The usefulness of fecal hemoglobin and calprotectin tests in diagnosing significant bowel diseases: a prospective study

Abstract Objectives Although colonoscopy remains the gold standard for determining bowel diseases, it’s invasive and expensive. New non-invasive diagnostic methods are urgently needed as an initial screening modality. We aimed to investigate the value of fecal calprotectin (FC) and fecal immunochemical test (FIT) in differentiation of significant and non- significant bowel diseases. Methods In this prospective study, consecutive individuals were included if they underwent colonoscopy for symptoms of lower gastrointestinal (GI) tract, positive fecal occult blood test, surveillance for IBD or colorectal cancer (CRC) screening. Diagnostic value of FC and FIT in discriminating significant bowel diseases (advanced neoplasia, active inflammatory bowel diseases or bowel inflammation due to other causes) and non-significant bowel diseases (normal, asymptomatic diverticulum, non-adenomatous polyp, or non-advanced neoplasia) were evaluated. Results Among 201 individuals included, 107 patients had significant bowel diseases. FC and FIT had an area under the curve (AUC) of 0.722 (95% confidence interval [CI] 0.653–0.792) and 0.797 (95%CI 0.734–0.860), respectively, for determining significant bowel diseases. Combination of FC and FIT predicted significant bowel diseases with an AUC, sensitivity, specificity, and accuracy of 0.832 (95% CI 0.775–0.890), 77.6%, 74.5%, and 76.1%, respectively. Moreover, combination of FC and FIT was more sensitive among patients with lower GI symptoms than asymptomatic individuals (80.8% vs. 74.1%) to identify significant bowel diseases. Conclusions A single measurement of FC or FIT is not sufficiently accurate to identify patients with significant bowel disease. However, combination of FC and FIT can help increase the sensitivity, especially in patients with lower GI symptoms.


Introduction
Gastrointestinal symptoms are common reasons for outpatient visits. Most of these patients will be diagnosed as having a functional gastrointestinal disease [1], and they are rarely caused by significant bowel diseases, such as inflammatory bowel disease (IBD), advanced adenomas (AAs) or colorectal cancer (CRC) [2]. However, it is important to identify patients at risk for the latter diseases as they require further investigations. Although colonoscopy is the gold standard for diagnosing organic colorectal diseases, unnecessary referrals should be avoided because it is invasive and costly. Reliable diagnostic tests that are easy to use as an initial screening method in clinic would be of considerable help.
Calprotectin, a member of the S100 family, is a calciumand zinc-binding protein, and it comes predominantly from neutrophils [3]. Fecal calprotectin (FC) correlates well with bowel inflammation, and it has been proved to be a valuable tool to distinguish IBD from irritable bowel syndrome, and also to monitor known IBD [3]. Additionally, although not well studied, FC also has evidence for detecting other disease, such as infectious colitis, eosinophilic colitis, and gastrointestinal (GI) neoplasms [4][5][6]. However, FC is not sensitive or specific enough to detect non-cancerous tumor formation and is not recommended for routine use in intestinal screening patients to detect cancer [7].
Fecal immunochemical test (FIT) is a quantitative test for human hemoglobin in feces and has been recommended to guide referral of patients considered at risk of CRC [8]. However, due to intermittent bleeding from lesions, the sensitivity of a single measurement of FIT for detecting CRC is low, and repeated FIT tests are always warranted [9,10]. In addition, the sensitivity of FIT is much unsatisfactory in advanced adenoma and early CRC detection [11]. However, there is insufficient data on the application of FIT as a diagnostic test in clinical settings other than CRC screening.
Herein, we conducted this study to evaluate the diagnostic yield of the two fecal biomarkers, FIT and FC, in the identification of significant bowel diseases and more importantly, to assess the value of the combination of the two tests with the aim of avoiding unnecessary colonoscopies.

Design and participation
A total of 201 consecutive individuals, who underwent colonoscopy at our center between November 2020 and March 2021 were prospectively included. Inclusion criteria were as follows: (1) age !18 years old; (2) patients with any lower GI symptoms (including hematochezia, abdominal pain, abdominal distension, change of bowel habits, etc.); (3) subjects with positive fecal occult blood test; (4) active IBD or colonoscopic surveillance for patients with inactive IBD; (5) individuals undergoing screening for CRC; (6) patients signed the informed consent. Those who met one of the following conditions were excluded: (1) patients with obvious upper GI symptoms or diagnosed with upper GI diseases; (2) patients with suspected or diagnosed with small bowel diseases; (3) patients taking regular antiplatelet and/or anticoagulant drugs, nonsteroidal anti-inflammatory agents, proton pump inhibitors, or corticosteroids.

