Background : Osteoporosis is a frequent complication in Crohn's disease. Although the efficacy of both sodium fluoride and aminobisphosphonates in postmenopausal osteoporosis has been investigated in long-term therapy studies, no long-term results are available regarding the effect of these agents in the management of osteoporosis in patients with Crohn's disease.
Methods : Eighty-four patients with Crohn's disease and pathological bone mineral density findings were randomized to receive either vitamin D3 (1000 IU) and calcium citrate (800 mg) daily (group A) or sodium fluoride (25 mg b.d., group B) or intravenous ibandronate (1 mg every 3 months, group C) in addition to daily calcium/vitamin D substitution. On admission to the study and after 12 and 27 months, patients underwent dual-energy X-ray absorptiometry and radiological examination of the spine.
Results : Sixty-eight patients completed the 1-year observation period and were available for the intention-to-treat analysis. No new vertebral fractures were diagnosed. In group A, lumbar bone density increased by 2.6% (P = 0.066, N.S.), in group B by 5.7% (P = 0.003) and in group C by 5.4% (P = 0.003). Therapy with sodium fluoride was associated with an increase in osteocalcin (N.S.), whereas administration of ibandronate was associated with a decrease in the resorption parameter, carboxy-terminal cross-linked type-I collagen telopeptide (P < 0.05). Both sodium fluoride and ibandronate resulted in significant decreases in the serum concentration of osteoprotegerin after 9 months (P < 0.001).
Conclusions : The findings of the present study show that both sodium fluoride and ibandronate are effective in combination with calcium and vitamin D substitution in the management of osteopenia and osteoporosis in patients with Crohn's disease. Both agents are safe and well tolerated, and induce continuous increases in lumbar bone density.
Osteoporosis occurs as a complication of chronic inflammatory bowel disease in 5.3–38% of patients.1–5 An increased risk of fractures has been reported for children and adults.6–8 Pathological vertebral fractures are present in 21.7% of patients with Crohn's disease and reduced bone density.9
Therapy with systemic glucocorticoids and a low body mass index are significant risk factors.4,5,10 A recently defined molecular concept of osteoclasts and bone metabolism has assigned a central role in osteoclast function to osteoprotegerin and its ligand (receptor activator of NK-κB ligand, RANKL).11 By induction of RANKL, several pro-inflammatory cytokines, such as interleukin-1 and tumour necrosis factor-α, stimulate the proliferation of osteoclasts, while at the same time inhibiting the apoptosis of mature osteoclasts. By contrast, osteoprotegerin, which acts as a receptor antagonist, blocks these effects, reduces these osteoclast functions and prevents bone resorption and bone loss.11 Glucocorticoids have been demonstrated to inhibit osteoprotegerin and stimulate RANKL production in vitro and in vivo.12
In addition, the association of polymorphisms in interleukin-1 receptor antagonist and interleukin-6 genes with reduced bone density points to the importance of genetic factors in bone metabolism in inflammatory bowel disease.13
Both hormone replacement therapy14 and the administration of aminobisphosphonates15,16 are established methods in the management of postmenopausal and steroid-induced osteoporosis, and reduce the risk of fracture. Sodium fluoride can also increase bone density; its efficacy in reducing the rate of fracture, however, remains controversial.17–19 To date, only a few studies have evaluated the management of osteoporosis in patients with inflammatory bowel disease. Vogelsang et al. have shown that vitamin D and calcium substitution can inhibit the rate of bone loss.20 In a previous study, we were able to demonstrate the efficacy of sodium fluoride in increasing bone mineral density in Crohn's disease,21 and another recent publication reported a significant increase in bone density in response to the administration of alendronate (10 mg/day).22 However, both studies were performed in small patient cohorts. In both studies, follow-up was limited to 1 year and the primary clinical end-point was bone density; the incidence of vertebral fractures was not evaluated.
The objective of this study was to compare the effects of a combined anabolic and anti-resorptive therapeutic regimen with vitamin D and calcium substitution alone. In addition, this study included a larger patient population and a longer period of follow-up, which should permit the comparison of various therapeutic modalities with regard to fracture incidence.
Eighty-four patients with confirmed Crohn's disease and osteopenia/osteoporosis of the lumbar spinal column (t-value < − 1) were recruited prospectively between April 1998 and June 1999. Female patients planning pregnancy, patients with renal insufficiency and those who had undergone previous treatment with either sodium fluoride or aminobisphosphonates were excluded from the study. The patients' data are summarized in Table 1.
