Search

Contains fulltext : 291051.pdf (Publisher’s version ) (Open Access) Conventional esophagogastroduodenoscopy (cEGD) is currently the gold standard endoscopic procedure for diagnosis and surveillance of Barrett's esophagus (BE). This procedure is however less suitable for widespread screening because of its invasiveness and costs. An alternative endoscopic procedure is unsedated transnasal endoscopy (uTNE). We performed a systematic review and meta-analysis to evaluate the diagnostic accuracy, patient tolerability, technical success rate, and safety of uTNE compared with cEGD for detecting BE and related neoplasia. PubMed, EMBASE, and Cochrane Library were searched for studies that reported the diagnostic accuracy of uTNE compared with cEGD for detecting BE and related neoplasia. Eight prospective studies were included, in which 623 patients underwent both uTNE and cEGD. Pooled sensitivity and specificity of uTNE for detecting columnar epithelium were 98% (95% CI 83-100%) and 99% (95% CI 82-100%), respectively. Pooled sensitivity and specificity of uTNE for detecting intestinal metaplasia in biopsies were 89% (95% CI 78-95%) and 93% (95% CI 71-98%), respectively. In three of the six studies that reported patient tolerability, a higher patient tolerability of uTNE compared with cEGD was reported. The technical success rate of uTNE ranged from 89% to 100% and no (serious) adverse events were reported. This systematic review and meta-analysis provides evidence that uTNE is an accurate, safe, and well-tolerated procedure for the detection of columnar epithelium and can be considered as screening modality for BE.

Anastomotic leakage (AL) is a common and severe complication after esophagectomy. It is largely unknown which primary treatments are most effective for which type of leak. We aimed to investigate the effectiveness of different primary treatments of AL. Methods International retrospective cohort study, in which patients with AL after esophagectomy with gastric tube reconstruction were included in the period 2011-2019. Detailed data regarding case mix, resection, leakage characteristics (e.g. organ failure, leak circumference, contamination, drains present) and leakage treatment (e.g. hours from diagnosis to treatment, primary and secondary treatment modalities) were collected. Primary outcome was 90-day mortality and secondary outcomes included length of stay and leak healing time. Different clinically relevant leakage groups have been defined. Efficacy of different treatment strategies adjusted for leakage severity will be analyzed in these clinical groups. The study protocol is accessible at www.tentaclestudy.com. Results Detailed data of 1451 patients with AL was collected from 71 centers in 20 countries. Data accuracy was 96.5%. Preliminary results showed that the overall 90-day mortality was 11.6%. The analysis of TENTACLE—Esophagus data is currently being performed and efficacy of different leakage treatment strategies is being assessed. The efficacy of initial leakage treatment strategies will be ready to be presented at the ISDE meeting. Conclusion This is the largest study on effectiveness of AL treatments. The final results of initial leak treatments, which will be available for presentation at the ISDE meeting, could provide an evidence-based basis that can be used by clinicians to determine the preferred primary treatment strategy in patients with a given type of anastomotic leakage.

Anastomotic leakage (AL) is a common and potentially life-threatening complication after esophagectomy. In patients with AL it is largely unknown which patient parameters and leakage characteristics are associated with leak severity and mortality. We aimed to identify prognostic factors for mortality in patients with anastomotic leakage after esophagectomy. Methods The TENTACLE—Esophagus study is an international retrospective cohort study, in which 1451 patients with AL after esophagectomy between 2011 and 2019 were included in 71 centers from 20 countries. Potential prognostic factors were selected from literature and a hypothetical association with mortality. Confounders of (individual) prognostic factors were identified using a directed acyclic graph approach to minimize bias. Primary outcome was 90-day mortality. Logistic regression analysis was performed to estimate crude and adjusted odds ratios (AOR) and 95% confidence intervals (95%CI). The study protocol is accessible at www.tentaclestudy.com. Results Overall 90-day mortality rate was 11.6%. Leakage characteristics with the largest prognostic effect on mortality were gastric conduit ischemia/necrosis (AOR 2.23, 95%CI 1.43-3.49), defect circumference ≥ 25% (AOR 2.10, 95%CI 1.32-3.36) and intrathoracic fluid collections (drained AOR 1.98, 95%CI 1.05-3.75; undrained AOR 2.43, 95%CI 1.57-3.75). Patient parameters with the largest prognostic effect were ASA-score ≥ 3 (AOR 4.18, 95%CI 1.67-10.51), ECOG-score ≥ 2 (AOR 2.83, 95%CI 1.56-5.14) and respiratory failure (AOR 3.89, 95%CI 2.67-5.66), hemodynamic failure (AOR 3.09, 95%CI 1.96-4.88) or renal failure (AOR 4.08, 95%CI 2.20-7.59) at time of AL diagnosis. Conclusion Defect circumference, intrathoracic fluid collections, gastric conduit condition and several patient parameters were identified as prognostic factors for mortality in patients with AL. Adjusting for these prognostic factors may reduce confounding bias in future studies assessing efficacy of AL treatments. The identified prognostic factors contribute to the understanding of the severity of anastomotic leakage after esophagectomy and may be used to recognize the severity of an anastomotic leak in individual patients.

Geographical variation in cancer care has been observed between and within countries. Variation in treatment may occur at any point along the cancer care continuum attributing to potentially avoidable disparities in patient outcomes such as survival. Other studies have shown a large variation in the proportion of patients undergoing treatment with curative intent for esophageal(EC) and gastric cancer(GC). Due to these studies and thus an increased awareness, we hypothesize that variation has decreased over time. Methods Patients with potentially curable EC (n = 10,115) or GC (n = 3,983) diagnosed between 2012-2017 were selected from the Netherlands Cancer Registry. Multilevel multivariable logistic regression was used to examine the differences in the probability of treatment with curative intent between hospitals of diagnosis, comparing 2012-2014 with 2015-2017. We compared relative survival (RS) and relative excess risk of death (RER) between hospitals with different probabilities of treatment with curative intent. Results The range of proportions of patients undergoing treatment with curative intent per hospital of diagnosis was 45-95% in 2012-2014 and 54-89% in 2015-2017 for EC, and 52-100% and 45-100% for GC. The adjusted variation declined for EC with Odds Ratios (ORs) ranging from 0.50 to 1.72 between hospitals in the first period to 0.70–1.44 in the second period (p Conclusion Although substantial variation between hospitals of diagnosis in the probability in receiving treatment with curative intent still exists for both malignancies, it has decreased for EC, but did not change for GC. A low probability of receiving curative treatment was associated with a worse survival. Further research is needed to elucidate the factors associated with the observed variability.

