In THA, the choice of surgical approach remains controversial [10,12,14,18,19,27,28]. In the 2000s, the miniposterior approach (MPA) saw wide adoption based on studies demonstrating that it provided rapid recovery with a low risk of complications . More recently, there has been a suggestion that the direct anterior approach (DAA) leads to less muscle damage [14,20]. However, there is little high-quality evidence to indicate whether the DAA accelerates recovery, or whether this approach—which may be technically more demanding—is associated with more complications [16,20,23].
To shed some light on the ongoing controversies with respect to clinical and patient-reported outcomes, radiographic results, and complications among patients, we performed a randomized trial to compare the DAA and the MPA. Because our primary endpoints involved questions about short-term recovery, rather than reconstructive durability, we felt it reasonable to report our results at 1 year of followup rather than the more conventional 2 years.
In this study, we asked four specific research questions: (1) Does the DAA result in faster return to activities of daily living than the MPA? (2) Does DAA have superior patient-reported outcome measures than the MPA? (3) Does the DAA result in improved radiographic outcomes than the MPA? (4) Does the DAA have a higher risk of complications than the MPA?
Patients and Methods
A nonblinded, prospective, randomized clinical trial was designed to compare the outcomes of the DAA and the MPA for primary THA. One hundred sixteen patients undergoing primary THA between March 1, 2013, and May 31, 2016, at our tertiary medical center were recruited. Fifteen patients did not receive the allocated intervention as a result of patient withdrawal from the study. As a result of an inability to follow the patients who withdrew as a result of an institutional review board rule, a per-protocol analysis was performed. Therefore, the final cohort consisted of 101. The institutional review board of our institution approved this study (12-001341) and it was registered with clinical trials.gov (NCT01613508); all investigations were conducted in conformity with ethical principles of research. The patients were followed, and data were collected, from the preoperative setting, in-hospital, at 2 weeks, 8 weeks, and at 1 year after THA. The primary focus of this study was to determine if there is a difference in early function or complications between the DAA and MPA, not survivorship of implants. Therefore, we chose to end the study at minimum 1 year of followup.
Patients having a consultation appointment for unilateral hip osteoarthritis (OA) were identified as potential study recruits by study coordinators. After identification, diagnosis of end-stage unilateral hip OA was confirmed by interpretation of radiographs and clinical examination by the study investigators. At the time of consultation, the study was described to the patient by the surgeon and study coordinator. A consent form was provided to the patient that stated the background and reasoning for the study. To minimize potential patient bias, each patient was counseled that both techniques are clinically successful. The hypothesis of the study was not discussed with the patients, only that a comparison of the techniques was being conducted (Table 1).
The 101 patients randomized and included in this study were recruited from the practices of the four participating surgeons. After randomization, all anterior approaches were done by one high-volume DAA surgeon (MJT), regardless of who consulted on the patient initially. Conversely, posterior approaches were done by a MPA surgeon (RTT, RJS, or MWP) even if they were the DAA surgeon’s patient. The initial consulting surgeon continued to care for the patient postoperatively regardless of the approach that was performed.
Direct Anterior THA
For the DAA technique, a specialized table with fluoroscopy was utilized. The specific technique is as described by Taunton et al.  with capsulotomy and repair. Before initiation of this study, the surgeon in the DAA arm of the trial had performed > 500 THAs with the DAA technique.
For the MPA technique, the hip capsule and external rotators were incised as one layer and repaired formally at conclusion of THA as described by Pagnano et al. . Before initiation of this study, all of the surgeons in the MPA arm of the trial had performed > 500 THAs with the MPA technique.
All patients received the same hemispherical uncemented acetabular component (Pinnacle®; DePuy Orthopaedics Inc, Warsaw, IN, USA) and the same uncemented hydroxyapatite-coated femoral stem (Corail®; DePuy Orthopaedics Inc) with a Biolox® delta ceramic femoral head (CeramTec GmbH, Plochingen, Germany) and highly crosslinked polyethylene acetabular bearing surfaces. All components were US FDA-approved.
All patients received 1 g tranexamic acid at incision and at closure. All wounds were closed by the midlevel provider for that surgeon using the same suture and closure method.
