Although total hip arthroplasty is one of the most successful surgeries in orthopedics, the number of revision procedures is predicted to increase by 137% during the next 2 decades.1 Implant failure modes such as instability, infection, loosening, and wear are becoming more prevalent.2 Instability, infection, extensive bony defects, and soft tissue damage are the most important concerns and complications associated with revision surgery. More than 50% of revisions involve the acetabular component.2 Paprosky et al3 described a classification of acetabular defects that occur in cases of implant failure. Treating type 2 and type 3 uncontained defects can be technically challenging because the surgeon must use extensive reconstruction techniques to adequately restore the biomechanics of the hip, structural stability, and leg length. Furthermore, neurovascular structures can be in jeopardy when complex pelvic reconstructive procedures are being conducted.
To optimize access to the pelvic bone, minimize soft tissue damage, and protect the pelvic neurovascular structures, the authors used an extensile anterior approach to the acetabulum. This approach has been described by Ganz et al4 to conduct periacetabular osteotomies.5 This approach uses the Smith–Petersen interval and exposes the anterior column and the acetabulum along with its defects. To the authors' knowledge, this approach has not been used or described for complex reconstructive surgeries for extensive acetabular defects in total hip arthroplasty. The aims of this study were to describe a modified extensile surgical technique and to report its outcomes when applied to challenging acetabular defects and reconstructions. The authors hypothesized that it could provide excellent exposure and be safely performed.
Materials and Methods
Institutional review board approval was obtained for this study. Between April 2010 and January 2014, a total of 49 hips (48 patients) underwent reconstruction and arthroplasty through an extensile anterior approach to the acetabulum. During this time period, 426 revision hip arthroplasties were performed at the authors' institutions, with this being the preferred approach. The posterior approach was used in 4.3% of revisions (eg, when posterior hardware removal or reconstruction of the posterior column was indicated). The anterior approach was used for the exposure and reconstruction in the remainder of the hip revisions. The Paprosky classification was used to define each case's acetabular defect.3 There were 5 Paprosky type 1, 12 type 2A, 11 type 2B, 7 type 2C, 11 type 3A, and 3 type 3B acetabular defects. Intraoperative assessments of bone deficiency and rim support were correlated with preoperative radiographic evaluation (Table 1). Two of the authors (J.T.M., K.C.) also used preoperative computed tomography scans when posterior deficiency or hardware from previous surgeries was a concern. Of the 5 Paprosky type 1 acetabular defects, 3 had a preoperative diagnosis of traumatic osteoarthritis that developed after previous open reduction and internal fixation via an ilioinguinal approach to the acetabulum (Figure 1), 1 had a previous hip fusion preformed through an anterior approach (Figure 2), and 1 had a previous en bloc resection of a chondrosarcoma of the pelvis and acetabulum with a residual defect (Figure 3). There were 23 female and 26 male patients with a mean age of 71 years (range, 42–90 years). All patients were followed at 2 weeks, 6 weeks, 3 months, 1 year, and then annually. Harris hip and Western Ontario and McMaster Universities Osteoarthritis Index scores were recorded preoperatively and at final follow-up. Preoperative diagnoses included isolated aseptic loosening of the acetabular component (N=23), aseptic loosening of both the acetabular and femoral components (N=12), infection (N=9) treated with 2-stage revision arthroplasty, and other (N=5) (eg, progressive traumatic osteoarthritis following open reduction and internal fixation of an acetabular fracture, take down of previous hip fusion, and previous tumor resection). During primary surgeries, a variety of approaches were used: lateral (N=8), ilioinguinal (N=3) (2 cases of open reduction and internal fixation of an acetabular fracture and 1 en bloc resection of a chondrosarcoma), anterior (N=1) (Smith–Peterson for hip arthrodesis), and posterior lateral (N=37).
Paprosky Classification of Preoperative Acetabular Defects
Anteroposterior radiograph of a 60-year-old woman after open reduction and internal fixation of a comminuted left anterior column and wall acetabular fracture performed via an ilioinguinal approach progressed to advanced traumatic osteoarthritis (A). Anteroposterior radiograph 5 years postoperatively (B).
