Patients
Institutional review board approval was obtained for this retrospective study. From September 1997 to December 2004, a total of twenty-seven consecutive patients (thirty-seven hips) with developmental dysplasia of the hip were treated with cementless total hip arthroplasty with use of the S-ROM modular stem (DePuy, a Johnson & Johnson Company, Warsaw, Indiana). Two patients (four hips) were lost to follow-up two to three years after the surgery. Complete clinical and radiographic data obtained more than five years after the surgery were available for the remaining thirty-three hips in twenty-five patients (two men and twenty-three women).
All patients had developmental dislocation of the hip. The hip dislocations were not associated with any known syndromes or neuromuscular disorders. All hips had complete dislocation corresponding to Group IV in the Crowe classification4, which divides developmentally dysplastic hips into four groups according to the percentage of proximal displacement of the femoral head with respect to the height of the pelvis. The vertical distance from the inferior margin of the ischial tuberosity to the tip of the iliac crest is used as the height of the pelvis. According to Crowe et al.4, the vertical diameter of the normal femoral head is approximately 20% of the height of the pelvis. The vertical distance of the femoral head-neck junction from the interteardrop line connecting the lower parts of the radiographic teardrops is defined as the proximal displacement of the femoral head. A dysplastic hip is considered to be <50% subluxated (Crowe Group I) if the proximal displacement of the head is <10% of the height of the pelvis, 50% to 75% subluxated (Crowe Group II) if the proximal displacement is 10% to 15%, 75% to 100% subluxated (Crowe Group III) if the displacement is 15% to 20%, and completely dislocated (Crowe Group IV) if the displacement is >20%.
The indication for surgery in the twenty-five patients was pain unresponsive to nonoperative management. Complete hip dislocation was present bilaterally in eight patients and unilaterally in seventeen patients, four of whom showed osteoarthritis in the contralateral hip. A Schanz osteotomy had been performed in two hips (one patient) and a tectoplasty25, in two hips (one patient). A tectoplasty is a type of acetabuloplasty, the aim of which is to provide an extra-articular weight-bearing surface in patients with a dysplastic acetabulum or a hip subluxation or dislocation with a false acetabulum. The lateral wall of the iliac bone at the lateral edge of the affected acetabulum is raised as a proximally based flap, and massive bone grafts are inserted to provide a congruous, nonabsorbable roof for the hip capsule and femoral head25.
Surgical Planning and Procedures
For preoperative planning, an acetabular template was placed at the position of the anatomical hip center. The angle between the vertical line and a line drawn from the cup center to the lateral edge of the acetabulum (the cup center-edge angle26) was measured. If the cup center-edge angle was <10°, bulk bone-grafting with the resected femoral head was planned. A femoral template was placed with the prosthetic femoral head center at the height of the tip of the greater trochanter to preserve femoral bone stock, optimize the abductor moment arm, and avoid osseous impingement. The vertical difference between the center of the prosthetic femoral head and the center of the acetabular cup represents the maximum amount of lengthening that would result without shortening of the femur. We planned the amount of femoral shortening so that the lower limb would be lengthened no more than 3 cm.
All procedures were performed through a posterior approach with the patient in the lateral position. After femoral head resection and total capsulectomy, we removed osteophytes covering the true acetabulum with the use of curved gouges in order to recognize the position of the true acetabulum before reaming. In most of these hips, the true acetabulum is covered with fibrous scar tissue and large osteophytes and looks like a wall. As acetabular walls are inherently fragile here, the acetabulum was reamed carefully with hemispherical reamers to obtain an interference fit between the anterior and posterior columns, with bleeding cancellous or subchondral bone. The acetabulum was underreamed by 1 to 2 mm, depending on the bone quality. The acetabular cup was implanted in the position of the anatomical hip center with use of the press-fit technique, with the surgeon aiming for 40° of inclination and 15° of anteversion as seen radiographically. Bulk bone-grafting with the resected femoral head was performed in three hips.
