Introduction

Irritable bowel syndrome (IBS) is one of the most frequent disorders in clinical practice, with a mean 7.6–10.8% worldwide prevalence using Rome III criteria. It accounts for many consultations and greatly impacts patients' quality of life [1].

In clinical practice, IBS is characterized by symptoms of recurrent abdominal pain and defecation disorder, mainly affecting women [2, 3].

IBS is not only considered a diagnosis achieved after the exclusion of organic disease, but also requires a positive diagnosis using a symptom-based criterion with the most recently Rome IV classification [4].

Nevertheless, there are some clinical conditions that can simulate IBS symptoms such as colon cancer, inflammatory bowel disease, celiac disease or microscopic colitis; therefore, is important to perform a careful and deep evaluation in selected patients to rule out these confounding conditions [5]. Knowing which patients will benefit from further investigations is challenging because the diagnostic yield in IBS patients may be low. The ACG guidelines strongly recommend that serologic testing be performed to rule out celiac disease (CD) in patients with IBS diarrhea symptoms and fecal calprotectin for inflammatory bowel disease [6].

Some patients with irritable bowel syndrome might have other underlying pathologies including pancreatic disease, which is difficult to diagnose, especially in the early stages. Even if the patient can be diagnosed as having IBS by the Rome IV classification, some symptoms may overlap with those of exocrine pancreatic insufficiency (EPI). Some studies have demonstrated a prevalence of 5 to 6% of severe exocrine pancreatic insufficiency in IBS diarrhea dominant patients (IBS-D) [7].

To date, this is the first study to our knowledge evaluating the prevalence of EPI in patients with IBS-D in Latin America.

Objective

We aimed to identify the prevalence of EPI based on fecal elastase testing in patients with diarrhea-predominant IBS in the outpatient setting and identify the clinical characteristics of patients with D-IBS that may predict the diagnosis of EPI.

Methods

Study Design

We undertook this prospective, cross-sectional study in a single tertiary gastroenterology center at a university hospital in Buenos Aires, Argentina, from May 2018 to March 2020.

All authors had access to the study data and reviewed and approved the final manuscript.

Patient Selection

The inclusion criteria were hospital outpatients aged > 18 years referred to the Neurogastroenterology Division who fulfilled the Rome IV criteria for D-IBS (Table 1).

Table 1 Baseline data: demographic and clinical data
Full size table

All patients underwent an exhaustive evaluation including full blood test (complete blood count, basic metabolic panel, LDL and HDL cholesterol, triglyceride levels, liver function test, calcium, magnesium, vitamin B12, prothrombin time, vitamin D), thyroid function test, erythrocyte sedimentation rate, celiac disease antibodies, upper endoscopy with duodenal biopsies and colonoscopy (in patients > 50 years or with warning symptoms) including intubation of the terminal ileum with biopsies for microscopic colitis.

Patients were excluded if they did not fulfill Rome IV criteria for D-IBS, had an organic disease, acute abdominal pain or rectal bleeding, or were unwilling to consent or unwilling to give a stool sample.

Measures

Patients completed questionnaires regarding stool frequency, stool consistency (Bristol Stool form Score) [8], IBS Severity Score [9], Hospital Anxiety and Depression Scale [10] and PHQ 15 [11] for somatization. Subject baseline body mass index and alcohol/tobacco use were also assessed. Functional dyspepsia was investigated in all patients according to Rome IV classification [2]. To assess IBS symptoms such as distention, abdominal pain and global IBS score, we used a patient-defined 6-point Likert scale ranging from very satisfied to very unsatisfied.

Fecal Elastase-1 Test

To investigate exocrine pancreatic insufficiency, a random stool sample was requested in all patients and assayed for fecal elastase (Fel-1).

Participants were instructed to collect their stool sample in a sterile plastic disposable container and deliver it to the Gastroenterology Chemical Laboratory for their process and analysis. According to previously published analyte stability reports [12, 13], the samples were stored refrigerated at 4–8 °C for no more than 48 h.

