Clinical Gastroenterology and Hepatology 2014;12:978–985
Endoscopic Assessment and Treating to Target Increase the Likelihood of Mucosal Healing in Patients With Crohn’s Disease Guillaume Bouguen,*,‡,§ Barrett G. Levesque,* Suresh Pola,* Elisabeth Evans,* and William J. Sandborn* *Division of Gastroenterology, University of California San Diego, La Jolla, California; ‡Department of Gastroenterology, University Hospital of Rennes, Pontchaillou, France; and §Inserm U991 University of Rennes 1, Rennes, France BACKGROUND & AIMS:
Mucosal healing has been proposed as a goal for treatment because it is associated with improved clinical outcomes of patients with Crohn’s disease (CD). However, little is known about the feasibility or probability of achieving mucosal healing in clinical practice. We evaluated the feasibility of treating patients to achieve mucosal healing based on endoscopic evaluation (treating to target).
METHODS:
We reviewed the endoscopic outcomes of 67 patients with CD who had lesions detected by endoscopy. Patients underwent 2 to 4 subsequent endoscopic evaluations at the University of California San Diego and were followed up from 2011 through 2012; data were collected on therapies and patient management. The cumulative incidences of mucosal healing and endoscopic improvement were estimated using the Kaplan–Meier method. Factors independently associated with mucosal healing were identified using a Cox proportional hazards model.
RESULTS:
After a median follow-up period of 62 weeks, 34 patients (50.7%) had mucosal healing and 41 patients (61.1%) had endoscopic improvement. The cumulative probabilities of mucosal healing were 12.7% and 45.0% at 24 and 52 weeks of treatment, respectively. Factors associated with mucosal healing were as follows: fewer than 26 weeks between endoscopic procedures (hazard ratio, 2.35; 95% confidence interval, 1.15–4.97; P [ .035) and adjustment to medical therapy when mucosal healing was not observed (hazard ratio, 4.28; 95% confidence interval, 1.9–11.5; P [ .0003).
CONCLUSIONS:
In an endoscopic study of patients with CD, we found that assessment of endoscopic disease activity and adjustments to medical therapy (treat to target) increase the likelihood of mucosal healing.
Keywords: Treat-to-Target; Prognostic Factor; Ulcer Detection; Inflammatory Bowel Disease.
rohn’s disease (CD) is a chronic but dynamic inflammatory disease of the gastrointestinal tract that leads to progressive bowel damage. Current evidence suggests that treatment strategies aimed at controlling symptoms do not alter the course of CD.1,2 The natural history of CD, as defined by population-based cohort studies from the past 30 years, is a progression from inflammatory disease to the development of complications such as stricture, fistula, abscess, and, ultimately, the requirement of surgery in a majority of patients.3,4 The medical management of CD and the treatment goals at that time, which were based exclusively on control of symptoms and relied heavily on corticosteroids, were unable to prevent bowel damage or slow the progressive destructive course of the disease in the majority of patients.5 Based on this unfavorable natural history, and the introduction of immunosuppressives and biologics into
C
clinical practice, new therapeutic goals and strategies that potentially could reduce complications and thus modify the course of the disease have emerged. Subsequently, the idea of mucosal healing (MH) as a treatment goal began to receive increasing attention after studies showed that it might improve long-term clinical outcomes. For example, among 130 cases in the Inflammatory Bowel Disease in South-Eastern Norway (IBSEN) population-based cohort of patients with newly diagnosed CD who had endoscopy follow-up evaluation, the
Abbreviations used in this paper: CD, Crohn’s disease; CI, confidence interval; CRP, C-reactive protein; HR, hazard ratio; IBD, inflammatory bowel disease; IQR, interquartile range; MH, mucosal healing; TNF, tumor necrosis factor; UCSD, University of California San Diego. © 2014 by the AGA Institute 1542-3565/$36.00 http://dx.doi.org/10.1016/j.cgh.2013.11.005
June 2014
absence of ulceration 1 year after the diagnosis was associated with less inflammatory activity (P ¼ .02), less corticosteroid use, and a trend toward fewer surgeries (P ¼ .1).6 In another study of a cohort of patients with CD who were starting infliximab, the absence of ulceration or the presence of clear endoscopic improvement after starting infliximab predicted fewer complications and a lower rate of hospitalizations and surgeries.7 In the ACCENT-I (a Crohn’s disease clinical trial evaluating infliximab in a new long-term treatment regimen in patients with Crohn’s disease) trial of infliximab as maintenance therapy for CD, a decrease in the rates of CDrelated hospitalizations and surgery was observed among patients who achieved MH (defined as no ulceration).8 In the step-up/top-down trial, MH at 2 years (defined as a simple endoscopic severity of CD score of 0) was predictive of sustained clinical- and steroid-free remission through years 3 and 4.9 Finally, a decision analysis model that explored treatment strategies for the management of moderate-to-severe CD with infliximab showed that MH as an end point was a cost-effective strategy as compared with a strategy based on clinical symptoms.10 Although the idea of MH as a treatment goal or outcome measure increasingly is being considered in patients with CD, at present there is no study that shows the feasibility and effectiveness of using a medical therapy algorithm to reach the goal of MH, a so-called treatto-target strategy. We therefore evaluated the feasibility of treat to target according to endoscopic disease activity to reach MH in clinical practice.
