Another Look at Fracture Risk in IBD

Hip fractures in patients with inflammatory bowel disease and their relationship to corticosteroid use. Card T, West J, Hubbard R, Logan RFA. Gut 2004;53:251–255. This study used the General Practice Research Database (GPRD) of the UK, which samples approximately 6% of the general practitioners’ practices in the UK and facilitates creation of longitudinal medical histories of the subjects in the database and identification of certain morbidities within the database as identified by ICD-9 codes. All patients with inflammatory bowel disease (IBD) within the database (n = 16550) were identified and a matched control group, matching 5:1 by age and gender from within the GPRD was created. The unadjusted relative risk of hip fracture in the IBD group compared to the control group was 1.62 (95% confidence interval, C.I., 1.24–2.11). In ulcerative colitis (UC) the relative risk was 1.49 (95% C.I. 1.04–2.15), and in Crohn’s disease was 2.08 (95% C.I. 1.36–3.18). After adjusting for confounding the relative risk was 1.41 (95% C.I. 0.94–2.11) for UC and for Crohn’s disease was 1.68 (95% C.I. 1.01–2.78). The increased risk was evident in females only. Multivariate modeling showed that both current and cumulative use corticosteroids and the use of opioid analgesics confounded this relationship. The relationship between corticosteroid use and fracture was nonlinear. The only statistically significant risk was for users of 1–5 courses (relative risk = 1.92, 95% CI 1.05–3.52), whereas use of more courses was not significantly associated with an increased risk. The authors concluded that the risk for hip fracture was significantly increased in Crohn’s disease and corticosteroid use contributed to this risk. For those without any corticosteroid use among incident cases of IBD the hazard ratio for hip fracture was 1.8 (95% C.I. 1.06–3.08). Hip fracture was rare in both IBD patients and controls less than 60 years of age. Overall these authors found that while corticosteroids contributed to the excess risk of fracture among patients with IBD, they only accounted for less than half of the excess fracture risk in Crohn’s disease patients and in less than 20% of ulcerative colitis patients. Comments: This population-based general practice database has provided a unique tool to explore disease related morbidity and drug and demographic associations in the UK. The strength of any population-based database lies in the validity of the diagnoses under study. This group has reported an accuracy for diagnosis of Crohn’s disease and UC of over 90% (Pharmacoepidemiol Drug Saf 2002;11:211–8) and a similarly high accuracy for hip fracture identification (Pharmacoepidemiol Drug Saf 2000;9:359–66). The validity of hip fracture diagnosis from administrative data has been assessed in both Baltimore and Manitoba hospitals and found to be highly accurate (J Am Geriatr Soc 1998; 46:745–750, Soc Sci Med 1989;28:175–82). The GPRD collects prescription data prospectively, however it is unclear whether the data reflect prescriptions written by physicians or actually filled by patients (and presumably taken). Nonetheless, the GPRD has developed a lengthy track record of work using its database. The assessment of fracture risk in IBD has been explored in 3 previous population-based studies. The first published report from Manitoba utilized the population-based University of Manitoba Inflammatory Bowel Disease Epidemiology Database a validated database of all Manitobans with IBD (Am J Epidemiol 1999;149:916–924) and an age, gender and geographically matched cohort drawn from the administrative databases of Manitoba Health, the single provincial health insurer. The reported relative risk was 1.41 (95% CI, 1.27–1.56), with a comparable risk for Crohn’s disease as for UC. The most notable association in this study was with advanced age, with the highest incidence rate of fractures in the Manitoba IBD population occurring amongst subjects over 60 years of age. The relative risk for hip fractures, specifically, was 1.47 (95% CI, 1.03–2.10) in Crohn’s disease and 1.69 (95% CI 1.26–2.28) in UC with an overall relative risk for IBD of 1.59 (95% CI 1.27–2.00). In this study males had a similar increased risk to females. In a later study these researchers created another matched case control study whereby they linked their IBD database with a comprehensive provincial prescription drug database and explored the use of corticosteroids among subjects with fractures compared to age, gender, disease duration and geographically matched IBD subjects without fractures (Am J Gastroenterol 2003;98:1797–1801). This latter study found that Crohn’s disease subjects who fractured were significantly more likely to be using corticosteroids than those who did not fracture. However, this relationship did not hold up in UC. The second population-based IBD fracture study was from the Olmsted County database, a population-based research registry directed by researchers at Mayo Clinic with chart-validated diagnoses. The IBD database is small with several hundred IBD patients and may not have had the power to show an increased fracture risk in IBD. It did show a trend in that regard with an increased fracture risk in the elderly. The hazard ratio per 10-year increase in age was 1.3 (95% CI, 1.1–1.5). The third study was from Denmark using the Danish national