Table 1. Baseline data of the 84 patients included in the study. Data are presented as mean ± S.E.M.
Bone mineral density measurements were conducted by an independent, specialized radiologist. Bone density measurements were acquired from the non-fractured lumbar vertebrae L1–L4 and from the right femur using dual-energy X-ray absorptiometry (QDR-1000, Hologic Inc., Waltham, MA, USA). The individual bone density findings of each patient were compared with the mean value established for a control collective of subjects of about 30 years of age compiled by Hologic, Inc. and provided as a data bank for the bone densitometry unit (t-value). In contrast, the z-score compares the individual result with the mean value of an age-matched healthy control collective.
Patients were randomized to the following three treatment groups: (i) 1000 IU vitamin D3 (Vigantoletten, Merck, Darmstadt, Germany) and 800 mg calcium citrate (Calcitrat, Merckle, Ulm, Germany) daily (group A); (ii) 1000 IU vitamin D3 (Vigantoletten) and 800 mg calcium citrate (Calcitrat) daily, plus 25 mg of slow-release sodium fluoride (Nafril, Merckle, Ulm, Germany) b.d. (group B); (iii) 1000 IU vitamin D3 (Vigantoletten) and 800 mg calcium citrate (Calcitrat) daily, plus 1 mg ibandronate (Bondronat, Roche, Basle, Switzerland) intravenously every 3 months (group C).
Patients with osteopenia of the lumbar spinal column (t-values of non-fractured vertebrae L1–L4 of < − 1 but > − 2.5) were randomized to groups A, B or C. Patients with osteoporosis (t-values of non-fractured vertebrae L1–L4 of < − 2.5) were randomized to groups B or C. Baseline examination in all patients included plain radiographic imaging of the thoracic and lumbar vertebrae in two planes. Morphometric evaluation for the detection of osteoporosis-related vertebral deformities was performed on the basis of criteria proposed by the European Vertebral Osteoporosis Study.23 Qualitative visual evaluation was followed by a semi-automatic quantitative calculation of the height of the vertebral bodies on lateral radiographs. Evaluation of the vertebral deformities detected by this method was conducted on the basis of various published algorithms.24,25
At the time of study recruitment, patients also underwent complete physical examination and venepuncture. Patients' haematocrit was determined for the calculation of the Crohn's disease activity index, and further parameters related to inflammation (erythrocyte sedimentation rate, C-reactive protein) and bone metabolism (osteocalcin, carboxy-terminal cross-linked type-I collagen telopeptide) were obtained.
Osteoprotegerin was measured using a highly sensitive enzyme-linked immunoabsorbent assay from Immundiagnostik (Bensheim, Germany) as reported previously.26
Follow-up examinations were conducted at 3-month intervals. In treatment group B, sodium fluoride was taken daily for 12 months, followed by 3 months of a fluoride-free period. The second 12-month cycle was started at month 15.
Follow-up bone densitometry and plain radiography of the thoracic and lumbar spinal column were performed after 12 and 27 months.
An intention-to-treat analysis was performed for all patients with at least one bone mass densitometry measurement during follow-up. All data are presented as the mean ± standard error of the mean (S.E.M.). The Mann–Whitney rank sum test was used to test the effect of each therapy on the bone density and biochemical markers after 12 and 27 months compared to baseline. Student's t-test for unpaired observations was used to compare changes in variables between all regimens. A P value of < 0.05 was considered to indicate a significant difference. Microsoft EXCEL and WinSTAT statistical programs were used for all analyses.
Fifty-five of the 84 patients (65.5%) completed the 27-month study period. The intention-to-treat analysis was performed on 68 patients (81.0%) who completed at least 1 year of follow-up. The reasons for withdrawal from the study are shown in Figure 1. The most frequent reasons for withdrawal were failure to attend for follow-up (17 patients) and personal reasons (nine patients). One patient withdrew due to an intercurrent malignancy (testicular cancer), which was retrospectively demonstrated to be present before randomization. The lower number of patients in treatment group A was due to the randomization criteria, which permitted only patients with osteopenia (t-score: < − 1 but > − 2.5) to be assigned to this group. This was also the reason why the baseline bone density of the lumbar vertebrae in patients in groups B and C was significantly lower than that in group A. Similarly, fracture of a vertebral body was observed at the time of study recruitment in only one patient in group A, but in eight patients in group B and in four patients in group C (see Table 1). No new vertebral fractures were observed in any of the 68 patients during follow-up.
