Beruflich Dokumente
Kultur Dokumente
Lukas Van Oudenhove,1,2 Hans Trnblom,3,4 Stine Strsrud,3 Jan Tack,1 and
Magnus Simrn3,4
1
Translational Research Center for Gastrointestinal Disorders, Department of Clinical and Experimental Medicine, University of
Leuven, Leuven, Belgium; 2Consultation-Liaison Psychiatry, University Psychiatric Centre KU Leuven, University Hospitals
Leuven, Leuven, Belgium; 3Department of Internal Medicine and Clinical Nutrition, Institute of Medicine, 4University of
Gothenburg Centre for Person-Centered Care, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden
BACKGROUND & AIMS: Patients with irritable bowel syndrome
(IBS) have increased postprandial symptom responses and more
psychosocial morbidities than healthy individuals. However, the
relationship between psychosocial status and postprandial
symptom responses in patients with IBS is unclear. We investigated this relationship in a prospective study of patients with
IBS. METHODS: A total of 193 IBS patients, diagnosed according
to Rome II (n 126) or Rome III (n 67) criteria, consumed a
standard breakfast (540 kcal: 36% fat, 15% proteins, 49% carbohydrates, and 8.9 g ber). They completed visual analogue
scales assessing the severity of 5 gastrointestinal symptoms
(abdominal pain, bloating, nausea, gas, and fullness) before
breakfast and every 30 minutes, up to 240 minutes after eating.
All patients completed validated self-report questionnaires for
their current levels of anxiety and depression; patients diagnosed based on Rome II criteria also completed a somatization
questionnaire. The relationship between these variables and the
course of gastrointestinal symptom scores over time was
analyzed using linear mixed models, controlling for comorbid
functional dyspepsia. RESULTS: We observed a main effect of
anxiety levels on fullness and bloating (P < .04), and of
depression levels on abdominal pain (P .007), reecting a
general upward shift of the entire symptom curve. Depressionby-time interactions were seen for nausea and gas (P < .03).
Somatization levels had a main effect on all 5 symptoms
(P < .0001), independent of anxiety and depression. We
observed somatization-by-time interaction effects for bloating
(P .005), and nausea (P .02), and a nonsignicant trend was
found for pain (P .054), reecting a steeper early postprandial
increase in symptoms among subjects with higher levels of
somatization. CONCLUSIONS: Based on a prospective study of
patients with IBS, psychosocial morbidities are associated with
increased levels of gastrointestinal symptoms in general.
Depression and somatization levels are associated specically
with increased postprandial symptoms.
867
Psychosocial Assessment
For the assessment of anxiety and depression symptom
levels during the past week, patients completed the validated
Hospital Anxiety and Depression Scale.26 The Hospital Anxiety
and Depression Scale consists of 14 items in total (7 anxiety
items and 7 depression items), each scored on a Likert scale
ranging from 0 to 3. Subscale scores for anxiety and depression
are calculated by summing the 7 respective items.
Somatization (the tendency to report multiple somatic
symptoms) during the past 4 weeks was assessed using the
somatic symptom severity module of the Patient Health Questionnaire (PHQ), consisting of the 15 most common somatic
symptoms (PHQ-15).27 The severity of each of these symptoms
and their associated distress is scored on a Likert scale ranging
from 0 to 2, resulting in a total somatization score ranging from
0 to 30. To avoid confounding of the relationship between
somatization and our GI symptom outcome variables as well as
with the comorbid functional dyspepsia (FD) variable (see
later) by the GI items of the PHQ-15, these 3 items were
omitted as previously described and validated (PHQ-12).28
Because to the best of our knowledge no validated cut-off
values for the PHQ-12 exist, we also report the PHQ-15
scores for descriptive purposes. Somatization was measured
only in the Rome II subsample (n 127; 65.3%).
Statistical Analysis
SAS 9.3 software (SAS Institute, Cary, NC) was used to
analyze the data, which are expressed as means SEM. Signicance was set at a P value of less than .05.
