Predictors for subsequent need for immunosuppressive therapy in early Crohn's disease

Transcription

1 Journal of Crohn's and Colitis (2012) 6, Available online at Predictors for subsequent need for immunosuppressive therapy in early Crohn's disease Sandra Wenger a, Susanna Nikolaus b, Stefanie Howaldt c, Bernd Bokemeyer d, Andreas Sturm e, Jan C. Preiss f, Alain M. Schoepfer g, Andreas Stallmach a,, Carsten Schmidt a a University Clinic Jena, Clinic of Internal Medicine II, Jena, Germany b Christian Albrechts-University, Clinic of Internal Medicine, Kiel, Germany c Gastroenterologie am Rothenbaum, Hamburg, Germany d Gastroenterologische Gemeinschaftspraxis, Minden, Germany e Charité, Department of Gastroenterology and Hepatology, Campus Virchow Clinic, Berlin, Germany f Charité, Campus Benjamin Franklin, Medizinische Klinik I, Berlin, Germany g Department of Gastroenterolog and Hepatology, University hospital Lausanne, Switzerland Received 14 February 2011; received in revised form 12 June 2011; accepted 14 June 2011 KEYWORDS Clinical; Parameter; Prediction; Crohn; IBD Abstract Background and aims: The clinical course of Crohn's disease (CD) is highly variable with a subgroup of patients developing a progressive disease course necessitating immunosuppressive therapy (IT). However, reliable, stable and non-invasive individual clinical parameters in order to identify patients at risk for undergoing subsequent IT have not been sufficiently established. We therefore aimed to identify such clinical parameters. Methods: A retrospective, multicenter analysis of CD patients from 6 German tertiary IBD centers was performed. Patients were classified into two groups depending on requiring IT or not. Personal data, clinical and laboratory parameters during the first 3 months after CD diagnosis and effects of initial medical therapy were compared between these two groups. Results: In 218 (61.8%) of the 353 patients the CD course necessitated IT. Those patients were significantly younger at symptom onset and diagnosis, and required significantly more often a systemic corticosteroid therapy. Furthermore, significant differences in serological markers of inflammation were observed. Age, gender and the effect of initial steroid therapy were used to develop a prognostic model predicting the individual probability of necessitating IT. Corresponding author at: Klinik für Innere Medizin II, Abteilung für Gastroenterologie, Hepatologie und Infektiologie, Friedrich-Schiller- Universität Jena, Erlanger Allee Jena, Germany. Tel.: ; fax: address: andreas.stallmach@med.uni-jena.de (A. Stallmach) /$ - see front matter 2011 European Crohn's and Colitis Organisation. Published by Elsevier B.V. All rights reserved. doi: /j.crohns

2 22 S. Wenger et al. Conclusions: The simple clinical items age at diagnosis, gender, and need for systemic steroid therapy can predict a progressive disease course in early CD. Our model based on these parameters allows an individualized estimation of each patient's risk to develop a progressive disease course. Thereby, our model can help in deciding if patients will need immunosuppressive drugs early in the disease course or if a careful watch and wait strategy is justified European Crohn's and Colitis Organisation. Published by Elsevier B.V. All rights reserved. 1. Introduction The clinical course of Crohn's disease (CD) is highly variable and ranges from a single episode to a potentially life-threatening continuous disease. In a population-based prospective study with a follow-up of five years performed by the IBSEN group, four different types of disease course have been differentiated: patients with a decrease in intensity following the first flare of disease (44%), patients with an increase in intensity (3%), patients with chronic continuous symptoms (24%) and patients with chronic intermittent symptoms (29%), resp. 1 While the localization of CD is relatively stable, the disease type changes from an inflammatory pattern to a stricturing or penetrating disease over time, giving rise to a complicated disease course. 2 The early introduction of immunosuppressive drugs (thiopurines or methotrexate, e.g.) or anti-tnf-α antibodies may have an impact on the occurrence of severe flares, on the development of complications such as stenoses and perforations, on the necessity of surgery and finally on the patients quality of life. 3 Therefore, it is of major importance to identify those patients who will exhibit a severe disease course as early as possible, as these therapeutic measures may alter the natural course of CD and thereby the occurrence of these complications. However, early introduction of immunosuppressive treatment in all CD patients could result in an overtreatment of those patients showing a less disabling disease course not requiring treatment intensification. 