Several reports on the bone mineral density (BMD) of adults with IBD have been published. In the vast majority, the investigators used dual-energy x-ray absorptiometry (DXA), the current gold standard for evaluating BMD, but the number of patients included in each study was rather small. Both controlled9-19 and uncontrolled1,20–31 cross-sectional studies in IBD or in patients with Crohn's disease (CD) only32–41 produced a wide variation in prevalence rates. The prevalence of significantly low BMD, defined by a T score of less than −2.5 or a z score of less than −2.0, has been reported to be as high as 59%.42 However, the overall prevalence of osteoporosis is estimated at ≈15% and is strongly affected by age. Therefore, the effects of IBD on BMD are considered modest. In newly diagnosed IBD patients, the prevalence of reduced BMD is low,17,19,22,29 and longitudinal BMD changes are not excessive.15,22,25,32,35,43–46
BMD is affected to the same extent in both male and female IBD patients, and between them, females exhibit slightly higher T and z scores.1,4,13-15,18,20,22,24-27,29,32,42 There is no consistent pattern of low BMD found exclusively in the spine or in the hip. However, a number of studies reported that the hip could be affected more frequently than the spine.9,14,16,26,28,32,34,36 Generally, in some studies, both osteopenia and osteoporosis (as determined by the T score) have been observed with similar frequencies in CD and ulcerative colitis (UC); 9,16,17,25,26,28,43 other studies suggested that BMD may be lower in CD.10,13,15,22,24 Disease duration has not been established as a significant risk factor for low BMD because some of the studies reported no effects,9,14,15,18,32,35,46,47 other indicated a positive relationship between longer disease duration and lower BMD,13,16,17,27,28,42 and 1 study showed a markedly shorter duration in female patients with sacroiliac involvement.41 In addition, disease activity had no effect on BMD according to findings from some studies,16,22,33 but a study in 137 patients (64 with UC, 73 with CD) reported that age-matched (Z score) BMD was higher with increasing duration of disease remission.48 Furthermore, in 47% of 34 patients with a history of active IBD, biochemical bone markers suggested increased bone degradation.27 Disease site may contribute to low BMD because there is a report of lower BMD in CD patients with jejunal disease,33 although other studies failed to show such effects.9,10,14,15,22,36,43 Previous small bowel surgery, even after taking into account the length of the resection, is not considered a significant risk factor, according to most of the published studies.9,10,14,16,21,23,26,32 However, 2 studies assessing patients who underwent ileal resection,42,49 a cross-sectional analysis of 117 patients with CD and partial small bowel resection,33 and 2 more studies1,49 suggested otherwise. Actually, van Hogezand et al49 reported that ileum resection is the most predictive factor for osteoporosis in patients with CD. Interestingly, colectomy has been reported to stabilize or improve BMD.21,25,51,52 Young age of onset of IBD initially had been reported to be associated with lower BMD,34 but a subsequent study failed to confirm this claim.30
Hip fractures are documented more reliably than spinal fractures, the vast majority of which are not clinically evident. In small case series, the reported incidence of new fractures varies considerably, from no difference up to 27% greater in IBD patients compared with the general population.1,10,23,34,46 However, findings of 3 North American population-based studies,53–55 coupled with 2 Danish studies with a respectable number of patients, failed to establish that the risk of fracture in IBD patients is increased.56,57
In more detail, the largest study was conducted in the Canadian province of Manitoba, where comprehensive healthcare coverage is provided for all residents.53 Analysis of data from 6027 IBD patients and an age-, gender-, and geographic residence-matched control group of 60,270 individuals showed a small increased risk of fractures (relative risk [RR] 1.41; and 95% CI 1.27–1.56). These results were confirmed by 2 population-based studies that are smaller (243 CD patients, 273 UC patients)54,55 but involve a relatively homogeneous population from the Olmsted County (Minn) database. In addition, these North American population-based studies confirmed once again that age is an independent risk factor for fractures.
In a nationwide follow-up study of 16,416 patients in Denmark that included patients with CD, UC, and celiac disease,56 CD was associated with a minor increase in overall fracture risk compared with UC and the general population. The second Danish study was based on a survey mailed to members of the Danish Crohn's/Colitis Association.57 The overall fracture rate in UC was similar to that of control subjects, but the RR for CD patients was increased at 1.7 (95% CI 1.7–2.3), with female patients at slightly higher risk (RR 2.5). Vertebral fractures were more common in CD patients (RR 6.7; 95% CI 2.1–21.7), and a similar trend was identified in UC patients (RR 2.4; 95% CI 0.5–11.9). Furthermore, the risk of femur fractures was similar to that of the control subjects. However, these findings should be interpreted with caution because poor matching of control subjects and failure to take corticosteroid use into account may have biased the results.
