ORIGINAL CONTRIBUTIONS 1

see related editorial on page x

A Cross-Sectional Study of the Prevalence of

COLON/SMALL BOWEL
Gastrointestinal Symptoms and Pathology in Patients
With Common Variable Immunodeficiency
Silje F. Jørgensen, MD1, 2, 3, Henrik M. Reims, MD, PhD4, Didrik Frydenlund, MD, PhD4, Kristian Holm, MSc1, 5, Vemund Paulsen, MD6,
Annika E. Michelsen, PhD1, 9, Kristin K. Jørgensen, MD, PhD5, 7, Liv T. Osnes, MD, PhD8, Jorunn Bratlie, BLS1, Tor J. Eide, MD, PhD4, 9,
Christen P. Dahl, MD, PhD1, Ellen Holter, MD10, Rune R. Tronstad, MD11, 12, Kurt Hanevik, MD, PhD11, Hans-Richard Brattbakk, PhD11, 13,
Fatemeh Kaveh, PhD14, Torunn Fiskerstrand, PhD11, 13, Anne-Marte B. Kran, MD, PhD9, 15, Thor Ueland, PhD1, 16,
Tom H. Karlsen, MD, PhD1, 3, 5, 9, Pål Aukrust, MD, PhD1, 2, 3, 9, Knut E.A. Lundin, MD, PhD6, 9, 17 and Børre Fevang, MD, PhD1, 2, 9

OBJECTIVES: The objective of this study was to study the prevalence of gastrointestinal (GI) symptoms and
histopathology in patients with common variable immunodeficiency (CVID) as well as linking the
findings to GI infections and markers of systemic immune activation.

METHODS: In this cross-sectional study, we addressed GI symptoms in 103 patients and GI histopathological
findings in 53 patients who underwent upper and lower endoscopic examination. The most frequent
histopathological findings were linked to GI symptoms, B-cell phenotype, and markers of systemic
immune activation (soluble (s)CD14, sCD25, and sCD163). Microarray analysis compared “celiac-
like disease” in CVID to celiac disease. Screening for selected bacterial and viral infections in fecal
samples and gut mucosal biopsies was performed.

RESULTS: The main findings of this study were as follows: most common GI symptoms were bloating (34%),
pain (30%), and diarrhea (26%). The most frequent histopathological findings were increased
intraepithelial lymphocytes in the descending part of the duodenum, i.e., “celiac-like disease”
(46% of patients), decreased numbers of plasma cells in GI tract mucosa (62%), and lymphoid
hyperplasia (38%), none of which were associated with GI symptoms. Reduced plasma cells in GI
mucosa were associated with B-cell phenotypic characteristics of CVID, and increased serum levels
of sCD14 (P=0.025), sCD25 (P=0.01), and sCD163 (P=0.04). Microarray analyses distinguished
between CVID patients with “celiac-like disease” and celiac disease. Positive tests for bacterial and
viral infections were scarce both in fecal samples and gut mucosal biopsies, including PCR test for
norovirus in biopsy specimens (0 positive tests).

CONCLUSIONS: In conclusion, GI pathology is common in CVID, but does not necessarily cause symptoms. However,
reduced plasma cells in GI mucosa were linked to systemic immune activation, “celiac-like disease”
in CVID and true celiac disease appear to be different disease entities, as assessed by gene
expression, and infections (including norovirus) are rarely a cause of the CVID enteropathy.
SUPPLEMENTARY MATERIAL is linked to the online version of the paper at http://www.nature.com/ajg

Am J Gastroenterol advance online publication, 16 August 2016; doi:10.1038/ajg.2016.329

1
Research Institute of Internal Medicine, Oslo University Hospital, Rikshospitalet, Oslo, Norway; 2Section of Clinical Immunology and Infectious Diseases, Oslo
University Hospital Rikshospitalet, Oslo, Norway; 3K. G. Jebsen Inflammation Research Centre, Institute of Clinical Medicine, University of Oslo, Oslo, Norway;
4
Department of Pathology, Oslo University Hospital, Rikshospitalet, Oslo, Norway; 5Norwegian PSC Research Center, Oslo University Hospital Rikshospitalet,
Oslo, Norway; 6Department of Transplantation Medicine, Section of Gastroenterology, Oslo University Hospital Rikshospitalet, Oslo, Norway; 7Department of
Gastroenterology, Akershus University Hospital, Lørenskog, Norway; 8Department of Immunology, Oslo University Hospital, Rikshospitalet, Oslo, Norway; 9Institute
of Clinical Medicine, University of Oslo, Oslo, Norway; 10Department of Microbiology, Oslo University Hospital, Rikshospitalet, Oslo, Norway; 11Department of
Clinical Science, University of Bergen, Bergen, Norway; 12Department of Pediatrics, Haukeland University Hospital, Bergen, Norway; 13Center for Medical
Genetics and Molecular Medicine, Haukeland University Hospital, Bergen, Norway; 14Medical Genetics Department, Oslo University Hospital, Ullevål, Oslo,
Norway; 15Department of Microbiology, Oslo University Hospital, Ullevål, Oslo, Norway; 16K.G. Jebsen Thrombosis Research and Expertise Center, University
of Tromsø, Tromsø, Norway; 17Centre for Immune Regulation, University of Oslo, Oslo, Norway. Correspondence: Silje F. Jørgensen, MD, Research Institute of
Internal Medicine, Oslo University Hospital, Rikshospitalet, postbox 4950, Nydalen, Oslo 0424, Norway. E-mail: s.f.jorgensen@medisin.uio.no
Received 29 February 2016; accepted 6 July 2016

© 2016 by the American College of Gastroenterology The American Journal of GASTROENTEROLOGY

