Body mass index as a predictor of fracture risk: A meta-analysis

ORIGINAL ARTICLE

Body mass index as a predictor of fracture risk: A meta-analysis

C. De Laet Æ J.A. Kanis Æ A. Oden Æ H. Johanson

O. Johnell Æ P. Delmas Æ J.A. Eisman Æ H. Kroger

S. Fujiwara Æ P. Garnero Æ E.V. McCloskey

D. Mellstrom Æ L.J. Melton 3rd Æ P.J. Meunier

H.A.P. Pols Æ J. Reeve Æ A. Silman Æ A. Tenenhouse

Received: 1 December 2004 / Accepted: 27 January 2005 / Published online: 1 June 2005� International Osteoporosis Foundation and National Osteoporosis Foundation 2005

Abstract Low body mass index (BMI) is a well-docu-mented risk factor for future fracture. The aim of thisstudy was to quantify this effect and to explore theassociation of BMI with fracture risk in relation to age,gender and bone mineral density (BMD) from aninternational perspective using worldwide data. Westudied individual participant data from almost 60,000men and women from 12 prospective population-basedcohorts comprising Rotterdam, EVOS/EPOS, CaMos,Rochester, Sheffield, Dubbo, EPIDOS, OFELY, Kuo-pio, Hiroshima, and two cohorts from Gothenburg, witha total follow-up of over 250,000 person years. The ef-fects of BMI, BMD, age and gender on the risk of anyfracture, any osteoporotic fracture, and hip fracture

alone was examined using a Poisson regression model ineach cohort separately. The results of the differentstudies were then merged. Without information onBMD, the age-adjusted risk for any type of fractureincreased significantly with lower BMI. Overall, the riskratio (RR) per unit higher BMI was 0.98 (95% confi-dence interval [CI], 0.97–0.99) for any fracture, 0.97(95% CI, 0.96–0.98) for osteoporotic fracture and 0.93(95% CI, 0.91–0.94) for hip fracture (all p <0.001). TheRR per unit change in BMI was very similar in men andwomen (p >0.30). After adjusting for BMD, these RRbecame 1 for any fracture or osteoporotic fracture and0.98 for hip fracture (significant in women). The gradi-ent of fracture risk without adjustment for BMD was

C. De LaetScientific Institute of Public Health,Brussels, Belgium

J.A. Kanis (&) Æ E.V. McCloskeyWHO Collaborating Centre for Metabolic Bone Diseases,University of Sheffield Medical School,Beech Hill Road, Sheffield, S10 2RX, UKE-mail: [email protected].: +44-114-2851109Fax: +44-114-2851813

A. Oden Æ H. JohansonConsulting Statistician, Gothenburg, Sweden

O. JohnellDepartment of Orthopaedics,Malmo General Hospital, Malmo, Sweden

P. DelmasINSERM Unite 149, Villejuif, France

J.A. EismanBone and Mineral Research Program,Garvan Institute of Medical Research,St Vincent’s Hospital and University of New South Wales,Sydney, Australia

H. KrogerDepartment of Surgery, Bone and Cartilage Research Unit,Kuopio University Hospital, Kuopio, Finland

S. FujiwaraDepartment of Clinical Studies,Radiation Effects Research Foundation,Hiroshima, Japan

P. GarneroINSERM Unite 403, Lyon, France

D. MellstromDepartment of Geriatric Medicine, Goteborg University,Gothenburg, Sweden

L.J. Melton 3rdDivision of Epidemiology, Mayo Clinic,Rochester, Minnesota, USA

P.J. MeunierINSERM Unit 403, Faculty R Laennec, Lyon, France

J. ReeveStrangeway’s Research Laboratory,Wort’s Causeway, Cambridge, UK

A. SilmanARC Epidemiology Unit, University of Manchester,Manchester, UK

A. TenenhouseDivision of Bone Metabolism,The Montreal General Hospital,Montreal, Canada

Osteoporos Int (2005) 16: 1330–1338DOI 10.1007/s00198-005-1863-y

not linearly distributed across values for BMI. Instead,the contribution to fracture risk was much more markedat low values of BMI than at values above the median.This nonlinear relation of risk with BMI was most evi-dent for hip fracture risk. When compared with a BMIof 25 kg/m2, a BMI of 20 kg/m2 was associated with anearly twofold increase in risk ratio (RR=1.95; 95% CI,1.71–2.22) for hip fracture. In contrast, a BMI of 30 kg/m2, when compared with a BMI of 25 kg/m2, wasassociated with only a 17% reduction in hip fracture risk(RR=0.83; 95% CI, 0.69–0.99). We conclude that lowBMI confers a risk of substantial importance for allfractures that is largely independent of age and sex, butdependent on BMD. The significance of BMI as a riskfactor varies according to the level of BMI. Its valida-tion on an international basis permits the use of this riskfactor in case-finding strategies.

Keywords BMI Æ Fractures Æ Meta-analysis ÆOsteoporosis Æ Prospective studies Æ Risk

Introduction

Low weight, or low body mass index (BMI), is a well-documented risk factor for future fracture, whereas ahigh BMI appears to be protective [1–10]. The increasingprevalence of overweight and obesity in Western socie-ties [11,12] might at first seem a promising developmentfrom the point of view of osteoporosis and fractureprevention. From a public health point of view, how-ever, the story is more complicated. Obesity is associatedwith increased morbidity from diabetes, hypertension,and cardiovascular diseases, and is also associated withincreased mortality. The same may also be true foroverweight women. Recently, it was estimated thatoverweight (BMI>25 kg/m2) 41-year-old female non-smokers lost on average 3.3 years of life, whereas, obese(BMI>30 kg/m2) female nonsmokers lost 7.1 life years[13]. It is important, therefore, to quantify the associa-tion between BMI and fracture risk and to explore itsrelationship to age, gender and bone mineral density(BMD) with the aim of being able to give balanced ad-vice on lifestyle to patients. These relationships are alsoimportant when using BMI to assess fracture risk in casefinding [2,14–16].

