The Journal of Arthroplasty 32 (2017) 883e890

Contents lists available at ScienceDirect

The Journal of Arthroplasty

journal homepage: www.arthroplastyjournal.org

Primary Arthroplasty

A Prospective Randomized Clinical Trial in Total Hip

ArthroplastydComparing Early Results Between the Direct Anterior
Approach and the Posterior Approach
Tze E. Cheng, MBBS, MS a, *, Jason A. Wallis, BPhty, MPhty (MSK) b,
Nicholas F. Taylor, BSc, BAppSc (Physio), PhD b, c, Chris T. Holden, MBBS, FRANZCR d,
Paul Marks, MBBS, FRANZCR d, Catherine L. Smith, MPH (Biostatistics) e,
Michael S. Armstrong, MBBS, FRACS f, 1, Parminder J. Singh, FRCS (Tr & Ortho), MS, FRACS f
a
Eastern Health Clinical School, Monash University, Victoria, Australia
b
Department of Physiotherapy, Eastern Health, Victoria, Australia
c
Department of Rehabilitation, Nutrition and Sports, La Trobe University, Victoria, Australia
d
Imaging Associates Box Hill, Victoria, Australia
e
Department of Epidemiology, School of Public Health and Preventive Medicine Monash University, The Alfred Centre, Victoria, Australia
f
Department of Orthopaedics, Eastern Health, Monash University, Victoria, Australia

a r t i c l e i n f o a b s t r a c t

Article history: Background: We report a prospective randomized study comparing early clinical results between the
Received 18 May 2016 direct anterior approach (DAA) and posterior approach (PA) in primary hip arthroplasty.
Received in revised form Methods: Surgeries were performed by 2 senior hip arthroplasty surgeons. Seventy-two patients with
25 July 2016
complete data were assessed preoperatively 2, 6, and 12 weeks postoperatively. The primary outcomes
Accepted 22 August 2016
Available online 31 August 2016
were the Western Ontario McMasters Arthritis Index and Oxford Hip Scores. Secondary outcome mea-
sures included the EuroQoL, 10-meter walk test, and clinical and radiographic parameters.
Results: Data analyses showed no difference between DAA (n ¼ 35) and PA (n ¼ 37) groups when
Keywords:
total hip arthroplasty
comparing total scores for primary outcomes. No significant differences were observed for 10-meter walk
THA test, EuroQoL, and radiographic analyses. Subgroup analysis for surgeon 1 identified that the DAA group
direct anterior approach had shorter acute hospital stay, less postoperative opiate requirements, and smaller wounds. However,
posterior approach this was offset by increased operative time, higher intraoperative blood loss, and weaker hip flexion at 2
prospective randomized trial and 6 weeks. Subgroup analysis of items on the Western Ontario McMasters Arthritis Index and Oxford
Hip Score identified that hip flexion activity favored the DAA group up to 6 weeks postoperatively. There
was an 83% incidence of lateral cutaneous nerve of thigh neuropraxia at the 12-week mark in the DAA
group. No neuropraxias occurred in the PA group. One dislocation occurred in each group. A single
patient from the DAA group required reoperation for leg-length discrepancy.
Conclusion: DAA total hip arthroplasty (THA) has comparable results with PA THA. Choice of surgical
approach for THA should be based on patient factors, surgeon preference, and experience.
© 2016 Elsevier Inc. All rights reserved.

Primary total hip arthroplasty (THA) is one of the most suc-

One or more of the authors of this paper have disclosed potential or pertinent
conflicts of interest, which may include receipt of payment, either direct or indirect,
institutional support, or association with an entity in the biomedical field which
may be perceived to have potential conflict of interest with this work. For full
disclosure statements refer to http://dx.doi.org/10.1016/j.arth.2016.08.027.
* Reprint requests: Tze E. Cheng, MBBS, MS, Eastern Health Clinical School,
Monash University, 5 Arnold Street, Box Hill, Victoria 3128, Australia.
1
Michael S. Armstrong was one of the operating surgeons in the study. Michael
sadly passed away after completion of the study.
cessful surgeries in modern medicine. The arthroplasty of the hip
joint has revolutionized treatment of osteoarthritis of the hip with
excellent long-term results [1]. Demand for THA continues to in-
crease with 32,306 primary conventional THAs performed in
Australia in 2014 [2]. Technological advancements paired with low
revision rates have seen candidates for THA becoming younger [3].
The demands of patients for THA are no longer solely related to the
resolution of hip pain and restoration of function but also a quicker
recovery [4].

