The Effect of Intra- vs. Extramedullary Tibial Guides on the Alignment of Lower Extremity and Functional Outcomes following Total Knee Arthroplasty: A Randomized Clinical Trial

Background: Total knee arthroplasty (TKA) has been known as a denitive treatment of advanced knee osteoarthritis. Both intra- (IM) and extramedullary (EM) tibial guides have been used to restore the desired extremity alignment. However, controversy exists regarding the superiority of either technique. We aimed to compare their functional outcomes and accuracy in providing neutral alignment after TKA. Methods: In a randomized, double-blinded clinical trial, we studied 98 patients undergoing primary TKA in two groups of IM and EM. We measured the medial proximal tibial angle (MPTA), varus angle (VA), and joint-line convergence angle with normal ranges of 90°±3°, 0-2°, and 0±3°, respectively, on a three-joint alignment view after three months. We also assessed functional outcomes at the last follow-up. Finally, we compared these outcomes between groups. Results: Eighty-four patients (IM=42, EM=42) were included in the nal analysis (16 males, 68 females; mean age: 63.9±8.6 years; mean follow-up: 27±2.9 months). The mean postoperative alignment angles showed no signicant difference, although MPTA outliers were signicantly more frequent in the EM group (26.2% vs. 9.5% in IM, P=0.04). None of the functional outcomes showed a signicant difference between groups. However, the mean ROM increase was signicantly higher in VAs within ±3° of normal than those outside it (30.8 vs. 27.4, respectively; P=0.039). Conclusions: We conclude that both techniques were not different in terms of the mean alignment angles and functional outcomes. However, fewer MPTA outliers can be seen with IM. A postoperative mechanical axis within ±3° of neutral can result in a more ROM increase after one year.


Background
Total knee arthroplasty (TKA) has been known as an end to analgesics use in patients with knee osteoarthritis and as a means of restoring the physical activity and function (1). Several factors are in uential in the long term outcome of a TKA, including patient selection, the prosthesis of use, and surgical technique (2). The surgical technique is of great importance among knee surgeons as it should implement proper lower extremity alignment for a good long-term prognosis (1)(2)(3). Several studies have reported poor outcomes of the lower extremity malalignment, as prosthesis mispositioning can lead to loosening, recurrence of pain, and compromised physical performance (4-8).
Among surgical techniques, both intra-(IM) and extramedullary (EM) tibial guides have been used to provide the desired lower extremity alignment. However, there are still controversies over the outcome of TKA using either technique to provide a neutral alignment (1,9). Some studies have preferred one for providing a more accurate tibial alignment (9)(10)(11), while the others found no signi cant difference between either technique (1,(12)(13)(14). Many studies have demonstrated that most knees are suitable for both techniques; however, this is not always the case. EM guides are unreliable in cases of soft tissue or ankle abnormality, whereas IM technique is not preferred in patients with excessive tibial bowing, previous fracture, or retained metalwork (9). Moreover, the functional outcomes following both techniques were compared only by few studies (10,15). In the present study, we decided to compare the accuracy of IM and EM techniques in providing neutral lower extremity alignment in patients undergoing TKA. We also compared the functional outcomes following TKA between both techniques.

Methods
This randomized clinical trial was designed and reported based on the Consolidated Standards of Reporting Trials (CONSORT) principles. The study protocol was reviewed and approved by the institutional review board of our university. The Ethics Committee Board declared no ethical concern in the current study. All the patients were informed of the fact that the results of this study are to be published in scienti c medical journals and gave their informed consent. Moreover, the Ethics Committee Board allowed the publishing of study results. This trial was registered in 21/04/2020 in the clinical trial registry of our country with the identi cation number of IRCT20160809029286N5.

Study design and participants
This study was a randomized controlled clinical trial. A total of 98 consecutive patients who met the eligibility criteria of the study were enrolled from the orthopedic clinic of our institute. The inclusion criteria were patients with primary knee osteoarthritis indicated for TKA, who had a varus or neutral knee alignment. The exclusion criteria were: 1. hemophilia, 2. in ammatory knee arthritis like rheumatoid arthritis, 3. previous tibial fracture, 4. genu valgum, and 5. inadequate radiographs.

