Article Text

FRI0302 Analysis of the Defra Algorithm for Definition of Risk Fracture
  1. L. Wolenski1,
  2. B. Frediani2,
  3. A. Marchi1
  1. 1Libero Professionista, cesena
  2. 2Rheumatology Unit, University of Siena, Siena, Italy


Background To define the transition, in the study of osteoporosis, including diagnostic and therapeutic thresholds, algorithms have been built, which assess the risk of fracture within 10 years, with an examination of variables, both clinical and instrumental.

The first developed was the FRAX. It is similar to DEFRA (derived FRAX), whose advantages are essentially:

  1. The weighting is quantified of the individual clinical risk factors, reduced to dichotomous variables in FRAX.

  2. As instrumental variable is added over the femoral BMD, exclusive of FRAX, even the lumbar BMD. In addition it's possible to insert the parameters of QUS (phalangeal and heel).

Clinical risk factors of Defra: Age, smoking, steroids, alcohol, family history of fractures, previous femoral fractures and/or vertebral, previous non-traumatic fractures, rheumatoid arthritis and other connective tissue.


  • Define the validity of the algorithm, with the evaluation of an alternative with the only clinical variables.

  • Define the weighing of the different variables.

Methods 1408 patients at a Rheumatology outpatient clinic were analyzed with the Defra questionnaire, using the calcaneal QUS (Achilles Insight GE), with analysis of the variables of which are annexed descriptive statistics.

The methodology used in order to achieve the objective of the study, was articulated on one hand by analysis of binary correlations between the variables investigated, on the other in the development of a dual model of multiple linear regression (with and without ultrasound variables).

Table 1 for MEN

Table 2 For WOMEN


  1. Bivariate correlation analysis:

    • Negative correlation between risk and Stiffness (46.6%)

    • Positive correlation with previous vertebral fractures and or femoral(74%)

    • Positive correlation with age factor (50.8%)

  2. Analysis of the multiple linear regression:

    • the model containing all the variables analyzed detects an R2 equal to 91.6%, with a very high degree of risk prediction.

    • the model without ultrasound variables can explain 81.2% of the risk.


Conclusions Defra is an extremely valuable algorithm in the prediction of fracture risk, even with the clinical variables, although in a slightly lower rate compared to the classical model.

The higher weighting, among the parameters studied, is attributed to vertebral and/or femoral fractures and in second place to age.

Disclosure of Interest None declared

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.