FRAX: A nuanced tool for predicting fracture risk

Eugene McCloskey
Eugene McCloskey

According to WHO, chronic noncommunicable diseases pose a significant threat to global public health and currently account for 70% of deaths worldwide. In the 1990s, as part of an increased focus on the burden of noncommunicable diseases, WHO identified osteoporosis as a future major health problem. To address this problem, researchers sought to create a tool designed to predict those at future risk for fragility fracture.

“At that time, they identified a need to build a risk-assessment tool to predict fracture risk,” Eugene McCloskey, MD, professor in adult bone disease and honorary consultant in the department of oncology and metabolism at the University of Sheffield, told Endocrine Today. “Then, at the University of Sheffield, from the 1990s through 2008, we collated prospective cohorts of populations that looked at risk factors for fracture and combined those into what became the FRAX tool.”

McCloskey spoke with Endocrine Today about the factors FRAX evaluates, the benefits FRAX has brought to clinical practice and factors that may be added to the tool in the future.

What impact has FRAX had on clinical practice?

McCloskey: The major change is that we’ve moved away from largely basing our treatments on bone mineral density. A BMD T-score of –2.5, which was actually developed in the early 1990s as an epidemiologic tool, almost accidentally became an intervention threshold, but there is much more to your fracture risk than your BMD. It’s important to take other factors into account. Age is one example. A BMD of exactly the same value has a different meaning when you’re 85 years old vs. when you’re 55 years old.

What are some of the other factors you take into account ?

McCloskey: We take gender into account. We use height and weight to calculate BMI because BMI has a good correlation with bone density. We look at whether the patient has had a prior fracture and if either of their parents broke a hip. We ask whether the patient has been exposed to glucocorticoid treatment, whether they have rheumatoid arthritis and whether they drink 3 or more units of alcohol daily. We ask if they smoke daily, and so on. You can then take all the risk factors and compute a patient’s risk for breaking a bone in the next 10 years, and then you can use that value to decide whether you should intervene or not.

Certain factors that appear to affect fracture risk, such as diabetes, are not considered in the FRAX calculation. Why is that?

McCloskey: That’s because at the time we built the tool, there were no good data that diabetes conferred a risk for fracture independent of bone density or other FRAX risk factors. That has changed over time, and, of course, the FRAX tool is something that we continue to work on and develop. It may be that diabetes or other risk factors will be entered in the future.

Another big risk factor that is partially missing, although not completely missing, from FRAX is previous falls. We know that FRAX predicts falls. This shouldn’t be a surprise because many of the risk factors for fracture are the same as those for falls. It does predict falls, but we also have increasing evidence that previous falls add an independent contribution to fracture risk prediction beyond the variables that we have got in FRAX at the moment.

If a patient had previous falls, how much greater would their likelihood be of having a fracture?

McCloskey: We don’t have a precise answer to that question, yet, because even in the falls world, there’s a debate about what actually constitutes a fall, so there is some confusion. Many studies have used different definitions of what is a fall. The easiest thing to say is that if a patient has recurrent falls, that is, two or more falls, they’re at particularly high risk.

If you do a FRAX calculation in such a patient today and that patient lies just below the intervention threshold, for example, the doctor should be considering that that is probably an underestimate, and move them to above the intervention threshold and take action.

Over the past 10 years of using FRAX, what lessons have been learned?

McCloskey: That’s a little bit difficult to answer, again, because health care systems vary enormously around the world. Within the U.K., for example, we’ve moved away from T-score intervention thresholds to FRAX-based intervention thresholds that we know will decrease the use of BMD resources. This targets treatment to high-risk individuals.

We just published a study in the The Lancet at the end of December, where we actually ran a screening program in women aged 70 to 85 years. Half of them were randomized to normal clinical care and the other half were randomized to having a FRAX assessment, and then treatment targeted to those of highest risk. We actually reduced the number of hip fractures by 28%.

FRAX is at least mentioned in about 100 or more guidelines around the world in terms of osteoporosis and fracture risk management. It’s increasingly being seen as a tool that allows us to make more sensible decisions about who and when to treat.

Do you think it will ever become the standard of care?