3 7 In a population-based study from Olmsted county, Minnesota, it has been shown that 57% of CD patients did not require corticosteroid therapy at all. And among patients receiving corticosteroids, 32% of CD patients were corticosteroid free without operation after 1 year of follow-up. 8 Moreover, there is an ongoing debate about potentially harmful side-effects of immunosuppressors as well as anti-tnf-α antibodies. Therefore, it is necessary to identify those patients who will eventually develop a severe disease course and in whom the early introduction of immunosuppressors and anti-tnf-α agents will be necessary. Though, at present, it is difficult to predict the individual patients' course of CD. Reliable, non-invasive and easily available clinical parameters in order to identify patients at risk for a complicated disease course early in a case history have not been established sufficiently and have neither been introduced into clinical algorithms so far. Already identified parameters associated with a complicated disease course such as fistulizing disease, e.g., are less useful in clinical practice because these parameters reflect a complication rather than predicting it. Therefore, we intended to identify simple clinical parameters during an early period after initial diagnosis of CD pointing towards a progressive disease course and the subsequent need for immunosuppressive therapy. 2. Materials and methods 2.1. Patients We performed a retrospective, multicenter analysis in CD patients from six German IBD centers. Besides gastroenterological outpatient clinics from four university clinics (University Clinic Jena, Charité Universitätsmedizin Berlin: Campus Virchow-Klinikum and Campus Benjamin Franklin, respectively, and Christian Albrechts-University, Kiel) there were two specialized IBD private practices (Gastroenterologie am Rothenbaum, Hamburg and Gastroenterologische Gemeinschaftspraxis Minden) involved in the study. Patients diagnosed with CD between January 2003 and February 2008 with a follow-up period of at least six months after diagnosis or with initiation of immunosuppressive therapy after at least three months after diagnosis of the disease were included in the study. The need for immunosuppressive therapy served as a surrogate parameter of a progressive disease course in CD patients. Our patients were classified into two groups depending on requiring immunosuppressive therapy or not. Immunosuppressive therapy contains thiopurines, methotrexate, anti-tnf-α antibodies, as well as cyclophosphamide, and tacrolimus. Of note, topical (budesonide) and systemic corticosteroids were not regarded as immunosuppressive therapy. Further details on the frequency of the specific medication prescribed are provided in the section immunosuppressive therapy in the Results section. Immunosuppressive therapy was begun by study doctors if CD patients underwent two or more flares within a time period of 12 months. The definitions of active disease and remission followed the guidelines from the European Crohn's and Colitis Organization (ECCO). 9 Patients were excluded from the study in case of missing informed consent, or in case of loss of follow-up. Medical records were investigated in order to identify patients who subsequently will exhibit a severe course of the disease requiring immunosuppressive therapy. Personal data such as date of first diagnosis, date of first symptoms, gender, smoking habits and family history of IBD as well as clinical parameters available during an early phase after the initial diagnosis of CD were recorded. Moreover, extension of the disease, extraintestinal manifestations (articular, ocular and cutaneous manifestations, resp.), fever at the first flare of CD, abdominal tenderness, existence of granulomas in

3 Predictors for immunosuppressant s need in CD 23 histology and laboratory parameters (hemoglobin, thrombocytes and CRP level) were investigated, too. Furthermore, we evaluated necessity, time of initiation, kind and effect of oral steroid therapy Statistical considerations Data were analyzed using SPSS In order to identify significant differences between patients requiring immunosuppressive medication and those not requiring this therapy each variable was investigated by using univariate Cox regression analysis. The level of significance was set at α 0.05 (2-sided). Oneway Anova and Welch test were performed to calculate temporal differences. Survival curve analysis was performed according to the Kaplan-Meyermethod. The simultaneous influence of several risk factors on the need for immunosuppressive therapy has been examined using the Cox model. The event in this survival analysis is the necessity for an immunosuppressive therapy. The