Study protocol and interventions
The enrolled patients were admitted and underwent pre-operative imaging and TKA according to the study protocol. The patients participated voluntarily and had signed the informed consent formula. An assistant researcher did the data collection and clinical assessment of the functional outcomes following the surgery. A single experienced knee surgeon performed all the TKAs.

Imaging protocol
A full standing three-joint alignment view (3-JV) of the lower limb together with the anteroposterior and lateral knee projections were performed for each patient before and three months after the surgery. For the 3-JV, the patient must stand with the knees in full extension and no aid. The patella should face forward in the direction of the x-ray beam. The collimator was set superiorly on the iliac crest and inferiorly on the sole. The radiograph is composed automatically after three projections done by CARESTREAM DRX-Evolution device (Carestream Health Inc., US). The alignment angles of the lower limb were measured on 3-JV using the MediCAD software by a single experienced knee surgeon. The acceptable (inlier) range for MPTA was considered 90°±3°, JLCA 0° to 2°, and VA 0 ± 3°.

Surgical technique
After spinal anesthesia and standard prepping and draping, TKA was performed using a standard anterior midline incision with the anteromedial approach. A pneumatic tourniquet had been applied above the knee prior to the operation. Releasing the medial soft tissue of the knee, we subluxated the patella and subsequently exposed the proximal tibia. For femoral cuts, we used an intramedullary jig. To perform tibial cuts, we used an intra-or extramedullary jig in each group accordingly. The entry point for IM jig was the center of the tibial axis, according to preoperative planning, on tibial plateau anterior to the insertion of the anterior cruciate ligament. The EM jig was mounted on the leg parallel with the tibial axis and leveled proximally with the tibial crest in the coronal plane and distally with the talus center in the axial plane (3 mm medial to ankle center). The tibial bone cut was done with a posterior slope of 3°. The tibial component was placed in a 3° external rotation. In all the patients, a cemented posterior stabilized (PS) NexGen© LPS-Flex Knee prosthesis (Zimmer Biomet ®, Warsaw, Indiana, US) was used. Finally, standard wound closure and care were done.

Outcomes
The primary outcome measures were the radiologic and functional outcomes, which were assessed both pre-and postoperatively. The radiologic outcomes included the medial proximal tibial angle (MPTA), mechanical femoral mechanical tibial or varus angle (MFMTA or VA), and joint-line convergence angle (JLCA). The functional outcomes included the knee society score (KSS), functional knee society score (fKSS), pain visual analog scale (VAS), and the measurement of knee range of motion (ROM). The postoperative assessment of the radiologic and functional outcomes was at three months, and the last follow-up visit, respectively.

Sample size
The sample size was calculated based on the study of Chin et al. (16) that compared the three techniques of IM, EM tibial guides, and computer-navigated surgery in the patients who underwent TKA. We used the risk ratio of postoperative MPTA angle outliers ( = 43.34% IM vs. =13.34% EM) as a reference value for power analysis and assumed a β value of 20% and α of 5%. We found that 35 patients per each group (70 patients in total) were required to achieve statistical signi cance. The calculations are presented below. We considered a sample size of at least 90 to compensate for possible losses during follow-up. The patients were randomly assigned to two groups of intramedullary (n = 48) and extramedullary (n = 48) based on the tibial guide used during the surgery. The patients had an equal chance of being randomly assigned to each of the two arms of the study. The randomization of the patients was done using the permuted balanced block method. Six blocks of four were assumed, and the patients were divided into 21 sequentially numbered groups. The groups were randomized using a list of random numbers generated by Microsoft Excel 365, and the patients of each group were allocated to each intervention arm accordingly. The randomization sequence was concealed before the enrollment until the patient was transferred to the operation room. An independent researcher, who was not involved in the process of data collection and outcome assessment, performed the randomization. This study was tripleblinded, as neither the patient nor the assistant researcher nor the analyzer researcher was aware of the technique used during the surgery.

