INTRODUCTION
Crohn’s disease (CD) is a chronic, progressive, inflammatory disease of the bowel with manifestations involving any part of gastrointestinal (GI) tract [1]. The pathogenesis of CD involves an inflammatory damage to the intestine due to an aberrant immune response against the environment, primarily the dysbiotic gut microbiome, in genetically predisposed individuals. Diet, through its influence on the gut microbiome plays a major role in the pathogenesis of CD, and studies have correlated certain dietary constituents with the risk of CD onset as well as progression [2-5]. Dietary beliefs play an important role in disease perception, as many patients with CD make dietary changes which can lead to malnutrition [6]. Another important role of diet is its influence on the body composition, the dynamics of which varies with disease severity and duration [7,8]. The major treatment strategies in CD aim to arrest the aberrant immune response, but because of their nonspecific nature, these are limited by adverse effects, cost and limited efficacy. The adverse effects assume importance particularly with the ever increasing use of anti-tumor necrosis factor (TNF) agents for the management of CD in the current era [9]. Dietary manipulation would be a more physiological and safer strategy to manage the CD associated inflammation, with an additional advantage of improving the nutritional status of these patients.
There are various dietary strategies which have been tested for the management of CD, but the principle behind all these strategies involves the enrichment and depletion of anti-inflammatory and pro-inflammatory components respectively [3]. Exclusive enteral nutrition (EEN) is an extreme form of dietary manipulation, defined as the use of a complete liquid diet, with the exclusion of normal dietary components for a defined period of time, as a therapeutic measure to induce remission [10]. EEN in the form elemental diet, semi-elemental diet and or polymeric formula-based diet has been shown to be as effective as steroids in the management of pediatric CD [11-13]. Because of the side effects of steroids on growth and development, EEN is the recommended treatment for induction of remission in pediatric CD [14,15]. There is an emerging but limited evidence of EEN in adults. Recent studies have demonstrated its efficacy in inflammatory stricturing disease, in complicated disease like abscess, postoperative recurrence and in fistulizing variants of CD [16-18] although, the randomized controlled trials performed till date have found it inferior to steroids [19]. Presently, EEN is recommended only as an adjunctive treatment in adults with CD [20].
In view of limited evidence regarding use of EEN in adults with CD, we undertook this study to ascribe the role of EEN in the management of CD.
METHODS
1. Patients
Consecutive patients with CD who received EEN and were under follow-up at inflammatory bowel disease (IBD) clinic, All India Institute of Medical Sciences, New Delhi, India between October 2017 and May 2019 were included.
2. Study Design
It was a retrospective analysis of a prospectively maintained database of patients with CD who received EEN. The decision to start EEN was taken by physicians managing patients in IBD clinic, and the major indications were active disease despite being on immunosuppressive therapy, active disease and contra-indication to immunosuppressive therapy like abscess, active enterocutaneous fistula with features of sepsis, and as an adjunct to immunosuppressive therapy (for patients who required nutritional built up prior to administration of biologicals). Active CD was defined as Crohn’s Disease Activity Index (CDAI) > 150 [21], and the following treatment regimens were utilized: corticosteroids, immunomodulators like azathioprine and methotrexate and anti-TNF-α agents like infliximab/adalimumab. The disease location and behavior was classified as per Montreal classification [22]: B1-inflammatory disease; B2-stricturing disease, B3-fistulizing disease; L1-terminal ileal disease; L2-colonic disease; L3-ileocolonic disease. Complicated CD was defined as development of stricture, fistula or abscess or requirement of surgery in patients with CD [23]. Antibiotics were administered in patients who had evidence of active infection.
