Correlation between Femoral Head Lateralization and Bone Morphology in Primary Hip Osteoarthritis

Background Osteoarthritis (OA) is the most common disease of the hip in adults, and its etiology is divided into two groups: primary and secondary. Although acetabular dysplasia is the most frequent reason for total hip arthroplasty (THA) in Japan, primary OA has increased recently. Although there are two types of femoral head migration in primary OA: superior and medial, there are some patients with prominent femoral head lateralization. This study aimed at evaluating the relationship between femoral head lateralization and bone morphology of the acetabulum and proximal femur using radiographic factors in primary OA of the hip. Methods A retrospective study was conducted between 2008 and 2017 to assess 1308 hips with OA who underwent primary THAs at our institute. The diagnostic criteria for primary OA were Crowe type 1, Sharp's angle <45°, and center-edge (CE) angle >25°. We classified patients with primary OA into two groups based on femoral head lateralization: group L with lateralization or group N without. Radiographic factors included Sharp's angle, CE angle, acetabular inclination, acetabular depth ratio (ADR), acetabular head index (AHI), and femoral neck-shaft angle (FNA), all examined on an anteroposterior pelvic radiograph. Femoral neck anteversion was calculated using computerized axial tomography. Results Primary OA was diagnosed in 210/1308 hips (16.1%) (group L: 112 hips (8.6%); group N: 98 (7.5%)). Patient demographics were not significantly different. Radiographic factors with observed significant differences between group L and group N were the average CE angle (33.0° vs. 35.1°, respectively, p = 0.009), ADR (251.6 vs. 273.4, p < 0.001), AHI (77.2 vs. 80.4, p < 0.001), and FNA (136.9° vs. 134.8°, p = 0.012). Conclusions This investigation suggests that primary OA with femoral head lateralization demonstrated specific identifiable radiographic characteristics in the acetabulum and proximal femur that might contribute to hip joint instability such as the dysplastic hip.


Background
Osteoarthritis (OA) is the most common disease of the hip in adults. Its etiology is commonly divided into two groups: (1) primary or idiopathic OA, in which the underlying cause cannot be defned, and (2) secondary OA in which the predisposing cause is well defned [1]. Primary OA of the hip is a frequent reason for total hip arthroplasty (THA) in Western countries [2,3], whereas secondary OA of the hip due to acetabular dysplasia (AD) is a more typical reason for THA in Japan [4]. Several reports in the literature indicate that the prevalence of secondary OA due to AD in Asians is high [5][6][7], but primary OA of the hip has increased in Japan recently [4,8]. Although primary OA is generally idiopathic, several reports reveal that femoroacetabular impingement (FAI) has been increasing and appears to be a potential precursor of idiopathic hip OA [9][10][11]. Tere is a perception that bone morphology is associated with primary OA, and several authors have reported that two types of femoral head migration can occur: (1) superior (or eccentric) and (2) medial (or concentric) [1,[12][13][14]. Tere are some patients with prominent femoral head lateralization but the correlation between femoral head lateralization and bone morphology of the acetabulum and femur has not been sufciently evaluated in the past.
Tis study aimed at evaluating the relationship between femoral head lateralization and bone morphology of the acetabulum and proximal femur using radiographic factors in primary OA of the hip for the population undergoing THA.