The duration of disease was significantly shorter in group B than in group A (P = 0.024) or group C (P = 0.001). There was no significant difference with regard to body mass index or Crohn's disease activity index between the groups at baseline. The mean steroid dosage during the study and the number of patients requiring surgical bowel resection were comparable in each group. A slight increase in the mean body mass index and an improvement of the Crohn's disease activity index were observed during the study period in each treatment group, with no significant differences between the groups (Table 2).
Table 2. Clinical parameters of 68 patients with Crohn's disease during the course of 27 months of treatment with calcium/vitamin D (n = 12), calcium/vitamin D and sodium fluoride (n = 28) or calcium/vitamin D and ibandronate (n = 28). Data are given as mean ± S.E.M.
BMI, body mass index; CDAI, Crohn's disease activity index.
Number of patients with bowel resection during study
In the 12 patients in group A, the mean bone density in the lumbar vertebrae increased by 2.2% after 1 year (from − 1.58 ± 0.08 to − 1.39 ± 0.09; t-value ± S.E.M.), and by a further 0.4% after 2.25 years to − 1.33 ± 0.11 (N.S.; Figure 2).
In the sodium fluoride group (group B), the bone density in the lumbar spine improved over the first treatment year by 4.1% (from − 2.22 ± 0.15 to − 1.83 ± 0.16; P = 0.025), and by a further 1.6% to − 1.69 ± 0.18 (N.S.) during the second treatment phase. The serum fluoride level, which was determined at 0, 6 and 12 months, increased significantly during therapy in comparison with the baseline value, and was in the therapeutically effective range of 0.095–0.19 mg/L.27
In the 28 patients who received intravenous ibandronate (1 mg) once every 3 months (group C), the bone density measured in the lumbar vertebrae increased during the first year by 3.5% (from − 2.35 ± 0.12 to − 2.01 ± 0.13; P = 0.033), and by a further 1.9% to − 1.83 ± 0.13 (N.S.) during the second phase of treatment.
There was no significant change in femoral bone density in any of the three treatment groups during the entire follow-up period (Figure 2), and no significant difference between the t-score and z-score in the lumbar spine or total femur (Table 3).
Table 3. Bone mineral density of the lumbar spine and the femoral neck (given as t-score and z-score) in 68 patients with Crohn's disease after 12 months and 27 months of treatment with calcium/vitamin D (n = 12), calcium/vitamin D and sodium fluoride (n = 28) or calcium/vitamin D and ibandronate (n = 28)
A comparison of the three groups after 1 year revealed superior results for the group receiving sodium fluoride (group B) in comparison with the group receiving only vitamin D and calcium citrate (group A) (P = 0.042). After 2 years, however, both the group receiving sodium fluoride (group B; P = 0.034) and the group receiving ibandronate (group C; P = 0.022) showed significantly better results than the group receiving only vitamin D and calcium (group A) (Figure 2). Comparison between the sodium fluoride and ibandronate groups showed no significant difference during the observation period.
When patients with pre-existing osteopenia (lumbar t-score of < − 1.5 and > − 2.5) and patients with osteoporosis (lumbar t-score of < − 2.5) were analysed separately, the increase in the lumbar spine bone mineral density after 2 years in the fluoride group was more pronounced in osteoporotic patients (N.S.). No significant differences between osteopenia and osteoporosis were found with regard to the other treatment groups or the femoral bone mineral density (Table 4).
Table 4. Effect of calcium/vitamin D substitution alone or a combination of calcium/vitamin D with sodium fluoride or ibandronate on the lumbar bone density in relation to pre-existing osteopenia or osteoporosis
Thirty-five of the 55 patients who completed the 27-month study period were treated with systemic glucocorticoids at least once during the treatment period. The change in lumbar bone density in this patient group (+ 6.28 ± 1.13%) did not differ significantly from the change observed in patients who had not received systemic steroids during this period (+ 5.05 ± 0.97%). Bowel resection, which was required in 11 patients during the follow-up period, also did not exert a measurable effect on bone density. On the other hand, in the sodium fluoride group, the increase in bone density correlated with the patients' baseline bone density measurements.
Carboxy-terminal cross-linked type-I collagen telopeptide, a parameter specific for bone resorption, decreased significantly in comparison with baseline values in the ibandronate group at 12, 18 and 21 months (P < 0.05; Figure 3), although no comparable change was observed in the sodium fluoride group. On the other hand, osteocalcin, a specific marker of bone formation, increased in the fluoride group after 6 months of treatment (N.S.) and, at that point, was significantly higher than that in the ibandronate group (P = 0.01).
