Detection of colorectal neoplasms by the highly sensitive

Detection of colorectal neoplasms by the highly sensitive

Int J Colorectal Dis (1999) 14:267–271
© Springer-Verlag 1999
O R I G I N A L A RT I C L E
Andreas Sieg · Christine Thoms · Kai Lüthgens
Markus R. John · Heinrich Schmidt-Gayk
Detection of colorectal neoplasms
by the highly sensitive hemoglobin-haptoglobin complex in feces
Accepted: 25 October 1999
Abstract Screening for fecal occult blood by means of
guaiac tests has an unsatisfactory sensitivity for the detection of colorectal neoplasms. The immunological determination of human hemoglobin in feces has a higher
sensitivity and specificity, but hemoglobin is degraded
during its transport through the gastrointestinal tract. We
compared the hemoglobin test to a newly developed immuno-chemiluminometric (ILMA) assay for quantifying
the hemoglobin-haptoglobin complex in feces which
shows high stability against degradation. From each of
621 patients with gastrointestinal complaints before
scheduled colonoscopy we collected two 1-ml samples
from a single stool; there were no dietary restrictions.
The sensitivity for detecting colorectal carcinomas
proved 87% with hemoglobin. With the hemoglobinhaptoglobin complex it was 87% at a cutoff level of
1.5 µg/g feces, 83% at 2.0 µg/g feces, and 78% at 2.5 and
3.0 µg/g feces. The sensitivity for detecting large adenomatous polyps was 54% with hemoglobin, 76% with the
hemoglobin-haptoglobin complex at a cutoff point of
1.5 µg/g feces, 73% with the hemoglobin-haptoglobin
complex at 2.0 and 2.5 µg/g feces, and 65% with the hemoglobin-haptoglobin complex at 3.0 µg/g feces. The
optimal cutoff point for the hemoglobin-haptoglobin
complex was estimated to be 2.0 µg/g stool. The specificity for hemoglobin (99%) was significantly higher
than that for the hemoglobin-haptoglobin complex at
2.0 µg/g feces (96%). Immunological determination of
the hemoglobin-haptoglobin complex in feces has a
comparable sensitivity as the fecal hemoglobin assay for
A. Sieg (✉)
Practice of Gastroenterology, Hauptstrasse 45,
D-76684 Östringen, Germany
e-mail: [email protected],
Tel.: +49-6221-21806, Fax: +49-6221-27454
K. Lüthgens · M.R. John · H. Schmidt-Gayk
Laboratory Group, University of Heidelberg,
Heidelberg, Germany
A. Sieg · C. Thoms · M.R. John · H. Schmidt-Gayk
Department of Medicine, University of Heidelberg,
Heidelberg, Germany
colorectal carcinomas and a significantly higher sensitivity for adenomatous polyps but a significantly lower
specificity. Its use for colorectal cancer prevention is
currently being evaluated in a screening study.
Key words Colorectal cancer · Colonic polyps ·
Hemoglobin-haptoglobin complex · Occult blood ·
Hemoglobin
Introduction
Guaiac tests are widely used in colorectal cancer screening to detect fecal occult blood (FOB), although they
have an unsatisfactory sensitivity for detecting colorectal
neoplasms in asymptomatic patients [1–4]. In a group of
248 patients with positive fecal guaiac tests who underwent colonoscopy followed by esophago-gastro-duodenoscopy upper gastrointestinal lesions were identified
more frequently than colonic lesions [5]. FOB screening
with guaiac tests reveal significantly fewer colorectal
neoplasms than sigmoidoscopy [6]. A major disadvantage of guaiac tests is the fact that they are not specific
for human blood and can produce positive results when
meat, fruit, or vegetables containing peroxidase have
been ingested [7]. The rate of false-negative results is increased by large amounts of vitamin C [8]. Nevertheless,
the mortality rate from colorectal cancer could reportedly be reduced by 15–18% with biennial screening using
nonrehydrated guaiac tests [9, 10] and by 33% with annual screening using rehydrated guaiac tests [11]. A meta-analysis with four randomized and two nonrandomized trials of Hemoccult screening showed a 16% reduction in mortality from colorectal cancer [12]. When adjusted for attendance at screening, this reduction was
23% for persons actually screened [12]. Given the more
sophisticated immunological screening techniques to detect constituents of human blood in feces, it should be
possible to reduce mortality from colorectal cancer even
further. A higher sensitivity and specificity of immunological tests for hemoglobin (Hb) in the detection of co-
268
um azide, and 0.4 g EDTA per liter. The diluted samples were mixed
thoroughly and centrifuged for 15 min at 3000 g, and the supernatants were measured in luminescence immunoassays specific for human Hb or the HbHp complex. The luminescence immunoassay for
Hb was set up according to Gao et al. [21] and that for the HbHp
complex according to Lüthgens et al. [20]. The upper limit of normal
for human hemoglobin was set at 10 µg/g feces, according to our earlier findings [16]. For the HbHp complex the optimal threshold value
was to be found. The detection limit was about 1 µg/g feces for human Hb and 0.03 µg/g feces for the HbHp complex.
