The purpose of the present study was to investigate the incidence, characteristics and causes of the “reverse wedge effect” during application of the IN in TF. In contrast to the previously reported “wedge effect”, which was defined as an intraoperative secondary displacement at the superior part of the pertrochanteric fracture line [6], we observed another IN-related intraoperative complication, which was demonstrated as an inferior opening at the already reduced primary fracture line. We also identified that all the cases occurred in basicervical TF variants combined with a disrupted GT. By reviewing intraoperative fluoroscopy, we noticed an impingement between the reamer/IN and the superolateral edge of the cephalocervical fragment, which caused this secondary displacement (Figure 6). Because all the 414 reviewed cases were treated by PFNA/TRIGEN INTERTAN nailing but the incidence of the “reverse wedge effect” was nearly 7.97%, we believe that anatomic factors other than implantation should play a major role.
The basicervical fracture line is the most significant feature of this cohort of patients. Strictly, a basicervical fracture is defined as a 2-part fracture where the fracture line originates from the base of the femoral neck and exits above the lesser trochanter [9]. Despite the concomitant PLF or LT fracture, the primary fracture lines in this cohort of patients met this description exactly. There is little controversy about the fact that a basicervical fracture is an unstable fracture and prone to varus deformity [10]. Sliding screws were initially proposed as a treatment for these fractures [11, 12], but the recent trend favors IN fixation [9, 13]. Anatomically, the piriformis fossa is the transitional region between the hard cortex of the femoral neck and the relatively weak cancellous bone of the GT [14]. Our study demonstrated that the primary fracture lines just crossed the piriformis fossa center. This indicated that the superolateral corner of the cephalocervical fragment was primarily made from the hard cortex, which is hard enough to resist reaming and colliding with the reamer and PFNA/TRIGEN INTERTAN nail insertion. Eventually, the reamer or EFNA/TRIGEN INTERTAN nail would push the cephalocervical fragment, internally rotating and inferiorly displacing, resulting in the “reverse wedge effect”.
Notably, in this cohort, the PLF and LT fragment were the other two prominent anatomical features. As Cho’s study found that the incidence of the PLF in TF was as high as 88.4% [15], we observed a similar prevalence among all the 414 patients, and every “reverse wedge effect” case had a PLF. We believe that further classification and comparison would facilitate understanding and exploring the influences of those two fragments on “reverse wedge effect” formation and treatment/strategy design.
Although Van Embden stated that AO/OTA classification was more comprehensive and reliable compared to the Jensen-Evans classification [16], we found that it failed to cover this special fracture pattern. According to AO/OTA classification (2018 version), AO-31A1 and A2 were designated as pertrochanteric fractures in which the main fracture line propagated through the trochanters. Saarenpaa, Watson and other authors described basicervical TF variant as fracture at the base of femoral neck that is medial to the intertrochanteric line [9, 17, 18]. So basicervical TF variant could not be classified into pertrochanteric (AO - 31A1 and A2) or intertrochanteric fracture(AO - 31A3). In contrast, in 1949, even before the clinical application of computed tomography, Evans had proposed 3- and 4-fragmentary fracture patterns in his classification scheme [19]. Considering the complexity of the fracture morphology, a three-dimensional classification system originating from Evans’ would be more pragmatic [16]. The principle from Babhulkar and Shoda’s classifications of GT fracture [2, 7] could be incorporated to further classify basicervical TF variants according to the morphology of the PLF.
Although the sample size was not large enough to make statistical comparisons between the four subgroups, the measured displacements of “reverse wedge effect” and ultimate NSA were similar among them. Therefore, we could infer that the volume of the PLF minimally influenced the extent of the secondary displacement in this cohort of patients. From a mechanical view, it is the existence of the PLF contributes to the development of the “reverse wedge effect”.
When the GT region is intact, the guide wire could be constrained around the tip of the GT, and the trajectory of the reamer/IN was centralized into the femoral canal. However, an incompetent GT made the guide wire and reamer float at the start site. Under an image intensifier, we observed that the reamer was prone to skew medially, especially when attempting to over-ream the superolateral corner of the cephalocervical fragment (Figure 6D). As a result, impaction between the cephalocervical fragment and the reamer/nail occurred. Since the volume of the PLF plays a relatively minor role, this fragment could be excluded during the management of the “reverse wedge effect”.
In this cohort of patients, the incidence of PLF was 100% and the LT disruption was 60.6%. To investigate the role of LT disruption in this operative complication, comparisons between 3- and 4-fragmentary subgroups in the magnitude of displacement and postoperative NSA were made and revealed no significant differences. We could conclude that the appearance of LT fragment contributed minimally to the occurence and degree of the “reverse wedge effect”. When preventing the reverse wedge effect in basicervical TF variant, the lesser trochanter fragment did not have to be reduced and temporarily stabilized.
It is unclear how this intraoperative complication negatively impacts the treatment outcome. The postoperative measurement demonstrated that the magnitude of the secondary displacement in the “reverse wedge effect” was significant. Zhang et al. advocated the “medial positive reduction” concept to validate a stable medial cortex mismatch, but they did not quantify the threshold of the acceptable diastasis at the medial cortex [20]. It is reasonable that an obvious opening at the primary fracture line represents instability in the fracture. Thereafter, postoperative excessive collapse would occur because of failure to restore cortex interdigitation. Early postoperative weight bearing had to be postponed. The postoperative radiograph demonstrated a mild tendency of valgus reduction in the majority of patients. Although mild valgus reduction is more preferred than varus reduction for allowing interfragment compression and reducing bone-implant stresses, Ciufo et al. recently showed that residual basicervical gapping was closely associated with fixation cutout [21]. Considering that the magnitude of the secondary displacement exceeded 4 mm, which was proposed by Baumgaertner as the lowest threshold for malreduction [22], we believe that this newly reported intraoperative complication is worth further investigation.
There were some limitations to the present study. First, this study was a retrospective analysis. Therefore, the robustness of the analysis is undetermined. Second, a comparison between the different trochanteric fracture patterns in the incidence rate of the “reverse wedge effect” was not carried out. However, we rigorously reviewed all the intraoperative fluoroscopy findings during the study period and did not find any “reverse wedge effect” in other fracture patterns. Thus, we believe that this complication is closely correlated with the nailing basicervical TF variant combined with the PLF. Third, the treatment outcomes of this cohort of patients were not evaluated for several reasons. The primary reason is that this cohort of patients was highly heterogeneous, and some very young patients were included, which made building an age- and sex-matched control group difficult.