Overdiagnosis of bone fragility in the quest to prevent hip fracture

Jarvinen et al are to be commended for their comprehensive analysis of fragility fracture evaluation in hip fracture prevention(1). It is to be welcomed. It is vitally important that we consistently question apparently accepted medical orthodoxy. However, there is a significant gravamen against fragility fracture treatment algorithms, not highlighted in the article. Bisphosphonates have been proven to be efficacious at preventing fragility fracture in those with osteoporosis or severe osteopenia (BMD <-2.0SD)(2). Their effectiveness in those who do not fall within this category remains to be definitively determined. The authors correctly identify that most fragility fractures occur in those who are not osteoporotic. They also note that the UK’s National Osteoporosis Guideline Group (NOGG) would advocate treatment, for example, in a 55 year-old UK woman if her 10-year fracture risk exceeds 1.5%, as determined by FRAX. However, FRAX is not dependent on bone mineral density in its evaluation of fracture risk. Bone mineral density is only a facultative variable for this calculator. Hence some of those for whom treatment is recommended may not be osteoporotic. This is problematic as it is not clear that bisphosphonates or anti-resorptives are effective at preventing fragility fracture in this context.

Fragility fracture is a much more sophisticated phenomenon than originally thought. Bone mineral density and bone mineral content may not be the sole or even the most important determinative factors. Even after correction for BMD, weight and other covariates, Barret-Connor et al found that Asian American and Black post-menopausal women had 0.3 and 0.5 the risk of fragility fractures respectively of white post-menopausal women. Similarly Uitterlinden et al observed that even following adjustment for BMD and a number of other potential confounders, post-menopausal women who were heterozygotes of a collagen 1 gene polymorphism have a significantly higher risk of fracture(3). Consider an analogy in cardiology. One might determine the 10-year risk of myocardial infarction in a patient and find this to be high. If the patient were normotensive, his blood pressure would not be targeted for intervention, however high the risk of ischaemic heart disease. However, some of the fragility fracture recommendations seem to suggest a similar logic with regard to bisphosphonates and modifying fragility hip fracture risk.

The authors do identify a real problem, however; this is that osteoporotic fractures are often thought to be synonymous with fragility (low energy) fractures. Hence there is the temptation to treat all fragility fractures with agents that improve bone mineral density. However, the root cause may not lie here. This may explain the limited efficacy observed by the authors of bisphosphonates. Abraham Maslow, psychologist, famously stated

“it is tempting, if the only tool you have is a hammer, to treat everything as if it were a nail”.(4)

My emendation would be

“it is tempting, if the only tool you have is a bisphosphonate, to treat every fracture as if it were osteoporosis”

1.Järvinen TL, Michaëlsson K, Jokihaara J, Collins GS, Perry TL, Mintzes B, Musini V, Erviti J, Gorricho J, Wright JM, Sievänen H.Overdiagnosis of bone fragility in the quest to prevent hip fracture. BMJ. 2015 May 26;350:h2088
2.Barrett‐Connor E, Siris ES, Wehren LE, Miller PD, Abbott TA, Berger ML, Santora AC, Sherwood LM. Osteoporosis and fracture risk in women of different ethnic groups. J Bone Miner Res. 2005;20:185–194.
3.Uitterlinden AG, Burger H, Huang Q, Yue F, McGuigan FE, Grant SF, Hofman A, van Leeuwen JP, Pols HA, Ralston SH.Relation of alleles of the collagen type Ialpha1 gene to bone density and the risk of osteoporotic fractures in postmenopausal women. N Engl J Med. 1998 Apr 9;338(15):1016-21
4.Abraham H. Maslow (1962), Toward a Psychology of Being:

Hip fracture risk is never a sectional diagnosis but an important indicator of potential skeletal deficiencies in many other sites which are not easily accessible. Conversely, treatments ameliorate bone function and stability in the whole skeleton, probably improving the global health status of a person having “hip fracture risk”.

I was impressed by the responding letter saying “I suspect very few will choose a bisphosphonate, given the quoted figure of 175 needed to treat for three years to prevent one hip fracture - I certainly wouldn't.” The potential benefits from many vaccines are much lower than that, and with such kind of isolated arguments nobody should accept a vaccine.