Contains fulltext : 225282.pdf (Publisher’s version ) (Open Access) Barrett's esophagus (BE) is a metaplastic condition of the distal esophagus, resulting from longstanding gastroesophageal reflux disease (GERD). BE predisposes for the highly malignant esophageal adenocarcinoma (EAC). Both BE and EAC have the highest frequencies in white males. Only a subset of patients with GERD develop BE, while

Background: A low vitamin D status has been put forward as a potential risk factor for the development of inflammatory bowel disease (IBD). This study investigated the association between prediagnostic circulating vitamin D concentrations and dietary intakes of vitamin D, and the risk of Crohn’s disease (CD) and ulcerative colitis (UC).Methods: Among 359,728 participants of the European Prospective Investigation into Cancer and Nutrition cohort, individuals who developed CD or UC after enrollment were identified. Each case was matched with2 controls by center, gender, age, date of recruitment, and follow-up time. At cohort entry, blood samples were collected and dietary vitamin D intakes were obtained from validated food frequency questionnaires. Serum 25-hydroxyvitamin D levels were measured using liquid chromatography-tandem mass spectrometry. Conditional logistic regression was performed to determine the odds of CD and UC.Results: Seventy-two participants developed CD and 169 participants developed UC after a median follow-up of 4.7 and 4.1 years, respectively. Compared with the lowest quartile, no associations with the 3 higher quartiles of vitamin D concentrations were observed for CD (p trend = 0.34) or UC (p trend = 0.66). Similarly, no associations were detected when serum vitamin D levels were analyzed as a continuous variable. Dietary vitamin D intakes were not associated with CD (p trend = 0.39) or UC (p trend = 0.83).Conclusions: Vitamin D status was not associated with the development of CD or UC. This does not suggest a major role for vitamin D deficiency in the etiology of IBD, although larger studies are needed to confirm these findings.

Aim The aim of this study is to investigate which factors contribute to anastomotic leakage severity and to compose an evidence based anastomotic leakage severity score. Secondly, we aim to investigate which anastomotic leakage characteristics are associated with success of different anastomotic leakage treatments and to compare the effectiveness of different initial anastomotic leakage treatments for anastomotic leakage classified according to severity and leakage characteristics. Background Anastomotic leakage occurs in 0%-30% after esophagectomy for cancer. It is a severe complication with mortality rates approximately ranging from 2%-12%. In addition, it is associated with a prolonged ICU treatment and hospital stay. Anastomotic leakage severity is currently graded according to how it is treated (grade I: conservative treatment, grade II: endoscopic/radiologic intervention and grade III: surgical intervention). However, this scoring system cannot be used to guide decision making when anastomotic leakage is diagnosed in a clinical setting. Factors that may influence the severity of the anastomotic leakage are (amongst others) location of the anastomosis, estimated surface of the defect, estimated circumference of the defect, extent of contamination, degree of sepsis and time from diagnosis until therapy. However, little is known about to what extent these and other factors contribute to anastomotic leakage severity. In addition, there is a paucity of data on what leakage characteristics dictate the success of a specific treatment. Methods We will perform an international multicenter retrospective cohort study. All adult patients with anastomotic leakage according to the ECCG definition after esophagectomy and gastric conduit reconstruction for esophageal cancer are suitable for inclusion. We aim to include at least 1000 patients. The primary outcome parameter is 90-day mortality. Secondary outcome parameters are in-hospital mortality, 30-day mortality, 180-day mortality, comprehensive complications index, total number of reinterventions, hospital and ICU length of stay and hospital related costs. Results Data collection will take place from April 2019 until December 2019. Study results are expected in June 2020. Conclusion We hypothesize that we will be able to identify factors that influence the severity of the anastomotic leakage and to compose a leakage severity score. Furthermore, we expect to identify leakage characteristics dictating the success of a specific treatment.

Esophageal and gastric cancer is associated with a poor prognosis since many patients develop recurrent disease. Treatment requires specific expertise and a structured multidisciplinary approach. In the Netherlands, this type of expertise is mainly found at the University Medical Centers (UMCs) and a few specialized nonacademic centers. Aim of this study is to implement a national infrastructure for research to gain more insight in the etiology and prognosis of esophageal and gastric cancer and to evaluate and improve the response on (neoadjuvant) treatment. Clinical data are collected in a prospective database, which is linked to the patients' biomaterial. The collection and storage of biomaterial is performed according to standard operating procedures in all participating UMCs as established within the Parelsnoer Institute. The collected biomaterial consists of tumor biopsies, blood samples, samples of malignant and healthy tissue of the resected specimen and biopsies of recurrence. The collected material is stored in the local biobanks and is encoded to respect the privacy of the donors. After approval of the study was obtained from the Institutional Review Board, the first patient was included in October 2014. The target aim is to include 300 patients annually. In conclusion, the eight UMCs of the Netherlands collaborated to establish a nationwide database of clinical information and biomaterial of patients with esophageal and gastric cancer. Due to the national coverage, a high number of patients are expected to be included. This will provide opportunity for future studies to gain more insight in the etiology, treatment and prognosis of esophageal and gastric cancer.

Background Inflammatory bowel disease (IBD) is associated with high costs to society. Few data on the impact of IBD on work disability and potential predictive factors are available. Aim To assess the prevalence of and predictive factors for work disability in Crohn's disease (CD) and ulcerative colitis (UC). Methods A web-based questionnaire was sent out in seven university hospitals and seven general hospitals in the Netherlands. Initially, 3050 adult IBD patients were included in this prospective, nationwide cohort study, whereof 2629 patients were within the working-age (18–64 years). We used the baseline questionnaire to assess the prevalence rates of work disability in CD and UC patients within working-age. Prevalence rates were compared with the Dutch background population using age- and sex-matched data obtained from Statistics Netherlands. Multivariable logistic regression analyses were performed to identify independent demographic- and disease-specific risk factors for work disability. Results In CD, 18.3% of patients was fully disabled and 8.8% partially disabled, compared to 9.5% and 5.4% in UC patients (p Conclusion We report high work disability rates in a large sample of IBD patients in the Netherlands. CD patients suffer more frequently from work disability than UC patients. A combination of demographic and disease-related factors is predictive of work disability.