 2 Jørgensen et al.
INTRODUCTION
Common variable immunodeficiency (CVID) is the most com-
mon symptomatic primary immunodeficiency in adults, with a
prevalence of 1 in 25,000 in Caucasians. CVID patients are char-
COLON/SMALL BOWEL
acterized by a B-cell defect with reduced antibody production
as the immunological hallmark. However, other immunological
abnormalities including T-cell deficiency, monocyte/macrophage
hyperactivity, and signs of low-grade systemic inflammation are
also described (1).
Most of the patients suffer from recurrent bacterial infections in
the respiratory tract. A large proportion (68–71%) (2,3) of CVID
patients, however, also have various forms of autoimmune and
inflammatory disorders; such as immune-mediated cytopenia,
sarcoid-like granulomas, immune-mediated hepatitis, and various
forms of non-malignant and malignant lymphoid disorders (2,4).
These non-infectious complications in CVID also include various
forms of inflammatory and autoimmune manifestations of the
intestine (5).
Intermittent or persistent diarrhea has been described as the
most common gastrointestinal (GI) symptom in CVID patients,
reported in 20–60% of cases, depending on the study (6–10).
Several types of GI infections seem to occur more commonly
in CVID patients, including Giardia, Campylobacter, certain
Salmonella species (11), and Cytomegalovirus (CMV) (5), and in a
recent relatively small study (n=8) chronic norovirus infection was
suggested as cause of severe CVID enteropathy (12). In addition,
there is a wide spectrum of non-infectious GI pathology among
CVID patients, including nodular lymphoid hyperplasia, atrophic
gastritis, inflammatory bowel disease, lymphocytic colitis, and
celiac-like findings in the duodenum (5,6,8). All of these condi-
tions seem to share an abnormal immune response in the GI tract,
supported by the fact that many of the patients with GI abnormali-
ties on endoscopic examination have autoimmune manifestations
in other organs (10). However, the true frequency of these mani-
festations in CVID, as well as their relation to GI symptoms, and
other clinical and immunological characteristics of CVID patients
are largely undefined.
Here, we performed a cross-sectional study in unselected
CVID patients addressing the prevalence of GI symptoms in
103 patients and GI histopathological findings in 53 patients
who agreed to undergo endoscopic examination. By including
patients irrespective of GI symptoms, we sought to reveal the
true prevalence of GI symptoms and pathology in CVID patients
as well as their link to clinical and immunological features in
these patients.
METHODS
Study design
All the adult CVID patients (n=112) registered at the Section of
Clinical Immunology and Infectious Diseases at Oslo University
Hospital, Rikshospitalet, Oslo, Norway received a letter invit-
ing them to participate in the study. The Section functions as a
national center for diagnosis and treatment of primary immuno-
deficiency diseases for adults in Norway. The CVID diagnosis was
The American Journal of GASTROENTEROLOGY
defined as decreased serum levels of immunoglobulin (Ig)G, IgA,
and/or IgM by at least two standard deviations below the mean
for age, and exclusion of other causes of hypogammaglobuline-
mia (13,14). The actual concentration used as cutoff levels were
for IgG≤5 g/l (15) and for IgA<0.7 g/l. For males and females <50
years the lower IgM cutoff value was 0.4 g/l and 0.6 g/l, respec-
tively. The study included an upper and lower endoscopy (GIF-
HQ190, Olympus, Hamburg, Germany), and the completion of
a GI-symptoms questionnaire. Patients undergoing endoscopy
also had blood tests and fecal samples taken prior to endoscopy.
In addition, duodenal biopsies from 17 healthy adults without
abdominal symptoms and 10 patients with untreated celiac dis-
ease (16) were obtained for gene-expression profiling analyses.
These biopsies were compared with duodenal biopsies from 12
CVID patients with increased intraepithelial lymphocytes (IEL),
and 8 CVID patients with normal number of IEL in the duode-
num (pars descendens).
The study was approved by the Regional Committee for Medical
and Health Research Ethics and conforms to the principles out-
lined in the “Declaration of Helsinki”. Written informed consent
was obtained from all participants.
GI symptoms questionnaire
We used the well-validated “GI symptoms” questionnaire Gas-
trointestinal Symptom Rating Scale IBS (GSRS-IBS); a 13-item
questionnaire to assess the presence of different GI symptoms
using a 7-point Likert scale (17). We considered a few ques-
tionnaires suitable for the task of mapping the GI symptoms in
CVID. The question regarding diarrhea was in our opinion, at
pre-analysis, the most important symptom to record and the
GSRS-IBS addressed this symptom (as well as pain, bloating,
satiety, and constipation) in a way that was transportable to other
disease phenotypes (like CVID), however, the questionnaire is
developed for irritable bowel syndrome (IBS) and is not CVID
specific.
The patients indicated from 1–7; where 1 was “no discomfort at
all” and 7 was ”very severe discomfort”. The 13 questions could be
further sub-scaled into symptom clusters of: pain, bloating, con-
stipation, diarrhea, and satiety. The patients with more than 3 on
the Likert scale (more than “mild discomfort”) were considered
positive for the symptom.
Histopathological assessment
Intestinal biopsies were collected according to protocol (Supple-
mentary Figure 1 online) at the Section for Gastroenterological
Endoscopy at Oslo University Hospital, Rikshospitalet, Oslo,
Norway. Prior to endoscopy, fecal samples were analyzed for
pathogenic microbes (Supplementary Methods). Hematoxylin
and eosin-stained slides were reviewed independently by two
pathologists (H.M.R and D.F.) using a structured reporting sheet.
The cutoff level for increased IEL in duodenal biopsies has var-
ied over the years from 40 IEL/100 enterocytes (EC) to as low as
20 IEL/100 EC for different classification schemes (18–21). In
this study, an increased number of IELs in the duodenal mucosa
was defined by a number of IEL ≥25 per 100 EC based on the
www.nature.com/ajg
 Gastrointestinal Disease in CVID 3

CVID patients
most recent European (16) and the American College of
Gastroenterology guidelines (22). See Supplementary Informa- 112
tion (online) for methods and further classification into diagnos-

COLON/SMALL BOWEL
tic categories. Endoscopy
Yes No

Gene expression by microarray analysis

Total RNA was extracted from the tissues and each sample was 53 59
biotin-labeled and amplified. The raw data were first imported
into GenomeStudio Data Analysis Software (Illumina, San Diego, Of which....
CA) and then to J-Express 2012 (Bergen, Norway) (23). Corres-
pondence analysis was applied to study global trends in the data
Gastroscopy Colonoscopy GI symptoms GI symptoms
(Supplementary Methods) (24). questionnaire questionnaire