To test our hypothesis that levels of psychological symptoms (anxiety, depression, and somatization) are associated
with an increased postprandial symptom response in IBS patients, we used marginal linear mixed models with main effects
CLINICAL AT
April 2016
868
Van Oudenhove et al
CLINICAL AT
Results
Descriptive Results
We included a total of 193 IBS patients (mean age, 41.3
1.0 y; body mass index, 24.0 0.4 kg/m2; 143 [74.1%]
women). A total of 126 (65.3%) patients were recruited
before the introduction of the Rome III criteria1 and, hence,
were diagnosed according to the Rome II criteria.25 Somatization data were available only in this subgroup. The
remaining 67 (34.7%) patients were diagnosed according to
the Rome III criteria because these patients were included
after their introduction.1 Rome III patients had slightly,
although signicantly, higher IBS-SSS total scores (270.0
97.4 vs 303.5 88.3; t[189] 2.41; P .017). Given the
small effect size for this difference (Cohens d 0.36), we
believe this does not represent a major problem for pooling
subjects from both subgroups in our mixed-model analysis
on the effect of anxiety and depression on postprandial
symptoms reported later, especially because controlling for
IBS-SSS score in these models did not change the effects of
anxiety and depression for any of the GI symptom outcomes
(data not shown). The fact that patients were recruited
prospectively from a single center by 2 gastroenterologists
experienced in IBS (H.T., M.S.), and that Rome II and Rome
III criteria for IBS are relatively similar, further ensures
n (%)
41.3 1.0
143 (74.1)
50 (25.9)
24.0 0.4
126 (65.3)
67 (34.7)
54 (28.3)
81 (42.4)
56 (29.3)
43.2
55.9
57.6
68.9
65.0
291.5
2.0
2.4
1.9
1.9
1.8
6.7
13
86
94
116
(6.7)
(44.6)
(48.7)
(60.1)
5.9 0.3
5.8 0.3
51 (26.4)
21 (10.9)
49 (25.4)
9 (4.1)
7.1 0.4
April 2016
869
F
b SEa
Time
Anxiety
Anxiety-by-time
Depression
Depression-by-time
Functional dyspepsia
3.1
0.8
1.3 1.5
1.1
7.35
3.9 1.5
0.8
4.4
5.1 2.5
P
.003
.37
.36
.007
.64
.038
Fullness
F
b SEa
24.4
6.7
4.8 1.9
0.3
0.1
-0.7 1.9
1.4
10.3
9.9 3.1
Bloating
P
<.0001
.010
.96
.70
.19
.002
F
b SEa
8.3
4.3
3.6 1.8
0.9
0.3
0.9 1.7
1.5
4.0
5.7 2.8
Nausea
P
<.0001
.039
.55
.59
.16
.047
F
b SEa
2.5
0.02
0.21 1.5
1.81
2.8
2.5 1.5
2.4
0.8
2.2 2.5
Gas
P
.012
.89
.077
.097
.017
.38
F
b SEa
3.18
2.2
2.4 1.7
1.2
0.0
0.06 1.6
2.3
1.5
3.2 2.6
P
.002
.14
.28
.97
.021
.23
Abdominal Pain
A signicant main effect of depression was found (P
.007), reecting an upward shift of the entire curve with
increasing levels of depression, as illustrated in Figure 1A.
For somatization, a main effect was found (P<.0001), as
well as a nonsignicant trend (P .054) for the interaction
effect with time. Figure 2A indeed shows that the entire
curve not only is shifted upward with increasing levels of
somatization, but that increasing somatization levels also
are associated with a different shape of the curve, particularly a steeper increase in pain rating in the early postprandial period (with the curve being virtually at in
subjects with low somatization levels). The b coefcients for
the time-by-somatization interaction effect conrm that the
effect of somatization is signicantly stronger compared
with the preprandial time point up to the 90th minute
postprandially (details not shown). The effect of comorbid
FD was signicant in the model with anxiety and depression, but not in the model with somatization.
Fullness
Signicant main effects were found for anxiety (P .01)
and somatization (P < .0001), as shown in Figures 1B and
2B, respectively. The effect of comorbid FD was signicant
in both models.
