Periacetabular Osteotomy (Ganz/Bernese PAO)

The Bernese periacetabular osteotomy (PAO) is a joint-preserving realignment procedure that reorients the acetabulum to improve femoral head coverage in patients with symptomatic acetabular dysplasia. Developed by Reinhold Ganz and colleagues in Bern, Switzerland, and first described in 1988, the PAO has become the gold standard surgical treatment for acetabular dysplasia in the young adult with a preserved, congruent joint.

The fundamental principle: dysplasia results in an acetabulum that is too shallow and/or too anteverted, providing inadequate coverage of the femoral head. This concentrates contact pressures on a small area of articular cartilage, leading to labral overload, chondral damage, and premature osteoarthritis. By reorienting the acetabulum, PAO redistributes forces over a larger articular surface area, relieves pain, and aims to delay or prevent the onset of osteoarthritis.

Key advantages over earlier osteotomies (triple, spherical, rotational):

• Posterior column remains intact - preserving pelvic ring stability and allowing early mobilisation
• Large corrections possible in all three planes (lateral coverage, anterior coverage, version)
• Preserved vascularity of the acetabular fragment via the inferior gluteal artery
• No external immobilisation required (unlike some triple osteotomies)
• Spherical acetabulum maintained - the congruent joint remains congruent

Indications:

• Symptomatic acetabular dysplasia (groin pain, activity-related pain, limp) despite conservative management (activity modification, physiotherapy, NSAIDs)
• Age typically 15-45 years (some centres extend to 50-55 in suitable candidates)
• LCEA less than 20 degrees (dysplastic range) on well-centred AP pelvis radiograph
• Congruent joint - the femoral head reduces into the acetabulum on functional abduction views
• Preserved joint space (Tonnis grade 0 or 1)
• Good range of motion (flexion greater than 90 degrees, minimal fixed deformity)
• Skeletally mature (or near-mature; combined with other procedures in the skeletally immature)
• Realistic expectations and motivation for rehabilitation

Relative indications:

• Borderline dysplasia (LCEA 18-25 degrees) with labral pathology requiring bony correction
• Mild coxa valga amenable to combined PAO and varus derotational femoral osteotomy (VDRO)
• Residual dysplasia after prior childhood pelvic osteotomy (salvage PAO, technically demanding)

View larger

Acetabular anatomy and dysplasia:

• The acetabulum is a hemispherical socket formed by the confluence of the ilium, ischium, and pubis at the triradiate cartilage (fused in adults). In dysplasia, the socket is shallow (under-coverage) and often excessively anteverted.
• The lateral centre-edge angle of Wiberg (LCEA) measures lateral acetabular coverage on an AP pelvis radiograph. Normal is 25-39 degrees; less than 20 degrees is dysplastic; less than 0 degrees indicates severe dysplasia with subluxation.
• The Tonnis acetabular index (acetabular roof angle) measures the slope of the weight-bearing acetabular surface. Normal is 0-10 degrees; greater than 10 degrees indicates acetabular dysplasia (the sourcil slopes upward rather than being horizontal).
• The anterior centre-edge angle (ACEA) of Lequesne and de Seze measured on the false-profile view assesses anterior coverage. Normal is approximately 25 degrees or greater; less than 20 degrees indicates anterior deficiency.

Biomechanics of dysplasia:

• In a dysplastic hip, the contact area between the femoral head and acetabulum is reduced, leading to elevated contact pressures per unit area of articular cartilage. Pauwels calculated that acetabular dysplasia increases resultant joint forces by reducing the weight-bearing surface.
• Labral overload occurs as the labrum becomes the primary restraint to femoral head subluxation, leading to labral tears, degeneration, and ganglion formation.
• The femoral head may be eccentrically loaded anterosuperiorly, producing a characteristic pattern of anterosuperior chondral wear seen arthroscopically and on MRI.
• Over time, this mechanical overload initiates a degenerative cascade: labral failure, chondral delamination, osteophyte formation, and progressive osteoarthritis.

Vascular anatomy critical to PAO:

• The inferior gluteal artery (branch of the internal iliac artery) is the primary blood supply to the mobilised acetabular fragment, entering via the short external rotators and the posterior soft-tissue pedicle. Preservation of this pedicle during the posterior column cut is essential to prevent AVN.
• The superior gluteal artery exits the pelvis via the greater sciatic notch above piriformis and supplies the gluteus medius/minimus; it must be protected during the iliac cut.
• The medial femoral circumflex artery (MFCA) is the primary blood supply to the femoral head; it is not typically at risk during PAO but must be considered if concurrent femoral osteotomy is performed.

