Content uploaded by Douglas A. Drossman
Author content
All content in this area was uploaded by Douglas A. Drossman on Jan 15, 2019
Content may be subject to copyright.
Content uploaded by Douglas A. Drossman
Author content
All content in this area was uploaded by Douglas A. Drossman on Jan 15, 2019
Content may be subject to copyright.
SPECIAL REPORT
Neuromodulators for Functional Gastrointestinal Disorders
(Disorders of GutLBrain Interaction): A Rome Foundation
Working Team Report
Douglas A. Drossman,
1,2
Jan Tack,
3
Alexander C. Ford,
4,5
Eva Szigethy,
6
Hans Törnblom,
7
and
Lukas Van Oudenhove
8
1
Center for Functional Gastrointestinal and Motility Disorders, University of North Carolina, Chapel Hill, North Carolina;
2
Center
for Education and Practice of Biopsychosocial Care and Drossman Gastroenterology, Chapel Hill, North Carolina;
3
Translational
Research Center for Gastrointestinal Disorders, University of Leuven, Leuven, Belgium;
4
Leeds Institute of Biomedical and
Clinical Sciences, University of Leeds, Leeds, United Kingdom;
5
Leeds Gastroenterology Institute, St James’s University
Hospital, Leeds, United Kingdom;
6
Departments of Psychiatry and Medicine, University of Pittsburgh, Pittsburgh, Pennsylvania;
7
Departments of Internal Medicine and Clinical Nutrition, Institute of Medicine, Sahlgrenska Academy, University of
Gothenburg, Gothenburg, Sweden; and
8
Laboratory for BrainGut Axis Studies, Translational Research Center for
Gastrointestinal Disorders, University of Leuven, Leuven, Belgium
BACKGROUND & AIMS: Central neuromodulators (antide-
pressants, antipsychotics, and other central nervous
systemtargeted medications) are increasingly used for treat-
ment of functional gastrointestinal disorders (FGIDs), now
recognized as disorders of gutbrain interaction. However, the
available evidence and guidance for the use of central neuro-
modulators in these conditions is scanty and incomplete. In this
Rome Foundation Working Team report, a multidisciplinary
team summarized available research evidence and clinical
experience to provide guidance and treatment recommenda-
tions. METHODS: The working team summarized the literature
on the pharmacology of central neuromodulators and their ef-
fects on gastrointestinal sensorimotor function and conducted
an evidence-based review on their use for treating FGID syn-
dromes. Because of the paucity of data for FGIDs, we included
data for non-gastrointestinal painful disorders and specific
symptoms of pain, nausea, and vomiting. This information was
combined into a final document comprising a synthesis of
available evidence and recommendations for clinical use guided
by the research and clinical experience of the experts on the
committee. RESULTS: The evidence-based review on neuro-
modulators in FGID, restricted by the limited available controlled
trials, was integrated with open-label studies and case series,
along with the experience of experts to create recommendations
using a consensus (Delphi) approach. Due to the diversity of
conditions and complexity of treatment options, specific rec-
ommendations were generated for different FGIDs. However,
some general recommendations include: (1) low to modest
dosages of tricyclic antidepressants provide the most convincing
evidence of benefit for treating chronic gastrointestinal pain and
painful FGIDs and serotonin noradrenergic reuptake inhibitors
can also be recommended, though further studies are needed; (2)
augmentation, that is, adding a second treatment (adding que-
tiapine, aripiprazole, buspirone a2dligand agents) is recom-
mended when a single medication is unsuccessful or produces
side effects at higher dosages; (3) treatment should be continued
for 612 months to potentially prevent relapse; and (4) imple-
mentation of successful treatment requires effective communi-
cation skills to improve patient acceptance and adherence, and to
optimize the patientprovider relationship. CONCLUSIONS:
Based on systematic and selectively focused review and the
consensus of a multidisciplinary panel, we have provided
summary information and guidelines for the use of central
neuromodulators in the treatment of chronic gastrointestinal
symptoms and FGIDs. Further studies are needed to confirm
and refine these recommendations.
Keywords: Functional Gastrointestinal Disorders; Central Neu-
romodulators; Antidepressants; Antipsychotics; Disorders of
Gut Brain Interaction; Chronic Abdominal Pain.
This Rome Foundation Working Team Report pro-
vides guidance in central nervous system (CNS)
targeted pharmacotherapy for functional gastrointestinal
symptoms and disorders (FGIDs). We recognize that the
value and utility of antidepressants and other neuro-
modulators in treating patients with these disorders are not
well understood by many gastroenterologists and other
clinicians. This may occur because their application is not
well taught in training programs or because these agents
may have stigmatizing features that result from mindbody
dualistic thinking.
1,2
New evidence is changing the thinking about these dis-
orders and their treatments. With the 2016 publication of
Rome IV, the FGIDs have been redefined as disorders of
gutbrain interaction,
2
characterized by any combination of
motility disturbance, visceral hypersensitivity, altered
mucosal and immune function, altered gut microbiota, and
Abbreviations used in this paper: CBT, cognitive behavioral therapy; CNS,
central nervous system; CVS, cyclic vomiting syndrome; D
2
, dopamine 2;
DA, dopamine; EPS, epigastric pain syndrome; FD, functional dyspepsia;
FGID, functional gastrointestinal disorder; GI, gastrointestinal; H1, hista-
mine-1; 5-HT, 5-hydroxytryptamine; IBS, irritable bowel syndrome; IBS-C,
irritable bowel syndrome with constipation; IBS-D, irritable bowel syn-
drome with diarrhea; NA, noradrenalin; NBS, narcotic bowel syndrome;
PDS, postprandial distress syndrome; SNRI, serotonin noradrenalin re-
uptake inhibitor; SSRI, selective serotonin reuptake inhibitor; TCA, tricy-
clic antidepressant.
Most current article
© 2018 by the AGA Institute
0016-5085/$36.00
https://doi.org/10.1053/j.gastro.2017.11.279
Gastroenterology 2018;154:1140–1171
SPECIAL REPORT
altered CNS processing. Dysfunction in this braingut axis
(the bidirectional neurohumoral communication between
the gastrointestinal [GI] tract and CNS) is the biologic basis
for these disorders and symptoms. The braingut axis de-
rives from a common embryologic basis: in the developing
fetus, the neural crest differentiates into the brain and spi-
nal cord, and sends down ganglia to populate the developing
endoderm, which ultimately becomes the enteric nervous
system. Thus, the nervous systems of the brain and gut are
“hardwired”; they share the same neurotransmitters and
receptors. These neurotransmitters have actions that
depend on their location, so increased serotonin in the CNS
can treat depression and in the gut can cause diarrhea. The
braingut axis with its noradrenergic, serotonergic, and
dopaminergic neurotransmitter systems is particularly
relevant with regard to gut motor functioning and visceral
pain. Thus, antidepressants will have effects not only on
psychiatric disorders, but also on chronic GI symptoms.
With this evolving understanding of gutbrain in-
teractions, it is necessary to redefine and relabel the ter-
minology for medications acting within this system because
patients may be reluctant to use “antidepressants”for GI
symptoms. Similarly, clinicians not well trained in their use
may prescribe them solely to treat comorbid psychiatric
disease or to reduce stress. In the light of modern research,
this terminology can limit their potential clinical value.
Consistent with the Rome Foundation’s new definitional
guidelines, we relabel agents working both in the brain
and gut as gutLbrain neuromodulators. This term
includes the primarily central neuromodulators (eg,
antidepressants and antipsychotic or other centrally
acting agents, such as buspirone) and the primarily
peripheral neuromodulators, including serotonergic,
chloride channel, a2d(delta) ligand agents, and others.
We believe this new terminology will improve under-
standing of their pharmacologic value, reduce stigma,
and likely improve treatment adherence.
Methodological Approach
The Rome Foundation creates multidisciplinary working
teams to evaluate areas where there is scientific uncertainty
or a lack of evidence to answer clinical questions or make
treatment recommendations. When the knowledge acquired
is unclear or controversial, discussions ensue to achieve
consensus (ie, Delphi approach).
3,4
For this working team,
committee members were selected representing gastroen-
terology, GI motility, psychiatry, pain management,
evidence-based data acquisition, and psychopharmacology.
An outline was created to cover basic pharmacology of the
central neuromodulators (Table 1,Figures 15), effects on
GI physiology (Table 2), available clinical studies relating to
chronic pain, non-GI painful disorders and FGIDs, and
treatment approaches. A systematic evidence-based review
was conducted to include the major classes of central
modulators used for treating specific FGID syndromes
(functional heartburn and functional chest pain, functional
dyspepsia, irritable bowel syndrome [IBS], and cyclic vom-
iting syndrome [CVS]) (Table 3). However, we were aware
Table 1.Action of Neuromodulators on Transporters and Receptors
Transporter or receptor Stimulate or inhibit Action Clinical Adverse effects Drug class
SERT (t) Inhibit serotonin reuptake Increase serotonin AD and anti-anxiety Nausea, diarrhea SSRI, SNRI, TCA
NET (t) Inhibit norepinephrine reuptake Increase norepinephrine AD and analgesic Dry mouth, sweats, constipation SNRI, TCA
DAT (t) Inhibit dopamine reuptake Increase dopamine Increase activation Nausea Buproprion, sertraline
D
2
Receptor antagonist Decrease dopamine Antipsychotic and antiemetic EPS galactorrhea All antipsychotics
5-HT
1
Receptor agonist Stimulate 5-HT
1
AD and improves gastric compliance Nausea, headache, nervousness Buspirone
5-HT
2A
Receptor antagonist Increase dopamine in
striatum and pituitary
Antipsychotic without EPS or galactorrhea Akathisia agitation Atypical antipsychotics
5-HT
3
Receptor antagonist Inhibit 5-HT
3
Less nausea, diarrhea, pain Constipation Mirtazapine, olanzapine
NOTE. Reprinted with permission from Sobin et al.
27
AD, antidepressant; DAT, dopamine transporter; SERT, serotonin transporter.
March 2018 GutBrain Neuromodulators: A Rome Working Team Report 1141
SPECIAL REPORT
that there would be a lack of studies adequately addressing
their use in some FGIDs, so our analysis also included
painful non-GI disorders (eg, fibromyalgia, low back pain,
and chronic headache), given the understanding that cen-
trally targeted agents would have similar mechanisms of
action on pain. We also looked at specific symptoms, such as
pain, nausea, and vomiting. The committee then proceeded
to evaluate smaller studies and case reports on painful
FGIDs. This information, combined with clinical experience
and expert opinion, was used to generate recommendations
for use of central neuromodulators (Table 4,Figure 5),
including strategies for relapse prevention and avoidance of
opioids and the use of communication skills to improve
patient acceptance and adherence. The committee inter-
acted through conference calls, finalizing the document
during a 15-month period between April 2016 and
July 2017.
The final document is a synthesis of available evidence
guided by the research and clinical experience of the experts
on the committee. It highlights a newer understanding of the
value of central neuromodulators for FGIDs that we believe
will ultimately help clinicians in the care of their patients
with these GI disorders.
Mechanisms of Action of Central
Neuromodulators and Impact on
Brain and Gut Physiology
Mechanisms of Action
The most accepted mechanism of action of the available
antidepressants remains based on the “monoamine hy-
pothesis,”in which depression is believed to result from a
deficiency in 1 or more of the 3 highly interacting mono-
amines: serotonin (5-hydroxytryptamine [5-HT]),
noradrenalin (NA), and, to a lesser extent, dopamine (DA), in
various brain circuits (whether or not accompanied by
compensatory up-regulation of their post-synaptic re-
ceptors). According to this hypothesis, antidepressants work
by rapidly boosting the synaptic actions of 1 or more of
these monoamines, followed by a slower adaptive down-
regulation and/or desensitization of post-synaptic
Figure 1. Hypothesized
mechanism of action of
antidepressants. The vast
majority of currently avail-
able antidepressants work
by blocking the presynap-
tic reuptake pump of 1 or
more of the 3 main mono-
amine neurotransmitters
(serotonin, noradrenalin,
and dopamine), causing
the respective neurotrans-
mitter to accumulate in the
synaptic cleft (A), which in
turn leads to a delayed
down-regulation or
desensitization of post-
synaptic receptors for the
respective neurotrans-
mitter (B). The latter effect
is believed to account for
the antidepressant actions
of the drugs.
