Fecal calprotectin variability in Crohn's disease

To the Editor: We read with interest the articles by Gisbert et al1,2 and Sipponen et al.3,4 Calprotectin in feces has been a widely used biological marker in screening for inflammatory bowel disease (IBD) and clinical follow-up in Crohn's disease (CD) and ulcerative colitis (UC). The predictive value of fecal calprotectin (Fc) has been good to predict relapse before patients experience symptoms, as well as when IBD patients experience symptoms indicative of activity. The specificity of Fc, however, has been questioned in CD,3 and where to set the Fc baseline level could be an explanation for the conflicting results. The fecal samples can be collected at home, mailed to the laboratory, and analyzed with a simple enzyme-linked immunosorbent assay (ELISA) method for a quantitative evaluation, at low cost. In a cohort of patients with IBD there will be some degree of variability based on disease activity. The issue of intraindividual variability must be considered within the context of the population measured, and there is variability of Fc concentrations within the IBD population as well as variability individually and between the diagnosis of UC and CD. We do not, however, find anyone describing the variability between test and retesting aimed at quality assurance in individual patients with disease activity. In our clinical practice the Fc test has been widely used in the last few years, both helping to select those who should have an endoscopic examination in diagnosing IBD (fromthose with normal Fc) and in follow-up and in monitoring treatment response, especially with biologicals. In a study to evaluate the reproducibility of Fc in a clinical setting, we asked patients with CD with mild to moderate activity assessed by the Harvey–Bradshaw Index (HBI) to collect a stool sample on 2 consecutive days. All patients gave informed consent before being included in the study. The study was approved by the Regional Ethics Committee, HR SouthEast, Norway. All the Fc were analyzed at the same laboratory. The patients were all familiar with the sampling and mailing procedure. As shown in Figure 1, a significant difference was found in those 63 pairs of stool samples examined (P < 0.0002; Wilcoxon paired test). As seen, the largest differences occurred with increasing concentrations of Fc. Fecal calprotectin values (“Value”) in stools from the same individual taken on 2 consecutive days (t2 was arbitrarily chosen as the lowest paired value and presented first) in 63 patients with CD. Measure of agreement, kappa 0.355 (SD 0.115; P < 0.0002). To evaluate if this difference could be explained by differences in collection of the 2 samples or by variation within stool analysis, we performed 5 subsamples in 10 new fecal samples of calprotectin, and the results are given in Figure 2. Five subsamples were taken from 10 stools in CD patients; extracts prepared with 4 mg extracted in 2 mL standard buffer and run in triplicate on an ELISA kit. The results are given as mg/kg after correction for the 1:500 dilutions. For each 5 subsamples the coefficient of variation (%) is given below. Our results, surprisingly, show large differences in Fc values in stool samples from 2 consecutive days in CD patients (Fig. 1). The importance of clinical investigation and retesting of Fc if conflicting findings are present is essential. Single values in the interval of 50–200 mg/L (where normal is less than 50 mg/L) is critical, since this can indicate starting or stopping treatment on a misapprehension diagnosis. Fortunately, our findings are reassuring, since few of the variations occur in the interval between positive and negative values. Six out of 63 (9%) had a value below and the other more than 200 and only 3 out of 63 (5%) had 1 negative and 1 value more then 50 mg/L. Since patients included should have mild or moderate HBI, relatively few cases had low values. This will also in most cases be the indication for testing, namely, to clarify if symptoms are consistent with disease activity and therefore a relapse that necessitates medical treatment. There could be several reasons for the variation in Fc. Sampling “error” from the patients (much mucus or much water in the stool sample), sampling “error” when preparing the stool test in the laboratory, and variation in Fc assays. In our evaluation regarding Fc assay variation, the results seems satisfying (Fig. 2). The time from patient sampling, included sending it to the laboratory and to the start of analysis, were in all test fewer than 4 days. Some variation may also be caused by the broad expression pattern of calprotectin, which is found in granulocytes as well as monocytes and is also inducible in epithelial cells. Finally, the observed variation could be caused by the varying presentation of epitopes on the calprotectin complex under different conditions. Although more studies are needed to fully define the role of Fc, and especially variability patterns, convincing studies and growing clinical experience point to an expanded role in the diagnosis and management of IBD.