Association of atypical femoral fractures with bisphosphonate use by patients with varus hip geometry

Transcription

1 Washington University School of Meicine Digital Open Access Publications 2014 Association of atypical femoral fractures with bisphosphonate use by patients with varus hip geometry Jennifer E. Hagen University of Marylan - Baltimore Anna. Miller Wake Forest Baptist Health Susan M. Ott University of Washington - Seattle Campus Michael Garner Washington University School of Meicine in St. Louis Saam Morshe University of California - San Francisco See next page for aitional authors Follow this an aitional works at: Recommene Citation Hagen, Jennifer E.; Miller, Anna.; Ott, Susan M.; Garner, Michael; Morshe, Saam; Jeray, Kyle; Alton, Timothy B.; Ren, Dennis; Abblitt, W. Parker; an Krieg, James C.,,"Association of atypical femoral fractures with bisphosphonate use by patients with varus hip geometry." The journal of bone an joint surgery.96, (2014). This Open Access Publication is brought to you for free an open access by Digital Commons@Becker. It has been accepte for inclusion in Open Access Publications by an authorize aministrator of Digital Commons@Becker. For more information, please contact engeszer@wustl.eu.

2 Authors Jennifer E. Hagen, Anna. Miller, Susan M. Ott, Michael Garner, Saam Morshe, Kyle Jeray, Timothy B. Alton, Dennis Ren, W. Parker Abblitt, an James C. Krieg This open access publication is available at Digital

3 1905 COPYRIGHT Ó 2014 BY THE JOURAL OF BOE AD JOIT SURGERY, ICORPORATED Association of Atypical Femoral Fractures with Bisphosphonate Use by Patients with Varus Hip Geometry Jennifer E. Hagen, MD, Anna. Miller, MD, Susan M. Ott, MD, Michael Garner, MD, Saam Morshe, MD, Kyle Jeray, MD, Timothy B. Alton, MD, Dennis Ren, BS, W. Parker Abblitt, MD, an James C. Krieg, MD Backgroun: There is increasing evience associating atypical femoral fractures with prolonge exposure to bisphosphonate therapy. The cause of these fractures is unknown an likely multifactorial. This stuy evaluate the hypothesis that patients with primary osteoporosis who sustain atypical femoral fracture(s) while on chronic bisphosphonate therapy have a more varus proximal femoral geometry than patients who use bisphosphonates for primary osteoporosis but o not sustain a femoral fracture. Methos: The femoral neck-shaft angle was measure on the raiographs of 111 patients with atypical femoral shaft fracture(s) an thirty-three asymptomatic patients; both groups were on chronic bisphosphonate therapy. Patients with characteristic lateral cortical thickening, stress lines, an thigh pain were inclue in the fracture group. Results: The mean neck-shaft angle of the patients who sustaine atypical femoral fracture(s) while taking bisphosphonates (case group) iffere significantly from that of the patients on bisphosphonate therapy without a fracture (129.5 versus ; p < 0.001). Fifty-three (48%) of the patients in the case group ha a neck-shaft angle that was lower than the lowest angle in the control group (128 ). Sie-to-sie comparison in patients with a unilateral pathologic involvement an an asymptomatic contralateral lower limb i not emonstrate any significant ifference between the neck-shaft angles in the two limbs. Conclusions: Patients on chronic bisphosphonate therapy who presente with atypical femoral fracture(s) ha more varus proximal femoral geometry than those who took bisphosphonates without sustaining a fracture. Although no causative effect can be etermine, a fining of varus geometry may help to better ientify patients at risk for fracture after long-term bisphosphonate use. Level of Evience: Prognostic Level III. See Instructions for Authors for a complete escription of levels of evience. Peer Review: This article was reviewe by the Eitor-in-Chief an one Deputy Eitor, an it unerwent bline review by two or more outsie experts. The Deputy Eitor reviewe each revision of the article, an it unerwent a final review by the Eitor-in-Chief prior to publication. Final corrections an