Diagnostic accuracy of faecal occult blood tests used in screening for colorectal cancer: a systematic review.

OBJECTIVE: To determine the accuracy of guaiac and immunochemical faecal occult blood tests (FOBTs) for the detection of colorectal cancer in an average-risk screening population. METHODS: Fifteen electronic databases, the internet, key journals and reference lists of included studies were searched. We included diagnostic accuracy studies that compared guaiac or immunochemical FOBTs with any reference standard, for the detection of colorectal cancer in an average-risk adult population, with sufficient data to construct a 2 x 2 table. RESULTS: Fifty-nine studies were included. Thirty-three evaluated guaiac FOBTs, 35 immunochemical FOBTs and one evaluated sequential FOBTs. Sensitivities for the detection of all neoplasms ranged from 6.2% (specificity 98.0%) to 83.3% (specificity 98.4%) for guaiac FOBTs, and 5.4% (specificity 98.5%) to 62.6% (specificity 94.3%) for immunochemical FOBTs. Specificity ranged from 65.0% (sensitivity 44.1%) to 99.0% (sensitivity 19.3%) for guaiac FOBTs, and 89.4% (sensitivity 30.3%) to 98.5% (sensitivity 5.4%) for immunochemical FOBTs. Diagnostic case-control studies generally reported higher sensitivities. Sensitivities were higher for the detection of CRC, and lower for adenomas, in both the diagnostic cohort and diagnostic case-control studies for both guaiac and immunochemical FOBTs. CONCLUSIONS: Immudia HemSp appeared to be the most accurate immunochemical FOBT, however, there was no clear evidence to suggest whether guaiac or immunochemical FOBTs performed better, either from direct or indirect comparisons. Poor reporting of data limited the scope of this review, and the use the Standards for Reporting of Diagnostic Accuracy guidelines is recommended for reporting future diagnostic accuracy studies.

A colon cancer cell line (LIM1215) derived from a patient with inherited nonpolyposis colorectal cancer.

A new human colon cancer cell line (LIM1215) has been derived from a tumor arising in a member of a family known to have a high incidence of colorectal cancer. The tumor cell line is comprised of small pleomorphic cells that clone in liquid medium and form tumors in immunosuppressed mice. Ultrastructurally, the cells are capable of differentiation, with cells with multiple microvilli and cells resembling goblet cells being present in the one culture. The cells are pseudodiploid and contain a 13p+ marker chromosome.

Improved survival in esophageal cancer in the period 1978 to 1983.

The outcome of therapy of adenocarcinomas and squamous-cell carcinomas of the esophagus is so poor that the results of new approaches to therapy, such as the addition of radiotherapy or chemotherapy, are often compared with those achieved in historical controls. To determine the validity of this approach in cancers with a poor outcome, the results of therapy were analyzed at our institution from Jan 11, 1978 to Aug 9, 1981 (77 patients) and compared with the results achieved in the period from Aug 14, 1981 to Feb 19, 1984 (77 patients). The patients were evenly matched for prognostic factors. It was found that the median survival of the first group of patients (4 months) was significantly less (P less than .01) than that of the recently treated group (10 months). This was due to the better median survival of patients treated surgically from 4 months in the early group to 29 months at present (P less than .01). There was no change in the survival of the other patients. The major improvement in the outcome of surgery was due to the reduction of the perioperative mortality rate to less than 5%. There was no detectable change in patient selection for surgery. The results indicate that even in tumors with a very poor outcome, such as esophageal cancer, large changes in results can occur without a specific change in therapy, and that historical controls may be misleading.

Profuse familial adenomatous polyposis with an adenomatous polyposis coli exon 3 mutation.

The attenuated form of familial adenomatous polyposis coli (AAPC) is associated with mutations in the adenomatous polyposis coli (APC) gene which cluster in the 5' region of the gene. It has been proposed that a 'genotype-phenotype boundary' exists at codons 159-163, and mutations that are 5' of this boundary will produce AAPC. Herein we document a three-generation family with an exon 3 mutation well to the 5' side of the proposed boundary, in which two affected individuals have had, in their 40s, a profuse form of familial adenomatous polyposis coli. We conclude that the codon 159-163 'boundary' is indicative rather than definitive. These two patients also had postoperative intra-abdominal adhesions, severely so in one.

Chromosome studies in inherited nonpolyposis colon cancer syndrome.

