CORR® Synthesis: To What Degree Does the Direct Anterior Approach Improve Outcomes in THA? A Systematic Evaluation of Meta-analyses

In the Beginning… Contrary to recent claims that the direct anterior approach is an innovative and novel surgical technique, this approach has been used in hip surgery or arthroplasty for more than 140 years. The direct anterior approach was first described in 1881 by Carl Hueter in the textbook, Grundriss der Chirurgie [44] and subsequently popularized by Smith-Peterson in 1917 and then by the Judet brothers and O’Brien in the 1950s [22, 39, 51]. The approach fell out of favor over the next few decades due to the increased risk of femoral complications and the growing popularity of Charnley’s transtrochanteric osteotomy [10]. The direct anterior approach had a resurgence after Light and Keggi [27] reported their experience with this approach in 1980 and Judet and Judet [21] published their outcomes with this approach using an orthopaedic operating table in 1985. The basic premise of the direct anterior approach is that it utilizes the internervous and intermuscular intervals between the tensor fascia lata and the sartorius muscles [30]. Conceptually, using this interval avoids detachment of the muscles from bone and thereby limits the trauma of surgery. It was concluded in an earlier review that “the orthopaedic community will embrace this technique” and that it should “be introduced into routine practice” [36]. Currently, a growing number of surgeons are adopting the direct anterior approach in THA. The 2018 survey by the American Association of Hip and Knee Surgeons showed that 56% of respondents used the direct anterior approach in THA, and those who performed the direct anterior approach reported increased patient market share [40]. The Argument Although some randomized controlled trials (RCTs) have shown that the direct anterior approach was associated with improved clinical outcomes such as less postoperative pain, less muscle damage and inflammatory response, faster rehabilitation, and decreased usage of opiate analgesics [6, 8, 29, 32, 63], these results were reported without much discussion about the clinical relevance, and the minimum clinically important differences (MCIDs) of these parameters were not discussed. At the same time, others have shown either the opposite or mixed effects or no differences in outcomes when compared with conventional approaches, making a conclusion on these benefits difficult to draw [11, 14, 19, 34, 35, 48]. Importantly, some have reported that the direct anterior approach was associated with increased risk of major complications such as nerve injury, femoral fracture, increased risk of femoral loosening, and infection [1, 2, 8, 13, 31, 42, 52]. Such serious complications are also uncommon, so most studies were not powered to detect a difference, lending a false sense of safety in using the direct anterior approach when a difference was not detected in a particular study or review. The absence of a noted difference, however, does not mean that differences did not exist. Moreover, the direct anterior approach has been associated with increased operating time, use of fluoroscopy leading to additional radiation exposure to surgeons and patients, and a steep learning curve, often estimated at approximately 100 cases [15, 28, 30, 37, 53]. Despite the many meta-analyses that have been published in the past decade to analyze the existing evidence, disagreement persists. This disagreement, in our opinion, is related to the potentially poor methodology in these meta-analyses and to biases within the studies from which the data were synthesized in the meta-analyses. Our objective was to summarize the results and evaluate the quality of existing meta-analyses of RCTs to provide an overview of study findings and to identify the relative merits and limitations of using the direct anterior approach versus conventional approaches in THA. The results should facilitate more informed discussions between patients and healthcare professionals. Essential Elements Search Strategy We developed a written protocol before we began our search. Because PROSPERO did not accept registration of literature reviews using a systematic search or systematic critical appraisals, this study was not registered with PROSPERO. The search terms used for PubMed were (meta-analysis) AND ((total hip replacement) OR (hip arthroplasty) OR (hip replacement) OR (total hip arthroplasty)) AND (approach). Aside from PubMed, Web of Science, Embase, Cochrane reviews, and Google Scholar were searched using the same search terms, adapted as necessary. The eligible period for the search was from database inception through April 2024, with no language restriction being placed on the search. The bibliographies of the eventually included papers were screened to reduce the possibility of missing relevant studies. The inclusion criteria were meta-analyses that compared surgical time, pain after surgery, speed of recovery, patient-reported outcome scores, reoperation rates, and/or complications between the direct anterior approach and conventional approaches, including the posterior approach (also called the posterolateral approach), the direct lateral approach (also called the lateral approach or the Hardinge approach), and the anterolateral approach (also called the Watson-Jones approach) for primary THA. In the first phase of screening, all meta-analyses on the relevant subject were included to get a sense of the number of meta-analyses on the subject; in the second phase, only meta-analyses of RCTs were included, including meta-analyses with study designs other than RCTs, as long as analyses stratified for RCTs only had been performed. Ultimately, the analysis in this review included only data from RCTs and not data from studies with a lower level of evidence. Because of claims that surgical approach may affect early recovery, it was important to include publications that included early results even when these may not have had a minimum follow-up time of 2 years. Also, certain complications such as fractures and dislocations happen early; in order not to miss capturing