Ileum resection is the most predictive factor for osteoporosis in patients with Crohn's disease.

Abstract

Crohn's disease is associated with a host of factors potentially increasing the risk for osteoporosis and fractures. The aim of our study was to identify the most predictive factors for skeletal pathology in this patients.
Using a cross-sectional study design, 146 randomly selected patients with Crohn's disease of variable disease activity who were given standard therapy to control disease activity, including glucocorticoids, and who attended the outpatient clinic of the Gastroenterology Unit on regular follow-up visits were studied. Bone mineral density (BMD) measurements and lateral X-rays of the spine were performed, and biochemical parameters of bone turnover, gonadal hormones and C-reactive protein (CRP) as markers of disease activity were measured in all patients.
There were 61 men and 85 women, with a mean age of 43 years and mean disease duration of 20 years. The majority of patients (86%) had been treated with glucocorticoids at some stage during their illness at a median dose of 7.5 mg/day, 43% were currently using these agents and 66% had undergone an intestinal resection. Twenty-one percent of patients had below-normal 25-hydroxy vitamin D levels. Osteoporosis was documented in 26% of patients, predominantly at the femoral neck, but also at the lumbar spine or at both sites; osteopenia was documented in 45% of patients. Prevalence of vertebral and non-vertebral fractures was, respectively, 6% and 12%. Ileum resection was the most predictive factor for osteoporosis: RR 3.84 (CI 1.24-9.77, p=0.018), followed by age: RR 1.05 (CI 1.02-1.08, p<0.001) and current or past glucocorticoid use: RR 1.94 (CI 0.92-4.10, p=0.08).
Our data suggest that in patients with Crohn's disease, the risk of osteoporosis is best predicted by a history of ileum resection.

