Delayed presentation of bilateral central hip protrusio acetabuli: evidence-based management

Abstract

This is a case of a previously healthy 51-year-old man who sustained bilateral central hip dislocations following a sudden presentation of epileptic seizures. The patient was initially treated conservatively for a period of 9 months. On presentation, he had gross disability due to stiffness in both hips and left peroneal nerve paresis. Through minimally invasive direct anterior approaches, bilateral total hip arthroplasties were performed using tripolar head articulations. These were cemented into a biologic acetabular buttress constructed out of autologous bone graft. The femoral heads and necks were used as plugs and pressed into the acetabular defects, putting the medial acetabular walls under tension. At 24 months’ follow-up, there was a good clinical outcome, and the acetabular walls remodelled bilaterally. In conclusion, in traumatic protrusio acetabuli, a functional, biologic reconstruction of the acetabular wall can be facilitated with the application of distraction osteogenesis (tension-stress) principles while using minimally invasive surgical techniques.

Background

Twenty-eight cases of protrusio acetabuli as a result of seizure attacks have been reported in literature since 1970.^{1 2} Managing traumatic central hip dislocation include either conservative or surgical methods.^3–8 The aim of this article does not concentrate on the clinical outcome, which was excellent, but explores the decision-making process that led to the reconstruction of both medial acetabular walls. This was done in a way which produced the soundest biomechanical environment and provided two functional and stable arthroplasties, preventing socket migration and maintained adequate bone stock for future revisions, if needed, in a young adult.

Case presentation

A previously healthy 51-year-old man sustained bilateral central hip dislocations following the sudden presentation of epileptic seizures. Initially, he was treated conservatively for a period of 9 months with the dislocations unreduced and then presented with gross disability due to stiffness of both hips and muscle atrophy. On physical examination, range of motion (ROM) of both hips was: 5°–30° of flexion, 15° of extension lag and 5°–10° of abduction and adduction on both sides. The patient was unable to sit. Preoperatively, the patient was evaluated to have a Harris Hip Score (HHS) of 2.05.⁹

Investigations

On CT scan (figure 1), the right hip centre had migrated medially to Kohler’s line by 2.07 cm, the columns were intact, however, the rim was non-united and deficient by up to one-third. On the left side, the hip centre was displaced superiorly, by 1.8 cm, and medially, by 2.8 cm, and the rim was non-united and deficient by up to one-half. There were islands of cartilaginous callus in front of the medial wall openings bilaterally (figure 1, figure 2).

Figure 1.

Preoperative CT scan of the pelvis, frontal plane: bilateral hip central dislocations shown with respect to Kohler’s line. On the right side, the displacement is 2.07 cm medially to the Kohler’s line. On the left side, the displacement is 2.80 cm medially to the Kohler’s line and 1.80 cm superiorly to the teardrop.

Figure 2.

Preoperative CT scan of the pelvis, coronal plane: bilateral hip central dislocations shown with presence of islands of callus and cartilage-like tissue adjacent to the defects of the medial walls bilaterally.

At the beginning of the decision-making process, we carefully evaluated the risks involved and prepared accordingly:

The risk of intraoperative neurovascular damage due to the relationship of the femoral head to pelvic neurovascular structures was addressed with preoperative CT angiography. In addition, we ensured a vascular surgeon was on site, the patient was positioned supine and the abdomen was prepared and draped, in case abdominal access for vascular repair was needed.

The risk of difficult exposure due to soft tissue contractions was overcome by preparing to perform bilateral trochanteric slides and to extend the anterior hip approach to a lateral approach, if needed. The risk of cup malpositioning was overcome by using intraoperative fluoroscopy.

In the event of an intraoperative fracture or discovery of a gross acetabular deficiency, instrumentation for femoral reconstruction and antiprotrusio rings to perform major acetabular reconstruction were available.

Treatment

Bilateral total hip arthroplasty (THA) with minimally invasive direct anterior approach was performed on a normal fracture table. The legs were prepped and draped simultaneously. An image intensifier was used. On the right side, the neck of femur was divided by a double osteotomy (one at the base of the neck and one higher, at the level where the neck was impinging against the lip of the acetabulum). On the left side, where the hip protrusion was more severe, a trochanteric slide was performed that was left unfixed prior to a similar double osteotomy of the neck of femur. Both femoral heads were sequentially excised and denuded from the articular cartilage

Intraoperative findings: There was difficulty in extracting the femoral heads, the left side being worse. Femoral heads were held by a cork-screw and following extensive and meticulous capsular releases by the use of diathermy and thin osteotomes, they were brought out of the wound for graft preparation.

