Comparison of direct anterior and posterior approaches in total hip arthroplasty within a mature surgical practice: A single surgeon analysis of complications and perioperative metrics

INTRODUCTION

Total hip arthroplasty (THA) is a frequently performed orthopedic procedure. Its annual volume is predicted to continue increasing over the next several decades.^[1,2] The two most common surgical approaches used in THA are the posterior approach (PA) and the direct anterior approach (DAA).^[3] Historically, the PA has been more commonly performed; however, the DAA is gaining popularity due to reported advantages in complication rates, short-term functional outcomes, and shorter discharge times.^[4-10]

Although both approaches are associated with relatively low complication rates, there is no definitive consensus on which surgical approach results in fewer overall complications and the breakdown of complication types by approach. Several studies have found no significant difference in dislocation rates between the two approaches,^[11-13] while a recent study reported a higher overall complication rate with the DAA compared to the PA.^[14] Findings on operative time are also inconsistent: Some studies report shorter times with the DAA, whereas others support the opposite.^[15-17] Many of these studies use aggregated data from procedures performed by multiple surgeons at different institutions and locations, introducing variability in surgical technique and expertise that may contribute to inconsistent findings regarding perioperative parameters, complications, and outcomes in THA.

To address this variability and the ongoing debate over the optimal approach in THA, we aimed to analyze and compare perioperative parameters, post-operative and intraoperative complication rates and types, readmission frequency, and discharge times up to 1-year post-operation between the two approaches in a mature hip practice. All operations were performed by a single, highly experienced, fellowship-trained surgeon at a single institution. Patient demographics were also recorded to assess any differences between the two groups.

MATERIALS AND METHODS

This was a single-surgeon, single-center, retrospective review of medical records from patients who underwent THA through either the PA between January 2011 and December 2013 or the DAA between January 2021 and December 2023.

The operating surgeon transitioned from the PA to the DAA in October 2013. During October through December 2013, both surgical approaches were performed, and only the PA cases were included in the study. To ensure consistency, all surgical cases included in this analysis were performed after the surgeon had completed more than 2000 cases of each approach type. All THAs were performed by the senior author.

Inclusion criteria consisted of patients aged 18–85 who underwent either an anterior or PA for THA. Exclusion criteria were patients with prior open procedures on the same hip and surgery performed at other healthcare centers. Due to incomplete medical records in the posterior THA cohort, some data were absent: American Society for Anesthesiologists (ASA) Physical Status Classification class (29 patients), body mass index (BMI) at time of operation (21 patients), operative time (47 patients), intraoperative blood loss (1 patient), and implant size (1 patient). Data on readmissions and complication events were available for all patients.

Based on our inclusion criteria, we included 244 posterior THA cases (2011–2013) and 250 anterior THA cases (2021– 2023) in the final analysis for age, sex, readmission rates, and complications. The final sample sizes for these parameters in the PA group were ASA class (n = 215), BMI (n = 223), operative time (n = 197), intra-operative blood loss (n = 243), and implant size (n = 243), for both the outer diameter of the femoral head and the internal diameter of the acetabular cup.

We examined patient demographics, including age, sex, BMI, and ASA class. Peri-operative data collected included operative time (from surgical incision to procedure completion), estimated blood loss, implant size, and days to discharge. Medical records were reviewed for any intraoperative complications. Admissions to the hospital for any reason were reviewed up to 90 days postoperatively. Readmissions to the hospital for any reason were included in this study, whether they were directly related to the THA procedure or not. Complications were assessed by manual chart review for up to 1 year post-surgery. A complication was defined by a noted event in the electronic medical record that occurred within 1 day up to 1 year postoperatively. Events must have had a documented encounter or hospital admission and were appropriately classified based on the type of complication. Complications were determined to be events related to the surgery, including any documented visit for prosthesis-related complications, infection, dislocation, and fracture. Complications related to surgery, including sepsis, pneumonia, or other acute events, were also determined to be complications per our methods, but were not found in this study.

