By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
Dr T. M. Trebble, Institute of Human Nutrition, School of Medicine, University of Southampton, Tremona Road, Southampton SO16 6YD, UK. E-mail: firstname.lastname@example.org.
Background : Crohn's disease is associated with reduced bone density. The power of simple markers of systemic inflammation to identify higher rates of bone loss, in Crohn's disease, is uncertain. This relationship and the role of circulating (peripheral blood) mononuclear cells were investigated in a case–control study.
Methods : Urinary deoxypyridinoline/creatinine and serum osteocalcin concentrations were compared in male and premenopausal females with ‘active’ Crohn's disease (C-reactive protein ≥ 10 and/or erythrocyte sedimentation rate ≥ 20) (n = 22) and controls with ‘quiescent’ Crohn's disease (C-reactive protein < 10 and erythrocyte sedimentation rate < 20) (n = 21). No patients were receiving corticosteroid therapy. Production of tumour necrosis factor-α, interferon-γ and prostaglandin E2 by peripheral blood mononuclear cells were measured.
Results : Active Crohn's disease was associated with a higher deoxypyridinoline/creatinine (P = 0.02) and deoxypyridinoline/creatinine:osteocalcin ratio (P =0.01) compared with quiescent Crohn's disease, but similar osteocalcin (P = 0.24). These were not explained by vitamin D status, dietary intake or nutritional status. However, production of interferon-γ by concanavalin A-stimulated peripheral blood mononuclear cells was lower in active Crohn's disease (P = 0.02) and correlated negatively with the deoxypyridinoline/creatinine:osteocalcin ratio (r = −0.40, P = 0.004).
Conclusion : In Crohn's disease, raised C-reactive protein and erythrocyte sedimentation rate may indicate higher rates of bone loss and, if persistent, the need to assess bone mass even where disease symptoms are mild. This may be partly explained by altered production of interferon-γ by peripheral blood mononuclear cells.
Crohn's disease (CD) is a relapsing inflammatory condition within the gastrointestinal tract, associated with an increased prevalence of osteoporosis. Reduced bone density in CD results from pathological rates of bone remodelling and turnover1–3 because of multifactorial but incompletely characterized mechanisms that may include corticosteroid use and malnutrition.2, 4 Recognition and treatment of osteoporosis, in CD, remains suboptimal even amongst patients who have already sustained a fracture.5 There is a need, in CD, for an improved understanding of the pathophysiological processes underlying bone loss and for the identification of surrogate markers of altered bone turnover. This may, potentially, lead to the stratification of patients with respect to their risk of developing osteoporosis and the targeting of resources for investigation, monitoring and treatment.
Biochemical markers of bone metabolism are dynamic tests of bone turnover (resorption and formation) that identify ‘fast’ and ‘slow’ bone losers and are predictive for osteoporotic fracture.6–8 Studies in the published literature fail to demonstrate a relationship in CD between markers of clinical disease severity and bone turnover9, 10 or bone density.4, 11 This may reflect the reliance of such studies on clinical disease scores, i.e. Crohn's Disease Activity Index (CDAI)12 that are subjective assessments of disease symptoms with a poorly documented relationship to inflammation in the gut or extra-intestinal tissues. By comparison, simple laboratory markers of inflammation that include C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR), are objective markers of the acute phase or ‘systemic’ inflammatory response.13–15 CRP and ESR may, therefore, identify patients with ‘active’ and ‘quiescent’ systemic inflammation, in CD, even in the absence of symptomatic gastrointestinal disease. Potential mechanisms through which a systemic inflammatory response may modify bone turnover in CD include altered cytokine production by circulating (peripheral blood) mononuclear cells (PBMC)16, 17 and malnutrition.4, 18
In this study, the relationship between simple laboratory markers, CRP and ESR, and biochemical markers of bone turnover were investigated, and compared with production of tumour necrosis factor (TNF)-α, interferon (IFN)-γ and prostaglandin (PG)E2 by PBMC and nutritional status.