http://dx.doi.org/10.1016/j.arth.2016.08.027
0883-5403/© 2016 Elsevier Inc. All rights reserved.
884 T.E. Cheng et al. / The Journal of Arthroplasty 32 (2017) 883e890
The common surgical approaches utilized for hip arthroplasty
are broadly divided into the posterior, lateral, and anterior ap-
proaches. Each approach has evolved to remain viable for clinical
practice. Direct anterior approach (DAA) THA has an acceptable
complication profile, a low dislocation rate, and potentially earlier
return to function [5]. Improvements in outcomes associated with
DAA THA are supported by visual, radiological, and biological
findings of reduced muscle damage owing to the intermuscular and
internervous interval utilized [6-8].
Early functional outcomes following DAA THA when compared
with other approaches have been reported. Retrespo et al [9]
identified significant differences in Western Ontario McMasters
Arthritis Index (WOMAC), and Short Form 36 (SF-36) results up to
1-year postoperatively favoring the DAA compared to the lateral
approach (LA). Gait analysis studies have also identified differences
in gait parameters when comparing the DAA to other approaches.
Mayr et al [10] reported that the DAA THA resulted in a larger
number of improved gait parameters than the anterolateral group
6-12 weeks after surgery. Rathod et al [11] reported that the DAA
group had better improvement in hip internal and external rotation
following THA when compared to the posterior approach (PA).
With objective evidence favoring the DAA over the LA, the
current emphasis on research has been to examine potential dif-
ferences in functional outcomes between the DAA and PA. Recent
randomized clinical trials reported significant improvements in
length of hospital stay, pain, early mobility, and cessation of gait
aids in favor of the DAA when compared to the PA [12-14].
Despite its potential benefits, current DAA research reports
increased complication rates with early surgeon experience
particularly during the learning curve [15]. Complications reported
include iatrogenic fractures, neuropraxia, implant migration, and
early revision [16-20]. Registry data from the United Kingdom and
New Zealand identify that the majority of primary THA surgery is
performed using the PA and LA with less than 5% of surgeons using
the DAA [21,22].
Therefore, there has been growing interest in the DAA approach
due to perceived advantages compared to conventional surgical
approaches. The primary aim of our study was to compare patient-
reported outcome measures for patients undergoing DAA and PA
THA. The secondary aims were to compare objective physical
function, health-related quality of life, musculoskeletal impair-
ments, and radiological and clinical outcomes.
Methods
A randomized clinical trial was conducted between March 2014
and March 2015. Two senior hip arthroplasty surgeons experienced
in both DAA and PA THA surgery performed all surgeries. Both
surgeons were trained in the PA and have adopted the DAA into
their routine practice since 2010.
Ethical approval was obtained through the Eastern Health
Human Research Ethics Committee. The trial was prospectively
registered in the Australian New Zealand Clinical Trials Registry
ACTRN12614000131651.
The inclusion criteria for the study were unilateral symptomatic
hip osteoarthritis, Dorr's femur classification A/B, American Society
of Anesthesiologists (ASA) score 3 or less, a body mass index (BMI)
less than 35 kg/m2, and age between 40 to 75 years. Participants
were excluded if they had Dorr's femur classification C, previous
hip surgery (excluding arthroscopy), anticipated complex primary
THA, previous joint arthroplasty, were unwilling to accept
randomization and blinding, or had severe pathology that would
affect postoperative participation such as neurologic, psychiatric, or
other confounding preexisting musculoskeletal disorders.
Participants were recruited from the health service's outpatient
clinic and elective surgical waiting list.
All participants meeting the inclusion criteria provided written
informed consent for the research and both surgical approaches.
Participants were then stratified by age (<65 or 65 years), surgeon
(1 or 2), and randomly allocated to either DAA or PA groups within
strata using a concealed method. An independent researcher not
involved in participant recruitment, treatment, or assessment
prepared the randomization sequence with allocation prepared in
sequentially numbered opaque envelopes. Surgeons and the pri-
mary investigator were blinded to the approach until the preop-
erative planning meeting while participants were blinded
preoperatively.
Surgical Technique
Implants utilized were the R3 acetabular system and Anthology
femoral stem. Weight-bearing surfaces used were either ceramic on
ceramic Biolox Delta or Oxinium on polyethylene (Smith &
Nephew, Memphis, TN).
Preoperative templating was performed using IMPAX Ortho-
paedic Tools 2.52 by AFGA Healthcare NV, Belgium. Acetabular
component templating aimed to achieve 40 ± 10 of inclination.
Femoral component templating aimed to maintain a neutral
alignment within the femoral canal and adequate offset.
An orthopedic traction table (Maquet, Rastatt, Germany) was
utilized for all DAA THAs. The anterior incision begins 3-cm
posterior and distal to the anterior superior iliac spine,
extending distally approximately 10 cm over the tensor fascia
lata. Hueter's interval was then identified and developed to gain
access to the hip joint. A capsulotomy and femoral neck
osteotomy was performed. This was followed by the retrieval of
the femoral head and repositioning of retractors to expose the
acetabulum. Sequential reaming and acetabular component
implantation was conducted and verified under fluoroscopy.
Femoral preparation was undertaken with the leg extended,
externally rotated, and adducted. A superior capsulotomy was
performed to aid in femoral exposure. Femoral broaching and
trials were performed with fluoroscopic assistance. Definitive
implantation of the remaining prosthesis was undertaken with
routine capsular and wound closure.
PA surgery was performed with the patient adopting a lateral
position on a standard surgical table. The curvilinear incision 10- to
15-cm long centers over the posterior third of the greater
trochanter. Dissection through the fascia in line with the fibers of
gluteus maximus was conducted to reach the short external rota-
tors. With the piriformis muscle identified, the short external ro-
tators and hip capsule were tagged and reflected. Subsequent hip
joint dislocation was followed by a femoral neck osteotomy at the
templated level. Acetabular and femoral preparations were then
performed in a routine manner. Definitive implants were trialed
and inserted under direct vision. An enhanced intraosseous short
rotator and capsular repair was performed for all cases.
Perioperative and Postoperative Protocol
Similar intraoperative local-infiltration anesthetic protocols
were utilized in both DAA and PA groups based on a modification of
Kerr's technique [23]. A concoction of 0.2% ropivocaine with 30-mg
ketorolac and 1% adrenaline was used. Ketorolac was not used in
patients with evidence of renal impairment. Continuous infusion
pumps were employed on the ward up to 24 hours postoperatively.
All participants received prophylactic antibiotics (cephazolin) and
thromboprophylaxis (daltaparin) in accordance with the health
service's protocols.
 T.E. Cheng et al. / The Journal of Arthroplasty 32 (2017) 883e890 885

Fig. 1. Consort diagram of patient enrollment characteristics. DAA, direct anterior approach; PA, posterior approach.

