Gastric dysmotility and gastrointestinal symptoms in myalgic encephalomyelitis/chronic fatigue syndrome

Abstract Background Gastrointestinal symptoms are common in ME/CFS, but there is a knowledge gap in the literature concerning gastrointestinal motility features and detailed symptom description. Objective In this study, we aimed to characterize gastric motility and gastric symptoms in response to a liquid meal. Methods We included 20 patients with ME/CFS with abdominal complaints who were recruited to a double-blind randomized placebo-controlled trial of Rituximab. The patients of this sub study were examined with an ultrasound drink test, and gastrointestinal symptoms were evaluated using the Rome III questionnaire and Irritable Bowel Syndrome Symptom Severity Scale (IBS-SSS) questionnaire. Results We found that patients commonly reported fullness/bloating (75%), abdominal pain (45%) and nausea (35%). Ultrasound measurements revealed lower proximal measurements of the stomach after a meal (p < 0.01) and larger fasting antral area (p = 0.019) compared to healthy controls. The patients had a stronger symptomatic response to the liquid meal compared to healthy controls regarding epigastric pain, discomfort and nausea (p < 0.05). Ninety percent of the patients reported bowel movement frequencies within the normal range but scored high on bowel habit dissatisfaction and life disruption. Conclusion The patients presented with fullness/bloating, nausea and epigastric pain, showed signs of impaired gastric accommodation and visceral hypersensitivity, showing that the gastrointestinal symptoms of ME/CFS patients are similar to functional dyspepsia. Key summary  Gastrointestinal symptoms are common in ME/CFS, but there is a knowledge gap in the literature concerning gastrointestinal motility features and detailed symptom description.  • In this study, patients with ME/CFS had signs of impaired gastric accommodation after a liquid meal.  • Out of 20 patients, 15 patients reported fullness/bloating, 9 reported abdominal pain, and 7 reported nausea. The patients showed signs of visceral hypersensitivity on a drink test.  • Our findings suggest that patients with ME/CFS share many similarities with patients with Functional Dyspepsia. The findings were not typical for Irritable Bowel Syndrome.


Functional dyspepsia;
ultrasound; ME/CFS; impaired accommodation; irritable bowel syndrome Background Myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) is a disease characterized by postexertional malaise, fatigue, cognitive symptoms, sleep disturbances, sensory hypersensitivity, and symptoms related to the immune system. Following the Canadian consensus criteria of 2003 [1], it affects approximately 0.2-0.8% of the population, and the majority of the patients are women. The condition is often accompanied by gastrointestinal, genitourinary and skeletal muscular symptoms [2]. An epidemiological survey from severity in DGBIs correlates to the degree of visceral hypersensitivity as Simr en et al. found in a multicenter study [5].
Other key mechanisms in the DGBIs are enteric dysmotility such as delayed gastric emptying [6][7][8] and impaired gastric accommodation [9,10] in functional dyspepsia, and low-grade inflammation [11,12], and intestinal dysbiosis [13] in irritable bowel syndrome. Gastric motility has not been extensively described in ME/CFS, but Burnet et al. found delayed gastric emptying for liquids in 72% of the patients, and for solids in 38% of the patients [14]. To our knowledge, no studies have explored gastric accommodation in ME/CFS.
Oncologists at Haukeland University Hospital observed that some patients with ME/CFS and malignant disease reported an improvement in ME/CFS symptoms after receiving chemotherapy and/or rituximab, an anti-CD20 monoclonal antibody. With this in mind, open label studies were performed to systematically evaluate Rituximab treatment response in patients with ME/CFS, with promising results. This was the background for the double-blinded randomized clinical trial ('RituxME') 15 . In this study we aimed to characterize gastrointestinal symptoms in response to a liquid meal, and to investigate gastric motility features of ME/CFS with an ultrasound meal test. A secondary aim was to investigate if B-cell depletion therapy influenced upper gastrointestinal (GI) symptoms or gastric motility.

