Enteric-Release Budesonide May Be Useful in the Management of Non-Responsive Celiac Disease

Abstract

Background:

Non-responsive celiac disease (NRCD) has many aetiologies, including gluten exposure. Budesonide may be used for refractory celiac disease (RCD) and celiac crisis.

Aims:

We reviewed the effectiveness of budesonide to induce clinical and histological response in NRCD with villous atrophy (VA).

Methods:

Case series of adult cases with NRCD and VA prescribed budesonide at two Celiac Centers. Clinical variables and mucosal recovery (i.e. normal villous architecture within 1 year of treatment) were evaluated.

Results:

Forty-two cases (77% female, median age 45.0 (IQR 28.3–60.0) years) were included. Most common symptoms were diarrhea (64%) and abdominal pain (62%). Budesonide was initiated at 9 mg (83%) for a median duration of 16.0 weeks (IQR 6.8–25.0 weeks). 57% exhibited a clinical response, positively associated with diarrhea (Adjusted OR 6.08 95CI 1.04–35.47) and negatively with fatigue (Adjusted OR 0.18 95CI 0.03–0.98). Clinical response was higher among those with dietitian counseling prior to budesonide (n=29, 70 vs 23% p<0.01). Mucosal recovery was observed in 11/24 with follow-up duodenal biopsies. There was no association between clinical response and mucosal recovery and 79% of clinical responders had a symptomatic relapse. RCD (48%) and chronic gluten exposure (24%) were the main suspected aetiologies of NRCD. Most individuals without a clinical response subsequently received an IBS-related diagnosis.

Conclusions:

Budesonide may be effective to induce clinical response in NRCD presenting with diarrhea and VA, but clinical recurrence and lack of mucosal recovery are frequent after tapering. Other diagnoses, including coexisting IBS may be considered in non-responders to budesonide therapy.

Introduction

Celiac disease (CD) is a common gastrointestinal condition, with a prevalence of 0.7 to 1.4% [1]. The treatment for CD - a gluten free diet (GFD) - generally induces clinical remission in a majority of patients [2]; however, 7 to 30% of CD patients may experience non-responsive celiac disease (NRCD), a condition defined by persistent symptoms or laboratory abnormalities despite being on a GFD for at least one year [3]. NRCD has many aetiologies, with gluten exposure being the most common as minimal exposure may perpetuate symptoms and immune activation in some patients [3,4]. Other recognized causes of NRCD include irritable bowel syndrome (IBS), microscopic colitis, disaccharidase deficiency, small intestinal bacterial overgrowth (SIBO) and refractory celiac disease (RCD) [3]. RCD type 1 is a diagnosis of exclusion reserved for persisting symptoms and villous atrophy despite a GFD in the absence of other causes of villous atrophy, whereas RCD type 2 is characterized by aberrant T cells showing T cell receptor gene rearrangement clonality [5].

Budesonide is a micronized synthetic corticosteroid with a high first pass effect that is effective in accelerating clinical remission among individuals presenting with celiac crisis [6], acute symptoms following gluten exposure [7] and among individuals with RCD type 1 and 2 [8,9]. Budesonide was also reported to be effective in combination with GFD to reduce stool weight and bowel movement frequency, as well as to increase overall well-being when taken twice daily for four weeks at the time of diagnosis in an Italian CD cohort [10]. Budesonide has been suggested as a second-line therapy for those who do not respond to a gluten contamination elimination diet (GCED) [11], yet detailed experience of the use of budesonide in the management algorithm of NRCD has not been reported. The primary aim of this study is to review the effectiveness of budesonide to induce a clinical response among CD patients with persisting symptoms and villous atrophy. A secondary aim is to assess the rate of mucosal recovery associated with the use of budesonide in these patients.