Diagnosis was based on the endoscopic and histopathological
findings. Carcinomas were classified according to the UICC stages I–IV of the TNM system, and location. Adenomatous polyps
were classified as to histopathological characteristics, size (large
polyps: >1 cm; small polyps <1 cm) and location. We defined the
left colon as consisting of the rectum, sigmoid, descending colon,
and the splenic flexure; and the right colon as consisting of the
cecum, ascending colon, hepatic flexure, and transverse colon.
The results were classified as true-positive if a neoplasm (carcinoma or large adenomatous polyp) was found, and as false-positive if a normal colonic mucosa and no cause of extracolonic gastrointestinal bleeding was found. In patients with upper abdominal
complaints and normal colonic mucosa the upper gastrointestinal
tract was examined by esophago-gastro-duodenoscopy. Sensitivity
and specificity are expressed below as percentages defined in the
standard manner [22–24]. Fecal Hb and the HbHp complex were
compared by McNemar’s χ2 test [25]; P<0.05 was considered as
statistically significant.
lorectal neoplasms in comparison to guaiac tests has
been reported in screening [1, 4, 13] and evaluation studies [14, 15], and Hb has proven more sensitive than albumin in detecting colorectal neoplasms [16].
The Hb molecule, however, is partly degraded during
passage through the gastrointestinal tract and may become
unrecognizable to the antibody in the assay. This is supported by the finding in our previous study that 2 of 12
right-sided cancers were not detected by the Hb assay, in
contrast to 31 left-sided cancers which all were detected
[16]. This problem can be overcome by employing an immunoassay for α1-antitrypsin [17] which degrades at a
much lower rate [18], or an immunological test of the Hbhaptoglobin (Hp) complex which shows high stability in
gastric juice and fecal extracts [19]. The aim of this study
was to compare a newly developed immuno-chemiluminometric assay (ILMA) of the HbHp complex [20] with our
standard luminescence immunoassay of Hb in feces in an
evaluation study for the detection of colorectal neoplasms.
Methods
Human Hb and the HbHp complex in feces were determined immunologically in 621 patients (280 men, 341 women; aged 15–85
years, median 59) scheduled for colonoscopy. The patients had
been referred to a practice for gastroenterology by their family
physicians for the investigation of gastrointestinal symptoms. The
patients collected 1-ml samples into 1-ml air-tight stool specimen
tubes (Sarstedt, Nümbrecht-Rommelsdorf, Germany) from two
different sites of a single stool at least 2 days prior to examination.
Collection into vials provides a more representative amount of feces than paper smears and allows a quantitative determination of
the constituents. No dietary restrictions were made. The samples
were stored in the deep-freeze for 1 week before examination.
Those handling the laboratory results were blinded to the colonoscopic results, and vice versa.
The laboratory samples were thawed, weighed, and diluted tenfold with a phosphate buffer pH 7.4 containing 2 g gelatin, 1 g sodi-
Table 1 Sensitivity of immunologically determined fecal
human Hb and the HbHp complex (at various cutoff values:
µg/g stool) in patients with carcinomas and large adenomatous polyps: percentages of
positive test results
Results
In 23 patients with colorectal carcinoma the sensitivity
of fecal Hb was 87% (Table 1). The sensitivity of the
HbHp complex depended on the cutoff value and was
87% at 1.5 µg/g stool, 83% at 2.0 µg/g stool, and 78% at
2.5 and 3.0 µg/g stool. The sensitivity of fecal Hb in 37
patients with large colorectal adenomas was 54% and
that of the HbHp complex was 76% at 1.5 µg/g stool,
73% at 2.0 and 2.5 µg/g stool, and 65% at 3.0 µg/g stool.