Fel-1 concentrations of all samples were measured with a commercially available enzyme linked monoclonal based immunosorbent assay (ELISA) kit (ScheBo-Pancreatic Elastase 1™, Giessen, Germany), and the fecal elastase was extracted and determined according to the manufacturer’s instructions.

EPI was defined as Fel-1 concentration < 100 µg/g or Fel-1 between 100 and 200 µg/g and an abnormality in additional pancreatic pathology testing such as serum albumin, vitamin E, vitamin D, vitamin A, folic acid, iron, transferrin, calcium, magnesium and/or malnutrition identified with anthropometric measurements performed by an expert nutritionist. Fel-1 ≥ 200 µg/g was considered normal [14].

All patients with Fel-1 < 200 µg/g were studied for α-1-antitrypsin deficit and IgG4 for autoimmune disease.

EPI Follow-Up Evaluation and Intervention

Endoscopic ultrasound evaluation was performed by an expert endoscopist in all the EPI patients (Fig. 1). Chronic pancreatitis was diagnosed according to the Rosemont criteria [15].

Fig. 1
figure 1

Flowchart

Full size image

Treatment of the EPI group included support to stop alcohol and smoking consumption and dietary evaluation. All EPI patients received treatment with PERT (Creon®, Abbott Laboratories GmbH, Germany) and underwent a 12-week follow-up to evaluate the treatment effect. The pancreatic enzyme therapy initial dose was 50,000 Eur.Ph.U. with each main meal and 25,000 Eur.Ph.U. with snacks [16]; we additionally added PPIs (proton pump inhibitors) twice a day to improve enzyme treatment efficiency.

After 12-week follow-up, stool frequency, stool consistency and questionnaires were re-assessed, as well as any side effects, to evaluate treatment response.

Ethics

The study protocol and informed consent document were approved by the institutional review board, and the study was conducted in accordance with the Declaration of Helsinki. Informed and written consent was obtained from all participants. Participant data were encrypted and confidential, and only the study authors had access to them.

Statistical Analysis

Quantitative variables were defined in mean (standard deviation, SD) or median (minimum–maximum range) in accordance with the statistical distribution (Kolmogorov-Smirnov test). Qualitative variables were defined in frequencies (%) and 95% CI. p < 0.05 was considered statistically significant, and < 0.02 for the multivariate analysis. The database was analyzed by the institutional statistician on R v3.6.0 (R Foundation for Statistical Computing; Vienna, Austria).

We analyzed different variables between IBS non-EPI and EPI patients (< 100 µg/g) and performed a sub-analysis between those patients with a fecal elastase 100–200 µg/g and EPI patients on specific variables. The contrast between quantitative variables was performed with Student’s t-test or Mann-Whitney U test (in correspondence with statistical distribution) and between qualitative variables with Pearson’s chi-squared test or Fisher’s exact test.

For those variables with significant difference between D-IBS non-EPI and EPI, we performed a univariate and multivariate analysis to confirm the statistical association between EPI presence among individuals with D-IBS.

Sample Size

We used the formula for estimating proportions from infinitive populations. We considered a 95% confident interval (CI), 5% error margin and 6.1% exocrine pancreatic insufficiency (EPI) rate (95% CI 3.7–9.3%) among 314 cases with D-IBS, reported by Leeds et al. [7]. Considering the upper limit of the previously cited CI, we estimated a sample of at least 130 cases.

Results

During the recruitment period, 140 patients who met Rome IV criteria for D-IBS were included (mean age 60 years, females 75.7%). Most of our patients were between moderate and severe according to the IBS severity score (55.7% moderate, 27.8% severe). For complete baseline data, refer to Table 1.

Fel-1 was < 100 µg/g stool in 7 of 140 (5%; 95% CI 2.2–10.4) of the D-IBS patients, who were therefore diagnosed with EPI; smoking was significantly higher in this group (22.6% vs. 71.4%, p = 0.013).

Comparing the laboratory tests between the EPI versus the D-IBS group, vitamin B12 was significantly lower in the EPI patients (p < 0.001) (Table 2), as were vitamin A (p = 0.04) and vitamin E (p = 0.04) (Table 3). None of the other tests showed significant differences between the tow groups.