1142 Drossman et al Gastroenterology Vol. 154, No. 4
SPECIAL REPORT
monoamine receptors (Figure 1). The same mechanisms of
action are believed to underlie the well-documented anxi-
olytic effects of antidepressants. Most classes of antide-
pressants boost monoamine activity by blocking the
presynaptic transporters for 1 or more monoamines, which
terminate their synaptic action by reuptake into the pre-
synaptic neuron.
5
Information on the synaptic actions of the
3 main monoamine neurotransmitter systems is provided in
Figure 2.
The overall boost of monoaminergic neurotransmission,
resulting in acute stimulation of pre- and post-synaptic re-
ceptors in the CNS, may also induce some of the centrally/
autonomically mediated side effects of antidepressants.
Boosting 5-HT neurotransmission, for example, may induce
agitation, anxiety, insomnia, and sexual dysfunction (due to
stimulation of 5-HT
2A
and 5-HT
2C
receptors at various sites
in the CNS), as well as nausea and vomiting (due to stimu-
lation of 5-HT3 receptors in the brainstem).
5
Tolerance to
most of these side effects usually develops relatively rapidly,
although sexual dysfunction is more likely to persist.
5
Stimulating NA receptors may induce cardiovascular side
effects, including alterations in heart rate and blood pres-
sure, as well as motor activation/agitation.
5
Nevertheless,
the fact that antidepressants are typically used in somewhat
lower dosages in the treatment of FGIDs compared with
treatment of anxiety or mood disorders (especially with
tricyclic antidepressants [TCAs]),
6
may limit the risk of side
effects.
The same pharmacologic properties that explain the
antidepressant action of these drugs may also account for
their analgesic effects via the braingut axis, which con-
stitutes the biologic basis of visceral pain perception.
7
First,
through their monoaminergic actions, antidepressants may
interfere with the function of pain-related brain circuits,
especially as emotional and cognitive circuits targeted by
antidepressants are highly intertwined with pain-
processing regions. This may also account for the pro-
found psychological modulation of pain experience,
8,9
particularly relevant for the large group of FGID patients
with comorbid mood or anxiety disorders. Second, antide-
pressants interfere with the complex mechanisms of pain
transmission at the level of the dorsal horn of the spinal
cord (ie, the first synapse in the afferent pain transmission
cascade
10
). There are important descending projections
from brainstem nuclei, including the peri-aqueductal gray,
raphe nuclei, locus ceruleus, and rostrolateral ventral me-
dulla, to the dorsal horn of the spinal cord, modulating
ongoing afferent pain transmission at the level of the first
synapse (Figure 3). These descending pathways are
controlled in a topdown fashion by brain regions,
including the amygdala and perigenual anterior cingulate
cortex. Importantly, these projections are primarily opioi-
dergic, noradrenergic, and serotonergic in nature and, as a
result, antidepressants can profoundly interfere with these
modulatory processes.
6,10,11
Over the last decade, there has been increasing evidence
for another mechanism of action for antidepressants rele-
vant to treating painful FGIDs. Neuroplasticity, the loss of
cortical neurons with chronic pain, traumatic life events,
and psychiatric disease, and neurogenesis (or regrowth) of
neurons with clinical treatment,
12
are improving our un-
derstanding of how antidepressants can help reduce GI
symptoms. The age-old concept that neural cells are
Figure 2. Overview of presynaptic transporters, and pre- and
postsynaptic receptors for the 3 main monoamine neuro-
transmitter systems. This figure summarizes transporters and
receptors for (A). Dopamine (post-synaptic receptors D
15
,
presynaptic D
2
, dopamine transporter [DAT]); (B). Seroto-
nin (5-HT) (postsynaptic receptors 5-HT
1A-7
, presynaptic
5-HT
1B/D
, serotonin transporter SERT); (C) norepinephrine
dopamine (postsynaptic receptors Alpha
1
-Beta
3
, presynaptic
Alpha
2
, norepinephrine transporter [NET]). X, Y, Z: unknown
receptors.
March 2018 GutBrain Neuromodulators: A Rome Working Team Report 1143
SPECIAL REPORT
established at birth or soon after and die, mostly in relation
to major events like an ischemic stroke or brain hypoxemia,
is now being revisited by evidence that CNS neurons are
plastic and capable of new growth. Brain cells in regions
such as the hippocampus can die after severe psychological
trauma, and this is associated with developing post-
traumatic stress disorder
13
or chronic pain.
14
Reduced
cortical density after trauma is seen in other brain regions
involved in emotional and pain regulation,
14,15
and relevant
here to pain control regions such as the cingulate cortex, in
chronic and painful GI conditions including IBS
16
and even
chronic pancreatitis.
17
Adding to this is the evidence that antidepressant, and
possibly psychological, treatments, appear to increase pre-
cursor neuronal growth in these regions. Brain-derived
neurotrophic factor levels increase with antidepressant
treatment, and this correlates with longer periods of treat-
ment and with the degree of recovery from depression.
12,18
Furthermore, the longer patients are treated with antide-
pressants, the lower the frequency of relapse or recurrence
of the depression.
19,20
This may help explain why these
treatments have more than immediate effects of symptom
reduction; over time they may help “rewire”the brain to
approach a premorbid state of functioning.
Figure 3. Simplified over-
view of ascending (A) and
descending (B) neural
pathways involved in
visceral perception and
pain regulation. The
descending modulatory fi-
bers from brain stem cen-
ters can alter the
sensitivity of the dorsal
horn neuron signaling and
can serve as a central
control of pain perception
during visceral stimulation.
Reprinted with permission
from Rome Foundation.
1144 Drossman et al Gastroenterology Vol. 154, No. 4
SPECIAL REPORT
Besides these central effects, antidepressants can exert
profound effects on peripheral GI physiology by boosting
serotonergic and noradrenergic neurotransmission (and
peripheral neurogenesis), which could also account for
some of their beneficial effects in FGIDs, as well as some of
their GI side effects (see section “Antidepressants”).
These mechanisms of action provide the theoretical
rationale for the use of central neuromodulators in FGIDs
and other painful conditions. Given their analgesic and pe-
ripheral GI effects, which can occur independently of their
antidepressant and anxiolytic effects, there is a basis for
their use in patients with FGIDs or other painful somatic
symptoms, regardless of anxiety or mood disorder
comorbidity.
The different classes of central neuromodulators rele-
vant to FGID treatment discussed in the next section share a
common effect for treating depression, and some can reduce
pain, and have propensities for side effects, depending on
the pharmacologic properties of their specific receptors, as
illustrated in Figure 4 and summarized in Table 1.
Antidepressants
Tricyclic antidepressants. The hallmark feature of
TCAs, believed to be primarily responsible for their anti-
depressant (and analgesic) properties, is a variable combi-
nation of 5-HT and NA reuptake inhibition properties.
5,6
Due to this dual action, TCAs theoretically have a stronger
potential for analgesic effects compared with other antide-
pressant classes targeting only 1 monoamine system, such
as selective serotonin reuptake inhibitors (SSRIs) (see
Figure 4). However, this can also make these drugs more
prone to the potential side effects induced by boosting 5-HT
and NA neurotransmission, as outlined. Most of the TCAs
have additional receptor affinities,
5
some of which (5-HT
2A
and 5-HT
2C
receptor antagonism, for example) can
contribute to their antidepressant and/or analgesic prop-
erties, while others may be primarily responsible for their
side effect profile. Examples of the latter include muscarinic-1
receptor antagonism (may cause classic anticholinergic side
effects, including dry mouth, constipation, drowsiness, and
blurred vision), a1 adrenergic receptor antagonism (may
lead to dizziness, drowsiness, and orthostatic hypotension),
and histamine 1 (H1) receptor antagonism (may lead to
weight gain, especially in combination with 5-HT
2C
antago-
nism, as well as drowsiness). Finally, most TCAs have weak
sodium channel blocking properties, which leads to a risk of
arrhythmias, and coma or seizures upon overdosing.
5
Thus,
TCAs should be avoided in patients with bundle branch
block or prolonged QT intervals. In the context of FGID
treatment, some of the side effects may actually be benefi-
cial, such as slowing of GI transit due to anticholinergic
properties in patients with IBS with diarrhea, and increased
appetite and weight gain in patients with functional
dyspepsia (FD) with early satiation and weight loss.
The general pharmacologic properties of TCAs are
shown in Figure 4A, but vary slightly between different
drugs within this class.
Selective serotonin reuptake inhibitors. SSRIs are
characterized by selective blockade of the presynaptic 5-HT
transporter (Figure 4B), thereby boosting 5-HT neurotrans-
mission. Their primary serotonergic effect, without norad-
renergic effect, leads to greater expected benefit in treating
anxiety, obsessivecompulsive disorder, and phobic-related
behaviors, rather than for chronic painful symptoms or dis-
orders. This class includes the drugs fluoxetine, fluvoxamine,
paroxetine, sertraline, and (es)citalopram. Besides the com-
mon feature of 5-HT reuptake inhibition, each of them can
have various secondary pharmacologic properties, including
5-HT
2C
antagonism for fluoxetine and mild anticholinergic
action for paroxetine.
5
It remains unclear, however, how
clinically relevant these secondary properties are, particu-
larly at the rather low dosages commonly used to treat FGIDs.
By boosting 5-HT neurotransmission in general, SSRIs can
induce centrally mediated side effects, as explained in the
section on “Mechanism of Action”as well as impact on gut
physiology, as explained in the following section.
Serotonin noradrenalin reuptake inhibitors. Like
the TCAs, serotonin noradrenalin reuptake inhibitors
(SNRIs) primarily block both 5-HT and NA reuptake
(Figure 4C), thereby boosting 5-HT and NA neurotrans-
mission.
5
Again similar to the TCAs, the degree of seroto-
nergic, relative to noradrenergic reuptake inhibition activity
differs somewhat between individual drugs in this class.
5
Venlafaxine only exerts significant NA reuptake inhibition
effects at doses of 225 mg or more. Duloxetine has a strong,
Figure 4. Pharmacologic
properties of the 4 most
important classes of anti-
depressants: serotonin
reuptake inhibition (SRI);
noradrenalin reuptake
inhibition (NRI); 5-
hydroxytryptamine (sero-
tonin) receptor (5-HT);
muscarinic acetylcholine
receptor (M); histamine re-
ceptor (H); a-noradrenalin
receptor (a).
March 2018 GutBrain Neuromodulators: A Rome Working Team Report 1145
SPECIAL REPORT
Table 2.Summary of GutBrain Neuromodulators by Class, Mode of Action, Actions on Gastrointestinal Sensorimotor Function, Relevance to Gastrointestinal Symptom,
and Side Effects
Drug class, drug Mode of action
Actions on GI sensorimotor
function Relevance to symptom control Side effects
TCA
Amitriptyline,
imipramine,
desipramine,
nortriptyline
Presynaptic SRI and NRI.
Antagonism/inhibition of multiple
post-synaptic (5-HT
2
, 5-HT
3
,H1,
muscarinic-1, a1) and presynaptic
(a2) receptors.
Motility: slow GI transit, largely
related to their anticholinergic and
noradrenergic properties
Sensitivity: limited and inconsistent
evidence that TCAs decrease
visceral sensitivity
Pain reduction. Best documented for
IBS, but also FD (EPS). Potential
usefulness in all FGIDs where pain
is a prominent feature. Side effect
profile can be useful in order to
reduce diarrhea and improve
sleep.
Drowsiness, dry mouth,
constipation, sexual
dysfunction,
arrhythmias, and
weight gain
SSRI
Citalopram,
escitalopram,
fluoxetine,
paroxetine,
sertraline
Presynaptic SRI. Motility: enhancement of gastric and
small bowel propulsive motility
Sensitivity: no major impact on
visceral sensitivity in healthy
subjects or patients with FGIDs
Treatment of associated anxiety,
phobic features, and OCD in
FGIDs.
