The impact of intestinal resection on serum levels of anti-Saccharomyces cerevisiae antibodies (ASCA) in patients with Crohn’s disease

Authors


Dr W. Reinisch, Department of Internal Medicine III, Division of Gastroenterology and Hepatology, Medical University of Vienna, Waehringer Guertel 18-20, A-1090 Vienna, Austria.
E-mail: walter.reinisch@meduniwien.ac.at

Abstract

Aliment Pharmacol Ther 2012; 35: 292–299

Summary

Background  Antibodies to Saccharomyces cerevisiae (ASCA) are highly prevalent in sera of patients with Crohn’s disease and have been proposed to identify subgroups of patients with a disabling disease course.

Aim  To investigate the impact of intestinal resection on serum levels of ASCA in patients with Crohn’s disease and the predictive value of ASCA levels on surgical recurrence.

Methods  Sera from 60 patients who underwent ‘curative’ intestinal resection due to stricturing and/or penetrating complications were collected preoperatively and during post-operative follow-up (week 2, months 4, 8 and 11 ± 1). Measurement of ASCA IgG and IgA isotypes were performed using ELISA. Re-operation rate was associated with ASCA status and serum levels.

Results  At baseline 44/60 (73%) of patients were rated as positive for ASCA IgG, 45/60 (75%) for ASCA IgA and 52/60 (87%) as positive for at least one of both. ASCA serum levels remained stable during first year from resection. After a median of 106 months 10 of 40 (25%) patients with long-term follow-up underwent one or more intestinal re-operations. Neither ASCA positivity nor absolute ASCA serum levels were predictive of surgical recurrence.

Conclusions  Serum ASCA levels remain stable after curative intestinal resection in Crohn’s disease. This indicates the persistence of both stimulus and immunological mechanism operative in the production of ASCA even after complete surgical resection of macroscopically inflamed intestinal tissue. After intestinal resection, neither ASCA positivity nor ASCA serum levels predict the risk of surgical recurrence during long-term follow-up.

Introduction

Crohn’s disease (CD) represents a major phenotype of inflammatory bowel disease, characterised by chronic remitting flares from primarily ileocolonic lesions. Evidence is accumulating that luminal bacteria, impairment of intestinal immunity and environmental factors contribute to the inflammatory response, predisposed by a genetic susceptibility.1 During the course of disease intestinal resection is required in 70–80% of patients, owing to stricturing or penetrating complications.2, 3 However, even after complete resection of inflamed intestinal segments, CD recurs endoscopically in up to 90% of patients within the first post-operative year. A clinical relapse follows in 34% of patients within 3 years after major surgery.4–6 So far, particularly the assessment of post-operative endoscopic lesions is used to predict the risk of relapse.6

In CD, anti-Saccharomyces cerevisiae antibodies (ASCA) of the isotypes IgG or IgA have been demonstrated in sera of 60–70% of patients, particularly in those with ileal involvement.7 Anti-Saccharomyces cerevisiae antibodies have been associated with complicated disease course, increased hospital costs and an increased risk of small bowel surgery in CD.8–13Saccharomyces cerevisiae is an ascomycetous yeast extensively used in the brewing and baking industry.14 Therefore, the human alimentary tract is widely exposed to its antigens. However, increased levels of ASCA are only found in up to 5% of the healthy population, indicating a general mucosal tolerance to the widespread yeast and its antigens.7 As among the existing serologic markers in CD, ASCAs are most prevalent and of greatest predictive value for complicated disease course,7 we were interested in their long-term stability. We investigated the influence of intestinal resection on ASCA serum levels of both IgG and IgA isotypes by performing serial pre- and repetitive post-operative measurements and subsequently by long-term follow-up evaluated ASCA as prognostic variable to predict the risk of surgical recurrence.

Materials and methods

Patient population

Serum samples were collected prospectively from 60 patients with CD who had undergone an intestinal resection at the Department of Surgery at the University of Vienna between September 1992 and April 2001. Criteria for inclusion were documented complete removal of all macroscopically inflamed intestinal segments during index surgery. For the current study, residing macroscopic inflammation was excluded by ileocolonoscopy, small bowel imaging and/or inspection by the surgeon. Microscopic examination had to exclude residual CD at the resection margin. Patients with ileo- or colostoma, receiving antibiotics more than 2 weeks during the post-operative follow-up were excluded from this study.