In the most recently published population-based cohort study,58 subjects within the General Practice Research Database in the United Kingdom with a diagnosis of IBD were matched with up to 5 control subjects for each patient. Seventy-two hip fractures were recorded in 16,550 IBD cases and 223 in 82,917 control subjects, with the rate of hip fracture being increased ≈60% in IBD patients. The risk was 1.5-fold higher in CD compared with UC patients. After the IBD cases were subdivided by disease and after correction for confounding (age, sex, corticosteroid use [both current and cumulative], and opioid use), the hazard ratio remained greater for CD at 1.68 (95% CI 1.01–2.78) than for UC (1.41; 95% CI 0.94–2.11). Therefore, the authors concluded that most hip fracture risk in IBD (>50% in CD and >80% in UC) patients cannot be attributed to steroid use.
Using data from the same database, another group of investigators published a case-control study.59 A total of 2130 IBD patients were among 231,778 case patients with a history of fracture plus an equal number of age- and sex-matched control subjects without recorded fractures. A total of 1134 of the patients with a history of fracture had a diagnosis of IBD compared with 896 IBD patients of the control group (no history of fracture). Furthermore, the risk of hip fracture was increased by 86% in patients with CD and by 40% in patients with UC. The adjusted odds ratio (OR) for fractures at all sites was estimated at 1.21 (95% CI 1.10–1.32). This increased risk was attributed to a combination of disease activity and use of oral corticosteroids.
In summary, the current understanding is that the overall risk of fracture may be slightly increased in IBD patients. Furthermore, risks for fracture are comparable among patients with CD and UC and among male and female IBD patients.
In general, bone mass in adults >30 years old reflects the bone mass accumulated during growth minus any bone mass lost since the adult peak was attained. In the general and the IBD populations, the risk factors for developing low bone mass, characterized either as osteoporosis or osteopenia, for both men and women can be separated into 2 groups: those that are measurable and those that are modifiable. Measurable risk factors include BMD, serological markers and urinary markers of bone formation and resorption, age, and genotype. Modifiable factors include glucocorticoid therapy, treatment with drugs that could affect bone metabolism, sex hormone and vitamin D status, exercise status, nutritional status, including dietary calcium intake, weight, smoking status, and risks for trauma/fall. The following details the disease states and conditions associated with low bone mineral mass:
Parathyroid hormone excess
Primary and secondary thyroxin excess, endogenous and exogenous
Cortisol excess, endogenous and exogenous
Estrogen deficiency (premenopausal state may be linked with anorexia, bulimia, athletic amenorrhea, premature menopause, prolactinoma, or hypopituitarism)
Estrogen deficiency, postmenopausal
Testosterone deficiency, primary and secondary testicular failure
Vitamin D metabolite deficiency, inadequate intake, or malabsorption
Miscellaneous (not necessarily mediated by hormonal abnormalities):
Medical conditions, including postgastrectomy states, idiopathic hypercalciurie, systemic mastocytosis, and prolonged immobilization (paraplegia and quadriplegia)
Lifestyle factors, including cigarette smoking, excessive ingestion of caffeine, and excessive sodium intake (promotes hypercalciuria)
Most of these conditions/states associated with low bone mass are not influenced by changes in diet or lifestyle. Each of the circumstances listed could have an impact on the skeleton at any time in life. The effect is aggravated if it occurs in conjunction with low estrogen levels after menopause. Furthermore, as with low BMD, risk factors for fracture may be considered modifiable or nonmodifiable, as follows:
Below-normal body weight (especially those with low body weight <127 lb)
Premenopausal amenorrhea for >1 year
Lifelong low intake of calcium
Excessive consumption of ethanol
Visual impairment despite adequate correction (may increase risk of falling)
Inadequate physical activity
Frailty or poor overall physical condition
Nonmodifiable risk factors:
History of fracture as an adult
Family history of fractures, especially among first-degree relatives
Frailty or poor overall physical condition
A strong genetic component is apparent in peak bone mass.60–62 In addition, diet (calcium and protein intake) and exercise contribute to maximum bone mass. Children and adolescents with disease states or conditions that interfere with growth (including sexual maturation), nutrition, and exercise typically have suboptimal bone mass. In women, BMD is stable from the mid 20s to the earlier stages of climacteric and then declines, initially sharply, as estrogen production falls. In men, there is a steady slow decline probably after age 50. Although the priming and mechanism of bone loss are not as extensively studied in men as in women, it is known that estrogen also plays an important role in bone regulation and maturation in males.63
Different risk for bone loss has been reported in people of different ethnic backgrounds. White women have been identified as the racial group with the highest risk. Black women have higher BMD than white non-Hispanic women throughout life and experience lower hip fracture rates. Japanese women have lower peak BMD than white non-Hispanic women but have a lower hip fracture rate, the reasons for which are not fully understood. Mexican American women have bone densities intermediate between those of white non-Hispanic women and black women. Limited available information on Native American women suggests that they have lower BMD than white non-Hispanic women.3
Low bone mass is the most important measurable predictor of fragility fractures. In animal models, ≈80% of the variance in bone strength and resistance to fracture could be explained by bone mineral content per cubic centimeter. However, population-attributable risk for osteoporosis and fracture among older women is modest, ranging from ≈15% for all types of fractures to <10% to 44% for specific types of fractures.64
A meta-analysis of large, well-designed prospective studies confirmed the relationship between BMD and fracture risk in which a decrease in BMD is associated with an increased risk of fracture65 Site-specific BMD measurements could more accurately predict fracture risk because the proportion of trabecular versus cortical bone varies in the different parts of the skeleton. Therefore, it is not surprising that the correlation between femoral neck and lumbar spine BMD ranges between 0.5 and 0.7. It is now known that postmenopausal women in the lowest quartile of bone mass at the femoral neck have the highest incidence of hip fractures. Furthermore, the higher bone mass is related to a lower risk of fracture. These effects were less obvious at sites other than the hip. The same data have historically been extrapolated to the setting of IBD because only limited information is available in patients with CD or UC.
