Hip Preservation—Complications of Periacetabular Osteotomy


Acetabular dysplasia is a major cause of hip pain and dysfunction in young patients. Moderate to severe dysplasia, when left untreated, is an established cause of arthritis requiring total hip arthroplasty (THA) at an early age. , Congenital malformation of the acetabulum and femoral head results in malalignment and increased joint reactive forces that disrupt the labrochondral junction and cartilage, which may eventually result in osteoarthritis. , Significant pain and disability can result from acetabular dysplasia prior to osteoarthritis; therefore, joint preservation techniques such as the Bernese periacetabular osteotomy (PAO) have been developed to better optimize femoral head coverage, reduce patient pain, and prevent secondary osteoarthritis. ,

The Bernese PAO, developed by Ganz et al. in 1988, is a commonly utilized technique that involves the Smith-Petersen approach and multiple hexagonal cuts around the acetabulum (while leaving the posterior column of the pelvis intact) to reorient the acetabulum. , This technique preserves stability and allows for early patient mobility. Postoperative clinical outcomes indicate major pain reduction and improved function in both young and older patients.

The predominant indication for PAO is symptomatic acetabular dysplasia in an adolescent or young adult with a correctable deformity and preserved range of motion. More severe acetabular dysplasia has also been found to be successfully treated with PAO; therefore, dysplasia secondary to neuromuscular disorders and Legg-Calvé-Perthes disease are appropriate indications. , Other indications for PAO include acetabular retroversion and posttraumatic dysplasia. , Borderline dysplasia and instability patterns, commonly associated with increased versional/torsional abnormalities, can be treated with PAO and concomitant open techniques. However, some patients are managed with concurrent or closely staged hip arthroscopy to manage the intraarticular pathology. The optimal treatment strategy has not been determined for this patient population.

Hip arthritis is a relative contraindication to PAO, as increased Tönnis grade 2 or 3 has been found to be a predictor of failure. However, some patients have had improved outcomes, and it may represent a viable alternative to THA in very young patients. Other contraindications include incongruence on radiographs, as well as young patients at risk of damaging the triradiate cartilage. ,

The complication rate of PAO ranges from 4% to 42% depending on the report and the particular complication. Complications are further classified as major or minor depending on severity and impact on the patient. Major complications of PAO include nonunion requiring surgery, heterotopic ossification requiring surgery, permanent lateral femoral cutaneous or major sensorimotor nerve damage, and deep infection. Minor complications include asymptomatic pubic nonunion, lateral femoral cutaneous nerve dysfunction/paresthesia, superficial infection, and asymptomatic heterotopic ossification. Some studies have found the incidence of minor complications to be as high as 95%, with surgeon learning curve commonly acknowledged as a cause. The Clavien-Dindo-Sink classification is a more objective complication schema commonly used with young hip surgery that has demonstrated high interrater and intrarater reliability for hip-preservation surgery. Given that PAO is a complex surgical procedure with a variable complication rate, more exploration of complications—and ways to prevent or mitigate complications—is warranted. The purpose of this chapter is to review the complications of PAO, including the preoperative and patient factors, intraoperative technical issues, and postoperative ramifications.

Preoperative Factors

Patient selection and surgeon experience are important factors for a successful PAO outcome. Preoperative factors for PAO complications include:

  • (1)

    Surgeon Experience

  • (2)


  • (3)

    Patient Age

  • (4)

    Smoking Status

Surgeon Experience

An inverse relationship between complication risk and surgeon experience is noted, with multiple studies indicating surgeon inexperience with PAO technique as a significant factor in complications. , , Various authors reported more complications in the first PAO procedures of their series, likely due to the technically challenging nature of the surgery. , Hussell et al. found that 85% of the technical complications occurred during the first 50 procedures, suggesting that an important learning curve is associated with PAO. However, in a more recent study, only 6% of patients who underwent PAO by a group of 10 surgeons with an average of 9 years (range, 1–37 years) of experience suffered a major complication, which demonstrates that surgeons with sufficient experience can perform PAO with a low complication risk profile. In an effort to investigate the impact of training on PAO complications, Novais et al. reported the complication rate from two young surgeons from their first 4 years of practice. They found that with case exposure greater than 40 PAOs and progressive surgical responsibility during contemporary structured fellowship training, two young surgeons were able to perform PAO with a low risk of complications. Due to the challenging nature of the surgery, orthopaedic residency training alone is not sufficient to perform PAO. Rather, exposure in fellowship, cadaver training, visiting high-volume hip centers, and/or attending specialty courses may be necessary to improve PAO success. A mentored introduction to PAO surgery, with surgeries performed in conjunction with a senior experienced surgeon, may further mitigate the learning curve.