The diagnosis of CD was based on established clinical, radiological, endoscopic and histopathological criteria. Indications for surgery were symptomatic stricturing or internally penetrating disease due to CD. From each patient age at diagnosis, duration of disease, location and behaviour according to the Montreal classification,15 nicotine consume, number and type of surgery, length of the resected bowel as well as pre- and post-operative treatment were recorded.

Blood was drawn at our out-patient clinic immediately preoperatively and during post-operative follow-up at 2 weeks and at 4 ± 1, 8 ± 1 and 11 ± 1 months and thereafter upon clinical consultation. Sera were stored at −20°C until measurement, which was done within approximately 4 weeks after sample acquisition.

According to our clinical practice, patients, who consented to colonoscopy, were examined around 1 year after index operation. Severe endoscopic recurrence was defined as i3 or i4 according to the endoscopic score by Rutgeerts et al.6

Surgical recurrence was defined as the need of re-laparotomy due to CD related stricturing and/or perforating disease complications.

Long-term follow-up after the first post-operative year was obtained via telephone interview or during outpatient clinic visits, conducted on the basis of a standardised questionnaire. Data were crosschecked with patient records and the institution’s administrative database (KIS-System, Vienna, Austria). This study and the written informed consent were approved by the local Ethics committee.

ASCA ELISA assay

Anti-Saccharomyces cerevisiae antibodies.  QUANTA Lite (INOVA Diagnostics, Inc., San Diego, CA, USA) ASCA IgG and IgA enzyme-linked immunosorbent assays (ELISA) were used for determination of serum ASCA levels as described by the manufacturer’s instructions. In short, one hundred microlitres of patient’s serum at a dilution of 1:101 were added to 96-well polystyrene microwell plates adhered with partially purified and disrupted Saccharomyces cerevisiae antigen. Bound ASCAs were detected by incubation with horseradish peroxidase IgG or IgA conjugate (goat anti-human) and addition of tetramethylbenzidine chromogen. The absorbance (optical density, OD) was read at 450 nm on a photometer (Anthos HT2; Anthos Labtech Instruments, Vienna, Austria). On each plate a high and a low positive, as well as a negative control were added. ASCA reactivity was determined by the formula: sample OD/low positive OD × 25.

Anti-Saccharomyces cerevisiae antibodies cut-off values were established as described elsewhere by our group.16 The calculated cut-off values for ASCA IgG were 24.0 U, with a corresponding sensitivity of 78.0%, and a specificity of 95.0%, respectively. For ASCA IgA a cut-off value of 14.0 U was determined with a corresponding sensitivity of 62.6% and specificity of 95.0%, respectively.

The technician was blinded for the clinical outcome and the clinical investigators had no access to ASCA results at time of clinical data acquisition.

Statistical analysis

Anti-Saccharomyces cerevisiae antibodies were used as quantitative (serum levels) and qualitative variables (positive vs. negative). To compare the post-operative time course of ASCA IgA and ASCA IgG serum levels, a mixed model was calculated with patients as random factor. Independent variable was time. To clarify if disease location, post-operative treatment and the length of resected small bowel (<50 cm vs. ≥50 cm) take significant impact on the post-operative course of ASCA serum levels an analysis of variances with repeated measurements was carried out. In addition, for each time point we analysed if the influence factors, length of resection, location of disease and post-operative therapy, have a significant impact on ASCA serum levels of both isotypes at each of the follow-up measurements.

For the detection of a potential association of ASCA status and levels with the occurrence of endoscopic recurrence Chi-squared test and Mann–Whitney U-test were performed.

To investigate a relationship between preoperative ASCA serum levels of both isotypes, age, gender, onset of disease, smoking habits, disease behaviour, length of resection, post-operative therapy and the outcome surgical recurrence univariate logistic regressions were calculated. In case of associations with a P-value < 0.1, multivariate testing was conceived. If a patient underwent more than one re-operation after the index surgery, he was considered only once in the analysis. Statistical analyses were performed using sas Software, Version 9.1 (SAS Institute Inc. Cary, NC, USA) and spss Version 17 (IBM SPSS, Ehningen, Germany).

Results

At index surgery, most patients underwent intestinal resection with ileocolonic anastomosis resulting in the loss of <50 cm of intestine (Table 1).

