Evaluation of Cortical Microarchitecture and Biomechanical Properties 6-12 Months after Atypical Femoral Fracture

Atypical Femoral Fractures (AFF) are low energy femoral fractures associated with long-term bisphosphonate therapy. Our aim was to evaluate the cortical bone parameters and biomechanical proprieties by high-resolution peripheral quantitative computed tomography (HR-pQCT) at distal tibia and radius and by iliac crest histomorphometry, 6-12 months after AFF. Twelve patients with AFF were evaluated. Cortical volumetric bone mineral density (Ct.vBMD), cortical thickness (Ct.Th), cortical porosity (Ct.Po), Stiffness (S) and failure load (F.Load) were assessed by HR-pQCT. Histomorphometry of transiliac bone biopsy evaluated Ct.Th, Ct.Po and bone remodeling. All bone parameters were compared with healthy controls. The mean age of the patients was 60.5±12.8 years, 91.6% women, 91.6% used BPs (median: 6 years[1-10yr]). Rheumatic diseases were observed in 50% patients and glucocorticoid (GC) use in 41.6%. At distal tibia and radius, Ct.vBMD was low/normal in 66.6%, Ct.Th low/normal in 77.7% and Ct.Po high/ normal in more than 66.6%. More than 50% of the patients had low S and F.Load at these sites. S and F.Load were positively correlated with Ct.Th (r=0.724, p=0.028; r=0.718, p=0.02, respectively) at tibia. At radius, S and F.Load were also positively correlated with Ct.Th (r=0.735, p=0.024; r=0.717, p=0.030, respectively). Iliac crest histomorphometry exhibited low Ct.Th in 60%, low Ct.Po in 100%, and suppressed bone remodeling in 100% of patients. Our data demonstrated that among patients with AFF, rheumatic diseases and GC use are common. Cortical and strength bone parameters were deteriorated at distal tibia/radius and iliac crest, suggesting the presence of significant bone fragility after 6-12 months of AFF follow-up.

Typically, patients with AFF demonstrate some peculiar findings, such as prodromal thigh pain for weeks to months prior to the fracture, minimal or no trauma, transverse or short oblique configuration, and little or noncomminuted and bilateral fractures (either simultaneous or sequential) [1,2]. Moreover, patients with AFF exhibit cortical bone thickening and periosteal reaction on radiographs of the femoral shaft [2], suggesting that AFF may occur due to cortical stress. The cortical bone exerts an important role in bone stiffness and contributes to whole bone mechanical competence. In fact, cortical thickness is a strong predictor of bone strength and resistance to fracture [8,9]. Recently, the involvement of cortical bone has been studied and assessed by new imaging methods such as high-resolution peripheral quantitative computed tomography (HR-pQCT) [10] and bone histomorphometry [11]. HR-pQCT is a noninvasive technique that assesses separately the cortical and trabecular components at peripheral sites such as the distal tibia and radius. HR-pQCT is a method for assessing bone microarchitecture and bone stiffness [10]. Bone histomorphometry is the quantitative histological analysis of the bone obtained from transiliac bone biopsies and is the gold standard in the static and dynamic evaluation of bone parameters. It has recently been used to evaluate the cortical compartment [11,12]. No study has evaluated cortical microstructure and bone strength parameters using HR-pQCT and bone histomorphometry assessments. Therefore, the aim of this study is to evaluate patients with AFF after 6-12 months of follow-up, from a rheumatology service, with emphasis on (1) the cortical bone microarchitecture and bone strength parameters using HR-pQCT and (2) the cortical parameters by iliac crest histomorphometry.

Study Design
This study had the approval of the Ethics Committee of the University of Sao Paulo. We conducted a cross-sectional, case-control study of 12 subjects who met the AFF criteria according to the American Society for Bone and Mineral Research (ASBMR) [2] and were regularly followed-up at the outpatient Osteometabolic Diseases and Osteoporosis, Rheumatology Service, Clinics Hospital of the University of Sao Paulo between January 2010 and December 2018. The demographic and clinical data, including race, age, BP indication, BP duration, clinical manifestations of AFF, diagnosis of rheumatic disease, glucocorticoid (GC) use and treatment post-AFF were obtained through interviews with the patients and reviews of their medical records. The bone parameters obtained by dual-energy X-ray absorptiometry (DXA), HR-pQCT and iliac crest biopsy histomorphometry were evaluated 6-12 months after AFF.