All patients were mobilized the day after surgery. Routine hip
precautions (avoidance of combined hip flexion >90 and internal
rotation past the neutral plane) were instituted for the PA group. The
DAA group did not have restrictions to hip movement. The target day
of discharge for home or transfer to rehabilitation was the third
postoperative day. This was assessed daily by physiotherapists and
physicians supporting the orthopedic team. Patients not meeting the
discharge requirements were transferred to a rehabilitation facility.
Outcome Measures
Participants were evaluated preoperatively and at 2 weeks, 6
weeks, and 12 weeks postoperatively following THA. Follow-up
was conducted by the same investigator. Primary outcome mea-
sures were the WOMAC and Oxford Hip Score (OHS) question-
naires. A low WOMAC score and high OHS score are indicative of
better hip function [24,25].
Secondary outcome measures included assessments of health-
related quality of life using the EuroQol five dimensions question-
naire (EQ5D), functional performance using the 10-meter walk test
(10mwt), hip flexors and abductors strength using the Medical
Research Council grading scale [26-28]. EQ5D utility scores were
Table 2
Surgical Characteristics of Participants by Treatment Group.
Characteristic DAA Frequency (%) PA Frequency (%) P Valuea

Anesthetic .3
Table 1 General 17 (49) 17 (45)
Demographic Characteristics of the Participants by Treatment Group. Spinal 10 (29) 17 (45)
General and spinal 8 (22) 4 (10)
Characteristic DAA (%) PA (%)
Anesthetic rating .3
Male 33 (45%) 15 (43) 18 (47) 1 5 (14) 9 (24)
Female 40 (55%) 20 (57) 20 (53) 2 28 (80) 24 (63)
3 2 (6) 5 (13)
DAA Median (IQR) PA Median (IQR)
Transfusion (no. of packs) .9
Age at surgery (y) 59 (54, 69) 62.5 (55, 69) 0 32 (91) 35 (92)
Preoperative height (cm) 170 (162, 177) 166 (162, 174.5) 1 1 (3) 0 (0)
Preoperative weight (kg) 81 (71, 87) 82 (67, 91) 2 2 (6) 2 (5)
Preoperative BMI 27.7 (25.8, 30.0) 28.3 (24.8, 31.1) 3 0 (0) 1 (3)

DAA, direct anterior approach; PA, posterior approach; IQR, interquartile range; DAA, direct anterior approach; PA, posterior approach.
a
BMI, body mass index. P value based on Fisher's exact test.
886 T.E. Cheng et al. / The Journal of Arthroplasty 32 (2017) 883e890

Table 3
Comparison of WOMAC, OHS, and Secondary Outcomes Between Treatment Groups at Each Time Point.

Variable Baseline: Mean (SEM) 2 wk: Mean (SEM) 6 wk: Mean (SEM) 12 wk: Mean (SEM) Test for
(Group  Time)
DAA PA DAA PA Pa DAA PA Pa DAA PA Pa
Interactionb

WOMAC pain (0-20) 13.1 (0.6) 14.6 (0.57) 7.5 (0.71) 7.5 (0.68) .94 3.8 (0.56) 3.7 (0.55) .86 1.7 (0.46) 2.3 (0.45) .33 0.35
WOMAC stiffness (0-8) 5.4 (0.29) 6.1 (0.28) 3.3 (0.33) 3.6 (0.31) .64 2.4 (0.28) 2.0 (0.27) .39 1.4 (0.28) 1.8 (0.27) .27 0.11
WOMAC function (0-68) 44.5 (1.86) 50.5 (1.78) 29.5 (2.16) 33.4 (2.08) .2 13.0 (1.78) 16.3 (1.72) .2 6.0 (1.4) 8.7 (1.36) .17 0.76
WOMAC total (0-96) 63.0 (2.59) 71.2 (2.48) 40.3 (3.01) 44.5 (2.89) .33 19.2 (2.47) 22.0 (2.4) .43 9.1 (2.05) 12.8 (1.99) .2 0.70
OHS (0-48) 19.1 (1.13) 14.5 (1.08) 28.5 (1.56) 26.8 (1.5) .44 39.8 (1.05) 37.3 (1.01) .10 43.8 (0.87) 42.8 (0.84) .39 0.14
EQ5D 0.4 (0.05) 0.3 (0.05) 0.6 (0.04) 0.5 (0.04) .16 0.8 (0.03) 0.8 (0.03) .86 0.9 (0.02) 0.9 (0.02) .57 0.53
EQ5D VAS (0-100) 61.2 (3.29) 59.1 (3.16) 74.0 (2.7) 74.1 (2.59) .98 86.6 (1.63) 87.0 (1.58) .84 91.6 (1.31) 91.9 (1.27) .87 0.95
10mwt normal (m/s) 1.1 (0.04) 1.1 (0.04) 0.9 (0.04) 0.8 (0.04) .45 1.2 (0.04) 1.2 (0.04) .55 1.3 (0.03) 1.3 (0.03) .85 0.50
10mwt fast (m/s) 1.5 (0.06) 1.4 (0.06) 1.1 (0.05) 1.1 (0.05) .48 1.6 (0.04) 1.6 (0.04) .9 1.7 (0.04) 1.7 (0.04) .78 0.75

WOMAC, Western Ontario McMasters Arthritis Index; OHS, Oxford Hip Score; SEM, standard error of the mean; DAA, direct anterior approach; PA, posterior approach; mwt,
meter walk test.
a
P value obtained from pairwise comparison of treatment groups at each time point using linear mixed effects model.
b
P value based on F test for the group  time interaction.