Femoral preparation was performed before subtrochanteric osteotomy unless the femoral canal deformity prevented straight axial reaming, which was done to determine the diameter of the distal part of the stem. Conical and triangular reaming of the metaphysis was then performed to prepare the metaphyseal cone. A shortening osteotomy was performed with a step-cut design of 1 cm in height at 2 cm below the end of the proximal sleeve, corresponding to 1 to 2 cm below the lesser trochanter. If trial reduction was impossible, additional shortening was performed. If there was femoral canal deformity from prior surgery, the femoral preparation was carried out after performance of a subtrochanteric step-cut osteotomy at the level of the deformity. Conical and triangular reaming of the metaphysis was performed with the osteotomy fragments held with forceps. If trial reduction was impossible, additional shortening was performed.
The average amount of shortening was 4.1 cm (range, 2.5 to 6.8 cm). After implantation of the proximal sleeve, the straight stem was inserted with 30° of anteversion, with the osteotomy fragments held with forceps. Any gaps at the osteotomy site were filled with autogenous morselized bone. Wiring at the osteotomy site was performed in the initial five cases and in some hips with poor bone quality and intraoperative fracture—i.e., in a total of twenty-two hips.
The S-ROM ZTT I acetabular cup (DePuy) was used in twenty-four hips and the Trilogy Acetabular System (Zimmer, Warsaw, Indiana), in nine hips. The median diameter of the acetabular cup was 42 mm (range, 40 to 48 mm). The median stem length was 130 cm (range, 125 to 165 cm), and the median diameter of the distal part of the stem was 11 mm (range, 8 to 15 mm). No long-stem designs were used. Eighteen hips were treated with a 22-mm cobalt-chromium femoral head; twelve hips, with a 26-mm head; and three hips, with a 28-mm head. A conventional ultra-high molecular weight polyethylene liner was used in twenty-four hips, and a highly cross-linked polyethylene liner (Longevity; Zimmer) was used in nine hips. Constrained liners were not used.
The initial ten patients were allowed partial weight-bearing at eight weeks after the surgery, and full weight-bearing was permitted at twelve weeks. Subsequent patients were allowed partial weight-bearing from three days to three weeks after the surgery and then full weight-bearing at one to six weeks, depending on the intraoperative stability of the femoral component and osteotomy site.
Clinical Assessment and Radiographic Evaluations
Clinical follow-up was performed at three, six, and twelve months after the surgery and annually thereafter. Clinical assessment was performed with use of the Merle d'Aubigné and Postel hip score, with a maximum score of 6 points each for pain, range of motion, and gait function27. Complications were reviewed.
Preoperative and postoperative serial anteroposterior and lateral radiographs of the hip were reviewed. All radiographic measurements were adjusted for magnification. The magnification of preoperative radiographs was calculated with use of a calibration marker placed lateral to the patient at the height of the greater trochanter. The magnification of postoperative radiographs was calculated by measuring the size of the prosthetic femoral head. Radiographs were evaluated with respect to the zones around the femoral component described by Gruen et al.28. The cortical index, preoperative bone quality, level of neck resection, amount of filling of the distal part of the canal by the stem, orientation of the femoral component, stability of the implant, and degree of subsidence were documented. Preoperative bone quality was classified according to the system of Dorr et al.29. The level of the neck resection was measured as a vertical distance from the base of the lesser trochanter. The orientation of the component was classified as valgus, slight valgus, neutral, slight varus, or varus according to the method described by Christie et al.30.
The mode of fixation of the proximal sleeve was classified as bone-ingrown, fibrous stable, or unstable with use of the classification system of Engh et al.31. Migration of the femoral component was assessed by measuring the vertical distance from the shoulder of the stem to the tip of the greater trochanter and the varus angle of the stem formed by the stem axis and the neutral axis of the femoral canal of the distal fragment. Femoral subsidence of >4 mm or a change in the varus angle of >2° was considered to indicate stem migration. The acetabular cup was considered to be loose when the vertical distance from the medioinferior edge of the cup to the interteardrop line had changed by >2 mm or the opening angle of the cup had changed by >5°32.
