• BMI;
  • WOMEN;


  1. Top of page
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

Several recent studies suggest that obesity may be a risk factor for fracture. The aim of this study was to investigate the association between body mass index (BMI) and future fracture risk at different skeletal sites. In prospective cohorts from more than 25 countries, baseline data on BMI were available in 398,610 women with an average age of 63 (range, 20–105) years and follow up of 2.2 million person-years during which 30,280 osteoporotic fractures (6457 hip fractures) occurred. Femoral neck BMD was measured in 108,267 of these women. Obesity (BMI ≥ 30 kg/m2) was present in 22%. A majority of osteoporotic fractures (81%) and hip fractures (87%) arose in non-obese women. Compared to a BMI of 25 kg/m2, the hazard ratio (HR) for osteoporotic fracture at a BMI of 35 kg/m2 was 0.87 (95% confidence interval [CI], 0.85–0.90). When adjusted for bone mineral density (BMD), however, the same comparison showed that the HR for osteoporotic fracture was increased (HR, 1.16; 95% CI, 1.09–1.23). Low BMI is a risk factor for hip and all osteoporotic fracture, but is a protective factor for lower leg fracture, whereas high BMI is a risk factor for upper arm (humerus and elbow) fracture. When adjusted for BMD, low BMI remained a risk factor for hip fracture but was protective for osteoporotic fracture, tibia and fibula fracture, distal forearm fracture, and upper arm fracture. When adjusted for BMD, high BMI remained a risk factor for upper arm fracture but was also a risk factor for all osteoporotic fractures. The association between BMI and fracture risk is complex, differs across skeletal sites, and is modified by the interaction between BMI and BMD. At a population level, high BMI remains a protective factor for most sites of fragility fracture. The contribution of increasing population rates of obesity to apparent decreases in fracture rates should be explored. © 2014 American Society for Bone and Mineral Research.


  1. Top of page
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

Fractures are an important cause of morbidity in the population, especially in women. Hip fractures in particular are a major cause of pain, loss of function, and increased mortality, and are associated with very high costs to society.[1-3] Because fracture incidence increases with age, the burden from fracture is predicted to increase in the future due to an increase in the elderly population.[3-5]

In addition to low bone mineral density (BMD), many risk factors for fragility fractures have been identified.[2, 6, 7] Strong risk factors include a prior fragility fracture, a family history of fracture, exposure to glucocorticoids, and low body mass index (BMI).[8-11] Low BMI has been considered a risk factor for fracture, and obesity has been considered a protective factor for fracture,[11-13] but this association has recently been challenged.[14, 15] Compston and colleagues[15] reported that obesity was not protective against fracture in postmenopausal women and, indeed, was associated with an increased risk of ankle and upper leg fractures. Similarly, Prieto-Alhambra and colleagues[16] concluded that obesity, though protective against hip and pelvis fracture, was associated with an increase in risk for proximal humerus fractures. In a recent review, Nielson and colleagues[17] stated that the importance of fractures occurring in the overweight and obese elderly may have been lost in the message that being underweight increases the risk of fracture.

The aim of this study was to investigate the association between BMI and future fracture risk at different skeletal sites in 25 international prospective cohorts comprising almost 400,000 women.

Subjects and Methods

  1. Top of page
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

Cohorts studied

We used baseline and follow-up data from 25 prospective cohorts, the majority of which were population based (20/25). Details of each of the cohorts are published elsewhere, but are summarized briefly below and in Tables 1, 2, and 3.

The Adult Health Study (AHS) at the Radiation Effects Research Foundation was established in 1958 to document the late health effects of radiation exposure among atomic bomb survivors in Hiroshima and Nagasaki, Japan. The original AHS cohort consisted of about 15,000 atomic bomb survivors and 5000 controls selected from residents in Hiroshima and Nagasaki using the 1950 national census supplementary schedules and the Atomic Bomb Survivors Survey. AHS subjects have been followed through biennial medical examinations since 1958.[18, 19] In the Aberdeen Prospective Osteoporosis Screening Study from the UK (APOSS),[20] women were randomly selected from a community-based register and invited to participate in a population-based screening program for osteoporotic fracture risk. The Canadian Multicentre Osteoporosis study (CaMos) is an ongoing prospective age-stratified cohort of men and women ages 25 to 80+ randomly selected from regional residential telephone listings. The sampling frame was a 50-km radius around nine study centers in seven provinces, and participants are representative of 41% of the population of Canada.[21] The Dubbo Osteoporosis Epidemiology Study (DOES) is a population-based study from Dubbo, Australia.[22] The Ecografía Osea en Atención Primaria (ECOSAP) study was a referral population recruited in 58 primary care center throughout Spain, regardless of the reason for consultation.[23] The Norfolk cohort of the European Prospective Investigation into Cancer (EPIC-Norfolk) comprises men and women aged 40 to 79 years who were resident in Norfolk, UK, at the time of recruitment and were recruited from general practice listings.[24] The Epidemiologie de l'osteoporose (EPIDOS) study comprises a population-based cohort from five French centers (Amiens, Lyon, Montpellier, Paris, and Toulouse)[25]: women were recruited through mailings using large population-based listings such as voter registration rolls. The European Vertebral Osteoporosis Study (EVOS) comprised age- and sex-stratified random samples from 36 centers in 19 European countries.[26] Equal numbers of men and women were drawn in each center within six 5-year age bands (50–74 and 75+ years). BMD was measured in 13 centers. This sample provided the framework for the European Prospective Osteoporosis Study (EPOS), in which repeated assessment was undertaken in 29 of the centers.[27, 28] The Gothenburg I subjects were drawn randomly from the population register in Gothenburg, Sweden, by the date of birth to provide cohorts aged 70, 76, 79, and 85 years at the time of investigation.[29] The Gothenburg II study comprised a randomly drawn population that attended for mammography screening.[30] The Geelong Osteoporosis Study (GOS) is an age-stratified sample of women drawn randomly from the electoral roll of Geelong and surrounding districts in south eastern Australia.[31] The Manitoba cohort is a referral population of all women attending for BMD measurements in the Province of Manitoba, Canada, where health services are provided to residents through a single public healthcare system.[32] The Miyama study is a population-based cohort drawn from inhabitants born in Miyama, Japan, between 1910 and 1949.[33] Of 1543 inhabitants, an age-stratified sample of 400 men and women was drawn by birth decade. The MsOS study is a cohort study on osteoporosis in a convenience sample of ambulant Asian women recruited from the community in Hong Kong.[34] The Os des Femmes de Lyon (OFELY) cohort comprised an age-stratified female cohort randomly selected from the regional section of a large health insurance company (Mutuelle Generale d'Education Nationale, Lyon, France).[35] The Osteoporosis and Ultrasound Study (OPUS) comprises five age-stratified population-based female cohorts drawn from different European centers (Sheffield and Aberdeen in the UK; Berlin and Kiel in Germany; and Paris in France).[36] The Kuopio osteoporosis risk factor and prevention (OSTPRE) study in Finland comprised a postal inquiry sent to all 14,220 women who were residents of Kuopio province.[37] The Prospective Epidemiological Risk Factors (PERF) study was a population-based cohort in Copenhagen, Denmark.[38] The survey invited women to participate in screening for various placebo-controlled clinical trials and epidemiological studies in Copenhagen. The Rochester cohort was recruited from two random population samples of women from Minnesota, USA, stratified by decade of age.[39, 40] The Rotterdam Study is an ongoing prospective cohort study that aimed to examine and follow all residents aged 55 years and older living in Ommoord, a district of Rotterdam, the Netherlands.[41-43] The Swiss Evaluation of the Methods of Measurement of Osteoporotic Fracture Risk (SEMOF) study is a prospective multicenter study (10 centers in Switzerland).[44] Women were randomly selected from an address register. The Sheffield cohort comprised women aged 75 years or more selected randomly from the population of Sheffield, UK, and surrounding districts, identified from general practitioner listings. The women willing to participate and meeting inclusion criteria were randomly allocated to treatment with placebo or the bisphosphonate, clodronate, to study its effects on fracture risk. The subjects for this study comprised 2171 women allocated to treatment with placebo only.[45, 46] The Study of Osteoporotic Fractures (SOF) is a multicenter cohort study of risk factors for osteoporosis and fracture.[47] Participants were ambulatory white women selected by convenience and recruited at four clinical centers from the United States (Baltimore, MD; Minneapolis, MN; Pittsburgh, PA; and Portland, OR, USA). The Health Improvement Network (THIN) research database was derived from computerized records of a sample of general practitioners in the UK, similar to the General Practice Research Database.[48] The study population comprised all women aged 50 years or more. The Women's Health Initiative (WHI) study comprises three overlapping randomized controlled studies and an observational study in a convenience sample of postmenopausal women.[49, 50] The trials comprised dietary modification (low-fat diet) (n = 48,836), hormone replacement therapy (HRT) in women with or without a uterus (n = 27,347), and supplementation with calcium and vitamin D (n = 36,282). The total sample size was 161,808. For this analysis women taking bone active medication (HRT, bisphosphonates, and calcitonin) were excluded, leaving a sample size of 81,377.


