Circulating Cytokines Reflect Mucosal Inflammatory Status in Patients with Crohn’s Disease

ORIGINAL ARTICLE

Circulating Cytokines Reflect Mucosal Inflammatory Statusin Patients with Crohn’s Disease

Farid Ljuca • Adnan Gegic • Nermin N. Salkic •

Nada Pavlovic-Calic

Received: 9 March 2009 / Accepted: 24 September 2009

� Springer Science+Business Media, LLC 2009

Abstract

Background There is a great need for a simple activity

assessment tool that can reliably predict activity in patients

with Crohn’s disease (CD).

Aim To investigate the relationship between serum

cytokines and endoscopic activity of CD using Crohn’s

Disease Endoscopic Index of Severity (CDEIS) as a gold

standard.

Methods We prospectively evaluated 32 firmly estab-

lished CD patients using ileocolonoscopy, CDEIS score,

and Crohn’s Disease Activity Index (CDAI) score. Blood

samples for cytokine analysis were obtained 1 day prior to

procedure.

Results The correlation between CDEIS and CDAI was

moderate (r = 0.43; P = 0.01); however, the correlation

between CDEIS and inflammatory cytokines was excellent,

with the highest coefficients for tumor necrosis factor alpha

(TNFa) and interleukin-6 (IL-6) (r = 0.96 and r = 0.96,

respectively; P \ 0.001). CDEIS and anti-inflammatory

cytokines were correlated nonlinearly (power function).

We identified two separate models for predicting CDEIS

value, based on the best performing pro-inflammatory

[CDEIS = 0.445 9 (IL-6) - 5,143] and anti-inflammatory

[CDEIS = 27.478 9 (IL-10)-0.71] cytokines. Both IL-6

and IL-10 models had high adjusted R2 values (0.916 and

0.954, respectively). IL-6 had excellent diagnostic accuracy

for detecting patients with CDEIS[7 (active disease), with

area under the receiver operating characteristic (ROC)

curve of 1.0 [95% confidence interval (CI) = 0.89–1.0;

P \ 0.001].

Conclusion Serum cytokine levels are excellent predic-

tors of endoscopic activity in patients with CD.

Keywords Crohn’s disease � Cytokine �Tumor necrosis factor � Interleukin �Crohn’s Disease Endoscopic Severity Index �Crohn’s Disease Activity Index

Introduction

Treatment of CD is a clinically demanding process that

quite often has to be individualized for each patient. For

this reason clinicians have the need for constant monitoring

and evaluation of disease status, usually through various

clinical, biochemical, and endoscopic assessments.

Various instruments for measuring disease activity have

been developed over the last two decades, but few of them

have been widely adopted by gastroenterology practitio-

ners. Some of these instruments are based mostly on clinical

features of the patient, some are based on quality-of-life

parameters, some take into account biochemical markers,

and some are based exclusively on the endoscopic appear-

ance of intestinal mucosa [1]. However, most of them are

associated with various problems; they either poorly predict

the level of mucosal inflammation (e.g., CDAI score) or

they are hard to use in clinical settings or in clinical trials

(e.g., CDEIS score) [2, 3]. Crohn’s Disease Endoscopic

Farid Ljuca, Adnan Gegic, and Nermin Salkic contributed equally to

this paper.

F. Ljuca

Faculty of Medicine, University of Tuzla, Univerzitetska 1,

Tuzla, Bosnia and Herzegovina

A. Gegic � N. N. Salkic (&) � N. Pavlovic-Calic

Department of Gastroenterology, Internal Medicine Hospital,

University Clinical Center Tuzla, Trnovac bb, 75000 Tuzla,

Bosnia and Herzegovina

e-mail: [email protected]

123

Dig Dis Sci

DOI 10.1007/s10620-009-1016-9

Severity Index (CDEIS) score is currently the only tool that

can reliably evaluate the degree and extent of mucosal

inflammation [4].

The role of cytokines in inflammatory bowel disease

pathogenesis is firmly established. The activation of central

immune-cell populations in inflamed intestinal mucosa is

accompanied by the production of a wide variety of

inflammation mediators, including various cytokines, e.g.,

interleukin-1 (IL-1), IL-6, tumor necrosis factor alpha

(TNFa), interferon gamma (IFNc), and others [5]. Some of

these released cytokines have pro-inflammatory properties

(e.g., IL-1, IL-6, IL-12, TNFa, and IFNc) whereas others

have anti-inflammatory properties (e.g., IL-4, IL-10, and

IL-11).

Although cytokine concentrations are highest at the

inflammation site, they are also easily detectable in plasma

as well. Several reports have established the relationship of

mucosal or plasma levels of cytokines with disease activity

in CD and ulcerative colitis patients [6–9]. Some studies

reported increased levels of pro-inflammatory cytokines

such as IL-6 in active CD and increased levels of anti-

inflammatory cytokines such as IL-10 during recovery

phases of CD [10, 11]. However, all of those studies either

omitted to use CDEIS as a gold standard or investigated

only a limited set of cytokines.

The need for a CD activity assessment tool that is

accurate, simple, easy to use, and readily available in

routine clinical and research practice is obvious. Therefore

we aimed to investigate the relationship between inflam-

matory/anti-inflammatory serum cytokines and mucosal

activity in patients with firmly established CD, using

CDEIS as a gold standard. We hypothesized that serum

cytokines might be used as serum markers of mucosal

activity of CD.

Subjects and Methods

During a 2-year period (2004–2006) we prospectively

recruited all patients referred to our institution for clinical

reevaluation of established CD. The study protocol con-

formed to the Declaration of Helsinki and was approved by

the University Clinical Center Tuzla Ethics Committee.