Methods Study Population All medical records of consecutive patients referred to the Inflammatory Bowel Disease (IBD) center at the University of California San Diego (UCSD) between January 2011 and December 2012 were reviewed. The International Classification of Diseases, 9th revision, codes were used to identify all patients with a diagnosis of CD or regional enteritis who were referred to UCSD. The diagnosis of CD then was confirmed based on radiologic, endoscopic, and/or histologic evidence. Only patients with CD who had ulcers seen at the initial endoscopic procedure and who had at least 2 consecutive endoscopic procedures performed during the study period were included. The following demographic and clinical characteristics were abstracted from the electronic medical records of the patients: sex, birth date, age at diagnosis, smoking status, CD phenotype at diagnosis according to the Montreal classification,11 previous history of surgery and hospitalization, and previous and concurrent treatment for CD. At the time of each endoscopic procedure, the medical records were reviewed for the type and findings of the
Treat to Target in Crohn’s Disease
979
procedure, medical therapies being used at the beginning of the study period, any adjustments of medical therapy after endoscopic procedures were performed, and the presence of clinical symptoms at the time of each endoscopic procedure and within 3 to 6 months after each endoscopic procedure. Adjustments of medical therapy after endoscopic procedures were defined as follows: the introduction or switch of immunosuppressives; the introduction, optimization, or switch within the class or out of the class of biologics; or changes in both immunosuppressives and biologics. Patients were defined as having clinical symptoms if they had symptoms of diarrhea and/or abdominal cramping. Endoscopic disease activity was defined by the presence of at least superficial ulcers (including aphthous ulceration) in any segment of the gastrointestinal tract at the time of the endoscopic procedure. To perform sensitivity analyses, we also analyzed separately the presence of deep ulcers and superficial ulcers. Endoscopic lesions were identified by a review of the findings of standardized endoscopy reports, which routinely included color photographs. Of note, the second and subsequent endoscopic assessments were usually a priori planned within 6 months by the treating physician to assess response to therapy (ie, scheduled endoscopic follow-up evaluation to assess for MH). This study was approved by the UCSD Institutional Review Board.
Statistical Analysis Quantitative variables were described as medians and percentiles (interquartile range [IQR], 25%–75%). Categoric variables were presented as the number and percentage of the cohort. Two events were defined: (1) MH was defined as the absence of any ulcers in any segment of the gastrointestinal tract during the endoscopic procedure, and (2) endoscopic improvement was defined by the downgrading of deep ulcers to superficial ulcers or the disappearance of superficial ulcers. Patients who underwent surgery during the study period were classified as treatment failures. The events were analyzed using survival analysis. The cumulative probabilities of MH and endoscopic improvement were estimated using the Kaplan–Meier method. For both, the time to MH or improvement was considered to begin at the date of the first endoscopic assessment and end at the last known follow-up evaluation or at the date of the first procedure during which MH/improvement was observed. To identify factors predictive of each event, we performed a univariate analysis using the log-rank test. We performed univariate analysis to identify the predictors of the outcomes described earlier (characteristics of patients at referral) and the factors associated with outcomes (factors related to management of patients after referral). We then performed a multivariate analysis to identify the independent predictors and factors associated with MH and
980
Bouguen et al
Clinical Gastroenterology and Hepatology Vol. 12, No. 6
endoscopic improvement: all variables with P values of less than .05 in the log-rank test were retained and integrated into a Cox proportional hazards regression model. The results are shown as hazard ratios (HRs) with 95% confidence intervals (CIs). Statistical analyses were performed using JMP Pro 10.0.0 software (Cary, NC).