Material and Methods

Retrospective cohort study of patients with biopsy-confirmed CD who received enteric-release budesonide either at Beth Israel Deaconess Medical Center or Massachusetts General Hospital (Boston, USA) between June 2006 and January 2019. Subjects were identified through a clinical query of electronic medical records with the search terms “budesonide” and “celiac disease”. Charts were reviewed to confirm the diagnosis of CD and the indication for budesonide therapy. Only cases with biopsy-proven celiac disease on a GFD for at least 12 months who received budesonide for persisting celiac disease symptoms were examined. Those prescribed budesonide for coexisting microscopic colitis, eosinophilic gastrointestinal disorders, acute symptomatic gluten exposure [7], celiac crisis [6] or ulcerative jejunitis (UJ) were excluded. Included patients received enteric-release budesonide, either as coated pellets or a powder compounded into a gelatin capsule. They all had villous atrophy on duodenal biopsies or videocapsule endoscopy (VCE) performed within the year prior to initiation of budesonide. NRCD was defined as presenting with symptoms or signs suggestive of CD after having been on a GFD for at least 12 months [3]. Refractory celiac disease (RCD) was defined as persistent villous atrophy, with signs and symptoms of malabsorption after at least 12 months of GFD and for which other causes of NRCD were excluded [12]. Those who underwent a GCED before budesonide were identified. Our centers propose a GCED to some NRCD patients with Marsh 3 histology on follow-up biopsy after being on a GFD for 12 months irrespective of celiac antibody levels [11]. A GCED requires elimination of all processed foods, dairy products and grains other than rice for two weeks with stepwise reintroduction over a 12-week period[11]. Additionally, fructose, lactose and FODMAPs (fermentable oligo-, di-, monosaccharides and polyols) intolerance were assessed at the time of dietitian visit. Clinical response was defined as patient-reported clinical improvement. To be consistent with RCD literature, complete response was a complete resolution of symptoms as opposed to partial response or absence of response [9,13]. For instance, complete response of the two most frequent symptoms (diarrhea and abdominal pain) would mean having formed bowel movements and no abdominal pain, while partial response would mean the patient reported that at least one symptom improved but not all symptoms resolved completely. Clinical response was assessed by the treating physician while the patient was still on therapy, unless side effects appeared, and the patient self-discontinued budesonide.

Pathology reports of small intestinal biopsies performed during treatment or the year following budesonide therapy were used to assess mucosal recovery. Specifically, mucosal recovery was defined as absence of villous atrophy. Additionally, for cases without villous atrophy on duodenal biopsies but a videocapsule endoscopy (VCE) showing distal villous atrophy, mucosal recovery would be defined as a normal follow-up VCE exam [11]. Longitudinal medical history of each included case was reviewed through February 2019.

Statistical analysis

Continuous variables are presented as median and interquartile range (IQR). Dichotomous variables are expressed as frequencies and proportions. Mann-Whitney U tests were performed to compare continuous data, while Fisher’s exact tests were conducted for dichotomous data. Odds ratio and 95% confidence intervals were calculated to identify clinical characteristics that predict a clinical response to budesonide. A multivariate logistic regression model was computed to adjust for microscopic colitis, as well as predictors that were initially significant on the univariate analysis. Statistical significance was defined as a p-value < 0.05. Statistical analyses were performed using STATA 15.1IC (College Station, Texas, USA). The study was approved by the Institutional Review Boards of both centers.

Results

The search query identified 242 patients, of whom 42 individuals who had received budesonide met our inclusion criteria (Figure 1). Budesonide was prescribed on a case-by-case basis, because of patient distress regarding persisting symptoms and enteropathy and presumed strict adherence to the GFD.

Figure 1. Flow diagram.

^† Acute symptoms following a gluten exposure (n=14), celiac crisis (n=5), ulcerative jejunitis (n=5), treatment within the first year of GFD (n=10), symptoms without documented enteropathy (n=9), refractory celiac disease without recent duodenal biopsies before initiation of budesonide (n=6).

Clinical characteristics

The cohort was predominantly female (77%) with a median age 45.0 (IQR 28.3 – 60.0) years and median GFD duration of 2.0 (IQR 1.0–6.0) years (Table 1). Classical symptoms predominated, including diarrhea (64%). Fifteen (36%) had elevated tissue transglutaminase antibodies (tTG), 29 (69%) met with a celiac dietitian before initiating budesonide and 11 (26%) tried a GCED (3 of whom did not have a serologic response to their GCED)[11]. Recent duodenal biopsies showing Marsh 3 histology were available (n= 39) except for three patients who did not have villous atrophy on biopsies, but a videocapsule showing patchy villous atrophy. Four patients had both a videocapsule endoscopy and duodenal biopsies showing villous atrophy. Twenty-two (52%) had a colonoscopy to rule out active microscopic colitis. Seven patients had a history of microscopic colitis that was either in remission on recent colonic biopsies or judged not to be responsible for the clinical presentation (individuals presenting with bloating, abdominal pain). Additionally, 23 (55%) individuals had a breath test to assess for small intestinal bacterial overgrowth (SIBO) and it was diagnosed in nine individuals. Fructose, lactose and FODMAPs intolerance were assessed at the time of dietitian visit and were diagnosed in six individuals (Table 1). In these patients, symptoms persisted despite treatment of the conditions (SIBO, non-gluten food intolerances) and budesonide was subsequently offered.