calculated using United Kingdom weights. Length of stay, surgical
time, 2-week opiate analgesic requirements, postoperative hemo-
globin levels, and adverse outcomes were also analyzed. Adverse
outcomes were classified as major or minor. Major adverse events
included fractures, dislocations, and reoperations. Minor adverse
events included cutaneous nerve neuropraxias; deep vein throm-
bosis; bursitis, tendonitis, and minor wound pathology. Standard-
ized weight-bearing anteroposterior pelvic, Charnley, and lateral
radiographs were obtained for all participants at the immediate
postoperative and 6-week time points. Analysis focusing on
acetabular inclination and femoral stem orientation was performed
by 2 blinded independent consultant radiologists.
Statistical Analysis
The study was powered to detect a 10-point difference in
WOMAC score, assuming a standard deviation of 13.7 with 80%
power and a 2-tailed significance level of 5% [29]. Allowing for 10%
attrition, the study aimed to recruit a total of 70 participants.
Statistical analyses were conducted according to the intention to
treat principle, with all available data analyzed according to group
allocation. Missing outcome data were handled via restricted
maximum likelihood estimation within linear mixed models.
Linear mixed effects models were used to analyze continuous
outcomes at all time points. The outcomes (WOMAC, OHS, EQ5D,
and 10mwt) were categorized as the dependent variable.
Group (DAA or PA), time (preoperative, 2, 6, and 12 weeks), and
group  time interactions were categorized as the predictor
variables. Group and time were considered fixed effects while pa-
tients were classified as random effects. The P value for the
treatment  time interactions was reported together with pairwise
comparisons between treatment groups at each time point.
Abductor and straight leg raise function were analyzed as ordinal
outcomes using ordered logistic regression (Odds Ratios [ORs])
adjusting for “clustering” of repeated measures.
The percentage of participants using gait aids was compared
between groups at each time point using risk ratios obtained from
binomial regression (using generalized estimating equations). The
model included group, time, and a (group  time) interaction as
predictors.
Clinical parameters and radiographic data were compared be-
tween groups using Fisher's exact test for categorical outcomes. The
2 sample t test or Wilcoxon's rank-sum test were used for clinical
parameters and radiological analysis, depending on data skew.
Incidences of adverse outcomes were compared using Fisher's
exact test with risk ratios and 95% confidence intervals (CIs)
calculated if possible.
Subgroup analysis was performed to explore the effect of sur-
geon and the effect of hip movement precautions on hip-related
activity based on relevant items of the WOMAC and OHS
questionnaires.
All statistical tests were 2 sided with a statistical significance
level of 5%. No adjustments for multiple testing were made. Find-
ings with P values of less than .05 were considered statistically
significant. All analyses were performed using Stata version 12
(StataCorp LP, College Station, TX).

Table 4
Summary of Clinical Parameter Outcomes by Treatment Group.

Clinical Parameters Combined Results Surgeon 1 Surgeon 2

DAA Median PA Median P DAA Median PA Median P DAA Median PA Median P Valuea
(IQR) (n ¼ 35) (IQR) (n ¼ 38) Valuea (IQR) (n ¼ 26) (IQR) (n ¼ 28) Valuea (IQR) (n ¼ 9) (IQR) (n ¼ 10)

Operative time (min) 125 (111, 138) 100 (93, 113) <.001 120 (110, 133) 99 (90, 110) <.001 135 (125, 145) 114 (100, 135) .05
Wound length (mm) 107 (88, 120) 135 (127, 155) <.001 111 (104, 125) 148 (132, 162) <.001 79 (74, 90) 125 (114, 135) <.001
Hemoglobin drop (g/L) 35 (29, 45) 31 (24, 37) .04 35 (29, 45) 30 (23, 36) .03 35 (31, 37) 36 (27, 37) .84
Length of stay in acute hospital (h) 77 (73, 118) 95 (76, 120) .15 77 (73, 118) 101 (78, 121) .04 80 (76, 100) 77 (75, 77) .25
Total hospital stay (h) 96 (74, 127) 100 (76, 190) .15 99 (74, 147) 103 (78, 206) .11 80 (76, 100) 77 (75, 79) .57
Postoperative opiate analgesia 264 (193, 476) 406 (275, 565) .04 257 (218, 476) 354 (271, 544) .15 325 (180, 385) 530 (478, 668) .07
(morphine/mg)
Dischargedb home 29 (83%) 28 (74%) .4 20 (77%) 19 (68%) .55 9 (100%) 9 (90%) 1.0
Dischargedb rehab center 6 (17%) 10 (26%) .4 6 (23%) 9 (32%) .55 0 (0%) 1 (10%) 1.0
Rehab stay duration (h) 155 (72, 169) 166 (95, 211) .48 155 (72, 169) 166 (95, 189) .68 d 264 d