50 52
Statistical analysis 103 total
Both 52 51
Univariate analyses were performed using parametric (t-tests)
or non-parametric methods (Mann–Whitney’s U- and Kruskal– 49
Wallis tests) for continuous variables, and χ2-tests and Fisher’s
Figure 1. Overview of number of common variable immunodeficiency
exact test for categorical variables. Correlation analysis was (CVID) patients in the study.
performed using parametric (Pearson’s) or non-parametric
(Spearman’s) tests. P values are two-sided and considered signifi-
cant when <0.05.
Detection of norovirus, rotavirus, adenovirus, astrovirus, CMV,
Table 1. Patient characteristics
Epstein–Barr virus (EBV), and Helicobacter pylori in tissue speci-
mens by PCR and/or immunohistochemical assessment, blood Total No endoscopy Endoscopy P value
(n=104)a (n=51)b (n=53)
sampling protocol, measurements of soluble (s)CD14, sCD25,
and sCD163 in serum and The human leukocyte antigen (HLA) Male, n (%) 53 (51) 28 (53) 25 (47) 0.55c
analyses are described in Supplementary Methods. Age, mean 47.5 (19–83) 45.3 (20–83) 49.6 (19–81) 0.16d
years (range)
Splenomegaly, 47 (45) 23 (45) 24 (45) 1.00c
RESULTS n (%)
Study population Organ-specific 20 (19) 9 (17) 11 (21) 0.80c
In all, 112 adult CVID patients were invited to participate in autoimmunity,
n (%)
the study, of whom 53 underwent endoscopy and 103 patients
completed the GI symptom score questionnaire (the 13-item Autoimmune 22 (21) 12 (24) 10 (19) 0.64c
cytopenia,
GSRS-IBS; Figure 1). Forty-nine patients underwent both gas- n (%)
troscopy and colonoscopy, three patients colonoscopy only and
HLA DQ2.5e 34 (38) 19 (37) 15 (28) 0.39c
one patient gastroscopy only (based on the patient preferences).
e
In total, 50 gastroscopies and 52 colonoscopies were performed. HLA DQ8 26 (29) 12 (24) 14 (26) 0.81c
Of the 53 patients undergoing endoscopy, 52 were on Ig replace- Enteropathy, lymphoid hyperplasia, and therefore “Infection only” are not
included because these are influenced by the study. NRH liver (3), LIP (1), and
ment therapy (10 on intravenous IG, 37 on subcutaneous IG, 5 lymphoma (1) are not included due to low numbers.
on a combination thereof, and one refused to take IgG replace- a
The total include: 103 patients filling in the questionnaire plus 1 patient that
ment therapy). None were newly diagnosed CVID patients performed endoscopy but did not fill out the questionnaire.
b
Excluding 8 missing.
and except for IgG substitution, none of the patients were on c
Fisher’s exact test.
any immunomodulatory medication with relevance to the GI d
Independent t-test.
tract. Patient characteristics regarding age, sex, selected non- e
n=90 for the HLA data, 14 (13%) missing.
infectious complications, and the celiac disease-specific HLA
phenotype are presented in Table 1 for the whole cohort n=104
and separately for the group undergoing endoscopy, and the
group not undergoing endoscopy. The recruitment of patients Prevalence of GI symptoms in CVID patients
undergoing endoscopy was based on volunteerism and not The most prevalent symptom in our CVID cohort was bloating
stratified. There was therefore a risk that the cohort undergoing (34%) followed by pain (30%), and diarrhea (26%; Table 2). In
endoscopy would be biased in terms of the patient characteris- total, 48 patients (47%) suffered from one or more pre-defined
tics, but importantly, when comparing the patient characteristics GI symptoms (bloating, pain, diarrhea, satiety, and constipation)
between the patient who had undergone endoscopy and those with no significant difference between the “endoscopy group” and
patients who had not, there were no significant differences. the “no endoscopy group” (Table 2).

© 2016 by the American College of Gastroenterology The American Journal of GASTROENTEROLOGY

 4 Jørgensen et al.

Table 2. Symptoms from GI tract in the study group

Symptomsa CVID %, n (total)b CVID mean (95% CI Mean±s.d. Mean±s.d. no P value endoscopy/
for mean)c endoscopyc endoscopyc no endoscopy
COLON/SMALL BOWEL

Bloating 34%, 34 (99) 2.49 (2.25, 2.74) 2.56±0.18 2.40±0.17 0.52d

Diarrhea 26%, 26 (102) 2.19 (1.94, 2.44) 2.03±0.14 2.35±0.20 0.51e
Constipation 13%, 13 (102) 1.60 (1.37, 1.83) 1.73±0.20 1.51±0.14 0.63e
Pain 30%, 31 (103) 2.12 (1.85, 2.38) 1.99±0.29 1.95±0.16 0.68e
Satiety 14%, 14 (100) 1.55 (1.26, 1.84) 1.59±0.12 1.74±0.15 0.31e
CI, confidence interval; CVID, common variable immunodeficiency; GI, gastrointestinal.
a
GSRS-IBS.
b
≥3 on GSRS-IBS scale.
c
1=no discomfort at all, 2=minor discomfort, 3=mild discomfort, 4=moderate discomfort, 5=moderately severe discomfort, 6=severe discomfort, and 7=very severe
discomfort.
d
t-test.
e
Mann–Whitney’s test for CVID endoscopy patients compared with CVID no endoscopy.

The most frequent histopathological findings in CVID Table 3. Percentage of patients with histopathological finding
The most frequent findings were increased IEL (n=32), reduced
number of plasma cells (n=33), and lymphoid hyperplasia (n=20; Histopathological finding Number of patients, n (%)

Table 3 and Figure 2). In addition, gastric metaplasia of the duo- Increased IEL total 32 (60%)
denal bulb (n=13) and other findings in the stomach (n=19; i.e., Descending part of duodenum 23 (46%)
intestinal metaplasia (n=6), atrophic gastritis (n=9), and fibrosis
Reduced number of plasma cells 33 (62%)
(n=13); Table 3) were frequently present. Inflammation was noted
in 24 patients, representing a heterogeneous group with regard Lymphoid hyperplasia 20 (38%)

