This study is the first to investigate CCP in DDH patients during CR and explore the distribution of CCP with different abduction and flexion angles. During CR, the gravity of the lower extremities was offset by the plaster. Thus, the force maintaining the reduction was only the tension of muscles and force exerted by the surgeon during the reduction.
In case of complete hip dislocation, femoral head and acetabulum grow noninteractively, yielding a dysmorphia head and a shallow acetabulum (Fig. 3). Other researches have indicated that in case of lateral hip subluxation, the pressure on the femoral head concentrates in the medial aspect of the femoral head as the hip hinges along the edge of the acetabulum. Likewise, concentric pressure on the acetabular floor is reduced while it is increased along the lateral border. Then, the lateral aspect of the femoral head continues to grow and thus flattens the head. The acetabular growth cartilage fills the acetabular floor and arrests its lateral growth, forming a progressively shallower and more oblique acetabulum [10]. Liqun Duan et al. [11] established healthy children hip model, including complete cartilage and bone in hip joint. By applying FEA, they found that the cartilage in femoral head was thicker in the center and thinner in periphery, whereas the acetabular cartilage was the opposite. Shefelbine et al. [12] found that under loading conditions of the dysplastic hip, the octahedral shear stress was much more obvious on the medial side than the lateral, which promoted growth on the medial side and resulted in coxa valga. Shefelbine [13] also used FEA to compare the morphology between the normal side and affected side of the proximal femur, and concluded that the stress distribution of the femoral head epiphysis in DDH patients was significantly different from that in normal population, which contributes to the abnormal morphology of the femoral head. Recently, Vafaeian et al. [5] reported hip joint contact pressure distribution during Pavlik Harness (PH) treatment of infant. It turned out that PH position generates a horseshoe-shaped articular contact area involving most of the acetabulum but relatively sparing the superolateral portion, and a rapid increase in pressure with increasing leg abduction in harness may result in AVN.
Although we did not catch the finding of coxa valga of proximal femur from our 3-D model, we did find that the structure of the femoral head between two sides had a significant difference. Through the construction of the cartilage model of hip, it perfectly showed us the “plate-shaped” deformity of the affected side as compared with the “cup-shaped” acetabulum of the normal side, which to some extent explained why a severe dislocation hip was tough to reduce (Fig. 3).
By increasing the abduction angles, the simulated CCP demonstrates a “butterfly-shaped” distribution around the U-shaped notch of the acetabulum, due to more pressure absorbed in the anterior and superior acetabulum than the superior portion. Under that condition, it might deepen the acetabulum to guarantee stability during CR. This deepening may become morphologically permanent if femoral head position and pressure on the acetabulum are preserved during continuing growth of the cartilage [14]. Acetabular index (AI) was an important and most commonly used parameter to assess acetabular dysplasia. Li Y et al. [15] retrospectively compared AI, centre-edge angle of Wiberg (CEA), Reimer’s index and centre-head distance discrepancy over time among groups divided by final outcomes, to find out the best predictor of late residual acetabular dysplasia in DDH after CR as well as whether it indicates the necessity for secondary surgery. According to their results, AI was the best post-surgical predictor. If AI > 28° for 1 year following CR or AI > 25° for two to four years after CR, the secondary surgery was warranted. Shin et al. [16] also recommended secondary surgery, if AI > 32° and CEA < 14° when patients older than 3 years. According to our results, the muscle force pushed the femoral head forward on Y-axis. Because the contact area between the femoral head and the superior acetabulum was small, on the Z-axis, the contact stress was minor between the acetabulum and the femoral head. Thus, during CR, inadequate pressure leads to minor stimulating effect, which explained the phenomenon of lack of improvement on the top of the acetabulum deformity after CR. It also explained why the AI is a good marker for failure of the CR.
