Role of fecal calprotectin test in the work-up of IBS patients

Abstract

Dear Sir:We read with great interest the article from Tibble et al.,1Tibble JA Sigthorsson G Foster R Forgacs I Bjarnason I. Use of surrogate markers of inflammation and Rome criteria to distinguish organic from nonorganic intestinal disease.Gastroenterology. 2002; 123: 450-460Abstract Full Text Full Text PDF PubMed Scopus (340) Google Scholar in the August 2002 issue of Gastroenterology. Based on the sensitivity (Se) and specificity (Sp) of the calprotectin test to differentiate organic gastrointestinal disease from IBS, the authors propose the use of this test in combination with Rome I criteria to “screen” tertiary referred IBS patients for further work-up.There are a number of methodological issues in the study that we believe are worth further analysis.First, the use of an inadequate criterion standard (“gold standard”).2Ransohoff DF Feinstein AR. Problems of spectrum and bias in evaluating the efficacy of diagnostic tests.N Engl J Med. 1978; 299: 926-930Crossref PubMed Scopus (1315) Google Scholar In the study from Tibble et al., the criterion standard varies from individual to individual, since patients were seen by 4 different gastroenterologists and the tests performed in each individual was determined by each investigating clinician's choice. This may have significantly influenced the final diagnosis in each individual and impacts negatively on the internal validity of the study.Second, the inclusion of a narrow spectrum of patients.2Ransohoff DF Feinstein AR. Problems of spectrum and bias in evaluating the efficacy of diagnostic tests.N Engl J Med. 1978; 299: 926-930Crossref PubMed Scopus (1315) Google Scholar Thus, patients with diseases outside the gastrointestinal tract were excluded from the study. This may lead to a biased estimate of the performance characteristics of the test and hence, limits the external validity of the results.Third, the authors describe the performance characteristics of the different tests using Se, Sp, positive, and negative predictive values. The use of predictive values can be misleading since they depend on the prevalence of the disease in the population in whom the test is performed.3Weissler AM. A perspective on standardizing the predictive power of noninvasive cardiovascular tests by likelihood ratio computation: 1. Mathematical principles.Mayo Clin Proc. 1999; 74: 1061-1071Abstract Full Text Full Text PDF PubMed Scopus (25) Google Scholar On the other hand, likelihood ratios (LR), which depend only on the operating characteristics of the test, are thought to be more appropriate and allow assessment of the usefulness of the test in all possible scenarios.4Jaeschke R Guyatt GH Sackett DL. Users' guides to the medical literature. III. How to use an article about a diagnostic test. B. What are the results and will they help me in caring for my patients? The Evidence-Based Medicine Working Group.JAMA. 1994; 271: 703-707Crossref PubMed Scopus (2132) Google ScholarThus, using the Se and Sp reported in the article, we calculated the LRs for each test under different scenarios, and reached conclusions that differed from those in Tibble's article. For example, in the tertiary referral population in this study (pretest probability of organic disease = 44%), the posttest probability of organic disease based solely on absence of Rome I criteria is 79% (positive LR = 4.7). This is similar to the posttest probability of organic disease based solely on a positive calprotectin test, 77% (positive LR = 4.2). The combination of the absence of Rome I criteria plus a positive C-reactive protein test (positive LR = 2.1) yields a posttest probability of organic disease of 91%. This probability of organic disease, with a much simpler blood screening test that is universally applicable, is high enough to make it reasonable to perform further investigation regardless of the results of fecal calprotectin. This information limits the usefulness of the calprotectin test to “rule in” organic disease in a tertiary care practice.On the other hand, if Rome criteria are present (negative LR = 0.27), the posttest probability of organic disease is 21%, and the addition of a negative calprotectin test (negative LR = 0.11) decreases that probability to 4% compared with the 14% in case of a negative C-reactive protein test (negative LR = 0.6). Thus, it appears that in the presence of Rome I criteria for IBS, fecal calprotectin might be useful to “rule out” organic disease. It would have been of interest to include in the analysis the combination of presence/absence of “red flags” or alarm symptoms and Rome I criteria, since these are of critical value to discriminate organic from nonorganic patients with GI complaints on formal analysis5Vanner SJ Depew WT Paterson WG et al.Predictive value of the Rome criteria for diagnosing the irritable bowel syndrome.Am J Gastroenterol. 