Height and weight were measured using standard techniques in all cohorts. BMI was calculated as weight in kilograms divided by height squared in meters and used as a continuous variable or categorized according to the WHO criteria[51]: underweight (BMI < 18.5 kg/m2); normal (18.5–24.9 kg/m2); overweight (25.0–29.9 kg/m2); obese I (30.0–34.9 kg/m2); and obese II (≥35.0 kg/m2). BMD was assessed in 27% of the women using several different techniques summarized in Table 1 and converted to standardized cohort-specific Z-scores. The proportion of women with BMD measurement varied by cohorts from 0% to 100% (Table 2).

Table 1. Cohorts Studied
CohortYear for baselineBone densitometryFracture report
  • AHS = Adult Health Study; BMD = bone mineral density; DXA = dual-energy X-ray absorptiometry; FN = femoral neck; QDR = quantitative digital radiography; APOSS = Aberdeen Prospective Osteoporosis Screening Study; CaMos = Canadian Multicentre Osteoporosis study; DOES = Dubbo Osteoporosis Epidemiology Study; ECOSAP = Ecografía Osea en Atención Primaria; QUS = quantitative ultrasound; EPIC-Norfolk = Norfolk cohort of the European Prospective Investigation into Cancer; EPIDOS = Epidemiologie de l'osteoporose; EVOS = European Vertebral Osteoporosis Study; EPOS = European Prospective Osteoporosis Study; GBG I = Gothenburg I; GBG II = Gothenburg II; GOS = Geelong Osteoporosis Study; Manitoba = Province of Manitoba, Canada; ICD = International Classification of Diseases; Miyama = Miyama, Japan; MsOs HK = osteoporosis in Asian women in Hong Kong; OFELY = Os des Femmes de Lyon; OPUS = Osteoporosis and Ultrasound Study; OSTPRE = osteoporosis risk factor and prevention, Kuopio, Finland; PERF = Prospective Epidemiological Risk Factors; Rochester = two random population samples of women, Minnesota, USA; Rotterdam = ongoing study in Ommoord district, Rotterdam, the Netherlands; SEMOF = Swiss Evaluation of the Methods of Measurement of Osteoporotic Fracture Risk; Sheffield = women ≥75 in Sheffield, UK; THIN = The Health Improvement Network; WHI = Women's Health Initiative.

  • a

    Denotes that the cohort was not population-based.

  • b

    EPIC Norfolk collected QUS data on approximately 15,000 men and women between 1997 and 2000; fractures were ascertained by hospital record linkage.

AHS1958 (BMD: 1994)DXA FN, Hologic QDR 2000Spinal radiographs and self-report
APOSS1990–1994DXA left FN, Norland (Cooper Surgical)Self-report, computer reports from radiologists, hospital record, primary care physicians' record
CaMos1996–1997DXA FN, Hologic QDR and Lunar DPX Alpha phantom-calibrated across centers and machinesSelf-report. Radiographic or medical report verification of incident fractures was obtained when information was available.
DOES1989DXA FN, GE-Lunar, DPX and ProdigyRadiologists' report
ECOSAPa2000–2001QUS right calcaneus, Sahara (Hologic)Self-report, confirmed by investigator by X-ray or radiological or surgical reports
EPIC-Norfolkb1997–2000Hospital record linkage
EPIDOS1992–1993DXA FN, Lunar DPXSelf-report, family, or physician
EVOS/EPOS1989DXA FN, cross-calibrated using European Spine PhantomSelf-reported fractures were confirmed where possible by radiograph, attending physicians or subject interview
GBG I1985–1993Dual photon absorptiometry right heelRadiology departments servicing the region
GBG IIa1992–1997Distal forearm, Osteometer DTX-200Radiology departments servicing the region
GOS1994–1997DXA FN, Lunar DPX-LRadiographically confirmed from hospital records
Manitobaa1990–2007DXA FN, Lunar DPX or Lunar prodigyAscertained using ICD codes, where two or more hospitals or physicians ICD fracture codes had to be present to confirm a fracture. Site-specific orthopedic intervention codes for hip and forearm fractures.
Miyama1989–1990DXA FN, Lunar DPXSelf-report, confirmed by X-ray
MsOs HKa2001DXA FN, Hologic QDR-4, 500-WSelf-report, confirmed by X-ray or medical record
OFELY1992–1993DXA FN, Hologic QDR 2000Radiography, X-rays, surgical reports
OPUS1999–2001DXA FN, Hologic QDR 4500 or Lunar ExpertSpinal radiograph; verification of non-vertebral incident fractures when information was available.
OSTPRE1989DXA FN, Lunar DPXSelf-report
PERF1977–1997DXA FN, Hologic QDR-2000Spinal radiographs and self-report
Rochester1980DXA FN, Hologic QDR 2000 and dual-photon absorptiometry cross-calibrated to DXASelf-report combined with review of the in-patient and outpatient medical records of all local care providers
Rotterdam1990–1993DXA FN, Lunar DPX-LAutomatic link with general practitioner computer systems and hospital admission data. Validated by two independent research physicians.
SEMOF1997–1999DXA FN, Hologic QDR 4500Questionnaire and confirmed from medical records
Sheffield1993–1999DXA FN, Hologic QDR 4500Self-report at home visits
SOFa1986–1988 (BMD: 1990–1991)DXA FN, Hologic QDR 1000Telephone or correspondence and confirmed from X-ray reports
THIN1995–2004General practitioners' records
WHIa1990DXA FN, Hologic 2000Hip fractures by medical records and adjudicated at a central facility. Other fractures were adjudicated locally (clinical trials) and by self report (observational study for patients without BMD).
Table 2. Details of Cohorts Studied
CohortaSubjects (n)Length of follow-up (years), mean (maximum)Age (years), mean (range)BMI (kg/m2) mean (SD)BMD (n)b
  • BMI = body mass index; BMD = bone mineral density.

  • a

    The cohort abbreviations are defined in detail in the Cohorts studied section of Subjects and Methods, and are defined in brief in the footnotes for Table 1.

  • b

    Subjects with BMD data available.