McCloskey: It’s hard to say. There are lots of other tools out there, but I don’t think any of them are as well-validated as FRAX. It took 10 years to develop it — there was a lot of thought that went into it. I think it’s going to be an important tool going forward. Whether it becomes the global standard of care is not a call I’d like to make.

Do you have any tips for providers in using this tool most effectively?

McCloskey: That does depend on the health care resource. I think the best way to implement it is to embed it in family physician software. It sits alongside the electronic patient record and can take information from the patient record that can then be supplemented by the physician just asking the patient a couple of questions. This would streamline the whole management process.

FRAX is now embedded in one of the major manufacturers of general practice software in the U.K. It is also embedded in a variety of electronic patient record programs, both here and in the U.S. and elsewhere. It’s something that’s increasingly happening. It’s a very sensible way forward.

Is there anything else clinicians should know about FRAX?

McCloskey: There are currently about 63 different models of the FRAX tool for different countries and regions around the world. This is because the epidemiology of fracture differs around the world, so you need to know where you’re using it. For example, if I see a patient who has had a previous fracture, I know that roughly doubles their risk of fracture in the next 10 years, but it will double it from a different number depending on where the patient lives. Scandinavia has a very high fracture risk, whereas places like China and Turkey have lower fracture risks.

Are these differe nces due to genetics, geography or another cause?

McCloskey: We don’t know the answer to that, yet. We know there’s quite a variability. It’s almost certainly genetic, but we don’t know to what extent nutritional, environmental or epigenetic factors play a role.

References:

Shepstone L, et al. Lancet. 2017;doi:10.1016/S0140-6736(17)32640-5.

University of Sheffield. International FRAX calculation tools. Available at: www.sheffield.ac.uk/FRAX/index.aspx. Accessed Feb. 26, 2018.

WHO. Noncommunicable diseases fact sheet. www.who.int/mediacentre/factsheets/fs355/en. Accessed Feb. 26, 2018.

For more information:

Eugene McCloskey, MD, can be reached at Metabolic Bone Centre, Sorby Wing, Northern General Hospital, Herries Road, Sheffield, S57 AU, United Kingdom; email: e.v.mccloskey@sheffield.ac.uk.

Disclosure: McCloskey reports no relevant financial disclosures.

Eugene McCloskey
Eugene McCloskey

According to WHO, chronic noncommunicable diseases pose a significant threat to global public health and currently account for 70% of deaths worldwide. In the 1990s, as part of an increased focus on the burden of noncommunicable diseases, WHO identified osteoporosis as a future major health problem. To address this problem, researchers sought to create a tool designed to predict those at future risk for fragility fracture.

“At that time, they identified a need to build a risk-assessment tool to predict fracture risk,” Eugene McCloskey, MD, professor in adult bone disease and honorary consultant in the department of oncology and metabolism at the University of Sheffield, told Endocrine Today. “Then, at the University of Sheffield, from the 1990s through 2008, we collated prospective cohorts of populations that looked at risk factors for fracture and combined those into what became the FRAX tool.”

McCloskey spoke with Endocrine Today about the factors FRAX evaluates, the benefits FRAX has brought to clinical practice and factors that may be added to the tool in the future.

What impact has FRAX had on clinical practice?

McCloskey: The major change is that we’ve moved away from largely basing our treatments on bone mineral density. A BMD T-score of –2.5, which was actually developed in the early 1990s as an epidemiologic tool, almost accidentally became an intervention threshold, but there is much more to your fracture risk than your BMD. It’s important to take other factors into account. Age is one example. A BMD of exactly the same value has a different meaning when you’re 85 years old vs. when you’re 55 years old.

What are some of the other factors you take into account ?

McCloskey: We take gender into account. We use height and weight to calculate BMI because BMI has a good correlation with bone density. We look at whether the patient has had a prior fracture and if either of their parents broke a hip. We ask whether the patient has been exposed to glucocorticoid treatment, whether they have rheumatoid arthritis and whether they drink 3 or more units of alcohol daily. We ask if they smoke daily, and so on. You can then take all the risk factors and compute a patient’s risk for breaking a bone in the next 10 years, and then you can use that value to decide whether you should intervene or not.

PAGE BREAK

Certain factors that appear to affect fracture risk, such as diabetes, are not considered in the FRAX calculation. Why is that?