Results Characteristics of the Population at Referral Of 510 patients with CD seen at UCSD during the study period, 199 patients were referred to the IBD center, leading to 171 endoscopic procedures performed at the time of referral. A total of 110 patients had at least 2 consecutive endoscopic procedures performed during the study period. Among those 110 patients, 67 patients who had ulcers seen at the initial endoscopic procedure constituted the study population. The baseline characteristics of these 67 patients are listed in Table 1. Of these patients, 32 (47.7%) were male. The median age at referral was 39 years (IQR75, 26.6–53 y) and the median disease duration was 9.8 Table 1. Baseline Characteristics Variable Sex, M:F Active smoker, n (%) BMI, median (IQR75) Median disease duration, y (IQR75) Familial history of IBD, n (%) Age at diagnosis, y (IQR75) Age at referral, y (IQR75) Montreal classification at CD diagnosis, n (%) L1: ileal L2: colonic L3: ileocolonic L4: upper digestive tract P: perianal lesion B1: nonpenetrating nonstricturing B2: stricturing B3: penetrating Disease outcomes before baseline, n (%) Extraintestinal manifestation History of hospitalization Penetrating disease Stricturing disease Anal fistula Previous treatments, n (%) Prior surgery Prior medical treatment Steroid Immunosuppressives (AZA/6MP or MTX) TNFa antagonists Treatment at referral Steroid Immunosuppressives (AZA/6MP or MTX) TNFa antagonists
N ¼ 67 5 24.4 9.8 13 25 39
32:35 (7.4) (22.3–26.7) (1.2–21.12) (19.4) (18.9–32.5) (26.6–53)
16 33 18 7 7 63 1 3
(23.8) (49.2) (26.8) (10.4) (10.4) (94) (1.4) (4.4)
13 37 9 23 15
(19.4) (55.2) (13.4) (34.3) (22.8)
26 (38.8) 63 (94) 34 (50.7) 32 (47.7) 16 (23.8) 16 (23.8) 17 (25.3)
AZA, azathioprine; BMI, body mass index; 6MP, 6-mercaptopurine; MTX, methotrexate.
years (IQR75, 1.2–21.1 y). According to the Montreal classification, 49.2% of the patients had isolated colonic disease at the time of diagnosis. Before referral, 47.7% of the patients had experienced disease complications (defined as the occurrence of stricturing or luminal fistulizing disease behavior) and 22.8% of the patients had experienced perianal fistulizing disease. Baseline endoscopic assessment showed deep ulcers in 23 patients (34.3%) and superficial ulcers in 44 patients (65.6%). Only 26 patients (38.8%) were naive to both immunosuppressives and biologics at referral. Thiopurine analogs were the main previous immunosuppressive and were used in 31 patients (42.3%). Among 32 patients (47.8%) who previously were treated with tumor necrosis factor (TNF) antagonists, 14 patients, 16 patients, and 2 patients were treated with 1, 2, and 3 different TNF antagonists, respectively. Medical therapies used at the time of referral are listed in Table 1.
Endoscopic Procedures and Outcomes After a median follow-up period of 62 weeks (IQR75, 44–79 wk), a total of 161 endoscopic procedures were performed, including 151 ileocolonoscopies or flexible sigmoidoscopies, 41 esophagogastroduodenoscopies, and 2 double-balloon enteroscopies. According to inclusion criteria, all 67 patients had at least 2 consecutive endoscopic procedures, 21 patients (31%) underwent 3 consecutive endoscopic procedures, and 6 patients (9%) underwent 4 consecutive endoscopic procedures. All endoscopies were performed during the follow-up period for each patient by the same endoscopist except for the 2 double-balloon enteroscopies. The overall median time between the 2 endoscopic procedures assessments was 24 weeks (IQR75, 17–38 wk). Of the 67 patients, 13 (19.4%), 28 (41.8%), and 34 (50.7%) achieved MH at 24 weeks, 52 weeks, and at the end of the follow-up period, respectively, and 15 (22.4%), 33 (49.2%), and 41 (61.1%) patients showed a clear improvement, with disappearance of baseline deep ulcers or baseline superficial ulcers at 24 weeks, 52 weeks, and at the end of the follow-up period, respectively. The cumulative probabilities of MH, defined by no ulcers at weeks 24 and 56 from the time of initial endoscopic procedure, were 12.7% and 45%, respectively (Figure 1). The cumulative probabilities of endoscopic improvement were 15.4% and 54% at 24 and 56 weeks, respectively. Overall, MH was described in 44 of the 161 endoscopic procedures. At the time of the procedure, patients experienced significantly less clinical symptoms when MH was achieved (n ¼ 26 of 44; 59.1%; 95% CI, 44.4–73.3) as compared with those with persistent endoscopic disease activity (n ¼ 22 of 117; 18.8%; 95% CI, 12.7–26.8; P < .0001). Of note, 40.9% (n ¼ 18 of 44) of patients had clinical symptoms despite MH. The C-reactive protein (CRP) concentration at the time of the endoscopic procedure was available for 108 procedures.
June 2014
Figure 1. Cumulative probability of achieving MH.