This study was designed to determine if any constitutional chromosomal markers were linked with the expression of colorectal neoplasms in the inherited nonpolyposis colon cancer syndrome, using a number of cytogenetic techniques. High resolution G-banding in 12 affected and 17 unaffected family members did not reveal a structural chromosome abnormality. Increased C-band heteromorphism was not seen in either affected or unaffected individuals, and no heritable fragile sites were detected. Mean baseline and mitomycin C-induced sister chromatid exchanges were not elevated in affected patients compared with controls. Mapping of sister chromatid exchanges did not reveal any hot spots of exchange. A tumor cell line with the karyotype 46,XY,der(13),t(13;?)(p11;?) was established from one patient, but no constitutional abnormality of chromosome #13 was found. In addition, 11 patients with familial polyposis coli were studied with high resolution G-banding and no heteromorphism of chromosome #2 in the region 2q21.3 was detected.

Small intestinal malabsorption of vitamin B(12) in iron-deficient rats.

Rats were rendered iron deficient by a combination of diet and bleeding to study its effects on vitamin B(12) absorption. Small intestinal loops were isolated in vivo and the absorption of -57Co-vitamin B(12) bound to a known quantity of intrinsic factor was measured. Iron deficiency resulted in the impairment of both uptake and transport of B(12). This malabsorption was corrected within 5 days by parenteral iron repletion. The findings were not due to a non-specific effect of anaemia since no correlation existed between haemoglobin levels and B(12) absorption in rats anaemic as a result of acute blood loss. No evidence was found for an altered small-intestinal microflora, bacterial counts being similar in iron-deficient and control rats. It is concluded that iron deficiency in the rat results in impaired absorption of B(12) by the small intestine, probably as a result of some defect produced in the enterocyte.

Prescreening evaluation of a brush-based faecal immunochemical test for haemoglobin.

OBJECTIVES: To undertake a prescreening evaluation of a new brush-based faecal immunochemical test for haemoglobin, relative to a traditional spatula-sampling immunochemical test. SETTING: Patients aged between 24 and 90 years, scheduled to undergo diagnostic colonoscopy in two major urban hospitals, for a range of clinical indications. DESIGN: Patients sampled three stools using a spatula for the reference FlexSure OBT test and two stools using a brush for the InSure test; order of sampling was randomised. Faecal haemoglobin was quantified by a modified InSure in a subset of patients to determine whether brush-sampling allowed discrimination between groups. MAIN OUTCOME MEASURES: Sensitivity for cancer or adenoma; false-positive rate in normals. Faecal haemoglobin levels. Preference for sampling method. RESULTS: InSure and FlexSure OBT did not differ in their sensitivities for cancer (27/36, 75% vs 29/36, 80.5%, respectively), adenomas >or= 10 mm (12/29, 41.4% vs 13/29, 44.8%) or adenomas <10 mm (each 8/56, 14.3%). Likewise, false-positive rates in normals were similar: 4/179 (2.2%) and 5/179 (2.8%) respectively (specificities of 97.8% and 97.2%, respectively). Levels of faecal haemoglobin were highest in those with cancers; those with adenomas had intermediate levels which were also significantly higher than those in normals. The brush sampling method was preferred by 38/46 (82.6%), while 4/46 (8.7%) preferred the spatula (p<0.00001). CONCLUSIONS: InSure is as sensitive and specific as FlexSure OBT for faecal haemoglobin. The novel stool-sampling method of InSure allows discrimination between normals and classes of neoplasia, and is highly preferred. The brush-sampling faecal immunochemical test InSure should now be evaluated in a screening population.

Aspirin-induced glandular dysplasia of the stomach: histologic and histochemical studies in rats.

Stomachs of 22 rats treated with aspirin for six months were examined histologically at times ranging from 6 to 18 months after completion of treatment. Healed chronic gastric ulcers were found in 20 rats. Glandular dysplasia was present at the sites of healed ulcers in 12 rats, the glands extending into the submucosa, muscularis propria, and even subserosal fat. The apparent lack of progression with time and the absence of metastases suggest that the changes are not neoplastic but are a consequence of repeated injury and regeneration, with entrapment of glands below the level of the muscularis mucosae. The lesions in the aspirin-treated animals closely resemble those of gastritis cystica polyposa and colitis cystica profunda in the human subject.

Screening for rectal cancer.

Although screening for rectal cancer remains a controversial topic, the case for screening has been strengthened by the demonstration that screening can reduce mortality from the disease. Three randomized controlled trials based on fecal occult blood testing have shown a significant reduction in overall colorectal cancer mortality in those offered testing. The case for screening based on flexible sigmoidoscopy is not as strong. Evidence from randomized controlled trials will not be available for another 8-10 years. Whatever screening method is chosen, factors such as the level of participation and the cost-effectiveness of protocols for adenoma follow-up will have major effects on the success of screening.