Patients came into the hospital on the day of their procedures. The patients received one preoperative dose and three postoperative doses of IV antibiotics separated by a 6- to 8-hour period. All patients received aspirin for deep vein thrombosis prophylaxis unless they reported a history of previous venous thromboembolism or were on preoperative anticoagulation medication.
The mean operative time for the DAA was 70 ± 16 minutes versus MPA at 61 ± 18 minutes (p = 0.01). The mean anesthesia time for the DAA group was 137 ± 19 minutes versus the MPA at 129 ± 21 minutes (p = 0.07).
During the study period, hospitalization of 2 to 3 nights was routine. The patients were discharged when they were independent and safe with gait aids, were able to transfer out of and into bed from a standing position, rise to and from a chair to a standing position, and ambulate 100 feet.
Every patient received the same formal preoperative class educating them on perioperative expectations. Patients received the same comprehensive multimodal pain management approach, including an indwelling psoas nerve catheter for 36 hours postoperatively, and an oral pain regimen, including scheduled acetaminophen with tramadol and short-acting opioid medication on an as-needed basis.
Both treatment groups had identical postoperative care. Patients were treated on the same ward and seen by the same physical therapy (PT) team; no specific hip precautions were given to either group. Structured PT began the day after surgery and continued during the hospitalization. Patients were encouraged to sit up at the bedside the evening of their surgery. On postoperative Day 1, the patients began ambulation with the assistance of PT with a walker or crutches as well as active ROM. Weightbearing was progressed as tolerated. A home therapy program was given to the patient although formal PT did not continue on an outpatient basis.
The patients were instructed to progress ambulation from a walker when they were able to walk stable without pain and then to continue with a crutch or cane until they were able to walk without a limp. The patients were encouraged to maximize independent ambulation and increase daily distance ambulated.
All patients received a phone call at 2 weeks to discuss progression of activities, pain control, and any postoperative issues or complications. At that time, the patients were also mailed the activity monitors for 3 days. The first postoperative visit was at 8 weeks and the second was at 1 year (Table 2).
In hospital, the amount of morphine equivalents was recorded. The distance ambulated with each PT session, average visual analog scale (VAS) pain score at rest while in-hospital, and any complications during hospitalization were recorded.
Each patient received a milestone diary to take home to record the postoperative day when the patient discontinued using crutches or a walker, discontinued the use of a cane, the date they stopped taking oral narcotics, resumed various independent activities of daily living, and was able to ambulate 6 blocks. If patients who returned were employed, the day returned to work was recorded. All patients in both groups filled out the diaries completely.
Activity in study participants was measured with custom, wearable, activity-monitoring sensors designed and developed at the Mayo Clinic [6,11]. Each self-contained sensor contains a triaxial MEMS accelerometer (analog, ± 16 g; Analog Devices, Norwood, MA, USA), microcontroller (12-bit ADC; Texas Instruments, Dallas, TX, USA), power source (Tadiran battery; semiconductor voltage regulator; Tadrian Batteries, Lake Success, NY, USA), and onboard data storage (NAND flash memory, 4-Gb memory chip; Micron, Boise, ID, USA). The battery life capacity for each device is 10 continuous days with a sampling frequency of 100 samples per second. Five activity-monitoring sensors were worn by the participant. Monitors are secured to both ankles and both thighs using Velcro straps, and one additional monitor is placed at the torso level on the participant’s pants with a belt clip.
Within the 2 months immediately before the scheduled date of their surgery, each participant received in the mail five activity-monitoring sensors. Instructions for correct placement and orientation of the activity monitors were included and participants were asked to wear the monitors continuously for 3 days, removing them only for sleeping and bathing. Participants were not required to restrict or enhance their activities during wear time. The monitors collected data over a 3-day period. After wearing the activity monitors for 3 days (1 weekday and 2 weekend days), the participants were instructed to return the monitors to the study site in a prepaid mailing package. Data from the monitors were downloaded and activity levels were determined for each minute epoch throughout the day and separated into bins of inactivity, low-activity, medium-activity, and high-activity levels. Activity monitors were sent home (or mailed) with each participant at the preoperative and 2-, 8-, and 52-week followup tests of lower extremity function. The data from the monitors were analyzed to determine three outcomes for each time point for the study: number of steps per day, percent of day active, and gait entropy. Gait entropy is a measure of randomness of gait. As entropy increases, gait becomes more fluid, random, and normal.