to localization and type of inflammation and was therefore not Gastric metaplasia in duodenal bulb 13 (26%)
ranked as a distinct group (Table 3 and Supplementary Table S1). Fibrosis in the gastric mucosa 13 (26%)
Supplementary Table S2 (online) shows the positive findings in Intestinal metaplasia in gastric mucosa 6 (12%)
relation to anatomical localization. The prevalence of increased
Subacute inflammation 2 (4%)
IELs was highest in biopsies taken from the descending part of
the duodenum. Chronic/chronic active inflammation, total 24 (45%)
For the three most common histopathological findings in CVID; Stomach 20 (40%)
increased IEL in the duodenum (pars descendens), reduced num- Colon 11 (22%)
ber of plasma cells. and lymphoid hyperplasia, we looked for asso- Atrophic gastritis 9 (18%)
ciations with certain characteristics of CVID: GI symptoms, B-cell
GVHD-like 1 (2%)
parameters in blood according to the Euro classification (25), and
markers of systemic immune activation, i.e., serum levels of sCD14 Eosinophilic inflammation 4 (8%)
and sCD163 as markers of monocyte activation and sCD25 as a Lymphocytic enteritis/colitis 4 (8%)
marker of T-cell activation. The results are represented below and Collagenous enteritis/colitis 3 (6%)
in Table 4.
Granulomatous inflammation 3 (6%)
GVHD, graft-vs.-host disease; IEL, intraepithelial lymphocytes.
Increased IEL in duodenal biopsies without villous blunting
Text in bold relates to percentage of patients at each anatomical site.
An increased number of IELs was found in one or more biopsy
sites in 32 patients (60%). The presence of increased IEL was most
frequent in duodenal biopsies pars descendens (46%) and bul-
bus (28%), and to a lesser extent in the stomach, ileum, colon, constipation), B-cell phenotype, or markers of systemic inflam-
and rectum (Supplementary Table S2). Of the 23 patients who mation (Table 4).
had IEL≥25 (three patients had IEL above 40), only one had vil-
lous atrophy (Marsh grade 3a. IEL 56). This patient had been on Gene profiling shows little overlap between celiac disease
a gluten-free diet for several years at the time of the biopsy, was and CVID patients with increased IEL
HLADQ2.5 positive, and negative to celiac-specific autoantibod- On the basis of the findings of increased IEL in duodenal biopsies
ies (as one would expect in a patient with antibody deficiency). in CVID, an overlap with celiac disease has been suggested
The finding of increased IEL in pars descendens was not asso- (26). We therefore examined the gene-expression profile in pars
ciated with GI symptoms (diarrhea, bloating, satiety, pain, or descendens of duodenum in CVID patients with increased IEL

The American Journal of GASTROENTEROLOGY www.nature.com/ajg

 Gastrointestinal Disease in CVID 5

COLON/SMALL BOWEL
Figure 2. The most common histopathological findings in common variable immunodeficiency (CVID) (a–c) compared with normal histopathological findings
in the same tissue (d–f). (a) Duodenal mucosal villi with an increased number of intraepithelial lymphocytes (arrows). (b) Colonic mucosa with a reduced
number of plasma cells in the lamina propria. (c) Duodenal mucosa with a lymphoid aggregate (arrows). (d) Normal duodenal mucosal villi illustrating nor-
mal number of intraepithelial lymphocytes. (e) Normal colonic mucosa with a plasma cells (arrows) in the lamina propria. (f) Normal duodenal mucosa.

2.0
Table 4. Comparing the most common histopathological finding CVID
CVID
in CVID with GI symptoms, B-cell phenotype, and systemic CVID
CVID CVID

inflammation H
1.0 CVID

H
CVID
CVID
CVID
CVID
CVID
Marsh 3a
CVID subtypes Histopathological findings H
CVID CVID
CVID
CVID CVID
H H H
H
Reduced plasma Lymphoid Increased IEL in H H
CVID
CVID
CVIDH
cells hyperplasia pars descendens 0.0 H
H
CD
CDCD
GI symptoms No No No H Marsh 1
H

B-cell phenotype ↓switched memory, ↑B-cells No –1.0

H
CD
CVID

(Euroclass) ↑CD21low, H
CD
CD

Systemic inflam- ↑sCD14, ↑sCD163, No No CD CDCD

mation (sCD14, ↑sCD25
sCD163, sCD25) –2.0

H
CVID, common variable immunodeficiency; GI, gastrointestinal. Celiac disease CVID-IEL
CVID-normal Healthy CD

(n=12, IEL mean 33 (range 25–56) IEL/100 EC), CVID with nor-
mal levels of IEL (n=8), untreated celiac disease (n=10, Marsh
grade 3a or above), and healthy controls (n=17) by cDNA micro-
array assessment. With regard to the gene-expression profile,
CVID patients with increased IEL were more similar to CVID
patients with normal IEL than to celiac disease patients (Figure 3).
Even the CVID patient with Marsh grade 3a did not cluster
with the other celiac disease patients (Figure 3). This difference
between CVID and celiac disease was also confirmed by dis-
criminating pathways analyses (Supplementary Figure 2a–d and
Supplementary Table S3a–h). In fact, 544 genes (q=0) are signifi-
cantly different expressed between CVID with increased IEL and
celiac disease (Supplementary Table 4; see URL).
We also investigated the occurrence of two HLA genotypes com-
mon in celiac disease: HLA DQ2.5 and HLA DQ8 in our cohort
–3.0
–3.0 –2.0 –1.0 0.0 1.0 2.0
Figure 3. Principal component plot showing the difference in gene expres-
sion between common variable immunodeficiency (CVID) patients with
increased intraepithelial lymphocytes (IEL), CVID patients with normal
IEL, patients with celiac disease, and healthy controls. In the CA plot, the
microarray data for samples are projected onto a two-dimensional plane
defined by the first and second principal components. The first principal
component (along the x-axis) explains most of the total χ2 (6.9%), the
second principal component explains second most of the total χ2 (6.1%).
Samples that are close together in the plot are more similar than samples
further apart. The lines drawn from the origin and through the group mean
gives the direction of the sample group.
(Table 1). Forty-two patients who underwent gastroscopy had
HLA DQ2.5 and HLA DQ8 data available. Of these, 13 patients
carried the HLA DQ2.5 genotype (26%) of which 7 had increased

© 2016 by the American College of Gastroenterology The American Journal of GASTROENTEROLOGY