Agitation, diarrhea,
insomnia, night
sweats, headache,
weight loss, and
sexual dysfunction.
SNRI
Duloxetine,
milnacipran,
venlafaxine
Pre-synaptic SRI and NRI. Equally
strong for duloxetine. NRI for
venlafaxine in higher doses.
Milnacipran stronger NRI than SRI
effects.
Motility: inhibitory effect on gastric
and colonic tone, but not to the
degree of TCAs; more studies are
needed
Sensitivity: few studies available; area
requiring further research
Treatment of associated pain (based
on efficacy in fibromyalgia, back
pain, and headache) in FGIDs.
Potential use for painful FGIDs;
however, formal evidence in
treatment of specific FGID-related
pain is lacking.
Nausea, agitation,
dizziness, sleep
disturbance, fatigue,
and liver
dysfunction
NA and specific
serotonergic
antidepressants
Mirtazapine,
mianserin,
trazodone
Indirect effects resulting in increased
NA and serotonergic activity
through a2 antagonism on NA
and 5-HT neurons. Also 5-HT
2
,5-
HT
3
, H1, muscarinic-1
antagonism
Motility: lack of detailed studies
Sensitivity: lack of detailed studies
Potential use for treatment of early
satiation, weight loss, and chronic
nausea/vomiting. Side effect
profile can be useful to improve
sleep.
Sedation, headache,
dry mouth, and
weight gain
Azapirones
Buspirone,
tandospirone
Partial pre- and post-synaptic 5-HT
1
agonists
Motility: enhanced esophageal
contractions and increased
gastric accommodation in health
and FD
Sensitivity: limited data suggest
no effect
Treatment of associated anxiety.
Potential use for treatment of early
satiety, fullness, and nausea, but
consistent evidence in FGIDs is
lacking.
Sedation, headache,
and vertigo
1146 Drossman et al Gastroenterology Vol. 154, No. 4
SPECIAL REPORT
and roughly equal, affinity for the 5-HT and NA transporter,
thereby acting as a true SNRI even at lower doses. Finally,
milnacipran has stronger NA reuptake inhibition compared
with 5-HT reuptake inhibition properties. These drugs are
largely devoid of additional receptor affinities,
5
as can be
seen in Figure 4C. Therefore, they have a more favorable
side effect profile compared with the TCAs, while still
maintaining their potential analgesic benefits. However, the
side effects related to boosting 5-HT and NA neurotrans-
mission per se (see in the section on “Mechanism of action”)
can obviously occur, with hypertension under venlafaxine
and nausea under duloxetine being common. This makes
them good candidates for treating conditions characterized
by chronic painful physical symptoms (not limited to the
context of depression), including neuropathic pain and
functional somatic syndromes, such as fibromyalgia,
2123
which are often comorbid with FGIDs.
Noradrenergic and specific serotonergic (tetra-
cyclic) antidepressants. Mirtazapine (and the older
related agent mianserin) is the prototypical antidepressant
in this class. It boosts both 5-HT and NA neurotransmis-
sion, not by blocking their reuptake pumps, but by
blocking presynaptic a2 noradrenergic auto- and hetero-
receptors on NA and 5-HT neurons, respectively, which act
as brakes on both NA and 5-HT release from these
respective neurons.
5
Inaddition,likesomeoftheTCAs,it
has 5-HT
2A
and 5-HT
2C
receptor antagonist properties,
which may account for some additional antidepressant
properties, as well as a more favorable side effect profile
by blocking some of the unwanted receptor actions of
boosting 5-HT transmission (see section on “Mechanism of
action”). The same applies to its 5-HT
3
antagonist prop-
erties, which may explain its more favorable GI side effect
profile, which would include reduction in nausea, pain, and
diarrhea. However, through its H1 and 5-HT
2C
antagonist
properties, mirtazapine may cause increased appetite and
weight gain (a possible advantage in some FGID pop-
ulations), as well as sedation.
5
Its receptor affinity profile
is shown in Figure 4D.
5-Hydroxytryptamine 1A Receptor (Partial)
Agonists (Azapirones)
These agents, including buspirone and tandospirone,
have been developed as non-benzodiazepine anxiolytics,
which can dampen activity in fear circuitry in the brain,
centered around the amygdala, through their partial agonist
action at pre- and post-synaptic 5-HT
1A
receptors.
5
Through
the same receptor affinity at the peripheral level, this class
of drugs can also directly affect GI physiology (see section
on “Actions on Gastointestinal Motility and Sensitivity”).
Which of these effects (central and/or peripheral) primarily
accounts for the putative effect of these agents on functional
GI symptoms remains unclear.
Atypical Antipsychotics
Dopamine 2 (D
2
) receptor antagonist activity is the
hallmark of antipsychotics as a class of drugs, which is
responsible for both the desired antipsychotic effect and
Table 2.Continued
Drug class, drug Mode of action
Actions on GI sensorimotor
function Relevance to symptom control Side effects
Atypical antipsychotics
Aripiprazole,
levosulpiride,
olanzapine,
quetiapine,
sulpiride
D
2
receptor antagonism as main
mechanism. Partial D
2
agonism
for the sulpirides. Various profiles
of 5-HT
2A
antagonism
(olanzapine, quetiapine), 5-HT
1A
agonism (quetiapine), H1, a1, a2,
muscarinic-1 receptor
antagonism.
Motility: lack of data
Sensitivity: limited data suggest
decreased gastric sensitivity in
functional dyspepsia
Potential use in augmentation for
pain reduction; however, formal
evidence in treatment of specific
FGID pain currently lacking. Low
evidence in FGIDs. Potential use
of sulpirides for nausea and
dyspepsia, but formal evidence is
lacking. Improved sleep.
Sedation, dizziness,
weight gain,
hyperlipidemia, and
diabetes
Delta ligand agents
Gabapentin,
pregabalin
a2dsubunit blockage of (mostly
presynaptic) voltage-sensitive
calcium channels
Motility: no data
Sensitivity: decreased sensitivity to
rectal distention in IBS
Treatment of associated general
anxiety disorder or fibromyalgia/
abdominal wall pain. Potential use
for treatment of neuropathic pain
in FGIDs. However, formal
evidence in FGIDs is lacking.
Sedation, headache,
vertigo, weight gain,
and peripheral
edema.
NRI, noradrenaline reuptake inhibitor; SRI, serotonin reuptake inhibitor.
March 2018 GutBrain Neuromodulators: A Rome Working Team Report 1147
SPECIAL REPORT
Table 3.Summary of Evidence for Different Classes of Central Neuromodulators as Treatments for Painful Non-Gastrointestinal Disorders and Functional Gastrointestinal
Disorders
Drug Condition
Highest level
of evidence
No. of
participants End point studied Effect
GRADE
quality of
evidence Adverse events Comments
TCA FD Meta-analysis
of 3 RCTs
339 Persistence of global
symptoms
Pain scores
RR favored TCAs
(RR ¼0.74; 95%
CI, 0.61 to 0.91),
NNT ¼6
No effect
Moderate More common with
TCAs, NNH ¼7
Only amitriptyline
and imipramine
studied
IBS Meta-analysis
of 11 RCTs
744 Persistence of global
symptoms
Persistence of
abdominal pain
RR favored TCAs
(0.69), NNT ¼4
RR favored TCAs
(RR ¼0.69; 95%
CI, 0.58 to 0.82),
NNT ¼3
Moderate More common with
TCAs, NNH ¼8
Noncardiac
chest pain
Meta-analysis
of 2 RCTs
of imipramine
58 Percentage reduction
in episodes of
chest pain
Mean number of
episodes of chest pain
Reduced with TCAs
Reduced with TCAs
Very low More common
with TCAs
No evidence for
other TCAs
Functional heartburn 1 RCT of imipramine 83 50% improvement in
GERD scores
Chest pain scores
No effect
No effect
Low Constipation
commoner with
imipramine
Only 1 RCT, no
evidence for
other TCAs
Functional
anorectal pain
1 case series 26 Pain status at discharge 38.5% improved Very low Not reported No RCTs, specific
TCAs used not
reported
Functional nausea
and vomiting
1 case series 37 Complete remission
of symptoms
Response to therapy
51.4% achieved
remission
83.8% achieved
response
Very low Agitation, drowsiness,
and anticholinergic
side effects
No RCTs
CVS Pooled analysis of 4
case series
237 Response to therapy 75.5% achieved
response
Very low Not reported No RCTs
Fibromyalgia
a
Meta-analysis
of 4 RCTs
of amitriptyline
275 Pain relief 50% RR favored amitriptyline
(RR, 2.9; 95%
CI, 1.7 to 4.9),
NNT ¼4
Very low More common
with TCAs, NNH ¼3
No evidence for
other TCAs
Low back pain
a
Meta-analysis of 3 148 Pain scores No effect Very low Not reported
Chronic headache
a
Meta-analysis
of 8 RCTs
608 Number of days per
month with headache
50% improvement
in headaches
SMD in number of
days favored TCAs
RR favored TCAs
(RR, 1.41; 95%
CI, 1.02 to 1.89)
Very low More common
with TCAS
1148 Drossman et al Gastroenterology Vol. 154, No. 4
SPECIAL REPORT
Table 3.Continued
Drug Condition
Highest level
of evidence
No. of
participants End point studied Effect
GRADE
quality of
evidence Adverse events Comments
SSRI FD Meta-analysis
of 2 RCTs
388 Persistence of global
symptoms
Pain scores
No effect
No effect
Moderate No more common
with SSRIs
Only sertraline and
escitalopram
studied
IBS Meta-analysis
of 7 RCTs
356 Persistence of global
symptoms
Persistence of
abdominal pain
RR favored SSRIs
(RR ¼0.74; 95%
CI, 0.58 to 0.95),
NNT ¼6
No effect
Moderate No more common
with SSRIs
Noncardiac
chest pain
Meta-analysis
of 4 RCTs
184 Chest pain scores No effect Moderate No more common
with SSRIs
Only sertraline and
paroxetine
studied
Functional heartburn 1 RCT of citalopram 75 Relief of predominant
symptom at baseline
Favored citalopram,
NNT ¼4
Low No more common
with citalopram
Only 1 RCT, no
evidence for
other SSRIs
Fibromyalgia
a
Meta-analysis
of 7 RCTs
383 Global symptoms
Reduction in pain
by 30%
RD favored SSRIs
(RD, 0.14; 95%
CI, 0.06 to 0.23),
NNT ¼7
RD favored SSRIs
(RD, 0.10; 95%
CI, 0.01 to 0.20),
NNT ¼10
Very low No more common
with SSRIs
Low back pain
a
Meta-analysis
of 3 RCTs
199 Pain scores No effect Very low Not reported Only paroxetine
and fluoxetine
studied
Chronic headache
a
Meta-analysis
of 2 RCTs
108 Analgesic requirement
Number of days per
month with headache
Mean difference favored
SSRIs (mean
difference, 1.87;
95% CI, 2.09
to 1.65)
No effect
Very low Not reported Only citalopram
and sertraline
studied
SNRIs FD 1 RCT of venlafaxine 160 Absence of global
symptoms
Pain scores
No effect
No effect
Low Not reported Only 1 RCT, no
evidence for
other SNRIs
IBS Case series of
duloxetine
15 Global symptom
improvement
Pain scores
40% achieved response
Pain scores reduced
Very low Nausea, insomnia,
constipation, and
drowsiness
No RCTs, no
evidence for
other SNRIs
Noncardiac
chest pain
1 crossover RCT
of venlafaxine
50 >50% improvement
in chest pain scores
OR favored venlafaxine
(26.0)
Very low More common
with venlafaxine
Only 1 RCT, no
evidence for
other SNRIs
March 2018 GutBrain Neuromodulators: A Rome Working Team Report 1149
SPECIAL REPORT
Table 3.Continued
Drug Condition
Highest level
of evidence
No. of
participants End point studied Effect
GRADE
quality of
evidence Adverse events Comments
Biliary pain and
sphincter of Oddi
dysfunction
1 case series of
duloxetine
18 Global symptom
improvement
Pain scores
50% achieved response
Pain scores reduced
Very low Fatigue, nausea,
headaches,
constipation,
and insomnia
No RCTs, no
evidence for
other SNRIs
Fibromyalgia
a
Meta-analysis
of 6 RCTs of
duloxetine
Meta-analysis
