402 Gut 2001;49:402–408 Faecal and faecal occult blood tests in the diagnosis of colorectal carcinoma and adenoma

J Tibble, G Sigthorsson, R Foster, R Sherwood, M Fagerhol, I Bjarnason

Abstract cancer registrations and 55 000 deaths each Background and aims—Testing for faecal year1 while in the UK there are an annual occult blood has become an accepted 28 000 cancer registrations and 19 000 deaths technique of non-invasive screening for due to this disease.2 Survival rates are closely colorectal neoplasia but lack of sensitivity related to the stage of cancer at the time of remains a problem. The aim of this study diagnosis and the most promising approach to was to compare the sensitivity and specifi- reducing mortality rates is early detection of city of faecal calprotectin and faecal precancerous or cancerous lesions. There is occult blood in patients with colorectal now overwhelming epidemiological evidence cancer and colonic polyps. and molecular biological data to substantiate Methods—Faecal calprotectin and occult previous suggestions of the colonic adenoma- blood were assessed in 62 patients with carcinoma progression.3–5 Collectively, such colorectal carcinoma and 233 patients data have increased the pressure to develop referred for . The range of novel approaches for colon cancer detection, normality for faecal calprotectin (0.5–10.5 critical for secondary prevention through mass mg/l) was determined from 96 healthy population screening whereby early diagnosis subjects. of will detect tumours with Results—Median faecal calprotectin con- the best prognosis and result in improved sur- centration in the 62 patients with colorec- vival rates. tal carcinoma (101 mg/l, 95% confidence The most widely accepted non-invasive interval (CI) 57–133) diVered significantly method for detecting colorectal cancer is faecal from normal (2.3 mg/l, 95% CI 1.6–5.0) occult blood (FOB) testing. Screening in with 90% of patients having elevated levels asymptomatic populations have, at best, re- 6–8 (normal <10 mg/l) whereas only 36/62 duced mortality rates by 15–33%. There are (58%) had positive faecal occult bloods. however many problems with screening using There was no significant diVerence in fae- FOB. The sensitivity of the most commonly cal calprotectin levels when considering used guaiac based FOB tests may be as low as 9 location or Dukes’ staging of tumour. Per- 26% which means that 74% of patients with centage positivity of faecal occult bloods malignant lesions will remain undetected by was significantly higher for Dukes’ stage C this method, presumably because blood loss and D cancers compared with Dukes’ A from the tumour may be intermittent or below the detection threshold (2–4 ml of blood/100 g Department of and B. In the colonoscopy group, 29 Medicine, Guy’s, patients with adenomatous polyps were stool). The test may also not be suitable for Kings, St Thomas’ screening of precancerous adenomas which detected in whom the median faecal 10 11 Medical School, calprotectin was 12 mg/l (95% CI 2.9–32). often do not bleed. There are also practical Bessemer Road, diYculties with certain types of FOB tests London, UK Sensitivity for detection of adenomatous polyps was 55% using the calprotectin which require patients to provide three stool J Tibble samples while subject to some dietary G Sigthorsson method and 10% using faecal occult blood 12–14 R Foster testing. The overall sensitivity and specifi- restrictions. I Bjarnason city of calprotectin for colorectal cancer In order to increase the detection rate, clini- and adenomatous polyps as a combined cians have sought alternative methods for Department of Clinical detection of early colorectal cancer/adenomas. Biochemistry, Guy’s, group was 79% and 72%, respectively, compared with a sensitivity and specifi- A large trial of flexible sigmoidoscopy (com- Kings, St Thomas’ promising the non-invasive nature of investiga- Medical School, city of faecal occult blood of 43% and 92%. tion) in an asymptomatic population is now Bessemer Road, Conclusions—Faecal calprotectin is a 15 16 London, UK simple and sensitive non-invasive marker underway in the USA and the UK to assess R Sherwood of colorectal cancer and adenomatous whether this will lead to significantly improved survival for colorectal cancer, but are not Department of polyps. It is more sensitive than faecal occult blood tests for detection of colorec- expected to yield results until 2008, while flex- Immunology, Ullevaal ible sigmoidoscopy is now widely used as an University Hospital, tal neoplasia at the cost of a somewhat initial examination in symptomatic patients. Oslo, Norway lower specificity. M Fagerhol (Gut 2001;49:402–408) Colorectal cancer is associated with a local acute inflammatory reaction so that in some Correspondence to: Keywords: colorectal cancer; faecal occult blood cases it can be visualised by white cell neutro- Dr J Tibble, Department of testing; calprotectin 17 Gastroenterology, 1st Floor phil scanning. Calprotectin is a stable neutro- College House, St Thomas’ phil specific marker which can be assayed in Hospital, Lambeth Palace stool with high precision and ease.18 Within the Road, London SE1 7EH, Colorectal cancer is the second commonest calprotectin is found in the extra UK. [email protected] cause of death from malignancy in the Western Accepted for publication world. In the USA it accounts for 14% of can- Abbreviations used in this paper: FOB, faecal 26 February 2001 cer deaths with about 134 000 colorectal occult blood.

