Between 1995 and 2006, we performed a curved periacetabular osteotomy on 388 hips for the treatment of symptomatic acetabular dysplasia in adolescents and adults. All of the surgical procedures were performed by the senior author (M.N.). We evaluated forty-six of these hips in forty-two patients who were fifty years of age or older (the older group) at the time of surgery and who had a minimum clinical follow-up period of two years. No patient fifty years of age or older who underwent a curved periacetabular osteotomy was excluded from this study. The mean age of the patients at the time of surgery was 54.6 years (range, fifty to sixty-five years), and the mean follow-up period was 4.6 years (range, 2.0 to 11.4 years). We also evaluated fifty hips in forty-four patients who were younger than fifty years (the younger group) at the time of surgery. The mean age of the patients in the younger group at the time of surgery was 32.3 years (range, nineteen to forty-nine years), and the mean follow-up period was 4.4 years (range, two to eleven years). The patients in the younger group were matched to the patients in the older group according to sex and Tönnis grade12 within a similar time period. Patients who had undergone previous surgery were excluded from the study. The study was approved by the institutional review board. All patients gave informed consent prior to their participation in the study.
Surgical indications for the curved periacetabular osteotomy included acetabular dysplasia with symptoms, such as pain that was tolerable but made the patient feel uncomfortable and caused some limitations in his or her daily activities for more than five months; a lateral center-edge angle13 of <16° on anteroposterior radiographs; and an improvement in joint congruency on the anteroposterior radiograph with the hip in abduction3,4. Preoperatively, the appearance of the hip following osteotomy was simulated, and anteroposterior radiographs of the pelvis with the hip in abduction and a false-profile view of the affected hip in flexion were made with the patient in the supine position. In addition, improvement in joint congruency and widening of the cartilage space with the hip in the abducted position were evaluated. This type of osteotomy is not recommended for patients with aggravation of joint incongruency; narrowing of the cartilage space; and, especially, partial disappearance of the cartilage space in the abducted position. Instead, we recommend that such patients with more advanced osteoarthritis be treated with a salvage osteotomy, such as the Chiari osteotomy and/or an intertrochanteric valgus osteotomy, or total hip arthroplasty. During this same time period, between 1995 and 2006, we performed a total hip arthroplasty in 918 hips in patients with dysplasia who were fifty years of age or older. Preoperatively, all of the patients were cautioned not to increase their body weight, carry heavy items, or perform high-impact sports.
Surgical Technique
The curved periacetabular osteotomy was performed with use of a direct anterior approach with the patient in the supine position as has been described previously4. The skin incision of this osteotomy was relatively small, approximately 9 cm long. The surgical exposure was similar to that in previous studies14,15, and little damage was caused to the hip abductor muscles because the gluteal muscles were not stripped from the bone14,16. Before the osteotomy of the anterior superior iliac spine, two holes were drilled for later fixation. The anterior superior iliac spine was then osteotomized with the inguinal ligament and sartorius muscle attached and retracted medially. The supra-acetabular portion of the iliacus muscle was detached, and the inner table of the pelvis was sharply stripped. A c-shaped osteotomy line was marked with use of a power drill from the anterior inferior iliac spine to the distal part of the quadrilateral surface, which was located one finger width anterior to the greater sciatic notch (Figs. 1-A, 1-B, and 1-C). To keep the posterior column of the acetabulum intact and avoid creating an intra-articular osteotomy, the most posterior part of the c-shaped osteotomy line in the quadrilateral surface was placed in front of the sciatic notch but not more than 1.5 cm anterior to it17. After the most posterior point, which comprised the midpoint of the c-shaped osteotomy, had been determined, the proximal half of the osteotomy was confirmed with use of an intraoperative image intensifier. The distal half of the osteotomy was then created symmetrically with the proximal half. The end of the osteotomy extended >2.5 cm below the pelvic brim to connect with the ischial osteotomy line. The superior ramus of the pubis was also scored just medial to the iliopubic eminence with use of a power drill. The actual osteotomy along the scored line was carried out with use of a curved osteotome, which was designed to correspond approximately to the circumferential curvature of the acetabulum11. The iliopsoas muscle was retracted medially with the hip in flexion and slight adduction. The osteotomy of the quadrilateral space was initially carried out by directing the curved osteotome posteriorly in the proximal portion and distally in the distal portion. The distal portion of the quadrilateral surface was deep in the incision and could be partially visualized with use of a wide retractor. The ischium was palpated with an elevator, and the osteotome was introduced into the space between the distal joint capsule and the psoas tendon. When the blade contacted the corpus of the ischium, the direction of the blade toward the infracotyloid