This was an unmatched case–control study of bone turnover in male and premenopausal females with CD. Patients with ‘active’ and unmatched controls with ‘quiescent’ CD were recruited using a comprehensive database of out-patients with inflammatory bowel disease under follow-up at a large university hospital. The diagnosis of CD was based on endoscopic, histological or radiological findings. ‘Active’ disease was indicated by a CRP equal to or greater than 10 mg/L19 and/or an ESR equal to or greater than 20 mm/h, ‘quiescent’ CD was indicated by a CRP less than 10 mg/L and an ESR of less than 20 mm/h. A secondary analysis was performed by stratifying the CD cohort by the CDAI,12 a composite clinical disease score consisting of predetermined measures of symptoms (stool frequency, abdominal pain and general well-being), extra-intestinal manifestations of CD, use of antidiarrhoeals, presence of an abdominal mass, weight differential from predicted weight and haematocrit. A score of greater than 150 indicated active disease.12
All recruited patients underwent a questionnaire-based assessment of previous gastroenterological and medical history, and medical notes were reviewed for further information. An abbreviated questionnaire of habitual physical activity assessed estimated daily activity at work, sport and as daily walking, with each category scoring from 0 (low) to 3 (high). A validated semiquantitative, self-administered food frequency questionnaire20 was completed at the time of sample collection to estimate habitual dietary nutrient intake.
Subjects were excluded if they required nutritional support in any form, had undergone major small bowel intestinal resection, total or subtotal colectomy; exhibited features of short bowel, inflammatory disease unrelated to CD or fever or sepsis, had received oral or intravenous corticosteroid within the previous 4 weeks, or pharmacological interventions for the prevention or treatment of osteoporosis including hormone replacement therapy or bisphosphonates; had a history of endocrinological, liver or renal disease; or had recently consumed calcium or vitamin D dietary supplements.
The study was approved by the Southampton and South West Hampshire Joint Research Ethics Committee. All subjects gave informed consent.
Assessment of nutritional status
Body composition was determined directly by anthropometry and indirectly by bioelectrical impedance. All measurements were made by a single investigator using standard methods and with the subject wearing only light clothes and without shoes. Body height was measured using a free standing CMS Stadiometer (Chasmore Ltd, London, UK); body weight was measured using electronic scales (Seca, Hamburg, Germany); body mass index (BMI) was calculated as body weight (kg) divided by height (m) squared. Skinfold thickness (SFT) was measured on the non-dominant side of the body in triplicate at four predetermined sites using a single set of callipers (Holtain Ltd, Crymych, UK). The sum of the mean values for each site represented the value for SFT in each subject. Mid-arm circumference (MAC) was measured on the non-dominant side using a flexible steel tape; mid-arm muscle circumference (MAMC) was calculated as MAC [mean triceps SFT × Pi (3.14159)]. Bioelectric impedance was measured electronically (Biostat 1500, Bodystat, Isle of Man, UK) using standard methods. All measurements were made in the morning following an overnight fast. Age- and sex-matched healthy control subjects were recruited from amongst academic, clinical and general staff at the same university hospital for comparison of dietary intake, BMI, body weight and height.
Laboratory analysis and assessment of cytokine production and bone turnover
For assessment of full blood count, ESR, CRP concentration, and plasma vitamins D and synthetic function of PBMC, peripheral venous blood samples were taken by standard methods following an overnight fast. The methods of preparation and culture of PBMC and cytokine analysis have been previously described.21 In brief, TNF-α, IFN-γ and PGE2 synthesis by purified PBMC (1 × 106 cells/mL) were measured following incubation with and without the monocyte/macrophage-stimulant lipopolysaccharide (LPS) (TNF-α and PGE2) or the T-cell stimulant concanavalin A (Con A) (IFN-γ). TNF-α and IFN-γ concentrations were determined using EASIA ELISA kits (Biosource Europe S.A., Nivelles, Belgium). PGE2 concentrations were determined using NEOGEN ELISA kits (Neogen Corporation, Lexington, KY, USA). Coefficient of variation was <10% for both cytokine and prostaglandin assays, and the limits of detection were 3 ng/L for TNF-α and 30 IU/L for IFN-γ.
Deoxypyridinoline (DPD), is a peptide derived from non-reducible pyridinium crosslinks within mature collagen and excreted in the urine following bone degradation, and is, therefore, a marker of bone resorption.8 DPD was measured in urine samples collected as a second early morning void, following an overnight fast. Urinary DPD was measured by a Heterogeneous Competitive Magnetic Separation Assay (MSA) a competitive immunoassay, and corrected for urinary creatinine (Cre) concentrations (DPD/Cre). The minimum detectable concentration was 0.9 nm and the manufacturer's reference ranges were 2.3–5.4 μm/m Cre in males and 3.0–7.0 μm/m Cre in females. Osteocalcin, is a non-collagenous protein synthesized by osteoblasts and incorporated within the bone matrix, and is a marker of bone formation.8 Osteocalcin, was measured in serum extracted from non-heparinized, clotted blood samples following centrifugation (720 g for 10 min at 4 °C), by a chemiluminescence assay (Nichols Institute Diagnostics Ltd, Heston, UK). The minimal detectable concentration was 0.5 ng/mL and the manufacturer's reference ranges were 1.1–7.2 μg/L in males and 0.5–7.0 μg/L in females. All samples were stored at −70 °C prior to analysis and analysed in batches using commercially available immunoassays.