Overall, the median CRP concentration at baseline of 7.25 mg/L (range, 0.6–213 mg/L) decreased to 2 mg/L (range, 1–111 mg/L) at the last follow-up evaluation. With MH, the median CRP concentration (1 mg/L; IQR75, 1–2 mg/L) was significantly lower than for patients without MH (9 mg/L; IQR75, 3–27 mg/L) (P < .0001). None of the patients with MH had a CRP concentration greater than 8 mg/L.
Baseline Predictors of Endoscopic Outcomes For predictors of endoscopic outcomes, all baseline factors were evaluated in univariate analysis using the Mantel–Cox log-rank test (Table 2). CD duration less than 2 years (HR, 2.3; 95% CI, 1.1–4.7), as well as female sex (HR, 2.14; 95% CI, 1.1–4.4), were associated with MH. Conversely, previous immunosuppressive use (HR, 0.43; 95% CI, 0.2–0.9), and previous abdominal surgery (HR, 0.32; 95% CI, 0.1–0.7), were associated negatively with MH. By using multivariate analysis, none of the baseline factors independently predicted MH (Table 2). The same baseline factors were analyzed to identify predictors of endoscopic improvement. Previous abdominal surgery (HR, 0.38; 95% CI, 0.16–0.78; P ¼ .01) and a history of bowel stricture (HR, 0.4; 95% CI, 0.15–0.91; P ¼ .04) predicted a lower probability of achieving endoscopic improvement by univariate analysis. None of the baseline characteristics independently predicted MH using a Cox model for multivariate analysis.
Medical Therapy Adjustments During the Study Period After the endoscopic procedures, a total of 72 adjustments in medical therapy were performed as a result of finding ulcers at endoscopy (Figure 2), and 49 endoscopic procedures were not followed by adjustments in medical therapy despite the presence of ulcers. From an individual perspective, 27 of the 34 patients who
Treat to Target in Crohn’s Disease
981
achieved MH had one or more adjustments in medical therapy after endoscopic procedures until MH was achieved (ie, they were treated to a target of MH). Conversely, 26 of the 33 patients without MH at the end of the follow-up period did not have adjustments in medical therapy despite the findings of ulcers at an endoscopy (ie, they were not treated to a target of MH). Of the 72 adjustments in medical therapy, 9 (12.5%) were made in the absence of clinical symptoms. Overall, during the study period, TNF antagonists were introduced in 37 patients, optimized with dose escalation or interval shortening in 17 patients, and switched within the class in 9 patients. Seven patients underwent an out-of-class switch in biologic therapy to ustekinumab. Immunosuppressive therapy was initiated after referral in 35 patients and 1 patient was switched from azathioprine to methotrexate. A total of 8 patients underwent surgery during the follow-up period. At the last follow-up evaluation, 52 patients were treated with biologics, 45 patients were treated with immunosuppressives, including 39 patients receiving combination therapy, and 18 patients were treated with steroids (prednisone or budesonide).
Factors Associated With Endoscopic Outcomes We then assessed factors associated with MH during the follow-up period. Individually, biologics or immunosuppressive treatment modifications were not associated with MH. However, repeated endoscopic procedures within 26 weeks from the previous endoscopic procedure and any adjustments in medical therapy after each endoscopic procedure in the setting of persistent ulcers were associated with MH (Table 3 and Figure 3). With regard to endoscopic improvement, similar findings were observed. In multivariate analysis, endoscopic procedures within 26 weeks (HR, 2.21; 95% CI, 1.16–4.26; P ¼ .016) and any adjustment in medical therapy after each endoscopic procedure in the case of persistent ulcers (HR, 2.35; 95% CI, 1.2–4.94; P ¼ .012) were associated with endoscopic improvement. When the predictors and factors associated with medical management linked to MH by univariate analysis were included in a multivariate analysis, only adjustment of medical therapy was associated with MH (HR, 3.59; 95% CI, 1.41–10.4; P ¼ .0064). For endoscopic improvement, multivariate analysis showed that both adjustment of medical therapy (HR, 2.08; 95% CI, 1.04–4.50; P ¼ .037) and endoscopic procedures within 26 weeks (HR, 1.95; 95% CI, 1.02–3.81, P ¼ .043) were significant predictors.
