Anatomical Science International https://doi.org/10.1007/s12565-018-0450-1

ORIGINAL ARTICLE

Differences in femoral morphology among the Orientals and Caucasians: a comparative study using plain radiographs Najam Siddiqi1,2 · Antonio Valdevit1,3 · E. Y. S. Chao1,4 Received: 6 November 2017 / Accepted: 8 June 2018 © Japanese Association of Anatomists 2018

Abstract The purpose of this study was to identify the differences in femoral dimensions among Caucasian and Oriental populations. A total of 268 femora were collected from China, Japan, Korea, Taiwan and the United States. Firstly, the dimensional parameters for measuring femur were identified. These were initially measured on bone specimens to determine the methodology, followed by measuring the same parameter on plane radiographs of the same bone specimen using a board, and digitized with the aim of verifying the repeatability and reliability of the data. Data were analyzed using ANOVA, paired students t test and Pearson’s correlation analysis. The results revealed that Caucasian femora are significantly larger in maximum bone length (BL), head-neck length (HNL), lesser trochanter width and the total width of the distal epiphysis (Wdf). The Beijing femora were found to be the longest and the Japanese femora constituted the shortest bone lengths and smallest angle alpha among the Oriental populations. A strong correlation was observed between Wdf and HD, HNL, Wmc and Wlc in all the populations; however, correlation between Wdf and BL was mild. The angle alpha showed no correlation with BL. This study generated a large database of femoral geometry, which may help pharmaceutical companies to design orthopedic implants for Oriental populations. Keywords  Caucasians · Dimensional parameters · Femora · Orientals · Radiograph

Introduction It is well understood that not only genetic factors, but environmental, lifestyle and cultural factors all affect long bone development in humans (Siddiqi 2013; Rawal et al. 2012). It is therefore logical to assume that long bones in Oriental populations would differ in dimensional shapes and geometry with respect to Caucasian populations (Khang et al. 2003). Few published reports exist in the Western medical literature on the * Najam Siddiqi [email protected] 1



Biomechanics Laboratory, School of Medicine, Johns Hopkins University, 720 Rutland Avenue, Baltimore, MD 21205, USA

2



Present Address: Department of Anatomy, National University of Science and Technology, College of Medicine and Health Sciences, Al‑Tareef, PO Box 391, Zip Code: 321 Sohar, Sultanate of Oman

3

SEA Limited, 7001 Buffalo Parkway, Columbus, OH 43229, USA

4

Southern California, USA



geometric measurements, and most of these are on Caucasian femora (Husmann et al. 1997; Yoshioka et al. 1987; Rubin et al. 1989; Wright et al. 2014; Eckrich et al. 1994; Laine et al. 2000; Noble et al. 1995; Husmann et al. 1997; Pujol et al. 2014, 2016). This information, though neither extensive nor complete, has long been utilized by surgeons, design engineers and manufacturers of prosthetic implants. The complexities associated with joint motion and stability require optimal prosthesis design and size match (Lew and Lewis 1982). A need for a new hip prosthesis system to be designed for the Japanese and Korean populations has been reported because the prosthesis design based on Caucasian femoral data does not show an optimal fit (Khang et al. 2003). The more closely matched the geometry of the prosthesis components to the actual geometry of the articular surfaces, the more compatible the motion of the joint becomes with its ligament function. If a prosthesis design based on Caucasian geometry is simply scaled for use in the Oriental population, these discrepancies may be further enhanced, possibly leading to catastrophic post-operative results (Rawal et al. 2012). Because most studies have been carried out on Caucasian bone specimens, data for the Oriental population is limited. Few studies have attempted to