The primary outcome variables consisted of markers of bone turnover (DPD/Cr and osteocalcin), cytokine and PGE2 production and markers of nutritional status. Statistical significance of the differences in outcome variables between active and quiescent CD groups was assessed by the two-sample t-test for normally distributed variables. Non-normally distributed data was log-transformed prior to analysis. Linear regression was used to test the differences in CD groups for bone markers and nutritional markers adjusted for the effect of gender. Differences in outcome variables between CD patients and matched healthy controls were tested by the paired t-test. Associations between TNF-α, IFN-γ and PGE2 production and markers of bone turnover and nutritional status were assessed by Pearson's correlation coefficient. The κ test was used to assess the relationship of laboratory to clinical methods of assessments of CD activity. Analyses were performed using SPSS for Windows (SPSS Inc., Chicago, IL, USA).
A total of 22 ‘active’ CD patients and 21 unmatched ‘quiescent’ CD controls were identified by laboratory markers of inflammation (CRP and ESR). Disease groups were well-matched except for CRP, ESR and CDAI (Table 1). Mean values for CDAI in the ‘active’ CD groups were consistent with mild clinically symptomatic disease, and in the ‘quiescent’ group with clinically quiescent disease. There were no significant differences in dietary intake between active and quiescent groups, including intakes of vitamin D, calcium and magnesium (Table 2) or compared with age- and sex-matched healthy controls. There were no significant differences in plasma concentrations of vitamin D (25-hydroxycholecalciferol) between active (21.2 ± 10.1 μg/L) and quiescent (25.2 ± 10.4 μg/L) disease groups (P = 0.22) and values from all CD subjects were within the accepted normal range for the local population (4–40 μg/L). There were no significant differences between active and quiescent groups in mean age of menarche or contraceptive pill use in female patients (all were premenopausal) and there was no history of androgen deficiency in male patients.
Table 1. Baseline characteristics of patients with ‘quiescent’ [C-reactive protein (CRP) < 10 and erythrocyte sedimentation rate (ESR) < 20] and ‘active’ (CRP ≥ 10 or/and ESR ≥ 20) Crohn's disease
Quiescent (n = 21)
Active (n = 22)
IQR, interquartile range.
Male : female
Mean age (years) (s.d.)
Mean disease duration (years) (s.d.)
Small and large bowel
Current extra-intestinal manifestations
Small and large bowel
Current drug history
Combined oral contraception
Markers of disease activity
Mean Crohn's Disease Activity Index (CDAI) (s.d.)
Median CRP (IQR)
3.9 (2.0, 6.8)
15.0 (10.6, 22.5)
Mean ESR (s.d.)
Current smoking (n)
Mean Activity Index score (0–9) (s.d.)
Table 2. Habitual dietary intake in patients with ‘quiescent’ [C-reactive protein (CRP) < 10 and erythrocyte sedimentation rate (ESR) < 20] and ‘active’ (CRP ≥ 10 or/and ESR ≥ 20) Crohn's disease and matched healthy controls
Healthy controls (n = 43)
Quiescent (n = 21) [mean (s.d.)]
Active (n = 22) [mean (s.d.)]
a Paired t-test: comparison of Crohn's disease patients and matched healthy controls.
b Two-sample t-test: comparison of patients with quiescent and active Crohn's disease.
Vitamin D (µg/day)
Bone turnover and inflammatory makers
DPD/Cre was significantly lower in CD males compared with CD premenopausal females (P < 0.01), therefore comparison of values for DPD/Cre in all groups were adjusted for patient gender. DPD/Cre was higher in active compared with quiescent CD (P =0.02) (Figure 1), consistent with higher rates of bone resorption. Values for DPD/Cre were above the reference range in 13 of 21 (62%) quiescent CD patients and 17 of 22 (77%) active CD patients, but this was not significantly different (P = 0.27).
There was some evidence of differences in osteocalcin concentrations between males and female CD patients (P = 0.08). However, there were no effects on differences between active and quiescent CD groups of adjustment for patient gender. There was no difference in serum osteocalcin between active and quiescent CD (P = 0.24) (Figure 2). Values in the CD cohort were above the reference range, adjusted for gender, in five of 21 (23.8%) with quiescent and five of 22 (22.7%) with active disease, with the remaining (majority) patients in each group within the reference ranges.