DAA, direct anterior approach; PA, posterior approach; IQR, interquartile range.
a
P value based on Wilcoxon rank-sum test.
b
P value based on Fisher's exact test.
 T.E. Cheng et al. / The Journal of Arthroplasty 32 (2017) 883e890
Table 5
Odds Ratios of Having a Higher Grade of Hip Strength Comparing the PA Group to the DAA Group.
Hip Activity Baseline: Odds 2 wk: Odds P 6 wk: Odds
Ratio (95% CI) Ratio (95% CI)
Straight leg raise 1.0 (0.4, 2.5) 2.7 (1.2, 6.5) .02
Hip abduction 0.8 (0.3, 1.8) 1.2 (0.5, 2.8) .76
PA, posterior approach; DAA, direct anterior approach; CI, confidence interval.
Results
Seventy-five patients (75 hips) were recruited and provided
written consent to participate in the study. Thirty-seven hips were
randomized into the DAA group and 38 hips were randomized into
the PA group. Surgeon 1 operated on 54 hips (DAA 26, PA 28) and
surgeon 2 operated on 21 hips (DAA 11, PA 10). Two patients from
the DAA group breached trial protocols and were excluded from
analysis. The 2 patients excluded from the DAA group were due to
equipment failure requiring conversion to PA (n ¼ 1) and an
emergency requiring cessation of surgery (n ¼ 1). A patient from
the PA group sustained a periprosthetic fracture at the 4-week
postoperative time point requiring revision. Preoperative and
2-week data for this patient were included for analysis. Results for
73 patients (35 DAA, 38 PA) were analyzed, 72 (35 DAA, 37 PA) were
complete. Barring exclusions, no participants were lost to follow-up
(Fig. 1).
Both DAA and PA groups appeared well matched for de-
mographic characteristics (Table 1). The median age of participants
was 61 years and BMI was 28 kg/m2. There were no statistical
differences between American Society of Anesthesiologists scores
and modality of anesthesia between both groups (Table 2).
Primary Outcomes
There were no statistically significant differences in OHS and
WOMAC scores across time. Very similar patterns of recovery were
exhibited in both groups following THA surgery (Table 3).
Secondary Outcomes
There were no statistically significant differences between both
groups at all time points for EQ5D and 10mwt. Statistically
significant secondary outcomes associated with the DAA group
included longer operative times (P < .001), smaller surgical wounds
(P < .001), higher blood loss (P ¼ .04), lower 2-week postoperative
opiate analgesic usage (P ¼ .04), and weaker hip flexion at 2 weeks
(P ¼ .03) and 6 weeks (P ¼ .01) compared with PA (Tables 4 and 5).
Three participants from each group required blood transfusions for
anemia (Table 2).
Ordinal analyses identified increased odds of having improved
straight leg raise strength in favor of the PA at 2 weeks (OR 2.7, 95%
CI 1.2, 6.5: P ¼ .02) and 6 weeks (OR 3.2, 95% CI 1.3, 8.2: P ¼ .01) after
surgery (Table 5). There were no statistically significant differences
in gait aid usage between both groups. Two weeks postoperatively,
14% and 18% of participants did not require a gait aid the DAA and
PA groups, respectively. Gait aids used at this time were mostly
Table 6
Gait Aid Usage After THA.
Baseline: n (%) 2 wk: n (%) 6 wk: n (%)
DAA PA DAA PA P DAA
3 (9) 6 (16) 30 (86) 31 (82) .64 4 (11)
887
P 12 wk: Odds P Treatment  Time
Ratio (95% CI) Ratio (95% CI) Interaction
3.2 (1.3, 8.2) .01 2.2 (0.8, 5.8) .13 0.19
1.8 (0.7, 4.6) .22 0.8 (0.3, 2.3) .61 0.19
single crutches (DAA 40% and PA 45%) or 2 crutches (DAA 37% and
PA 29%). Six weeks postoperatively, 89% of the DAA group and 87%
of the PA group did not require gait aids. At 12 weeks, 97% of the
DAA group and 100% of the PA group did not require gait aids
(Table 6).
Subgroup Analysis
Surgeon 1 had a significantly shorter acute length of hospital
stay for DAA group when compared to the PA group (P ¼ .04). There
was no significant difference in total length of hospital stay be-
tween both groups for combined surgeon results (Table 4).
Subgroup analysis of WOMAC and OHS scores detected statis-
tically significant differences favoring hip flexion activity items on
these scales in the DAA group up to the 6-week postoperative mark
when compared with the PA group. This occurred in domains of
putting on socks, taking off socks, and bending to the floor (Table 7).
Radiological Analysis
There were no statistically significant differences in radiological
positioning of implants between the DAA and PA group. The mean
acetabular inclination for the DAA group was 46.2 and for the PA
group 45.9 . Acetabular anteversion was 24.6 for the DAA and
20.3 for the PA group, respectively. The DAA group had 20 (57%)
outside of Lewinnek's safe zone vs 13 (34%) for the PA group
(P ¼ .06). The mean femoral stem orientation was within 2 of
varus/valgus in both groups; 5/35 hips (14%) of the DAA femoral
stems had radiological evidence of subsidence of >3 mm compared
to 1/38 (3%) for the PA group at 6 weeks (P ¼ .1) (Table 8). Femoral
implants that subsided, stabilized without symptoms and did not
require revision.
Adverse Events
The number of major adverse events was observed to be higher
in the DAA group 4/35 (11%) compared to the PA group 2/38 (5%)
but was not statistically significant (Fisher's P ¼ .418). Two fractures
were identified in the DAA group and managed conservatively. The
first was a recognized intraoperative femoral perforation on
broaching that was managed with only protected weight bearing of
6 weeks. The second was an asymptomatic avulsion tip fracture of
the greater trochanter postoperatively identified at 6 weeks that
did not require surgical fixation. A single dislocation occurred in
both groups. An anterior dislocation occurred for a DAA patient
while a posterior dislocation occurred for a PA patient. Both dislo-
cations were treated with closed reduction and had no further
12 wk: n (%) Group  Time Interaction
PA P DAA PA P
5 (13) .93 1 (3) 0 (0) d <0.001
888 T.E. Cheng et al. / The Journal of Arthroplasty 32 (2017) 883e890

Table 7
Comparison of Hip Flexion Activities Between Treatment Groups at Each Time Point.

