Abnormal bone turnover in long-standing Crohn’s disease in remission

Authors


Dr E. J. Schoon, Department of Gastroenterology and Hepatology, University Hospital Maastricht, PO Box 5800, NL-6202 AZ Maastricht, the Netherlands. E-mail: ESCH@sint.azm nL

Abstract

Background:

A high prevalence of osteoporosis is found in patients with Crohn’s disease. The pathogenesis of this condition seems to be multifactorial and its pathophysiology is still not completely understood.

Aim:

To elucidate the pathophysiology of osteopenia in quiescent Crohn’s disease.

Methods:

Bone turnover was studied in 26 patients (13 males and 13 females) with long-standing quiescent Crohn’s disease and small bowel involvement. Bone mineral density was assessed by dual energy X-ray absorptiometry. Biochemical markers for bone formation (osteocalcin and bone-specific alkaline phosphatase) and for bone resorption (deoxypyridinoline and collagen type I C-terminal crosslinks) were measured. Urinary calcium excretion was determined.

Results:

Markers for bone formation were significantly lower in patients than in controls (osteocalcin: P= 0.027, bone-specific alkaline phosphatase: P < 0.001), but both bone resorption markers were not significantly different. Urine calcium excretion was significantly decreased in patients (P=0.002) compared to controls. Bone mineral density of the lumbar spine was significantly and inversely correlated with bone-specific alkaline phosphatase and collagen type I C-terminal crosslinks.

Conclusions:

Bone turnover in long-standing Crohn’s disease in clinical remission is characterized by suppressed bone formation and normal bone resorption. Urine calcium excretion is decreased. Hence, interventions and therapy should be directed towards the improvement of bone formation.

INTRODUCTION

Patients with Crohn’s disease are at high risk of developing osteopenia and osteoporosis.1[2][3][4]–5 The pathogenesis and pathophysiology of these conditions in Crohn’s disease are still not completely understood. A number of factors are considered to contribute to the reduced bone density. These include: steroid use, malnutrition, vitamin D and calcium deficiency, immobilization, smoking, sex hormone deficiency, hyperparathyroidism, and the inflammatory process itself.6 It has been demonstrated that clinical risk factors are poor diagnostic predictors of actual bone mass.7 In a large controlled study, low bone mineral density was found in patients with Crohn’s disease, but not in those with ulcerative colitis.8

The pathophysiological process can be clarified by studying bone turnover by means of biochemical markers that reflect bone turnover in the entire skeleton and have the advantage of being non-invasive, relatively inexpensive and of allowing repeated evaluation.9 Biochemical markers of bone resorption are: serum osteocalcinl total and bone-specific alkaline phosphatase; and procollagen I extension peptides. Markers of bone resorption are: urinary hydroxyproline and hydroxylisine glycosides; urinary pyrinoline and deoxypyriniline; and serum tartrate-resistant alkaline phosphatase and pyrodinoline peptides. Generally, these markers correlate poorly with current bone mineral density and therefore are not appropriate to diagnose low bone mineral density.10[11]–12 Published data are conflicting with regard to whether osteopenia in Crohn’s disease is due to increased bone resorption, suppressed bone formation or both.10[11][12][13][14]–15 Some studies have failed to indicate significant changes in biochemical markers in patients compared to controls.16[17]–18

In patients with recently diagnosed inflammatory bowel disease, bone mineral density was not different from that in controls, indicating that the subsequent development of osteoporosis must be related to the disease process and/or the treatment modalities of inflammatory bowel disease.19 The use of corticosteroids in patients with Crohn’s disease is considered to be an important risk factor for low bone mineral density. In cross-sectional bone mineral density studies in Crohn’s disease patients, the correlation between the cumulative corticosteroids dose and low bone mineral density is not unanimous. One of the pathophysiological mechanisms of the effect of corticosteroids on bone in Crohn’s disease is the suppression of bone formation.13, 14 Most of the studies on bone turnover in inflammatory bowel disease have been flawed by the inclusion of heterogeneous patient populations regarding the type of disease (Crohn’s disease and ulcerative colitis), the disease activity (active disease and disease in remission), the administration of corticosteroids and the menopausal status of female patients. In two studies, only Crohn’s disease patients were included; in one study including 20 male patients with long-standing quiescent Crohn’s disease, no difference in biochemical markers was demonstrated between patients and controls.20 Another large study, comprising a population of 117 Crohn’s disease patients showed increased bone resorption. However, patients using corticosteroids and post-menopausal female patients were included in this study.21 To date, the effects of disease activity on bone turnover in Crohn’s disease have only been studied in vitro.22

The aim of this study was to evaluate bone turnover using biochemical markers, in long-standing Crohn’s disease in the absence of active disease, of significant corticosteroid use and of the influence of menopausal status in order to further elucidate the pathophysiology of this condition.