Background Fecal calprotectin can be used as a noninvasive tool to assess inflammation in patients with inflammatory bowel disease (IBD). However, the diagnostic accuracy of calprotectin is modest, and therefore additional markers are needed. We compared the efficacy of fecal hemoglobin and calprotectin as markers for endoscopic inflammation in patients with IBD. Methods Consecutive patients with Crohn's disease or ulcerative colitis scheduled for surveillance colonoscopy collected a stool sample before bowel preparation. Experienced endoscopists assessed the presence of inflammation in each colonic segment. Fecal calprotectin and hemoglobin were analyzed with an enzyme-linked immunosorbent assay. Receiver operator characteristic statistics were used to determine cutoff values for calprotectin and hemoglobin. Results A total of 176 surveillance colonoscopies were performed in 164 patients, of which 83 patients had Crohn's disease, 74 had ulcerative colitis, and 7 IBD-unclassified. Median (interquartile range) calprotectin and hemoglobin concentrations were 137 mg/kg (interquartile range, 33-494) and 0.51 μg/g (interquartile range, 0.18-8.50), respectively. For calprotectin, a cutoff value of 140 mg/kg predicted endoscopic inflammation with 86% sensitivity, 72% specificity, 64% positive predictive value, 90% negative predictive value, and an area under the curve of 0.87. For hemoglobin, a cutoff value of 1.51 μg/g indicated endoscopic inflammation with 74% sensitivity, 84% specificity, 72% positive predictive value, 84% negative predictive value, and an area under the curve of 0.81. Combining both tests did not increase the predictive accuracy substantially compared with calprotectin or hemoglobin alone (area under the curve, 0.88). Conclusions Fecal hemoglobin can identify patients with IBD with active inflammation with a predictive accuracy similar to calprotectin.

Background Neoadjuvant chemoradiotherapy (nCRT) induces a pathologically complete response in approximately 30% of patients with oesophageal cancer. To explore the possibility of safe postponement of surgery, accurate clinical response evaluations are needed to exclude residual disease. The present study aims to assess the value of F-18-FDG-PET/CT for the detection of residual tumour (> 10% tumour cells = TRG3–4 vs. no vital cells = TRG1) or metastases after nCRT. Methods FDG-PET/CT at baseline and 12 weeks after nCRT was performed according to the European Association of Nuclear Medicine guidelines 1.0 (2.3MBq/kg F-18-FDG; scanning 60 ± 5min.) and the protocol of the preSANO study. Qualitative analysis included sensitive reading of presence of residual tumour and/or metastases. A lesion was considered FDG-positive, when any uptake in the lesion itself was above the adjacent oesophageal background uptake. Quantitatively, SUV/lean body mass (SUL) measurements at tumour, lymph nodes, oesophagus, liver and bloodpool were recorded and compared with pathology (resection specimen: gold standard). Results Some 129 of 207 patients with FDG-avid tumours at baseline proceeded to FDG-PET/CT at around 12 weeks after nCRT just before surgery. Forty-one of 129 patients had TRG3–4, of whom 6 were missed on FDG-PET/CT (15% false negative) with SULmax 2.07 ± 0.25, SUL-ratio tumour/oesophagus (SULR) 1.35 ± 0.14. Sensitivity for TRG2–3-4 vs. TRG1 was 57/71 (80%). SULmax and SULR of FDG-positives were 3.76 ± 1.33 and 1.82 ± 0.69 respectively, compared to SULmax 2.21 ± 0.42 and SULR 1.31 ± 0.22 in FDG-negatives. Distant metastases were detected in 18 of 190 (10%) patients. Of all patients with postponed surgery, 12 had ≥ 1 additional FDG-PET/CT during follow-up (25–49.7 weeks after nCRT). Eventually, 4 patients underwent surgery. Three of 4 had increased FDG-signal and TRG3–4; 1 patient had decreased FDG-signal and no tumour left (TRG1). Conclusion FDG-PET/CT at around 12 weeks after nCRT misses TRG3–4 tumours in 15% and detects residual TRG2–3-4 in 80%. Furthermore, PET-CT detects distant metastases in 10% of patients after nCRT. These data indicate that serial FDG-PET may become valuable in an active surveillance approach. Disclosure All authors have declared no conflicts of interest.

Barrett's esophagus (BE) is a metaplastic condition of the distal esophagus that occurs because of chronic gastroesophageal reflux. Previous studies have identified BE-specific microRNAs (miRNAs) in comparison with normal squamous epithelium (SQ). We hypothesized that BE-specific miRNAs could be induced in esophageal SQ cells by exposure to acid and/or bile salts. We aimed to determine whether BE-specific miRNAs are upregulated in an esophageal SQ cell line (Het-1A) in an environment with acid and/or bile salts and whether this is nuclear factor-κB (NF-κB) dependent. Acid and/or bile salt incubations were performed in Het-1A cells. Experiments were performed with or without inhibiting the NF-κB pathway. Quantitative reverse transcriptase polymerase chain reaction was performed to determine expression of miRNA-143, -145, -192, -194, cyclo-oxygenase-2 (COX2), mucin 2 (MUC2), and sex determining region Y-box 9. For validation, we determined levels of these miRNAs in biopsies from patients with reflux esophagitis and normal SQ. Significantly increased expression levels of miRNA-143 (2.7-fold), -145 (2.6-fold), -192 (2.0-fold), -194 (2.2-fold), COX2, MUC2, and sex determining region Y-box 9 were found upon acidic bile salt incubation, but not upon acid or bile salt alone. NF-κB pathway inhibition significantly decreased miRNA-143, -192, -194, COX2, and MUC2 expression. Additionally, miRNA-143, -145 and -194 expression was increased in reflux esophagitis biopsies compared with normal SQ, but no changes were found in miRNA-192 expression. Our findings suggest that upregulation of BE-specific miRNAs by acidic bile may be an early event in the transition of SQ to BE and that their expression is partly regulated by the NF-κB pathway.

In this study, the relation between fruit and vegetable consumption and mortality was investigated within the European Prospective Investigation Into Cancer and Nutrition. Survival analyses were performed, including 451,151 participants from 10 European countries, recruited between 1992 and 2000 and followed until 2010. Hazard ratios, rate advancement periods, and preventable proportions to respectively compare risk of death between quartiles of consumption, to estimate the period by which the risk of death was postponed among high consumers, and to estimate proportions of deaths that could be prevented if all participants would shift their consumption 1 quartile upward. Consumption of fruits and vegetables was inversely associated with all-cause mortality (for the highest quartile, hazard ratio = 0.90, 95% confidence interval (CI): 0.86, 0.94), with a rate advancement period of 1.12 years (95% CI: 0.70, 1.54), and with a preventable proportion of 2.95%. This association was driven mainly by cardiovascular disease mortality (for the highest quartile, hazard ratio = 0.85, 95% CI: 0.77, 0.93). Stronger inverse associations were observed for participants with high alcohol consumption or high body mass index and suggested in smokers. Inverse associations were stronger for raw than for cooked vegetable consumption. These results support the evidence that fruit and vegetable consumption is associated with a lower risk of death.