Full text

Page 1

Osteoporos Int (2006) 17: 535–542
DOI 10.1007/s00198-005-0016-7
ORIGINAL ARTICLE
Ileum resection is the most predictive factor for osteoporosis
in patients with Crohn’s disease
R. A. van Hogezand.D. Bänffer.A. H. Zwinderman.
E. V. McCloskey.G. Griffioen.N. A. T. Hamdy
Received: 29 April 2005 / Accepted: 23 September 2005 / Published online: 4 January 2006
# International Osteoporosis Foundation and National Osteoporosis Foundation 2006
Abstract Introduction: Crohn’s disease is associated with
a host of factors potentially increasing the risk for osteo-
porosis and fractures. The aim of our study was to identify
the most predictive factors for skeletal pathology in these
patients. Methods: Using a cross-sectional study design,
146 randomly selected patients with Crohn’s disease of
variable disease activity who were given standard therapy
to control disease activity, including glucocorticoids, and
who attended the outpatient clinic of the Gastroenterology
Unit on regular follow-up visits were studied. Bone min-
eral density (BMD) measurements and lateral X-rays of the
spine were performed, and biochemical parameters of bone
turnover, gonadal hormones and C-reactive protein (CRP)
as markers of disease activity were measured in all patients.
Results: There were 61 men and 85 women, with a mean
age of 43 years and mean disease duration of 20 years. The
majority of patients (86%) had been treated with glucocor-
ticoids at some stage during their illness at a median dose
of 7.5 mg/day, 43% were currently using these agents and
66% had undergone an intestinal resection. Twenty-one
percent of patients had below-normal 25-hydroxy vitamin
D levels. Osteoporosis was documented in 26% of patients,
predominantly at the femoral neck, but also at the lumbar
spine or at both sites; osteopenia was documented in 45%
of patients. Prevalence of vertebral and non-vertebral frac-
tures was, respectively, 6% and 12%. Ileum resection was
the most predictive factor for osteoporosis: RR 3.84 (CI
1.24–9.77, p=0.018), followed by age: RR 1.05 (CI 1.02–
1.08, p<0.001) and current or past glucocorticoid use: RR
1.94 (CI 0.92–4.10, p=0.08). Conclusion: Our data sug-
gest that in patients with Crohn’s disease, the risk of oste-
oporosis is best predicted by a history of ileum resection.
Keywords Crohn’s disease.Fractures.Ileum resection.
Osteoporosis.Vitamin D deficiency
Introduction
In patients with Crohn’s disease several factors may be
implicated in the development of osteoporosis and frac-
tures. These include the cytokine-mediated nature of the
inflammatory bowel disease [1], the intestinal malabsorp-
tion resulting from disease activity or from extensive
intestinal resection [2, 3] and the use of glucocorticoids to
control disease activity [4–8]. Inability to achieve peak
bone mass when the disease starts in childhood, malnutri-
tion,immobilisation, lowbody mass index (BMI), smoking
and hypogonadism may also play a contributing role in the
pathogenesis of bone loss [8, 9]. The relationship between
long-term use of glucocorticoids and increased risk for
osteoporosis andfracturesiswellestablished[10–19].Prior
and current exposure to glucocorticoids confers an in-
creased risk of fracture that is of substantial importance
beyond that explained by bone mineral density (BMD)
measurement[10],butquestionsstillremainconcerningthe
relationship between glucocorticoid use and increased
fracture risk in different disease pathologies requiring
treatment with these agents [5, 18]. Moreover, in patients
with Crohn’s disease, opinions are still largely divided on
the magnitude of the problem of osteoporosis, its long-term
clinicalconsequencesintermsofincreasedriskforfractures
R. A. van Hogezand.G. Griffioen
Department of Gastroenterology and Hepatology,
Leiden University Medical Center,
Leiden, The Netherlands
D. Bänffer.N. A. T. Hamdy (*)
Department of Endocrinology and Metabolic Diseases,
Leiden University Medical Center,
Albinusdreef 2,
2333 Leiden, ZA, The Netherlands
e-mail: N.A.T.Hamdy@lumc.nl
Tel.: +31-71-5263300
Fax: +31-71-5248136
A. H. Zwinderman
Department of Medical Statistics,
Leiden University Medical Center,
Leiden, The Netherlands
E. V. McCloskey
Metabolic Bone Centre,
Northern General Hospital,
Sheffield, UK

