Use of Surgical Approach Is Not Associated With Instability After Primary Total Hip Arthroplasty

A Meta-Analysis Comparing Direct Anterior and Posterolateral Approaches

Elina Huerfano, MD; Maria Bautista, MD, MSc; Manuel Huerfano, MD, MSc; Juan M. Nossa, MD


J Am Acad Orthop Surg. 2021;29(22):e1126-e1140. 

In This Article


The present meta-analysis was reported according to the methodology outlined in the Cochrane Handbook[9] and was not registered.

A systematic review of the literature in PubMed, Ovid MEDLINE, and Embase databases was performed. The following are some Medical Subject Headings and free terms used: "Arthroplasty, Replacement, Hip," "Hip Dislocation/complications," "direct anterior approach hip arthroplasty," and "posterolateral approach hip." Terms were combined using Boolean operators (see Supplemental Digital Content 1, Appendix,

The search did not include the year of publication restriction, but only articles with full text available in English were included in the analysis. Contact with authors, additional manual or gray literature searches were not performed. The date of the last search is March 2020.

We included clinical randomized controlled trials (RCTs) and non-RCT prospective or retrospective studies comparing direct anterior and posterolateral surgical approaches in primary THA, in which the investigators reported the following surgical outcomes: postoperative dislocation, acetabular implant positioning (anteversion, inclination, or acetabular cups positioned within the Lewinnek safe zone), and postoperative leg length discrepancy (LLD). Studies with incomplete data, including hemiarthroplasty or resurfacing hip arthroplasty and studies in which prespecified outcomes were not calculable, were excluded. Studies that used clinical registries and administrative claims databases were also excluded.

Two authors (E.H. and J.M.N.) independently assessed the quality of each study. The Cochrane risk of bias tool was used to appraise the risk of bias of the RCTs that graded the risk as high, low, or unclear in the following six domains of bias: selection bias, performance bias, detection bias, attrition bias, reporting bias, and other bias.[10]

The Newcastle-Ottawa Quality Assessment Scale (NOS) was applied for the assessment of non-RCTs and was scored from 0 to 9; the scale includes the following contents: selection of cohorts, comparability of cohorts, and evaluation of outcome.[11]

Continuous data were recorded as mean and SD, with the treatment effect being reported as the mean difference (MD) and its corresponding 95% confidence intervals (CIs). For binary outcomes, the pooled effect measure used was expressed as a difference of proportions (risk difference [RD]) with 95% CIs for dislocation variable; this allows the inclusion of zero total event maintaining analytic consistency and incorporation of all available data.[12] For the number of acetabular cups positioned within the Lewinnek safe zone variable, the pooled effect measure used was expressed as a ratio of proportions (relative risk [RR]) with 95% CIs. P values < 0.05 were considered statistically significant. We used the I[2] statistic to assess statistical heterogeneity: values of 25%, 50%, and 75% represented a low, moderate, and high degree of inconsistency.[13] When notable heterogeneity was indicated, a random-effects model was applied using the inverse variance method. Otherwise, the Mantel-Haenszel method was used to combine studies using a fixed-effects model. We performed all analyses by using software RevMan 5.3 (The Cochrane Collaboration).