There are previous published biomechanical studies [23, 24] that directly compared the stability of single and dual lag screw implants used for treatment of intertrochanteric hip fractures. The authors of these prior studies concluded that the two constructs showed equivalent rigidity and stability, that biomechanical properties were as favourable as conventional hip screws, and that the dual lag screw implants had a greater ultimate failure load. However their loading parameters in these prior studies did not reflect the physiological forces that act on the hip during level walking since loading was applied to the specimens uni-axially in the coronal plane. Coronal plane loading represents the forces across the hip in single leg stance of the gait cycle . The HIPS model was specifically designed to simulate loading vectors experienced by the proximal femur during ambulation . It employs BRM, a well-validated protocol for producing hip motion using a dynamic, multi-planar double peak loading regimen. Ehmke et al. demonstrated that, cut-out mechanisms differed between multi-planar BRM loading and uni-axial loading. Only the BRM loading model resulted in cut-out that occurred by combined varus collapse and neck rotation. Moreover, they found that the initial motion was rotation about the lag screw, followed by varus collapse .
The DHS lag screws were placed in near perfect position, according to Baumgaertner et al.  with a tip-to-apex distance of 20 mm to 32 mm, depending on whether the steel shell is considered part of the articular layer, or part of the femoral head, respectively. According to the same principles, the proximal Endovis screw was placed closer to the central axis of the femoral head leaving space in the distal third of the femoral head and neck for the distal screw. The tip-to-apex distance was 20 mm for the proximal and 23 mm for the distal screw. This relatively eccentric placement of the Endovis screws preserves bone stock in the upper part of the femoral head which may further improve cut-out resistance.
Both implants were inserted into the surrogate specimens according to the manufacturer's instructions using the suggested instrumentation in the same manner as in real surgery. Differences in insertion methods between the two implants are due to differences in design of the implant's tip. The two cephalic screws had a self drilling and shelf taping design. It is known that insertion torque and pull-out forces for these screws are similar to pre-tapped screws . Moreover there are no data from the literature to support that such differences in insertion methods may have influenced the cut-out resistance of our screws in order to affect our results. In a recent biomechanical comparison of conventional DHS and DHS Blade the authors noticed an enhanced cut out resistance of the DHS Blade. The main difference between the two implants was the implantation technique, pre-drilling for the DHS and impaction for the blade. They suggested that the underlying mechanism for improved purchase of the blade implant is unclear, but bone-compaction is deemed to play a major role. However they finally concluded that maximizing the bone content around the implant forgoing pre-drilling does not necessarily enhance the cut-out resistance, since mainly elastic deformation seems to contribute to the implant anchorage. The importance of the implantation technique with or without pre-drilling is therefore decreased .
In addition to differences in insertion between screws, there are some small differences in the geometry of the screws as well. It is accepted that implant development and design remains a major approach in the efforts for developing superior treatment concepts for osteoporotic proximal femur fractures . A plethora of published biomechanical studies have attempted to define the optimum shape and size of the ideal implant for these challenging fractures. It is obvious that this ideal implant is still unknown up to date since cut-out remains one of the major clinical challenges in the field of osteoporotic proximal femur fractures . With these thoughts in mind the small differences in shaft diameter, and outer thread diameter of our implants, in relation to fixation strength and migration, is extremely difficult to investigate. The only parameter, in our model that can explain the superior biomechanical properties of dual cephalic screws seems to be the presence of the second screw, which can better control torsional forces.
Clinical studies have consistently failed to find significant differences between implant designs with regard to lag screw cut-out [11, 39, 40]. The clinical incidence of implant-related cut-out is masked by the high variability in bone quality, fracture pattern, quality of reduction, and implant placement. Using surrogate specimens in a controlled laboratory model provided sufficient reproducibility to enable detection of significant differences in migration resistance between the two implants tested. Similarity in migration kinematics and cut-out failure modes between cadaveric and surrogate specimens tested in the HIPS simulator has been demonstrated previously , and further supports the relevance of the present findings obtained within surrogate specimens.