Page 2

and the means to prevent it [3, 20–31]. Whether patients
with Crohn’s disease sustain bone loss in the long term
remains also controversial [31–35]. The availability of
therapeuticinterventionsthatbeneficiallyinfluenceskeletal
morbidity in post-menopausal women, men or glucocorti-
coid users with osteoporosis [36–41], raises the question
whether all patients with Crohn’s disease should be treated
with bone-protective agents on the assumption that they are
allatriskforosteoporosisorwhethertheseagentsshouldbe
restricted to patients in whom an increased risk can be
clearly identified.
In our study, we set out to identify which patients were at
increased risk for osteoporosis and fractures first by
establishing the prevalence of osteoporosis and fractures
in a relatively large, representative population of Crohn’s
patients with variable disease activity, and second, by
identifying factors associated with an increased prevalence
of osteoporosis and fractures in these patients.
Patients and methods
Study population
In a cross-sectional study design, we studied 146 patients
with Crohn’s disease of variable disease activity attending
regular outpatient clinics of the Department of Gastroen-
terology of the Leiden University Medical Center for
routine follow-up visits. All patients with Crohn’s disease
attending a scheduled out-patient clinic held on a weekly
basis (four to five patients per clinic) were asked to take
part in the study over a 1-calendar-year period. The only
inclusion criterion was the willingness of the patient to
participate in the study. There were no exclusion criteria.
The diagnosis of Crohn’s disease was established on the
basis of clinical, radiological, endoscopic and histological
findings. Disease duration was calculated as time from start
of symptoms, which may have preceded clinical diagnosis
by a number of years. Data on glucocorticoid use (past or
present), disease localisation (ileum, colon, or both) and
details of surgical intervention (ileum and/or colon resec-
tion) were obtained from patients’ records. Patients were
directly questioned about current medication, smoking
habits, alcohol use and history of non-vertebral fractures.
Standard management policies for Crohn’s disease were
adhered to in all cases [42]. The study was approved by the
Ethics Committee of the Leiden University Medical
Center, and informed consent was obtained from all
patients.
Laboratory investigations
Blood and second-void morning urine specimens were
collected from all patients. Patients were not required to be
fasting, but the out-patient clinics were held in the morning.
Serum was measured for calcium (corrected for an albumin
of 42 g/l), phosphate, creatinine and alkaline phosphatase
activity using semi-automated techniques. Serum 25-hy-
droxy vitamin D, 1,25-dihydroxy vitamin D, parathyroid
hormone and estradiol or testosterone concentrations were
measured using standard radio-immuno-assays (RIA).
Vitamin D levels were not corrected for seasonal variation.
Hypogonadism was defined in men by a serum testosterone
concentration <10 nmol/l and in women by a serum
estradiol concentration <70 pmol/l in the absence of men-
strual cycles, associated in both men and women with
follicle-stimulating hormone (FSH) and leuteinising hor-
mone (LH) concentrations >30 U/l and >60 U/l, respec-
tively. Diseaseactivitywasindirectlyevaluatedbymeasuring
C-reactive protein (CRP) using a sandwich enzyme immu-
noassay (Kordia, Leiden, The Netherlands; normal range
7–20 mg/l). Serum bone-specific alkaline phosphatase was
measured by an immuno-radiometric assay (Tandem-R
Ostase, Hybritech, San Diego, CA, USA) and used as a
marker of bone formation. The detection limit of the assay
was 2 μg/l, and the intra-assay coefficient of variation was
7–8% (all samples were measured in the same assay).
Normal values are 3.7–20.9 μg/l (men), 2.9–14.5 μg/l (pre-
menopausal women), and 3.8–22.6 μg/l (post-menopausal
women). Urine specimens were measured for the cross-
linked telopeptides of type I collagen N-telopeptide (NTx)
(Ostex International, Inc., Seattle, WA, USA) as marker of
bone resorption. Results are expressed as urinary N-
telopeptide/creatinine ratio. The intra-assay variation coef-
ficient was 6–7% (all samples weremeasured in oneassay),
and normal values ranged from 3–63 (men) and 5–65
(women) nmol/mmol creatinine.
Bone mineral density measurements
BMD was measured at the lumbar spine and femoral neck
using dual X-ray absorptiometry (DXA) (Hologic QDR-
1000, Waltham, MA, USA). The coefficient of variation of
BMD measurements was 1%, and the machine was cal-
ibrated at regular intervals. The Hologic used in this study
was equipped with reference values based on the National
Health and Nutrition Examination Surveys from 1988 to
1991 (NHANES), which are compatible with those of
Dutch control populations [43–45]. World Health Organi-
zation (WHO) criteria were used to define osteopenia (T
score between −1 and −2.5) and osteoporosis (T score <
−2.5) [46].
Radiological investigations
Lateral X-rays of the thoracic and lumbar spine were
performed at baseline and blindly evaluated in the Meta-
bolic Bone Centre of the University of Sheffield, UK, using
the highly specific semi-quantitative McCloskey method
for assessment of vertebral deformities and fractures [47].
Using this method, the anterior (A), central (C) and pos-
terior (P) vertebral heights are evaluated by means of a
semi-automated technique comprising a back-lit digitising
tablet and a computer database. The definition of deformity
requires the fulfilment of two criteria for each of the three
536