Obesity is associated with an increased risk of complications in patients undergoing PAO. , Novais et al. found that the odds of a complication after PAO in an obese pediatric patient were 10 times higher than that of a patient with a normal body mass index (BMI). Complications seen more frequently included superficial and deep infections and wound hematomas. In another study, the average probability of a patient developing a major complication was 22% for an obese patient compared with 3% for a nonobese patient. Postoperative radiographic acetabular correction was nonsignificant between obese and nonobese patients in both studies, indicating that PAO surgery is reliable in improving radiographic acetabular coverage of the femoral head in both populations. , Increased wound healing complications are likely due to both an increased soft-tissue pannus anteriorly in these patients and an incision located over the anterior flexion crease. Other reasons proposed for increased complications in obese patients include difficult retraction and exposure, particularly for the posterior column osteotomy due to extensive soft tissue in the abdomen and pelvis, challenging repositioning after osteotomy due to obscured intraoperative anatomic landmarks, and poor intraoperative radiographs due to soft-tissue shadowing. Proper patient optimization, including weight management strategies and nutritional counseling, remain important in the preoperative period prior to surgery.


Proper patient selection regarding age for PAO surgery is important for successful outcomes. Increased age and Tönnis grade 2 or 3 are relative contraindications for PAO. Likely due to increased osteoarthritis, older age has been found to be associated with poor outcomes. The increased rehabilitation demands may also lessen the results in older populations. Steppacher et al. assessed 20-year follow-up of patients treated with PAO and found increased age and Tönnis grade at time of surgery to be independent predictors of poor outcomes. Failure was 100% in patients with Tönnis grade 3 and 87% with Tönnis grade 2. Advanced age has been found to be an independent risk factor for conversion to THA in other long-term follow-up studies on rotational acetabular osteotomies. Similarly, in a more recent study of long-term PAO survivorship, hips of patients over 25 years old at time of surgery were three times more likely to be symptomatic after surgery than hips of patients under 25 years old, indicating that increased age has a significant impact on quality of life and subsequent reoperation. Older age has also been shown to negatively impact bone healing and has also been found to be an independent predictor of nonunion in PAO. ,

Smoking Status

Smoking negatively impacts bone healing, leading to delayed union and nonunion in orthopaedic surgeries. , Matsunaga et al. found that superior pubic osteotomy nonunion in patients undergoing curved pelvic osteotomy was 10.7 times higher in smokers compared with nonsmokers, indicating that smoking status is a significant factor in postoperative healing. They recommended guidance on smoking cessation preoperatively as necessary to decrease nonunion risk. Similarly, a case series of painful nonunions after triple pelvic osteotomies found the majority of the patients to be heavy smokers; the authors recommended advising patients that smoking preoperatively is a risk factor for nonunion. Smoking has a well-known negative impact on soft-tissue wound healing postoperatively; we recommend cessation of smoking at least 4 to 6 weeks preoperatively to minimize the complications.

Intraoperative Complications

Vascular Injury

Patients undergoing PAO are at risk for vascular injury to the iliac and femoral vessels and their branches to the hip due to their proximity. The obturator artery, a branch of the iliac, is especially at risk because it runs near the pubic osteotomy site in an acetabular reconstructive osteotomy. Toki et al. reported a case of obturator artery injury during PAO near the medial base of the iliopectineal eminence when the pubic osteotomy was performed. The patient went into shock 3 hours postoperatively and required urgent embolization of the vessel. The authors recommend performing the pubic osteotomy several millimeters medial to the medial base of the iliopectineal eminence, with great care taken during the posterior-distal part of the osteotomy. In a cadaver study, Kalhor et al. found that obturator artery and nerve injury was more likely during unprotected osteotomy of the pubis if the far cortex was penetrated by >5 mm. This can be avoided by inclining the osteotome 45 degrees medially and performing the osteotomy at least 2 cm medial to the iliopectineal eminence.

Major postoperative bleeding is a relatively rare complication and was seen in 1%, 3%, and 8% of cases in three studies. Thawrani et al. reported a major bleeding complication from an aberrant artery in one patient during elevation of the periosteum on the medial aspect of the iliac wing toward the sciatic notch, which required urgent embolization. The patient had previously undergone a hip osteotomy, and the authors believe that neovascularization occurred.