 6 Jørgensen et al.
IEL in duodenal pars descendens and 6 did not, showing no asso-
ciation with increased IEL in duodenal pars descendens (Fisher’s
exact test, P=0.75; Supplementary Table S5). Also, HLA DQ8 was
present in 14 CVID patients (26%) and was not associated with
COLON/SMALL BOWEL
increased IEL in duodenal pars descendens (Fisher’s exact test,
P=0.76).
Reduced numbers of plasma cells are associated with an
inflammatory phenotype
A paucity or lack of plasma cells was found in 33 of the patients
(62%), most frequently in the colon (56%) followed by rectum
(50%), ileum (42%), duodenum (42%), and—to a much lesser
extent—in the stomach (16%). Reduced number of plasma cells
at one or more site were associated with the B-cell characteristics
in peripheral blood, i.e., decreased switched memory (P=0.0004)
and CD21low B-cells (P=0.045), but not with percentage of CD19+
B-cells (P=0.47), or total number of circulating B-cells (P=0.24;
Table 4). Patients with reduced plasma cells in GI mucosa, had
increased serum levels of sCD14 (3,620 ng/ml (3,185–4,282) vs.
2,939 ng/ml (2,482–3,945), P=0.025) and sCD163 (1,055 ng/
ml (825–1,314) vs. 581 ng/ml (420–1,382), P=0.04) as mark-
ers of monocyte activation and sCD25 (1.49 ng/ml (0.87–2.05)
vs. 0.61 ng/ml (0.52–1.29), P=0.01) as marker of T-cell activa-
tion, with and without reduced plasma cells, respectively, given
in median (25–75 percentiles; Table 4). We found no association
with reduced plasma cells (at one or more site in the GI tract)
and GI symptoms (i.e., pain, bloating, constipation, diarrhea, and
satiety).
Lymphoid hyperplasia is associated with increased number of
circulating B-cells
Lymphoid hyperplasia was present in one or more biopsy sites in
20 patients (38%; Table 3), being most frequent in the duodenum
(26%, Supplementary Table S2). Lymphoid hyperplasia was asso-
ciated with increased numbers of circulating B-cells (238 cells/μ l
(105–408) vs. 104 cells/μ l (31–221), P=0.019, with and without
lymphoid hyperplasia, respectively, given in median (25–75 per-
centiles)), but not with GI symptoms or signs of systemic inflam-
mation (i.e., sCD14, sCD25, and sCD163; Table 4).
Chronic inflammation in GI biopsies from CVID patients is
associated with systemic inflammation and immune activation
In total, 24 (45%) CVID patients had chronic/chronic active
inflammation on histological examination (Table 3 and Supple-
mentary Table S1), of which 11 had chronic active or chronic
inflammation of the colon. In our study, 17 CVID patients had
diarrhea syndrome (>3 on the GSRS-IBS scale), and in only 6 of
these, diarrhea was associated with inflammatory manifestation
in the colon (“GVHD (graft-vs.-host disease)-like” (n=1), eosino-
philic inflammation (n=1), collagenous colitis (n=2), lympho-
cytic colitis (n=1), and unspecified chronic active inflammation
(n=1); Table 3 and Supplementary Table S1). Of the five other
patients with chronic inflammation of the colon, but not diarrhea
syndrome, two had “IBD-like” findings, two had granulomas,
and one patient had eosinophilic inflammation. The causes of the
The American Journal of GASTROENTEROLOGY
remaining 11 patients with diarrhea, but without chronic inflam-
mation of the colon, were difficult to pin-point with two excep-
tions; one patient with infection with Cryptosporidium (found
on fecal screening) and one patient with collagenous enteritis
(antrum; Supplementary Table S6).
Compared with patients without any histological inflamma-
tion, those categorized in the heterogeneous inflammation group,
irrespective of site, were associated with increased level of sCD25
(1.93 ng/ml (0.96–2.24) vs. 0.78 ng/ml (0.62–1.34); P=0.03),
and inflammation of the antrum (stomach) was associated with
increased sCD14 (3,816 ng/ml (3,332–5,354) vs. 3,281 ng/ml
(2,785–3,964); P=0.02), and sCD25 (2.02 ng/ml (0.91–2.24) vs.
0.90 ng/ml (0.63–1.37); P=0.03), given as median and (25–75) per-
centiles, with and without inflammation, respectively.
Malignancy and acute inflammation
None of the biopsies showed malignancy, acute infection, or
inflammation. However, intestinal metaplasia in the gastric
mucosa was present in 6 (12%) patients.
Gastrointestinal infections in CVID
Screening for GI infections prior to endoscopy revealed the fol-
lowing outcomes: one patient with Clostridium difficile (positive
PCR for toxin A and B and culture), one had Cryptosporidium
(rapid immunoassay and positive microscopy including Ziehl–
Neelsen), and one had Campylobacter jejuni infection (stool cul-
ture). The patients with infections generally had few pathological
findings on endoscopy (data not shown) and only the patient with
Cryptosporidium had significant symptoms (i.e., diarrhea).
In all, 3 out of 50 patients were found positive for H. pylori rapid
test directly on the biopsy at the time of endoscopy. In addition,
immunohistochemical staining for H. pylori was performed in
biopsies from gastric corpus and antrum on all patients (n=50),
detecting only one patient with positive staining (also positive
rapid test).
In the light of the recent publication (12) of the potentially prom-
inent role of chronic norovirus infection leading to GI enteropathy
in CVID patients, we investigated the presence of norovirus RNA
and other viruses with a potential to cause GI infection (i.e., rotavi-
rus, adenovirus, and astrovirus) in biopsies from 52 CVID patients
from three different sites (stomach (n=40), duodenum (n=49),
and colon (n=46) by PCR (Supplementary Methods). None of
the biopsies were positive for norovirus, adenovirus, or astrovirus.
One patient had positive rotavirus PCR both in the duodenal and
colon biopsy. He suffered from diarrhea, pain, bloating, and satiety
syndrome, and histopathology showed increased IEL in the duode-
nal biopsy and GVHD-like finding in the colon (Supplementary
Table S6, patient 13). The healthy controls from the microarray
analysis (duodenal biopsies, n=17) were used as controls and were
negative for all the above enteropathogenic viruses.
We also investigated the presence of EBV and CMV from the
same three sites in the GI tract. Twelve (10 patients) of a total of
127 biopsies (41 in the duodenum, 40 in the stomach, and 46 in the
colon) had positive PCR test for EBV; six positives from stomach,
five positives from duodenum (including two positives in both
www.nature.com/ajg