Page 3

vertebral height ratios (A/P, C/P and P/predicted P), with at
least one ratio reduced to more than 3 standard deviations
(SDs) below their expected mean values. No further grad-
ing of deformed vertebrae was undertaken.
Statistical analysis
Results areexpressedasmean±SD. TheSPSSforWindows
software package (Chicago, IL, USA) was used for
statistical analysis. Student’s t test, chi-square test or
Mann-Whitney test were used as appropriate. Non-nor-
mally distributed parameters were log-transformed before
analysis. A step-wise logistic regression analysis was used
to evaluate relative predictive value of various factors for
the presence or absence of osteoporosis and fractures. A
p value of <0.05 (two tailed) was considered statistically
significant.
Results
Demographic and clinical characteristics
Patients studied were 61 men and 85 women, with a mean
age of 43.5 (range 16–78) years. Mean disease duration
was 20 years, ranging from 9 months to 56 years. The vast
majority of patients had been followed up in our Gastro-
enterology Unit for the duration of their illness. The disease
was localised in the ileum in 37 patients, in the colon in 26
and in both in 83; 66% (n=96) had an intestinal resection,
most frequently of the ileum, either in isolation (n=48) or
combined with colon resection (n=26). Mean time since
intestinal resection was 17.06±11.94 (SD) years. There was
no significant difference in time elapsed between ileum or
colon resection and time of evaluation. The majority of
patients were treated with glucocorticoids at some stage
during the course of their illness (n=126, 86%). At the time
of study, 63 patients (43%) were current glucocorticoids
users receiving doses of prednisone varying from 2.5 to
15 mg/day (median 7.5 mg/day), despite which nine pa-
tients still had CRP concentrations >20 mg/l denoting
persisting disease activity. Thirty-nine patients were using
calcium and/or vitamin D supplements. None had been
previously treated with bisphosphonates or other anti-
resorptive agents. None used excessive alcohol; 48% were
smokers (n=68) (Table 1).
Biochemical measurements
Mean serum creatinine, calcium, phosphate, alkaline phos-
phatase activity, 25-hydroxy vitamin D, 1,25-dihydroxy
vitamin D and parathyroid hormone (PTH) were within the
normal laboratory reference range (Table 2). Serum 25-
hydroxy vitamin D concentration was below normal
(<30 nmol/l) in 21% of patients, the majority of whom
had disease localisation at the ileum or had undergone an
ileum versus a colon resection (38% vs. 0%, p<0.02).
Corrected serum calcium concentrations were below
normal in 45 of the 146 patients (31%), most of whom
also had low 25-hydroxy vitamin D levels. Serum PTH
Table 1 Demographic and clinical characteristics of 146 patients with Crohn’s disease with or without evidence of osteoporosis as defined
by a T score=<−2.5 at any measured site
No osteoporosis (n=108)Osteoporosis (n=38)p value*Odds ratio (95% CI)
Age: mean years (SD)
Female gender: n (%)
Male gender: n (%)
Menopausal state: n (%)
Pre-menopausal
Post-menopausal
Current smoking: n (%)
Disease duration: median years (range)
Disease localisation: n (%)
Ileum
Colon
Ileum + colon
Intestinal resection: n (%)
None
Ileum
Colon
Ileum + colon
Ileum ± colon
Current prednisone therapy: n (%)
Ever prednisone therapy: n (%)
41 (13)
65 (60)
43 (40)
51 (15)
20 (53)
18 (47)
<0.001
0.42
1.05 (1.02–1.08)
0.74 (0.35–1.55)
1.00 (–)
51 (47)
14 (13)
52 (50)
17 (1–52)
15 (40)
5 (13)
16 (44)
25 (1–56)
0.39
0.79
0.60
0.006
0.30
0.70 (0.32–1.56)
0.85 (0.27–2.72)
0.82 (0.38–1.75)
1.05 (1.02–1.08)
24 (22)
21 (19)
63 (58)
13 (34)
5 (13)
20 (53)
2.28 (0.70–7.45)
1.00 (–)
1.33 (0.45–4.00)
0.005
44 (40)
34 (32)
16 (12)
14 (16)
48 (44)
42 (39)
93 (86)
6 (16)
17 (45)
3 (8)
12 (32)
29 (76)
21 (55)
33 (87)
1.00 (–)
3.67 (1.31–10.30)
1.38 (0.31–6.16)
6.29 (1.99–19.85)
4.43 (1.68–11.68)
1.94 (0.92–4.10)
1.07 (0.36–3.16)
0.080
0.91
CI confidence interval; SD standard deviation; n sample size
*p value calculated by Student’s t test, chi–square test, or Mann-Whitney test, where appropriate
537