www.gutjnl.com Faecal calprotectin and colorectal carcinoma 403

lysosomal cytosol and constitutes up to 60% of shown to cause an and raised fae- the total protein content.19 Levels of faecal cal calprotectin,24 and faecal sampling was per- calprotectin are increased in patients with formed when patients had no evidence of colorectal cancer but immunohistochemical respiratory tract infection.18 23 examination of colorectal cancer specimens has All patients attended a preoperative or shown reactivity confined to neutrophilic precolonoscopic assessment clinic where the granulocytes with no reactivity seen in neoplas- nature of the study was explained. Patients tic cells,20–22 suggesting that elevated faecal lev- provided stool samples prior to surgery/ els may be due to neutrophil shedding from an colonoscopy and those in the colonoscopic ulcerated tumour. The purpose of this study group underwent additional investigation if was to assess whether faecal calprotectin is an relevant where the cause for symptoms after improvement on the sensitivity and specificity undergoing colonoscopy were not determined of FOB in current use as a non-invasive and therefore all patients were given a final biochemical marker for colorectal cancer and diagnosis. All stool samples were provided colorectal polyps. prior to any administered bowel preparation. The study was approved by the King’s Patients and methods Healthcare local research ethics committee and AIMS all patients gave informed consent. The aims of the study were to (i) assess and compare the sensitivity of faecal calprotectin Faecal occult blood tests and FOB for colorectal carcinoma in sympto- For 48 hours prior to collection of samples for matic patients with colorectal cancer and relate FOB testing, all patients were required to the results to cancer site or degree of invasion; observe a red meat free diet, avoid eating raw (ii) assess and compare the sensitivity of the fruits/vegetables containing peroxidase-like two tests for colorectal polyps subsequently substances such as turnips, broccoli, and pars- found at colonoscopy; and (iii) assess and nips, and avoid ingestion of preparations compare the specificity of the two tests for containing vitamin C. Patients completed a detecting colorectal cancer/premalignant Hema screen card (Immunostics. Inc., 3505 polyps. Sunset Ave Ocean, New Jersey 07712, USA), a guaiac based FOB, on three consecutive stool PATIENTS samples which was returned on the day of the Ninety six healthy volunteers (51 males, 45 last sample and analysed unrehydrated on the females, mean age 41 years) provided single day of arrival for FOB. Patients were asked to stool samples to obtain a reference range for follow the instructions on the information faecal calprotectin. Controls were predomi- sheet provided with the test pack. nantly healthcare workers and their relatives, who had no symptoms of gastrointestinal or Faecal calprotectin other diseases, in particular respiratory tract Patients provided a single stool sample for cal- infections,18 23 on direct questioning. Other protectin measurement and returned it on the exclusion criteria for controls were as for same day. Stool samples were frozen on receipt patients studied. at −20°C. After thawing, a single 5 g aliquot18 25 Sixty two consecutive patients (22 females, was suspended in 10 ml of faecal extraction 40 males; mean age 68 years (range 36–92)) buVer (Tris buVered isotonic (150 mM) saline,

diagnosed with colorectal cancer by barium with 10 mM CaCl2, and 0.25 mM thiomersal enema or colonoscopy were recruited to as an antimicrobial agent, pH 8.4) and homog- provide faecal samples for calprotectin and enised for 45 seconds at 20 000 rpm with an occult blood testing; 56/62 (90%) of these Ultra Turrax (Ika Werke, Germany) mechani- patients underwent operative resection of the cal homogeniser. The homogenates were cen- tumour and six patients had tumours which trifuged for 15 minutes at 10 000 g at room were considered inoperable due to advanced temperature. The upper portion of the super- metastatic spread. natants were pipetted oV, frozen, and stored at A total of 233 consecutive patients (130 −20°C until quantitation by ELISA. females, 103 males; mean age 56 years (range Microtitre plates were coated by adding 17–91)) referred for colonoscopy for polyp fol- 200 µl of an IgG fraction of a rabbit anticalpro- low up, colorectal cancer surveillance, iron tectin26 diluted 1:2000 in phosphate buVered deficiency anaemia, or investigation of colonic saline to each well. The plates were covered symptoms thought to be compatible with or with mylar foil and kept at 4°C until use. Cal- suggestive of colonic neoplasia were also protectin standards22 23 3.75–60 mg/l were pre- recruited to provide samples for faecal calpro- pared by diluting purified calprotectin in assay tectin and occult blood testing. No patient had buVer: Tris (50 mM) buVer containing 150