Anteroposterior radiograph of a 45-year-old man 25 years after open arthrodesis of the right hip performed via an anterior approach with debilitating low back pain (A). Anteroposterior radiograph 3 years postoperatively (B).
Anteroposterior radiograph of a 34-year-old female patient after en bloc resection of a chondrosarcoma of the pelvis and acetabulum. Arrows indicate defect (A). Computed tomography scan (B) and 3-dimensional reconstruction (C) showing residual acetabular and pelvic defect. Intraoperative view prior to femoral head resection (D). Acetabular cup placement and fixation (E). Modular augment fixation (F). Unitizing cup and augment with bone cement (G). Final reduction and construct (H). Anteroposterior radiograph 4 years postoperatively (I).
With the technique used in this series, the patient is placed in the supine position on either a regular operating room table or a special table, depending on surgeon preference. Spinal or general anesthesia is administered. A modified Smith–Petersen approach to the hip is used for exposure.4,5 The incision is started along the iliac crest, over the anterior superior iliac spine (ASIS), and is directed distally over the tensor fascia lata (TFL) (Figure 4). Subcutaneous flaps are raised medially and laterally, with care taken to avoid injuring the lateral femoral cutaneous nerve. The nerve remains included in the soft tissue sleeve that is created. The abdominal muscles along with the inguinal ligament and the sartorius are released from the ASIS. The nerve is included in this sleeve, and thus traction on the nerve is minimized. The TFL fascia is incised and peeled off the TFL fibers. The interval between the TFL and the rectus femoris is identified, and the lateral circumflex vessels are coagulated. Proximally, the aponeurosis of the external oblique muscle is subperiosteally peeled off the iliac crest and reflected medially along with the oblique abdominal muscles. The aponeurosis of the sartorius and the inguinal ligament is then peeled off the ASIS. The hip is slightly flexed, and the medial muscle envelope is lifted off the inner iliac table with a Hohmann retractor that subperiosteally rests on the pelvic brim; in cases of an extensive medial defect, it rests on the inner surface of the sciatic spine. The iliopsoas muscle is thus retracted medially. The anterior inferior iliac spine and the rectus femoris are identified (Figure 4). The interval between the iliopsoas medially and the insertion of the rectus femoris and iliocapsularis laterally is identified and opened. A Langenbeck retractor is used to lift the iliopsoas off the iliopectineal eminence, and a sharp-tipped Hohmann retractor is placed medially to the eminence. The tip of this retractor is fixed into the pubic bone to safely retract the psoas medially (Figure 5). If the view on the anterior column is insufficient, the rectus femoris tendon can be tenotomized as originally described by Ganz et al.4,5 In 35 of the hips in this series, the rectus femoris tendon was tenotomized to improve exposure and allow the desired reconstruction. In these cases, the tendon was sutured back at the end of the procedure. The hip capsule is then incised. The superior part of the capsule can be removed if it is hindering visualization. The pubo-femoral ligament (ie, the inferior capsule) is tagged and retracted inferiorly by a posteroinferior retractor. The superior capsular release is done with the hip located. The femur is then dislocated, which is followed by dislodging the femoral head component in isolated acetabular revisions, extracting the femoral component in reconstructions requiring revision of both the femoral and the acetabular components, or performing the femoral neck cut at the desired level in a primary arthroplasty. The femur is then lifted to the level of the TFL. A superior retractor is placed at the level of the superior release just in front of the gluteus minimus. The femoral exposure is achieved either by external rotation, extension, and adduction when using a special table or by placing the ipsilateral leg underneath the contralateral leg when using a regular operating table (Figure 6). Additional femoral exposure can be obtained, if desired, with 15° of hyperextension. In the 21 hips for which the femoral stem also had to be revised, the anterior 1 to 2 cm of the TFL origin at the level of the ASIS was gently released to prevent the TFL from tearing. Attention is then directed toward socket removal and debridement of the bony defects.
The modified extensile Smith–Petersen approach to the anterior column of a right hip is shown on a cadaver specimen. The iliac crest and anterior superior iliac spine (ASIS) (¶) are identified. The incision runs over the lateral side of the crest, toward the ASIS, and then distally over the muscle belly of the tensor fascia lata (°) (A). The ASIS is identified after the abdominal muscles have been peeled off subperiosteally. Similarly, the insertion of the inguinal ligament is subperiosteally peeled off (B). With the hip flexed, a Hohmann retractor is placed at the inner table of the ileum underneath the iliacus muscle (★) (C). The retractor is placed at the pelvic brim, and the interval between the iliacus–iliopsoas medially and the rectus femoris direct head laterally is opened (arrow) (D).
