Study design and patients
Between November 2001 and December 2020, 71 consecutive revision hip surgeries for recurrent dislocation following THA were performed by four experienced surgeons at our institution. We conducted a retrospective study of all 65 patients (71 hips), who were followed for a mean of 4.7 years (range, 1–14 years). The subject population consisted of 48 female patients (52 hips) and 17 male patients (19 hips), with a mean age of 71 years (range, 34–92 years) at the time of surgery. Dislocation had followed primary THA in 45 hips (63%) and revised THA in 26 hips (37%). The mean number of previous surgeries was 1.6 (range, 1–5).
Before surgery, anteroposterior and lateral radiography and computed tomography were performed to check for implant malposition, raised bone, inclusion, fractures, and non-union. Joints were examined under traction or motion with X-ray imaging to assess soft tissue imbalance and impingement and to identify the cause of dislocation. Gluteus medius failure was diagnosed if the hip dislocated readily in response not only to ordinary traction, but also to the application of lateral stress with the hip in adduction and lateral or backward stress with the hip flexed at 90 degrees. Causes of revision hip surgeries for recurrent dislocation following THA were categorized based on Dorr’s classification [15]: type I (positional) in 2 hips, type II (soft tissue imbalance) in 58 hips, and type III (component malposition) in 11 hips (Table 1). Dislocations were classified as type III if they were caused by cup migration and loosening over time or by the implant itself (thick stem neck, etc.).
Table 1
Preoperative patient characteristics.
Characteristics | Value |
Number of hips | 71 |
Age at surgery (years), mean (range) | 71 (34 − 92) |
Sex, male:female | 19:52 |
Follow-up period (years), mean (range) | 4.7 (1 − 14) |
Previous surgery (times) 1 (Primary THA) 2 ≥3 | 45 17 9 |
Previous approach Anterolateral Posterior Greater trochanteric osteotomy Unknown | 41 17 1 12 |
Previous operating institution Our hospital Other hospital | 22 49 |
Previous fixation method Cement Cementless Hybrid | 55 12 4 |
Causes of revision surgery for recurrent dislocationa Type Ⅰ (positional) Type Ⅱ (soft tissue imbalance) Type Ⅲ (component malposition) | 2 58 11 |
THA, total hip arthroplasty |
a Causes for recurrent dislocation were categorized according to Dorr’s classification |
Our institutional review board (2021153) approved this prospective cohort study. Each patient provided informed consent for data included in the published findings.
Revision Hip Surgery
According to intra-operative findings, revision hip surgery that used implants for recurrent dislocation following THA were classified into 6 groups: group A, open reduction and internal fixation (ORIF); group B, only head change or only liner change; group C, only cup change with head size up; group D, only stem change; group E, cup and stem change; and group F, conversion to constrained cup (Fig. 1). In Group A, ORIF was performed for dislocation caused by greater trochanteric fracture associated with gluteus medius deficiency (Fig. 2). In group B, the head was changed to a longer neck length or the liner was changed. In Group C, the head was changed to a larger diameter, and the cup was changed to match the new head (Fig. 3). In Group D, stem version or depth of insertion was adjusted by changing stems using the “cement-in-cement” technique [16]. In Group E, both cup and stem were changed. Techniques were the same as for Groups B, C, and D, using a conventional cup (Fig. 4). In Group F, the cup was converted to a constrained cup. This technique was used in patients who had either insufficient soft tissue tension or impingement that was unlikely to be managed by the techniques used in Groups A–E. In some cases, stem version or depth of insertion were adjusted by changing stems, as was done in Group D (Fig. 5). Additionally, the stem was replaced in some patients in Groups D–F to make the legs of equal length.
The transgluteal approach in the lateral position was used in all patients. The implant was deployed, and a direct visual check was made to confirm that there was no component failure, component malposition, or impingement, and that no problems with soft tissue were anticipated. After this visual check, the details of treatment strategy were determined. In revised THA, a cemented cup and stem were implanted. For hips with greater trochanteric fracture, osteosynthesis was performed by implanting a trochanteric claw plate (CMK Trochanteric plate, Zimmer Biomet Holdings Inc., Warsaw, IN). This procedure was also used in Group A hips. Revision of cup was performed using a K-MAX CLHO flanged cup with 26-mm head (Kyocera Medical, Osaka, Japan). If necessary, structural allografts and KT plates (Kyocera Medical) were used for massive bone defects [17, 18]. For conversion to a constrained cup, we used the Physio-Hip System Reconstruction Cup (Kyocera Medical), and for stem revision surgery, we used an SC stem (Kyocera Medical) in 14 hips, an HS-3 stem (Kyocera Medical) in 9 hips, a C stem (DePuy International, Leeds, United Kingdom) in 8 hips, a PHS long stem (Kyocera Medical) in 2 hips, and an Exeter stem (Stryker Orthopaedics, Mahwah, NJ) in 2 hips. Long stems were used if a normal stem was considered potentially unstable because of bone fragility or defect. A dislocation test was performed using a trial stem and head to determine the stem version and the depth of insertion. All components were fixed with Endurance Bone Cement (DePuy CMW, Blackpool, United Kingdom) using a third-generation cement technique. Full weight-bearing was permitted as soon as possible, although the use of a cane was encouraged for up to 3 months.
Follow-up Protocol
Postoperative follow-ups were performed at 2 weeks, 3 months, 6 months, 1 year, and annually thereafter. A retrospective analysis was performed by 2 blinded orthopedic surgeons. Relationship between the causes and revision hip surgeries for recurrent dislocation were assessed. Postoperative complications, including re-dislocation, implant breakage, implant loosening, periprosthetic fracture, and periprosthetic infection, were recorded.
Statistical Analyses
For prosthesis survival, we used the Kaplan-Meier method with 95% confidence intervals (CIs). The study end points were repeat revision surgery for re-dislocation, implant breakage, or implant loosening. Univariate logistic regression analysis using a Cox proportional hazards model was performed for risk factors of re-revision surgery due to re-dislocation, implant breakage, and implant loosening. Patient variables included age, gender, number of previous surgeries, implant (cement or cementless), Dorr’s classification [15], and revision hip surgery. Data were analyzed using SAS 9.2 (SAS institute Inc., Cary, NC).