But the therapeutic efect of bisphosphonates is not just the avoidance of a hip fracture but complex changes in skeletal strength. Whether the patient fractures or not is a consequence of simultaneous events, including occasional exposure to minimal traumas, in turn promoted by other organ deficiencies. In patients at risk, taking appropriately long term bisphosphonates reduces many non-hip fractures and in addition may have the benefit of improving skeletal structure and increase life expectancy by an average of around 2 years or more compared with no treatment.[1] Likely “over-diagnosis” seems not to be the overall effect. Now, I believe that many will accept the treatment if a complete picture of potential benefits and risks are duly stated.

Center JR, Bliuc D, Nguyen ND, Nguyen TV, Eisman JA. Osteoporosis medication and reduced mortality risk in elderly women and men. Clin Endocrinol Metab 2011, 96 :1006-1014. doi: 10.1210/jc.2010-2730. Epub 2011 Feb 2.

It is a pity that Järvinen et al (1) did not pay attention to the paper by Carter et al (2) and more precisely define the term ‘overdiagnosis’. This has angered many osteoporosis experts, who point out that under-diagnosis and undertreatment dominate the field.

Underdiagnosis and overdiagnosis can coexist in arbitrarily large measure, if by ‘overdiagnosis’ we mean that many people will be accorded a ‘disease-label’, whilst only a few will benefit from treatment. Our risk assessment is necessarily imperfect, diagnostic thresholds inevitably include false positives, and treatments are often of low efficacy. I think we all agree that underdiagnosis occurs when a disease goes unrecognised. To use a fishing metaphor, we may catch large numbers of the wrong fish, whilst at the same time missing a goodly proportion of the right fish !

Järvinen is concerned that we do not treat too many people to too little benefit. Expert responses reveal much more concern to ensure that we DO treat as many as can reasonably benefit from treatment to prevent fractures.

Both aims are laudable, and require more sensitive AND more specific trawling nets. The accurate estimation of NNT and NNH is necessary.

Some time ago we (3) proposed the valuable estimation of NETT (Numbers expected to be treated - Prim Care Cardiovasc J doi:10.3132/pccj.2012.060) in relation to Statin prescribing, which permits comparison of the efficiency of competing strategies at the population level. NETT computes the number of people eligible for treatment using the chosen screening method. For example, offering statin above a CVD risk threshold of 10% instead of 20% will greatly increase the population NETTed. Good strategies require not only a low NNT, a high NNH, but also achievable and sustainable NETTs based on accurate population risk profiling.

References:
1. Järvinen TLN, Michaëlsson K, Jokihaara J, et al. Overdiagnosis of bone fragility in the quest to prevent hip fracture. BMJ 2015;350:h2088

2. Carter S M, Rogers W, Heath I, Degeling C,Doust J, Barratt A et al. The challenge of overdiagnosis begins with its definition BMJ 2015;350:h869

3. Lewis LS, Gollop ND. The Numbers expected to be treated: a way forward in CVD prescribing. Prim Care Cardiovasc J doi:10.3132/pccj.2012.060

Although Jarvinen et al highlight some of the areas that are still somewhat unclear and demand further scientific research, the outcome of this article can have serious consequences and long lasting negative effects on efforts that are directed to find people with high risk of fractures. Inarguably, fragility fracture is a serious health concern and use of multiple risk factor approach in case finding is common to most of the non-communicable diseases. It is worth noting that the association between BMD and fracture is much stronger than that between lipids and heart attacks.

Fragility fractures are a major health issue in Asian countries. In these countries, osteoporotic fractures are inadequately treated and high risk patients are largely undetected and undertreated leading to a high care gap. Although the current hip fracture incidence is lower in Asian countries compared to European and Scandinavian countries, more hip fractures are predicted to occur in the Asian region. Unless those responsible for the health of the population do take appropriate preventive measures, the health and socio-economic consequences of this could be enormous. The take home message of this article can cause a serious damage to the efforts that are being taken to minimize fracture burden in these countries.

Despite certain limitations, FRAX model has many advantages. It allows multiple risk factor approach in case finding and can be used in countries with no or limited DXA access. Although we have witnessed many advances in this area of medicine there are areas which need more accurate knowledge. The optimal period of treatment is certainly an area which needs further studies and clear answer. Until we find those answers, I sincerely hope that final message of the article by Jarvinen et al would not cause deleterious effects on osteoporosis related care services which are emerging in these developing countries.

Professor Sarath Lekamwasam (MD, FRCP, PhD)
Professor of Medicine
Center for Metabolic Bone Diseases
Faculty of Medicine
Galle, Sri Lanka