Study overview
The RituxME trial The RituxME trial was a multicenter, national, randomized, parallel-group, double-blind, placebo-controlled, phase 3 trial. The objective of the study was to verify or disprove the association between B-cell depletion and clinical responses in patients with ME/CFS. Patients were recruited from September 2014 until September 2015, and follow-up was completed in September 2017. The participants in the study received induction treatment with rituximab or placebo, given as 2 infusions, 2 weeks apart, followed by maintenance infusions at 3, 6, 9, and 12 months. (The Rituximab treatment study is registered at ClinicalTrials.gov: NCT02229942; EudraCT: 2014-000795-25). The results from the RituxME trial has been published elsewhere [15].

The RituxME gastro study
Patients included at Haukeland University Hospital (n ¼ 40) were interviewed upon inclusion, and patients reporting any symptoms from the gastrointestinal tract, such as abdominal pain or problems with diarrhea or constipation, were invited to participate in the sub-study. Twenty-two participants were considered for inclusion. At baseline, one patient was excluded due to a gastric ulcer on endoscopy, and one was excluded because of probable celiac disease, leaving 20 patients (90% female, mean age 35 years) at baseline ( Figure  1). At follow-up at 15-18 months after the first treatment, one patient was too ill to participate, one patient withdrew from the sub-study and one patient had withdrawn from the RituxME study, leaving 17 patients for the follow-up visit.
The included patients were examined with the ultrasound meal accommodation test, including a routine abdominal ultrasound scan, at baseline and 15-18 months following the first infusion of Rituximab/placebo. A physician (EKS) did an anamnestic semi-structured interview before the ultrasound examination, including questions about symptom duration, previous attempts of diet intervention, their main GI complaints as well as association between GI symptoms and ME/CFS symptoms.

Healthy controls
Healthy controls (HC) were recruited among hospital staff at Haukeland University Hospital and the University of Bergen using e-mail and bulletin boards, and among nursing students at the University College of Western Norway. The participants received oral and written information about the study and signed consent before inclusion. Inclusion and exclusion criteria are presented in Supplementary material. As the healthy control study was designed to use as a control group for other studies as well, we included 66% women in the study. The participants were examined with the same ultrasound protocol as the patients, and answered the following questionnaires; visual analogue scales for symptom load at fasting, 1 min, 10 min and 20 mins, as well as IBS-SSS, Rome III and EPQ-N 12.

The Ultrasound meal accommodation test
The patient met in the morning in a fasting condition, and was examined in a seated position, leaning slightly backwards. After reporting upper gastrointestinal symptoms on a visual analogue scale (0-100 mm), the patient ingested 500 mL commercial meat soup ('Toro klar kjøttsuppe', Orkla Foods, Bergen, Norway, containing 84 kJ, 1.8 g protein, 1.1 g carbohydrate, 0.9 g bovine fat, per 100 g of soup). A cross-sectional antral area was obtained by scanning the epigastrium in a sagittal section, using the left liver lobe, the aorta and the superior mesenteric vein as anatomical landmarks ( Figure 2). The antral area was measured in a fasting state and 1, 10 and 20 min postprandially. The proximal stomach was studied at 1, 10 and 20 min postprandially in 2 sections: an oblique frontal diameter ('Proximal diameter') and the area was measured in a sagittal section ('Proximal area'). Normal values have been published previously [16]. At the same time as each ultrasound measurement, the patient's symptoms were registered on a visual analogue scale (VAS) ranging from 0-100 mm. We measured nausea, epigastric pain, fullness/bloating, satiety, and upper abdominal discomfort in a fasting state as well as at 1, 10 and 20 min postprandially.