Discussion This study specifically investigated the feasibility of a treat-to-target approach in patients with CD, in which the
982
Bouguen et al
Clinical Gastroenterology and Hepatology Vol. 12, No. 6
Table 2. Predictors of MH by Univariate Analysis (Log-Rank Test) and by Multivariate Analysis (Cox Model) Univariate analysis
Multivariate analysis
Baseline factors
P
HR (95% CI)
P
HR (95% CI)
Sex, female Active smoking BMI,
Endoscopic Assessment and Treating to Target Increase the Likelihood of Mucosal Healing in Patients With Crohn’s Disease Guillaume Bouguen,*,‡,§ Barrett G. Levesque,* Suresh Pola,* Elisabeth Evans,* and William J. Sandborn* *Division of Gastroenterology, University of California San Diego, La Jolla, California; ‡Department of Gastroenterology, University Hospital of Rennes, Pontchaillou, France; and §Inserm U991 University of Rennes 1, Rennes, France BACKGROUND & AIMS:
Mucosal healing has been proposed as a goal for treatment because it is associated with improved clinical outcomes of patients with Crohn’s disease (CD). However, little is known about the feasibility or probability of achieving mucosal healing in clinical practice. We evaluated the feasibility of treating patients to achieve mucosal healing based on endoscopic evaluation (treating to target).
METHODS:
We reviewed the endoscopic outcomes of 67 patients with CD who had lesions detected by endoscopy. Patients underwent 2 to 4 subsequent endoscopic evaluations at the University of California San Diego and were followed up from 2011 through 2012; data were collected on therapies and patient management. The cumulative incidences of mucosal healing and endoscopic improvement were estimated using the Kaplan–Meier method. Factors independently associated with mucosal healing were identified using a Cox proportional hazards model.
RESULTS:
After a median follow-up period of 62 weeks, 34 patients (50.7%) had mucosal healing and 41 patients (61.1%) had endoscopic improvement. The cumulative probabilities of mucosal healing were 12.7% and 45.0% at 24 and 52 weeks of treatment, respectively. Factors associated with mucosal healing were as follows: fewer than 26 weeks between endoscopic procedures (hazard ratio, 2.35; 95% confidence interval, 1.15–4.97; P [ .035) and adjustment to medical therapy when mucosal healing was not observed (hazard ratio, 4.28; 95% confidence interval, 1.9–11.5; P [ .0003).
CONCLUSIONS:
In an endoscopic study of patients with CD, we found that assessment of endoscopic disease activity and adjustments to medical therapy (treat to target) increase the likelihood of mucosal healing.
Keywords: Treat-to-Target; Prognostic Factor; Ulcer Detection; Inflammatory Bowel Disease.
rohn’s disease (CD) is a chronic but dynamic inflammatory disease of the gastrointestinal tract that leads to progressive bowel damage. Current evidence suggests that treatment strategies aimed at controlling symptoms do not alter the course of CD.1,2 The natural history of CD, as defined by population-based cohort studies from the past 30 years, is a progression from inflammatory disease to the development of complications such as stricture, fistula, abscess, and, ultimately, the requirement of surgery in a majority of patients.3,4 The medical management of CD and the treatment goals at that time, which were based exclusively on control of symptoms and relied heavily on corticosteroids, were unable to prevent bowel damage or slow the progressive destructive course of the disease in the majority of patients.5 Based on this unfavorable natural history, and the introduction of immunosuppressives and biologics into
C
clinical practice, new therapeutic goals and strategies that potentially could reduce complications and thus modify the course of the disease have emerged. Subsequently, the idea of mucosal healing (MH) as a treatment goal began to receive increasing attention after studies showed that it might improve long-term clinical outcomes. For example, among 130 cases in the Inflammatory Bowel Disease in South-Eastern Norway (IBSEN) population-based cohort of patients with newly diagnosed CD who had endoscopy follow-up evaluation, the
Abbreviations used in this paper: CD, Crohn’s disease; CI, confidence interval; CRP, C-reactive protein; HR, hazard ratio; IBD, inflammatory bowel disease; IQR, interquartile range; MH, mucosal healing; TNF, tumor necrosis factor; UCSD, University of California San Diego. © 2014 by the AGA Institute 1542-3565/$36.00 http://dx.doi.org/10.1016/j.cgh.2013.11.005
June 2014
absence of ulceration 1 year after the diagnosis was associated with less inflammatory activity (P ¼ .02), less corticosteroid use, and a trend toward fewer surgeries (P ¼ .1).6 In another study of a cohort of patients with CD who were starting infliximab, the absence of ulceration or the presence of clear endoscopic improvement after starting infliximab predicted fewer complications and a lower rate of hospitalizations and surgeries.7 In the ACCENT-I (a Crohn’s disease clinical trial evaluating infliximab in a new long-term treatment regimen in patients with Crohn’s disease) trial of infliximab as maintenance therapy for CD, a decrease in the rates of CDrelated hospitalizations and surgery was observed among patients who achieved MH (defined as no ulceration).8 In the step-up/top-down trial, MH at 2 years (defined as a simple endoscopic severity of CD score of 0) was predictive of sustained clinical- and steroid-free remission through years 3 and 4.9 Finally, a decision analysis model that explored treatment strategies for the management of moderate-to-severe CD with infliximab showed that MH as an end point was a cost-effective strategy as compared with a strategy based on clinical symptoms.10 Although the idea of MH as a treatment goal or outcome measure increasingly is being considered in patients with CD, at present there is no study that shows the feasibility and effectiveness of using a medical therapy algorithm to reach the goal of MH, a so-called treatto-target strategy. We therefore evaluated the feasibility of treat to target according to endoscopic disease activity to reach MH in clinical practice.