AHS1,8103.8 (6.8)66 (47–95)23.1 (3.6)1,797
APOSS5,1107.0 (12.3)48 (44–56)25.5 (4.6)5,102
CaMos6,3156.0 (8.6)63 (25–103)26.9 (5.2)5,719
DOES1,2707.8 (13.6)71 (57–94)25.4 (4.6)1,259
ECOSAP5,1282.9 (4.5)72 (65-100)29.2 (4.7)
EPIC-Norfolk8,8565.4 (6.9)62 (42–81)26.6 (4.4)
EPIDOS7,5933.4 (5.0)80 (70-100)25.4 (4.2)7,560
EVOS/EPOS9,0133.0 (5.9)64 (41–93)27.2 (4.6)2,761
GBG I1,1587.9 (16.3)79 (69–85)25.3 (4.2)947
GBG II7,06512.4 (16.2)59 (21–89)24.6 (3.6)7,056
GOS1,8636.3 (10.9)63 (35–95)26.8 (5.3)1,805
Manitoba43,8605.3 (18.4)62 (40–102)26.6 (5.4)43,186
Miyama4008.6 (13.0)59 (40–79)22.1 (2.8)400
MsOs HK2,0003.5 (5.3)73 (65–98)23.9 (3.5)2,000
OFELY66810.9 (14.2)62 (50–89)24.0 (3.5)663
OPUS2,8816.0 (8.2)61 (20–81)26.3 (4.6)2,836
OSTPRE3,05810.0 (10.0)52 (47–57)26.1 (4.3)1,743
PERF5,4337.2 (24.0)63 (44–81)25.5 (3.9)2,305
Rochester6558.1 (19.0)58 (21–94)25.5 (4.9)650
Rotterdam4,0685.9 (9.4)70 (55–99)26.7 (4.1)3,325
SEMOF7,0622.8 (4.9)75 (70–91)25.9 (4.3)908
Sheffield2,1703.8 (5.8)80 (74–96)26.7 (4.5)2,150
SOF9,70411.9 (20.6)72 (65–99)26.4 (4.6)7,963
THIN180,0934.7 (13.9)60 (50–105)26.0 (5.1)
WHI81,3777.4 (11.2)64 (49–79)28.6 (6.2)6,132
Totals398,6105.7 (24.0)63 (20–105)26.6 (5.4)108,267

For fracture outcomes, we used information on fractures only at sites considered to be associated with osteoporosis[52]; ie, fractures of the spine, coccyx, ribs, pelvis, humerus, forearm, elbow, hip, other femoral, tibia and fibula, clavicle, scapula, and sternum. Fractures of the skull, face, hands and fingers, feet and toes, ankle, and patella were excluded. In addition to “osteoporotic fractures,” incident hip, distal forearm, lower leg (tibia and/or fibula), and upper arm (humerus and/or elbow) were considered separately.

Statistical methods

Correlation tests between BMI and other variables used nonparametric Pitman's permutation test; Pearson correlation coefficients were also calculated.

The association between BMI and the risk of fracture was examined using an extension of the Poisson regression model[53] in each cohort. The observation period of each participant was divided in intervals of 1 month. The first fracture per person was counted for each relevant outcome. Covariates included current age and time since start of follow-up, and analyses were performed with and without adjustment for BMD. Interactions between BMD and BMI were also studied. The β-coefficients from each cohort were weighted according to the variance, and then merged to determine the weighted mean of the coefficient and its SD. The associations between BMI and risk of fracture were described as the hazard ratio (HR) for fracture per 1-unit change in BMI together with 95% confidence intervals (CIs).

Heterogeneity between cohorts was tested by means of the I2 statistic.[54] Heterogeneity was found for the osteoporotic fracture outcome (I2 = 75%; 95% CI, 63% to 83%) and the hip fracture outcome (I2 = 86%; 95% CI, 81% to 90%). When the interaction between BMI and current age was included, there was no significant heterogeneity between cohorts for BMI (I2 = 14%; 95% CI, 0% to 48%) for the outcome of osteoporotic fracture. For the outcome of hip fracture there was a moderate heterogeneity between cohorts for BMI (I2 = 61%; 95% CI, 39% to 75%). Because we had a moderate heterogeneity for the outcome of hip fracture even when including an interaction with age, we performed both a fixed and a random effect model when merging the result from the different cohorts. Overall the weighted β-coefficient describing the association between BMI and the outcome of osteoporotic fracture was –0.0215 when using a fixed-effect model and –0.0210 when using a random effect model (with a SD describing the variance between cohorts of 0.013), resulting in the same HR per 1-unit of 0.98. When describing the association between BMI and the outcome of hip fracture the β-coefficient was –0.0740 when using a fixed-effect model and –0.0719 when using a random effect model (with a SD of 0.014) resulting in the same HR per 1-unit of 0.93. Because the estimates were so similar, we used the fixed-effect model to present the results.

In order to study the association between BMI and fracture risk in more detail, a spline Poisson regression model was fitted using cohort specific knots at the 10th, 50th, and 90th percentiles of BMI, as recommended by Harrell.[55] The splines were second order functions between the breakpoints and linear functions at the tails, resulting in a smooth curve. When the comparisons between two points at the curve was done, a piecewise linear model with knot at BMI = 25 kg/m2 were used to study the relationship between BMI and the risk of fracture.

In sensitivity analyses, we repeated the calculations (1) in those cohorts that were population-based (see Table 1); (2) in cohorts without excluding women that received treatments for osteoporosis; and (3) using a random-effect rather than a fixed-effect model.


  1. Top of page
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

The cohorts comprised 398,610 women aged 20 to 105 years with an average age of 63 years, who were followed for approximately 2.26 million person-years (Tables 2 and 3). During an average follow-up of 5.7 years 30,280 osteoporotic fractures were documented, of which 6457 were at the hip (Table 3). The mean BMI was 26.6 kg/m2 and approximately one-half of the women were overweight or obese (56%), with 22.1% being obese (Table 4). Approximately 7700 women (1.9%) were underweight. There was a weak but significant negative correlation between age and BMI (p < 0.001; r = −0.01; 95% CI, −0.01 to −0.01). For example, in women aged 55 to 59 years, 1.3% of women were underweight and the proportion increased progressively with age, so that 5.8% of women aged 85 to 89 years were underweight. Conversely, the prevalence of obesity decreased with age from 25.3% in the age group 55 to 59 years to 10.9% between the ages of 85 and 89 years. There was a significant positive correlation between BMI and BMD (p < 0.001; r = 0.33; 95% CI, 0.32–0.33). In underweight women, the mean BMD femoral neck Z-score was –0.89 and for the obese II category it was 0.67 (Table 4).

Table 3. Details of Incident Fractures by Cohort
CohortaPerson-yearsIncident fracture
OsteoporoticHipDistal forearmTibia/fibulaHumerus/elbow
  • – = site of fracture not given.

  • a

    The cohort abbreviations are defined in detail in the Cohorts studied section of Subjects and Methods, and are defined in brief in the footnotes for Table 1.

GBG I9,191255198
GBG II87,5778871164433198
MsOs HK6,9759621438
Age at fracture (years), mean (SD) 72.7 (10.4)79.5 (8.8)71.0 (9.6)69.6 (8.5)73.6 (9.7)
Table 4. Baseline Characteristics by BMI Category
 Underweight (BMI <18.5)Normal (BMI 18.5–24.9)Overweight (BMI 25.0–29.9)Obese I (BMI 30.0–34.9)Obese II (BMI ≥35.0)
  1. Values are mean (SD).