Bloating
Signicant main effects were found for anxiety (P
.039) (Figure 1C) and somatization (P < .0001). For somatization, in addition to the main effect, a signicant time-bysomatization interaction effect was found (P .005).
Figure 2C shows that this effect again reects a steeper increase in bloating ratings in the early postprandial period as
well as an additional increase in bloating ratings in the later
postprandial period in subjects who score higher on somatization, with the curve being virtually at in subjects with
low somatization levels. The b coefcients for the time-bysomatization interaction effect conrm that the effect of
somatization is signicantly stronger compared with the
preprandial time point up to the 60th minute postprandially
and again at the 180-minute postprandial time point (details
not shown). The effect of comorbid FD was (borderline)
signicant in both models.
Adding the main effect of subtype and its interaction
effect with time to the models did not change the signicance of any of the other effects (details not shown). Similar
to the model without psychological variables reported
earlier, the subtype-by-time interaction effect was signicant in both models, whereas the main effect of subtype was
not, indicating that the effect of subtype and the effects of
psychological variables on the time course of the bloating
response are independent.
Nausea
A signicant depression-by-time interaction effect was
found (P .017), reecting an enhanced early (up to the
60th minute) and late (from the 210th minute) symptom
response in patients with higher depression scores
(Figure 1D). For somatization, signicant main and interaction effects were found (P < .0001 and P .018,
respectively). The interaction effect reects steeper early
postprandial increases in nausea ratings with increasing
levels of somatization (Figure 2D). The b coefcients for the
time-by-somatization interaction effect show that the effect
of somatization is signicantly stronger compared with the
preprandial time point at the 60th minute postprandially
CLINICAL AT
Abdominal pain
870
Van Oudenhove et al
2.72
23.4
7.9 1.6
2.0
0.6
2.2 2.9
P
.009
<.0001
.054
.45
Fullness
F
b SEa
11.7
29.6
11.5 2.1
1.5
8.0
10.6 3.8
Bloating
P
<.0001
<.0001
.18
.005
F
b SEa
4.0
20.1
9.0 2.0
3.0
3.9
6.7 3.4
Nausea
P
.0003
<.0001
.005
.052
F
b SEa
1.1
29.4
9.8 1.8
2.4
0.02
0.5 3.3
Gas
P
.41
<.0001
.018
.88
F
b SEa
1.8
19.6
8.3 1.9
1.8
1.2
3.4 3.1
P
.086
<.0001
.079
.28
NOTE. The somatization analyses were performed on the Rome II subgroup only (n 125; 65.3% of the total sample); all
models controlled for main effects of anxiety and depression, as well as anxiety-by-time and depression-by-time interaction
effects. Signicant effects (P < .05) are shown in bold.
a
Shown when a single b was available (ie, for main effects of continuous or dichotomous variables).
Gas
A signicant time-by-depression interaction effect was
found (P .021), reecting a late postprandial increase in
gas rating in subjects scoring high on depression, with the
inverse pattern observed during the early postprandial
period (Figure 1E). Furthermore, a signicant main effect of
somatization was found (P < .0001) (Figure 2E). The effect
of comorbid FD was not signicant in any of these models.
Discussion
In this study we demonstrate that increasing levels of
anxiety, depression, and somatization are associated with
higher preprandial and/or postprandial GI symptom levels
in IBS patients. More specically, increasing levels of anxiety
were associated with an upward shift of the timesymptom
curve over its entire course (ie, to the same extent at the
preprandial baseline and the postprandial time points) for
fullness and bloating. Increasing depression levels were
associated with abdominal pain levels in a similar way. For
nausea, however, increasing levels of depressive symptoms
Figure 1. Overview of the signicant associations between anxiety, depression, and preprandial and postprandial gastrointestinal symptoms ratings in Rome II/Rome III IBS patients (n 193). (A) Depression and abdominal pain, (B) anxiety and
fullness, (C) anxiety and bloating, (D) depression and nausea, and (E) depression and gas. VAS, visual analog scale. Error bars
represent the standard error of the mean; time 0 represents the preprandial time point, and other time points are in minutes
after the meal.
