Osteoporosis and inflammatory bowel disease

Authors

  • C. N. Bernstein,

    1. University of Manitoba Inflammatory Bowel Disease Clinical and Research Centre, and Manitoba Osteoporosis Programme, Winnipeg, Man., Canada
    Search for more papers by this author
  • W. D. Leslie

    1. University of Manitoba Inflammatory Bowel Disease Clinical and Research Centre, and Manitoba Osteoporosis Programme, Winnipeg, Man., Canada
    Search for more papers by this author

Professor C. N. Bernstein, Section of Gastroenterology, University of Manitoba, 804F-715 McDermot Avenue, Winnipeg, Man., Canada R3E 3P4.
E-mail: cbernst@cc.umanitoba.ca

Summary

Studies using dual-energy X-ray absorptiometry have suggested a high prevalence of osteoporosis in inflammatory bowel disease. However, population-based data on fracture incidence suggest only a small increased risk of fracture amongst patients with inflammatory bowel disease compared with the general population. Therefore, it would be helpful to identify patients with inflammatory bowel disease at particularly high risk for fracture so that these risks might be modified or interventions might be undertaken. The data on calcium intake as a predictor of bone mineral density are conflicting. Although there are data suggesting that a one-time survey to determine current calcium intake will not help to predict bone mineral density in inflammatory bowel disease, persistently reduced calcium intake does appear to lead to lower bone mineral density. In the general population, body mass is strongly correlated with bone mineral density, which also appears to be true in Crohn's disease. Hence, subjects with inflammatory bowel disease and considerable weight loss, or who are obviously malnourished, could be considered for bone mineral density testing, and the finding of a low bone mineral density would suggest the need for more aggressive nutritional support. Although vitamin D is undoubtedly important in bone health, vitamin D intake and serum vitamin D levels do not correlate well with bone mineral density. Sex hormone deficiency can also adversely affect bone health, although a well-developed strategy for sex hormone measurements in patients with inflammatory bowel disease remains to be established. Ultimately, the determination of genetic mutations that accurately predict fracture susceptibility may be the best hope for developing a simplified strategy for managing bone health in inflammatory bowel disease.

The therapy of osteoporosis in inflammatory bowel disease has been adapted from other osteoporosis settings, such as post-menopausal or corticosteroid-induced osteoporosis. To date, there remains no therapy proven to be efficacious in inflammatory bowel disease-related osteoporosis; however, calcium and vitamin D supplementation and bisphosphonates have their roles.

Low bone density in inflammatory bowel disease

Osteoporosis in gastrointestinal disease has become a hot topic in the past decade. With the advent of dual-energy X-ray absorptiometry, it is possible to measure the bone mineral density (BMD), a major determinant of bone strength, easily and non-invasively. BMD results are typically expressed as the number of standard deviations (s.d.) above or below the mean for a young adult population (called the T-score) or an age-matched population (called the Z-score). Early studies described a high prevalence of osteopenia (a T-score of − 1 or lower) and osteoporosis (a T-score of − 2.5 or lower) in inflammatory bowel disease (IBD).1, 2 This led some experts to recommend dual-energy X-ray absorptiometry testing in all patients with IBD,3–7 despite the absence of a proven correlation with fracture risk or proven therapies for osteoporosis in this disorder. Much of what has guided the approach in IBD has been extrapolated from post-menopausal osteoporosis, in which the pathophysiology is undoubtedly very different from that in young IBD sufferers, or corticosteroid-induced osteoporosis, for which connective tissue diseases dominate most clinical trials. Although many descriptive studies have defined the disease-specific and demographic correlates of reduced BMD in patients with IBD, there has emerged a pressing need to determine, first, whether fracture rates are increased (as would be predicted by the high prevalence of low BMD) and, second, whether therapies can impact significantly on bone health in patients with IBD.