Clinical profile in Barrett's esophagus: who should be screened for cancer?

A retrospective survey identified 96 patients (58 males) with Barrett's esophagus, diagnosed at the Royal Melbourne Hospital between 1978 and 1986. The age at presentation varied from 20 to 93 years, and 43% were greater than 70 years. Heartburn was a presenting symptom in 71%, regurgitation into the pharynx in 54%, dysphagia in 31% and hematemesis or melena in 29%. At endoscopy, the length of Barrett's epithelium ranged from 3 cm to 15 cm. Macroscopic esophagitis was observed in 69%, benign esophageal strictures in 14% and a co-existent adenocarcinoma of the lower esophagus in 10% of patients. Only 30% of the patients were cigarette smokers at the time of diagnosis, but 64% drank alcohol (9% greater than 80 g alcohol daily). Patients with esophageal cancer at presentation were more likely to be male and cigarette smokers (Fisher's exact probability test). It has been suggested that patients with Barrett's esophagus should be screened to detect the early development of esophageal cancer. If patients who already have cancer, the elderly (age greater than 70 years) and those with a chronic alcohol problem (greater than 80 g intake daily) are excluded from endoscopic cancer surveillance, only 42% of the patients described in this survey would be eligible for enrollment in such a program. This represents a recruitment of only 5 new patients yearly in a large teaching hospital endoscopy unit.

Colorectal cancer screening in general practice. A survey of current practice and attitudes in Victoria.

OBJECTIVE: To determine the views and practices of Victorian general practitioners (GPs) on the early detection of colorectal cancer (CRC). METHODS: A 1995/1996 quantitative mail survey was used to determine GP knowledge and practice in relation to CRC screening and diagnosis in asymptomatic and symptomatic patients. This includes the amount of investigation, the type and process of diagnostic testing and attitudes to CRC screening. RESULTS: GPs are investigating symptomatic patients but few asymptomatic patients. In general, only patients in the increased risk groups are being investigated. Colonoscopy is the most common method of examination. Most GPs prefer faecal occult blood tests (FOBTs) to be interpreted via laboratory and endoscope procedures to be performed by a specialist. CONCLUSION: GPs were unlikely to screen asymptomatic patients unless there is a family or personal history of CRC or adenomatous polyps. Colonoscopy is the preferred practice for examining such patients. The use of FOBTs for screening standard risk asymptomatic patients was lower than recent research recommends. If population screening for CRC is endorsed, professional and public education in CRC tests, particularly for asymptomatic standard risk patients would be required. Patient initiated screening was highly favoured, supporting the need for public awareness programs.

Serum carbenoxolone: ulcer-healing and metabolic effects.

The relationship of serum carbenoxolone to dosage, age, sex, efficacy and side-effects were studied in 92 patients. Log serum carbenoxolone was correlated with dose and with age, and was higher in women. Serum levels were not significantly different in those patients with duodenal (or gastric) ulcer whose ulcers had or had not healed after allowance was made by analysis of variance for the effects of age and sex. However, oedema and hypokalaemia were associated with higher serum levels. These results support the concept that the ulcer-healing effect of carbenoxolone is topical rather than systemic.

A comparative case study of bowel cancer screening in the UK and Australia: evidence lost in translation?

OBJECTIVES: (i) To document the current state of the English, Scottish, Welsh, Northern Irish and Australian bowel cancer screening programmes, according to seven key characteristics, and (ii) to explore the policy trade-offs resulting from inadequate funding. SETTING: United Kingdom and Australia. METHODS: A comparative case study design using document and key informant interview analysis. Data were collated for each national jurisdiction on seven key programme characteristics: screening frequency, population coverage, quality of test, programme model, quality of follow-up, quality of colonoscopy and quality of data collection. A list of optimal features for each of the seven characteristics was compiled, based on the FOBT screening literature and our detailed examination of each programme. RESULTS: Each country made different implementation choices or trade-offs intended to conserve costs and/or manage limited and expensive resources. The overall outcome of these trade-offs was probable lower programme effectiveness as a result of compromises such as reduced screening frequency, restricted target age range, the use of less accurate tests, the deliberate setting of low programme positivity rates or increased inconvenience to participants from re-testing. CONCLUSIONS: Insufficient funding has forced programme administrators to make trade-offs that may undermine the potential net population benefits achieved in randomized controlled trials. Such policy compromise contravenes the principle of evidence-based practice which is dependent on adequate funding being made available.