The anterior superior iliac spine (¶) and the anterior inferior iliac spine (^) are identified. The iliopsoas (arrow) is retracted medially. The hip remains flexed (A). The pubic eminence (★) is identified and marks the medial extension of the acetabulum (B). A sharp retractor is fixed in the pubic bone, just medial to the eminence. The psoas and the neurovascular structures are retracted medially (C).
Anteroposterior radiograph of an 83-year-old female patient who presented with a severe acetabular defect and a peri-Vancouver type B3 periprosthetic femoral fracture 20 years following a cemented hip replacement (A). A retractor was placed at the ischial spine and at the pubis. The posterior column was intact, but the anterior column and the medial wall were completely disrupted (B). A lateral augment was applied to provide support to the trabecular revision shell (C). A cage was applied to provide support (D), and a liner was cemented in (E). The incision was extended distally, cerclage wires were applied around the femur, and a modular revision stem was inserted (F). Postoperative anteroposterior radiograph (G).
Overall reconstruction constructs (Table 2) included highly porous hemispherical acetabular components being used in 40 cases, along with particulate graft in 12 cases and titanium augments in 19 cases. A cage was used in 9 hips. In these cases, a slot was created in the ischium for anchorage of the distal flange of the cage, and the anterior insertion of the gluteus minimus was subperiosteally peeled off the outer table of the ilium to allow fixation of the proximal flange of the cage to the ilium (Figure 6). The polyethylene liner was cemented in either the revision shell or the cage, which was placed over the revision shell (Figure 7).
Surgical Reconstruction Constructs
Anteroposterior radiograph (A) of a 45-year-old female patient who presented with a Paprosky type 2 lateral defect of the left hip (B). The intraoperative view of the cup-augment trials (C). The lateral augment was inserted and loosely fixed (D). The revision shell was inserted and the augment was fixed (E). A trial reduction was done and allowed for extensive stability testing (F). Anteroposterior radiograph 4 years postoperatively (G).
Testing for stability or impingement was performed in deep flexion with 30° of internal/external rotation as well as hyperextension with external rotation. In cases for which a suture repair of the rectus femoris tendon was performed at the end of the procedure, this was followed by transosseous suturing of the muscle insertions at the level of the anterior inferior iliac spine. The oblique abdominal muscle is sutured onto the conjoined fascia of the TFL and the maximus at the level of the iliac crest. The TFL fascia is then closed. A drain is placed and subsequently removed within 24 to 48 hours. Patients were immediately mobilized, with partial weight bearing (50% to 75%) for 6 weeks. Open chain exercises were prohibited to allow the rectus femoris tendon to heal. Antibiotics were started immediately preoperatively and continued for 24 hours postoperatively (intravenous cefazolin).
Clinical and radiographic assessment was performed at 2 weeks, 6 weeks, 3 months, 1 year, and 2 years and at final follow-up. The components were deemed to be loose when a radiolucent line of greater than 2 mm was present or if the socket had migrated during the follow-up evaluation. The restoration of the biomechanics was measured via anteroposterior pelvic radiograph. The perpendicular distance to the inter-teardrop line and Kohler's line was measured and compared with the contralateral side. Similarly, leg lengthening was determined as the distance of the lesser trochanters relative to the inter-teardrop line or the line connecting both ischial tuberosities in cases where the teardrop was absent.
For statistical analysis, the Student's t test was used to compare functional outcomes pre- and postoperatively. SPSS Statistics Desktop version 23 software (IBM, Armonk, New York) was used.