a previous diagnosis of colorectal disease mM NaCl, 0.5 mM MgCl2, 2.5 mM KCl, 0.25 except for those undergoing polyp/cancer mM thiomersol, 0.05% Tween 20, and 10 g/l surveillance where the most recent colonos- bovine serum albumin (pH 8.0). Before use the copy was at least 12 months before recruit- microtitre plates were washed four times in this ment. All polyps had to have been removed at buVer, less the bovine serum albumin. The fro- the previous colonoscopy. Patients taking non- zen faecal extracts were thawed and diluted steroidal anti-inflammatory drugs and those 1:20 and 1:200 (and dilutions of 1:400–6400 with a high alcohol intake (>14 units/week for when required) in the assay buVer. Standards females, >21 units/week for males) were and diluted samples (100 µl) were added to the excluded from the study as these have been plates in duplicate which were covered and

www.gutjnl.com 404 Tibble, Sigthorsson, Foster, et al

1000 1.0

0.8 + 100 0.6 _ + 0.4 Sensitivity 10 0.2 + + _ _ 0.0 _ 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1 _ Log calprotectin concn (mg/l) Log 1 specificity Controls Normal/ Adenomatous Cancer irritable polyps Figure 2 Receiver operating curve for faecal calprotectin bowel in detecting patients with colorectal neoplasia (carcinoma syndrome and adenoma). The curve is statistically significant (p<0.0001) with the area beneath it being greater than Figure 1 Log faecal calprotectin concentration (mg/l) in 50%, and shows the eVect of increasing the cut oV level for the diVerent diagnostic groups. Median values with 95% calprotectin on the sensitivity and specificity for detecting (+) and 5% (–) confidence intervals are shown. All three colorectal neoplasia. diagnostic groups diVered significantly from controls (normal/, p=0.01; adenomatous Statistics polyps, p=0.0003; cancer p<0.0001). Statistical analysis of calprotectin levels be- tween groups was performed using the non- incubated at room temperature for 45 minutes parametric Mann-Whitney test, as the results were not normally distributed. The 2 test was on a plate shaker with an agitating speed of ÷ used to compare the percentage positivity of 600/minute. The wells were then washed four FOB between the diVerent groups. Sensitivities times with buVer and 100 µl of alkaline and specificities of the FOB and calprotectin phosphatase conjugated anticalprotectin (dilu- tests were calculated, with those for calprotec- tion 1:1600) was added to each well, the plate tin at diVerent cut oV values being determined covered, and incubated at room temperature by construction of an ROC curve (plot of sen- for 45 minutes on the plate shaker with an agi- sitivity against 1−specificity at diVerent levels tating speed of 600/min. Thereafter the wells of calprotectin—see fig 2) including all patients were washed four times and 100 µl of substrate in the colorectal cancer and colonoscopy solution was added (2×5 mg p-nitro phenol- groups but not those in the control group. The phosphate tablets dissolved in 10 ml of ROC curve enables determination of sensitiv- substrate buVer which contained 10% dieth- ity and specificity for disease at diVerent cut oV anolamine with 0.5 mM MgCl, 0.25 mM values for a continuous variable. The specificity thiomersal, pH adjusted to 9.6 with HCl). and sensitivities of the two tests diVer depend- Optical density (405 nm: measured on a Micro ing on the disease in question and we have Tracer plate reader; Syva, Milton Keynes, UK) therefore calculated these for colorectal cancer of the highest standard was monitored and the and adenomatous polyps as a combined group. reaction stopped by adding 50 µl ofa1M In our view an ideal screening test detects pre- NaOH solution to each well when its optical malignant disease. Therefore, a positive test in density read between 1.2 and 1.5, similar to the the premalignant group was not regarded as a method previously described.18 false positive. All statistics were performed using SPSS for Windows 95. Histopathological assessment Each carcinoma specimen obtained at opera- Results tion was examined histopathologically by the The median calprotectin concentration in the same pathologist and classified according to 96 normal controls was 2.3 mg/l (95% CI 1.6– 5.0). The 95 percentile indicates an upper ref- Dukes’ stage and location. Polyps in all cases erence limit of 10.0 mg/l which we have chosen were obtained by endoscopic snare removal as our cut oV value for normality. and classified, on the basis of pathology Median faecal calprotectin concentration in reports, in terms of their histopathological the 62 patients with colorectal carcinoma (101 characteristics, size, and location. The loca- mg/l, 95% CI 57–133 mg/l) (fig 1) diVered sig- tions were classified for both polyps and nificantly from controls (p<0.0001). The cancers as rectum, sigmoid, descending, trans- sensitivity of calprotectin for colorectal cancer verse (including hepatic and splenic flexures), (90%) was significantly greater (÷2=26.6, and right colon (ascending colon and caecum). p<0.0001) than that for FOB testing (58%). Table 1 Faecal calprotectin levels and percentage of cancers detected by calprotectin and There was no significant diVerence in calpro- faecal occult blood testing according to Dukes’ staging of the tumour tectin levels (p>0.2) according to the degree of invasion of the tumour using Dukes’ staging Calprotectin (mg/l) (table 1) whereas there was a significantly Positive by Positive by higher positivity rate of FOBs in Dukes’ stages Dukes’ stage n Median Range calprotectin (%) FOB (%) C and D compared with Dukes’ A and B 2 A: Confined to bowel wall 10 62.5 7–933 90 20 (÷ =14.4, p<0.0001). All patients detected by B: Spread through bowel wall 24 115 2–3770 88 46 the FOB test had elevated levels of faecal C: Involvement of regional lymph nodes 14 62 1.5–314 86 100 calprotectin. The location of cancers within the D: Distant metastases 14 132 10.5–3388 100 71 colon was 31 rectal, 21 sigmoid, one descend- FOB, faecal occult blood. ing, four transverse, and five in the right colon.