Inclusion criteria were firmly established diagnosis of CD

(based on a characteristic combination of history data,

laboratory tests, radiology, endoscopy, and histology) and

referral for ileocolonoscopy by the gastroenterologist in

charge of the patient. We excluded all patients without

complete ileocolonoscopy and those with previous history

of surgical procedures related to CD, consequently evalu-

ating only patients with intact bowel and small intestine.

We also excluded all patients treated with anti-TNFamedication(s) (infliximab) during the year prior to

evaluation and patients with clinical or laboratory param-

eters indicating other concomitant disease that might have

systemic inflammatory response (respiratory and genito-

urinary infections, other autoimmune diseases, etc.).

Each patient scheduled for evaluation was given a 7-day

diary with items of relevance to CDAI score (number of

liquid stools, abdominal pain, general well-being, use of

antidiarrheal medications, fever [37.8�C). The diary was

completed with a final entry on the day before procedure,

and then returned to investigators. A blood sample for

standard laboratory tests and cytokines analysis was taken

from each patient upon admission to the diagnostic site,

1 day before ileocolonoscopy. Body weight and height

were measured after thorough physical examination, during

which any presence of extraintestinal manifestations or

abdominal mass were also recorded. A single experienced

gastroenterologist performed all of the mentioned exam-

inations and recorded findings immediately.

The following day, the same gastroenterologist per-

formed total colonoscopy with ileoscopy at least 10 cm in

length in a patient prepared for the procedure in the usual

fashion. Ileum was reached and explored in all patients.

The intestinal mucosa was carefully evaluated for presence

and length of lesions with relevance to CDEIS score and

findings were recorded immediately after the procedure.

For calculating CDEIS, as described by score authors,

the intestine was divided into five segments: ileum, right

colon, transverse colon, left and sigmoid colon, and rec-

tum [4]. The extent of disease involvement in each of the

five segments was determined by examining the presence

of mucosal lesions (pseudopolyps, healed ulcerations,

frank erythema or mucosal swelling, aphthoid ulcers,

superficial or deep ulcers, and nonulcerated or ulcerated

stenosis) and by estimating the percentage of both that

involvement and the ulcerated surface on a 10-cm linear

analog scale for the segment (score between 0 and 10,

with a score of 0 meaning no lesions or ulcerations at all,

and a score of 10 meaning 100% involvement of the

segmental surface; for partially explored segments and for

the ileum, the 10-cm linear scale represented the surface

effectively explored). In each segment the presence of a

superficial ulcer was scored as 6 and a deep ulcer was

scored as 12 (0 if absent). The sum of the score of

involved surface, the score of ulcerated surface, and the

scores of superficial and deep ulcers of each segment was

divided by the number of segments examined, and a score

of 3 was added if either ulcerated or nonulcerated stenosis

was present. The lowest possible CDEIS was 0, repre-

senting an intestine without any lesions, and higher scores

(up to a maximum of 44) indicated more severe disease

[4].

CDAI score was also calculated in a standardized

manner, as previously described [12].

Dig Dis Sci

123

Cytokines Assay

Cytokine levels were determined in the University of

Ljubljana Faculty of Medicine Laboratory (Ljubljana,

Slovenia). Enzyme-linked immunosorbent assay (ELISA)

kit for IL-1 was obtained from Biosource; IL-4, IL-6, IL-10,

IL-11, IL-12, tumor necrosis factor alpha (TNFa), and

gamma interferon (IFNc) ELISA kits were obtained from

Sanquin (Amsterdam, The Netherlands). ELISA was per-

formed by coating 96-well polystyrene microtiter plates

(Nalge Nunc International, Roskilde, Denmark) with a

specific monoclonal antibody. Standards and samples were

added after the plates were blocked, and the plates were

incubated overnight at 4�C. A specific biotinylated antibody

was added to all wells after they were washed, and they

were incubated for 1 h at room temperature. The plates

were washed and incubated for 30 min with horseradish

peroxidase-conjugated streptavidin. After removal of non-

bound horseradish peroxidase conjugate by washing,

3-30,5,50-tetramethylbenzidine substrate reagent solution

(ICN Biomedicals Inc., Aurora, OH) was added to the wells.

The reaction was stopped by addition of 1.8 M H2SO4. The

absorbance of all ELISAs was read at 450 nm with a

Milenia microtiter plate reader (Diagnostic Products Cor-

poration Nederland BV, Breda, The Netherlands). Standard

curves for the various cytokines ranging from 0.25 to 5 pg/

ml and up to 300–1,000 pg/ml were constructed by a four-

parameter regression formula and plotted as a linear curve

(log–log). Cytokine concentrations of experimental samples

were calculated with Elisa plus software version 3.01

(Meddata Inc., New York, NY) [13, 14].

Statistical Analysis

All statistical analysis procedures were performed using

SPSS version 12.0 (SPSS, Chicago, IL). Standard parameters

of descriptive statistics were used for determination of

baseline characteristics. All variables satisfied criteria for

accepting a normal distribution on testing with the

Kolmogorov–Smirnov test. Student’s T-test with correction

for unequal variances, where appropriate, or analysis of

variance (ANOVA) test for more than two variables, was

used to compare quantitative variables. Post hoc analysis

after performing ANOVA was made by using Tukey test.

Chi-square test was used to compare categorical variables.

Pearson’s correlation with two-tailed significance was used

to investigate the strength of links between analyzed

variables.