871
CLINICAL AT
April 2016
Figure 2. Associations between somatization and preprandial and postprandial gastrointestinal symptom ratings in Rome II
IBS patients (n 125). (A) Abdominal pain, (B) fullness, (C) bloating, (D) nausea, and (E) gas. VAS, visual analog scale. Error
bars represent standard error of the mean; time 0 represents the preprandial time point, and other time points are in minutes
after the meal. All analyses are controlled for the effects of anxiety and depression (main effects and interaction effects with
time).
were associated with a change in the shape of the postprandial symptom curve, more specically, with a steeper
early postprandial increase as well as the presence of a
smaller late postprandial increase. For gas, increasing
depressive symptom levels were associated with increased
symptom responses in the late postprandial phase only. The
strongest effects, however, were found for somatization.
These effects were independent of those of anxiety and
depression, but it should be noted that somatization data
were available in the Rome II subsample only. Besides signicant main effects of somatization for all 5 GI symptoms
(indicating an overall higher symptom level over all time
points), increasing levels of somatization were associated
with steeper early postprandial increases for abdominal
pain (although this effect did not reach signicance, P
.054), bloating, and nausea. Taken together, these results
indicate that, irrespective of time relative to meal intake (ie,
both preprandially and postprandially), increasing anxiety
and depression levels are associated with higher levels of
specic individual GI symptoms (fullness and bloating for
anxiety, pain for depression), whereas increasing somatization levels are associated with higher levels of all GI
symptoms. In addition, however, for some symptoms, higher
depression and somatization levels are associated with
increased postprandial symptom levels specically. These
relationships occurred independently of comorbid FD, with
FD comorbidity being on the high end of the range reported
in the literature.32 Because interaction effects were
found primarily for symptoms that often are present in FD
(fullness, bloating, nausea), and these effects were driven
872
Van Oudenhove et al
CLINICAL AT
effects on symptom reporting in general (ie, over the preprandial and postprandial time period together) and postprandial symptom responses specically.
We believe that the present ndings are clinically
important because they suggest that assessing levels of
depression and somatization and, to a lesser extent, anxiety,
may be indicated in case of postprandial worsening of
symptoms in IBS. It obviously remains to be determined
whether successful treatment of psychological morbidity
would normalize postprandial symptom responses, as has
been shown for mechanosensitivity.37
Given the lack of earlier data on this matter and the lack
of physiological measurements in the present study, we can
only speculate on the potential mechanisms involved in the
associations we found. It is known from the pain literature
that normal psychological processes such as emotions
(anxiety, sadness) and, a fortiori, their pathologic equivalents such as anxiety and depressive disorders can inuence
pain perception and symptom reporting by interfering with
systems involved in processing and modulating noxious
(visceral) afferent signals.24,3840 It is conceivable that
similar mechanisms are involved in modulating the processing and, hence, the perception of physiological gut
brain signals such as nutrient-induced signals. However, this
speculative interpretation would require testing using, for
example, functional brain imaging approaches. Furthermore,
it remains to be elucidated why depression levels are
associated specically with postprandial symptom levels for
some symptoms, whereas anxiety levels only seem to be
associated with symptom reporting irrespective of time
relative to meal intake. In any case, these ndings are in line
with the observation that both depressive and anxiety disorders are associated with increased levels of GI symptoms
in general.41 Another candidate mechanism to explain the
association between depression levels and postprandial GI
symptom levels in IBS patients is the (afferent and efferent)
autonomic nervous system (ANS), which acts as the principal interface of the gutbrain axis. Abnormal ANS
responses to a meal indeed have been described in IBS,42
and depressive disorders are associated with abnormal
(efferent) ANS function,43 which may in turn inuence GI
symptom perception by interfering with motor and other
responses of the GI tract to meal ingestion. Finally, we may
speculate that gastrointestinal hormones such as peptide
tyrosine tyrosine (PYY) and/or cholecystokinin (CCK) may
play a role in the associations we found because IBS patients
have abnormal levels of PYY44 and abnormal CCK responses
to a meal.45 Furthermore, CCK may increase rectal mechanosensitivity,46 and mood disorders have been associated
with CCK genetic polymorphisms47 as well as abnormal PYY
levels.48
The mechanisms underlying the associations between
somatization level and (postprandial) GI symptom levels
(which was studied in the Rome II subsample only) may be
even more difcult to speculate. Evidence is growing that
increased somatic symptom reporting (ie, somatization) is a
multidimensional phenomenon. More specically, interrelated psychobiological processes such as negative affectivity (including personality traits such as neuroticism as well
as mood and anxiety disorders), somatosensory amplication, hypervigilance, fear learning, symptom-associated anxiety, attributions, cognitions, and behaviors may constitute the
key underpinnings of somatization.4951 Applied to our results, we may speculate that most of these processes may
increase the perception of postprandial GI signals, as already
has been shown convincingly for (somatic) pain signals and
symptoms.3,5255 For example, sensory amplication of
physiological nutrient-induced signals as well as hypervigilance, anxiety, or catastrophizing thoughts toward the sensations elicited by these signals may increase their perception
as symptoms. However, because the PHQ-12 operationalizes
somatization as the distress elicited by somatic symptoms,
the associations with somatization we found in our study
may reect a more general association between (psychological) distress and (postprandial) GI symptom reporting, even
though we controlled for levels of anxiety and depression in
the somatization analyses.
