Detection of Coeliac Patients at Risk of an Osteoporotic Fracture: A Two-Cycle Clinical Audit
Osteoporotic fractures are a common problem presenting to health services in the UK: Around 300,000 osteoporotic (or ‘low-impact’) fractures, defined as those resulting from a force equivalent to falling from a standing position, occur each year . In their lifetime, one in three women and one in five men can expect to experience such a fracture and, aside from the obvious pain and suffering resulting from the injury itself, these events are also associated with considerable mortality and long-term morbidity, particularly fractures of the hip which carry a mortality of 15% . In addition to the human cost of these injuries, they also carry a significant monetary cost to the NHS, estimated to have totalled £1.8 billion in the year 2000, a figure that is set to rise with our aging population . Many of these fractures will be primary osteoporotic fractures, resulting from the natural loss of bone density experienced as one ages, but others, perhaps more worryingly, arise from inflammatory conditions of the gut which result in osteoporosis through the complex interplay of calcium and vitamin D deficiency, and the production of pro-inflammatory cytokines .
A number of inflammatory conditions affecting the small bowel can result in low bone density and subsequent low-impact fractures, such as Coeliac disease or Crohn’s disease. The focus of this audit is on Coeliac disease, which is now acknowledged as a common condition, with recent estimates placing the prevalence at between 0.1 and 1% of the population [5, 6]. Although not as pronounced a risk factor as originally thought, there is good evidence that a diagnosis of Coeliac disease carries with it an increased risk of fracture. Of the eight studies so far conducted on this subject, the majority favour a small increased risk of fracture amongst Coeliac patients [7-12]. The largest study (~13,000 patients in Sweden) showed a hazard ratio of 1.4 , although two earlier studies failed to show an effect [13, 14]. Two of the eight suggested that fracture risk was higher pre-diagnosis than post-diagnosis [8, 11], supporting the conclusion that untreated Coeliac patients are at higher risk, although the Swedish study did not support this , and two others found that risk of fracture in Coeliac disease did not decline even many years after diagnosis [7, 9]. No study to date appears to have stratified fracture risk against time spent on a gluten-free diet.
In terms of potential treatment options, both oral vitamin D and bisphosphonates have been suggested , although neither treatment has been supported by trial data. Only oral vitamin D has been tested in this way, and only in one small study, which failed to show a benefit . However, treatment with a gluten-free diet (if not already instigated) has proven effective in reversing osteopenia in the majority of trials [16-19]. Recent evidence that low bone density does not necessarily correlate with perceived disease activity [20, 21] suggests that there may be some patients with few or no symptoms, who are not strictly adherent to a gluten-free diet, that are still at risk of osteopenia. Furthermore, some Coeliac patients already well-established on a gluten-free diet but with a high risk factor profile may benefit from scanning with a view to the addition of a bisphosphonate if they are shown to be osteoporotic.
Guidelines published in 2007 by the British Society of Gastroenterology (BSG)  emphasised the need for patients diagnosed with Coeliac disease who have more than one other risk factor for osteoporosis to receive a dual energy X-ray absorptiometry (DEXA) scan to determine if they require treatment to prevent an osteoporotic fracture. The relevant risk factors are shown in Table 1. Previous guidelines have suggested that scanning at diagnosis, regardless of the risk factor profile, is warranted, although work by the same authors cast doubt on this . Therefore, as the criterion for this audit, the accepted standard of DEXA scanning in Coeliac patients who have more than one additional risk factor has been chosen. The aim of this audit is to determine if this current standard is being adhered to in a typical General Practice setting in a town in the south west of England (population approximately 7,000).