Methods Study Population All medical records of consecutive patients referred to the Inflammatory Bowel Disease (IBD) center at the University of California San Diego (UCSD) between January 2011 and December 2012 were reviewed. The International Classification of Diseases, 9th revision, codes were used to identify all patients with a diagnosis of CD or regional enteritis who were referred to UCSD. The diagnosis of CD then was confirmed based on radiologic, endoscopic, and/or histologic evidence. Only patients with CD who had ulcers seen at the initial endoscopic procedure and who had at least 2 consecutive endoscopic procedures performed during the study period were included. The following demographic and clinical characteristics were abstracted from the electronic medical records of the patients: sex, birth date, age at diagnosis, smoking status, CD phenotype at diagnosis according to the Montreal classification,11 previous history of surgery and hospitalization, and previous and concurrent treatment for CD. At the time of each endoscopic procedure, the medical records were reviewed for the type and findings of the
Treat to Target in Crohn’s Disease
979
procedure, medical therapies being used at the beginning of the study period, any adjustments of medical therapy after endoscopic procedures were performed, and the presence of clinical symptoms at the time of each endoscopic procedure and within 3 to 6 months after each endoscopic procedure. Adjustments of medical therapy after endoscopic procedures were defined as follows: the introduction or switch of immunosuppressives; the introduction, optimization, or switch within the class or out of the class of biologics; or changes in both immunosuppressives and biologics. Patients were defined as having clinical symptoms if they had symptoms of diarrhea and/or abdominal cramping. Endoscopic disease activity was defined by the presence of at least superficial ulcers (including aphthous ulceration) in any segment of the gastrointestinal tract at the time of the endoscopic procedure. To perform sensitivity analyses, we also analyzed separately the presence of deep ulcers and superficial ulcers. Endoscopic lesions were identified by a review of the findings of standardized endoscopy reports, which routinely included color photographs. Of note, the second and subsequent endoscopic assessments were usually a priori planned within 6 months by the treating physician to assess response to therapy (ie, scheduled endoscopic follow-up evaluation to assess for MH). This study was approved by the UCSD Institutional Review Board.
Statistical Analysis Quantitative variables were described as medians and percentiles (interquartile range [IQR], 25%–75%). Categoric variables were presented as the number and percentage of the cohort. Two events were defined: (1) MH was defined as the absence of any ulcers in any segment of the gastrointestinal tract during the endoscopic procedure, and (2) endoscopic improvement was defined by the downgrading of deep ulcers to superficial ulcers or the disappearance of superficial ulcers. Patients who underwent surgery during the study period were classified as treatment failures. The events were analyzed using survival analysis. The cumulative probabilities of MH and endoscopic improvement were estimated using the Kaplan–Meier method. For both, the time to MH or improvement was considered to begin at the date of the first endoscopic assessment and end at the last known follow-up evaluation or at the date of the first procedure during which MH/improvement was observed. To identify factors predictive of each event, we performed a univariate analysis using the log-rank test. We performed univariate analysis to identify the predictors of the outcomes described earlier (characteristics of patients at referral) and the factors associated with outcomes (factors related to management of patients after referral). We then performed a multivariate analysis to identify the independent predictors and factors associated with MH and
980
Bouguen et al
Clinical Gastroenterology and Hepatology Vol. 12, No. 6
endoscopic improvement: all variables with P values of less than .05 in the log-rank test were retained and integrated into a Cox proportional hazards regression model. The results are shown as hazard ratios (HRs) with 95% confidence intervals (CIs). Statistical analyses were performed using JMP Pro 10.0.0 software (Cary, NC).