The number patients with early carcinoma (UICC stage
Findings
n
Hb
HbHp 1.5
HbHp 2.0
HbHp 2.5
HbHp 3.0
Carcinomas (all cases)
Left-sided
Right-sided
23
22
1
87
86
100
87
86
100
83
82
100
78
77
100
78
77
100
Large adenomas (all cases)
Left-sided
Right-sided
37
30
7
54
53
57
76
77
71
73
77
57
73
77
57
65
70
57
5
60
60
40
40
40
Early cancers (UICC stage I)
Table 2 Specificity of fecal human Hb and the HbHp complex (at various cutoff values: µg/g stool) in 357 patients with normal colonic
mucosa
Hb
n
Positive test results
Positive test results
without reason for GI bleeding
Specificity
19
3
%
5
1
99
HbHp 1.5
HbHp 2.0
HbHp 2.5
HbHp 3.0
n
%
n
%
n
n
47
25
13
7
36
15
10
4
31
12
93
96
%
9
3
97
25
8
%
7
2
98
269
1.0
0.9
0.8
Sensitivity
0.7
Hemoglobin
Hemoglobin - haptoglobin
0.6
0.5
0.4
0.3
0.2
0.1
0
0
0.1
0.2
0.3
a
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1- Specificity
1.0
0.9
0.8
Sensitivity
0.7
Hemoglobin
Hemoglobin - haptoglobin
0.6
0.5
0.4
0.3
0.2
I; n=5) was too small for statistical evaluation, as was
the number of right-sided carcinomas and adenomas.
Of 357 patients with normal colonic mucosa 19 had
positive hemoglobin (5%), 47 (13%) positive HbHp
complex at 1.5 µg/g stool, 36 (10%) at 2.0 µg/g stool, 31
(9%) at 2.5 µg/g stool and 25 (7%) at 3.0 µg/g stool
(Table 2). In 26 of these 47 patients we found the following causes of gastrointestinal bleeding: gastric ulcer, 1;
erosive gastritis, 2; duodenal ulcer, 3; enlarged hemorrhoids, 18; coumarine therapy, 1; nonsteroidal antiinflammatory drugs, 2; Oslers’ disease with epistaxis,
1. Excluding these patients from the group with normal
colonic mucosa and positive FOB tests, only 21 patients
remained with false-positive findings. Thus the specificity of the test, defined by false-positive results if a normal
colonic mucosa and no other cause of gastrointestinal
bleeding was found, was 99% for Hb (Table 2). For the
HbHp complex the specificity was 93% at 1.5 µg/g stool,
96% at 2.0 µg/g stool, 97% at 2.5 µg/g stool, and 98% at
3.0 µg/g stool. The receiver operating characteristic
curves of Hb and the HbHp complex for the detection of
carcinomas and adenomatous polyps are shown in Fig. 1.
The optimal cutoff point for the HbHp complex was
calculated at 2.0 µg/g stool. With this value we compared
the sensitivity and specificity for Hb, the HbHp complex, and the combined test (Table 3). The positive predictive value of the combined test was 14% for carcinomas, 19% for large adenomas, and 33% for colorectal
neoplasms (carcinomas and large adenomas). The results
for miscellaneous diagnoses other than colorectal cancer
and large adenomatous polyps are shown in Table 4.
