Chronic continuous abdominal pain: a large cohort-specific evaluation of diagnostic features, iatrogenesis and drug treatments.

Kilgallon E 1, Vasant D.H. 2,3, Green D 4 Shields P.L. 5, Hamdy S1,2, Lal S 1,2, Paine

P.A. 1,2

1 Gastroenterology, Salford Royal NHS Foundation Trust, Salford, United Kingdom2Division of Diabetes, Endocrinology and Gastroenterology, University of

Manchester, Manchester, United Kingdom3 Manchester University Foundation Trust, Wythenshawe Hospital, Manchester,

United Kingdom4 Division of Cardiovascular Sciences, University of Manchester5 Lancashire Teaching Hospitals, Preston, United Kingdom

Short title : Chronic continuous abdominal pain

Total word count (excluding abstract/summary): 4,554

Grant Support: No external grant support was obtained for this study.

Abbreviations: CCAP – chronic continuous or nearly continuous abdominal pain,

CAPS – Centrally mediated abdominal pain syndrome, SD – Standard deviation,

Abdominal Allodynia is short form for ‘cutaneous abdominal mechanical dynamic

allodynia’.

Correspondence: Dr Peter A Paine PhD MRCP, Gastroenterology Department,

Salford Royal NHS Foundation Trust, Stott Lane, Salford, Greater Manchester,

United Kingdom. E-mail: peter.paine@srft.nhs.uk, Telephone number: +44 (0)161

2064560

Disclosures: None of the authors have any relevant financial disclosures.

Writing Assistance: No writing assistance was required.

Author Contributions: EK gathered data and helped write the manuscript, DHV

collected data, helped with data analysis and helped write the manuscript, DG

helped with data analysis and helped write the manuscript, PLS SH and SL helped

with study design and manuscript appraisal, PAP conceived the study, helped with

data collection, analysis and manuscript writing.

2

Summary:Background: Chronic continuous abdominal pain (CCAP) is characteristic of centrally

mediated gastrointestinal pain disorders. It consumes significant healthcare

resources yet is poorly understood, with minimal cohort-specific data in the literature.

Aims: In a large cohort of CCAP patients, we examined; 1. diagnostic features, 2.

iatrogenic impact of opioids and surgery. 3. drug treatment effects and tolerance.

Methods: Consecutive tertiary CCAP referrals to a neurogastroenterology clinic

(2009-2016) were reviewed for Rome IV and neuropathic pain criteria. Medical,

surgical and drug histories, interventions and outcomes were correlated with clinical

diagnosis and associated opioid use.

Results: Of 103 CCAP patients (mean age 40±14, 85% female), 50% had

physiological exacerbations precluding full Rome IV Centrally Mediated Abdominal

Pain Syndrome criteria. However, there were no significant differences between

patients who satisfied Rome IV criteria and those who didn’t. Overall, 81% had

allodynia (light brush strokes evoking pain on abdominal examination). Opioid use

was associated with allodynia (P=0.003). Prior surgery was associated with further

operations post CCAP onset (p<0.001). Although 68% had undergone surgical

interventions, surgery did not resolve pain in any patient and worsened pain in 35%.

Whilst duloxetine was the most effective neuromodulator (P=0.003), combination

therapy was superior to monotherapy (P=0.007).

Conclusions: This currently largest cohort-specific CCAP dataset supports eliciting

neuropathic features, including allodynia, for a positive clinical diagnosis, to guide

treatment. Physiological exacerbation of CCAP may represent visceral allodynia, and

need not preclude central origin. Use of centrally acting neuromodulators, and

avoidance of detrimental opioids and surgical interventions appears to predict

favourable outcomes.

Keywords: neuropathic pain, allodynia, hyperalgesia, opioids, chronic abdominal

pain, centrally acting neuromodulators, centrally mediated gastrointestinal pain

3

IntroductionSignificant advances have been made in understanding, classifying and diagnosing

common episodic painful functional gastrointestinal disorders (FGIDs) such as

irritable bowel syndrome (IBS) 1. However, less common chronic continuous

abdominal pain (CCAP), with no clear underlying structural or metabolic cause,

remains a challenging clinical entity for which very little is known to guide practice.

Such patients may present repeatedly to secondary and tertiary services, with

significant psychological co-morbidity, multiple repeated negative investigations and

heavy socioeconomic costs including healthcare utilisation and the burden of loss of

working days due to illness 2-6. These patients may be subjected to futile or counter-

productive treatments including opioid analgesia 7. Indeed, the detrimental effects of

opioids in this setting are increasingly being recognised and include nausea,

vomiting, constipation, and the development of opioid-induced hyperalgesia 2, 8-9.

Despite the significant severity and morbidity attributed to CCAP, there remain

substantial gaps in the clinical evidence base in several key areas:

Firstly, there is very limited published cohort specific data for the clinical diagnostic

and putative pathophysiological features in CCAP patients. In order to distinguish

CCAP from episodic painful functional gut disorders such as IBS, the ‘centrally

mediated abdominal pain syndrome’ (CAPS) construct has been formulated. This is

defined by Rome IV diagnostic criteria as CCAP that is minimally related to the

physiological events of eating, defecation and menses and results in a loss of day-to-

day functioning2. However, direct clinical evidence of how frequently these diagnostic

criteria are present in a 'real world' CCAP cohort is currently lacking.

Furthermore the nature of CCAP pathophysiological attribution and treatment

recommendations, including a putative role of ‘neuropathic’ pain (i.e. pain arising

from peripheral/central nerve dysfunction as opposed to normal ‘nociceptive’ pain

provoked by tissue damage) remains speculative and derivative 10. In CAPS, it is

notable that the clinical features, postulated pathophysiology of central sensitisation 11-12 and recommended centrally acting neuromodulator therapy significantly overlap

that of neuropathic pain 13 and evidence is partially extrapolated from studies in

other, non-gastrointestinal, painful neuropathic conditions including fibromyalgia,

chronic back pain and headache 2, 14-15.

4

Moreover, key clinical diagnostic features of neuropathic pain, including allodynia (a

non-painful stimulus evoking pain sensation) and hyperalgesia (pain experienced as

more painful for a given stimulus), have recently been defined by a consensus study

using a Delphi approach 13. Additionally, lower abdominal cutaneous electrical pain

thresholds as a quantitative measure of allodynia have been shown to be predictive

of response to centrally acting neuromodulators in CCAP secondary to chronic

pancreatitis 16. As an alternative to electrical pain thresholds, mechanical dynamic

allodynia is a more readily elicited physical sign in the clinic setting using simple

gentle brush strokes over the skin, and is attributed to central sensitisation in

patients with neuropathic pain 17-19. However, the prevalence of neuropathic pain

diagnostic features and the objective sign of mechanical dynamic allodynia in CCAP

are unknown.

Secondly, another major gap in the literature concerns more detailed information as

to the possible iatrogenic impacts of opioid use and of surgical interventions on

clinical outcomes in CCAP.

