By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
Low body weight has long been recognized as a risk factor for fracture in the elderly. The increased risk associated with low body mass index (BMI) may be the result of several factors, including low bone mineral density (BMD), less soft tissue that may protect bone from impact forces, and increased fall risk resulting from muscle weakness. Most recommendations for the clinical care of osteoporosis appropriately emphasize the need to consider low weight as a risk factor for fracture.
On the other hand, attention has only recently turned to fractures in overweight or obese individuals. With worldwide increases in both BMI and age, it has never been more important to understand the risks of osteoporosis in that segment of the population. Here we review reports that suggest that fractures in overweight and obese individuals are a substantial public health problem, that the interaction between high BMI and fracture risk factors is complex, and that developing methods for gauging fracture risk in this group should be an important goal. Furthermore, we propose research and policy goals with the aim of developing fracture prevention strategies in the overweight and obese elderly. Other recent reviews have discussed the mechanisms that might link obesity and adiposity to bone and fracture risk.1–6
Large Proportion of Elderly That Are Overweight or Obese, Compared to Underweight
We are currently faced with the highest numbers of overweight and obese older persons (BMI ≥ 25 kg/m2) ever observed; 78% of men and 59% of women ≥60 years old are affected in the United States7; 34% of women and 37% of men are frankly obese (BMI ≥ 30 kg/m2).7 Similarly, the majority of people over 65 years old are overweight or obese in European countries and Australia.8, 9 On the other hand, those who are underweight make up a very small proportion of Americans (0.7% of people aged 60–69 years and 2.4% of those ≥70 years old).7 There is a lower prevalence of obesity in Asian populations but the proportion affected is rapidly increasing.10
Fracture Risk Associations With BMI
Older people with low BMI (<18.5 kg/m2) are clearly at increased risk for fracture, particularly hip fracture. This association has been confirmed in several studies that have included both sexes and a range of populations.11–16 The effect of high BMI on fracture incidence is less clear. An important meta-analysis of seven studies that evaluated BMI as a risk factor for fracture demonstrated a nonlinear pattern in the relationship between fracture risk and BMI.11 There was a marked increase in fracture risk from normal to low BMI (<25 kg/m2), and the reduction in risk was particularly large for hip fracture (approximately one-quarter the risk of those with normal BMI). However, there was a much more modest reduction in the risk of fracture in overweight and obese people compared to those of normal BMI.11 For instance, the risk of hip fracture was 17% lower with BMI of 30 kg/m2 compared to those with a BMI of 25 kg/m2 (De Laet and colleagues11) (Fig. 1). Armstrong and colleagues17 also found that in women the protective effect of increasing BMI on hip fracture risk was most notable at low BMI ranges, and that increasing BMI above the normal range had much less impact on risk, emphasizing that the BMI-fracture risk relationship is nonlinear. Even more recently, an analysis in the Global Longitudinal Study of Osteoporosis in Women (GLOW), involving >60,000 postmenopausal women in 10 countries, confirmed the increased risk of fracture in underweight women but showed no protective effect of obesity in women.18
In large part, risk reduction at higher levels of BMI in the meta-analysis by De Laet and colleagues11 appeared to be related to a trend toward higher BMD in those with greater BMI, since the risk of fracture (all fractures, osteoporotic fracture, or hip fracture) was not reduced in those with high BMI compared to those of normal weight after adjustment for BMD (Fig. 1). In an analysis by Johansson and colleagues19 of women in the WHO combined cohorts, there was a slight increase in osteoporotic fracture risk at high BMI after BMD adjustment.
The influence of high BMI on fracture risk may differ by anatomic site. For example, Compston and colleagues18 recently reported that obese women had more ankle and upper leg fractures than non-obese women (GLOW). And Gnudi and colleagues16 reported that obese postmenopausal women were at higher risk than those with normal BMI for humerus fracture but not other osteoporotic fracture types. Of particular interest, due to its relatively high prevalence among older adults, is vertebral fracture. A recent study of men and women in Australia found multiple measures of adiposity to be associated with vertebral deformity in women but not in men.20 And a report from Italy demonstrated that postmenopausal, osteoporotic women with vertebral fractures had a higher mean BMI (26.5 kg/m2) than those without fractures (24.8 kg/m2, p = 0.002).21 However, the majority of studies have found lower mean BMI in vertebral fracture cases or no independent association with BMI for vertebral fracture.22–26 The subclinical nature of most vertebral fractures and the difficulty of accurately measuring thoracic spine BMD by dual-energy X-ray absorptiometry (DXA) present particular challenges to the evaluation of adiposity and obesity on vertebral BMD and fracture risk.