Various treatments are available for the palliation of esophageal cancer, but the optimal therapeutic approach is unclear. This study aimed to assess the palliative treatment modalities used in patients with inoperable esophageal cancer and to identify factors associated with treatment decisions. A population-based, retrospective cohort study was conducted using data from the nationwide Netherlands Cancer Registry and medical records of seven participating hospitals. Patients diagnosed with stage III-IV inoperable esophageal or gastric cardia cancer in the central part of the Netherlands between 2001 and 2010 were included. Logistic regression analyses were performed to identify determinants of treatment choices. In total, 736 patients were initially treated with best supportive care (21%), stent placement (19%), chemotherapy (18%), external beam radiotherapy (EBRT) (16%), brachytherapy (6%), a combination of EBRT and brachytherapy (6%), a combination of chemotherapy and EBRT (5%) or another treatment (9%). The palliative approach varied for disease stage (P

Background: The increased risk of colorectal cancer (CRC) in patients with inflammatory bowel disease (IBD) is well established. The incidence of IBD-related CRC however, differs markedly between cohorts from referral centers and population-based studies. In the present study we aimed to identify characteristics potentially explaining these differences in two cohorts of patients with IBD-related CRC. Methods: PALGA, a nationwide pathology network and registry in The Netherlands, was used to search for patients with IBD-associated CRC between 1990 and 2006. Patients from 7 referral hospitals and 78 general hospitals were included. Demographic and disease specific parameters were collected retrospectively using patient charts. Results: A total of 281 patients with IBD-associated CRC were identified. Patients from referral hospitals had a lower median age at IBD diagnosis (26 years vs. 28 years (p = 0.02)), while having more IBD-relapses before CRC diagnosis than patients from general hospitals (3.8 vs. 1.5 (p < 0.01)). In patients from referral hospitals, CRC was diagnosed at a younger age (47 years vs. 51 years (p = 0.01)). However, the median interval between IBD diagnosis and diagnosis of CRC was similar in both cohorts (19 years in referral hospitals vs. 17 years in general hospitals (p = 0.13)). Conclusions: IBD patients diagnosed with CRC treated in referral hospitals in The Netherlands are younger at both the diagnosis of IBD and CRC than IBD patients with CRC treated in general hospitals. Although patients from referral centers appeared to have a more severe course of IBD, the interval between IBD and CRC diagnosis was similar. * Abbreviations: RH : Referral hospital GH : General hospital

SUMMARY Barrett's esophagus (BE) is a premalignant condition with an increased risk of developing esophageal adenocarcinoma (EAC). Risk factors for EAC overlap with those for esophageal squamous cell carcinoma (ESCC), but ESCC is surprisingly rare in BE. We report two cases of ESCC directly surrounded by BE. Both patients had a previous medical history of cancers, i.e., head and neck squamous cell carcinomas, and were using alcohol and smoking tobacco. Using immunohistochemistry for p63, CK5, CK7, and CDX2, it was confirmed that these carcinomas were pure squamous cell carcinomas, and not EACs or esophageal adenosquamous carcinomas arising from BE. Using TP53 mutation and loss of heterozygosity analysis, we established that the ESCCs in BE were not metastases of the previously diagnosed head and neck squamous cell carcinomas but de novo primary ESCCs. This study shows the strength of molecular analysis as an adjunct to the histopathologic diagnosis for distinguishing between metastases of prior cancers and primary cancers. Furthermore, these cases imply that presence of BE is not protective with regards to developing ESCC in the lower one third of the esophagus. We suggest that their ESCCs arose from islets of squamous epithelium in BE.

Background Anastomotic leakage affects up to 30% of patients after esophagectomy each year and leads to considerable morbidity and mortality. The aim of this study was to determine which treatment for anastomotic leakage after esophagectomy has the best clinical outcome, based on currently available literature. Methods A systematic literature search was performed in Medline, Embase and Web of Science until April 2017. All studies reporting on the treatment of anastomotic leakage following esophagectomy with gastric tube reconstruction for esophageal or cardia cancer were included. The primary outcome parameter was postoperative mortality. Methodological quality was assessed by the Newcastle-Ottawa Quality Assessment Scale. Results Nineteen retrospective cohort studies including 273 patients were identified. Methodological quality of all studies was poor to moderate. Regarding intrathoracic anastomotic leakages, mortality rates in the conservative, endoscopic stent, endoscopic drainage, endoscopic vacuum assisted closure system and surgery treatment group were 14%, 8%, 8%, 0%, and 50%, respectively. Regarding cervical anastomotic leakages, mortality rates in the conservative, endoscopic stent and endoscopic dilatation group were 8%, 29%, and 0%, respectively. Conclusion Due to small cohorts, heterogeneity between studies, and lack of data regarding leakage characteristics, no evidence supporting one treatment for anastomotic leakage after esophagectomy was found. A severity score based on leakage characteristics instead of treatment given is essential for determining the optimal treatment of anastomotic leakage. A prospective registration study could provide answers to issues as which leakage characteristics determine its severity and which treatment options have the best outcomes for a given anastomotic leakage severity. Disclosure All authors have declared no conflicts of interest.

Patients with longstanding achalasia have an increased risk of developing esophageal cancer. Surveillance is hampered by chronic stasis. We investigated whether aberrant expressions of tumor suppressor gene p53 and proliferation marker ki67 are early predictors for progression to malignancy. In 399 achalasia patients, 4% died of esophageal cancer despite surveillance. We performed a cohort study, using surveillance biopsies from 18 patients (11 carcinoma, one high-grade dysplasia [HGD], and six low-grade dysplasia [LGD]) and 10 controls (achalasia patients without cancer or dysplasia development). One hundred sixty-four biopsies were re-evaluated and studied for p53 and ki67 expression using immunohistochemistry. Eighty-two percent of patients with cancer/HGD showed p53 overexpression in surveillance biopsies at a mean of 6 (1-11) years prior to cancer development. In 67% of patients with LGD and only in 10% of the controls p53 overexpression was present. The proportion of samples with p53 overexpression increased with increasing grades of dysplasia. We found no difference for ki67 overexpression. p53 overexpression may identify achalasia patients at increased risk of developing esophageal carcinoma. Further study is needed to determine if patients with p53 overexpression would benefit from intensive surveillance to detect esophageal neoplasia at a potential curable stage.