Fit and FC analyses
In this study, the stool samples were collected for FC and FIT measurement before bowel preparation and within 3 days of colonoscopy. Stool samples were stored at room temperature and delivered to our center within 3 days of excretion. FC specimens were stored at À80 C in a freezer till tested. FC was measured using a qualitative enzyme-linked immunosorbent assay kit (BHLMANN, Switzerland), according to the manufacturer instructions. The analytical sensitivity of calprotectin was 12.6 lg/g. For FIT measurement, the OC-Sensor FIT kit (EIKEN CHEMICAL, Japan) was used. A specifically designed sampling probe [EIKEN CHEMICAL, Japan] was digged six times in the feces and reinsert the probe into the respective device containing 2 mL buffer. FIT samples were stored at 4 C within one week till analysis. Specimens were tested on a qualitative FIT system [OC-SENSOR io, EIKEN, Japan] with an analytical working range of 50-1000 ng/mL. The FC and FIT tests were performed by two independent technicians who were blinded to the clinical details of the involved individuals.

Colonoscopy and histological evaluation
Polyethylene glycol solution was used for bowel cleansing under the guidance of nurses. Colonoscopy was performed by experienced endoscopists who were blinded to the results of FC and FIT. When lesions were found, they were biopsied if necessary. Endoscopic treatment was further scheduled for patients with colorectal neoplasms. Pathologists were not informed of the FC and FIT results of the subjects. The final diagnosis was determined based on one's medical history, colonoscopic findings, and pathological results.
Advanced adenoma was defined as an adenomatous polyp meeting one or more of the following criteria: (1) size !10 mm in diameter; (2) containing !25% villous component; (3) high-grade intraepithelial neoplasia [12]. For participants with multiple adenomatous polyps, the most advanced lesion was counted. Although every adenoma has the capacity of malignant evolution, most adenomas stabilize their progression or even regress, and only few kinds of adenomas (advanced adenomas) actually develop invasive cancer [13]. Thus, we defined advanced adenoma, CRC, active inflammatory bowel disease (IBD) and colorectal inflammation due to other causes (e.g., infection, ischemia, etc.) as significant bowel diseases. On the other side, asymptomatic diverticulum, non-adenomatous polyp (including inflammatory polyp and hyperplastic polyp), non-advanced adenoma, inactive IBD, and a normal colon were defined as non-significant bowel diseases.

Statistical analysis
Continuous variables were described as median and interquartile range (IQR) if they did not conform to a normal distribution. Comparisons between groups were performed using the non-parametric Mann-Whitney U test. The Spearman rank correlation test was used to analyze the correlation between FC and FIT. McNemar test was used to compare the sensitivity between different methods. Receiver operating characteristic (ROC) curves were constructed for assessment of the diagnostic performance of FC and FIT in the discrimination between significant and non-significant bowel diseases. The ideal cutoff level for each tool was calculated by maximizing the Youden index. If the ideal cutoff level fell below the lower limit of the working range of measurement, the lower limit level was determined as the actual cutoff. A two-sided p-values <0.05 was set as statistically significant. Analyses were performed with Statistical Package for the Social Sciences (SPSS, version 23.0, SPSS Inc. Chicago, Illinois, USA) and Prism (version 8.0.2, GraphPad Software, Inc. San Diego, CA, USA).

Ethical considerations
This study was performed in line with the principles of the Declaration of Helsinki and the protocol was approved by the Ethics Committee of Beijing Friendship Hospital, Capital Medical University (No. 2020-P2-207-02). Informed consent was obtained from each patient.

Characteristics of the study population
Two hundred and one individuals were consecutively enrolled (males 51.7%, and the median age 56 (IQR 41-65) years). Indications for colonoscopy and the diagnoses were summarized in Table 1. Among the included subjects, 107 were categorized as significant bowel diseases (advanced adenoma n ¼ 33, CRC n ¼ 36, active IBD n ¼ 25, other colorectal inflammation n ¼ 13), and the other 94 were categorized as non-significant bowel diseases (normal colon n ¼ 50, asymptomatic diverticulum n ¼ 3, non-adenomatous polyp n ¼ 7, non-advanced adenoma n ¼ 25, inactive IBD n ¼ 9).