BMI and total body weight are the most common body size metrics evaluated in association with fracture risk. However, these variables are imperfect markers of adiposity. Fat mass, hip girth, and waist:hip ratio may be useful as more refined measures of body composition. For instance, Pluijm and colleagues27 reported that fat mass and muscle mass were both associated with BMD in women, but fat mass was particularly important. However, in the Study of Osteoporotic Fractures, similar associations with hip fracture were found for hip girth, fat mass, percent body fat, and BMI. As for total body weight, these measures were inversely associated with fracture risk before BMD adjustment and demonstrated a U-shaped risk pattern after adjustment for BMD.28 In women age 75 and older, lower fat mass, but not lean mass, was associated with greater risk of hip fracture; however, linearity of the fat mass association was assumed, and evidence of independence from BMD was not presented.29 How these components of body size individually contribute to the associations of BMI with skeletal outcomes remains uncertain but is of great interest.
In general, these analyses establish a steep gradient of risk of nonvertebral fracture associated with low BMI but also that the effect of BMI on fracture risk is nonlinear. With increasing levels of BMI above normal there is a weak protective effect that may disappear in frank obesity. The protective effect further weakens or even reverses with BMD adjustment. Most of the data available to date are in women (see “Associations Between BMI and Fracture Risk by Age and Sex” section for a discussion of sex differences).
Predominance of Overweight and Obesity Among Fracture Cases
The small decrease in fracture risk in those with high BMI coupled with the fact that the majority of elderly are overweight or obese suggest that a high proportion of fractures may occur in those with BMI > 25 kg/m2. In fact, the mean BMI for postmenopausal women presenting with a fracture in a UK study was 27.4 kg/m2, and 28% of cases were obese (BMI ≥ 30 kg/m2).30 In the Million Women Study in the UK, almost one-half of all postmenopausal hip fractures occurred in women who were overweight or obese: 40% were overweight (BMI 25–29.9 kg/m2) and 9% were obese (BMI ≥ 30 kg/m2).17 In a large longitudinal study of older U.S. men, 68% of all incident clinical fractures and 62% of incident hip fractures occurred in those who were overweight or obese; 19% of all clinical fractures and 13% of hip fractures occurred in obese men.31 For this Perspective we performed an analysis of the National Health and Nutrition Examination Survey I (NHANES) (see Supporting Information Appendix for analytical methods) and again found that the majority of men over 65 years old who experience hip fracture were overweight or obese (Fig. 2, Table 1). Similarly, among women in the NHANES study, 46% of hip fractures occurred in those who were overweight or obese (Fig. 2, Table 1). From Figure 2 it is clear that although the incidence of fracture is higher in those with the lowest BMI, the preponderance of fractures occur in heavier individuals. Thus, although the individual risk for fracture is higher in those with low weight, the population burden of fracture is now in large part the result of fractures in the overweight and obese. This has enormous practical implications, as the challenge of identifying older men and women at highest risk of fracture must be directed at the heaviest segments of the population.
Table 1. Hip Fracture by BMI Category in NHANES and NHEFS (Ages 65–74 Years)
Underweight (<18.5 kg/m2)
Normal (18.5–24.9 kg/m2)
Overweight (25–29.9 kg/m2)
Obese (≥30 kg/m2)
Incidence is per 10,000 person-years.
BMI = body mass index; CI = confidence interval; NHANES = National Health and Nutrition Examination Survey I; NHEFS = NHANES I Epidemiologic Follow-Up Study.
