A Multivariate Analysis to Predict Total Hip Arthroplasty Dislocation With Preoperative Diagnosis, Surgical Approach, Spinal Pathology, Cup Orientation, and Head Size

There are several risk factors for dislocation after total hip arthroplasty (THA), but few studies include radiographic assessment of implants, with spine pathology and patient characteristics. This study estimates the rate of dislocation by patient gender, age, race/ethnicity, body mass index, Charlson Comorbidity Index, spine pathology, prior spine fusion, levels affected, radiographic Kellgren-Lawrence score of spine osteoarthritis, THA indication, surgical approach, and femoral head size.

Seventy-six primary THA patients between January 2007 and 2020 with a dislocation were matched on age and gender to subjects without a known history of dislocation using a 2:1 allocation. Univariable and multivariable survival models that account for matched strata were used to estimate the rate of dislocation.

Median follow-up of patients at risk for dislocation was 26.48 months (95% confidence interval [CI] 23.75-36.40). On multivariable analysis, patients with an indication other than primary osteoarthritis were 3.69 (95% CI 2.22-6.13, P < .001) times more likely to dislocate than those with osteoarthritis. Patients with a spine pathology were also nominally more likely to dislocate (hazard ratio 1.76, 95% CI 0.97-3.18, P = .06), and patients receiving a posterior surgical approach were 2.74 (95% CI 1.11-6.76, P = .03) times more likely than those receiving a non-posterior approach to dislocate.

Patients with THA indication other than primary osteoarthritis and receiving a posterior surgical approach, and to a lesser degree spinal pathology, were identified as affecting the rate of dislocation. After correcting for other variables, femoral head size, cup orientation, and patient factors were not predictive.

Level IV, case-control study.