Curve estimation regression analysis was used to

investigate nonlinear relationships between anti-inflam-

matory cytokines and CDEIS score. Simple linear regres-

sion analysis was used to evaluate the best performing

pro-inflammatory cytokine for the prediction of CDEIS

value. A similar approach was used for anti-inflammatory

cytokines, only we used nonlinear regression analysis

based on a power model.

In order to validate the regression models in the setting

of a limited sample size, we used the ‘‘bootstrap’’ method

[15]. We randomly split the sample 300 times into test and

validation groups. Each time, after splitting the sample, a

regression analysis (linear or nonlinear) was performed

with CDEIS as a dependent and IL-6 or IL-10 as an

independent variable. Cross-validation between test and

validation groups was performed each time after splitting.

Coefficients for each particular regression analysis and

cross-validation were recorded. In this way we were able to

analyze the distribution of the obtained regression and

correlation coefficients with a final aim of improving

robustness and validity of analysis. The entire procedure

was performed automatically by using a custom-made

script created in SPSS Syntax Editor.

A 95% level of significance was accepted for all tests.

Results

We initially recruited 37 patients, but 5 of them were

excluded due to various exclusion criteria. Finally, a

sample of 32 patients was analyzed. Median age of the

study sample was 40 years (25–75 percentile: 26–49 years)

with a range of 18–71 years. There were 21/32 (65.6%)

males and 11/32 (34.4%) females (ratio 1.91:1). Summary

statistics for CDAI, CDEIS, and cytokine levels are pre-

sented in Table 1.

Median time since diagnosis was 22 months (25–75

percentile: 16–32 months), ranging from 9 to 53 months.

All of the patients were receiving therapy with mesalazine at

the time of evaluation. There were also 22/32 (68.75%)

patients on concomitant therapy with systemic corticoste-

roids and 6/32 (18.8%) patients on concomitant azathio-

prine, but there were no patients on methotrexate or

anti-TNFa drugs. Patients treated with concomitant corti-

costeroids were on a mean (±standard deviation, SD) daily

dose of 15 mg (±10 mg) of prednisone equivalent. We did

not find any significant difference in values of cytokines,

CDEIS or CDAI score between patients on versus off

corticosteroids (P [ 0.05).

Half of our patients (16/32) had ileal localization of the

disease. An additional 12/32 patients (37.5%) had ileoco-

lonic localization, while the remaining 4/32 patients

(12.5%) had colonic localization. We compared average

values of measured cytokines according to disease locali-

zation. As seen from Table 2, there was a significant

between-group difference in cytokine levels, with a clear

tendency for lower levels of pro-inflammatory cytokines

and higher levels of anti-inflammatory cytokines in patients

Dig Dis Sci

123

with ileal localization when compared with those with

ileocolonic/colonic localization. Post hoc group-by-group

analysis with Tukey test also confirmed this (Table 3).

There were 18 (56.3%) patients with CDAI B150 and 14

(43.8%) patients with CDAI [150. Patients with CDAI

[150 (mean ± SD: 17.54 ± 11.46) had significantly

higher (corr. t = 19.11; df = 14; P = 0.01) CDEIS than did

patients with CDAI B150 (8.26 ± 2.58), with a mean dif-

ference of 9.27 (95% CI = 2.58–15.97). There were 13/18

(72.2%) patients with CDEIS [7 (endoscopically active

disease) within the group of those with CDAI B150 (clinical

remission), in comparison with 13/14 (92.9%) patients

within the group of those with CDAI [150 (v2 = 1.06;

df = 1; P = 0.30).

Table 2 Comparison of mean cytokine levels according to disease localization: analyzed by ANOVA

Cytokine Disease localization Mean SD Minimum Maximum P-value (ANOVA)

IL-1 Ileal 27.46 13.97 9.80 49.70 0.004

Ileocolonic 46.11 16.16 29.80 80.90

Colonic 55.40 29.16 35.10 98.70

IL-4 Ileal 7.45 4.78 1.80 20.40 0.011

Ileocolonic 3.33 1.87 0.90 5.80

Colonic 2.88 1.94 0.80 5.50

IL-6 Ileal 29.11 12.62 10.80 51.80 0.006

Ileocolonic 47.38 17.33 30.40 82.60

Colonic 55.05 28.31 33.80 96.80

IL-10 Ileal 7.47 5.12 2.00 22.80 0.014

Ileocolonic 3.31 1.80 0.70 5.80

Colonic 2.68 1.97 0.70 5.40

IL-11 Ileal 7.74 5.93 1.90 26.60 0.026

Ileocolonic 3.33 1.86 0.90 5.70

Colonic 2.80 1.83 0.90 5.30

IL-12 Ileal 29.30 14.08 11.50 50.90 0.006

Ileocolonic 47.94 17.31 31.10 89.00

Colonic 56.78 28.97 34.20 99.30

TNFa Ileal 33.05 12.47 10.50 54.60 0.007

Ileocolonic 51.04 19.39 31.30 97.90

Colonic 60.58 27.86 35.90 100.60

INFc Ileal 32.37 13.25 7.40 56.90 0.005

Ileocolonic 51.39 17.77 34.80 92.30

Colonic 60.40 28.95 37.40 102.80

Cytokine levels are expressed as picograms per milliliter (pg/ml)

IL interleukin, IFNc interferon gamma, TNFa tumor necrosis factor alpha, SD standard deviation, ANOVA analysis of variance

Table 1 Summary statistics for

values of CDAI, CDEIS, and

cytokines

Cytokine levels are expressed as

picograms per milliliter (pg/ml)