Cases, n (% of all cases)
Age-standardized incidence (95% CI)
Cases, n (% of all cases)
Age-standardized incidence (95% CI)
Associations Between BMI and Fracture Risk by Age and Sex
The association between obesity and fractures in children has also received recent attention; Goulding and colleagues32 first noted that childhood fractures (predominantly forearm) are more frequent in heavier children. In some reports, obese children with fracture had somewhat higher levels of radial bone mineral content (BMC)32 while in others, fractured kids had lower BMC.33, 34 Increased risk of fracture in overweight children may be because of greater fall forces or a greater tendency to fall, but other effects of fat on bone maturation have also been invoked as a potential explanation.1, 4
In older adults, the meta-analysis by De Laet and colleagues11 found that after BMD adjustment there was a small but significant protective effect of higher BMI on hip fracture in women but not in men. The lack of a clear effect in men may have been due to lower sample size (men represented 25% of the overall sample of 59,644 participants) or to a difference in the effect of BMI on BMD and fracture risk in men. In the Osteoporotic Fractures in Men Study (MrOS), obese men were actually at significantly higher risk than normal weight men after BMD adjustment. This was particularly obvious with hip fracture: obese men were five-fold more likely to experience a hip fracture than normal weight men when BMD was taken into account.31 In NHANES women, as in the reports by De Laet and colleagues11 and Johansson and colleagues,19 obesity, unadjusted for BMD, was associated with a reduced risk of fracture (Table 1, Fig. 2) but instead, as in MrOS, obese men in NHANES were at slightly higher risk. These results from MrOS and NHANES seem to be at odds with those from De Laet and colleagues,11 but in the latter report the relative risk of sex specific BMI-fracture associations as compared to normal BMI were not included.
Other reports in men similarly concluded that there was an increased risk of osteoporotic fracture for those with greater percent body fat35 and for those with higher waist circumference or waist:hip ratio.36 A small nested case-control study of men concluded there was no effect of abdominal fat, weight, or BMI on hip fracture risk.37 However, results were presented only for linear associations among body size and adiposity variables, but as discussed above, those associations appear in fact to be nonlinear.
Why obese men may be at higher risk of fracture than normal weight men, whereas obesity slightly mitigates risk in some studies of women, is uncertain. Trochanteric soft tissue thickness is greater in obese women than thinner women, which may reduce hip fracture risk by reducing the forces experienced by the proximal femur during a fall.38 In a longitudinal study of men, however, tissue thickness was not associated with hip fracture risk.39 A potential explanation for the differences between men and women may be that trochanteric soft tissue thickness in men appeared to be considerably lower than in women regardless of BMI39 and hence the increased fall forces associated with greater weight are not effectively dissipated even in obese men. Thus, sex-specific differences in weight distribution (greater hip fat in women) may afford more protection against hip fracture in women and provide at least a partial explanation for a sex difference in the BMI–hip fracture relationship.
Obesity and Temporal Trends in Hip Fracture Rates
Is the obesity epidemic driving changes in the incidence of hip fractures? In Western societies the rates of hip fracture have been noted to be gradually declining in recent decades, particularly in women, and increasing rates of obesity have been postulated to be one of the reasons for the trend.40, 41 Although the data we review here would suggest that overweight/obesity exerts only a modest protective effect on hip fracture risk in women, it may explain some of the trend in declining fracture rates. Obesity may increase the risk for fracture in men, potentially providing some of the reason fracture rates have appeared to decline less steeply in men. Importantly, temporal trends in obesity prevalence have co-occurred with trends in smoking cessation, osteoporosis screening and treatment, and perhaps other lifestyle and health care changes that may have contributed to lower fracture risk.40, 41 It would be difficult to attribute the trend in hip fracture to any of these ecological observations individually. The complexity of risk factors for fracture should be more carefully evaluated in obese people.
Factors Associated With Fracture Risk in the Overweight and Obese
An increase in the number of fractures can be expected as populations age. Assuming a constant age-specific risk of hip fracture, the expected increase in the number of Americans over 65 years old could be accompanied by a three-fold rise in the number of hip fractures from 1990 to 2050.42 If the incidence of fracture in the obese elderly also remains constant during that period, a large proportion of fractures can be expected to occur in overweight and obese individuals. Thus, approaches for the identification of those overweight and obese individuals at higher risk must be developed. Although traditional risk factors for fragility fracture, such as low BMD and history of previous fractures, have been linked to incident fracture in obese women,43 there may also be differences in fracture risk factors for those with high BMI.