Table 1.   Clinical characteristics of study population. Continuous data are given as median; and range. Categorical data as number (n) and percent (%)
Number of patients (f/m)60 (32/28)
 Age (years)33 (20–70)
 Age at onset (years)24 (7–58)
Disease location (number; n)
 Ileal (L1)18 (30)
 Colonic (L2)2 (3)
 Ileocolonic (L3)40 (67)
Indication of index surgery (n)
 Stricturing35 (58)
 Penetrating25 (42)
Type of surgery (n)
 Intestinal resection with ileocolonic anastomosis54 (90)
 Small bowel resection only6 (10)
Length of resected small bowel (n)
 <50 cm38 (63)
 >50 cm15 (25)
 Not documented7 (12)
Preoperative medical treatment (n)
 No treatment18 (30)
 5-ASA13 (22)
 Prednisolone10 (17)
 Azathioprine15 (25)
 Azathioprine and Prednisolone3 (5)
 Infliximab1 (2)

Ileocoecal resection was performed in 46/60 (77%) patients, right-sided hemicolectomy in 5/60 (8%), subtotal colectomy in 3/60 (5%) and small bowel resection only in 6/60 patients (10%). Preoperatively, more than half of the patient population was on no treatment or 5-aminosalicylates. After index operation azathioprine was administered in 50% of patients (Table 2).

Table 2.   Post-operative treatment
Medical treatment (n)Study population
No treatment22
5-ASA8
Azathioprine30

Serum samples for ASCA measurement were obtained preoperatively, at week 2, month 4 ± 1, month 8 ± 1 and month 11 ± 1 on 60 (100%), 49 (81.7%), 48 (80%), 45 (75%) and 43 (72%) patients respectively. In addition, ASCA serum levels were available after a median follow-up of 29 months (11–102 months) in 25 patients.

Prevalence of ASCA

Judging from preoperative serology, 44/60 (73%) patients were positive for ASCA IgG and 45/60 (75%) were positive for ASCA IgA. Thirty-seven of sixty (62%) patients were concordantly positive, 8/60 (13%) patients were concordantly negative, whereas 7/60 (12%) patients were positive only for ASCA IgG and 8/60 (13%) only for ASCA IgA.

Changes of ASCA serum levels during post-operative follow-up

During follow-up, ASCA IgA serum levels converted from negative to positive in four patients. One patient turned negative. ASCA IgG serum levels converted from negative to positive in one patient and from positive to negative in two. Two patients became seronegative transiently only to return to their baseline levels during long-term follow-up. None of these patients were on a pre- or post-operative CD – related treatment. No effect of time on serum ASCA levels of both isotypes during the post-operative course could be discerned (ASCA IgA: preoperative median 38.6 U/mL [range 4.1–3686.0 U/mL], week 2 39.4 U/mL [4.3–2291.2], month 4 44.8 U/mL [4.3–4416.0], month 8 33.4 U/mL [4.3–2524.8], 56.7 U/mL [2.8–2421.8], last follow-up 44.3 U/mL [4.0–1689.6], = 0.13; ASCA IgG: preoperative median 42.7 U/mL [3.4–655.2], week 2 48,0 U/mL [3.9–651.2], month 4 52.1 U/mL [3.3–584.8], month 8 43.0 U/mL [2.4–569.6], month 12 52.0 U/mL [2.8–503.2], last follow-up 35.0 U/mL [10.4–232.0] = 0.47; Figure 1a,b) which also applied to patient subgroups according to disease location, post-operative treatment and length of resected small bowel (data not shown).

Figure 1.

 (a) Post-operative course of ASCA IgA, Values are given as log U/mL, dashed black line gives cut-off (14.0 U/mL) for ASCA seropositivity. Last follow up was obtained after a median of 29 months (11–102 months). Log scale was used to reduce data overlap. (b) Post-operative course of ASCA IgG, values are given as log U/mL, dashed black line gives cut-off (24 U/mL) for ASCA seropositivity. Last follow-up was obtained after a median of 29 months (11–102 months). Log scale was used to reduce data overlap.

Endoscopic and surgical recurrence

In 18/32 (56%) patients, who had agreed upon ileocolonoscopy around 1 year post-operatively, mucosal lesions were scored as severe (Rutgeerts score i3: 2/32; 6%, i4: 16/32; 50%). In another 14 patients no (i0: 6/32; 19%), mild lesions (i1: 4/32; 12.5%) or intermediate grade lesions (i2: 4/32; 12.5%) were detected. There was no difference in the course of ASCA serum levels in patients with and without severe endoscopic recurrence (ASCA IgG = 0.35, ASCA IgA = 0.24) (Table 3). No surgical recurrence occurred within the first year after index surgery.