clarifications occurre uring one or more exchanges between the author(s) an copyeitors. An estimate 10 million Americans have osteoporosis, an more than 1.5 million fractures per year are attribute to this isease 1. Since their introuction in 1995, bisphosphonates have become the stanar of care for treatment of osteoporosis, an more than 4 million women in the Unite States were taking bisphosphonates in In 2005, reports began to appear of atypical femoral fractures in patients on bisphosphonate therapy for a prolonge perio of time. Many stuies have evaluate the association between the use of bisphosphonates an the occurrence of an atypical fracture 3-7, an a recent meta-analysis suggeste an increase risk for atypical femoral fractures in patients taking a bisphosphonate 3.Ina stuy of 716 femoral shaft fractures in patients on bisphosphonate therapy, Park-Wyllie et al. foun that patients on bisphosphonate therapy for longer than five years ha a 2.74 times increase risk of iaphyseal femoral fracture as compare with a control group 4. Disclosure: One or more of the authors receive payments or services, either irectly or inirectly (i.e., via his or her institution), from a thir party in support of an aspect of this work. In aition, one or more of the authors, or his or her institution, has ha a financial relationship, in the thirty-six months prior to submission of this work, with an entity in the biomeical arena that coul be perceive to influence or have the potential to influence what is written in this work. Also, one or more of the authors has ha another relationship, or has engage in another activity, that coul be perceive to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitte by authors are always provie with the online version of the article. J Bone Joint Surg Am. 2014;96:

4 1906 B ISPHOSPHOATE U SE BY PATIETS WITH VARUS H IPS TABLE I Patient Characteristics Case Control o. of patients o. of hips available for measurement Mean age (range) (yr) 68.7 (46-91) 66.8 (39-85) Mean bisphosphonate exposure (range) (yr) 7.4 (1-20)* 5 (1-14) Fracture characteristics (no. of patients) Complete 88 Stress reaction/beaking 23 Bilateral 42 (38%) Bilateral complete 22 Complete an contralateral stress 15 Bilateral stress 5 *ot quantifie for thirteen patients. Schilcher et al. foun the os ratio of atypical fracture in patients on bisphosphonate therapy to be 33.3 after ajusting for comorbi conitions 8. The risk increase with prolonge use an was ten times higher than normal after just two years 8. We believe that other risk factors must also be present to explain why (1) only a small percentage (0.02% to 0.2%) of patients taking bisphosphonates on a chronic basis evelop these fractures an (2) atypical femoral fractures sometimes occur in patients who are not on long-term therapy 5. It is possible that patients who evelop atypical femoral fractures have an anatomic biomechanical preisposition. We hypothesize that patients on chronic bisphosphonate therapy who sustain atypical femoral fracture(s) or isplay raiographic characteristics consistent with a stress fracture or lateralcorticalbeaking 7 are more likely to have varus proximal femoral anatomy than exposure-matche controls. Materials an Methos We performe a multicenter retrospective case-control stuy of patients from six institutions. Institutional review boar approval was obtaine at each iniviual site. A series of 111 patients who ha been treate for a complete or incomplete atypical femoral fracture were ientifie. Each patient ha raiographic characteristics of the atypical femoral fracture associate with chronic bisphosphonate use. To be inclue in the stuy group, those without a complete fracture ha to have thigh or hip pain in aition to raiographic evience of either lateral cortical beaking or a transverse stress line base on ASBMR (American Society for Bone an Mineral Research) criteria 6. These 111 patients constitute the case group. All raiographs were reviewe an approve for inclusion by senior authors (J.C.K., A..M., S.M.O, S.M., K.J., an