2.1 Risk factors for hip fracture

Risk factors for hip fracture can be neatly, but perhaps simplistically, separated into those which relate to the increased prevalence of falls in the elderly and those which relate more specifically to changes in bone mass. However, there are some factors, such as smoking and immobility, which may have direct effects on bone mass while also increasing the risk of falls.

An American study of osteoporotic fractures defined 16 risk factors which, when present in middle-aged women, led to an increased incidence of fracture during a follow-up period of three years.³ Of these 16 factors, the four most prevalent and which define the largest number of at risk women are shown in Table 1. Evidence level 2+

A more recent population-based cohort study from the Netherlands with a follow-up period of 2.8 years⁴ identified a history of immobility as another significant factor. Low body weight was also a specific risk factor, but this study did not find that current smokers had an increased risk. However, the balance of evidence from other cohort studies does suggest that current smoking at least doubles the risk and it may be up to 10 years after cessation of smoking before the excess risks disappear.⁵ Evidence level 2+

Table 1: Risk factors for fracture

2.2 Risk factors for falls

The tendency to fall increases with age. Risk factors for falling tend to increase in prevalence with age, leading to more frequent falls. Evidence from cohort studies strongly suggests that the direction of the fall (to the side rather than forward) is critical in determining hip fracture and is also an age-related effect. Poor visual acuity, use of hypnotics, neurological disease and slow gait speed have also been shown to be significant risk factors.^6,7,8 Evidence level 2+

Risk factors for falls are of most significance if they are (1) easily identifiable and if they are (2) potentially reversible. Potentially reversible risk factors (see Table 2) are an obvious target for intervention, provided that the preventative approach is cost-effective. Identifiable risk factors which cannot be reversed might be used to target protective devices.

Table 2: Risk factors for falls

2.2.1 BONE-RELATED FACTORS

Assessment of bone mass is the most studied and probably the most powerful bone-related predictor of future hip fracture. Several prospective cohort studies have shown that a reduction in bone density, measured at the hip or heel, of one standard deviation, is associated with at least a doubling in risk of hip fracture.⁹ This has been confirmed in two recent meta-analyses.^10,11 Evidence level 1+

However, bone mineral density (BMD) is not the only measurable characteristic of bone that predicts hip fracture. Examination of the Study of Osteoporotic Fractures (SOF) prospective database has demonstrated an excess risk, independent of BMD and age, of an increase in the length of the femoral neck.¹² However, femoral length is set at an early age. The EPIDOS (Epidémiologie de l'Ostéoporose Study) prospective cohort study from France has demonstrated that a one standard deviation reduction in quantitative bone ultrasound parameters, measured at the heel, is also associated with a doubling of fracture risk,¹³ an effect which was partly independent of BMD. Similar findings have been demonstrated from the SOF database.¹⁴ Finally, an elevated rate of bone turnover, as assessed by measurement of bone resorption and formation markers, has also been shown to predict future hip fracture independently of BMD in the EPIDOS^15,16 and SOF studies.¹⁷ At present it is not possible to quantify fracture risk using a validated scale based on the presence of these markers, but the evidence for this may continue to develop. Evidence level 2++

Bone-related clinical risk factors may also be independent predictors of hip fracture and may be identifiable without recourse to clinical measurement.^{3, 4} These include:

• maternal history of hip fracture
• a previous low trauma fracture after the age of 50
• low body weight (i.e. BMI <18.5) is one of the strongest risk factors.¹⁸
  Evidence level 2+

Assess the risk of hip fracture in older people using the identified risk indicators and base any intervention on this risk assessment (patient and environment).

2.3 Non-drug interventions to prevent falls and fracture

2.3.1 EXERCISE AND ASSOCIATED INTERVENTIONS

A number of randomised controlled trials (RCTs) have studied the use of exercise programmes in the prevention of hip fracture and two high quality meta-analyses of these studies have been carried out.^8,19 In most of the trials, exercise was combined with other interventions such as home assessment, dietary change, use of hip protectors, education, cognitive intervention or medication change. It is not clear which of these interventions, or which subset, is effective; however, exercise programmes alone have not been shown to be effective.¹⁹

There is insufficient evidence to support the use of individual interventions such as exercise or balance training in fall or fracture prevention. However, interventions which target multiple, identified risk factors in individuals can be effective in reducing falls, as can behavioural interventions targeting environmental hazards plus other risk factors. Home-based programmes to improve strength and balance can reduce significantly the number of falls and injuries experienced by women in community settings aged 80 years and older.²⁰ Evidence level 1+

Older people should have their risk of falls and fracture assessed. Those at increased risk should be offered multiple interventions* aimed at reducing the identified individual and environmental risks.