of 3 RCTs of
milnacipran
2249
1925
Pain relief 50%
Pain relief 30%
RR favored duloxetine
(1.57), NNT ¼8
RR favored milnacipran
(1.4), NNT ¼9
Low
High
More common
with duloxetine
More common with
milnacipran,
NNH ¼11 to 15
Only duloxetine a
nd milnacipran
studied
Low back pain
a
3 RCTs of duloxetine 1263 30% reduction in
pain scores
Pain scores
Favored duloxetine in
1 RCT
Reduced with duloxetine
in 2 RCTs
Low More common
with duloxetine
No evidence for
other SNRIs
Chronic headache
a
1 RCT of venlafaxine 60 Number of days
per month with
headache
Favored venlafaxine Very low More common
with venlafaxine
Only 1 RCT, no
evidence for
other SNRIs
TCA FD 1 RCT of mirtazapine 34 Global symptom scores
Epigastric pain scores
Favored mirtazapine
No effect
Very low Not reported Only 1 RCT, no
evidence for
other tetracyclic
antidepressants
IBS 1 case series
of trazodone
and amoxapine
47 Global symptom
improvement
Global symptom
remission
41.7%87.5% achieved
response
25.0%33.3% achieved
remission
Very low Not reported No RCTs, no
evidence for
other tetracyclic
antidepressants
Noncardiac
chest pain
1 RCT of trazodone 29 Global symptom
improvement
Chest pain scores
Favored trazadone
No effect
Very low Dizziness,
drowsiness,
and fatigue
Only 1 RCT, no
evidence for
other TCAs
Functional nausea
and vomiting
1 case report 1 Response to therapy Patient had complete
response
Very low Not reported No RCTs or case
series
Fibromyalgia
a
1 RCT of mirtazapine 40 30% reduction in
pain scores
Pain scores
No effect
Reduced with mirtazapine
Very low Increased appetite,
weight gain, and
constipation
Only 1 RCT, no
evidence for
other tetracyclic
antidepressants
Low back pain
a
2 RCTs of maprotiline
and trazodone
111 Pain scores Favored maprotiline
No effect with trazodone
Very low Constipation, dizziness,
drowsiness, dry
mouth, and confusion
Only 2 RCTs, no
evidence for
other tetracyclic
antidepressants
Chronic headache
a
1 RCT of mianserin
and 1 crossover
RCT of mirtazapine
102 Pain scores
Headache frequency,
duration, and intensity
Favored mianserin
Reduced with mirtazapine
Very low Drowsiness, dizziness,
and weight gain
Only 2 RCTs, no
evidence for
other TCAs
1150 Drossman et al Gastroenterology Vol. 154, No. 4
SPECIAL REPORT
Table 3.Continued
Drug Condition
Highest level
of evidence
No. of
participants End point studied Effect
GRADE
quality of
evidence Adverse events Comments
Azapirones FD Meta-analysis
of 3 RCTs
220 Persistence of global
symptoms
Pain scores
No effect
Reduced with
tandospirone
Moderate No more common
with azapirones
Tandospirone
beneficial for
global symptoms
and pain in the
largest RCT
IBS 1 RCT of tandospirone 200 Global symptom
improvement
Abdominal pain
improvement
Favored tandospirone
Favored tandospirone
Low No more common
with tandospirone
Only 1 RCT, no
evidence for
other azapirones,
all patients also
received
pinaverium
Chronic headache
a
One case series
of buspirone
26 >50% reduction in
number of days per
month with headache
54.5% achieved
response
Very low Dizziness, drowsiness,
and nausea
No RCTs, no
evidence for
other azapirones
Fibromyalgia
a
Meta-analysis
of 2 RCTs
of quetiapine
155 Reduction in pain
by 50%
Reduction in pain
by 30%
No effect
RD favored quetiapine
(RD, 0.12; 95%
CI, 0.00 to 0.23),
NNT ¼8
Very low Weight gain commoner
with quetiapine,
NNH ¼12
Only 2 RCTs, no
evidence for
other atypical
antipsychotic
drugs
Atypical
antipsychotics
FD Meta-analysis
of 3 RCTs
172 Persistence of global
symptoms
Pain scores
RR favored atypical
antipsychotics
(RR, 0.50; 95%
CI, 0.37 to 0.67),
NNT ¼3
No effect
Very low No more common
with atypical
antipsychotics
Only sulpiride and
levosulpiride
studied
IBS 1 case report 1 Remission of symptoms Patient had complete
remission
Very low Not reported No RCTs or
case series
Chronic headache
a
1 case series
of olanzapine
50 Number of days per
month with headache
Headache severity
Improvement in
headaches
Reduced with olanzapine
Reduced with olanzapine
74.0% achieved response
Very low Weight gain and
drowsiness
No RCTs, no
evidence for
other atypical
antipsychotic
drugs
Delta ligand
agents
FD 1 case series
of gabapentin
and pregabalin
41 Global symptom scores
Pain scores
Reduced
Reduced
Very low Not reported No RCTs
IBS 1 RCT of pregabalin 85 Adequate relief of global
symptoms
Pain scores
No effect
Reduced with pregabalin
Very low Not reported Only 1 RCT, no
evidence for
other delta
ligand agents
March 2018 GutBrain Neuromodulators: A Rome Working Team Report 1151
SPECIAL REPORT
Table 3.Continued
Drug Condition
Highest level
of evidence
No. of
participants End point studied Effect
GRADE
quality of
evidence Adverse events Comments
Fibromyalgia
a
1 RCT of gabapentin
Meta-analysis
of 3 RCTs
of pregabalin
150
2027
Reduction in pain
by 30%
Pain scores
Reduction in pain
by >30%
Favored gabapentin
Reduced with gabapentin
RR favored pregabalin
(depending on
the dosage
range, RR, 1.53; 95%
CI, 1.18 to 1.98 for
300 mg/d to RR, 1.92;
95% CI, 1.49
to 2.48 for 450 mg/d),
NNT ¼7to11
Very low
Low
Dizziness, drowsiness,
and light-headedness
more common with
gabapentin
Dizziness, drowsiness,
dry mouth, weight
gain, and peripheral
edema commoner
with pregabalin
Only 1 RCT of
gabapentin
Low back pain
a
1 RCT of gabapentin
1 RCT of pregabalin
108
44
Pain scores
Pain scores
No effect
Reduced with pregabalin
Very low
Very low
Fatigue, dry
mouth, difficulty
concentrating,
and dizziness
commoner with
gabapentin
No more common
with pregabalin
All patients
also received
buprenorphine
Chronic headache
a
1 crossover RCT of
gabapentin
133 Mean difference in
headache-free rates
>50% reduction in
headache frequency
Favored gabapentin
Favored gabapentin,
NNT ¼4
Very low Dizziness, drowsiness,
and ataxia
Only 1 RCT, no
evidence for
other delta
ligand agents
NOTE. The table summarizes the level of evidence, number of subjects and end points studied, the effect and effect size where available, quality of evidence grading, and
adverse events occurrence.
GRADE, grading of recommendations assessment, development and evaluation; GERD, gastroesophageal reflux disease; NNT, number need to treat; RCT, randomized
controlled trial; RD, risk difference; RR, relative risk; SMD, standardized mean difference.
a
Because the more robust studies have been done in painful non-GI conditions, they are included in this table. More detailed information is included in the Supplementary
Material.
1152 Drossman et al Gastroenterology Vol. 154, No. 4
SPECIAL REPORT
Table 4.Summary of Recommended Use of GutBrain Neuromodulators by Condition, Based on Existing Literature and Clinical Expertise in Treating Chronic Painful and
Non-Gastrointestinal Conditions
Condition Recommendation
Chronic GI pain,
general
considerations
1. Low to modest dose regimens of TCAs have the most convincing evidence of benefit for treating chronic GI pain
2. SNRIs may have at least equal benefit as TCAs based on data in other chronic, painful disorders like fibromyalgia, migraine headaches, widespread body pain, and
peripheral neuropathy, though they have not been adequately tested for chronic GI pain
3. In the patient with anxiety or multiple somatic symptoms (somatization) or where there are incomplete benefits from TCAs or SNRIs, adding an augmenting agent is
the recommended next option for treating chronic GI pain
4. Delta ligand agents treat neuropathic pain, or pain associated with fibromyalgia, a condition commonly associated with FGIDs, particularly IBS and may be of help
with chronic GI pain, though studies have not adequately addressed their benefits in this condition.
IBS 1. Generally, when the pain is mild to moderate and intermittent, peripherally acting agents may be sufficient, but when pain is more severe or persistent, central agents
may be added or substituted
2. The TCA class of drugs is the first-line central neuromodulater for treating IBS, especially for IBS-D. In particular, the tertiary amine TCAs (amitriptyline and
imipramine) can reduce diarrhea, and also improve poor sleep quality. A secondary amine TCA (desipramine and nortriptyline) may be selected if less anticholinergic
or antihistaminic effect is desired (eg, for treating pain with IBS-M or IBS-C)
3. The SNRI class of medications (duloxetine, venlafaxine, or milnacipran) has potential to improve the pain component of IBS based on data from treating other pain
disorders, and has fewer side effects than TCAs, but these agents have not been adequately studied in IBS
4. SSRIs can be considered in IBS if anxiety states are present, and the abdominal pain and diarrhea are not the dominant clinical features
Functional
heartburn and
functional
chest pain
1. In patients with troublesome heartburn or chest pain where gastroesophageal reflux disease has been excluded, a treatment trial with a centrally acting pharma-
cologic agent may be considered.
2. There is insufficient evidence to recommend a particular class of central agent, though SSRIs have shown some benefit for esophageal pain along with TCAs and
SNRIs.
FD 1. When treating functional dyspepsia, it is best to consider the Rome IV symptom-specific subgroups: PDS and EPS
2. Buspirone, an anxiolytic azapirone may be used for PDS where early satiety, fullness, and nausea predominate
3. Mirtazapine is a good treatment option for PDS when there is chronic nausea and vomiting, or weight loss, and it may also help coexisting abdominal pain
4. When the dyspeptic symptoms are consistent with EPS, studies mainly support the use of TCAs, either initially or after an unsuccessful response to a proton pump
inhibitor
5. As with other painful FGIDs, the SNRI group of medications can also be considered for patients with EPS who do not tolerate TCA treatment though confirmatory
studies are lacking
CVS 1. First, for treating acute episodes, the effort is to reduce the symptom severity and duration of the attacks with anti-emetic agents like ondansetron or promethazine,
minimizing the use of opioids for the abdominal pain, and intravenous hydration. Benzodiazepines can be used to treat acute anxiety and distress and have
independent effects on nausea reduction.
2. Second, to prevent future episodes, cannabinoids must be eliminated and prophylactic treatment using central neuromodulators (TCAs, SNRIs, tetracyclics, atypical
antipsychotics and perhaps anticonvulsants) should be instituted to reduce the severity and frequency of acute attacks.
IBS-M, irritable bowel syndrome with diarrhea and constipation.
March 2018 GutBrain Neuromodulators: A Rome Working Team Report 1153
SPECIAL REPORT
the undesired side effects, such as extrapyramidal symp-
toms/tardive dyskinesia, prolactin elevation, and affective
(apathy, anhedonia) and cognitive symptoms.
5
Unlike the
older conventional or typical antipsychotics, in addition to
D
2
antagonist properties, the newer class of atypical an-
tipsychotics is characterized by 5-HT
2A
receptor antago-
nist properties (olanzapine, quetiapine), rapid dissociation
from the D
2
receptor (most atypical antipsychotics), D
2
partial agonism (amisulpiride, (levo)sulpiride, and/or 5-
HT
1A
partial agonism (quetiapine).