  2. BMI = body mass index (kg/m2); BMD = bone mineral density.

Subjects (n)7,699166,087136,87358,91929,032
Age (years)65.7 (14.0)62.2 (11.6)63.6 (10.7)63.2 (10.1)61.2 (9.3)
BMI (kg/m2)17.2 (1.3)22.5 (1.6)27.2 (1.4)32.0 (1.4)39.3 (4.5)
Femoral neck BMD (Z-score)–0.89 (0.97)–0.25 (0.93)0.12 (0.94)0.41 (0.96)0.67 (1.0)
Subjects with BMD values (n)2,30946,79637,74115,0516,370

BMI and risk of fracture

A total of 30,280 osteoporotic fractures were reported during follow-up (Table 3). A minority (19%) of all osteoporotic fractures occurred in obese women (Table 5) and the observed number was lower than expected (5798 versus 6691, respectively) if BMI was assumed to exert no influence on fracture risk. Thus obesity was a protective factor for osteoporotic fractures as a whole. Similar results were found when hip fracture or distal forearm fractures were considered individually (Table 5). In contrast, the observed incidence of lower leg fractures was not reduced, and the risk of upper arm fractures was higher than expected in obese women.

Table 5. Number of Fractures According to Fracture Outcome and Category of Baseline BMI
Fracture outcomeBMI categoriesaObese versus non-obese
Underweight (1.9%)Normal (41.7%)Overweight (34.3%)Obese I (14.8%)Obese II (7.3%)HR95% CIp
  • Values are the number of fractures in each BMI category and in parentheses are the expected number of fractures according to the percentage of women in each BMI category.

  • BMI = body mass index; HR = hazard ratio; CI = confidence interval.

  • a

    BMI categories (kg/m2): Underweight, BMI <18.5; Normal, BMI 18.5–24.9; Overweight, BMI 25.0–29.9; Obese I, BMI 30.0–34.9; Obese II, BMI ≥35.0. Percentages are the proportion of women in each BMI category.

Osteoporotic806 (575)13,293 (12,627)10,383 (10,386)4119 (4481)1679 (2210)0.850.82–0.88<0.001
Hip320 (123)3257 (2693)2062 (2215)628 (956)190 (471)0.630.59–0.68<0.001
Distal forearm126 (150)3424 (3424)2990 (2816)1202 (1215)468 (599)0.810.76–0.86<0.001
Tibia/fibula10 (36)608 (801)704 (659)361 (284)237 (140)1.040.94–1.14>0.30
Humerus/elbow76 (75)1452 (1649)1399 (1357)694 (585)334 (289)1.211.11–1.31<0.001

When BMI was used as a continuous variable, there was a significant association between BMI and fracture risk (p < 0.001). In the case of all osteoporotic fractures, the HR per unit increase of BMI was 0.98 (95% CI, 0.98–0.98) and for hip fracture it was 0.93 (95% CI, 0.92–0.94). The HR was not, however, uniform across BMI; low BMI was associated with a greater risk than would be predicted from a uniform HR and, conversely, a high BMI contributed less to fracture prevention than expected. Thus, when studying the relationship in more detail with spline functions, the function was steeper below a BMI of 25 kg/m2 than above this value (Fig. 1). When a woman with a BMI of 15 kg/m2 was compared with a woman with a BMI of 25 kg/m2 using piecewise linear functions, the HR was 1.5 (95% CI, 1.4–1.6) for osteoporotic fracture and 2.9 (95% CI, 2.6–3.3) for hip fracture (Table 6). By contrast, if a woman with a BMI of 25 kg/m2 was compared to one with a BMI of 35 kg/m2, the HR was 0.9 (95% CI: 0.9–0.9) for osteoporotic fracture and 0.7 (95% CI = 0.6–0.8) for hip fracture.


Figure 1. Relationship between BMI and risk of fracture (HR versus BMI 25 kg/m2) for osteoporotic fracture (solid line) and hip fracture (dashed line), adjusted for age and time since baseline. BMI = body mass index; HR = hazard ratio.

Download figure to PowerPoint

Table 6. HRs for Fracture and 95% CIs Comparing a BMI of 25 kg/m2 With BMIs of 15 kg/m2 and 35 kg/m2, Respectively, According to Different Fracture Outcomes
Fracture outcomeNot adjusted for BMDAdjusted for BMD
BMI 15 versus 25BMI 35 versus 25BMI 15 versus 25BMI 35 versus 25
  1. Values are HR (95% CI), adjusted for age and time since baseline.

  2. HR = hazard ratio; CI = confidence interval; BMI = body mass index; BMD = bone mineral density.

Osteoporotic1.54 (1.44–1.64)0.87 (0.85–0.90)0.89 (0.80–0.99)1.16 (1.09–1.23)
Hip2.88 (2.56–3.25)0.68 (0.62–0.75)1.41 (1.16–1.72)0.99 (0.86–1.15)
Distal forearm1.05 (0.91–1.20)0.76 (0.71–0.81)0.72 (0.60–0.86)0.97 (0.87–1.07)
Tibia/fibula0.64 (0.45–0.89)1.03 (0.94–1.14)0.34 (0.16–0.74)1.14 (0.87–1.49)
Humerus/elbow1.13 (0.92–1.37)1.18 (1.04–1.27)0.70 (0.54–0.90)1.60 (1.42–1.80)

The use of BMI as a continuous variable also confirmed the different patterns between fracture sites. In the case of upper arm fractures, a BMI of 35 kg/m2 conferred a significantly higher risk than a BMI of 25 kg/m2, whereas a BMI of 15 kg/m2 had a similar risk to that at 25 kg/m2 (Table 6). The lower BMI was associated with a significant reduction in lower leg fractures, whereas the risk was similar at 25 and 35 kg/m2 (Table 6).

Adjustment for BMD

When the association between BMI and hip fracture risk was adjusted for BMD, the association was weaker than in the absence of BMD but was still significantly negative. The HR was 0.99 per 1 kg/m2 increase (95% CI, 0.98–0.99; p = 0.0014). When the relationship was examined with spline functions, the relationship was much flatter with BMD adjustment (Fig. 2) than without (Fig. 1). Not withstanding, the risk of hip fracture with low BMI was greater than the protective effect of a high BMI. Thus, a BMI of 15 kg/m2 had an HR of 1.4 (95% CI, 1.2–1.7) compared to a BMI of 25 kg/m2 (Table 6), but a BMI of 35 kg/m2 conferred no greater hip protection than a BMI of 25 kg/m2 (HR = 1.0; 95% CI, 0.9–1.2).


Figure 2. Relationship between BMI and risk of fracture (HR versus BMI 25 kg/m2) for osteoporotic fracture (solid line) and hip fracture (dashed line), adjusted for age, time since baseline, and BMD. BMI = body mass index; HR = hazard ratio; BMD = bone mineral density.

Download figure to PowerPoint

Interestingly, the association between BMI and osteoporotic fracture risk was weaker but inverted when adjusted for BMD, so that a higher BMI was now associated with a small but significant increase in fracture risk (HR per 1-unit increase in BMI = 1.01; 95% CI, 1.01–1.02; p < 0.001). For example, the HR for all osteoporotic fracture was 1.16 (95% CI, 1.09–1.23) when comparing a BMI of 35 kg/m2 with a BMI of 25 kg/m2; at a BMI of 15 kg/m2, the risk was reduced. Thus, for all osteoporotic fractures a higher BMI was, if anything, a modest albeit significant risk factor following adjustment for BMD. A similar pattern was observed for distal forearm fractures. The association of high BMI with increased fracture risk following adjustment for BMD was most marked for upper arm fractures (Table 6). For lower leg fractures, fracture risk was increased and decreased at high and low BMIs, respectively, compared to 25 kg/m2 (Table 6).