stomach and duodenum), and one positive in the colon biopsy. Six
(4 patients) out of 127 biopsies were positive for CMV PCR, two
at each site of stomach, duodenum, and colon. However, none of
these positive specimens were positive on immunohistochemistry
(CMV) or in situ hybridization (EBV). There was no association
between histopathological findings or symptoms with positive
CMV and/or EBV by PCR.
DISCUSSION
In the present study, 103 CVID patients were screened for GI
symptoms and 53 unselected CVID patients underwent endo-
scopic examination. Our major findings were as follows: (i)
bloating (34%), pain (30%), and diarrhea (26%) were the most
common gastrointestinal symptoms. (ii) The most frequent histo-
pathological findings were increased IEL in the descending parts
of the duodenum, decreased plasma cells in GI tract mucosa, and
lymphoid hyperplasia. (iii) Decreased number of plasma cells was
associated with immunological features of the CVID patients.
(iv) The heterogeneous inflammation group, irrespective of site,
was associated with signs of systemic immune activation. (v) The
CVID patients with increased IEL in pars descendens of duode-
num showed little or no overlap with celiac disease, as assessed
by gene-expression microarray analyses and HLA typing. (vi) GI
tract infection, such as H. pylori, was seen only in one patient and
none of the patients had evidence of chronic norovirus infection.
The strengths of this study lie in the cross-sectional design
within a given time frame, standardized biopsy procedure, the
structured reporting questionnaire, the dual endoscopy (gastros-
copy and colonoscopy) investigation in the majority of patients,
and the inclusion of patients with and without GI tract symptoms.
Moreover, the CVID patients who underwent endoscopy were rep-
resentative for our total CVID cohort at our hospital with regard
to clinical characteristics and GI symptoms. Previous studies have
been inconsistent in the reporting of GI disease in CVID, mainly
due to inadequate investigation of the GI tract, biased study popu-
lation in terms of symptoms, retrospective study design, and low
number of participants in each study (5,27–29). Based on our
study design, we believe that our data give a good indicator of the
prevalence of GI disease and symptoms in the CVID population
representing a disease model of the interaction between immuno-
deficiency, autoimmunity, and the GI tract.
Herein, we showed increased IEL in a large proportion (i.e.,
46%) of biopsies from duodenum. With one exception, however,
this was not accompanied by villous atrophy. This may seem in
contrast to previous publications where villous atrophy was pre-
sent in duodenal biopsies in 31–42% of CVID patients (26,27,30).
However, these studies give no indication of the true prevalence of
villous atrophy in CVID due to their retrospective study design,
and that the individuals studied were patients with severe GI
symptoms. On the other hand, Maarschalk et al. (29) found in
their cross-sectional study only 1 patient out of 30 (3%) with vil-
lous atrophy, but unfortunately, this study lacked information on
IEL. The apparent co-existence of CVID and celiac disease has pre-
viously been described (31), but it is not clear if this represents a
© 2016 by the American College of Gastroenterology
Gastrointestinal Disease in CVID 7
common pathogenic mechanism of celiac disease and “celiac-like”
disease findings in CVID patients. The unreliable auto-antibody
tests owing to B-cell deficiency, inconsistent response to gluten-
COLON/SMALL BOWEL
free diet (31,32), and inconsistency in the HLA profile in CVID
patients with celiac-like disease contribute to making the diagno-
sis of concomitant celiac disease challenging in CVID. This is also
illustrated for the patient with increased IEL and villous atrophy
in our study were the histopathology and HLA phenotype sup-
ports the diagnosis of celiac disease, whereas the lack of response
to gluten-free diet and the unreliable interpretation of the autoan-
tibodies weakens the likelihood that this is a true celiac disease.
It is in fact difficult in all CVID patients to truly confirm or rule
out if they have celiac disease or if the histopathological finding is
part of the immunodeficiency syndrome representing an immune-
mediated complication as seen at other sites in the GI tract and in
other organs in CVID. The large difference in gene-expression data
shown in this study supports that celiac disease and CVID with
increased IEL are different disease entities. Other possible triggers
for the increased IEL than gluten may be infectious, but as shown
in this study, the rate of infection is low and therefore the trigger is
unlikely to be infectious. In our opinion the celiac-like findings in
CVID most probably reflect an immune-dysregulated phenotype
in these patients with immunodeficiency co-existing with persis-
tent immune activation and inflammation. On the basis of this
hypothesis, the treatment approach could be immunomodulatory
rather than a gluten-free diet. Notwithstanding, systematic studies
that explore the role of gluten-free diets in CVID with increased
IEL, in particular randomized intervention trial, could be
warranted.
A reduced number of plasma cells in mucosal tissue were found
to be associated with signs of increased systemic inflammation
(i.e., increased serum levels of sCD14 and sCD163) and immune
activation (sCD25), but not with GI symptoms. Moreover, this
histological feature was significantly associated with certain B-cell
phenotypes in peripheral blood, i.e., decreased class-switched
memory and increased CD21low B-cells, the latter is known to be
associated with non-infectious complications in CVID (33). The
reason for these associations is at present not clear, but underscore
that reduced plasma cells in GI tract mucosa is not an “obligate”
finding in all patients with B-cell deficiency. The finding of reduced
plasma cells in mucosal tissue could therefore potentially be used
as an indicator of a “systemic” phenotype, (a state of low-grade
systemic inflammation and immune activation) which in turn is
related to inflammatory and autoimmune complication (2).
In this study we show that a large proportion of the CVID
patients had chronic inflammation of the GI tract, associated
with signs of systemic immune activation in both monocytes and
T cells. Perreau et al. (34) have reported that increased systemic
endotoxin levels are associated with systemic T-cell pathology in
CVID, and our findings herein may suggest a link between GI