Recently, a systematic review of the medical literature relating to bone health in IBD was published.8 For patients with established IBD, uncontrolled studies gave a prevalence of severe demineralization by dual-energy X-ray absorptiometry (Z-score of lower than − 2 or T-score of lower than − 2.5) that ranged from 18% to 42%.9–12 Larger studies (group sizes of at least 60 and, in some cases, well over 100 subjects) that included a healthy control group showed prevalence rates of severely reduced BMD of only 2–16%.13–16

Dual-energy X-ray absorptiometry and dual-photon absorptiometry (an older but similar isotope-based technology) results of the spine and hip from patients with Crohn's disease and ulcerative colitis were pooled to give a mean lumbar T-score of − 1.45, mean lumbar Z-score of − 0.48, mean hip T-score of − 1.94 and mean hip Z-score of − 0.58.8 This analysis also revealed that 14% of patients with IBD had a lumbar T-score of lower than − 2.5 and 6% had a lumbar Z-score of lower than − 2. Sixteen per cent had a hip T-score lower than − 2.5 and 13% had a hip Z-score lower than − 2.

Thus, although the magnitude of bone disease in IBD, as determined by dual-energy X-ray absorptiometry, is significant, it is important to bear in mind that most of these data were from speciality IBD clinics and may overestimate the problem due to the possibility of selection bias. In general, longitudinal changes in BMD in patients with IBD were similar to those in the general population.8, 17–22

Fracture risk in Inflammatory bowel disease

There have been four large population-based studies published describing fracture risk in IBD and all have shown similar results. A survey was mailed to members of the Danish Crohn's/Colitis Association regarding fractures.23 The relative risks for fracture amongst Crohn's disease and ulcerative colitis patients were calculated; however, the controls were not well matched and were significantly more likely to be male and older. Furthermore, the study was prone to bias as subjects at greater fracture risk may have been more likely to respond. The authors concluded that ulcerative colitis patients had an overall fracture rate similar to controls. However, patients with Crohn's disease had higher fracture rates, with a relative risk of 1.7 [95% confidence interval (CI), 1.7–2.3] for all fractures, 2.5 (95% CI, 1.7–3.6) for fractures in females and 2.9 (95% CI, 1.8–4.8) for fractures in pre-menopausal females. The relative risk in post-menopausal females was 1.8 (95% CI, 1.0–3.3) and in males was 0.6 (95% CI, 0.3–1.3). Additional risk factors for fracture in Crohn's disease included a family history of fracture (2.4; 95% CI, 1.4–4.1), and this was particularly true of a paternal fracture history (3.6; 95% CI, 1.9–6.8). A maternal fracture history led to an increased fracture risk for ulcerative colitis patients (2.4; 95% CI, 1.2–4.7). Current smoking increased the fracture risk in ulcerative colitis (3.8; 95% CI, 1.9–7.8), but not in Crohn's disease. There was no correlation for either disease group and fracture risk based on corticosteroid use. This same group published a population-based study of IBD and fractures using hospital discharge abstracts.24 Studies relying on hospital discharge abstracts will not capture out-patient fracture diagnoses and are limited to severe fractures, such as those of the hip or incapacitating spinal fractures. Subjects with IBD and age- and gender-matched controls randomly selected from the population were compared (matching ratio, 1 : 3). The authors identified 7072 Crohn's disease patients and 8323 ulcerative colitis patients from administrative data sources, but very few cases underwent direct validation of the diagnosis from the patient files. The relative risk of sustaining a fracture requiring hospitalization in Crohn's disease was 1.19 (95% CI, 1.06–1.33), whereas in ulcerative colitis it was not significantly increased at 1.08 (95% CI, 0.97–1.20).