The following parameters were analyzed from the database of these patients: baseline clinical and demographic profile including age, sex, disease duration, and phenotype, type of treatment received prior to initiation of EEN, and laboratory parameters like hemoglobin and serum albumin, which were also assessed at 4 and 8 weeks. CDAI was assessed at baseline and at 4 and 8 weeks after starting EEN. Clinical response was defined as decline in CDAI by 70 points or more from the baseline [24]. Clinical remission was defined as absolute reduction of CDAI to less than 150 post-intervention [24]. All procedures performed in studies were in accordance with the ethical standards of the institutional ethics committee and in accordance with the 1975 Helsinki Declaration and its later amendments or comparable ethical standards (Ref No. IECPG-219/27.03.2019). Written informed consent was obtained from all included patients.
3. Intervention
The formulation used in our patients was both semi-elemental based (Peptamen; Nestle Health Science) and polymeric based (Fresubin; Fresenius Kabi India), and depended upon the availability and patient preference. The patients were instructed about the method of preparation of EEN and the calories and protein requirement were calculated according to standard formulae by the IBD dietician. The patients were also instructed about abstaining from any other food items during the duration of EEN intake. The dietician kept a track of patient intake, compliance, and tolerance to EEN, both for admitted patients (daily patient rounds) and out-patients (regular telephonic interviews and in-person interview at clinic visits). The amount of formula being given was recorded and any subsequent intolerance or other side effects were noted. The route of administration was per orally or through naso-enteric tube in all patients. The patients were followed at 4 and 8 weeks or in-between in case of exacerbation of disease activity or intolerance to EEN.
4. Outcomes
The primary outcome assessed was improvement in CDAI post EEN at 4 and at 8 weeks. Secondary outcomes were: (1) changes in hemoglobin and serum albumin post EEN institution at 4 and at 8 weeks. (2) Proportion of patients having clinical response and remission while on EEN assessed at 4 and at 8 weeks respectively. (3) Identification of predictors of clinical response to EEN. In addition, a subgroup analysis to compare outcomes was done with respect to type of EEN, previous treatment strategies and compare when EEN was used as a sole treatment modality.
5. Statistical Analysis
Baseline parameters with normal distribution were expressed as mean ± standard deviation, and those with skewed distribution were expressed using median (interquartile range). Qualitative data were presented as number (%). Comparison of disease response parameters like CDAI, hemoglobin and serum albumin pre and post EEN intervention was made using paired t-test or Wilcoxon rank sum test as appropriate. A test was considered as significant if P-value < 0.05. Change in CDAI over time was modelled using linear mixed effects model, keeping subject ID as random effects and time after initiating EEN as fixed effects. Significance of change over time was estimated. Profile plots of change in CDAI, hemoglobin and albumin over time was presented stratified according to disease behavior along with individual patient parameters. Cumulative clinical response rates were calculated and represented using Kaplan-Meier plot for overall group and were also stratified according to disease behavior. The data was entered using Microsoft Excel 2011 (Microsoft Corp., Redmond, WA, USA) and was analyzed using R studio (Boston, MA, USA). In addition to the base packages in R, ggplot2, lmer, cart, dplyr, and tidyr packages were used.
RESULTS
EEN was started in 35 patients of complicated CD during the study period of October 2017 to May 2019. In view of initial intolerance within first 7 days, 4 patients discontinued EEN and 31 patients were included for final analysis (Fig. 1).
1. Baseline Demographic and Clinical Characteristics
The baseline characteristics of the cohort is summarized in Table 1. Mean age of the cohort was 34 ± 16 years with a slight male preponderance (54.8%). Stricturing disease (58.1%) was the most common phenotype followed by penetrating (29.0%) and non-stricturing, non-penetrating disease (12.9%). Terminal ileal (38.7%) was the most common site of involvement followed by ileocolonic (32.3%), colonic (16.1%) and combined ileal and upper GI tract involvement (12.9%).
2. Indications of EEN
Semi-elemental based preparation of EEN was started in 23 patients (74.2%) whereas 8 patients (25.8%) were started on polymeric based preparation. The median duration of EEN administration was 4 weeks (interquartile range, 2-6 weeks). The most common indication of EEN was active disease despite being on immunosuppressive therapy (53.1%) followed by active disease in the presence of contraindications for immunosuppression (34.3%). Other patients were given EEN for build-up of nutrition in patients with active disease who required immunosuppression. EEN was started as an adjunct to steroids (19.3%), immunomodulators (9.7%), anti-TNF-α therapy (35.5%) while it was used as the only modality of treatment in 35.5% of patients.
