2.1 Patient groups

A consecutive group of 200 patients with end stage osteoarthritis of the hip were operated using either the ALA (Watson-Jones approach) in the technique described by Roettinger⁴ or the DLA (modified Hardinge approach). For this retrospective, single-institutional, single surgeon study, 100 patients with ALA were propensity matched to 100 patients with DLA. For both groups the age, gender, body mass index (BMI), blood loss (pre. - and postoperative red blood cells (RBC)), transfusion rates and complications were noted. Blood loss was defined as the difference between the preoperative hemoglobin (g/dl) to postoperative value at the first postoperative day (POD1). In our clinical routine the patients outcome was compiled using the Western Ontario und McMaster Universities Osteoarthritis Index (WOMAC) in the German version with 24 questions, ranging from 0.0 (good clinical result) to 4.0 points (poor clinical result) and was converted to a maximum of 100.0 points (worst result). The self-administered score eliminates observer bias and was evaluated preoperatively and postoperatively at one year follow-up examination.

2.2 Surgical technique

All patients were operated in the lateral decubitus position with the surgeon standing at the back (DLA) or in front of the patient (ALA). The prepping and draping were identical. With the ALA, minimal invasive surgery (MIS) retractors and a leg holder were used instead of a Charnley retractor and a bolster for leg positioning (DLA). After hip dislocation, the leg was brought to the front (DLA) or to the back (ALA) for stem placement. No image intensifier was used neither for cup nor for stem placement. The cup was positioned using the transverse acetabular ligament. There were two offset options with the Corail® stem (DePuy Orthopaedics, Inc., Warsaw, USA). In all cases 32 mm heads and neutral liners were applied. With the DLA, the straight capsule incision was closed. With the ALA, the H-shaped incision was adapted. No drain was used. In addition, local infiltration analgesia (LIA with ropivacaine) was applied in all patients.

Whenever the DLA was used, an additional device was applied for intraoperative measuring of femoral offset and leg length (Smith and Nephew, Memphis, USA) by placing a pin in the ilium and performing a relative measurement of leg length and offset prior to dislocation.³ With the ALA in MIS setting, the application of the device proved to be not feasible and was discontinued. In these cases (ALA), stem and cup position were solely based on preoperative planning and clinical assessment.

2.3 Radiological assessment

X-rays were evaluated pre. - and postoperatively. An anterior-posterior (a.p.) pelvic and an axial Lauenstein view of the treated hip were taken preoperatively and after mobilizing with full weight bearing prior to discharge. Additionally, x-rays were taken immediately postoperatively (in the operating room (OR)). The standardized pre. - and postoperative radiographs were digitally acquired with the Carestream®. A calibration using a 25 mm radiopaque sphere and Orthoview™ was subsequently performed. Both hips of each patient were depicted for comparison. All measurements were performed by two institutional surgeons. After evaluation of the test-retest reliability, several radiological parameters were measured and blinded to clinical results. For comparison the leg length discrepancy (LLD, mm), the vertical and horizontal center of rotation (v/h COR, mm), the femoral offset (FO, mm), the body weight lever arm (BWLA, mm) and the femoral offset ratio (FO/BWLA, = FOR, %) were measured before and after surgery according to Mahmood.^5,6 The cup inclination and anteversion were calculated according to Ming Lu.⁷ The femur type (A, B, C) was described according to Dorr's classification.⁸-The leg length discrepancy (LLD, mm) was defined as distance between the basis of both teardrop lines and the lesser trochanter base before and after THA. The values of the operated hip were subtracted from the non-treated side. Negative results represented a shortened post-treatment leg compared to the untreated leg.-The v/hCOR (mm) represented the distance between the center of the ceramic head and the teardrop line (horizontal) resp. the transischal line (vertical).-The femoral offset (mm) was defined as distance between the COR and the femoral axis at the 90° rectangular crossing.-The body weight lever arm (mm) was measured between the body axis (a line through the symphysis intersecting rectangular the vCOR) and the COR.

The calculated intraclass correlation for the cup inclination/anteversion was 0.89/0.83, for v/hCOR 0.91/0.93 and for the FO/BWLA 0.86/87.

2.4 Statistical analysis

The collected data underwent pseudonymization and were transferred to IBM® SPSS® Statistics (IBM® SPSS® Statistics, Version 24.0.0.0, 64-Bit-Version) for statistical analysis.

Metric parameters were described with the mean value, standard deviation (SD) and the range (minimum, maximum). Non-metric values were expressed as the absolute number of counts and the percentage.

First a propensity matching was performed with the approach (DLA/ALA) as dependent and age, gender and BMI as independent variables. The propensity score was used for further linear regression to compare radiological and clinical parameters between both approach groups. The level of significance was 0.05 with a corresponding confidence interval (CI) of 95%. If difference was statistically significant, the CI did not include the “0”.