* E.g. exercise programmes (focusing on strength, flexibility and which are weight-bearing), balance training, and modification of identified hazards.

2.3.2 HIP PROTECTORS

A recent systematic review²¹ of seven RCTs has found that hip protectors worn by older people in institutional settings who are at high risk of hip fracture appear to reduce the risk of fracture by 50-66%. A recent study in care homes in Japan confirms this strong protective effect.²² Evidence level 1+

Compliance with wearing hip protectors in older people living in care homes is likely to be only 25-30%, mainly due to problems with fit and skin irritation.²³

Hip protectors are recommended in men and women at high risk of hip fracture, particularly older people in care homes, although problems with compliance should be recognised.

2.4 Drug therapies

The risk of hip fracture may be reduced by a number of dietary and pharmacological agents that decrease bone turnover and reduce fracture incidence. Calcium (alone or with vitamin D), hormone replacement therapy (HRT), bisphosphonates, calcitonin, fluoride and thiazides all have effects on bone mass, and some of these drugs have been promoted for use in the primary and secondary prevention of osteoporosis. Evidence is still emerging regarding the role of selective oestrogen receptor modulators (SERMs), and they have not been addressed in this guideline.

The focus in this guideline is on those treatments where data exists on hip fracture as an endpoint and, when this is absent, where there is good data on bone density at the hip. Cost-effectiveness has been taken into account in deriving recommendations on treatment, which are summarised in section 2.5.

Additional information about the cost-effectiveness of interventions to prevent falls and hip fractures is available on the SIGN website: http://www.sign.ac.uk.

2.4.1 CALCIUM

No RCTs were found that looked at the effect of calcium supplementation alone on hip fracture as an endpoint. One RCT²⁴ showed no benefit of calcium supplementation on bone loss during the first five years postmenopause, but supplementation produced a significant increase in BMD at the hip in the late menopause. Calcium supplementation appeared most effective in those with lowest calcium intakes. A second RCT,²⁵ using high dose calcium supplements in late menopausal women, produced significant retention of BMD at all hip sites. This was confirmed by a four year RCT which also demonstrated an associated reduction in the total fracture rate.²⁶ Evidence level 1+

A Department of Health report on nutrition and bone health²⁷ concluded that there was insufficient evidence to recommend an increase in calcium intake in the elderly from the recommended daily allowance (RDA) of 700 mg/day, although expert opinion admitted this might be inadequate. The report concluded there was evidence that calcium intake below 400 mg/day might not be compatible with good bone health. Evidence level 4

2.4.2 CALCIUM PLUS VITAMIN D

Vitamin D acts to enhance calcium absorption and also corrects the secondary hyperparathyroidism found in deficiency states, which are common in the elderly due to lack of sun exposure and reduced dietary intake of vitamin D.

Calcium plus vitamin D has been shown to reduce significantly the incidence of all fractures, including hip, in both elderly women with a high risk of hip fracture living in institutions²⁸ and in independently living men and women over 65 years of age.²⁹ This has been confirmed in a systematic review,³⁰ although it was unclear whether vitamin D alone offered protection. Vitamin D also appeared to reduce death rates in treatment groups, suggesting age-related actions other than on bone. Evidence level 1+

Calcium plus vitamin D is cheap and safe at prescribed dosages and could be a cost-effective intervention in groups such as institutionalised elderly people and those with poor diets. This is an evolving area and further trials will inform practice in the future.

2.4.3 HORMONE REPLACEMENT THERAPY

A prospective controlled study³¹ concluded that postmenopausal women with the lowest total oestrogen levels (and highest sex hormone binding globulin levels) were most at risk of a hip fracture. Evidence level 2++

The median age for hip fracture in the UK is 79 years, more than 25 years post menopause, yet fewer than 20% of women continue HRT for more than 10 years, even in the most committed groups. Long term compliance is poor because women take HRT for climacteric symptoms and not bone loss. Continuous combined preparations now overcome cyclical bleeding but there is a recognised risk of breast and endometrial carcinoma (the latter only with unopposed oestrogen use) with treatment longer than 10 years.³² Evidence level 2++

There are no RCTs on the effect of HRT on hip fracture. A systematic review³³ of all types of fracture, concluded that there was fair evidence that HRT reduced the incidence of fracture while treatment was continued, and good evidence that bone mass was preserved during treatment. Evidence level 2++