13

Vol.:(0123456789)



compare these groups quantitatively or compare the anatomic geometric differences between femurs from Koreans, Japanese and Americans (Khang et al. 2003; Fuku 1994; Yeung et al. 2006; Yang et al. 2014; Zhang et al. 2011). A study comparing Indian with Caucasian, Swiss and French populations revealed marked differences in dimensions of the femur in neck-shaft angle, femoral neck length, head diameter and head position (Rawal et al. 2012). Significant differences were also reported in another study on femora between South Africans of European descent and other ethnic groups in South Africa (Siddiqi 2013). In this latter study, Siddiqi reported marked differences among the Africans of European descent and native Africans with respect to bone length, bicondylar breadth of distal femur and head maximum diameter. It was also revealed that the Zulu and Ndebele tribes of South Africa have bigger femur than other tribes. Recently a three-dimensional (3D) computed tomogrophy (CT) scan using CAD software was used to reconstruct the morphology of femur for patients in order to develop a custom made prosthesis (Khang et al. 2003; Miura et al. 1998; Mahaisavariya et al. 2002). However, this is costly and time consuming. In order to establish optimal design of orthopedic implants for Oriental populations, there is a vital need to develop a database among several population groups from different ethnic origins.

Hypothesis It was hypothesized that this experiment would demonstrate that direct measurements on bones may not be necessary; instead, measurement on properly taken radiographs would be as accurate as direct measurement, thereby greatly simplifying data collection.

Objectives The first objective of this study was to develop a methodology for taking radiographs of femora that can then be used to measure bone geometry accurately and can be correlated well with direct measurements. The second objective of this study was to obtain the basic dimensional parameters of a large sample of human femora in Caucasian and Oriental populations on plane radiographs that bear strong correlation with actual measurements of the bone and can be measured reliably and consistently to determine bone geometry.

Materials and methods A total of 267 femurs was analyzed (Table 1). The femurs were grouped into either Caucasian (US) or Oriental (China, Korea, Japan, Taiwan) depending on the country in which the specimens were collected.

13

N. Siddiqi et al.

Caucasian data During the early project period, two Caucasian groups of bone specimens (69 femora) were examined: Group “CME” (C = Caucasian, M = Mayo, E = Embalmed) contained 47 femora (Right = 15, Left = 24, unknown = 8) all embalmed with no record of their source or associated demographic information; Group “CJH” (C = Caucasian, JH = Johns Hopkins) consisted of 22 cadaveric femora (side = unknown) for distal femoral dimensional analysis and the establishment of measurement techniques and key sizing parameters. In most specimens, sex was unknown. Bone specimens after contact radiograph and direct measurement, were used for biomedical testing projects or disposed of.

Oriental data A total of 198 femoral specimens was collected from four Asian countries; China, Taiwan, Korea and Japan and five cities, Beijing (27 femora, right = 17, left = 10) Shanghai (50 femora, right = 24, left = 26), Taipei (27 femora, right = 12, left = 15), Seoul (50 femora, right = 44, left = 6) and Kurume (44 femora, right = 21, left = 23). Most specimens were without sex identification. Long bones were collected from anatomy, orthopedic research laboratories and other research institutes. The majority of these specimens were without identification. The long bones of interest were not matched from the same donor. Fresh, frozen and embalmed bones were included. Specimens with prior fracture or visible deformities were excluded from measurement analysis. When bilateral specimens of a particular bone type were available, only one was selected as the representative sample. This procedure ensured true random unbiased selection. The research was approved by the Institutional Research Board.

Table 1  Bone specimen (femora) group description Group name Source

Number Ethnicity

1 2

CME CJH

47 22

Caucasian Caucasian

3 4 5 6 7

OCS OCB OTT OKS OJK

50 27 27 50 44

Asia Asia Asia Asia Asia

United States (Mayo) United States (Johns Hopkins) China (Shanghai) China (Beijing) Taiwan (Taipei) Korea (Seoul) Japan (Kurume)

Differences in femoral morphology among the Orientals and Caucasians: a comparative study…

3

1

.

.