There have been two North American population-based studies of fracture risk in IBD.25, 26 The largest study reported on 6027 patients with IBD and an age-, gender- and geographical residence-matched control group of 60 270 subjects (matching ratio, 1 : 10) using an IBD case definition carefully validated through an extensive chart review.27 Hospital and out-patient fractures were identified using the administrative databases for the Canadian province of Manitoba, which provides comprehensive health care coverage for all residents. The overall fracture rate for patients with IBD was found to be approximately 1 per 100 patient-years and this was increased compared with controls, with a relative risk of 1.41 (95% CI, 1.27–1.56), although the absolute excess risk was small. Most of the spine and hip fracture risk was in patients over 60 years of age. This study reported no significant differences between males and females or between Crohn's disease and ulcerative colitis. The observed fracture rate is consistent with the modest BMD reduction reported earlier (a pooled Z-score of − 0.5 would translate into a relative risk of fracture of 1.4 assuming a doubling in fracture rate for each s.d. reduction in BMD).

The second North American study used the Olmsted County population-based database of 243 Crohn's disease patients.26 Compared with controls, the overall risk ratio for any fracture was 0.9 (95% CI, 0.6–1.4), whereas the relative risk for an osteoporotic fracture was 1.4 (95% CI, 0.7–2.7). The risk ratio for thoracolumbar vertebral fracture was 2.2 (95% CI, 0.9–5.5). Cox proportional hazards regression identified only age as a significant clinical predictor of fracture risk (hazard ratio per 10-year increase in age, 1.3; 95% CI, 1.1–1.5). Thus, the risk of fracture was no greater than that in the general population, except in the elderly.

The fourth study was a primary care-based nested case control study using a General Practice Research Database in the UK that included 683 practices accounting for 6% of the UK population.28 This study has the advantage of accessing patients' records and hence recording corticosteroid use and symptomatology. Patients with Crohn's disease had a higher risk of fracture than patients with ulcerative colitis at all sites; however, after accounting for a measure of disease severity (based on number of gastrointestinal related symptoms recorded) there were no significant differences in fracture risk between patients with Crohn's disease and ulcerative colitis. The overall odds ratio for IBD patients compared with a matched control group was 1.21 (95% CI 1.1–1.32). The odds ratio for patients with at least 2 gastrointestinal symptoms was as high as 1.76 (95% CI 1.27–2.44) for patients with Crohn's disease and 1.70 (95% CI 1.32–2.19) for patients with ulcerative colitis. The increased risk of fracture did not differ between males and females. The risk of fracture increased with number of symptoms and with age.

The collective messages of these studies are consistent: patients with IBD may have increased fracture rates, but the magnitude of the excess risk is small and most evident in the elderly. This should obviously temper any enthusiasm for pursuing widespread dual-energy X-ray absorptiometry assessments in all patients. These data also suggest that fracture rates are generally similar in Crohn's disease and ulcerative colitis. Finally, although age-related osteoporosis is predominantly a female problem, it is gender neutral amongst patients with IBD.

The preceding findings are at odds with those from a study in a speciality IBD clinic in which 293 patients with Crohn's disease were screened with dual-energy X-ray absorptiometry.29 Subjects with T-scores of − 1 or lower (n = 156) underwent thoracolumbar spine X-rays with combined visual and quantitative vertebral morphometry to identify compression fractures. Of these, 34 (22%) had 63 osteoporotic vertebral fractures, the vast majority of which (88%) were asymptomatic. The average spine T-score amongst those with fractures was − 2.5 and this was slightly lower than that in the non-fracture patients (T-score of − 2.1; P < 0.025), although hip T-scores were the same. Of note, 35% of the subjects with fractures were less than 30 years of age and only 38% of the fracture cases actually had a spine BMD that was in the osteoporotic range. In retrospect, it would have been interesting if the investigators had performed spine X-rays on all subjects, including those with a normal bone density. One cautionary note is that prevalent vertebral deformities in younger individuals may be due to unrecognized previous trauma, as there is a higher prevalence of thoracic vertebral fractures in men compared with women below the age of 50 years, but thereafter the rate of increase is much greater in women.30–33 This study requires confirmation and extension to a general IBD population with inclusion of age-matched controls. The important message from this study is that clinicians should use a low threshold for obtaining spine X-rays, which still have an important place in the investigation of osteoporosis, particularly as the majority of vertebral fractures have few (if any) symptoms. Unfortunately, morphometric spine analysis is not generally available outside of a research setting, and severe under-reporting of vertebral fractures by radiologists is distressingly common.34