Materials and Methods
A retrospective radiographic evaluation was performed at our institute (Matsudo City General Hospital). Te research protocol for this study was approved by the Institutional Review Board of the authors' afliated institutions in compliance with the principles of the Helsinki. Written informed consent was obtained from all participating subjects. A total of 1308 consecutive patients who underwent primary THA for hip OA from January 2008 through December 2017 were included in the study. Te diagnosis of OA was determined using the criteria of Altman et al. which is hip pain, along with two of the following: (1) erythrocyte sedimentation rate (ESR) <20 mm, (2) femoral or acetabular osteophytes, and (3) joint space narrowing [15]. Two examiners (I.S. and S.C.) determined the diagnosis after discussion.
Primary OA was defned by commonly used diagnostic criteria as follows: Crowe's classifcation type 1 [16], Sharp's angle <45° [17], and a CE angle >25° [18]. Computed tomography (CT) was used in all cases for preoperative planning. Te subjects were divided into two groups based on the lateralization of the femoral head. Te lateralization group (group L) included patients in whom the distance between the ilioischial line and the medial aspect of the femoral head was ≥10 mm. In the group without lateralization (group N), the distance was <10 mm (Figure 1) [19].
Cases of secondary OA were excluded, and the following items were examined in cases classifed as primary OA: age, sex, afected side, height, body weight, and body mass index (BMI). Sharp's angle [17], CE angle [18], acetabular inclination (AI) [20], acetabular depth ratio (ADR) [21], acetabular head index (AHI) [22], and femoral shaft-neck angle (FNA) were measured from an anteroposterior (AP) pelvic radiograph (Figures 2 and 3), and femoral neck anteversion (FNAV) was measured from an axial plane CT ( Figure 4) [23,24]. Te measurement methods for each radiographic factor were as follows: Sharp's angle was the angle between the line joining the lateral aspect of the weight-bearing zone and the inferior point of the teardrop, parallel to the transverse axis of the pelvis. Te CE angle was the angle between the line joining the lateral aspect of the weight-bearing zone and the center of the femoral head with the line perpendicular to the transverse axis of the pelvis. AI was the angle between the line joining the medial and lateral aspects of the weight-bearing zone and the line parallel to the transverse axis of the pelvis. ADR was calculated by dividing : a (mm) is the distance between the ilioischial line and the medial aspect of the femoral head. When the distance was ≥10 mm, the hip was categorized as group L, and when it was <10 mm it was categorized as group N. In this case, a is 12.2 mm, so this hip is categorized as group L. Sharp's angle was the angle between the line joining the lateral aspect of the weight-bearing zone and the inferior point of the teardrop, parallel to the transverse axis of the pelvis. Te CE angle was the angle between the line joining the lateral aspect of the weight-bearing zone and the center of the femoral head with the line perpendicular to the transverse axis of the pelvis. FNA was the angle between the axis of the femoral neck and femoral shaft. ADR was calculated by dividing the depth of the acetabulum by the length between the inferior teardrop point and the lateral weightbearing zone of the center of the femoral head in the coronal plane, then multiplying by 1000. the depth of the acetabulum by the length between the inferior teardrop point and the lateral weight-bearing zone of the center of the femoral head in the coronal plane, then multiplying by 1000. AHI was calculated by dividing the distance from the medial margin of the femoral head to the lateral side of the weight-bearing zone by the width of the femoral head, then multiplying by 100. FNA was the angle between the axis of the femoral neck and the femoral shaft. FNAV was the angle between the femoral neck axis and the posterior condylar line of the distal femur. An AP pelvic radiograph was performed with the patient supine and their lower extremities internally rotated by approximately 15°so that the patella could be positioned in the frontal plane to maximize the length of the femoral neck. CT was performed with the patient in a supine position with the lower extremities oriented in natural rotation. Two board-certifed orthopedic surgeons specializing in the hip joint evaluated all radiographic factors and used the mean value.

Statistical
Analysis. An independent-sample Student's ttest and Fisher's exact test were used to compare groups L and N in terms of patient demographics and radiographic factors. We compared FNA, ADR, and FNAV between the afected and unafected sides of unilateral OA using a paired t-test. Te interclass correlation coefcient (ICC) of each radiographic factor and its 95% confdential interval (CI) with an absolute agreement defnition were calculated. p < 0.05 was considered signifcant in all tests of statistical inference. All statistical analyses were performed using EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan), which is a graphical user interface for R (Te R Foundation for Statistical Computing, Vienna, Austria).