.2

HNL 5

.4 6

BL

Wdf

Fig. 1  Radiographic parameters: maximum femoral length (BL), total width of distal epiphysis AP plane (Wdf), head diameter (HD) (1–2), head neck length (HNL) (3–4), lesser trochanter width (LTw) (5–6). Draw the anatomical and mechanical axis of the femur. Draw two lines perpendicular to the anatomical axis, one passing from the highest point of the head and the other passing from the lowest points of the two condyles. The distance between these two lines is

the BL. Draw the head neck center axis passing through the center of the head and joining points 3–4, which gives the HNL. HD is measured by joining points 1 and 2 and passing through the center of the head. LTw is measured by joining points 5 and 6. Wdf is measured by drawing two tangent lines at the medial and lateral margins of the two condyles parallel to the anatomical axis. The distance in between these two lines is Wdf

Definition of key measurements

5. Lesser trochanter width (LTw): draw a line perpendicular to the anatomical axis of the proximal femur from the most medial point 5 on the lesser trochanter to the lateral cortex of the shaft of the proximal femur at point 6. The distance between point 5 and 6 is the LTw (Fig. 1). 6. Total width of distal epiphysis AP plane (Wdf): draw two tangent lines to the most medial and most lateral points of the medial and lateral condyles. The distance between these points is Wdf (Fig. 1). 7. Total width of medial condyle in sagittal plane (Wmc): distance between the tangent line to the most anterior and posterior points of the anterior and posterior curvatures of the medial condyle (Figs. 2, 3). These tangent lines should be parallel with the anatomical axis of the distal femur in sagittal plane. 8. Total width of lateral condyle in sagittal plane (Wlc): distance between the tangent lines to the most anterior and posterior points of the anterior and posterior curvatures of the lateral condyle (Figs. 2, 3). These tangent lines should be parallel with the anatomical axis of the distal femur in sagittal plane.

1. Maximum femoral length (BL): draw the anatomical axis and mechanical axis of the femur. Draw a line connecting the lowest points on the femoral condyles and perpendicular to the anatomical axis. Draw another line passing from the highest point of the femoral head and perpendicular to the anatomical axis. The distance between these two lines is the length of the femur (BL) (Fig. 1). 2. Head diameter (HD): draw a line joining two points 1 and 2 crossing the mechanical axis at the center of the head (Fig. 1). 3. Head neck length (HNL): draw a line connecting points 3 and 4 and passing through the head center and neck center. The distance between the outermost point 3 on the head circumference and point 4 at lateral border of the greater trochanter is the head neck length (Fig. 1). 4. Neck shaft angle (α): the angle between the anatomical axis of the femur and anatomical axis of the neck (line between points 3 and 4).

13



N. Siddiqi et al.

Fig. 2  a Direct measurement: distal end of the femur (Wdf), b width of distal femoral condyle in sagittal plan (Wlc, Wmc)

Step 1: Comparison of direct measurements on the bone with the radiographic measurements The first objective of this study was to develop a method of taking bone radiographs so that measuring the dimensions on radiographs correlate well with the dimensions on the actual bone. For the purpose of ensuring that all the measurements made on long bone radiographs were accurate and consistent, 22 distal femurs (Group CJH) harvested from cadavers were utilized. Contact AP and ML radiographs of the distal femora were obtained (Fig. 3). Contact radiographs (bones lying directly on the cassette with X-ray beam source six feet away to minimize image distortion and magnification) and the beam focusing in the center of the two condyles. Modified ruler and caliper were manufactured in order to obtain consistent physical measurements from the bone specimens (Fig. 2). A digitizer was used to measure distances between two points on radiographs. Student’s t test and Pearson’s linear regression analyses were used to compare the corresponding measurements. The following three parameters were used to compare the two measuring techniques:

Fig. 3  Radiographic parameters: lateral condyle sagittal width (Wlc) and medial condyle sagittal width (Wmc). Draw two tangent lines to the most anterior and posterior points of the anterior and posterior curvatures of the lateral medial condyles. The distance between these two lines will give the sagittal width of the condyle

13

1. Total width of distal epiphysis in AP plane (Wdf): measuring the distance between the two condyles in the anterio-posterior plane as shown in the Fig. 2a. 2. Total width of lateral (Wlc) (Figs. 2b, 3). 3. Total width of lateral condyle (Wlc): in the sagittal plan: measuring the sagittal width of the lateral and medial condyles (Fig. 2b).