There is consistent observational evidence from a meta analysis of case controlled and cohort studies³⁴ that oestrogen reduces the risk of hip fracture in post menopausal women by approximately 25% and that risk decreased with duration of treatment. A prospective cohort study³⁵ of HRT in women older than 65 years showed the greatest reduction in hip fracture rates in those over 75 years of age. A large population-based case-control study by the Swedish Hip Fracture Group³⁶ showed that the hip fracture rate reduced by 6% per year with use of HRT, with combined oestrogen/progestogen products being more effective. Progestins also allowed a lower dose of oestrogen to be equally effective. HRT started even nine years or more postmenopause gave some reduction of hip fracture risk, and most of this acquired protection was lost within five years of stopping therapy.

2.4.4 BISPHOSPHONATES

Bisphosphonates such as alendronate, risedronate and etidronate act by inhibiting the dynamic resorption of bone by osteoclasts, reducing the rate of bone turnover and preserving bone mass. There are RCTs on the use of the alendronate in both primary and secondary prevention.^37,38 These trials, on women with and without pre-existing vertebral fractures, showed a statistically significant reduction in hip fractures over three years of treatment but contained only small numbers of fractures in a highly selected group of women. Both trials showed statistically significant increases in bone density at hip sites with duration of treatment. Evidence level 1-

In a large phase III RCT of risedronate, designed specifically to prevent hip fracture, the bisphosponate reduced hip fracture rates by 40% in women aged 70-79 years, with low bone density at the femoral neck. However, it was no more effective than calcium and vitamin D alone in women aged 80 years or more.³⁹ This is consistent with the non-vertebral fracture rate of alendronate which is also only seen in those with low femoral neck bone density.³⁸

A retrospective population-based cohort study⁴⁰ of the General Practice Research Database on the effect of cyclical etidronate showed a significant reduction in hip fracture rate, greatest in those over the age of 76 years. Control and treatment groups were not well matched, but bias would have favoured a reduced effect of treatment.

2.5 Cost-effective targeting of interventions

Modification of environmental risk factors, use of HRT and treatment with calcium and vitamin D targeted at those with relevant risk factors all may result in reductions in hip fracture rates. The costs associated with these interventions are lower in the longer term compared to the cost of no treatment to reduce risk and the cost of managing a later hip fracture. However, some sustained treatments (e.g. HRT) may be less clinically desirable and should be assessed for each patient and related to lifestyle issues. Evidence level 2+

The quality of the cost-effectiveness evidence for some interventions is relatively poor (e.g. modification of environmental risk factors, HRT and vitamin D).^{41,42,43,44,45,46,47,48}

The most cost-effective intervention is calcium and vitamin D. The more costly bisphosphonates start to become cost-effective when their use is targeted to high risk individuals (see sections 2.1 and 2.2).^49,50,51 Evidence level 1+

Targeting therapy to high risk individuals - by using either BMD measurement or an assessment of clinical risk factors for bone related risk factors during routine visits - greatly improves the cost-effectiveness of hip fracture prevention. Targeting those with low bone mineral density gives a cost per hip fracture prevented of approximately £11,000 for bisphosphonates (excluding cost savings from avoiding treatment). The cost per hip fracture prevented and the total cost to the health service are even more favourable for calcium and vitamin D, and hip protectors. Evidence level 2+

BMD measurement appears to be a less cost-effective method of targeting therapy with calcium and vitamin D than assessing clinical risk factors. However, it may be the only realistic way to target the use of bisphosphonates to reduce hip fractures.³⁸

A number of factors are recognised as indicators of increased risk of hip fracture in older people (see sections 2.1 and 2.2). At present it is not possible to quantify risk using a validated scale based on the presence of these markers, but it would seem reasonable to assume that higher risk is associated with the presence of more markers.

Assessment of recognised risk factors for low bone density is the most cost-effective method of targeting interventions that act on low bone density. Mass screening for low BMD is less cost-effective and is not recommended.

All patients who are assessed as being at risk of hip fracture should be treated with calcium and vitamin D.

All patients who are assessed as being at high risk of hip fracture should be treated with: hip protectors, when patients are living in a care home setting and are assessed as being compliant or the bisphosphonates, alendronate or risedronate, when risk is assessed by measuring BMD.

Where access to BMD measurement is impractical, bisphosphonates may be considered in patients with strong evidence of pre-existing osteoporosis (see Table 1). There is no evidence at present that these drugs are effective in preventing hip fracture in patients over the age of 80 years.