CDAI Crohn’s Disease Activity

Index, CDEIS Crohn’s Disease

Endoscopic Index of Severity,

IL interleukin, IFNc interferon

gamma, TNFa tumor necrosis

factor alpha

Mean SD Median Minimum Maximum 25–75 Percentile

CDAI 151.81 82.24 124.50 56.00 370.00 91.00–205.00

CDEIS 12.32 8.98 9.20 4.20 40.40 7.45–14.10

IL-1 37.95 19.79 37.35 9.80 98.70 30.30–43.90

IL-10 5.31 4.36 4.50 0.70 22.80 2.30–5.60

IL-11 5.47 4.89 4.65 0.90 26.60 2.50–5.65

IL-12 39.73 20.08 39.10 11.50 99.30 31.25–46.80

IL-4 5.33 4.16 4.75 0.80 20.40 2.60–5.65

IL-6 39.20 19.29 36.65 10.80 96.80 30.80–44.80

IFNc 43.01 20.07 39.65 7.40 102.80 35.35–50.70

TNFa 43.24 19.99 38.35 10.50 100.60 32.10–52.90

Dig Dis Sci

123

Table 3 Comparison of mean cytokine levels according to disease localization: post hoc group-by-group analysis by Tukey test

Cytokine Localization (I) Localization (J) Mean difference (I–J) P-value 95% Confidence interval of mean difference

Upper bound Lower bound

IL-1 Ileal Ileocolonic -18.65 0.02 -2.64 -34.65

Colonic -27.94 0.02 -4.51 -51.36

Ileocolonic Ileal 18.65 0.02 34.65 2.64

Colonic -9.29 0.61 14.90 -33.48

Colonic Ileal 27.94 0.02 51.36 4.51

Ileocolonic 9.29 0.61 33.48 -14.90

IL-4 Ileal Ileocolonic 4.12 0.02 7.59 0.65

Colonic 4.58 0.08 9.65 -0.50

Ileocolonic Ileal -4.12 0.02 -0.65 -7.59

Colonic 0.46 0.97 5.70 -4.79

Colonic Ileal -4.58 0.08 0.50 -9.65

Ileocolonic -0.46 0.97 4.79 -5.70


Colonic -25.94 0.03 -2.87 -49.01


Colonic -7.67 0.71 16.15 -31.50

Colonic Ileal 25.94 0.03 49.01 2.87

Ileocolonic 7.67 0.71 31.50 -16.15


Colonic 4.79 0.09 10.17 -0.58


Colonic 0.63 0.96 6.19 -4.92

Colonic Ileal -4.79 0.09 0.58 -10.17

Ileocolonic -0.63 0.96 4.92 -6.19


Colonic 4.94 0.13 11.09 -1.22


Colonic 0.53 0.98 6.89 -5.82

Colonic Ileal -4.94 0.13 1.22 -11.09

Ileocolonic -0.53 0.98 5.82 -6.89


Colonic -27.48 0.02 -3.45 -51.50


Colonic -8.83 0.66 15.99 -33.65

Colonic Ileal 27.48 0.02 51.50 3.45

Ileocolonic 8.83 0.66 33.65 -15.99

TNFa Ileal Ileocolonic -17.99 0.03 -1.56 -34.42

Colonic -27.53 0.02 -3.48 -51.57


Colonic -9.53 0.61 15.30 -34.37

Colonic Ileal 27.53 0.02 51.57 3.48

Ileocolonic 9.53 0.61 34.37 -15.30

Dig Dis Sci

123

A correlation matrix of CDAI, CDEIS, and cytokines is

presented in Table 4, showing that all correlations were

significant. The correlation between CDAI and CDEIS

appeared to be positive and low to moderate in strength.

Correlation of CDEIS with anti-inflammatory cytokines

(IL-4, IL-10, and IL-11) was negative and also low to

moderate in strength. However, correlation between

CDEIS and pro-inflammatory cytokines (IL-1, IL-6,

IL-12, IFNc, and TNFa) appeared to be excellent. Nota-

bly, the correlation between CDAI and all cytokines was

similar in strength to the correlation between CDAI and

CDEIS, with negative coefficients for anti-inflammatory

cytokines.

Additional analysis showed that, in contrast to pro-

inflammatory cytokines that exhibited linear correlation

with CDEIS score, anti-inflammatory cytokines exhibited

nonlinear correlation with CDEIS, corresponding to the

power function (y = a 9 xb). Relationships between each

of the evaluated cytokines and CDEIS are depicted in

Figs. 1, 2, 3, and 4.

We used regression analysis to construct a prediction

model for CDEIS based on cytokine values. Since there

were significant correlations between cytokine variables,

which would produce undesirable multicollinearity in a

final model if entered together, we tested each cytokine

separately in order to choose the parameter with best pre-

dictive power.

On testing pro-inflammatory cytokines using linear

regression analysis, each was a significant predictor of

CDEIS value, but IL-6 showed the best performance. The

model that included IL-6 (b = 0.957; P \ 0.001) as an

independent variable was highly significant (F = 325.69;

df = 1; P \ 0.001). The equation identified for this model

was CDEIS = 0.445 9 (IL-6) - 5.143. The adjusted R2

value was 0.916, which indicates that 91.6% of the vari-

ance in CDEIS was explained by TNFa value (a very large

effect). The standard error of the estimate by this model

was 2.65. A graphical representation of this regression line

in relation to observed data is depicted in Fig. 1.

Nonlinear regression analysis was conducted to inves-

tigate how well anti-inflammatory cytokines predicted

CDEIS value. Each of the anti-inflammatory cytokines

exhibited excellent performance for prediction of CDEIS

value, but the model that included IL-10 (b = -0.977;

P \ 0.001) performed the best (F = 617.18; df = 1;

P \ 0.001). The equation identified for this model was

CDEIS = 27.478 9 (IL-10)-0.71 with adjusted R2 value of

0.954 and standard error of the estimate of 0.126. A

graphical representation of this regression line in relation to

observed data is depicted in Fig. 2.

