DO ALL PATIENTS WITH CELIAC DISEASE AND INFLAMMATORY BOWEL DISEASE NEED DUAL-ENERGY X-RAY ABSORPTIOMETRY?

In this study, the incidence of fracture was analyzed among patients living in Denmark and diagnosed with celiac disease (n = 1,021), Crohn's disease (n = 7,072), and ulcerative colitis (UC; n = 8,323) between 1983 and 1996. Disease and fracture diagnosis were identified by ICD-8 and ICD-10 codes within the National Patient Discharge Register, a computer-based registry of all admissions to and discharges from Danish hospitals. The Register covered only inpatient admissions until 1994 and has covered both inpatient and outpatient admissions since 1995. Each patient has a unique identifying number within the Register, allowing for longitudinal tracking within the hospital system. Fractures were captured by ICD codes, and only fractures identified in the hospital-based registry were captured. Each patient was matched by age and gender with three controls from the general population. The investigators assessed fracture risk prior to the diagnosis (accounting for patients who may have undiagnosed disease but are still at risk for fracture) and after the diagnosis. In celiac disease, there was no increased risk of fracture either before or after the diagnosis. In Crohn's disease, the fracture risk was increased before diagnosis (incidence rate ratio [IRR], 1.15; 95% confidence interval [CI], 1.00–1.32) and after diagnosis (IRR, 1.19; 95% CI, 1.06–1.33). Fracture risk was not increased in UC, except for a 5-year period just after diagnosis. Increasing age increased the risk of sustaining a new fracture after diagnosis for all three diseases, whereas having a fracture before diagnosis increased the risk of sustaining a new fracture after diagnosis in Crohn's disease and in UC. The age at first fracture after diagnosis was significantly older for females than males for all three diseases. The highest increase in IRR compared with controls for Crohn's disease was for spine, rib, and pelvis fractures (1.87; 95% CI, 1.24–2.82) and for UC was for hand and finger fractures (1.31; 95% CI, 1.01–1.71), but for celiac disease, no category of fracture was found to be increased compared with controls. Although there has been much written over the past decade on osteoporosis and gastrointestinal disorders, most studies have relied upon dual-energy x-ray absorptiometry (DEXA) data to determine bone mineral density (BMD) and thereby estimate fracture risk. It is known from the literature on postmenopausal osteoporosis that BMD, as measured by DEXA, correlates with fracture risk (BMJ 1996;312:1254–9); however, studies quantifying fracture risk using DEXA in gastrointestinal disorders are lacking. In fact, there are few reports of fracture incidence in gastrointestinal disorders. Nonetheless, gastroenterologists have begun ordering DEXA tests on most patients thought to be at some risk for osteoporosis, such as those with celiac disease and inflammatory bowel disease (IBD), and even prescribing bisphosphonates when DEXA results suggest osteopenia. No other population-based fracture risk data in celiac disease exist in the literature. In the past 2 years, data have emerged regarding fracture risk in IBD. For clinicians who only occasionally encounter fractures in these patients, the results of the fracture incidence studies, including this one by Vestergaard and Mosekilde, are reassuring. For clinicians who have adopted the “DEXA for all” approach and who will institute treatment, including treatment with bisphosphonates at T scores of less than −1 (osteopenia), the results have been confusing. In general, the data, particularly for IBD, suggest that fracture rates are low. There are several methodological issues that warrant discussion regarding the Vestergaard and Moskilde study. The first is regarding patient accrual. The authors suggest that all patients with celiac disease or IBD are hospitalized early in their disease to undergo diagnostic testing. If true, this would substantiate that it is a population-based registry at least to include all disease cases. However, considering universal approaches at cost saving in health care systems, it seems reasonable to question whether patients diagnosed in the 1990s were all hospitalized to facilitate diagnoses. Hence, it is possible that this system did not capture all cases. Nonetheless, whether it captured all cases, the data could still produce relative risks for fracture of the cases that were captured compared with controls. Arguably, the more severely affected patients would be hospitalized rather than solely seen as outpatients, and, if disease activity does in fact correlate with reduced BMD, any relative risks for fracture might have erred toward overestimates. Independent of capture of all celiac disease and IBD cases, there were likely fracture instances that were not captured, if relying only on hospital discharge diagnoses of fracture. Undoubtedly, this approach would capture all hip fractures, but spine fractures (including some that are asymptomatic or at least undiagnosed) and other appendicular fractures would be treated without admission to a hospital (or without even necessarily being seen in a hospital outpatient department). This may have led to an overestimation of fractures among IBD patients if it could be argued that inpatients may have more fractures picked up (particularly spine fractures) since they may be more likely to have x-rays than the general