Table 1:

Baseline Characteristics of the NRCD Cohort and Association with a Clinical Response to Budesonide

	NRCD n = 42	Clinical response to budesonide OR (95%CI)¤
Female, n (%)	32 (76)	7.0 (0.78–63.1)
Age at CD diagnosis, {years, median (IQR)}	41 (24.0–55.0)	1.01 (0.98–1.05)
Age at first budesonide, {years, median (IQR)}	45 (28.3–60)	1.01 (0.97–1.05)
Time on a GFD, {years, median (IQR)}	2 (1.0–6.0)	1.03 (0.84–1.26)
Previously known conditions
Irritable bowel syndrome, n (%)	5 (12)	0.12 (0.01–1.20)
Small intestinal bacterial overgrowth, n (%)	9 (21)	1.33 (0.28–6.39)
Non-gluten food intolerances, n (%)	6 (14)	1.02 (0.35–2.98)
Microscopic colitis, n (%)	7 (17)	0.80 (0.15–4.20)
Trisomy 21, n (%)	4 (10)	1.27 (0.11–15.38)
NRCD after initial celiac crisis, n (%)	2 (5)	-
Clinical presentation
Diarrhea, n (%)	27 (64)	4.50 (1.13–18.00) *
Abdominal pain, n (%)	26 (62)	0.43 (0.11–1.74)
Nausea / vomiting, n (%)	16 (38)	0.47 (0.12–1.80)
Bloating, n (%)	19 (45)	0.69 (0.19–2.54)
Weight loss, n (%)	16 (38)	2.33 (0.57–9.42)
Constipation, n (%)	5 (12)	0.38 (0.06–2.61)
Extraintestinal manifestations,† n (%)	10 (24)	0.10 (0.02–0.61) **
Fatigue, n (%)	17 (40)	0.12 (0.03–0.53) **
Elevated tissue transglutaminase antibodies, n (%)	15 (36)‡	2.18 (0.57–8.41)
Secondary non response to GFD¥, n (%)	16 (38)	0.63 (0.18–2.12)
Postpartum, n (%)	1 (2)	-
Interventions
Celiac dietician evaluation, n (%)	29 (69)	10.50 (2.14–51.53) **
Gluten contamination elimination diet, n (%)	11 (26)	0.67 (0.16–2.75)
Previous treatment,§ n (%)	25 (60)	1.11 (0.30–4.17)

CD: Celiac disease, GFD: Gluten-free diet, NRCD: Nonresponsive celiac disease

^†Brain fog, non-specific joint pain, headaches, rash (including one case of dermatitis herpetiformis), paresthesia,

^‡60% had a less than 3x ULN elevation, median (range) 2.2 (1.2–10.9) × ULN

^§5-Aminosalicylic Acid, pancreatic enzymes, probiotics, dicyclomine, fibers, laxatives, antibiotics (ciprofloxacin, metronidazole, rifaximin), prednisone, low FODMAP diet, proton pump inhibitors, cholestyramine, tricyclic antidepressants, loperamide, metoclopramide, budesonide for previous celiac crisis, thiopurine.

^¤OR of the association between clinical characteristics and a clinical response to budesonide compared to the subjects that did not have a clinical response

^¥Recurrence of symptoms after an initial response to the GFD [3]