Osteoprotegerin, an anti-resorptive cytokine synthesized by osteoblasts, decreased significantly in both treatment groups after 9 months (P < 0.001), and stayed at this level for the remaining study period. There was no significant difference between osteoprotegerin levels in the fluoride and ibandronate groups (Figure 3).
Clinical adverse events occurred in 23 patients (group A, n = 3; group B, n = 11; group C, n = 9). Most adverse events were related to a worsening of the underlying inflammatory bowel disease (16 patients). Two female patients in the ibandronate group had to be withdrawn due to pregnancy and one male patient in the fluoride group due to testicular cancer (see above and Figure 1). The study medication was generally well tolerated by all patients. Two patients who received sodium fluoride found undigested pills in their faeces, and two patients in group C reported minor and completely reversible bone pain (< 2 h) after the infusion of ibandronate.
The aim of this study was to compare the efficacy of osteo-anabolic and anti-resorptive therapeutic regimens in Crohn's disease patients with osteopenia or osteoporosis.
The data show that both cyclic treatment with sodium fluoride and intermittent treatment with ibandronate, as adjuncts to calcium and vitamin D substitution, significantly increased the lumbar bone mineral density. After the 2.25-year treatment period, the changes in bone mineral density were comparable between the sodium fluoride group (+ 5.7%) and the ibandronate group (+ 5.4%), and both regimens were superior to calcium/vitamin D substitution alone.
This effect was particularly pronounced during the first year of treatment. Thereafter, the bone density continued to increase, but at a lower rate. The increase in bone density in patients receiving sodium fluoride was 4.1% after 1 year and 5.7% after 2.25 years. This is somewhat less than in our recently published pilot study in 15 patients with Crohn's disease.21 The difference is most probably due to the lower sodium fluoride dose in the present study (50 mg vs. 75 mg). We chose the lower dose for the present study on the basis of published data from an investigation using the same dosage and the same slow-release form of the agent.17 This study of postmenopausal women, published by Pak et al. in 1995, reported increases in bone density of 4–5% after 1 year.17
The administration of intravenous ibandronate (1 mg every 3 months) also resulted in a continuous increase in the lumbar bone density of 3.5% and 5.4% after 1 and 2.25 years, respectively. This increase was less pronounced than that in the fluoride group, and corresponds to recently published data on the efficacy of alendronate for the treatment of osteoporosis in Crohn's disease,22 and of ibandronate in postmenopausal osteoporosis.28 The effect on bone density appears to be dose dependent, and is most pronounced at an ibandronate dose of 2 mg, associated with an annual bone density increase of 5.2%.28
Unlike earlier studies,20,21 our data showed that the bone density also tended to increase during the study period in patients receiving only calcium and vitamin D. In addition to the small number of cases in the earlier studies, another reason for this finding could be the reduction in disease activity, which was observed in all groups, but significant only in the ibandronate-treated group (P = 0.016). Furthermore, the bone mineral density seemed to be influenced by the body weight and body mass index during the study. Although the increase in the body mass index did not reach statistical significance in any group, there was a significant correlation between the change in the body mass index and the bone mineral density during the observation period (r = 0.46, P < 0.001). These findings emphasize that effective treatment of the underlying bowel disease and an associated improvement of the body weight may be important factors in the management of osteoporosis.
In planning our study, we did not include a placebo group for ethical reasons. For this reason, however, the findings of the present study do not permit conclusions to be drawn on the natural progression of change in the patients' bone density. As other studies have reported an annual decrease in bone density of 3–7%,20,29 the present findings support the hypothesis that calcium and vitamin D substitution is partially effective in preventing further bone loss in patients with Crohn's disease-related osteopenia/osteoporosis.