Results
A total of 210 hips (16.1%) were diagnosed with primary OA out of 1308 hips. Among these, 112 (53.3%) hips were in group L and 98 (46.7%) in group N. Tere was no signifcant diference in demographics between the two groups (Table 1). Radiographic factors are summarized in Table 2. Te average values of CE angle (33.0°vs. 35.1°, p � 0.009), ADR (251.6 vs. 273.4, p < 0.001), and AHI (77.2 vs. 80.4, p < 0.001) in group L were signifcantly lower compared to group N. On the other hand, the average value of FNA in group L was signifcantly larger than in group N (136.9°vs. 134.8°, p � 0.012). Tere were no signifcant differences in other radiographic factors including Sharp's angle, AI and FNAV. Bilateral primary OA afected 106 hips, whereas 104 hips had unilateral disease. With unilateral OA, 63 patients (60.6%) were in group L and 41 (39.4%) were in group N. Te results of radiographic factors between the afected and unafected sides in unilateral OA patients are shown in Table 3. Te average ADR of the afected side was signifcantly less than that of the unafected side in both groups (group L: 251.6 vs. 286.2, group N: 271.3 vs. 297.5) (p < 0.001). Te FNA of the afected side in group L was larger than that of the unafected side, however, it was not signifcantly diferent (p � 0.072). FNAV was not signifcantly diferent between the afected and unafected sides in either group.
Te ICC of each radiographic factor was as follows:

Discussion
In this study, the prevalence of primary OA of the hip was 16.1% based on the diagnostic criteria of Crowe's classifcation type 1 [16], Sharp's angle <45° [17], and a CE angle >25° [18]. Nakamura et al. reported in 1989 that primary OA was detected in 13 cases (0.9%) out of 2,000 consecutive cases diagnosed with hip OA [24]. Te diagnostic criteria included the absence of femoral head deformity, a CE angle >19°, a Sharp's angle <45°, and an acetabular roof obliquity <15° [25]. Tis low prevalence could be infuenced by the patient population, which was collected from a specifc outpatient clinic for the treatment of acetabular dysplasia. Hoagland et al. evaluated 200 consecutive Japanese patients in Japan and 199 consecutive white American patients in the USA, all of whom were admitted for hip surgery [26]. Tey reported in 1985 that the prevalence of primary OA was 18% in Japanese patients and 90% in white American patients [26]. Recently, primary OA of the hip was reported to be increasing in Japan [8]. Te Japanese Arthroplasty Register reported that the percentage of primary OA patients referred for primary total hip arthroplasty was 16.3% in 2013, 21.5% in 2015, and 26.6% in 2017 [8].
In 2010, Jingushi et al. conducted a multi-institutional examination of patients with hip OA who were newly admitted to the orthopedic outpatient clinic in Japan. Tey reported that the prevalence of primary OA was 9% (44 out of 485 hips) [4]. Tus, the prevalence of primary OA of the hip has been reported variously in Japan. Te prevalence of primary OA might difer depending on the diferences in diagnostic criteria and patient populations. In this study, the prevalence was 16.1% but if the hips in group L were excluded from the diagnosis of primary OA, the prevalence dropped to 7.5%, similar to the data reported by Jingushi et al. [4].
Femoral head lateralization in our study was recognized in approximately half of all primary OA cases (53.3%). Hartoflakidis and Karachalios reported that 80% (218/272) of primary hip OA patients had eccentric type (or superior migration) [1], which included superolateral or superomedial migration. Nakamura et al. reported that eight hips (62%) demonstrated superolateral types [14]. In our study, femoral head lateralization was defned as the distance between the ilioischial line and the medial aspect of the femoral head ≥10 mm. Te prevalence of femoral head lateralization in primary OA in our study was similar to the result reported by Nakamura et al. [14] and less than Hartoflakidis et al., which may suggest the possibility that a racial diference might infuence the prevalence of femoral head lateralization.
Te etiology of femoral head lateralization in patients with primary OA is uncertain. Nakamura et al. reported that the superolateral type of primary OA develops in the subset of normal hips with a greater degree of acetabular roof obliquity [14]. Our radiographic evaluation demonstrated that the CE angle, ADR, and AHI of group L were significantly less than those of group N, and the FNA of group L was signifcantly larger than that of group N. Furthermore, the radiographic evaluation of the patients with unilateral primary OA demonstrated that the FNA of the afected side was slightly larger than that of the unafected side, and the ADR of the afected side was signifcantly less than that of the unafected side. Te CE angle and AHI are reduced by femoral head lateralization. Regarding ADR, dysplastic hips usually have a smaller ADR than normal hips, and femoral head lateralization often occurs in dysplastic hips [27,28]. Tus, primary OA with femoral head lateralization might be a boundary condition between primary and dysplastic OA, and the diagnostic criteria for primary OA using only the Sharp and CE angles might miss the condition. Regarding FNA, Pauwels and Maquet reported that larger femoral neck-shaft angles might induce a laterally directed joint reaction force to potentiate hip instability [27,29]. Terefore, a larger FNA might have the potential role of femoral head lateralization in primary OA. Eventually, the smaller ADR (shallow acetabulum) and larger FNA (coxa valgus) might induce hip joint instability and correlate with the development of OA associated with femoral head lateralization. In clinical practice, cup position and stem   Te efect of femoral head lateralization on the clinical course in the primary OA is unclear in this cross-sectional study. In dysplastic hips, femoral head lateralization correlates strongly with the development of hip OA [1,28]. Mimura et al. propose that femoral head lateralization induces greater hip joint pressure to maintain stabilization of the joint [28]. Hartoflakidis and Karachalios reported that hips with concentric idiopathic OA underwent THA on average 10 years after symptom onset; by contrast, eccentric hips underwent THA on average four years after symptom onset [1]. Tus, femoral head lateralization might accelerate the osteoarthritic change of the hip and advance the timing of THA. Terefore, femoral head lateralization might strongly correlate with the instability of the hip and lead to the development and progression of OA.
Our study has several limitations. First, the radiographic review was based only on AP radiographs. However, we consider that it is essential for physicians to acquire common and reliable radiographic views as well as parameters for plain radiographic assessment that can serve as a foundation for accurate diagnosis, disease classifcation, and surgical decision-making. Second, osteophytes of the proximal femur and acetabulum are variable, such that measurement errors in radiographic factors could occur. However, the data were reviewed by two experienced orthopedic surgeons, and measurements for cases with complex imaging fndings were made following discussion. Furthermore, most of the ICCs of each radiographic factor were good. Tird, the lateralization of the femoral head was determined by the distance between the ilioischial line and the medial aspect of the femoral head. Te distance of 10 mm should be considered a general reference number as opposed to a strict parameter, as magnifcation errors and variability in patient size can infuence this measurement. Fourth, FNA might be infuenced by the rotation of the lower extremity [30]. However, FNA was measured using AP pelvic radiographs performed with the lower limb in internal rotation, thus placing the patella in the frontal plane. Furthermore, FNAV measured by CT in both groups was approximately 15°, and there was no signifcant diference between the two groups. Tese relatively small values of FNAV were considered to be negligible in terms of the measurement of FNA. Fifth, the stages of OA varied, and we did not distinguish among them. Tere is a possibility that femoral head lateralization can progress over time in accordance with the stage of OA. Sixth, the diagnostic criteria for primary OA did not include AI or AHI in this study. However, the diagnosis of primary OA was made using commonly used diagnostic criteria: Sharp's angle and CE angle.

Conclusions
Our investigation suggests that primary OA with femoral head lateralization demonstrated specifc radiographic characteristics in the acetabulum (signifcantly smaller ADR, CE angle, and AHI) and proximal femur (signifcantly larger FNA), which might contribute to some hip joint instability such as the dysplastic hip.

Data Availability
Te datasets used and/or analyzed during the current study are available from the corresponding author upon request.

Ethical Approval
Te research protocol for this study was approved by the Institutional Review Board of authors' afliated institutions in compliance with the principles of the Helsinki Declaration (Number: R5-3).

Consent
Written informed consent was obtained from all participating subjects.

Conflicts of Interest
Te authors declare that they have no conficts of interest.