Finally, a third major gap in the CCAP literature is regarding pharmacological

treatment. Whilst there is increasing interest in the use of gut-brain neuromodulators

for treating functional gastrointestinal pain 14-15, there remains a paucity of 'real world'

cohort-specific data to support their use in patients with CCAP. Given the relative

rarity of the diagnosis compared to the more prevalent episodic painful functional

gastrointestinal disorders, it is unlikely that a randomised controlled trial will ever be

performed. Therefore there is an urgent need for 'real world' data evaluating

outcomes from these drugs in patients with CCAP. Interestingly, linaclotide is not

currently considered to be a gut-brain neuromodulator, but has been shown to have

some emerging evidence for visceral analgesic effects, possibly through

neuroenteric mechanisms 20-21, but its effects in CCAP are currently unknown and

were therefore evaluated in this study.

This study has therefore sought to address these evidence gaps by reviewing a large

CCAP patient dataset. In particular, this study had three main aims;

1. To elicit clinical diagnostic and putative pathophysiological features in CCAP

patients, vis-à-vis both current Rome IV Centrally Mediated and also Delphi

consensus Neuropathic Pain criteria.

5

2. To solicit any iatrogenic effects of opioids and surgery in CCAP patients.

3. To determine the responses to drug treatments in CCAP patients.

Materials and MethodsWe conducted a retrospective review of the electronic medical records of all patients

who attended a regional tertiary functional gastrointestinal clinic in the North West of

England between 1 January 2009 and 31 May 2016. Ethics approval was not

required for this study, since this was a service evaluation of existing practice

compared with Rome guidelines.

Definition of CCAP

All CCAP patients included met the same chronic continuous pain definition as used

in Rome IV - that the pain should be “continuous or nearly continuous” “with pain

occurring more or less every day” and should have been present for at least the

preceding 3 months and onset at least 6 months before diagnosis. Questionnaires

were not used, but all patients were asked the same questions in clinic as to the

frequency and duration of their pain. These were documented in the patients’ case

notes and retrospectively elicited from all patients’ charts.

Patient population:

Patients with unexplained CCAP were identified from a prospectively maintained

database of tertiary clinic referrals. Patients whose symptoms were attributable to an

alternative active primary structural or metabolic diagnosis were excluded. In line

with Rome IV CAPS criteria, patients with previous surgery, adhesions or inactive

organic diagnosis were included. Patients with the anterior cutaneous nerve

entrapment syndrome were excluded on the basis of a positive Carnett’s sign and

very focal cutaneous allodynia along the rectus sheath 22.

Data collection:

Data collected for each patient from case note review included: age, gender, clinical

diagnoses; relationship of pain to physiological events including menses, bowel

habit, stool frequency and consistency and relationship with meals; presence of

6

additional gastrointestinal symptoms; details of previous surgical interventions; co-

existing gastrointestinal, functional or chronic pain diagnoses; use of motility drugs,

opioid analgesics and patient reported response to centrally acting neuromodulators

(amitriptyline, nortriptyline, gabapentin, pregabalin, duloxetine) and linaclotide.

It was recorded as to whether the patient had indicated that the drug had helped

their pain, whether it had not helped their pain and whether they had needed to stop

due to intolerance of side effects. These were standard questions asked about drug

efficacy and tolerance to all patients in clinic.

All patients initiated on neuromodulators received the same advice, together with

their GP, to start at a low dose and up-titrate according to tolerance and response

and to use for a minimum of two to three months before concluding there was no

effect.

Whilst we were unable to systematically capture the data on dose and duration, the

standard dosing advice for these drugs from the clinic was as follows:

For amitriptyline to start at 20mg nocte and up-titrate in 10mg increments per week

to a dose between 30-50mg according to tolerance and response. Higher doses up

to 100mg nocte were permitted. If limiting daytime somnolence side effects were

encountered then the recommendation was made to switch to an equivalent dose of

nortriptyline.

For duloxetine 30mg once daily was recommended as the starting dose, up-titrating

to 30mg twice daily if required.

For gabapentin, a 300mg three times daily starting dose was recommended, up-

titrating in 100mg three times daily increments per week according to tolerance and

response to a maximum dose of 1.2g three times daily. It was recommended that

most patients would likely need between 600mg to 900mg three times daily to get

some meaningful response. If side effects were encountered at the 300mg three

times daily starting dose, then it was suggested to re-start at 100mg three times daily

and up-titrate.

For pregabalin it was recommended to start at 75mg twice daily and up-titrate in

75mg twice daily weekly increments to a maximum dose of 300mg twice daily.

Data were also collected on investigations performed and details of engagement with

nutrition, chronic pain and psychology services. Background pain scores were

recorded which had been physician recorded from verbal patient report, of the level

7

of constant background pain, with the patients given a reference range by the

physician of zero (no pain) to ten (worst pain ever).

Outcome measures

Diagnostic criteria:

Case notes were reviewed retrospectively, by investigators not involved in the

patients care, for the following diagnostic criteria: Rome IV the CAPS criteria and the

neuropathic pain criteria described by Searle et al 13 (Table 1).

Definition and examination for “abdominal allodynia”:

“Allodynia” is defined as a non-painful innocuous stimulus evoking a painful

sensation. It is a characteristic feature and biomarker of neuropathic pain (i.e. pain

due to nerve dysfunction). There are different types of allodynia including mechanical

dynamic allodynia, mechanical static allodynia, cold allodynia and heat allodynia that

reflect different types of nerve dysfunction 17. In particular, dynamic mechanical

allodynia on physical examination refers to gentle/light brush strokes over the skin

evoking pain (using either a brush or cotton swab), which is due to the central

sensitisation component of neuropathic pain 17. This has previously been used for

abdominal and pelvic skin examination 18. Derivative from the well established and

validated brush or cotton swab light pressure moving stimulus on the skin,

mechanical dynamic allodynia was assessed for in all patients by clinician fingertip

brush strokes in unaffected compared with maximally affected abdominal cutaneous

pain areas 17-18. This sign may be missed using the traditional “light pressure”

palpation abdominal examination technique. The patient was first asked to indicate

an area of abdominal skin that was not painful or felt normal. This was lightly brush-

stroked with the clinician’s fingertips to help the patient calibrate their baseline

normal sensation. The patient was then instructed to indicate if the sensation

became painful or otherwise altered as the fingertip brushstrokes were moved

systematically around the abdomen, in the same distribution as covered by

traditional light palpation. Note was made of areas where the patient described

increased pain to light finger brushstroke together with the patient reported pain

level. Patients sometimes volunteered other sensory changes such as dysasthesia

or numbness. When volunteered, these were recorded but were not otherwise

8

systematically solicited. Patients may frequently exhibit non-verbal fingertip

brushstroke evoked pain behaviours and autonomic signs in areas of cutaneous

abdominal allodynia (including moaning, grimacing, breath-holding, sweating, pallor

and nausea) but these were not systematically recorded.