M.G.) at each institution. Exclusion criteria inclue bisphosphonate therapy for iseases other than primary osteoporosis, previous hip arthroplasty, an raiographs not being available. A control group of thirty-three patients was ientifie. These patients ha a ocumente uration of bisphosphonate use for primary osteoporosis; no history of fracture; no history of proromal thigh or hip pain; an hip, pelvic, or femoral raiographs alreay in the meical recor (Table I). One hunre an forty-four patients were inclue in the analysis, an raiographs of 255 hips were available for measurement. All patients are female. The raiographs were compile an reviewe at three sites where the senior authors (J.C.K., A..M., an M.G.) practice. One reviewer was a trauma traine attening surgeon. The other two were a senior resient an a meical stuent, who coul consult with trauma traine attening surgeons for guiance. The reviewers were aware of the hypothesis of the stuy. A sample of raiographs was sent to all four reviewers, an the interobserver ifference between the neck-shaft angle measurements was <1 (stanar eviation [SD], 1.6 ). Intraobserver reliability was not calculate. Measurements were mae on anteroposterior raiographs of both hips, when raiographs of both were available. The neck-shaft angle was efine as the angle forme by the intersection of a line own the center of the femoral neck an a line through the center of the femoral shaft 9. We i not arbitrarily assign a cutoff neck-shaft angle to efine varus geometry; rather, it was recore as a continuous variable. Measurements of each hip were recore as inepenent ata, an measurements in the case group were ivie into those on the pathologic sie(s) an those on the asymptomatic sie. The measurements from the pathologic sie(s) were compare with the control group as well as the asymptomatic sies of the patients with a unilateral pathologic involvement. The asymptomatic sies of these patients were measure to etermine if there was a sie-to-sie variation in anatomy that correlate with unilateral pathologic involvement. For patients for whom only post-fracture raiographs were available, the center of the femoral shaft was measure proximal to the level of the fracture to eliminate improper measurement ue to any post-fixation malalignment. When rotation of the limb preclue visualization of the neck-shaft junction, the hip was exclue from measurement. Statistical Analysis An inepenent two-sample t test for continuous variables was use to compare the results between the control an stuy populations an within the stuy population. All statistical assessments were two-sie an were evaluate at the 0.05 level of significance. Source of Funing One senior author (M.G.) receive research coorinator support from Synthes. Results The mean neck-shaft angle in the case group (129.5 ) was significantly lower (p < 0.001) than that in the controls (133.8 ). The mean neck-shaft angle on the pathologic sies(s) in the case group (129.9 ) was also significantly lower (p < 0.001) than the mean angle in the control group.

5 1907 B ISPHOSPHOATE U SE BY PATIETS WITH VARUS H IPS Fig. 1 Plot of neck-shaft angles (SA) in each group. Sie-to-sie comparison of patients with a unilateral pathologic involvement i not emonstrate any significant ifference in the neck-shaft angle between the pathologic sie (129.8 ) an the asymptomatic, contralateral lower limb (128.2 ). The lowest recore neck-shaft angle was 128 in the control group an118 in the case group. Fifty-three (48%) of the patients in the case group ha a neck-shaft angle of <128 on the pathologic sie (Fig. 1). Discussion Atypical femoral fractures represent a challenge for patients with osteoporosis an osteopenia. Many have speculate about the etiology, an it is likely multifactorial, although no unerlying mechanism has been emonstrate. This stuy was esigne to etermine if patients with atypical fractures have a more varus proximal femoral geometry than an age an exposurematche control group. The mean neck-shaft angle of the patients with a fracture