Atypical Femoral Fracture
AFF was defined according to the second report of the ASBMR Task Force 2013 [2]. All five major criteria were required for the definition of AFF: (1) located anywhere along the femur from just distal to the lesser trochanter to just proximal to the supracondy-lar flare, (2) minimal trauma, (3) noncomminuted or minimally comminuted, (4) fracture line originating at the lateral cortex and substantially transverse in orientation and (5) complete fractures that may have medial spikes, incomplete only involving the lateral cortex. The following minor criteria, although not decisive for the diagnosis, were also evaluated: the presence of localized periosteal or endosteal thickening of the lateral cortex ("beaking" or "flaring"), prodromal symptoms such as pain in the groin or thigh and bilateral fractures [2].

Reference Curves of Healthy Individuals of HR-pQCT and Histomorphometric Parameters
Patients' parameters obtained using HR-pQCT were compared to data from a HR-pQCT reference curve of Brazilian healthy individuals, according to sex and age [10]. Moreover, patients' histomorphometric static parameters were compared to histomorphometric variables from a reference curve of Brazilian healthy individuals [13], and the dynamic variables were compared to Melsen et al data [14], according to sex and age.

Radiological Assessment
Two rheumatologists (MOP and DSD) independently evaluated the bilateral femoral radiographs in accordance with the ASBMR definition [2]. Subtrochanteric or femoral shaft fracture sites were assessed. Major features such as complete fracture and the presence or absence of a medial spike were evaluated, considering also that focal cortical thickening and a transverse fracture on the lateral side are the elements with a high accuracy for the diagnosis of AFF [2,15]. When the two rheumatologists had discordant evaluations, another rheumatologist (RMRP) made the final decision after an independent evaluation. A technetium bone scintigraphy or magnetic resonance imaging were performed to detect bilateral AFF when there was no visible radiographic alteration in the contralateral femur [15]. SCANCO Medical, Switzerland) were performed after immobilizing the nondominant distal tibia and forearm in a carbon fiber shell as previously described [17]. The region of interest was defined on a scout film by manual placement of a reference line at the endplate of the tibia or radius, with the first slice 22.5 mm and 9.5 mm proximal to the reference line at the tibia and radius, respectively [17][18][19]. The standard protocol for image acquisition and analyses was applied, and a detailed description of the measurement protocol at our center was previously described [17]. The standard and advanced cortical (auto segmentation) methods of the scanner were used for the analysis with appropriate references [10,20]. All examinations were conducted by a trained biomedical scientist (JCA), who also carefully examined each scan for motion artifacts. In the case of significant motion artifacts, a second examination was performed (by RMRP). All HR-pQCT images were scored based on a motion scale ranging from 0 (no movement) to 4 (significant blurring of the periosteal surface, discontinuities in the cortical layer) [21]. For this study, scans with a score of 4 were excluded. Quality control was monitored daily using a phantom calibration provid- [kN]) [17,23].

Transiliac Bone Biopsy
Five patients with AFF underwent transiliac bone biopsy. All patients received double tetracycline labeling before the biopsy (20 mg/kg/day for 3 days, administered in two separate doses, separat-

Bone Histomorphometry
The bone fragments were composed of two cortices (internal and external). Bone material designated for histomorphometry was fixed in 70% ethanol and processed according to Malluche and Faugere [24]. Using a Polycut S equipped with a tungsten carbide knife (Leica, Heidelberg, Germany), nondecalcified bones were cut into 5 µm and 10 µm thick sections. Static histomorphometric data were obtained using the software Osteomeasure (Osteometrics Inc., Atlanta, GA, USA). Unstained 10 µm tissue slices were prepared for analysis of the dynamic parameters under a microscope with ultraviolet light. All biopsies were coded, and the histomorphomet-ric analysis was performed blindly by two observers (VJ and MOP). All histomorphometric data are described according to the ASBMR nomenclature [11]. The histomophometric parameters evaluated were as follows: a) Histomophometric Cortical Parameter: Both cortices were analyzed, and the results for each cortex were averaged.