Hip disability and Osteoarthritis Outcome Score
Hip disability and Osteoarthritis Outcome Score (HOOS)  is a self-administered patient-reported outcome questionnaire containing adequate measurement qualities for test–retest reliability, floor and ceiling effects, construct validity, responsiveness, and interpretability, including patients with hip OA and THA.
The SF-12  is a self-administered questionnaire that has been validated for measuring and monitoring health status in large group studies. It has been published as the best measure for assessing general health for patients undergoing arthroplasty as noted in the analysis of the Swedish Registry .
Harris Hip Score
The Harris hip score (HHS)  is a clinician-based outcome tool that is frequently used for the evaluation of patients after THA. The indication for THA is particularly pain and impaired physical function, which are the two dominating domains in the HHS. However, there are ceiling effects that severely limit its validity.
Radiographic Outcome Parameters
Hip radiographs including an AP view and true lateral view plus a full pelvis profile were recorded preoperatively, postoperatively, at 2 months, and at the 1-year followup appointment. The radiographs were evaluated for component fixation, component position, and alignment with special emphasis on offset restoration and leg length measurement. Measurements of leg length, offset, acetabular abduction, and anteversion were calculated using validated techniques [24,28]. Two independent reviewers (TMM, MPA) who had not performed the surgical procedures used calibrated 8-week postoperative radiographs and digital templating software (OrthoCase™; Merge Healthcare, Chicago, IL, USA) for analysis.
Complications/Loss to Followup
Any serious complications that occurred from the surgery were documented. Sepsis, embolism, failure of primary wound healing, hemorrhage, prosthesis loosening, hip dislocation, skin necrosis, hematoma, approach extension, and periprosthetic fracture are possible complications. If for any reason a patient was lost to followup (will not return for office visits), there was a form completed to indicate this event.
Groups did not differ in mean age, gender, or mean body mass index (Table 3).
The CONSORT diagram outlines the progress through the phases of the parallel randomized trial of the two groups (Fig. 1). Initial power analysis indicated a sample size of 64 patients in each group would provide 80% power to detect a difference of > 6.0 days in the time to discontinue all walking aids. As a result of slow enrollment, the statistics department was directed by the investigators to unblind the data for interim analysis. Four patients withdrew from the study after randomization from the DAA group, and 11 withdrew after randomization from the MPA group. The institutional review board precluded further research on these patients. Therefore, a per-protocol analysis was performed. At a time when 101 patients were allocated, there was a 6.7-day difference in discontinuation of all gait aids between the two groups and therefore we elected to end enrollment. All patients completed 2-week and 8-week followup, and 100 patients have completed 1-year followup; one patient died at 6 months from a stroke unrelated to the procedure. The minimum followup was 365 days (mean ± SD, 627 ± 369 days).
Several factors determine the appropriate sample size for a scientific clinical investigation. The following criteria are believed to be relevant for this study: (1) selection of the level of difference between treatment results, if it exists, that the study desires to detect; (2) a sample of sufficient size to provide statistical validity at a power level of 80% and an α level of 0.05; (3) a study and database size that is manageable to ensure good data quality; and (4) consideration of expected participant attrition.
The two patient cohorts were followed prospectively and evaluated with specific functional, clinical, and radiographic outcome measures at 2 weeks, 2 months, and 1 year from surgery. The principal outcomes include the rate of early functional recovery after primary THA. A power analysis was performed on the basis of the milestone functional outcomes found in a previous study of the early results of miniposterior THA . In that study, the mean time to discontinue all walking aids (± SD) was 15 ± 12 days, the mean time to climb stairs was 8 ± 7 days, and the mean time to walk a half mile (0.8 km) was 28 ± 18 days. We chose the number of days to discontinue all walking aids as the primary outcome variable for the study. Assuming that similar variability would be observed in the current study, assuming a two-sided two-sample t-test at α = 0.05, with a sample size of 64 patients in each group would provide 80% power to detect a difference of > 6.0 days in the time to discontinue all walking aids. Allowing for an additional 10% as a result of patient dropout and loss to followup, the accrual goal for the study was 142 patients. However, as a result of slow recruitment, the study was suspended after enrollment of 101 patients. Therefore, this study was powered to detect a difference in days to discontinuation of all walking aids of ≥ 6.8 days.