Questionnaires
To assess symptom severity of irritable bowel syndrome, we used the Irritable Bowel Symptom severity score (IBS-SSS), with possible scores from 0-500. A total score <75 is considered normal, scores of 75-149 is considered mild IBS, 150-299 is considered moderate IBS, and scores >300 is considered severe IBS [17].
The personality trait of neuroticism was measured using the Eysenck 12 item Personality Questionnaire (EPQ-N; Eysenck's Personality Questionnaire Revised, Short Form) [18].
We used a Norwegian translation of the Rome III criteria including questions for functional dyspepsia and irritable bowel syndrome [19,20].
The patients answered the questionnaires between the ultrasound measurements. Some patients needed extra time, and completed the questionnaires later the same day.

Ethical considerations
The study was conducted according to the Declaration of Helsinki. All participants received oral and written information and signed consent forms before any studyrelated procedures. The RituxME study was registered at ClinicalTrials.gov (NCT02229942) and EudraCT (2014-000795-25), and approved by the Health Research Ethics Committee in Norway (REK 2014/365), as well as the Norwegian Medicines Agency. The substudy was included in the Regional Ethics Committee approval (REK 2014/365), and the Healthy Controls study was approved by the Regional Ethical Committee of South-Eastern Norway (REK 2014/222-20).

Statistical analysis
We used IBM Statistics SPSS version 25 for statistical analyses, and to make Figures 3 and 4. We tested normality by Shapiro-Wilks test and by evaluating histograms. The ultrasound data and some of the symptom measures were normally distributed, but other symptom data were not. We calculated a gastric distribution coefficient based on ultrasound measurements, dividing the proximal gastric area by the antral area. For comparison of continuous variables between two groups, we used Student's t-test for normally distributed data, and Mann-Whitney-U test for non-normally distributed data. Results are presented as median and interquartile range (IQR) or mean and standard deviation. P-values <0.05 were considered statistically significant. Associations between the gastric distribution coefficient and symptoms were analyzed using Spearman's correlation analysis, results are presented as Spearman's rho.
Ultrasound measurements, individual symptoms and IBS-SSS scores were stratified by treatment status, and baseline values compared to post-treatment values by paired samples t-tests. To compare the change in the mentioned variables before vs after treatment in the Rituximab-group to the placebo-group, we calculated the difference and compared the differences with Student's t-test. Figure 1. Flow diagram presenting the RituxME gastro study, a sub-study of the RCT study RituxME, where patients with chronic fatigue syndrome (ME/CFS) were randomized to placebo or rituximab treatment. In this sub-study, the patients were examined with a drink test combined with ultrasound measurements of the stomach, as well as symptom registration of gastrointestinal symptoms. Using the left liver lobe, aorta and the superior mesenteric vein (SMV) as landmarks, the antral area can easily be measured by transabdominal ultrasound. This image was acquired 1 min after the patient had ingested a low-calorie liquid meal.

Patient characteristics
Out of 40 patients included in the RituxME study at Haukeland University Hospital, patients with unspecified gastrointestinal complaints were invited to participate in this substudy. We included 20 patients and 30 healthy controls. Of the patients, 18 were women (90%), compared to 20 (66%) in the healthy controls group (p ¼ 0.091). The patients had ME/CFS by the Canadian consensus criteria [1], and selfreported gastrointestinal complaints. Patients were on average 35.0 (11.8) years old, healthy controls 32.7 (11.8) years old (p ¼ 0.478), and body mass index were 23.7 (3.4) and 23.5 (2.6) kg/m 2 in the patient group and healthy controls groups, respectively (p ¼ 0.857). We present anamnestic information about the patients' gastrointestinal symptom history, based on a semi-structured interview upon inclusion, in Table 1. We found that 55% of the patients reported that their gastrointestinal symptoms had increased after the debut of ME/CFS symptoms. The most frequent reported GI symptom was fullness/bloating in 75% of cases, followed by abdominal pain (45%) and nausea (35%). The patients scored higher on the personality trait neuroticism (mean 3.15, SD 2.4) compared to healthy controls (mean 1.5, SD 1.4) on a scale from 0-12 (p ¼ 0.009).

