Risky business: the debate over osteoporosis diagnosis

What is the best strategy to mitigate the risk of bone fractures as Australia's population ages?

---

 

According to osteoporosis specialists, osteoporosis is generally undiagnosed and untreated in primary care, with one in two Australian women over the age of 60, and more than one-quarter of men, at risk of osteoporotic fracture.

Nevertheless, they say fewer than one in three postmenopausal women with a fracture, and fewer than one in five men, are on drug treatment for the condition in Australia.

Australian guidelines and the PBS don’t currently use risk calculators to determine who receives pharmacological therapy, but many clinicians have supported the use of calculators in general practice to better communicate the need for interventions.

One of the most commonly used, web-based osteoporosis risk calculators, known as the WHO Fracture Risk Assessment Tool (FRAX), is incorporated into international guidelines.

In countries such as the US, patients are recommended to consider treatment if they have low bone-mineral density and a FRAX 10-year risk of hip fracture of 3% or more.

Osteoporosis drugs may also be recommended if the 10-year risk of major osteoporotic fractures, including proximal humeral, wrist, hip and clinical vertebral fractures is 20% or greater.

Almost 18 million patients or doctors have used FRAX to assess their fracture risk since 2011.

Until recently, emblazoned across the top of the FRAX website were the words “FRAX: WHO Fracture Risk Assessment Tool”.

Except that it wasn’t developed by the WHO, the organisation now says.

A few months ago, WHO released a statement saying it wished to “make clear that the FRAX tool is not a WHO tool and has not been developed, endorsed, evaluated or validated by WHO”.

In fact, the organisation has never given permission for its name or logo to be used in connection with the tool, the statement says.

This may surprise those who have read scientific articles, and patient and doctor groups, who have been describing FRAX as being developed by WHO. Repeatedly. Over many years.

But the WHO statement goes on to say it has no access to the algorithms or underlying data that underpin the tool – and that any treatment recommendations coming from the tool should not be construed as WHO-endorsed.

Despite the organisation’s firm language distancing itself from the tool, there has been very little media coverage of the WHO bulletin.

Professor Teppo Järvinen, an orthopaedic surgeon at the department of orthopaedics and trauma at Helsinki University, says this false advertising is a major disgrace in the field of orthopaedics.

“The FRAX group has completely falsely used the name of the WHO and its logo, and it refuses to share the algorithm at the core of the tool to the public,” he says.

For Professor Järvinen, this controversy feeds into what he sees as a larger problem of overdiagnosis in the field of osteoporosis, driven by commercially vested interests.

HISTORY

So how did this situation come about?

In the early 1990s, the WHO commissioned a group of experts at the University of Sheffield in the UK to tackle the problem of osteoporosis having no uniform definition, and thus a disease burden which was difficult to pinpoint. This group was known as a WHO Collaborating Centre.

Noting the threshold for osteoporosis was “somewhat arbitrary”, in 1994 the group defined postmenopausal osteoporosis as a bone mineral density 2.5 or more standard deviations below the mean of healthy young adult women. Low bone mass, or osteopenia, was defined as a bone mass more than one standard deviation below the mean.

But what had initially begun as an epidemiological classification quickly became a diagnostic one, with one in three white women having the “disease” under this new classification.

For Professor Järvinen and others concerned about widening disease definitions, the creep from a risk factor to disease, and the large swathe of the population suddenly becoming medicalised, raised alarm bells that women were being overdiagnosed and unnecessarily treated.

FRAX predicts the risk of a major osteoporotic fracture in the following 10 years, based on characteristics such as sex, age, weight, smoking, alcohol intake and fracture history.

But unlike something like the Framingham cardiovascular risk calculator, which has been repeatedly evaluated by independent researchers with access to the underlying algorithm, the creators of the FRAX tool have blocked requests from independent researchers hoping to do the same.

Back in the 2000s, the Sheffield group developed a model of risk factors that could be used to case-find without relying on densitometry. The WHO report on that work said these factors were at least as good at case-finding as that provided by peripheral assessment of bone-mineral density.

We have framed the problem as being thinning of bones – osteoporosis – while the problem should really be the fractures.

Nevertheless, this was not tantamount to an endorsement of the model, the WHO said in its recent statement.

Then in 2008, the University of Sheffield reformulated the earlier model into a tool and released it as FRAX.

Around the same time, new tools, such as the Garvan tool from Australia, and later the QFracture calculator, were created.

The collaboration centre was discontinued at the start of 2010, according to the WHO.

But the director of the University of Sheffield team Professor John Kanis, emeritus professor in human metabolism, and his colleagues, were referring to themselves as part of the WHO Collaborating Centre in journal articles as late as 2014.

The issue is complicated further by the financial relationships many patient groups and osteoporosis researchers have with companies that produce osteoporosis drugs or devices, both here in Australia and abroad. Osteoporosis groups continue to take money from calcium and vitamin D supplement companies, despite a lack of evidence for their benefit.

The expanded osteoporosis definition came at a fortuitous time for industry, with the first bisphosphonate, alendronic acid (Fosamax, Merck) hitting the markets in 1995.

By 2015, the global market for osteoporosis drugs hit an estimated $US11.2 billion, and is expected to reach $US14.3 billion by 2021, according to Zion Market Research.

BURDEN

According to Osteoporosis Australia, an estimated 4.7 million over-50s have osteoporosis or osteopenia, which are to blame for more than 144,000 fractures.

If nothing is done, this figure is set to increase by 30% by 2022, the group says.

Hip fractures are the most debilitating and dangerous, with patients twice as likely to die over the following decade compared with their peers. They are also costly to the healthcare system, and these costs look set to increase as the population ages.