A model that included IFNc was also evaluated

[CDEIS = 0.424 9 (IFNc) - 5.926]; however, the adjus-

ted R2 value, although rather high, was somewhat lower

(0.89) than that of the other two models and it was not

further explored.

Receiver operating characteristics analysis was per-

formed in order to test the diagnostic performance of IL-6

for detecting patients with CDEIS [7, which was sug-

gested as a threshold to distinguish between patients with

active disease and those in remission [16]. The area under

the ROC curve (AUROC) was 1.0 (95% CI = 0.89–1.0;

P \ 0.001) for a cutoff value of 15.9 pg/ml. This threshold

had 100% sensitivity (95% CI = 86.7–100), specificity

(95% CI = 54.1–100), and positive and negative predictive

values. A dot diagram depicting the observed cases in

regard to the proposed cutoff value is presented in Fig. 5.

In order to test the validity of the IL-6-based predictive

model for CDEIS and to increase the robustness of the

results in the setting of a limited sample size, we used the

bootstrap method, as described in the ‘‘Methods’’ section.

After splitting the sample 300 times (into test and valida-

tion groups) and performing linear regression and internal

cross-validation each time, we analyzed the distribution of

all of the obtained coefficients.

The mean (±SD) R2 value was 0.91 ± 0.03 and each of

the 300 regression models was statistically significant

(P \ 0.001). The average correlation coefficient obtained

for the validation groups was 0.95 ± 0.037.

Table 3 continued

Cytokine Localization (I) Localization (J) Mean difference (I–J) P-value 95% Confidence interval of mean difference

Upper bound Lower bound

INFc Ileal Ileocolonic -19.02 0.02 -2.76 -35.28

Colonic -28.03 0.02 -4.23 -51.84


Colonic -9.01 0.64 15.58 -33.59

Colonic Ileal 28.03 0.02 51.84 4.23

Ileocolonic 9.01 0.64 33.59 -15.58

Cytokine levels are expressed as picograms per milliliter (pg/ml)

IL interleukin, IFNc interferon gamma, TNFa tumor necrosis factor alpha

Dig Dis Sci

123

Similar analysis was performed in order to test the

validity of IL-10 as a predictor of CDEIS. The only dif-

ference was that, instead of linear regression, we performed

nonlinear, power model regression after each split of the

sample. The mean (±SD) R2 value was 0.95 ± 0.02 and

each of the 300 regression models was statistically signif-

icant (P \ 0.001). The average correlation coefficient

obtained for the validation groups was 0.98 ± 0.014.

Discussion

Mucosal healing is considered the ultimate aim of treat-

ment in CD. It has been demonstrated that achieving this

specific goal reduces the number of hospitalizations and

surgical procedures, which additionally emphasizes its

clinical relevance [17, 18]. Consequently, clinicians need a

reliable yet easy to use disease activity index that is cheap

and simple, preferably consisting of widely available and

easy-to-collect parameters [3].

The Crohn’s Disease Activity Index (CDAI), being one

of the first developed clinical monitoring tools, is to date

the most widely accepted index for clinical monitoring and

the instrument most frequently used in clinical trials [12].

The score is based on eight independent variables (number

of liquid stools, severity of abdominal pain, general well-

being, presence of extraintestinal manifestations, need for

antidiarrheal drugs, presence of abdominal mass, hemato-

crit value, and body weight) monitored via 7-day diary.

Nevertheless, CDAI score is prone to interobserver vari-

ability [19]. It also correlates poorly with inflammatory

status of the intestinal mucosa, which is one of its greatest

Table 4 Correlation matrix for CDAI, CDEIS, and cytokines

CDEIS CDAI IL-4 IL-10 IL-11 IL-1 IL-6 IL-12 IFNc TNFa

CDEIS

r N/A 0.434 -0.627 -0.601 -0.559 0.939 0.957 0.943 0.948 0.957

P 0.0130 0.0001 0.0003 0.0009 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001

CDAI

r 0.434 N/A -0.442 -0.419 -0.412 0.453 0.445 0.466 0.466 0.467

P 0.0130 0.0114 0.0170 0.0192 0.0093 0.0108 0.0072 0.0071 0.0071

IL-4

r -0.627 -0.442 N/A 0.992 0.981 -0.801 -0.793 -0.798 -0.809 -0.788

P 0.0001 0.0114 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001

IL-10

r -0.601 -0.419 0.992 N/A 0.994 -0.769 -0.762 -0.765 -0.786 -0.764

P 0.0003 0.0170 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001

IL-11

r -0.559 -0.412 0.981 0.994 N/A -0.723 -0.719 -0.718 -0.745 -0.723

P 0.0009 0.0192 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001

IL-1

r 0.939 0.453 -0.801 -0.769 -0.723 N/A 0.996 0.996 0.989 0.978

P \0.0001 0.0093 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001

IL-6

r 0.957 0.445 -0.793 -0.762 -0.719 0.996 N/A 0.993 0.991 0.985

P \0.0001 0.0108 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001

IL-12

r 0.943 0.466 -0.798 -0.765 -0.718 0.996 0.993 N/A 0.989 0.987

P \0.0001 0.0072 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001

IFNc

r 0.948 0.466 -0.809 -0.786 -0.745 0.989 0.991 0.989 N/A 0.990

P \0.0001 0.0071 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001

TNFa

r 0.957 0.467 -0.788 -0.764 -0.723 0.978 0.985 0.987 0.990 N/A

P \0.0001 0.0071 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001 \0.0001

CDAI Crohn’s Disease Activity Index, CDEIS Crohn’s Disease Endoscopic Index of Severity, IL interleukin, IFNc interferon gamma,

TNFa tumor necrosis factor alpha, r Pearson correlation coefficient, P significance level

Dig Dis Sci

123

limitations [20, 21]. Given the fact that mucosal healing is

presently considered a main treatment goal in CD, recent

reviews suggest that CDAI cannot remain as primary

endpoint in clinical trials [3].

