Search in the Abstract Database

Search Abstracts 2012

* = Presenting author

P114. What factors might contribute to borderline faecal calprotectin levels?

O.M. Demir1, R. Appleby2, Y. Wang3, R.P.H. Logan1

1King's College Hospital, Gastroenterology, London, United Kingdom; 2West Middlesex Hospital, Gastroenterology, London, United Kingdom; 3King's College, Primary Care and Public Health Sciences, London, United Kingdom

Background: In adults, Faecal Calprotectin (FC) ELISA has a sensitivity and specificity of >93% for detecting IBD (1). However, small case series have shown that other factors (eg NSAIDs, cirrhosis, obesity etc) may lead to false +ve results. It is also unclear if immuno-suppressants, taken for non-GI indications can give false −ve FC results.

Aim: to determine the size and effect of possible factors which might contribute to borderline FC values to cause false +ve or −ve results.

Methods: Study population: any patient with a FC result held in the pathology database from May 2005–09 (n = 5943). Patient data were cross referenced to electronic records, radiology (PACS), endoscopy & pathology data sets (SnoMed CT). Exclusions: previous diagnosis of IBD or CRC, or aged <16 yr at study entry. To increase the chance of discovering a significant effect, only patients with FC result of >50 µg/g and <150 µg/g were included. Multivariate analysis (generalised linear model) was performed sequentially using SPSS with age, gender, BMI, ethnicity, medication, family history of IBD, smoking/alcohol consumption, liver disease. Histological severity of colonic inflammation was only included in the final model.

Results: From 5943 patients, 569 (10%) were between 50–150 and met the inclusion criteria. 62% were female, mean age 51 yr, range 16–91 yr). Diarrhoea (50%) was the most common indication for testing.

In the preliminary model FC values increased with age (B = 0.15, p < 0.04) and with a positive trend for liver disease (B = 7.7, p = 0.06); however neither variable remained significant in the final model, which was overwhelmingly dominated by the effect of mucosal inflammation (B = 41, p < 0.001). Immunosuppressant use (mainly for Rharthritis/SLE) was associated with lower FC values (p = 0.027). There was no effect from any of the other variables at any stage in either model.

Final Model
 calprotectin change (slope)p value
liver disease4.00.2
low dose aspirin use−5.90.1
regular NSAID use−5.10.42
longterm steroid use−1.20.86
immunosuppressant use−13.10.027
age (per annum change)0.130.064
mucosal inflammation41.0<0.001

Conclusions: These data show that FC levels are almost entirely determined by the presence of mucosal inflammation. Previously suggested confounding factors have no effect. Concurrent immunosuppressant use may slightly lower FC levels. These data show the FC is a robust measure of mucosal inflammation.