inflammation and T-cell as well as monocyte pathology. GI inflam-
mation may contribute to microbial translocation with subsequent
systemic immune activation as seen in other diseases like HIV
infection (35). We have recently shown that lipopolysaccharide (a
marker of microbial translocation) is increased in plasma in CVID
The American Journal of GASTROENTEROLOGY
 8 Jørgensen et al.
patients—associated with altered gut microbiota in these patients
(36). It is tempting to hypothesize that altered gut microbiota and
GI tract inflammation, causing a breach in the mucosal surface
of the GI tract results in a “leaky gut” where lipopolysaccharide
COLON/SMALL BOWEL
leaks into the blood stream, could contribute to the inflammatory
phenotype seen in a large proportion of CVID patients.
Of particular interest with regard to the unexplained diarrhea,
we were not able to replicate the recently shown association of
CVID enteropathy and chronic norovirus infection in our cohort.
The difference in results may be due to patient selection and that
Woodward et al. (12) did a more detailed investigation. However,
although Woodward et al. (12) examined 8 patients, we analyzed
samples from 52 patients, and the fact that we are not able to
identify any patients with norovirus in our cohort suggests that
norovirus is unlikely to be a common cause of the GI enteropathy
of unknown etiology in CVID.
H. pylori infection has previously been associated with gastritis
in CVID (37), but none of the patients with atrophic gastritis or intes-
tinal metaplasia in our study had positive H. pylori tests, nor did the
single patient with positive H. pylori test have atrophic gastritis. This
suggests another or an additional trigger of atrophic gastritis in CVID.
In summary, GI pathology is common in CVID, but all patho-
logy does not necessarily correspond with symptoms. However,
certain findings (i.e., reduced plasma cells in mucosa and GI tract
inflammation) were associated with increased levels of sCD25
and sCD14, suggesting a link between GI pathology, and systemic
inflammation and immune activation in CVID. Although similari-
ties between patients with celiac disease and CVID patients with
increased IEL on duodenal biopsies, our findings suggest that these
two groups represent different disease entities. Moreover, very few
patients had positive tests for microbes associated with GI patho-
logy including norovirus and H. pylori. Future studies should aim
to identify the triggers for CVID enteropathy as well as to charac-
terize the molecular mechanisms for increased IEL and lymphoid
hyperplasia in GI mucosa of CVID patients.
ACKNOWLEDGMENTS
Mona Bjørnstad, Liv Wenche Thorbjørnsen, Hege Dahlen Sollid,
and other members of the Norwegian PSC Research Center are
acknowledged for support on sample collection and logistics. We
acknowledge the technical support and service from the Genomics
Core Facility at the Department of Clinical Science, the University of
Bergen. Anusha Aravinthan and the Department of Microbiology at
Oslo University Hospital are acknowledged for technical assistance.
CONFLICT OF INTEREST
Guarantor of the article: Silje F. Jørgensen, MD.
Specific author contributions: Recruited subjects, obtained human
samples, and clinical data collection: Silje F. Jørgensen, Børre
Fevang, Kurt Hanevik, Jorunn Bratlie, Christen P. Dahl, Kristin K.
Jørgensen, and Pål Aukrust; reviewed the pathology slides: Henrik
M. Reims and Didrik Frydenlund; performed the endoscopies: Knut
E.A. Lundin and Vemund Paulsen; performed the flow cytometry
analysis: Liv T. Osnes; performed PCR analysis: Anne-Marte B. Kran
and Ellen Holter; contributed to the microarray analysis:
The American Journal of GASTROENTEROLOGY
Rune R. Tronstad, Hans-Richard Brattbakk, and Torunn Fiskerstrand;
performed the pathway analysis: Fatemeh Kaveh; analyzed the data
and performed statistical analysis: Silje F. Jørgensen, Kristian
Holm, Hans-Richard Brattbakk, Annika E. Michelsen, and Thor
Ueland; coordinated and supervised the project: Silje F. Jørgensen,
Børre Fevang, Tor J. Eide, Pål Aukrust, and Tom H. Karlsen; and
drafted the manuscript Silje F. Jørgensen, Henrik M. Reims, Didrik
Frydenlund, Børre Fevang, and Pål Aukrust. All authors revised the
manuscript for critical content and approved the final version.
Financial support: S.F.J. was funded by a grant from the South-East-
ern Norway Regional Health Authority (project number 2012063).
Potential competing interests: None.
Data and material availability: Microarray data are deposited in
the GEO archives with data accession number: GSE72625 URL:
http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE72625.
Study Highlights
WHAT IS CURRENT KNOWLEDGE
✓ Gastrointestinal symptom and histopathology in common
variable immunodeficiency (CVID) patients varies greatly
between studies, potentially reflecting that a low number of
patients have been included.
✓ There is a wide spectrum of non-infectious gastrointestinal
(GI) pathology among CVID patients, but neither their fre-
quency in an unbiased CVID cohort in terms of symptoms,
nor their relationship to other clinical and immunological
characteristics is established.
✓ The apparent co-existence of CVID and celiac disease
has previously been described, but it is not clear if this
represents a common pathogenic mechanism of true celiac
disease and “celiac-like” disease findings in CVID patients.
✓ A recent study suggests chronic norovirus as the cause of
severe CVID enteropathy.
WHAT IS NEW HERE
✓ In the largest study of GI symptoms and GI histopathology,
performing upper and lower endoscopies in CVID patients
with and without GI symptoms, we found that the most
frequent histopathological findings in GI mucosa of CVID
patients were increased intraepithelial lymphocytes (IEL)
in descending duodenum—celiac-like disease—(46%
of patients), decreased plasma cells in GI tract mucosa
(62%), and lymphoid hyperplasia (38%).
✓ Reduced plasma cells in GI mucosa were associated with
B-cell phenotypes and signs of low-grade systemic immune
activation and inflammation reflecting phenotypic charac-
teristics of CVID.
✓ CVID patients with “celiac-like disease” and true celiac disease
are different disease entities as assessed by gene-expression
analysis and The human leukocyte antigen (HLA) genotyping.
✓ Norovirus infection was not found as a cause of CVID
enteropathy in our cohort.
REFERENCES
1. Yong PF, Thaventhiran JE, Grimbacher B. "A rose is a rose is a rose," but
CVID is not CVID common variable immune deficiency (CVID), what do
we know in 2011? Adv Immunol 2011;111:47–107.
www.nature.com/ajg