0.1
Discussion
0
0
b
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1- Specificity
Fig. 1a,b Receiver operating characteristic curves for fecal human Hb and the HbHp complex. a In 23 patients with colorectal
carcinoma. b In 37 patients with large adenomatous polyps
We found the sensitivity of the HbHp complex in detecting colorectal cancers to be comparable or, at higher cutoff levels, slightly below that of the Hb assay. However,
the HbHp complex and the combined test were signifi-
Table 3 Sensitivity and specificity of fecal human Hb (normal value <10 µg/g stool), the HbHp complex (normal value <2 µg/g stool),
the combined test, and the significance of the χ2 test for Hb vs. HbHp at a cutoff of 2.0 µg/g feces
Sensitivity for colorectal carcinoma (n=23)
Sensitivity for large adenomas (n=37)
Specificity
Table 4 Sensitivity of fecal Hb
(normal value <10 µg/g stool),
the HbHp complex (normal
value <2 µg/g stool), and the
combined test in patients with
miscellaneous diagnoses: positive test results
Hb (%)
HbHp (%)
P
Hb+HbHp (%)
87
54
99
83
73
96
NS
<0.05
<0.05
87
73
96
Diagnoses
Small adenomas (<10 mm)
Hyperplastic polyps
Diverticulosis
Diverticulitis
Ulcerative colitis
Crohn’s disease
Rectal ulcer
Unspecific colitis
n
94
35
47
5
12
8
1
2
Hb
HbHp
Hb+HbHp
n
%
n
%
n
%
11
1
1
3
9
6
0
1
12
3
2
60
75
75
19
3
5
5
11
6
1
2
20
9
11
100
92
75
22
3
6
5
11
6
1
2
23
9
13
100
92
75
270
cantly superior to Hb in detecting large adenomatous
polyps. This offers important progress in screening since
adenomas are precursors of most of the carcinomas and
systematic removal of adenomas may reduce the incidence of colorectal cancers by up to 90% [26]. The optimal cutoff point for the HbHp complex was determined
to be 2 µg/g feces. The sensitivity in the present study
was in the same range or even better than that reported
by other authors for hemoglobin tests [1, 4, 13–15], and
far higher than the values reported for guaiac tests without rehydration [1–4, 13–15], although we examined only two samples from a single stool. This is more convenient for the participants in a screening program and, according to a previous study [27], results in better compliance of 83% – compared to 30–67% reported for guaiac
tests [4, 9–11]. However, our study was not performed
under screening conditions, under which the values for
sensitivity would possibly be lower due to the small proportion of participants with symptoms. In the specific
setting of a gastroenterological practice the positive predictive value of the combined test, describing the possibility of having the disease if the test is positive, was
14% for colorectal carcinomas, 19% for large adenomas,
and 33% for colorectal neoplasms (carcinomas and large
adenomas). These values are comparable to those obtained in screening studies with Hb, in which the respective positive predictive values were 7.6% [27] and 5.0%
[4] for colorectal cancer and 49.6% [27] and 20.5% [4]
for all neoplasias, and comparable to nonrehydrated Hemoccult tests with positive predictive values of 6.6% [4]
and 12% [9] for colorectal cancer and 23.2% [4] and
46% [9] for all neoplasias, but superior to studies using
the rehydrated Hemoccult test, with a positive predictive
value of only 2.2% for colorectal cancer [11]. The low
probability of detecting cancer after a positive FOB test
is due to the wide variety of diseases that may produce a
positive FOB test. Especially with the immunological
tests a positive result may, on the one hand, be false-positive for cancer but, on the other, true-positive for another disease of the gastrointestinal tract. This is not atypical for screening tests.
Specificity is an important determinant of the cost of
FOB screening. Even small changes in test specificity result in large effects on the number of false-positives in a
screening study, which are responsible for unnecessary
colonic examinations [28]. A 33% annual reduction in
colorectal cancer mortality has been reported possible
for patients screened by a rehydrated Hemoccult II test
[11]. Rehydration, however, decreases the specificity of
the test to 90.4% resulting in a high false-positive rate;
colonoscopy has been reported as necessary in 38% of
screened individuals [11]. In our study the specificity –
defined by false-positive results if a normal colonic mucosa and no other cause of gastrointestinal bleeding was
found – was 99% for hemoglobin. This was significantly
higher than the values for the HbHp complex and the
combined test (96%, P<0.05). This rather high specificity of immunological FOB tests was obtained under conditions of an evaluation study in patients with symptoms.
The specificity in a screening study is expected to be
higher. For the Hb test the specificity under screening
conditions was 99.5% [27].
Whether Hb alone, the HbHp complex, or the combined test should be used for colorectal cancer screening
depends on the decision of public health authorities. In
general, compliance and sensitivity of FOB tests are important determinants of the cost-effectiveness of colon
screening programs [29–31]. A major advantage of immunological FOB tests is the possibility of varying the
normal range and thereby affecting sensitivity and specificity according to the risk of the population screened
and according to the financial resources of the country.