Levels of FC and FIT in different groups
There was a moderate correlation between FIT and FC con- Figure S1). In Figure 1 (Figure 1(a)) As for the FIT test, except for diverticulum ( (Figure 1(c,d)). According to the ROC analysis, the ideal cutoff of FC for identifying significant bowel diseases was 114.3 lg/g, which had a sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy of 57.0%, 83.0%, 79.2%, 62.9%, and 69.2%, respectively. (Table 2) We also set 100 lg/g as the cutoff value of FC, for this is one the most widely used cutoff in previous studies [3,14]. FC level ! 100 lg/g had a sensitivity, specificity, PPV, NPV and accuracy of 59.8%, 78.7%, 76.2%, 63.2% and 68.7%, respectively, in the diagnosis of significant bowel diseases. (Table 2).

Overall diagnostic performance of FIT and FC
According to the ROC analysis, the ideal cutoff of FIT for determining significant bowel diseases was 22.5 ng/mL, which was smaller than the lower limit of the working range of FIT measurement. Therefore, we set 50 ng/mL as the FIT cutoff value. FIT concentration ! 50 ng/mL had a sensitivity, specificity, PPV, NPV and accuracy of 59.8%, 94.7%, 92.8%, 67.4% and 76.1%, respectively, in the diagnosis of significant bowel diseases. (Table 2). Furthermore, we assessed whether a strategy of combination of FC testing and FIT testing could improve the diagnostic performance of identifying patients with significant bowel diseases. When FC and FIT were combined, the AUC increased to 0.832 (95%CI 0.775-0.890, p < 0.001) (Figure 2), and the sensitivity increased markedly to 77.6% as compared with 59.8% for FC alone (p < 0.001) and 59.8% for FIT alone (p < 0.001). Moreover, the NPV was increased when compared with FC or FIT alone. (Table 2).

Diagnostic performance of FIT and FC in symptomatic and asymptomatic patients
Patients with lower GI symptoms had higher levels of FC (237.7 (38.8-928.4) lg/g vs. 54.9 (19.0-171.1) lg/g, p ¼ 0.002) when compared with asymptomatic patients. Among patients with symptoms, a cutoff value of FC as 237.7 lg/g resulted in 73.1% sensitivity, 83.3% specificity, 86.4% PPV, and 68.2% NPV in identifying patients with significant bowel diseases with a corresponding AUC of 0.817 (95% CI 0.690-0.945, p < 0.001) (Table 3, Figure 3(a)), which was better than that in overall patients. Though the median FIT in symptomatic patients was higher than that in asymptomatic patients (26.5(3.0-245.0) ng/mL vs. 10.0 (1.0-138.0) ng/mL), the difference was not significant statistically (p ¼ 0.177). For FIT, a cutoff value of 50 ng/mL resulted in 61.5% sensitivity, 83.3% specificity, 84.2% PPV, and 60.0% NPV in identifying patients with significant bowel diseases with a corresponding AUC of 0.854 (95% CI 0.740-0.967, p < 0.001) ( Table 3, Figure  3(a)). The sensitivity was also better than that in overall patients. Combination of FIT and FC resulted in 80.8% sensitivity, 72.2% specificity, 80.8% PPV, and 72.2% NPV in identifying patients with significant bowel diseases. (Table 3).
As Table 4 and Figure 3