Patient demographics and outcomes are described using mean ± SD if continuous and distributed approximately Gaussian or median (25th percentile, 75th percentile) if continuous but not Gaussian. Categorical variables are described as count (percent). We were also interested in the early functional outcome and quality of life as measured by the SF-12 and HOOS at 2 and 12 months. Given the total sample size of 101 patients, the Central Limit Theorem applies and the two treatment groups are compared on these outcomes using a two-sample t-test. For activity monitoring, HHS, HOOS assessment, and SF-12 assessments made over the followup period, p values were adjusted for multiple comparisons using a false discovery rate with the linear step-up method of Benjamini and Hochberg. In-hospital and perioperative complications such as periprosthetic fracture (intraoperative and postoperative), deep vein thrombosis, pulmonary embolus, neurovascular complications, infection, and mortality are compared using chi-square tests or Fisher’s exact tests if necessary and appropriate. Analysis of time to event outcomes such as dislocation, complications related to the surgical procedure, the need for revision surgery, and survival utilizes survival techniques such as the method of Kaplan-Meier and Cox proportional hazard models. All statistical tests are two-sided and p values < 0.05 are considered statistically significant. SAS Version 9.1 (SAS Institute Inc, Cary, NC, USA) was utilized.
Early Functional Recovery
There were small differences in early functional recovery, as measured by the number of days to achieve functional milestones, with slightly fewer days required to achieve some milestone functions in the DAA group versus the MPA group. In the primary outcome measure, days to walk without a gait aid, the DAA group achieved that milestone 7 days earlier than the MPA group (17 ± 8 days versus 24 ± 14, mean difference 7 [0-12], p = 0.04). Additionally, when comparing the DAA and the MPA groups, the number of days to walk without a walker was 10 ± 7 versus 15 ± 9 days (mean difference 5 [1-8], p = 0.01), and the number of days to ascend and descend stairs with a gait aid was 5 ± 5 versus 10 ± 12 (mean difference 5 [1-8], p < 0.01). There were no other differences in the other early functional milestones (Table 4).
There were slight differences in the in-hospital outcomes. The number of morphine equivalents taken in the hospital for the DAA patients was 100 ± 73 mg (0-278) versus 145 ± 102 mg (4-525) for the MPA patients (mean difference 45 [10-80], p = 0.01). The mean VAS pain scores in the hospital were 2 ± 1 (0-5) for the DAA versus 3 ± 1 (1-7) for the MPA (mean difference 1 [0-1], p < 0.01), but this difference is unlikely to be clinically important . The distance ambulated in the first PT session was 100 ± 75 feet (0-260) for the DAA versus 68 ± 62 feet (2-300) for the MPA (mean difference -32 [-60 to -5], p = 0.02). There were no differences in the other in-hospital outcomes (Table 5).
Activity monitoring was completed over a 3-day period preoperatively, at 2 and 8 weeks postoperatively, and at 1 year postoperatively. Two weeks after surgery, there were some differences between the cohorts, which slightly favored the DAA. Patients in the DAA group took more steps per day at 2 weeks: 3897 ± 2258 (737-11,010) versus 2235 ± 1688 (27-7450) (mean difference -1662 [-2453 to -871], p < 0.01). Additionally, the percentage of days active was higher for the DAA at 2 weeks and at 8 weeks: 11% ± 5% (4-25) for the DAA versus 7% ± 4% (2-17) for the MPA (mean difference -4% [-5% to -2%], p < 0.01) and 17% ± 6% (4%-31%) for the DAA versus 13% ± 6% (4%–25%) for the MPA (mean difference -4% [-5% to -2%], p = 0.01), respectively. There were no other differences in activity monitoring outcomes between the cohorts (Table 5).
There was no difference in the HOOS at any time point (Table 6). At 2 months, the HOOS scores for the DAA and MPA, respectively, were: symptoms 60 ± 12 (10-75) versus 57 ± 10 (30-75) (mean difference 3 [-1 to 8], p = 0.13), pain 63 ± 12 (28-75) versus 61 ± 12 (25-75) (mean difference 2 [-3 to 6], p = 0.54), ADLs 62 ± 11 (27-75) versus 61 ± 11 (27-75) (mean difference 1 [-3 to 6], p = 0.61), sports/recreation 52 ± 20 (-6 to 75) versus 51 ± 19 (8-75) (mean difference 0 [-8 to 8], p = 0.94), and quality of life 49 ± 19 (6-75) versus 45 ± 19 (-6 to 75) (mean difference 4 [-4 to 11], p = 0.34). There was no difference in the HHS at any time point (Table 5). There were no differences in SF-12 mental or physical component scores at 2 months and 1 year (Table 7).