Cox regression is based on the assumption that the effects of variables on survival (necessity for immunosuppressants) are constant over time (proportional hazards assumption). From the estimated regression parameters β 1, β 2 and β 3 for the three risk factors sex (x 1 ), age (x 2 ) and effect of steroid therapy (x 3 ), the survival probability at time point t, S(t), was estimated for selected combinations by use of with the survival function Ŝ t ðþ¼e H 0 ðþ t eˆβ 1 x1 þ ˆβ 2 x2 þ ˆβ 3 x H 0 (t) is the baseline hazard at time point t. Table 1 Baseline characteristics of the entire study population (n = 353). The sum of involved intestinal areas exceeds 100% as there was more than one involved area in several patients. Baseline characteristics Number of patients Median age/ percentage Age at onset of disease years Age at diagnosis years Sex Male Female Smoking status Smoker Non-smoker Former smokers Unknown Involved intestinal areas Upper GI tract Ileal disease Ileocolonic disease Colonic disease Perianal disease Extraintestinal manifestation Articular manifestation Cutaneous manifestation 7 2 Ocular manifestation Ethical statement The study was approved by the ethics committee of the Friedrich-Schiller-University Jena ( /07). The study has been performed in accordance with the principles laid down in the Declaration of Helsinki. 3. Results The study population consisted of 353 CD patients from both gastroenterological University hospital outpatient clinics and private practices. The main baseline characteristics of our patients are summarized in Table Immunosuppressive therapy In 218 patients (61.8%) the course of CD necessitated an immunosuppressive therapy. Among these patients 204 (93.6%) received azathioprine, 57 (26.1%) infliximab, 31 (14.2%) adalimumab, 23 (10.6%) 6-mercaptopurine, 16 (7.3%) methotrexate, 4 (1.8%) certolizumab pegol, 1 (0.5%) cyclophosphamide, 1 (0.5%) visilizumab and 1 (0.5%) tacrolimus, respectively. The total counts more than 100% as several immunosuppressive drugs were combined. The number of different immunosuppressive drugs in an individual patient as well as the cumulative probability of receiving immunosuppressive treatment in the whole population regardless of possible predictors is illustrated in Fig. 1. The decision to initiate immunosuppressive therapy was based on several reasons. Most patients requiring immunosuppressive treatment showed a chronic active disease course including steroid dependence and steroid resistance as well as frequently recurring flares of CD requiring corticosteroid therapy. Furthermore, in several cases the development of a complicated disease course, the presence of perianal fistulizing disease or extraintestinal manifestations supported the initiation of immunosuppressive agents. The reasons for the initiation of immunosuppression are summarized in Table 2. In total, the sum exceeds 100% as there was more than one reason to initiate IT in individual patients. In general, immunosuppressive drugs were well tolerated. Side effects were documented in only 19.6% of patients Demographic parameters Patients with a progressive disease course were significantly younger at disease onset (28.3 vs years, P=0.036) and diagnosis (29.1 vs years, P = 0.008) than patients not requiring immunosuppressors. To be more specific, patients 10 years older at diagnosis had a 14% lower risk requiring immunosuppressive medication in their individual disease course compared to patients being 10 years younger. In total we evaluated 44.5% male and 55.5% female patients with CD. Male gender was found to be a significant

4 24 S. Wenger et al. interest, active smoking was not predictive for a progressive disease course with the need for immunosuppressive drugs (P=0.561). However, a significant difference could be observed between non-smokers and former smokers with former smokers having a higher risk to require immunosuppressors (P = 0.032). In our cohort 7.3% of patients reported about having first degree-relatives with IBD. However, there was no significant difference in the need for immunosuppressive treatment in patients with a family history of inflammatory bowel disease (P=0.949) Disease specific parameters Figure 1 Representation of the number of immunosuppressive drugs per patient during follow-up (a) and the cumulative probability of receiving immunosuppressive treatment in the whole population regardless of possible predictors (b). risk factor with a higher probability for immunosuppressive treatment (male gender hazard ratio (HR) 1, female gender HR 0.751, P=0.043). 