METHODS

Subjects

Patients were asked to participate in this study while attending the Gastroenterology out-patient clinic of the University Hospital Maastricht. Inclusion criteria were: Crohn’s disease; clinical remission; prednisolone dose ≤ 5 mg/day; and duration of disease of 5 years or longer. Exclusion criteria were: post-menopausal status in female patients; past or current active treatment for osteoporosis (bisphosphonates, calcitonine, fluorides, hormone replacement therapy apart from oral contraceptives); and concomitant disease predisposing to secondary osteoporosis (e.g. ankylosing spondylitis, liver disease, renal insufficiency). Twenty-six patients with Crohn’s disease (13 males, age range 18–68 years and 13 females, age range 22–46 years) with an overall mean age of 38 ± 12 years (s.d.) were included in the study. Basic characteristics of the patient population are given in Table 1. Controls were healthy, age- and gender-matched persons (males, 20–70 years and pre-menopausal females, 20–50 years) selected from a community registry.

Table 1.   Basic characteristics of 26 patients with long-standing quiescent Crohn’s disease Thumbnail image of

Crohn’s disease was diagnosed by a combination of clinical symptoms and endoscopic, radiological and histological data for which the Lennard–Jones criteria were applied.23 At the time of the study, all patients had been in clinical remission for at least 3 months before inclusion. Disease activity was measured using the Crohn’s disease activity index (CDAI).24 All patients had small bowel involvement and in 13 patients inflammation also involved the colon. Four patients were taking prednisolone (≤ 5 mg/day). Steroid doses were kept stable over at least 1 month before inclusion. All patients were using 5-amino salicylic acid (5-ASA) in a dosage of 2–3 g/day. Five female patients were taking oral contraceptive medication. Five patients used physiological doses of vitamin D (400 IE/day), and three of them were also taking a low-dose calcium supplementation (500 mg/day) which, at the time of the study, was not considered to be an active treatment of osteoporosis. Thirteen patients were cigarette smokers; for all patients the number of pack-years was calculated. The patients’ lifetime physical activity and their activities during the last 6 months were evaluated according to Baecke.25 The study protocol was approved by the Ethics Committee of the University Hospital of Maastricht, and all subjects gave their informed consent before the start of the study.

Biochemical assessment

Blood samples taken at baseline in the morning after an overnight fast were immediately centrifuged and stored at − 70 °C. The second morning urine portion was collected on the same day and stored at − 70 °C. Two biochemical markers for bone formation and two markers for bone resorption were determined. For the assessment of bone formation, serum osteocalcin and bone-specific alkaline phosphatase were measured. For the assessment of bone resorption, collagen type I C-terminal crosslinks and deoxypyridinoline were determined in urine.

Serum osteocalcin was measured using the Osteometer test kit (Osteometer Bio Tech A/S, Copenhagen/Denmark). Bone specific alkaline phosphatase was measured using the IRMA test kit (Hybritech, Liege, Belgium).26 The results were compared to those of a group of population controls consisting of 90 healthy men (age range 20–70 years) and women (age range 20–50 years).

Collagen type I C-terminal crosslinks is a product of type I collagen, that is degraded during remodelling of the skeleton and is excreted in the urine.27 Collagen type I C-terminal crosslinks was measured using the Crosslabs TM ELISA technique (Osteometer Bio Tech A/S, Copenhagen/Denmark).28 Deoxypyridinoline was measured using Pyrilinks-D, a competitive enzyme immunoassay for measuring deoxypyridinoline in urine (Metra Biosystems Inc., Mountain View, California/USA).29 Results for both resorption parameters were calculated as the collagen type I C-terminal crosslinks/creatinine and deoxypyridinoline/creatinine ratio to correct for small differences in renal function. Data were compared to those of a group of 12 controls, six healthy males and six healthy females ranging in age from 18 to 28 years.

Calcium excretion in urine was measured using atom absorption spectrometry and calculated as the calcium/creatinine ratio. The results were compared to the same control population as osteocalcin and bone-specific alkaline phosphatase.

Serum 25-hydroxyvitamin D concentration was measured using a 125I radioactive immunoassay (Incstar Corporation, Stillwater, Minnesota, USA) in specimens obtained in April for which the winter reference value was applied (25–70 nmol/L).30 Routine laboratory parameters were measured including serum creatinine and haematocrit, in order to calculate the Crohn’s disease activity index.