Acetabular deficiencies were classified intraoperatively, using the American Academy of Orthopaedic Surgeons Acetabular Deficiency Classification, as type IIIb (combined deficiency of the rim, the medial wall and the dome).¹⁰ The defect of the left medial wall was between 35 and 45 mm and the right medial wall between 20 and 30 mm (figure 2). If one was to use the Chandler-Penenberg’s acetabular deficiency classification system,¹¹ the right acetabular deficiency would be classified as V-A (medial wall perforation) and the left acetabular deficiency as V-C (medial wall perforation+superior rim+superior intra-acetabular defect). Only loose fibrous tissue was curetted off the medial walls.

The femoral heads were used as solid autografts. They were fashioned as plugs and were pressed into the medial wall openings. Conventional reamers, in forward mode, were used to shape the grafts in order to accept the cemented cups. Corticocancellous bone from the femoral necks together with bone from the reamings were used and pressurised around the grafts in order to create a biologic buttress. Tripolar head articulations were used bilaterally which offered additional joint stability. These were deliberately implanted into the grafts without interposition of any mesh in 5° of retroversion. The level of the true acetabular floor was identified by visual inspection and was confirmed by intraoperative fluoroscopy during cementation. The femoral stems were also cemented into place due to osteoporosis, existing secondarily to the absence of load bearing. The implants used were the ImplantCast EcoFit stem and EcoFit 2M cup.

Routine postoperative care consisted of anticoagulation and a double dose of antibiotics due to the extended surgical time (5½ hours). Two units of blood were transfused perioperatively and postoperatively. The patient sat upright on the second postoperative day, and his hip flexion had immediately increased to 100°, and there was no need for abduction pillows. He spent the first 5 days in bed performing isometric quadriceps strengthening exercises, together with passive ROM hip exercises and electric muscle stimulation, starting from day 1.

Outcome and follow-up

On the 10th day, he was transferred to a rehabilitation centre. He mobilised from bed to chair during the first postoperative month. He could stand in a frame by the end of the third month and could walk with aids at the end of the fourth month. His torso and upper arms were rehabilitated accordingly. The equinus deformity of his left foot, due to the peroneal nerve palsy, was addressed with vigorous physiotherapy and a walking shoe with stirrup.

On repeat CT scan of the pelvis postoperatively (figure 3, figure 4), full reconstruction of the acetabular bed with presence of abundant callus was noted on both sides, both sockets were in place, just at the teardrop (the left side being initially slightly higher by 0.7 mm), with no acetabular migration noted. The trochanteric slide healed with fibrous union.

Figure 3.

Postoperative CT scan of the pelvis, frontal plane: The reconstructed acetabular beds with no migration and well-positioned implants and bilateral callus formation can be seen. The left-sided trochanteric slide has healed with fibrous union.

Figure 4.

Postoperative CT scan of the pelvis, coronal plane: full reconstruction of the acetabular bed with presence of abundant callus is noted on both sides.

At 24 months’ follow-up, both right and left hips had flexion of more than 140°, extension about 10°, adduction 35°, abduction 45°, internal and external rotation 20° and 90° accordingly. The HHS was 95, and the patient was walking aid free without Trendelenburg gait. At this point, bilateral acetabulums and the trochanteric slide were fully ossified (figure 5).

Figure 5.

Two-year follow-up with anteroposterior pelvis view. Note the bilateral fully ossified acetabular bed.

Discussion

Type of acetabular fixation remains a subject of debate in THA. A reliable cementless acetabular fixation requires an environment with adequate biologic potential. Mechanical stability is an essential prerequisite for bone ingrowth.¹² In this case, the size of defects in the non-united acetabular rims, the amount of bone loss of both medial walls and the potential of osteolysis due to an unstable cementless fixation had to be considered. Therefore, two cemented acetabular components were placed on a buttress of biologic grafts. Keeping in mind that there were enough bone and cartilage islands adjacent to both medial walls, putting the medial walls under optimal tension-stress would create a possibility of remodelling. In that respect, authors aimed at restoring bone and providing new bone for future reconstruction if needed, as the patient was quite young. Moreover, a systematic review and meta-analysis concluded that the available literature suggests that the fixation of cemented acetabular components is more reliable than that of cementless components beyond the first postoperative decade.¹³

Strain fields in bone tissue may stimulate biological bone resorption, compromising implant effectiveness and bone production. In a finite element analysis of stress distribution during acetabular reconstruction in protrusio acetabuli, the von Mises stress on the medial wall was found to be lower with metal-backed polyethylene acetabular components and higher with a medialised metal shell reinforcement of the medial wall.¹⁴ The stress on the medial wall was also high when antiprotrusio rings or cups were used (figure 6). What seems to contribute most to the biologic reconstruction of the bone of the medial wall and consequently to the longevity of the THA and the prevention of the cup migration is the restoration of the cup in the normal inferolateral position. This position serves to transfer the weight-bearing loads to the strong anterior and posterior columns.^{7 8 14} Additionally, it is shown that when the diameter of an implant is increased, within morphological range, it produces the least von Mises stress when loaded within the surrounding bone.¹⁵ Therefore, the best option was a large diameter, metal-backed polyethylene lined shell, lateralised and positioned at the level of the teardrop, as proximal as possible to its anatomic position. Using 50 mm metal backed polyethylene cups bilaterally, and pressuring the femoral head grafts against the acetabular defects, the medial acetabular walls were put under optimal tension-stress, thus any cartilage present underwent ossification, and the bone was encouraged to remodel.

