Posterior tilt/posterior neck collapse is frequently encountered after fixation of femoral neck fractures and is associated with further neck shortening and nonunion [14–16]. Additionally, a preoperative posterior tilt of more than 20° was reported to be a significant predictor for reoperation . It is likely that preoperative posterior tilt of femoral head would largely damage the bony mechanical transduction of the posterior cortex, and the inclination of femoral head retroversion is less likely to be counteracted with the use of normal fixation construct (parallel partially threaded screws), especially in the presence of posterior comminution [17, 18]. Thus, a more specialized fixation construct is needed to resist posterior tilt in retroverted femoral neck fractures.
Shin et al.  recently reported that in the classic configuration of parallel cannulated screws, replacing a partially threaded cannulated screw with a posterior fully threaded positioning screw can prevent femoral neck shortening and posterior tilt in Garden Ⅰ and Ⅱ femoral neck fractures, as this hybrid construct was regarded to be more length- and angle-stable, which had already been demonstrated previously by a biomechanical study . Similarly, in our study, the anterior partially threaded compression screw was replaced with a long-threaded positioning screw (Fig. 5), which would also provide length- and angle-stability to prevent posterior tilt. Just as was shown in our results, lower increased posterior tilt and smaller amount of FNS were observed in the ALTS group (Table 1). In addition, lateral plating offers better integral property by combining the three screws into a whole one construct. What’s more, the screw purchase of the anterior long-threaded screw was larger than that of the other two screws, which was helpful to counteract the tendency of posterior tilt of femoral head.
There might be concerns of inducing nonunion by replacing compression screws with non-sliding positioning screws. However, a study compared short- and long-threaded cancellous screws in the fixation of femoral neck fractures in a randomized trial of 432 patients, and no difference was found regarding fracture healing complications . Thus, we speculate that the absolute length of screw thread may not be determinant in femoral neck fracture fixation. Other factors, including fracture geometry (which largely determines the tendency of fracture displacement), patient characteristics, and reduction quality, may be more important. In this perspective, the strength of our study is that we focused on femoral neck fractures that presented posterior tilt before surgery (in consideration of homogeneity of this study), in whom femoral head retroversion would largely occur postoperatively due to damaged posterior cortical transduction. In this situation, anterior partiality of screw purchase in the ALTS configuration was demonstrated to significantly resist posterior tilt of femoral head. Consequently, we hold the view that in the fixation of femoral neck fractures, implant construct with biomechanical partiality that specifically counteracts the inclination of fracture displacement is of paramount importance to create a balanced mechanical environment for fracture healing (Fig. 5).
Chiang et al.  recently reported that three fully threaded headless compression screws, which were normally regarded as a non-sliding length-stable construct, failed to prevent FNS and varus collapse in Garden Ⅰ and Ⅱ femoral neck fractures compared with partially threaded screws. We think that though the patient population were all non-displaced fractures, the tendency of fracture displacement (in three-dimensional orientation) may not be the same due to potential heterogeneity in fracture geometry, bone quality, etc., and thus, the universally used three fully threaded screws may not be effective in all patients. In comparison, the ALTS construct in our study with anterior partiality of screw purchase was used to treat retroverted femoral neck fractures (which was more targeted and specific), and yielded favorable results in terms of decreasing posterior tilt of femoral head. However, precise prediction and evaluation of three-dimensional stability of femoral neck fractures (or inclination of displacement) is still a difficult task, which deserves further study in the future.
FNS was also frequently encountered after fixation of femoral neck fractures. Of note, FNS was reported to be related with length discrepancy of the lower extremity, decreased abductor length, femoral head collapse, hip impingement, and inferior hip function [22–25]. Worse still, femoral neck compaction after surgery has been reported as an important risk factor for avascular necrosis . In our study, the ALTS configuration showed statistically significant difference in decreasing the amount of FNS in the fixation of retroverted femoral neck fractures, one reason may be that by creating a more balanced biomechanical environment, further collapse of femoral neck would be largely hampered. Unfortunately, when using 5 mm as the cut-off value, the difference was not significant, which implies limited clinical significance. Small sample size and specific study design (anteriorly-partialized screw purchase to resist posterior tilt) may be two explanations for this. Regarding the risk factors for developing posterior tilt of femoral head, we identified posterior comminution as a risk factor, which implies that posterior bony transduction is of vital importance to achieve fracture stability; NTS configuration (in reference to ALTS configuration) was also identified as a risk factor, which proved superiority of the ALTS configuration in resisting femoral head retroversion; also, inferior reduction quality was identified as a risk factor, which was in line with the literature. Altogether, anterior partiality of screw purchase that specifically counteracts the inclination of retroversion and good reduction quality were two important factors for achieving favorable outcomes.
Several limitations existed in this study. First, this retrospective study may contain potential selection bias. Second, the sample size in each group is relatively small. Third, a minimum follow-up of 12 months is relatively short, and long-term complications remain to be evaluated.