5
As 5-HT
1A
and 5-
HT
2A
receptors act as brakes and accelerators, respec-
tively, on dopaminergic neurons, all of these additional
mechanisms of action reduce the impact of DA blockade in
some of the pathways mentioned in the section on
“Mechanism of action”and, hence, the side effects
mentioned there. This includes a reduction in risk of
extrapyramidal side effects, such as dystonic reactions or
Parkinsonism, when compared with older typical anti-
psychotics like haloperidol. However, even the atypical
antipsychotics are not devoid of side effects. Moreover, as
with some of the typical antipsychotics, due to additional
H1, 5-HT
2C
,a1- and/or a2-noradrenalin receptor antago-
nist, anticholinergic properties (particularly olanzapine
and, to a lesser extent, quetiapine), as well as some un-
known properties, additional side effects, such as
increased appetite; weight gain; cardiometabolic illness
(dyslipidemia/diabetes); and sedation may occur. There-
fore, like the typical antipsychotics, these agents should be
used with sufficient care and monitoring of side effects,
especially when used chronically.
5
However, in the context
of FGID treatment, far lower doses than the antipsychotic
doses are often used, which may limit the risk of side
effects.
Whether and how these central pharmacologic proper-
ties underlie the potential effect of these agents on FGID
symptoms in general, and nausea and abdominal pain in
particular (either as monotherapy or augmentation therapy,
see section on “Actions on Gastointestinal Motility and
Sensitivity”), remains, at present, unclear. However, inter-
ference with the complex neurotransmission of ascending
and descending pain pathways at the level of the dorsal
horn through various receptor affinities is a theoretical
possibility. In addition, atypical antipsychotics may have a
profound effect on gut physiology through the receptor af-
finities described in the section on “Actions on Gastointes-
tinal Motility and Sensitivity,”which may also account for
their putative effects.
Olanzapine and quetiapine have combined D
2
/5-HT
2A
antagonist properties characteristic of most atypical anti-
psychotics, with in addition H1, 5-HT
2C
, and a1-antagonist,
as well as anticholinergic properties. Quetiapine (or its
active metabolite norquetiapine) has additional 5-HT
1A
partial agonist properties, as well as noradrenalin reuptake
inhibitory effects, which may provide a rationale for its use
in FGIDs through mechanisms explained in the section on
“Actions on Gastointestinal Motility and Sensitivity”. Ami-
sulpride and the older related compound, sulpiride, on the
contrary, can be considered atypical antipsychotics through
their partial D
2
agonist rather than DA antagonist
properties, particularly at the lower doses typically used in
the treatment of FGIDs.
5
These agents do not have signifi-
cant affinities for other receptor systems, making them less
prone to side effects.
Delta Ligand Agents
These agents, with gabapentin and pregabalin being the
prototypical examples, exert their effect by blocking the
a2dsubunit of (mostly presynaptic) voltage-sensitive cal-
cium channels, which can, in turn, result in reduction of
the excessive release of excitatory neurotransmitters such
as glutamate. This mechanism of action accounts for their
anticonvulsant properties but can also underlie their
anxiolytic properties.
5
Probably most importantly in this
context, however, as signal transduction in nociceptive
pathways critically depends on voltage-sensitive calcium
channels, these agents may dampen activity in overly
active pain circuitry (from the dorsal horn to the brain).
5
This mechanism likely underlies their well-established
efficacy in neuropathic pain as well as, although to a
lesser extent, functional somatic syndromes putatively
characterized by central sensitization, such as fibromyal-
gia,
21,24
which then can lead to benefitinIBSandother
FGIDs.
25
These agents are classified as peripheral neuro-
modulators. However, pregabalin has also been shown to
alter brain connectivity, leading to a central analgesic
response in experimental pain.
26
By virtue of these prop-
erties, these agents are promising treatment options for
pain-predominant FGIDs.
Table 1 illustrates the action of the transporters and
receptors as discussed, their clinical and adverse effects, and
the drug class most closely linked to the transporter or
receptor.
27
Actions on Gastointestinal Motility and
Sensitivity
This section provides an overview of the physiological
actions of central neuromodulators specifically on GI
motility and sensitivity related to their use in the treatment
of FGIDs (summarized in Table 2).
Tricyclic Antidepressants (Amitriptyline,
Imipramine, Desipramine, Nortriptyline, and
Doxepin)
Motility. As noted, TCAs have 5-HT and NA reuptake
inhibition properties, but also have variable antimuscarinic
effects that affect motility. Several studies have addressed
the actions of TCAs on GI motility, but the extent to which
observed effects are attributable to serotonin transporter
inhibition or to other aspects of their pharmacology is
unclear.
28
Amitriptyline 25 mg for 2 weeks slowed solid gastric
emptying in healthy volunteers,
29
but a dosage of 12.5 mg 3
times daily did not alter liquid gastric emptying rate
30
and
nortriptyline up to 50 mg daily for 14 days did not alter
solid gastric emptying.
31
Treatment with desipramine 50
mg for 4 days slowed both orocecal and whole-gut transit
1154 Drossman et al Gastroenterology Vol. 154, No. 4
SPECIAL REPORT
times in healthy volunteers.
32
Similarly, imipramine and
amitriptyline 12.5 mg 3 times daily were shown to slow
orocecal transit time in healthy subjects, as measured with
the lactulose breath test.
30,33
Acute administration of
amitriptyline 80 mg to healthy volunteers did not alter
rectal compliance but reduced pressures in the anal canal.
34
Taken together, TCAs slow GI transit, largely related to their
anticholinergic and noradrenergic properties.
Sensitivity. TCAs are often advocated for the treatment
of visceral hypersensitivity. Using 3 different TCAs, a rat
study showed a dose-related attenuation of visceral afferent
pelvic nerve signaling to noxious colonic distension.
35
However, the evidence that they are effective in reducing
visceral sensitivity in humans is limited. Amitriptyline 80
mg acutely administered did not alter rectal sensitivity in
healthy controls.
34
In healthy volunteers, amitriptyline 50
mg had no effect on sensitivity to esophageal or rectal
balloon distention.
36
Imipramine in ascending doses up to
75 mg did not alter the level of distention needed to induce
first sensation, but increased the volume, although not the
pressure, needed to induce pain during esophageal balloon
distention.
37
It is unclear whether anticholinergic effects
underlie the lower volume sensitivity. Amitriptyline 25 mg
for 2 weeks did not alter nutrient volume tolerance.
29
In a controlled study in patients with noncardiac chest
pain, 3 weeks treatment with imipramine 50 mg improved
symptoms and decreased sensitivity to esophageal balloon
distention compared with placebo.
38
In a small cross-over
study in FD, amitriptyline improved symptoms, but this
was not accompanied by a change in visceral hypersensi-
tivity, as assessed by gastric balloon distention.
39
In a cross-
over study in 19 IBS patients, amitriptyline did not alter
perception ratings during rectal balloon distention, but
reduced the effect of stress on rectal distention-induced
activation of brain areas involved in visceral perception.
40
In an open-label study, amitriptyline up to 50 mg given
for 3 months to IBS patients decreased the stress-induced
increase in rectal sensitivity to electrical stimulation.
41
Taken together, there is limited and inconsistent evidence
that TCAs decrease visceral sensitivity in healthy humans
and in some patients with FGIDs.
Selective Serotonin Reuptake Inhibitors
(Paroxetine, Fluoxetine, Sertraline, Citalopram,
and Escitalopram)
Motility. Acute administration of SSRIs prolongs the
availability of physiologically released serotonin, thereby
potentially enhancing effects of serotonin released from the
GI tract, but also from the CNS.
28
Acute administration of
citalopram did not alter esophageal motor function in healthy
volunteers.
42
Five days of pretreatment with paroxetine
enhanced gastric accommodation in healthy volunteers.
43
Paradoxically, acute intravenous administration of cit-
alopram enhanced fasting gastric volumes, inhibited accom-
modation, and significantly enhanced solid gastric emptying
rate.
44
In the inter-digestive state, citalopram stimulated
occurrence of small intestinal phase 3 propulsive activity,
while suppressing gastric phase 3 propulsive activity.
44
Two
days of pretreatment with paroxetine 20 mg inhibited gall-
bladder emptying in healthy volunteers.
45
Four days of par-
oxetine 30 mg reduced orocecal transit time, but did not
significantly affect whole-gut transit time in healthy con-
trols,
32
and 11 days of treatment with paroxetine 20 mg
enhanced small bowel transit in healthy volunteers.
46
Acute
serotonin transporter inhibition in humans increased colonic
phasic contractility and the occurrence of high-amplitude
propagated contractions, increased colonic compliance, and
suppressed the colonic tonic response to a meal.
47
Taken together, these data show SSRIs enhances gastric
and small bowel propulsive motility.
Sensitivity. Administration of citalopram to healthy
controls does not alter sensitivity to multimodal esophageal
stimulation.
48
In contrast, in hypersensitive healthy volun-
teers, acutely administered citalopram decreased sensitivity
to acid perfusion and to balloon distention.
42
Intravenously
administered citalopram did not alter sensitivity to gastric
distention, but decreased meal-induced satiation scores and
increased the amount of nutrient ingested until satiation.
44
The latter may relate more to motor than to sensory ef-
fects. Citalopram intravenously did not alter sensitivity of
the rectum and the colon to distention.
47,49
In IBS, cit-
alopram also did not alter rectal or colonic sensitivity, and
fluoxetine did not alter rectal sensitivity.
47,49,50
Taken
together, SSRIs have no major impact on visceral sensitivity
in healthy subjects or patients with FGIDs.
Serotonin Noradrenalin Reuptake Inhibitors
(Duloxetine, Venlafaxine, and Milnacipran)
Motility. The actions of SNRIs on esophageal motility
have not been studied to date. Venlafaxine 75 mg did not
affect gastric emptying rate in healthy volunteers,
51
although venlafaxine overdose has been associated with
gastric bezoar formation.
52
Venlafaxine increased the meal-
induced change in gastric volumes in healthy volunteers,
suggesting an effect on gastric accommodation.
51
Venlafax-
ine 75 mg increased colonic compliance and decreased
fasting colonic tone and the tonic response to a meal, but did
not affect colonic transit in healthy controls.
51
Taken
together, there are some indications of an inhibitory effect of
SNRIs on gastric and colonic tone, but not to the degree of
TCAs; more studies are needed.
Sensitivity. SNRIs are known to have somatic analgesic
properties and are generally accepted to exert visceral anal-
gesic properties as well.
53
However, few studies have
addressed the effects of SNRIs on visceral sensitivity in health
or disease. In a colonic barostat study in healthy controls,
venlafaxine 75 mg increased colonic compliance, decreased
tone, reduced postprandial colonic contractions, and reduced
pain intensity ratings during graded distensions.
51
Noradrenergic and Specific Serotonergic
(Tetracyclic) Antidepressants (Trazodone,
Mirtazapine, and Mianserin)
Motility. Tetracyclic antidepressant drugs are used for
some functional disorders; trazodone mainly for esophageal
disorders and mianserin/mirtazapine mainly for FD. In
March 2018 GutBrain Neuromodulators: A Rome Working Team Report 1155
SPECIAL REPORT
healthy volunteers, mianserin 10 mg did not alter gastric
emptying rate, but inhibited gastric accommodation to a
meal.
54
Mirtazapine did not alter gastric emptying rate and
gastric accommodation in healthy volunteers.
55
The effect of
tetracyclic antidepressants on motility has not been studied
in much detail.
Sensitivity. The effects of trazodone on visceral sensi-
tivity in health have not been tested. Neither mianserin 20
mg, nor mirtazapine 15 mg, altered sensitivity to gastric
distention in healthy volunteers.
54,55
In FD patients, mirta-
zapine increased nutrient volume tolerance, while gastric
emptying rate was not significantly altered.