Step 2: Radiographs of femora and measurement on radiographs using a digitizer Data collection methods We developed a bone-holding jig in order to ensure consistent radiographic quality and reproducibility and avoidance of rotation of the femur (Eckrich et al. 1994) (Fig. 4). The bone can be fixed by two clamps to hold the bone parallel to the X-ray plate placed under the bone. Rotation of the bone can be controlled by rotation of the clamp and locking it in 0° position. A radio-opaque ruler was used to determine the magnification. For each bone three radiographs were taken. Two radiographs were taken in lateral and one in AP plane; the X-ray beam was focused at the middle of the shaft of the femur.

Differences in femoral morphology among the Orientals and Caucasians: a comparative study… Fig. 4  Specially designed jig to hold the bone in neutral position thus preventing rotation during radiography

All data were measured and collected using a background light digitizer table. This is a device consisting of an X-ray table with an integrated electrostatic micro grid, which facilitates accurate point position (0.01 mm accuracy) through a movable mental crosshair. A software program was written for digital data collection through the serial port of the digitizer. A series of subprograms were used to compute the X-ray magnification factor, prompt the use for data digitization, and format the data for transfer to a database software package. At the beginning of digitizing of each X-ray, the distance between the digitizer grid and the X-ray was calibrated using a radio-opaque scale on the X-ray film. Although the distance between the specimens, the radiation source, and the cassette were maintained constant, the exact magnification was computed. The program calculates distances through two selected endpoints on the digitizer. All the measurements were repeated three times with the mean and standard deviation reported. Percentage errors between the mean and standard deviation exceeding 5% required the user to re-digitize the parameter. Typical distance and angle measurements displayed percentage errors of 1% and 3%, respectively.

Repeatability and reliability test The purpose of conducting this experiment was to verify the consistency and accuracy of measuring bone dimension and geometry directly from the specimen and from the radiographs, and to assure repeatability and reliability of the data. The data were analyzed using ANOVA, paired students t test, and Pearson’s product–moment correlation analysis.

Results Step 1: Comparison of direct measurements with radiographs Distal femur specimens (n = 22) were used to compare direct versus radiographic measurements on the three linear dimensional parameters. AP measurements were closely correlated with high r2 value, and were significant (P = 0.001),however, there was no significant difference between the direct and radiographic measurements (P = 0.22) (Table 2).

Step 2: Measurements on radiographs

Parameter

r value

P value

Data of all the parameters measured on radiographs are given in Table 4. Of the five measured parameters, three were found to correlate significantly. Wdf has strong cross correlation with Wlc (r = 0.837, P < 0.001) and Wmc (r = 0.864, P < 0.001), based on directly measured values (Table 3). Similarly, a strong correlation was found between the direct and radiographic measurement (Table 2). Data from AP radiographs of normal subjects showed similar trends.

Wdf Wmc Wlc

0.94 0.86 0.96

0.001 0.001 0.001

Comparative results between Caucasian and Asian populations

Table 2  Correlation of direct vs radiographic measurements of total width of distal epiphysis parameters in bone group “CJH”. Direct measurements were strongly correlated with the radiographic measurementstotal width of distal femoral condyle in AP and ML views (n = 22)

1 2 3

Direct measurements were strongly correlated with the radiographic measurements of total width of distal epiphysis in AP and ML views (n = 22)

Next, 47 Caucasian and 198 Asian embalmed femora were collected; the overall bone dimension parameters were

13

N. Siddiqi et al.

Strong correlation was also found between Wdf and HD, HNL, Wmc and Wlc in all the populations studied. However, correlation of bone length with Wdf, HD, LTw, HNL, Wmc and Wlc was mild. Alpha angle has no correlation with bone length (Table 5).