Since the term ‘cutaneous abdominal mechanical dynamic allodynia’ is overly

cumbersome, it will be referred to elsewhere in the paper as the short form

‘abdominal allodynia’.

We have also occasionally referred to visceral allodynia. This is defined as painful

sensation due to nerve dysfunction arising from the lumenal viscera (as opposed to

cutaneous) provoked by innocuous physiological events or stimuli. This was not

directly tested for during this study as there is no standardised clinically available test

for this and is confined to research settings using invasive methodologies.

Statistical analysis

A comparison of patient characteristics was made between patients with chronic

continuous abdominal pain who fulfilled Rome IV CAPS criteria and those who did

not (non-CAPS). Parameters compared were age, pain score, surgery post CCAP,

previous surgery, number of tertiary appointments, gender, other functional

diagnoses, other gastrointestinal diagnoses, cutaneous abdominal allodynia, nausea

and vomiting, bloating, weight loss, and the concurrent use of opioid drugs. This was

followed by a comparison of the tolerance and efficacy of different neuromodulating

drugs between CAPS and non-CAPS patients, and of the respective symptom

burden in CAPS versus non-CAPS patients on each drug.

Within group analyses were then made for CAPS and non-CAPS patients

independently, comparing characteristics of those patients who did and did not

tolerate specific prescribed medication. These analyses were limited to cases where

each medication or combination of medications was prescribed.

In all analyses above, for continuous variables normality of distribution was

determined by Shapiro-Wilk test. Normally distributed descriptive data are presented

as mean ± standard deviation (SD), and non-normally distributed data as median

(minimum, maximum). Statistical comparisons between normally distributed

variables were made using unpaired t-tests. Statistical comparisons between non-

normally distributed variables were made using Mann-Whitney U test.

9

To compare the association of non-pharmacotherapeutic variables with CCAP

symptom burden, univariate binary logistic regression was used. Dependent

variables were the presence of abdominal allodynia and any surgery after CCAP

diagnosis, independent variables are those listed above. Analyses were performed

for CAPS and non-CAPS patients separately.

A similar comparison was performed for CCAP patients who gave a history of

concomitant (quiescent) structural gastrointestinal disease and those who did not

and for CCAP patients who gave a history of surgery prior to CCAP onset and those

who did not.

For between group comparison of clinical characteristics between patients on opioid

therapy, patients previously on opioid therapy, and opioid naive patients, univariate

binary logistic regression was performed with opioid naïve patients as the reference

group. Odds ratios were generated for the presence of each symptom or clinical

outcome in patients on or previously on opioids compared to opioid naïve patients.

For all analyses, statistically significant results are presented as p<0.05, and results

that remained significant after Bonferroni correction are highlighted as such, the α

values after correction being specific to each analysis. All analyses were performed

using the statistical software package SPSS v22.0, IBM.

ResultsClinical characteristics

103 patients with CCAP (mean age 40 14; 88/103 (85%) female) were included. At

first presentation to the clinic, only 18/103 (17%) had never used opioids (opioid

naïve), 13/103 (13%) had previously used opioids but were not currently on opioids,

and 72/103 (70%) were currently using opioids.

Just over half of patients (57/103, 55%) reported concomitant nausea and vomiting,

whilst 45/103, 44% reported symptoms of bloating and abdominal distension and

24/103, 23% reported weight loss. Most patients (90/103. 87%) tolerated an oral diet,

two patients received oral supplement drinks and three patients adhered to a liquid

diet. Six patients required enteral tube feeding, whereas two received parenteral

nutrition.

10

Background pain scores were available for 59/103 patients based on a verbally

solicited analogue score, given anchors by the clinician of 0 (no pain) to 10 (worst

pain). The mean background pain score was 4.8 1.7.

Rome IV centrally mediated diagnostic criteria

52/103 (50%) patients fully satisfied all of the Rome IV CAPS criteria (Table 1). Of

the 51/103 (50%) who did not fulfil all Rome IV CAPS criteria (non-CAPS), 50/51

reported that their constant background pain was frequently exacerbated by

'physiological events' (post-prandial (n=36), on defecation (n=8), preceding

defecation (n=5) during menstruation (n=2), with some patients reporting more than

one type of physiological trigger for exacerbation and two also reporting no loss of

function) and one reported no loss of function without physiological event

exacerbation.

A comparison was made between CCAP patients who fulfilled all Rome IV CAPS

criteria and those (non-CAPS) who did not. There were no significant differences

between these two groups in clinical characteristics (Table 2); drugs used (Table 3);

drug effectiveness (Table 3) or tolerance (Table 3). Furthermore, there were no

significant differences between the two groups with respect to effects of the

presence of abdominal allodynia, pain scores or number of operations after

diagnosis on the different pharmacotherapy regimes (supplementary tables 1-3).

There were also no differences between the two groups with respect to non-

pharmacotherapy parameters associated with abdominal allodynia (supplementary

table 4).

Neuropathic pain criteria

Data were available for four of the neuropathic pain criteria (spontaneous, difficult to

manage, abdominal allodynia, poor response to opioids) in all patients. Two

neuropathic features were present in almost all patients (‘spontaneous’ pain 100/103

(97%), 'difficult to manage' 102/103 (99%), Table 1). Notably, 83/103 (81%) had

evidence of abdominal allodynia on physical examination. Another important

neuropathic feature 'poor response to opioids', was found in the majority of those

who had taken opioids (61/81, 72%).

11

Impact of opioids

Intriguingly, the presence of abdominal allodynia on physical examination was highly

significantly associated with current opioid use (p = 0.003, Table 4). There was a

trend for previous opioid use to be associated with previous surgery, and both

previous and current opioid use to be associated with bloating symptoms, but these

did not retain significance after Bonferroni correction for multiple comparisons.

Co-morbidities, aetiological triggers and other functional disorders

Whilst the majority of patients could not identify a clear trigger for onset of their

CCAP symptoms, 35/103 patients (34%) reported that their symptoms began as a

direct result of an initiating trigger. Surgery triggered symptoms in n=19 (abdominal

surgery (n=17); spinal surgery (n=1); knee surgery (n=1)). The onset of pain was

also reported following an episode of gastroenteritis (n=7) and following both

abdominal surgery and gastroenteritis (n=1).

Additionally, other specific clear triggers of onset included (n=8); haemorrhagic

ovarian cyst, viral illness, colonoscopy, grief reaction, diverticular disease associated

colitis, road traffic accident, vasculitis, and pyelonephritis. There was no association

of opioid with identification of a clear trigger.

Overall, fifteen patients (15%) had concurrent functional or chronic pain diagnoses.

Fibromyalgia was present in 7/103 (7%) patients. Other functional or chronic pain

diagnoses included; chronic fatigue syndrome (n=4), chronic back pain (n=3),

pseudoseizures (n=1), chronic knee pain (n=1) and chronic regional pain syndrome

(n=1). Three patients had hypermobility type Ehlers Danlos Syndrome (hEDS), one

had postural orthostatic tachycardia syndrome (POTS) and a further 3 patients had

both hEDS and POTS concomitantly.