To account for baseline demographic differences in age and sex between the cohorts, a binary logistic regression was done for the overall complication rate and 90-day hospital readmission rates. Surgical approach, age, and sex were chosen as independent variables. Results are reported as odds ratios (OR) with 95% confidence intervals (CI).

RESULTS

Table 1 presents the patient demographics for the two groups. There was a statistically significant difference in age between the DAA and PA cohorts (66.1 vs. 64.1, p = 0.03). A significantly higher proportion of female patients was in the PA group compared to the DAA group (62.7% vs. 51.2%, p = 0.01). No significant differences were found between the groups in terms of BMI or ASA class.

Compared to the PA group, the DAA group had a significantly shorter operating time (66.2 vs. 70.1, p = 0.0007), lower estimated blood loss (286.2 vs. 315, p < 0.0001), and a shorter time to discharge (1.1 vs. 3.2 days, p < 0.0001). There was no significant difference in femoral head diameter. A 40 mm femoral head was used in 27 (11.1%) PA cases and 0 (0%) DAA cases, a statistically significant difference (p < 0.0001). The 36 mm femoral head was more frequently used in the DAA group (186 cases, 74.4%) compared to the PA group (151 cases, 62.1%). A 28 mm acetabular liner was used in 4 (1.6%) PA cases and in none of the DAA cases, a significant difference (p = 0.042). Differences in the use of 32 mm femoral heads were not statistically significant. These findings are summarized in Table 2.

Hospital readmission rates were higher in the PA group across all time points, but did not reach statistical significance. At 1 week post-operatively, 1% of DAA patients were readmitted to the hospital, compared with 4% in the PA group. Within 3 weeks of their operation, 3% of DAA patients had been readmitted, and 7% of PA patients. Within 90 days, 8% of DAA patients had been readmitted, compared with 14% in the PA group. These differences in readmission rates were not statistically significant for any of the three time points [Table 3].

Post-operative infection occurred in 3 patients (1.2%) in the DAA group, and in none in the PA group; this was not statistically significant [Table 4]. There were no deep infections in either group. There were 3 (1.2%) dislocations that did not occur in the post-anesthesia care unit (PACU) in the DAA group and 3 (1.2%) with the PA (p > 0.05). Two (0.8%) DAA patients experienced PACU dislocations, whereas none occurred with the PA (p > 0.05). Periprosthetic fractures occurred in 3 patients (1.2%) who underwent DAA and 4 patients (1.6%) who underwent the PA, a non-significant difference. Deep vein thrombosis (DVT) occurred in 1 DAA patient (0.4%) and 1 PA patient (0.4%). One patient in the PA group had a cardiac complication not categorized above (0.4%). In total, 12 complications (4.8%) occurred in the DAA group and 9 (3.7%) in the PA group; this difference was not statistically significant.

Table 5 shows the surgical approach, age, and sex-adjusted regression analysis. None of the variables was independently associated with overall complication rates. Surgical approach was not associated with overall complication rates (OR 0.52, 95% CI 0.19–1.29) or 90-day hospital readmission rates (OR 2.28, 95% CI 0.92–6.17). Age and sex were not significant predictors of either outcome in the adjusted models.

DISCUSSION

Several reports have shown an increased incidence of periprosthetic fractures with the DAA compared with the PA, potentially due to reduced femoral exposure.^[19,20] A recent study found no difference in overall fracture incidence between approaches but did report a shorter time to fracture with the DAA.^[21] Although periprosthetic fracture rates are generally low for both approaches, the DAA appears to carry a higher risk. In our study, the DAA group had three periprosthetic fractures, and the PA group had four, a difference that was not statistically significant. One study analyzing cases performed by two surgeons also found no difference in the rate of periprosthetic fracture between approaches.^[22] More large-scale studies are needed to determine if a true difference in periprosthetic fracture risk exists. Reduced femoral exposure is a plausible explanation for the increased fracture risk with DAA, but additional variables, including surgical expertise, implant design, and patient type, may also influence this association.

