High Risk of Neck-liner Impingement and Notching Observed with Thick Femoral Neck Implants in Ceramic-on-ceramic THA

Recently, impingement between the femoral stem and ceramic liner, which appears as a notch in the stem neck on radiographs, has emerged as a new complication. However, the proportion of impingement, potential risk factors, and related complications are poorly elucidated.

In patients who underwent ceramic-on-ceramic (CoC) THA and had at least 10 years of follow-up, we asked: (1) What proportion had radiographic evidence of stem neck-ceramic liner impingement (notching of the stem), and what implant design, implantation factors, or complications such as ceramic fracture or metallosis are associated with impingement? (2) How common are the complications of noise and ceramic fracture? (3) What are the radiologic changes (including fixation of femoral and acetabular components, osteolysis, and heterotopic ossification) and clinical outcomes as determined by the modified Harris hip score (mHHS) for this cohort? (4) What is the survivorship with implant revision as the endpoint and with reoperation for any reason as the endpoint after CoC THA using a thick-neck stem design?

Between May 2003 and April 2010, 643 patients underwent primary THA at a tertiary referral hospital. After excluding patients with metal-on-polyethylene and with ceramic-on-polyethylene implants, 621 patients were considered eligible for this study. All patients received the same hemispherical titanium cup, a standard-length tapered titanium stem, and a CoC bearing. Of those, 19% (115) were lost to follow-up before 10 years, and 8% (50) were excluded because they died before the minimum follow-up duration of 10 years, leaving 73% (456) for analysis. Patients had a mean age of 50 ± 14 years and were followed for a median (range) of 13 years (10 to 17). The mean cup abduction was 38° ± 5°, and the mean cup anteversion was 26° ± 7°. To determine the proportion of patients with the neck-liner impingement, we analyzed the plain radiographs of every patient during follow-up to detect notches around the stem. The detection of stem neck notches on the radiographs was reliable (intraobserver reliability: κ = 0.963; p < 0.001 and interobserver reliability: κ = 0.975; p < 0.001). To evaluate factors related to notching, we compared the possible confounding factors including gender, age, BMI, implant position, neck length, and head diameter. Complications such as ceramic fracture, noise, dislocation, and periprosthetic joint infection were recorded. Noise was evaluated via interview and with the Hip Noise Assessment Questionnaire, which assessed the noise qualitatively. For clinical outcome, we assessed the mHHS, which includes pain and function scales (0 [worst] to 100 [best]), every visit. Tilting of at least 4° or migration of at least 4 mm was the criteria for cup loosening; subsidence more than 3 mm, any change in position, or a continuous radiolucent line greater than 2 mm was the criteria for stem loosening. To evaluate osteolysis, we performed CT scans in 57% (262 of 456) of patients. Kaplan-Meier survivorship analysis was performed using the endpoints of survivorship free from implant revision and survivorship free from reoperation for any cause.

The proportion of stem neck notching was 11% (49 of 456). There were no differences in cup abduction and anteversion between hips with notches and those without notches. Notched hips were more likely to have 28-mm than 32-mm heads (90% [44 of 49] versus 70% [285 of 407]; odds ratio 3.77 [95% CI 1.46 to 9.73]; p = 0.004). None of the 49 notched hips had a ceramic head or liner fracture or evidence of metallosis. A ceramic head fracture was reported in 2% (9 of 456); all fractures occurred in 28-mm short neck heads. A ceramic liner fracture occurred in 0.2% (1 of 456), and noise was noted in 6% (27 of 456). Acetabular osteolysis developed in 2% (7 of 456). The mHHS was 91 ± 12 points at the final follow-up. The survivorship free from implant revision was 97% (95% CI 96% to 99%), and the survivorship free from reoperation for any cause was 96% (95% CI 95% to 98%) at 13 years.

The proportion of stems with neck-ceramic liner impingement and ceramic component fracture were unacceptably high after the use of a thick-neck stem design, especially when a 28-mm head was used. We have discontinued the use of this stem design and we recommend that such stems should not be used when CoC bearings are used. As these findings might be generalized to other bearing couples, further studies focused on polyethylene liner wear and local metallosis due to thick stem neck are warranted.

Level III, therapeutic study.