Bone mineral density

Bone mineral density was measured using dual energy X-ray absorptiometry (Lunar DPX-L, Lunar software version DPX-L 4.7; Lunar Radiation Corp., Madison, WI) of the lumbar spine (L2–L4), femoral neck and total body.31 Bone mineral density was expressed in absolute values (g/cm2), T-score (one standard deviation compared to the mean of a young adult gender-matched reference population) and Z-score (one standard deviation compared to the mean of an age- and gender-matched reference population), respectively. Reference data were based on populations from the United States, United Kingdom and Northern Europe.32[33]–34 There was a 1.3% s.d. in the average density values among various geographical areas. The diagnosis of osteopenia and osteoporosis was based on T-scores according to the WHO criteria.35

Statistics

Continuous data were presented as mean (± s.d.) when normally distributed or as median (range) when not. If continuous data were normally distributed, a Student’s t-test was applied. In other cases, the non-parametric Mann–Whitney U-test was used. Correlations between continuous variables were assessed using the Pearson’s correlation test or Spearman’s rank test, respectively. Two-tailed tests for significance were used in all the statistical analyses and P ≤ 0.05 was considered statistically significant. The Statistical Package for the Social Sciences (SPSS) was used for the analysis (version 7.5, SPSS Inc. 1998).

RESULTS

Mean bone mineral density of the lumbar spine, of the femoral neck, and of the total body, as well as the T- and Z-scores and the prevalence of osteopenia and osteoporosis are given in Table 2.

Table 2.   Mean bone mineral density, T- and Z-scores, and prevalence of osteoporosis and osteopenia (according to the WHO criteria) in 26 patients with long-standing Crohn’s disease in remission35Thumbnail image of

In the patients, bone-specific alkaline phosphatase and osteocalcin, as biochemical markers of bone formation, were significantly decreased compared to the control population (P < 0.001 and P=0.027, respectively), while collagen type I C-terminal crosslinks and deoxypyridinoline as bone resorption markers were not significantly different from those of controls. Urinary calcium excretion was significantly decreased (P=0.002) compared to controls (Table 3, Figures 1, 2 and 3).

Table 3.   Results of the biochemical markers of bone turnover in patients with Crohn’s disease (n = 26) compared to controls Thumbnail image of
Figure 1.

 (A) Osteocalcin in 26 patients with long-standing quiescent Crohn’s disease vs. 90 healthy controls. Lines indicate median values. Mann–Whitney U-test. (B) Bone-specific alkaline phosphatase in 26 patients with long-standing quiescent Crohn’s disease vs. 90 healthy controls. Lines indicate median values. Mann–Whitney U-test.

Figure 2. (A) Collagen type I C‐terminal crosslinks (CTX)/creatinine (nmol/μmol) urinary excretion in 26 patients with long‐standing quiescent Crohn’s disease vs. 12 healthy controls. Lines indicate median values. Mann–Whitney U‐test. (B) Deoxypyridinoline (DPD)/creatinine (nmol/μmol) urinary excretion in 26 patients with long‐standing quiescent Crohn’s disease vs. 12.

Figure 2. (A) Collagen type I C-terminal crosslinks (CTX)/creatinine (nmol/μmol) urinary excretion in 26 patients with long-standing quiescent Crohn’s disease vs. 12 healthy controls. Lines indicate median values. Mann–Whitney U-test. (B) Deoxypyridinoline (DPD)/creatinine (nmol/μmol) urinary excretion in 26 patients with long-standing quiescent Crohn’s disease vs. 12.

healthy controls. Lines indicate median values. Mann–Whitney U-test.

Figure 3.

 Calcium/creatinine urinary excretion in 26 patients with long-standing quiescent Crohn’s disease vs. 90 healthy controls. Lines indicate median values. Mann–Whitney U-test.

A significant inverse correlation was found between the serum level of bone-specific alkaline phosphatase and the lumbar spine bone mineral density (R=−0.386, P=0.047), the lumbar spine T-score (R=−0.408, P=0.035; Figure 4), but not between bone-specific alkaline phosphatase and the lumbar spine Z-score (R=−0.321, P=0.103). The collagen type I C-terminal crosslinks/creatinine ratio also inversely correlated with lumbar spine bone mineral density (R=−0.389, P=0.030), lumbar spine T-score (R=−0.390, P=0.049; Figure 5) and lumbar spine Z-score (R=−0.400, P=0.043). A significant inverse correlation was found between the serum level of bone-specific alkaline phosphatase and the total body T-score (R=−0.421, P=0.029; Figure 6). No significant correlations were found between the biochemical markers and femoral neck bone mineral density scores.