Five patients died prior to the 2-year follow-up, with death being unrelated to the index operation. Thus, 44 hips (43 patients) with a minimum follow-up of 2 years remained. Mean follow-up was 40 months (range, 24–60 months). Postoperative radiographic analysis revealed that, compared with the contralateral hip, the mean horizontal and vertical centers of rotation were restored by 4 mm each to Kohler's line (range, −8 to +12 mm) and the inter-teardrop line (range, −8 to +15 mm). The mean leg length discrepancy was −1.3 mm (range, −9 to +9 mm). Harris hip score and all Western Ontario and McMaster Universities Osteoarthritis Index sub-scores had improved significantly (Table 3). Clinical function of the rectus femoris was determined by having patients extend the knee with the hip flexed with manual resistance. Additionally, the hip was flexed in the supine position against resistance with the knee extended and patients were requested to squat. In all cases, the function of the rectus femoris was clinically normal 3 months postoperatively
Overview of Pre- and Postoperative Functional Scores
There were no intraoperative complications. Ten hips (20%) had postoperative complications; 8 occurred within the first 3 months and 2 occurred between 3 and 12 months (Table 4). Three hips (10%) of the 30 type 2 defects had complications: 2 had an early superficial wound dehiscence that required debridement and 1 had a dislocation that was treated by closed reduction with bracing. Seven hips (50%) of the 14 type 3 defects had complications. Three required reoperation: 1 wash out for an acute deep infection in a Paprosky type 3A defect and a revision to a constrained liner for recurrent instability in 2 of the 3 type 3B defects. Of the remaining 4 complications in the type 3 defects, 3 cases had a dislocation that was successfully treated with closed reduction and in 1 type 3A case, the socket had initially migrated for 5 mm but the patient was pain free and the socket was radiographically stable at latest follow-up. There were no complications in the type 1 defects. At minimum 2-year follow-up, there were no additional complications and all implants were radiographically stable.
Complications by Defect Type
The all-cause reoperation rate was 10% (N=5), all within the first 12 months postoperatively. In 2 of the Paprosky type 2 cases, there was an early superficial wound dehiscence that required debridement. There was 1 wash out for a deep infection in a Paprosky type 3A case, and 2 of the 3 type 3B cases required a revision to a constrained liner for recurrent instability.
Dislocation accounted for 60% (6 of 10) of all complications. Five (83%) of 6 dislocations occurred in patients with a type 3 defect. There were 2 anterior and 4 posterior dislocations. Four dislocations were successfully treated with closed reduction. Two dislocations occurred in 1 patient with a type 3A defect on the left side and a type 3B defect on the right side. Preoperatively, she had bilateral resection Girdlestone arthroplasties because of recurrent infection in both hips and an absent abductor musculature on the right. The right hip with the type 3B defect had a cup-cage reconstruction that was subsequently converted to a constrained liner for recurrent dislocations. The other patient with a type 3B defect had a history of preexisting instability prior to the index revision procedure. In both patients with a type 3B defect, the abductor mechanism had been significantly disrupted or was deficient due to multiple transgluteal surgeries.
The Smith–Petersen and Ganz approaches have been mainly described for pelvic fracture treatment, periacetabular osteotomies, or primary total hip arthroplasty. To the authors' knowledge, this is the first report of a modification of the approach for complex revision surgery of the acetabulum. The approach allows excellent access to the anterior column and the inner table of the ilium. From their experience with periacetabular osteotomy and primary direct anterior total hip arthroplasty, the authors hypothesized that this approach had some features that would facilitate complex reconstructive surgery of the pelvis.4 During these complex procedures, it can be difficult to fully appreciate the extent of the bony defects and their relationship to the intrapelvic neurovascular structures. Theoretically, these concerns would be offset if the muscle envelope around the pelvis was extensively freed to fully access and assess the pelvic and acetabular bony defects from anterior. The intact muscular tissue envelope helps to protect the neurovascular structures. This technique allows for the reconstruction of complex acetabular defects in a supine position, and theoretically decreases jeopardizing the neurovascular structures. Finally, the procedure can be performed through a soft tissue plane that most likely has not been previously violated. This may minimize additional soft tissue trauma for patients with multiple surgeries.