Our results also suggest that CDAI correlates poorly

with mucosal status (Table 2). The development of CDAI

was based on laboratory markers that later failed to be

proved reliable predictors of mucosal inflammation, and a

considerable part of the score is based on purely subjective

symptoms, thus leading to its limitations and failure to

predict inflammation [3]. In our sample we also found over

70% of patients with CDEIS[7 (those with endoscopically

active disease) within those defined to be in remission

according to CDAI (CDAI B150), which further under-

lines the imperfection of CDAI.

At present, severity of mucosal inflammation assessed by

endoscopy is considered the gold standard for disease activity

in CD. Crohn’s Disease Endoscopic Index of Severity

(CDEIS) proposed by a group of French investigators

(GETAID group) in 1989 emerged in the following years as a

gold standard for evaluation of endoscopic activity [4]. This

index is based on endoscopic findings in five segments of the

gut (ileum, right colon, transverse colon, left and sigmoid

colon, and rectum). The elaboration of the score requires

analog scale transformation. Based on data from five pro-

spective GETAID studies that included 562 colonoscopies

performed in 231 patients, it has been suggested that endo-

scopic remission might be defined as CDEIS score lower than

6 or 7 [16]. The CDEIS, although reliable and reproducible, is

rather time consuming and technically demanding.

However, CDEIS is currently the only tool that can

reliably evaluate activity of mucosal inflammation. The

score is based on ileocolonoscopy findings, which can lead

to numerous practical problems that are a serious obstacle

for routine application, and its use can be complex in

clinical trials. Ileocolonoscopy is invasive, expensive, and

time consuming; the major part of the small bowel is not

within the reach of endoscopy, therefore remaining outside

of this particular scoring system, which might lead to

potential underestimation. Although there are proposed

threshold values for remission or response following

medical therapy, their validity is still under evaluation [16].

In the recent literature there have been several attempts

to identify serum indices of mucosal inflammation. To date,

several serum markers of CD activity, alone or in various

combinations, have been evaluated. Most of the published

studies reported significant but rather weak correlations

with severity of mucosal inflammation [20–24]. Some of the

evaluated parameters include hemoglobin, hematocrit,

erythrocyte sedimentation rate (ESR), albumin, C-reactive

protein (CRP), platelets, a1-antitrypsin, orosomucoid, and

sialic acid. One study reported a high correlation rate of

serum albumin (r = 0.8) with mucosal inflammation, but

the study sample was not uniform and included 28 patients

with CD colitis and ulcerative colitis [25].

Several studies have established a connection between

disease activity and levels of circulating cytokines (both

pro-inflammatory and anti-inflammatory), implying their

potential use as monitoring tools [6, 10, 11]. Mitsuyama

et al. reported that serum concentrations of IL-10 increased

Fig. 1 Scattergram depicting the linear relationship between IL-6

and CDEIS. Solid line represents the equation identified by linear

regression: CDEIS = 0.445 9 (IL-6) - 5,143

Fig. 2 Scattergram depicting the nonlinear (power) relation-

ship between IL-10 and CDEIS. Solid line represents the equa-

tion identified by nonlinear (power function) regression:

CDEIS = 27.478 9 (IL-10)-0.71

Dig Dis Sci

123

during phases of disease resolution, while serum levels of

IL-6 and CRP increased during the active phase of CD

[11]. However, disease activity of CD in this particular

study was assessed according to scores established by the

International Organization for the Study of Inflammatory

Bowel Disease (IOIBD), which is also a clinical and not an

endoscopic activity index [26]. Our study is the first

attempt to correlate CDEIS (as the most reliable mucosal

activity index) with serum cytokine levels.

Our results suggest that circulating cytokines could be

used as reliable and noninvasive markers of endoscopic

activity in CD. According to our findings, levels of pro-

Fig. 3 Scattergrams depicting

the nonlinear (power)

relationships between anti-

inflammatory cytokines (IL-4

and IL-11) and CDEIS. Solidline in each diagram represents

the nonlinear (power) regression

line

Fig. 4 Scattergrams depicting

the linear relationship between

pro-inflammatory cytokines

(IL-1, IL-6, IL-12, and INF-c)

and CDEIS. Solid line in each

diagram represents the linear

regression line

Dig Dis Sci

123

inflammatory and anti-inflammatory cytokines predict the

degree of endoscopic activity with more than adequate

accuracy. Pro-inflammatory cytokines, due to their positive

linear relationship with mucosal inflammation, are perhaps

easier to use in routine clinical practice, especially after

CDEIS thresholds for remission and therapy response are

validated. ROC analysis of the diagnostic accuracy of IL-6

for predicting patients with endoscopically active disease

(CDEIS [7) demonstrated excellent diagnostic perfor-

mance, thus proposing another possible clinical application.

Apparently, anti-inflammatory cytokines provide even

better prediction of CDEIS score, with standard error of

estimate of 0.126 for IL-10 and very high R2 values.

However, due to their nonlinear and negative correlation

with CDEIS, they are less straightforward markers,

although they could be used in clinical trials or in any other

situation where accurate prediction of mucosal status is of

value.