^*p<0.05

^**p<0.01

Budesonide therapy

Budesonide was commonly initiated at a dose of 9 mg (83%) (Table 2 and Figure 2). Median treatment duration was 16.0 (IQR 6.8–25.0) weeks; 14 (33%) patients received budesonide for 12 to 16 weeks and 12 (29%) had a longer course. Fourteen (33%) patients were considered “budesonide-dependent”, necessitating a slower taper than initially planned; of these, eight had courses of more than 4 months. Otherwise, the duration of initial therapy was according to clinical response and physician’s assessment. In most cases, after assessment of a clinical response on 9mg daily, budesonide was tapered to 6mg daily then 3 mg daily (Figure 2). Six (14%) individuals were instructed to open the budesonide capsule. Overall, 24 (57%) had a clinical response to budesonide, of whom two received open-capsule budesonide, and 16 (38% of total cohort) had a complete resolution of the symptoms. Interestingly, individuals presenting with diarrhea were more likely to have a clinical response to budesonide (OR 4.50 95CI 1.13–18.00 p < 0.05), as opposed to those reporting extraintestinal manifestation (OR 0.10 95CI 0.02–0.61 p < 0.01) or fatigue (OR 0.12 95CI 0.03–0.53 p<0.01) (Table 3). In the multivariate analysis, the association between clinical response and diarrhea (OR 6.08 95% CI 1.04–35.47 p<0.05) and fatigue (OR 0.18 95CI 0.03–0.98 p < 0.05) were still significant (Table 3). Among the responders, 19 (79%) experienced a clinical recurrence after tapering budesonide and 13 (54%) had at least one additional course. There was no difference in terms of clinical response between the subjects who received other treatments at the same time as budesonide (Table 2) than those who only received budesonide (63 vs 50%). Additionally, 19 (45%) were eventually given another therapy, either because of non-response to budesonide (n=10), or an eventual recurrence of symptoms after tapering budesonide (n=9) (Table 2).

Table 2:

Budesonide Treatment and Mucosal Recovery among the NRCD Cohort

	NRCD cohort	MR	No MR
	n = 42	n = 11	n = 13
Treatment duration
Treatment duration in weeks median (IQR)	16.0 (6.8–25)	15 (5–28)	16 (12–44)
Treatment duration on 9mg in weeks median (IQR)	12 (5–16)	12 (7.5–19)	12 (10–18)
Treatment duration 3–4 months n (%)	134 (33)	5 (45)	6 (46)
More than 4 months n (%)	12 (29)	3 (27)	45 (38)
Less than 3 months n (%)	16 (38)	3 (27)	2 (15)
Posology
Initial dose 9 mg n (%)	35 (83)	10 (91)	9 (69)
Open capsule n (%)	6 (14)	1 (9)	2 (15)
Cotreatment † n (%)	23 (55)	5 (45)	9 (69)
Response
Clinical response n (%)¶	24 (57)	6 (55)	9 (69)
Clinical Recurrence after stopping budesonide among the responders n (%)	19/24 (79)	3/6 (50)	8/9 (89)
Budesonide-dependent n (%)	14 (33)	2 (18)	5 (38)
Side effects ‡ n (%)	24 (57)	6 (55)	8 (62)
Subsequent courses of budesonide n (%)	13 (31)	1 (9)	6 (46)
Switch to another therapy § n (%)	19 (45)	1 (9)*	7 (54)*

^*Fisher’s exact test between the subgroups with and without MR p<0.05

MR: Mucosal recovery, NRCD: Non-responsive celiac disease

^†Loperamide, lubiprostone, tegaserod, linaclotide, 5-Aminosalicylic Acid, cholestyramine, fibers, thiopurine, pancreatic enzymes, dicyclomine, proton pump inhibitor, gluten contamination elimination diet, bismuth, antibiotics (ciprofloxacin, metronidazole, rifaximin), amitriptyline, hyoscyamine, ondansetron

^‡Anxiety, constipation, irritability, agitation, insomnia, weight gain, headaches, acne, high blood pressure, vision changes, easy bruising, hirsutism, increased blood glucose levels, flushing, nausea, vomiting, muscular weakness, rash, flu-like symptoms.

^§Thiopurine, amitriptyline, low FODMAPS diet, antibiotics (vancomycin, rifaximin, ciprofloxacin, metronidazole) pancreatic enzymes, loperamide, bismuth, hyoscyamine, peppermint oil, PEG laxative, 5-Aminosalicylic Acid, prednisone, lubiprostone, gluten contamination elimination diet, octreotide, anti-TNF, methotrexate, omalizumab.

^¶We could not clearly assess clinical response for 3 subjects.

Figure 2: Initial budesonide dose and duration of treatment.

a) Repartition of initial dose among the cohort b) Duration on each dosing (median) 9 mg n = 35, 6mg n = 23, 3mg n = 19.