these potential risks, we did not insist on a minimum follow-up time of 2 years. The trade-off of including studies with a shorter follow-up time is the potential of underestimating the longer-term failure due to loosening that may potentially be related to surgical approach. The exclusion criteria were narrative reviews, systematic reviews without meta-analyses, meta-analyses not including outcomes described in the inclusion criteria, and meta-analyses comparing the direct anterior approach to approaches that were not the posterior approach, the direct lateral approach, or the anterolateral approach. Data from non-RCTs were not included. Abstracts from conference proceedings, preprint server papers, and papers published in nonindexed journals were not considered. One author (MC) and a colleague (not a study author) independently screened the titles and abstracts based on the inclusion and exclusion criteria to determine the eligibility of the publications. In case of disagreement, the differences were resolved by consensus. Data Extraction Before data extraction, two authors (MC, BAM) agreed on the outcomes to be extracted. These were: (1) study characteristics such as the number of RCTs included, quality assessment (that is, risk of bias [RoB] assessment), inclusion of learning cases, and declaration of conflict of interests; (2) functional and patient-reported outcomes such as the Harris hip score (HHS), Oxford Hip Score, Hip Disability and Osteoarthritis Outcome Score, and pain, plus the follow-up time for these measures; (3) complications including infection, periprosthetic fractures, intraoperative fractures, femoral loosening, dislocation, and nerve injury; (4) operative time; and (5) blood transfusions (which was used as a surrogate for blood loss). One reviewer performed the data extraction, and the results were scrutinized and confirmed by a second reviewer (MC). In case of disagreement, a third reviewer acted as the tiebreaker. We did not include outcomes such as length of hospital stay, incision length, opiate use, and blood loss because these endpoints commonly were not measured in robust, standardized ways and/or because they are prone to subjective influences. For example, the length of hospital stay may be influenced by hospital occupancy rate, the surgeon’s personal preference and habit, or other circumstantial rather than medical reasons. In addition, with any surgical approach, there has been a trend toward a reduction in hospital length of stay and in outpatient hip arthroplasty over the timespan during which many of our source studies were published, which is not necessarily related to the approaches used. We did not look at blood loss but used blood transfusion as a surrogate because estimates of blood loss are notoriously inaccurate and should not be used as a definitive proof of superiority of one approach over another. Quality Assessment We assessed the quality of the methodology of the included meta-analyses using the updated version of the “A Measurement Tool to Assess Systematic Reviews” (AMSTAR2) tool [49]. The updated tool includes 16 questions (domains) on topics such as the protocol and the methods of the systematic search, statistical methodology, assessment of RoB and heterogeneity, discussions of RoB and heterogeneity, and potential conflicts of interest. As described by Shea the are to first identify the that may affect the of a review and Shea then a for the level that have in the that is, the have or in the results and conclusion of the review. In with this two authors (MC, first the critical that is, that were most relevant in the clinical in the meta-analyses were by the same two (MC, to the and tool were and resolved by consensus. the level of of meta-analyses based on the number of was is to have when it has to when it has when it has critical with or without additional and it has critical with or without additional [49]. the same search terms, an was this to publications without to search for all the publications the search terms an was to In addition, a search of PubMed was performed using the same search terms before the was in April from the and the PubMed search were and for papers with or important We did not data and quality assessment for publications from these We We Systematic Search and We the systematic search between and The of for database were as PubMed Google Cochrane and Web of were to be meta-analyses comparing the direct anterior approach with other of these meta-analyses that meta-analyses were to include only RCTs or had stratified analyses for RCTs a number we to be to provide a of the of from studies of the level of evidence. We a of the meta-analyses The for exclusion were (1) the study was not for RCTs only and (2) the study compared direct anterior approach with of a conventional approach for Systematic and for this included study designs other than RCTs or those not comparing the direct anterior approach with conventional of the included meta-analyses associated with operative time in to RCTs had improved at RCTs mean difference blood transfusion at and pain at associated with risk of surgery RCTs and nerve RCTs compared with femoral risk and risk associated with operative time in with RCTs associated with operative time in with RCTs associated with improved at RCTs mean difference and at RCTs mean difference compared with nerve pain associated with less pain compared with at postoperative 2 and 2 of studies and patients associated with at compared with mean difference compared with mean difference at pain at postoperative and risk between and associated with increased operative time in to and RCTs and less pain at RCTs associated with in RCTs mean difference and score RCTs mean difference compared with the other approaches associated with operative time in to RCTs at and pain at associated with pain in with number of studies and and number of studies associated with improved mean difference compared with at risk of dislocation, fracture, infection, and reoperation associated with operative time RCTs less pain RCTs blood complications associated with operative time RCTs and lower at postoperative and 2 in with associated