The findings of Klaus et al.29 are consistent with recent data showing that most fractures, including those of the spine and hip, occur in individuals who do not have a BMD in the osteoporotic range. In a large prospective cohort study of 9704 post-menopausal women followed for 8.5 years following baseline dual-energy X-ray absorptiometry, the proportion of fractures attributable to osteoporosis (based on a T-score of lower than − 2.5) was 25–39% for spine fractures, 21–28% for hip fractures and 16–26% for wrist fractures.35 A T-score cut-off of − 1.5 was only marginally better (attributable risk of 31–51% for these same sites). Although BMD is an important predictor of fractures, it is clear that other risk factors are involved,35 and can be combined to form a fracture risk index.36 Previous vertebral fracture is the strongest risk factor for future vertebral fractures (relative risk, 4.4; 95% CI, 3.6–5.4), and the risk is high even when BMD is normal.37, 38 Vertebral fractures also predict non-vertebral fractures (including hip fractures with a relative risk of 2.3; 95% CI, 2.0–2.8). Many therapies have now been shown to prevent vertebral fractures, and these are most beneficial in those with pre-existing vertebral insufficiency fractures.39

The relationship between calcium intake and bone health

Adequate calcium intake has been suggested to be an important determinant of optimal bone mass.40–42 Lactose intolerance in lactose malabsorbers has been associated with reduced calcium intake and low BMD in post-menopausal women. However, post-menopausal women with lactose malabsorption who tolerate lactose do not show alterations in calcium intake or BMD.43 Recent data in children have also shown that a low lactose diet is associated with a reduced dietary calcium intake and low bone mass.44 Lactose intolerance in women leads to sub-optimal calcium intake and slightly more fractures (odds ratio, 1.33; 95% CI, 1.09–1.62).45 In a study of 103 healthy adults aged 25–33 years, lactose malabsorption by breath testing was confirmed in 55. Lactose malabsorbers did not differ from normal absorbers in BMD or calcium intake. However, in the 29 lactose malabsorbers with symptomatic intolerance, there was significantly lower BMD and calcium intake, and calcium intake correlated directly with BMD (r = 0.69 at the spine and r = 0.73 at the hip).46

Recently, a population-based study was undertaken to determine the BMD achieved as adults amongst pre-menopausal women who developed IBD prior to 20 years of age.47 Surprisingly, the average BMD was no different from that in controls and only 3% had BMD in the osteoporotic range. Hence, despite considerable fears about bone health in children with IBD, as adults, most still achieve normal BMD. It is possible that, during periods of disease remission and/or corticosteroid ‘holidays’, there is sufficient opportunity for catch-up bone deposition with normalization in bone mass. In this same study, oral calcium and vitamin D intake were measured by a 4-day food record and no correlation was observed between the intake of these nutrients and BMD;48 this is similar to a previous report in 168 Crohn's disease patients.49

How can we reconcile what to do with oral calcium intake and fracture risk for patients with IBD? The cross-sectional study of pre-menopausal women with IBD suggests that a one-time survey to determine current calcium intake will not help to predict BMD. However, persistently reduced calcium intake does appear to lead to lower BMD. Hence, the 4-day food record can be beneficial to determine the quantity of calcium supplementation required. Furthermore, subjects with lactose intolerance usually require calcium supplementation.

Lean body mass and bone mineral density

In the general population, body mass is strongly correlated with BMD.50 The same appears to be true in Crohn's disease. In a recent population-based study of 60 subjects with Crohn's disease, 60 subjects with ulcerative colitis and 60 healthy subjects, those with Crohn's disease had both a lower BMD as well as a lower lean body mass.51 This effect was only seen in those with current or past systemic corticosteroid exposure, which is consistent with their known adverse actions on both bone and muscle tissue. Amongst 168 subjects with Crohn's disease, the body mass index was significantly correlated with BMD (P < 0.0001) on multivariate analysis.49

