Comparative postoperative prognosis of ceramic-on-ceramic and ceramic-on-polyethylene for total hip arthroplasty: an updated systematic review and meta-analysis

Protocol and registration

This systematic review and meta-analysis followed a preprint registered in the International Prospective Register of Systematic Reviews (PROSPERO; CRD42023400537) and was designed and conducted according to the guidelines in the Cochrane Handbook.

Search strategy and data sources

We conducted a meta-analysis of all studies that compared CoC and CoP for THA, identified by searching PubMed, Embase, Cochrane Library, and Web of Science up to March 2023. References were managed using Endnote X9 software (Clarivate Analytics). In addition, relevant reviews and the references of selected articles were examined for potentially relevant trials.

Eligibility criteria

The inclusion criteria were as follows: (1) prospective randomised studies; (2) retrospective randomised studies (3) randomised controlled trials (RCTs) (3) studies of patients undergoing primary THA; (3) studies comparing COC-THA and COP-THA; (4) studies reporting complications or revisions or Harris Hip Score; (5) studies in which the mean age of the patients was >40 years; (6) studies published in the English language.

The following studies were excluded: (1) non-human studies; (2) non-original studies (letters, reviews, editorials); (3) studies with patients who had no THA; (4) studies of revision surgery; (5) interim follow-up of the same study; and (6) studies without available data.

Study selection and data extraction

After removing duplicates, each article was screened independently by two reviewers who were blind to the journal, author, institution at which the study was performed and the date of publication, to identify potentially eligible studies and relevant clinical trials. One reviewer (TYW) was responsible for extracting the data, and another (YPJ) checked the data for accuracy against the source material. The final eligibility of the retrieved full-text articles for inclusion was assessed independently by two reviewers. Differences of opinion were resolved by discussion, and if no agreement was reached, the third reviewer (TL) made the final decision.

Quality assessment

The quality of included studies was evaluated using the Cochrane risk of bias assessment tool. The standards implemented by the Oxford Centre for Evidence-based Medicine were applied to assess the level of evidence. Two investigators independently evaluated the quality and level of evidence for eligible studies. An arbiter was consulted to reconcile any disagreements.

Statistical analysis

Statistical analysis was performed with Review Manager 5.4 (Cochrane Collaboration, Oxford, UK) and Stata 12.0 (StataCorp LP, College Station, Texas) software. The weighted mean difference (WMD) and 95% CI were calculated for the continuous outcome (Harris Hip Score). The medians and interquartile ranges of continuous data were converted to means and standard deviations. The results of binary variables (dislocation, fracture, deep vein thrombosis, infection, wear debris or osteolysis and revision) are expressed as odds ratios (OR) with 95% confidence intervals (CI). For meta-analyses, the Cochrane Q p-value and I² statistic were applied to assess heterogeneity. A p < 0.05 or I² > 50% indicated significant heterogeneity; in such cases, a random-effects model was used to estimate the combined WMD or OR. Otherwise, a fixed-effects model was used. We performed one-way sensitivity analyses to evaluate the effects of included studies on the combined results for outcomes with significant heterogeneity. A value of p < 0.05 was considered indicative of statistical significance.

Study selection and characteristics

Figure 1 outlines study selection and reasons for exclusion. A total of 3,640 relevant articles in PubMed (n = 888), Embase (n = 1,272), Cochrane Library (n = 99), and Web of Science (n = 1,381) were identified. After removing repeated studies, and reading titles and abstracts, 10 full-text articles involving 1,946 patients and 2,098 THA surgeries were included in the pooled analysis. These articles comprised seven prospective randomised studies (Van Loon et al., 2021; Amanatullah et al., 2011; Atrey et al., 2018; Bal et al., 2005; Feng et al., 2019; Lewis et al., 2010; Ochs et al., 2007), two retrospective studies (Giuseppe et al., 2022; Epinette & Manley, 2014) and one RCT (Beaupre, Al-Houkail & Johnston, 2016). All studies were in English and were published between 2005 and 2022. The mean follow-up period was 2–15 years. HXLPE liners were used in two of the studies (Feng et al., 2019; Epinette & Manley, 2014), and conventional polyethylene liners in the others (Giuseppe et al., 2022; Van Loon et al., 2021; Amanatullah et al., 2011; Atrey et al., 2018; Lewis et al., 2010; Ochs et al., 2007; Beaupre, Al-Houkail & Johnston, 2016). All the articles presented baseline age, sex, and body mass index values. The characteristics of the studies are listed in Table 1. We did not assess publication bias because <10 studies were included in the required observation measures.

Demographic characteristics

There was no significant difference between the two groups in terms of gender (female/total, OR: 0.90; 95% CI [0.74–1.09] p = 0.27), BMI (WMD: 0.04; 95% CI[−1.20 to 1.29]; p = 0.94). However, the two groups were significantly different in baseline age (WMD: −4.07; 95% CI [−5.93 to −2.20]; p < 0.0001) (Table 2). The mean age of patients in the CoP-THA group was 4 years older than in the CoC-THA group (Giuseppe et al., 2022; Van Loon et al., 2021; Amanatullah et al., 2011; Atrey et al., 2018; Bal et al., 2005; Feng et al., 2019; Lewis et al., 2010; Ochs et al., 2007; Epinette & Manley, 2014; Beaupre, Manolescu & Johnston, 2013).