The overall risk of osteoporosis-related fractures in IBD, as in otherwise healthy postmenopausal women, is not entirely reflected by BMD measurements alone. Patients with IBD and normal BMD still may be at increased risk for fractures on the basis of other risk factors.
BMD is determined by genetic and environmental factors. Genetic factors are known to play a role in the pathogenesis of osteoporosis.60 Epidemiological and twin studies suggest that heritable factors account for 65% to 92% of the variability in BMD. Several candidate genes thought to play a role in determining BMD include tumor necrosis factor (TNF) receptor, the interleukin (IL)-1 receptor antagonist gene, and more recently the IL-6 gene.66 It is well established that IL-1 and IL-6 have a central role in the paracrine stimulation of osteoclast development and regulation and the process of bone resorption.67 Enhanced expression of these cytokines in immune-mediated diseases such as IBD may be important. A recent study identified the lack of a 240-bp allele of the IL-1 receptor antagonist gene and the presence of a 130-bp allele of IL-6 as independent variables associated with bone loss in IBD patients.66 However, there is no consistent association between all of these reported polymorphisms and phenotypes of bone mass across studies. The most convincing finding so far is the identification in normal healthy individuals of a gain of function mutation in the LDL receptor-related protein 5 (LRP5) gene that results in the autosomal dominant high-bone-mass trait68 and a loss-of-function mutation that maps to the same genomic region that contains LRP5 and causes the osteoporosis pseudoglioma syndrome.69 Therefore, Wnt-mediated signaling via LRP5 affects bone accrual during growth and is important for the establishment of peak bone mass.
Inflammation has now moved to the center of the pathophysiological mechanisms involved in the process of bone loss in IBD. Transmural inflammation limited to the large intestine of an experimental model of IBD caused rapid and substantial (33% compared with age-matched, pair-fed control animals) cancellous bone loss by reducing the formation rate to <30% of that in control animals. Furthermore, the bone volume returned to control levels after the colitis healed.70 In a bone organ culture system, serum from children with active UC reduced the noncollagen protein synthesis.71 In the same system, serum from untreated children with CD caused disorganization of mineral and osteoid and morphologically abnormal osteoblasts; it also decreased bone dry weight and calcium content.72
Cytokine profiles specific to IBD have been well identified. In CD, the principal cytokines released by the inflammatory cells of the intestine are TNF-α, interferon-γ (IFN-γ), and IL-6.73 These cytokines, particularly TNF-α and IL-6, stimulate osteoclast activity, leading to increased bone resorption and resulting in net bone loss. Recent studies reported improvement of bone turnover markers74,75 and BMD76,77 in CD patients treated with infliximab, an anti-TNF-α antibody. In contrast, the predominant cytokines in UC include IL-4 and IL-10. These are not known to be significantly involved in bone remodeling. In the colon, however, there is an obligatory loss of calcium into the lumen, which is why “net calcium balance” is used in relation to the intestine: balance = absorption − colonic loss.78 This loss could be exaggerated in the inflamed colon and prevented in cases of colectomy. This could explain the positive effects of ileoanal anastomosis on BMD.51
Accumulating evidence suggests that TNF and TNF receptor-related superfamily proteins mediate basic immunological and bone remodeling functions by directly participating in signaling pathways for cell proliferation, differentiation, and survival.79,80 Some of the receptors and ligands of this superfamily are upregulated in IBD.81–83 Activated T cells produce receptor activator of nuclear factor-kB ligand (RANKL), which, through RANK, its receptor, plays an integral role in osteoclast differentiation and activation.84 Osteoprotegerin (OPG), a soluble RANKL decoy receptor, counteracts these effects.85 Indeed, treatment with recombinant human OPG prevented or reversed bone loss in a mouse IBD model.86
Glucocorticoid use, a known cause of bone loss, remains commonplace in the treatment of IBD, but the true incidence of osteoporosis in corticosteroid-treated IBD patients is unknown. It is estimated that one fourth to one half of patients on long-term glucocorticoids will experience bone fractures.87 Glucocorticoids are known mainly to reduce bone formation and, to a lesser extent, enhance bone resorption by means of several mechanisms.88,89 They may inhibit osteoblast maturation and osteoblast bone-forming ability. They may decrease gonadotropin-releasing hormone, leading to decreased estrogen and androgen concentrations. Glucocorticoids also are known to suppress circulating estrogen, thus reducing its role in inhibiting IL-6, a stimulator of osteoclast activity. In men, glucocorticoids have been reported to suppress serum testosterone concentrations, leading to a similar effect on bone. Glucocorticoids also could inhibit intestinal absorption of calcium and increase urinary calcium losses. Together, these effects may lead to negative calcium balance and secondary hyperparathyroidism.