In order to determine the extent to which the BSG’s guideline on the use of DEXA scanning to identify those patients with Coeliac at risk of developing an osteoporotic fracture, the notes of all Coeliac patients at a GP practice in the south-west of England were examined to determine the prevalence of DEXA scanning in those who met the BSG criteria for this investigation (two or more risk factors; moderate or high – see Table 1). Patients with Coeliac were identified by searching for the Coeliac disease read code in the Vision software that was used at the Practice. Their demographics and risk factor characteristics are shown in Table 3. The notes of these patients were then reviewed (by a single investigator, A.N.) to identify any of a list of nine risk factors for osteoporosis (modified risk factor list – Table 2). Some of the risk factors outlined in the BSG’s guidelines were excluded from our modified list, since it was felt that they were not ascertainable from the medical notes alone; these were weight loss >10% (no timeframe is given), physical inactivity (subjective/difficult to define), late menarche (not routinely recorded) and short fertile period (requires age at menarche data). No patients with Coeliac were excluded from the study.
Based on the guidance of the BSG, the criterion selected for the audit was that all those patients with a diagnosis of Coeliac disease, who had two or more risk factors for osteoporosis, should receive a DEXA scan. A systematic search of the literature to determine the rates of bone density scanning in other centres, performed by searching PubMed for articles containing either the word ‘coeliac’ or ‘celiac’ together with either ‘bone density’, ‘densitometry’, ‘absorptiometry’ or ‘dexa’, revealed 270 articles but only one in which the prevalence of bone density scanning in Coeliac patients was addressed . This study, carried out in a secondary care setting in Hungary, showed a scanning rate of 68.2%. In discussions with D.P., who is a senior GP, and other members of the Practice staff, a scanning rate of 80% was felt to be a reasonable ‘standard’.
A search of the patient database at a Practice in the south-west of England revealed 54 patients whose records contained the Coeliac disease read code (55.6% female; median age 54.1 [range 8-88]). The demographics of this population are shown in Table 3. Among this population, only 37% were also coded as having received a DEXA scan (Figure 1). This was some way below our standard of 80%, and below the level seen in the Hungarian study by Juhász and colleagues (68.2%) . Having said this, there was clear evidence that the possibility of metabolic bone disease was being considered at our centre, since the DEXA’d population of Coeliac patients tended to have more risk factors than the unDEXA’d (Figure 2). Further breaking down the risk factors seen in those Coeliac patients with ≥2 risk factors (who according to the BSG guideline should have received a DEXA scan) allowed us to determine which risk factors were being ‘missed’ (Table 4). The most commonly overlooked risk factor was female gender, followed by smoking, age >70 and low BMI or low weight. Reassuringly, these are considered lesser risk factors by the BSG, and there is evidence to suggest that ‘moderate’ risk factors were much more robustly detected; for instance, a previous low-impact fracture did generally ensure a scan, with three of four patients in whom a low-impact fracture had been confirmed by history (from discharge summary or A&E letter) receiving an appropriate scan. Of more concern, two patients taking oral or intramuscular steroids had not been DEXA’d.
The results of this audit show that the prevalence of DEXA scanning amongst Coeliac patients deemed to be at risk of an osteoporotic fracture fell somewhat below that which was anticipated (80%) and that seen in other, albeit secondary care, settings. Although there have been some conflicting results in the literature, the majority of studies support an increased risk of fracture in patients with Coeliac disease [7-12]. Evidence also exists that fracture risk decreases following commencement of a gluten-free diet . Detection of Coeliac patients at risk of fragility fracture allows intervention such as advocating stricter adherence to a gluten-free diet or instigation of a bisphosphonate. Although no studies have addressed the question of whether DEXA scanning in at risk patients reduces fracture incidence, this does appear highly likely, and in the absence of evidence to the contrary, the BSG have recommended that Coeliac patients with ≥2 risk factors receive a DEXA scan to exclude low bone density.