3. Changes in Parameters Post EEN
The changes in parameters post EEN administration are summarized in Table 2. The primary outcome assessed by change in CDAI showed a significant improvement post administration of EEN at 4 weeks (baseline 290 [260-320] vs. 240 [180-280], P=0.001) and at 8 weeks (baseline 290 [260-320] vs. 186 [160-240], P=0.001) respectively (Fig. 2A). Hemoglobin showed a significant rise at 4 weeks (baseline 7.7 [6.5-8.0] g/dL vs. 9.2 [8.0-10.0] g/dL, P=0.001) and at 8 weeks (baseline 7.7 [6.5-8.0] g/dL vs. 9.8 [9.0-10.8] g/dL, P=0.001) (Table 2, Fig. 2B). Serum albumin also showed a significant improvement post administration of EEN at 4 weeks (baseline 2.5 [1.8-3.0] g/dL vs. 3.2 [2.8-3.4] g/dL, P=0.001) and at 8 weeks (baseline 2.5 [1.8-3.0] g/dL vs. 3.2 [3.0-3.6] g/dL, P=0.001) (Fig. 2C).
4. Response to EEN in Whole Cohort and in Subgroups
Clinical response (CDAI reduction > 70) to EEN was seen in 22 out of 31 patients at 8 weeks in the whole cohort (70.9%). The cumulative clinical response rates at 4 and 8 weeks were 37.3% and 80.4% respectively (Fig. 3). Five out of 31 patients (16.1%) achieved clinical remission (CDAI < 150) at 8 weeks. Among different phenotypes, 1 patient of B1 disease (25%), 4 patients of B2 disease (27.7%) and none of the patients in B3 disease could achieve clinical remission.
The efficacy of EEN across different subgroups and treatment arms was analyzed. The clinical response rates at 8 weeks across B1, B2, and B3 disease were 50%, 78.8%, and 100% respectively (log-rank test, P=0.093) (Fig. 4). EEN when used as the only treatment for CD (n = 11), showed significant improvement in CDAI at 4 weeks (280 [260-320] vs. 220 [180-240], P<0.050) and at 8 weeks (190 [170-220], P=0.050) (Supplementary Fig. 1). Clinical response was present in 7 patients (63.6%) while clinical remission was seen in 1 patient who were only on EEN without concurrent pharmacological therapy. The response rates at 8 weeks with polymeric and semi-elemental diet were 75% and 82.6% respectively (log-rank test, P=0.49) (Supplementary Fig. 2). In the subgroup analysis comparing different arms of treatment received, the response rates at 8 weeks in patients who previously received immunomodulators and anti-TNF therapy were 59% and 80% respectively.
5. Predictors of Clinical Response on EEN
A linear mixed effect model was created using variables like type of EEN, phenotype of disease, disease duration, disease location, baseline CDAI, baseline serum albumin and hemoglobin levels and EEN used alone/adjunct to other treatment modalities to analyze which factor remains predictive of clinical response with EEN in CD. In this model, on univariate analysis, only baseline CDAI (odds ratio, 1.008; 95% confidence interval, 1.002-1.017; P=0.046) could predict clinical response while all other variables were non-significant (Table 3).
6. Adverse Events
None of the patients developed serious adverse events. Overall, GI intolerance was the most common adverse event seen in 7 out of 35 patients (20%) (initial cohort) (Fig. 1). Of these, 4 patients discontinued EEN within 1st week and 3 additional patients stopped EEN after a median duration of 10 days. All these patients were given semi-elemental based preparation of EEN.