56
No conclu-
sions can be drawn on the effects of tetracyclic antide-
pressants on visceral sensitivity due to limited studies.
Azapirones (Buspirone and Tandospirone)
Motility. Only a handful of studies have evaluated the
actions of this class of agents on GI function. Buspirone
enhances contractile amplitude in the esophagus.
57
Also,
probably through 5-HT
1A
receptormediated inhibition of
acetylcholine release from cholinergic nerve endings in the
enteric nervous system, buspirone relaxed the proximal
stomach and slowed gastric emptying rate in healthy con-
trols in a dose-dependent manner.
58
In a cross-over pilot
study of FD, buspirone improved symptoms, and this was
associated with enhanced gastric accommodation.
57
There
are limited data to suggest that azapirones enhance esoph-
ageal contractions and increase gastric accommodation in
health and in FD.
Sensitivity. Buspirone did not alter sensitivity to
esophageal multimodal stimulation,
48
sensitivity to gastric
distention,
58
or rectal distention.
59
Buspirone did not
significantly alter colonic compliance, tone or sensitivity in
healthy controls.
46
Limited data suggest that azapirones do
not alter GI sensitivity.
Atypical Antipsychotics (Sulpiride, Levosulpiride,
Quetiapine Aripiprazole, and Olanzapine)
Motility. Sulpiride and levosulpiride are occasionally
used for the treatment of FD and gastroparesis, based on
their beneficial effects on gastric emptying rate in these
patients.
60
No data for atypical antipsychotics or other an-
tipsychotics have been published on motility effects in
healthy volunteers.
Sensitivity. Levosulpiride was shown to decrease
sensitivity to gastric distention in FD patients.
61
No studies
are available on the effects of atypical antipsychotics on GI
sensitivity in healthy volunteers and very limited informa-
tion suggests that it reduces gastric sensitivity in patients.
Delta Ligand Agents (Pregabalin and Gabapentin)
Motility. The effect of delta ligands on GI motility has
not been studied.
Sensitivity. The effect of delta ligands on sensitivity in
the upper GI tract has not been studied.Both pregabalin and
gabapentin increased distension sensory thresholds in IBS
patients with rectal hypersensitivity.
62,63
Systematic Review of Clinical Trials of
the Effects of GutLBrain
Neuromodulators on Symptoms in
Chronic Painful NonLGastrointestinal
Disorders and Functional
Gastrointestinal Disorders
A systematic literature review was undertaken to
determine the effect of gutbrain neuromodulators in
FGIDs. However, we were aware there would be a very
limited amount of controlled trials in some FGIDs, pre-
cluding broad conclusions and recommendations. For this
reason, we decided to also address the effects of centrally
acting neuromodulators on non-GI chronic painful condi-
tions, as they contribute to the concept of using these agents
for chronic pain management. The complete literature re-
view report, and a summary of the methodology used and
the results of the review, are available in the Supplementary
Material. The findings, grouped by class of agents, assessing
evidence in specific non-GI and FGID chronic pain condi-
tions are summarized in Table 3.
We also searched the literature for evidence in favor of
augmentation therapy with centrally acting neuro-
modulators (see section on “Use of augmentation treat-
ment”). We found 2 meta-analyses supporting the
augmentative effects of central neuromodulators in
depression (atypical antipsychotic or second antidepressant
added to a single antidepressant that was deemed insuffi-
cient).
64,65
There are no controlled studies to identify the
effects of augmentation in FGIDs, but 1 case series showed a
benefit of adding quetiapine in patients with IBS or func-
tional abdominal pain who failed TCA or SNRI
monotherapy.
66
The outcome of the systematic review, as well as clinical
expertise and consensus within the panel, were used to
generate recommendations for clinical application (see
section on ”Clinical Applications”).
Clinical Applications
As discussed, well-designed studies that evaluate the use
of gutbrain neuromodulators for specific FGIDs or symp-
toms are limited, so recommendations made herein blend
available data with well-designed smaller studies targeting
GI patients, the clinical experience of experts, and group
consensus. These therapeutic recommendations are
directed toward patients with chronic and treatment re-
fractory and/or painfully severe or overlapping FGIDs, often
with comorbidities, where the approach is to reduce
symptom burden and improve quality of life rather than
achieve resolution of specific symptoms. Therefore, these
recommendations are difficult to incorporate into controlled
drug trials because one cannot know whether a clinical
response is due to a specific drug or another therapeutic
intervention of augmenting character made in parallel un-
less there is a prominent drug effect.
The presence of a comorbid dominant psychiatric diag-
nosis (based on clinical assessments or questionnaires, such
1156 Drossman et al Gastroenterology Vol. 154, No. 4
SPECIAL REPORT
as the Hospital Anxiety Depression scale) may need to be a
primary consideration when selecting a neuromodulator.
Neuromodulators can also be selected based on their spe-
cific peripheral effects, such as to treat a co-occurring
disturbed bowel habit or chronic nausea. Also there are
no scientific data to support the common use of opioids in
chronic visceral pain
67
and their use carries a substantial
risk of unwanted side effects, including opioid-induced
constipation and opioid-induced central hyperalgesia, also
known as narcotic bowel syndrome (NBS).
6871
In the sections that follow, treatment advice is based on
the predominant GI symptom, FGID diagnosis, concurrent
non-GI symptom profiles, and underlying assumptions of
pathophysiological mechanisms of relevance, as illustrated
in Table 2. The key recommendations are boldfaced in the
text and summarized in Table 4.
Selection of Treatment Options Based on
Clinical Profiles
Chronic gastrointestinal pain. Abdominal pain is a
key symptom in many FGIDs and, depending on its bodily
location and time course, is a central part of the diagnostic
criteria in several disorders. Examples of painful FGIDs
include functional heartburn, epigastric pain syndrome
(EPS) (eg, FD), IBS, centrally mediated abdominal pain
syndrome, biliary pain, and anorectal pain or levator ani
syndrome.
27
In those patients with an FGID where pain is
dominant and of frequent occurrence, a central neuro-
modulator would be a logical part of the treatment profile.
Tricyclic antidepressants. Low to modest dosage
regimens of TCAs have the most convincing evidence of
benefit for treating chronic GI pain. TCA dosages in the
range of 2575 mg/d have been used in most studies, apart
from 1 study where a wider dosage regimen was allowed
with desipramine up to 150 mg/d.
72
Evidence for analgesic
effects from doses <25 mg is lacking. Only 1 study for IBS,
which is referred to frequently (Vahedi et al
73
) used a 10-
mg dose. The authors reported a significant improvement
from baseline to end of treatment on the primary outcome
in the intention-to-treat analysis, but this was also observed
with placebo. Of note, there was no significant difference
between treatment and placebo in the between group
comparison. We believe these data are not sufficient to
justify using what is likely a nonpharmacologic dose. How-
ever, a low starting dose of 10 mg can help patients over-
come early side effects, which can then lead to increasing to
a more desirable dose thereafter. This strategy can also help
address symptoms reported as side effects that are present
before treatment, or which relate to anxiety about TCA
treatment.
74
One can adjust the TCA dose within the first
46 weeks of treatment within the range of 2575 mg at
night, and then further increase to 100150 mg if no dis-
turbing side effects develop. Most often the occurrence of
anticholinergic and anti-histaminic side effects sets a natural
upper limit, although it should be recognized that the ter-
tiary amine TCAs (eg, amitriptyline and imipramine) are
more likely to produce these side effects compared with the
secondary amine agents (eg, desipramine, nortriptyline).
The antihistaminic and anticholinergic actions that might
cause drowsiness, dry mouth, palpitations, or constipation
are sometimes clinically useful for employing the tertiary
amine agents when the abdominal pain is associated with
sleep problems or diarrhea.
Serotonin noradrenalin reuptake inhibitor. SNRIs
may have at least equal benefit as TCAs for treating
chronic GI pain based on data for treating other chronic
painful disorders like fibromyalgia, migraine headaches,
widespread body pain, and peripheral neuropathy, though
they have not been adequately tested for chronic GI
pain.
7581
Although studies are lacking in the FGIDs, SNRIs
have empiric value based on the previously discussed
mechanisms. The advantage of SNRIs compared with TCAs
is that they do not have antihistaminic or anticholinergic
side effects, which can preclude adequate dosing with TCAs.
However, nausea is a common side effect. The risk for sig-
nificant side effects can be reduced by starting treatment in
the low-dose range, and supporting patients through the
first week of treatment before increasing to full doses. Other
side effects include palpitations, sweating, sleep disorders,
dizziness, and visual impairment.
81,82
Some patients can
experience anticholinergic side effects mediated through a
modulation of sympathetic tone by norepinephrine.
5
Comparing the different SNRIs, venlafaxine is probably
more prone to side effects when treating pain, including
increased diastolic blood pressure, which requires moni-
toring, at least in the higher dosage range of treatment.
83
The serotonergic effects dominate in the lower dosage
range for venlafaxine, and the dosage needs to be escalated
to 225 mg/d in order to reach norepinephrine effects suf-
ficient for pain modulation. Thus, duloxetine may be easier
to use in FGIDs, as it has clinically meaningful noradrenergic
even in the low dose range at start of treatment. Milnacipran
is an SNRI that is marketed for fibromyalgia and widespread
body pain and is not used for depression in the United
States, but is in European countries. Although there are no
studies to support its use in FGIDs, we recommend
considering this agent if there are difficulties in using the
other SNRIs due to side effects.
Augmentation by combining central neuro-
modulators. In the patient with anxiety or multiple
somatic symptoms (somatization) or where there are
incomplete benefits from TCAs or SNRIs, adding an
augmenting agent is the recommended next option for
treating chronic GI pain. For further details see the sec-
tion: “Augmentation with Central or Peripherally Acting
Treatments.”
Delta ligand agents. Delta ligand agents treat
neuropathic pain, or pain associated with fibromyalgia,
a condition commonly associated with FGIDs, particu-
larly IBS and may be of help with chronic GI pain,
though studies have not adequately addressed their
benefits in this condition. Pregabalin or gabapentin may
reduce visceral hypersensitivity.
63,84,85
Theoretically, in
clinical practice these agents might also be helpful for
abdominal wall pain, although no studies to assess this ef-
fect are available. There are data showing some benefit for
postoperative patients taking gabapentin.
86
Aside from its
March 2018 GutBrain Neuromodulators: A Rome Working Team Report 1157
SPECIAL REPORT
peripheral effects, pregabalin has been shown to have cen-
tral effects in patients with chronic fibromyalgia by
improving brain connectivity, which was associated with a
reduction in pain.
26
Brain imaging studies suggest that the
analgesic effects of pregabalin may also have a central
component, involving reductions in brain insula glutamate
levels.
26
Studies in FGIDs are largely lacking, but its clinical
use in certain situations, like when a general anxiety dis-
order or fibromyalgia/abdominal wall pain coexist, is
reasonable in a dosage between 150 and 600 mg/d with
effects expected within a month.
Other Potential Agents
The following classes are of potential value in treatment
of these disorders, but data are not sufficient to make
stronger recommendation.
Aminoketones. There is no formal evidence for using
bupropion in the treatment of abdominal pain, but it may be
considered from extrapolation of SNRI effects on descending
inhibitory nerve fibers, which can lead to an anti-
hyperalgesic effect. It is also associated with lower fre-
quencies of sexual dysfunction compared with other
antidepressants. A theoretical advantage could be to use
bupropion if there is a problem with fatigue and sleepiness
because the treatment effects of bupropion from this
respect are better when compared with the SSRIs.
87
It has
also been used as an augmenting agent in treating depres-
sion, when other antidepressants are not successful.
88
Dosage is the same as for the psychiatric indications, that
is, 150300 mg/d.
N-methyl-D-aspartate receptor antagonists. Me
mantine, ketamine, and dextromethorphan are N-methyl-D-
aspartatereceptor antagonists that can help reduce pain of
presumed neuropathic origin.
89
Memantine in particular has
shown analgesic effects in the clinical setting of fibromyalgia
and migraine headache, which are commonly associated
with the FGIDs, as well as for peripheral neuropathic
pain.