The question of which therapeutic regimen is most effective in treating Crohn's disease-related osteoporosis can only be answered on the basis of the relative fracture incidence. Therapy studies focusing on this factor have underscored the efficacy of bisphosphonates in postmenopausal and steroid-induced osteoporosis.15,16
Less information is available from therapy studies investigating sodium fluoride. Most of these studies have been performed in postmenopausal osteoporosis, and the efficacy of sodium fluoride in preventing new fractures remains controversial.17,18 Nevertheless, Rubin et al. have reported the efficacy of sustained-release sodium fluoride in the prevention of vertebral fractures in postmenopausal osteoporosis.19
Unlike postmenopausal osteoporosis, in which, due to oestrogen deficiency, osteoclast resorption of bone predominates, the pathogenesis in patients with inflammatory bowel disease is multi-factorial and not yet fully understood. Because of the decrease in osteoblast activity caused by systemic glucocorticoids,10 and the actions of pro-inflammatory cytokines (interleukin-1, interleukin-6), a therapeutic strategy that includes osteoblast stimulation and the inhibition of osteoclast resorption would appear to be justified. Nonetheless, the findings of the present study have not demonstrated any differences in terms of fracture incidence. Because our patient cohort included relatively young patients (mean age, 38.6 years), the effects of other age-related factors, such as general fragility and poor vision, have no influence, and this may be one factor leading to a fracture incidence lower than that observed in postmenopausal osteoporosis. In order to study the fracture incidence in patients with chronic inflammatory bowel disease more reliably and to compare the effects of different treatment strategies, larger patient numbers and longer follow-up periods than those used in the present study are required.
It should be noted that serum levels of the anti-resorptive cytokine osteoprotegerin were decreased during treatment with both sodium fluoride and ibandronate. The measurement of osteoprotegerin serum concentrations has recently been used as an additional biochemical marker of bone metabolism in a variety of metabolic bone diseases, including postmenopausal osteoporosis,30,31 rheumatoid arthritis,32 juvenile Paget's disease33 and various malignant bone diseases, such as myeloma bone disease.34 In one of these studies, serum concentrations of osteoprotegerin were highest in those postmenopausal women with the most pronounced rate of bone loss, and has been interpreted as an insufficient counter-regulatory mechanism to prevent further bone loss.30 The clinical potential of osteoprotegerin serum measurement and its limitations in assessing bone metabolism have been reviewed recently.35 It should be noted that osteoprotegerin is produced by various skeletal and extra-skeletal tissues, and serum measurement of osteoprotegerin may only partly reflect the regulation pattern of osteoprotegerin occurring within the bone micro-environment.35 Thus far, the diagnostic value of the osteoprotegerin serum level, its correlation with other biochemical markers of bone turnover and its predictive value in osteoporosis risk assessment have not been clarified.
Factors identified in epidemiological studies as risk factors for the development of osteoporosis in patients with chronic inflammatory bowel disease, such as systemic glucocorticoid administration or surgical bowel resection,4,5,10 had no significant effect on the bone mineral density in our study. On the other hand, baseline bone density did have an effect in patients treated with sodium fluoride. The increase in bone density during therapy was particularly pronounced in patients with osteoporosis. One explanation for this observation is the possibility that sodium fluoride is only effective in bone which has already been affected by osteoclast resorption.
It is interesting that a significant increase in bone density with both sodium fluoride and ibandronate was documented only for the lumbar spinal column. Failure to increase bone density in the femur has already been reported for fluoride therapy.19,36 One explanation may be that the turnover in mixed-type bone, such as in the femoral neck, is much lower than that in cancellous bone and, consequently, the magnitude of response to therapy would be smaller in the femur (and the sample size required to show an effect larger) than in the spine. It is, however, surprising that the bone density in the femur does not undergo a significant increase in patients receiving ibandronate. As the effect of the drug on the femoral bone density may be less than that in the lumbar vertebrae, the dose (1 mg) applied in the present study may have been too low to observe a significant increase.28 On the other hand, the mean standard error in the measurement of the femoral bone density was significantly higher than that associated with measurements of the lumbar spine, suggesting a reduced precision of the method when applied to the femur.
In conclusion, the present data show that both sodium fluoride and ibandronate are effective adjuncts to calcium and vitamin D substitution in the treatment of osteopenia and osteoporosis associated with Crohn's disease. Both are safe and well-tolerated substances which produce a continuous increase in lumbar bone density. The findings, however, do not permit an unequivocal statement to be made regarding the efficacy of the methods in preventing new fractures in patients with Crohn's disease. This can only be determined from future, multi-centre, long-term therapy studies.
Morphometry of vertebral radiographs was supported by the Osteoporosis Study Group of the Clinic for Radiology and Nuclear Medicine, Klinikum Benjamin Franklin, Berlin, Germany. Bone mineral densitometry was performed by the Division of Radiology, University of Ulm, Germany. Serum fluoride levels were measured by André Scholer, Department of Clinical Chemistry, University Hospital, Basle, Switzerland.
The study was supported by a grant from the ‘Kompetenznetzwerk chronisch entzündliche Darmerkrankungen’ and the German Crohn's and Colitis Association (DCCV).