There was no difference in any radiographic outcome between the two groups (Table 8). There was no subsidence, migration, or loosening of either the acetabular or femoral components noted in either group on 2-month or 1-year followup radiographs. In the DAA group, there were four stems in varus. In the MPA group, there were six stems in varus and two in valgus.
There was no difference between the DAA and the MPA groups in terms of complications (8% [four of 52] versus 10% [five of 49], odds ratio, 0.71; 95% confidence interval, 0.18-2.83; p = 0.33).
In the MPA group, we observed one dislocation (treated with closed reduction with no redislocations), one wound dehiscence (treated with dressing changes and oral antibiotics), two intraoperative calcar fractures treated with a cable, and one deep vein thrombosis treated with Xarelto® (Janssen Pharmaceuticals, Beerse, Belgium). Both intraoperative calcar fractures were healed at 1-year followup with normal bony ingrowth of the femoral prosthesis in all zones with no evidence of subsidence. In the DAA group, there was one dislocation (treated closed with no further dislocations), two wound dehiscences (both treated with open wound débridement and primary closure), and one fall at 3 months resulting in a pubic ramus fracture, which healed with protected weightbearing. No patients underwent blood transfusions in either group. There were no patients with postoperative in-hospital complications in either group.
The demands of patients, surgeons, and payers have led to a search for improved surgical approaches for primary THA. To make a fair comparison between surgical approaches for THA—such as the DAA and the MPA—studies should focus on clinical and patient-reported outcomes, radiographic outcomes, and the frequency of complications. This randomized clinical trial has demonstrated that both the direct anterior and posterior approaches provided excellent early postoperative recovery with a low risk of complications. The patients undergoing the DAA had slightly shorter times to achieve some milestones of function and some differences as measured by advanced, quantitative activity monitoring at 2 weeks postoperatively. There were no substantive differences in outcomes by 8 weeks postoperatively and no differences at 1 year postoperatively.
This study has a number of limitations. First, although the study has demonstrated a small difference in early functional recovery, the focus of the study is on that early outcome alone, and the followup of this study is just 1 year. Because the DAA is a less studied approach, longer term (> 1 year) complications may accrue, will be important to quantify, and may offset early benefits. To our knowledge, no other study has specifically looked prospectively at the long-term survivorship of THA through the DAA in comparison to the MPA. There could be an inherent increased risk for failures of the implant or increased risk of revision resulting from other factors when performed through the DAA. Further long-term followup is needed to assess for any accrued differences in survivorship or complications. Second, although the milestone diaries encourage patients to record their milestones as they are attained and help reduce recall bias, the research team is not at their home to directly observe the actual achievement of the various levels of independence. For those patients who are employed, variations in return to work could lead to variations in activity. To supplement those findings, we did record more objective findings in the early recovery period, including in-hospital outcomes and gait monitoring. Despite the randomized trial design, there could remain differences between the groups in regard to sociologic parameters (work status, social support, and living environment) that impact early recovery and were not accounted for in our methodology.
Additionally, without blinding of the patients, surgeons, or other care providers, expectations may have confounded the early functional improvements seen. This study was specifically designed to help combat some parts of that issue. The surgeon-crossover design ensured that the postoperative instructions and recommendations remained homogenous between the approaches performed. This also helped reduce referral bias because each surgeon had patients randomized. Additionally, the postoperative recovery instructions that were given to the patients were also the same for both approaches from all surgeons. We also note that only one surgeon performed all of the DAA approaches, whereas three surgeons performed the MPA approaches. All surgeons had been performing the approach of their choice and had performed > 500 THAs through that approach before study initiation. The surgeon-crossover design allowed surgical expertise to be focused on the approach, not an individual surgeon’s ability to perform different approaches equally as well. There was little clinical difference in operative or anesthesia time between the groups. There were no transfusions in either group. The results and complications of the three surgeons in the MPA group were homogenous among them, and no trend was noted in any preoperative characteristics or outcomes.
