Search in the Abstract Database

Search Abstracts 2011

* = Presenting author

P128. Serum IL8 concentration improves clinical activity index in ulcerative colitis

M.L.R. Perálvarez, V. García, E. Iglesias, C. Villar, R. González, J. Jurado, J. Muntané, F. Gómez-Camacho

Hospital Reina Sofía, Córdoba, Spain

Aims: To set the value of combining Truelove Witts modified index (TWm) with serum cytokine levels in predicting the histological severity of patients with ulcerative colitis (UC).

Methods: Transversal study including 67 UC consecutive patients who came to our center to perform a colonoscopy. Clinical activity was assessed using TWm index and blood samples were taken for determination of cytokines (IL1β, IL2, IL6, IL8, IL 10, IL13, IL17, IFNγ and TNFα). Bio Plex multi assay technology (Bio Rad laboratories, USA) was used to measure plasma cytokines concentration. Biopsies were taken from inflamed mucosa and histological activity was determined (GEBOES scale). All biopsies were examined by the same pathologist who was blinded for clinical information. Statistical analysis was performed by using SPSS 15.0 for Windows.

Results: Mean age was 43.3±12 years and 42 patients (62.7%) were men. The extent of the disease was proctitis in 12 (17.9%), distal colitis in 24 (35.8%) and pancolitis in 31 patients (46.3%). Clinical activity was inactive-mild in 41 patients (61.2%) and moderate-severe in 26 (38.8%). Endoscopic activity was inactive-mild in 22 patients (32.8%) and moderate-severe in 45 (67.2%). With regard to histological activity, 9 patients (13.3%) were classified like inactive disease (GEBOES = 0), 5 patients (GEBOES 1–2) like mild, 27 patients (40.8%) like moderate (GEBOES = 3) and 26 patients (38.6%) like severe disease (GEBOES = 4–5). Although significant correlation was found between clinical and histological activity (table 1), it was not perfect (table 2). Low specificity founded suggests that TWm underestimates histological activity. Serum cytokines were included in a multiple regression model identifying IL8 as a significant independent variable (OR = 1.27 CI95% 1.01–1.61; p = 0.038) (table 3). Including IL8 in the model implied an improve of specificity (from 33.3% to 58.3%), sensitivity (from 92.5% to 94.3%), global precision (from 81.5% to 87.7%) and fitting accuracy of the model (from 0.25 to 0.33) (table 4).

Table 1: Basal logistic regression without cytokines (fit goodness R2 = 0.25)
VariableORCI 95%p
Clinical activity (TWm)1.801.22–1.660.03
Table 2: Correlation between TWm and histological activity
ObservedPrognosedPercent correct
 Inactive-mild (GEBOES 0–2) (n)Moderate-severe (GEBOES 3–5) (n) 
(GEBOES 0–2)
4833.3% (Specificity)
(GEBOES 3–5)
44992.5% (Sensitivity)
   81.5% (global precision)
Table 3: Multiple Logistic regression including IL8 serum concentration (fit goodness R2 = 0.32)
VariableORCI 95%p
Clinical activity (TWm)1.891.15–3.080.011
Table 4: Correlation of TWm-IL8 index with histological activity
ObservedPrognosedPercent correct
 Inactive-mild (GEBOES 0–2) (n)Moderate-severe (GEBOES 3–5) (n) 
Inactive-mild (GEBOES 0–2)7558.3% (Specificity)
Moderate-severe (GEBOES 3–5)35094.3% (Sensitivity)
   87.7% (global precision)

Conclusion: Serum IL8 concentration improves specificity of clinical scoring system in predicting the histological activity of the disease in UC patients.