The aim of the present study was to explore therelationship of BMI with fracture risk (any fracture, anyosteoporotic fracture and hip fracture alone) in men andwomen using data from 12 prospective population-basedcohort studies in an international perspective. BMI waschosen rather than weight to explore this association,because of the wide variation in average weight andheight between different countries, which is reduced byadjusting weight for height. Moreover, BMI is as good apredictor of fractures as weight in most studies of hipfracture outcomes [17,18].

Methods

Participants

We used baseline and follow-up data from 12 prospec-tive population-based cohorts comprising the Rotter-dam Study, The European Vertebral Osteoporosis Study(later the European Prospective Osteoporosis Study(EVOS/EPOS), The Canadian Multicentre OsteoporosisStudy (CaMos), Rochester, Sheffield, Dubbo, Epidemi-ologie de l’Osteoporose Study (EPIDOS), a cohort fromFrance (OFELY), Kuopio, Hiroshima and two cohortsfrom Gothenburg. Details of each of the cohorts arepublished elsewhere, but are summarized briefly belowand in Table 1.

The Rotterdam Study, begun in 1990, is an ongoingprospective cohort study that aimed to examine andfollow-up all residents aged 55 years and older living inOmmoord, a district of Rotterdam [19]. By 1993, 7,983residents had been included (response rate 78%). Frac-ture follow-up was achieved through an automatic linkwith general practitioner computer systems and hospitaladmission data [20]. Fracture data were collected andvalidated by two independent research physicians. Forthis analysis, validated fracture follow-up was availablefor 6,851 participants (2,793 men) with an average fol-low-up time of 6 years. Femoral neck BMD was mea-sured in 5,731 individuals (2,414 men) by dual X-rayabsorptometry (DXA) (Lunar DPX-L).

The European Vertebral Osteoporosis Study (EVOS)comprised age- and sex-stratified random samples from36 centers in 19 European countries [21]. Equal numbersof men and women were drawn in each center within six5-year age bands (50–74 and 75+ years). BMD wasmeasured in 3,461 men and women from 13 centers byDXA at the femoral neck using pencil beam machinesthat were cross-calibrated using the European SpinePhantom. This sample provided the framework for theEuropean Prospective Osteoporosis Study (EPOS),where repeated assessment was undertaken in 29 of thecenters [22,23]. For this analysis, validated fracture fol-low-up was available for 13,490 participants (6,521 men)with an average follow-up time of 3 years. Femoral neckBMD was measured in 4,746 individuals (2,141 men).

The Canadian Multicentre Osteoporosis study (Ca-Mos) is an ongoing prospective age stratified cohort. Thestudy is documenting the incidence of fractures and riskfactors in a random sample of 9,424 men and womenaged 25 years or more selected by telephone listings. Thesampling frame is from nine study centers in sevenprovinces [24]. Characterization of individuals was byinterview. BMD was measured by DXA at the femoralneck with Hologic QDR in seven centers and the LunarDPX Alpha in two centers in 8,297 individuals (2,589men). Machines were cross-calibrated using the sameEuropean Spine Phantom. For this analysis, validatedfracture follow-up was available for 9,101 participants(2,801 men) with an average follow-up time of 3 years.

1331

The Rochester cohort was recruited from two ran-dom population samples stratified by decade of age, oneof women who were subsequently followed for up to20 years [25], and another sample of women and menfollowed for 8 years [26]. BMD of the right femoral neckwas measured by dual photon absorptiometry in the firstcohort (cross-calibrated to DXA) and by DXA (HologicQDR 2000) in the second group. Fractures were ascer-tained by periodic interview combined with review of thein-patient and out-patient medical records of all localcare providers. For this analysis, validated fracture fol-low-up was available for 1,001 participants (348 men)with an average follow-up time of 6 years. Femoral neckBMD was measured in 993 individuals (345 men).

The Sheffield cohort comprised women aged 75 yearsor more selected randomly from the population ofSheffield, UK, and surrounding districts between 1993and 1999. Approximately 35,000 women, identified fromgeneral practitioner listings, were contacted by letter andinvited to attend for the assessment of skeletal status;5,873 women were willing to attend. Of these, 281 wo-men were excluded and the remainder randomly allo-cated to treatment with placebo or a bisphosphonate,clodronate, to study its effects on fracture risk. Thematerial for this study comprised 2,172 women allocatedto treatment with placebo only [27]. All women hadbaseline assessment of BMD undertaken at the femoralneck using the Hologic QDR 4500. Outcomes were as-sessed by 6-monthly home visits. For this analysis, val-

idated fracture follow-up was available for 2,170participants with an average follow-up time of 6 years.Femoral neck BMD was measured in 2,150 individuals.

The Dubbo Osteoporosis Epidemiology Study(DOES) is a population-based study with multipleassessments of skeletal status in men and women aged60 years or more from Dubbo, Australia [28]. Partici-pation in the study was 56% of the population. Baselinemeasurements included BMD at the femoral neck as-sessed using DXA (GE-Lunar, DPX and Prodigy).Fractures are identified through radiologists’ reportsfrom the two centers serving the region. For this anal-ysis, validated fracture follow-up was available for 2,089participants (819 men) with an average follow-up time of8 years. Femoral neck BMD was measured in 2,060individuals (801 men).

The EPIDOS cohort is a prospective multicenterstudy of risk factors for hip fractures that included 7,575French women age 75 years or older [29]. Participantswere recruited through mailings using large population-based listings such as voter-registration rolls. Baselinecharacteristics were obtained through a structuredquestionnaire as well as through clinical and functionalexaminations and BMD measurement (Lunar DPX).For this analysis, validated fracture follow-up wasavailable for 1,183 participants comprising 291 hipfracture cases and age-matched controls with an averagefollow-up time of 3 years [30]. Femoral neck BMD wasmeasured in 1,180 individuals.