groove was ensured with use of an image intensifier. The anatomic guide point for the osteotomy site was approximately 14 mm inferior to the distal margin of the hip11,17. The blade was hammered 10 mm into the bone without attempting to break it. A pubic osteotomy was performed with use of an inclination of 30° to the horizontal line for medialization of the femoral head18. To cut the pubis at 30° to the horizontal line, a Kirschner wire was placed on the transteardrop line under an image intensifier and was used as a horizontal line landmark. The angle made by the Kirschner wire and the chisel was checked under the image intensifier. The final step of the osteotomy was carried out in the supra-acetabular portion of the ilium, into which the blade was hammered in a proximal-lateral direction. Typically, the bone was very thick and dense around the linea arcuata. The blade was hammered in the posterior-lateral direction in this area, and the osteotome was allowed to follow its own curve. When the blade penetrated the ilium, a loss of resistance was felt. If it was difficult to connect the most distal portion of the c-shaped osteotomy with the ischial osteotomy, the involved hip was placed in external rotation and abduction with use of a figure-of-four maneuver, and the osteotomy line was extended into this area. The osteotomized acetabular fragment usually began to move as a result. If it failed to move, a spreader was driven into the osteotomy site to complete the most distal part of the c-shaped osteotomy and then was hammered a few times until the acetabular fragment broke loose. The c-shaped osteotomy enabled acetabular reorientation because the osteotomy surfaces had the same curvatures. The acetabular fragment was redirected and temporarily fixed with a Kirschner wire. The hip was flexed and rotated internally to verify that there was no excessive anterior or lateral correction that might lead to femoroacetabular impingement. To confirm a lack of femoroacetabular impingement, the anterior part of the hip joint capsule was palpated when the hip was placed in flexion and internal rotation. If abutment of the anterior femoral head and neck on the anterior rim of the acetabulum was detected, the reoriented acetabular fragment was rotated posteriorly. The image intensifier or intraoperative radiography was used to ensure that the desired goals were achieved, namely, that the femoral head was adequately covered by the reoriented acetabular fragment and that the hip had been medialized.
All of the curved periacetabular osteotomies in this study were performed with use of the same technique, with the exception of a change from cannulated cancellous screws to poly-L-lactic acid screws after these screws became available, and two or three screws were used to fix the reoriented acetabular fragment. Specifically, one or two screws were inserted from the reoriented acetabular fragment into the ilium in a superior-lateral direction, and another screw was inserted from the ilium into the reoriented acetabular fragment in a distal direction. Because direct contact between the osteotomy surfaces was achieved, a bone graft was not used. Passive movement of the hip was used to confirm the stability of the acetabular fragment. No resection of the degenerative limbus or osteophytes was carried out. The osteotomized anterior superior iliac spine was then returned to its original position and fixed with two cannulated cancellous screws.
Active motion exercises were initiated on postoperative day 2, and partial weight-bearing with crutches was allowed on postoperative day 3. Full weight-bearing was allowed at eight weeks after the surgery.
Clinical Evaluation
All patients were evaluated clinically with use of the Harris hip score by one of two investigators (T.T. and T.K.)19. In this system, pain, function, and range of motion are included. In addition, the surgical time and amount of blood loss during surgery were recorded for all patients.
Radiographic Evaluation
The severity of osteoarthritis was classified into four grades with use of the Tönnis12 system. Specifically, hips with subchondral sclerosis were classified as grade 1; hips with subchondral cyst formation and partial cartilage interval narrowing, as grade 2; hips with severe or complete but localized cartilage interval narrowing, as grade 3; and hips with extensive and severe or complete nonlocalized loss of the cartilage interval, as grade 4.
Radiographic measurements included the lateral center-edge angle, acetabular roof obliquity20, acetabular head index21, anterior center-edge angle22, and acetabular head lateralization index23. The ratio of femoral head lateralization was calculated with use of the acetabular head lateralization index and modified Ninomiya measurements24 (Fig. 2). The distance between the two Kohler ilioischial lines (t) was referred to as the interhip distance. The acetabular head lateralization index was calculated with use of the formula (a)/(t)/2, and the ratio of femoral head lateralization was calculated by dividing the postoperative acetabular head lateralization index by the preoperative acetabular head lateralization index. Anteroposterior radiographs of the pelvis were made with the patient in a supine position. The tube-to-film distance was 120 cm, and the tube was perpendicular to the table. The center beam was directed toward the midpoint between the upper border of the symphysis and a horizontal line connecting both anterior superior iliac spines. The method of Siebenrock et al.25 was used to judge the extent of pelvic inclination. No corrections were made for radiographic magnification.