with improved at RCTs mean difference compared with of blood risk of dislocation, fracture, and nerve injury; at associated with improved compared with the lateral approach mean difference and mean difference in two RCTs but than RCTs score between and at the follow-up between and or associated with operative time in with RCTs at and pain score at associated with operative time in to RCTs at and pain score at operative time; at and and In to operative time, for blood pain at and In to lower pain at RCTs operative time, for blood pain at 2 and associated with less pain at RCTs RCTs and RCTs compared with associated with improved at 2 RCTs mean difference and RCTs mean difference compared with operative time, intraoperative risk associated with in with at RCTs mean difference operative time and at and nerve in with RCTs less pain at RCTs and 2 RCTs operative time, intraoperative risk associated with compared with difference to operative time in with and RCTs associated with improved at RCTs mean difference at and risk no difference was noted most studies were the absence of a noted difference does not the absence of a difference the study had been powered for the direct anterior posterior direct lateral conventional approaches, including and anterolateral results from and were only the results from are reported For meta-analyses, only the with relevant direct are and two studies by included RCTs of from the first and of from the second We that this does not affect the results because the authors This with the studies of and which were studies and had For the published an of to the because this of papers results in operative time and no difference in we the of the results number of patients included in the two RCTs is most Although and included patients in their this number in their many studies learning cases this mean that learning cases were not included in the and not that any study learning in should be as that is, who reported on the hip in the included noted that studies included only patients with only one included patients with of of the femoral of and of the included two were published in abstracts In these cases, data was performed using the abstracts by one reviewer and the publications and by the second reviewer (MC). We did not an assessment for these two publications. The most from the included meta-analyses was an increased surgical time for the direct anterior approach Although we had to include meta-analyses on primary many of these meta-analyses did not that only studies on primary were we have included such studies and that they did not include We had also that most of the studies be on patients with but most of the meta-analyses did not the inclusion criteria on surgical the of their inclusion criteria to for example, fractures, even it is not many or any fractures were included in the studies. We to include all meta-analyses of RCTs in our review it was to include only other than The was on all included meta-analyses for the two publications written in We as of the and 2 before the of for the RoB in or the an or the bias and on the and 16 of potential conflict of studies had in at two critical all studies a level of We have a table including all of our of and critical of critical the questions and inclusion criteria for the review include the of the of the review an that the review methods were to the of the and did the any from the the review authors use a for RoB in studies that were included in the the review authors for RoB in studies when the results of the the review authors provide a and discussion any in the results of the they performed did the review authors out an of bias study and on the results of the the review authors any potential of conflict of including any they for the 2 2 2 and 2 2 and were by two authors (MC, before the studies were are that may affect the of a review and was in the on a of criteria should be In our we did not a score of of a critical as a is to have when it has to when it has when it has critical with or without additional and it has critical with or without additional two studies by included RCTs of from the first and of from the second We that this does not affect the results because the authors This with the studies of and which were studies and had For the published an of to the The noted that and by only one the were and The results were from the included meta-analyses The was the most used outcome for As it was the functional outcome used by most of the meta-analyses for data Although there are in the results for to were reported but not for time The is more with studies that the direct anterior approach in less pain on postoperative also 2 and/or but not others One study that patients had less pain on postoperative 2 and 2 but not on or study that patients had less pain at but not at 2 or number of studies detected a risk of complications as nerve or when the direct anterior approach was used but most reported no differences in a number of complications that had been in our review the Essential Elements As the absence of detected differences in complications does not necessarily that those differences are of the primary studies used in these meta-analyses were not powered to detect differences in these complications and were not in the to that a approach was compared with the studies shorter operative in conventional approaches studies detected no difference The difference in operative time from a mean difference of approximately to most of the studies detected a mean difference We however, that two of the studies that detected no difference in operative time had in their one included a study as a randomized study and one placed the data on the of the transfusion was reported to have no difference between the direct anterior approach and conventional approaches The learning was a often not or not in the studies. however, some of these results in For example, a number of the meta-analyses included as an even the authors in their that the study was not Because this study had a much shorter mean operative time for the direct anterior approach than the posterior approach it had influenced the and the of the operative time