Table 1 Details of cohorts studied (see Methods for cohort abbreviations, BMI body mass index)

Cohort Gender Samplesize

Women(%)

Person-years

Anyfracture

Hipfracture

Osteoporoticfracture

Meanage (years)

Mean BMI(kg/m2)

Meanheight (cm)

Meanweight (kg)

EVOS-EPOS Men 6,521 52 19,736 213 20 213 64.2 27 171 79.1Women 6,969 20,945 506 30 506 63.5 27.1 159 68.4Total 13,490 40,681 719 50 719

CaMos Men 2,801 69 8,002 124 9 59 59.9 27 174 81.4Women 6,300 17,832 447 31 248 62.9 26.9 160 68.6Total 9,101 25,834 571 40 307

Rochester Men 348 65 1,160 38 0 25 55.4 27.3 175 84.3Women 653 5,067 251 42 219 57.6 25.5 161 66.3Total 1,001 6,227 289 42 244

Rotterdam Men 2,793 59 16,150 185 52 130 68.4 25.7 175 78.3Women 4,058 23,443 676 168 516 69.9 26.7 161 69.4Total 6,851 39,593 861 220 646

DOES Men 819 61 6,365 138 27 110 70.1 26 173 78.2Women 1,270 9,629 381 76 297 70.7 25.4 160 64.8Total 2,089 15,994 519 103 407

Gothenburg 1 Men 812 59 6,010 95 73 95 77 25.4 173 75.7Women 1,158 9,191 255 198 255 78.6 25.3 159 63.9Total 1,970 15,201 350 271 350

Hiroshima Men 793 70 3,004 44 7 12 63.2 22.7 163 60.7Women 1,810 6,821 143 25 78 65.9 23.1 150 52.3Total 2,603 9,825 187 32 90

OFELY Women 430 100 2,144 50 – – 64.1 24.2 158 60.5Sheffield Women 2,170 100 6,894 292 63 243 80 26.7 156 64.8Kuopio Women 11,691 100 56,091 1,043 – – 52.3 26.2 161 68.2Gothenburg 2 Women 7,065 100 29,603 440 29 312 58.9 24.6 165 67.2EPIDOS Women 1,183 100 3,947 – 291 – 82.4 25.4 153 59.5All men 14,887 60,427 837 188 644 66.4 26.2 172.6 77.9All women 44,757 75 191,607 4,484 953 2674 62.2 25.9 160.4 66.9Overall 59,644 252,034 5,321 1,141 3318 63.2 26.0 163.3 69.5

1332

The OFELY cohort comprises a cohort of 1,039women aged 31–89 years stratified by age randomly se-lected from the regional section of a large health insur-ance company (Mutuelle Generale d’EducationNationale, Lyon) [31]. Eighteen percent of women con-tacted participated in the study. Baseline characteristicswere obtained using a standardized questionnaire,including the documentation of prior wrist, humeral,vertebral and hip fracture that occurred after the age of45 years. Only low-trauma fractures (falls from astanding height or less) were recorded, but not the site offracture. BMD was measured at the spine (L1–L4), atthe proximal femur, distal radius and whole body byDXA using a QDR 2000 (Hologic). Women were re-viewed annually and incident fractures registered.Peripheral fractures were confirmed by radiography. Forthis analysis, validated fracture follow-up was availablefor 430 participants with an average follow-up time of5 years. Femoral neck BMD was measured in 427individuals.

The Kuopio osteoporosis risk factor and prevention(OSTPRE) study in Finland comprised a postal enquirysent to all 14,220 women aged 47–56 who were residentsof Kuopio province in 1989 [32]. Of 13,100 womenresponding to the enquiry, 938 were excluded forincomplete information. For this analysis, validatedfracture follow-up was available for 11,691 participantswith an average follow-up time of 5 years. The site offracture was not recorded other than those at the fore-arm. Femoral neck BMD was measured in 1,743 indi-viduals with the Lunar DPX [32].

The Gothenburg I study comprised four birth cohortsof 2,375 randomly sampled men and women aged70 years or more followed for up to 20 years after abaseline BMD measurement [33]. The participants weredrawn randomly from the population register in Goth-enburg by the date of birth to provide cohorts aged 70,76, 79 and 85 years at the time of investigation. Partic-ipation rate was 73%. Bone mineral density was mea-sured at the right heel using dual photonabsorptiometry. For this analysis, validated fracturefollow-up was available for 1,970 participants (812 men)with an average follow-up time of 8 years. BMD wasmeasured in 1,633 individuals (686 men).

The Gothenburg II study comprised a randomlydrawn population cohort of over 7,000 women aged 50–70 years followed for 4 years [34]. The participation ratewas 67%. Assessment included a standardized ques-tionnaire that recorded information on risk factors forosteoporosis. Fractures were identified prospectivelythrough the radiology departments serving the region.BMD was assessed at baseline at the distal forearm byusing the Osteometer DTX-200. For this analysis, vali-dated fracture follow-up was available for 7,065 partic-ipants with an average follow-up time of 4 years. BMDwas measured in 7,056 individuals.

The Adult Health Study (AHS) was established in1958 to document the late health effects of radiationexposure among atomic bomb survivors in Hiroshima

and Nagasaki. The original AHS cohort consisted ofabout 15,000 atomic bomb survivors and 5,000 controlsselected from residents in Hiroshima and Nagasakiusing the 1950 national census supplementary schedulesand the Atomic Bomb Survivors Survey. AHS subjectshave been followed through biennial medical examina-tions since 1958. The participation rate has been around80% throughout this period. BMD was measured at theproximal femur by DXA in 1994 (Hologic QDR-2000)in 2,588 individuals (791 men). Self-reported fractureswere documented at 6-monthly intervals [35,36]. For thisanalysis, validated fracture follow up was available for2,603 participants (793 men) with an average follow-uptime of 4 years.

