Reprinted from JAMA ® The Joumal of the Amencan Medical Association Janual}' 12, 1994 Volume 271 Copyright 1994, American Medical AssoclaNon Fall Severity and Bone Mineral Density as Risk Factors for Hip Fracture in Ambulatory Elderly Susan L. Greenspan, MD; Elizabeth R. Myers, PhD; Lauri A. Maitland, MPH; Neil M. Resnick, MD; Wilson C. Hayes, PhD Objective.- To determine the relative importance of fall characteristics, body habitus, and femoral bone mineral density (BMD) in predicting hip fracture in community-dwelling elderly. Design.-Prospective case-control stUdy. Setting.-Community-based academic hospital. Participants.-A total of 149 ambulatory, community-dwelling fallers (126 women, 23 men) aged 65 years and older, including 72 case patients (fallers with hip fracture) and 77 control fallers (fallers with no hip fracture). Main Outcome Measures.-Fall characteristics, body habitus, femoral BMO. Results.-Significant and independent risk factors for hip fracture in both sexes were direction of the fall (adjusted odds ratio [OR], 5.7; 95% confidence interval [CI], 2.3 to 14.0; P<.001); femoral neck BMO (a decrease of 1 SO; adjusted OR, 2.7; 95% CI, 1.6 to 4.6; P<.001); potential energy of the fall (an increase of 1 SO; ad- justed OR, 2.8; 95% CI, 1.5 to 5.2; P<.001); and body mass index (a decrease of 1 SO; adjusted OR, 2.2; 95% CI, 1.2 to 3.8; P<.01). Importantly, the OR for the fall direction was unaffected by the addition or removal of BMO from the model. Conclusions.-We conclude that among elderly fallers-in most of whom hip BMO is already less than the fracture threshold-fall characteristics and body habitus are important risk factors for hip fracture and touch on a domain entirely missed by knowledge of BMO. These data suggest new targets for preventive therapy. In addition to the maintenance of bone density, reductions in fall severity (eg, by use of trochanteric padding or enhancement of muscle strength) may pro- vide additional strategies for prevention of hip fracture in this age group. EACH year, more than 250000 Ameri- cans fracture their hips.l.2 In addition to the $8.7 billion cost associated with such fractures,2 patients suffer acute morbid- From the D,v,s,ons of Bone and Minerai Metab- olism and Gerontology. Department of MediCine (Dr Greenspan and Ms Maitland). OrthOpaedic Biome- chaniCS Laboratory. Department 01 OrthopaediC Sur- gery (Drs Myers and Hayes). Charles A. Dana Re- search Institute (Drs Greenspan and Hayes). Beth Israel Hospital and Harvard Medical School. Boston. Mass: D,v,s,on of Gerontology. Brigham and Womens Hospital and Harvard Medical School (Drs Greenspan and Resnick). Boston. Mass: and Geriatric Research, Education and Clinical Center. BrocktonM'est Roxbury VA Medical Center (Dr Resnick). West Roxbury. Mass. Repnnt requests to D,v,s,on of Bone and Minerai Melabolism. Beth Israel Hospital. 330 Brookline Ave. Boston. MA 02215 (Dr Greenspan) (JAMA. 1994;271:128·133) ity, long-term loss of function, substan- tial risk for institutionalization, and an increased risk of death. 3 However, al- though more than 90% of hip fracture patients are over 70 years of age,4 little is known about how such fractures could be prevented in this 'age groUp.5 While several studies have found that low bone mass and increasing age are associated with an increased risk of hip fracture in this population,G .• virtually all investi- gators have also found a considerable overlap in bone density values between hip fracture patients and age- and gen· der-matched controls.H.g Other factors have also been associated with hip frac- ture, including use of long-acting ben- zodiazepines,lO impaired vision,lI lower limb dysfunction, neurological condi- tions, and barbiturate use. 12 The most commonly cited factor for increased risk of hip fractures in the elderly is falling, because more than 90% of hip fractures occur following a fall. 12.13 However, less than 5% of falls result in a hip fracture,14-16 and the factors that protect the remaining 95% of falls from resulting in hip fractures are unknown. Thus, falling appears to be necessary but not sufficient for the occurrence of a hip fracture. Moreover, although at- tempts have been made to decrease the risk of falling, to date these have been largely ineffective in controlled trials. An alternative hypothesis proposed by us and others is that fracture risk is related to both the characteristics of the fall and the individual's vulnerability.17.18 Thus, the most appropriate predictive model would be one that incorporates the characteristics ofthe fall, use of pro- tective responses (eg, breaking the force of the fall \\;th an outstretched arm), and the energy-absorbing capacity of soft tissue over the hip, in addition to femoral bone mass. This approach would examine how people fall, rather than whether they fall. 19 This hypothesis has important clinical implications for future preventive strategies, but to date there have been few data available to support it. Recently we examined the relation- ship of mechanics of falling to the risk of hip fracture. 1u In this study of elderly nursing home residents, we found that the adjusted odds ratio (OR) of hip fracture from a fall involving direct im- pact on the hip region was 21.7 (95% confidence interval [CI], 8.2 to 58).lY The potential energy associated with the fall was also important, as was the body mass index (BMIl. However, bone min- eral density (BMD) was not assessed. Therefore, to corroborate our previous 128 JAMA. January 12. 1994-Vol 271. No 2 Fall Seventy. Bone Minerai Density. and HIp Fracture--Greenspan et al