For the late presenting or diagnosed patients, Nikolaos G et al. [17] reported a modified Hoffmann-Daimler functional method. By placing the dislocated hip in flexion (100°-120°), it allows the femoral head to gradually move into the acetabulum by the redirected action of the adductor and the flexor muscles followed by use of abduction splint (abduction angle 90°). It demonstrated satisfying outcomes with high success rate and low AVN rate among 69 patients (95 hips) with average of 11.5 years follow-up time. The acetabular index will be corrected as a result of the mechanical forces placed on the acetabulum. Some researchers [18] believe that extreme flexion angle (i.e.,>120°) during CR could produce a femoral nerve palsy as the nerve could be compressed by the diapers between the thigh and abdomen. Hyperflexion may also cause the femoral head to dislocate inferiorly. Alternatively, inadequate flexion (i.e., < 90°) will fail to reduce the hip. In our study, the pressure against the femoral head decreased with the flexion angle changing from 90° to 100°, without apparent displacement of acetabulum (Fig. 6), suggesting that a better success rate of CR can be achieved by increasing the flexion angle appropriately.
AVN is a common and severe complication after the DDH treatment, occurring as high as in 60% of patients after CR [19]. The most common cause is the immobilization in a position that places excessive pressure on the femoral head. Thus, Ramsey et al. [20] recommended creating a “safe zone” to prevent AVN. In certain situation, an adductor tenotomy will increase the safe zone by allowing a wider range of abduction. However, extreme abduction should never be used because this has been shown to cause AVN. The relationship between hip abduction and blood-flow velocity in the femoral head has been established with Doppler ultrasound. In normal volunteers, the blood flow in femoral head drops significantly when abduction angle increases: with their hips in neutral position, mean blood flow was 13 cm/sec; at 30 degrees of abduction, it was 10.3 cm/sec; and at 45 degrees, 3.8 cm/sec[21]. The incidence of AVN varies widely (0 ~ 92.4%), mainly due to the lack of unified diagnostic criteria and some different understandings. Bradley et al. [22] conducted a retrospective study on AVN incidence after CR, and it was 10% among 441 DDH children (538 hips) on an average follow-up time of 7.6 years.
In our study, among the simulated five muscles, adductor magnus and adductor longus have the most potent restrictive effect on the abduction of the hip joint. Therefore, cutting adductor longus or adductor magnus during surgery can increase the abduction angle and maintain a stable reduction. As seen from Fig. 4, when the hip was flexed at 90° and abducted at 45°, the stress was mainly concentrated posteriorly. In this case, the contact area between the acetabulum and the femoral head was small, and the reduction was unstable. When the abduction angle was increased to 65°, the stress on the anterior area increased correspondingly, but the stress on the posterior had not increased significantly, which could result in more stable reduction, and distribute the pressure over the contact area more uniformly between the femoral head and the u-shaped notch, without excessive pressure on the femoral head. However, when the abduction of the hip was increased to 80°, CCP on the anterior and posteromedial acetabulum increased significantly, and the corresponding anterolateral and lateral stresses of the femoral head increased significantly, which would undoubtedly increase the pressure on the supportive and epiphyseal arteries of the femoral head, which is very important for the growth and development of femoral head in childhood.
There were multiple necessary improvements and limitations in the model. 1) Since anesthesia is administrated, it may cause of data collection problem, because 3D MRI will take much longer time than the usual one. 2) The unaffected hip was constructed to simulate the distribution of CCP during CR at various abduction angles and flexion angles, which might be incomparable to the affected side. 3) Only adductor muscle and operator force were applied to simulate the force of the joint, which ignores the possibility of soft tissues involvement in the hip joint obstructing reduction, as well as the effect of hip capsule and surrounding ligaments on CCP. Using unaffected side hip may help to control for this involvement. 4) There is no quantitative validation of the muscle forces and CCP, which is currently unsolvable, since in vivo measurements are not possible due to technological challenges and ethical concerns. In the future work, we will continue to conduct simulation analysis on the affected side and load more boundary conditions, or using baroreceptor to simulate the biomechanical changes in the process of CR from the most realistic perspective.