1999; 94: 2912-2917Crossref PubMed Scopus (234) Google Scholar and are used extensively in clinical practice. Thus, it is important to know whether adding absence of “red flags” to the presence of Rome criteria for IBS would decrease the value of fecal calprotectin test to “rule out” organic disease.Finally, Tibble et al. state that the use of calprotectin test in a primary care population should be deferred until the test is validated in such population. We beg to disagree with the authors in this respect. The Se and Sp of a diagnostic test do not change from population to population, as the authors claim, provided that the assay or the cut-off values are not changed, and the performance characteristics of the test have been adequately assessed.2Ransohoff DF Feinstein AR. Problems of spectrum and bias in evaluating the efficacy of diagnostic tests.N Engl J Med. 1978; 299: 926-930Crossref PubMed Scopus (1315) Google ScholarThe potential for bias in the assessment of the Se and Sp of the calprotectin test in the study by Tibble et al., as well as the lack of consideration of “red flags” associated to Rome I criteria as part of their analysis raise some questions about the potential role of fecal calprotectin in the work-up of IBS patients. Dear Sir:We read with great interest the article from Tibble et al.,1Tibble JA Sigthorsson G Foster R Forgacs I Bjarnason I. Use of surrogate markers of inflammation and Rome criteria to distinguish organic from nonorganic intestinal disease.Gastroenterology. 2002; 123: 450-460Abstract Full Text Full Text PDF PubMed Scopus (340) Google Scholar in the August 2002 issue of Gastroenterology. Based on the sensitivity (Se) and specificity (Sp) of the calprotectin test to differentiate organic gastrointestinal disease from IBS, the authors propose the use of this test in combination with Rome I criteria to “screen” tertiary referred IBS patients for further work-up.There are a number of methodological issues in the study that we believe are worth further analysis.First, the use of an inadequate criterion standard (“gold standard”).2Ransohoff DF Feinstein AR. Problems of spectrum and bias in evaluating the efficacy of diagnostic tests.N Engl J Med. 1978; 299: 926-930Crossref PubMed Scopus (1315) Google Scholar In the study from Tibble et al., the criterion standard varies from individual to individual, since patients were seen by 4 different gastroenterologists and the tests performed in each individual was determined by each investigating clinician's choice. This may have significantly influenced the final diagnosis in each individual and impacts negatively on the internal validity of the study.Second, the inclusion of a narrow spectrum of patients.2Ransohoff DF Feinstein AR. Problems of spectrum and bias in evaluating the efficacy of diagnostic tests.N Engl J Med. 1978; 299: 926-930Crossref PubMed Scopus (1315) Google Scholar Thus, patients with diseases outside the gastrointestinal tract were excluded from the study. This may lead to a biased estimate of the performance characteristics of the test and hence, limits the external validity of the results.Third, the authors describe the performance characteristics of the different tests using Se, Sp, positive, and negative predictive values. The use of predictive values can be misleading since they depend on the prevalence of the disease in the population in whom the test is performed.3Weissler AM. A perspective on standardizing the predictive power of noninvasive cardiovascular tests by likelihood ratio computation: 1. Mathematical principles.Mayo Clin Proc. 1999; 74: 1061-1071Abstract Full Text Full Text PDF PubMed Scopus (25) Google Scholar On the other hand, likelihood ratios (LR), which depend only on the operating characteristics of the test, are thought to be more appropriate and allow assessment of the usefulness of the test in all possible scenarios.4Jaeschke R Guyatt GH Sackett DL. Users' guides to the medical literature. III. How to use an article about a diagnostic test. B. What are the results and will they help me in caring for my patients? The Evidence-Based Medicine Working Group.JAMA. 1994; 271: 703-707Crossref PubMed Scopus (2132) Google ScholarThus, using the Se and Sp reported in the article, we calculated the