in the meta-analyses that included this a study outcomes that compared intraoperative outcomes and outcomes to after and reported only from this however, were included in the of pain at 2 and and at of third an that included patients but in a this patients in These are not the only we but are some to the that of meta-analyses in the the many we in the meta-analyses and the level of to our we it to that based on the evidence, in the of an using the direct anterior approach may in with the at to and less pain during time to approximately 2 after surgery. the of these differences may not be as important as and the of the direct anterior approach may be This bias is not when this is in the literature and in and patients may be by the and of a The operative time of with the direct anterior approach is to be even with surgeons who are with the for surgeons who have of these however, it is that with the direct anterior approach may the same of time as with the posterior approach. additional meta-analyses two were meta-analyses on RCTs, of which in or than the studies we have included The systematic search of PubMed in April for publications from to April in two additional meta-analyses of RCTs that and meta-analyses are to provide the level of to through the of there have been many meta-analyses on the of the of the direct anterior approach in to conventional approaches, because of the surgeons are with the of the direct anterior approach is the on our in meta-analyses was often not discussed. Aside from the such as the and the characteristics of the patient and differences in surgical other than approach, one major was the of learning curve, which has not been standardized in studies. RoB in the RCTs was often it was biases may affect the was that the primary studies used in these meta-analyses were not powered to detect differences in these when these studies reported no difference in an outcome that the studies were not powered to it not be to that the approach was This has been many studies that on faster without the studies to look at or but potentially these potential with patients include of not clinically differences in of and surgeons should that the approach may be associated with a of potentially important should so on differences of a few or of recovery, and studies should their analyses to look at clinically differences in patient-reported outcomes using and to look at less to that we are not of at the of a more risk of a serious and to The level of we in our based on the is but this was not the only we The that the RoB and are commonly is more the study of the included primary studies was and there have been not such as the learning of the direct anterior approach, no serious had been to the of RoB and heterogeneity, which it more difficult to the One of the to definitive on the of one approach versus has been the of one over another. The posterior approach has been a commonly used approach for for early it was associated with a the of the direct anterior approach, and with the of the and there has been a toward adopting this approach. This important bias in patient discussion of potential of the evidence, and the of review and systematic reviews, as in this is the of clinical to the meta-analyses included in this the direct anterior approach was associated with after of approximately 2 to compared with other conventional approaches that are no published data is a clinically important difference of the others have the of the to be to without clinical others that an early of in a clinical these it is not that the reported in of less than are clinically it is not to patients from they had the direct or lateral approach, meta-analyses are prone to the and bias because of the with the direct anterior approach. the of one approach is to the first be to for a surgical approach from the and to the We that this study is a first in that the in this the orthopaedic community to determine is important for the patient and the healthcare and then on an that is powered for the functional outcomes for the first to more or is less use of analgesics for to 2 more we this with the potentially surgical time and a there in which a conventional approach may be more for some patients the direct anterior approach may be more for It as that an powered will be on topics such as In the absence of evidence, based on the clinical results we have shown surgeons should not to to a approach but should in their to their The of faster does not without a and the patient to be informed of objective and to be in the of making a on this we that surgeons with their patients the clinical of an of 2 to in at after surgery and the to this with the potential trade-off of a risk of nerve injury, that of the lateral nerve of the and the risk of femoral systematic reviews of RCTs, we also that there is a for analysis based on of these and on the of the surgical time it to a hip using the direct anterior approach. on the differences in the recent in from conventional approaches to the direct anterior approach is not We that and of studies may have this It is our that this study will these results in and will surgeons who are with conventional approaches to the for and also to their patients to the of the by the direct anterior approach. Despite we that there will be a in the direct anterior approach due to with some surgeons to their approaches they with complications or with a steep learning One potential toward the direct anterior approach is the that this is the only approach that after THA, which is not as the of our patients who and are on the of surgery have their surgery using the posterior approach, and only a with the direct anterior approach. this may increased of the direct anterior approach. it is of surgery, such as will the between the posterior and direct anterior For example, reduce the risk of of the direct anterior approach), this may the more toward the posterior approach, which has a shorter surgical at this time, there is no that that this is an as a for arthroplasty the number of that may or may not but to be surgeons and the clinical that they

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.