Evidence of spot welding, osteolysis, stress shielding, and heterotopic ossification was evaluated on serial radiographs. Stress shielding was classified with the system of Engh et al.31. Heterotopic ossification was classified according to the system of Brooker et al.33. Whether and when bone union at the osteotomy site occurred were also assessed.
The vertical distance between the interteardrop line and the tip of the greater trochanter was measured on preoperative and postoperative anteroposterior radiographs. The height of the tip of the greater trochanter in the neutral position was calculated by adjusting for the degree of abduction or adduction on each radiograph. Preoperative limb-length discrepancy was evaluated by comparing these distances between the left and right sides. Postoperative limb-length discrepancy was measured with the amount of shortening taken into account. The amount of advancement of the tip of the greater trochanter was calculated by subtracting the postoperative height of the greater trochanter from the preoperative height. The amount of limb lengthening was calculated by subtracting the amount of femoral shortening from the amount of translation of the tip of the greater trochanter.
Survivorship Analysis
The end points for survival were defined as radiographic loosening and revision. Survival of implants was determined with actuarial life-table constructs described by Kaplan and Meier.
Source of Funding
There was no external funding source for this study.
Clinical Results
The mean age at the time of surgery was sixty years (range, forty-five to seventy-six years). The mean duration of follow-up was eight years (range, five to eleven years). The mean patient height was 147 cm (range, 136 to 157 cm), and the mean patient weight was 51 kg (range, 42 to 63 kg).
The mean Merle d'Aubigné and Postel hip score improved from 9 points (range, 3 to 14 points) preoperatively to 16 points (range, 13 to 18 points) at the time of the latest follow-up. Two patients had a positive Trendelenburg sign, and three had a slight limp at the time of the latest follow-up. Intraoperative fracture occurred in the proximal femoral fragment in four hips and in the distal part of the diaphysis in four hips, and all of the fractures were successfully stabilized by wiring. No cases of infection, symptomatic deep vein thrombosis, symptomatic pulmonary embolism, or nerve palsy were encountered. Two hips treated with a 22-mm femoral head had postoperative dislocations without malalignment of the acetabular cup. In one case, anterior dislocation occurred twice within one week after the surgery and was treated successfully with closed reduction and cast immobilization for three weeks. In the other case, anterior dislocation occurred at ten days postoperatively and was initially treated with bracing. Since dislocation recurred five years postoperatively, revision surgery was being considered at the time of writing.
Radiographic Results
One hip showed progressive radiolucent lines around the proximal femoral sleeve within two years after the surgery, and this was followed by progressive stem subsidence (Fig. 1). In this hip, resection had been performed at the base of the femoral neck because of severe hypoplasia of the neck. The femur was classified as Dorr type C29, while filling of the distal part of the canal by the stem was >90% and no initial malalignment of the stem was apparent (Table I). Stem revision with a cemented femoral stem was performed 4.8 years after the surgery. No intraoperative findings of gross metallic debris or blackening of the periprosthetic tissue was identified. This was the only femoral stem revision. No loosening of acetabular components was seen, and no acetabular revisions were required.
Spot welds around the inferior border of the metaphyseal sleeve were observed in twenty-one hips (64%) (Fig. 2). First-degree stress shielding was observed in four hips (12%); second-degree stress shielding, in twenty-four hips (73%); and third-degree stress shielding, in four hips (12%). Small focal areas of osteolysis of the greater trochanter were seen in only one hip. Brooker class-I heterotopic ossification was seen in one hip. In those hips with a gap at the osteotomy site, the gap disappeared at an average of 3.6 months (range, two to eight months) postoperatively, and this was considered to represent bone union at the osteotomy site. There were no cases of stem fracture or dissociation of a modular component.
