Assessment of faecal occult blood loss by qualitative and quantitative methods

M.D. BARBER*, A. ABRAHAM*, W.G. BRYDON†, B.M. WALDRON* and A.J.K. WILLIAMS#

Departments of *Surgery and #Medicine, St John’s Hospital, Livinsgton, Scotland, U.K. †Gastrointestinal Laboratory, Western General Hospital, Edinburgh, Scotland, U.K.

Introduction Methods Results

Discussion Conclusion References

J.R.Coll.Surg.Edinb., 47, April 2002, 491-494

Background: Various methods exist for the assessment of faecal occult blood loss in a patient with suspected gastrointestinal blood loss. Methods: The present study examined the effectiveness and financial implications of a qualitative guaiac-based method (Haemoccult) of faecal occult blood detection and a quantitative measure of haemederived porphyrins (Hemoquant) in 184 patients who underwent assessment of faecal blood loss by both methods over a three year period during assessment of iron deficiency anaemia. Main findings: At least one Haemoccult test was positive in 72.2% of patients while Hemoquant was suggestive of significant blood loss (>2mg haemoglobin/g faeces) in 29.9%. Patients underwent a total of 324 further endoscopic or radiological investigations of which 76.5% demonstrated no abnormality. A diagnosis was reached in 60 patients (32.6%). A significant potential source of gastrointestinal bleeding was found in 48 patients (26.1%). Hemoquant achieved a sensitivity of 62.5% and a specificity of 81.6% while with Haemoccult it was 85.4% and 32.4%, respectively. Hemoquant was normal in 18 patients with significant gastrointestinal conditions including peptic ulcers and colonic polyps. While Haemoccult only missed 7 lesions, two of these were colonic cancers. The quantitative nature of the Hemoquant test gave little clue as to diagnosis. Conclusion: Neither of the tests examined was ideal but Hemoquant had an overall better performance and further investigation of patients with evidence of blood loss from this test should be mandatory.

Keywords: faecal occult blood testing, gastrointestinal bleeding, iron deficiency anaemia

INTRODUCTION

Iron deficiency anaemia due to suspected gastrointestinal blood loss is a common cause of referral to medical and surgical outpatient clinics.¹ Assessment of faecal occult blood loss has an important role in guiding the investigation of a patient presenting with iron deficiency anaemia.² Qualitative guaiac-based methods are cheap, easily performed, give instant results and have a relatively high sensitivity.³ Quantitative measures of faecal haeme-derived porphyrins are more costly, complicated and time-consuming and may have a lower sensitivity in screening.⁴ However, guaiac methods may have a poor positive predictive value and are sensitive to factors such as local trauma and stool hydration while faecal porphoryin measurement may have a better specificity and so may prevent further expensive and invasive investigations.⁵

The present study aimed to assess the ability of a guaiacbased faecal occult blood test, Haemoccult, and a quantitative porphyrin test, Hemoquant, to detect significant gastrointestinal pathology in patients with iron deficiency anaemia. This information was then used to examine the cost implications of guiding radiological and endoscopic investigations based on various strategies of faecal occult blood measurement.

METHODS

Patients attending the clinic of a single consultant gastroenterologist for investigation of iron deficiency anaemia with suspected gastrointestinal blood loss, between 1/7/95 and 30/9/98, underwent assessment of faecal occult blood loss by Haemoccult and Hemoquant methods. A total of 199 patients underwent both tests and the case notes of 184 patients were examined. Results of further investigations and long-term outcome until 30/9/99 were recorded. Symptomatology provided the initial guide to further investigation.

Haemoccult tests (SKD Pharma, Griesham, Germany) were performed on specimens obtained during rectal examination. Samples of stool were smeared onto the windows of the test card and developed. A positive result was recorded if a blue colour appeared within 30 seconds. Equivocal Haemoccult tests were recorded when results from repeated samples differed. Hemoquant tests were performed on single, spontaneously passed stool specimens as described by Schwarz et al (1983). ⁶ A faecal haemoglobin of >2mg/g faeces was used to define significant blood loss. No dietary restrictions were imposed before faecal specimens were taken. The reference range used was determined from 20 healthy volunteers and 10 patients with irritable bowel syndrome on an unrestricted diet.

