There Are No Differences in Short- to Mid-term Survivorship Among Total Hip-bearing Surface Options: A Network Meta-analysis

Abstract

Background: Total hip arthroplasty (THA) is increasingly being performed in patients with long life expectancies and active lifestyles. Newer implant bearing surfaces, with superior wear characteristics, often are used in this cohort with the goal of improving longevity of the prosthesis, but comparisons across the numerous available bearing surfaces are limited, so the surgeon and patient may have difficulty deciding which implants to use.

Questions/purposes: The purpose of this study was to answer the following question: Is there a short- to mid-term survivorship difference between common THA bearings used in patients younger than age 65 years?

Methods: We conducted a systematic review to identify randomized clinical trials (RCTs) published after 2000 that reported survivorship of ceramic-on-ceramic (CoC), ceramic-on-highly crosslinked polyethylene (CoPxl), or metal-on-highly crosslinked polyethylene (MoPxl) bearings. To qualify for our review, RCTs had to have a minimum 2-year followup and study patients were required to have an average age younger than 65 years. Direct-comparison meta-analysis and network meta-analysis were performed to combine direct and indirect evidence.

Results: Direct-comparison meta-analysis found no differences among the bearing surfaces in terms of the risk of revision; this approach demonstrated a risk ratio for revision of 0.65 (95% confidence interval [CI], 0.19-2.23; p = 0.50) between CoC and CoPxl and a risk ratio for revision of 0.40 (95% CI, 0.06-2.63; p = 0.34) between CoC and MoPxl. Network meta-analysis (with post hoc modification) likewise found no differences in survivorship across the three implant types, demonstrating the following probabilities of most effective implant with 95% credible intervals (CrI): CoC = 64.6% (0%-100%); CoPxl = 24.9% (0%-100%); and MoPxl = 9.9% (0%-100%). The CrIs ranged from 0% to 100% for all three bearing surfaces. Direct-comparison meta-analysis allowed for pooling of five RCTs, including 779 THAs, whereas network meta-analysis (before post hoc analysis) enabled pooling of 18 RCTs, including 2599 THAs.

Conclusions: Current published evidence does not support survivorship differences among commonly used bearing surfaces in patients younger than age 65 years undergoing THA at short- to mid-term followup. Long-term RCT data will be needed to determine if a survivorship benefit is realized in younger, more active patients over time.

Level of evidence: Level I, therapeutic study.

Fig. 1

This algorithm details the process of study selection.

Fig. 2

This is a direct-comparison meta-analysis with a forest plot of the RRs for revision of CoC versus CoPxl bearings. The horizontal bars represent the CIs for individual studies, and the black diamond represents the cumulative risk ratio of the set of studies. The black diamond crosses the vertical line, indicating no statistical difference in risk of bearing failure between the groups. M-H = Mantel-Haenszel.

Fig. 3

This is a direct-comparison meta-analysis with a forest plot of the RRs for revision of CoC versus MoPxl bearings. The horizontal bars represent the CIs for individual studies, and the black diamond represents the cumulative RR for the set of studies. The black diamond crosses the vertical line, indicating no statistical difference in risk of bearing failure between the groups. M-H = Mantel-Haenszel.

Fig. 4

This is a schematic of the constructed linkage system enabling a network meta-analysis to be performed on all 18 RCTs. The CoPc studies were excluded for the sensitivity analyses. Circle size is proportional to the number of hips receiving each implant type and line thickness is proportional to the number of RCTs.