was significantly lower than that of the controls, an no control patient ha a neck-shaft angle <128. Overall, we foun an association between proximal femoral geometry an the presence of atypical femoral fractures in this population. In the recently upate ASBMR task force consensus statement, it was conclue that atypical femoral fractures may represent stress fractures that progress over time 7. These stress fractures are unique in that the transverse component of the fracture begins in the lateral cortex, whereas exercise-inuce stress fractures typically initiate in the meial cortex. Very little is known about femoral geometry in this patient population. Proximal femoral strength is ecrease with a varus mechanical axis 10,11. Biomechanical stuies have shown that trochanteric an femoral shaft fractures are more common in patients with low neck-shaft angles 12,13. Koh et al. reviewe the raiographs of forty-eight patients with atypical fractures an foun that the fractures clustere in the lateral cortex at the region of maximal tensile loaing 14. Sasaski et al. measure the variations in the anatomic axes of the femora of patients with low-energy femoral shaft fractures 15. These patients were taking meications for osteoporosis but not specifically bisphosphonates. The patients with fractures ha a significant increase in the lateral an anterior bow of the femur, an this geometry was thoughttoresultinanimbalanceinstrainsseeninthefemur. We believe that this imbalance also occurs with an alteration in the mechanical axis of the lower extremity prouce by a varus femoral neck. A large number of patients in our case group ha a neckshaft angle of <128 onthepathologicsie.thestuywasnot powere to calculate the sensitivity an specificity of a neckshaft angle of <128 preicting the risk of fracture, but stuying this risk in a larger cohort woul be useful. A large proportion (38%) of the patients in our stuy ha bilateral pathological involvement. This is consistent with the finings from the stuy performe by Lo an colleagues, in which 40% of the population ha bilateral complete fracture, an an aitional 21% ha contralateral stress-reaction changes 16. They foun a higher proportion of fractures in women of Asian escent, a factor not controlle for in our stuy. akamura et al. foun that women of Japanese escent ha lower neckshaft angles than white American women, an Japaneseescent women coul be an at-risk subset of the population 17. This stresses the importance of obtaining imaging of both femora in patients with a history of atypical femoral fracture an thigh pain.

6 1908 B ISPHOSPHOATE U SE BY PATIETS WITH VARUS H IPS Our stuy is limite by all of the factors that routinely affect retrospective reviews. We i not control for patient factors, incluing boy mass inex, smoking status, activity level, an bone mineral ensity. Also, we stuie only a small sample of the more than 4 million women on bisphosphonate therapy in the Unite States. The reliability of hip measurements has been stuie with a variety of moalities in patients with various pathological conitions. In general, the reliability of the neck-shaft angle has been shown to be goo. A recent stuy of patients with cerebral palsy showe an overall intraobserver an interobserver reliability of neck-shaft-angle measurements on hip raiographs of an 0.912, respectively 18. Marmor et al. foun that neck-shaft-angle measurements varie <5 with <35 of limb rotation, an the measurements were always greater than the true value 19. Thus, it is unlikely that our measurements overestimate the actual varus in our patients; if anything, they unerestimate it. Many of the measurements in the group with a complete fracture were performe on injury raiographs, for which rotation is the most ifficult to control, an it is likely the true neck-shaft angle in this group was in even more varus than we reporte. Bias coul have been introuce into the measurements as the majority of those in the case group were one on injury raiographs or raiographs showing the stress reaction; therefore, the