Risk of bias

Overall, the risk of bias was low across the trials. All trials used allocation concealment, followed a blinded design, and no selective reporting occurred. One trial (Bal et al., 2005) did not describe the sources of the exposed and non-exposed groups, and the exposed group in one trial (Beaupre, Al-Houkail & Johnston, 2016) was not representative, all of whom were volunteers. The rate of loss to follow-up was >10% in five trials (Amanatullah et al., 2011; Atrey et al., 2018; Bal et al., 2005; Ochs et al., 2007; Epinette & Manley, 2014), and loss to follow-up was not clearly described in one trial (Van Loon et al., 2021). The study groups of four trials (Van Loon et al., 2021; Amanatullah et al., 2011; Bal et al., 2005; Feng et al., 2019) were significantly different in age from the control group and were not comparable in age, so other biases may have been introduced. Figures 2 and 3 summarise the risks of bias of the included studies.

Primary outcome: rates of common postoperative complications and revision surgery

Dislocation rate

Postoperative dislocations were reported in seven of the 10 studies, involving 1,742 patients (1,086 CoC-THA versus 807 CoP-THA) (Amanatullah et al., 2011; Bal et al., 2005; Feng et al., 2019; Lewis et al., 2010; Ochs et al., 2007; Epinette & Manley, 2014; Beaupre, Al-Houkail & Johnston, 2016). There was no significant difference between the two groups (OR: 0.85; 95% CI [0.51–1.42]; p = 0.54), and no significant heterogeneity (I² = 0%, p = 0.77) (Fig. 4A).

Revision surgery

Seven of the ten studies reported information on the revision surgery (Van Loon et al., 2021; Amanatullah et al., 2011; Atrey et al., 2018; Bal et al., 2005; Ochs et al., 2007; Epinette & Manley, 2014; Beaupre, Al-Houkail & Johnston, 2016). Revision surgery was performed in 949 patients (1,026 hips) in the CoC-THA group and 723 patients (760 hips) in the CoP-THA group. The mean follow-up duration was 2.0–15 years. Common causes included hip instability, loose components, wear and osteolysis, infection, recurrent or multiple dislocation and implant or periprosthetic fracture, all of which were common postoperative complications (Jungwirth-Weinberger et al., 2022; Ledford et al., 2019). A small number of patients experienced stem subsidence, implant tilt, and persisting pain of unknown cause after surgery, so revision surgery was performed. The revision surgery rate in the CoC-THA group was non-significantly lower (RR: 0.77; 95% CI [0.45–1.32]; p = 0.35). No significant heterogeneity (I² = 38%, p = 0.14) was observed (Fig. 5).

Secondary outcome: postoperative Harris Hip Score

Because few of the included studies evaluated hip function after surgery, there was insufficient information to compare hip function among implant combinations. The Harris Hip Score was used in six studies. Because one study did not differentiate between CoC-THA and CoP-THA scores (Amanatullah et al., 2011), and another did not clearly describe the standard deviation or range of scores (Feng et al., 2019), only four sets of data were analysed, involving 420 patients (239 CoC-THA vs. 186 CoP-THA) (Van Loon et al., 2021; Atrey et al., 2018; Ochs et al., 2007; Epinette & Manley, 2014). The mean follow-up duration was 8.1–15 years. The Harris Hip Score improved significantly in both groups compared to that preoperatively (Table 3), it was non-significantly higher in the CoC-THA group (WMD: 2.50; 95% CI [−1.26 to 6.26]; p = 0.19). However, statistically significant heterogeneity was observed (I² = 55%, p = 0.08) (Fig. 6).

Because of the >50% heterogeneity in postoperative Harris Hip Score, we conducted one-way sensitivity analyses to evaluate the influence of each individual study on the combined WMD by removing the studies one-by-one. Exclusion of Epinette & Manley (2014) eliminated the heterogeneity of the Harris Hip Score (I² = 0%, p = 0.78) (Fig. 7) and the postoperative Harris Hip Score was significantly higher in the CoC-THA group than in the CoP-THA group (WMD: 5.17; 95% CI [1.38–8.96]; p = 0.007) (Fig. 8), suggesting that this study accounted for most of the heterogeneity.

Strengths and limitations

To the best of our knowledge, previous studies on the selection of bearing surfaces for THA are all from the perspective of doctors, focusing on the outcome of the operation. However, this current report is the first to focus on postoperative outcomes from the patient’s perspective and to provide an important reference for patients in preoperative selection of bearing surfaces for THA.

Although we used the Cochrane collaborative-recommended GRADE system to evaluate the results, this study has several limitations. First, the clinical trials in the references we included did not take into account factors such as BMI, comorbidities, and they tended to use CoC bearings in younger patients, so these variables were not controlled well. Second, we analysed a small number of trials, precluding the generation of funnel plots to assess publication bias. The inclusion of studies with small sample sizes may lead to smaller study effects, leading to large standard deviations (SDs). Third, there were insufficient data to perform a subgroup analysis according to type of prosthesis, which may have introduced bias. Finally, since only two of the cited literatures used crosslinked polyethylene, there was insufficient data. We could not distinguish conventional polyethylene from crosslinked polyethylene. However, it is worth noting that in the analysis of wear debris, the use of highly cross-linked polyethylene liners was not included in the study, so the accuracy of the data is guaranteed.