This study had a number of limitations that should be
addressed. First, this was a cross-sectional study, thereby
inherently limiting the conclusions that can be drawn
regarding the directionality or causality of the associations
found. Therefore, it remains possible that the higher (postprandial) GI symptom levels are the cause rather than the
effect of increased levels of depression, anxiety, or somatization. However, based on the literature on psychological
amplication of bodily signals outlined earlier, as well as
epidemiologic studies,20,21 we consider the latter interpretation (ie, increasing psychological distress levels induce
increased [postprandial] GI symptom responses) most likely
to be correct. Second, the study was performed on a mixed
Rome II/Rome III sample. Therefore, our results on subtype
should be interpreted with sufcient caution. Furthermore,
somatization data were available only in the Rome II subgroup, representing 65.3% of the total sample. Therefore,
our ndings on somatization cannot be generalized to Rome
III IBS patients, and should be interpreted with somewhat
more caution given the smaller sample size. However, even
this subgroup still can be considered a large sample for this
type of research, which should be considered a strength.
Third, patients were recruited in secondary and tertiary
care, which implies that caution is warranted when
extrapolating these results to other populations of IBS
patients (primary care, nonhealth careseeking patients).
Fourth, because we only included IBS patients in this study,
we cannot exclude the possibility that the associations between anxiety, depression, and somatization levels and
postprandial GI symptom responses are not specic for IBS.
Testing this would require performing a similar study
in healthy volunteers and/or other patient cohorts.
Regarding healthy volunteers, however, we previously
showed that postprandial symptom responses to the standard meal, which also was used in the present study, are
very limited except for fullness,12 leaving very little variability to be putatively explained by psychological symptom
levels. Finally, the time-by-somatization interaction effect
for abdominal pain only showed a nonsignicant trend
(P .054) and therefore should be interpreted with more
caution.
References
1. Longstreth GF, Thompson WG, Chey WD, et al. Functional bowel disorders. Gastroenterology 2006;
130:14801491.
2. Lovell RM, Ford AC. Global prevalence of and risk factors
for irritable bowel syndrome: a meta-analysis. Clin Gastroenterol Hepatol 2012;10:712721.e4.
3. Elsenbruch S. Abdominal pain in irritable bowel syndrome: a review of putative psychological, neural and
neuro-immune mechanisms. Brain Behav Immun 2011;
25:386394.
4. Ohman L, Simren M. Pathogenesis of IBS: role of
inammation, immunity and neuroimmune interactions.
Nat Rev Gastroenterol Hepatol 2010;7:163173.
5. Whitehead WE, Palsson O, Jones KR. Systematic review
of the comorbidity of irritable bowel syndrome with other
disorders: what are the causes and implications?
Gastroenterology 2002;122:11401156.
6. Henningsen P, Zimmermann T, Sattel H. Medically
unexplained physical symptoms, anxiety, and depression: a meta-analytic review. Psychosom Med 2003;
65:528533.