Thirty-five patients (34%), had a history of organic diagnoses. Quiescent

inflammatory bowel disease (Crohn’s disease n=8, Ulcerative Colitis n=1) and

diverticular disease (n=8) were the most prevalent diagnoses. Other organic co-

morbidities included endometriosis (n=5), gallstones (n=4), haemorrhoids (n=3),

hernia (n=2), chronic pancreatitis (n=2) and peripheral neuropathy (n=2). In all

12

patients, the referring clinician had indicated that the concomitant organic diagnosis

was not active and did not explain the chronic severe pain.

There were no significant differences between CCAP patients who gave a history of

concomitant (quiescent) gastrointestinal diseases and those who did not (Table 5).

Surgical interventions performed in patients with CCAP

Seventy patients (68%), had a background of previous abdominal surgery and 27%

of these patients identified surgical intervention as an initiating trigger to CCAP.

Overall, the most common surgical interventions were; appendicectomy (n=26/70,

37%), cholecystectomy (n=15/70, 21%) and hysterectomy (n=19/70, 27%). Histology

was not available from the referral centres in the majority of patients’ case records.

Interestingly, the majority of patients who had surgical interventions 44/70 (63%) had

at least one surgical procedure in order to identify the cause of, treat, or as a

consequence of, their pain. Twenty five of these patients had multiple surgeries, with

n=10 having 4 or more procedures. Surgery did not resolve the symptoms in any

patient and 13/44 (35%) reported that surgery resulted in a direct worsening of their

pain.

Exploratory laparoscopy/laparotomy was the most common surgical intervention

performed post-symptom onset. 10/44 patients (22%) had exploratory surgery with

no abnormal findings and 25/44 patients (56%) had a subsequent interventional

procedure.

Adhesiolysis was the most frequently performed surgical intervention, in 14/44

patients (31%), but only one patient was found to have obstruction due to adhesions.

Four patients had multiple adhesiolysis procedures, with no resolution of symptoms.

Other surgical procedures post CCAP onset included; appendicectomy (n=3, normal

histology in n=1 for whom histology was available), cholecystectomy (n=1 with

normal histology), bowel resections (n=2; including repair of an inadvertent

enterotomy during the division of adhesions) and hernia repairs (n=5). One patient

had an ERCP with sphincterotomy as their pain was thought to be attributable to

13

Sphincter of Oddi dysfunction; however, this resulted in post-sphincterotomy

pancreatitis with no improvement of symptoms.

There was a highly significant association between CCAP patients having had

surgery prior to CCAP onset going on to have more operations post CCAP onset

than CCAP patients who had never had a previous operation (p < 0.001). There

were otherwise no significant differences between CCAP patients with or without

previous surgery (Table 5).

Clinical response to centrally acting neuromodulators

Almost all patients (95/103, 92%) received centrally acting neuromodulators, with

clinical outcomes available for those that were followed-up in the tertiary

neurogastroenterology clinic (51/95, 54%), as opposed to those that were solely

discharged to their local chronic pain service. A positive response was seen in 34/51

(67%), non-response 10/51(19%) and intolerance 7/51 (14%). The tricyclic

antidepressants (TCAs) had the lowest response rates, with 7/32 (22%) responding

to amitriptyline and 5/21 (24%) responding to nortriptyline. 15/40 (38%) responded to

gabapentin and 10/26 (38%) responded to pregabalin.

Duloxetine was the only neuromodulator significantly better at achieving a response

than amitriptyline (P = 0.003 Table 6), with 12/16 (75%) responding, and was the

only drug to be tolerated by all patients in which it was tried.

Linaclotide was used in 29/103 patients (28%) with outcomes available for 21/29.

12/21 (57%) reported a positive pain symptom response, 8/21 reported no response

or worsening of symptoms and 1/21 did not tolerate linaclotide due to side effects.

There was a trend towards linaclotide being more effective and better tolerated than

amitriptyline but this did not retain significance after Bonferroni correction for multiple

comparisons (Table 6).

Patients with known outcomes were stratified into two groups according to the

number of drugs they had taken concurrently. Eight patients were excluded as data

regarding the relative time periods of the drug prescribing were not available.

Combination therapy with >1 centrally acting neuromodulator (n=12/14, 86%

14

response) was superior to centrally acting neuromodulator monotherapy (n=10/19,

53% response,), P = 0.007 (Table 7).

DiscussionIn this largest cohort-specific analysis of the diagnosis and management of CCAP to

date, our data are the first to elicit the clinical diagnostic and putative

pathophysiological features vis-à-vis the Rome IV CAPS and the Delphi neuropathic

pain criteria. In particular, we have found the presence of neuropathic pain features

including abdominal allodynia to be a frequent finding. Moreover, we have found

frequent physiological exacerbations of the chronic pain, which may reflect visceral

allodynia. Other novel and important findings include the detrimental effects of prior

surgical interventions increasing the risk for further futile surgical interventions and

opioid analgesia in CCAP associated with increased abdominal allodynia, as a

possible biomarker of opioids adversely altering pain neurobiology. To the authors’

knowledge, our outcome data on the use of centrally acting neuromodulators are

also the first in a CCAP population which support recently published empirical

guidance 14-15.

Our findings suggest the full current Rome IV CAPS criteria may exclude a

significant number of patients with CCAP. In particular, almost half of the patients in

our cohort reported constant background pain, but were also frequently exacerbated

by physiological events. We appreciate that the reason underpinning the CAPS

criteria exclusion of physiological event exacerbation is in order to try and

discriminate more effectively from the more common episodic painful functional

bowel disorders, and to highlight likely central as opposed to potentially more

peripheral neural mechanisms. We submit however that the chronic and continuous

nature of the pain itself is a sufficient discriminator from the episodic functional pain

disorders, and moreover chronic and continuous pain per se is likely to have a

central mechanism. Moreover, extensive comparisons between CCAP patients who

met all the CAPS criteria and those who did not (primarily due to frequent

physiological exacerbations) did not yield any significant differences between the two

groups.

15

Given the frequent finding of abdominal allodynia (a biomarker of central

sensitisation), and our knowledge of viscerosomatic convergence at the level of the

dorsal horn of the spinal cord, we think it is a reasonable interpretation that in the

CCAP population, physiological event exacerbation may represent visceral allodynia

whether due to peripheral and/or central sensitisation. In any case we would suggest

that excluding nearly half of patients with CCAP due to physiological event

exacerbations, from being considered as centrally mediated and thereby perhaps

less consideration of the use of centrally acting gut-brain neuromodulators, may be

clinically unhelpful as patients with CCAP do not readily fit into any other functional

gastrointestinal disorder category due to the continuous nature of their pain.

By contrast, evaluation of this patient group against neuropathic pain criteria

identified that abdominal allodynia was present in most patients with CCAP (81%).