Table 2 Baseline data: laboratory tests
Full size table
Table 3 Additional evaluation in patients with Fel-1 between 100 and 200
Full size table

Fel-1 between 100 and 200 µg/g was found in 11 patients (7.8%). After additional laboratory and nutritional evaluation (Table 3), none showed any abnormality so they were not considered to have EPI. IgG4 and α-1-antitrypsin were within normal levels in all patients.

In this cohort, 81 patients (56%) had an increased BMI, with 47 (33.6%) overweight, 18 (12.9%) obese class I, 9 (6.4%) obese class II and 7 (5%) obese class III (Table 1).

Patient-Reported Outcomes

When evaluating the baseline characteristics of our patients, most of them (89.3%) had a Bristol stool scale between 5 and 6. As for gastrointestinal symptoms, dyspepsia was reported significantly higher on EPI patients (5.3% vs. 71.4%, p < 0.01). After multivariate analysis (adjusted by age and sex), dyspepsia was an independent factor associated with EPI (OR 34.7; 95% CI 4.95–366.37, p = 0.0007).

There was no significant difference between bowel movements per day between EPI and non-EPI patients.

Regarding validated questionnaires (Table 4), depression was significantly higher among IBS patients than EPI (28.6% vs. 14.3%, p < 0.0069), but we found no differences between PHQ-15 and anxiety scores (p = 0.14 and 0.55). Likert scores for distension, pain, IBS and severity score showed no differences (p = 0.90, 0.84 and 0.54, respectively).

Table 4 Baseline data: patient-reported outcome scores
Full size table

Intervention and Follow-Up Evaluation

All EPI patients (n = 7) underwent endoscopic ultrasound evaluation, where pancreatic steatosis was the main finding (71.4%), followed by chronic pancreatitis (14.2%, Rosemont criteria: suggestive of CP; 1 major B + 3 minor features) and a normal pancreas (14.2%) (Fig. 1). The chronic pancreatitis patient admitted high tobacco exposure (40 pack-year).

All EPI patients started on pancreatic enzyme supplementation and were reassessed after 12 weeks (Fig. 2), showing a significant improvement in the Bristol stool scale (p < 0.0001), bowel movements per day (p < 0.005), distension score (0.0009), pain score (0.0277) and IBS severity (0.0034). Some patients reported improvement in the dyspeptic symptoms, but it was not significative (p = 1.0).

No adverse events were reported in any patient taking enzyme supplementation.

Fig. 2
figure 2

EPI cases (n = 7) most significant pre-versus post-treatment outcomes differences

Full size image

Discussion

This study describes a prospective, cross-sectional analysis for exocrine pancreatic insufficiency in patients fulfilling the Rome IV criteria for D-IBS referred to a university hospital in Buenos Aires, Argentina. The prevalence of a low Fel-1 level (< 100 µg/g stool) in D-IBS was 5%, showing similar results as previously reported by Leeds et al. [7] in the UK (6.1%) and Talley et al. [17] in Australia (4.6.%).

Similar to Talley et al. [17], most of our patients had a moderate IBS severity score (55.7%) and almost one third of the patients (27.8%) severe, and they were therefore referred to secondary and tertiary level gastroenterologists.

When evaluating the patients for functional disorders with Rome IV classification, it is important not only to rule out a systemic disease, but also to have positive criteria to identify the patient with IBS. To achieve a correct diagnosis, some confounder entities should be previously studied, such as celiac disease or microscopic colitis. As we can see from these study results, underlying pancreatic disease could also be considered a confounding factor. To identify symptoms that may help us to differentiate this group of patients, we performed a clinical evaluation in association with several validated patient-reported questionnaires, finding functional dyspepsia to be an independent symptom related to exocrine pancreatic insufficiency.