www.gutjnl.com Faecal calprotectin and colorectal carcinoma 405

There was no significant diVerence (p>0.5) in mg/l (95% CI 8.5–33.5) which diVered signifi- calprotectin values or positivity of FOBs cantly from normal (p<0.001). In these between these diVerent tumour locations. patients 32 (64%) had elevated levels of In the colonoscopy group 51 patients with a calprotectin while 13 (26%) had at least one total of 86 polyps were detected of whom 22 positive FOB. had hyper/metaplastic (median 2 polyps, range The sensitivities and specificities of faecal 1–5) and 29 had tubular/tubulovillous adeno- calprotectin and occult bloods in diVerent mas (median 1 adenoma, range 1–3). The diagnostic groups are shown in table 3. When range in size of the adenomas was 0.5–5.5 cm considering the colonoscopy and colorectal (median 2). Median calprotectin was 4.5 mg/l cancer groups as a whole, the sensitivity of cal- (95% CI 2.9–14) in the hyper/metaplastic protectin for colorectal cancer and premalig- group and 12 mg/l (95% CI 2.9–32 mg/l) in the nant (adenomatous) polyps (that is, malignant adenoma group (fig 1) both of which diVered and premalignant disease as a combined group) was significantly greater than for FOB significantly from normal (p=0.003 and 2 p=0.0003, respectively). There was no signifi- testing (79% v 43%, ÷ =27.3, p<0.0001). The specificity of calprotectin for colorectal cancer cant diVerence (p>0.2) in calprotectin levels and adenomas as a joint group at a cut oV level between the diVerent sizes of adenomatous of 10 mg/l was significantly lower than for FOB polyps, between adenoma locations (13 proxi- testing (72% v 92%, ÷2=5.7, p=0.01). How- mal to the splenic flexure, 16 distal to the ever, if the sensitivity of calprotectin in splenic flexure—most distal polyp recorded identifying patients with significant colorectal where more than one present), or between disease is considered—that is, all patients with patients with diVerent numbers of polyps. The colorectal cancer and those in table 2 except sensitivity for detecting tubular/tubulovillous those with haemorrhoids, fissures, and meta- adenomas using faecal calprotectin was 55%. plastic polyps—then the sensitivity of calpro- The FOBs were positive in only three cases of tectin at a level of 10 mg/l is 77% with a tubular/tubulovillous adenomas (sensitivity specificity of 84% compared with FOB testing 10%). with a sensitivity of 39% and specificity of Of the 233 patients who underwent colonos- 97%. copy, a final diagnosis of normal or irritable bowel syndrome was made in 128 patients, of whom 109 (85%) had normal calprotectin lev- Discussion els while 98% had negative FOBs. Median fae- We have made a direct comparison of FOB cal calprotectin in this group was 5.0 mg/l tests and faecal calprotectin in detecting colo- (95% CI 2.0–6.5), which as a group diVered rectal neoplasia. The sensitivity for colorectal significantly from normal controls (p=0.01) cancer using calprotectin was 90% in compari- (fig 1) but mostly remained well below the cut son with 58% for FOB testing. This accords oV value for normality. The remaining 54 well with the findings of Kristinsson and patients had a variety of diagnoses (table 2). In colleagues21 who demonstrated a sensitivity of this group (excluding the four patients with 94% for colorectal cancer using calprotectin. colorectal cancer who are analysed in the can- The 60% sensitivity of our FOB test is similar cer group) median faecal calprotectin was 23 to that seen in a number of previous studies11 27 using guaiac based non-rehydrated FOB tests Table 2 Diagnoses in the 233 consecutive patients who underwent colonoscopy. The four in symptomatic patients with colorectal cancer. patients with colorectal cancers are included in the colorectal cancer group when assessing sensitivity of calprotectin and faecal occult blood (FOB) testing for colorectal cancer In a recent review assessing the sensitivity of FOB testing in asymptomatic non-referred No with faecal populations,9 Ahlquist found that the overall calprotectin >10 mg/l No with positive FOB sensitivity for colorectal cancer was 26%. Sen- Diagnosis Total No No % No % sitivity estimates in large Haemoccult screen- ing trials has been reported to be as high as Normal/IBS 128 19 15 3 2 81%.8 However, these are calculated using an Adenomatous polyps 29 16 55 3 10 Hyper/metaplastic polyps 22 6 27 2 9 approximation of the prevalence of cancer Colorectal cancer 4 4 100 4 100 among test negative screenees based on cancers Diverticulitis 24 15 63 3 13 Crohn’s/ulcerative 13 12 92 8 61 that present during an arbitrary period of Haemorrhoids 6 0 0 0 0 follow up. The poor sensitivity of the guaiac Angiodysplasia 4 2 50 1 25 based tests have led a number of authors28–30 to Infective diarrhoea 1 1 100 0 0 Pneumatosis coli 1 1 100 1 100 recommend the immunological based occult Anal fissure 1 1 100 0 0 blood tests with higher chemical sensitivity for Hb (2 mgHb/100 g faeces).31 However, be- IBS, irritable bowel syndrome. cause of the greater cost and complexity of the Table 3 Sensitivities and specificities (%) of faecal calprotectin (at a level of 10 mg/l) and test, their use in screening programmes for faecal occult bloods (FOBs) in the diVerent diagnostic groups colorectal cancer has not been widespread, although in time immunochemical tests for Calprotectin FOBs blood may be automated, reducing their com- Sensitivity Specificity Sensitivity Specificity plexity considerably. (%) (%) (%) (%) It is estimated that blood loss of >20 ml/day Colorectal cancer 90 58 is required to achieve 80–90% positivity of Adenomatous polyps 55 10 FOB tests32–34 and the fluctuating levels of Colorectal cancer +adenomatous polyps 79 72 43 92 blood loss and intermittent bleeding of cancers Significant intestinal disease 77 84 39 97 compromises the sensitivity of the FOB test.