Cutaneous mechanical dynamic allodynia to brush strokes is a well-established and

validated clinical biomarker for neuropathic pain of central origin and is a direct

consequence of central sensitization indicating the recruitment of A fibres as

primary nociceptive afferent neurons and therefore the development of a

neuropathic, amplified pain state 17-19. Fingertip brush strokes is a readily clinician-

elicited physical sign on bedside examination and allows for a positive, objective

diagnostic feature of neuropathic pain, to supplement solely symptom-based criteria.

It was sufficiently common that we believe this might be added as a positive

corroborating feature for centrally mediated abdominal pain disorders. It does

however require a change in examination technique for patients with CCAP to

include fingertip brushstroke examination to precede gentle palpation. Without

recourse to concurrent central brain imaging technology, it will not be possible to

discount possible contributions of hypervigilance to the patient sensation reports,

and this may be a useful future avenue for research.

In addition to clinical characteristics of CCAP patients, our study also explored

iatrogenesis relating to opioids and surgery. Importantly, almost three quarters of

patients who had used opioids in our study reported a poor response to these

medications. Moreover, abdominal allodynia was significantly associated with opioid

use, with patients on opioids more likely to exhibit the altered nociceptive behaviour

underpinning allodynia. Allodynia in this context therefore may also be a bio-marker

16

for opioid-induced hyperalgesia in the context of a pre-existing neuropathic pain

state. This supports the notion that opioids are counterproductively altering pain

neurobiology and should be avoided in the management of CCAP.

Increased prevalence of abdominal surgeries in patients with functional gut disorders

is well documented, as is the onset of post-surgical neuropathic pain and these

findings are mirrored in our CCAP cohort 23-26. In particular we have identified that

surgery prior to CCAP onset is associated with further surgery post-CCAP onset. We

have also identified that surgical intervention is associated with poor clinical

outcomes, and worsening symptoms in a third of patients. Moreover, our data have

shown no clinical improvement with adhesiolysis and would support findings from

published randomised controlled trials, which suggest avoidance of adhesiolysis

unless bowel viability is under threat 27-28. It should be recognised that a limitation of

this study was that we did not have access to the pathology reports of the surgical

specimens from the referring centres.

Finally after exploring clinical characteristics and iatrogenesis in CCAP patients, our

study explored response to treatments. In particular, our data lend support to the use

of centrally acting neuromodulators in patients with CCAP. Duloxetine appeared to

be the most efficacious single agent neuromodulator, whilst neuromodulator

combination therapy appeared to be more effective than monotherapy. Despite the

lack of cohort-specific data on clinical efficacy and utility of these agents, current

pharmacotherapeutic recommendations for centrally mediated gastrointestinal pain

disorders are to use neuromodulatory medications derivatively from evidence in

other painful functional gut disorders 16, 29-31. Our data thus provide additional clinical

evidence to support the recent empirical guidelines on centrally acting

neuromodulators and suggest benefit in 42% of patients with monotherapy,

increasing to 52% response using augmentation therapy with more than one agent in

combination. These 'real world' data are particularly important in light of the new

recommendations 14-15, given that most gastroenterologists have traditionally not

been accustomed to nor trained to prescribe these types of medications 14.

The efficacy of linaclotide for centrally mediated pain, both as monotherapy and in

combination with other neuromodulators in our cohort is another interesting and

17

novel finding. Linaclotide has emerging evidence as a neuromodulator, with

analgesic effects demonstrated in clinical trials in IBS patients 20, and earlier work in

animal models suggesting this may be via neuroenteric mechanisms21. Most recently

in a study using cortical evoked potentials, linaclotide was found to improve

abdominal pain in IBS by modulating afferent nerve signalling pathways to the brain 32. Our data suggest that it might be considered as an option for patients with CCAP,

particularly when chronic constipation is a co-existing feature.

We recognise that there are significant limitations to an uncontrolled retrospective

single-centre case series review; however, given the rarity and difficulties in

evidence collation for CCAP patients, we present a number of interesting and

important new observations in addition to reinforcing already existing consensus.

Interpretation of the centrally acting neuromodulator data needs to be with caution as

prescriber bias cannot be excluded. Furthermore there may be some degree of

patient selection bias in that some patients who did not respond to treatments may

be likely to present for care optimisation than those who did. Additionally, outcome

data were not available for all patients that received these medications. After initial

diagnosis of a centrally mediated pain disorder, for many patients the distance of

travel involved for this tertiary setting, meant local follow up with their local chronic

pain team to review and modify the anti-neuropathic and opioid pain medication was

more appropriate. Outcome data for neuromodulator treatment was therefore not

available for these. Moreover we were unable to collect dose and duration of

treatment data, although all patients received standard advice in these respects, as

outlined in the methods section above. Given the rarity of these cases, it will be

challenging to embark on any placebo controlled or head-to-head randomised and

blinded studies in CCAP. Furthermore, our study did not evaluate the efficacy of

psychological 6 or individualised pain management programmes, which would be

important in the future given that pain does not respond to drug therapy in nearly half

of patients. Future study of other non-drug interventions therefore, such as

neurostimulation devices 33, may also be valuable for these patients. Other areas of

future research include determining whether there are additional genotypic or

psychophysiological endophenotype markers for pain vulnerability; and more

18

thorough prospective attempts to elicit other neurobiological signs such as

cutaneous dysaesthesia/numbness and visceral allodynia/dysaesthesia.

In conclusion, this large cohort series of CCAP patients has shed light on some of

the clinical evidence gaps, and found risk factors for both poorer outcomes but also

optimised therapy. We believe these findings can help inform future iterations of

centrally mediated pain syndrome criteria as well as clinical decision making to help

avoid harm and optimise positive outcomes for this group of patients.

References:1. Drossman DA, Hasler WL. Rome IV -Functional GI Disorders: Disorders of

Gut-Brain Interaction. Gastroenterology 2016;150(6):1257-1261.

2. Keefer L, Drossman DA, Guthrie E, et al. Centrally Mediated Disorders of

Gastrointestinal Pain. Gastroenterology 2016.

3. Maxton DG, Whorwell PJ. Use of medical resource and attitudes to health

care of patients with chronic abdominal pain. Br J Med Econ 1992;2:75-79.

4. Drossman DA, Li Z, Andruzzi E, et al. U.S. householder survey of functional

gastrointestinal disorders. Prevalence, sociodemography, and health impact. Dig Dis

Sci 1993;38(9):1569-80.

5. Koloski NA, Talley NJ, Boyce PM. Epidemiology and health care seeking in

the functional GI disorders: a population-based study. Am J Gastroenterol

2002;97(9):2290-9.

6. Keefer L, Mandal S. The potential role of behavioral therapies in the

management of centrally mediated abdominal pain. Neurogastroenterol Motil

2015;27(3):313-23.

7. Szigethy E, Knisely M, Drossman D. Opioid misuse in gastroenterology and

non-opioid management of abdominal pain. Nat Rev Gastroenterol Hepatol

2018;15(3):168-180.