9092
The tolerability was good, with a number needed
to treat of 6 in a double-blind, placebo-controlled, ran-
domized trial in fibromyalgia with a dosage titrated to 20
mg/d within the first month of treatment.
90
With the limited
evidence at hand, this should be considered as a third-line
treatment in patients with FGID-associated abdominal
pain, where other more common treatment options have
failed.
Irritable Bowel Syndrome
IBS can have both peripheral (ie, visceral hypersensi-
tivity) and central contributions to the pain. Generally,
when the pain is mild to moderate and intermittent,
peripherally acting agents may be sufficient, but when pain
is more severe or persistent, central agents may be added
or substituted. For IBS, there are first-line treatment op-
tions that act peripherally on gut function for use in pa-
tients without psychiatric comorbidity and moderate to
severe symptom intensity. These agents are briefly
mentioned for completeness as putative peripheral neu-
romodulators, and their indications and pathophysiology
can be reviewed elsewhere. When diarrhea dominates
(IBS-D), alosetron, a highly selective 5-HT
3
antagonist has
been shown to increase the thresholds for visceral sensa-
tion and slow-down intestinal transit, by inhibiting the
extrinsic visceral pain pathway.
93,94
This class of drug,
while effective for IBS-D, is available under restricted li-
cense in the United States. A recent alternative, eluxado-
line, is approved by the FDA and the European Medicines
Agency as a new class of drug interacting with peripheral
opioid receptors and reducing symptoms in IBS-D with
more convincing effects on bowel habit compared with the
reduction in abdominal pain.
95
In patients where con-
stipation is the predominant bowel disturbance (IBS-C),
linaclotide has shown to be safe and with convincing ef-
fects on the combined end point that includes significant
reductions in both abdominal pain and relief of con-
stipation.
96,97
This is mediated by a prosecretory effect
from binding to the guanylate cyclase C receptor on in-
testinal epithelial cells, which activates intracellular for-
mation of cyclic guanosine monophosphate that both
stimulates chloride secretion through interaction at the
CFTR channel, as well as increases the threshold for
colonic sensation through effects on the primary afferent
gut neuron. Yet another secretagogue, lubiprostone, is
indicated for treatment of IBS-C, but with a mode of action
that involves activation of type-2 chloride channels.
66
The TCA class of drugs is the first-line central
neuromodulater for treating IBS, especially IBS-D. In
particular, the tertiary amine TCAs (amitriptyline and
imipramine) can reduce diarrhea, and also improve poor
sleep quality. In selected cases of IBS-D, the combined
central and peripheral effects of a TCA can suffice as a single
therapy option. A secondary amine TCA (desipramine and
nortriptyline) may be selected if less anticholinergic or
antihistaminic effect is desired (eg, for treating pain with
IBS with diarrhea and constipation or IBS-C).
The SNRI class of medications (duloxetine, ven-
lafaxine, or milnacipran) has potential to improve the
pain component of IBS, based on data from treating
other pain disorders, and has fewer side effects than
TCAs, but these agents have not been adequately stud-
ied in IBS. Bowel habits may be influenced by both the
serotonergic effect and the change in noradrenergic tone
that might have an indirect anticholinergic effect, but in
general these effects are fewer than with TCAs and may be
favored for treating IBS-C. These agents are desired partic-
ularly where abdominal pain is the major or primary
problem of significance to the patient.
SSRIs can be considered in IBS if anxiety states,
including hypervigilance, somatic symptom disorder,
visceral anxiety, and maladaptive cognitions, are pre-
sent, and the abdominal pain and diarrhea are not the
dominant clinical features. It can also be helpful when
constipation is present. Citalopram has been shown to in-
crease colonic contractility and reduce colonic tone during
fasting conditions and reduce the colonic tone increase after
meal ingestion.
47
In IBS, the same drug decreased scores for
abdominal pain, as well as bloating, independent of anxiety,
depression, and colonic sensorimotor function in a small
1158 Drossman et al Gastroenterology Vol. 154, No. 4
SPECIAL REPORT
crossover study involving 23 patients at a tertiary referral
center.
47
Also, improvements in overall well-being, regard-
less of coexisting depression, with paroxetine treatment in
IBS patients
98
and decreased abdominal discomfort in IBS-C
patients treated with fluoxetine
99
lends support that sub-
groups of patients benefit from mechanisms other than the
antidepressive and anxiolytic ones.
Functional heartburn and functional chest
pain. In patients with troublesome heartburn or chest pain
where gastroesophageal reflux disease has been excluded, a
treatment trial with a centrally acting pharmacologic agent
may be considered. There is insufficient evidence to
recommend a particular class of central agent, though SSRIs
have shown some benefit for esophageal pain along with
TCAs and SNRIs.
TCAs involving low-dose imipramine
38
or amitripty-
line
100
can be considered. Both studies are small, and the
latter study that showed superior effects compared with
double-dose proton pump inhibitor was open label.
The SSRIs have been shown to benefit some patients
with functional chest pain
101
or in situations where
esophageal hypersensitivity is suspected as a pathophysio-
logical mechanism of importance.
42
Like the considerations
in previous diagnostic groups, coexisting anxiety, depres-
sion, and phobic features strengthens a decision for a
therapeutic trial with a SSRI.
Regarding SNRIs,
102
a small dose (75 mg) of venlafaxine
in an extended-release formulation at night was superior to
placebo in a study involving a young patient group only
(2029 years) with functional chest pain, which indicates a
putative positive effect with a good tolerance at this low-
dose regimen.
Functional dyspepsia. When treating FD, it is best to
consider the Rome IV symptom-specific subgroups: post-
prandial distress syndrome (PDS) and EPS. With PDS,
meal-induced symptoms like fullness and early satiation
dominate, and in EPS, epigastric pain and a burning sensa-
tion that may or may not be associated with meal intake is
prominent. Where basic measures, like lifestyle adaptations
and eradication of Helicobacter pylori, if present, do not
result in sufficient symptom control, 1 or more of the
treatments outlined here can be used.
Postprandial distress syndrome. Buspirone, an
anxiolytic azapirone may be used for PDS where early
satiety, fullness, and nausea predominate. It has rela-
tively few side effects and no potential for physical depen-
dence. The mode of action involves 5-HT
1A
agonism, which
may improve receptive relaxation of the gastric fundus. In a
4-week study of buspirone in FD patients, dyspeptic
symptoms significantly decreased over placebo with regard
to nonpainful discomfort, such as early satiety, fullness,
bloating, and nausea. Interestingly, this was associated with
greater increases in postprandial gastric volumes, suggest-
ing that gastric accommodation may be the physiological
basis for this positive effect.
57
The same symptom response
has been seen after a nutrient satiety drink test as well.
46
The buspirone dosage should be the same as in the treat-
ment of anxiety, that is, 30 mg divided 2 or 3 occasions
daily, but can be increased to 60 mg/d.
Mirtazapine is a good treatment option for PDS
when there is chronic nausea and vomiting, or weight
loss, and it may also help coexisting abdominal pain.
The antihistaminic action may reduce sleep latency as
shown when treating depression.
103
Among FD patients
with weight loss, without coexisting anxiety or depression,
15 mg of mirtazapine in the evening for 8 weeks was su-
perior to placebo in improving overall symptom scores,
early satiation, and nutrient tolerance, and resulted in
weight recovery.
56
The regular dosage range is from 15 to
45 mg/d given in the evening, to reduce daytime sedation.
In those patients having too much sedation from mirtaza-
pine or with an incomplete response, olanzapine is an
alternative treatment option, as shown in anesthesiology
and oncology.
104
Using a dosage range of 2.5 to 10 mg/d, the
risk for its potential to cause neurologic side effects, aka-
thisia and dystonic reactions included, is low.
Epigastric pain syndrome. When the dyspeptic
symptoms are consistent with EPS, studies mainly support
the use of TCAs, either initially or after an unsuccessful
response to a proton pump inhibitor. A small study
including 38 patients showed benefit for amitriptyline over
placebo in improving total symptom score and nausea, as
well as upper abdominal pain.
105
A more recent multicenter
study also indicated a positive effect of treatment with
amitriptyline in FD patients with epigastric pain.
106
Patients
labeled as ulcer-like dyspepsia (Rome II definition for EPS)
had the most favorable response, with a more than 3-fold
higher likelihood of adequate relief compared with pla-
cebo or escitalopram, which did not show symptomatic
benefit in any FD subgroup. This effect was not seen in the
subgroup dysmotility-like dyspepsia (Rome II definition for
PDS). The American College of Gastroenterology guideline
also advocates the use of amitriptyline in FD patients when
proton pump inhibitors have failed to provide relief.
107
As with other painful FGIDs, the SNRI group of medi-
cations can also be considered for patients with EPS who
do not tolerate TCA treatment, although confirmatory
studies are lacking. The side effect of nausea may limit the
usefulness in this group of patients, particularly in the group
having clinical features that overlap with PDS.
Cyclic vomiting syndrome. CVS is characterized by
sudden, stereotypical episodes of intense nausea and
vomiting lasting up to days at a time with symptom-free
intervals.
108
The condition varies in severity from a few,
self-limiting episodes per year to a debilitating situation
with need for frequent hospital admissions. There are 2
treatment components for CVS. First, for treating acute
episodes, the effort is to reduce the symptom severity
and duration of the attacks with anti-emetic agents like
ondansetron or promethazine, minimize the use of
opioids for the abdominal pain, and intravenous hy-
dration. Benzodiazepines can be used to treat acute
anxiety and distress and have independent effects on
nausea reduction. Second, to prevent future episodes,
cannabinoids must be eliminated, and prophylactic
treatment using central neuromodulators (TCAs, SNRIs,
tetracyclics, atypical antipsychotics, and anticonvul-
sants as discussed in the next paragraph) should be
March 2018 GutBrain Neuromodulators: A Rome Working Team Report 1159
SPECIAL REPORT
instituted to reduce the severity and frequency of acute
attacks to every 4 months or less.
Acute episodes of CVS are often triggered by physio-
logical or emotional stress, so it is important to identify
comorbid psychiatric problems and treat them accord-
ingly. This is in line with the general approach to FGID
management. In patients with frequent attacks of intense
vomiting, the first factor to consider is whether the patient
may have the cannabinoid hyperemesis syndrome, which
can mimic CVS. In that situation, treatment involves
eliminating this substance.
108
Following this, TCAs can be
considered as first-line treatment for prophylaxis. Among
them, amitriptyline has some weak evidence in support
from open-label trials,
109
and the doses reported are
within the same range as for treatment of GI pain. SNRIs
can also be used as second-line therapy, but without any
formal evidence base.
Other prophylactic agents include mirtazapine, a tetra-
cyclic antidepressant because of its potential effects on re-
lief of complex nausea and pain, and augmenting agents that
concurrently reduce anxiety, including atypical antipsy-
chotics, in particular olanzapine due to its antiemetic effects,
and SSRIs in addition to TCAs or SNRIs. Furthermore, anti-
convulsants, such as zonisamide and levetiracetam, have
been reported to be effective in a retrospective follow-up in
adults failing TCA therapy, where 75% had at least a mod-
erate clinical response and 20% achieved symptomatic
remission during a time period of <1 year.
110
Use of Augmentation Treatment
When a single treatment, as discussed in the section on
clinical applications, is either ineffective and/or produces
side effects or is not well tolerated, augmentation treat-
ment can be considered. This concept was demonstrated in
psychiatry when a second antidepressant improved
depression after monotherapy was unsuccessful,
111
has
also been reported for medical disorders like chronic
headache
112
and premenstrual dysphoric disorder,
113
and
has been recommended empirically for functional GI pain.
6
Augmentationcanbeachievedbyaddingtoanantide-
pressant used for pain, a second centrally acting agent, or
one that acts peripherally, or a behavioral treatment
(Figure 5). It is believed that benefit relates to recruiting
additional neural receptors or pathways (eg, dopaminergic
receptors with an atypical antipsychotic added to an SNRI
or TCA), and dosages may be lower, thus minimizing side
effects.