Baseline and outcome variables

Height and weight were measured using standard tech-niques in all cohorts. BMI was calculated as weight inkilograms divided by height squared in meters. For thepurposes of this analysis, we utilized BMD assessed atthe femoral neck by DXA, with the exception of the twoGothenburg cohorts where BMD was assessed by DXAat the distal forearm or by DPA at the right heel. Weadditionally analyzed the BMD data excluding thosetwo cohorts.

Fracture ascertainment was undertaken by self-report(Sheffield, EVOS/EPOS, Hiroshima, Kuopio, EPIDOS,OFELY) and/or verified from hospital or central data-bases (Gothenburg, CaMos, DOES, Kuopio, Sheffield,EVOS/EPOS, Rochester, Rotterdam). The EPOS andthe Rotterdam study also included sequential systematicradiography to define incident vertebral deformities, butthese were not used in this analysis. In the analysis, weused information on any clinical fracture and on frac-tures considered to be osteoporotic. In addition, hipfracture alone was considered separately. An osteopo-rotic fracture was one considered to be due to osteo-porosis by the investigator in the EVOS/EPOS studyand in CaMos. For the EVOS/EPOS study, osteoporoticfractures comprised hip, forearm, humeral or spinefractures. For the CaMos Study they comprised frac-tures of the spine, pelvis, ribs, distal forearm, forearmand hip. In the other cohorts, fractures at sites consid-ered to be characteristic for osteoporosis were extracted[37].

Statistical methods

The association of BMI with the risk of any fracture,osteoporotic fracture and hip fracture was examinedusing a Poisson regression model in each cohort sepa-rately. Covariates included current age and time sincestart of follow-up, and we performed analyses for bothsexes separately, with and without taking BMD infor-mation into account. BMD was expressed as sex- andcohort-specific Z -scores. BMI was analyzed either

1333

continuously or using specific thresholds. The b-coeffi-cients of each cohort and the two sexes were weightedaccording to the variance and merged to determine theweighted mean of the coefficient and its standard devi-ation. The risk ratios (RR) at different BMI levels arethen given by e(weighted mean coefficient). We also analyzedthe RR per unit difference of BMI for fracture at variouslevels of BMI, taking a BMI of 25 kg/m2 as the refer-ence, since a BMI of 25 is the internationally acceptedthreshold between normal weight and overweight [11].This was slightly below the average in most of the co-horts (see Table 1). To optimally describe the associa-tion of RR with level of BMI, we analyzed thisrelationship using spline functions.

Heterogeneity between cohorts was tested by meansof the I2 statistic [38]. Low heterogeneity was noted forhip fracture outcomes (I2=8%; p>0.30) and wasmoderate for osteoporotic fracture outcomes (I 2=49%). When the interaction between BMI and currentage (BMIÆcurrent age) was included, there was no sig-nificant heterogeneity between cohorts (p >0.30) forBMI (I 2 =0) nor for the interaction term (I 2 =0) and afixed-effects rather than a random-effects model wasused.

Results

The total sample included almost 60,000 men andwomen from the 12 cohorts with a total follow-up ofmore than 250,000 person years. The distribution ofBMI is shown in Table 2. Information on any frac-ture, on osteoporotic fracture and on hip fracturealone was available for about 58,000, 46,000 and47,000 participants, respectively. During follow-up,there were 5,321 fractures registered including 3,318osteoporotic fractures and 1,141 hip fractures. Bonemineral density measurements were available in 65%of individuals. Details by cohort and gender are shownin Table 1.

Without information on BMD, low BMI in men andwomen combined was associated with a significantlyincreased age-specific risk of fracture, while at higherBMI values the risk of fracture was decreased. The riskratio per unit increase in BMI (gradient of risk; GR) wasfor any fracture 0.98 (95% confidence interval [CI],

0.97–0.99), for osteoporotic fracture 0.97 (95% CI, 0.96–0.98) and for hip fracture 0.93 (95% CI, 0.91–0.94)(Fig. 1A). As the figure shows, the RRs per unit changeof BMI in men and women were very similar and notsignificantly different (p >0.30). Since all osteoporoticfractures also included hip fractures, we analyzed oste-oporotic fractures excluding hip fractures. As expected,the GR increased slightly (i.e., predictive value de-creased) but remained significantly lower than 1 in menand women combined (data not shown).

When information on BMD was included (Fig. 1B),the GRs changed markedly and remained significantlydifferent from unity only for hip fractures in women.When the two studies from Gothenburg were excluded(because BMD was not measured at the femoral neck),the GR for hip fracture in women was not significantlydifferent from unity.

Since the GRs were similar in men and women bothbefore and after correction for BMD and not signifi-cantly different, further results are presented for menand women combined, but all analyses were also carriedout for men and women separately.

The effect of age on gradient of risk is shown inFig. 2. For any fracture and for any osteoporotic frac-ture, the GR per unit of BMI increased with advancingage (without adjustment for BMD). In contrast, for hipfractures the gradient of risk decreased with age, al-though this trend was not significant. When hip frac-tures were excluded from the osteoporotic fractures, asimilar trend with age was observed as seen for allosteoporotic fractures. After correction for BMD, therisk gradients showed nonsignificant trends with age,and at most ages were not significantly different from 1.