8. Szigethy E, Schwartz M, Drossman DA. Opiate-Induced Bowel Dysfunction.

Handbook of Gastrointestinal Motility and Functional Disorders 2015;Chapter 25:323-337.

19

9. Grunkemeier DM, Cassara JE, Dalton CB, Drossman DA. The narcotic bowel

syndrome: clinical features, pathophysiology, and management. Clin Gastroenterol

Hepatol 2007;5(10):1126-39; quiz 1121-2.

10. Clouse RE, Mayer EA, Aziz Q, et al. Functional Abdominal Pain Syndrome.

Gastroenterology 2006;130(5):1492-1497.

11. Mayer EA, Tillisch K. The brain-gut axis in abdominal pain syndromes. Annu

Rev Med 2011;62:381-96.

12. Nozu T, Kudaira M. Altered rectal sensory response induced by balloon

distention in patients with functional abdominal pain syndrome. BioPsychoSocial

Medicine 2009;3(1):13.

13. Searle RD, Howell SJ, Bennett MI. Diagnosing postoperative neuropathic

pain: a Delphi survey. Br J Anaesth 2012;109(2):240-4.

14. Sobin WH, Heinrich TW, Drossman DA. Central Neuromodulators for Treating

Functional GI Disorders: A Primer. Am J Gastroenterol 2017;112(5):693-702.

15. Drossman DA, Tack J, Ford AC, Szigethy E, Tornblom H, Van Oudenhove L.

Neuromodulators for Functional Gastrointestinal Disorders (Disorders of Gut-Brain

Interaction): A Rome Foundation Working Team Report. Gastroenterology

2018;154(4):1140-1171 e1.

16. Olesen SS, Graversen C, Bouwense SA, van Goor H, Wilder-Smith OH,

Drewes AM. Quantitative sensory testing predicts pregabalin efficacy in painful

chronic pancreatitis. PLoS One 2013;8(3):e57963.

17. Baron R. Mechanisms of disease: neuropathic pain--a clinical perspective.

Nat Clin Pract Neurol 2006;2(2):95-106.

18. Jarrell J. Demonstration of Cutaneous Allodynia in Association with Chronic

Pelvic Pain. Journal of Visualized Experiments : JoVE 2009(28):1232.

19. Lolignier S, Eijkelkamp N, Wood JN. Mechanical allodynia. Pflügers Archiv -

European Journal of Physiology 2015;467(1):133-139.

20. Quigley EM, Tack J, Chey WD, et al. Randomised clinical trials: linaclotide

phase 3 studies in IBS-C - a prespecified further analysis based on European

Medicines Agency-specified endpoints. Aliment Pharmacol Ther 2013;37(1):49-61.

21. Eutamene H, Bradesi S, Larauche M, et al. Guanylate cyclase C-mediated

antinociceptive effects of linaclotide in rodent models of visceral pain.

Neurogastroenterol Motil 2010;22(3):312-e84.

20

22. Glissen Brown JR, Bernstein GR, Friedenberg FK, Ehrlich AC. Chronic

Abdominal Wall Pain: An Under-Recognized Diagnosis Leading to Unnecessary

Testing. J Clin Gastroenterol 2016;50(10):828-835.

23. Borsook D, Kussman BD, George E, Becerra LR, Burke DW. Surgically

induced neuropathic pain: understanding the perioperative process. Ann Surg

2013;257(3):403-12.

24. de Leon-Casasola O. A review of the literature on multiple factors involved in

postoperative pain course and duration. Postgrad Med 2014;126(4):42-52.

25. Shipton E. Post-surgical neuropathic pain. ANZ J Surg 2008;78(7):548-55.

26. Longstreth GF. Avoiding unnecessary surgery in irritable bowel syndrome.

Gut 2007;56(5):608-10.

27. Molegraaf MJ, Torensma B, Lange CP, Lange JF, Jeekel J, Swank DJ.

Twelve-year outcomes of laparoscopic adhesiolysis in patients with chronic

abdominal pain: A randomized clinical trial. Surgery 2017;161(2):415-421.

28. Swank DJ, Swank-Bordewijk SCG, Hop WCJ, et al. Laparoscopic

adhesiolysis in patients with chronic abdominal pain: a blinded randomised

controlled multi-centre trial. The Lancet 2003;361(9365):1247-1251.

29. Ford AC, Quigley EM, Lacy BE, et al. Effect of antidepressants and

psychological therapies, including hypnotherapy, in irritable bowel syndrome:

systematic review and meta-analysis. Am J Gastroenterol 2014;109(9):1350-65; quiz

1366.

30. Calandre EP, Rico-Villademoros F, Slim M. Alpha2delta ligands, gabapentin,

pregabalin and mirogabalin: a review of their clinical pharmacology and therapeutic

use. Expert Rev Neurother 2016:1-15.

31. Bouwense SA, Olesen SS, Drewes AM, van Goor H, Wilder-Smith OH.

Pregabalin and placebo responders show different effects on central pain processing

in chronic pancreatitis patients. J Pain Res 2015;8:375-86.

32. Rao SS, Xiang X, Patcharatrakul T, et al. 1021 - Investigation of the

Mechanism of Action of Linaclotide on Abdominal Pain in Ibs-C: An Evaluation of the

Afferent Gut-Brain Axis. Gastroenterology 2018;154(6):S-193-S-194.

33. Kovacic K, Hainsworth K, Sood M, et al. Neurostimulation for abdominal pain-

related functional gastrointestinal disorders in adolescents: a randomised, double-

blind, sham-controlled trial. Lancet Gastroenterol Hepatol 2017;2(10):727-737.

21

Tables:

Table 1 – Proportions of patients with chronic continuous abdominal pain who met the diagnostic criteria for ‘centrally mediated abdominal pain syndrome (CAPS)’ and neuropathic pain.

CCAP, total

Criteria N=103

Rome IV CAPS criteria

(Nearly) Continuous, n (%) 103 (100)

No relationship to gastrointestinal physiological events, n (%) 53 (51)

Function loss, n (%) 100 (97)

Not feigned, n (%) 103 (100)

Not explained by an alternative diagnosis, n (%) 103 (100)

3-6 months, n (%) 103 (100)

Neuropathic Pain criteria

Spontaneous, n (%) 100 (97)

Difficult to manage, n (%) 102 (99)

Allodynia, n (%) 83 (81)

Poor opioid response, n (%) 61 (59)

Shooting, n (%) 13 (13)

Burning, n (%) 12 (12)

Dysaesthesia, n (%) 9 (9)

Hyperalgesia, n (%) 7 (7)

Good response to neuropathic agents, n (%) 35 (34)

22

Table 2. Comparison of characteristics between chronic continuous abdominal pain patients

who fulfilled all Rome IV CAPS criteria and those who did not (non-CAPS). * = statistically

significant at p<0.05 (value did not retain significance after Bonferroni correction). Key:

CAPS = centrally mediated abdominal pain syndrome; N & V = nausea and vomiting.

