In clinical practice, post-operative dislocations after hip hemi-arthroplasty (HA) receive much attention, but intra-operative fractures (IOFs) are often overlooked. The risk factors of IOFs include female sex, increasing age, osteoporosis, rheumatoid arthritis, bone morphology, and cementless stem123456). Supine-position HA with the anterolateral approach in the supine position (ALS) or direct anterior approach (DAA) has become prevalent. The difficulty of femoral preparations when using ALS or DAA often leads to IOFs78910). We also experienced some IOFs and stem mal-alignment in supine-position HA for hip fractures. Before using the supine-position approach, we had performed HA with the conventional posterior approach (PA) and experienced IOFs in 3.1% of the patients (4/128) and dislocations in 2.3% (3/128) over 5 years. All four IOFs were periprosthetic fractures requiring cerclage wiring. In contrast, supine-position HA caused IOFs of another type (great trochanter fractures) that we had never experienced when using PA. Based on this experience, we had returned to lateral-position HA (PA and superior approach [SA]11)), until we took measures to reduce the incidence of IOFs the supine-position HA. No studies have focused on the incidence of IOFs in supine-position and lateral-position HA. We aimed to clarify the incidence and types of IOFs in HA with the supine and lateral positions.

We retrospectively investigated cases of HA for acute femoral neck fractures from June 2013 to May 2018. Since June 2013, we had used mainly tapered-wedge stem implants. We collated data regarding the approaches used, operation duration, intra-operative blood loss, and the following complications: IOFs (we used Vancouver classification6)), treatment for IOFs, post-operative dislocations, deep infections, symptomatic deep venous thrombosis, cerebrovascular infarction, and death. In addition, data on patient characteristics (age at operation, sex, body height, body weight, and body mass index [BMI]) and radiographic indices (limb length discrepancy [LLD] and varus angle), healthy (contralateral) center-edge angle, and canal flare index (CFI). This research was approved by the ethics review board of the authors' affiliated institution (H27-11) and written informed consents from the patients were obtained.

This study included 194 patients. The mean age at HA was 82.6±8.4 years (Table 2). The number of female patients was 148 (76.3%). The mean BMI was 20.3±3.6 kg/m². ALS was used in 25 (12.9%) patients, DAA in 21 (10.8%), DLA in 9 (4.6%), PA in 50 (25.8%), and SA in 89 (45.9%) patients. Young surgeons performed 142 HAs (73.2%). The mean operation duration was 66.2±21.7 minutes and the mean operative blood loss was 140.3±133.2 g. IOFs occurred in 11 (5.7%) patients: six VA_GT fractures, one VB1, and four VB2. IOFs occurred in 4.2% (6/142) of young surgeons' HAs and 9.6% (5/52) of experienced surgeons' HAs (P=0.17).

In the radiographic analysis, we excluded four IOFs (conversion to a cement stem), five straight stems, and one planned cement stem. The 184 patients for whom a tapered-wedge stem was used were evaluated by post-operative radiographs. The mean LLD was 3.2±5.5 mm and the mean varus angle was 0.1±1.6°. We excluded three hip arthroplasties and three multiple pinnings. The radiographic indices on the healthy side for the patients (n=188) were as follows: mean center-edge angle was 31.5±6.5°, mean CFI was 3.6±0.6.

In our study, IOFs occurred in 5.7% of the patients, which was higher than the frequency of post-operative dislocations (1.6%). IOFs were frequently (in 17.4% of the cases) encountered with the use of supine-position HA, this rate was higher than that with lateral-position HA (2.0%). IOFs were not related to surgical experience, in line with the previous studies18).

In ALS, the difficulty of implant insertion causes femoral shaft penetration (0.3%). In our patients, three VB2 fractures occurred during rasping or implantation. One VA_GT occurred during elevation of the femur. In DAA, fractures of the great trochanter occur at a frequency of 1.4% to 4.5%92122). An insufficient release of the capsule9) and elevation of the femur with powerful levers10) result in great trochanter fractures. Additionally, the reduction maneuver results in periprosthetic fractures4). In our patients, all four VA_GT fractures were diagnosed on post-operative radiographs. In DLA, VA_GT occurs at a frequency of 0.7% to 3%, VA_LT at 2.8% to 8.1%, and VB fractures at 6.4%123). These two reports did not mention whether some particular procedures of DLA cause these fractures. Generally, VB fractures occur because of a mismatch between the stem and canal during rasping in an osteoporotic femur24). In PA, periprosthetic fractures occur at a frequency of 1.8% to 2.6%2122). In our cases, one VB2 fracture occurred during final stem insertion and one VA_GT occurred during removal of the rasp. In SA, periprosthetic fractures occur at a frequency of 5.8%14). In our cases, one VB1 fracture was diagnosed on a radiograph at 1 week post-operatively. Theoretically, rasping before the neck cut would reduce hoop stress and prevent VB fractures14).

We thought that the hyperextended (over-range) position of the femoral rasping and soft tissue sparing would overload the tip of the great trochanter and the proximal femur in the supine position. Insufficient release of the hip capsule would result in excessive leverage momentum of the retractor behind the greater trochanter during hip extension, which may lead to VA_GT. It would also lead to an oblique fracture of the proximal femur due to excessive external rotation torque applied to the distal femur to maintain the hip at the desired position to insert the stem. Lastly, an insufficient release of the capsule would lead to penetration of the femur by the femoral stem. In ALS and DAA, a release of the capsule and short external rotators as necessary would prevent IOFs and facilitate femoral preparation. The femoral preparation positions over a normal range of motion would require a release of muscles and the hip capsule. Fewer complications and correct implantation should precede muscle sparing.

Several limitations of this study should be acknowledged. First, this study was a retrospective case series. Second, we did not evaluate bone quality and disease-specific osteoporosis-like rheumatoid arthritis and further studies should be required. In our hospital (500-bed, secondary-level), lateral-position HAs decreased the incidence of IOFs. Meanwhile, the incidence of post-operative dislocations remained low. The post-operative LLD and varus angle of lateral-position HAs were more favorable than those of supine-position HAs. These results may be useful for young and experienced surgeons to employ new or unused approaches.

The supine position in HA surgery is an IOF risk factor and significantly increases the incidence of great trochanter fractures of Vancouver type A.