Radiographic parameters:

• LCEA (Wiberg angle): angle between a vertical line through the centre of the femoral head and a line to the lateral edge of the acetabular sourcil on AP pelvis. Target after PAO: 25-35 degrees.
• Tonnis angle (acetabular index): angle between a horizontal line and a line from the medial to the lateral edge of the acetabular sourcil on AP pelvis. Target after PAO: 0-10 degrees.
• ACEA (Lequesne): measured on the false-profile (Lequesne) view. Target: greater than 25 degrees.
• Shenton's line: should be intact (or nearly so) in a congruent dysplastic hip; broken Shenton's line suggests subluxation.
• Extrusion index: percentage of the femoral head uncovered by the acetabulum; less than 25% is normal.
• Femoroacetabular impingement (FAI) risk: overcorrection of LCEA to greater than 40 degrees or excessive retroversion creates iatrogenic FAI, which is increasingly recognised as a cause of poor outcome after PAO.

The Bernese PAO is performed through one of two main approaches:

Modified Smith-Petersen (Hueter) approach - most commonly used:

• Interval: between the tensor fasciae latae (TFL) (superior gluteal nerve) laterally and the sartorius (femoral nerve) medially - the classic Smith-Petersen internervous plane.
• The lateral femoral cutaneous nerve (LFCN) crosses this interval variable distances below the anterior superior iliac spine (ASIS) and is at risk; it should be identified and protected or intentionally sacrificed if it obstructs the exposure.
• The direct head of rectus femoris arises from the anterior inferior iliac spine (AIIS); in the classic approach it is released or detached to access the anterior column and superior pubic ramus. In the rectus-sparing modification, the rectus is left attached and the exposure proceeds medial to it.
• The hip joint capsule is exposed anteriorly and medially; a limited capsulotomy may be performed for intra-operative arthrotomy to inspect the joint, though many surgeons perform this arthroscopically before the open procedure.
• The iliopsoas tendon is retracted medially; the femoral nerve and vessels lie further medial within the femoral triangle and must be protected.

Rectus-sparing (Roberge) / intra-pelvic approach:

• This approach accesses the inner table of the ilium and the true pelvis through a medial subperiosteal dissection, preserving the direct head of rectus femoris and minimising abductor disruption.
• The interval is developed between the iliacus (femoral nerve) medially and the gluteal muscles (superior gluteal nerve) on the outer table.
• All four osteotomy cuts are performed from within the pelvis (intra-pelvic), which reduces the risk of intra-articular extension but limits direct visualisation of the lateral ilium.
• The LFCN is still at risk during the superficial dissection.

• Position: supine on a radiolucent table. The affected leg is draped free to allow full hip range of motion. Some surgeons prefer the lateral decubitus position, but supine is more common for the modified Smith-Petersen approach and facilitates fluoroscopic imaging.
• Anaesthesia: general anaesthesia with or without regional supplementation (lumbar plexus block). Epidural or spinal anaesthesia may be used. Nerve monitoring (sciatic nerve) is employed by some surgeons but is not universal.
• Fluoroscopy: essential. The image intensifier is positioned to obtain AP pelvis, obturator oblique (Judet), and iliac oblique views. Intra-operative assessment of correction requires a true AP pelvis view (symmetrical obturator foramina, coccyx aligned with symphysis) and a false-profile equivalent.
• Pre-operative planning: a well-centred AP pelvis radiograph, false-profile (Lequesne) view, and Dunn/lateral views are mandatory. CT with 3D reconstruction is increasingly used for precise planning of correction in all three planes. The planned correction is templated to achieve LCEA 25-35 degrees, Tonnis angle 0-10 degrees, and ACEA greater than 25 degrees.
• Blood conservation: PAO involves significant blood loss (average 500-1500 mL). Cell saver autotransfusion, tranexamic acid, and controlled hypotensive anaesthesia are commonly employed. Type and crossmatch is mandatory (2-4 units).
• Prophylaxis: pre-operative antibiotics (cefazolin or alternative), thromboprophylaxis (LMWH or mechanical), and heterotopic ossification prophylaxis (indomethacin 75 mg daily for 2-6 weeks) per surgeon preference.

PAO is a major reconstructive procedure with a significant complication profile. Understanding structures at risk is essential for safe execution and for exam answers.

Nerve injuries:

• Lateral femoral cutaneous nerve (LFCN): the most commonly injured nerve (10-30%); traverses the TFL-sartorius interval. Most result in anterolateral thigh numbness or dysaesthesia; some resolve, others are permanent.
• Sciatic nerve: the most feared nerve injury (1-3%); at risk during the ischial and posterior column cuts, which are performed close to the nerve as it exits the greater sciatic notch. Meticulous subperiosteal dissection and use of Ganz osteotomes (which deflect away from soft tissues) reduce risk. Post-operative sciatic nerve palsy may also result from haematoma compression or excessive limb lengthening.
• Femoral nerve: at risk during medial retraction of the iliopsoas and exposure of the anterior column; neuropraxia is rare but reported.
• Obturator nerve: at risk during the superior pubic ramus cut as it traverses the obturator foramen; protected by subperiosteal dissection.