CAPS non-CAPS p

Age (years) 40 (18 - 80) 33 (18 - 82) 0.294

Male Gender (%) 17 8 0.17

Abdominal Allodynia (%) 26 12 0.065

N&V (%) 50 38 0.295

Bloating (%) 58 54 0.579

Weight loss (%) 85 68 0.046*

Pain score 4.9 ± 2.0 4.7 ± 1.4 0.59

Trigger (%) 74 58 0.095

Other gastrointestinal disease (%) 65 64 0.828

Other functional diagnosis (%) 85 86 0.875

Psychology referral (%) 55 40 0.135

Any previous surgery (%) 26 36 0.293

Any surgery after diagnosis (%) 47 59 0.711

Surgery after diagnosis (n) 2.2 ± 1.5 2.3 ± 1.4 0.762

Total surgery (n) 1 (0 - 9) 1 (0 - 19) 0.927

Tertiary appointments (n) 1 (1 - 12) 1 (1 - 11) 0.616

23

Table 3 A comparison of the frequency of prescription of neuromodulatory drugs and drug

combinations, the effectiveness of neuromodulatory drugs and drug combinations and the

tolerance of pharmacotherapy prescribed between chronic continuous abdominal pain

patients who fulfil Rome IV CAPS criteria and those who did not (non-CAPS). Analyses

were performed using chi square test on an intention to treat basis. Patients whose symptoms

improved after using each drug were compared to those who either did not tolerate the drug

or did not improve after its use. Patients who had never been prescribed the drug were

excluded. * = statistically significant at p<0.05 (values did not retain significance after

Bonferroni correction). Key: CAPS = centrally mediated abdominal pain syndrome; N & V =

nausea and vomiting, NM = neuromodulatory drug, L = linaclotide.

CAPS non-CAPS p

The frequency of prescription of drugs and drug combinations

amitriptyline (%) 64 36 0.097

nortriptyline (%) 53 47 1.000

gabapentin (%) 47 54 0.429

pregabalin (%) 62 38 0.250

duloxetine (%) 44 56 0.591

linaclotide (%) 52 48 1.000

Combination NM (%) 69 31 0.238

Combination NM+L (%) 56 44 0.788

Short acting opioids (%) 54 46 0.695

Long acting opioids (%) 54 46 0.823

Transdermal opioids (%) 56 43 0.641

tramadol (%) 54 46 0.823

codeine (%) 44 56 0.607

Combination morphine (%) 56 44 0.680

Combination opioids (%) 52 49 1.000

The effectiveness of neuromodulatory drugs and drug combinations

amitriptyline (%) 33 17 0.144

nortriptyline (%) 50 0 0.047*

gabapentin (%) 50 39 0.540

pregabalin (%) 60 11 0.035*

duloxetine (%) 71 78 0.766

24

linaclotide (%) 54 70 0.764

Combination NM (%) 89 100 0.318

Combination NM+L (%) 100 86 0.466

The tolerance of pharmacotherapy

N % tolerated N %

tolerated

amitriptyline 21 71 12 67 0.775

nortriptyline 10 80 9 56 0.252

gabapentin 20 75 23 78 0.801

pregabalin 15 80 9 89 0.572

duloxetine 7 100 9 100 n/a

linaclotide 11 91 10 100 0.329

NM Monotherapy 12 83 10 40 0.035*

NM combination 9 100 4 100 n/a

NM combination + linaclotide 9 100 7 100 n/a

25

Table 4 A between group comparison of clinical characteristics between patients on opioid

therapy, patients previously on opioid therapy, and opioid naive patients. Analyses were

performed using binary logistic regression with opiate naïve patients as the reference group. *

= statistically significant at p<0.05, # = retained significance after Bonferroni correction.

Key: N & V = nausea and vomiting. OR = odds ratio compared to reference group; CI =

confidence intervals.

OR 95% CI p

Abdominal Allodynia Current 7.63 1.98 – 29.42 0.003*#

Previous 1.05 0.22 – 5.00 0.951

N&V Current 1.15 0.34 – 3.71 0.810

Previous 0.54 0.11 – 2.55 0.433

Bloating Current 5.31 1.10 0 25.31 0.037*

Previous 8.40 1.27 – 55.39 0.027*

Weight loss Current 1.38 0.34 – 5.50 0.653

Previous 0.33 0.03 – 3.72 0.372

Above median pain score (4.5) Current 0.86 0.13 – 5.68 0.873

Previous 0.48 0.11 - 1.95 0.290

Trigger Current 2.38 0.61 – 9.37 0.214

Previous 1.83 0.31 – 9.37 0.498

Other gastrointestinal disease Current 0.96 0.29 – 3.21 0.951

Previous 0.90 0.18 – 4.56 0.899

Other functional diagnosis Current 0.95 0.18 – 4.95 0.959

Previous 1.20 0.14 – 10.11 0.867

Psychology referral Current 1.20 0.38 – 3.80 0.757

Previous 1.40 0.29 – 6.62 0.671

Any previous surgery Current 0.75 0.21 – 2.73 0.663

Previous 0.27 0.09 – 0.78 0.016*

Any surgery after diagnosis Current 2.76 0.78 – 9.82 0.117

Previous 0.35 0.11 – 1.17 0.089

2 or more tertiary appointments Current 1.43 0.43 – 4.79 0.561

Previous 0.72 0.25 – 2.02 0.528

26

Table 5 Comparison of characteristics between chronic continuous abdominal pain patients

who have concomitant (quiescent) gastrointestinal disease and those who do not, and who

have undergone previous surgery and those who have not. * = statistically significant at

p<0.05 (value did not retain significance after Bonferroni correction), # = retained

significance after Bonferroni correction. Key: N & V = nausea and vomiting.

yes no

Concomitant quiescent gastrointestinal disease

Age (years) 44 ± 17 38 ± 12 0.041*

Male Gender (%) 22 8 0.102

Abdominal Allodynia (%) 81 82 1.000

N&V (%) 38 47 0.658

Bloating (%) 58 54 0.509

Weight loss (%) 16 28 0.207

Pain score 4.3 ± 1.6 5.1 ± 1.8 0.114

Trigger (%) 40 33 0.503

Other functional diagnosis (%) 12 15 1.000

Psychology referral (%) 59 52 0.517

Any previous surgery (%) 72 72 1.000

Any surgery after diagnosis (%) 50 35 0.338

Surgery after diagnosis (n) 1 (0 – 5) 1 (0 – 5) 0.902

Total surgery (n) 1 (0 – 19) 1 (1 – 9) 0.954

Tertiary appointments (n) 2 (1 – 11) 1 (1 – 12) 0.337

Current opioids (%) 72 720.993

Previous opioids (%) 13 13

Previous surgery

Age (years) 41 ± 13 38 ± 18 0.378

Male Gender (%) 15 8 0.498

Abdominal Allodynia (%) 80 85 0.770

N&V (%) 53 369 0.170

Bloating (%) 41 50 0.488

Weight loss (%) 24 23 1.000

Pain score 4.9 ± 1.7 4.9 ± 2.1 0.967

Trigger (%) 74 58 0.095

27

Other gastrointestinal disease (%) 42 19 0.053

Other functional diagnosis (%) 9 27 0.043*

Psychology referral (%) 50 65 0.246

Any surgery after diagnosis (%) 62 0 <0.001*#

Surgery after diagnosis (n) 1 (1 – 5) 0 (0 – 0) -

Tertiary appointments (n) 1 (1 – 12) 1 (1 – 7) 0.616

Current opioids (%) 76 62 0.146

Previous opioids (%) 14 12

Table 6 A comparison of the tolerance and effectiveness of different neuromodulatory drugs

and linaclotide. Analyses were performed using binary logistic regression with amitriptyline

as the reference drug. * = statistically significant at p<0.05, # retained significance after