Table 3.   ASCA IgG and ASCA IgA quantitatively and qualitatively vs. endoscopic recurrence. ASCA status and serum levels are preoperative levels. Recurrence was scored according to Rutgeerts et al. Continuous variables are given as median and range, categorical data as number
Rutgeerts ScoreASCA IgG status (pos: n, %)ASCA IgA status (pos: n, %)ASCA IgG quantitative (U/mL)ASCA IgA quantitative (U/mL)
  1. ASCA, anti-Saccharomyces cerevisiae antibodies.

Rutgeerts i0/i1/i2 (= 14)11 (78.5)12 (85.7)42.6 (6.5–552.8)40.0 (4.5–3686.0)
Rutgeerts i3 or i4 (= 18)11 (61.1)12 (66.6)37.1 (6.0–243.8)24.1 (5.0–1984.0)

Long-term outcome

Forty of the initially included patients were followed up for a median duration of 106 months (range 91–194 months). Ten subjects (25%) underwent surgery (five patients were operated once, four had two and one patient underwent three operations) after a median follow-up of 70.5 months with an indication of stricturing disease in eight (80%) patients and penetrating disease in one (10%) patient for first re-operation. In one patient (10%) both stricturing and penetrating disease were present. As serum ASCA levels were stable after intestinal resection, we decided to associate preoperative levels and status with surgical recurrence. Neither ASCA status (positive vs. negative for IgA, IgG or both) nor ASCA serum levels were predictive for the risk of CD-related re-operation (ASCA IgG = 0.78, ASCA IgA = 0.66) (Figure 2a,b). Only smoking was identified as predictive for surgical recurrence in uni-variate logistic regression (OR 17, CI 95%, 1.75–164.97, = 0.02).

Figure 2.

 (a) ASCA IgG baseline serum levels, surgical recurrence vs. recurrence free cohort, Values are given as log U/mL. Follow-up was obtained for a median period of 106 months (range 91–194 months), dashed black line gives cut-off (24 U/mL). (b) ASCA IgA baseline serum levels, surgical recurrence vs. recurrence free cohort, values are given as log U/mL. Follow-up was obtained for a median period of 106 months (range 91–194 months), dashed black line gives cut off (14.0 U/mL).

Discussion

In this study, we were able to demonstrate the stability of ASCA serum levels in patients with CD after intestinal surgery. None of the initially ASCA positive patients became negative in both isotypes. In our series neither ASCA status nor serum levels were predictive of surgical recurrence after a median follow-up period of 9 years.

Data on the stability of ASCA in CD after surgical resection are scarce. Ruemmele et al. noticed ASCA serum levels being negative after surgery in 7 of 11 patients in a paediatric cohort.17 This study was limited by its small number of patients who were resected upon and the lack of preoperative measurements. Our data indicate, for the first time, the persistence of ASCA after complete surgical removal of macroscopically inflamed intestinal segments in CD patients. This suggests that the production of antibodies against epitopes of Saccharomyces cerevisiae is not dependent on active inflammation and a consecutive breach in the mucosal barrier, but may rather reflect the disposition to disease. This notion is in line with prior observations by our group and others, that ASCA serum levels are generally stable hallmarks of disease. Israeli et al. were able to demonstrate the seroconversion for ASCA occurring long before the first CD specific symptoms develop.18 This indicates a loss of tolerance for Saccharomyces cerevisiae antigens or cross reactives, that occurs independently of and prior to the actual breach of the mucosal barrier by inflammation. ASCAs not only appear before inflammation of the mucosa is clinically detectable, they also remain largely uninfluenced by anti-inflammatory therapy. Specifically, Landers et al. were able to demonstrate the stability of ASCA and other biomarkers during Infliximab treatment.19 Teml et al. found similar results in patients under 5- aminosalicylates and corticosteroids.16 In line with this, Müller et al. did not detect significant changes in ASCA levels in a cohort of CD patients observed over a period of 3 years.20 This group, however, only included patients on medical therapy and did not observe patients post-operatively. Interestingly, Vermeire et al. could not associate ASCA serum levels with the degree of intestinal permeability in patients with CD as measured by 51Cr-EDTA intestinal permeation,21 which supports the hypothesis that loss of tolerance towards Saccharomyces cerevisiae in CD is independent of a breach of intestinal integrity. Concordantly, in our cohort there was neither a difference in ASCA seropositivity nor ASCA serum levels in patients experiencing mild vs. severe post-operative endoscopic recurrence. This contrasts findings from coeliac disease patients, where ASCA serum levels drop to normal after strict gluten free diet.22, 23 Thus, whereas in coeliac disease causal treatment results in normalisation of ASCA serum levels, medical and surgical therapy in CD remains associated with unchanged levels of serum ASCA.