administrative health databases and hospital discharge abstracts (Am J Epidemiol 2002;156:1–10). Studies relying on hospital discharge abstracts will not capture outpatient fracture diagnoses and are limited to severe fractures such as those of the hip or incapacitating spinal fractures. IBD subjects and age- and gender-matched controls randomly selected from the population were compared. The authors identified over 15000 cases of IBD, 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 UC it was not significantly increased at 1.08 (95% CI, 0.97–1.20). The study from GPRD in the UK adds to the growing body of evidence that fractures are increased in IBD but the risk relative to the general population is only mildly increased. There is an added risk to the elderly, a group at risk for osteoporosis and fracture regardless of whether they have IBD. Hence clinicians should be extra vigilant with elderly IBD patients at assessing their bone health and attempting to control modifiable risk factors. The GPRD study like the Manitoba study suggests an association between corticosteroid use and fracture in Crohn’s disease with a lesser or non-association between corticosteroids and fracture in UC. Neither the GPRD or Manitoba studies can resolve the dilemma as to whether the corticosteroid associated fracture risk is secondary to the drugs or secondary to the underlying inflammatory condition being treated. In fact the increased risk for fractures in the GPRD study that correlated with opioid use may be a proxy measure of active disease. An interesting interleukin-2 knockout mouse model of colitis and osteopenia has underscored the complex interaction between osteoprotegerin, and receptor activator of nuclear factor kappa-b, regulators of bone health that interact with T cells and colon dendritic cells, and intestinal inflammation (Immunity 2003;19:849–861). The therapeutic message for clinicians is to be concerned about the bone health of patients with active Crohn’s disease who are using corticosteroids, but it is less clear whether low dose corticosteroids in patients in remission poses a bone health risk and even less clear as to whether the use of corticosteroids in active UC poses any additional risk for fracture. The mildly increased fracture risk among IBD patients is at odds with findings from a specialty IBD clinic study in which 293 patients with Crohn’s disease were screened with dual energy x-ray abosrptiometry (DXA) (Gut 2002;61:654–58). Subjects with T scores <−1 underwent thoracolumbar spine x-rays with combined visual and quantitative vertebral morphometry to identify compression fractures. Of these, 22% had osteoporotic vertebral fractures, the vast majority of which (88%) were asymptomatic. Of note, 35% of the subjects with fractures were less than 30 years old and only 38% of the fracture cases actually had spine bone density that was in the osteoporotic range. In retrospect, it would have been interesting if the investigators had done spine x-rays on all subjects including those with normal bone density. 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 have 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. This study also suggests that population-based studies may have underestimated the risk for spine fractures in IBD patients since many may be asymptomatic or even symptomatic but not diagnosed. These findings (Gut 2002;61:654–58) are consistent with recent data showing that most fractures, including those of the spine and hip, occur in individuals who do not have DXA bone density in the osteoporotic range. In a large prospective cohort study of 9,704 postmenopausal women followed for 8.5 years following baseline DXA, the proportion of fractures attributable to osteoporosis (based on a T-score <−2.5) was 25–39% for spine fractures, 21–28% for hip fractures and 16–26% for wrist fractures (J Bone Miner Res 2003;18:1947–1954). Prior 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 bone density is normal (Ann Intern Med 1991;114:919–23, J Bone Miner Res 2000;15:721–39). Vertebral fractures also predict non-vertebral fractures (including hip fractures with relative risk = 2.3, 95% CI 2.0–2.8). In the recent AGA technical review and position paper on the subject of osteoporosis in gastrointestinal diseases it was recommended to be selective with DXA testing in IBD patients (Gastroenterology 2003;124:795–841). This review encouraged clinicians to be aware that DXA testing alone could not fully predict fracture risk and hence should not be used as the sole determining factor in whether patients with IBD should start more advanced osteoporosis related therapy (beyond calcium and vitamin D supplementation). It may be that spine x-rays or spine morphometry will have an equally if not greater role in identifying patients either at risk or with underlying osteoporosis/fractures who truly require intervention. Fractures, including high morbidity ones like hip fractures, are mildly increased in IBD patients, particularly the elderly with IBD. The GPRD study adds important information to the growing body of evidence in this regard. Strategies to assess fracture risk need to include a rational approach to DXA testing, a complete medical history that includes past fractures and an evolution toward including routine spine assessment in patients at increased risk.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.