In case of a power >0.9 (r = 0.5, alpha = 0.05, n = 100 per group) it was a normal data distribution presupposed. For comparison between pre – and postoperative values of each patient, a paired t-test was used.

3.1 Discussion

This study's aim was to investigate the possible contradiction of recommended surgical approaches for THA. Registry data still show a widespread use of standard hip approaches. Data from the Norwegian Arthroplasty register revealed that the DLA or PA were used in close to 80% for hip arthroplasties with a minority of the ALA (12.8%) and or the direct anterior approach (DAA) (7.6%).⁹ Longterm registry data concerning the DAA or ALA are pending or present preliminary data with small series. Meanwhile, promoting these minimal invasive surgery approaches is also seen critical, especially the undiscriminating advertisements as seen online. To lessen potential influencing, we maintained the same patient position, prepping, draping, indications and implants. Although there is no randomization, propensity matching was performed to reduce confounding parameters and to derived two homogenous groups (BMI, age, gender). The question was to compare the potential for anatomical hip reconstruction and the clinical results with the DLA to those of the minimal invasive ALA.

The radiographic analysis revealed that the general anatomical reconstruction was well within limits of the DLA and ALA. There were similar, almost identical results for the cup position and reconstruction of the COR. The x-rays, taken in an educational institution for conventional skeletal radiology, were evaluated following the Sundsvall method, which was tested before use, to have the required validity and reliability.⁵ The intra- and interobserver reliability was tested before, being aware of the methodological limits compared to CT evaluation.¹⁰ However, as described in previous studies, the method to derive cup inclination and anteversion from plain antero-posterior radiographs was deemed to be accurate.⁷

In our patient group the one-year clinical follow-up showed an improvement of the WOMAC about 45.8 points to 3.1 (p < 0.001). The further analysis indicated a comparable benefit for patients treated through DLA or ALA (p > 0.05). Other studies supported these results, insofar as surgical approach seemed not to influence the clinical results.¹¹

The mean cup inclination was 0.2° higher and the anteversion 2° lower in the DLA group, but the differences were not statistically significant (p > 0.05) and the values in both groups were in line with the general recommendations.¹² Further, in a prospective multicenter study a comparable cup position between the ALA and an alternate minimal invasive approach was described.¹¹ The COR in DLA patients was placed 4.2 mm more medial and 0,9 mm more cranial after surgery, resp. 3.5 mm more medial and 1.9 mm more cranial in ALA patients. The displacement was comparable to the values previously published.¹³ In addition, displacement was not statistically different between the approaches, nor between the treated and non-treated side (p > 0.05). Due to a higher FO and reduced BWLA after surgery for both approaches the postoperative FOR increased by an average of 0.3% (DLA) resp. 0.2% (ALA) (p > 0.05). As a result, we observed differences in the mean values of the cup position and the anatomical reconstruction of the resurfaced hip joint, which might be influenced by the chosen approach. Nevertheless, a hint for clinical consequences was not observed.

In terms of early aseptic loosening with the anatomic Corail® stem (DePuy Orthopaedics, Inc., Warsaw, USA), there were no radiographic signs in our study (ALA/DLA). From the five patients with symptoms of limping in the DLA group, four are likely to be related to the approach itself. One of them resolved over time, leaving three patients with temporary lateral hip pain. Two patients in the ALA group with abductor weakness were probably caused by technical problems (one fracture, one decreased offset about 7 mm). According to the findings of Renkawitz et al., patients developed an altered gait pattern if the offset was more than 5 mm out of range.¹⁴ No dislocation was observed in both groups, so the previously described relatively high rate of dislocations using the modified Watson-Jones could not be reproduced.¹⁵ Surgeons used to perform procedures in the lateral position theoretically can proceed using the ALA in the technique described with nearly equivalent setting (i.e. draping, positioning). It is thought that proceeding to the DAA in the dorsal decubitus position, likely in combination with a motorized positioning table is more prone for an extended learning curve.

Although conventional hip approaches still represent the norm, minimal invasive techniques were promoted intensively and might turn out to be the new standard incision in future, being muscle preserving approaches that support the trend of fast track surgery with speedier rehabilitation. This retrospective study confirmed that a reduced overview in a minimal invasive ALA setting without controlling devices for FO and LL can be used for primary total hip arthroplasty to gain similar reconstruction of the hip geometry. Clinical and radiological parameters were comparable between standard DLA and ALA. The effect of lesser abductor issues opposed to the DLA might drive the surgical community towards a more extensive use of the anterior approach alternatives.