The ratio of DPD/Cre to osteocalcin was compared in patients with active and quiescent CD (Figure 3). Ratios were significantly higher in active CD (P = 0.01).
Bone turnover and nutritional status
Comparisons of nutritional markers between active and quiescent CD groups were adjusted for patient gender. In the CD cohort, values for BMI suggested normal levels of nutrition (BMI 19–25 kg/m2) in 19 patients (44%), overweight (BMI 25–30 kg/m2) in 19 patients (44%) and obesity (BMI > 30) in five patients (12%), but no patients were malnourished (BMI < 19). There were no differences in BMI between CD patients and healthy controls (P = 0.47) or between active and quiescent CD groups in BMI (P = 0.15), percentage total body fat mass (P = 0.20), percentage total body lean mass (P = 0.20), SFT (P = 0.15) or MAMC (P =0.26).
There were no consistent associations between markers of bone turnover and nutritional status.
Cytokine and prostaglandin production and bone turnover
The IFN-γ by Con A-stimulated PBMC was lower in active compared with quiescent disease (Figure 4). There were no significant differences between active and quiescent CD in production of PGE2 or TNF-α by unstimulated or LPS-stimulated PBMC (Table 3) or production of IFN-γ by unstimulated PBMC (P = 0.67). IFN-γ by Con A-stimulated PBMC moderately and inversely correlated with the ratio of DPD/Cre: osteocalcin (r = −0.40, P = 0.004) but not DPD/Cre (r = −0.26, P = 0.05) or osteocalcin (r = 0.29, P = 0.03).
Table 3. Synthesis of tumour necrosis factor-α (TNF-α) and prostaglandin E2 (PGE2), by unstimulated or lipopolysaccharide (LPS)-stimulated peripheral blood mononucelar cell (PBMC) in patients with ‘quiescent’ [C-reactive protein (CRP) < 10 and erythrocyte sedimentation rate (ESR) < 20] and ‘active’ (CRP ≥ 10 or/and ESR ≥ 20) Crohn's disease
Mean differenceb (95% CI)
Quiescent (n = 21) [mean ± s.d.]a
Active (n = 22) [mean ± s.d.]a
b Geometric mean.
c Adjusted for gender.
2.81 ± 1.17
2.72 ± 1.20
0.09 (−0.64, 0.82)
3.66 ± 0.75
3.28 ± 0.80
0.38 (−0.10, 0.86)
0.86 ± 0.59
1.01 ± 0.97
−0.15 (−0.64, 0.35)
1.01 ± 0.49
0.93 ± 0.69
0.08 (−0.29, 0.45)
Analysis of variables following stratification by Clinical Disease Score (CDAI)
The CD cohort was stratified by clinical disease score (CDAI) into active and quiescent disease groups for a secondary analysis of baseline and outcome variables. There were no significant differences between active and quiescent CD groups in dietary intake or nutritional status and no significant differences in DPD/Cre or serum osteocalcin concentration (P = 0.64 and P = 0.72, respectively) or the ratio of DPD/Cre: osteocalcin (P = 0.31).
There was poor concordance of patients identified as active and quiescent CD activity by CDAI score and by inflammatory markers (κ = 0) (Table 4).
Table 4. Crohn's disease (CD) patients with active and quiescent disease by laboratory markers of inflammation [C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR)] and clinical disease score (Crohn's Disease Activity Index, CDAI) (κ = 0)
Disease activity by laboratory markers of inflammation
Disease activity by clinical disease activity
The cohort in the current study represented CD out-patient populations with quiescent and mildly active symptomatic disease as indicated by mean CDAI values of 133 and 170, respectively. In both of these populations there may generally be a low index of suspicion for altered bone turnover, malnutrition or other extra-intestinal manifestations of CD. To our knowledge, the current study is the first in the literature to suggest that rates of bone loss in this CD population may be significantly higher amongst patients with increased CRP and/or ESR. In the CD population in the current study, patients with a mildly raised CRP or ESR demonstrated higher rates of bone resorption, trends towards low rates of bone formation and an increase in the resorption/formation ratio consistent with relative bone loss. There was no observed relationship between bone turnover and clinical symptom score (CDAI).
The influence of a number of other possible factors on bone turnover were examined but there were no demonstrated differences between active and quiescent CD groups that could simply explain the contrasting rates of bone turnover. These included plasma vitamin D levels, dietary intakes of calcium, magnesium and vitamin D (that were also similar to age- and sex-matched controls), physical activity score and oral contraceptive use. Furthermore, there was no evidence that poor nutritional status accounted for differences between CD groups in bone turnover as measures of nutritional status were similar, were consistent with normal or over nutrition (by BMI) compared with the general population, and were not significantly different from matched healthy controls.