There are other studies that evaluate the association between pancreatic disease and dyspepsia; Lariño-Noia et al. [18] published a prospective, cross-sectional study to evaluate the morphological changes of chronic pancreatitis in patients with dyspepsia and concluded that this symptom is frequent in this group of patients (21%). There is another study that utilizes EUS to detect evidence of pancreatic disease in patients with persistent dyspepsia, this symptom being an independent predictor of severe EUS abnormalities (OR 7.21: 95% CI 1.99–26.26) [19]. Another prospective study in Turkey investigated the association between low Fel-1 and non-ulcer dyspepsia, finding in these patients that mean Fel-1 levels were significantly lower than in non-dyspeptic patients (367.47 ± 43.27; 502.48 ± 50.94, respectively; p = 0.04) [20].

Therefore, we think that dyspepsia should be an important symptom to explore in these patients, as it may be of help in clinical practice to further investigate whether patients have an underlying pancreatic disease. As for the blood samples, vitamins B12, A and E were significantly decreased in patients with EPI compared with the rest of the patients. In our population, tobacco smoking had a higher prevalence than alcohol intake, and it was significantly higher in this sub-group of patients.

To further evaluate the pancreas, endoscopic ultrasound was performed in all seven Fel-1 patients, with pancreatic steatosis being the main finding (71%, 5/7) in this study's population. A cross-sectional study in Texas evaluated the EUS findings in patients with EPI diagnosed by low Fel-1, and from a total of 40 patients, chronic pancreatitis was the main finding (72.5%) followed by fatty pancreas (22.5%) [21]. Another article in Turkey found that pancreatic steatosis can cause EPI; in this research, Fel-1 levels were significantly lower in patients with fatty pancreas vs. controls (319.76 ± 45.7 vs. 549.31 ± 69.4, respectively, p = 0.003) [22].

Other publications confirm the relationship between metabolic syndrome and nonalcoholic fatty pancreas disease [23].

From our baseline population report, we found a high prevalence of overweight and obese patients, and this could be the cause of the pancreatic steatosis. A recently published article evaluated the frequency of EPI in nonalcoholic fatty liver disease (NAFLD), finding that NAFLD patients had significantly lower Fel-1 levels than controls (297 [204–517] vs. 500 [298‐678] μg/g, p < 0.01) and identified NAFLD as an independent predictor of EPI (OR = 4.892, p = 0.021) [24]. Nevertheless, this was not our main study focus, and we think that more research is needed to investigate this topic.

Furthermore, pancreatic enzyme supplementation in EPI patients led to significant improvements in stool consistency, stool frequency, abdominal pain, distension and IBS severity score and was well tolerated with no adverse events reported by any of the treated patients. After the follow-up, we found an improvement in dyspeptic symptoms in this group of patients (40%, 2/5), but it was not statistically significant, probably related to the low number of EPI cases. This could be further investigated in future studies.

One of the main weaknesses of our study is that it was developed in a single tertiary gastroenterology center as opposed to Talley et al.'s [17] multicenter study. Also, our sample size (n = 140) was lower than other similar studies like Leeds [7] (n = 300) and Talley et al. (n = 218). Another weak point is that we only evaluated pancreatic insufficiency by low Fel-1 levels because we did not perform fecal fat quantification and secretin pancreatic function tests in our hospital. Regarding Fel-1, false-positive results may occur in these patients because of sample dilution for watery diarrhea. Nevertheless, we think this was not the case as all included patients had diarrhea and only seven had Fel-1 < 100 μg/g in association with an abnormal pancreatic laboratory test, and all of them showed improvements after enzyme replacement treatment.

The strengths of this study are that we used the most recent Rome IV classification [2], and that all patients were fully evaluated using blood tests for celiac disease, autoimmune disease, complete upper and lower endoscopic studies with duodenal and colonic biopsies, and several validated questionnaires for symptoms and comorbidities.

In conclusion, exocrine pancreatic insufficiency was present in 5% of patients who fulfill Rome IV criteria for D-IBS in a university hospital in Buenos Aires, Argentina. Smoking habit was higher in the EPI group, and dyspepsia was an independent symptom strongly associated with EPI. This could be a useful clinical tool to identify D-IBS patients with underlying pancreatic disease. In this cohort, pancreatic steatosis was the main endoscopic ultrasound finding. After PERT therapy, all EPI patients significantly improved their stool frequency, stool consistency, abdominal pain, distension and IBS severity score.