Bone morphometry has been performed in prior studies. We attempted to comprehensively interpret the measurements from such studies to obtain basic data for designing TKA, but we encountered difficulty because of obscurities and variations in measuring methods. Comparisons of parameters obtained using different methods may produce apparent racial differences.
In the present study, bone resection face shapes likely to occur in TKA were measured to enable design of artificial knee joints. To this end, we used the measuring criteria considered to be most commonly used in TKA. For the resection planes, we assumed a coronal plane at a right angle with regard to the functional axis and a sagittal plane at a right angle with regard to the distal axis for the femur and without posterior tilt for the tibia. For circumflex alignment, the SEA served as the reference for the femur, and the Insall line served as the reference for the tibia. Because artificial knee joints involve component circumflex assumed at the time of design, the rotational relationship is reproduced at the time of reduction of the femoral and tibial components to maximize the functioning of the artificial knee joint. In this survey, measurements were taken with alignment of each of the femur and tibia models on CT images taken in a non-loaded supine position; therefore, the angle between the femoral SEA and tibial Insall line remained unknown. Wernecke et al.  reported that the Insall line in extension exhibits an external rotation by a mean of 2.7 degrees with respect to the SEA and that individual differences exist. When designing an artificial joint, it seems necessary to make a dimensional correction for this angle and to have a design accommodating circumflex mismatches between the femur and tibia.
With regard to sex-related differences in measurements, Lonner et al.  reported that the femoral MAP/ML was narrower in women than in men (0.84 vs. 0.81, respectively). Two types of prior study involving the AP/ML ratio were identified: one with AP as the denominator and the other with ML as the denominator. Because AP dimensions occur on both the medial and lateral sides and are susceptible to circumflex alignment, we considered that ML was more useful as a size reference than AP, and ratios were calculated with ML as the denominator. In the present study, there were no significant sex-related differences in F-MAP/F-ML; however, F-LAP/F-ML was significantly narrower in women than in men, and the tendency was thus the same. For the tibia, there were no significant sex-related differences in T-MAP/T-ML or T-LAP/T-ML.
If we assume an artificial knee joint fabricated with an F-MAP of 60.0 and an F-ML of 70.7, which correspond to a femoral F-MAP/F-ML of 0.849 determined as the mean of both sexes in the present study, and given an overhang/underhang of 3 mm on one side (6 mm on both sides), for example, F-MAP/F-ML in the range of 0.78 to 0.93 will be covered. However, some knees had F-MAP/F-ML measurements outside the range of 0.78 to 0.93 (21 in 188 knees = 11%), suggesting that the proportional use of a single size can cause mismatches even in normal joints. To ensure full coverage for the bone, it seemed necessary to make at least two variations with different ratios or to provide custom-made artificial knee joints. In addition, because the medial and lateral condyles are asymmetrical, artificial knee joints with different thicknesses are more physiological. However, individual differences exist even in normal knees, and nothing more than average joint shape can therefore be attained with proportional design.
With regard to the relationship between the AP-ML aspect ratio and size, Hitt et al.  and Kwak et al.  demonstrated a correlation between AP size and ML/AP, showing that the width narrowed with decreasing size in both the femur and tibia. In the present study, to confirm size-related differences, F-MAP/F-ML and F-LAP/F-ML relative to F-ML dimensions and T-MAP/T-ML and T-LAP/T-ML relative to T-ML dimensions were determined in both sexes. In both sexes and in both the femur and tibia, the value increased with decreasing size and vice versa; therefore, the width narrowed with decreasing size, which is similar to the trends reported by Hitt et al.  and Kwak et al. . It was considered that the width may be narrower in women than in men because women have smaller bone sizes than men. The matching ratio was considered to improve as the aspect ratio is adjusted according to the size when an artificial joint is designed.
There were no significant differences in the measured sagittal flexion angle between the Japanese and Caucasian women. A prior study showed that the Japanese femur was more anteriorly bowed . Despite this greater bowing in the Japanese population, we found no significant differences in the present study; therefore, racial differences need not be taken into account when surgery is performed using bowing-free intramedullary rods approximately 150 mm in length. Nor was there any significant difference in F-MAP/F-ML and F-LAP/F-ML, unlike the results reported by Kim et al. . In the present study, the individual differences were greater than the racial differences; therefore, we consider that there is no significant need to change the shape of the artificial knee joint according to racial differences in bone morphology.