Our study found a non-significant increase in post-operative superficial infection rate in the DAA group compared to the PA group (1.2% vs. 0%, p = 0.09), and there were no deep infections in either group. Other studies comparing infection rates between the two approaches have found that the DAA is associated with a higher infection rate.^[23,24] Interestingly, one study found no difference in overall wound complication rates but found that obesity was a stronger risk factor for wound infection in DAA patients than in those receiving PA THA.^[25]One possible explanation for the increase in infection with the DAA is the anatomical proximity of the incision to the inguinal area. This may be an important consideration in deciding on the surgical approach in obese patients with abdominal overhang. Future research should include multivariate analyses to better identify patient populations at higher risk for wound complications and characterize infection types.

Dislocation is one of the most common complications post-THA. We observed a dislocation rate of 1.2% in both groups. PACU dislocations were analyzed separately, occurring in 0.8% of DAA cases and 0% of PA cases. The current literature is heterogeneous regarding dislocation rates between the two approaches.^[5,11,26] Our findings align with studies that show no significant difference between the two approaches in terms of dislocation rate. A recent study reported higher dislocation rates in female patients compared to males following THA.^[27]In contrast, a recent meta-analysis found no difference in dislocation rates based on sex.^[28] Given the sex distribution disparity between our two experimental groups, it remains unclear to what extent this may have influenced dislocation rates in our study. As the DAA continues to gain popularity and surgical expertise with the approach grows, larger-scale studies may be possible, allowing the potential detection of differences in dislocation rates.

Although DVT is a known complication of THA, comparative data on DVT rates between the DAA and the PA remain limited. To the best of the authors’ knowledge, no data have demonstrated a significant difference in DVT incidence when comparing these two approaches. In our study, the DVT rate was similar in the two groups; the difference was not statistically significant, consistent with the existing literature.^[29,30] At present, there does not appear to be an increased risk of DVT associated with either approach.

In this study, the DAA was associated with a significantly shorter operating time (66.2 vs.70.1 min, p = 0.0007), lower estimated blood loss (286.2 vs. 315 mL, p < 0.0001), and a shorter time to discharge (1.1 vs. 3.2 days, p < 0.0001). A recent meta-analysis that pooled data from four randomized controlled trials found no significant difference in blood loss between the approaches.^[31] However, another study that examined cases performed by a single surgeon who had surpassed the learning curve reported significantly less blood loss with the DAA, findings that align with our study.^[32]Perioperative anemia is a known risk factor for post-surgical complications and mortality, highlighting the importance of blood loss in THA.^[33] In our study, TXA was administered only in DAA cases and not in PA cases. As an antifibrolytic agent, TXA may have contributed to the observed differences between the two groups. In addition, the procedures used in this study were drawn from two distinct time periods which may limit the robustness of findings related to surgical skill and expertise. Future research is warranted to determine whether there are significant differences in blood loss between DAA and PA.

Two recent meta-analyses have found that the DAA is associated with a longer operating time compared to the PA.^[6,16] However, procedure length is highly dependent on surgeon experience, institutional setting, and support staff, all of which can substantially influence operating time. While meta-analyses provide valuable aggregated data, they may not fully account for these key confounding variables. Our study, one of the largest single-surgeon, single-site analyses to date, likely minimizes these variables, as all THA procedures were performed by a fellowship-trained surgeon well beyond the learning curve with similar support staff present for all cases. It is worth noting that cases in this study were drawn from different time periods, and there is the possibility that surgical technique and skill may have biased procedure lengths. Future research should prioritize similar single-surgeon, single-site designs when comparing operative duration.