Gastric dysmotility assessed by ultrasound
Ultrasound measurements and P-values are presented in Table 2. Ultrasound of the stomach after drinking 500 mL soup revealed a smaller proximal diameter in patients with ME/CFS compared to healthy controls, at 1 min (p ¼ 0.001) and 10 min (p ¼ 0.003). However, when we only included female participants (patients and HC) in the analyses, the measurements at 10 min were no longer significant at a <.05 level (p ¼ 0.053). Furthermore, patients with ME/CFS had larger antral areas in a fasting condition (p ¼ 0.019) and a tendency toward larger antral areas 1 min postprandially (p ¼ 0.078) compared to healthy controls. When only including females in these analyses, the differences were no longer significant.

Dyspeptic symptoms before and after a meal
The participants registered symptoms of epigastric pain, nausea, satiation, fullness/bloating and total upper abdominal discomfort on a visual analogue scale from 0-100 mm, in a fasting condition as well as 1, 10, and 20 min postprandially. Results are presented in Table 3. We found that patients with ME/CFS reported higher levels of epigastric pain, nausea and upper abdominal discomfort compared to healthy controls at all time points. At 10 and 20 min, the patients also reported higher levels of fullness/bloating.
A distribution coefficient was calculated by dividing the proximal area by the antral area at each time point and are presented in Supplementary table 1. As demonstrated in Figures 3 and 4, we found significant correlations between the distribution coefficient and nausea (r ¼ 0.70, p ¼ 0.003) and upper abdominal discomfort (r ¼ 0.65, p ¼ 0.003) at 10 min postprandially.

Symptoms of irritable bowel syndrome and functional dyspepsia
Seven patients fulfilled Rome III criteria of Functional Dyspepsia and Postprandial Distress Syndrome. Only 12 of 20 patients answered all necessary questions for the IBS diagnosis, and seven patients (58% of the 12) fulfilled the criteria for the IBS diagnosis. At follow-up, 12 of 16 (75%) patients fulfilled the IBS diagnostic criteria, and 9 of 16 (56%) had functional dyspepsia.
We present results from some of the items from the Rome III and IBS-SSS questionnaires in Table 4. We found that 63% of the patients reported bloating/distention of the abdomen once per week or more, and 58% reported nausea once per week or more. None reported bowel movements >4 times per day, and only 10% reported less than 3 bowel movements per week (constipation).
At baseline, the mean IBS-SSS score was 180 (87.6), ranging from 7 -350. When exploring the separate questions from the IBS-SSS questionnaire, we found that the patients generally scored higher on 'bowel habit dissatisfaction' and 'life disruption' compared to the questions regarding abdominal pain. Bloating was reported during the previous 10 days in 70% of the patients, and abdominal pain in 55%.

Treatment effects of rituximab
The patients were treated with placebo or rituximab for 12 months and re-examined at 15-18 months with the Ultrasound Meal Accommodation test. In total, 17 patients participated in the control visit at 15-18 months (Figure 1). Nine patients (53%) had received active treatment with rituximab, and eight patients (47%) had received placebo treatment. There was no overall change in IBS-SSS scores in either group (p ¼ 0.838 in the placebo group, p ¼ 0.516 in the Rituximab group), and the difference in IBS-SSS scores before vs. after treatment were not different between placebo and Rituximab groups (p ¼ 0.812). Furthermore, we found no difference in the ultrasound measurements in Table 1. Anamnestic information obtained during a semi-structured physician interview upon inclusion of 20 patients with ME/CFS with self-reported GI symptoms.