The mean advancement of the greater trochanter was 60 mm (range, 36 to 97 mm). As seen radiographically, the mean limb lengthening was 19 mm (range, 0 to 41 mm). The mean limb-length discrepancy seen radiographically in the seventeen patients with unilateral involvement was reduced from 5.1 cm (range, 3.7 to 6.5 cm) to 2.8 cm (range, 1.4 to 4.6 cm).
Survivorship Analysis
Kaplan-Meier survivorship with an end point of stem revision was 97% (95% confidence interval [CI], 80% to 99%) at eight years (Fig. 3). Kaplan-Meier survivorship with an end point of aseptic loosening of the stem was 97% (95% CI, 80% to 99%) at eight years (Fig. 3).
There have been some studies of the mid-term results of cementless total hip arthroplasty with use of a nonmodular standard stem in combination with subtrochanteric femoral osteotomy for hips with a high dislocation11-16,19,20,22. Inadequate fit and fill of a nonmodular standard stem causes stem loosening11,18,22, stem fracture due to distal fixation11, and nonunion of the osteotomy site due to insufficient torsional stability15,18,22. In our consecutive series, no stem fractures or nonunions at the osteotomy site were encountered, while one hip showed stem loosening (Table II). In the hip showing stem loosening, femoral neck resection had been performed at the base of the femoral neck because of severe hypoplasia of the femoral neck. Whiteside et al. examined the effect of femoral neck retention on torsional stability of modular femoral components with a tight fit in the proximal and distal parts of the femur and reported that the base of the femoral neck was the most important factor34. If femoral neck resection was performed at the base of femoral neck, micromotion caused by torsional loading increased substantially. In addition, the bone quality of the hip showing stem loosening in our series was poor (Dorr type C), although filling of the distal part of the femoral canal by the system was >90% and no initial malalignment of the stem was apparent. We thus surmised that the femoral stem loosening was caused by insufficient proximal rotational stability due to hypoplasia of the femoral neck and insufficient distal rotational stability due to poor bone quality. Use of a cemented femoral component might be preferable in hips with poor bone quality and a developmentally short femoral neck, when femoral neck resection needs to be performed at the base of femoral neck.
In the 1980s, we treated completely dislocated hips with total hip arthroplasty with cement by osteotomizing the greater trochanter and excising a portion of the proximal part of the femur5. However, of thirty-four hips followed for a mean of nine years, ten (29%) showed nonunion of the greater trochanter with an impaired functional status of the hip, manifesting as pain and limping with a significantly increased frequency of a Trendelenburg sign (44%). Moreover, five hips (15%) showed focal osteolysis due to a defect in the cement mantle and three hips (9%) showed stem loosening due to an inadequate cement mantle and insufficient rotational stability of the straight, small stem. These results suggested that successful management of high hip dislocation requires a method of femoral shortening without impairment of the abductor function and the use of a cementless stem with contours that can fit a narrow and straight medullary canal and a severely distorted femur. In the present study, subtrochanteric shortening osteotomy combined with anatomical placement of the cup led to good active abduction strength postoperatively, as most patients walked without a limp and displayed a negative Trendelenburg sign. By comparing the results in the present study with those in our historical control group, we could conclude that cementless, modular total hip arthroplasty combined with subtrochanteric shortening osteotomy resulted in satisfactory outcomes for the treatment of developmental dysplasia of the hip.
Modular femoral components raise concerns about fretting, corrosion, and mechanical failure. Tanzer et al. reported that femoral osteolysis had occurred around twenty-five of fifty-nine modular femoral components evaluated six to twelve years postoperatively35. However, they could not determine whether femoral osteolysis was influenced by fretting debris from the stem-sleeve modular junction, as they used standard polyethylene and twenty of the twenty-five hips with femoral osteolysis had an eccentric position of the femoral head in the acetabular liner. In our consecutive series, femoral osteolysis was seen around only one of thirty-three modular femoral components. The patient with the stem revision in this series showed no intraoperative findings of gross metallic debris or blackening of periprosthetic tissue.