Time to discharge is a less objective metric of recovery, as it can be influenced by patient preference, hospital protocols, insurance factors, and other institutional trends. Existing studies present mixed results on discharge timing, with some reporting no difference or even longer stays with the DAA.^[34-37] In contrast, our study found a shorter average time to discharge with the DAA (1.1 days) compared to PA (3.2 days), a finding that aligns with select prior reports.^[10,38,39] A study analyzing 2.8 million THA procedures found that the average length of stay decreased from 4.06 days in 2002 to 2.75 days in 2013.^[40] Grosso et al. later reported an average length of stay of 2.7 days between 2014 and 2016, significantly shorter than that of a 2006–2009 cohort.^[41] In 2020, THA was removed from the Centers for Medicare and Medicaid Services inpatient-only list. Subsequently, a study found that outpatient THA cases increased by 1,392% between 2018 and 2021, with the average length of stay dropping from 1.91 to 1.36 days.^[42] The average discharge time of our DAA cohort was 1.1 days, consistent with recent trends. However, the two cohorts in our study were treated a decade apart, which may limit the validity of comparisons regarding discharge time due to changes in hospital protocols and practices over time. There is a clear trend in THA toward shorter lengths of stay and an increase in outpatient procedures. Future research should examine time to discharge within the context of consistent admission and discharge protocols to isolate the effect of surgical approach.

There was no statistically significant difference in acetabular component size between groups. While the overall distribution of femoral head sizes was similar, the PA group had significantly greater use of 28 mm and 40 mm heads. The surgeon stopped using 40 mm heads between 2013 and 2021 due to the associated increased incidence of trunnionosis. However, complication and dislocation rates were not significantly different between the two groups, which indicates that head size likely did not influence stability outcomes. Given the low number of dislocation events, multivariate modeling including head size was not statistically feasible. Overall, implant components were comparable between groups and unlikely to have influenced study outcomes.

In our study, there were no cases of major nerve palsy or significant limb length discrepancy (LLD) requiring treatment, both of which are known complications of THA. Although no patients were treated for LLD, we did not analyze individual outcomes. The absence of major nerve palsy may be due to the random distribution of patients between cohorts. In addition, the lack of findings may reflect a mature hip practice. In our study, there were no differences in complications between approaches. However, the PA remains the widely accepted strategy for THA. Due to limited femoral exposure with the DAA, complications such as femoral canal perforation and component undersizing are more common. In our series, these were not noted as complications, but more robust studies with larger sample sizes may be able to elucidate these.

Our study has several limitations. It is a single-surgeon, single-center, retrospective review of patient medical records. While this design allows for consistent surgical skill and minimizes variability, it is unclear how well these results may relate to findings outside of a single surgeon’s practice. The PA cohort was drawn from 2011 to 2013, and the DAA cohort from 2021 to 2023, introducing potential bias due to changes in medical recording and perioperative practices over time. A single surgeon performed all operations, and there is a 10-year difference between the cases used for analysis, which may lead to differences in surgical skill and expertise. This also introduces potential biases in surgical efficiency, surgical team proficiency, and surgical maturity, all of which may affect operative length and complication rates between the two distinct periods. Further, perioperative measures differed between the two groups: Fluoroscopy and TXA were used only in the DAA cohort, which limits the strength of our findings, particularly regarding blood loss.

In addition, the DAA group was significantly older and had a lower proportion of female patients than the PA group, which may have affected the results. To address this disparity, we conducted a binary logistic regression that adjusted for age and sex. After adjustment, the surgical approach remained non-significantly associated with both complication and readmission rates, supporting our primary findings. We only reviewed complications occurring within 12 months postoperatively; longer-term follow-up data are needed to better assess the relative effectiveness of the approach type. Finally, although our study’s sample size is comparable to other single-surgeon retrospective studies, it does not match the scale of registry-based analyses, which may limit the ability to determine “true” complication rates.

CONCLUSION

This study found that the DAA was associated with shorter operative time, reduced blood loss, and a shorter time to discharge than the PA. Overall complication rates were similar between the two approaches, with no significant differences in complication rates or types. Hospital readmission rates did not differ significantly. Results of this study should be interpreted as an association rather than causation due to the significant time gap in the surgical era in which cases for each approach were drawn.