A previous meta-analysis has concluded that a low body weight is a significant and consistent risk factor for fractures in the general population, although the 51 studies used in this determination were dominated by older women and extrapolation to younger individuals may not be justified.35 Hence, subjects with IBD and considerable weight loss, or who are obviously malnourished, could be considered for BMD testing, and the finding of a low BMD would suggest the need for more aggressive nutritional support. The clinical experience with anorexia nervosa may be relevant, as this is predominantly a disorder of young women. Low BMD is common in anorexia nervosa, but traditional pharmacological therapies for post-menopausal osteoporosis (including hormonal products) are generally ineffective.52, 53 A recent randomized trial in 61 young women with anorexia nervosa (dehydroepiandrosterone vs. conventional hormone replacement therapy) found that, after controlling for weight gain, no treatment effect was detectable.54 It therefore remains unclear whether non-nutritional intervention is warranted for malnutrition-related osteoporosis unless this also occurs in a setting in which clinical trials have shown a beneficial role for drug treatment (e.g. steroid therapy, fragility fractures or older age).

Does corticosteroid use in Inflammatory Bowel Disease lower bone mineral density?

Data have been published both in support of and against the notion that corticosteroids are primarily at fault for the low BMD observed in IBD.8 An age-old problem is whether corticosteroids are causally related to the low BMD, as users typically have more active disease and, therefore, the role of active inflammation in lowering BMD cannot be easily distinguished from the role of corticosteroids. The confusion is compounded by the fact that most reports are derived from specialized centres and may not be representative of the general IBD population. The most recent study to assess this question selected 91 Crohn's disease patients with high risk for osteoporosis from a larger clinic population of 350 Crohn's disease patients.55 The lumbar spine BMD Z-score was inversely correlated with the cumulative dose of corticosteroids (r = − 0.24, P < 0.05), whereas the femoral neck Z-score was not related to the dose of corticosteroids. After adjustment for the body mass index and history of bowel resections, corticosteroids were no longer a factor related to spine BMD.

Recently, a population-based database of IBD was used to identify those IBD subjects with fractures, who were then matched to a cohort without fractures.56 Medication use was compared through linkage with a comprehensive population-based prescription drug database. Subjects with Crohn's disease with fractures were significantly more likely to have used corticosteroids in the 2 years prior to the fracture than those without fractures. Corticosteroid use was not more common amongst ulcerative colitis patients with fractures. These data suggest that, when Crohn's disease subjects are sufficiently ill to warrant corticosteroids, they are at increased risk for fracture.

It may be the active inflammation or other factors related to disease severity, and not necessarily corticosteroids, that lead to fractures. It is important to note that, in a large mixed general practice patient research database, those who stopped taking oral corticosteroids showed a rapid decrease in fracture risk towards baseline shortly after stopping corticosteroids.57, 58 Even very low doses of corticosteroids (prednisolone 2.5–7.5 mg/day) are associated with an increase in fracture rates, although the risk clearly increases with higher doses.59 It is not clear whether these findings relate primarily to corticosteroid use or to the underlying disease, which has obvious implications in terms of the use of bone prevention therapies. This does not diminish the importance of corticosteroid use as a marker for fracture risk.

In a non-randomized retrospective study, investigators assessed BMD in 31 female subjects with Crohn's disease treated predominantly by corticosteroids (total dose of prednisolone of at least 5 g; mean total dose of 12.9 g), 33 females treated by dietary manipulation (elemental feeds, followed by a sequential re-introduction of food groups) with a low lifetime dose of corticosteroid (mean total dose of 1.2 g) and 31 females treated by therapies other than diets who also had a low lifetime use of corticosteroids (mean total dose of 1.5 g).60 In subjects with a low lifetime dose of corticosteroids, BMD was no different from that in controls (mean hip T-scores of − 0.56 to − 0.65), whereas BMD was decreased in those who used high doses of corticosteroids (mean hip T-score of − 1.34). Although the authors suggested that the use of dietary manipulation rather than corticosteroids can positively impact on bone health, in such a non-randomized trial, it may be that patients who were more unwell with more inflammatory disease used the higher doses of corticosteroids. Once again, the reader is left to ponder the relative importance of corticosteroids vs. active inflammatory disease in the genesis of low BMD (although, surprisingly, disease activity was not significantly associated with BMD in this study). Nonetheless, as corticosteroids are known to be deleterious to bones, it is prudent to minimize their use and possibly consider enteral diets as primary therapy.