Other complications:

• Intra-articular extension: the osteotomy enters the hip joint if cuts are placed too close to the acetabulum; reported in approximately 2-5% of cases. Risks chondral damage and post-traumatic arthritis. Requires intra-operative recognition and management.
• Avascular necrosis (AVN) of the acetabular fragment: rare (less than 2%) but catastrophic; caused by disruption of the inferior gluteal artery pedicle during the posterior column cut. The entire acetabular fragment becomes avascular, leading to collapse and necessitating THA.
• Nonunion or delayed union: reported in approximately 2-5% of cases, most commonly at the ischial or pubic osteotomy sites. Smoking is a significant risk factor. Symptomatic nonunion may require bone grafting or revision fixation.
• Heterotopic ossification (HO): reported in 5-20% of cases (most Brooker grade 1-2 and asymptomatic). Prophylaxis with indomethacin or single-dose radiation is used in many centres. Significant HO (grade 3-4) may restrict motion and require excision.
• Loss of correction: migration of the acetabular fragment post-operatively if fixation is inadequate; more common with screw-only fixation than plate-and-screw constructs.
• Overcorrection and iatrogenic FAI: excessive lateral coverage (LCEA greater than 40 degrees) or excessive retroversion creates pincer-type FAI, causing impingement pain and progressive labral damage. This is increasingly recognised as a cause of persistent post-operative pain.
• Wound complications: infection (approximately 1-3%), haematoma, wound dehiscence, and prominent hardware are reported. The anterior approach places hardware in a subcutaneous position.
• Thromboembolic events: DVT/PE risk is approximately 1-3%; routine thromboprophylaxis is recommended.
• Blood loss: average 500-1500 mL; transfusion rates vary from 15-50% depending on technique and blood conservation strategies.

PAO needs no postoperative external immobilisation (a key advantage of the intact posterior column), but protected weight-bearing is essential while the four osteotomies unite.

• Weeks 0-6 (protection): toe-touch or partial weight-bearing (typically around 15-20 kg) with two crutches to protect the ischial and pubic cuts; isometric and gentle active-assisted hip range-of-motion exercises, avoiding active hip flexion against resistance (protecting the rectus/iliopsoas and the anterior fixation). Continue thromboprophylaxis and heterotopic-ossification prophylaxis per unit protocol.
• Weeks 6-12 (progression): confirm early osseous union radiographically, then progress to full weight-bearing over 2-4 weeks and wean the crutches. Begin abductor strengthening (gluteus medius is the key weak link after the approach) with core and gait re-education.
• Months 3-6: restore symmetric strength and a normal gait; low-impact conditioning (cycling, swimming, hydrotherapy) before any impact loading.
• Months 6-12: graded return to running and sport once radiographic union, full pain-free range and near-symmetric abductor strength are achieved.

There is no single universal guideline for the management of acetabular dysplasia; practice follows the evidence-based principles of patient selection and correction targets described above. Several international societies have published consensus statements and clinical practice guidelines:

Key guidance principles:

• AAOS (US): the AAOS clinical practice guideline on the treatment of osteoarthritis of the hip (non-arthroplasty) gives a moderate-strength recommendation for periacetabular osteotomy in symptomatic acetabular dysplasia with preserved joint space. The AAOS emphasises shared decision-making and referral to high-volume centres.
• BOA/BSR (UK): national guidance supports referral to specialist young adult hip services for consideration of PAO in symptomatic dysplasia. The Getting It Right First Time (GIRFT) programme in the NHS recommends centralisation of PAO to high-volume units.
• EFORT/European consensus: European practice has strongly favoured PAO since its development in Bern. The European Hip Society has published position statements supporting PAO as the gold standard joint-preserving procedure for acetabular dysplasia, with an emphasis on early intervention before arthritis develops.

Registry evidence:

• Large national and multicentre registries are increasingly capturing PAO data. The Danish Hip Arthroplasty Register and the Scandinavian PAO collaboration have provided real-world survivorship data, confirming approximately 75-85% 10-year survivorship free of THA in well-selected cohorts.
• Registry data consistently demonstrate that surgeon volume matters: outcomes are better and complication rates lower in centres performing more than 20-30 PAOs per year.

Global practice variation:

• European centres (particularly Bern, Boston/Copenhagen collaboration, and several German and French centres) have the largest PAO experience globally. The technique was developed in Europe and has been refined over three decades.
• North American adoption has increased significantly since the 2000s, with dedicated young adult hip fellowship training programmes producing surgeons skilled in PAO. However, access remains limited in many regions, with long waiting lists at high-volume centres.
• Asian centres (particularly Japan and Korea) have extensive experience with PAO for developmental dysplasia, which has higher prevalence in many Asian populations.
• In low- and middle-income settings, PAO may be the only joint-preserving option available to young patients with dysplasia, but the technical demands of the procedure limit its availability to major teaching hospitals.

Conversion to THA after PAO:

• Approximately 10-25% of PAO patients require THA by 10-20 years, depending on pre-operative arthritis grade.
• THA after PAO is technically more demanding than primary THA due to altered acetabular anatomy (medialised socket, acetabular bone graft, altered version), retained hardware, and possible HO.
• Despite the added complexity, implant survivorship after THA conversion from PAO is generally comparable to primary THA when performed by experienced surgeons.
• Patients should be counselled that PAO aims to delay THA (often by 10-20 years), not necessarily prevent it entirely.