There are few studies that specifically address the issue of risk factors for fracture in the overweight and obese. Previous reports (eg, Cummings and colleagues44, 45 and Nguyen and colleagues46) have demonstrated the importance of several characteristics that are associated with fracture risk independent of BMI, including increased age, lower BMD, history of previous fracture, family history, and others. In other words, these traits are important regardless of BMI and can be useful in identifying those at higher risk.
However, known fracture risk factors may be more prevalent among overweight or obese people or may have different associations with fracture depending on BMI. An early report on hip and forearm fractures in women supports the likelihood that effects of BMI also depend on other risk factors, such as hormone use and smoking. Investigators found that forearm fracture was not significantly associated with BMI among estrogen users, but, among estrogen nonusers, being underweight, particularly among smokers, greatly raised the risk of forearm fracture.47 Geographic and temporal trends in smoking and hormone use may also explain differences we observe in the associations between BMI and fracture.
Some conditions that increase fracture risk are particularly germane in the overweight and obese. For instance, diabetes is more common in those with higher BMI and is associated with increased fracture risk48 and vitamin D levels are lower in obesity.49–52 In MrOS, an adjustment for physical performance reduced the strong association between obesity and hip fracture in men, suggesting that part of the increased fracture risk associated with obesity may be related to increased fall risk.31 In fact, adiposity is known from other studies to impair physical performance.53, 54 Compston and colleagues18 reported that among obese women with incident fracture there was a greater frequency of falls, use of arms to assist standing from a seated position, fair or poor self-rated health, and early menopause than among nonobese women with incident fracture. In men, obesity is associated with lower testosterone levels that are in turn linked to more frequent falls.55
Adiposity itself has been linked to abnormalities in muscle mass and function, a complex pathology that results in “sarcopenic obesity.”56, 57 Obesity is associated with fat infiltration in muscle, and Lang and colleagues58 reported that greater fat infiltration into thigh muscle, as assessed by CT scanning, is independently associated with increased fracture risk. These studies suggest risk factors that may differ in their strength or prevalence by BMI category and deserve further evaluation to inform fracture prevention targets in overweight and obese individuals.
Low BMD as a Risk Factor for Fracture in Those With High BMI
Bone mineral density (BMD) tends to be higher in those with greater weight or BMI.59, 60 In classical longitudinal studies, Heaney and colleagues61 reported that increases in weight in women over 21 years of age were associated with increases in cortical area, and most studies suggest a correlation between BMI and BMD of 0.4 to 0.6.62, 63 In the Unite States. the increase in the prevalence of overweight between NHANES III (1988–1994) and NHANES 1999–2002 was 6 percentage points, and resulted in an estimated 1.3 percentage point decrease in osteoporosis.63 In relation to fracture risk, it is important to note that only 6% of fracture cases are estimated to have osteoporosis (BMD more than 2.5 SD below the reference mean).64 Our evaluation of National Health and Nutrition Examination Survey (NHANES III) data (Fig. 3, Table 1) revealed that approximately one-third of overweight and obese Americans age 45 and older have low bone mass (total hip BMD between 1 and 2.5 SD below the young reference mean), and many osteoporotic fractures occur at those levels of BMD.64, 65 Within these higher BMI categories, the prevalence of osteoporosis and low bone mass was higher in women than men and higher in whites and Mexican Americans than in African Americans (Fig. 3). Thus, although higher BMI may provide the benefit of greater BMD in many individuals, it does not guarantee that bone mass is ideal, and in a proportion of the overweight and obese population bone fragility may contribute to fracture propensity. This may be particularly germane in light of the greater mechanical forces generated by a fall in heavier people. In fact, low bone density and previous fracture history were recently identified as a risk factor for nonvertebral fracture in obese women.43
Understanding the utility of low BMD as a risk factor for fracture in overweight and obese people is important because it demands the development of efficient, cost-effective bone density screening strategies in that segment of the population. Identification of additional factors that predict low bone mass in the overweight and obese population is also necessary to design more effective screening approaches and enhance the ability to develop targeted preventive strategies. The use of the WHO FRAX tool should be valid in assessing fracture risk in the obese if the complex associations between BMI and fracture risk are adequately represented in the WHO reference populations and the FRAX algorithm itself.