Table 3:

Predictors of Clinical Response to Budesonide among NRCD Patients

	Univariate analysis OR 95% Cl	Multivariate analysis† OR 95% Cl
Diarrhea	4.50 (1.13–18.00) *	6.08 (1.04–35.47) *
Fatigue	0.12 (0.03–0.53) **	0.18 (0.03–0.98) *
Extraintestinal manifestation	0.10 (0.02–0.61) **	0.15 (0.02–1.18)

^†Model including the covariates diarrhea, fatigue and extraintestinal manifestations and adjusting for microscopic colitis

^*p < 0.05

^**p <0.01

Side effects were reported by 57% of the cohort, with anxiety, weight gain, headaches and constipation being the most common. Budesonide was stopped because of adverse events in six patients (14%).

Aetiologies of NRCD

Following retrospective review of the medical course before and after budesonide therapy, the main aetiologies of the NRCD symptoms were RCD type 1 (48%) and suspected gluten contamination (24%) (Table 4). One case had a postpartum flare of symptoms which resolved rapidly with budesonide. There were nine (20%) patients with enteropathy at the time of NRCD diagnosis whose symptoms were ultimately attributed to non-inflammatory conditions (i.e., IBS, SIBO, other food intolerance, pancreatic exocrine insufficiency) and not to their enteropathy. Among these individuals, only one had mildly increased tTG (1.6 times upper limit of normal {ULN}) in the three months preceding budesonide therapy. The alternate diagnoses were based upon additional testing and/or response to specific therapies and were supported by normalisation of histology in six of the seven cases who had a follow-up biopsy.

Table 4:

Clinical and Histologic Response According to Suspected Aetiology of NRCD

Suspected aetiology	Clinical response %	Clinical recurrence among responders %	Mucosal recovery within one year of therapy %	Mucosal recovery on follow-up biopsies within or more than one year after therapy %
Refractory Celiac Disease type I n = 20	70†	86	25 (3/12)	27 (4/15)
Gluten contamination n = 10	50‡	80	83 (5/6)	67 (6/9)
Irritable bowel syndrome/food intolerance/pancreatic exocrine insufficiency/small intestinal bacterial overgrowth n = 9	11	50	75 (3/4)§	86(6/7)§
Post-partum flare n = 1	100	100	Not performed
Other form of enteritis¶ n = 2	100	100	0 (0/2)

^†Three cases (14%) were subsequently diagnosed with coexisting IBS considering the lack of CR to budesonide

^‡Two cases (18%) without any CR were also diagnosed with IBS and FODMAP intolerance

^§The case with persistent mucosal damage was not considered to have RCD because her predominant symptoms were constipation, abdominal pain and heartburn, without signs and symptoms of malabsorption. Initial improvement on budesonide therapy but rapid recurrence and she did not respond to a second course. A subsequent glucose breath test for SIBO was positive, possibly explaining the persisting mild villous atrophy.

^¶Pouchitis (n=1), prolonged course of GI symptoms following a suspected viral enteritis (n=1)

Overall, the rate of clinical response varied from 71 to 100% across the aetiologies, except for those with symptoms attributed to a non-inflammatory condition, for which the rate of clinical response was poor (11%) (Table 4). Most non-responders to budesonide (56%) had non-inflammatory conditions that were eventually felt to explain their symptoms. Finally, the patients identified as having RCD type 1 (n=20) received budesonide for a median duration of 16 (IQR 12–40) weeks. The majority (70%) showed significant initial clinical response, but 86 % of the responders had a clinical recurrence and 35% required at least one additional course.

Follow-up histology / videocapsule

In total, 24 individuals had duodenal biopsies and/or VCE either on budesonide therapy (n=14, median 22.4 weeks after initiation of budesonide) or within one year of completing budesonide therapy (n=10). Overall, 46% (11/24) had normal villous architecture (mucosal recovery) within one year of budesonide treatment, including two patients who had a VCE to follow-up villous identified by VCE. Expanding the analysis to include individuals for whom biopsies were performed more than one year after budesonide (n = 31) did not significantly improve the rate of mucosal recovery (52%). Only three individuals who received open-capsule budesonide had follow-up biopsies, and 33% (1/3) had mucosal recovery.