In their cornerstone publication, Forcione et al. were able to identify ASCA IgA as risk factor for early surgery in a case control study.24 ASCA positivity was associated with early surgery with an odds ratio of 8.5 (95% CI 2.0–75.9; = 0.0013). However, as their study was based on post-operative measurements of ASCA only, the authors acknowledge the possibility of ASCA levels decreasing after resection, thereby confounding their effect size estimation. Recently, in a paediatric cohort, Amre et al. detected a shortened time between diagnosis and occurrence of complications in paediatric CD patients considered positive for ASCA vs. those being negative.25 The same group also attempted to evaluate ASCA as biomarker for the natural course of CD. ASCA predicted a relapsing disease behaviour, but were not useful for the prediction of other clinical endpoints, such as occurrence of fistula or need for intestinal surgery.26 Similar results were also obtained by Russell et al.9 Recently, Schoepfer et al. were able to associate ASCA with fistulising disease behaviour and small bowel surgery in an adult cohort of CD patients.8

In our study, we were able to demonstrate the stability of ASCA serum levels post-operatively, regardless of the length of the resected small bowel and post-operative treatment with azathioprine. However, after a median clinical follow-up of 9 years in forty of the initially included patients, we could not detect an association between ASCA serum levels and the likelihood of surgical recurrence in our cohort. Thus, by our findings ASCAs do not appear to be helpful in the identification of patients with CD patients who are at increased risk of a severe post-operative course and, therefore, in the selection of those patients who would benefit most from a progressive post-operative treatment approach. In the context of the existing literature, ASCA serum levels could help in the risk stratification of patients with a new diagnosis of CD and an increased risk for first intestinal resection, whereas their value for post-operative management is not supported by our results.

In good accordance with the existing literature, smoking was strongly associated with surgical recurrence in our cohort,27, 28 again emphasising the need to encourage tobacco cessation in high risk CD patients.

Among the strengths of our study are its longitudinal prospective design, its uniquely long follow-up period and the fact, that only CD patients with ‘curative’ intestinal resections were included. Remaining macroscopic inflammation was excluded by ileocolonoscopy, small bowel imaging and/or inspection by the surgeon, whereas resection margins had to be free of residual inflammation by microscopic examination. However, we acknowledge the possibility of microscopic inflammation in remaining intestinal segments distant from the site of surgical resection. Inflammation can be detected by wireless capsule endoscopy within 6 months after surgery, as shown by A. Bourreille et al.29 As to whether the detected lesions did reflect recurrence of CD or persistent inflammation after surgery was not addressed. Thus, the post-operative model by itself is only an approximation of the ideal inflammation-free condition, but seemingly the closest available. Limitations lie in the thus limited sample size and the fact that long-term follow-up beyond 1 year was only available for two-thirds of patients.

It is concluded that ASCA serum levels of both IgG and IgA isotypes are not influenced by intestinal resection during post-operative long-term follow-up, which indicates a breach of oral tolerance towards Saccharomyces cerevisiae epitopes in CD, likely to be driven by an intrinsic patho-mechanism remaining operative in patients after complete resection of macroscopically inflamed tissue. However, other biomarkers than serum ASCA levels need to be revealed to guide post-operative management in CD.

Acknowledgement

Declaration of personal interests: A. Eser has received speaker’s fees from AESCA and Abbott, and is holding a research grant from the Medizinsch-wissenschaftlicher Fonds des Buergermeisters der Stadt Wien, Vienna, Austria. W. Reinisch has served as a speaker, a consultant and/or an advisory board member for Abbott Laboratories, Aesca, AstraZeneca, Biogen IDEC, BMS, Cellerix, Chemocentryx, Centocor, Danone Austria, Elan, Ferring, Genentech, Lipid Therapeutics, Millenium, MSD, Novartis, Ocera, Otsuka, PDL, Pharmacosmos, Pfizer, Procter & Gamle, Prometheus, Schering-Plough, Shire, Therakos, UCB, Vifor, Yakult Austria and 4SC. Declaration of funding interests: None.

Ancillary