Interactions with BMI

There was a significant interaction between age and BMI for osteoporotic fracture (p < 0.001). This age interaction was significant both below and above a BMI of 25 kg/m2 (p = 0.042 and p < 0.001, respectively). Thus, when BMI was set at 15 kg/m2 and compared with a BMI of 25 kg/m2 using piecewise linear functions, the HR was 1.4 at the age of 50 years and 1.7 at the age of 80 years, suggesting that low BMI was a stronger risk factor for osteoporotic fractures in elderly women. The same age-BMI interaction was true for BMI greater than 25 kg/m2, in that high BMI was a stronger protective factor for elderly women. A significant interaction between age and BMI was seen for hip fracture below a BMI of 25 kg/m2 (p < 0.001), but not for BMI above 25 kg/m2 (p = 0.058). Thus, when BMI, set at 15 kg/m2, was compared with a BMI of 25 kg/m2 using piecewise linear functions, the HR was 9.2 at the age of 50 years and 3.1 at the age of 80 years, indicating that low BMI was a stronger risk factor for hip fracture in younger women than in elderly women.

Because there was a significant correlation between BMD and BMI, and BMD affected the relationship between BMI and the risk of fracture, the interaction between BMI and BMD was investigated with both linear and cubic models. No such interactions were found, indicating that the correlation between BMI and fracture risk did not change for different values of BMD. There were also no significant interactions between BMI and time since baseline; ie, the predictive value of BMI did not change with time (p > 0.20 for both osteoporotic and hip fracture outcomes).

When women allocated to treatments for osteoporosis in the WHI cohort were included, the results were similar. So, too, were the results when the analysis was confined to population-based cohorts.


  1. Top of page
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

The principal finding of the present meta-analysis of predominantly prospective population-based cohorts of women is the significant association between BMI at baseline and future osteoporotic fracture, in that a low BMI was a significant risk factor for all osteoporotic fractures, including hip and forearm fractures. These finding are very consistent with an earlier but smaller meta-analysis,[11] though it should be acknowledged that 11% of the women over a shorter time appeared in both meta-analyses. As previously reported in that study, a high BMI was a protective risk factor for osteoporotic fracture, including hip fracture, but a high BMI was weaker as a protective factor than low BMI was as a risk factor. An important conclusion is that obesity itself is not a risk factor for osteoporotic fracture, hip fracture, or forearm fracture. As also seen in the earlier analysis,[11] the association between BMI and fracture risk was dependent on BMD. In the subset of women in whom femoral neck BMD was measured, the association of BMI with hip fracture risk was attenuated and was not evident for all osteoporotic fractures combined. It should be noted that the HRs with and without adjustment for BMD are not strictly comparable; a minority of women (27%) had a BMD test and there was a significant cohort bias in the proportion of women with a BMD test. With this caveat, the results are consistent with the earlier meta-analysis.

Our results also suggest that the association between BMI and risk of future fracture is site-specific. Whereas low BMI was a risk factor for all osteoporotic fractures, a low BMI was a protective factor for lower leg fracture. In this regard, several of the cohorts did not adequately distinguish fractures of the lower leg that are associated with low BMD (eg, proximal tibial fractures) from ankle fractures which are not regarded as being associated with osteoporosis.[52] Exclusion of these cohorts from the analysis still showed a similar pattern of association of lower leg fractures with BMI (data not shown). In the present study, a high BMI was a significant risk factor for humerus fractures and this persisted after adjustment for BMD. The finding is consistent with a recent short-term (1 year) prospective analysis in 832,775 Spanish women aged 50 years or more visiting general practitioners (SIDIAP),[16] in which a protective effect of obesity was found on future hip fracture and forearm fracture (relative risk [RR] = 0.49; 95% CI, 0.44–0.55, and RR = 0.83; 95% CI, 0.75–0.91, respectively), but obese women were at significantly higher risk of future proximal humeral fracture than the rest of the study population (RR = 1.28; 95% CI, 1.04–1.58). These findings are also consistent with an earlier report that obese women had a higher prevalence of a prior humeral fracture (odds ratio [OR] = 3.48; 95% CI, 0.18–6.68).[56] The reasons for the site-specific association between high BMI and humeral fracture risk are not known, though it may conceivably reflect a different pattern of falling or a greater load upon bones in the upper extremity in falls among the obese population. Moreover, a different padding effect of the soft tissues in different skeletal regions may produce diverse energy dissipation after trauma and, therefore, a different protection of the underlying bone.

Our results are at first sight at variance with the conclusions of Compston and colleagues,[15] who state that that obesity is not protective against fracture in postmenopausal women. That study, however, included a large number of non-adjudicated ankle and tibial fractures. Ankle fractures are not generally regarded as being associated with osteoporosis[51, 56] and, as implied above, the accuracy of a self-reported distinction between ankle and other lower leg fractures is questionable. In their report, ankle fractures were significantly more frequent in obese compared with non-obese women. Given that the incidence of forearm, hip, pelvic, upper leg, and spine fractures was higher in underweight women than in obese women, their report is not inconsistent with our findings. Moreover, the present study also found a protective effect of low BMI for future lower leg fracture.

The question arises whether our findings have implications for the Fracture Risk Assessment Tool (FRAX®), which predicts the probability of a hip and a major fracture based on clinical risk factors such as sex, age, BMI, previous fracture, family history, glucocorticoid use, smoking, alcohol use, and secondary osteoporosis.[57] BMI is used as a continuous variable in FRAX, and BMD can be optionally entered into the model. Data from the meta-analysis of De Laet and colleagues[11] were used in the construct of FRAX. The association between BMI and the risk of hip fracture and other osteoporotic fractures in the present study is nearly identical to that described by De Laet and colleagues[11] in the absence of BMD. After adjustment for BMD, the risk of hip fracture associated with low BMI was attenuated in the same way as that described.[11] In the case of osteoporotic fractures, we have shown a slight though significant increase in risk with increasing BMI (see Table 6). This finding is consistent with the earlier meta-analysis, though the increase in risk was not statistically significant because of the smaller sample size. These considerations indicate that modifications of the FRAX algorithm are not warranted based on the present analysis; a view consistent with a recent report from the SOF study that FRAX is of value predicting fractures in obese women, particularly when used with BMD.[58]

The present study has several limitations, some of which we have discussed. These include the limited sampling frame for BMD measurements, inaccuracies in the estimate of BMD in the presence of a high fat mass, and uncertainties in the coding of some fractures. With regard to the first limitation, our results were similar when HRs not adjusted for BMD were calculated in those 27% of women in whom BMD was measured. The different settings of the cohorts are also a limitation, but that would weaken, not strengthen, an association between BMI and fracture. Conversely, the different settings increase the generalizability of our findings. The greatest limitation is that the present analysis is confined to women. Several lines of evidence suggest that the relationship between BMI and fracture risk may differ in men.[11, 59]

A limitation in the understanding of possible mechanisms is that we have not been able to examine all potential confounding factors (eg, smoking, previous fracture, alcohol, comorbidities). Of possible relevance is the association of type 2 diabetes with high BMI. In a recent large clinical database in Manitoba, Canada, individuals with diabetes had a BMI approximately 3 kg/m2 higher than those without diabetes.[60] Of particular interest, diabetes was associated with a 60% increased risk for major osteoporotic fracture when adjusted for clinical risk factors for fracture including BMI and BMD (HR = 1.61; 95% CI, 1.42–1.83). Thus, the higher risk for osteoporotic fracture for obese women (BMI 35 kg/m2 versus 25 kg/m2) in this report could be related in part to diabetes. Diabetic status was recorded in the present analysis for only 9% of women. In the women that had information on diabetes, the prevalence of diabetes was 3.4% in women with a normal BMI and 6.7% in obese women (data not shown). The small size of the available sample meant that we were unable to examine the impact of diabetes on the relationship between BMI and future fracture risk in more detail. The age interactions, the result with and without BMD and some of the fracture-specific findings might suggest an important role for low physical function and frailty in explaining these associations; but, as was the case for diabetes, we were unable to examine this further.