A recent study of BMD in patients who had undergone total colectomy with ileo-anal pouch anastomosis for ulcerative colitis demonstrated the potential risk posed by inflammation on BMD.61 Eighty-eight patients who had undergone a J-pouch for ulcerative colitis underwent dual-energy X-ray absorptiometry testing at least 5 years after pouch operation. Patients with villous atrophy and signs of inflammation in the pouch were significantly more likely to have low BMD scores than those with normal pouch histology, and this could not be explained by current or past corticosteroid use. In fact, none of the subjects with normal villous architecture in their pouches were osteopenic. The number of episodes of pouchitis correlated inversely with BMD. Further support for the importance of circulating inflammatory mediators comes from a study of children with newly diagnosed IBD who had not received any type of corticosteroid for 1 year. Patient serum reduced osteoblast nodule formation and osteoblast differentiation in a pure culture of rat osteoblasts.62 Hence, a certain circulating factor in these paediatric IBD patients interferes with osteoblast function, and it is plausible, but as yet unproven, that this is one of the pro-inflammatory mediators.

Do vitamin d levels predict osteopenia?

Normal vitamin D metabolism is critical in developing and maintaining normal bone and, in conjunction with adequate calcium, has been shown to prevent fractures in older individuals.39, 63, 64 The role of vitamin D in IBD is more controversial, and most data suggest that overt osteomalacia, the skeletal hallmark of vitamin D deficiency, is infrequent.8 That does not diminish the importance of vitamin D as part of a general approach to osteoporosis prevention and treatment, however. Serum levels of 25-(OH)-vitamin D in 60 subjects with Crohn's disease and 60 subjects in a speciality clinic in Norway revealed that 27% of Crohn's disease subjects and 15% of ulcerative colitis subjects had subnormal levels (< 30 nmol/L).63 Although these levels suggest that supplementation with vitamin D is necessary in a considerable number of IBD patients, there was no correlation between the levels of serum vitamin D metabolites and BMD. No healthy controls were included and these measurements were obtained during the winter months; therefore, it is unclear to what extent the normal seasonal changes in vitamin D status that occur at northern latitudes are responsible.66, 67 Finally, vitamin D intake did not correlate with BMD in a study of pre-menopausal females with IBD.48

Sex hormone status and bone mineral density in Inflammatory bowel disease

Sex steroid hormones play an important role in regulating bone turnover and bone mass in males and females. Malnutrition and chronic illness can lead to menstrual dysfunction in women with IBD,68 probably contributing to low BMD, although it is overly simplistic to believe that this is the only (or even a major) cause (see section on ‘Lean body mass and bone mineral density’). No large studies have investigated systematically the relationship between ovarian function, serum oestrogen and osteoporosis in women with IBD. One non-randomized prospective 2-year cohort study concluded that hormone replacement therapy was effective in the prevention of bone loss in post-menopausal women with IBD.69 The enthusiasm for hormone replacement therapy as prevention or treatment of osteoporosis has been tempered by the publication of data from the Women's Health Initiative showing that the risks from hormone replacement therapy (more cardiovascular, cerebrovascular, thrombotic and breast cancer events) outweigh the benefits (fewer colorectal cancers and hip fractures) in generally healthy post-menopausal women.70

Testosterone has been measured in males with IBD in two studies.71, 72 One study found low testosterone levels in 44%, but this did not correlate with BMD;72 the second study found normal testosterone levels in most males with Crohn's disease, of whom only 8% had subnormal levels.71 Three studies assessed dehydroepiandrosterone sulphate levels, and all found low levels in a high proportion of patients.72–74 There was a significant association between dehydroepiandrosterone sulphate and BMD.72, 73