Overall, we observed a discordance between clinical response to budesonide therapy and mucosal recovery. Within one year of therapy, 25% (6/24) had both clinical response and mucosal recovery and 17% (4/24) had both persistent symptoms and mucosal damage. Persistent mucosal damage was associated with clinical recurrence, need for subsequent budesonide course and use of additional therapy (Table 2). Among the 15 individuals without a clinical response, 11 (73%) had follow-up biopsies either within one year (n = 8) or more than one year (n = 3) after budesonide therapy. Six (55%) of these individuals had mucosal recovery and their symptoms were considered to be related to an alternative diagnosis such as motility disorder, IBS, SIBO or food intolerance. Among the individuals with a diagnosis of RCD1, 12 (60%) patients had subsequent duodenal biopsies during or within one year of budesonide therapy and only three (25%) had mucosal recovery. Again, a clinical response was not predictive of mucosal recovery (only 38% of individuals with clinical response had mucosal recovery) and absence of mucosal recovery was associated with more clinical recurrence, subsequent budesonide courses and switch to other therapies.

Individuals with prior celiac dietitian evaluation

We then examined the subjects who had an evaluation by a celiac dietitian prior to budesonide therapy (n = 29, 69% of the cohort). This subgroup was similar to the main cohort in terms of clinical characteristics (Supplemental Table 1). Notably, most (80%, 12/15) of the individuals with positive celiac serologies saw a dietitian at one of the celiac centers before initiating budesonide. Clinical response rate to budesonide was higher among the subjects who saw a dietitian prior to the medication compared to those who did not (70 vs 23% p<0.01). Adding dietitian counseling to the multivariate model showed a significant association with clinical response (Adjusted OR 80, 95%CI 3.3–1930) and still kept a significant association for diarrhea (Adjusted OR 16.9 95%CI 1.3–216), absence of fatigue (Adjusted OR 26.7, 95%CI 2.1 – 336). However, the rate of mucosal recovery and relapse after tapering budesonide were the same among individuals with and without dietitian counseling (Supplemental Table 2). The distribution of etiologies of the NRCD symptoms in the group with a dietitian evaluation was also the same as the main cohort (Supplemental Table 3).

Discussion

This retrospective study included subjects with persistent symptoms and villous atrophy despite being on a GFD. Therefore, budesonide was given in the context of possible RCD. In the recent series from Mayo Clinic [9], uncertain diagnosis of RCD or possible gluten contamination were excluded. We chose to present the use of budesonide among a cohort that included individuals eventually diagnosed with other NRCD diagnosis, which has not been reported so far. We found that clinical response to budesonide was more likely among individuals with diarrhea and less likely with individuals with fatigue or other extraintestinal manifestations. Acknowledging Mayo Clinic’s experience with budesonide in RCD [9], we may consider that most of our participants received relatively short courses of budesonide (less than 4 months). A lack of response to these short courses was suggestive of a functional disorder. We also observed that in individuals with RCD, short courses were associated with high risk of clinical recurrence and lack of mucosal recovery, emphasizing that in these patients, a longer course could be more effective, as shown in Mayo Clinic’s recent series [9].

Surprisingly, after careful and thorough medical evaluation, only 48% of this cohort was a posteriori considered probable RCD, although the cohort received budesonide for this clinical suspicion. Despite initial dietary changes and even GCED, 36% of our cohort still had elevated celiac serologies at the time of budesonide initiation, potentially suggesting ongoing gluten exposure [14]. Recent studies involving stool and urine testing for gluten immunogenic peptides have established that except perhaps for a GCED, a 100% gluten-free diet is utopian [15]. There were twenty-nine subjects who met with a celiac dietitian and did not have a significant source of gluten identified to explain the severity of symptoms. Gluten ingestion was suspected a posteriori in eight of them, either because it was later identified (e.g daily use of a herbal supplements that contained barley) or not divulged by the patient. Most of those with elevated tTG were seen by a dietitian for suspicion of gluten exposure. Interestingly, seeing a dietitian was associated with higher odds of clinical response. It is unclear whether this reflects a lag between dietitian intervention and clinical response or if dietitian involvement affects response to budesonide. Alternately, the rate of decrease of tTG is highly variable from one individual to another [16], and although the presence of circulating tTG implies on-going immune activation, it may not be a sign of active gluten exposure, especially when the levels are trending down[16]. Finally, the rate of positive serology observed in this study is not unusual, being comparable to a previous cohort of individuals with RCD treated with budesonide [8].

Discrepancies between histologic and clinical improvement, especially among those with a final diagnosis of IBS, are a reminder that CD patients may present with gastrointestinal symptoms that are unrelated to their enteropathy. Although mucosal recovery is expected in adults with CD, delays to mucosal recovery of more than one year despite a GFD have been reported and may be related to the initial histologic severity and degree of tTG elevation [17]. Consequently, symptomatic patients with persistent villous atrophy must not be diagnosed with RCD until a thorough and longitudinal evaluation is completed. The terminology “slow-responders” was recently proposed to describe individuals with residual mild immune activation who eventually achieve complete clinical remission [18]. In the context of significant patient distress, budesonide therapy is a helpful tool as it may help to dampen any residual activation of the immune system.

