Search in the Abstract Database

Search Abstracts 2015

* = Presenting author

P120 Consistently low albumin level is associated with subsequent bowel stenosis, fistula or perforation in patients with Crohn's disease.

G. Radford-Smith*1, E. Ferguson2, K. Hanigan3, L. Simms3, J. Irwin1

1Royal Brisbane and Women's Hospital, Gastroenterology, Brisbane, Australia, 2Royal Brisbane & Women's Hospital, Department of Gastroenterology, Brisbane, Australia, 3QIMR Berghofer Medical Research Institute, Complex Diseases, Brisbane, Australia


Laboratory tests are used longitudinally in the management of patients with Crohn's disease. Whether the Results of these tests correlate with the subsequent development of bowel stenosis, perforation or fistula formation is unknown.

Aim: To study the correlation between longitudinal laboratory testing and subsequent outcome in patients with Crohn's disease.


Patients diagnosed at a tertiary referral centre with Crohn's disease between 1994 and 2014, with more than five years of clinical follow-up, had objective clinical, laboratory and genetic data recorded. An objective poor outcome was defined as the development of a fistula, a bowel stenosis or a bowel perforation. Outcomes occurring within 6 months of diagnosis were excluded. Laboratory data was represented by the area under the curve of values measured when patients were well, in the complication free period leading up to development of the outcome. Cox regression was used to analyse the association between outcome and laboratory values; C reactive protein (CRP), platelet count, albumin level, faecal calprotectin, serum ferritin, serum haemoglobin and erythrocyte sedimentation rate (ESR). Laboratory values were converted to categorical variables with optimized cut-offs. Recognized predictors of poor outcome were added to the model to assess independence of identified associations.


The data of 382 patients were reviewed, 325 patients had more than five years of clinical follow-up, and 268 had a complete clinical, biochemical and genetic record. 144 outcome events (120 stenoses, 15 perforations and 9 fistulae) were observed over a median of 10.77 (IQR (interquartile range) 7.65 - 13.91) years of follow-up. Blood testing was performed a median of 4.79 (IQR 3.23 - 7.66) times per year for each patient, over a median of 3.95 (IQR 1.76 - 7.37) years prior to each event. After multivariate analysis with inclusion of recognised predictor variables, an albumin level < 36 (HR 3.19, p<0.001) maintained an independent association with outcome. ATG16L1 AG or GG genotype (HR 2.32, p=0.031) , continued smoking (HR 1.57, p=0.018) and L1 or L3 Montreal location at diagnosis (HR 1.73, p=0.032) were also independently associated with a poor outcome in the final model.


A longitudinally measured albumin level consistently<36 g/L correlates with subsequent development of an objective poor outcome - predominantly small bowel stenosis - in patients with Crohn's disease. This observation may represent malnutrition from sub clinical bowel obstruction, or a chronic inflammatory state. Serial monitoring of albumin may aid in identification of patients at risk of progression to bowel stenosis, peroforation or fistula formation.