2. Resnick ES, Moshier EL, Godbold JH et al. Morbidity and mortality in com-
mon variable immune deficiency over 4 decades. Blood 2012;119:1650–7.
3. Chapel H, Lucas M, Lee M et al. Common variable immunodeficiency
disorders: division into distinct clinical phenotypes. Blood 2008;112:
277–86.
4. Chapel H, Cunningham-Rundles C. Update in understanding common
variable immunodeficiency disorders (CVIDs) and the management of
patients with these conditions. Br J Haematol 2009;145:709–27.
5. Daniels JA, Lederman HM, Maitra A et al. Gastrointestinal tract pathology
in patients with common variable immunodeficiency (CVID): a clinico-
pathologic study and review. Am J Surg Pathol 2007;31:1800–12.
6. Cunningham-Rundles C, Bodian C. Common variable immunodeficiency:
clinical and immunological features of 248 patients. Clin Immunol 1999;92:
34–48.
7. Hermans PE, Diaz-Buxo JA, Stobo JD. Idiopathic late-onset immuno-
globulin deficiency: clinical observations in 50 patients. Am J Med 1976;61:
221–37.
8. Hermaszewski R, Webster A. Primary hypogammaglobulinaemia: a survey
of clinical manifestations and complications. QJM 1993;86:31–42.
9. Quinti I, Soresina A, Spadaro G et al. Long-term follow-up and outcome of
a large cohort of patients with common variable immunodeficiency. J Clin
Immunol 2007;27:308–16.
10. Washington K, Stenzel TT, Buckley RH et al. Gastrointestinal pathology
in patients with common variable immunodeficiency and X-linked
agammaglobulinemia. Am J Surg Pathol 1996;20:1240–52.
11. Oksenhendler E, Gérard L, Fieschi C et al. Infections in 252 patients with
common variable immunodeficiency. Clin Infect Dis 2008;46:1547–54.
12. Woodward JM, Gkrania-Klotsas E, Cordero-Ng AY et al. The role of
chronic norovirus infection in the enteropathy associated with common
variable immunodeficiency. Am J Gastroenterol 2015;110:320–7.
13. International Union of Immunological Societies. Primary immunodefi-
ciency diseases. Report of an IUIS Scientific Committee. Clin Exp Immunol
1999;118:1–28.
14. Al-Herz W, Bousfiha A, Casanova JL et al. Primary immunodeficiency
diseases: an update on the classification from the international union of
immunological societies expert committee for primary immunodeficiency.
Front Immunol 2014;5:162.
15. Ameratunga R, Woon ST, Gillis D et al. New diagnostic criteria for com-
mon variable immune deficiency (CVID), which may assist with decisions
to treat with intravenous or subcutaneous immunoglobulin. Clin Exp
Immunol 2013;174:203–11.
16. Ludvigsson JF, Bai JC, Biagi F et al. Diagnosis and management of adult
coeliac disease: guidelines from the British Society of Gastroenterology. Gut
2014;63:1210–28.
17. Wiklund IK, Fullerton S, Hawkey CJ et al. An irritable bowel syndrome-
specific symptom questionnaire: development and validation. Scand J
Gastroenterol 2003;38:947–54.
18. Oberhuber G, Granditsch G, Vogelsang H. The histopathology of coeliac
disease: time for a standardized report scheme for pathologists. Eur J
Gastroenterol Hepatol 1999;11:1185–94.
© 2016 by the American College of Gastroenterology
Gastrointestinal Disease in CVID 9
19. Mahadeva S, Wyatt J, Howdle P. Is a raised intraepithelial lymphocyte count
with normal duodenal villous architecture clinically relevant? J Clin Pathol
2002;55:424–8.
20. Hayat M, Cairns A, Dixon M et al. Quantitation of intraepithelial lympho-
cytes in human duodenum: what is normal? J Clin Pathol 2002;55:393–4.
COLON/SMALL BOWEL
21. Veress B, Franzén L, Bodin L et al. Duodenal intraepithelial lymphocyte-
count revisited. Scand J Gastroenterol 2004;39:138–44.
22. Rubio-Tapia A, Hill ID, Kelly CP et al. ACG clinical guidelines: diagnosis
and management of celiac disease. Am J Gastroenterol 2013;108:656–76.
23. Dysvik B, Jonassen I. J-Express: exploring gene expression data using Java.
Bioinformatics 2001;17:369–70.
24. Fellenberg K, Hauser NC, Brors B et al. Correspondence analysis applied to
microarray data. Proc Natl Acad Sci USA 2001;98:10781–6.
25. Wehr C, Kivioja T, Schmitt C et al. The EUROclass trial: defining subgroups
in common variable immunodeficiency. Blood 2008;111:77–85.
26. Luzi G, Zullo A, Iebba F et al. Duodenal pathology and clinical-immuno-
logical implications in common variable immunodeficiency patients. Am J
Gastroenterol 2003;98:118–21.
27. Malamut G, Verkarre V, Suarez F et al. The enteropathy associated with
common variable immunodeficiency: the delineated frontiers with celiac
disease. Am J Gastroenterol 2010;105:2262–75.
28. Khodadad A, Aghamohammadi A, Parvaneh N et al. Gastrointestinal
manifestations in patients with common variable immunodeficiency.
Dig Dis Sci 2007;52:2977–83.
29. Maarschalk-Ellerbroek LJ, Oldenburg B, Mombers IM et al. Outcome of
screening endoscopy in common variable immunodeficiency disorder and
X-linked agammaglobulinemia. Endoscopy 2013;45:320–3.
30. Teahon K, Webster AD, Price AB et al. Studies on the enteropathy associ-
ated with primary hypogammaglobulinaemia. Gut 1994;35:1244–9.
31. Venhoff N, Emmerich F, Neagu M et al. The role of HLA DQ2 and DQ8
in dissecting celiac-like disease in common variable immunodeficiency.
J Clin Immunol 2013;33:909–16.
32. Biagi F, Bianchi PI, Zilli A et al. The significance of duodenal mucosal
atrophy in patients with common variable immunodeficiency: a clinical
and histopathologic study. Am J Clin Pathol 2012;138:185–9.
33. Mouillot G, Carmagnat M, Gerard L et al. B-cell and T-cell phenotypes in
CVID patients correlate with the clinical phenotype of the disease. J Clin
Immunol 2010;30:746–55.
34. Perreau M, Vigano S, Bellanger F et al. Exhaustion of bacteria-specific CD4
T cells and microbial translocation in common variable immunodeficiency
disorders. J Exp Med 2014;211:2033–45.
35. Vazquez-Castellanos JF, Serrano-Villar S, Latorre A et al. Altered metabo-
lism of gut microbiota contributes to chronic immune activation in HIV-
infected individuals. Mucosal Immunol 2015;8:760–72.
36. Jorgensen SF, Troseid M, Kummen M et al. Altered gut microbiota profile
in common variable immunodeficiency associates with levels of lipopoly-
saccharide and markers of systemic immune activation. Mucosal Immunol;
advance online publication, 16 March 2016; doi:10.1038/mi.2016.18.
37. Zullo A, Romiti A, Rinaldi V et al. Gastric pathology in patients with com-
mon variable immunodeficiency. Gut 1999;45:77–81.
The American Journal of GASTROENTEROLOGY