7. Henningsen P, Herzog W. Irritable bowel syndrome and
somatoform disorders. J Psychosom Res 2008;
64:625629.
8. Morcos A, Dinan T, Quigley EMM. Irritable bowel syndrome: role of food in pathogenesis and management.
J Dig Dis 2009;10:237246.
9. Tomba C, Baldassarri A, Coletta M, et al. Is the subjective perception of lactose intolerance inuenced by the
psychological prole? Aliment Pharmacol Ther 2012;
36:660669.
10. Dainese R, Galliani EA, De Lazzari F, et al. Discrepancies
between reported food intolerance and sensitization test
ndings in irritable bowel syndrome patients. Am J
Gastroenterol 1999;94:18921897.
11. Monsbakken KW, Vandvik PO, Farup PG. Perceived food
intolerance in subjects with irritable bowel syndrome etiology, prevalence and consequences. Eur J Clin Nutr
2005;60:667672.
12. Posserud I, Strid H, Strsrud S, et al. Symptom pattern
following a meal challenge test in patients with irritable
bowel syndrome and healthy controls. United European
Gastroenterol J 2013;1:358367.
13. Simren M, Castedal M, Svedlund J, et al. Abnormal
propagation pattern of duodenal pressure waves in the
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
873
CLINICAL AT
April 2016
874
Van Oudenhove et al
CLINICAL AT
31. Tack J, Talley NJ, Camilleri M, et al. Functional gastroduodenal disorders. Gastroenterology 2006;130:14661479.
32. Cremonini F, Talley NJ. Review article: the overlap between functional dyspepsia and irritable bowel syndrome
- a tale of one or two disorders? Aliment Pharmacol Ther
2004;20:4049.
33. Clauwaert N, Jones MP, Holvoet L, et al. Associations
between gastric sensorimotor function, depression, somatization, and symptom-based subgroups in functional
gastroduodenal disorders: are all symptoms equal?
Neurogastroenterol Motil 2012;24: 1088e565.
34. Lipowski ZJ. Somatization: the concept and its clinical
application. Am J Psychiatry 1988;145:13581368.
35. Elsenbruch S, Lovallo WR, Orr WC. Psychological and
physiological responses to postprandial mental stress in
women with the irritable bowel syndrome. Psychosom
Med 2001;63:805813.
36. Le Neve B, Posserud I, Bohn L, et al. A combined nutrient
and lactulose challenge test allows symptom-based
clustering of patients with irritable bowel syndrome. Am
J Gastroenterol 2013;108:786795.
37. Guthrie E, Barlow J, Fernandes L, et al. Changes in
tolerance to rectal distension correlate with changes in
psychological state in patients with severe irritable bowel
syndrome. Psychosom Med 2004;66:578582.
38. Bushnell MC, Ceko M, Low LA. Cognitive and emotional
control of pain and its disruption in chronic pain. Nat Rev
Neurosci 2013;14:502511.
39. Goesling J, Clauw D, Hassett A. Pain and depression: an
integrative review of neurobiological and psychological
factors. Curr Psychiatry Rep 2013;15:18.
40. Asmundson GJG, Katz J. Understanding the cooccurrence of anxiety disorders and chronic pain:
state-of-the-art. Depress Anxiety 2009;26:888901.
41. Mussell M, Kroenke K, Spitzer RL, et al. Gastrointestinal
symptoms in primary care: prevalence and association
with depression and anxiety. J Psychosom Res 2008;
64:605612.
42. Ng C, Malcolm A, Hansen R, et al. Feeding and colonic
distension provoke altered autonomic responses in irritable bowel syndrome. Scand J Gastroenterol 2007;
42:441446.
43. Kemp AH, Quintana DS, Gray MA, et al. Impact of
depression and antidepressant treatment on heart rate
variability: a review and meta-analysis. Biol Psychiatry
2010;67:10671074.
44. Simren M, Stotzer PO, Sjovall H, et al. Abnormal levels of
neuropeptide Y and peptide YY in the colon in irritable
bowel syndrome. Eur J Gastroenterol Hepatol 2003;
15:5562.