34.8% of patients in our study group were smokers, while 32.3% were non-smokers and 2.5% were former smokers. Smoking status was unknown in 30.3% of patients. Of Table 2 Reasons for initiation of immunosuppressive therapy in 182 patients. Reason for immunosuppressive therapy Number of patients Steroid-dependent course of the disease Steroid-refractory course of the disease Severe disease course requiring early immunosuppression Perianal disease Severe disease course with complications requiring early immunosuppression Extraintestinal manifestations Percentage We investigated if there exists a correlation between involved intestinal areas according to the Vienna Classification and the subsequent need for immunosuppressive therapy. Ileocolonic disease was the most common localization (51.7%), followed by colonic disease (25.9%) and ileal disease (21.9%) while perianal disease (9.9%) and manifestation of upper GI tract (9.1%) occurred more rarely. There was no significant difference between patients requiring immunosuppressive therapy and those not requiring this kind of therapy in either subgroup. Articular (5.5%), cutaneous (2.0%) or ocular (0.6%) manifestations as extraintestinal manifestations of CD at diagnosis were described rarely. However, the frequency increased during the subsequent course of disease. There was no significant difference in the subsequent use of immunosuppressors depending on the presence of extraintestinal manifestations at diagnosis Steroid therapy In total 91.4% of patients required corticosteroids at any time in the course of CD (either budesonide or systemic acting corticosteroids). Table 3 summarizes data on steroid treatment in our population. Regarding steroid therapy there were many significant differences between patients requiring subsequent immunosuppressive medication and those without. At the first flare significantly more patients were treated with systemic steroids (prednisone, prednisolone, methylprednisolone; Table 3 Descriptive statistics concerning kind and effect of steroid therapy. Number of patients Percentage Steroid therapy At any time 320/ At first flare 256/ Kind of steroid therapy at first flare Budesonide 72/ Systemic steroids 189/ Effect of steroid therapy Remission 33/ Improvement 247/ No effect 30/

5 Predictors for immunosuppressant s need in CD 25 Figure 2 Patients with a progressive course of CD and the subsequent need for immunosuppressive therapy significantly more often required corticosteroids at any time (P b 0.001). Figure 3 Remission upon steroid therapy at the first flare of CD is associated with a reduced need for immunosuppressive medication (P = 0.014). HR 1.647, P = 0.011). Patients consecutively being treated with immunosuppressors significantly more often required corticosteroids (systemic and locally acting) at the first flare of CD (HR 1.560, P=0.01) and at any time during the disease course (HR 6.359, P b 0.001; please see Fig. 2). Patients requiring immunosuppressive therapy significantly more often showed no effect upon steroids (HR 1.898, P=0.008) while use of budesonide therapy at the first flare was associated with a reduced rate of subsequent immunosuppressive therapy (HR 0.593, P = 0.007). Moreover, remission upon steroid therapy was associated with a reduced need for immunosuppressive treatment (HR 0.491, P = 0.014). Time duration between onset of CD symptoms and first steroid therapy was not significantly different between patients requiring immunosuppressive therapy and patients not requiring this therapy (14.5 vs months, P = 0.141), except for the patients who required steroids at the first flare of the disease, as mentioned above. As expected, the time span between a steroid treated flare and a subsequent flare was significantly shorter in those patients with a future need for immunosuppressive drugs (10.3 vs months, P = 0.015). The patient's response to steroid treatment after diagnosis (remission, response or no effect, resp.) was significantly associated with subsequent use of immunosuppressive treatment in univariate (P b 0.001) as well as in multivariate Cox regression analysis (P b 0.001); please see Fig. 3 for further details. (15.4%), resp., were not significantly different expressed between CD patients with a severe or with a mild disease course (PN0.05). Patients with CD necessitating immunosuppressive medication significantly more often showed anemia at diagnosis (HR 1.895, P=0.009; please see Fig. 4) compared to patients without immunosuppressive treatment. Anemia was defined 3.5. Clinical, histopathological and laboratory parameters Fever at onset (7.4%) or at first flare (15.3%), abdominal tenderness at first physical examination (7.7%) and existence of granulomas in histopathological specimens at diagnosis Figure 4 CD patients with anemia at diagnosis showed a more severe disease course and had a higher probability to require immunosuppressive drugs than patients with a hemoglobin value within the reference range (P= 0.009).