In the initial description of the surgical technique of the periacetabular osteotomy, Ganz et al4 described an osteotomy of the ASIS along with a tenotomy of the direct head of the rectus femoris.5 The current authors slightly modified this technique by subperiosteally peeling the soft tissues off the pelvis and only performing a tenotomy of the rectus tendon when additional visualization of the anterior column was insufficient. Similar to the periacetabular osteotomy technique, a retractor is placed on the sciatic spine and in the pubic bone. This allowed for an excellent exposure of the anterior column, which was often disrupted in the current series of cases. The remaining muscle envelope is left intact and not mobilized unless additional assessment of the extent of the defect and the residual bone was required. This facilitated safe and accurate reaming of the acetabulum. In addition, direct visualization for screw insertion was simplified, especially when multiple screws in multiple directions had to be used.
No intraoperative complications were encountered. On clinical examination at 3 months, functional testing of the rectus femoris yielded normal results. Except for 2 cases of superficial wound breakdown, there were no complications directly related to this specific surgical approach. Furthermore, the approach allowed for an accurate restoration of leg length and hip biomechanics, with the center of rotation within the range of 5 mm compared with the contralateral side in 63% of the cases.
The all-cause reoperation rate of 10% is within previously reported ranges.6–11 Dislocation was the most common complication, accounting for 60% (6 of 10) of all complications. It was initially theorized that an anterior approach for this complex revision surgery would minimize the risk of dislocation. A detailed analysis of the results allowed the authors to draw several conclusions. First, similar to the more traditional approaches, the dislocation rate is directly influenced by the complexity of the revision procedure and the associated soft tissue envelop deficiencies. Among 30 type 2 defects, only 1 patient (3.3%) had a postoperative dislocation. Five (83%) of 6 dislocations occurred in patients with type 3 defects. Four of the 6 dislocations were successfully treated with closed reduction and bracing. Two dislocations occurred in 1 patient with a type 3A defect on the left side and a type 3B defect on the right side. She previously had bilateral resection Girdlestone arthroplasties because of recurrent infections and an absent abductor mechanism on the right. The right hip with the type 3B defect had a cup-cage reconstruction that was subsequently converted to a constrained liner for recurrent dislocations. The other patient with a type 3B defect had a history of preexisting instability prior to the revision procedure. In both patients with a type 3B defect, the abductor mechanism had been significantly disrupted or was deficient due to multiple transgluteal surgeries. Excluding the type 3 defect reconstructions, there was a dislocation rate of 2.8% (1 of 35 cases). Second, the authors now believe that, in addition to the extent of the bony defect, a deficient soft tissue envelope around the hip secondary to previous surgeries is a strong predictor of postoperative dislocation and instability. In 4 of 6 dislocations, the patients had multiple previous surgeries, with 1 patient having bilateral resection Girdlestone arthroplasties for several years. The authors now believe that it is better to simultaneously incorporate a constrained liner or a dual mobility socket in the reconstruction if intraoperative testing of stability confirms instability, especially in Paprosky type 3 defects. Also, they no longer osteotomize the ASIS. Finally, the current series represents the initial learning curve of the authors.
Except for 1 socket that showed 5 mm of migration following medial impaction grafting, all of the other sockets were deemed radiographically stable, having bony ingrowth on all of the postoperative follow-up radiographs.
One additional potential advantage of performing the procedure with the patient in the supine position is that fluoroscopy can be easily used intraoperatively to assess component position, limb length, center of rotation, offset, and restoration of hip biomechanics, as well as the fit and fill of both the acetabular and the femoral components. This provides an opportunity to make any desired modifications or adjustments in real time before completing the surgical procedure. The authors think that confirmation with intraoperative imaging better helps achieve preoperative goals and avoids outliers, which may increase complications and compromise results.
The modified extensile anterior approach to the acetabulum and pelvis is safe and allows for excellent exposure and successful reconstruction of both acetabular and pelvic bony defects. The exposure is less successful in addressing instability caused by extreme soft tissue deficiency with an absent or significantly deficient abductor mechanism or soft tissue constraints, with a re-revision rate of 4% to a constrained liner for recurrent instability. Seventy percent of all complications and 83% of all dislocations occurred in the Paprosky type 3 defect group. To decrease complications in patients with a severe acetabular defect and an associated abductor mechanism deficiency, the authors recommend consideration of using either a constrained liner or a dual mobility socket during the reconstruction, to better address both bony defects and soft tissue deficiencies simultaneously.