Bonferroni correction, $ = OR value >1,000 likely spurious due to small numbers. Key: OR =

odd’s ratio for effectiveness or tolerance compared to reference drug; CI = confidence

intervals.

OR 95% CI p

Effectiveness

amitriptyline (%) ref ref ref

duloxetine (%) 8.00 2.04 31.37 0.003*#

gabapentin (%) 2.11 0.80 5.59 0.133

linaclotide (%) 4.33 1.35 13.92 0.014*

nortriptyline (%) 0.95 0.27 3.41 0.940

pregabalin (%) 1.90 0.62 5.81 0.258

Tolerance

amitriptyline (%) ref ref ref

duloxetine (%) $ - - 0.998

gabapentin (%) 1.43 0.51 4.00 0.490

linaclotide (%) 8.70 1.02 74.00 0.048*

nortriptyline (%) 0.94 0.28 3.19 0.924

pregabalin (%) 2.17 0.59 8.02 0.244

28

Table 7 A comparison of the effectiveness of and clinical burden associated with combination

neuromodulatory (comb), combination therapy with linaclotide (comb + L), and

monotherapy. Analyses were performed using binary logistic regression with monotherapy as

the reference group. * = statistically significant at p<0.05, # = retained significance after

Bonferroni correction, $ = OR value >1,000 likely spurious due to small numbers. Key: N &

V = nausea and vomiting. : OR = odd’s ratio for effectiveness and clinical burden compared

to reference; CI = confidence intervals.

OR 95% CI pEffectiveness Comb 21.0 2.28 - 192 0.007*#

Comb + L $ - 0.998Abdominal Allodynia Comb 1.22 0.19 – 7.81 0.832

Comb + L 0.49 0.10 – 2.22 0.354N&V Comb 2.25 0.53 – 9.54 0.271

Comb + L 1.67 0.45 – 6.19 0.446Bloating Comb 1.40 0.35 – 5.54 0.632

Comb + L 2.64 0.69 – 10.81 0.159Weight loss Comb 1.020 0.20 – 5.21 0.981

Comb + L 1.13 0.25 – 5.12 0.871Above median pain score (4.5) Comb 1.27 0.24 – 6.82 0.778

Comb + L $ - 0.999Trigger Comb 0.64 0.15 – 2.74 0.550

Comb + L 0.87 0.23 – 3.25 0.832Other gastrointestinal disease Comb 0.43 0.09 – 2.03 0.289

Comb + L 0.87 0.23 – 3.25 0.832Other functional diagnosis Comb 0.53 0.05 – 5.68 0.598

Comb + L 1.46 0.25 – 8.40 0.671Psychology referral Comb 0.49 0.12 – 1.97 0.316

Comb + L 0.74 0.20 – 2.74 0.646Any previous surgery Comb 1.27 0.29 – 5.56 0.736

Comb + L 1.71 0.41 – 7.14 0.459Any surgery after diagnosis Comb 1.92 0.48 – 7.77 0.360

Comb + L 1.20 0.33 – 4.36 0.7822 or more tertiary appointments Comb 2.31 0.49 – 10.82 0.298

Comb + L 1.52 0.39 – 5.91 0.543

29

Supplementary Table 1. The presence of abdominal allodynia in patients with chronic

continuous abdominal pain on different pharmacotherapy regimens, comparing those who

fulfil Rome IV CAPS criteria and those who did not (non-CAPS). * = statistically significant

at p<0.05 (value did not retain significance after Bonferroni correction). Key: CAPS =

centrally mediated abdominal pain syndrome; NM = neuromodulator, L = linaclotide.

CAPS Non-CAPSp

Cases % allodynia Cases % allodynia

amitriptyline 15 67 8 88 0.278

nortriptyline 8 63 7 86 0.310

gabapentin 15 67 18 89 0.120

pregabalin 12 83 8 75 0.648

duloxetine 7 86 9 100 0.242

linaclotide 53 74 50 88 0.065

NM monotherapy 10 80 4 100 0.334

NM combination 9 78 4 100 0.305

NM combination + L 9 44 7 100 0.016*

30

Supplementary Table 2. Pain scores in patients with chronic continuous abdominal pain on

different pharmacotherapy regimens, comparing those who fulfil Rome IV CAPS criteria and

those who did not (non-CAPS). Key: CAPS = centrally mediated abdominal pain syndrome;

NM = neuromodulator, L = linaclotide.

CAPS Non-CAPSp

Mean SD Mean SD

amitriptyline 5.67 2.33 4.50 0.71 0.517

nortriptyline 3.25 1.77 4.50 1.73 0.454

gabapentin 5.05 1.92 4.86 1.03 0.784

pregabalin 5.00 2.58 4.88 0.85 0.929

duloxetine 6.10 2.97 4.79 0.76 0.28

linaclotide 4.94 1.99 4.67 1.38 0.558

NM monotherapy 5.00 2.39 4.38 0.95 0.632

NM combination 4.20 1.30 5.00 0.71 0.31

NM combination + L 7.00 1.00 5.83 0.76 0.184

31

Supplementary Table 3. Number of operations after chronic continuous abdominal pain

diagnosis in patients on different pharmacotherapy regimens, comparing those who fulfil

Rome IV CAPS criteria with those who did not (non-CAPS). * = statistically significant at

p<0.05 (value did not retain significance after Bonferroni correction). Key: CAPS = centrally

mediated abdominal pain syndrome; NM = neuromodulator, L = linaclotide.

CAPS Non-CAPSp

Mean SD Mean SD

amitriptyline 2.70 1.83 1.00 0.00 0.399

nortriptyline 1.20 0.45 1.00 0.00 0.482

gabapentin 2.20 1.75 2.63 1.19 0.566

pregabalin 2.25 1.89 2.33 1.15 0.949

duloxetine 1.67 0.58 2.50 1.97 0.51

linaclotide 2.17 1.49 2.30 1.38 0.763

NM monotherapy 1.75 1.50 1.50 0.71 0.841

NM combination 1.83 1.60 4.00 1.41 0.142

NM combination + L 3.75 1.26 1.75 0.96 0.045*

32