Osteopenia and osteoporosis are frequently encountered in patients with Inflammatory Bowel Disease (IBD). Our aims were to evaluate the actual practice of screening for low bone mineral density (BMD) by dual energy X-ray absorptiometry (DEXA), to determine the prevalence of low BMD and to investigate the risk factors associated with a low BMD in the IBD population of a regional Dutch hospital.A retrospective chart review was performed in 474 patients (259 with ulcerative colitis, 210 with Crohn's disease and 5 with indeterminate colitis). DEXA results and potential predictive factors of low BMD were documented. Predictive factors of low BMD were assessed by logistic regression.DEXA was performed in 168 IBD patients (35.4%). A low BMD (T-score-1) was present in 64.3%. Osteoporosis (T-score-2.5) was found in 23.8%. Low BMI, older age at the moment of diagnosis and male gender were found to be predictive factors of low BMD. For patients with osteoporosis, disease duration was an additional predictive factor. After subgroup analysis predictive factors were found to be the same in patients with Crohn's disease.The prevalence of osteopenia and osteoporosis in IBD patients in a regional centre is as high as the prevalence rates reported from tertiary referral centres. A low BMI, an older age at the moment of diagnosis and male gender were predictive factors of low BMD. Prediction of osteoporosis and osteopenia using risk factors identified in this and previous studies is presently not feasible.

Porphyria cutanea tarda (PCT) results from a metabolic block in heme synthesis at the level of uroporphyrinogen decarboxylase. We measured the activity of one of the enzymes preceding it in the heme biosynthetic pathway, porphobilinogen deaminase (PBGD; EC 4.3.1.8), in erythrocytes of 47 patients with symptomatic or asymptomatic familial or sporadic PCT. PBGD activity was significantly increased in all four PCT groups, compared with controls. To study the mechanism of this increased PBGD activity, we determined, using polyclonal antibodies, the amount of immuno-detectable PBGD per 100 units of PBGD activity (Ig PBGD/100 U) and the total amount of immuno-detectable PBGD (Ig PBGD) in erythrocytes from all 47 patients and from controls. In both familial and sporadic PCT, Ig PBGD/100 U was decreased compared with that in controls (P less than 0.05). Especially in asymptomatic patients of the familial PCT group there was an inverse correlation between increasing PBGD activity and Ig PBGD/100 U (r = -0.90). In familial PCT, and to a minor degree in sporadic PCT, an increase in PBGD activity was accompanied by an increased Ig PBGD, compared with controls (familial PCT: P less than 0.001, sporadic PCT: P less than 0.05). In familial and sporadic PCT an increase in erythrocyte PBGD activity can, at least partly, be explained by a diminished degradation of PBGD. In familial PCT, in the symptomatic group more than in the asymptomatic group, and to a minor degree in sporadic PCT, there is in addition an increase in the absolute amount of PBGD.

BACKGROUND Active colitis impairs neoplasia detection during colonoscopic surveillance for colorectal cancer in patients with inflammatory bowel disease. We investigated whether fecal calprotectin testing before surveillance colonoscopy might identify ineffective surveillance procedures. METHODS All consecutive patients with Crohn's disease or ulcerative colitis scheduled for surveillance colonoscopy were asked to collect a stool sample before the start of bowel cleansing. Ineffective surveillance was defined as at least 1 colonic segment with moderate or severe inflammation. Calprotectin was analyzed using an enzyme-linked immunosorbent assay (Ridascreen; R-Biopharm). Receiver operator characteristics statistics were used to determine the optimal cutoff for calprotectin. RESULTS A total of 176 surveillance colonoscopies were performed in 164 patients, of which 83 had Crohn's disease and 81 had ulcerative colitis or inflammatory bowel disease-unclassified. Complete endoscopic remission or mild inflammation categorized as effective surveillance was observed in 151 colonoscopies (86%), whereas moderate or severe inflammation categorized as ineffective surveillance was observed in 25 colonoscopies (14%). Median calprotectin levels for the effective and ineffective surveillance group were 84 mg/kg (range, 20-4609) and 1605 mg/kg (range, 66-26,336), respectively (P < 0.01). A cutoff of 539 mg/kg identified patients with ineffective surveillance with 84% sensitivity, 89% specificity, 55% positive predictive value, 97% negative predictive value, and an area under the curve of 0.89. CONCLUSIONS Low fecal calprotectin accurately identifies inflammatory bowel disease patients without colonic inflammation in whom colorectal cancer surveillance is most effective.

Background: Faecal calprotectin is a non-invasive tool to asses colonic inflammation in patients with inflammatory bowel disease (IBD). However, the diagnostic accuracy of calprotectin is modest and analysis is relatively costly and therefore additional markers are needed. We compared the efficacy of faecal haemoglobin and calprotectin in the prediction of active endoscopic inflammation in IBD patients. Methods: Consecutive patients with either Crohn’s colitis (CD) or ulcerative colitis (UC) scheduled for surveillance colonoscopy collected a stool sample prior to the start of bowel preparation. Three experienced endoscopists assessed the presence of active inflammation in each colonic segment. Stool samples were analyzed for calprotectin and haemoglobin, both with a Ridascreen® enzyme-linked immunosorbent assay (R-Biopharm, Germany) and were analyzed without reference to colonoscopy findings and vice versa. ROC statistics were used to determine cutoff values for calprotectin and haemoglobin. Results: A total of 119 patients were included, of which 54 patients had CD, 59 had UC and 6 indeterminate colitis. Median (interquartile range [IQR]) calprotectin and haemoglobin concentrations were 118 mg/g (IQR 31 367) and 0.48 mg/g (IQR 0.27 5.73) respectively. Active inflammation was encountered in 43 patients (36%) (5 severe, 9 moderate and 29 mild). Calprotectin and haemoglobin levels were significantly higher in patients with active inflammation compared to patients with no inflammation [451mg/g vs 52mg/g, p < 0.01 and 10.34mg/g vs 0.34mg/g, p < 0.01 respectively (Mann Whitney U test)]. A cutoff value of 172mg/g for calprotectin predicted endoscopic inflammation with 79% sensitivity, 80% specificity, 69% positive predictive value and 87% negative predictive value. The predictive accuracy of calprotectin (area under the curve [AUC]) was 0.88 which was similar for patients with CD and UC (AUC 0.86 and 0.86 respectively). For haemoglobin, a cutoff value of 1.45mg/g indicated endoscopic inflammation with 77% sensitivity, 88% specificity, 79% positive predictive value and 87% negative predictive value. The predictive accuracy of haemoglobin was 0.88 (AUC) which was similar for patients with CD and UC (AUC 0.87 and 0.81 respectively). A combination of both tests did not increase the predictive accuracy compared to each test alone (AUC 0.88). Conclusions: Faecal haemoglobin can identify IBD patients with active inflammation with a predictive accuracy similar to faecal calprotectin. P262 Demographic characteristic of children with early clinical manifestation of the inflammatory bowel disease S. Szymanska1 *, M. Szczepanski1, P. Landowski2, G. CzajaBulsa3, E. Jarocka-Cyrta4, B. Korczowski5, E. Krzesiek6, E. Szymanska1, J. Kierkus1. 1The Children Memorial Health Institute, Hepatology and Feeding Disorders, Warsaw, Poland, 2Medical University of Gdansk, Poland, 3Pomeranian Medical University, Poland, 4Medical University of Bialystok, Poland, 5Medical College. University of Rzeszow, Poland, 6Medical University of Wroclaw, Poland

