Background
In total knee arthroplasty (TKA), the prosthetic placement and overall limb alignment has been demonstrated to be most influential in determining implant survival [
1‐
4]. The ideal position for the components recreates a neutral mechanical axis. Most surgeons currently favor intramedullary (IM) alignment for its ease of use and accuracy as compared with extramedullary (EM) alignment. Previous comparisons of IM and EM femoral alignment systems have shown the former to be more accurate in performing the distal femoral cut [
5‐
7]. In these studies, the IM alignment technique has ranged from 85 to 96% in the normal range as compared with 69 to 86% for the EM alignment technique.
However, the IM femoral alignment system does not always guarantee the accuracy of the component position in TKA. Previous studies used only the anterosuperior iliac spine as an intraoperative landmark for the EM referencing instruments [
5‐
7]. Intraoperative visual assessment of the longitudinal femur axis and the anterosuperior iliac spine by EM rods is difficult owing to the large soft tissue cover and tourniquets. TKAs using extramedullary femoral alignment guides have been extensively studied in recent years. Extramedullary instruments using newly designed mechanical axis marker systems have provided as accuracy in reproducing a neutral distal femoral resection on the coronal and sagittal planes during TKA as standard IM instruments [
8‐
10].
Recent advances in TKA have focused on the reduction of damage during the procedure [
11‐
13]. One of the most invasive parts of TKA is the violation of the intramedullary femoral canal and the subsequent use of IM instruments. The use of an intramedullary guide for the femur can result in various complications such as blood loss, postoperative hypoxia, intraoperative fractures, and fat embolism [
11,
14,
15], while the EM method has less morbidity in terms of blood loss because it does not invade bone marrow. In addition, the IM instrument may not be applicable when a long-stemmed femoral component implanted during a previous surgery remains or a rod cannot be inserted due to severe deformity of the femur [
9].
In this study, we conducted a meta-analysis of pooled the data from relevant RCTs to evaluate whether an IM or EM femoral guide is more accurate in assuring correct femoral positioning. Moreover, the blood loss and operation time were also compared between these two techniques. Our hypothesis was that the EM femoral guide provided similar accuracy for femoral positioning and less blood loss compared with the IM femoral guide during TKAs. If the hypothesis is confirmed, the EM femoral guide may provide an alternative approach for femoral cuts and display clinical benefit for particular cases, such as a previous surgery remains and severe deformity of the femur.
Discussion
Our meta-analysis compared the radiographic outcomes between the EM and the IM femoral guiding technique in patients undergoing TKA. No significant differences were found between the two groups in terms of the lower limb coronal alignment, the coronal alignment of femoral component, and the sagittal alignment of femoral component. The EM guide was associated with a less blood loss and exhibited a similar operation time as the IM guide.
Long-term success after TKA is dependent especially on proper intraoperative component positioning [
1‐
4]. Most authors have reported favorable femoral cuts made with the IM guide and it is considered to be the most accurate with a statistically significant increase in the percentage of distal femoral cuts [
6,
7]. Several studies found 85–96% of IM femoral cuts to be acceptable compared with 69–86% of EM cuts. The EM femoral alignment technique had inferior accuracy with approximately 10% more outliers on the coronal plane compared with the IM technique [
9]. Historically, all the authors were using the EM instruments referring only to the anterosuperior iliac spine (ASIS) intraoperatively [
7,
16]. However, the use of the ASIS to locate the femoral head center (FHC) might not be an accurate method since the FHC was indirectly determined by and dependent on anatomical structures adjacent to the femoral head. Identification of the FHC using the two finger-breadths medial to the ASIS method was unreliable and a wide variation of inter-ASIS distances was found among patients [
17].
Recently, some studies have introduced new techniques to improve the accuracy of the extramedullary alignment guide system. Baldini and Adravanti [
8] developed a set of EM instruments calibrated with preoperative templating radiograph measurements of inter-femoral head center distance (IFD), which allowed one to perform distal femoral resection without violating the femoral canal. They reported that the femoral component coronal alignment was within 0° ± 2° of the mechanical axis in 84% of the IM group and 86% of the EM group. Matsumoto et al. [
18] reported that the femoral component coronal alignment was within 0° ± 3° of the mechanical axis in 98% of patients by using a similar EM instrument. Seo et al. [
19] using preoperative templating radiograph measurements of IFD showed that outliers (±3°) of the femoral component coronal alignment were observed in 9.4% of all cases. However, the proximal reference point in the coronal plane was considered to be incorrect when the lower limb was abducted or adducted during surgery. Therefore, Seo et al. [
20] reported the extramedullary technique assisted by a mechanical axis marker, which could easily identify the center of femoral head and result in 98.2% of patients achieving acceptable alignment in the range of 0° ± 3° in the coronal plane. Jung et al. [
9] reported that the femoral component coronal alignment was within 90° ± 5° in 98.4% of patients using the mechanical axis marker system. This new EM alignment guide system was accurate since it included proximal and distal coronal axis markers to indicate the IFD, which was independent of leg posture. The use of an intramedullary guide for the femur was associated with increased risks of fat embolism, blood loss, postoperative hypoxia, and intraoperative fractures [
11,
14,
15]. In addition, the IM femoral alignment system did not always guarantee accuracy of the component position in the TKA. The femoral bowing was a common phenomenon and could affect axial alignment of TKA when IM alignment systems were used, especially in East Asian populations [
21,
22]. In these cases, the EM femoral alignment system was a useful alternative surgical option to adjust femoral component alignment [
9]. In patients with a more bowed femur, malalignment of lower limb may occur in IM technique, while Computer-assisted techniques could improve the accuracy in both sagittal and coronal planes and EM technique could improve the accuracy in coronal planes. Although the computer-assisted TKA enabled more accurate component alignment [
23,
24], the 10-year outcomes of computer-assisted TKA are not superior to that of the conventional technique in function, patient satisfaction and implant survivorship [
25,
26].
The opening of the medullary canal using intramedullary jigs was postulated to cause significant blood loss during TKA, although most surgeons have closed the femoral canal opening with bone plug. The application of EM technique was associated with minimal invasiveness since the femoral canal was not breached and the blood loss could be reduced by 145–396 mL [
8,
14,
27‐
29]. Computer-assisted TKA and patient-specific instrumentation (PSI) were recently introduced with the aim of improving alignment without violating the femoral canal [
30‐
32]. TKA using PSI did not result in significantly better femoral component alignment in the sagittal and axial planes than TKA using conventional instrumentation [
32]. However, the blood loss was significantly reduced by using the PSI system compared with the IM system for femoral cut. While, the computer-assisted TKA enabled more accurate component alignment. However, it did not reduce the hidden blood loss since the blood loss avoided by not opening the canal might be compensated by greater post tourniquet bleeding due to greater tourniquet time [
31].
This present meta-analysis has several limitations. First, only four studies were included and the sample size of the included studies was small, which might have affected our results. Second, we could not perform a valid statistical comparison of the functional outcomes between the two groups. Therefore, further high-quality RCTs with long-term follow-up should be designed to assess radiographic outcomes, knee function, and implant survival rate.
Acknowledgements
We thank all authors of the included studies and the authors listed have made substantial contributions to the study HXL conceived the study design.