Results
Twelve patients with AFF were evaluated. The clinical characteristics of patients are shown in Table 1. The mean age of the patients was 60.5 ± 12.8 years old; 91.6% were women and 83.3% Caucasian. At the time of fracture, 91.6% used BPs, with a median BP use of 6 years (range, 1-10 years). Approximately seventy-seven percent of patients were on alendronate. One patient was on denosumab but had received BP for 6 years before starting denosumab.
In all patients, the use of antiresorptive therapy was indicated due to their osteoporosis diagnosis. All AFF were associated with minimal trauma, and 75% of the patients reported thigh pain before the AFF. Rheumatic diseases were observed in 50% of the patients, most of them with rheumatoid arthritis. Approximately 42% of the patients used oral glucocorticoid, with a median prednisone dose of 5 mg/day (range, 5-10 mg/day). All patients presented with a diaphyseal fracture, of which 11 (91.6%) were complete fracture and 2 (16.6%) were bilateral. Typically, the fractures were noncomminuted and originated at the lateral cortex with a transverse orientation. *Data expressed as median (range)

Discussion
To the best of our knowledge, this is the first study to analyze the cortical bone microarchitecture and bone stiffness at peripheral sites by HR-pQCT and the iliac crest bone histomorphometry in patients with AFF. A significant deterioration of the cortical bone and of the bone strength in these AFF patients was observed. This study has several strengths, such as the evaluation of the cortical bone at peripheral sites (distal tibia and radius) using concomitantly HR-pQCT and iliac crest bone histomorphometry and the comparison with references curves of HR-pQCT and histomorphometric parameters obtained from healthy Brazilian subjects. Unlike other studies, we assessed the patients during a specific time period (6-12 months after AFF). Although AFF is clearly defined as a complication of long-term antiresorptive therapy with BPs, the median period of BPs use in this study (6 years) was lower than in other studies (8-10 years) [29,30]. This difference might be related to the greater risk of fracture of our patients, given the high prevalence of risk factors for osteoporosis and bone fragility in this patient's cohort. ters at the distal radius and tibia compared to healthy controls [29].
These differences could also be explained by differences of the populations at risk of fragility fracture. One patient had AFF in the first 3 months of use of denosumab, but she had previously received BPs for 6 years. Of note, AFF has also been reported as a complication arising from denosumab therapy, but it calls into question whether AFF cannot be derived from the previous use of BPs [32,33]. Interestingly, there was a significant positive correlation of S and F. Load with Ct.Th at the distal radius and tibia, suggesting that the cortical microarchitecture could be a fundamental parameter of bone fragility of the peripheral skeleton in AFF patients. Indeed, stiffness and cortical bone play an important role in bone strength and have been described as independent predictors of fracture [8,9]. Characteristically, the patients with AFF have cortical thickening of the femoral diaphysis and local periosteal reaction, raising the possibility that atypical fractures could occur from a cortical stress.
Conversely, this finding was not observed in the present study at the distal radius and tibia after 6 to 12 months after AFF. More than half of the individuals had lower cortical thickness by iliac crest bone histomorphometry. Regarding bone turnover, some studies have measured bone remodeling in the iliac crest or biopsies taken at the fracture site at various intervals after AFF. Consistent with most of these studies, our patients presented with low bone remodeling at the iliac crest, a finding consistent with BP treatment [1,2,34]. The mean cortical porosity was lower as measured by histomorphometry than by HR-pQCT. This result can be due to the iliac crest histomorphometric parameters may not be representative of the whole skeleton. Additionally, comparisons between HR-pQCT and histomorphometry were assessed at different anatomical regions and this difference may be secondary to specific site. Other authors have reported either nonsignificant or weak correlations when different sites were compared and is important to emphasized that HR-pQCT analyzed distal peripheral sites. Furthermore, the lack of agreement between histomorphometry and HR-pQCT for cortical porosity is likely related to the size of pores detectable by each technique: the resolution of the HR-pQCT device (82 μm) is insufficient to depict accurate parameters such as small cortical pores. Moreover, cortical thinning associated with cortical porosity in hip fractures, so-called trabecularized cortex, may partly contribute to their similar or lower cortical porosity [35]. This study has some limitations that hamper the interpretation of results, such as the limited number of patients and the cross-sectional design.

Conclusion
In conclusion, our data demonstrated that among patients with AF, rheumatic diseases and GC use were common. Cortical and strength bone parameters were deteriorated at distal tibia/radius and iliac crest, suggesting the presence of significant bone fragility after 6-12 months of AFF follow-up.

Acknowledgments
We would like to thank the contribution of the biomedical Jackeline C Alvarenga in performing the HR-pQCT exams.

Conflicts of Interest
The authors declare no conflicts of interest related to this article.