Budesonide is a micronized corticosteroid with high topical effect that is used in celiac crisis [6], RCD [8,9] and for acute symptoms related to gluten exposure [7], although it is only approved by the FDA as a treatment for mild-to-moderate ileocolonic Crohn’s disease. In addition, enteric-release budesonide is also used as a treatment for microscopic colitis[19], autoimmune enteropathy [20] and eosinophilic gastrointestinal disorders [21]. Here we report that budesonide may also be indicated to induce symptomatic improvement among individuals with NRCD. Budesonide has direct anti-inflammatory properties on the duodenal mucosa [10], triggers translation of anti-inflammatory genes and modulates Th1 and Th2 immunity [22]. Accordingly, most clinical improvement occurred in individuals whose symptoms were suspected to be enteropathy-related. As such, clinical improvement was more significant among individuals presenting with diarrhea, even after adjusting for those with microscopic colitis. Conversely, those with extraintestinal manifestations and fatigue did not respond to budesonide. This may be related to the high first pass metabolism of budesonide which may attenuate its effectiveness for CD manifestations resulting from systemic inflammation. Alternatively, others have reported that extraintestinal manifestations may persist for more than two years even among those considered adherent to a GFD [23]. However, a placebo effect cannot be excluded without a randomized controlled trial.

While there was a good initial clinical response rate, symptom recurrence and need for additional courses of budesonide therapy were common. Nevertheless, most of those with MR did not experience a clinical recurrence, suggesting that as for inflammatory bowel disease, achieving MR in symptomatic CD patients with active enteropathy may lead to prolonged clinical remission [24]. Consequently, the lack of follow-up biopsies in 43% of our subjects is a concern and histologic assessment should be considered as an important endpoint before tapering budesonide, as recently reported by another celiac center [9].

RCD type 1 is a difficult diagnosis to establish, because there are no specific biomarkers and gluten can be ubiquitous in the diet. Although gluten ingestion can not be completely excluded, the routine use of a GCED and longitudinal medical follow-up of the patients included in our study has facilitated the identification of a subgroup of patients for which RCD I was more likely. However, this subgroup had a significantly lower rate of mucosal recovery with budesonide (25%) than was reported in a recent cohort study (67%) [9]. The formulation of budesonide (not open capsule) and significantly shorter treatment duration in our cohort are probably the main factors explaining our results [9], which were more comparable to the previous experience of budesonide in RCD [8]. Alternately, not all the patients had a colonoscopy prior to the initiation of budesonide, so it is possible that some clinical responses be due to treatment of undiagnosed microscopic colitis. Finally, two individuals with a celiac crisis subsequently developed RCD; this is an interesting finding since long-term follow-up of patients presenting with a celiac crisis has not been reported [6] and our observations suggest a risk of persistent activation of the immune system in these patients.

We used strict inclusion criteria for our study population with NRCD, excluding individuals for whom symptoms were most likely attributable to active microscopic colitis or who had been on a GFD for less than one year. We also present a well-characterized cohort with thorough investigations, including recent duodenal biopsies showing villous atrophy. Nevertheless, this study has limitations due to retrospective and observational design, small sample size, lack of standardized clinical and histological evaluation and heterogeneity of the cases and treatment durations. Thirty-three percent of patients did not meet with a celiac dietician before initiating budesonide, but this may be an overestimate since we only considered visits with celiac dieticians from our centers. There were no differences in the frequency of patients with dietitian visits between the centers. Furthermore, we cannot assess precisely the rapidity of the effect of budesonide and optimal dosage regimen. Nevertheless, we report novel and relevant data on the use of budesonide to induce or accelerate clinical response remission in NRCD. Budesonide may attenuate enteropathy symptoms such as diarrhea and contribute to determine whether ongoing symptoms are or are not related to active CD. Our observations also suggest that short courses of budesonide do not lead to mucosal recovery in RCD type 1 patients, and that in these patients, tapering the drug according to clinical response only may be associated with clinical relapse. Future studies are warranted to investigate the benefits and optimal dosage regimen of budesonide in the management algorithm of NRCD.