Introduction

The correct alignment of the implant components has been cited as the one of the most important aspects of a successful knee arthroplasty [2, 9, 12]. Although numerous bone and soft tissue landmarks have been advocated [5, 6, 15–17], it is difficult to align an extramedullary guide to the mechanical axis (MA) of the tibia [8, 10, 14]. The same is true for patient-specific positioning guides and it is recommended to perform the alignment check using anatomical landmarks before making the cuts [4, 14].

The second metatarsal (MT2) is a well-known distal landmark for the extramedullary tibial cutting guide in total knee arthroplasty (TKA). Most of the manufacturers’ instruction manuals recommend that the alignment rod through the tibial cutting block should point to the MT2 when an extramedullary alignment check is performed. However, to our knowledge, there is no study on the accuracy of predicting proximal tibial cutting using the MT2 as a distal landmark for the extramedullary guide in TKA in RA patients or in osteoarthritis (OA) patients.

Because any rotational foot abnormality would affect the accuracy of the MT2 as a landmark for tibial cutting, it was hypothesized that the accuracy of the MT2 is not high, especially in RA patients whose foot joints are apt to be involved.

This study was conducted to answer the following questions. (1) How often do outliers occur when MT2 is used as a distal landmark for the extramedullary tibial cutting guide in TKA? (2) Is the accuracy of the predicted tibial cutting using MT2 as a landmark in RA patients inferior to that in OA patients? (3) Is the accuracy of the predicted tibial cutting affected by foot involvement in RA patients?

Materials and methods

This computer simulation study involved a total of 93 knees: 45 knees in 39 consecutive female patients with OA and 48 knees in 34 consecutive female patients with RA. All patients with OA or RA who were candidates for primary TKA were considered for inclusion in this study. Exclusion criteria included evidence of trauma, infection, tumor, or any congenital disorder. The demographic data of the patients are shown in Table 1. For computed tomography (CT) scans, the patient was placed in the supine position on a table and the knee was extended while the ankle was maintained at 0° of flexion with a leg holder. Preoperative high-resolution CT scans of all the affected lower limbs, including the whole tibia and foot, were performed as a standard exam for TKA preoperative planning using a 16-detector CT unit (Toshiba Medical, Japan) in the helical mode in a 512 × 512 matrix, and setting slice thickness at 1 mm. Three-dimensional (3D) CT-based preoperative TKA planning software (ZedKnee® LEXI Co., Ltd., Tokyo, Japan) was used to determine the tibial MA and perform the measurements. The tibial MA was defined as a straight line from the center of the appropriate-sized tibial component without posterior slope to the center of the distal tibial plafond [17]. The AP axis of the tibia was defined as a straight line connecting the mid-posterior cruciate ligament (PCL) attachment with the medial edge of the patellar tendon attachment (Akagi line [1]).

Table 1 Demographic data of the patient groups
Full size table

For RA patients, the Larsen grades [11] were applied to each of the five joints related to the MT2 axis (i.e., ankle, subtalar, talonavicular, navicular–second cuneiform and second cuneiform–second metatarsal joints) using reconstructed CT images. Foot involvement was defined as a foot that had a Larsen grade 3 or a high-grade arthritic change in at least one of these five joints.

Two simulations were performed using the virtual extramedullary cutting guide, which was 8 cm long between the proximal fixation spike and the rod. The rotational alignment of the guide was the Akagi line and the rod was pointed toward (1) the base of the MT2 and (2) the distal part of the MT2 (Fig. 1). The spike of the extramedullary guide was affixed to the point of the plateau on the line that passed through the center of the keel and parallel to the component axis (Fig. 1a, b). The angle between the projected lines of the MA of the tibia and the longitudinal axis of the tibial cutting guide in the plane perpendicular to the Akagi line and the MA of the tibia was measured (Fig. 1c, d). A positive value indicated varus of the predicted tibial alignment. Preoperative TKA planning automatically measures the angle by dotting the reference points on reconstruction CT images (Fig. 2). To measure the test–retest reliability, each set of measurements was repeated on 20 randomly selected subjects with 6 months intervening. The outliers were defined as deviations >3° from the MA of the tibia [12]. The accuracy of the predicted tibial osteotomy using the MT2 as a landmark for the extramedullary cutting guide in RA patients with or without foot involvement was also evaluated. The angle between the MT2 axis and the Akagi line was measured. A positive value indicated that the MT2 axis was rotated externally compared with the Akagi line (Fig. 3). This study was approved by the Chiba Rehabilitation Center Institutional Review Board (ID number of approval: Iryou 24-5).

Fig. 1
figure 1

Two simulations of the predicted tibial cutting using the Akagi line as the anteroposterior axis of the guide. a Figure as seen from above. b Lateral view. The alignment rod was pointed toward c the base of the second metatarsal (MT2) and d the distal part of the MT2. MA mechanical axis of the tibia

Full size image
Fig. 2
figure 2

Angle measurement by preoperative TKA planning. a Lateral view. b Anteroposterior view. c The plane including the distal part of the MT2

Full size image
Fig. 3
figure 3

The angle between the second metatarsal (MT2) axis and Akagi line

Full size image

Statistical analysis

Chi square tests were used to analyze the differences in frequency of outlier occurrences between OA and RA groups. Paired t tests and Chi square tests were used to analyze the differences of predicted alignment and frequency of outlier occurrences between the base and distal part of the MT2, respectively. The frequencies of outlier occurrence and foot involvement in RA patients were analyzed by Fisher’s exact tests. The relationship between the deviation from the MA and the angle between the MT2 and the Akagi line was analyzed by Pearson’s correlation. A p value <0.05 was considered significant. A sample size calculation showed that for the Chi square tests, a total of 88 knees would allow detection of a moderate effect size (ω = 0.3), with a power of 0.80 and α = 0.05.