Bone Density Testing in Overweight and Obese Individuals
DXA artifacts may hinder scan interpretation in obesity because large amounts of soft tissue can increase the apparent density of bone and may obscure bone edges, making the analysis of DXA data more problematic.66 These artifacts are more problematic in assessing spinal bone mass, but an overlying fat panniculus can affect femur BMD and must be lifted during scanning to ensure proper measurement.67 Other relevant potential sources of error include marrow fat and distance of the femur from the table.68
Osteoporosis Screening in Overweight and Obese Individuals
With a substantial obese/overweight population in Western societies, there is a need to focus on recommendations for appropriate risk screening and fracture prevention strategies for overweight and obese elderly. A mistaken belief that overweight or obese individuals are unlikely to experience fracture may affect health care providers' referrals for risk screening and reduce patients' acceptance of screening recommendations. Considering the existing focus on the risk of underweight individuals for osteoporosis, it comes as no surprise that the rate of DXA screening declines as BMI increases.69 This is likely to be due to a combination of provider and patient inattention to fracture as a legitimate health risk. The challenge is to devise fracture risk detection approaches that recognize the relatively lower risk of fractures in those with high BMI with the need to reduce the personal and population burden of their fractures.
In the United States, osteoporosis screening is recommended for women aged 65 years and over, men aged 70 years and over, and for younger people who have risk factors for fracture.70–72 However, many patient and provider factors lead to noncompliance.73 In a survey of older adults, increased weight was associated with both level of belief in susceptibility to osteoporosis and efficacy of screening.74 Although screening acceptance for osteoporosis has not been as well-studied as for cancer screening, for instance, it is likely that weight presents a similar patient barrier for DXA screening as for mammography.75 Comorbidities that are more frequent in obese people may also play a role in a lower frequency of screening.69, 73
While DXA screening is one area of focus, screening among overweight and obese populations for low and declining physical function, fall risk, and insufficiency of vitamin D and other nutrients may have even greater returns in improving not just bone health but also mobility and quality of life.
Implications of High BMI for Adverse Outcomes Following Fractures
Postfracture outcomes, including medical complications, mortality, and disability, are known to be worse among those with more comorbidities.76–78 Although there is a gap in research regarding the effects of overweight and obesity on fracture outcomes, we would expect obese patients to have lower physical functioning79 and more comorbidities80 than normal-weight patients. However, in a study of hip fracture cases in Rochester, MN, USA, the rate of cardiac events in the year after hip fracture repair was found to be more common in patients with low BMI than among those with normal BMI and was not higher in obese patients,81 potentially because of poor caloric reserve in those with low BMI. The risk of all noncardiac medical complications following hip fracture did not differ by BMI level.82 There appears to be little information concerning other potential adverse consequences of fracture in the obese elderly (eg, functional limitation or loss of independence), and the issue deserves further examination, particularly in light of the prevalence of fracture in those with obesity. Losses in BMD and lean body mass and gains in fat have been observed for hip fracture cases.83 If obese cases start with relatively lower lean and greater fat, the impact on survival, physical function, and repeated fractures may be greater than for normal-weight cases.
Effect of Obesity on Osteoporosis Treatment
Few analyses have specifically examined the associations between obesity and the response to pharmacological intervention, but some studies suggest that treatment effects are not worsened by greater levels of BMI. In fact, Eastell and colleagues84 reported that overweight and obese women had greater vertebral fracture risk reduction than normal weight women in response to zoledronate therapy in the HORIZON Trial. Although additional experience may be needed, these results suggest that pharmacological intervention can be effective in the overweight and obese, and should be considered to reduce fracture risk in those at higher risk of fracture. Unfortunately, obese women may receive therapy less often than others. In the GLOW study, obese women with incident fractures were receiving osteoporosis therapies less often than were normal or underweight women with fractures,18 emphasizing the challenge in identifying heavier women who should be treated.
