PMO-243 Faecal calprotectin analysis: does the method matter?. (28th May 2012)

• Record Type: Journal Article
• Authors: Tomkins, C; Zeino, Z; Nwokolo, C; Smith, S C; Arasaradnam, R
• Is Part Of: Gut. Volume 61(2012)Supplement 2
• Page Start: A173
• Page End: A174
• Publication Date: 2012-05-28
• Subjects: Gastroenterology -- Periodicals; 616.33
• DOI: 10.1136/gutjnl-2012-302514b.243
• Languages: English
• ISSNs: 0017-5749

Abstract

Abstract : Introduction: Faecal calprotectin (FC) is a sensitive marker of intestinal inflammation and is useful to help distinguish between organic and non-organic (functional) disease. The increasing popularity of this test, with various analytical methods available, potentially leads to confusion in interpreting results. The aim of this study was to technically evaluate FC measured by different ELISA methods in secondary/tertiary care. Methods: 62 stool samples were collected from sequential out-patients presenting with chronic diarrhoea. All participants had a colonoscopy with biopsy, to which FC results were compared. FC was measured by ELISA assays: Immundiagnostik PhiCal (version 1) and Buhlmann EK-CAL. A subset were also measured by PhiCal (version 2). Stool was weighed and extracted, and ELISAs performed manually. Results: 38 patients with IBD/other organic bowel disease (mean 36 yrs, range 15–49) and 24 patients with IBS (mean 36 yrs, range 20–48) were sampled. Sensitivity and specificity for active IBD vs IBS using manufacturers' cut-offs of 50 μg/g were: Buhlmann EK-CAL 86% (95% CI 42 to 99) and 60% (95% CI 33% to 83%), PPV 50% (95% CI 22% to 78%), NPV 90% (95% CI 54% to 99%); PhiCal1 78% (95% CI 40% to 96%) and 92% (95% CI 60% to 100%), PPV 88% (95% CI 47% to 99%) and NPV 86% (95% CI 56% to 97%). Correlation across full range of results were PhiCal1 vs EK-CAL, R 2 =0.45; PhiCal2 vs PhiCal1, R 2 =0.54. However for results <100 μg/g by PhiCal1, correlations improved that is, R 2 =0.64 and R 2 =0.83 respectively. Intra-batch imprecision of the whole process, including extraction of native stool, was assessed at a range of clinically relevant concentrations: Buhlmann EK-CAL 17% (10 μg/g), 12% (47 μg/g), 19% (62 μg/g); PhiCal1 21% (8 μg/g), 24% (10 μg/g), 18% (27 μg/g); PhiCal2 19.5% (18.9 μg/g). Inter-batch imprecision of ELISA analysis was lower: Buhlmann EK-CAL 8.6% (30 μg/g), 5.8% (129 μg/g); PhiCal1 6.2% (39 μg/g), 10.8% (135 μg/g); PhiCal2 8.9% (33 μg/g). Functional sensitivity: Buhlmann EK-CAL 10 μg/g; both PhiCal 20 μg/g. Assays were found to be linear (without further sample dilution) up to 600 μg/g for EK-CAL, PhiCal1 400 μg/g, PhiCal2 800 μg/g. Mean recovery in spiked stool samples: Buhlmann EK-CAL 98%, PhiCal1 83%, PhiCal2 79%. Conclusion: All three ELISA assays evaluated have relatively high coefficients of variation compared to other laboratory tests, due to heterogeneity of stool material and manual extraction/analysis. Results from different FC methods are not directly comparable, despite widespread adoption of single cut-offs. Using 50 μg/g cut-off, PhiCal1 performed better than Buhlmann EK-CAL in distinguishing IBD from IBS in our study. There is improved assay linearity using PhiCal2. Clinicians should be aware of type of ELISA methods employed when interpreting FC results, and cut-offs used should be fully evaluated. Competing interests: None declared.