Variable Baseline: Mean (SEM) 2 wk: Mean (SEM) 6 wk: Mean (SEM) 12 wk: Mean (SEM) Test for
(Group  Time)
DAA PA DAA PA Pa DAA PA Pa DAA PA Pa
Interactionb

WOMAC bending to floor 4.09 (0.17) 4.39 (0.16) 3.60 (0.19) 4.11 (0.19) .07 2.06 (0.18) 2.79 (0.17) .01 1.60 (0.13) 1.73 (0.13) .49 0.05
WOMAC getting in/out of car 3.94 (0.14) 4.34 (0.14) 2.86 (0.18) 3.08 (0.18) .38 1.77 (0.13) 1.78 (0.12) .94 1.57 (0.14) 1.52 (0.14) .78 0.40
WOMAC putting on socks 4.20 (0.15) 4.58 (0.14) 3.80 (0.17) 4.39 (0.17) .02 2.29 (0.21) 3.09 (0.20) .01 1.71 (0.21) 2.05 (0.17) .16 0.28
WOMAC taking off socks 3.97 (0.15) 4.39 (0.14) 3.20 (0.22) 4.00 (0.21) .01 2.17 (0.19) 2.54 (0.19) .17 1.54 (0.14) 1.62 (0.13) .69 0.11
OHS socks 1.54 (0.15) 1.24 (0.14) 1.66 (0.22) 0.79 (0.21) .01 3.03 (0.18) 2.05 (0.18) .01 3.31 (0.16) 3.17 (0.16) .52 0.01

SEM, standard error of the mean; DAA, direct anterior approach; PA, posterior approach; WOMAC, Western Ontario McMasters Arthritis Index.
a
P value obtained from pairwise comparison of treatment groups at each time point using linear mixed effects model.
b
P value for the F test of the time by treatment interaction.

sequelae. A single patient in the DAA group had an unexpected
return to theater on the first postoperative day to revise a leg-
length discrepancy of 3 cm.
The minor adverse event profile was skewed significantly to-
ward the DAA group due to the high incidence of lateral cutaneous
nerve of thigh (LCNT) neuropraxia. The incidence of LCNT neuro-
praxia at 12 weeks was 29/35 (83%) in the DAA group. LCNT was
defined by the investigators as the absence of normal sensation
when compared to the contralateral side. Excluding LCNT neuro-
praxia, the complication rate the DAA group was 31% compared to
the 21% (P ¼ .42). Three participants from each group had minor
wound issues that did not require surgical intervention. A single
case of below knee deep vein thrombosis was recorded in the PA
group. One episode of trochanteric bursitis was identified in the PA
group and treated with a cortisone injection. Two cases of iliopsoas
tendonitis were documented in the follow-up of the DAA group.
There were no complications of heterotopic ossification or post-
operative hematoma observed 12 weeks postoperatively (Table 9).
Discussion
Primary outcomes following DAA and PA THA were similar. This
extended to secondary outcome measures including the EQ5D,
10mwt results, and gait aids. Total length of hospital stay was also
similar for both groups. Despite the overall similarity of the DAA
and PA outcomes, some significant differences were identified in
other domains. These differences occurred within the first 6 weeks
after surgery. Differences in favor of DAA include a shorter acute
hospital length of stay for surgeon 1, smaller surgical wounds, and
lower 2-week postoperative total opiate analgesic usage. Differ-
ences in favor of the PA included shorter operative times, less blood
loss, and lower complication rates. These initial findings are similar
to previous reports in the literature [12,18].
Analysis investigating the effect of differing hip movement
precautions in both groups demonstrated that early functional
outcomes for the WOMAC and OHS favored the DAA group in hip
flexionetype activities. In randomized studies comparing hip pre-
cautions using the anterolateral approach, patients were able to
cease gait aids, drive and return to work sooner compared to pa-
tients in the restricted group. The benefits were without increased
risk of dislocations [30,31]. Restrepo et al [32] similarly concluded
in his retrospective study of 2386 patients that postoperative re-
covery following anterolateral and DAA THA did not require routine
hip precautions.
Theoretically, the DAA minimizes muscle damage and weakness
through the utilization of an internervous and intermuscular plane.
However, assessment of anterior thigh muscle strength post-
operatively was not conducted by previous randomized studies. In
this study, the DAA group had statistically significant weakness
compared with the PA during clinical straight leg-raise test
assessment in the supine position. This difference in hip flexor
strength did not persist beyond the 6-week postoperative mark. A
study by Winther et al [33] that compared muscle strength
following THA identified no significant differences in quadriceps
strength using a leg press machine when comparing the DAA and
PA up to 3 months postoperatively. The muscle groups activated to
perform a leg press are different to that of a straight leg raise
making it difficult to ascertain if there is indeed a true measurable
impairment in quadriceps function following DAA surgery.
Winther's study also found better abductor strength in favor of the
PA group at 6 weeks. We were unable to identify any clinically
significant abductor weakness between both groups.
Meticulous use of the traction table by the study surgeons
prevented intraoperative femoral shaft and ankle fractures. The risk
of minor adverse events in our study favored the PA owing to the
high 83% incidence of LCNT neuropraxia for both surgeons in the

Table 8
Radiological Analysis.