With these caveats, we conclude that low BMI remains an important clinical risk factor for hip and all osteoporotic fractures combined and that obesity in women is associated with a significant, albeit modest, reduction in fracture risk. In contrast, obese postmenopausal women appear to be at higher risk for humeral fractures than those with normal BMI. Moreover, after adjustment for BMD there is a slight increase in osteoporotic fracture risk with increasing BMI.


  1. Top of page
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

HJ was supported by an ESCEO-AMGEN Osteoporosis Fellowship Award. Amgen had no input into the analysis plan or in the writing of this report. The EPIC-Norfolk study is supported by the Medical Research Council UK (G0401527) and Cancer Research UK (8257). CPRD has received funding from the MHRA, Wellcome Trust, Medical Research Council, NIHR Health Technology Assessment program, Innovative Medicine Initiative, UK Department of Health, Technology Strategy Board, Seventh Framework Programme EU, various universities, contract research organizations, and pharmaceutical companies. The Department of Pharmacoepidemiology & Pharmacotherapy, Utrecht Institute for Pharmaceutical Sciences has received unrestricted funding for pharmacoepidemiological research from GlaxoSmithKline, Novo Nordisk, the private-public funded Top Institute Pharma (; includes co-funding from universities, government, and industry), the Dutch Medicines Evaluation Board, and the Dutch Ministry of Health. The AHS has been conducted at the Radiation Effects Research Foundation (RERF), Hiroshima and Nagasaki, Japan, which is a private, nonprofit foundation funded by the Japanese Ministry of Health, Labour and Welfare (MHLW) and the U.S. Department of Energy (DOE), the latter in part through DOE Award DE-HS0000031 to the National Academy of Sciences. The ECOSAP Study was sponsored by Eli Lilly. The Rochester study was supported by the National Institute of Musculoskeletal and Skin Diseases (R01 AR27065), U.S. Public Health Service. CaMOS was funded by the Canadian Institutes for Health Research with contributions from pharmaceutical and other entities who made no contribution to the study design or execution.

Authors' roles: Study design: HJ, AO, JAK, and EMC. Study conduct: HJ, AO, JAK, and EMC. Data collection: RDC, CC, SRC, ADP, JAE, SF, CCG, DG, DH, KTK, MAK, HK, AZL, EL, WBL, DM, LJM, TWON, JAP, MCZ, FR, JCP, DMR, TvS, and NY. Data analysis: HJ. Data interpretation: HJ, AO, JAK, and EMC. Drafting manuscript: HJ, JAK, and EMC. Revising manuscript content: HJ, AO, JAK, EMC, RDC, CC, SRC, ADP, JAE, SF, CCG, DG, DH, KTK, MAK, HK, AZL, EL, WBL, DM, LJM, TWON, JAP, MCZ, FR, JCP, DMR, TvS, and NY. Approving final version of manuscript: HJ, AO, JAK, EMC, RDC, CC, SRC, ADP, JAE, SF, DG, CCG, DH, KTK, MAK, HK, AZL, EL, WBL, DM, LJM, TWON, JAP, MCZ, FR, JCP, DMR, TvS, and NY. HJ takes responsibility for the integrity of the data analysis.