In our paper we are proposing the utilization of two

cytokines: IL-6 and IL-10. However, the scattergrams for

the relationships between CDEIS and other anti-inflam-

matory and pro-inflammatory cytokines evaluated in this

study (Figs. 3, 4) look almost identical to those in Figs. 1

and 2. Our choice of IL-6 and IL-10 was based on the fact

that their regression models had the best R2 values and also

on the fact that their relationships with disease activity

were already established in previous studies, although

without validation against CDEIS [11]. However, in any

future research with a larger sample, this choice can be

reevaluated and perhaps any of the cytokines evaluated in

our research might prove to be better.

There is reasonable consensus that mucosa of patients

with established Crohn’s disease is dominated by CD4?

lymphocytes with type 1 helper-T cell (Th1) phenotype,

characterized by production of IFNc and IL-2 [5]. There-

fore, one may assume that IL-2 and/or IFNc levels could

have a significant predictive potential for prediction of

mucosal status. Although we have not measured IL-2, we

did evaluate the prognostic performance of IFNc and found

it to be excellent, although slightly inferior to that of IL-6

and IL-10, which led us to prefer the IL-6 and IL-10

models.

In a recent paper, Italian authors reported significantly

higher levels of serum IL-6 and TNFa in patients with

perianal CD when separately compared with groups of

patients with small bowel CD and healthy controls [27].

We have also observed a clear difference in serum cytokine

levels in three different CD localization groups (Tables 2,

3) with higher levels of anti-inflammatory cytokines and

lower levels of pro-inflammatory cytokines in patients with

ileal localization of the disease. Whether this is a result of

the extent of intestinal surface involved or perhaps a result

of different disease behavior and different immunological

response in ileojejunal phenotype of the disease (which is

already recognized as a separate and distinct form of CD)

remains to be answered [28–30].

There are several important limitations of this study that

need to be addressed. Sample size is the obvious issue, but

CD is a relatively rare disease in our region, with 140 new

cases detected in the 12-year period of 1995–2006 [31].

Therefore, obtaining a reasonably large sample would

significantly prolong this research. One could also raise

questions about validation of the proposed score(s).

Although we used the bootstrap method in order to increase

the robustness of our internal validation, we agree that

external validation would provide more reliable results.

Another important issue is the value of cytokines as

markers of mucosal inflammation in patients receiving

different types of treatment, particularly anti-TNFa ther-

apy. In order to test this, one would need a significantly

larger sample and a substantial number of patients on anti-

TNFa drugs. Bearing in mind the relative prevalence of the

CD in our region and the fact that anti-TNFa treatment is

very expensive and not affordable for most of our patients,

achieving these conditions was not possible in our setting.

However, in this context, it is very important to note that

our sample included patients receiving different types of

treatment (except anti-TNFa), with a variable time since

initiating treatment and since diagnosis. In a diverse sam-

ple such as this, cytokines predicted the degree of mucosal

inflammation with more than good accuracy. Nonetheless,

the true value of cytokines in noninvasive monitoring of

therapy response is the question of great clinical signifi-

cance that needs to be answered in future studies.

Despite limitations, we believe that this study brings

very important information about the relationship between

circulating cytokines and endoscopic activity of CD. This

Fig. 5 Dot diagram depicting the observed cases in regard to the

proposed cutoff value of 15.9 pg/ml for IL-6

Dig Dis Sci

123

fact, on its own, might be the first step in the direction of

creating the simple and reliable CD activity index that we

all need so much.

To conclude, serum cytokines are excellent predictors of

mucosal inflammation in patients with CD. The best per-

forming predictor for mucosal inflammation among the

pro-inflammatory cytokines was IL-6, while the best per-

forming anti-inflammatory cytokine appeared to be IL-10,

although others were almost as good as those two. We

propose predictive models based on IL-6 and/or IL-10 that

might be used as accurate and simple mucosal activity

monitoring tools in CD.

References

1. Sostegni R, Daperno M, Scaglione N, et al. Review article:

Crohn’s disease: Monitoring disease activity. Aliment PharmacolTher. 2003;17(Suppl 2):11–17.

2. Biancone L, De Nigris F, Del Vecchio Blanco G, et al. Review

article: Monitoring the activity of Crohn’s disease. AlimentPharmacol Ther. 2002;16(Suppl 4):29–33.

3. Minderhoud IM, Samsom M, Oldenburg B. What predicts

mucosal inflammation in Crohn’s disease patients? InflammBowel Dis. 2007;13:1567–1572.

4. Mary JY, Modigliani R. Development and validation of an

endoscopic index of the severity for Crohn’s disease: A pro-

spective multicentre study. Groupe d’Etudes Therapeutiques des

Affections Inflammatoires du Tube Digestif (GETAID). Gut.1989;30:983–989.

5. Podolsky DK. Inflammatory bowel disease. N Engl J Med.

2002;347:417–429.

6. Nancey S, Hamzaoui N, Moussata D, et al. Serum interleukin-6,

soluble interleukin-6 receptor and Crohn’s disease activity. DigDis Sci. 2008;53:242–247.

7. Yamamoto T, Nakahigashi M, Saniabadi AR, et al. Impacts of

long-term enteral nutrition on clinical and endoscopic disease

activities and mucosal cytokines during remission in patients with

Crohn’s disease: A prospective study. Inflamm Bowel Dis.

2007;13:1493–1501.

8. Yamamoto T, Umegae S, Kitagawa T, Matsumoto K. Systemic

and local cytokine production in quiescent ulcerative colitis and

its relationship to future relapse: A prospective pilot study. In-flamm Bowel Dis. 2005;11:589–596.