Discussion The main focus of this study was to establish a database for femur in Oriental populations and to study in detail the geometric differences between Caucasians and Orientals. We developed a method of plain radiography to collect data for a large number of bone specimens, which not only reduces cost but is feasible. The availability of such databases will provide the essential parameters and size ranges for implant design for Oriental patients. The database will also provide a reference for anatomical information associated with

13

399.57 ± 18.6 44.28 ± 2.79 88.85 ± 5.89 131.71 ± 5.60 42.78 ± 3.27 74.47 ± 3.99 56.54 ± 3.81 55.92 ± 3.55 396.64 ± 21.9 45.02 ± 3.91 90.93 ± 7.19 130.90 ± 5.31 42.35 ± 3.19 75.60 ± 5.79 59.72 ± 4.48 55.99 ± 4.32 382.95 ± 23.9 44.03 ± 3.56 90.48 ± 7.40 127.75 ± 4.74 40.92 ± 3.49 75.10 ± 5.41 57.44 ± 4.61 58.26 ± 2.66 397.73 ± 23.7 44.42 ± 3.71 92.52 ± 6.87 134.12 ± 7.29 43.04 ± 4.04 75.07 ± 5.71 59.00 ± 4.49 57.14 ± 4.25 408.81 ± 20.2 45.95 ± 3.4 92.88 ± 3.4 130.29 ± 6.6 42.67 ± 3.4 76.92 ± 5.1 60.63 ± 4.0 58.8 ± 4.1

Statistical correlation analysis

452.44 ± 26.1 DNA 102.11 ± 8.08 DNA 48.47 ± 5.15 84.24 ± 7.25 DNA DNA

Bone length (BL): the Beijing population possessed the longest and Japanese specimens the shortest (P < 0.05), while bone length was similar among Shanghai, Seoul and Taipei (Fig. 5). No significant difference was found in HD and Wdf among the Oriental populations (Figs. 6, 7). Head neck length: Taiwanese bones displayed the smallest head neck length (Fig. 8). Alpha angle: Japanese bone showed smallest angle (127°) while the other populations revealed an angle above 130° (Fig. 10). HD, Wdf, LTw, Wlc and Wmc: No significant difference was observed among the populations (Figs.  6, 7, 9). Wlc was larger than Wmc in all populations except Japanese.

BL HD HNL α LTw Wdf Wlc Wmc

Comparison among different Asian populations

OTT

presented with as mean and standard deviation. The results are given in Figs. 5–10 and Table 4). BL, HNL, LTw and Wdf were significantly larger (P < 0.05) in Caucasian populations with respect to Asian populations Measurements for HD, Wmc, Wlc and α angle were not available for the Caucasian population.

OKS

0.001 0.001

OJK

0.864 0.837

OCS

Wmc Wlc

OCB

P value

CME

r value

Parameters

1 2

Parameter

Table 4  Measurements results of Caucasian group (CME) and the Oriental groups Chinese–Beijing (OCB) and Shanghai (OCS), Japanese (OJK), Korea (OKS) and Taiwan (OTT)

Table 3  Correlation of width of distal femur (Wdf) with other distal condyle parameters using radiographs in bone group “CJH” (n = 22)

BL Maximum femoral length, HD head diameter, HNL head neck length, α neck shaft angle, LTw lesser trochanter width, Wdf total width of distal epiphysis AP plane, Wlc total width of lateral condyle in sagittal plane, Wmc total width of medial condyle in sagittal plan



Differences in femoral morphology among the Orientals and Caucasians: a comparative study… 480

100

Femoral Bone length (FBL)

440 400

CME OCS

320

OJK

280

OKS

200

50

OJK OKS OTT

50 40

Groups

Fig. 7  Caucasians (CME) showing significant different (P = 0.05) in Wdf with all the Oriental populations

Femoral head (HD)