  1. Top of page
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  • 1
    Cooper C, Atkinson EJ, Jacobsen SJ, O'Fallon WM. Melton LJ 3rd. Population-based study of survival after osteoporotic fractures. Am J Epidemiol. 1993; 137(9):10015.
  • 2
    Kanis JA. on behalf of the World Health Organization Scientific Group. Assessment of osteoporosis at the primary health-care level. Technical Report. Sheffield, UK: WHO Collaborating Centre, University of Sheffield; 2008.
  • 3
    Ström O, Borgström F, Kanis JA, Compston J, Cooper C, McCloskey EV, Jönsson B. Osteoporosis: burden, health care provision and opportunities in the EU. A report prepared in collaboration with the International Osteoporosis Foundation (IOF) and the European Federation of Pharmaceutical Industry Associations (EFPIA). Arch Osteoporos. 2011 Dec; 6(1–2):59155. doi: 10.1007/s11657-011-0060-1
  • 4
    Cooper C, Campion G, Melton LJ 3rd. Hip fractures in the elderly: a world-wide projection. Osteoporos Int. 1992; 2(6):2859.
  • 5
    Gullberg B, Johnell O, Kanis JA. World-wide projections for hip fracture. Osteoporos Int. 1997; 7(5):40713.
  • 6
    Kanis JA, Oden A, Johnell O, Johansson H, De Laet C, Brown J, Burckhardt P, Cooper C, Christiansen C, Cummings S, Eisman JA, Fujiwara S, Gluer C, Goltzman D, Hans D, Krieg MA, La Croix A, McCloskey E, Mellstrom D, Melton LJ 3rd, Pols H, Reeve J, Sanders K, Schott AM, Silman A, Torgerson D, van Staa T, Watts NB, Yoshimura N. The use of clinical risk factors enhances the performance of BMD in the prediction of hip and osteoporotic fractures in men and women. Osteoporos Int. 2007; 18(8):103346.
  • 7
    Kanis JA. Diagnosis of osteoporosis and assessment of fracture risk. Lancet. 2002; 359(9321):192936.
  • 8
    Kanis JA, Johansson H, Oden A, Johnell O, De Laet C, Eisman JA, McCloskey EV, Mellstrom D, Melton LJ 3rd, Pols HA, Reeve J, Silman AJ, Tenenhouse A. A family history of fracture and fracture risk: a meta-analysis. Bone. 2004; 35(5):102937.
  • 9
    Kanis JA, Johnell O, De Laet C, Johansson H, Oden A, Delmas P, Eisman J, Fujiwara S, Garnero P, Kroger H, McCloskey EV, Mellstrom D, Melton LJ, Pols H, Reeve J, Silman A, Tenenhouse A. A meta-analysis of previous fracture and subsequent fracture risk. Bone. 2004; 35(2):37582.
  • 10
    Kanis JA, Johansson H, Oden A, Johnell O, de Laet C, Melton IL, Tenenhouse A, Reeve J, Silman AJ, Pols HA, Eisman JA, McCloskey EV, Mellstrom D. A meta-analysis of prior corticosteroid use and fracture risk. J Bone Miner Res. 2004; 19(6):8939.
  • 11
    De Laet C, Kanis JA, Oden A, Johanson H, Johnell O, Delmas P, Eisman JA, Kroger H, Fujiwara S, Garnero P, McCloskey EV, Mellstrom D, Melton LJ 3rd, Meunier PJ, Pols HA, Reeve J, Silman A, Tenenhouse A. Body mass index as a predictor of fracture risk: a meta-analysis. Osteoporos Int. 2005; 16(11):13308.
  • 12
    Albala C, Yanez M, Devoto E, Sostin C, Zeballos L, Santos JL. Obesity as a protective factor for postmenopausal osteoporosis. Int J Obes Relat Metab Disord. 1996; 20(11):102732.
  • 13
    Felson DT, Zhang Y, Hannan MT, Anderson JJ. Effects of weight and body mass index on bone mineral density in men and women: the Framingham study. J Bone Miner Res. 1993; 8(5):56773. doi: 10.1002/jbmr.5650080507
  • 14
    Premaor MO, Pilbrow L, Tonkin C, Parker RA, Compston J. Obesity and fractures in postmenopausal women. J Bone Miner Res. 2010; 25(2):2927. doi: 10.1359/jbmr.091004
  • 15
    Compston JE, Watts NB, Chapurlat R, Cooper C, Boonen S, Greenspan S, Pfeilschifter J, Silverman S, Díez-Pérez A, Lindsay R, Saag KG, Netelenbos JC, Gehlbach S, Hooven FH, Flahive J, Adachi JD, Rossini M, Lacroix AZ, Roux C, Sambrook PN, Siris ES. Glow Investigators. Obesity is not protective against fracture in postmenopausal women: GLOW. Am J Med. 2011; 124(11):104350. doi: 10.1016/j.amjmed.2011.06.013
  • 16
    Prieto-Alhambra D, Premaor MO, Fina Aviles F, Hermosilla E, Martinez-Laguna D, Carbonell-Abella C, Nogues X, Compston JE, Díez-Pérez A. The association between fracture and obesity is site-dependent: a population-based study in postmenopausal women. J Bone Miner Res. 2012; 27(2):294300. doi: 10.1002/jbmr.1466
  • 17
    Nielson CM, Srikanth P, Orwoll E. Obesity and fracture in men and women: an epidemiologic perspective. J Bone Miner Res. 2012; 27(1):110.
  • 18
    Fujiwara S, Kasagi F, Masunari N, Naito K, Suzuki G, Fukunaga M. Fracture prediction from bone mineral density in Japanese men and women. J Bone Miner Res. 2003; 18(8):154753.
  • 19
    Fujiwara S, Kasagi F, Yamada M, Kodama K. Risk factors for hip fracture in a Japanese cohort. J Bone Miner Res. 1997; 12(7):9981004.
  • 20
    Judson RN, Wackerhage H, Hughes A, Mavroeidi A, Barr RJ, Macdonald HM, Ratkevicius A, Reid DM, Hocking LJ. The functional ACTN3 577X variant increases the risk of falling in older females: results from two large independent cohort studies. J Gerontol A Biol Sci Med Sci. 2011; 66(1):1305. doi: 10.1093/gerona/glq189
  • 21
    Kreiger N, Tenenhouse A, Joseph L, Mackenzie T, Poliquin S, Brown JP, Prior JC, Rittmaster RS. The Canadian multicentre osteoporosis study CaMos: background, rationale, methods. Can J Aging. 1999; 18:37687.
  • 22
    Jones G, Nguyen T, Sambrook PN, Kelly PJ, Gilbert C, Eisman JA. Symptomatic fracture incidence in elderly men and women: the Dubbo Osteoporosis Epidemiology Study (DOES). Osteoporos Int. 1994; 4(5):27782.
  • 23
    Diez-Perez A, Gonzalez-Macias J, Marin F, Abizanda M, Alvarez R, Gimeno A, Pegenaute E, Vila J. Prediction of absolute risk of non-spinal fractures using clinical risk factors and heel quantitative ultrasound. Osteoporos Int. 2007; 18(5):62939.
  • 24
    Khaw KT, Reeve J, Luben R, Bingham S, Welch A, Wareham N, Oakes S, Day N. Prediction of total and hip fracture risk in men and women by quantitative ultrasound of the calcaneus: EPIC-Norfolk prospective population study. Lancet. 2004; 363(9404):197202.
  • 25
    Schott AM, Cormier C, Hans D, Favier F, Hausherr E, Dargent-Molina P, Delmas PD, Ribot C, Sebert JL, Breart G, Meunier PJ. How hip and whole-body bone mineral density predict hip fracture in elderly women: the EPIDOS Prospective Study. Osteoporos Int. 1998; 8(3):24754.
  • 26
    O'Neill TW, Felsenberg D, Varlow J, Cooper C, Kanis JA, Silman AJ. The prevalence of vertebral deformity in European men and women: the European Vertebral Osteoporosis Study. J Bone Miner Res. 1996; 11(7):10108. doi: 10.1002/jbmr.5650110719
  • 27
    Felsenberg D, Silman AJ, Lunt M, Ambrecht G, Ismail AA, Finn JD, Cockerill W, Banzer D, Benevolenskaya LI, Bhalla A, Bruges Armas J, Cannata JB, Cooper C, Dequeker J, Eastell R, Ershova O, Felsch B, Gowin W, Havelka S, Hoszowski K, Jajic I, Janott I, Johnell O, Kanis JA, Kragi G, Lopez Vaz A, Lorenc R, Lyritis G, Masaryk P, Matthis C, Miazgowski T, Parisi G, Pols HA, Poor G, Raspe H, Reid DM, Reisinger W, Scheidt-Nave C, Stepan J, Todd C, Weber K, Woolf AD, Reeve J, O'Neill TW. Incidence of vertebral fracture in Europe: results from the European Prospective Osteoporosis Study Epos. J Bone Miner Res. 2002; 17:71624.
  • 28
    Ismail AA, Pye SR, Cockerill WC, Lunt M, Silman AJ, Reeve J, Banzer D, Benevolenskaya LI, Bhalla A, Bruges Armas J, Cannata JB, Cooper C, Delmas PD, Dequeker J, Dilsen G, Falch JA, Felsch B, Felsenberg D, Finn JD, Gennari C, Hoszowski K, Jajic I, Janott J, Johnell O, Kanis JA, Kragl G, Lopez Vaz A, Lorenc R, Lyritis G, Marchand F, Masaryk P, Matthis C, Miazgowski T, Naves-Diaz M, Pols HA, Poor G, Rapado A, Raspe HH, Reid DM, Reisinger W, Scheidt-Nave C, Stepan J, Todd C, Weber K, Woolf AD, O'Neill TW. Incidence of limb fracture across Europe: results from the European Prospective Osteoporosis Study (EPOS). Osteoporos Int. 2002; 13(7):56571. doi: 10.1007/s001980200074