reviewers were not bline to the presence of the pathological conition. It was impossible to avoi this as many of the patients ha not sought treatment at our facilities prior to their injury or i not have any reason to have pelvic raiographs prior to the evelopment of symptoms. However, there was not a significant ifference between the measurements on the pathologic an nonpathologic sies of our stuy group, thus proviing a marker for internal valiity of our measurement techniques. Another limitation of this stuy is that the average bisphosphonate exposure times of the stuy an control patients were not ientical. Thirteen patients in the case group i not have numerical ocumentation of their uration of bisphosphonate use; charts inicate it was long term. Despite this, the mean exposure times for both groups were greater than those previously ocumente to incur risk 4,8. The link between atypical femoral fracture an bisphosphonates is an association; it has not been efinitively shown to be causation. The exact uration of exposure that is neee for an increase fracture risk is unknown. Schilcher et al. foun the os ratio of fracture to be ten times higher after just two years of exposure 8.Asourstuy was a retrospective analysis an we ocumente fracture or stress reaction in all of the patients in the case group, we can assume that it was an at-risk population. It is possible that the control group was evaluate early in the time course of their isease. We have no current metho for monitoring their progression to fracture. A prospective an longituinal stuy coul better control for this factor by ocumenting the uration of patients exposure to bisphosphonates at the time of enrollment an more thoroughly tracking their actual use. The ability to monitor for progression to fracture woul help etermine the relative risk of fracture. In conclusion, we foun an association between varus proximal femoral geometry an a propensity to sustain atypical femoral fractures in patients on long-term bisphosphonate therapy. We agree with the ASBMR task force s conclusion that more nees to be one to evaluate the impact of femoral geometry on this patient population 16. The neck-shaft angle is only one variable, but it appears to have a high correlation with atypical femoral fractures, an follow-up measurements are easy, low cost, an noninvasive. The strength of this stuy was insufficient for us to conclusively avise a change in management of these patients at this time, but we have mae a strong argument to use this as a pilot stuy to prompt further investigation. At a minimum, patients shoul be counsele to be aware of symptoms that accompany these injuries. Larger stuies shoul be one in a longituinal fashion to etermine the true risk of progression to fracture. While this is likely only one piece of the picture, we think that our finings a a useful clinical marker that coul help ientify an at-risk subset of this population. n OTE: The authors acknowlege Justin Langan for his contribution to the ata-collection portion of this project. Jennifer E. Hagen, MD University of Marylan, 22 South Greene Street, Baltimore, MD aress: Jeh29@case.eu Anna. Miller, MD Wake Forest Baptist Health, 131 Miller Street, Winston-Salem, C Susan M. Ott, MD University of Washington, 4245 Roosevelt Way.E., Box 35470, Seattle, WA Michael Garner, MD Washington University, 4921 Parkview Place, St. Louis, MO Saam Morshe, MD University of California San Francisco, r Street, San Francisco, CA Kyle Jeray, MD Steaman Hawkins Clinic, 200 Patewoo Drive, Suite C100, Greenville, SC Timothy B. Alton, MD University of Washington, 1959.E. Pacific Street, #356500, Seattle, WA 98195