www.gutjnl.com 406 Tibble, Sigthorsson, Foster, et al

Calprotectin, although present in blood, even less and more intermittent than for carci- enters the bowel lumen as part of an inflamma- nomas. A number of authors38–41 have reported tory process rather than bleeding from the negative Haemoccult reactions in >75% of tumour, as it has been calculated20 that blood patients with endoscopically proved adenomas, losses of over 300 ml/day would be required to while MacRae and St John11 found that only produce a median calprotectin level of 33 mg/l. polyps larger than 2 cm produced consistent Calprotectin is a cytosolic protein in neu- detectable bleeding. trophilic granulocytes and macrophages26 35 36 To maximise the mortality benefits of colo- and immunohistochemical examination of rectal cancer screening it will be necessary to colorectal cancer specimens has shown calpro- detect and treat adenomatous lesions, which tectin reactivity confined to such cells with no may result in a fall in colorectal cancer reactivity in neoplastic cells.22 Thus the high incidence, as well as detecting early malignant levels of calprotectin seen in patients with lesions with the best prognosis. In our study the colorectal carcinoma are likely to be due to overall sensitivity of calprotectin for colorectal polymorphonuclear cell infiltration of the cancer and adenomatous polyps was 79% tumour and subsequent shedding into the compared with 43% for FOB testing, accepting intestinal lumen. Recruitment of such cells to that this is in a symptomatic group of patients the tumour is almost certainly due to local pro- possibly overestimating the sensitivity values of duction of chemotactic factors possibly pro- both tests. The reduction in colorectal cancer mortality seen in the University of Nottingham duced in response to a breach in the protective 6–8 mucosal lining. Hence calprotectin sensitivity Minnesota trials (15–33%, respectively) for colorectal cancer is not likely to be aVected using guaiac based FOB screening were due to significantly by variations in tumour blood loss, detection of cancers at an earlier stage and not and may account for the non-significant diVer- due to the detection of premalignant lesions. The colorectal cancer incidence was the same ence in faecal calprotectin levels between the in both the screened and non-screened groups. diVerent Dukes’ stages of cancer, results which If we consider case control studies using would need to be confirmed in an asympto- sigmoidoscopy which detects both cancers and matic population. Patterns of bleeding may adenomas,42–45 they have demonstrated a 50% significantly aVect the sensitivity of FOB tests reduction in the risk of developing colorectal and may explain why we observed higher cancer and a 59% reduction in mortality from sensitivities for Dukes’ stages C and D than for cancers reached by the sigmoidoscope. Thus stages A and B. Similar observations were with the much higher overall sensitivity of cal- made by Frommer et al who found improved protectin for colorectal cancer/adenomatous sensitivity for Dukes’ A tumours compared polyps it has the potential for a greater with Dukes’ D when using non-rehydrated 28 reduction in mortality than that provided by Hamoccult II. St John et al demonstrated the currently recommended FOB tests. sensitivities of 63.6% for stage A and B If calprotectin is to be considered for use in tumours compared with 89.7% for stages C large screening programmes the issue of 37 and D when using the HaemoQuant FOB. A specificity has to be addressed. Sensitivity and logical explanation for FOB sensitivity diVer- specificity have an inverse relationship—the ing according to stage of disease might be that sensitivity of the calprotectin test may be more advanced tumours bleed more consist- improved by lowering the cut oV point but at ently and to a greater extent, thereby improving the cost of a lowered specificity for the disease the number of positive FOB tests. However, in in question. The specificity of the test has the study of St John et al assessing levels of important implications on the overall cost of a blood loss in patients with colorectal cancer, screening programme. From our study we can mean daily blood loss was dependent only on only make an assessment of specificity in a site of tumour (right sided greater than left) group of patients most of whom had symptoms and was not aVected by tumour stage.37 of colorectal disease and were therefore more It may be argued that comparison between likely to have intestinal inflammation from a faecal calprotectin and FOBs is unfair based on variety of causes than an asymptomatic popu- awareness of the inherent problems associated lation. The comparative specificities of calpro- with guaiac based tests, in particular their lower tectin and FOBs for detecting colorectal cancer sensitivities for rectal and right sided tumours,7 and adenomatous polyps as a combined group especially as 50% of our tumours were rectal. were 72% and 92%. By analysis of the receiver Our objective however was simply to compare operating curve (fig 2), the specificity of the the two diVerent stool tests (calprotectin v calprotectin test could be improved to 80% FOB) for the detection of colorectal cancer, with only a small reduction in the sensitivity irrespective of location. It is of interest that (75%) using a cut oV level of 15 mg/l. Such an despite the accepted inherent limitations of the exercise would need to be performed when FOB tests it has been proposed to use them in assessing the calprotectin assay in an asympto- a national screening programme for colorectal matic population in order to determine the cancer in the UK. “best” cut oV value, depending on the relative We have demonstrated a marked diVerence importance given to sensitivity and specificity. in the sensitivity of calprotectin and FOB for If of equal importance then the value which colorectal adenomas—55% versus 10%, re- gives the greatest sum of these two parameters spectively. The limitations of FOBs are even should be chosen. more marked when screening for these prema- However, if we consider the specificities for lignant lesions where quantitative blood loss is detecting all significant colorectal disease