Radiological Analysis DAA PA Risk Ratio (PA to DAA)

n % N % RR 95% CI LL 95% CI UL P Valuea

 
Acetabular implant inclination (<30 or >50 ) 11 31 8 22 0.7 0.3 1.5 .43
Femoral implant alignment (<3 varus or >3 ) 10 29 11 30 1.0 0.5 2.1 1.0
Lewinnek combined 20 57 13 34 0.6 0.4 1.0 .06
0- to 6-wk stem subsidence >3 mm 5 14 1 3 0.2 0.1 1.5 .10

Mean SD Mean SD Difference (PA-DAA)

Difference 95% CI LL 95% CI UL P Valueb

Acetabular inclination ( ) 46.2 6.1 45.9 8.0 0.3 3.6 3.0 .86
Acetabular anteversion ( ) 24.6 8.8 20.3 10.2 4.2 8.7 0.2 .06
Femoral stem position (varus) ( ) 1.1 1.8 1.6 2.4 0.5 1.5 0.5 .29

DAA, direct anterior approach; PA, posterior approach; CI, confidence interval; SD, standard deviation; RR, risk ratio; LL, lower limit; UL, upper limit.
a
P value based on Fisher's exact test.
b
P value based on 2 sample t test.
 T.E. Cheng et al. / The Journal of Arthroplasty 32 (2017) 883e890
Table 9
Summary of Complications by Treatment Groups at 12 wk.
Adverse Events DAA PA
n %a n %b
Dislocation 1 3 1 3
Fracture 2 6 1 3
Unplanned return to surgery 1 3 0 0
Nonoperative wound 3 9 3 8
LCNT neuropraxia 29 83 0 0
DVT 0 0 1 3
Bursitis 0 0 1 3
Iliopsoas tendonitis 2 6 0 0
DAA, direct anterior approach; PA, posterior approach; LCNT, lateral cutaneous
nerve of thigh; DVT, deep vein thrombosis.
a
Denominator is number in DAA group ¼ 35.
b 5. Anterior Total Hip Arthroplasty Collaborative Investigators, Bhandari M,
Denominator is number in PA group ¼ 38.
c Matta JM, et al. Outcomes following the single-incision anterior approach to
P value based on Fisher's exact test.
DAA group. The study's incidence is at the upper limit of the re-
ported range in the literature [34,35]. LCNT neuropraxia itself was
not associated with lower WOMAC, OHS, or EQ5D scores. This is
partly supported by Homma et al who reported no association of
LCNT neuropraxia with Harris Hip and Japanese Orthopaedic As-
sociation Scores. Patients affected by LCNT neuropraxia however
had a lower Forgotten Joint Score-12 [36].
The higher rate of femoral stem subsidence in the DAA group
may be due to errors of femoral component sizing. A male gender
and higher BMI are recognized factors for more challenging DAA
cases [37]. This could potentially affect exposure and explain this
finding.
Strengths of the study can be attributed to early assessment of
participants at the 2-week time point, stratified randomization, and
analysis using a linear mixed model. Straightforward and stan-
dardized methods of testing were employed to facilitate repro-
ducibility in a clinical setting.
Limitations of the study include a performance bias with one
surgeon performing more surgeries than the other. The study's se-
lection criteria excluded the recruitment of patients with potentially
higher risks of complications. The study reports short term post-
operative data of 3 months. Full blinding was also not possible due to
surgical incisions and hip precautions. Hip movement precautions
were not homogenous between both groups. The difference in
scores observed in hip flexion activities favoring the DAA group did
not result in statistically significant differences for total WOMAC and
OHS scores. This was potentially because the study was under-
powered to detect subtle changes of less than 5 points in the
WOMAC. This study was also not designed to capture the precise
date of gait aid cessation in its follow-up. Data collection and
assessment by a single investigator may be subject to observer bias.
DAA THA has comparable results with PA THA. Both techniques
yielded good short-term results. Minor differences in early hip flexion
strength and function were observed. Results obtained support cur-
rent evidence in the literature for early outcomes and complications.
Choice of surgical approach for THA should be based on patient fac-
tors, surgeon preference, and experience. Following review of this
study, the authors in their practice continue to offer suitable patients
both DAA and PA THA. However, where complex primary surgery is
anticipated, the authors have a preference for the PA.
Acknowledgments
The authors of the paper acknowledge the supervision and
support of Professor Ian Davis and the Monash University Eastern
Health Clinical School. The authors also acknowledge the generous
donations from the Bulley Fellowship and Box Hill Golf Club for this
889
research. Finally, this body of research is dedicated to the work and
memory of Mr Michael Armstrong, the late Director of Orthopae-
P Valuec dics at Eastern Health.
1.00 References
.60
.48 1. Learmonth ID, Young C, Rorabeck C. The operation of the century: total hip
1.00 replacement. Lancet 2007;370(9597):1508.
2. Australian Orthopaedic Association National Joint Registry. Annual report 2015.
1.00 Adelaide: AOA; 2015.
3. Kurtz SM, Lau E, Ong K, et al. Future young patient demand for primary and
1.00
revision joint replacement: national projections from 2010 to 2030. Clin Orthop
.23
Relat Res 2009;467(10):2606.
4. Dosanjh S, Matta JM, Bhandari M. The final straw: a qualitative study to explore
patient decisions to undergo total hip arthroplasty. Arch Orthop Trauma Surg
2009;129(6):719.
total hip arthroplasty: a multicenter observational study. Orthop Clin North Am
2009;40(3):329.
6. Meneghini RM, Pagnano MW, Trousdale RT, et al. Muscle damage during MIS
total hip arthroplasty: Smith-Petersen versus posterior approach. Clin Orthop
Relat Res 2006;453:293.
7. Bremer AK, Kalberer F, Pfirrmann CW, et al. Soft-tissue changes in hip abductor
muscles and tendons after total hip replacement: comparison between the direct