Radiographic measurements were taken with use of a digital caliper (Mitutoyo, Tokyo, Japan) with an accuracy of ±0.02 mm. Measurements of the radiographs were carried out by two authors (T.T. and A.M.) who completed repeated training sessions and were blinded to the clinical results. In addition, the radiographs were reviewed three times on different days by the same observers, and the average values were calculated to assess intraobserver reliability. We assessed the intraobserver and interobserver reliabilities for the radiographic measurements using the intraclass correlation coefficient. The respective intraobserver reliabilities for the lateral center-edge angle, acetabular roof obliquity, acetabular head index, anterior center-edge angle, and head lateralization index in the older group were 0.93, 0.92, 0.92, 0.90, and 0.94 preoperatively and 0.91, 0.92, 0.91, 0.93, and 0.92 postoperatively. The respective intraobserver reliabilities for the younger group were 0.92, 0.92, 0.93, 0.90, and 0.92 preoperatively and 0.91, 0.92, 0.91, 0.93, and 0.92 postoperatively. The respective interobserver reliabilities for the lateral center-edge angle, acetabular roof obliquity, acetabular head index, anterior center-edge angle, and head lateralization index for the older group were 0.86, 0.88, 0.89, 0.88, and 0.87 preoperatively and 0.87, 0.88, 0.87, 0.86, and 0.87 postoperatively. The respective interobserver reliabilities in the younger group were 0.85, 0.86, 0.87, 0.86, and 0.86 preoperatively and 0.88, 0.87, 0.87, 0.88, and 0.86 postoperatively.
Statistical Analysis
Mann-Whitney U tests were used to compare radiographic parameters and Harris hip scores, surgical times, and blood loss between the two groups. A Wilcoxon signed-rank test was used to compare changes in radiographic parameters within the same group. The probability of progression of the radiographic osteoarthritis stage was estimated with use of the Kaplan-Meier survivorship method. The percent confidence intervals were calculated. The association of preoperative degenerative joint disease with the end point of osteoarthritis stage progression was assessed with the log-rank test. Statistical significance was defined a priori as p < 0.05.
Source of Funding
There was no external funding for the study.
The clinical data for both patient groups are shown in Table I. There were no significant differences in sex (p = 0.48) or preoperative Tönnis osteoarthritis grades between the two groups. There was no significant difference in the surgical time (p = 0.85), blood loss during surgery (p = 0.71), or postoperative drainage (p = 0.38) between the two groups. The time to independent walking was two or three days in both groups, and partial weight-bearing with crutches began at less than seven days after surgery.
The results for the radiographic parameters in both groups are also shown in Table I. There were no significant differences in the lateral center-edge angles, acetabular roof obliquities, acetabular head indexes, anterior center-edge angles, acetabular head lateralization indices, or ratios of femoral head lateralization between the two groups preoperatively or postoperatively. The mean ratios of femoral head lateralization were 0.93 in the older group and 0.95 in the younger group. Six hips exhibited lateralization of the femoral head postoperatively in the older group, and five hips exhibited lateralization of the femoral head postoperatively in the younger group. Preoperatively, five of the six hips with lateralization of the femoral head in the older group had Tönnis grade-3 osteoarthritis and the remaining hip had Tönnis grade 4. Similarly, preoperatively, four of the five hips with lateralization of the femoral head in the younger group had Tönnis grade-3 osteoarthritis and the remaining hip had Tönnis grade 4.
In the older group, the Tönnis grade improved in two hips (4%; Figs. 3-A and 3-B), remained unchanged in forty-one hips (89%), and progressed in three hips (7%). One of the latter three hips showed progression from Tönnis grade 3 to grade 4, and the remaining two hips showed progression from Tönnis grade 2 to grade 3. In the younger group, the Tönnis grade improved in three hips (6%), remained unchanged in forty-four hips (88%), and progressed in three hips (6%). Two of the latter three hips showed progression from Tönnis grade 3 to grade 4, and the remaining hip showed progression from Tönnis grade 2 to grade 3.
The clinical scores were similar between the two groups (Table I). The clinical data for the separate parts of the Harris hip scores for both patient groups are shown in a table in the Appendix.