6 26 S. Wenger et al. according to recently published guidelines. 11 Concerning other laboratory parameters like CRP and platelets at diagnosis we could not observe significant differences between patients with or without future need for immunosuppressive therapy. However, patients requiring immunosuppressive therapy significantly more often showed thrombocytosis during the disease course (HR 1.655, P=0.005) Multivariate Cox regression analysis By use of multivariate Cox regression analysis we identified three independent parameters in the early course of CD predicting the probability for the individual patient's need for immunosuppressive therapy: male gender (P = 0.043), young age at diagnosis (P = 0.014) and insufficient effect (lack of response) of steroid therapy (P b 0.001). 19 patients had to be excluded from multivariate analysis due to insufficient data (please see Table 4). Based upon these 3 parameters we developed a prognostic model allowing to calculate the individual patient's probability to require immunosuppressors in the subsequent course of the disease ( de/kim+ii/gastroenterologie_+hepatologie_+infektiologie/ Forschung/Prognostisches+Modell+Morbus+Crohn.html). Following there are three examples given, illustrating the probability to require immunosuppressive treatment according to the model (please see Fig. 5). 4. Discussion The aim of our study was to identify clinical parameters in early CD that predict a progressive disease course, represented by the subsequent need for immunosuppressive therapy. Considering that our study population consisted of patients from specialized IBD centers only, a high rate of 61.8% of patients was treated with immunosuppressors. Therefore, we feel confident that our data predicting a severe disease course are highly reliable as the high frequency of immunosuppressors from the participating centers allows an excellent estimation of these prognostic parameters. Table 4 Independent parameters significantly associated with a severe course of CD, predicting the risk for subsequent necessity of immunosuppressive therapy. n Hazard ratio p-value Gender Male Female Effect of steroid therapy b0.001 Remission 27 1 Response No improvement No steroids after diagnosis Age at diagnosis Certain age 1 One year less risk for immunosuppressive therapy [%] time after diagnosis [months] male, 21 years, no effect of steroids male, 35 years, improvement of symptoms upon steroids female, 45 years, remission upon steroids Figure 5 Calculation of the individual probability to require immunosuppressors in three hypothetical patients with CD according to our model, based on gender, age at diagnosis and effect upon corticosteroids. The most important finding of our study was the identification of three independent clinical parameters, which were significantly differently expressed between patients requiring immunosuppressive treatment and those without: age at diagnosis, gender and effect of primary steroid therapy (remission, improvement or no effect). These parameters that are easy to obtain early in a patient's medical history were used to develop a prognostic model which enables us to predict a progressive course of CD and the subsequent probability for needing immunosuppressive treatment on an individual basis. However, several parameters such as smoking, manifestation of the upper GI tract and perianal disease that were expected to predict a severe disease course according to former studies were not confirmed to be predictive by our study. The first substantial result of our study was that CD patients necessitating immunosuppressive therapy required steroids more frequently. Especially, these patients have been treated significantly more often with systemic corticosteroids at the first flare, showed significantly lower remission rates upon steroids and significantly more often had no response upon steroid therapy compared to patients without subsequent immunosuppressive treatment. Consequently, we observed that budesonide therapy at the first flare was a predictive parameter against requiring subsequent immunosuppressive therapy. Some data in the literature suggest that kind and effect of steroid therapy might be useful to predict the individual disease course, as well. The initial requirement for steroids was described to be predictive of a subsequent severe disease course by Beaugerie and colleagues. 