We also analyzed the risk ratio (RR) for fracture atvarious levels of BMI, taking a BMI of 25 kg/m2 as thereference. As expected, the RR increased with decreasingBMI. The magnitude of the effect was greater for hipfracture than for any osteoporotic fracture or any frac-ture. The RR of fracture risk with BMI was, however,nonlinear (Fig. 3 and Table 3). RR was markedly higherat the lower values of BMI, particularly with a BMI of20 kg/m2 or less. By contrast, between a BMI of 25 kg/m2 and 35 kg/m2 the differences in RR were small. Forexample, given the overall risk gradient for hip fracture(0.93/unit BMI) an increase of 5 BMI units from 25 kg/m2 to 30 kg/m2 would be expected to correspond to a32% reduction in hip fracture risk. The observed dif-ference was, however, only 17% (1.00 vs 0.83, see Ta-ble 3). With a 10-unit BMI difference the expected riskreduction would be more than 50%, whereas, the dif-ference between a BMI of 25 kg/m2 and one of 35 kg/m2

was only 25% (1.00 vs 0.75, see Table 3). At the low endof the BMI spectrum, on the other hand, a change of5 units from a BMI of 25 kg/m2 to a BMI of 20 kg/m2,corresponded to a doubling of the hip fracture risk (1.95vs 1.00, see Table 2). For any osteoporotic fracture therewas a 27% difference in fracture risk comparing a BMIof 20 kg/m2 vs 25 kg/m2 and an 11% difference com-paring a BMI of 30 kg/m2 v 25 kg/m2.

Table 2 Distribution (%) of men and women categorized byintervals of BMI (BMI body mass index)

BMI

(kg/m2) Men Women Total

<20 7.5 8.9 8.520–24 30.9 38.5 36.525–29 47.2 35.8 38.830–34 12.4 12.9 12.835–39 1.7 3.1 2.740+ 0.2 0.8 0.7

1334

Re-analyzing the data using only those cohorts with auniform acquisition of data on fractures (Rochester,Rotterdam,DOES,Hiroshima andSheffield) did not alterthe relationship of BMI and hip fracture risk. For osteo-porotic fracture, a high BMI had a greater and significantprotective effect in the absence of BMD (data not shown).

The data for Table 3 (unadjusted for BMD) wererecalculated in the 65% of individuals who had also aBMD test. The findings did not differ from those uti-lizing the entire cohort.

After adjustment for BMD, BMI was not predictiveof fracture risk except for hip fracture at a BMI of

Fig. 1 Risk Ratio (RR) forfracture per unit increase inBMI; (A) adjusted for currentage and time since start offollow-up, and (B) additionallyadjusted for BMD (BMI bodymass index, BMD bone mineraldensity) [04Ca001]

Fig. 2 Relative fracture riskper unit increase in BMI by agefor men and women combined;(A) adjusted for time sincestart of follow-up, and(B), additionally adjusted forBMD (BMI body mass index,BMD bone mineral density)[04Ca002]

Fig. 3 Relative fracture risk at various levels of BMI (kg/m2) formen and women combined. The reference is a BMI=25, (A)adjusted for current age and time since start of follow-up, and (B)additionally adjusted for BMD. The bold solid line describes hip

fracture, the solid line any osteoporotic fracture, and the dotted lineany fracture (BMI body mass index, BMD bone mineral density)[04Ca003]

1335

20 kg/m2 or less (Table 3). There were no significantdifferences in these relationships between men and wo-men. The results were not changed when cohorts withuniform characterization of fractures were analyzed.After adjustment for BMD, the risk of hip fracture was33% higher comparing individuals with a BMI of 20 kg/m2 with 25 kg/m2 (42% higher in all cohorts; see Ta-ble 3).

For osteoporotic fractures, the I2 statistic was 49%(95% CI, 8–71), indicating that about half of the ob-served differences between studies was due to heteroge-neity. This heterogeneity disappeared when account wastaken of age (I 2 =0). For hip fractures there was alsolow heterogeneity (I 2 =8% (0–44)) indicating that thedata about the relationship between BMI and fracturerisk were homogenous between cohorts.

Discussion

The principal finding of the present study, undertaken inlarge and internationally drawn cohorts, is the confir-mation that low BMI is associated with a substantialincrease in fracture risk of similar magnitude in men andwomen, whereas, a high BMI is protective. This riskassociated with a low BMI was present at most ages andfor all types of fracture studied, but was strongest for hipfracture. Overall, the RR for hip fracture decreased 0.93per unit increase in BMI.

This risk gradient was, however, not constant acrossthe range of BMI. In the most common range of normalweight and overweight (the average in our populationswas 26.0 kg/m2), the gradient of risk per unit BMI wasrelatively low (see Fig. 3A). In contrast, the gradientswere steeper at low values of BMI, below 22 kg/m2. Thisthreshold of 22 kg/m2 and lower corresponded overall inour cohorts to about 10% of men and 17% of women. Anonlinear relationship of BMI with fracture risk, withthresholds of BMI of approximately 22 kg/m2 and26 kg/m2, has been described previously in case-control

studies of men and women [17,18]. In the present study,where the individual participant data were pooled fromseveral prospective studies, we could derive the form ofthis relationship. The present study additionally dem-onstrates that this gradient of risk is markedly reducedwhen adjusted for BMD, suggesting that BMD is animportant intermediary or confounder.

These findings have important consequences for case-finding strategies based on clinical risk factors. Firstly,obesity should not be regarded as an important protec-tive factor for hip fracture risk [39,40]. Rather, leannessshould be regarded as a significant risk factor. Secondly,the use of low BMI as a risk factor will identify popu-lations with a low BMD and hence a high risk of frac-ture. The finding that leanness is much more importantas a risk factor for hip fracture—than obesity is a pro-tective factor—means that advice concerning bodyweight and osteoporosis need not be inconsistent withthe weight control advocated for the prevention of car-diovascular disease or diabetes.