DISCUSSION
EEN as a therapy for CD was serendipitously discovered in 1973 by Voitk et al. [25] who observed that the patients given EEN for preoperative build up improved clinically and many could avoid surgery. This dual ability of EEN in inducing the clinical response as well as a supporting tool for nutritional built-up, makes it an important therapeutic option both as an adjunct or exclusively in the management of CD. This is especially true in the settings where other treatment options may be contraindicated (as in active infection with active disease) and in malnourished patients. EEN is already the first line therapy for induction of remission in pediatric CD [14] and has even demonstrated better efficacy than steroids in terms of improvements in growth and mucosal healing [26,27]. The evidence for adults is heterogeneous, with meta-analyses reporting it as inferior to steroids [28] and recent case series demonstrating improvement in complicated CD [29]. The treatment options for complicated CD in the developing countries become limited by the higher risks of infectious adverse events, especially when these patients are malnourished. The evidence and awareness for EEN in this setting is scarce, and we have found good results with the use of EEN in our patients for the past 18 months.
In the present study, there was a significant improvement in clinical activity and biochemical parameters post administration of EEN in patients with CD, most of whom had a complicated phenotype. More than 80% patients had either stricturing or fistulizing disease, > 90% were steroid experienced, 16.1% patients had received anti-TNF and 41.9% had undergone CD specific surgery in the past. Even in this cohort of patients, more than two-thirds experienced clinical response at 8 weeks. EEN was effective both as an adjunct and as well as the only treatment in inducing remission. Baseline disease activity was the most important factor predicting clinical response in these patients.
The results of the present study are in line with the recent studies evaluating the role of EEN in adults with CD. Wall et al. [18] in their pilot study found 2 weeks of EEN was effective in inducing clinical remission in adults with CD. Interestingly, the authors of this study demonstrated further treatment with either partial or EEN had comparable treatment outcomes. However, this study had a predominantly B1 disease (78.9%), unlike the present study, which could account for better outcomes even at 2 weeks. Moreover, the present study had a sicker subgroup of patients with 93% and 40% patients having received corticosteroids and anti-TNF agents, respectively, while in the study from Wall et al, no patient had received steroids and only 5.2% patients had received biologicals. In addition, duration of disease also differed with patients in our study having a longer 48 months (range, 18-72 months) disease duration compared with median of 7 months in this study.
Hu et al. [30] in their study of 59 adult CD patients with inflammatory strictures found EEN to be effective in inducing clinical remission in 64.6% of patients when given over 12 weeks, which was higher when compared with our study, where though the response rate was 78.8%, clinical remission was seen only in 27.7% in stricturing disease. Inclusion of only inflammatory strictures and longer duration of EEN administration (12 weeks) in this study could account for this difference in clinical remission. In the present study, patients with both inflammatory (44.4%) and fibrotic strictures (55.6%) were included and EEN was administered for median duration of 4 weeks (range, 2-6 weeks).
Enteral nutrition has also been used as an adjunct to infliximab in fistulizing CD. Wu et al. [31] found that infliximab in combination with EEN was more effective in management of fistulizing CD when compared with infliximab alone. In the present study, of the 9 patients with fistulizing CD, although none could achieve clinical remission (CDAI < 150), clinical response rates to EEN were 100%. The response rates did not differ when EEN was used exclusively or as an adjunct to anti-TNF agents.
EEN has also been found to be useful when used in patients who have lost response to anti-TNF, as evidenced by a recent study which demonstrated usefulness of EEN in patients on anti-TNF agents with loss of response (LOR) [32]. In the present study, out of 5 patients who had LOR on anti-TNF, 4 patients (80%) had clinical response (ΔCDAI > 70) at 8 weeks post EEN administration. A recent meta-analysis also supported the use of enteral nutrition to maintain response to anti-TNF molecules [33]. Restoration of increased intestinal permeability and decline in mucosal production of TNF-α could be one of the possible explanations for benefit of EEN in patients who had LOR to anti-TNF agents [34].