Significant gastrointestinal lesions which have the potential to cause blood loss were defined as malignant or adenomatous polyps, Barratt’s oesophagus, inflammatory bowel disease, peptic ulceration, angiodysplasia and coeliac disease. Diverticular disease was also considered where evidence of this as a source of blood loss existed.

Statistical analysis was performed using Mann-Whitney U test or Kruskal Wallis test, as appropriate.

RESULTS

Of the 184 patients studied, 62.5% were female. Mean age was 59 years (range 16-92). Median follow-up was 35 months (range 12-50). Patients underwent a total of 324 further endoscopic or radiological investigations. These investigations revealed no abnormality on 76.5% of occasions. Barium enema identified or was suspicious of colonic cancer in only 5 of the 9 patients with this condition on whom it was performed. A diagnosis was reached in 60 patients (32.6%). In 48 patients (26.1%), a significant cause of gastrointestinal bleeding was identified (Table 1).

* One patient had a gastric ulcer, colonic angiodysplasia and coeliac disease; Ba: barium

Table 1: Diagnosis achieved in 184 patients investigated for suspected gastrointestinal blood loss, results of faecal occult blood tests and ultimate investigation performed

At least one Haemoccult test was positive in 72.2% of patients while Hemoquant suggested significant blood loss in 29.9%. The two tests agreed in 85 patients (46.2%). Overall performance of the two tests separately and combined is shown in Table 2. Hemoquant was normal in 19 patients with significant gastrointestinal lesions, which may cause blood loss (Table 1). Haemoccult was negative in 7 such lesions but two of these were colonic cancers (Table1). Both tests were negative for 4 lesions.

Table 2: Performance of haemoccult and hemoquant tests in 184 patients investigated for suspected gastrointestinal blood loss

The quantitive nature of the Hemoquant test gave little clue as to the diagnosis but all patients with malignant disease had significant blood loss (range 2.2-86.9mg/g). True positive Hemoquant samples had significantly higher faecal haemoglobin levels than false positives (median 8.3 mg/g versus 4.8mg/g, p = 0.046).

It was estimated that endoscopic or radiological assessment of the upper and lower gastrointestinal tract cost around £500, Haemoccult measurement cost around £0.50 and Hemoquant measurement cost around £20. Using data from the present study, estimated costs of investigation strategies based on faecal occult blood results of 100 theoretical patients with suspected gastrointestinal blood loss are shown in Table 3.

Table 3: Estimated cost of investigation strategies based on faecal occult blood results of 100 patients with suspected gastrointestinal blood loss

DISCUSSION

The present study examined 184 patients with iron deficiency anaemia and suspected gastrointestinal blood loss to compare the usefulness of two methods of faecal occult blood detection.

Previous studies investigating patients with suspected gastrointestinal blood loss have identified a source of bleeding in a similar low proportion of patients (between 20 and 50%), the precise proportion depending on the criteria used to define pathology.7-9 The spontaneous healing of lesions, intermittent bleeding from significant lesions, the occurrence of significant lesions in asymptomatic individuals without anaemia and non-gastrointestinal causes of anaemia all influence the incidence and detection of pathology. It remains possible that pathology was missed in the present study as not all patients underwent bi-directional imaging in an attempt to minimise patient inconvenience and risk. However, the lengthy follow-up in the present study should have minimised this. A number of the pathological lesions identified would not necessarily be expected to bleed and, thus, the performance values obtained represent a worst case scenario.

It has been suggested that Hemoquant testing provides the highest likelihood of a positive faecal occult blood test when blood is lost throughout the gastrointestinal tract.10 Guaiac and immunochemical tests do not detect haemoglobin degraded by digestion. This would imply that the Hemoquant test should have a relatively high sensitivity for the detection of lesions in the upper gastrointestinal tract. This was not found in the present study with the majority of lesions missed by Hemoquant being found in the upper gut. However, in contrast to previous studies which have suggested Haemoccult to have a superior sensitivity for malignant lesions than Hemoquant the present study found a 100% sensitivity for malignant lesions with Hemoquant using a cut off of 2mg/g faeces. ^{4, 5} As with previous studies, Hemoquant showed a limited ability to identify patients with colonic polyps.5, 11 Previous studies have, however, not found such good sensitivity in detecting colonic cancer although results depend on the faecal haemoglobin threshold used.^{4, 11}

