Article Text


P25 Fragility fracture risk in cirrhosis: a comparison of the fracture risk assessment tool (FRAX_), BSG and NICE guidelines
  1. L Ayres,
  2. J Digby-Bell,
  3. S Clarke,
  4. A Dhanda,
  5. K Caddick,
  6. P Collins
  1. University Hospital Bristol


Introduction Low bone mineral density (BMD) is common in patients with chronic liver disease, and tends to be related to the severity rather than aetiology of cirrhosis. BMD alone correlates modestly with fragility fracture. The BSG recommend DXA scanning all cirrhotic patients. NICE guidelines apply only to postmenopausal women.

The FRAX® score is a web-based tool incorporating clinical risk factors to predict the risk of osteoporotic fracture. Risk may be estimated without BMD, and refined with DXA. Pre-BMD FRAX® categorises patients to low, intermediate or high risk according to set thresholds by the National Osteoporosis Guidelines Group (NOGG) and recommend lifestyle-advice, DXA or anti-osteoporosis treatment respectively (intermediate patients are rescored post-BMD to either high or low risk). Cirrhosis contributes to calculated fracture risk.

Aim To determine what proportion of our patients would be eligible for DXA and recommended treatment under BSG, NICE and FRAX®/NOGG guidelines. To calculate the correlation of the pre-BMD FRAX® scores with the post-BMD FRAX® scores.

Method We interrogated our outpatient database of cirrhotic patients and calculated the FRAX® score pre and post-BMD measurement. For each guideline we calculated the number of: patients eligible for treatment; high risk patients left untreated and low risk patients unnecessarily treated.

Results Of 132 patients (58% male) with cirrhosis (mixed aetiology), 43 had normal BMD, 64 were osteopaenic and 25 osteoporotic. 103/132 patients (43 postmenopausal women and 60 men >50 years) were eligible for FRAX®. Of these, 64/103 were estimated low, 33/103 intermediate, and 6/103 at high risk of fracture. Of those at low risk (and therefore not required to undergo DXA scanning (NOGG)), 8 were reclassified as high risk post-BMD (one patient was reclassified from high to low risk). Patients eligible for treatment: pre-BMD FRAX®/NOGG 10/103, post-BMD FRAX®/NOGG 17/103, BSG 25/132, NICE 3/45. Low risk patients unnecessarily treated: pre-BMD FRAX®/NOGG 1, BSG 16, NICE 2. High risk patients not recommended to receive treatment: 8, 11, and 4. Pre-BMD FRAX® correlates well with post-BMD FRAX: Spearman's coefficient r=0.76, sensitivity of 47.5%, specificity 98.8%, PPV 90% and NPV 89% for correctly risk stratifying patients.

Conclusion BSG guidance does not incorporate clinical risks and is likely to expose some patients to unnecessary tests, or treatment. NICE is not widely applicable and does not suggest when to DXA. Cirrhosis is not listed as a risk factor, thus underestimating fracture risk. FRAX®/NOGG accurately identifies high risk patients and which group to DXA. Scanning only intermediate risk patients saves unnecessary scans, but in our cohort fails to identify 8/94 (8.5%) high risk patients. Male smokers with alcoholic cirrhosis were overrepresented in this group. Similarly to other cohorts, Pre-BMD FRAX® has a low sensitivity but is highly specific when compared to post-FRAX® BMD.

Abstract P25 Table 1

Patients recommended treatment vary according to guideline

Abstract P25 Table 2

Comparison of pre-BMD and post-BMD FRAX

Abstract P25 Table 3

Prevalence of clinical risk factors

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.