12 Furthermore, Lakatos and coworkers showed that the necessity to use steroids is predictive of a change in disease behavior in initially non-stricturing and non-penetrating disease patients. 15 Gelbmann et al. observed an association between steroid resistance and a complicated disease course including perianal disease and bowel resection. 16 Second, younger age at diagnosis has been identified as an independent parameter that is associated with an increased risk for future immunosuppressive therapy. In different studies young age at diagnosis was described as a predictive parameter of a complicated disease course. Young age has been identified as a negative prognostic factor for surgery as well as for an increase in recurrence rates. 17 An association

7 Predictors for immunosuppressant s need in CD 27 of age below 40 years at diagnosis with a subsequent disabling disease course within 5 years has been described, 12 whereas age above 40 years was protective in regard to disease recurrence. 13 Third, we identified male gender as a risk factor for a severe disease course, as well. Interestingly, there are only a few reports in the literature highlighting gender differences in the course of IBD. While Gupta and colleagues showed that female gender in pediatric patients with CD was associated with an increased risk for surgery, 18 Wang and colleagues identified male gender as a risk factor for surgery and reoperation in CD patients. 19 In summary, in the presence of all three risk factors the probability to experience a progressive disease course requiring immunosuppressive therapy was almost 100% over a period of 5 years, whereas it was below 50% in the absence of any risk factor. Therefore, our data do not only show statistically relevant results, but our model also provides usefulness in daily clinical practice to calculate the future need for immunosuppressors. Laboratory parameters may also contribute to the prediction of an individual patient's disease course. We observed that anemia at diagnosis had a higher probability to require immunosuppressive drugs than patients with a hemoglobin value within the reference range. But, we could observe only a trend towards elevated CRP levels at diagnosis being predictive for a progressive disease course. Similarly, a previous study from the IBSEN study group showed a significant association between CRP levels at diagnosis and the risk of surgery in patients with terminal ileitis. 20 Our prognostic model allows to predict a progressive disease course early in a patient's history. The response to corticosteroid therapy serves as a distinguishing parameter besides age at diagnosis and gender in predicting an individual patient's disease course. Other parameters associated with a severe disease course in the literature (such as perianal disease, upper GI involvement or extraintestinal manifestations, e.g. 12,13,21 24 have been observed only rarely in our large cohort and were not significantly associated with subsequent immunosuppressive therapy. Therefore, these parameters may represent a later event in a patient's disease history rather than serving as a predictor in an early phase of CD. Patients identified thus are likely to benefit from an early introduction of immunosuppressive therapy that may change the natural course of CD, because high rates of clinical relapse, intestinal resections and steroid dependence may occur upon conventional therapeutic strategies without early introduction of immunosuppressors. 