anterior and the transgluteal approaches. J Bone Joint Surg Br 2011;93(7):886.
8. Bergin PF, Doppelt JD, Kephart CJ, et al. Comparison of minimally invasive direct
anterior versus posterior total hip arthroplasty based on inflammation and
muscle damage markers. J Bone Joint Surg Am 2011;93(15):1392.
9. Restrepo C, Parvizi J, Pour AE, et al. Prospective randomized study of two sur-
gical approaches for total hip arthroplasty. J Arthroplasty 2010;25(5):671.
10. Mayr E, Nogler M, Benedetti MG, et al. A prospective randomized assessment of
earlier functional recovery in THA patients treated by minimally invasive direct
anterior approach: a gait analysis study. Clin Biomech (Bristol, Avon)
2009;24(10):812.
11. Rathod PA, Orishimo KF, Kremenic IJ, et al. Similar improvement in gait pa-
rameters following direct anterior & posterior approach total hip arthroplasty.
J Arthroplasty 2014;29(6):1261.
12. Barrett WP, Turner SE, Leopold JP. Prospective randomized study of direct
anterior vs postero-lateral approach for total hip arthroplasty. J Arthroplasty
2013;28(9):1634.
13. Taunton MJ, Mason JB, Odum SM, et al. Direct anterior total hip arthroplasty
yields more rapid voluntary cessation of all walking aids: a prospective, ran-
domized clinical trial. J Arthroplasty 2014;29(9 Suppl):169.
14. Christensen CP, Jacobs CA. Comparison of patient function 6 weeks after direct
anterior or posterior THA: a randomized study. J Arthroplasty 2015;30(9 Suppl):
94.
15. de Steiger RN, Lorimer M, Solomon M. What is the learning curve for the
anterior approach for total hip arthroplasty? Clin Orthop Relat Res
2015;473(12):3860.
16. Jewett BA, Collis DK. High complication rate with anterior total hip arthro-
plasties on a fracture table. Clin Orthop Relat Res 2011;469(2):503.
17. Woolson ST, Pouliot MA, Huddleston JI. Primary total hip arthroplasty using an
anterior approach and a fracture table: short-term results from a community
hospital. J Arthroplasty 2009;24(7):999.
18. Martin CT, Pugely AJ, Gao Y, et al. A comparison of hospital length of stay and
short-term morbidity between the anterior and the posterior approaches to
total hip arthroplasty. J Arthroplasty 2013;28(5):849.
19. Muller DA, Zingg PO, Dora C. Anterior minimally invasive approach for total hip
replacement: five-year survivorship and learning curve. Hip Int 2014;24(3):277.
20. Seng BE, Berend KR, Ajluni AF, et al. Anterior-supine minimally invasive total
hip arthroplasty: defining the learning curve. Orthop Clin North Am
2009;40(3):343.
21. 12th Annual report 2015. Wales, Northern Ireland and the Isle of Man: National
Joint Registry for England; 2015.
22. The New Zealand Joint Registry. Fifteen year report (January 1999 to December
2013). New Zealand: New Zealand Joint Registry; 2014.
23. Kerr DR, Kohan L. Local infiltration analgesia: a technique for the control of
acute postoperative pain following knee and hip surgery: a case study of 325
patients. Acta Orthop 2008;79(2):174.
24. Bellamy N, Buchanan WW, Goldsmith CH, et al. Validation study of WOMAC: a
health status instrument for measuring clinically important patient relevant
outcomes to antirheumatic drug therapy in patients with osteoarthritis of the
hip or knee. J Rheumatol 1988;15(12):1833.
25. Dawson J, Fitzpatrick R, Carr A, et al. Questionnaire on the perceptions of pa-
tients about total hip replacement. J Bone Joint Surg Br 1996;78(2):185.
26. Dolan P. Modeling valuations for EuroQol health states. Med Care
1997;35(11):1095.
27. van der Leeden M, Fiedler K, Jonkman A, et al. Factors predicting the outcome of
customised foot orthoses in patients with rheumatoid arthritis: a prospective
cohort study. J Foot Ankle Res 2011;4:8.
28. Aids to the examination of the peripheral nervous system. London, UK: Medical
Research Council; 1981.
890 T.E. Cheng et al. / The Journal of Arthroplasty 32 (2017) 883e890
29. Ferrara PE, Rabini A, Maggi L, et al. Effect of pre-operative physiotherapy in
patients with end-stage osteoarthritis undergoing hip arthroplasty. Clin Rehabil
2008;22(10-11):977.
30. Ververeli PA, Lebby EB, Tyler C, et al. Evaluation of reducing postoperative hip
precautions in total hip replacement: a randomized prospective study. Ortho-
pedics 2009;32(12):889.
31. Peak EL, Parvizi J, Ciminiello M, et al. The role of patient restrictions in
reducing the prevalence of early dislocation following total hip arthro-
plasty. A randomized, prospective study. J Bone Joint Surg Am 2005;87(2):
247.
32. Restrepo C, Mortazavi SM, Brothers J, et al. Hip dislocation: are hip pre-
cautions necessary in anterior approaches? Clin Orthop Relat Res
2011;469(2):417.
33. Winther SB, Husby VS, Foss OA, et al. Muscular strength after total hip arthroplasty.
A prospective comparison of 3 surgical approaches. Acta Orthop 2016;87(1):22.
34. Bhargava T, Goytia RN, Jones LC, et al. Lateral femoral cutaneous nerve
impairment after direct anterior approach for total hip arthroplasty. Orthope-
dics 2010;33(7):472.
35. Goulding K, Beaule  PE, Kim PR, et al. Incidence of lateral femoral cutaneous
nerve neuropraxia after anterior approach hip arthroplasty. Clin Orthop Relat
Res 2010;468(9):2397.
36. Homma Y, Baba T, Sano K, et al. Lateral femoral cutaneous nerve injury with the
direct anterior approach for total hip arthroplasty. Int Orthop 2015;40(8):1587.
37. Frye BM, Berend KR, Lombardi Jr AV, et al. Do sex and BMI predict or does stem
design prevent muscle damage in anterior supine minimally invasive THA? Clin
Orthop Relat Res 2015;473(2):632.