Page 4

concentrations were increased at baseline in 18 patients, the
majority of whom also had decreased 25-hydroxy vitamin
D levels. Fifteen women aged 44–72 years were hypogo-
nadal (estradiol <70 pmol/l+FSH >30 U/l), with only one
receiving hormone replacement therapy. Mean serum bone
alkaline phosphatase and osteocalcin concentrations and
urine NTx excretion corrected for creatinine excretion were
within the normal laboratory reference range although
values were above normal in, respectively, 22 (15%), four
(3%) and 13 (9%) patients. Compared with past or never
glucocorticoid users, current glucocorticoids users had
significantly lower serum alkaline phosphatase activity (80
vs.100 U/l, p<0.001), bone alkaline phosphatase concentra-
tions(11.2vs.14.9μg/l,p<0.02)andosteocalcinlevels(1.86
vs. 2.70 μg/l, p<0.001). There was no significant difference
in any other biochemical parameter measured between
current and non-users although users had a trend for lower
bone resorption markers: urine NTx 24.5 versus 30.5 mmol
BCE/μmol creatinine, p=0.23). Although median value
for CRP was within the normal laboratory reference range,
nine patients had CRP concentrations above 20 mg/l
(Table 2). Mean 25-hydroxyvitamin D concentration was
significantly lower in patients who had an ileum resection
compared with patients having a colon resection or an
intact gut (46±26 vs. 62±24 and 56±28; p=0.003).
Prevalence of osteopenia and osteoporosis
Mean Z scores were −0.58±1.13 (range −3.59 to +2.51) at
the femoral neck and −0.44 ±1.43 (range −3.83 to +4.56) at
the lumbar spine. Mean T-scores were −1.43±1.16 (range
−4.29 to +2.13) at the femoral neck and −0.95±1.41 (range
−3.94 to +3.61) at the lumbar spine. The distribution of
individual absolute BMD values is shown in Fig. 1.
Osteopenia was documented in 33% of patients at the
lumbar spine and in 45% of patients at the femoral neck.
Osteoporosis was observed in 22 patients (15%) at the
lumbar spine, in 28 (19%) at the femoral neck and in 12
(8%) at both measured sites. In total, 38 patients (26%) had
osteoporosis at one or both sites. In patients who were ever
treated with glucocorticoids (n=126), 50 (40%) and 26
(21%), respectively, demonstrated osteopenia and osteo-
porosis at the femoral neck compared with 38(30%) and18
(14%) at the lumbar spine. Of the 20 patients having never
received glucocorticoids, only four (5%) had osteoporosis
at the femoral neck and/or at the lumbar spine. There was
no significant difference in the prevalence of osteoporosis
between patients having never received glucocorticoids
and current or past users of these agents. There was also no
significant difference in the prevalence of osteopenia or
osteoporosis between patients with disease localisation at
either ileum or colon. There was, however, a significant
increase in the prevalence of osteopenia and osteoporosis at
the femoral neck but not at the lumbar spine in patients who
had undergone an ileum resection (n=81) with or without a
concomitant colon resection compared with patients with
an intact small bowel (Fig. 2).
Prevalence of fractures
Lateral X-rays of the spine were evaluable in 133 of the
146 patients. Nine vertebral fractures (6%) were docu-
mented in eight patients, with one patient having sustained
two fractures. Three of the eight patients had osteoporosis,
three had osteopenia and two had a normal BMD. There
was no relationship between prevalent fractures and age,
gender, post-menopausal status, age at diagnosis, disease
activity, duration or localisation (ileum vs. colon) or prior
intestinal resection. During the whole course of their illness
(also before the diagnosis of Crohn’s disease was estab-
lished), 17 patients (12%) sustained a non-vertebral frac-
ture. Some trauma was involved in most cases although
Table 2 Biochemical parameters of bone and mineral metabolism expressed as mean±SD or median (min-max) in case of non-normal
distribution in 146 patients with Crohn’s disease with or without evidence of osteoporosis as defined by a T-score=<−2.5 at any measured
site
All patients (n=146) No Osteoporosis (n=108) Osteoporosis (n=38)Normal range
S. creatinine
S. CRP
S. calcium (corr.)
S. phosphate
S. 25 (OH)D3
S. 1,25(OH)2D3
S. PTH
S. alkaline phosphatase
S. BALP (ostase)
S. osteocalcin
Urine NTx/creat
83±14
5.5 (0.15–60)
2.32±0.15
1.07±0.20
53 (12–135)
67 (15–228)
4.5 (1–52)
92 (42–187)
13.8 (4.4–31.2)
2.33 (0–9)
28 (2–234)
85±14
4.9±8.3
2.33±0.16
1.05±0.20
50±25
68±27
4.4±3.4
94±29
14.2±6.5
2.36±1.49
27±30
79±13*
6.9±12.2
2.29±0.12
1.13±0.20
61±30
62±21
4.7±8.1
85±28
12.6±5.5
2.27±1.25
30±19
70–120 μmol/l
<20 μg/l
2.25–2.55 mmol/l
0.81–1.45 mmol/l
30–120 nmol/l
40–140 pmol/l
1–6.5 pmol/l
40–120 U/l
6.5–20 μg/l
1–5.5 μg/l
3–51 nmol BCE/μmol
S. serum, CRP C-reactive protein, 25(OH)D 25-hydroxy vitamin D, 1,25 (OH)21,25-dihydroxy vitamin D, PTH parathyroid hormone,
BALP bone alkaline phosphatase, NTx type I collagen N-telopeptide, Creat creatinine
*p value=0.02
538