Importantly, the clinical response and remission rates with EEN were similar across different phenotypes of disease and could improve disease activity as the only treatment in patients who could not be started on any standard treatment of CD in view of its contraindications. Previous studies had shown different results about the efficacy of EEN in different subtypes of CD. Afzal et al. [35] in their study of 65 patients with pediatric CD found colonic disease to be less responsive than ileocolonic and isolated ileal disease, whereas Buchanan et al. [36] in their study of 110 patients found that 80% of their patients achieved remission irrespective of site of disease.
There is an emerging yet unclear role of EEN in the management of adults with CD. A systematic review showed EEN to be inferior to steroids in inducing remission in adults with CD [19]. A Cochrane review evaluating the overall impact of EEN and its comparison with steroids found steroids to be more beneficial when compared with EEN in adults [28]. However, the overall quality of evidence from the above conducted systematic reviews is low with significant heterogeneities among the studies that were analyzed. Prospective uncontrolled trials have shown benefit of EEN in adults across various subtypes of disease, in the presence of complicated disease (abscess) and reduction in postoperative recurrence/complications [16-18]. In the absence of well controlled randomized trials, the current guidelines do not recommend EEN in adults patients with CD [20] and highlights the need of further studies in this aspect.
One of the major limitations of EEN is its GI intolerance (diarrhea, bloating, and flatulence) and unpalatability. The compliance to EEN especially in semi-elemental preparation often limits its clinical effectiveness. Overall, 7 patients (20%) had intolerance in whom EEN had to be discontinued. All these patients were on semi-elemental preparation. Another important limitation of EEN is its cost. The average daily cost of semielemental based preparation and polymeric based preparation in India is Rs 1,200 (17 USD) and Rs 800 (11 USD) respectively. We could not perform a formal cost-effective analysis comparing different treatment options.
Our study had few limitations. First, is its retrospective design with a small sample size. Although we retrieved all data from a prospectively maintained database for IBD-CD, it is possible that we might have missed capturing some data. The sample size of the present study was small. However, EEN in adults with CD is a relatively newer concept which is still evolving. This study is the first to report the outcomes of EEN in CD from this region. Further, all the consecutive patients started on EEN were included in the study. Secondly, the criteria for activity assessment in CD was clinical using CDAI. The endoscopic disease activity and biochemical parameters like fecal calprotectin were not assessed in this study. Thirdly, the limited duration of EEN administration and a limited duration of follow-up. The limited number of patients showing clinical remission could have been because of the shorter duration of EEN administration as compared to previous studies. It is not known whether the clinical improvement seen post EEN could be sustained post discontinuation of the therapy. Fourth, EEN as the only modality of treatment was used in only 35.5% patients. Thus, the results of efficacy for EEN could have been affected by response to other medicines. However, as mentioned clinical response seen in EEN alone patients was comparable to those of the whole cohort. Further studies using a larger sample size are required to test the efficacy of EEN alone in management of adults with CD. Fifth, the improvement in the disease activity in patients with CD with active infections could have been because of antibiotics which were administered with EEN. However, for disease complicated with infections, evaluating the efficacy of EEN alone in this setting might be difficult as administration of antibiotics in this setting is the standard of care. Previous studies demonstrating effectiveness of EEN in this setting had also administered antibiotics [29,37]. Sixth, the benefit sustained with EEN might reflect the changes in nutritional status before and after treatment evaluated. As seen in pediatric CD, the effect of EEN might be due to both improvement in disease activity and improvement in nutritional status. The efficacy of EEN in this setting can be conclusively demonstrated only in the presence of a controlled group given a calorie and protein matched normal diet. This remains a major limitation of our study. Further, prospective studies with a larger sample size are required to validate these findings and to address these limitations.
To conclude, EEN is an effective therapeutic option for patients with complicated CD, especially when they are malnourished and have contraindications which preclude the use of immunosuppressive therapy. EEN is effective both as an adjunct and as only therapy, across all subtypes of CD, and the baseline disease activity remains the most important predictor of response to therapy.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
CrossRef TDM
Supplement1