Evaluation of radiochromium blood loss studies in unexplained iron-deficiency anaemia.

Radiochromium measurement of gastrointestinal (GI) blood loss was performed on an outpatient basis on 57 patients with unexplained iron-deficiency anaemia. With the exception of pre-menopausal women, patients were only selected for study if the anaemia remained unexplained after careful GI radiological and endoscopic examination. Occult GI blood loss was confirmed in 31 patients and further investigation including laparotomy in some cases, led to a diagnosis in 17 of the 31. Carcinoma was present in five (colon four, stomach one) and benign lesions of the small intestine in five. There was no correlation between the severity or pattern of bleeding and the nature of the lesion. Of the 26 patients without demonstrable occult bleeding, anaemia was aspirin-induced in five and caused by menorrhagia in six. With the pre-menopausal women, measurement of both GI and menstrual blood loss resulted in detection of silent GI lesions in four and gynaecological disease in six. The cause of anaemia has not been elucidated in 23 of the 57 patients, including seven of 11 with aortic valve disease, even after follow-up for an average of 3.0 years. These results indicate that radiochromium studies effectively identify the patients with unexplained iron-deficiency anaemia who require continued active investigation for a focal GI lesion.

Spontaneous bacterial peritonitis: reversible cause of deterioration in patients with cirrhosis.

Spontaneous bacterial peritonitis (SBP) in cirrhotic patients with ascites carries a mortality rate ranging from 50% to 90%. Eleven patients were diagnosed at The Royal Melbourne Hospital over a 25-year period. Fever was present in eight, abdominal pain in eight, rebound tenderness in six and pre-coma in four. In two, the peritonitis was silent. All five diagnosed by paracentesis survived, but the three patients first diagnosed at laparotomy died. The experience shows that SBP can be successfully treated with antibiotics. Because of the variable clinical features, a diagnostic paracentesis is recommended in all cirrhotic patients presenting with ascites.

Relationship between patterns of bleeding and Hemoccult sensitivity in patients with colorectal cancers or adenomas.

Although Hemoccult (Smith Kline Diagnostics, Inc., Sunnyvale, Calif.) screening for colorectal cancer is based on detection of the peroxidase activity of blood, little is known about the relationship between patterns of bleeding from cancers or adenomas and sensitivity of the test slides. 51Cr-labeled blood loss was measured for a mean of 9.1 days in 46 patients with colorectal cancer and 28 with adenomas. Duplicate sets of Hemoccult II slides were prepared by the patients and mailed for testing with and without preliminary rehydration. Mean blood loss was significantly related to the site of the cancers, but was unrelated to Dukes' staging. Geometric mean levels of blood loss from the cancers were 9.3 ml/day for cecum and ascending colon, 1.5 ml/day for transverse and descending colon, 1.9 ml/day for sigmoid colon, and 1.8 ml/day for rectum. Rehydration increased Hemoccult sensitivity for blood independently of storage delay, almost doubling sensitivity at the levels of bleeding commonly encountered with cancers of the left side of the colon. Relating results to testing over 3 days, the Hemoccult false-negative rate for cancer was 31% without rehydration, falling to 9% with rehydration. With adenomas, blood loss and Hemoccult positivity were closely related to lesion size. The findings support the validity of 3-day occult blood testing as a method of screening for colorectal cancers and large adenomas, provided the test employed has a high sensitivity for blood. Further information is needed about the specificity of rehydrated Hemoccult slides in subjects following diets of differing peroxidase content. The degree to which rehydration of slides lowers the false-negative rate for cancer indicates that this modification of slide development should be evaluated in large population screening programs.

The Mallory-Weiss syndrome.

A policy of immediate investigation of patients with haematemesis or melaena or both led to the diagnosis of the Mallory-Weiss syndrome in 16 out of 121 patients admitted to a combined medical-surgical unit over three and a half years. A typical history suggestive of the diagnosis was obtainable in only nine of the 16 patients, though recent alcohol intake was high in another four. All patients survived the episode. Establishment of the diagnosis by oesophagogastroscopy was of special benefit when surgery was needed for control of continuing blood loss, but it also simplified the subsequent medical management of those patients in whom bleeding stopped spontaneously. The incidence of 13.2% in this series suggests that the Mallory-Weiss syndrome may be a relatively common cause of upper gastrointestinal bleeding.