Results Characteristics of the Population at Referral Of 510 patients with CD seen at UCSD during the study period, 199 patients were referred to the IBD center, leading to 171 endoscopic procedures performed at the time of referral. A total of 110 patients had at least 2 consecutive endoscopic procedures performed during the study period. Among those 110 patients, 67 patients who had ulcers seen at the initial endoscopic procedure constituted the study population. The baseline characteristics of these 67 patients are listed in Table 1. Of these patients, 32 (47.7%) were male. The median age at referral was 39 years (IQR75, 26.6–53 y) and the median disease duration was 9.8 Table 1. Baseline Characteristics Variable Sex, M:F Active smoker, n (%) BMI, median (IQR75) Median disease duration, y (IQR75) Familial history of IBD, n (%) Age at diagnosis, y (IQR75) Age at referral, y (IQR75) Montreal classification at CD diagnosis, n (%) L1: ileal L2: colonic L3: ileocolonic L4: upper digestive tract P: perianal lesion B1: nonpenetrating nonstricturing B2: stricturing B3: penetrating Disease outcomes before baseline, n (%) Extraintestinal manifestation History of hospitalization Penetrating disease Stricturing disease Anal fistula Previous treatments, n (%) Prior surgery Prior medical treatment Steroid Immunosuppressives (AZA/6MP or MTX) TNFa antagonists Treatment at referral Steroid Immunosuppressives (AZA/6MP or MTX) TNFa antagonists
N ¼ 67 5 24.4 9.8 13 25 39
32:35 (7.4) (22.3–26.7) (1.2–21.12) (19.4) (18.9–32.5) (26.6–53)
16 33 18 7 7 63 1 3
(23.8) (49.2) (26.8) (10.4) (10.4) (94) (1.4) (4.4)
13 37 9 23 15
(19.4) (55.2) (13.4) (34.3) (22.8)
26 (38.8) 63 (94) 34 (50.7) 32 (47.7) 16 (23.8) 16 (23.8) 17 (25.3)
AZA, azathioprine; BMI, body mass index; 6MP, 6-mercaptopurine; MTX, methotrexate.
years (IQR75, 1.2–21.1 y). According to the Montreal classification, 49.2% of the patients had isolated colonic disease at the time of diagnosis. Before referral, 47.7% of the patients had experienced disease complications (defined as the occurrence of stricturing or luminal fistulizing disease behavior) and 22.8% of the patients had experienced perianal fistulizing disease. Baseline endoscopic assessment showed deep ulcers in 23 patients (34.3%) and superficial ulcers in 44 patients (65.6%). Only 26 patients (38.8%) were naive to both immunosuppressives and biologics at referral. Thiopurine analogs were the main previous immunosuppressive and were used in 31 patients (42.3%). Among 32 patients (47.8%) who previously were treated with tumor necrosis factor (TNF) antagonists, 14 patients, 16 patients, and 2 patients were treated with 1, 2, and 3 different TNF antagonists, respectively. Medical therapies used at the time of referral are listed in Table 1.
Endoscopic Procedures and Outcomes After a median follow-up period of 62 weeks (IQR75, 44–79 wk), a total of 161 endoscopic procedures were performed, including 151 ileocolonoscopies or flexible sigmoidoscopies, 41 esophagogastroduodenoscopies, and 2 double-balloon enteroscopies. According to inclusion criteria, all 67 patients had at least 2 consecutive endoscopic procedures, 21 patients (31%) underwent 3 consecutive endoscopic procedures, and 6 patients (9%) underwent 4 consecutive endoscopic procedures. All endoscopies were performed during the follow-up period for each patient by the same endoscopist except for the 2 double-balloon enteroscopies. The overall median time between the 2 endoscopic procedures assessments was 24 weeks (IQR75, 17–38 wk). Of the 67 patients, 13 (19.4%), 28 (41.8%), and 34 (50.7%) achieved MH at 24 weeks, 52 weeks, and at the end of the follow-up period, respectively, and 15 (22.4%), 33 (49.2%), and 41 (61.1%) patients showed a clear improvement, with disappearance of baseline deep ulcers or baseline superficial ulcers at 24 weeks, 52 weeks, and at the end of the follow-up period, respectively. The cumulative probabilities of MH, defined by no ulcers at weeks 24 and 56 from the time of initial endoscopic procedure, were 12.7% and 45%, respectively (Figure 1). The cumulative probabilities of endoscopic improvement were 15.4% and 54% at 24 and 56 weeks, respectively. Overall, MH was described in 44 of the 161 endoscopic procedures. At the time of the procedure, patients experienced significantly less clinical symptoms when MH was achieved (n ¼ 26 of 44; 59.1%; 95% CI, 44.4–73.3) as compared with those with persistent endoscopic disease activity (n ¼ 22 of 117; 18.8%; 95% CI, 12.7–26.8; P < .0001). Of note, 40.9% (n ¼ 18 of 44) of patients had clinical symptoms despite MH. The C-reactive protein (CRP) concentration at the time of the endoscopic procedure was available for 108 procedures.
June 2014
Figure 1. Cumulative probability of achieving MH.