When corticosteroids are used, trabecular bone loss occurs early in the course of therapy; however, both trabecular loss and cortical loss occur over time. The rate of bone loss is greatest during the first 6 months of therapy, and bone loss could be identifiable by 6 months by a DXA scan; up to 15% of bone mass could be lost within the first year of therapy. Furthermore, the increase in fracture risk after oral corticosteroid therapy is begun could be rapid, with significant increases in risk of nonvertebral fractures becoming apparent within the first 3 months.90 During the second year of treatment, bone loss continues, but at a slower rate. Recovery after discontinuation of corticosteroids may occur, although it seems to be related to the dose of corticosteroid used and the duration of treatment.88,89
Cumulative glucocorticoid dose has been demonstrated to be inversely associated with BMD based on data from a number of studies,1,9,10,15,16,18,21,23,24,27,33,35,36,43,45 whereas only a few studies have reported no effect on BMD.22,26,28,32 Significant bone loss has been thought to occur when the daily dose of prednisolone exceeds 7.5 mg. When normal individuals are given a high dose of prednisolone (40 mg/d) for 1 week, significant increases in urinary hydroxyproline and calcium excretion could occur, suggesting an increase in bone resorption.91 Furthermore, daily administration of relatively low-dose glucocorticoid (10 mg prednisone) could have significant negative skeletal effects. Indeed, after 7 months of treatment, iliac crest bone biopsies performed in 10 healthy volunteers (5 male, 5 female) demonstrated a 34.3% reduction in trabecular bone volume from baseline values.92 In a recent large retrospective study involving 244,235 oral corticosteroid users and 244,235 control patients, an assessment of fracture risk for dose of corticosteroid use was attempted.93 There was an increased risk of fractures during oral corticosteroid treatment, with greater effects on the vertebral body (RR 2.60) and hip (RR 1.61) than on the forearm (RR 1.09). A dose dependence of fracture risk was observed. With a daily dose of prednisolone <2.5 mg, the RR for hip fracture was 0.99, with an RR of 1.77 at 2.5 to 7.5 mg, and of 2.27 for doses >7.5 mg daily. For the vertebral fractures, the RR was 1.55 for <2.5 mg/d, 2.59 for 2.5 to 7.5 mg/d, and 5.18 for >7.5 mg/d prednisolone. The fracture risks decreased toward baseline once corticosteroids were stopped.93 In addition, the relationship between use of corticosteroids and fracture risk was estimated in a meta-analysis of data from 7 cohort studies of ≈42,000 men and women. The authors concluded that prior and current exposure to corticosteroids confers an increased risk of fracture that is of substantial importance beyond that explained by the measurement of BMD.94 Finally, a case-control (3 controls for each case), large, community-based study in Denmark assessed all subjects with any fracture sustained during the year 2000 (n = 124,655) and concluded that >2.5 mg/d oral prednisolone (or equivalent) is associated with an increase in fracture risk.95
Vitamin D deficiency has been estimated to occur in 30% to 60% of patients with CD.27,31,40,50,96–99 Factors associated with the development of such deficiency include decreased dairy product intake (supplemented with vitamin D), malabsorption of vitamin D as a result of small bowel disease or short gut syndrome, and bacterial overgrowth, or use of cholestyramine with subsequent steatorrhea. A small cross-sectional study (152 unselected IBD patients, 73 healthy controls) concluded that in IBD patients the calcium intake is not associated with BMD.