An unexplored area is how serum oestrogen affects skeletal health in males with IBD. Surprisingly, recent evidence has suggested that serum oestrogen may be more important than androgens in regulating bone mass in men.75, 76 CYP19 (gene product aromatase) has recently emerged as an interesting candidate gene for osteoporosis. Aromatase has the ability to convert testosterone to estradiol and is the major source of endogenous oestrogen in post-menopausal women and men. In late post-menopausal women and men, even low serum oestrogen levels exert a restraining effect on bone turnover and may contribute to differences in bone loss.77, 78 A review concluded that genetic variations of CYP19 were important factors affecting the estradiol serum level and speculated that these may be related to the development of osteoporosis.79 In summary, the implications of low sex hormone status in males and females with IBD and the rationale for testing remain to be more fully established.

Genetics of osteoporosis

Bone mass shows strong heritability and explains up to 90% of the variance in BMD, but the major gene(s) remains elusive.80, 81 The common form of osteoporosis is currently believed to be a polygenic disorder arising from the interaction of multiple common polymorphic alleles and environmental factors.82 In IBD, there are many potential metabolic modifiers (vitamin D deficiency, sex hormone deficiency, corticosteroid use and systemic inflammation) as well as environmental modifiers (smoking, lack of physical activity and low body mass). The contribution of genetics to IBD-related osteoporosis is an area that is still largely unexplored.

In the last few years, the dominant, final mediator of osteoclastogenesis has been identified.83 The osteoprotegerin (OPG)/receptor activator for nuclear factor kappa B (RANK) ratio in binding to the osteoblast ligand RANK ligand (RANKL) is thought to determine whether osteoclastogenesis and hence osteoporosis is promoted (when the balance favours RANK) rather than osteoblastogenesis and inhibition of osteoporosis (when the balance favours OPG). A recent Danish study identified 12 polymorphisms in the OPG gene.84 Amongst the study population of 268 subjects with osteoporosis (defined by non-traumatic spinal fracture) and 327 healthy controls, three polymorphisms were found to affect vertebral fracture risk. The most interesting was the A163-G polymorphism in the promoter region, which was the most common and predicted both low BMD and increased fracture risk. Interleukin-1β (IL-1β) is a pro-inflammatory cytokine that stimulates osteoclast development and activity. Subjects with IBD who are carriers of the IL-1B-511*2 allele of the IL-1β gene, who are hypersecretors of IL-1β, have significantly lower BMD at the spine and hip than healthy controls.85 The IL-1β genotype has previously been shown to influence the course of IBD,86 suggesting that IBD and osteoporosis seen in IBD may share common genetic alterations. IL-6 is another pro-inflammatory cytokine that is considered to have a role in promoting osteoporosis,87 and is known to be over-produced in IBD.88, 89 A logical extension of these findings was to assess IL-6 genotyping and BMD in patients with IBD. However, a study in 105 subjects with IBD and 113 controls revealed no association of any IL-6 genotype (including the genotype leading to over-production) with BMD at baseline or at 2 years of follow-up.88, 90

Treatment of osteoporosis in Inflammatory bowel disease

Very few studies have specifically investigated the impact of therapies on osteoporosis in the IBD population and no study has used fractures as the primary outcome. Hence, current approaches must be extrapolated from what is known about post-menopausal and corticosteroid-induced osteoporosis.39, 91