It has also been suggested that Hemoquant sensitivity may be better for right-sided tumours (9 of the 13 colonic tumours in the present study). ¹¹ While Haemoccult has been found to be a sensitive predictor of gut pathology, this occurs at the cost of a poor specificity.^{3, 4} Similar findings were noted in the present study with apparently unnecessary investigations being performed in 50% of the patients in the present study group. Of more concern is that two patients with colonic cancers (15%) were not identified using Haemoccult. A substantial number of lesions of the upper gastrointestinal tract were identified in patients with a positive Haemoccult test, despite the theoretical inability of this test to detect upper gastrointestinal bleeding. It is possible that these cases represented false positives who happened to be among the 10% of healthy subjects who have abnormalities on upper gastrointestinal endoscopy.¹² In addition, the acquisition of samples from rectal examination may have increased the number of false positive tests.

The sensitivity of detection of gastrointestinal pathology would have been improved by performing both the Hemoquant and Haemoccult tests and acting on any positive result. However, this would have increased the proportion of unnecessary investigation to 55% with a resultant increase in costs and risks associated with investigation. Acting only if both tests agreed on a positive result would reduce the number of unnecessary investigations but would miss lesions including cancers. It would appear that using Hemoquant as a screening test and investigating those with a positive result or significant symptoms may offer the best compromise in terms of workload, cost, unnecessary investigation and missed lesions.

CONCLUSION

The Hemoquant test would appear to a useful guide to those patients with suspected gastrointestinal blood loss requiring further radiological or endoscopic investigation.

REFERENCES

1. Frewin R, Henson A and Provan D. Iron deficiency anaemia. Br Med J 1997; 314: 360-3
2. Itzkowitz SH. Heads or tails of a positive faecal occult blood test. Lancet 1998; 352: 1490
3. Towler B, Irwig L, Glaszion P, Kewenter J,Weller D and Silagy C. A systematic review of the effects of screening for colorectal cancer using the faecal occult blood test, Hemoccult. Br Med J 1998; 317: 559-65
4. St John DJB, Young GP, Alexeyeff MA, Deacon MC, Cuthbertson AM, Macrae FA and Penfold JCB. Evaluation of new occult blood tests for detection of colorectal neoplasia. Gastroenterol 1993; 104: 1661-8
5. Ahlquist DA, Wieland HS, Moertel CG, McGill DB, Loprinzi CL, O’Connell MJ, Mailliard JA, Gerstner JB, Pandaya K and Ellefson RD. Accuracy of faecal occult blood screening for colorectal neoplasia. A prospective study using Hemoccult and Hemoquant tests. J Am Med Assoc 1993; 269: 1262-7
6. Schwarz S, Dahl J, Ellefson M and Ahlquist DA. The “HemoQuant” test: a specific and quantitative assay of haeme (hemoglobin) in faeces and other materials. Clin Chem 1983; 29: 2061-7
7. Hsia PC and Al-Kawas FH. Yield of upper endoscopy on the evaluation of asymptomatic patients with haemoccult-positive stool after a negative colonoscopy. Am J Gastroenterol 1992; 87: 1571-4
8. Zuckerman G and Benitez J. A prospective study of bidirectional endoscopy (colonoscopy and upper endoscopy) in the evaluation of patients with occult gastrointestinal bleeding. Am J Gastroenterol 1992; 87: 62-6
9. Chen YK, Gladden DR, Kestenbaum DJ and Collen MJ. Is there a role for upper gastrointestinal endoscopy in the evaluation of patients with occult blood-positive stool and negative colonoscopy. Am J Gastroenterol 1993; 88: 2026-9
10. Rockey DC. Occult gastrointestinal bleeding. N Engl J Med 1999; 341: 38-46
11. St John DJB, Young GP, McHutchison JG, Deacon MC and Alexeyeff MA. Comparison of the specificity and sensitivity of Hemoccult and HemoQuant in screening for colorectal neoplasia. Ann Intern Med 1992; 117: 376-82
12. Akdamar K, Ertan A, Agrawal NM, McMahon FG and Ryan J. Upper gastrointestinal endos-copy in normal asymptomatic volunteers. Gastrointest Endosc 1986; 32: 78-80

Copyright: 11 January 2002

Correspondence: M.D. Barber, Breast Unit, Western General Hospital, Edinburgh EH4 2XU, Scotland, U.K. E-mail: barbermd@hotmail.com