We also note that the activity monitoring was only performed over 3 days at each time point and without direct supervision. There was no formal gait analysis performed on any patient. There is a possibility that the patients altered their activity while wearing the monitors. The monitors were worn over 1 weekday and 2 weekend days to minimize the differential effects of return to work. The patients each met with a member of the gait laboratory and were personally instructed on how to wear the monitors and instructed to carry out activity just as they did on other days. Additionally, the activity monitors have been validated as a reliable method of measuring activity [6,11,15].
The final limitations pertain to methodological and statistical issues. First, the study was terminated prematurely because of slow enrollment. As a result of this, we directed the statistics department to unblind the data for interim analysis. A per-protocol analysis was performed. At a time when 101 patients were allocated, there was a 7-day difference in discontinuation of all gait aids between the two groups, which was the primary outcome variable that the study was initially powered on, and therefore we elected to end enrollment on a pragmatic basis. Lastly, we analyzed the data on a per-protocol basis; four patients withdrew from the study after randomization from the DAA group, and 11 withdrew after randomization from the MPA group; the institutional review board precluded further research on these patients. A per-protocol analysis refers to inclusion in the analysis of only those patients who strictly adhered to the protocol. Its results have been argued to show an exaggerated treatment effect. In contrast, the use of intention-to-treat analysis ensures maintenance of comparability between groups as obtained through randomization, maintains sample size, and eliminates bias .
The DAA patients in this study discontinued gait aids a mean of 7 days earlier than the MPA group (17 versus 24 days), answering the primary research question. Christensen et al.  in another randomized controlled trial (RCT) also found that the DAA patients discontinued assistive devices earlier than the MPA patients (33 versus 43 days), whereas another RCT  found a similar but smaller difference (22 versus 28 days). However, both of those studies demonstrated times to discontinue gait aids that were greater than in this study. The variation in the studies is likely multifactorial, including patient demographics, therapy recommendations, and regional variations among others. In our study, there were only slight differences favoring the DAA in terms of stairclimbing. By contrast, in Barrett et al. , 50% of patients undergoing DAA were able to “do stairs normally” at 6 weeks, but only 18% of patients undergoing MPA were able to do this. This may reflect the difference in study design, because the MPA patients in the Barrett et al. study were counseled to follow traditional hip precautions, whereas the DAA patients were not .
In this study, there were no differences observed on the HOOS, SF-12, or the HHS at any time point between the two approaches. Another RCT  also demonstrated no difference in HHS at 6 weeks with a score of 97 for the DAA versus 93 for the MPA. Barrett et al.  found a higher HHS among DAA patients than MPA patients (90 versus 81 points); however, that difference is below the minimal clinically important difference for the HHS. Multiple studies [1,3,7,23], including this study, show few differences in patient-reported outcome measures at any time point after THA performed with different surgical approaches.
Both surgical techniques were able to reliably and precisely reproduce patient anatomy as measured by leg length and offset. Additionally, the placement of the acetabular component was also found to be reliable as measured both by the abduction angle and anteversion. There was no evidence of component loosening, subsidence, or malpositioning in either group. Hamilton et al. , in a radiographic review of 100 DAA hips and 100 posterior approach (PA) hips, demonstrated that the DAA had 92% of THAs in the Lewinnek safe zone of 5° to 25° of acetabular anteversion, whereas PA only 64% of the time. Utilizing our software technique, we found no difference in the proportion of patients whose acetabular components were anteverted in the “safe zone,” with both at 60%. Hamilton et al.  found that 90% of the DAA hips had between 30° and 50° of acetabular abduction versus 79% of the PA hips. Our study demonstrated little difference in abduction angles with 97% of the DAA and 87% of the MPA in this “safe zone.” There were no hips with an abduction angle of ≥ 55°. This demonstrates again, with adequate surgical expertise, both surgical approaches lead to satisfactory radiographic outcomes.