LRs for each test under different scenarios, and reached conclusions that differed from those in Tibble's article. For example, in the tertiary referral population in this study (pretest probability of organic disease = 44%), the posttest probability of organic disease based solely on absence of Rome I criteria is 79% (positive LR = 4.7). This is similar to the posttest probability of organic disease based solely on a positive calprotectin test, 77% (positive LR = 4.2). The combination of the absence of Rome I criteria plus a positive C-reactive protein test (positive LR = 2.1) yields a posttest probability of organic disease of 91%. This probability of organic disease, with a much simpler blood screening test that is universally applicable, is high enough to make it reasonable to perform further investigation regardless of the results of fecal calprotectin. This information limits the usefulness of the calprotectin test to “rule in” organic disease in a tertiary care practice.On the other hand, if Rome criteria are present (negative LR = 0.27), the posttest probability of organic disease is 21%, and the addition of a negative calprotectin test (negative LR = 0.11) decreases that probability to 4% compared with the 14% in case of a negative C-reactive protein test (negative LR = 0.6). Thus, it appears that in the presence of Rome I criteria for IBS, fecal calprotectin might be useful to “rule out” organic disease. It would have been of interest to include in the analysis the combination of presence/absence of “red flags” or alarm symptoms and Rome I criteria, since these are of critical value to discriminate organic from nonorganic patients with GI complaints on formal analysis5Vanner SJ Depew WT Paterson WG et al.Predictive value of the Rome criteria for diagnosing the irritable bowel syndrome.Am J Gastroenterol. 1999; 94: 2912-2917Crossref PubMed Scopus (234) Google Scholar and are used extensively in clinical practice. Thus, it is important to know whether adding absence of “red flags” to the presence of Rome criteria for IBS would decrease the value of fecal calprotectin test to “rule out” organic disease.Finally, Tibble et al. state that the use of calprotectin test in a primary care population should be deferred until the test is validated in such population. We beg to disagree with the authors in this respect. The Se and Sp of a diagnostic test do not change from population to population, as the authors claim, provided that the assay or the cut-off values are not changed, and the performance characteristics of the test have been adequately assessed.2Ransohoff DF Feinstein AR. Problems of spectrum and bias in evaluating the efficacy of diagnostic tests.N Engl J Med. 1978; 299: 926-930Crossref PubMed Scopus (1315) Google ScholarThe potential for bias in the assessment of the Se and Sp of the calprotectin test in the study by Tibble et al., as well as the lack of consideration of “red flags” associated to Rome I criteria as part of their analysis raise some questions about the potential role of fecal calprotectin in the work-up of IBS patients. We read with great interest the article from Tibble et al.,1Tibble JA Sigthorsson G Foster R Forgacs I Bjarnason I. Use of surrogate markers of inflammation and Rome criteria to distinguish organic from nonorganic intestinal disease.Gastroenterology. 2002; 123: 450-460Abstract Full Text Full Text PDF PubMed Scopus (340) Google Scholar in the August 2002 issue of Gastroenterology. Based on the sensitivity (Se) and specificity (Sp) of the calprotectin test to differentiate organic gastrointestinal disease from IBS, the authors propose the use of this test in combination with Rome I criteria to “screen” tertiary referred IBS patients for further work-up. There are a number of methodological issues in the study that we believe are worth further analysis. First, the use of an inadequate criterion standard (“gold standard”).2Ransohoff DF Feinstein AR. Problems of spectrum and bias in evaluating the efficacy of diagnostic tests.N Engl J Med. 1978; 299: 926-930Crossref PubMed Scopus (1315) Google Scholar In the study from Tibble et al., the criterion standard varies from individual to individual, since patients were seen by 4 different gastroenterologists and the tests performed in each individual was determined by each investigating clinician's choice. This may have significantly influenced the final diagnosis in each individual and impacts negatively on the internal validity of the study. Second, the inclusion of a narrow spectrum of patients.2Ransohoff DF Feinstein AR. Problems of spectrum and