2.1 Introduction

This chapter summarises the evidence for the major risk factors that act as predictors for osteoporosis. Many other factors have been proposed as conveying risk of osteoporosis, but there is no consistent evidence for any risk factors other than those discussed below.

2.2 Falls

Not all fractures are associated with osteoporosis. Clinical risk factors related to physical function and falls are not a risk factor for osteoporosis per se , but are powerful predictors of fracture risk. This guideline, however, does not address the important issue of falls and their prevention. SIGN 56 on hip fracture briefly documents identifiable risk factors for falls,⁷ and NHS Health Scotland has conducted qualitative research into lay and professional attitudes to falling and how they can influence fall prevention programmes.³⁰ A much fuller review of this topic can be found in the guidelines developed by the American and British Geriatrics Societies.³¹

2.3 History of previous fracture

Women who have suffered a previous fragility fracture (defined as a fracture occurring after a fall from standing height or less) are at increased risk of further fractures, independent of BMD.³² Women who develop a vertebral fracture have a 19.2% (95% CI 13-6-24.8%) risk of a further vertebral fracture within one year.³³ Men and women aged 65 years or older with a vertebral fracture have a five year risk of femur or hip fracture of 6.7% and 13.3% respectively.³⁴ In women, the presence of one or two vertebral fractures increases the risk of further fracture 7.4 fold.³⁵ Evidence level 2+

Those at highest risk of fracture are those who have already fractured including those with loss of height or kyphosis which can indicate (what may have been painless) vertebral fractures.^36,37 Evidence level 1++

Patients who have suffered one or more fragility fractures should be priority targets for investigation and treatment of osteoporosis.

2.4 Non-modifiable risk factors

2.4.1 age

BMD decreases, and consequently the risk of osteoporosis increases with age. A significant increase in prevalence with each decade after age 60 has been demonstrated. The United States National Health and Nutrition Survey (NHANES) III survey of postmenopausal women showed that the prevalence of osteoporosis in non-Hispanic white American women was 27% (50-59 years), 32% (60-69 years) and 41% for those >=70 years.³⁸ A previous estimate based on data from Rochester, Minnesota indicated a lower (though still high) prevalence – 14.8% (age 50-59 years), 21.6% (aged 60-69 years), 38.5% (70-79 years) and 70% (>=80 years.)³⁹ A Yorkshire based study showed a prevalence of 24% in women aged 60-69 years.⁴⁰

2.4.2 sex

Women are at greater risk of osteoporosis as they have smaller bones and hence lower total bone mass. Additionally, women lose bone more quickly following the menopause, and typically live longer. Osteoporosis is less common in men but is still a significant problem.

The rate of bone loss in men is less than that in women. In the Framingham Osteoporosis Study annualised percent bone loss for women was 0.86% to 1.21% at different sites and for men, 0.04% to 0.90%.⁴¹ Secondary causes of osteoporosis are, however, more common in men, affecting approximately 40% of cases.^42,43

Excepting reproductive factors and taking into account the increased influence of secondary factors in men, the risk factors in women also apply to men.

2.4.3 ethnicity

Afro-Caribbean women have a higher BMD than white women at all ages due to a higher peak bone mass and slower rate of loss.^38,44 White women have a 2.5-fold greater risk of getting osteoporosis.³⁸

2.4.4 reproductive factors

A late menopause or short time from menopause to BMD measurement are associated with higher BMD. There is consistent evidence that low BMD is associated with early menopause.⁴⁵ Consequently, women with an early menopause should be considered at higher risk of osteoporosis than others at a similar age.

BMD decreases most rapidly in the early postmenopausal years.⁴⁶

There is no consistent evidence that tubal ligation, parity, number of previous miscarriages, or breast feeding affect BMD.

Current use of oestrogen replacement therapy is associated with a higher BMD.⁴¹ Those currently taking oestrogen therapy should therefore be considered as being at lower risk than others at a similar age, unless the therapy was prescribed for osteoporosis.

2.4.5 family history of osteoporosis

Lower BMD is found in women and men with a family history of osteoporosis, a family history being defined as a history of osteoporosis or brittle bones, kyphosis (“dowager's hump”), or low trauma fracture after age 50 years as reported by the offspring. Evidence level 2+

Individual BMD decreases as the number of family members with osteoporosis increases. Overall family history is a more sensitive predictor of osteoporosis risk than maternal or paternal history alone.⁴⁷ Prevalence of a positive history in sisters is similar to prevalence reported for mothers.^47,38 Evidence level 2+

In one epidemiological study the greatest risk of categoric osteopaenia (RR 2.16, CI=1.38-3.37) was in patients whose father had a history of osteoporosis.⁴⁷ Evidence level 2+

Use of family history in assessing risk of osteoporosis should include maternal, paternal and sister history.