A potential criticism of the current study is that the CD cohort was small (43 patients). This was due to the strict exclusion criteria used in the study and the exclusion of patients with characteristics that could confound the association between the systemic inflammatory response and bone metabolism including postmenopausal status, corticosteroid use or the use of any drug that could affect bone metabolism. Although this reduced subject numbers, and in particular excluded patients with very active disease, it allowed a more accurate assessment of the relationship between laboratory markers of inflammation and bone turnover. A second potential critism is that the relationship between altered production of bone markers and increased fracture risk has not been studied specifically in CD. There is, however, data in the literature to suggest that increased production of bone markers is associated with bone loss in inflammatory bowel disease22 similar to that seen in postmenopausal women.
The active and quiescent CD groups in the current study were identified by values for CRP and ESR. These simple inflammatory markers are objectively determined surrogate indicators of the systemic inflammatory response that reflect altered metabolic processes in the liver.13–15 In CD, CRP and ESR are predictive of disease relapse,23–25 correlate to TNF-α release by PBMC26 and may indicate subclinical intestinal inflammation.25, 27 By comparison, clinical scores, including the CDAI,12 are subjective, clinical assessments of gastrointestinal CD based on self-reporting of symptoms and physical examination. Raised laboratory markers of inflammation may occur in CD in the absence of symptomatic intestinal disease,27 however, the most likely drive for the systemic inflammatory response in CD remains inflamed intestinal tissue even in absence of clinical evidence of active CD.
The mechanism through which the systemic inflammatory response (and increased CRP and ESR) is associated with altered bone turnover remains unconfirmed. A possible explanation involves altered function of circulating mononuclear cells (PBMC). Increased production of CRP may be associated with altered production of cytokines by PBMC.28, 29 In the current study, we demonstrated in CD patients that lower production of IFN-γ by Con A-stimulated PBMC (T cells) was associated with increased production of CRP or/and ESR and a relative increase in the ratio of DPD/Cre (bone resorption) to osteocalcin (bone formation). Bone resorption reflects the function of osteoclasts that are derived from similar progenitor cell lines to monocyte/macrophages30 and, like monocyte/macrophages, respond to release of cytokines.31, 32 There is evidence to suggest that altered production of cytokines by PBMC is noted in osteoporosis associated with other inflammatory diseases33 and postmenopausal status.34 PBMC consist of circulating monocyte/macrophages and T and B lymphocytes with the capacity to synthesize prostaglandins and cytokines35–37 that may independently modulate osteoclast function.32 Activated T cells also directly promote osteoclast differentiation and function through expression of the membrane protein RANKL,38 that interacts with RANK on osteoclast precursor cells.17, 39 However, activated T cells simultaneously release the cytokine IFN-γ as a counter-regulatory mechanism to prevent uncontrolled bone resorption during the inflammatory response.40 Production of IFN-γ by circulating T cells is reduced in CD compared with healthy subjects41 and, in the current study, in CD patients with higher rates of bone resorption. The higher rates of bone resorption noted in CD patients with a systemic inflammatory response may therefore by explained by lower rates of production of IFN-γ by activated T cells and the subsequent loss of an inhibiting influence on osteoclast differentiation and function.
The results of the current study may have potentially important clinical implications in the assessment and monitoring of CD patients for the risk of developing osteoporosis. First, patients with persistent or frequently raised CRP and ESR, even in the presence of clinically mild or quiescent disease, may be at risk of a chronic increase in bone loss. These patients should be targeted for early investigation, monitoring and treatment. Secondly, there may be value in assessing the response of inflammatory markers, i.e. the ‘systemic’ inflammatory response, to treatments in CD, and in certain circumstances their failure to respond may indicate the need for more aggressive therapy. Thirdly, we have failed to demonstrate that disease symptoms alone (as the CDAI) in mild CD has power in identifying patients at increased risk of bone loss.
Further investigation of the associations noted in the current study are indicated. A clearer understanding of the relationship between the systemic inflammatory response and bone turnover would be provided by comparing CD patients with more severe disease, i.e. at disease relapse, with patients with quiescent disease and healthy volunteers. Moreover, an intervention study is indicated to determine the response of bone turnover to modulation of PBMC function, and in particular IFN-γ production, and therefore provide a clearer understanding of the possible role of circulating mononuclear cells in the pathogenesis of secondary osteoporosis in CD.
This study was supported by grants to TT from The Southampton Rheumatology Trust, South and East NHS Executive Research & Development; Nutricia Clinical Care, ARC.