After adjustment of fracture risk by BMD, a lowBMI was still a significant risk factor for hip fracture.Thus, low BMI can be used to enhance the predictivevalue of BMD in case finding. The mechanisms for thisare conjectural but might include muscle weakness [41],perhaps associated with nutritional deficiencies of pro-tein or vitamin D [41,42], decreased padding over thegreater trochanter [43], or a greater liability to fall [10].This BMD-independent risk was not observed, however,when only cohorts using femoral neck BMD wereanalysed.

BMI and other risk factors have been studied mainlyin relation to hip fracture risk. As expected, the riskgradient we found was higher for hip fractures than forall osteoporotic fractures. However, for any fracture orfor all osteoporotic fractures, the impact of BMI wasstill significant, even when hip fractures were excludedfrom the analysis. For BMD, it is customary to expressthe relation with fracture risk as a gradient of risk perSD change. For hip fracture, for example, the most

Table 3 Risk Ratio (RR) for fracture at various levels of BMI (kg/m2) for men and women combined, adjusted for current age and time instudy, without and with adjustment for BMD. The reference is a BMI of 25 kg/m2 (BMD bone mineral density, BMI body mass index, CIconfidence interval)

BMI Any fracture Osteoporotic fracture Hip fracture

RR 95% CI RR 95% CI RR 95% CI

Not adjusted for BMD15 1.66 1.31–2.09 1.79 1.35)2.37 4.48 3.11–6.4520 1.21 1.12–1.30 1.27 1.16–1.38 1.95 1.71–2.2225 1.00 reference 1.00 reference 1.00 reference30 0.92 0.85–1.00 0.89 0.81–0.98 0.83 0.69–0.9935 0.85 0.74–0.98 0.74 0.62–0.90 0.75 0.50–1.11

Adjusted for BMD15 1.00 0.75–1.33 1.07 0.78–1.48 2.16 1.42–3.2820 0.98 0.9–1.08 1.02 0.92–1.13 1.42 1.23–1.6525 1.00 reference 1.00 reference 1.00 reference30 1.01 0.91–1.11 0.96 0.86–1.08 1.00 0.82–1.2135 0.99 0.82–1.19 0.91 0.73–1.13 1.18 0.78–1.80

1336

commonly cited number is that of an RR of 2.6 per SDdecrease in BMD at the femoral neck [44]. In the pop-ulations that we studied, the average SD for BMI wasaround 4 kg/m2, corresponding to a RR for hip fractureof 1.4 per SD decrease in BMI, much lower than theestimate for BMD. The analogy is, however, not whollyappropriate in view of the nonlinearity of risk with BMI.

We found no significant differences in the gradients offracture risk between men and women; but, we found asignificant increase of risk associated with low BMI withage, for any fracture and for osteoporotic fractures, butan opposite, though not significant, trend for hip frac-tures. In the case of any fracture or osteoporotic frac-ture, an increased exposure with age to relative gonadaldeficiency might provide an explanation, at least inwomen [45]. A further possible explanation is that inyoung individuals, low BMI may be associated withphysical fitness and a lower risk of fracture. In contrast,in the elderly, where hip fractures are more common,low BMI may more likely be related to frailty.

The strength of the present study is that the estimateof risk is derived from several studies in an internationalsetting from population-based cohorts, using the origi-nal individual participant data. The large sample sizepermitted us to examine the general relationship of BMIwith fracture risk and also the detailed relationship withage and level of BMI. The study also has some limita-tions. The definition of what was considered an osteo-porotic fracture was not the same in all cohorts. For hipfractures, the definition is similar in all cohorts and thereis substantial homogeneity between studies; here wefound higher gradients of risk across the range of BMI.For osteoporotic fractures our results were essentiallysimilar when we analyzed only those cohorts with uni-form documentation of fracture.

The use of BMI, rather than weight, as a measurefor body composition has the great advantage thatthere is less variability across countries and betweensexes as can be seen from Table 1. A potential draw-back is that BMI can be influenced by the height lossassociated with vertebral deformities. Therefore, inindividuals with important height loss, the risk con-ferred through BMI on fracture risk could be under-estimated [46]. The use of maximal attained heightrather than current height might be a solution in clin-ical practice, if it could be shown that risk predictioncould thereby be improved.

We conclude that low BMI confers a risk of fractureof substantial importance that is largely independent ofsex. The significance of BMI as a risk factor variesaccording to the level of BMI and to a lesser extent onage. Its validation on an international basis permits theuse of this risk factor, at least in the absence of a BMDmeasurement, in case-finding strategies. Even withBMD, a low BMI may remain an independent riskfactor for hip fracture in those with a BMI of less than20 kg/m2. These data also show that there should be noconflict between advice for weight control with reason-able target values, such as for the prevention of diabetes

or cardiovascular disease, and the prevention of osteo-porotic and hip fractures.

Acknowledgements We are grateful to Drs. T.V. Nguyen and J.R.Center for their help with the DOES Study. We would like to thankthe Alliance for Better Bone Health, Hologic, IGEA, Lilly, Lunar,Novartis, Pfizer, Roche and Wyeth for their unrestricted support ofthis work. We are also grateful to the EU (FP3/5), the InternationalOsteoporosis Foundation, the International Society for ClinicalDensitometry and the National Osteoporosis Foundation forsupporting this study.