Fecal hemoglobin can be examined in every laboratory with standard equipment, although it is not a “bedside test.” According to our experience, a bedside test is
not required for a mass screening of healthy adults. The
time required to obtain a test result is similar for the Hemoccult and for our immunological test. The Hemoccult
test has a sample collection period of at least 3 days.
The immunological test has a sample collection period
of only 1 day, and analysis can be carried out in 1 day.
The costs of the immunological hemoglobin test or the
HbHp complex in Germany are about four times those
of the Hemoccult test. Previous data have shown that
cost of screening by immunological methods are less
than the savings from avoiding carcinoma costs by endoscopic polypectomy of Dukes’ A carcinomas [27,
32].
Immunological determination of the fecal HbHp complex exhibits a high sensitivity for colorectal neoplasms,
especially for adenomatous polyps, at the expense of decreased specificity. Whether immunological assays of fecal Hb, the HbHp complex, or a combined test should be
used in colorectal cancer screening can be decided after
evaluation of an ongoing screening study comparing fecal Hb and the HbHp complex.
Acknowledgements We are indebted to W. Bersch, M.D., and to J.
Bertling, M.D., Speyer, Germany for the histopathological evaluation.
References
1. Frommer DJ, Kapparis A, Brown MK (1988) Improved
screening for colorectal cancer by immunological detection of
occult blood. BMJ 96:1092–1094
2. Allison JE, Feldman R, Tekawa IS (1990) Hemoccult screening in detecting colorectal neoplasm: sensitivity, specificity,
and predictive value: long-term follow-up in a large group
practice setting. Ann Intern Med 112:328-333
3. Ahlquist DA, Wieand HS, Moertel CG, McGill DB, Loprinzi
CL, O’Connell MJ, Maillard JA, Gerstner JB, Pandya K,
Ellefson RD (1993) Accuracy of fecal occult blood screening
for colorectal neoplasia: a prospective study using Hemoccult
and Hemoquant tests. JAMA 269:1262–1267
4. Allison JE, Tekawa IS, Ransom LR, Adrain AL (1996) Comparison of fecal occult-blood tests for colorectal cancer screening. N Engl J Med 334:155–159
5. Rockey DC, Koch J, Cello JP, Sanders LL, McQuaid K (1998)
Relative frequency of upper gastrointestinal and colonic lesions in patients with positive fecal occult-blood tests. N Engl
J Med 339:153–159
271
6. Verne JECW, Aubrey R, Love SB, Talbot IC, Northover JMA
(1998) Population based randomised study of uptake and yield
of screening by flexible sigmoidoscopy compared with screening by faecal occult blood testing. BMJ 317:182–185
7. Macrae FA, St John DJB, Calagiore P, Taylor LS, Legge WJ
(1982) Optimal dietary conditions for Hemoccult testing. Gastroenterology 82:899–903
8. Jaffe RM, Kasten B, Young DS, MacLowry JD (1975) Falsenegative stool occult blood tests caused by ingestion of ascorbic acid (vitamin C). Ann Intern Med 83:824–826
9. Hardcastle JD, Chamberlain JO, Robinson MHE, Moss SM,
Satya SA, Balfour TW, James PD, Mangham CM (1996) Randomised controlled trial of faecal-occult-blood screening for
colorectal cancer. Lancet 348:1472–1477
10. Kronborg O, Fenger C, Olson J, Jörgensen OD, Söndergaard
O (1996) Randomised study of screening for colorectal cancer
with faecal-occult-blood test. Lancet 348:1467–1471
11. Mandel JS, Bond JH, Church TR, Snover DC, Bradley MG,
Schumann LM, Ederer F (1993) Reducing mortality from colorectal cancer by screening for fecal occult blood. N Engl J
Med 328:1365–1371
12. Towler B, Irwig L, Glasziou P, Kewenter J, Weller D, Silagy C
(1998) A systematic review of the effects of screening for colorectal cancer using the faecal occult blood test, Hemoccult.