www.gutjnl.com Faecal calprotectin and colorectal carcinoma 407

(table 2 diagnoses, except haemorrhoids, 3 Muto T, Bussey HJ, Morson BC. The evolution of cancer of the colon and rectum. Cancer 1975;36:2251–70. fissures, and hyper/meta plastic polyps), values 4 Morson B. President’s address. The polyp-cancer sequence increase to 84% and 97% with sensitivities of in the large bowel. Proc R Soc Med 1974;67:451–7. 77% and 39% for calprotectin and FOBs, 5 Stryker SJ, WolV BG, Culp CE, et al. Natural history of untreated colonic polyps. Gastroenterology 1987;93:1009– respectively. Due to the nature of calprotectin it 13. is likely that it will detect patients with signifi- 6 Kronborg O, Fenger C, Olsen J, et al. Randomised study of screening for colorectal cancer with faecal-occult-blood cant inflammation within the gastrointestinal test. Lancet 1996;348:1467–71. tract, including those with inflammatory bowel 7 Hardcastle JD, Chamberlain JO, Robinson MHE, et al. Ran- domised controlled trial of faecal-occult-blood screening disease, diverticulitis, and small bowel enter- for colorectal cancer. Lancet 1996;348:1472–7. opathies, explaining the levels of calprotectin 8 Mandel JS, Bond JH, Church TR, et al. Reducing mortality from colorectal cancer by screening for fecal occult blood. seen in such patients in our study. Although Minnesota Colon Cancer Control Study. N Engl J Med screening programmes for colorectal cancer are 1993;328:1365–71. 9 Ahlquist DA. Fecal occult blood testing for colorectal not intended to detect such conditions, they cancer: Can we aVordtodothis?Gastroenterol Clin North are often significant diseases for which eVective Am 1997;26:41–55. treatment is available and therefore identifying 10 Herzog P, Holtermuller KH, Preiss J, et al. Fecal blood loss in patients with colonic polyps: A comparison of measure- such patients who may benefit from colono- ments with (51)chromium-labeled erythrocytes and with scopic assessment could be regarded as a posi- the Haemoccult test. Gastroenterology 1982;83:957–62. 11 MacRae FA, St John DJB. Relationship between patterns of tive aspect of the calprotectin test. Further bleeding and Hemoccult sensitivity in patients with evaluation of the sensitivity and specificity of colorectal cancers or adenomas. Gastroenterology 1982;82: 891–8. faecal calprotectin will need to be performed in 12 MacRae FA, St John DJB, Caligiore P, et al. Optimal dietary asymptomatic non-referred populations but it conditions for Hemoccult testing. Gastroenterology 1982;82: 899–903. may be that in such a group there would be 13 JaVe RM, Kasten B, Young DS, et al. False-negative stool fewer patients with inflammatory conditions occult blood tests caused by ingestion of ascorbic acid (vitamin C). Ann Intern Med 1975;83:824–6. and the specificity of calprotectin for colorectal 14 JaVe RM, Zierdt W. A new occult blood test not subject to cancer/adenomas might be higher than the false-negative results from reducing substances. J Lab Clin Med 1979;93:879–86. 72% seen in our study. 15 Institute NC. Concept approval granted to trial of prostate, Compliance is an important factor to take lung, colorectal and ovarian screens. Cancer Lett 1989;15: into account when considering the use of a test 1–3. 16 Atkin WS. Flexible sigmoidoscopy as a mass screening tool. in a screening programme. Compliance rates Eur J Gastroenterol Hepatol 1998;10:219–23. for FOB screening in trials each oVering 17 Saverymuttu SH, Maltby P, Batman P, et al. False positive localisation of indium-111 granulocytes in colonic carci- screening to over 10 000 subjects were on aver- noma. Br J Radiol 1986;59:773–7. age 63%,46 ranging from 35% to 92%.47 48 The 18 Roseth AG, Fagerhol MK, Aadland E, et al. Assessment of the neutrophil dominating protein calprotectin in feces. A sensitivity of FOB testing is dependant on par- methodologic study. Scand J Gastroenterol 1992;27:793–8. ticipants providing three stool samples and it 19 Fagerhol MK, Andersson KB, Naess-Andresen CF, et al. Calprotectin (the L1 leukocyte protein). In: Smith VL, would seem probable that compliance rates Dedman JR, eds. Stimulus response coupling: the role of intra- would be improved if only one sample was cellular calcium-binding proteins. Boca Raton, Florida: CRC 18 Press Inc, 1990:187–210. required as is the case for faecal calprotectin. 