25,26 Our study has method-inherent strengths and also drawbacks. The major strength is the development of a simple model, based on easily accessible clinical items, to predict the individual patient's risk to develop a progressive disease course. This model has the potential to impact clinical decision making. Furthermore, the cohort represents a well balanced mixture of patients from tertiary referral centers and also private practices and is therefore not prone to selection bias. The involvement of dedicated IBD physicians has ensured a high data quality. One potential drawback lies in the retrospective design. In summary, young age at diagnosis, male gender, and insufficient response to initial corticosteroid therapy are associated with a progressive disease course in CD necessitating further therapy with immunomodulators. Our model, based on easily accessible clinical parameters, can predict the individual CD's patient risk to develop a progressive disease course and thereby assist in taking clinical decisions. Acknowledgments This study was supported by the Competence Network Inflammatory bowel disease and financially supported by Abbott International. The study sponsors had no involvement in the study design, in the collection, analysis and interpretation of data, in the writing of the manuscript and in the decision to submit the manuscript for publication. We thank Dr. Thomas Lehmann for assistance concerning statistical analysis. SW carried out the study, performed the statistical analysis and drafted the manuscript, SN, SH, BB, AS and JCP participated in data collection, AMS provided significant advice or consultation, AS designed the study and participated in interpretation of data and CS participated in the design of the study and interpretation of data and helped to draft the manuscript. All authors read and approved the final manuscript. References 1. Henriksen M, Jahnsen J, Lygren I, Aadland E, Schulz T, Vatn M, Moum B. Clinical course in Crohn's disease: results of a five-year population-based follow-up study (the IBSEN study). Scand J Gastroenterol 2007;42: Cosnes J, Cattan S, Blain A, Beaugerie L, Carbonnel F, Parc R, Gendre JP. Long-term evolution of disease behavior of Crohn's disease. Inflamm Bowel Dis 2002;8: Cosnes J. Can we modulate the clinical course of inflammatory bowel diseases by our current treatment strategies? Dig Dis 2009;27: Feagan BG, Panaccione R, Sandborn WJ, D'Haens GR, Schreiber S, Rutgeerts PJ, Loftus Jr EV, Lomax KG, Yu AP, Wu EQ, Chao J, Mulani P. Effects of adalimumab therapy on incidence of hospitalization and surgery in Crohn's disease: results from the CHARM study. Gastroenterology 2008;135: Holtmann MH, Krummenauer F, Claas C, Kremeyer K, Lorenz D, Rainer O, Vogel I, Böcker U, Böhm S, Büning C, Duchmann R, Gerken G, Herfarth H, Lügering N, Kruis W, Reinshagen M, Schmidt J, Stallmach A, Stein J, Sturm A, Galle PR, Hommes DW, D'Haens G, Rutgeerts P, Neurath MF. Long-term effectiveness of azathioprine in IBD beyond 4 years: a European multicenter study in 1176 patients. Dig Dis Sci 2006;51: Lichtenstein GR, Yan S, Bala M, Blank M, Sands BE. Infliximab maintenance treatment reduces hospitalizations, surgeries, and procedures in fistulizing Crohn's disease. Gastroenterology 2005;128: Cosnes J, Nion-Larmurier I, Beaugerie L, Afchain P, Tiret E, Gendre JP. Impact of the increasing use of immunosuppressants in Crohn's disease on the need for intestinal surgery. Gut 2005;54: Faubion Jr WA, Loftus Jr EV, Harmsen WS, Zinsmeister AR, Sandborn WJ. The natural history of corticosteroid therapy for inflammatory bowel disease: a population-based study. Gastroenterology 2001;121: Stange EF, Travis SP, Vermeire S, Beglinger C, Kupcinkas L, Geboes K, Barakauskiene A, Villanacci V, Von Herbay A, Warren BF, Gasche C, Tilg H, Schreiber SW, Schölmerich J, Reinisch W. European evidence based consensus on the diagnosis and