While not reported extensively in the PAO literature, there is risk of femoral artery injury with aggressive medial retraction as seen in THA with the anterior approach. The likely mechanism is compression or excessive traction by the tip of a retractor on the artery, causing intimal injury and thrombosis. This may occur with exposure of the pubis or posterior column and secondarily with mobilization of a large mobile segment correction in the setting of severe dysplasia. Adduction and flexion of the limb when exposing medial-based structures may mitigate undue tension on the femoral neurovascular bundle as well as aid with exposure due to less muscular soft-tissue tension (e.g., decreased tension on the psoas tendon and hip flexor muscles).

Lateral Femoral Cutaneous Nerve Injury

Lateral femoral cutaneous nerve (LFCN) injury is considered to be the most common complication of PAO. It has been found postoperatively in as few as 2% and upwards of 90% of patients postoperatively. , The injury occurs likely due to intraoperative retraction causing compression or stretch resulting in neural ischemia, direct injury, and/or inflammatory neuropathy. , , The superior branch sometimes traverses the operative field and must be sacrificed. Cates et al. recently reported the incidence of clinical LFCN injury as 90%, and 67% had confirmed LFCN injury with nerve conduction testing. Numbness was the most common reported symptom, with tingling, burning, and pain being other reported symptoms; 40% of patients had symptoms resolve by 4 months. Two-thirds of patients had symptoms up to 3 years postoperatively; however, this did not significantly impact any postoperative outcome scores. Various methods to reduce injury include making a C-shaped skin incision closer to the anterior superior iliac spine (ASIS), using blunt dissection as the LFCN commonly exits the pelvis medial to the ASIS, and flexing the hip and/or osteotomizing the ASIS, which can take tension off the LFCN and minimize thermal injury from electrocautery. ,

Major Nerve Injury

Sciatic and femoral nerve injuries after PAO are a serious complication and occur at a rate from 0% to 15%. , , , , Sierra et al. reported a 2.1% incidence of sciatic or femoral nerve injuries in their series; 47% of patients recovered completely, with a median time to recovery or plateau at 5.5 months. They did not identify a patient or surgical risk factor for injury. Femoral nerve injuries recovered in all patients and occurred due to excessive traction. Sciatic nerve injuries recovered in only half of the patients and occurred due to direct damage or stretch/praxis. Because the posterior column remains intact during a PAO, generally protecting the nerve, the risk of sciatic nerve injury is small. However, stretching the nerve can also occur during soft-tissue retraction, or direct injury can occur with aberrant osteotome placement. The obturator nerve is thought to be vulnerable during the pubic ramus osteotomy, as it travels on the inferior aspect of the ramus. Therefore, careful subperiosteal dissection and use of blunt reverse Hohmann retractors on both sides of the pubic ramus are recommended.

Historically, nerve injury is associated with a minimally invasive and abductor sparing technique. Therefore, wider exposure and direct observation of the nerve may decrease nerve injury. The use of intraoperative electromyography (EMG) is recommended by some to identify nerve injuries and as a prognostic tool should injury arise. , Advantages include identifying situations in which nerve damage can occur, such as placement of retractors and helping the surgeon decide whether exploration of the nerve is warranted if the EMG fires during a maneuver. , Disadvantages include cost, requirement of specialized personnel, not detecting sharp lacerations of the nerve that do not produce neurotonic discharge, and perhaps a false sense of security leading to inadvertent injury. We recommend the use of intraoperative EMG to monitor traction phenomena during surgery ( Fig. 25.1 ).

Fig. 25.1

Intraoperative neuromonitoring recording during periacetabular osteotomy by author. Evidence of traction neuropraxia of the right (operative) limb in the posterior tibial nerve (PTN) distribution. The signal change was seen in intraoperative neuromonitoring, as evidenced by poor replication on the right side (the red line signifies the baseline signals, the green line is the most recent average, and the black line is the current average at the time of the screenshot). While the signals recovered to reliable replication, the latency remained delayed slightly and the amplitude remained at about 50% to 75%. The patient had resulting mild motor weakness in the deep PTN distribution immediately postoperatively, which resolved completely by 3 weeks postoperatively.