20 Gilbert JA, Ahlquist DA, Mahoney DW, et al. Fecal marker In addition, compliance of screening for colo- variability in colorectal cancer: calprotectin versus hemo- globin. Scand J Gastroenterol 1996;31:1001–5. rectal cancer in asymptomatic high risk groups 21 Kristinsson J, Roseth A, Fagerhol MK, et al. Fecal using colonoscopy may be improved if patients calprotectin concentration in patients with colorectal carci- are prescreened with faecal calprotectin noma. Dis Colon Rectum 1998;41:316–21. 22 Roseth AG, Kristinsson J, Fagerhol MK, et al. Faecal whereby an elevated level may reinforce the calprotectin: a novel test for the diagnosis of colorectal can- need for a screening colonoscopy. cer? Scand J Gastroenterol 1993;28:1073–6. 23 Meling TR, Aabakken L, Roseth A, et al. Faecal calprotectin In summary, faecal calprotectin appears to shedding after short-term treatment with non-steroidal be a simple non-invasive surrogate marker of anti-inflammatory drugs. Scand J Gastroenterol 1996;31: 339–44. inflammation in patients with colorectal 24 Tibble JA, Sigthorsson G, Foster R, et al. High prevalence of cancer/adenomatous polyps. It has a number of NSAID enteropathy as demonstrated by a simple faecal test. Gut 1999;45:362–6. significant advantages over FOB testing in the 25 Tibble JA, Teahon K, Roseth A, et al. A simple method for detection of colorectal neoplasia, most notably assessing intestinal inflammation in Crohn’s disease. Gut 2000;47:506–13. a sensitivity of 79% compared with 43% when 26 Dale I BP, Fagerhol MK, Scott H. Distribution of a new considering malignancy and polyps as a whole. myelomonocytic antigen (L-1) in human peripheral blood leukocytes. Am J Clin Pathol 1985;84:24–34. Improved sensitivity however comes at the cost 27 Allison JE, Feldman R, Tekawa IS. Hemoccult screening in of diminished specificity which has cost impli- detecting colorectal neoplasm: Sensitivity, specificity, and predictive value: Long-term follow-up in a large group cations because of additional investigation. The practice setting. Ann Intern Med 1990;112:328–33. faecal calprotectin method seems to be a useful 28 Frommer DJ, Kapparis A, Brown MK. Improved screening adjuvant to the investigation of patients at high for colorectal cancer by immunological detection of occult blood. BMJ 1988;296:1092–4. risk for colorectal neoplasia but its value in a 29 St John DJB, Young GP, AlexeyeV MA, et al. Evaluation of truly asymptomatic population such as those new occult blood tests for detection of colorectal neoplasia. Gastroenterology 1993;104:1661–8. who would be considered for entry into a 30 Robinson MHE, Marks CG, Farrands PA, et al. Screening national screening programme for colorectal for colorectal cancer with an immunological faecal occult blood test: 2-year follow-up. Br J Surg 1996;83:500–1. cancer needs to be determined. 31 Frommer D, Kapparis A. Faecal occult blood testing. Lancet 1983;ii:738. Dr Tibble was funded by the NHS Executive South Thames 32 Morris DW, Hansell JR, Ostrow JD, et al. Reliability of Regional OYce, grant No RDF 039. chemical tests for fecal occult blood in hospitalized patients. Am J Dig Dis 1976;21:845–52. 33 Stroehlein JR, Fairbanks VF, McGill DB, et al. Hemoccult Conflict of interest. Professor Fagerhol shares patent rights on detection of fecal occult blood quantitated by radioassay. calprotectin. Am J Dig Dis 1976;21:841–4. 34 Doran J, Hardcastle JD. Bleeding patterns in colorectal 1 American Cancer Society. Cancer facts and figures. Atlanta: cancer: The eVect of aspirin and the implications for faecal American Cancer Society, 1996. occult blood testing. Br J Surg 1982;69:711–13. 2 Hart AR, Wicks AC, Mayberry JF. Colorectal cancer 35 Fagerhol MK, Dale I, Anderson I. Release and quantitation screening in asymptomatic populations. Gut 1995;36:590– of a leucocyte derived protein (L1). Scand J Haematol 8. 1980;24:393–8.