We were able to completely avoid neurological complications, which are common after total hip arthroplasties for completely dislocated hips, by determining the amount of femoral shortening required to avoid increasing limb length by >3 cm, which would have placed tension on the sciatic nerve4-7,9-25. The rate of intraoperative fracture was comparable with the rates reported in other studies of total hip arthroplasty for completely dislocated hips4-7,9-25. During preparation of the hypoplastic and distorted proximal part of the femur, it was difficult, in some cases, to obtain secure stability of the femoral component while avoiding intraoperative fracture. We encountered two cases of anterior dislocation without malalignment of the acetabular cup despite the posterolateral approach and the derotation of the stem neck and adjustment of femoral offset by modularity. Use of a small, 22-mm femoral head and total capsular release might cause joint instability. The joint capsule was, however, elongated in most of these dislocated hips, and it was necessary to excise redundant capsule to prevent soft-tissue impingement. All of the periarticular muscles were retained to prevent impaired abductor function postoperatively. Kerboull et al. reported that bringing down the hip to the level of the true acetabulum was possible provided that the entire articular capsule, scar tissues, osteophytes, and a shelf (when present) were removed36. We agree. A high rate of dislocation is an unresolved problem associated with total hip arthroplasty for treatment of developmental dysplasia of the hip. Use of a larger-diameter head coupled with third-generation highly cross-linked polyethylene might compensate for associated soft-tissue laxity.
Several osteotomy designs have been reported, including transverse11,12,14,16-19,21-23, v-shaped13, and step-cut techniques15,16,20. A step-cut design has several favorable effects fostering bone healing, including a larger bone contact surface and stability against torsional stresses. Onodera et al. reported that nonunion of a transverse osteotomy site occurred in one of fourteen hips treated with a modular stem (Table II)19. Sener et al. reported nonunion of a step-cut osteotomy site in two of twenty-eight hips treated with a nonmodular standard stem15. No nonunions were seen in our consecutive series of modular total hip arthroplasties combined with a step-cut osteotomy. These results indicate that bone union of the osteotomy site requires stability of the femoral prosthesis and osteotomy site and a larger bone contact surface.
It is difficult to determine if a shortening osteotomy is necessary preoperatively and intraoperatively. Kerboull et al. reported favorable results in a consecutive series of 118 total hip arthroplasties performed with cement for hips with a Crowe Group-IV dislocation36. In that series, in which the femoral component was implanted at the level of the lesser trochanter, only two hips required a shortening osteotomy and only one transient peroneal nerve palsy occurred despite the fact that thirty limbs were lengthened >4 cm. To our knowledge, there is no documented maximum amount that an extremity can be safely lengthened without a neurological complication37,38. Some authors have reported that overlengthening is a plausible etiology for nerve injury after total hip arthroplasty38. We believed that lengthening of >3 cm could increase the risk of nerve palsy in our patients with short stature (mean height, 147 cm). In addition, we believed that femoral shortening is necessary to reconstruct an anatomical hip center and to perform a femoral neck resection above the base of femoral neck.
This retrospective study had two limitations. It was an uncontrolled case series. We did not compare the results of total hip arthroplasty combined with subtrochanteric shortening osteotomy between nonmodular stems and modular stems. However, we concluded that our procedure resulted in satisfactory outcomes by comparing the present cohort with our historical control. The other limitation is a lack of power to analyze bilateral and unilateral cases separately.
In conclusion, cementless, modular total hip arthroplasty for the treatment of developmental dysplasia of the hip combined with subtrochanteric osteotomy resulted in satisfactory outcomes. However, obtaining sufficient rotational stability for bone ingrown fixation of the proximal sleeve was difficult in hips with poor bone quality and a developmentally short neck, in which resection is needed at the base of the femoral neck.