BMJ 317:559–565
13. Robinson MHE, Marks CG, Farrands PA, Whynes DK, Bostok
K, Hardcastle JD (1995) Is an immunological faecal occult
blood test better than Haemoccult? A cost-benefit study. Eur J
Surg Oncol 21:261–264
14. Hope RL, Chu G, Hope AH, Newcombe RG, Gillespie PE,
Williams SJ (1996) Comparison of three faecal occult blood
tests in the detection of colorectal neoplasia. Gut 39:722–725
15. St John DJB, Young GP, Alexeyeff MA, Deacon MC, Cuthbertson AM, Macrae FA, Penfold JC (1993) Evaluation of new
occult blood tests for detection of colorectal neoplasia. Gastroenterology 104:1661–1668
16. Sieg A, Scheida M, John MR, Hertel A, Schröter M, Lüthgens
K, Schmidt-Gayk H (1998) Validity of new immunological
human hemoglobin and albumin tests in detecting colorectal
neoplasms – an endoscopy-controlled study. Z Gastroenterol
36:485–490
17. Morrow RJ, Lawson N, Hussaini SH, Asquith P (1990) The
usefulness of faecal haemoglobin, albumin and alpha-1-antitrypsin in the detection of gastrointestinal bleeding. Ann Clin
Biochem 27:208–212
18. Florent C, L’Hirondel C, Desmazures C, et al (1981) Intestinal
clearance of alpha-1-antitrypsin. A sensitive method for the
detection of protein losing enteropathy. Gastroenterology
81:777–780
19. Meguro T (1994) Measurement of fecal hemoglobin-haptoglobin complex as a new diagnostic tool of lower gastrointestinal
tract diseases. Hokkaido Igaku Zasshi 69:995–1009
20. Lüthgens K, Maier A, Kampert I, Sieg A, Schmidt-Gayk H
(1998) Hemoglobin-haptoglobin complex. A highly sensitive
assay for the detection of fecal occult blood. Clin Lab 44:543–
551
21. Gao P, John MR, Schmidt-Gayk H, Arndt B, Scheida M,
Theuer D (1995) Solid-phase competitive luminescence immunoassay for lysozyme in faeces. Clin Chim Acta 239:167–
177
22. Abel U (1993) Die Bewertung diagnostischer Tests. Hippokrates, Stuttgart
23. Royall RM (1976) Current advances in sampling theory: implications for human observational studies. Am J Epidemiol
104:463–474
24. Saah AJ, Hoover DR (1997) Sensitivity and specificity reconsidered: the meaning of these terms in analytical and diagnostic settings. Ann Intern Med 126:91–94
25. Zöfel P (1992) Statistik in der Praxis. Fischer, Stuttgart
26. Winaver SJ, Zauber AG, Ho MN, O’Brien MJ, Gottlieb LS,
Sternberg SS, Waye JD, Schapiro M, Bond JH, Panish JF,
Ackroyd F, Shike M, Kurtz RC, Hornsby-Lewis L, Gerdes H,
Stewart ET, and the National Polyp Study Workgroup (1993)
Prevention of colorectal cancer by colonoscopic polypectomy.
N Engl J Med 329:1977–1981
27. Sieg A Hertel A, John MR, Lüthgens K, Schmidt-Gayk H
(1998) Screening for colorectal neoplasms with a new immunological human faecal haemoglobin and albumin test. Eur J
Cancer Prev 7:279–285
28. Ransohoff DF, Lang CA (1997) Screening for colorectal cancer with the fecal occult blood test: a background paper. Ann
Intern Med 126:811–822
29. Liebermann DA (1995) Cost-effectiveness model for colon
cancer screening. Gastroenterology 109:1781–1790
30. Toribara NW (1995) Sleisinger MH. Screening for colorectal
cancer. N Engl J Med 332:861–867
31. Winaver SJ, Fletcher RH, Miller L, Godlee F, Stolar MH,
Mulrow CD, Woolf SH, Glick SN, Ganiaths TG, Bond JH,
Rosen L, Zapka JG, Olsen SJ, Giardello FM Sisk JE, Van Antwerp R, Brown-Davis C, Marciniak DA, Mayer RJ (1997)
Colorectal cancer screening: clinical guidelines and rationale.
Gastroenterology 112:594–642
32. Sieg A, Schröter M, Hertel A, Schmidt-Gayk H (1996) Die
Prävention des kolorektalen Karzinoms nach dem Östringer
Modell ist kosteneffektiv. Z Gastroenterol 9:610(A)