Family history should include not only a given diagnosis of osteoporosis but also kyphosis and low trauma fracture after age 50.

2.5 Modifiable risk factors

2.5.1 weight

Weight loss or low body mass index (BMI) is an indicator of lower BMD.⁴⁸ In addition, those in the lowest tertile of BMI have a two-fold greater bone loss than those in the highest tertile over two years. Post menopausal women with below average BMI should be considered as being at increased risk of osteoporosis.⁴⁶

2.5.2 smoking

A meta-analysis of studies looking at the effect of smoking found that BMD in smokers was 2% lower with each increasing decade after the menopause than in non-smokers, with a 6% difference at 80 years.⁴⁹ Men who smoke show greater bone loss at the trochanter.⁴¹ Female smokers have been shown to be at greater risk of hip fracture than non-smokers, with the risk increasing in line with cigarette consumption. The level of risk declines on giving up smoking, but is not significantly reduced until 10 years after cessation.⁵⁰

Smokers should be considered at greater risk of osteoporosis than non-smokers, and advised to stop, for this and other reasons.

2.5.3 alcohol

Evidence for alcohol as a risk factor for low BMD is inconsistent, as the majority of studies do not include subjects with excessive alcohol intake.

2.5.4 exercise

A positive relationship between both current physical activity, physical activity in adolescence and BMD has been shown in young female Canadians (18-35 years)⁵¹ and in Italian middle aged women.⁵² Current exercise has been associated with higher bone density in postmenopausal English women⁵³ and in Norwegian women aged 50-75 years with fractures.⁴⁸ However a study of an Australian population has shown that current physical activity was positively associated with BMD but that after adjustment for age, BMI, calcium intake, and quadriceps strength the relationship did not remain statistically significant.⁵⁴ Consequently, individuals with a sedentary adolescent lifestyle should be considered at higher risk of osteoporosis. Those who currently have a sedentary lifestyle may also be at higher risk.

2.5.5 diet

Past dietary intake of milk in adult premenopausal women (45-49 years) has been positively associated with BMD.⁵⁵ Evidence of association between current calcium intake and low BMD is inconsistent. Vitamin D levels have been shown to be positively correlated with BMD in independent living men and women aged >80 years in Stockholm.⁵⁶ No consistent association has been found between other dietary factors and BMD.

2.6 Secondary causes of osteoporosis

A large number of clinical conditions and some categories of drugs have been associated with osteoporosis in adults. The most common conditions associated with osteoporosis are:

Anorexia nervosa Chronic liver disease Coeliac disease Hyperparathyroidism Inflammatory bowel disease

Male hypogonadism Renal disease Rheumatoid arthritis Long term corticosteroid use Vitamin D deficiency

The possibility of osteoporosis should be considered in patients with the conditions listed above.

Osteoporosis is also associated with the contraceptive Depo-Provera®. At the present time, there is no 'evidence based' answer to the concerns about adverse effects of Depo-Provera® on bone density as the available data are conflicting.^57,58,59

2.7 Assessing risk of osteoporosis

2.7.1 risk scores

Although there have been attempts to develop risk scores for osteoporosis most are not of satisfactory quality and have not been validated. One tool⁶⁰ was evaluated in 1,013 white American women and had a sensitivity of 98% and a specificity of 12% in that group. The positive predictive value was 69% and the negative predictive value was 75%. Other risk tools have assessed fracture risk. The best performing has a receiver operating characteristic (ROC) curve (graph indicating specificity and sensitivity at all possible cut-off points) of 0.88 (values >0.8 are generally required to indicate effectiveness of a test) but this has not been validated in a separate population.⁶¹

The development of a validated tool of this kind would be a useful addition to the diagnosis of osteoporosis.

2.7.2 who is at highest risk?

It must be clear which risk is being considered: the risk of osteoporosis, or the risk of falling or fracturing. Each of these is linked and osteoporosis is only one risk factor for fracture.

Priority should be given to finding and managing patients at the highest risk of falling and experiencing a fracture. Those who have already experienced a fracture are at high risk of a further fracture. Thus patients with a previous fracture are a key target group.

The next group to target are those with osteoporosis risk who have not yet sustained a fracture. The risk factors that have best evidence for increasing risk for this group are shown in Table 1. Although many are not modifiable, these factors contribute to a threshold for diagnostic testing, which helps prioritise which patients should be sent for a DXA scan (See section 3.5).

Table 1: Risk factors for osteoporosis (when no history of fracture)

It is difficult to offer evidence based advice about particular combinations of risk factors which justify further investigation since the evidence is lacking, but there seems to be an additive effect of risk factors - more present means greater risk.⁶⁰ A systematic approach to offering osteoporosis assessment to all such patients should be developed, though scarce resources should be targeted at those at highest risk to ensure the most efficient use of these resources.