McCloskey: That’s because at the time we built the tool, there were no good data that diabetes conferred a risk for fracture independent of bone density or other FRAX risk factors. That has changed over time, and, of course, the FRAX tool is something that we continue to work on and develop. It may be that diabetes or other risk factors will be entered in the future.

Another big risk factor that is partially missing, although not completely missing, from FRAX is previous falls. We know that FRAX predicts falls. This shouldn’t be a surprise because many of the risk factors for fracture are the same as those for falls. It does predict falls, but we also have increasing evidence that previous falls add an independent contribution to fracture risk prediction beyond the variables that we have got in FRAX at the moment.

If a patient had previous falls, how much greater would their likelihood be of having a fracture?

McCloskey: We don’t have a precise answer to that question, yet, because even in the falls world, there’s a debate about what actually constitutes a fall, so there is some confusion. Many studies have used different definitions of what is a fall. The easiest thing to say is that if a patient has recurrent falls, that is, two or more falls, they’re at particularly high risk.

If you do a FRAX calculation in such a patient today and that patient lies just below the intervention threshold, for example, the doctor should be considering that that is probably an underestimate, and move them to above the intervention threshold and take action.

Over the past 10 years of using FRAX, what lessons have been learned?

McCloskey: That’s a little bit difficult to answer, again, because health care systems vary enormously around the world. Within the U.K., for example, we’ve moved away from T-score intervention thresholds to FRAX-based intervention thresholds that we know will decrease the use of BMD resources. This targets treatment to high-risk individuals.

We just published a study in the The Lancet at the end of December, where we actually ran a screening program in women aged 70 to 85 years. Half of them were randomized to normal clinical care and the other half were randomized to having a FRAX assessment, and then treatment targeted to those of highest risk. We actually reduced the number of hip fractures by 28%.

FRAX is at least mentioned in about 100 or more guidelines around the world in terms of osteoporosis and fracture risk management. It’s increasingly being seen as a tool that allows us to make more sensible decisions about who and when to treat.

PAGE BREAK

Do you think it will ever become the standard of care?

McCloskey: It’s hard to say. There are lots of other tools out there, but I don’t think any of them are as well-validated as FRAX. It took 10 years to develop it — there was a lot of thought that went into it. I think it’s going to be an important tool going forward. Whether it becomes the global standard of care is not a call I’d like to make.

Do you have any tips for providers in using this tool most effectively?

McCloskey: That does depend on the health care resource. I think the best way to implement it is to embed it in family physician software. It sits alongside the electronic patient record and can take information from the patient record that can then be supplemented by the physician just asking the patient a couple of questions. This would streamline the whole management process.

FRAX is now embedded in one of the major manufacturers of general practice software in the U.K. It is also embedded in a variety of electronic patient record programs, both here and in the U.S. and elsewhere. It’s something that’s increasingly happening. It’s a very sensible way forward.

Is there anything else clinicians should know about FRAX?

McCloskey: There are currently about 63 different models of the FRAX tool for different countries and regions around the world. This is because the epidemiology of fracture differs around the world, so you need to know where you’re using it. For example, if I see a patient who has had a previous fracture, I know that roughly doubles their risk of fracture in the next 10 years, but it will double it from a different number depending on where the patient lives. Scandinavia has a very high fracture risk, whereas places like China and Turkey have lower fracture risks.

Are these differe nces due to genetics, geography or another cause?

McCloskey: We don’t know the answer to that, yet. We know there’s quite a variability. It’s almost certainly genetic, but we don’t know to what extent nutritional, environmental or epigenetic factors play a role.

References:

Shepstone L, et al. Lancet. 2017;doi:10.1016/S0140-6736(17)32640-5.

University of Sheffield. International FRAX calculation tools. Available at: www.sheffield.ac.uk/FRAX/index.aspx. Accessed Feb. 26, 2018.

WHO. Noncommunicable diseases fact sheet. www.who.int/mediacentre/factsheets/fs355/en. Accessed Feb. 26, 2018.

For more information:

Eugene McCloskey, MD, can be reached at Metabolic Bone Centre, Sorby Wing, Northern General Hospital, Herries Road, Sheffield, S57 AU, United Kingdom; email: e.v.mccloskey@sheffield.ac.uk.

Disclosure: McCloskey reports no relevant financial disclosures.