The fixation strength of the dual lag screw construct was found to be significantly greater than the classic DHS when multidirectional dynamic forces were used for loading. In their biomechanical study, Kubiak et al. found that TAN, a dual lag screw intramedullary implant, provided significantly stronger fixation than the IMHS when loaded to failure in an unstable intertrochanteric hip fracture model . These findings support Ingman's contention that the increased rotational control of the femoral head afforded by two screws would decrease femoral head cut-out . In the present study, the double screws constructs demonstrated significantly less rotation (1.2°) than the DHS constructs (15.8°). Assuming that rotation about the femoral neck contributes to a loss of reduction and fixation stability, the superior rotational stability of dual screws implants may therefore be in part responsible for its increased resistance to varus migration and cut-out failure.
All DHS implants failed before 20,000 loading cycles, on average after 6,638 ± 2,837 cycles. Three out of five DHS specimens experienced implant bending before producing cut-out. Ehmke et al. utilized the HIPS system and found that all implants tested survived 20,000 cycles in bone surrogates. A basic difference between the present study and that of Ehmke's was the mechanical characteristics of the implants each study tested. Ehmke et al. tested a gamma nail lag screw with 12 mm shaft and 12 mm outer thread diameter in contrast to our study that tested the classic DHS lag screw with only 7.8 mm shaft and 12.5 mm outer thread diameter. Implant bending of DHS screws has been very rarely reported in clinical studies. However, in biomechanical studies it has been reported that the failure mode associated with the DHS lag screw in hard bone was screw bending rather than cut-out . Bending is highly unusual and rarely if ever seen in osteoporotic patients and someone may say that our model is questionable at best. It is true that the surrogate foam chosen for our test proved to be harder that we expected. Moreover in biomechanical studies the high load levels such as 1.45 kN peak, or two times body weight, were chosen to reliably induce onset of implant migration within a clinically realistic number of loading cycles for each of the implants tested. When loading remained below a certain threshold, initiation of implant migration did not occur for a prolonged amount of time, during which fracture healing can occur clinically. Implant bending is therefore very rare in clinical practice but not so unusual in biomechanical studies.
However the critical point of our study is the amount of migration measured that is directly related to the fixation strength of the implants. Even in these "mild osteoporotic" surrogates dual screw implants shows superior stability compared to the classic DHS implants, and the measurements were comparable and reproducible.
An important result of dual lag screws is the substantial amount of axial medial migration of the inferior screw that was noted after 10,054 load cycles. The axial migration of lag screws has been described as the "Z effect" phenomenon . This is a rarely reported mechanism of implant failure of the femoral neck element and has been described primarily for two-screw devices such as the proximal femoral nail (PFN, Synthes, Switzerland) . The incidence of this phenomenon remains unknown and the biomechanical explanation for the medial migration of the femoral neck element has not been elucidated in the literature .
In our model both screws migrated medially by on average 0.3 ± 0.8 mm for the superior screw and substantially more (4.9 ± 3.0 mm) for the inferior screw. However only one inferior screw exhibited cut-out by medial axial migration, which occurred after 10,054 load cycles. To prevent the "Z effect" in clinical practice, Jinn Lin  emphasized inserting the inferior lag screw as close as possible to – or even right on – the inferior cortex of the femoral neck. Doing so could prevent this phenomenon and could also increase the bone mass to resist screw cut-out.
Limitations to the HIPS system must be recognized. Our neck constraint assumed that the fracture had undergone maximum collapse, but had not begun migration of the lag screw. This assumption has precedence in a study by Friedl and Clausen, which used neck constraints similar to those used in the HIPS model to simulate OTA 31-A3 pertrochanteric fractures . Additionally, we did not load all specimens to failure, but ceased loading at 20,000 cycles. Our previous studies have clearly shown that the onset and pathway of migration is a more sensitive tool to determine fixation strength than cut-out [29, 30]. As a further limitation, results of our study only describe implant performance in regard to cut out failure in absence of fracture healing and do not take into account alternative failure modes. A prospective randomized clinical trial would be required in order to determine if these favorable biomechanical results of double screw implants can be reproduced in clinical practice.
In conclusion, double screw construct provided significantly greater resistance against varus collapse and neck rotation in comparison to the gold-standard DHS implant when tested in the HIPS model under conditions representative of an unstable fracture and mild osteoporotic bone.