Page 5

usually inappropriate. Seven of these patients had osteopo-
rosis, eight had osteopenia at either or both sites measured,
and five had low serum concentrations of 25-hydroxy
vitamin D.
Predictive factors for the presence of osteoporosis
and fractures
Patients with osteoporosis were significantly older (51±15
vs. 41±13 years, p<0.001) and had longer disease duration
(median 25 vs. 17 years, p= 0.006) than those without
osteoporosis. Of the 38 patients with osteoporosis, 13 were
men (34%), ten were post-menopausal women (26%) and
15 (40%) pre-menopausal women. Menopausal status did
not appear to represent a significant risk factor for osteopo-
rosis in our population, but the number of post-menopausal
women was small. The prevalence of osteoporosis was
significantly higher in patients with an ileum resection
compared with those with a colon resection or an intact gut
(p=0.018). There was no correlation between time since
resection and BMD. At the time of the study, 55% of
patients with osteoporosis and 39% of patients without
osteoporosis were current glucocorticoid users. There was
no significant relationship between past use of glucocorti-
coids and prevalence of osteoporosis. There was no sig-
nificant correlation between CRP, the serologic marker of
disease activity used in this study, and the presence or
absence of osteoporosis. Of the nine patients with an
elevated CRP concentration suggesting disease activity,
only one had osteoporosis and no associated fractures. A
further three patients had fractures—one a vertebral
Fig. 2 Significantly higher prevalence of osteoporosis (T-score=<
−2.5) at the femoral neck (FN) but not at the lumbar spine (LS) in
patients with Crohn’s disease and an ileum resection compared with
patients with an intact small bowel
Fig. 1 Bone mineral density (BMD) measurements at the lumbar spine (upper panel) and femoral neck (lower panel) in 61 men (left
panels) and 85 women (right panels) with Crohn’s disease
539