References

1. Cummings SR, Nevitt MC, Browner WS, Stone K, Fox KM,Ensrud KE, Cauley J, Black D, Vogt TM (1995) Risk factorsfor hip fracture in white women. Study of Osteoporotic Frac-tures Research Group. N Engl J Med 332:767–773

2. Eddy DM, Johnston CC, Cummings SR, Dawson-Hughers B,Lindsay R, Melton LJ et al (1998) Osteoporosis: Review of theevidence for prevention, diagnosis and treatment and cost-effectiveness analysis. Osteoporos Int 8: [Suppl 4]

3. Honkanen RJ, Honkanen K, Kroger H, Alhava E, Tuppurai-nen M, Saarikoski S (2000) Risk factors for perimenopausaldistal forearm fractures. Osteoporos Int 11:265–270

4. Galvard H, Elmstahl S, Elmstahl B, Samuelsson SM, Roberts-son E (1996) Differences in body composition between femalegeriatric hip fracture patients and healthy controls: body fat ismore important as explanatory factors for the fracture thanbody weight and lean body mass. Aging (Milano) 8:282–286

5. Joakimsen RM, Fonnebo V, Magnus JH, Tollan A, SogaardAJ (1998) The Tromso study: body height, body mass indexand fractures. Osteoporos Int 8:436–442

6. Turner LW, Faile PA, Tomlinson R (1999) Osteoporosisdiagnosis and fracture. Orthop Nurs 18:21–27

7. Van der Voort DJ, Geusens PP, Dinant CJ (2001) Risk factorsfor osteoporosis related to their outcome: fractures. OsteoporosInt 12:630–638

8. Young Y, Myers AH, Provenzano G (2001) Factors associatedwith time to first hip fracture. J Aging Health 13:511–526

9. Roy DK, O’Neill TW, Finn JD, Lunt M, Silman AJ, Felsen-berg D, Ambrecht G, Banzer D, Benevolenskaya LI, Bhalla A,Bruges Armas J, Cannata JB, Cooper C, Dequeker J, Eastell R,Yershova O, Felsch B, Gowin W, Havelka S, Hoszowski K,Jajic I, Janot J, Johnell O, Kanis JA, Krag G, Lopez Vaz A,Lorenc R, Lyritis G, Masaryk P, Matthis C, Miazgowski T,Gennari C, Pols HAP, Poor G, Raspe HH, Reid DM, Rei-singer W, Scheidt-Nave C, Stepan J, Todd C, Weber K, WoolfAD, Reeve J (2003) Determinants of incident vertebral fracturein men and women: results from the European ProspectiveOsteoporosis Study (EPOS). Osteoporos Int 14; 19–26

10. Willig R, Luukinen H, Jalovaara P (2003) Factors related tooccurrence of hip fractures during a fall on the hip. PublicHealth 117:25–30

11. World Health Organization (1998) Obesity. Preventing andmanaging the global epidemic. World Health Organization,Geneva

12. Mokdad AH, Bowman BA, Ford ES, Vinicor F, Marks JS,Koplan JP (2001) The continuing epidemics of obesity anddiabetes in the United States. JAMA 286:1195–200

13. Peeters A, Barendregt JJ, Willekens F, Mackenbach JP, AlMamun AA, Bonneux L; NEDCOM, the Netherlands Epide-miology and Demographic Compression of Morbidity Re-search Group (2003) Obesity in adulthood and its consequencesfor life expectancy: a life-table analysis. Ann Intern Med138:24–32

14. Kanis JA, Delmas P, Burckhardt P, Cooper C, Torgerson D,on behalf of the European Foundation for Osteoporosis andBone Disease (1997) Guidelines for diagnosis and managementof osteoporosis. Osteoporos Int 7:390–406

1337

15. Royal College of Physicians (1999) Clinical Guidelines forstrategies to prevent and treat osteoporosis. Royal College ofPhysicians, London

16. Kwaliteitsinstituut voor de Gezondheidszorg CBP (2002) Os-teoporose: tweede herziene richtlijn. Utrecht, the Netherlands.ISBN 90–76906–23–8

17. Johnell O, Gullberg B, Kanis JA, Allander E, Elffors L, De-queker J, Dilsen G, Gennari C, Vaz AL, Lyritis G, Mazzuoli G,Miravet L, Passeri M, Cano RP, Rapado A, Ribot C (1995)Risk factors for hip fracture in European women—the MedosStudy. J Bone Miner Res 10:1802–1815

18. Kanis J, Johnell O, Gullberg B, Allander E, Elffors L, RanstamJ, Dequeker J, Dilsen G, Gennari C, Vaz A, Lyritis G, Maz-zuoli G, Miravet L, Passeri M, Perez CR, Rapado A, Ribot C(1999) Risk factors for hip fracture in men from southernEurope: the MEDOS study. Mediterranean OsteoporosisStudy. Osteoporos Int 9:45–54

19. Hofman A, Grobbee DE, de Jong PT, van den Ouweland FA(1991) Determinants of disease and disability in the elderly: TheRotterdam Elderly Study. Eur J Epidemiol 7:403–422

20. De Laet CEDH, Van Hout BA, Burger H, Hofman A, WeelAEAM, Pols HAP (1998) Hip fracture prediction in elderlymen and women: validation of the Rotterdam Study. J BoneMiner Res 13:1587–1593

21. O’Neill TW, Felsenberg D, Varlow J, Cooper C, Kanis JA,Silman AJ (1996) The prevalence of vertebral deformity inEuropean men and women: the European Vertebral Osteopo-rosis Study. J Bone Miner Res 11:1010–1017

22. Felsenberg D, Silman AJ, Lunt M, Ambrecht G, Ismail AA,Finn JD, Cockerill WC, Banzer D, Benevolenskaya LI, BhallaA, Bruges Armas J, Cannata JB, Cooper C, Dequeker J, EastellR, Ershova O, Felsch B, Gowin W, Havelka S, Hoszowski K,Jajic I, Janot J, Johnell O, Kanis JA, Kragl G, Lopez Vaz A,Lorenc R, Lyritis G, Masaryk P, Matthis C, Miazgowski T,Parisi G, Pols HAP, Poor G, Raspe HH, Reid DM, ReisingerW, Scheidt-Nave C, Stepan J, Todd C, Weber K, Woolf AD,Reeve J, O’Neill TW (2002) Incidence of vertebral fracture inEurope: results from the European Prospective OsteoporosisStudy, Epos. J Bone Miner Res 17:716–724