Central or peripherally acting treatment aug-
mentation. Augmentation treatment, that is, adding
a central with a peripheral or 2 central agents (eg,
atypical antipsychotic agent to an antidepressant), is
recommended when monotherapy is not successful.
Combining central agents is accepted in treating
psychiatric disorders,
64
but systematic studies that
evaluate combined central agents or central and pe-
ripheral agents in treating chronic pain and FGIDs are
lacking. There is, however, growing expert clinical
consensus to guide clinicians in this type of
treatment.
6,10,27
When a partial effect is obtained by a
specific pharmacologic agent (TCAs, SSRIs, and SNRIs), it
can be beneficial to add an additional agent to combine
effects, and to allow for use of a lower dosage to minimize
side effects.
66
Such synergistic positive effects are best
achieved with drugs that have complementary mecha-
nisms of action. For example, an SSRI can be added if a
TCA has resulted in some pain relief, but with insufficient
control of coexisting anxiety, because the usual TCA dose
most often is not sufficient to treat the psychiatric con-
dition or to produce serotonin-related side effects.
Another option is to add an atypical antipsychotic. These
agents, unlike the previous classes of antipsychotics, have
less risk for extrapyramidal side effects. There is some
experience from the use of quetiapine in the treatment of
chronicpain.Mostdataarefoundinthetreatmentoffi-
bromyalgia, where a controlled study reported effects
superior to placebo on the pain domains,
114
but with ef-
fects inferior to amitriptyline in yet another study.
115
Quetiapine has also been shown to improve pain when
used to augment the effects of a TCA or SNRI.
66
The
complex actions at the receptor level
116
can have added
clinical effects, like anxiety reduction and establishment
of a normal sleep pattern.
65,117
Its main metabolite also
has effects as a norepinephrine transporter inhibitor,
which is of theoretical advantage for analgesic effects.
118
Higher dosage ranges of 200400mg/dasseenin
psychiatry can lead to poor tolerability, with a high pro-
portion of patients experiencing excess sedation and
dizziness, in addition to metabolic side effects like weight
gain, hyperlipidemia, and diabetes (metabolic syndrome).
In balance, because most studies in psychiatry report
these adverse events with the higher doses, the recom-
mended range of 25200 mg for GI symptoms might be
considerably safer.
When combining medications, it is important to have a
familiarity with each agent’s side effect profile, and be aware
of potential hazardous side effects, such as serotonin syn-
drome, which is characterized by fever; hyper-reflexia;
spontaneous clonus; muscle rigidity; and, if not treated
immediately, increased risk of death. A simplified algorithm
of augmentation strategies for chronic abdominal pain is
summarized in Figure 5.
Psychological/behavioral augmentation. Psycho-
logical or behavioral treatment adds to the benefit to phar-
macologic agents working on deeper brain areas by their
action on more frontal “executive”areas of the brain.
119
Cognitive behavioral therapy (CBT) helps to reduce cata-
strophic thinking, increases one’s sense of control, and
reframes the sense of threat. These brain changes are asso-
ciated with reduced intensity of pain experiences and im-
proves psychological well-being and function. Many
randomized controlled trials support the efficacy of behav-
ioral interventions as an adjunct to pharmacologic strategies
in managing FGID symptoms.
120125
CBT, gut-directed hyp-
notherapy, and mindfulness meditation have been studied
most in patients with FGIDs and show empiric
support.
126132
The most significant improvements (mod-
erate effect size) occurred for abdominal pain intensity, IBS
1160 Drossman et al Gastroenterology Vol. 154, No. 4
SPECIAL REPORT
symptom severity, quality of life, anxiety, depression, and
daily functioning. Although CBT is associated with convincing
short-term improvement, gut-focused hypnotherapy has the
more consistent support in longer-term efficacy,
133136
although in a primary care study longer-term hypnotic
effects were not maintained.
137
Comprehensive
self-management interventions combining CBT with relaxa-
tion and dietary strategies have also shown promise both
short-term and long-term(12 months) in patients with IBS.
138
Virtual therapies delivered via the internet or on mobile
electronic platforms may be an effective option for patients
with FGIDs.
139141
Nurse-led hypnotherapy to improve IBS
care is showing promise.
142,143
Although these behavioral
approaches are often used to supplement pharmacologic
strategies, few studies compare behavioral interventions
alone to their combination with specific psychotropic
agents. Instead, this comparison has to be extrapolated from
studies where those in the treatment as usual arm were
receiving medications and little psychosocial care.
Relapse prevention. Relapse prevention relates to the
concept that continued treatment with a neuromodulator
beyond the period of achieving clinical benefitwillreducethe
likelihood of relapse or recurrence.
144
This is supported
through evidence that continued antidepressant treatment
may be associated with reversal of the clinical disorder
through neurogenesis.
12,18
It has also been demonstrated from
clinical experience of the authors when treating FGIDs. It offers
guidance for the FGIDs
6
because many of the same agents are
utilized. For major depression, it is recommended that anti-
depressants be continued for at least 4 to 9 months after an
initial positive response. To reduce the likelihood of relapse
when treating FGIDs, we empirically recommend that treat-
ment be continued 612 months after treatment response. In
1 meta-analysis of 31 randomized trials involving more than
4000 patients with depression, continuing treatment with
antidepressants reduced the odds of relapse by 70%.
20
Relapse is higher in patients where treatment is stopped,
compared with those where it is continued. Continuing or
starting behavioral interventions (CBT and hypnosis) also
reduced relapse risk.
121,145147
Where factors such as ongoing
psychosocial stress, positive family history, history of multiple
prior episodes, or current psychiatric comorbidities exist,
longer-term treatment can be empirically considered.
Risks of Using Opioids for Management of
Chronic Abdominal Pain and Availability of
Non-Opioid Alternatives
Opioids. The number of opioids prescribed around the
world for chronic pain, particularly in the United States and
Canada (see Supplementary Figure 1), is growing dramati-
cally.
148,149
This rate parallels increased heroin use,
Figure 5. Summary of the clinical characteristics that can be considered when selecting gutbrain neuromodulating phar-
macotherapy to treat FGIDs. Those drugs in the upper part of the figure can be considered as first-line options. In the lower
part of the figure, the pharmacologic options most often used to augment treatment effects are depicted, as well as some
nonpharmacologic treatment alternatives.
March 2018 GutBrain Neuromodulators: A Rome Working Team Report 1161
SPECIAL REPORT
incidence of HIV and hepatitis C, and numbers of overdose
deaths involving prescription narcotics and heroin.
150153
In the United States, more than 12 million people re-
ported misusing opioids, with at least 2 million endorsing
addiction in 2015
154
(www.samhsa.gov/data). Despite the
growing use of opioids to treat chronic non-malignant pain,
there is no evidence that such treatment leads to lasting
clinical response for FGIDs, and prolonged use is associated
with greater clinical harm, including opioid-induced con-
stipation and NBS.
70,71,155
Furthermore, nonmedical pre-
scription opioid misuse disorders have increased in the past
decade, and prevalence was highest in whites and Native
Americans, those with lower socioeconomic status, and in
patients with mood disorders, post-traumatic stress disor-
der, and personality disorders.
156
Risk factors for death
from prescription opioids include male sex, sedative hyp-
notic use, total number of prescriptions, and receiving a
daily average of >40 mg morphine equivalents.
157
Unfor-
tunately, there has been little change in the medical man-
agement of prescription-opioid consumers, even after
substance abuse is diagnosed.
158
The Surgeon General of
the United States strongly discourages the prescription of
opioids as first-line treatment for chronic non-cancer pain
in adults.
159
Approximately 20% of patients with FGIDs use opioids
chronically for management of their GI symptoms,
160,161
yet
there are no randomized controlled trials documenting the
efficacy of opioids for chronic abdominal pain. Furthermore,
the known GI side effects of chronic opioid use can be
particularly problematic, and include constipation, nausea,
abdominal pain, gaseousness, ileus, and acid reflux.
162
Opioid-induced constipation is associated with high medi-
cal utilization.
163
There is also an increasing recognition of
NBS, consisting of increased abdominal pain with chronic
opioids, which is thought to be a centrally mediated
hyperalgesia.
164
This is due to inflammatory processes in
the spinal cord and changes in the functioning of the opioid
mu receptor within the dorsal root ganglion.
165
Chronic
opioid use is also associated with increased risk for addic-
tion and other psychopathology, particularly mood and
anxiety disorders.
166,167
Non-opioid alternatives. For patients with FGIDs who
are on chronic opioids, a careful tapering is recommended
when there is insufficient relief of pain, evidence of opioid
misuse, and/or side effects, or other negative sequelaecovered
in the section above.
71,168,169
Patients with NBS showed a
reduction in pain intensity 3 months after detoxification,
although there was a high rate of opioid recidivism at 6
months.
168
Here it is important to offer patients nonopioid
alternatives, such as central neuromodulators and behavioral
interventions longer-term, to adequately manage chronic
abdominal pain. Clonidine has been used to prevent with-
drawal effects, although several studies show that buprenor-
phine may be even more effective during and immediately
after detoxification.
170,171
The most supported nonopioid al-
ternatives for chronic abdominal pain include TCAs and SNRIs,
which can also treat comorbid mood and anxiety symptoms.
Mood stabilizers, such as gabapentin, carbamazepine, and
topiramate, have shown efficacy in neuropathic chronic pain
syndromes, but their utility in FGIDs has not been studied
specifically.
CBT, gut-directed hypnotherapy, and mindfulness tech-
niques have a growing evidence base for treating GI symp-
toms of IBS and other FGIDs, and multidisciplinary
treatment of pain involving behavioral experts who are in-
tegrated into the medical care can help with longer-term
management of chronic pain.
172
Implementation: Use of Communication
Skills to Improve Patient Engagement,
Acceptance, Adherence, and the
PatientProvider Relationship
The decision to prescribe a neuromodulator for GI
symptoms is based on reasons that are not necessarily (and
often not) consistent with the patient’s understanding for
their use. This lack of concordance often relates to a dual-
istic perspective that the medications being recommended
to the patient are “psychiatric,”
1
rather than related to
treating disorders of gutbrain interaction.
2
Thus, the
clinician must educate the patient on the value of these
treatments, and this is best implemented through an effec-
tive patientprovider relationship. To merely recommend
treatment and not engage with the patient on this under-
standing may lead to refusal to take the medication, non-
adherence, or reporting side effects not related to the
medication but due to anxiety relating to taking it.
74
Establish the Therapeutic Relationship
An effective therapeutic relationship is essential
when caring for patients with disorders of gutLbrain
interaction, and when prescribing neuromodulator
treatment. Certain recommendations should be considered,
the first being to understand patient expectations. Some
patients may adhere to a more acute model of care: their
symptoms are presented, they expect the doctor to conduct
tests, make a diagnosis, and institute treatment that resolves
the problem. However, with chronic and painful illnesses,
the primary focus is adaptation to chronic or recurring
symptoms with little chance of cure. This must be addressed
on the first visit by asking: “What brought you to see me at
this time?”;“What worries or concerns do you have?”;
and “What do you expect from treatment?”More detailed
guidelines for establishing a therapeutic relationship can be
found elsewhere.
173
General guidelines to establish a
therapeutic relationship include: (1) listening actively
to determine the patient’s understanding of the illness
and his or her concerns, (2) providing a thorough
explanation of the disorder, (3) identifying and
responding to the patient’s concerns and expectations,
(4) setting realistic and consistent limits, (5) involving
the patient in the choice of treatment, and (6) estab-
lishing a long-term commitment to the care.
173,174
This
type of approach is associated with reductions in health care
visits
175
and nonadherence, and improvements in patient
satisfaction, symptom reduction, and other health out-
comes.
173
Guidelines are provided in a video link: http://
www.youtube.com/watch?v¼BeHPpvuB_mc.
1162 Drossman et al Gastroenterology Vol. 154, No. 4
SPECIAL REPORT
Educate to Legitimize the Disorder
Education is an iterative process in which the
physician assesses the pati