www.gutjnl.com 408 Tibble, Sigthorsson, Foster, et al

36 Steinbakk M, Naess Andresen CF, Lingaas E, et al. Antimi- mortality from colorectal cancer. N Engl J Med 1992;326: crobial actions of calcium binding leucocyte L1 protein, 653–7. calprotectin. Lancet 1990;336:763–5. 43 Winawer SJ, Flehinger BJ, Schottenfeld D, et al. Screening 37 St John DJB, Young GP, McHutchinson JG, et al. Compari- for colorectal cancer with fecal occult blood testing and son of the specificity and sensitivity of Haemoccult and sigmoidoscopy. J Natl Cancer Inst 1993;85:1311–18. HaemoQuant in screening for colorectal neoplasia. Ann 44 Muller AD, Sonnenberg A. Protection by endoscopy against Intern Med 1992;117:376–82. death from colorectal cancer: A case-control study among 38 Winawer SJ. Screening for colorectal cancer: An overview. veterans. Arch Intern Med 1995;155:1741–8. 1980; :1093–8. 45 Muller AD, Sonnenberg A. Prevention of colorectal cancer Cancer Philadelphia 45 by flexible endoscopy and polypectomy. A case-control 39 Crowley ML, Freeman LD, Mottet MD, . Sensitivity of et al study of 32 702 veterans. Ann Intern Med 1995;123:904– guaiac-impregnated cards for the detection of colorectal 10. neoplasia. J Clin Gastroenterol 1983;5:127–30. 46 Van Dam J, Bond JH, Sivak MV Jr. Fecal occult blood 40 Eddy DM. Computer assisted evaluation of screening programs. screening for colorectal cancer. Arch Intern Med 1995;155: Washington: US Department of Health, Education, and 2389–402. Welfare, 1979. 47 Morris JB, Stellato TA, Guy BB, et al. A critical analysis of 41 Lipshutz GR, Katon RM, McCool MF, et al. Flexible the largest reported mass fecal occult blood screening pro- sigmoidoscopy as a screening procedure for neoplasia of gram in the United States. Am J Surg 1991;161:101–6. the colon. Surg Gynecol Obstet 1979;148:19–22. 48 Zoubek V, Zoubkova H. Results of screening for colorectal 42 Selby JV, Friedman GD, Quesenberry CP Jr, et al.A carcinoma in the district of Bruntal using the Haemoccult case-control study of screening sigmoidoscopy and test in 1985–1988. Czech Med 1990;13:2–3.

www.gutjnl.com