In Australia, the criteria for treatment under the PBS is a T-score of  -2.5 or lower and a previous fracture, or age 70 or over with low bone density.

A PBS subsidy is also available for men and women with low bone density on corticosteroids at a dose of 7.5mg for at least three months.

The aim of risk calculators is to inform patients of their likelihood of having a major osteoporotic fracture in the next five to 10 years, experts say.

But according to Professor Järvinen, this relies on the assumption that bone fragility predicts hip fracture and that treatments do effectively reduce the risk of fractures.

While it may seem appealing to look at the absolute risk, Professor Järvinen says that it is a fundamentally flawed way of looking at health.

Ageing itself is many times more risky than bone fragility, Professor Järvinen says. Ageing does cause bone fragility, but without a fall even fragile hips don’t fracture.

Moreover, fewer than one in three hip fractures can be blamed on bone fragility, whereas falls account for most of the fractures in older adults, he says.

Using the rationale that patients with a risk fracture of 3% over 10 years should initiate therapy leads to overtreatment, Professor Järvinen says.

Research suggests that on this basis, almost three in four white women aged 65 and older, and 93% of over-75s, would be recommended to start drug therapy.

So while risk calculators were introduced to reduce the number of low-risk individuals receiving drug therapy, and better capture those high-risk individuals who would benefit most from drug therapy, they seem to have expanded the definition.

For this reason, more effort should be put into non-drug therapy, such as lifestyle interventions and strength and balance training, Professor Järvinen says. In a systematic review of 33 randomised-controlled trials of bisphosphonates, Professor Järvinen and his colleagues found the absolute risk reduction of hip fracture over three years was 0.57%. This means that 175 women would need to be treated for three years to prevent one fracture.

The study was considered controversial, and the authors were criticised for being sensationalist and selective with their reporting. Professor Järvinen disagrees, saying the argument they cherry-picked data was absurd.

“I find the claims that our estimate – based on a systematic review and as such, the totality of evidence – be selective or cherry-picked to be ironic,” he says.

Other research uses surrogate markers such as bone density or non-hip fractures, and these distract from the evidence that there is minimal benefit to hard outcomes, he says.

“The problem is really the framing of the problem. We have framed the problem as being thinning of bones – osteoporosis – while the problem should really be the fractures.

“When you consider fractures, you soon realise that osteoporosis doesn’t really represent that big of a problem in the whole picture. The problem is that people fall down and consequently, suffer fractures.”

For a patient, the point of a risk calculator is to determine whether the risk is high enough that action needs to be taken to reduce that risk, he says.  Unfortunately, almost anyone over the age of 75 is arguably at high risk, “the answer is always the same: start therapy”.

Some research suggests that what is considered high risk in the industry, an absolute risk of 3% over 10 years, is not necessarily high risk for patients, Professor Järvinen says.

In small focus groups, Dr Ray Moynihan, a health science researcher at Bond University, and his colleagues found that women were concerned about the risk-benefit ratio of the drugs and leaned towards referring to osteoporosis as a risk factor rather than a disease.

But there continues to be debate about the efficacy of the drugs in particular groups, and calls for more research into their effect on older adults and in men, who make up 30 to 40% of those who have hip fractures.

HARMS

Then there are the harms of patients being labelled with something that may not ever go on to cause problems in their lives.

Older patients often see various specialists, each with the idea that their specialty is the most important one, Professor Järvinen says.

One is concerned about their bones, the other about dementia and the next is hypertension, “and all of a sudden, these risks have matured into illnesses”, he says.

“The fact is, that even though these doctors refuse to believe it, labels are sticky.”

People do get anxious, and there is evidence that people getting a negative or low bone scan are less active and more anxious of falling, he says. “Which is exactly the opposite of what they should be doing.”

And while small, there is some risk of serious side effects with osteoporosis therapy, such as kidney damage, atypical femoral fracture and osteonecrosis of the jaw.

But Professor Järvinen’s argument that fragility is not the primary predictor of fracture, and that drug treatments do more harm than good, is strongly opposed.

Writing in response to Professor Järvinen’s BMJ paper, Dr Michael McClung, the founding director of the Oregon Osteoporosis Center, agreed that non-pharmacological options are often overlooked.

However, he pointed to several studies that found that fracture risk was higher in women aged over 75 with BMD-diagnosed osteoporosis compared to those without it.

In one study, the risk of hip fracture was two and a half times greater in women over-80 with BMD-criteria osteoporosis compared with their peers without osteoporosis. Ignoring these studies ignores the evidence that skeletal fragility acts independently and additively with falls, Dr McClung says.

There are multiple randomised control trials on medication and decreasing fractures, and much less on falls and fracture prevention.

And this interplay is exactly what the risk calculators try to assess. He also argues that the consistent findings that these drugs reduce osteoporosis risk by 40% are not modest, in fact a benefit of that magnitude would be welcomed by a cardiologist or neurologist.

Nevertheless, if a “high-risk” patient with a 5% risk of osteoporotic falls in 10 years was to achieve a 40% risk reduction, the absolute risk was still only 2%.

But Professor Jacqueline Center, the deputy director at the department of endocrinology at St Vincent’s Hospital and senior research fellow at the Garvan Institute, says risk calculators do have an important role and could be useful in general practice.

An absolute risk number is a good way of conceptualising the risk for a patient, and a way of reassuring low-risk patients too, she says.

Professor Center was a leading researcher on the Dubbo Osteoporosis Epidemiology Study, which informed the development of the Garvan Fracture Risk Calculator.

While it is important to look at non-pharmacologic interventions to preventing falls, she says it doesn’t replace osteoporosis treatment.

“There are multiple randomised control trials on medication and decreasing fractures, and much less on falls and fracture prevention,” Professor Centre said.