23. Ismail AA, Pye SR, Cockerill WC, Lunt M, Silman AJ, ReeveJ, 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, Hos-zowski 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 HAP, Poor G,Rapado A, Raspe HH, Reid DM, Reisinger W, Scheidt-NaveC, Stepan J, Todd C, Weber K, Woolf AD, O’Neill TW (2002)Incidence of limb fracture across Europe: results from theEuropean prospective osteoporosis study, EPOS. OsteoporosInt 13:565–571

24. Kreiger N, Tenenhouse A, Joseph L et al (1999) The CanadianMulticentre Osteoporosis Study CaMos: background, ratio-nale, methods. Can J Aging 18:376–387

25. Melton LJ III, Crowson CS, O’Fallon WM, Wahner HW,Riggs BL (2003) Relative contributions of bone density, boneturnover and clinical risk factors to long-term fracture predic-tion. J Bone Miner Res 18:312–318

26. Melton LJ III, Atkinson EJ, O’Connor MK, O’Fallon WM,Riggs BL (1998) Bone density and fracture risk in men. J BoneMiner Res 13:1915–1923

27. Johansson H, Oden A, Johnell O, Jonsson B, De Laet C,Oglesby A, McCloskey EV, Kayan K, Jalava T, Kanis JA(2004) Optimization of BMD measurements to identify highrisk groups for treatment—a test analysis. J Bone Miner Res19(6):906-913

28. Jones G, Nguyen TV, Sambrook PN, Kelly PJ, Gilbert C,Eisman JA (1994) Symptomatic fracture incidence in elderlymen and women. The Dubbo Osteoporosis EpidemiologyStudy, DOES. Osteoporos Int 4:277–282

29. Dargant-Molina P, Favier F, Grandjean H et al (1996) Fall-related factors and risk of hip fracture: the EPIDOS prospec-tive study. Lancet 348:145–149

30. Chapurlat R, Garnero P, Breart G, Meunier P, Delmas PD(2000) Serum estradiol and sex hormone-binding globulin andthe risk of hip fracture in elderly women: The EPIDOS study. JBone Miner Res 15:1835–1841

31. Garnero P, Sornay-Rendu E, Claustrat B, Delmas PD (2000)Biochemical markers of bone turnover, endogenous hormonesand the risk of fracture in postmenopausal women: the OFELYstudy. J Bone Miner Res 15:1526–1536

32. Honkanen R, Tuppurainen M, Kroger H, Alhava E, Saariko-ski S (1998) Relationships between risk factors and fracturesdiffer by type of fracture; a population-based study of 12,192perimenopausal women. Osteoporos Int 8:25–31

33. Svanborg A (1977) Seventy-year-old people in Gothenburg. Apopulation study in an industrialised Swedish city ii. JournalPresentation of Social and Medical Conditions. Acta MedicaScand 611 [Suppl]:5

34. Johansson C, Black D, Johnell O, Oden A, Mellstrom D (1998)Bone mineral density is a predictor of survival. Calcif Tissue Int63:190–196

35. Fujiwara S, Kasagi F, Yamada M, Kodama K (1997) Riskfactors for hip fracture in Japanese cohort. J Bone Miner Res12:998–1004

36. Fujiwara S, Kagasi F, Masunari N, Naito K, Suzuki G,Fukunage M (2003) Fracture prediction from bone mineraldensity in Japanese men and women. J Bone Miner Res18:1547–1553

37. Kanis JA, Oden A, Johnell O, Jonsson B, De Laet C, DawsonA (2002) The burden of osteoporotic fractures: a method forsetting intervention thresholds. Osteoporos Int 12:417–424

38. Higgins JPT, Thompson SG, Deeks JJ, Altman DG. (2003)Measuring inconsistency in meta-analyses. BMJ 327:557–60

39. Kiel DP, Felson DT, Anderson JJ, Wilson PWF (1987) Hipfracture and the use of estrogens in postmenopausal women.The Framingham Study. N Eng J Med 317:1169–1174

40. Malmivaara A, Heliovaara M, Knekt P, Reunanen A, AromaaA (1993) Risk factors for injurious falls leading to hospital-ization or death in a cohort of 19,500 adults. Am J Epidemiol138:384–394

41. Bischoff HA, Stahelim HB, Dick W, Akos R, Kuecht M, SalisC, Nebiker M, Thieler R, Pfeifer M, Bergerow B, Lew RA,Conzelmann M (2003) Effects of vitamin D and calcium sup-plementation on falls: a randomized controlled trial. J BoneMiner Res 18:343–351

42. Wootton R, Brereton PJ, Clark MB, Hesp R, Hodkinson HM,Klenerman L, Reeve J, Slavin G, Tellez M (1979) Fracturedneck of femur in the elderly: an attempt to identify patients atrisk. Clin Sci 57:93–101

43. Nilsson BE (1970) Spinal osteoporosis and femoral neck frac-ture. Clin Orthop Relat Res 68:93–95

44. Marshall D, Johnell O, Wedel H (1996) Meta-analysis of howwell measures of bone mineral density predict occurrence ofosteoporotic fractures. BMJ 312:1254–1259

45. Sirola J, Kroger H, Honkanen R, Sandini L, Tuppurainen M,Jurvelin JS, Saarikoski S (2003) Risk factors associated withperi- and postmenopausal bone loss: does HRT prevent weightloss-related bone loss? Osteoporos Int 14:27–33

46. Davison KS, Siminoski K, Chik C, Jen H, Warshawski R, LeeK (2003) Impact of height loss due to vertebral fractures onbody mass index. J Bone Miner Res [Suppl]:S243

1338