Results

The test–retest reliability coefficient was excellent (base of the MT2: 0.995; distal MT2: 0.993). The mean deviations from the MA of each simulation are shown in Table 2. A wide range of variability was found in both OA (base of MT2: range 2.9° to −6.9°, distal part of MT2: range 4.5° to −7.9°) and RA patients (base of MT2: range 8.1° to −5.0°, distal part of MT2: range 11.6° to −6.1°), especially when using the distal part of the MT2 (Figs. 4, 5). Statistically significant differences between the base and the distal part of the MT2 were identified with respect to the outlier frequency.

Table 2 Predicted alignment of tibial cutting (degrees ± SD) and number of outliers
Full size table
Fig. 4
figure 4

Histogram of the deviation from the mechanical axis of the tibia when the alignment rod pointed toward the base of the second metatarsal (MT2)

Full size image
Fig. 5
figure 5

Histogram of the deviation from the mechanical axis of the tibia when the alignment rod pointed toward the distal part of the second metatarsal (MT2)

Full size image

No significant difference between OA and RA was found with respect to the outlier rate (base of MT2; 11.1 vs 14.6 %, n.s., distal part of MT2; 33.3 vs 31.3 %, n.s.). However, foot involvement in RA patients demonstrated a trend toward significance (base of MT2; p = 0.078, distal part of MT2; p = 0.068).

The mean angles between the MT2 axis and Akagi line (Fig. 2) were −10.0° ± 10.8° (−40.6° to 18.5°) and −5.3° ± 16.5° (−41.6° to 45.6°) in OA and RA patients, respectively. Significant positive correlations (r = 0.84, p < 0.001 for both the base of the MT2 and the distal part of the MT2) were found between the alignment deviation and the angle between the MT2 and the Akagi line.

Discussion

The main finding of this study was that the outlier rate was relatively high. Outlier rates of more than 10 % were found in both OA and RA groups when the guide was pointed toward the base of the MT2 and more than 30 % when it was pointed toward the distal part of the MT2. A second important finding was that there was no significant difference between OA and RA with respect to the outlier rate. However, foot involvement in RA patients has a negative effect on the predicted tibial cutting alignment.

Many authors have proposed distal landmarks, such as points 3–5 mm medial to the malleolar center [5], the tibialis anterior tendon [15] and the extensor hallucis longus [16], for the extramedullary cutting guide in TKA. A 5 mm difference in the horizontal plane causes ≤1° change in the coronal alignment in patients with a 300 mm long tibia (300 mm × tan1° = 5.25 mm), so malalignment is not simply due to the difficulty of distal centering. The rotational mismatch between the proximal tibia and the ankle joint has a great impact on the accuracy of the proximal tibial cut in TKA [3, 13, 18]. This was not considered in these studies, or was there a clear definition of the AP axis at the ankle. If the definition of the AP axis changes, the distance between the projected ankle center on the skin and the points that are recommended as landmarks would also change (Fig. 6), but only by a small amount. The rotational error occurs mainly in the proximal tibia and is amplified by the distance between the bone and alignment rod [18]. Therefore, the outlier rate can be reduced by the surgeon’s meticulous attention to avoid the rotational mismatch regardless of which distal landmark on the ankle is chosen. On the other hand, there may be a large rotational mismatch of the MT2 with the tibial AP axis, which is unavoidable because the MT2 is an extra-articular landmark. This rotational mismatch is amplified by the distance between the ankle center and the point used for measurement on the MT2. This is why the outlier rate was significantly lower when the alignment rod pointed toward the base of the MT2 than toward the distal part of the MT2.

Fig. 6
figure 6

If the definition of the AP axis at the ankle changes (from A to B), the projected ankle center on the skin would also change (from a to b)

Full size image

Better landmarks than the MT2 have been identified in other studies [17, 18]. Using the middle one-third of the tibial crest [7, 17], a much lower outlier occurrence rate (2 %) should be possible. Another report found no outliers on the simulation TKA when the medial border of the tibial tubercle and the center of the distal soft tissue were used as the proximal and distal centers, respectively, and when the AP axis of the guide was matched to the Akagi line proximally and distally [18]. Considering the high outlier occurrence rate of over 10 %, the MT2 is not a useful landmark compared with these previously proposed landmarks. Furthermore, in a clinical setting, the foot is easy to rotate during surgery. This rotation may adversely affect the reliability of the MT2 as a landmark.

This study has limitations. First, anatomical features may differ with the ethnic origin of patients. Because all subjects in this study were ethnically Japanese, the findings might be difficult to directly extrapolate to a patient population of a different ethnic origin. Second, the results were derived from a simulation using the CT image of the patient with the knee extended and the ankle maintained at 0° of flexion with a leg holder, which was different from the operative posture. Third, this study did not take into account the fact that the foot is easy to rotate and it may adversely affect the reliability of the MT2 as a landmark. Despite these limitations, the results demonstrated that the outlier rate of the predicted tibial osteotomy using MT2 as a landmark was not less than 10 %.

Most of the manufacturers’ instruction manuals recommend the MT2 as a distal landmark in TKA, with no evidence of its accuracy. The results presented here serve to alert TKA surgeons to the possibility of outlier occurrence.

Conclusion

It has been demonstrated in this study that surgeons who use the MT2 as a landmark for the extramedullary tibial cutting guide in TKA should aim toward the base of the MT2, but avoid using it in RA patients with foot involvement. Moreover, they should be aware that the MT2 is a less accurate landmark in TKA than other previously reported landmarks. Therefore, the MT2 should not be used alone as a landmark for the extramedullary tibial cutting guide in TKA.