- Kurtz S, Ong K, Lau E, Mowat F, Halpern M. Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to 2030. J Bone Joint Surg Am. 2007; 89(4):780–785.
- Bozic KJ, Kurtz SM, Lau E, Ong K, Vail TP, Berry DJ. The epidemiology of revision total hip arthroplasty in the United States. J Bone Joint Surg Am. 2009; 91(1):128–133. doi:10.2106/JBJS.H.00155 [CrossRef]
- Paprosky WG, Perona PG, Lawrence JM. Acetabular defect classification and surgical reconstruction in revision arthroplasty: a 6-year follow-up evaluation. J Arthroplasty. 1994; 9(1):33–44. doi:10.1016/0883-5403(94)90135-X [CrossRef]
- Ganz R, Klaue K, Vinh TS, Mast JW. A new periacetabular osteotomy for the treatment of hip dysplasias: technique and preliminary results. Clin Orthop Relat Res. 1988; 232:26–36.
- Leunig M, Siebenrock KA, Ganz R. Rationale of periacetabular osteotomy and background work. Instr Course Lect. 2001; 50:229–238.
- Borland WS, Bhattacharya R, Holland JP, Brewster NT. Use of porous trabecular metal augments with impaction bone grafting in management of acetabular bone loss. Acta Orthop. 2012; 83(4):347–352. doi:10.3109/17453674.2012.718518 [CrossRef]
- Sporer SM, Paprosky WG. The use of a trabecular metal acetabular component and trabecular metal augment for severe acetabular defects. J Arthroplasty. 2006; 21(6)(suppl 2):83–86. doi:10.1016/j.arth.2006.05.008 [CrossRef]
- Sporer SM, Paprosky WG. Acetabular revision using a trabecular metal acetabular component for severe acetabular bone loss associated with a pelvic discontinuity. J Arthroplasty. 2006; 21(6)(suppl 2):87–90. doi:10.1016/j.arth.2006.05.015 [CrossRef]
- Sternheim A, Backstein D, Kuzyk PR, et al. Porous metal revision shells for management of contained acetabular bone defects at a mean follow-up of six years: a comparison between up to 50% bleeding host bone contact and more than 50% contact. J Bone Joint Surg Br. 2012; 94(2):158–162. doi:10.1302/0301-620X.94B2.27871 [CrossRef]
- Van Kleunen JP, Lee GC, Lementowski PW, Nelson CL, Garino JP. Acetabular revisions using trabecular metal cups and augments. J Arthroplasty. 2009; 24(6)(suppl):64–68. doi:10.1016/j.arth.2009.02.001 [CrossRef]
- Abolghasemian M, Tangsataporn S, Sternheim A, Backstein D, Safir O, Gross AE. Combined trabecular metal acetabular shell and augment for acetabular revision with substantial bone loss: a mid-term review. Bone Joint J. 2013; 95-B(2):166–172. doi:10.1302/0301-620X.95B2.30608 [CrossRef]
Paprosky Classification of Preoperative Acetabular Defects
|Paprosky 2A (superior)||12|
|Paprosky 2B (superolateral)||11|
|Paprosky 2C (medial)||7|
|Paprosky 3A (discontinuity superolateral, Kohler's line intact)||11|
|Paprosky 3B (discontinuity superomedial, Kohler's line disrupted)||3|
Surgical Reconstruction Constructs
|Impaction grafting (particulate grafting) allograft||12|
|Augments (trabecular metal)||19|
Overview of Pre- and Postoperative Functional Scoresa
|Time||Mean Score (Range)|
|Western Ontario and McMaster Universities Osteoarthritis Index||Harris Hip|
|Preoperative||38 (20–60)||52 (25–75) 37 (22–57)||40 (24–58)||41 (11–73)|
|Final follow-up||82 (50–100)||86 (50–100) 70 (49–100)||73 (54–89)||80 (53–98)|
Complications by Defect Type
|Defect Type||No. of Patients||No. of Complicationsa||Overall Complication Rate|
|Paprosky 2||30||3 (1 closed reduction; 2 reoperations [superficial wound debridements])||30%|
|Paprosky 3||14||7 (1 cup migration [stable]; 7 closed reductions; 3 reoperations [1 deep wash out and 2 revisions because of constrained liners])||70%|