Figure 4. Correlation between the lumbar spine T‐score and serum bone‐specific alkaline phosphatase level (bone‐specific alkaline phosphatase) in 26 patients with long‐standing quiescent Crohn’s disease. Pearson’s coefficient R=−0.4.

Figure 4. Correlation between the lumbar spine T-score and serum bone-specific alkaline phosphatase level (bone-specific alkaline phosphatase) in 26 patients with long-standing quiescent Crohn’s disease. Pearson’s coefficient R=−0.4.

08, P=0.035.

Figure 5.

 Correlation between the lumbar spine T-score and collagen type I C-terminal crosslinks/creatinine ratio (expressed in nmol/μmol) in 26 patients with long-standing quiescent Crohn’s disease. Pearson’s coefficient R=−0.390, P=0.049.

Figure 6. Correlation between the total body T‐score and serum bone‐specific alkaline phosphatase level in 26.

Figure 6. Correlation between the total body T-score and serum bone-specific alkaline phosphatase level in 26.

patients with long-standing quiescent Crohn’s disease. Pearson’s coefficient R=− 0.421, P=0.029.

The mean serum level of vitamin D was 28 ± 11 nmol/L; nine patients (35%) were considered vitamin D-deficient (serum 25-hydroxyvitamin D < 25 nmol/L), and only one patient had a serum vitamin D level below 10 nmol/L. Of the patients taking a vitamin D supplement (400 IU/day for more than 2 months), three patients still had serum vitamin D levels below 25 nmol/L (13, 14 and 23 nmol/L, respectively). No correlation was found between the extent of the small bowel resection (cm) and the vitamin D levels in those patients who did not take vitamin D supplements. Six patients had a Crohn’s disease activity index above 150. In these patients, the C-reactive protein, which is a sensitive indicator of disease activity in inflammatory bowel disease, ranged from 2 to 10 mg/L. The clinical characteristics of these six patients are given in Table 4. Exclusion of these patients did not change the results; a significant difference was still found in bone formation (osteocalcin and bone-specific alkaline phosphatase) but not in bone resorption (deoxypyridinoline and collagen type I C-terminal crosslinks).

Table 4.   Clinical characteristics of six patients with Crohn’s disease with a Crohn’s disease activity index (CDAI) above 150 Thumbnail image of

Body mass index (R=−0.631, P < 0.001), percentage body fat (R=−0.569, P=0.002) and serum vitamin D level (R=−0.450, P=0.018) inversely correlated with Crohn’s disease activity index.

No significant correlations were found between biochemical markers of bone turnover and physical activity or body mass index. No significant differences were found in markers of bone turnover between smokers and non-smokers.

DISCUSSION

In this study, biochemical markers of bone formation (bone-specific alkaline phosphatase and osteocalcin) in patients with long-standing quiescent Crohn’s disease were significantly decreased compared to those in a control population, while bone resorption markers (collagen type I C-terminal crosslinks and deoxypyridinoline) were not significantly different from those in controls. An inverse correlation was found between both bone mineral density of the lumbar spine and bone-specific alkaline phosphatase and bone mineral density of the lumbar spine and collagen type I C-terminal crosslinks/creatinine ratio. Urinary calcium excretion was also significantly decreased compared to that in normal controls.

This unbalanced bone metabolism in patients with longstanding Crohn’s disease is pathological and seems to be an ongoing process, even in the absence of clinical disease activity and significant corticosteroid use. Uncoupling of the bone degradation–formation cycle can be a risk factor for the progression of bone loss and can eventually lead to bone fractures. The inverse correlation between bone mineral density of the lumbar spine and one marker for bone formation (bone-specific alkaline phosphatase) as well as one marker for bone resorption (collagen type I C-terminal crosslinks/creatinine ratio) indicates that a higher bone turnover level is associated with a lower bone mineral density. If the turnover is unbalanced, bone loss occurs even during the ‘quiescent’ phase of the disease process. Just how much the unbalanced bone turnover contributes to bone loss has to be proven in a follow-up study on these patients. The trabecular bone in the lumbar spine is metabolically more active than the cortical bone of the femoral neck and this may explain the absence of correlation between biochemical markers and bone mineral density of the femoral neck.

The use of corticosteroids and disease activity with high levels of circulating pro-inflammatory cytokines are supposed to play a pivotal role in the bone metabolism of patients with inflammatory bowel disease. To exclude the confounding effects of active disease and corticosteroid use, we investigated a group of patients with inactive Crohn’s disease, all with small bowel involvement and/or a previous small bowel resection. They had no or very low stable doses of steroids. In six patients, Crohn’s disease activity index scores were higher than 150; all patients were in clinical remission and had a low C-reactive protein.