Anamnestic information about gastrointestinal symptoms
Presence of GI symptoms before ME/CFS diagnosis 9 (45%) Increased symptom load of GI symptoms after ME/CFS diagnosis 11 (55%) Meal related GI symptoms 16 (80%) Stress-related GI symptoms 12 (60%) Reduction in symptom load after diet intervention a 11 (55%) Fullness/bloating 15 (75%) Abdominal pain 9 (45%) Nausea 7 (35%) Early satiety 6 (30%) The patients were recruited from a cohort of 40 patients with ME/CFS participating in a multi-center randomized controlled trial. a The patients were asked if they had done any diet intervention previously to try alleviating their symptoms, and if this had any impact on symptom load.
either group before compared to after treatment. Regarding dyspeptic symptoms, there was no difference in symptomatic response to the soup meal nor in fasting symptoms in the treatment group. However, we found an increase in symptoms of satiety (p ¼ 0.012) and discomfort (p ¼ 0.012) at 1 min postprandially in the placebo group. We calculated the difference between baseline measurements and post-treatment measurements and compared the differences in patients who had received Rituximab treatment to patients who had received placebo. There were no differences in the change of ultrasound measurements or symptoms when we compared the placebo group to the treatment group ( Supplementary Tables 1 and 2).

Discussion
In this explorative study of 20 patients with ME/CFS we have characterized gastrointestinal symptoms as well as gastric motility with an ultrasound meal test. We found that these patients commonly reported fullness/bloating, abdominal pain and nausea, and that they had signs of impaired gastric accommodation and visceral hypersensitivity.
Visceral hypersensitivity is a common feature in disorders of gut-brain interactions. Gastric hypersensitivity can be tested by barostat, or non-invasively with a drink test [21]. We used a low-calorie commercial meat soup, as it has a high tolerability in healthy individuals but induces dyspeptic symptoms in patients with functional dyspepsia and diabetic gastroparesis [9,16]. The patients with ME/CFS reported higher symptoms of upper abdominal pain, discomfort, and nausea as a response to this meal, compared to healthy controls. The response to this meal test can be interpreted as visceral hypersensitivity. Furthermore, nausea and fullness/bloating are very common complaints in patients with functional dyspepsia [16].
In a paper from 2016 our research group presented scores from the ultrasound meal accommodation test in several different patient groups [22]. Comparing symptoms 1 min after the meal, the patients with ME/CFS in the present study had Table 2. Gastric Ultrasound measurements before and after (pp) intake of 500 mL low-calorie soup in patients with chronic fatigue syndrome (ME/CFS) and healthy controls. Differences between the groups (Student's t-test), results given as P-values. We also include the P-value for females only (in both patient and HC groups).  Results from a study of upper gastrointestinal symptoms in patients with chronic fatigue syndrome compared to healthy controls. Results are given as median values and interquartile range (IQR), and P-values after Mann-Whitney-U tests for non-parametric data, or student's t-test (marked with Ã ) for normally distributed data. a Student's t-test, others are Mann-Whitney U-test. higher levels of fullness/bloating and satiety compared to patients with functional dyspepsia, nausea of the same level, while discomfort and pain were lower in the ME/CFS population. Compared to IBS patients, the ME/CFS group had higher levels of nausea, fullness/bloating and satiety, and lower levels of discomfort and pain. Thus, in this regard the ME/CFS patients in our study had many common features with functional dyspepsia. Gastric motility has been extensively studied in functional dyspepsia, but although there seems to be an overlap between ME/CFS and FD, research on the mechanisms behind is scarce. We found that ME/CFS patients had impaired accommodation, which to our knowledge, is the first time ever reported. Gastric accommodation can be assessed by many different modalities, ranging from the invasive examination by gastric barostat, to imaging modalities such as MRI, SPECT and ultrasound, to the less specific but readily available maximum tolerance drinking test [23]. Gastric accommodation is a reflex triggered by food passing through the esophagus and entering the stomach [24,25], and also by gastric juice entering the duodenum [26]. It is largely depending on vagal stimuli via nitroxergic neurons through 5-hydroxytryptamine receptors [27,28]. This instigates relaxation of the circular muscular layer of the stomach, resulting in lower intragastric pressure, enabling the proximal stomach to function as a reservoir for food. Accommodation is impaired in conditions such as functional dyspepsia [9,16,29] and diabetic gastroparesis [30] and can be affected by emotional or psychological stress [31]. Our findings indicate that FD and ME/CFS may have some common pathogenetical factors.

CFS/ME
We found that the ME/CFS patients had larger fasting antral area compared to healthy controls, but the finding was not significant when adjusting for gender. The postprandial measurements were not different. Enlarged antral area is a well-known marker of functional dyspepsia and a sign of delayed gastric emptying postprandially [32]. Given the small scale of this study, one may hypothesize that given a larger study sample, this would be a stronger finding.
Furthermore, we found an association between the intragastric distribution at 10 min postprandially and symptoms of nausea and upper abdominal discomfort. This implies that the gastric dysmotility observed may be involved in generating the patient's symptoms.
Abdominal pain is a cornerstone symptom for the IBS diagnosis. However, only 55% of the included ME/CFS patients reported abdominal pain during the last 10 days (Table 3), and of a relatively mild severity. Abdominal distention/bloating was more frequent (70%). We found that only 10% of the patients had <3 bowel movements per week, indicative of constipation, and no one had >4 bowel  movements per day. Thus, the reported bowel habits were not similar to irritable bowel syndrome. This was surprising, as we had expected that this selected population of ME/CFS patients had classical IBS symptoms. Symptoms of bloating/distention, abdominal pain and nausea were described in previous literature [2,4], but to our knowledge, few authors have described abdominal complaints in ME/CFS in detail. Interestingly, the patients scored relatively high on the questions about life disruption because of their bowel habits, and bowel dissatisfaction. This is intriguing, given that 90% of the patients reported frequency of bowel movements within what is usually considered the normal range (<3 per day, >3 per week).
This study was part of a larger, double-blind RCT on the effects of depletion of CD20þ B-lymphocytes with Rituximab on ME/CFS. The results from the main study were negative with respect to effect of treatment [15]. We investigated the effects of Rituximab treatment on abdominal symptoms, gastric motility and gastric sensitivity. We found no significant treatment effects in the Rituximab group. Despite few patients in this sub-study, we believe that there were no clear effects of the treatment on the factors studied.
This study has some limitations. First, the low number of study participants makes it difficult to draw firm conclusions. Furthermore, the findings from this study may not be applicable to all patients with ME/CFS, as we recruited patients who reported symptoms from the gastrointestinal tract in this substudy of the RituxME study. An unfortunate misunderstanding in answering the Rome III questionnaire led to some missing data. And last, although the interobserver variability in ultrasound is good [33,34], the method is fairly user dependent. To reduce this limitation, the physician performing the ultrasound examinations conferred in all cases with a senior gastroenterologist with long experience in ultrasound meal testing.
We used a questionnaire developed for patients with IBS to grade the lower abdominal symptoms. We did this under the assumption that the patients had IBS-like symptoms. Although our results indicate that the abdominal complaints in this ME/CFS population may differ somewhat from that of IBS, we still believe that the IBS-SSS is a reasonable tool to use, in lack of a specific ME/CFS-specific gastrointestinal symptom questionnaire.

Conclusion and future perspectives
In this explorative study of 20 patients with ME/CFS, we found signs of gastric dysmotility expressed as impaired gastric accommodation and antral distention. Furthermore, we found that the patients reported increased nausea, discomfort and pain after drinking a low-calorie soup. Moreover, we observed an association between altered motility and nausea and discomfort. Together, these features may indicate that the gastrointestinal complaints commonly observed in ME/CFS may belong in the spectrum of disorders of gutbrain interactions, with more similarities to functional dyspepsia than to irritable bowel syndrome. Systematic registration and testing for gastrointestinal symptoms and motility disturbances in patients with ME/CFS can increase the knowledge about this complex disorder in future studies.