Previous short-term studies have indicated that bisphosphonates92 and slow-release sodium fluoride93 can improve BMD. No benefit was seen after 1 year with calcium and vitamin D administration alone in one randomized study of corticosteroid-using patients with IBD.94 However, all studies to date in IBD have involved small study numbers. In a more recent randomized study, Von Tirpitz et al. randomized 84 Crohn's disease patients with reduced BMD (defined as a lumbar spine T-score of lower than − 1) to receive either calcium citrate (800 mg/day) plus vitamin D (1000 IU/day) alone (n = 13) or in conjunction with either slow-release sodium fluoride (25 mg twice daily) (n = 36) or ibandronate (1 mg intravenously every 3 months) (n = 35).95 The randomization was unbalanced as those with osteoporosis (T-score of lower than − 2.5) were only allowed to be randomized to either fluoride or ibandronate, introducing a potential bias when comparing the outcomes in these groups with those in the calcium and vitamin D only group. Additional problems were a significantly shorter average disease duration in the fluoride group and incomplete follow-up (only 81% of the study participants were available for BMD at 12 months and 65% at 27 months). In the per protocol analysis at 27 months, BMD at the lumbar spine increased by 2.2% for the group given calcium and vitamin D only, 5.7% for those given fluoride and 5.4% for those given ibandronate. A significant benefit was evident at 1 year in the sodium fluoride and ibandronate groups and thereafter there was no significant change in BMD. Femoral BMD did not change in any group and no participant suffered a new vertebral fracture. Thirty-five subjects who received corticosteroids at least once during the study showed a mean increase in BMD similar to that observed in those who did not receive corticosteroids. A significant correlation was observed between the change in body mass index and the BMD during the observation period (r = 0.46, P < 0.001), raising the possibility that diminished disease activity (seen in all groups) and improved nutrition may have contributed to the noted improvement in BMD independent of the therapeutic interventions. Of interest, serum OPG was measured every 3 months in the fluoride and ibandronate groups and decreased significantly over time. This is contrary to what might have been predicted on the basis of the anti-osteoclastic action of OPG. This finding could indicate that OPG behaves as an acute phase reactant, perhaps up-regulated in order to attenuate the adverse skeletal effects of inflammatory mediators. Other studies have found that OPG is higher in conditions associated with lower BMD, and this may indicate an insufficient counter-regulatory attempt to prevent further bone loss.96, 97

In summary, for patients with IBD and low BMD, it is possible that calcium and vitamin D supplementation alone may be sufficient unless the patient is also receiving corticosteroids or has a history of fragility fractures. Slow-release sodium fluoride and bisphosphonates will improve BMD, but it remains unproven whether these (or any intervention for that matter) actually prevent fractures. In fact, BMD may be a poor surrogate for a treatment's anti-fracture effect.98 In the case of fluoride, increased BMD does not appear to be protective against vertebral fractures and actually increases the number of non-vertebral fractures.39 Recent guidelines have cautioned against the use of fluoride.99 Teriparatide, human recombinant parathyroid hormone-(1-34), is proving to be an exciting new development in the pursuit of a safe and effective skeletal anabolic agent. This agent produces large gains in BMD and favourable changes in bone geometry.100, 101 The bone that is produced is structurally sound, which translates into a reduction in vertebral and non-vertebral fractures in post-menopausal osteoporosis.102 Combination anti-resorptive therapy (typically a bisphosphonate combined with oestrogen or raloxifene) may produce greater gains in bone mass than either agent alone, but the use of two anti-resorptive agents is not recommended because the benefit on fracture risk has not been demonstrated and there is increased cost and side-effects.103 Simultaneous treatment with a bisphosphonate may actually interfere with the anabolic action of human recombinant parathyroid hormone-(1-34).104 Clearly, there is much left to learn about how and when to use this agent, and it should only be given under the direction of an osteoporosis specialist.

The level of reduced BMD at which intervention should be undertaken is still unknown. It is unlikely that any single cut-off will be appropriate for all situations, as age, current corticosteroid use and the existence of previous fragility fractures strongly affect fracture risk (independent of BMD) and should shift the intervention point. A shift from an arbitrary BMD T-score as the basis for intervention to the absolute fracture risk will probably be the next phase in the evolution of osteoporosis management.105

Acknowledgements

Charles N. Bernstein is supported in part by a Canadian Institutes of Health Research Investigator Award and by the Crohn's and Colitis Foundation of Canada Research Scientist Award.

Ancillary