No intraoperative or early postoperative complications, such as chisel penetration into the joint, major vascular injury, sciatic nerve injury, wound problems, infection, reflex sympathetic dystrophy, symptomatic pulmonary embolism, or prolonged abductor weakness, were encountered. Nonunion of the pubis was noted in one patient in the older group. One patient in the older group underwent conversion to a total hip arthroplasty at five years after the curved periacetabular osteotomy, and one patient in the younger group underwent conversion to total hip arthroplasty at ten years after the osteotomy.
Kaplan-Meier survivorship analysis, with radiographic evidence of progression of the osteoarthritis stage as the end point, predicted five-year survival rates of 94.3% (95% confidence interval, 86.5% to 100%) in the older group and 96.2% (95% confidence interval, 88.8% to 100%) in the younger group. There was no significant difference between the two groups (Fig. 4).
There are few reports of clinical results for periacetabular osteotomy performed on patients fifty years of age or older at the time of surgery. In the present study, no significant difference was identified between the older and younger groups with regard to the clinical scores, radiographic results, or the survival rates with radiographic evidence of osteoarthritis progression as the end point.
The results of various acetabular osteotomies for the treatment of dysplasia have been well described. Progression of secondary osteoarthritis was prevented in 63% of hips after seventeen years26, 79% of hips after eleven years27, 71% of hips after eight years24, and 70% of hips in patients with a mean age of 50.9 years and 94% of hips in patients with a mean age of 34.4 years after ten years23. In the present study, progression of osteoarthritis after five years, determined on the basis of the Kaplan-Meier survivorship method, was prevented in 94% of patients fifty years of age or older and 96% of patients younger than fifty years of age. Even though our follow-up periods were shorter than those in the other studies, our results compare well with their findings.
Few reports have reviewed the results of reconstructive osteotomy in dysplastic hips with degenerative joint disease. Trousdale et al.6 reported that five of nine patients who underwent a second major operation had Tönnis grade-3 osteoarthritis before the periacetabular osteotomy. Eight of our nine patients who had had Tönnis grade-3 or 4 osteoarthritis had a Harris hip score of <70 points at the last follow-up evaluation. Patients with severe osteoarthritis did not fare as well and had significantly lower Harris hip scores and an increased rate of subsequent surgery. We performed curved periacetabular osteotomy even in hips with Tönnis grade-3 or 4 osteoarthritis, if improved joint congruency could be demonstrated on anteroposterior radiographs in the abducted position preoperatively. However, in the older group, one hip showed progression from Tönnis grade 3 to grade 4 and two hips showed progression from Tönnis grade 2 to grade 3. In the younger group, two hips showed progression from Tönnis grade 3 to grade 4 and one hip showed progression from Tönnis grade 2 to grade 3. We performed the osteotomy in a fifty-one-year-old woman with Tönnis grade-2 osteoarthritis in 1996. The preoperative center-edge angle and acetabular roof obliquity were 10° and 20°, respectively. The center-edge angle and acetabular roof obliquity after the curved periacetabular osteotomy were 25° and 5°, respectively. At five years, the patient had progression to Tönnis grade-3 osteoarthritis and we performed total hip arthroplasty without any substantial difficulty in 2001. The difficulty associated with converting a healed periacetabular osteotomy to a total hip arthroplasty seems to be that the proper setting of the cup during total hip arthroplasty becomes difficult if the anatomy of the acetabulum has been changed by excessive rotation of the acetabular fragment following a curved periacetabular osteotomy.
Some hips had overcoverage in the present study. This can lead to femoroacetabular impingement. In recent procedures, we measured the lateral center-edge angle under image intensification during the osteotomy procedure and now take care to avoid overcoverage, which creates femoroacetabular impingement. The amount of acetabular version may also be an important factor affecting the clinical outcome, since postoperative retroversion may cause anterior femoroacetabular impingement or posterior wall deficiency. Previously, we investigated the clinical and radiographic results in dysplastic hips with acetabular retroversion and compared the clinical outcomes between hips with acetabular anteversion and retroversion after periacetabular osteotomy28. Our results indicated that preoperative acetabular retroversion did not influence the clinical results of periacetabular osteotomy when the acetabular retroversion was corrected to a neutral or anteverted position postoperatively. Therefore, we concluded that when a corrective osteotomy is performed in patients with developmental dysplasia who have a retroverted acetabulum, the surgeon should transfer the osteotomized acetabulum in a posterolateral direction while simultaneously moving the femoral head medially to prevent osteoarthritis of the posterior part of the hip owing to posterior wall deficiency28.