Figure 6.

von Mises effective stress occurring within the medial part of the pelvic cortex and maximum tension occurring within the cement of a protruded acetabulum reconstructed with a normally placed acetabular component (reproduced from Crowninshield et al¹⁴).

In the previous finite model, the lowest tensile stress within the cement occurred with the metal-backed polyethylene component.¹⁴ The highest stress, which is probably responsible for the fragmentation of the cement mantle, occurred with the polyethylene-only component. Additionally, the incidence of cement–bone demarcation increased with incorrect positioning of the acetabular component (away from the anatomic position: too far medially and superiorly).^{8 16} As already discussed, additional metallic reinforcement of the medial wall as long as the anterior and posterior columns are intact, does not offer any biomechanical advantage. The presence of metalwork, on the contrary, increases the risk for loosening, infection, non-union and heterotopic ossification due to the extensive stripping of soft tissues.4 12 17 18

The best source of graft remains the autologous bone graft taken from the femoral heads, or the ilium if absent. Mendes et al have shown that bone graft covered with methyl methacrylate remains viable and maintains its osteogenic capacity.^{4 19} There is some controversy as to whether the graft should be cut into strips and impacted into the defect due to the potential risk of mechanical weakening of solid grafts from the creeping substitution phenomenon.^{4 20} Graft incorporation has been reported to happen from 3 to 10 months following surgery.^{4 20} Heywood has shown that in certain cases, the femoral head can be fashioned into a solid graft which is anchored against the medial wall as a plug, minimising the use of cement and making the use of wire mesh and restraining rings about the acetabulum unnecessary.²¹ Mesh has not shown to provide any advantage for graft incorporation or prevention of cup migration for this size and type of acetabular defect (severe combined deficiency).^{5 22} Authors have used Heywood’s technique and obtained full graft incorporation in less than 12 months postoperatively.

The main advantage of using a bipolar socket is the stability of the THA articulation, especially for cases where extensive capsular releases have to be performed.⁷ The main disadvantage of the bipolar articulation is the resultant shear at the outer bearing interface that would erode the bone graft and result in early migration of the socket. The solution to both previous problems was resolved with the use of tripolar articulations cemented into the femoral head grafts, following conventional reaming in both forward and reverse directions to pressurise and shape the grafts.²³

Another matter of importance is the fixation of the trochanteric osteotomy. With trochanteric slide osteotomies, the vastus lateralis remains attached to the trochanter which results in a compressive force across the osteotomy site and relatively small shear forces.²⁴ Approximation of the osteotomy can be done with sutures instead of wires and cables that are related to higher complication rates. Ranawat et al reported that 23% of standard trochanteric osteotomies resulted in non-union and reduced abductor power, however without any disabling symptoms.¹⁶ For this patient, authors chose not to fix the trochanteric slide because the abductors were very short and fibrotic due to the long-standing retraction and disuse. The patient developed a fibrous (12-month follow-up) then osseous (24-month follow-up) union across the osteotomy side, without experiencing pain or Trendelenburg gait.

There is no optimal surgical approach for THA in patients with acetabular protrusion. Authors chose to combine the direct anterior approach, without a traction table, combined with a trochanteric slide on the right side only because access to the hip joint was found easier, and avoided the need to perform a trochanteric osteotomy on the less-affected side. Furthermore, both sides and the abdomen (for vascular access if needed) could be prepared and draped simultaneously, resulting in a shorter operating time. The anterior approach also allows less blood loss and lowers the risk of heterotopic ossification and surgical trauma to the soft tissues.²⁵ The exposure to the neck of the femur and the performance of the double neck osteotomy was easier compared with the lateral and posterior approaches. In addition, access to the acetabulum was easier, and positioning of the cup was more accurate due to direct visualisation. The use of a standard fracture table made it easier to perform the capsular releases sequentially and improved control over the femoral manipulation, avoiding excessive traction (through a traction table) that could result in intraoperative fractures.

Learning points.

• Applying transdifferentiation principles and using the ossification potential of bone tissue, instead of state-of-the-art metal implants, was deemed as the optimal way to manage this case.

• Despite severe preoperative disability, biological reconstruction and rehabilitation enabled great improvement in hip function and quality of life.

• Although a rare and devastating complication, bilateral protrusio acetabuli can occur from seizures.