120 110

45

Head neck length (HNL)

100

40

OCB OCS

mm 35 30

CME OCB

90 mm

OCS

OJK

80

OJK

OKS

70

OKS

OTT

25 20

OACS

60

Groups

Fig. 5  Femoral bone length (BL) of Caucasians (CME) showing significant difference (P = 0.05) with all the Oriental groups

OCB

mm 70

OTT

240

CME

80

OCB

360 mm

Distal width femur (Wdf)

90

50 Groups

Fig. 6  No significant difference was seen in femoral head (HD) among the Oriental populations. Caucasian data was not available

epidemiological studies involving musculoskeletal joint problems among Oriental populations, and help forensic experts and anthropologists. Furthermore, the present data can also be used as a guide for optimal fit allograft size matching in Asian populations. The results from this study revealed significant differences among the Caucasian and Oriental databases. According to the femoral data, with the exception of alpha angle, Wdf displayed a strong correlation with all the parameters measured. The investigators believe that Wdf is the key parameter to be considered for representative knee sizing. This finding is in keeping with a published report (Mensch and Amstutz 1975) that also proposed “Width of distal femur” as the key radiographic parameter that can aid the surgeon in obtaining more precise preoperative evaluation for selection of the correct prosthesis. Analysis of femoral condyle contours is crucial, since these affect ligament’s tensile balance and normal knee kinematics. Other studies (Lew and Lewis 1982; Goodfellow and O’Connor 1992) determined that, in order to maintain

OTT

60

Groups

Fig. 8  Caucasians (CME) showing significant different (P = 0.05) in HNL with all the Oriental populations

the normal length patterns in cruciate ligaments and avoid ligament tightening or laxity, less transforming prosthesis should be utilized. Data from this study revealed that femoral length in the United States was significantly larger than in all the Oriental populations. Similar results were published (Fuku 1994) comparing Japanese and Korean femoral length, and found Japanese to be smaller than Koreans; this study revealed a similar result. These Japanese bones were probably from the victims of the atomic bomb in World War II; at that time Japanese height was much smaller than other Oriental populations; however, with changes in culture and eating habits, Japanese are now taller than in the past. There was no significant different in femoral head diameter between all Oriental populations. Another study (Hoaglund and Djin 1980) compared Caucasian and Chinese femoral parameters using cadaveric bones; femoral length for the Caucasian femora was 451 mm compared to 452 mm in this study. Furthermore, head diameter and alpha angle in the Chinese populations

13



N. Siddiqi et al. 60

Lesser trochanter width (LTw)

50 CME

40

OCB OCS

mm 30

OJK

20

OKS OTT

10 0

Groups

Fig. 9  Caucasians (CME) showing significant different (P = 0.05) in LTw with all Oriental populations 150

Angle alpha (α)

140 130 Angle (α)

OCB

120

OCS OJK

110

OKS

100

OTT

90 80

Groups

Fig. 10  Significant difference (P = 0.05) in alpha angle was observed between OCS and OJK within the Oriental populations. Caucasian data not available

was again similar in both studies. Similarly, published Korean data (Khang et al. 2003) also showed similar values when compared with the Korean data in this study. It is important to realize that the current results were established using random Oriental and Caucasian bone specimens. Age, sex and chronological life span can have significant effect on data distribution due to drastic changes in nutrition and social habits in the last five decades (Pujol et al. 2014, 2016). Hence, the established correlative relationships may not exist in other populations, Table 5  Femoral correlation chart showing correlation coefficient (r value) of maximum bone length (BL) versus all the other parameters measured in different populations studies