Insufficient or Excessive Correction of Acetabulum

Appropriate orientation of the acetabulum is essential for a successful PAO. However, the difference between insufficient or excessive correction is small and challenging to assess intraoperatively. Overcorrection can lead to secondary femoroacetabular impingement requiring further surgery ( Fig. 25.2 ). Thawrani et al. reported on PAO complications and found that 2% of patients required revision of the acetabular fragment within 2 weeks of the index procedure due to overcorrection with excessive anterior coverage and retroversion of the acetabulum. Hussell et al. reported 1.4% (7 patients) overcorrection in their series with 1 patient requiring further surgery resulting from a lateral acetabulum stress fracture due to lateral acetabulum translation. Other surgical overcorrections reported included overcorrection of a supralateral defect resulting in greater trochanter impingement, center of rotation overmedialization, and anterior defect overcorrection resulting in impingement symptoms. Center of rotation overmedialization is seen radiographically with the teardrop medial to the ilioischial line, which clinically results in a protrusio deformity with weight bearing and can cause impingement symptoms.

Fig. 25.2

Anteroposterior plain radiograph view of the hip with overcorrection of periacetabular osteotomy resulting in global femoroacetabular overcoverage, overmedialization of the hip center, and secondary (iatrogenic) impingement. The patient had been referred to the author for management of recurrent labral tear and secondary pincer impingement morphology.

Undercorrection of the acetabular fragment results in recurrence of preoperative symptoms and subluxation. Hussell et al. reported 0.7% undercorrection, with only one patient requiring further osteotomy. This complication is thought to occur primarily due to inadequate intraoperative imaging; in addition, patients with severe dysplasia preoperatively are more likely to be undercorrected. Regarding under- and overcorrection, intraoperative image intensification and plain radiographs prior to final screw fixation can be used to further evaluate appropriate fragment position and have been shown to be useful in multiple studies. ,

Postoperative Complications


Osteonecrosis is a rare but serious postoperative complication involving the acetabular fragment or femoral head. The incidence of osteonecrosis of the acetabular fragment ranges from 0% to 3%; osteonecrosis of the femoral head ranges from 0% to 2%. , , Hussell et al. reported a 0.6% rate of acetabular osteonecrosis and a 0.2% rate of femoral osteonecrosis. Acetabular osteonecrosis may occur due to extensive soft-tissue stripping, scarring from previous pelvic osteotomies, and damage to the inferior branch of the superior gluteal artery and the acetabular branches from the inferior gluteal artery. Femoral osteonecrosis occurred in a patient who underwent a femoral osteotomy and PAO due to sequelae of Legg-Calvé-Perthes disease. THA was subsequently performed in all cases. More recently, Thawrani et al. reported acetabular osteonecrosis in 1 patient (1.2%) and femoral osteonecrosis in 1 patient (1.2%). Partial acetabular osteonecrosis was found radiographically in a patient with severe dysplasia; however, no further treatment was performed. Femoral osteonecrosis occurred in a patient with Legg-Calvé-Perthes disease that required a hinged external fixator, providing traction for 3 months; no further treatment was necessary. In summary, extensive soft-tissue disruption in patients with previous pelvic osteotomies and patients with Legg-Calvé-Perthes disease are at risk for acetabular and femoral osteonecrosis, respectively.


Nonunion of the PAO sites is one of the most common complications, with several studies reporting the prevalence ranging from 1% to 17%. , , , , , , , , , , , , Selberg et al. investigated factors contributing to pelvic nonunion and found that when controlling for confounding variables, older age, higher BMI, and more severe acetabular dysplasia were independently associated with nonunion at one or more osteotomy sites at 6 months’ postoperative follow-up, and older age was associated with nonunion at 12-month follow-up. Older age has been found to be an independent risk factor in bone healing. However, no difference in age was found between healed and nonunion pubic rami after curved PAOs in another study. , Smoking and a gap larger than 5.1 mm at the pubic osteotomy site were associated with nonunion. Selberg et al. also found that only 45% of patients had completed radiographic healing at all osteotomy sites at 6-month follow-up. However, 92% had complete radiographic healing at 1 year postoperatively. This may highlight that excessive intervention at 6-month follow-up is not needed. Additionally, at 1 year, there was no difference in postoperative patient-recorded outcome measures between patients with nonunion and patients with complete healing.

