Detecting the risks of osteoporotic fractures in coeliac disease

• osteoporosis
• fractures
• coeliac disease

The recent report by Thomason and colleagues (Gut 2003;52:518–22) which failed to detect a significant increase in fractures experienced by treated coeliacs might reassure many patients and physicians. However, this study, and the accompanying commentary by Compston (Gut 2003;52:459), need full and critical assessment before changes in practice are adopted and coeliacs are no longer targeted to be screened for osteoporosis.

It is not surprising that no significant increase in fracture could be detected in this population of well treated coeliacs, given previous findings. The American Gastroenterology Association recently reviewed studies of osteoporosis in gastrointestinal diseases, including coeliac disease, according to standard levels of evidence.¹ All such studies have shown low mean bone mineral density (BMD) around the time of diagnosis of untreated coeliac disease, with a pooled analysis showing very low bone mass (age and sex adjusted z scores below −2) in 40% in the spine and 15% at the hip. However, many reports, including our own,² have shown normal or near normal mean values after treatment. This reflects the great improvement in BMD³ and calcium absorption⁴ which occurs when enteropathy is reversed with a gluten free diet. The real issues are how to recognise previously undiagnosed cases, and how to identify potential patient subgroups who might still be at risk due to suboptimal treatment.²

The study also did not have sufficient power to detect any increase in those fractures most typical of osteoporosis which have a high prevalence late in life.⁵ Such fractures typically include vertebral collapse and deformity, causing significant morbidity, but which commonly are undiagnosed unless looked for radiologically. In a 50 year old woman, there is a 32% life time risk of subsequent vertebral fractures.⁶ However, these were not recorded in either coeliacs or controls in this study, indicating that the questionnaire method employed led to marked under reporting. Femoral neck (hip) fractures, the most serious complication of osteoporosis, have a population incidence of less than 1% by the age of 65 years but approaching 20% by the age of 90 years. In this study, only about one third of coeliacs were aged over 65 years and only one-tenth (n=20) over 75 years. Approximately 400 cases and controls would be needed in a prospective study to detect a 50% increase in risk from 20% to 30% with 90% power. Hence the ability of this study to detect a significant life time increase in fracture risk in the minority of coeliacs who are suboptimally treated is clearly limited. Much larger groups are needed in a prospective study which includes radiological ascertainment of vertebral fractures.

A relevant comparison can be made with the large US Study of Osteoporotic Fractures Research Group where over 9000 women older than 65 years of age were studied. In 1814 subjects with a daily calcium intake less than 400 mg, those below the median for fractional calcium absorption (so resembling untreated coeliacs) experienced an incidence of hip fractures of about 9 per 1000 person years—2.5-fold greater than those with absorption above the median.⁷

The absolute life time risk of osteoporotic fractures is reasonably large in the ageing population, even if the relative risk in coeliacs is not that great. Patients will be reassured by the knowledge that BMD has improved with dietary treatment and that individual risk has decreased. We hold that as the true risk of fracture in treated and untreated elderly coeliacs is unknown at present, judicious use of DXA measurements, following appropriate guidelines, to monitor calcium homeostasis and the benefits of treatment should continue.