13

Mayo Beijing Shanghai Japan Korea Taiwan

especially under different cultural environmental and ethnic backgrounds. Therefore, similar data measurements and correlative studies must be repeated for other identifiable populations. As with all data entry processes, bone grouping and analysis techniques have also been well established and can be applied to other long bone studies. Attempting to generate a large and wide database is a difficult and time-consuming undertaking, both in preparation and execution. Most investigators underestimate the effort and time required. Thus, limited published works in the literature only cite small sample sizes of bones and therefore are not able to justify their applicability. This study is not devoid of criticism or deficiencies. With more precisely defined aims and well established methodologies, additional databases specifically involving living subjects may be established. Unfortunately, age and gender were unknown for most of these bones and thus these parameters were not considered in this study. These variations require more in depth studies in which age-matched groups are employed to examine the effects of age as related to shaft geometry. From the view point of basic science, long bone development based on both genetic and cultural variations can be studied properly when databases involving different populations are established. All of these applications will require a more extensive investigation with more specifically defined goals and a well-established database from which to launch the study. It is anticipated that this study has helped to establish a standardized and practical methodology to explore additional databases in different ethnic populations including Caucasians.

Conclusions The present study has accomplished several important objectives. First, it identifies the key parameters for femur measurements: BL, HD, HNL, α, LTw, Wdf, Wmc and Wlc. These parameters, though not exhaustive, are relevant and, most importantly, measurable.

Wdf

HD

LTw

HNL

Alpha

Wmc

Wlc

0.75 0.538 0.732 0.757 0.591 0.733

DNA 0.682 0.713 0.783 0.570 0.665

DNA 0.364 0.686 0.764 0.505 0.711

0.594 0.801 0.793 0.773 0.711 0.757

DNA 0.051 0.178 0.157 0.060 0.106

DNA 0.493 0.638 0.761 0.563 0.466

DNA 0.386 0.665 0.793 0.673 0.553

Differences in femoral morphology among the Orientals and Caucasians: a comparative study…

Secondly, this study has developed a method to correctly obtain radiographs without rotation to measure different parameters reliably. Thirdly, a large database was established for Oriental populations from four different countries, and a comparison was initiated with respect to the Caucasian database. To our knowledge, this is the first and most expansive database of its kind for Asian populations. This rare resource will aid in validating that significant dimensional differences exist between population groups as well as within groups. Fourthly, this data clearly revealed that linear measurements of the distal femur are strongly correlated. Therefore not all linear measurements are required in order to characterize the morphology of normal long bones. Finally, this study demonstrated significant differences between the Caucasian and Oriental femoral bone geometry, thus emphasizing the fact that prostheses for Oriental populations should be developed based on Oriental bone dimensions for optimal fit and to reduce size mismatch. Acknowledgments  We are indebted to the following collaborators without them this research would never have been completed. Prof. Akio Inoue from department of Orthopedic Surgery, Kurume University School of Medicine, Japan, Prof. T.K. Lui, Department of Orthopedic Surgery, National Taiwan University, Taiwan, Prof. Young-Min Kim, and Prof. Myung-Sang Moon, Department of Orthopedic Surgery, Seoul National University, and Catholic University Medical College, Korea, Prof. Xian-Zheng Luo, Department of Orthopedic Surgery, Beijing Friendship Hospital, China, Prof. Kerong Dai, Department of Orthopedic Surgery, Shanghai Second Medical University, China. We are also thankful to Dr. Zhenyu Wang and Dr. Ai Guo from China, Dr. Jinn Lin and Gau-Tan Lin from Taiwan, Dr. Young-Koo Kang from Korea and Dr. Naoto Shiba and Dr. Kenichiro Miyazaki from Japan. We are also grateful for Ilka Lorenzen-Schmidt, Stephen Kraker and Veronika Bonin from Germany who were summer students and devoted their time to this project. Funding  This project was funded by grant from DePuy Company, USA.

Compliance with ethical standards  Conflict of interest  The authors declare that they have no conflict of interest.