Background: Mutations in the IL-10 and IL-10 receptor genes are associated with early onset of IBD. Patients with IBD have a higher risk of developing colorectal cancer. Also mice deficient in IL-10 develop spontaneous enterocolitis and colitis-associated cancer (CAC) and may serve as model to study CAC in humans. The mutational mechanisms of cancer development and neoplastic progression in IL-10-/mice, however, remain unclear (Sturlan, Carcinogenesis 2001). Methods: Seven week old C57BL/6 IL-10-/mice received either 200 ppm piroxicam for 7 days as part of their chow (n=43) or plain chow (n=45). After 30 weeks mice were euthanized, the intestine was dissected, coiled as Swiss rolls and inflammation was scored according to Berg et al. (Cell 1996) with a scale from 0 to 4. Immunohistochemistry was performed for 8-oxoguanine (8-oxoG) to identify oxidative DNA lesions and for phosphohistone H2AX (γH2AX) to evaluate DNA double-strand breaks (DSB). The induction of DSB repair was assessed by staining for MRE11, homologous recombination by BRCA1 and nonhomologous end joining (NHEJ) by Ku70. Apoptosis was detected by TUNEL assay. Global methylation status was assessed by analysis of the CpG islands within the B1 element. LOH fragment analysis of micro-dissected tissue is currently performed (data will be available at the meeting). Results: Mice in both groups developed colitis. The average inflammation grade was 2.86 ± 0.64 (95%CI=2.663.06) and 1.8 ± 0.82 (95%CI=1.562.04) for the piroxicam and the control group, respectively (p,0.001). 13/43 piroxicam-treated mice developed 22 invasive tumors (max. 5 per mouse) compared to 3/45 mice with 6 tumors (max. 4 per mouse) in the control group (p=0.004). In both groups, the number of invasive tumors increased with the severity of inflammation (R= 0.539, p ,0.001). 8-oxoG also increased with a higher inflammation score (R=0.923, p,0.001). γH2AX positive cells were found in crypts and the intermediate zone and varied with disease progression (19.6% γH2AX positivity in non-inflamed, 37.7% in inflamed, and 45.7% in invasive tumors, respectively; p,0.001) and inflammation (R=0.858, p,0.001). γH2AX staining was coupled with overexpression of MRE11, BRCA1 and most pronounced with Ku70 indicating presence of DSB with activation of NHEJ. Apoptosis and global methylation were not induced in this model of CAC. Conclusion: Piroxicam treatment increases intestinal inflammation and tumorigenesis in IL-10-/mice. Progression of inflammation to cancer is associated with oxidative DNA damage and the overt presence of DSB. We propose that the persistent upregulation of error-prone NHEJ is a mechanism of chromosomal instability in this model of CAC.