8 28 S. Wenger et al. management of Crohn's disease: definitions and diagnosis. Gut 2006;55: Gasche C, Berstad A, Befrits R, Beglinger C, Dignass A, Erichsen K, Gomollon F, Hjortswang H, Koutroubakis I, Kulnigg S, Oldenburg B, Rampton D, Schroeder O, Stein J, Travis S, Van Assche G. Guidelines on the diagnosis and management of iron deficiency and anemia in inflammatory bowel disease. Inflamm Bowel Dis 2007;13: Hosmer DW, Lemeshow S. Applied Survival Analysis: Regression Modeling of Time to Event Data. New York (NY): Wiley; Beaugerie L, Seksik P, Nion-Larmurier I, Gendre JP, Cosnes J. Predictors of Crohn's disease. Gastroenterology 2006;130: Wolters FL, Russel MG, Sijbrandij J, Ambergen T, Odes S, Riis L, Langholz E, Politi P, Qasim A, Koutroubakis I, Tsianos E, Vermeire S, Freitas J, van Zeijl G, Hoie O, Bernklev T, Beltrami M, Rodriguez D, Stockbrügger RW, Moum B. Phenotype at diagnosis predicts recurrence rates in Crohn's disease. Gut 2006;55: Cosnes J, Beaugerie L, Carbonnel F, Gendre JP. Smoking cessation and the course of Crohn's disease: an intervention study. Gastroenterology 2001;120: Lakatos PL, Czegledi Z, Szamosi T, Banai J, David G, Zsigmond F, Pandur T, Erdelyi Z, Gemela O, Papp J, Lakatos L. Perianal disease, small bowel disease, smoking, prior steroid or early azathioprine/biological therapy are predictors of disease behavior change in patients with Crohn's disease. World J Gastroenterol 2009;15: Gelbmann CM, Rogler G, Gross V, Gierend M, Bregenzer N, Andus T, Schölmerich J. Prior bowel resections, perianal disease, and a high initial Crohn's disease activity index are associated with corticosteroid resistance in active Crohn's disease. Am J Gastroenterol 2002;97: Romberg-Camps MJ, Dagnelie PC, Kester AD, MA Hesselink-van de Kruijs, Cilissen M, Engels LG, Van Deursen C, Hameeteman WH, Wolters FL, Russel MG, Stockbrügger RW. Influence of phenotype at diagnosis and of other potential prognostic factors on the course of inflammatory bowel disease. Am J Gastroenterol 2009;104: Gupta N, Cohen SA, Bostrom AG, Kirschner BS, Baldassano RN, Winter HS, Ferry GD, Smith T, Abramson O, Gold BD, Heyman MB. Risk factors for initial surgery in pediatric patients with Crohn's disease. Gastroenterology 2006;130: Wang JP, Huang MX, Yin L, Qian Q, Lan P, Wang T, He XS. Risk factors of operation and reoperation in patients with Crohn disease. Zhonghua Wei Chang Wai Ke Za Zhi 2007;10: Henriksen M, Jahnsen J, Lygren I, Stray N, Sauar J, Vatn MH, Moum B. C-reactive protein: a predictive factor and marker of inflammation in inflammatory bowel disease. Results from a prospective population-based study. Gut 2008;57: Siproudhis L, Mortaji A, Mary JY, Juguet F, Bretagne JF, Gosselin M. Anal-lesions: any significant prognosis in Crohn's disease? Eur J Gastroenterol Hepatol 1997;9: Veloso FT, Ferreira JT, Barros L, Almeida S. Clinical outcome of Crohn's disease: analysis according to the vienna classification and clinical activity. Inflamm Bowel Dis 2001;7: Chow DK, Sung JJ, Wu JC, Tsoi KK, Leong RW, Chan FK. Upper gastrointestinal tract phenotype of Crohn's disease is associated with early surgery and further hospitalization. Inflamm Bowel Dis 2009;15: Lakatos PL, Szalay F, Tulassay Z, Molnar T, Kovacs A, Gasztonyi B, Papp J, Lakatos L. Clinical presentation of Crohn's disease. Association between familial disease, smoking, disease phenotype, extraintestinal manifestations and need for surgery. Hepatogastroenterology 2005;52: Domenech E, Manosa M, Cabre E. Top-down therapy: is the evidence strong enough? Dig Dis 2009;27: Szamosi T, Banai J, Lakatos L, Czegledi Z, David G, Zsigmond F, Pandur T, Erdelyi Z, Gemela O, Papp M, Papp J. Lakatos PL. Early azathioprine/biological therapy is associated with decreased risk for first surgery and delays time to surgery but not reoperation in both smokers and nonsmokers with Crohn's disease, while smoking decreases the risk of colectomy in ulcerative colitis. Eur J Gastroenterol Hepatol.22: