Distinct Alterations in Central Pain Processing of Visceral and Somatic Pain in Quiescent Ulcerative Colitis Compared to Irritable Bowel Syndrome and Health

Abstract

Background and Aims

Despite relevance to pain chronicity, disease burden, and treatment, mechanisms of pain perception for different types of acute pain remain incompletely understood in patients with inflammatory bowel disease [IBD]. Building on experimental research across pain modalities, we herein addressed behavioural and neural correlates of visceral versus somatic pain processing in women with quiescent ulcerative colitis [UC] compared to irritable bowel syndrome [IBS] as a patient control group and healthy women [HC].

Methods

Thresholds for visceral and somatic pain were assessed with rectal distensions and cutaneous thermal pain, respectively. Using functional magnetic resonance imaging, neural and behavioural responses to individually calibrated and intensity-matched painful stimuli from both modalities were compared.

Results

Pain thresholds were comparable across groups, but visceral thresholds correlated with gastrointestinal symptom severity and chronic stress burden exclusively within UC. Upon experience of visceral and somatic pain, both control groups demonstrated enhanced visceral pain-induced neural activation and greater perceived pain intensity, whereas UC patients failed to differentiate between pain modalities at both behavioural and neural levels.

Conclusions

When confronted with acute pain from multiple bodily sites, UC patients’ responses are distinctly altered. Their failure to prioritise pain arising from the viscera may reflect a lack of adaptive behavioural flexibility, possibly resulting from long-lasting central effects of repeated intestinal inflammatory insults persisting during remission. The role of psychological factors, particularly chronic stress, in visceral sensitivity and disease-specific alterations in the response to acute pain call for dedicated mechanistic research as a basis for tailoring interventions for intestinal and extraintestinal pain symptoms in IBD.

Inflammatory bowel disease, pain perception, functional brain imaging

Glossary of Terms Related to Pain Research

Term	Definition/Explanation
Visceral pain	Pain arising from the viscera, herein induced by pressure-controlled rectal distension stimuli
Somatic pain	Pain arising from somatically innervated organs such as the muscles or the skin, herein induced using cutaneous thermal heat pain stimuli
Pain modality	A modality describes a particular mode in which something is experienced or expressed; visceral and somatic pain modalities are differentiated because of particularities in processing and experience of pain arising from inner organs [visceral] and pain arising from skin or muscle tissue [somatic]
Pain sensitivity	Sensitivity towards pain stimuli, related to pain thresholds [ie, lower thresholds reflect high pain sensitivity and vice versa]
Pain threshold*	The minimum intensity of a stimulus that is perceived as painful
Hypersensitivity	Enhanced sensitivity to sensory stimuli
Hyposensitivity	Reduced sensitivity to sensory stimuli
Hyperalgesia*	Increased pain from a stimulus that normally provokes pain [ie, lower pain threshold]
Hypoalgesia*	Diminished pain in response to a normally painful stimulus [ie, higher pain threshold]
Perceived pain intensity	Subjective perception of stimulus intensity [herein: rectal distensions; cutaneous heat], typically quantified using rating scales [eg, visual analogue scale, VAS]

*Definitions provided by the International Association for the Study of Pain [IASP; https://www.iasp-pain.org/resources/terminology/].

1. Introduction

Abdominal pain is reported by up to 80% of patients with inflammatory bowel disease [IBD],¹ typically in concert with other gastrointestinal [GI] symptoms, such as altered bowel habits, that characterise phases of acute disease. Even during clinical remission, however, pain persists in a substantial proportion of patients.² Furthermore, pain is not confined to the GI tract but may also arise from the musculoskeletal system or the joints as part of extraintestinal disease manifestations.³ Effective treatment of abdominal pain remains a major challenge.⁴ In fact, one in four patients reports their IBD-related pain to be untreated,¹ with detrimental effects on health-related quality of life.^5,6 Despite its clinical relevance, research on the mechanisms of pain perception and underlying neural circuits in IBD remains limited, especially in ulcerative colitis [UC].⁷

Most of the mechanistic research on pain in IBD has focused on pain sensitivity, ie, on thresholds for experimentally applied painful stimuli [for a definition of terms related to pain research, see Glossary]. For the visceral pain modality, pain thresholds assessed by rectal distensions are reportedly lower during active disease,⁸ indicating visceral hypersensitivity which appears to persist during remission mainly in patients presenting with ongoing symptoms despite adequate control of intestinal inflammation.⁹ In this subgroup, the clinical presentation mirrors that of irritable bowel syndrome [IBS], a primarily non-inflammatory disorder of gut-brain interactions, motivating research into distinct and common mechanisms in these conditions.¹⁰ Regarding sensitivity for pain arising from bodily regions like the skin or muscle [ie, somatic pain], higher pain thresholds [ie, hyposensitivity] but comparable perceived pain intensities have been observed in IBD.^11,12 Together, these data support modality-specific alterations in pain sensitivity possibly driven at least in part by recurring inflammation in IBD.

Despite relevance to pain chronicity, disease burden, and treatment, the subjective experience of pain, contributing factors, and underlying neural mechanisms which have been broadly studied in IBS¹³ remain incompletely understood in IBD. Experimental studies assessing behavioural responses to somatic or visceral pain stimuli are scarce and have yielded inconsistent results.^11,14,15 Putative alterations in the neural processing of experimental pain in IBD, which might persist during phases of disease remission, also remain elusive.^11,15,16 To investigate central pain mechanisms in the context of the gut-brain axis, novel experimental protocols with painful stimuli from different modalities applied within one brain imaging session, so called ‘multiple threat paradigms’, have more recently been implemented.^17–21 These were designed to model the experience of multiple symptoms arising from the viscera and other bodily sites, reflecting patients’ clinical reality of aversive intestinal and extraintestinal symptoms. For mechanistic research into brain mechanisms relevant to understanding the normal and adaptive specificity of pain responses for the visceral modality, these paradigms have also proven to be highly instrumental. Indeed, when repeatedly confronted with visceral and somatic painful stimuli matched to intensity, healthy volunteers perceive visceral stimuli as more unpleasant, fear-evoking, and threatening.^18,22 Distinct differences are also observable at the level of neural representations of visceral versus somatic pain,^13,17,18,21 in line with a proposed greater biological salience of visceral signals²⁰ which appears to shape perception and pain-related cognitive and emotional responses such as learning and memory processes.^19,23

Herein, for the first time, we report on behavioural and neural responses in a multiple-threat paradigm with visceral [ie, rectal distensions] and somatic [ie, thermal cutaneous] pain stimuli in patients with IBD [ie, UC in clinical remission], patients with IBS, and healthy controls. We directly compared responses to visceral and somatic pain stimuli to elucidate if altered pain processing in chronic visceral pain patients is confined to the GI tract [ie, the visceral modality] or applies more broadly to other pain modalities. To address the specificity of the pain experience to clinical condition, which can inform us about putative mechanisms underlying pain persistence beyond the resolution of acute inflammation in UC, we included IBS as a patient control group.

2. Materials and Methods

2.1. Participants

All volunteers were recruited by public advertisement between the years 2015 and 2019 at the University Hospital Essen, Germany, for a comprehensive study on different aspects of visceral and somatic pain, including behavioural testing and several brain imaging sessions. The screening process consisted of a standardised telephone screening, followed by a personal interview and a medical examination, including a digital rectal examination to exclude perianal tissue damage [eg, haemorrhoids, fissures, sphincter damage] which may interfere with rectal distensions used as visceral pain stimuli. Given the female preponderance of chronic visceral pain in disorders of gut-brain interactions²⁴ and evidence of sex/gender differences specifically for the visceral modality,²⁵ only women were recruited for all groups. Pregnancy was ruled out using a commercially available pregnancy test [Biorepair GmbH, Sinsheim, Germany, sensitivity 10 mIU/ml] on the day of the study. Besides male sex/gender, general exclusion criteria for all participants included age <18 or >65 years, body mass index <18 or >30, the usual magnetic resonance imaging [MRI]-specific exclusion criteria [eg, claustrophobia, ferromagnetic implants], and structural brain abnormalities [verified by neuroradiologist]. For the UC group, only patients in clinical remission were included to avoid interference of active disease with study-related procedures, and to minimise putative effects of acute inflammation [or medical treatments required during phases of disease exacerbation] on study measures. Treatment with systemic glucocorticoids within 4 weeks of the study was exclusionary. Clinical disease activity was assessed based on symptom reports, initially evaluated in a structured screening interview, and then quantified with the Clinical Activity Index [CAI],²⁶ which allows classification of current disease activity based on the total sum score into inactive [ie, remission; ≤4], mild activity [5–10], moderate activity [11–17], and high activity [≥18].²⁷ Here, the symptom-based version of the CAI was used, excluding one item concerning laboratory results [ie, erythrocyte sedimentation rate and haemoglobin; for a detailed description, see Öhlmann et al.^7]. Only patients with symptom-based CAI scores ≤4 [ie, clear remission] were enrolled. In addition, as a non-invasive biological marker of intestinal inflammation,²⁸ levels of faecal calprotectin served as an additional criterion, using the established reliable cut-off value of below 150 μg/g of stool indicating biomarker remission.²⁹ As time delays between initial screening and scheduling of the study day were not always avoidable, a standard procedure was implemented to ascertain that no changes in disease status or current medications occurred. To this end, every patient who fulfilled all criteria and had provided informed consent was contacted via telephone 7–8 days before the scheduled testing session. Symptoms were re-assessed to ensure clinical remission, and patients were instructed to collect the faecal sample, which was directly sent to the laboratory by the patients. In case clinical remission could not be ascertained based on either symptom report or faecal calprotectin levels, the study session was rescheduled. For the IBS group, symptom-based confirmation of diagnostic criteria was based on ROME IV criteria.³⁰For both patient groups, an existing and confirmed diagnosis of the respective GI disorder established at least 1 year prior to recruitment for this study was required. Minor and stable [or successfully treated] psychological symptoms, such as mild anxiety or depression symptoms [indicated by elevated scores on the Hospital Anxiety and Depression Scale, HADS,³¹ described in detail below] were not exclusionary, whereas patients with diagnosed, more severe psychiatric comorbidities [such as schizophrenia/psychotic disorders, substance use disorders, borderline personality disorder] were excluded. Note that given frequent reporting of additional extraintestinal pain symptoms in IBS and IBD,^32,33 patients who reported such symptoms in addition to symptoms of their primary GI diagnoses were not excluded. For healthy controls, additional exclusion criteria were any somatic or mental health conditions based on self-report or scores above the published cut-offs in HADS, frequent GI complaints,³⁴ or regular use of any medications [except hormonal contraceptives or thyroid medication]. Of note, data from a larger sample of healthy participants, recruited in parallel to patient cohorts, have previously been used within a published study on pain-related learning and memory in health,²⁰ which however did not contain data for the analysis carried out herein. The healthy sample used herein was a subset tested by the same study team and matched to the two patient groups based on age. Work was conducted in accordance with the Declaration of Helsinki, and the study was approved by the local Ethics Committee of the University Hospital Essen [protocol number 10-4493]. All volunteers gave written informed consent and received monetary compensation for participation.

2.2. Psychosocial and clinical questionnaires

Participants completed a comprehensive psychosocial and clinical symptom questionnaire battery that included assessment of GI symptoms, pain, anxiety, depression, and chronic stress. GI symptoms were quantified with a standardised questionnaire that we routinely use in our group, as it is applicable across visceral pain conditions as well as in healthy volunteers.³⁴ A range of typical GI symptoms [ie, diarrhoea, constipation, vomiting, nausea, lower abdominal pain, upper abdominal pain, heartburn, post-prandial fullness, bloating, loss of appetite] in the previous 3 months is assessed using a Likert-type response scale [ie, an ordinal scale where ordered categories are provided as options to choose from: 0 = experience never, 1 = experience once or twice per month, 2 = experience once or twice per week, 3 = experience more than twice a week], and total sum scores can be calculated for analyses. Additionally, the Brief Pain Inventory [BPI³⁵] was administered. In two subscales, pain severity and pain-related interference with function is assessed. For a mean pain severity score, the intensity of the least, average, and worst pain within the last 24 hours as well as current pain intensity are assessed and averaged. Symptoms of anxiety and depression were evaluated by the HADS,³¹ which consists of two subscales with seven items measuring anxiety [HADS-A] and depression [HADS-D], respectively. For each subscale, available cut-off values differentiate between non-cases [sub-scale score <8], potential cases [subscale score 8–10], and probable cases [subscale score ≥11] of anxiety and depression.³⁶ To assess chronic stress, the 12-item screening scale of the Trier Inventory of Chronic Stress [TICS-SSCS] was used.³⁷ The scale evaluates individual experiences with chronic stressors in everyday life and provides a reliable global measure of perceived stress during the past 3 months,³⁸ with higher scores indicating greater perceived stress. For TICS, norm values are available for healthy volunteers with a mean score of 13, corresponding to T = 50.³⁹

2.3. Experimental pain models

For the visceral modality, pressure-controlled rectal distensions were accomplished using a barostat system [modified ISOBAR 3 device, G & J Electronics, Toronto, ON, Canada]. The inflatable balloon allows for precise and graded distensions of the rectum inducing mild, intermediate or strong sensations of urgency, discomfort, and pain.⁴⁰ These sensations closely resemble the aversive visceral symptoms regularly experienced by patients with UC and IBS, which are also encountered by healthy controls, albeit less frequently and intensely. For the somatic modality, cutaneous thermal pain stimuli were applied on the left ventral forearm using a thermode [PATHWAY model CHEPS; Medoc Advanced Medical Systems, Ramat Yishai, Israel]. Note that for safety reasons, the maximum pressure applied with the rectal balloon was limited to 60 mmHg and for the thermode a temperature limit was set to 50°C.

2.4. Experimental procedures

An overview of experimental procedures is provided in Figure 1. On the study day, visceral and somatic pain thresholds were initially determined based on ratings of gradually increasing stimulus intensities, as previously accomplished.^18–20 Individual pain thresholds served as anchor for the subsequent calibration and matching procedures. During calibration, stimulus intensities for each pain modality were identified based on perceived pain intensity ratings within the target range of 60 to 80 mm on digitised visual analogue scales [VAS] with endpoints labelled ‘not painful’ [0] and ‘extremely painful’ [100], respectively. This procedure allows the determination of individual stimulation intensities [distension pressure for the visceral modality and temperature for the somatic modality] inducing adequately painful sensations for use in within-subject application of repeated pain stimuli during brain imaging. Afterwards, to ensure comparability across modalities, visceral and somatic perceived pain intensities were matched. For this purpose, visceral stimuli were presented together with thermal stimuli and participants were asked to compare the stimuli using Likert-type response options indicating more, less, or equally painful stimuli. If the rating showed a deviation, the intensity of thermal stimuli was successively adjusted until ratings indicated equal perception at least twice consecutively. As in our previous work,¹⁸ stimulus durations were adjusted for each individual, aiming at matched durations of ascending [increasing pressure and temperature, respectively] and plateau phases [stable pressure and temperature, respectively] of visceral and thermal stimulation [total stimulus duration per stimulus: 20 s]. The stimulation intensities resulting from the matching procedure in patients with UC were comparable to those in healthy volunteers and patients with IBS, for both thermal stimuli [UC: 45.43 ± 0.36°C; HC: 45.20 ± 0.30°C; IBS: 43.83 ± 0.70°C] and rectal distensions [UC: 32.60 ± 1.87 mmHg; HC: 38.16 ± 2.25 mmHg; IBS: 34.78 ± 2.15 mmHg].

Figure 1.

[A] Schematic and [B] narrative overview of experimental procedures. A detailed description of pain models and implementation [eg, pseudorandomisation, timing] is provided in Sections 2.3 and 2.4, respectively.

Subsequently, a structural brain scan was accomplished to exclude brain abnormalities and to ensure optimal settings [duration: ~5 min], immediately followed by the multiple-threat paradigm inside the MR scanner. Functional MRI was conducted to assess blood oxygenation level-dependent [BOLD] responses, indicating neuronal activity, to a total of 10 painful stimuli [five visceral, five somatic, implemented in pseudorandomised order], using the calibrated and matched pain stimulus intensities. Prior to stimulus presentation, a fixation cross was presented on the screen [with jittered durations 6–12 s], and after each stimulus, participants rated the perceived pain intensity on digitised VAS [presented for 9 s] as described above, using an MR-compatible hand-held fibre-optic response system [LUMItouchTM, Photon Control, Burnaby, BC, Canada]. An OFF phase [ie, black screen with a white frame] with jittered durations between 4.6 and 6.9 s followed each rating.

2.5. Brain imaging and analyses

Structural and functional brain images were acquired on a 3 Tesla MR scanner using a 32-channel head coil [Skyra, Siemens Healthcare, Erlangen, Germany]. For structural imaging, a 3D-MPRage T1-weighted sequence was used (repetition time [TR] 1900 ms, echo time [TE] 2.13 ms, flip angle 9°, field of view [FOV] 239 × 239 mm², 192 slices, slice thickness 0.9 mm, voxel size 0.9 × 0.9 × 0.9 mm³, matrix 256 × 256 mm², Generalized Partially Parallel Acquisitions [GRAPPA] r = 2). Note that the structural MR images acquired herein were used in a larger voxel-based morphometry analysis to examine structural brain alterations in chronic visceral pain conditions [for details, see Öhlmann et al.⁷]. Functional images for the assessment of BOLD responses were acquired with a single-shot echo-planar imaging [EPI] sequence [TR 2300 ms, TE 28.0 ms, flip angle 90°, FOV 220 × 220 mm², matrix 94 × 94 mm², GRAPPA r = 2 with 38 transversal slices angulated in the direction of the corpus callosum, slice thickness of 3 mm, slice gap 0.6 mm, voxel size 2.3 × 2.3 × 3.0 mm³]. The images were preprocessed and analysed with SPM12 [Wellcome Center for Human Neuroimaging, UCL Queen Square Institute of Neurology, London, UK] implemented in Matlab R2021b [MathWorks, Natick, MA, USA].

In preparation for the co-registration and normalisation of functional images, the CAT12 toolbox [Computational Anatomy Toolbox, Structural Brain Mapping group, Jena University Hospital, Jena, Germany] implemented in SPM12 was used for spatial registration to the Montreal Neurological Institute [MNI] template and segmentation into grey matter, white matter, and cerebrospinal fluid. Then, functional images were co-registered to the skull-stripped individual T1-weighted image, normalised to MNI space using a standardised template implemented in SPM12, and spatially smoothed with an isotropic Gaussian kernel of 6 mm. The standard motion correction and realignment procedure implemented in SPM12 was applied to the functional images, resulting in the estimation of six realignment parameters describing the rigid body transformation between each image and the mean image used as a reference image. A temporal high-pass filter of 128 s was used to correct for low frequency drifts in the data. Serial autocorrelations were accounted for by means of an autoregressive model first-order correction.

For statistical first-level analyses, a General Linear Model [GLM] was applied to each subject’s EPI images with regressors modelled based on a box car convolved with a canonical haemodynamic response function. For the first-level GLM [ie, computation of individual changes in BOLD responses], two regressors of interest were entered, ie, one for visceral stimulation [PAIN_VISC] and one for somatic stimulation [PAIN_SOM]. As regressors of no interest, presentations of the fixation cross prior to pain onset and ratings were included with the respective onsets and durations [see Section 2.3] and OFF phases were modelled as a stick function. Finally, realignment parameters were included as nuisance regressors for motor correction. To analyse BOLD responses to visceral compared tosomatic painful stimulation [ie, differential neural responses], the first-level contrasts PAIN_VISC > PAIN_SOM and PAIN_VISC < PAIN_SOM were computed and entered into second-level [ie, group-level differences in BOLD responses] analyses. One-sample t-tests were used to assess modality-specific effects within each group. Using small volume correction, these analyses were performed in a priori defined regions of interest [ROI], including the anterior and posterior insula, subregions of cingulate cortex [anterior cingulate cortex, midcingulate cortex, posterior cingulate cortex], thalamus, pre- and postcentral gyrus, supplementary motor area, basal ganglia [putamen, pallidum, caudate nucleus], amygdala, hippocampus, and prefrontal regions. All of these regions have previously been shown to be involved in the central processing and modulation of visceral pain,^21,41 with most regions being also involved in somatic pain processing.⁴² The unilateral anatomical templates used for the ROI-analyses were derived from the WFU Pick Atlas toolbox Version 2.5.2,⁴³ which incorporates the automatic anatomical labelling [AAL] atlas. Segmentation of the insula [anterior, posterior] was based on previous literature.⁴⁴ For all analyses, significant results are reported corrected for multiple comparisons (using family-wise error [FWE] of alpha, set at p <0.05). All coordinates refer to MNI space, and plots were prepared using the software MRIcroGL [version 1.2.20210317].

2.6. Statistical analyses of behavioural data

Data were analysed using IBM SPSS Statistics 27 [IBM Corporation, Armonk, NY]. Group comparisons of questionnaire data as well as of pain thresholds were accomplished using independent samples t-tests. For mean perceived pain intensity, a mixed analysis of variance [ANOVA] was computed with group [UC/HC/IBS] as between-subjects factor and modality [visceral/somatic] as within-subjects factor, using paired t-tests for post-hoc testing. Exploratory correlational analyses between behavioural pain-related measures and clinical variables [ie, GI symptoms, pain severity, anxiety, depression, and chronic stress] were accomplished using Pearson’s r. All data are reported as mean ± standard deviation [SD], unless indicated otherwise, and plots were prepared with the ggplot2 package⁴⁵ using R [Version 4.1.1; R Core Team, Vienna, Austria] in RStudio [Version 2021.09.0 + 351; RStudio, Boston, MA].

3. Results

3.1. Sociodemographic, clinical, and psychological characteristics

Sample characteristics for UC [N = 20] and IBS [N = 23] patients and for healthy controls [N = 25] are summarised in Table 1. In line with the recruitment strategy, symptom-based CAI of 0.60 [SD = 0.82; range: 0–3] and a mean faecal calprotectin concentration of 37.23 μg/g [SD = 23.33; range: 7.99–105.30] confirmed clinical and biomarker remission in UC patients. Colitis-related medications as prescribed by the treating physician included aminosalicylates [N = 11], local corticosteroids [N = 1], TNF-α blocker [N = 2], and azathioprine [N = 2]. The IBS cohort presented with different bowel habit disturbances [N = 9 diarrhoea-predominant, N = 4 constipation-predominant, N = 9 mixed, and N = 1 unspecified]. IBS-related medications included selective serotonin reuptake inhibitors [N = 1], muscarine receptor antagonists [N = 2], and loop diuretics [N = 1]. Numbers of women reporting vaginal deliveries [≥1] were as follows: UC [N = 7], IBS [N = 5], HC [N = 2], with perineal tears [UC: N = 2; IBS: N = 2; HC: N = 0] or episiotomy [UC: N = 2; IBS: N = 0; HC: N = 0]. Note that as per exclusion criteria, digital rectal examination revealed no abnormalities in any participant and no participant reported fecal incontinence. Reported disease duration was comparable in the UC and IBS cohorts [UC: 2–28 years; IBS: 1–26 years].

Table 1.

Clinical and psychological characteristics

	HC [N = 25]	UC [N = 20]	IBS [N = 23]	pUCvsHCa	pUCvsIBSa
Gender [% female]	100	100	100	N/A	N/A
Age, years	42.08 ± 12.99	39.20 ± 12.71	46.91 ± 10.92	0.459	0.038*
BMI	23.27 ± 2.67	23.03 ± 3.02	23.13 ± 4.02	0.787	0.926
Gastrointestinal symptoms	3.76 ± 3.03	9.45 ± 6.39	15.09 ± 4.88	0.001**	0.002**
Pain severity [BPI]	1.32 ± 3.39	5.53 ± 7.92†	9.48 ± 8.72	0.041*	0.136
Anxiety [HADS-A]	3.68 ± 2.48	6.95 ± 3.09	9.09 ± 3.27	<0.001***	0.034*
Depression [HADS-D]	3.00 ± 2.45	3.50 ± 2.42	6.00 ± 3.30	0.497	0.008**
Chronic stress [TICS]	14.88 ± 9.22	20.85 ± 7.51	25.17 ± 8.58	0.024*	0.089

^a Results of two-tailed independent samples t-tests comparing patients with ulcerative colitis [UC] to healthy controls [HC] and patients with irritable bowel syndrome [IBS], respectively. Data are shown as mean ± standard deviation. Published cut-off for HADS-A and HADS-D ≥11, respectively. For TICS, norm values are available with a mean score of 13 corresponding to T = 50.

BMI, body mass index; BPI, Brief Pain Inventory; HADS, Hospital Anxiety and Depression Scale; HADS-A, HADS Anxiety Subscale; HADS-D,  HADS Depression Subscale; N/A, not applicable; TICS, Trier Inventory of Social Stress.

†N = 19. *p < 0.05, **p < 0.01, ***p < 0.001.

As expected, groups differed significantly in the amount of GI symptoms and pain severity, with UC patients reporting more symptoms and more severe pain than healthy volunteers but less symptoms than patients with IBS. Regarding psychological characteristics, compared with healthy volunteers, UC patients reported higher levels of anxiety and chronic stress, whereas there was no significant difference in symptoms of depression. Compared with IBS patients, both anxiety and depression symptoms were significantly lower in UC patients. Levels of chronic stress were comparable in both patient groups.

3.2. Pain sensitivity in visceral and somatic modality

No significant group differences in pain thresholds were observed for the visceral [HC: 40.20 ± 12.12 mmHg; UC: 34.00 ± 10.46 mmHg; IBS: 37.83 ± 10.09 mmHg] or the somatic modality [Figure 2A; HC: 45.17 ± 2.33°C; UC: 45.93 ± 2.39°C; IBS: 44.18 ± 3.83°C]. However, a trend for lower visceral pain thresholds was observable in patients with UC compared with healthy volunteers (t[43] = 1.81, p = 0.077). Visceral and somatic pain thresholds were not intercorrelated in any group. Correlational analyses in patients with UC revealed negative correlations between visceral pain thresholds and GI symptoms [r = -.47, p = 0.037] and with levels of perceived chronic stress [r = -.48, p = 0.033], respectively [Figure 2B], whereas no significant correlations were found for somatic pain thresholds. In contrast, in both control groups [ie, IBS patients and healthy volunteers] there were no significant correlations of visceral and somatic pain thresholds with clinical and psychological characteristics.

Figure 2.

[A] Visceral and somatic pain thresholds. [B] For patients with UC, correlations are shown between visceral pain threshold and gastrointestinal symptoms and chronic stress, respectively. Error bars depict standard error. HC, healthy controls; IBS, irritable bowel syndrome, UC, ulcerative colitis.

3.3. Perceived pain intensity in visceral and somatic modality

With regard to perceived pain intensity, the mixed ANOVA revealed a significant interaction between group and modality (F[2,65] = 4.634, p = 0.013, η² = .13). Post-hoc paired t-tests demonstrated a significant difference between mean visceral and somatic perceived pain intensities for healthy volunteers (t[24] = 2.14, p = 0.042, d = 0.43) and patients with IBS (t[22] = 2.42, p = 0.024, d = 0.51). Both control groups reported visceral pain intensity [HC: 67.37 ± 13.55 mm; IBS: 62.87 ± 16.59 mm] to be higher than somatic pain intensity [Figure 3; HC: 63.18 ± 12.34 mm; IBS: 54.38 ± 18.13 mm]. In contrast, for patients with UC there was no significant difference between visceral [61.32 ± 15.85 mm] and somatic [66.36 ± 14.03 mm] perceived pain intensities (t[19] = -1.31, p = 0.208; Figure 3). Correlational analyses revealed no significant associations of visceral or somatic perceived pain intensities with clinical and psychological characteristics within either UC patients or healthy controls. In patients with IBS, visceral perceived pain intensity positively correlated with anxiety symptoms [r = 0.55, p = 0.007] and chronic stress [r = 0.44, p = 0.037]. Higher somatic perceived pain intensity was also associated with higher levels of anxiety symptoms in IBS [r = 0.49, p = 0.017].

Figure 3.

Visceral and somatic perceived pain intensities are depicted averaged across trials. Error bars depict standard error. Statistical labelling corresponds to results of post-hoc paired t-tests comparing mean visceral and somatic perceived pain intensities within each group [*p <0.05]. HC, healthy controls; IBS, irritable bowel syndrome; UC, ulcerative colitis; VAS, visual analogue scale.

3.4. Neural processing of visceral and somatic pain

Results regarding the neural processing of visceral and somatic pain are summarised in Tables 2 and 3, and a visualisation is provided in Figure 4. Briefly, enhanced differential neural responses to visceral compared to somatic pain [PAIN_VISC > PAIN_SOM] were observed in healthy volunteers, mainly in subregions of the insula and cingulum as well as sensory-motor related regions including the postcentral gyrus, thalamus, and supplementary motor area. Comparable response patterns involving several of these regions were seen in patients with IBS, whereas no significant differential neural responses in any ROI were found for this contrast in patients with UC [Table 2]. Conversely, UC patients showed enhanced differential responses to somatic compared with visceral pain [ie, in the reverse contrast PAIN_VISC < PAIN_SOM], involving the amygdala, hippocampus, subregions of the insula and cingulum, as well as sensory-motor related regions such as the postcentral gyrus, thalamus, and supplementary motor area [Table 2]. Enhanced differential responses to somatic compared with visceral pain stimuli were further observed in the hippocampus, posterior insula, and frontal regions in healthy volunteers, and altered involvement of the posterior insula, pre- and postcentral gyri was detectable in IBS patients [Table 3].

Table 2.

Differential neural responses to PAIN_VISC > PAIN_SOM

Group	ROI	H	x	y	z	t	pFWE	kE
HC	Anterior insula	L	-42	4	9	6.54	0.001	1309
		R	43	6	10	8.68	<0.001	1034
	Posterior insula	L	-33	3	8	5.78	0.003	585
		R	40	4	9	5.42	0.005	173
	Anterior cingulate cortex	L	-3	17	30	5.32	0.013	435
		R	1	22	26	5.02	0.023	165
	Midcingulate cortex	L	-10	-1	44	7.97	<0.001	2758
		R	13	9	39	6.22	0.003	1668
	Thalamus	L	-15	-14	1	5.15	0.014	251
	Postcentral gyrus	L	-61	-23	24	5.52	0.024	1240
	Inferior frontal gyrus [opercular part]	L	-42	5	9	6.9	<0.001	371
		R	39	9	7	8.17	<0.001	611
	Supplementary motor area	L	-11	-2	45	8.03	<0.001	1367
		R	2	2	47	6.11	0.004	1206
	Putamen	L	-27	-6	14	6.35	0.001	1013
UC	n.s.
IBS	Anterior insula	L	-33	5	13	5.6	0.008	361
	Posterior insula	L	-36	4	12	5.36	0.008	72
	Inferior frontal gyrus [opercular part]	R	40	13	8	5.07	0.027	107
	Supplementary motor area	L	-10	-7	66	5.99	0.007	563
	Putamen	L	-26	-15	1	5.42	0.01	39
	Pallidum	L	-25	-15	1	6.6	<0.001	46

FWE, family-wise error; H, hemisphere; HC, healthy control; IBS, irritable bowel syndrome; k_E, cluster size; L, left; R, right; ROI, region of interest; UC, ulcerative colitis; x/y/z, Montreal Neurological Institute [MNI] coordinates.

Table 3.

Differential neural responses to PAIN_VISC < PAIN_SOM

Group	Brain region	H	x	y	z	t	pFWE	kE
HC	Hippocampus	R	23	-14	-16	5.35	0.008	386
	Posterior insula	R	38	-15	20	5.32	0.007	77
	Precentral gyrus	R	30	-23	68	5.32	0.031	232
	Superior frontal gyrus [medial]	L	-11	34	57	5.05	0.047	68
	Superior frontal gyrus [dorsolateral]	L	-13	32	58	5.45	0.026	413
UC	Amygdala	L	-30	2	-23	5.37	0.005	65
	Hippocampus	L	-27	-16	-21	6.24	0.004	303
		R	28	-6	-25	6.19	0.005	250
	Anterior insula	R	50	8	-8	5.25	0.027	50
	Posterior insula	R	38	-10	16	6.3	0.003	558
	Midcingulate cortex	R	1	-6	33	5.54	0.034	22
	Precentral gyrus	L	-47	-7	30	5.68	0.043	291
	Inferior frontal gyrus [triangular part]	R	47	32	18	5.44	0.039	116
	Middle frontal gyrus	L	-46	33	29	5.8	0.048	134
	Superior frontal gyrus [dorsolateral]	R	26	-11	67	6.15	0.022	137
	Supplementary motor area	R	5	-23	57	5.69	0.028	193
	Putamen	R	18	7	-6	5.42	0.02	83
	Pallidum	R	18	6	-5	4.73	0.018	21
IBS	Posterior insula	L	-41	-17	13	5.47	0.007	63
		R	38	-17	22	6.44	0.001	196
	Postcentral gyrus	L	-65	-8	26	6.05	0.011	482
		R	66	-6	26	5.84	0.017	101
	Precentral gyrus	R	29	-22	58	6.24	0.007	1225

FWE, family-wise error; H, hemisphere; HC, healthy controls; IBS, irritable bowel syndrome; k_E, cluster size; L, left; R, right; ROI, region of interest; UC, ulcerative colitis; x/y/z, Montreal Neurological Institute [MNI] coordinates.

Figure 4.

Differential functional brain responses to visceral and somatic stimuli. For visualisation purposes, activations were superimposed on a structural T1-weighted MR-image, thresholded at p <0.01 uncorrected, and masked for the a priori defined regions of interest. Colour bars indicate t scores. HC, healthy controls; IBS, irritable bowel syndrome; UC, ulcerative colitis; z/x, Montreal Neurological Institute [MNI] coordinates.

4. Discussion

Although recurring symptoms of pain [especially yet not exclusively] arising from the viscera shape the clinical reality of patients with IBS and IBD alike, mechanisms relevant to pain experience and reporting remain incompletely understood, particularly in IBD. The present experimental study tested pain sensitivity for two pain modalities and employed a multiple-threat paradigm with visceral and somatic pain stimuli for the first time in patients, assessing pain responses at the behavioural and neural levels in women with UC in remission, compared to women with IBS as a patient control group, and healthy women. Regarding pain sensitivity, no significant group differences in pain thresholds emerged for either pain modality, although a trend was observed for lower visceral pain thresholds in UC patients. Further, lower visceral pain thresholds correlated with greater levels of self-reported GI symptoms and chronic stress in women with UC, suggesting that there may exist subgroup[s] of patients at risk for visceral hypersensitivity in phases of clinical remission, expanding on earlier findings supporting visceral hypersensitivity during active disease⁸ or in IBD with persistent symptoms during remission, ie, during inactive disease.⁹ The association between visceral sensitivity and chronic stress is novel and intriguing, as it adds to existing knowledge on the broad role of stress in shaping GI symptoms along the gut-brain axis in both acute and chronic visceral pain,⁴⁶ including IBD.^47,48 Longitudinal studies identified chronic stress and stress-related psychological symptoms as risk factors for disease exacerbation.^49,50 Given that the stress mediator cortisol specifically reduced visceral but not somatic pain thresholds in healthy volunteers, especially in women,²⁵ the present findings indicate that increased visceral pain sensitivity may be related to stress in UC in full remission, calling for larger studies to further elucidate mechanisms relevant to interindividual variability in persisting hyperalgesia beyond the resolution of acute inflammation in UC, as these may contribute to altered symptom perception and reporting.

Albeit pain sensitivity, as assessed by thresholding procedures, provides a clinically relevant marker in chronic pain conditions, it does not fully capture the multiple facets relevant to the experience of recurring pain as a salient and complex biopsychosocial threat. The present study is the first to employ an experimental paradigm modelling the experience of multiple aversive bodily threats from somatic and visceral pain modalities in patients with chronic visceral pain. For the UC cohort, our findings support distinct alterations in behavioural and neural measures of pain processing when compared with both IBS patients and healthy control groups. Specifically, upon experience of repeated acute pain stimuli from two modalities, control groups revealed higher pain reports and greater differential pain-induced neural activation in insular, cingulate, and somatosensory cortices in response to visceral pain stimuli, despite initial matching of stimulation intensities to perceived pain intensity. These corresponding findings at behavioural and neural levels corroborate earlier results supporting modality-specific responses for the visceral pain modality, thus far exclusively gathered in healthy samples.^{13,17,18,20,21,25} Together, these results indicate prioritised visceral over somatic pain processing, consistent with the notion of a greater biological salience of visceral pain. In contrast to both comparison groups, women with UC failed to show a similar pattern of responses, ie, lacked prioritisation of the visceral modality, indicating that visceral and somatic pain constituted equally salient threats. When interpreting this disease-specific finding and its possible clinical implications, it is important to consider that a prioritisation of visceral over somatic pain is likely an evolutionary-driven, adaptive response, which is fundamental to behavioural flexibility in the face of multiple threats.²⁰ In other words, adaptive responses, such as avoidance behaviour driven by fear of pain, are normally selectively scaled and dynamically changed based on the degree or imminence of danger.⁵¹ Our results in UC could indicate a maladaptive lack of behavioural flexibility in the face of pain from multiple bodily sites. This is consistent with the fact that the experience of clinical pain symptoms in IBD is not confined to the GI tract, but commonly also arises from the musculoskeletal system or the joints as part of extraintestinal disease manifestations which are related to disease activity.³ Hence, based on previous symptom experience and learning, somatic pain may comprise a higher threat value and greater salience in IBD than in IBS, where visceral pain is the defining symptom. Although speculative, this could imply that in IBD, the experience of somatic pain contributes to stress and worry, negatively impacting on coping strategies and quality of life. In turn, the experience of multiple symptoms could enhance patients’ vulnerability to being caught in the vicious feedback cycle of symptom perception, fear and stress responses, hypervigilance, and sensitisation, ultimately increasing relapse risk.⁵² Psychological factors such as chronic stress, depression, or anxiety symptoms, which constitute key components of this feedback cycle and demonstrably modulate the evaluation and central processing of pain in health and other chronic pain conditions including IBS,^41,46 could also play a role herein. Given low depression scores in this UC cohort, it is difficult to draw conclusions regarding the possible role of depression or other psychiatric comorbidities. The observed associations between psychological variables and pain intensity reporting in IBS may indicate a disease-specific impact of psychological factors which should be addressed in future studies in a patient cohort with a broader range of psychological symptoms. Importantly, the most obvious distinction between IBD and IBS pathophysiology is chronic inflammation, which may explain disease-specific alterations observed herein. Effects of acute peripheral inflammation on brain and behaviour are well documented, as illustrated by a wealth of research into sickness behaviour [such as mood disturbances, anhedonia, fatigue, mild cognitive impairment] with hyperalgesia as a key symptom.^53,54 Besides acute or short-term effects, which are likely less important to UC in full remission, there exist data supporting long-lasting effects of repeated intestinal inflammatory insults. Such long-term effects may persist beyond—or even emerge after—the resolution of acute inflammation, as suggested by our recent work supporting effects on visceral pain-related fear memory processing induced by pain-related conditioning during endotoxaemia [an experimentally-induced inflammatory state].²³ In animal models of colitis, behavioural responses associated with abdominal discomfort and anxiety-related behaviours have been documented in rats 50 days after repeated inflammatory challenges.⁵⁵ Changes in afferent sensory pathways and the central nervous system [CNS] have also been documented after the resolution of intestinal inflammation, including persistent central inflammation,⁵⁶ and widespread changes in the neural response to acute experimental visceral pain including thalamocortical sensitisation.⁵⁷ In patients with IBD in remission, functional brain imaging studies support altered task-related brain responses in the processing of emotional stimuli, ie, decreased sensitivity to positive emotions,⁵⁸ in the context of stress,^59,60 and during the anticipation of pain.⁶¹ Our data complement and expand on this work by indicating alterations in the processing of acute pain from multiple modalities, together extending existing knowledge about CNS alterations in IBD thus far mainly focused on altered brain structure [eg,^7,62–64] and resting state functional connectivity [eg,^62,65,66].

Our results should be interpreted in light of several limitations. Our UC patient cohort exclusively comprised women who were in full clinical remission. Hence, our findings may not generalise to male UC patients nor to patients with active disease. We included IBS as a patient control group with similar GI symptoms yet no history of overt intestinal inflammation, allowing indirect deductions regarding the putative role of inflammation. However, we had no endoscopic or histological data and did not measure additional inflammatory markers beyond faecal calprotectin. Low-grade inflammation, as it is detectable in IBD during clinical remission, eg, based on analyses of cytokine concentrations in plasma,⁶⁷ and has also been discussed to play a role in IBS⁶⁸ could not be evaluated. Given several possible immune-mediated mechanisms for the persistence of GI symptoms and hyperalgesia along the gut-brain axis, including increased mucosal barrier permeability, reduced bacterial diversity, proinflammatory microbiome, or ongoing subclinical mucosal inflammation,⁶⁹ future studies should ideally combine peripheral and central measures relevant to visceral pain, as recently accomplished in IBS.^70–72 Clearly, both IBD and IBS are complex and heterogeneous clinical conditions involving the gut-brain axis. Albeit a challenging endeavour, the comparison of two patient cohorts with similar pain-related phenotypes yet distinct pathophysiology holds much promise, as illustrated by work on alterations in central pain processing in other chronic inflammatory conditions unrelated to the gut-brain axis.⁷³ Therefore, future studies are needed to further clarify distinct and common pain mechanisms, ideally in larger samples that would also be more conducive to correlational or subgroup analyses addressing the putative role of disease characteristics, including disease extent and durations, or medications. Our first attempt to highlight possible disease-specific alterations in pain processing can contribute to the ongoing debate on overlap and differences of IBD and IBS.^10,74 Beyond a better knowledge about the gut-brain axis in patients, such research also has implications for treatment. Although a variety of evidence-based psychological interventions is available for IBS,⁷⁵ treatment of IBD is still mainly focused on the induction of remission in patients with active disease, and substantially less attention is given to improvement of persistent symptoms and quality of life during phases of disease remission.^69,76 Interventions proven to be effective in IBS are often adopted for patients with IBD in remission. However, at least from the perspective of acute pain processing, IBD and IBS show distinct features, questioning this practice and calling for interventions specifically tailored to the needs of IBD patients in remission, considering extraintestinal manifestations including somatic pain. In fact, previous studies often provide mixed evidence regarding psychological interventions for IBD,⁷⁷ and different coping styles have been shown to be beneficial in patients with IBD versus patients with IBS.⁷⁸ A better understanding of the long-lasting effects of repeated inflammatory bouts on the processing of aversive bodily threats might provide new starting points for the development of interventions specifically designed for patients with IBD in remission suffering from persistent pain and related GI symptoms. With respect to experimental research approaches, including brain imaging studies, we modelled the experience of multiple threats arising from the viscera and other bodily sites. Prospectively, it is important to integrate these different lines of research by implementing paradigms to investigate emotional and cognitive aspects specifically in the context of acute pain in IBD. This will help to generate a comprehensive understanding of the subjective experience of pain in IBD, broadening the hitherto existing focus on pain sensitivity, and further knowledge about the complex gut-brain axis at the interface of neurogastroenterology, psychological pain research, and the neurosciences.

Data availability

All fMRI data analysed for the current study are available in the neurovault repository [https://neurovault.org/collections/NOGXSZBF/] and behavioural data are provided in the main manuscript; additional data and information upon request.

Funding

This work was supported by funding from the Deutsche Forschungsgemeinschaft [DFG, German Research Foundation [grant numbers 316803389—SFB 1280, 422744262—TRR 289]. The funding organisation was not involved in study design; in collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the article for publication.

Conflict of Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Author Contributions

LRL, AI, NT: acquired data. SE, AI, HE, JL, and LRL: designed the study. HÖ, LRL, AI, and NT analysed the data. HÖ, AI, and SE: interpreted the data. HÖ, AI, and SE: wrote the first draft of the paper. SE and HE: acquired funding. All authors contributed to the interpretation of the data, revised the manuscript for critical content, and approved the final version of the manuscript.

Acknowledgements

The authors would like to thank Dr Marcel Gratz for excellent technical support and Sopiko Knuf-Rtveliashvili for support in data acquisition.

Author notes

Adriane Icenhour and Sigrid Elsenbruch These authors contributed equally to this work.