Complications of THA approaches, particularly of the DAA, have come to the forefront of academic debate. A recent American Association of Hip and Knee Surgeons poll found that the DAA now is the third most commonly used approach for routine THA, behind the posterior and anterolateral approaches . Although in this study we found no difference in the overall risk of complications between the DAA and the MPA, it is important to recognize that all surgical procedures in this study were done by high-volume, experienced hip surgeons. Therefore, the risk of complications reported in this study is not likely generalizable and is certainly not reflective of what might be expected when learning a new technique. Multiple prior reports from the early 2000s highlighted the incremental and sometimes unique complications that occurred during adoption of so-called minimally invasive THA approaches. With the DAA, there are concerns regarding wound healing, particularly in obese patients, because the incision is at or crosses the hip flexion crease. In this study, two patients underwent repeat surgery for wound healing problems after the DAA. Christensen et al.  found that 1.4% of hips developed wound complications after the DAA versus 0.2% with the MPA. Conversely, Watts et al.  found no overall difference in the risk of wound complications, but that females and patients who were obese (particularly those with a body mass index ≥ 40 kg/m2) were at increased risk for wound complications after DAA THA with a large proportion of wound complications resulting in reoperation. Recently Meneghini et al.  highlighted concerns about early femoral loosening with the DAA. Although we did not encounter that complication, it is certainly cause for concern. More robust data on complications in the future are likely to emerge from state, regional, and national joint registries and will have the advantage of both large numbers and a surgeon sample that is more reflective of broad practice. Complication data from subspecialized surgeons at academic centers may not be representative.
The patients who underwent the DAA achieved milestones of functional recovery earlier than those who underwent the MPA, but there were few differences observed by 8 weeks after surgery. In the context of the current debate around the relative merits of the DAA, the findings of this study likely represent a best case scenario for the DAA. The DAA procedures were performed by a skilled surgeon for patients who met study inclusion criteria, and yet the absolute difference in early outcome is modest. Surgeons and patients will have to interpret whether the potential to shorten the time to achieve some early functional milestones by 5 to 7 days is clinically meaningful. Given the relatively small size of the study and the expertise of the surgeons involved, it is unlikely that this study fully represents the risks of complications. Because the DAA is a relatively newer, less studied approach, longer term (> 1 year) complications may accrue and could offset early benefits. Further studies should identify means to improve early functional recovery without sacrificing long-term benefits of THA.
We thank William S. Harmsen HSR, Biomedical Statistics and Informatics, for assistance in study design, patient randomization, and statistical analysis. We thank Jennifer Krogman for assistance in operation of the study, study support, and research coordination. We thank Tad M. Mabry MD, and Matthew P. Abdel MD, for radiographic review.
1. Barrett WP, Turner SE, Leopold JP. Prospective randomized study of direct anterior vs postero-lateral approach for total hip arthroplasty. J Arthroplasty. 2013;28:1634–1638.
2. Berry DJ, Bozic KJ. Current practice patterns in primary hip and knee arthroplasty among members of the American Association of Hip and Knee Surgeons. J Arthroplasty. 2010;25:2–4.
3. Christensen CP, Jacobs CA. Comparison of patient function during the first six weeks after direct anterior or posterior total hip arthroplasty (THA): a randomized study. J Arthroplasty. 2015;30:94–97.
4. Christensen CP, Karthikeyan T, Jacobs CA. Greater prevalence of wound complications requiring reoperation with direct anterior approach total hip arthroplasty. J Arthroplasty. 2014;29:1839–1841.
5. Dunbar MJ, Robertsson O, Ryd L, Lidgren L. Appropriate questionnaires for knee arthroplasty. Results of a survey of 3600 patients from The Swedish Knee Arthroplasty Registry. J Bone Joint Surg Br. 2001;83:339–344.
6. Fortune E, Lugade V, Morrow M, Kaufman K. Validity of using tri-axial accelerometers to measure human movement–Part II: Step counts at a wide range of gait velocities. Med Eng Phys. 2014;36:659–669.
7. Graves SC, Dropkin BM, Keeney BJ, Lurie JD, Tomek IM. Does surgical approach affect patient-reported function after primary THA? Clin Orthop Relat Res. 2016;474:971–981.
8. Hamilton WG, Parks NL, Huynh C. Comparison of cup alignment, jump distance, and complications in consecutive series of anterior approach and posterior approach total hip arthroplasty. J Arthroplasty. 2015;30:1959–1962.
9. Harris WH. Traumatic arthritis of the hip after dislocation and acetabular fractures: treatment by mold arthroplasty. An end-result study using a new method of result evaluation. J Bone Joint Surg Am. 1969;51:737–755.
10. Howell JR, Masri BA, Duncan CP. Minimally invasive versus standard incision anterolateral hip replacement: a comparative study. Orthop Clin North Am. 2004;35:153–162.