population. Another problem with the study population was in the case validation. For celiac disease, only nine cases were reviewed, and the validation rate was 78%. For Crohn's disease, 19 cases were reviewed, and the validation rate was 95%; however, for UC, 22 cases were reviewed, and the validation rate was only 64%. Including non-UC cases (such as infectious colitis) as UC cases may have led to an underestimation of the fracture risk in UC. These same investigators have previously reported on a survey mailed to members of the Danish Crohn's/Colitis Association regarding fractures (Gut 2000;46:176–81). The controls were not well matched in that study, and that study suffers from the potential bias that those with more active disease, or even who have sustained fractures, may be more likely to respond to mailed surveys that include questions about fracture. UC patients had an overall fracture rate similar to controls. However, Crohn's disease had a relative risk for all fractures of 1.7 (95% CI, 1.7–2.3), for fractures among females of 2.5 (95% CI, 1.7–3.6), and for fractures among premenopausal females of 2.9 (95% CI, 1.8–4.8). Thus, the population-based data of this later study (Am J Epidemiol 2002;156:1–10) corroborate the increased risk in Crohn's disease found in the mailed survey study. The IRRs for all three diseases in this study (Am J Epidemiol 2002;156:1–10) were low and should be interpreted in the context of the Danish population. If either because of genetics or fitness, Danes have lower fracture rates than other Western populations, particularly North American ones, then the Danish fracture risk may not apply elsewhere. There have been two population-based studies of fracture risk in IBD from North America (Ann Intern Med 2000;133:795–9;Gastroenterology 2002;123:468–75). The largest study reported on 6,027 IBD patients and an age-, gender-, and geographic residence–matched control group of 60,270 (Ann Intern Med 2000;133:795–9). IBD patients were identified in the population-based University of Manitoba IBD Epidemiology Database (Am J Epidemiol 1999;149:916–24). Fractures were identified using the administrative databases for the Canadian province of Manitoba, which provides comprehensive health care coverage for all residents, and controls were identified using these data as well. The overall fracture rate for IBD patients was found to be approximately 1 per 100 patient-years, and this was increased compared with controls of (RR, 1.41; 95% CI, 1.27–1.56). An increased fracture risk at the spine and hip was mostly accounted for by fractures in patients over 60 years of age. This study reported no differences between males and females or between Crohn's disease and UC. The second study used the Olmsted County population-based database of 243 Crohn's disease patients (Gastroenterology 2002;123:468–75). Compared with controls, the overall risk ratio among Crohn's disease patients 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 not greater than that of the general population, except in the elderly. This study is of a small sample size of a relatively homogenous population in Olmsted County. However, the main message is similar to that of the Manitoba study, that is, IBD patients may have an increased risk of fracture over matched control populations, but it is a small increased risk, and the greatest increased risk is evident in the elderly with IBD. In summary, Crohn's disease is associated with a small, statistically significant increased risk of fractures, and, at least in Manitoba, UC is, as well. All studies show a correlation between increased fracture risk and age. Although the risk is increased, it is a small increase, and this should temper our enthusiasm for considering DEXA tests for all IBD patients and interventions that are unproven in IBD, particularly for BMD results that are not in the osteoporotic range. Although a T score of −2 to −1 is considered osteopenic and in the postmenopausal population does correlate with an increased fracture risk over those with higher T scores, it is important to recognize that this T-score range is in the lower end of the normal population curve. Since it appears that only a fraction of IBD patients will sustain fractures, perhaps gastroenterologists should attempt to stratify whom to screen with DEXA testing (and therefore ultimately whom to treat). All three population-based studies suggest an increased risk with aging. In the Manitoba population, this increased risk was among persons over the age of 60. Recent Manitoba data have shown that premenopausal females diagnosed with IBD below age 20 have generally normal BMD as adults (Gastroenterology 2002;122:A-607). Thus, perhaps gastroenterologists should focus their screening efforts on elderly IBD patients. The Danish study has also reported an increased risk of postdiagnosis fracture if patients sustained prediagnosis fractures. Thus, fracture history should be incorporated into the general gastroenterology history of IBD (and celiac disease) patients. Other risk factors, such as parental history of osteoporosis, smoking, corticosteroid use, malnutrition, disease activity status, and sex hormone status, should also be considered in the algorithm of whom to screen with DEXA. In conclusion, despite the potential flaws in design, the Danish study suggests that in Crohn's disease, fracture risk is significantly increased, albeit to a small extent.

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.