Overall, the median CRP concentration at baseline of 7.25 mg/L (range, 0.6–213 mg/L) decreased to 2 mg/L (range, 1–111 mg/L) at the last follow-up evaluation. With MH, the median CRP concentration (1 mg/L; IQR75, 1–2 mg/L) was significantly lower than for patients without MH (9 mg/L; IQR75, 3–27 mg/L) (P < .0001). None of the patients with MH had a CRP concentration greater than 8 mg/L.
Baseline Predictors of Endoscopic Outcomes For predictors of endoscopic outcomes, all baseline factors were evaluated in univariate analysis using the Mantel–Cox log-rank test (Table 2). CD duration less than 2 years (HR, 2.3; 95% CI, 1.1–4.7), as well as female sex (HR, 2.14; 95% CI, 1.1–4.4), were associated with MH. Conversely, previous immunosuppressive use (HR, 0.43; 95% CI, 0.2–0.9), and previous abdominal surgery (HR, 0.32; 95% CI, 0.1–0.7), were associated negatively with MH. By using multivariate analysis, none of the baseline factors independently predicted MH (Table 2). The same baseline factors were analyzed to identify predictors of endoscopic improvement. Previous abdominal surgery (HR, 0.38; 95% CI, 0.16–0.78; P ¼ .01) and a history of bowel stricture (HR, 0.4; 95% CI, 0.15–0.91; P ¼ .04) predicted a lower probability of achieving endoscopic improvement by univariate analysis. None of the baseline characteristics independently predicted MH using a Cox model for multivariate analysis.
Medical Therapy Adjustments During the Study Period After the endoscopic procedures, a total of 72 adjustments in medical therapy were performed as a result of finding ulcers at endoscopy (Figure 2), and 49 endoscopic procedures were not followed by adjustments in medical therapy despite the presence of ulcers. From an individual perspective, 27 of the 34 patients who
Treat to Target in Crohn’s Disease
981
achieved MH had one or more adjustments in medical therapy after endoscopic procedures until MH was achieved (ie, they were treated to a target of MH). Conversely, 26 of the 33 patients without MH at the end of the follow-up period did not have adjustments in medical therapy despite the findings of ulcers at an endoscopy (ie, they were not treated to a target of MH). Of the 72 adjustments in medical therapy, 9 (12.5%) were made in the absence of clinical symptoms. Overall, during the study period, TNF antagonists were introduced in 37 patients, optimized with dose escalation or interval shortening in 17 patients, and switched within the class in 9 patients. Seven patients underwent an out-of-class switch in biologic therapy to ustekinumab. Immunosuppressive therapy was initiated after referral in 35 patients and 1 patient was switched from azathioprine to methotrexate. A total of 8 patients underwent surgery during the follow-up period. At the last follow-up evaluation, 52 patients were treated with biologics, 45 patients were treated with immunosuppressives, including 39 patients receiving combination therapy, and 18 patients were treated with steroids (prednisone or budesonide).
Factors Associated With Endoscopic Outcomes We then assessed factors associated with MH during the follow-up period. Individually, biologics or immunosuppressive treatment modifications were not associated with MH. However, repeated endoscopic procedures within 26 weeks from the previous endoscopic procedure and any adjustments in medical therapy after each endoscopic procedure in the setting of persistent ulcers were associated with MH (Table 3 and Figure 3). With regard to endoscopic improvement, similar findings were observed. In multivariate analysis, endoscopic procedures within 26 weeks (HR, 2.21; 95% CI, 1.16–4.26; P ¼ .016) and any adjustment in medical therapy after each endoscopic procedure in the case of persistent ulcers (HR, 2.35; 95% CI, 1.2–4.94; P ¼ .012) were associated with endoscopic improvement. When the predictors and factors associated with medical management linked to MH by univariate analysis were included in a multivariate analysis, only adjustment of medical therapy was associated with MH (HR, 3.59; 95% CI, 1.41–10.4; P ¼ .0064). For endoscopic improvement, multivariate analysis showed that both adjustment of medical therapy (HR, 2.08; 95% CI, 1.04–4.50; P ¼ .037) and endoscopic procedures within 26 weeks (HR, 1.95; 95% CI, 1.02–3.81, P ¼ .043) were significant predictors.
Discussion This study specifically investigated the feasibility of a treat-to-target approach in patients with CD, in which the
982
Bouguen et al
Clinical Gastroenterology and Hepatology Vol. 12, No. 6
Table 2. Predictors of MH by Univariate Analysis (Log-Rank Test) and by Multivariate Analysis (Cox Model) Univariate analysis
Multivariate analysis
Baseline factors
P
HR (95% CI)
P
HR (95% CI)
Sex, female Active smoking BMI,