7 1909 B ISPHOSPHOATE U SE BY PATIETS WITH VARUS H IPS Dennis Ren, BS Tulane School of Meicine, 1430 Tulane Avenue, ew Orleans, LA W. Parker Abblitt, MD Wake Forest School of Meicine, Meical Center Boulevar, Winston Salem, C James C. Krieg, MD The Rothman Institute, Jefferson University, 925 Chestnut Street, Philaelphia, PA References 1. Keen RW. Buren of osteoporosis an fractures. Curr Osteoporos Rep Sep;1(2): Siris ES, Pasquale MK, Wang Y, Watts B. Estimating bisphosphonate use an fracture reuction among US women age 45 years an oler, J Bone Miner Res Jan;26(1): Gemintas L, Solomon DH, Kim SC. Bisphosphonates an risk of subtrochanteric, femoral shaft, an atypical femur fracture: a systematic review an metaanalysis. J Bone Miner Res Aug;28(8): Epub 2013 Feb Park-Wyllie LY, Mamani MM, Juurlink D, Hawker GA, Gunraj, Austin PC, Whelan DB, Weiler PJ, Laupacis A. Bisphosphonate use an the risk of subtrochanteric or femoral shaft fractures in oler women. JAMA Feb 23;305(8): Wells G, Cranney A, Peterson J, Boucher M, Shea B, Robinson V, Coyle D, Tugwell P. Riseronate for the primary an seconary prevention of osteoporotic fractures in postmenopausal women. Cochrane Database Syst Rev. 2008;(1):CD Epub 2008 Jan Shane E, Burr D, Ebeling PR, Abrahamsen B, Aler RA, Brown TD, Cheung AM, Cosman F, Curtis JR, Dell R, Dempster D, Einhorn TA, Genant HK, Geusens P, Klaushofer K, Koval K, Lane JM, McKiernan F, McKinney R, g A, ieves J, O Keefe R, Papapoulos S, Sen HT, van er Meulen MC, Weinstein RS, Whyte M; American Society for Bone an Mineral Research. Atypical subtrochanteric an iaphyseal femoral fractures: report of a task force of the American Society for Bone an Mineral Research. J Bone Miner Res ov;25(11): Shane E, Burr D, Abrahamsen B, Aler RA, Brown TD, Cheung AM, Cosman F, Curtis JR, Dell R, Dempster DW, Ebeling PR, Einhorn TA, Genant HK, Geusens P, Klaushofer K, Lane JM, McKiernan F, McKinney R, g A, ieves J, O Keefe R, Papapoulos S, Howe TS, van er Meulen MC, Weinstein RS, Whyte MP. Atypical subtrochanteric an iaphyseal femoral fractures: secon report of a task force of the American Society for Bone an Mineral Research. J Bone Miner Res Jan;29(1):1-23. Epub 2013 Oct Schilcher J, Michaëlsson K, Aspenberg P. Bisphosphonate use an atypical fractures of the femoral shaft. Engl J Me May 5;364(18): Kay RM, Jaki KA, Skaggs DL. The effect of femoral rotation on the projecte femoral neck-shaft angle. J Peiatr Orthop ov-dec;20(6): Antapur P, Prakash D. Proximal femoral geometry: a raiological assessment. J Arthroplasty Sep;21(6): Gong H, Zhang M, Fan Y, Kwok WL, Leung PC. Relationships between femoral strength evaluate by nonlinear finite element analysis an BMD, material istribution an geometric morphology. Ann Biome Eng Jul;40(7): Epub 2012 Jan El-Kaissi S, Pasco JA, Henry MJ, Panahi S, icholson JG, icholson GC, Kotowicz MA. Femoral neck geometry an hip fracture risk: the Geelong osteoporosis stuy. Osteoporos Int Oct;16(10): Epub 2005 Aug Thevenot J, Pulkkinen P, Kuhn V, Eckstein F, Jämsä T. Structural asymmetry between the hips an its relation to experimental fracture type. Calcif Tissue Int Sep;87(3): Epub 2010 Jun Koh JS, Goh SK, Png MA, g AC, Howe TS. Distribution of atypical fractures an cortical stress lesions in the femur: implications on pathophysiology. Singapore Me J Feb;52(2): Sasaki S, Miyakoshi, Hongo M, Kasukawa Y, Shimaa Y. Low-energy iaphyseal femoral fractures associate with bisphosphonate use an severe curve femur: a case series. J Bone Miner Metab Sep;30(5): Epub 2012 May Lo JC, Huang SY, Lee GA, Khanelwal S, Provus J, Ettinger B, Gonzalez JR, Hui RL, Grimsru CD. Clinical correlates of atypical femoral fracture. Bone Jul; 51(1): Epub 2012 Mar akamura T, Turner CH, Yoshikawa T, Slemena CW, Peacock M, Burr DB, Mizuno Y, Orimo H, Ouchi Y, Johnston CC Jr. Do variations in hip geometry explain ifferences in hip fracture risk between Japanese an white Americans? J Bone Miner Res Jul;9(7): Chung CY, Lee KM, Park MS, Lee SH, Choi IH, Cho TJ. Valiity an reliability of measuring femoral anteversion an neck-shaft angle in patients with cerebral palsy. J Bone Joint Surg Am May;92(5): Marmor M, ystuen C, Ehemer, McClellan RT, Matityahu A. Accuracy of in situ neck-shaft angle an shortening measurements of the anatomically reuce, varus malreuce an shortene proximal femur: can we believe what we see on the postoperative films? Injury Jun;43(6): Epub 2011 Oct 30.