Skip to content

Risk of Dislocation in Patients After THA Higher in Spinal Fusion Than Ankylosing Spondylitis

Key findings

  • This retrospective study compared dislocation rates and radiographic variables after total hip arthroplasty in 142 patients with ankylosing spondylitis (AS) and 135 with a history of spinal fusion (SF)
  • The dislocation rate was substantially higher in the SF group than in the AS group (12% vs. 3%; P=0.01)
  • In both groups, greater lumbar lordosis angle, and acetabular anteversion and inclination angles outside the Lewinnek safe zone, were associated with a higher risk of dislocation
  • Orthopedic surgeons should consider that AS patients may suffer from excessive acetabular anteversion and SF patients may suffer from excessive acetabular retroversion, possibly leading to a dislocation

Postoperative dislocation is a devastating complication of total hip arthroplasty (THA). Patients with ankylosing spondylitis (AS) or a history of spinal fusion (SF) are among the populations at higher risk, and their greater rate of dislocation has been attributed to spinopelvic immobility.

Now, Massachusetts General Hospital researchers have shown an association between greater lumbar lordosis angle (LLA) and increased dislocation risk in patients with AS or SF. In The Journal of Arthroplasty, surgical research fellow Akhil Katakam, MBA, orthopedic surgeons Christopher M. Melnic, MD, and Hany S. Bedair, MD, of the Center for Hip & Knee Replacement at Mass General, explain why the risk of dislocation is better accounted for by degree of lumbar lordosis than by spinopelvic stiffness itself.

Study Details

The retrospective comparison study involved patients who had THA at Mass General between January 2000 and December 2017. 142 had a preoperative diagnosis of AS affecting the lumbar or sacral spine and 135 had previously undergone arthrodesis of the lumbar spine.

Postoperative LLAs were measured from lateral standing radiographs, and acetabular anteversion (AA) and inclination angles were measured from anteroposterior and cross-table lateral radiographs. The average follow-up was 74 months for the AS cohort and 50 months for the SF cohort.


12% of the SF group, but only 3% of the AS group, experienced dislocation (P < 0.01). The average LLA was greater in the SF group (34°) than in the AS group (21°) (P < 0.01).

When the two groups were analyzed together via a Cox regression model:

  • An increase in LLA of 1° increased the probability of dislocation by 13% (HR, 1.13; 95% CI, 1.08–1.18; P < 0.001)
  • Hips beyond the Lewinnek safe zone parameters of 25° for AA angle and 40° for inclination angle were at substantially increased risk of dislocation (HR, 5.18; 95% CI, 1.64–16.33; P = 0.005)

A Delicate Balance

THA poses two contrasting problems for these patients: AS patients may suffer from excessive acetabular anteversion, and SF patients may suffer from excessive acetabular retroversion.

When a healthy person is sitting, the pelvis is tilted posteriorly, which increases AA and protects against posterior dislocation. In addition to spinopelvic immobility, SF patients have relatively retroverted positioning of the acetabulum, so they are at severe risk of posterior dislocation when sitting. That may explain their significantly higher dislocation rate compared with AS patients.

Orthopedic surgeons should consider the LLA and its relationship with AA angle when counseling patients considering THA. Collaborations between spine and hip surgeons are recommended, along with novel imaging techniques to position the acetabular cup precisely.

greater risk of dislocation after THA in patients with SF than those with ankylosing spondylitis

greater risk of dislocation after THA for every 1° of lumbar lordosis in patients with SF or AS

greater risk of dislocation after THA for patients with SF or AS who had hips beyond Lewinnek safe zone parameters

Learn more about the Center for Hip & Knee Replacement

Refer a patient to the Department of Orthopaedic Surgery

Related topics


A machine learning algorithm shows promise for determining which patients are at increased risk of opioid dependence after total knee arthroplasty.


A recently introduced-design of posterior-substituting, high-flexion knee prosthesis did not fully replicate the posterior femoral translation and ligament isometry of a healthy knee during in vitro robotic analysis.