Weight Loss Among the Elderly
The health risks of obesity have resulted in unequivocal recommendations for adults to maintain a healthy weight.80 However, in the elderly, weight loss has been clearly linked to concomitant losses of BMD and lean mass. Indeed, loss of weight in observational studies, even in the obese, has been demonstrated to increase the rate of BMD loss,85 regardless of whether weight loss is voluntary or involuntary. And the lean mass loss that most often dominates weight loss in older adults is independently associated with BMD loss.86 Weight loss has been linked to falls.87, 88 These adverse effects of weight loss on bone and muscle have raised concerns about recommending weight loss in the elderly who may be at increased fracture risk.85 Some randomized controlled trials of weight loss in the elderly have been proven effective and have demonstrated improvements in quality of life, cardiovascular health indicators, and physical function.89 Recently Villareal and colleagues90 reported that weight loss among obese older adults achieved through a regimen that included physical activity was accompanied by very small losses of lean mass (3%) and bone density (1%). These losses were much smaller than in the diet-only weight loss regimen and were accompanied by gains in strength, balance, and gait.90 The question remains whether such gains in physical function would ameliorate fall risk to the extent that they compensate for BMD loss in the fracture-risk calculus. We hypothesize that if weight loss can be achieved while maintaining or increasing lean mass and bone strength, the benefits to elderly participants would accrue in multiple domains of health and function, including fracture prevention. In some overweight/obese individuals at high risk of fracture, it may be reasonable to utilize pharmacological approaches to reduce bone loss (eg, bisphosphonates) while attempting weight loss, particularly if physical activity is restricted or unlikely.
Recommendations for Further Research and Policy Development
The contributions to research in the area of overweight/obesity and fracture risk to date have been predominantly incidental findings of observational studies or aging cohorts who have naturally accumulated participants with high BMI. We see a clear need for studies designed specifically to examine the natural history and risk factors for osteoporosis in the elderly with high BMI. Specific questions include: (1) the incidence of and risk factors for fracture, including diet, comorbidities such as type 2 diabetes, body composition, falls, and physical function; (2) efficient and effective approaches for the identification of overweight and obese individuals at highest risk of fracture, and most likely to benefit from prevention; (3) postfracture outcomes, including maintenance of physical function and independence, repeated fracture, and mortality; and (4) near-term and long-term effects on bone of intentional weight loss. Recruiting and retaining obese individuals in epidemiologic studies, which often underrepresent populations with comorbidities, will be a challenge for designing and interpreting future studies of high BMI and fracture in the elderly.91, 92
Many, if not most, osteoporotic fractures occur in the overweight and obese elderly, but the importance of those fractures may have been lost in the message that being underweight increases the risk of fracture. Additional research is needed to identify the unique mechanisms that link obesity to adverse effects on bone strength, physical performance, and fracture risk, and to develop clinical algorithms that are useful in detecting overweight and obese individuals at highest fracture risk. In parallel, a key public health goal should be to raise awareness among both patients and health providers of the potential for osteoporosis and fracture despite high BMI. The effectiveness of preventive and treatment strategies must be evaluated in the overweight and obese. Weight loss in the elderly continues to be a controversial topic due to concerns with concomitant losses of bone and muscle; however, we may do a disservice to the public if we counsel avoidance of weight loss for these reasons, when the health benefits in other domains are clear. Instead, our responsibility to overweight and obese elderly populations is to identify weight-loss interventions that allow them to maintain bone and muscle strength. Our hope is that future research will reveal osteoporosis prevention targets effective for the growing population of aging men and women with high BMI.
Dr. Nielson is supported by the Office of Research on Women's Health and the National Institute of Child Health and Human Development, Building Interdisciplinary Research Careers in Women's Health (BIRCWH) grant number HD043488-08. Dr. Orwoll is supported by grants from the National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS), the National Institute on Aging (NIA), and the National Center for Research Resources (NCRR).
Authors' roles: Drs. Nielson and Orwoll were involved in all aspects of the manuscript. Ms. Srikanth was involved in data analysis and manuscript preparation.