References Eckrich SG, Noble PC, Tullos HS (1994) Effect of rotation on the radiographic appearance of the femoral canal. J Arthroplasty 9:419–426 Fuku H (1994) Cementless hip prosthesis design: a basic study and analysis of the proximal femur in normal Japanese people. Nihon Seikeigeka Gakkai Zasshi 68:763–773

Goodfellow J, O’Connor J (1992) The anterior cruciate ligament in knee arthroplasty. A risk-factor with unconstrained meniscal prostheses. J Clin Orthop Relat Res 276:245–252 Hoaglund FT, Djin Low Weng (1980) Anatomy of the femoral neck and head with comparative data from Caucasians and Hong Kong Chinese. Clin Orthop Relat Res 152:10–16 Husmann O, Rubin PJ, Leyvraz PF, De Roguin B, Argenson JN (1997) Three-dimensional morphology of the proximal femur. J Arthroplasty 12:444–450 Khang G, Choi K, Kim CS, Yang JS, Bae TS (2003) A study of Korean femoral geometry. Clin Orthop Relat Res 406:116–122 Laine HJ, Lehto MU, Moilanen T (2000) Diversity of proximal femoral medullary canal. J Arthroplasty 15:86–92 Lew WD, Lewis JL (1982) The effect of knee-prosthesis geometry on cruciate ligament mechanics during flexion. J Bone Joint Surg Am 64:734–739 Mahaisavariya B, Sitthiseripratip K, Tongdee T, Bohez EL, Vander Stolen J, Oris P (2002) Morphological study of the proximal femur: a new method of geometrical assessment using 3-dimensional reverse engineering. Med Eng Phys 24:617–622 Mensch JS, Amstutz HC (1975) Knee morphology as a guide to knee replacement. Clin Orthop 112:231–241 Miura T, Matsumoto T, Nishino M, Kaneuji ATK (1998) A new technique for morphologic measurement of the femur. Its application for Japanese patients with osteoarthrosis of the hip. Bull Hosp Jt Dis 57:202–207 Noble PC, Box GG, Kamaric E, Fink MJ, Alexander JW, Tullos HS (1995) The effect of aging on the shape of the proximal femur. Clin Orthop Relat Res http://www.ncbi.nlm.nih.gov/pubme​ d/76347​21 316:31–44 Pujol A, Rissech C, Ventura J, Badosa J, Turbón D (2014) Ontogeny of the female femur. Geometric morphometric analysis applied on current living individuals of a Spanish population. J Anatom 225:346–357 Pujol A, Rissech C, Ventura J, Turbón D (2016) Ontogeny of the male femur: geometric morphometric analysis applied to a contemporary spanish population. Am J Phys Anthropol 159:146–163 Rawal BR, Ribeiro R, Malhotra R, Bhatnagar N (2012) Anthropometric measurements to design best-fit femoral stem for the Indian population. Indian J Orthop 46:46–53 Rubin PJ, Leyvraz PF, Heegaard JH (1989) Radiologic changes of anatomic parameters of the proximal femur as a function of its position in rotation. Rev Chir Orthop Reparatrice Appar Mot 75:209–215 Siddiqi N (2013) Comparison of osteometric femoral dimensions among the South Africans of different ethnic groups and South African whites. Egypt J Forensic Sci 3:8–14 Wright SJ, Boymans TA, Grimm B, Miles AW, Kessler O (2014) Strong correlation between the morphology of the proximal femur and the geometry of the distal femoral trochlea. 22:2900–2910 Yang Z, Jian W, Li ZH, Jun X, Liang Z, Ge Y, Shi ZJ (2014) The geometry of the bone structure associated with total hip arthroplasty PLoS One 9(3):e91058 Yeung Y, Chiu KY, Yau WP, Tang WM, Cheung WY, Ng TP (2006) Assessment of the proximal femoral morphology using plain radiograph—can it predict the bone quality? J Arthroplasty 21:508–513 Yoshioka Y, Sui D, Cooke TD (1987) The anatomy and functional axes of the femur. J Bone Joint Surg Am 69:873–880 Zhang H, Hu YQ, Zhang ZL (2011) Age trends for hip geometry in Chinese men and women and the association with femoral neck fracture. 22:2513–2522

13