GDH may regulate intramitochondrial ammonia concentration via direct ammonia consumption as well as by activation of carbamyl phosphate synthetase. 14-11 Ornithine carbamyl transferase (OCT, EC 2.1.3.3.), a key urea cycle enzyme, is also located within the mitochondrion. The intramitochondrial location of these key enzymes of ammonia metabolism, and the hyperammonaemia that accompanies mitochondrial damage in Reye’s syndrome, prompted our study of serum GDH and OCT activity in patients with Reye’s syndrome and in appropriate controls.

Subjects

Serum

and Methods

Patients Sera from 13 patients with Reye’s syndrome diagnosed at Strong Memorial Hospital and from 3 patients from the Massachusetts General Hospital were analysed for GDH and OCT activity. 13 of these samples had been frozen at -20°C for several days to 5 years before they were tested. All tests were done with surplus sera from routinely submitted laboratory specimens. The clinical and laboratory details of these 16 patients are presented in tablei.

Controls 10 consecutive paediatric in-patients with hyperammonaemia evaluated. Chart review confirmed that none of these patients met the clinical criteria for Reye’s syndrome. Most had transient neonatal hyperammonaemia. 20 consecutive paediatric patients without hyperammonaemia were also studied, as well as 8 patients with familial metabolic diseases, including 2 patients with arginosuccinic acidaemia, a homozygote and a symptomatic female heterozygote with ornithine carbamyl transferase deficiency, and 1

were