Discussion
To the best of our knowledge, this study is the first prospective study to investigate the use of a FIT and a FC test as diagnostic aids for physicians in differentiate significant and non-significant bowel diseases in patients scheduled for colonoscopy, and to assess whether combining these markers could improve the diagnostic performance. We included a variety of diseases, among which significant bowel diseases included not only IBD and CRC commonly identified by FC and FIT, but also advanced adenoma, early CRC and other colorectal inflammatory diseases, which are meaningful diseases that do require further colonoscopy to be early diagnosed. Our main finding was that the diagnostic performance of FC was comparable to that of FIT in differentiating significant from non-significant bowel disease with a relatively high specificity and PPV, and combination of FC and FIT could significantly increase sensitivity. Additionally, in patients with lower GI symptoms, FC and FIT both performed better and had a higher sensitivity in differentiating significant from non-significant bowel disease. In patients without lower GI symptoms, FC and FIT both had a higher specificity in differentiating significant from non-significant bowel diseases, and the combination of FIT and FC could also improve the sensitivity.
With increasing demand for gastroenterological consultations and endoscopic investigations, a diagnostic pathway that enables clinicians to confidently diagnose non-significant bowel disease will be of benefit for both patients and gastroenterologists. It is estimated that the outcome of nearly half of colonoscopies performed in patients are normal and therefore potentially avoidable [15]. The development of alternative non-invasive tests accessible to physicians is desirable. FC and FIT tests are two tools can be easily applied and analyzed in clinic, and there is growing evidence that they may fulfil this role.
FIT has been highly recommended as a test for CRC screening [16], while FC is a useful, evidence-approved adjunct in IBD screening and monitoring practices [17]. As the cutoff values varied from study to study, in general, FC reliably distinguished functional disease from IBD, with sensitivity ranging from 60.4% to 100% and specificity ranging from 44% to 100% [18]. Increased levels of FC have also been reported in patients affected by neoplastic disease of the GI tract [19,20]. Meucci et al. found that although single measurement of FC was not sufficiently accurate to identify those with significant colorectal disease, a normal result can help rule out organic disease among patients with diarrhea and those with abdominal pain and/or constipation [21]. In James and colleagues' study, FC also has a high NPV for colorectal cancer and significant polyps in patients referred from primary care with suspected cancer [22]. These evidences suggest that FC has diagnostic value not only for inflammatory diseases but also for GI neoplasms. Studies have also investigated the role of combination of FC and FIT in diagnosing CRC [23]. However, colorectal cancer and IBD are not the only diseases worthy of our attention. Colorectal diseases such as advanced adenoma and inflammation of other causes also require colonoscopy investigation. To the best of our knowledge, there are currently no prospective studies exploring the efficacy of combination of FC and FIT in the diagnosis of significant colorectal diseases, including advanced neoplasm and inflammation. The current work provides evidence-based data to this area.
Using FIT or FC alone in this cohort would potentially have avoided a referral of colonoscopy in 67.4% or 63.2% patients without serious pathology. Such measures could potentially ease pressures on already overburdened colonoscopy. The ability to identify those who do have significant bowel pathology is more important in terms of patient outcomes than limiting inappropriate referrals. In this context, FC and FIT alone was not perfect to be the diagnostic tool for their relatively low sensitivity. Yet, combination of FIT and FC could increase the sensitivity, and if only when both markers were lower than the cutoff values, it was considered as non-significant bowel diseases, the diagnostic sensitivity reached 77.6%. IBD always accompanied by typical clinical symptoms [24]. Studies also confirmed that some bowel symptoms increased the likelihood of being diagnosed with CRC or advanced adenoma [12]. As a result, we further investigated the usefulness of FC and FIT in patients with or without lower GI symptoms. Obviously, FC and FIT had a better diagnostic performance in symptomatic patients, although there was no significant difference in the proportion of significant bowel diseases between symptomatic and asymptomatic groups. Importantly, combination of FC and FIT had a high sensitivity both in symptomatic and asymptomatic patients, which is beneficial in avoiding missed diagnoses as much as possible.
The validity of the results of our study is supported by its prospective recording of symptoms. Nevertheless, some  limitations exist. First, the relatively small number of patients enrolled in a single center did not allow an ideal cut-off to be established for distinguishing between significant and non-significant bowel diseases. Instead, we set the most used cutoffs for FC (100 lg/g) and FIT (50 ng/mL) in our analysis to make our results be of more universality. Second, although we tried to exclude patients with upper gastrointestinal disease at the beginning of the study, it is impossible to rule out the possibility that in some of the patients with elevated FC or FIT and normal colonoscopy, a disease of the upper GI tract or of the small bowel was present, 5 which may bias the results.
In conclusion, the determination of FC and FIT is a necessary complementary tool in routine clinical practice for discriminating between significant and non-significant bowel diseases. The combination of FIT and FC would aid physicians in particular in decision-making as to which patients warranted referral.