The pubic ramus is the most common site of nonunion, and its impact on postoperative pain and function is unclear. Pubic nonunion may be an inevitable consequence in patients with moderate to severe dysplasia, as obtaining sufficient correction and maintenance of appropriate socket version requires greater corrective maneuvers of the acetabular fragment that displaces the ramus more, predisposing it to nonunion. , Several studies have reported no symptoms or effect on postoperative outcomes after pubic nonunion. , Other studies contend that pubic nonunion is symptomatic, with Matsunaga et al. reporting worse pain scores in the pubic nonunion group compared to a healed cohort. , , Surgical management of nonunion is uncommon, with rates seen of 1% and 4%. However, nonunion can be a source of continued instability and pain necessitating surgical management. Shearman et al. reported on a series of postoperative PAO patients with symptomatic pubic nonunion who subsequently were treated satisfactorily by mobilizing, grafting, and plating the nonunion at the superior pubic ramus. Early symptom improvement was noted, and they reported success utilizing a modified Stoppa approach to avoid risking neurovascular structures.

Stress Fracture/Postoperative Fracture

Stress fractures after PAO are not well recognized. They occur with a traditionally reported incidence of 2% to 3% and typically affect the inferior pubic ramus or ischiopubic ramus. , However, a more recent study by Malviya et al. reported an incidence of 18.4% and found that significant predictors for stress fractures were preoperative deformity requiring larger correction and pubic nonunion. Hamai et al. similarly found greater deformity correction associated with stress fracture. Both younger and older age have been associated with stress fractures. No difference in outcome measures were noted between patients with a fracture and those without, and no difference was found in long-term outcomes.

Stress fractures occur because PAO alters the force of weight bearing and adds more strain to the inferior pubic ramus and ischium, which can initiate a fracture. Force is typically dispersed through the sacroiliac joint via the ilium to the pubic symphysis across the superior pubic ramus more than the inferior ramus; however, PAO may alter this dispersion. Ischial stress fractures are more debilitating compared with pubic stress fractures and require double the amount of recovery time as reported by Espinosa et al., who found 2 of 5 requiring fixation and only 1 of 12 patients with a pubic stress fracture requiring fixation.

Stress fractures are typically nondisplaced and are found incidentally on postoperative films. They most likely occur intraoperatively during osteotomy or with immediate early weight bearing. , Accelerated postoperative rehabilitation protocols have also been found to cause pelvic stress fractures in PAO patients, with Ito et al. reporting significantly more postoperative fractures in an accelerated rehabilitation protocol group compared with a control. Underlying bone quality and patient age may create related risk as well.

Venous Thromboembolism

Venous thromboembolism (VTE) is rare after PAO and has a reported incidence of 0% to 2.1%. , , , , Zaltz et al. reported on 1067 patients from multiple centers using varying chemoprophylaxis regimens and found that the crude incidence of developing VTE was 0.94% with zero deaths. This is likely due to the younger patient population undergoing PAO and early postoperative mobilization. Yamanaka et al. similarly found PAO to have the lowest rates of VTE in comparison with other major hip surgeries at one institution.

The optimal prophylactic regimen for young patients undergoing elective hip surgery is unknown, as the routine use of chemoprophylaxis varies because the risk of VTE must be balanced with the risk of postoperative hematoma. Krauetler et al. reported on a series of PAO and/or derotational femoral osteotomy patients treated with postoperative mechanical compression and found excellent results with 0% incidence of VTE. Interestingly, 9 of 10 patients with VTE reported by Zaltz et al. were treated with chemoprophylaxis and mechanical compression. The combination of PAO with concomitant hip arthroscopy may increase the risk of thromboembolic events. Therefore, more aggressive VTE prophylaxis may be necessary. , We recommend aspirin in the senior author’s practice postoperatively to mitigate the risk of deep vein thrombosis.


Infection rates after PAO can vary from 0% to 3% for deep infections and as high as 15% for superficial wound infections. , , , Superficial wound infections did not require surgical treatment in the majority of cases, but deep infections did, with Zaltz et al. reporting 0.9% deep wound infections requiring incision and drainage. , Obese patients undergoing PAO were found to have increased rates of both superficial and deep infections and poor wound healing compared with normal BMI cohorts. , Wound healing and infection issues are seen in obese adolescent patients undergoing fracture fixations as well as in patients undergoing primary hip arthroplasty. , In PAO, this association is likely due to the anterior soft-tissue pannus and the incision over the anterior flexion crease. Smoking also has a negative impact on postoperative wound healing and, in conjunction with obesity, can lead to further poor wound healing. Preoperative medical counseling, weight management counseling, and smoking cessation may be strategies to further reduce infection rates ( Fig. 25.3 ).

Jun 18, 2022 | Posted by in ORTHOPEDIC | Comments Off on Hip Preservation—Complications of Periacetabular Osteotomy

Full access? Get Clinical Tree

Get Clinical Tree app for offline access