s of the 7th Congress of ECCO the European Crohn’s and Colitis Organisation S7 OP13 TLR4 signaling activates the Wnt pathway in a mouse model of colitis-associated neoplasia R. Santaolalla1 *, J. Ruiz1, M. Fukata1, C. Espana1, J.M. Davies1, C. Pastorini1, D.A. Sussman1, J. Ko2, R. Dirisina2, T.A. Barrett2, M.T. Abreu1. 1University of Miami, Miller School of Medicine, Division of Gastroenterology, Miami, United States, 2Northwestern University, Feinberg School of Medicine, Division of Gastroenterology, Chicago, United States Background: In patients with colitis cancers, we have shown that there is an increased expression of toll-like receptor-4 (TLR4) in intestinal epithelial cells (IECs). The Wnt pathway regulates intestinal cell proliferation and the development of neoplasia. We have previously shown constitutive expression of TLR4 in a mouse model increases inflammation in the mucosa and drives colitis-associated tumorigenesis. We hypothesized that cross-talk between TLR4 and the Wnt pathway could have a synergistic tumorigenic effect in an inflammatory environment. Methods: Villin-TLR4 mice, which have increased TLR4 signaling in the intestinal epithelium, were compared to wild-type (WT) littermates. To induce tumorigenesis, the mice were treated with an initial injection of the genotoxic agent azoxymethane (AOM) followed by two cycles of 3% dextran sodium sulfate (DSS). Proliferation was assessed by BrDU labeling. Western blots were performed on intestinal IEC lines for total and phosphorylated b-catenin, GSK3-b and AKT. Immunohistochemistry was done for total b-catenin, active bcatenin phosphorylated at Ser552 and the Wnt target cyclin-D1. Villin-TLR4 mice were also crossed to Lgr5-EGFP mice. Lgr5 is Wnt target gene and stem cell marker. Results: Compared to WT mice, villin-TLR4 mice have a higher cell proliferation rate in the colon crypts and show a higher expression of nuclear b-catenin and cyclin-D1, indicating Wnt pathway activation. Given this, we asked whether TLR4 activity regulated expression of Lgr5. Our data demonstrate that whereas Lgr5 expression is normally located at the base of the crypts in WT mice, there is a dramatic expansion of Lgr5+ cells in the crypts of villin-TLR4 mice at baseline. To understand the mechanism underlying Wnt pathway activation, we found that TLR4 activates the phosphatidylinositide 3-kinase (PI3K)/AKT pathway to phosphorylate b-catenin at activation sites (Ser552 and Ser675). In vivo, we also see that villin-TLR4 AOM-DSS mice show increased nuclear b-catenin and phosphorylation of bcatenin at Ser552, an event associated with stem cell activation and colitis-associated cancer. Conclusions: In the intestinal mucosa, tonic stimulation of TLR4 by its ligand LPS can lead to activation of the Wnt signaling pathway and thereby promote neoplasia. In addition, TLR4 is linked to Lgr5 expression in the colon and is associated with an increase in proliferation of the colon crypts. These results highlight the role of bacterial signaling in stimulation of oncogenic pathways. OP14 Adenomas in patients with inflammatory bowel disease: Increased risk of advanced neoplasia F. van Schaik1 *, E. Mooiweer1, M. van der Have1, T. Belderbos1, F. ten Kate2, G.J. Offerhaus2, M. Schipper3, G. Dijkstra4, M. Pierik5, P. Stokkers6, C. Ponsioen7, D. de Jong8, D.W. Hommes9, A. van Bodegraven10, P. Siersema11, M. Van Oijen11, B. Oldenburg12. 1University Medical Center Utrecht, Gastroenterology and hepatology, Utrecht, Netherlands, 2University Medical Center Utrecht, Department of Pathology, Utrecht, Netherlands, 3University Medical Centre Utrecht, Department of Pathology, Utrecht, Netherlands, 4University Medical Center Groningen, Gastroenterology and Hepatology, Groningen, Netherlands, 5Maastricht University Medical Center, Dept of Gastroenterology, Maastricht, Netherlands, 6Academic Medical Center, Gastroenterology and Hepatology C2, Amsterdam, Netherlands, 7Academic Medical Center, Gastroenterology and Hepatology, Amsterdam, Netherlands, 8Universitair Medisch Centrum St Radboud, Afd Maag Darm Leverziekten, Nijmegen, Netherlands, 9UCLA, Division of Digestive Diseases, Los Angeles, United States, 10VU University Medical Center, Gastroenterology, Amsterdam, Netherlands, 11University Medical Center Utrecht, Department of Gastroenterology and Hepatology, Utrecht, Netherlands, 12University Medical Centre Utrecht, Department of Gastroenterology, Utrecht, Netherlands Background: The increased risk of colorectal cancer (CRC) in longstanding colitis is well documented, but it is unclear whether colonic adenomas in patients with inflammatory bowel disease (IBD) carry a higher risk of progression to CRC than sporadic adenomas in non-IBD patients. The aim of the current study was to evaluate the risk of high-grade dysplasia (HGD) or CRC in IBD patients compared to controls, and to compare this with age-matched IBD patients without adenomas. Methods: The nationwide pathology database (PALGA) was used to identify IBD patients with a histological diagnosis of adenoma (IBD + adenoma), age-matched IBD patients without adenoma (IBD-nonadenoma) and patients with adenoma without IBD (nonIBD + adenoma) in seven university hospitals in the Netherlands between 1995 and 2005. Medical charts and endoscopy, pathology and surgery reports were reviewed for adenoma characteristics and development of advanced neoplasia (HGD or CRC). The endoscopic appearance of adenomas was characterized as typical (solitary sessile or pedunculated) or atypical (all other descriptions). Survival analysis was used to assess the development of advanced neoplasia. Results: A total of 110 IBD + adenoma patients, 123 IBDnonadenoma patients and 179 nonIBD + adenoma patients were included. Mean duration of follow-up was 88 months (SD ±41). Polypectomy was performed in 68 IBD + adenoma patients (62%) and 146 nonIBD + adenoma patients (82%) (p < 0.01). The cumulative risks of advanced neoplasia were 16%, 7% and 4% in IBD + adenoma, IBD-nonadenoma and nonIBD + adenoma patients, respectively (p < 0.01). Atypical adenomas were identified in 28 IBD + adenoma patients (25%), which were associated with a higher cumulative risk of advanced neoplasia compared to IBD + adenoma patients with typical adenomas (29% versus 12%, p = 0.03). Conclusions: IBD patients with a histological diagnosis of adenoma have a significantly higher risk of developing advanced neoplasia during follow-up than adenoma patients without IBD and age-matched IBD patients without adenomas. This is due to a higher prevalence of atypical adenomas in IBD patients, conferring an increased risk of advanced dysplasia in these patients. by gest on M arch 1, 2016 http://eccoxfordjournals.org/ D ow nladed from