On the basis of our data, we do not recommend a curved periacetabular osteotomy for patients with poor joint congruency or narrowing of the cartilage space, and especially not for patients with partial disappearance of the cartilage space with the hip in an abducted position. Instead, we recommend that patients with dysplastic hips of Tönnis grade 3 or 4 associated with narrowing of the cartilage space, and especially those with partial disappearance of the cartilage space in an abducted position, be treated with a salvage osteotomy29, such as the Chiari and/or intertrochanteric valgus femoral osteotomy, or total hip arthroplasty29.
Successful medialization of the femoral head was observed postoperatively in both groups. We were able to control the medialization of the femoral head safely because of the direct visualization of the medial side of the pubis. However, in patients with Tönnis grade-3 or 4 osteoarthritis and a nonspherical femoral head, medialization of the femoral head proved to be difficult. Two major causes of hip lateralization in advanced osteoarthritis were osteophyte formation of the acetabular fossa and deformation of the femoral head. Therefore, the acetabular fragment could not be rotated smoothly or medialization of the femoral head could not be achieved. Consequently, the nine patients in the older group and eleven patients in the younger group with preoperative Tönnis grade-3 or 4 osteoarthritis had lower mean hip scores postoperatively than did the patients with preoperative Tönnis grade-1 or 2 osteoarthritis. When the pubic osteotomy was performed vertically, the distal part of the pubis hindered rotation of the acetabular fragment. Furthermore, in patients with severe acetabular dysplasia of the hip requiring excessive rotation of the acetabulum, the contact area of the pubic osteotomy site was smaller and the risk of dehiscence and nonunion of the pubic osteotomy increased. Therefore, a modified pubic osteotomy procedure involving an inclination of 30° to the horizontal line to medialize the femoral head and decrease nonunion of the pubis has been used since 200518. When the pubic osteotomy was performed with use of an inclination of 30° and the teardrop was lifted up, the osseous obstruction for rotation of the acetabular fragment at the osteotomy site was decreased and the contact area at the pubic osteotomy was better preserved.
With regard to abductor muscle strength recovery, Ninomiya and Tagawa11 suggested that rotational acetabular osteotomy should not be performed in older patients in whom the hip abductor muscles may not regain sufficient strength postoperatively and who may thus have a severe limp develop. However, Ezoe et al.16 measured muscle strength after curved periacetabular osteotomy and investigated the factors influencing postoperative muscle strength recovery. They found that abductor muscle strength was greater in patients younger than forty years than in patients older than forty years because of differences in the activity levels between the groups, thereby accounting for the slight differences in recovery of abductor muscle strength. Moreover, their data suggested that the severity of preoperative osteoarthritis had a greater influence on postoperative recovery of muscle strength than did patient age30-32. Nevertheless, Ezoe et al.16 reported that the muscle strengths around the hip in both younger and middle-aged patients at twelve months postoperatively had recovered to such an extent that they surpassed the preoperative levels. In the present series, a direct anterior approach was used for surgical exposure, and no additional exposure of the outer table of the pelvis was required. The extent of soft-tissue dissection was limited, and the abductor muscles were left intact. These advantages seem to reduce postoperative complications, promote early postoperative rehabilitation, lead to an early recovery of muscle strength around the hip, and produce very good results even in older patients. Although the upper limit of chronological age for curved periacetabular osteotomy is not known, a sixty-five-year-old woman was the oldest patient at the time of surgery to date. Curved periacetabular osteotomy is not indicated for patients fifty years of age or older if they have severe osteoporosis or are inactive.
The limitations of the present study include the use of anteroposterior radiographs made in the supine position both preoperatively and postoperatively and the assessment of Tönnis grade from these supine anteroposterior radiographs. Some reports have noted that the joint space width of hips is narrower on standing radiographs than on supine radiographs33,34. In addition, a hip flexion contracture may affect the Tönnis grade on a supine radiograph, such that assessment of the Tönnis grade in the supine position may lack some accuracy. Therefore, anteroposterior radiographs should be evaluated in the standing position. Second, our follow-up periods were short and we may not have established whether performing an osteotomy in patients fifty years of age or older is preferable to proceeding directly to total hip replacement.
In conclusion, clinically and radiographically satisfactory results were obtained after curved periacetabular osteotomy in patients fifty years of age or older with Tönnis grade-1 or 2 osteoarthritis of the hip secondary to developmental dysplasia.