bias in evaluating the efficacy of diagnostic tests.N Engl J Med. 1978; 299: 926-930Crossref PubMed Scopus (1315) Google Scholar Thus, patients with diseases outside the gastrointestinal tract were excluded from the study. This may lead to a biased estimate of the performance characteristics of the test and hence, limits the external validity of the results. Third, the authors describe the performance characteristics of the different tests using Se, Sp, positive, and negative predictive values. The use of predictive values can be misleading since they depend on the prevalence of the disease in the population in whom the test is performed.3Weissler AM. A perspective on standardizing the predictive power of noninvasive cardiovascular tests by likelihood ratio computation: 1. Mathematical principles.Mayo Clin Proc. 1999; 74: 1061-1071Abstract Full Text Full Text PDF PubMed Scopus (25) Google Scholar On the other hand, likelihood ratios (LR), which depend only on the operating characteristics of the test, are thought to be more appropriate and allow assessment of the usefulness of the test in all possible scenarios.4Jaeschke R Guyatt GH Sackett DL. Users' guides to the medical literature. III. How to use an article about a diagnostic test. B. What are the results and will they help me in caring for my patients? The Evidence-Based Medicine Working Group.JAMA. 1994; 271: 703-707Crossref PubMed Scopus (2132) Google Scholar Thus, using the Se and Sp reported in the article, we calculated the LRs for each test under different scenarios, and reached conclusions that differed from those in Tibble's article. For example, in the tertiary referral population in this study (pretest probability of organic disease = 44%), the posttest probability of organic disease based solely on absence of Rome I criteria is 79% (positive LR = 4.7). This is similar to the posttest probability of organic disease based solely on a positive calprotectin test, 77% (positive LR = 4.2). The combination of the absence of Rome I criteria plus a positive C-reactive protein test (positive LR = 2.1) yields a posttest probability of organic disease of 91%. This probability of organic disease, with a much simpler blood screening test that is universally applicable, is high enough to make it reasonable to perform further investigation regardless of the results of fecal calprotectin. This information limits the usefulness of the calprotectin test to “rule in” organic disease in a tertiary care practice. On the other hand, if Rome criteria are present (negative LR = 0.27), the posttest probability of organic disease is 21%, and the addition of a negative calprotectin test (negative LR = 0.11) decreases that probability to 4% compared with the 14% in case of a negative C-reactive protein test (negative LR = 0.6). Thus, it appears that in the presence of Rome I criteria for IBS, fecal calprotectin might be useful to “rule out” organic disease. It would have been of interest to include in the analysis the combination of presence/absence of “red flags” or alarm symptoms and Rome I criteria, since these are of critical value to discriminate organic from nonorganic patients with GI complaints on formal analysis5Vanner SJ Depew WT Paterson WG et al.Predictive value of the Rome criteria for diagnosing the irritable bowel syndrome.Am J Gastroenterol. 1999; 94: 2912-2917Crossref PubMed Scopus (234) Google Scholar and are used extensively in clinical practice. Thus, it is important to know whether adding absence of “red flags” to the presence of Rome criteria for IBS would decrease the value of fecal calprotectin test to “rule out” organic disease. Finally, Tibble et al. state that the use of calprotectin test in a primary care population should be deferred until the test is validated in such population. We beg to disagree with the authors in this respect. The Se and Sp of a diagnostic test do not change from population to population, as the authors claim, provided that the assay or the cut-off values are not changed, and the performance characteristics of the test have been adequately assessed.2Ransohoff DF Feinstein AR. Problems of spectrum and bias in evaluating the efficacy of diagnostic tests.N Engl J Med. 1978; 299: 926-930Crossref PubMed Scopus (1315) Google Scholar The potential for bias in the assessment of the Se and Sp of the calprotectin test in the study by Tibble et al., as well as the lack of consideration of “red flags” associated to Rome I criteria as part of their analysis raise some questions about the potential role of fecal calprotectin in the work-up of IBS patients.