The results of this study are in agreement with the results of a histomorphometric study performed in 19 patients with inflammatory bowel disease and osteoporosis; reduced bone formation was found at cellular level with a negative remodelling balance.36 Although, in the present study, the biochemical markers for bone turnover bone-specific alkaline phosphatase and collagen type I C-terminal crosslinks/creatinine were significant and inversely correlated with the bone mineral density of the lumbar spine, these correlations were low. Recently, increased urinary N-telopeptide crosslinked type I collagen (NTx) excretion was found to be predictive of future spinal bone loss in inflammatory bowel disease patients.15 The value of these markers as a diagnostic tool for osteoporosis for the individual patient with Crohn’s disease needs to be further investigated, given the evidence that the level of bone turnover is as strong a predictor of future fractures as the level of bone mineral density in conditions other than inflammatory bowel disease.37[38][39]–40

The different results of other studies on bone turnover in Crohn’s disease may be explained by a different bone formation or bone resorption, responsible for bone loss at different phases of the disease process.18 The patients of the present study were homogeneous according to disease activity, disease duration, steroid use and menopausal status. In only one study a comparable population of 20 male Crohn’s disease patients with quiescent long-standing Crohn’s disease was studied.20 In this study, normal bone turnover was found. The differences in outcome with the present study can possibly be explained by demography and some clinical features (only male patients were involved; patients with more than five bowel movements per day were excluded; fewer patients had been resected [55% vs. 73%]; and the disease duration was markedly shorter, mean of 10 vs. 16 years, respectively). The biochemical markers used were different in both studies, except for osteocalcin. In the same study, fractional calcium absorption was not different from that of controls.

In patients with Crohn’s disease and with ileopathy, due either to inflammation or to previous resection, bile acid and fat malabsorption is prevalent, causing steatorrhoea and malabsorption of fat soluble vitamins. This is suggested in the present study by the inverse correlation between Crohn’s disease activity index and body mass index, body fat and serum vitamin D level, respectively. A low urinary calcium excretion can be a consequence of vitamin D deficiency, which was prevalent in this group of patients (35%). The mean age of the controls for the deoxypyridinoline and collagen type I C-terminal crosslinks measurements was lower than the mean age of the patients. If there had been a significant influence of increasing age, a higher bone resorption would have been expected; however, this was not the case.

It is not possible to determine, on the basis of dual energy X-ray absorptiometry measurements, whether mineralization defects also contribute to low bone mineral density. In the present patient group, clinical signs of osteomalacia as bone pain or muscle weakness were absent. Furthermore, the serum bone-specific alkaline phosphatase level, which is a sensitive indicator of osteomalacia, was generally low. The patient with the highest bone-specific alkaline phosphatase level had a normal serum 25-hydroxyvitamin D level.

The low bone formation, but normal bone resorption, found in this study indicates an unbalanced bone turnover, eventually leading to osteopenia and osteoporosis, which is already prevalent in about 40% of these patients. In Crohn’s disease, therapy should be directed to the prevention of bone loss and/or to the restoration of bone mineral density in patients with low bone mineral density values. Low bone formation points to a lower activity level of osteoblasts. Osteoblast growth and function, cellular life span and eventual apoptosis are influenced in a complex way by several hormones, circulating cytokines and growth factors.41, 42 A pharmacological agent capable of stimulating bone formation through a direct effect on osteoblastic activity is sodium fluoride, which has been extensively investigated in post-menopausal women.43 Recently, von Tirpitz et al. presented the results of a study that demonstrated a positive effect of a slow-release fluoride formulation vs. placebo on bone mineral density, also in Crohn’s disease.44 Therefore, this might be an option for patients with long-standing Crohn’s disease and osteopenia, although the quality of bone after fluoride use is still a subject of debate.45 It has been demonstrated in vitro that other agents, such as bisphosphonates and calcitonine, can prevent osteoblast and osteocyte apoptosis.46 Treatment with alendronate significantly improves lumbar spine bone mineral density in patients with Crohn’s disease compared to those taking placebo. Furthermore, biochemical markers of bone turnover decreased significantly in the alendronate group compared to those taking placebo.47

In summary, the finding of a pathological bone turnover in patients with long-standing Crohn’s disease in remission, in the absence of main risk factors for bone loss, i.e. corticosteroids and active disease, may be of importance for future preventive and therapeutic action. In these patients, therapy should be directed towards the stimulation of bone formation and the prevention and treatment of vitamin D deficiency.

Ancillary