9. Spoettl T, Hausmann M, Klebl F, et al. Serum soluble TNF

receptor I and II levels correlate with disease activity in IBD

patients. Inflamm Bowel Dis. 2007;13:727–732.

10. Mahida YR, Kurlac L, Gallagher A, Hawkey CJ. High circulating

concentrations of interleukin-6 in active Crohn’s disease but not

ulcerative colitis. Gut. 1991;32:1531–1534.

11. Mitsuyama K, Tomiyasu N, Takaki K, et al. Interleukin-10 in the

pathophysiology of inflammatory bowel disease: increased serum

concentrations during the recovery phase. Mediators Inflamm.

2006;2006:26875.

12. Best WR, Becktel JM, Singleton JW, Kern F Jr. Development of

a Crohn’s disease activity index. National cooperative Crohn’s

disease study. Gastroenterology. 1976;70:439–444.

13. de Jager W, Prakken BJ, Bijlsma JW, Kuis W, Rijkers GT.

Improved multiplex immunoassay performance in human plasma

and synovial fluid following removal of interfering heterophilic

antibodies. J Immunol Methods. 2005;300:124–135.

14. de Jager W, te Velthuis H, Prakken BJ, Kuis W, Rijkers GT.

Simultaneous detection of 15 human cytokines in a single sample

of stimulated peripheral blood mononuclear cells. Clin Diagn LabImmunol. 2003;10:133–139.

15. Fox J. Bootstrapping, In: Fox J, ed. Applied Regression Analysis,Linear Models, and Related Methods Thousand Oaks, CA: Sage;

1997.

16. Mary JY, Lemann M, Colombel JF, Lerebours E, Soule JC,

Gendre JP, Modigliani R, The GETAID. Endoscopic remission

and response in Crohn’s disease: An objective definition using the

CDEIS, In: 13th United European Gastroenterology Week.

Copenhagen: Denmark; 2005.

17. Hanauer SB, Feagan BG, Lichtenstein GR, et al. Maintenance

infliximab for Crohn’s disease: The ACCENT I randomised trial.

Lancet. 2002;359:1541–1549.

18. Rutgeerts P, Feagan BG, Lichtenstein GR, et al. Comparison of

scheduled and episodic treatment strategies of infliximab in

Crohn’s disease. Gastroenterology. 2004;126:402–413.

19. de Dombal FT, Softley A. IOIBD report no 1: Observer variation

in calculating indices of severity and activity in Crohn’s disease.

International organisation for the study of inflammatory bowel

disease. Gut. 1987;28:474–481.

20. Cellier C, Sahmoud T, Froguel E, et al. Correlations between

clinical activity, endoscopic severity, and biological parameters

in colonic or ileocolonic Crohn’s disease. A prospective multi-

centre study of 121 cases. The Groupe d’Etudes Therapeutiques

des Affections Inflammatoires Digestives. Gut. 1994;35:231–235.

21. Modigliani R, Mary JY, Simon JF, et al. Clinical, biological,

and endoscopic picture of attacks of Crohn’s disease. Evolution

on prednisolone. Groupe d’Etude Therapeutique des Affections

Inflammatoires Digestives. Gastroenterology. 1990;98:811–818.

22. Tromm A, Tromm CD, Huppe D, et al. Evaluation of different lab-

oratory tests and activity indices reflecting the inflammatory activity

of Crohn’s disease. Scand J Gastroenterol. 1992;27:774–778.

23. Simonis B, Heine M, Heene DL, Gladisch R. Evaluation and

validation of a Crohn’s disease inflammatory activity index

reflecting pattern of endoscopic severity. Scand J Gastroenterol.1998;33:283–288.

24. Solem CA, Loftus EV Jr, Tremaine WJ, et al. Correlation of C-

reactive protein with clinical, endoscopic, histologic, and radio-

graphic activity in inflammatory bowel disease. Inflamm BowelDis. 2005;11:707–712.

25. Moran A, Jones A, Asquith P. Laboratory markers of colono-

scopic activity in ulcerative colitis and Crohn’s colitis. Scand JGastroenterol. 1995;30:356–360.

26. Myren J, Bouchier IA, Watkinson G, et al. The O.M.G.E. Mul-

tinational Inflammatory Bowel Disease Survey 1976–1982. A

further report on 2, 657 cases. Scand J Gastroenterol Suppl.1984;95:1–27.

27. Ruffolo C, Scarpa M, Faggian D, et al. Cytokine network in

rectal mucosa in perianal Crohn’s disease: Relations with

inflammatory parameters and need for surgery. Inflamm BowelDis. 2008;14:1406–1412.

28. Freeman HJ. Long-term clinical behavior of jejunoileal

involvement in Crohn’s disease. Can J Gastroenterol. 2005;19:

575–578.

29. Freeman HJ. Long-term natural history of Crohn’s disease. WorldJ Gastroenterol. 2009;15:1315–1318.

30. Touze I, Gower-Rousseau C, Grandbastien B, et al. Diffuse je-

juno-ileitis of Crohn’s disease: a separate form of the disease?

Gastroenterol Clin Biol. 1999;23:307–311.

31. Pavlovic-Calic N, Salkic NN, Gegic A, Smajic M, Alibegovic E.

Crohn’s disease in Tuzla region of Bosnia and Herzegovina: A

12-year study (1995–2006). Int J Colorectal Dis. 2008;23:

957–964.

Dig Dis Sci

123

Circulating Cytokines Reflect Mucosal Inflammatory Status in Patients with Crohn’s Disease

Documents

Transcript of Circulating Cytokines Reflect Mucosal Inflammatory Status in Patients with Crohn’s Disease