  • 29
    Svanborg A. Seventy-year-old people in Gothenburg a population study in an industrialized Swedish city. II. General presentation of social and medical conditions. Acta Med Scand Suppl. 1977; 611:537.
  • 30
    Stenstrom M, Olsson J, Mellstrom D. Thyroid hormone replacement is not related to increased risk of osteoporosis. Osteoporos Int. 2000 Jun; 11(2 Suppl):S144.
  • 31
    Pasco JA, Nicholson GC, Kotowicz MA. Cohort profile: Geelong Osteoporosis Study. Int J Epidemiol. 2012; 41(6):156575. doi: 10.1093/ije/dyr148
  • 32
    Leslie WD, Caetano PA, Macwilliam LR, Finlayson GS. Construction and validation of a population-based bone densitometry database. J Clin Densitom. 2005; 8(1):2530.
  • 33
    Yoshimura N, Takijiri T, Kinoshita H, Danjoh S, Kasamatsu T, Morioka S, Sakata K, Hashimoto T, Takeshita T. Characteristics and course of bone mineral densities among fast bone losers in a rural Japanese community: the Miyama Study. Osteoporos Int. 2004; 15(2):13944. doi: 10.1007/s00198-003-1518-9
  • 34
    Wong SY, Kwok T, Woo J, Lynn H, Griffith JF, Leung J, Tang YY, Leung PC. Bone mineral density and the risk of peripheral arterial disease in men and women: results from Mr. and Ms Os. Hong Kong. Osteoporos Int. 2005; 16(12):19338. doi: 10.1007/s00198-005-1968-3
  • 35
    Garnero P, Sornay-Rendu E, Chapuy MC, Delmas PD. Increased bone turnover in late postmenopausal women is a major determinant of osteoporosis. J Bone Miner Res. 1996; 11(3):33749. doi: 10.1002/jbmr.5650110307
  • 36
    Gluer CC, Eastell R, Reid DM, Felsenberg D, Roux C, Barkmann R, Timm W, Blenk T, Armbrecht G, Stewart A, Clowes J, Thomasius FE, Kolta S. Association of five quantitative ultrasound devices and bone densitometry with osteoporotic vertebral fractures in a population-based sample: the OPUS Study. J Bone Miner Res. 2004; 19(5):78293. doi: 10.1359/jbmr.040304
  • 37
    Honkanen R, Kroger H, Tuppurainen M, Alhava E, Saarikoski S. Fractures and low axial bone density in perimenopausal women. J Clin Epidemiol. 1995; 48(7):8818.
  • 38
    Bagger YZ, Tankó LB, Alexandersen P, Hansen HB, Møllgaard A, Ravn P, Qvist P, Kanis JA, Christiansen C. Two to three years of hormone replacement treatment in healthy women have long-term preventive effects on bone mass and osteoporotic fractures: the PERF study. Bone. 2004; 34(4):72835.
  • 39
    Melton LJ 3rd, Crowson CS, O'Fallon WM, Wahner HW, Riggs BL. Relative contributions of bone density, bone turnover, and clinical risk factors to long-term fracture prediction. J Bone Miner Res. 2003; 18(2):3128. doi: 10.1359/jbmr.2003.18.2.312
  • 40
    Melton LJ 3rd, Atkinson EJ, O'Connor MK, O'Fallon WM, Riggs BL. Bone density and fracture risk in men. J Bone Miner Res. 1998; 13(12):191523. doi: 10.1359/jbmr.1998.13.12.1915
  • 41
    Hofman A, Grobbee DE, de Jong PT, van den Ouweland FA. Determinants of disease and disability in the elderly: the Rotterdam Elderly Study. Eur J Epidemiol. 1991; 7(4):40322.
  • 42
    De Laet CE, Van Hout BA, Burger H, Weel AE, Hofman A, Pols HA. Hip fracture prediction in elderly men and women: validation in the Rotterdam study. J Bone Miner Res. 1998; 13(10):158793. doi: 10.1359/jbmr.1998.13.10.1587
  • 43
    Hofman A, van Duijn CM, Franco OH, Ikram MA, Janssen HL, Klaver CC, Kuipers EJ, Nijsten TE, Stricker BH, Tiemeier H, Uitterlinden AG, Vernooij MW, Witteman JC. The Rotterdam Study: 2012 objectives and design update. Eur J Epidemiol. 2011; 26(8):65786. doi: 10.1007/s10654-011-9610-5
  • 44
    Krieg MA, Comuz J, Ruffieux C, Burckhardt P. [Role of bone ultrasound in predicting hip fracture risk in women 70 years or older: results of the SEMOF study and comparison with literature data]. Rev Med Suisse Romande. 2004; 124(2):5962.
  • 45
    Johansson H, Oden A, Johnell O, Jonsson B, de Laet C, Oglesby A, McCloskey EV, Kayan K, Jalava T, Kanis JA. Optimization of BMD measurements to identify high risk groups for treatment—a test analysis. J Bone Miner Res. 2004; 19(6):90613.
  • 46
    McCloskey EV, Beneton M, Charlesworth D, Kayan K, deTakats D, Dey A, Orgee J, Ashford R, Forster M, Cliffe J, Kersh L, Brazier J, Nichol J, Aropuu S, Jalava T, Kanis JA. Clodronate reduces the incidence of fractures in community-dwelling elderly women unselected for osteoporosis: results of a double-blind, placebo-controlled randomized study. J Bone Miner Res. 2007; 22(1):13541. doi: 10.1359/jbmr.061008
  • 47
    Cummings SR, Nevitt MC, Browner WS, Stone K, Fox KM, Ensrud KE, Cauley J, Black D, Vogt TM. Risk factors for hip fracture in white women. Study of Osteoporotic Fractures Research Group. N Engl J Med. 1995; 332(12):76773. doi: 10.1056/NEJM199503233321202
  • 48
    Walley T, Mantgani A. The UK General Practice Research Database. Lancet. 1997; 350(9084):10979. doi: 10.1016/S0140-6736(97)04248-7
  • 49
    Hays J, Hunt JR, Hubbell FA, Anderson GL, Limacher M, Allen C, Rossouw JE. The Women's Health Initiative recruitment methods and results. Ann Epidemiol. 2003; 13(9 Suppl):S1877.
  • 50
    Anderson GL, Manson J, Wallace R, Lund B, Hall D, Davis S, Shumaker S, Wang CY, Stein E, Prentice RL. Implementation of the Women's Health Initiative study design. Ann Epidemiol. 2003; 13(9 Suppl):S517.
  • 51
    WHO. Obesity: preventing and managing the global epidemic. WHO Technical Report Series 894. Geneva, Switzerland: World Health Organization; 2000.
  • 52
    Kanis JA, Oden A, Johnell O, Jonsson B, de Laet C, Dawson A. The burden of osteoporotic fractures: a method for setting intervention thresholds. Osteoporos Int. 2001; 12(5):41727.
  • 53
    Breslow NE, Day NE. Statistical methods in cancer research: the design and analysis of cohort studies. IARC Scientific Publications. No. 82. Vol II. Lyon, France: IARC; 1987 [cited 2013 Jul 1]. p. 1315. Available from:
  • 54
    Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003; 327(7414):55760. doi: 10.1136/bmj.327.7414.557
  • 55
    Harrell FJ. General aspects of fitting regression models: regression modeling strategies. New York: Springer Science + Business Media Inc; 2001.
  • 56
    Gnudi S, Sitta E, Lisi L. Relationship of body mass index with main limb fragility fractures in postmenopausal women. J Bone Miner Metab. 2009; 27(4):47984. doi: 10.1007/s00774-009-0056-8
  • 57
    Kanis JA, Hans D, Cooper C, Baim S, Bilezikian JP, Binkley N, Cauley JA, Compston JE, Dawson-Hughes B, El-Hajj Fuleihan G, Johansson H, Leslie WD, Lewiecki EM, Luckey M, Oden A, Papapoulos SE, Poiana C, Rizzoli R, Wahl DA, McCloskey EV. Interpretation and use of FRAX in clinical practice. Osteoporos Int. 2011; 22(9):2395411. doi: 10.1007/s00198-011-1713-z
  • 58
    Premaor M, Parker RA, Cummings S, Ensrud K, Cauley JA, Lui LY, Hillier T, Compston J. Predictive value of FRAX for fracture in obese older women. J Bone Miner Res. 2013; 28(1):18895. doi: 10.1002/jbmr.1729
  • 59
    Nielson CM, Marshall LM, Adams AL, LeBlanc ES, Cawthon PM, Ensrud K, Stefanick ML, Barrett-Connor E, Orwoll ES. BMI and fracture risk in older men: the osteoporotic fractures in men study (MrOS). J Bone Miner Res. 2011; 26(3):496502. doi: 10.1002/jbmr.235
  • 60
    Giangregorio LM, Leslie WD, Lix LM, Johansson H, Oden A, McCloskey E, Kanis JA. FRAX underestimates fracture risk in patients with diabetes. J Bone Miner Res. 2012; 27(2):3018. doi: 10.1002/jbmr.556