11. Lugade V, Fortune E, Morrow M, Kaufman K. Validity of using tri-axial accelerometers to measure human movement–Part I: Posture and movement detection. Med Eng Phys. 2014;36:169–176.
12. Maffiuletti NA, Impellizzeri FM, Widler K, Bizzini M, Kain MSH, Munzinger U, Leunig M. Spatiotemporal parameters of gait after total hip replacement: anterior versus posterior approach. Orthop Clin North Am. 2009;40:407–415.
13. Meneghini RM, Elston AS, Chen AF, Kheir MM, Fehring TK, Springer BD. Direct anterior approach: risk factor for early femoral failure of cementless total hip arthroplasty: a multicenter study. J Bone Joint Surg Am. 2017;99:99–105.
14. Meneghini RM, Pagnano MW, Trousdale RT, Hozack WJ. Muscle damage during MIS total hip arthroplasty: Smith-Peterson versus posterior approach. Clin Orthop Relat Res. 2006;453:293–298.
15. Morrow MM, Hurd WJ, Fortune E, Lugade V, Kaufman KR. Accelerations of the waist and lower extremities over a range of gait velocities to aid in activity monitor selection for field-based studies. J Appl Biomech. 2014;30:581–585.
16. Nakata K, Nishikawa M, Yamamoto K, Hirota S, Yoshikawa H. A clinical comparative study of the direct anterior with mini-posterior approach. Two consecutive series. J Arthroplasty. 2009;24:698–704.
17. Nilsdotter AK, Lohmander LS, Klassbo M, Roos EM. Hip disability and osteoarthritis outcome score (HOOS)–validity and responsiveness in total hip replacement. BMC Musculoskelet Disord. 2003;4:10.
18. Ogonda L, Wilson R, Archbold P, Lawlor M, Humphreys P, O'Brien S, Beverland D. A minimal-incision technique in total hip arthroplasty does not improve early postoperative outcomes: a prospective, randomized, controlled trial. J Bone Joint Surg Am. 2005;87:701–710.
19. Pagnano MW, Trousdale RT, Meneghini RM, Hanssen AD. Slower recovery after two-incision than mini-posterior-incision total hip arthroplasty. A randomized clinical trial. J Bone Joint Surg Am. 2008;90:1000–1006.
20. Poehling-Monaghan KL, Taunton MJ, Kamath AF, Trousdale RT, Sierra RJ, Pagnano MW. No correlation between serum markers and early functional outcome after contemporary THA. Clin Orthop Relat Res. 2017;475:452–462.
21. Ranganathan P, Pramesh CS, Aggarwal R. Common pitfalls in statistical analysis: intention-to-treat versus per-protocol analysis. Perspect Clin Res. 2016;7:144–146.
22. Salaffi F, Stancati A, Silvestri CA, Ciapetti A, Grassi W. Minimal clinically important changes in chronic musculoskeletal pain intensity measured on a numerical rating scale. Eur J Pain. 2004;8:283–291.
23. Taunton MJ, Mason JB, Odum SM, Springer BD. Direct anterior total hip arthroplasty yields more rapid voluntary cessation of all walking aids: a prospective, randomized clinical trial. J Arthroplasty. 2014;29:169–172.
24. Tiberi JV PN, Kertzner M, Schmalzried TP. A more reliable method to assess acetabular component position. Clin Orthop Relat Res. 2012;470:471–476.
25. Ware J Jr, Kosinski M, Keller SD. A 12-Item Short-Form Health Survey: construction of scales and preliminary tests of reliability and validity. Med Care. 1996;34:220–233.
26. Watts CD, Houdek MT, Wagner ER, Sculco PK, Chalmers BP, Taunton MJ. High risk of wound complications following direct anterior total hip arthroplasty in obese patients. J Arthroplasty. 2015;30:2296–2298.
27. Woolson ST, Mow CS, Syquia JF, Lannin JV, Schurman DJ. Comparison of primary total hip replacements performed with a standard incision or a mini-incision. J Bone Joint Surg Am. 2004;86:1353–1358.
28. Wright JM, Crockett HC, Delgado S, Lyman S, Madsen M, Sculco TP. Mini-incision for total hip arthroplasty: a prospective, controlled investigation with 5-year follow-up evaluation. J Arthroplasty. 2004;19:538–545.