s of the 7th Congress of ECCO the European Crohn’s and Colitis Organisation S5 OP09 Anti TNF-a therapy is a major cost driver in inflammatory bowel disease: Results from the COIN study M. Van der Valk1 *, M.-J. Mangen2, G. Dijkstra3, A. van Bodegraven4, H. Fidder5, D. de Jong6, M. Pierik7, C.J. van der Woude8, M. Romberg-Camps9, C. Clemens10, J. Jansen11, P. van de Meeberg12, N. Mahmmod13, C. Ponsioen14, C. Rogge-Wolf15, R. Vermeijden16, P. Siersema1, M. Van Oijen1, B. Oldenburg17. 1University Medical Center Utrecht, Department of Gastroenterology and Hepatology, Utrecht, Netherlands, 2University Medical Center Utrecht, Julius Center for Health Sciences and Primary Care, Utrecht, Netherlands, 3University Medical Center Groningen, Gastroenterology and Hepatology, Groningen, Netherlands, 4VU University Medical Center, Gastroenterology, Amsterdam, Netherlands, 5University Medical Center Utrecht, Department of Gastroenterology, Utrecht, Netherlands, 6Universitair Medisch Centrum St Radboud, Afd Maag Darm Leverziekten, Nijmegen, Netherlands, 7Maastricht University Medical Center, Dept of Gastroenterology, Maastricht, Netherlands, 8Erasmus Medical Center, Department of Gastroenterology & Hepatology, Rotterdam, Netherlands, 9Orbis Medical Center, Gastroenterology and Hepatology, Sittard, Netherlands, 10Diaconessenhuis, Gastroenterology and Hepatology, Leiden, Netherlands, 11Onze Lieve Vrouwe Gasthuis, Gastroenterology and Hepatology, Amsterdam, Netherlands, 12Slingeland Hospital, Gastroenterology and Hepatology, Doetinchem, Netherlands, 13Antonius Hospital, Gastroenterology and Hepatology, Nieuwegein, Netherlands, 14Academic Medical Center, Gastroenterology and Hepatology, Amsterdam, Netherlands, 15Reinier de Graaf Gasthuis, Gastroenterology and Hepatology, Delft, Netherlands, 16Meander Medical Center, Gastroenterology and Hepatology, Amersfoort, Netherlands, 17University Medical Centre Utrecht, Department of Gastroenterology, Utrecht, Netherlands Background: Inflammatory bowel disease (IBD) is associated with a high economic burden to society. It has been estimated that up to two third of the total costs of IBD are generated by productivity loss. Most costs studies, however, have been performed before the introduction of the effective but costly biological therapies. In the present study, we therefore aimed to assess the total costs of IBD in a large cohort of IBD patients and to identify the main cost drivers. Methods: Between October 1, 2010 and June 1, 2011 a total of 10,947 patients with IBD were invited to participate in the COIN study and to fill out a web-based questionnaire every 3 months during 2 years of follow up. The questionnaires included questions on demographics, healthcare costs (visits to outpatient clinic, diagnostic procedures, hospitalization, surgery and medication use) and productivity costs (work days lost due to sick leave). Costs were calculated by multiplying resource use by the unit costs as determined by the Dutch pharmaco-economic guideline by Oostenbrink et al. Results: A total of 2,554 patients completed both the baseline as well as 3 months of follow-up questionnaires, of which 1,304 (51%) reported to suffer from Crohn’s disease (CD), 928 (36%) patients from ulcerative colitis (UC), and 322 (13%) had unspecified type of IBD. In CD the mean total costs in 3 months were €1,738, of which €1,468 (84%) were healthcare costs and €270 (16%) productivity costs. For UC costs were significantly lower (€896 total costs, €553 (60%) healthcare costs, €363 (40%) productivity costs). Overall, the single most expensive resource was medication use, mainly due to prescription of anti-TNF-a therapy accounting for an average of €1,044 (71%) in CD patients and €186 (33%) in UC patients. Furthermore, 10% of the high-cost patients accounted for 40% of the total health care costs in CD and 59% in UC, respectively. Conclusions: The traditional cost profile in IBD has changed and health care costs are now even more driven by medication costs, most importantly due to anti-TNF-a therapy. Whether this is balanced by a decrease in surgery and hospital admissionrelated costs as suggested by recent studies, and/or resulting in a reduction of sick leave and work disability will be assessed in follow-up studies. OP10 Importance of trough levels and antibodies on the long-term efficacy of infliximab therapy in ulcerative colitis M.T. Arias1 *, N. Vande Casteele2, D. Drobne3, M. Ferrante4, I. Cleynen5, V. Ballet6, P. Rutgeerts4, A. Gils7, S. Vermeire4. 1Hospital Universitario Marques De Valdecilla, Santander, Spain, 2University Hospitals Leuven, Leuven, Belgium, 3University Medical Centre Ljubljana, Gastroenterology & Hepatology, Ljubljana, Slovenia, 4University Hospital Gasthuisberg, Department of Gastroenterology, Leuven, Belgium, 5Catholic University Leuven, Gastroenterology, Leuven, Belgium, 6University Hospitals Leuven, Department of Gastroenterology, Leuven, Belgium, 7Catholic University Leuven, Department of Pharmaceutical Biology, Leuven, Belgium Background: Infliximab (IFX) is an efficacious therapy for Crohn’s disease (CD) and ulcerative colitis (UC). Recent studies in CD have proposed that trough level (TL) and antibody measurement (ATI) may help to optimize therapy but data in UC, especially correlating TL to long-term outcome of treatment, are scarce. We studied the influence of TL and ATI formation on the long-term outcome of IFX in UC. Methods: 135 consecutive UC patients receiving induction and maintenance IFX therapy were analyzed retrospectively. TL were measured at predefined timepoints using an in-house developed ELISA. Clinical response, C-reactive protein (CRP), serum albumin and hemoglobin were assessed at the same timepoints. ATI were measured in samples with undetectable TL. Results: By week 10, 81% of patients responded to treatment and entered maintenance. 50% needed dose optimization for loss of response (LOR) during follow up (FU). Low IFX TL early after induction were associated with higher risk for LOR and discontinuation (p = 0.02). The ROC curve of TL at week 14 had an area under the curve of 0.672 (s.e = 0.06; p = 0.01), and a TL of 7.19mg/mLwas the cut-off value with 80% specificity and 57% sensitivity for sustained benefit. Patients with undetectable TL at least once during FU demonstrated a significantly shorter time to dose optimization (log-rank p = 0.02) and LOR (logrank p = 0.01). An increase in TL after dose optimization was associated with restoration of response (median, [IQR] 8.7mg/mL [3.9 17.4] vs 3.2 [0.3 16.4]; p = 0.05). In 26% patients (29) IFX was stopped for complete LOR and in 12 (11%), undetectable TL and ATI were measured at least once during FU. 58% patients were on immunomodulators (IMM) at baseline. Although concomitant IMM did not affect discontinuation rates, median TL throughout the study were higher while patients were on IMM (median, [IQR]; 10.5 mg/mL [4.4 21.0] vs 7.9 [2.4 19.8]; p = 0.02) and ATI were less frequently observed while on IMM (OR 0.6 [0.4 0.7]; p < 0.0001). Albumin, CRP and hemoglobin baseline values all correlated with TL (r = 0.152, r = 0.167, r = 0.162; p < 0.0001). Conclusions: LOR to IFX in UC patients is associated with low TL and formation of ATI. Low TL early after induction predict future LOR. Although concomitant IMM did not affect clinical LOR rates, the TL were higher in patients on combination therapy. Longer FU will learn how important these pharmacokinetic changes may be for sustained benefit. Our data further support a role for TL monitoring in optimizing IFX therapy in IBD. by gest on M ay 6, 2016 http://eccoxfordjournals.org/ D ow nladed from

number of hospitalizations, the body mass index (BMI), the presence of extraintestinal manifestations, the need for transfusion, transfer from another hospital and the frequency of Clostridium difficile infection were also compared and statistically analyzed between the two groups. P< 0.05 was considered statistically significant. Results: The records of 182 UC patients with a total of 276 hospitalizations because of severe relapse of UC were reviewed. 25.3% of the patients underwent colectomy after average disease duration of 10 years. Pancolitis, low hematocrit level and the need for blood transfusion were significantly more common in the colectomy group (p = 0.004 and p = 0.002). The number of hospitalization, disease duration, and a lower BMI value also proved to predispose to surgery (p = 0.004, p = 0.049 and p = 0.027). Women, younger age at the diagnosis and thrombocytosis showed borderline significance in the risk for colectomy. No association was found between the need for surgery and total parenteral nutrition, Clostridium difficile infection, transfer from another hospital, and electrolyte disturbances. None of the inflammatory laboratory parameters were associated with colectomy. Smoking and the presence of extraintestinal manifestations also did not predispose to surgery. Conclusions: Our results suggest that severe UC patients with pancolitis, anemia, longer disease duration and frequent hospital admissions have a higher likelihood of requiring colectomy.