Poor adherence to the BSG guideline in this centre may have arisen from lack of awareness of this guideline, scepticism regarding its evidence base or the belief that children and adolescents need not be scanned because fragility fractures are extremely rare in this group. Scepticism of the evidence base is to some extent justified, since no prospective studies of the efficacy of screening Coeliac patients in this way have been published, however, the increased risk of fracture in this group, and the fact that fracture risk falls following instigation of treatment support the stance of the BSG. Regarding the scanning of children, the BSG’s position is relatively strong. It is not the immediate risk of fracture in this group that justifies scanning, but the long-term risk of fracture if they do not attain their peak BMD. Low bone density has been consistently identified in studies of untreated children with Coeliac [19, 25]. Current guidelines advocate scanning children and adolescents who are not fully compliant with a gluten-free diet .
Despite the low overall scanning rate at this centre, the prevalence of DEXA scanning in patients with the strongest risk factors for osteoporotic fracture (such as previous fracture or low BMI) was very good, and the risk factor profiles of DEXA’d vs. unDEXA’d populations suggested that bone density had been considered in those at higher risk. However, an important finding of this audit is that the contribution of minor risk factors such as smoking and female gender are frequently overlooked. Although weak predictors of bone density in Coeliac patients, they form part of the risk factor profile, and should not be ignored.
Although it is quite possible that patients with Coeliac disease registered with this GP practice may have been omitted from the analysis due to coding errors, it is reassuring to note that the number of cases detected represented a prevalence of 1:275 (based on a list size of 14,830) which is similar to the expected level given the previously reported figures for the prevalence of Coeliac (somewhere between 0.1 and 1%, although the latter figure includes subclinical cases) [5, 6]. Aside from inaccurate coding, other potential sources of error may have arisen from patients who were DEXA’d in secondary care and the result not reported to the Practice, or those who were presumptively started on bisphosphonates as a result of steroid use or a low-impact fracture. Further to this, there are obvious limitations associated with any retrospective study relying on data recorded incidentally and not specifically for the purposes of the study. For instance, no patient in the study had a family history of osteoporotic fracture, which almost certainly reflects a lack of systematic recording of family history.
Recommendations and re-audit
As a consequence of this audit, the following recommendations were made: i) That steps were taken to raise awareness of the BSG guideline on this issue, by raising it at meetings and placing posters at prominent places within the practice; ii) that the computer system used at the Practice be set to notify the Practitioner that a patient who was due a DEXA scan had not received it, and; iii) that the GPs be informed of patients on their lists who should, according to the guideline have received a scan. Following the audit, education of Practice staff through a presentation on the BSG guidelines was carried out, together with lists of patients who met the criteria but had not been DEXA’d being distributed to GPs.
Upon re-auditing seven months later, two of the patients from our original cohort had been DEXA’d, and one had been found to be osteopenic. Unfortunately, the overall proportion of patients who met the BSG criteria for DEXA scanning and had received a scan was unchanged in the second cycle of audit (p = 1.0, Fisher’s exact test, Figures 1 and 3). It is possible that this reflects a lack of confidence in the validity of the BSG guidance in this area, possibly combined with the results of the DEXAs of the only two patients from the first cycle to have been DEXA’d, which did not in either case show significant osteoporosis (T-scores of -0.4 and -1.3). However, plotting the results of all the DEXA reports examined in this study against the number of risk factors identified from their notes, a negative correlation between these two variables is seen (Figure 4), providing some support for the BSG’s guidance.
In summary, there is good evidence that Coeliac disease is a risk factor, albeit a moderate one, for osteoporotic fracture, and metabolic bone disease resulting from Coeliac disease is easily treated by promoting adherence to a gluten-free diet or the addition of a bisphosphonate. Since BMD may not necessarily correlate with other symptoms of Coeliac, this may be most relevant to Coeliac sufferers with mild or subclinical disease. If this audit is typical of practice in other primary care settings, work needs to be done to promote bone density measurement in this group of patients, and since education alone had no effect on scanning here, it may be necessary to simplify current guidance by recommending routine scanning of all patients diagnosed with Coeliac (particularly children) to ensure that osteopenia, especially in relatively asymptomatic patients is not missed.