Data analysis
Data analysis was done by SPSS 25.0 software (SPSS Inc., Chicago, Illinois, US). The normality of the variables was tested by the skewness-kurtosis and Spearman Tests. The Student's independent t-test was used to compare continuous outcome variables such as the alignment angles in both groups. The chisquare and Fischer's Exact tests were used to compare the nominal outcome variables. The signi cance level was considered 0.05.

Results
A total of 98 patients who underwent TKA using either EM or IM tibial guide were enrolled in the study from September 2018 to May 2019. Two patients were excluded as they declined to participate in the study. The patients were randomized into two groups of IM (n = 48) and EM (n = 48) tibial guides. All the patients received the allocated intervention. There was no loss to follow-up. Twelve patients were excluded from the nal analysis due to inadequate postoperative radiographs. The CONSORT ow chart is shown in Fig. 1.
The mean age and body mass index (BMI) were 63.9 ± 8.6 and 29.6 ± 4.8, respectively. There were 16 (19%) males and 68 (81%) females. The mean follow-up duration was 27 ± 2.9 months (range = 7-19). As seen in Table 1, both groups had the same and matched demographics (P > 0.05). The mean ± standard deviation is reported for age, BMI, and follow-up (* independent t test).
The frequency is reported for sex (** chi-square test).
As seen in Table 2, the mean of the lower limb alignment angles, including MPTA, JLCA, and VA, was compared between both groups preoperatively and at three months postoperatively. Our data showed that there was no signi cant difference between EM and IM groups in terms of the lower limb alignment angles (P > 0.05). As mentioned earlier, we considered the normal range (inliers) for MPTA, JLCA, and VA as 90° ±3°, 0 to 2°, and 0 ± 3°, respectively. The outlier cases of MPTA, JLCA, and VA are presented in Table 3. The postoperative MPTA outliers were signi cantly more frequent in the EM than the IM group (26.2% vs. 9.5%, P = 0.04). However, no signi cant difference was found between the JLCA and VA outliers of the two groups (P > 0.05).  * independent t test The patients were followed for an average of 13 ± 2.9 months. The functional outcome measures, including KSS, fKSS, VAS, and the knee ROM were measured for all the patients preoperatively and in the last follow-up visit, as seen in Table 5. However, no signi cant difference was found between the EM and IM groups in terms of these outcome measures at a mean follow-up of 27 ± 2.9 months (P > 0.05). The preoperative and postoperative VAS pain scores showed no signi cant difference as well (1.09 ± 1 vs. 1.2 ± 1, P = 0.45). * independent t test The functional outcomes were also compared between the inlier and outlier ranges of postoperative VA, which demonstrated no signi cant difference except for the increase in ROM. In those patients with a VA within ± 3° of normal, the mean increase in ROM following TKA was signi cantly higher than that of those with a VA outside ± 3° of normal (30.8 vs. 27.4, respectively; P = 0.039) (Fig. 2).

Discussion
The controversy over IM and EM tibial guides has existed in the literature since the very introduction of both techniques (1,9,17). Based on a systematic review, the published literature includes nearly twenty original articles on this issue over the last three decades. Of these, 52.6% believed the two techniques had comparable accuracy, 36.8% preferred IM guides, and 10.5% found EM guides more accurate (1).
However, not all these studies had su cient sample size, proper methodology, or reported all the important radiologic or functional outcomes. Accordingly, only six of them were eligible for the metaanalysis by Zeng  The VA or MFMTA is the angle between the mechanical axis of femur and tibia in AP radiograph, which has also been designated as the mechanical axis angle (1,16,18 is also consistent with the results of these studies. The tibial slope is another angle, which also showed no signi cant difference in the metaanalysis study of Zeng et al. (1). We did not report this measure. However, we compared post-op JLCA between the two techniques, not reported before, which also showed no signi cant difference.
In the ancillary analysis of outlier data, we found that the mean weight of the patients was signi cantly lower in the MPTA outliers than that of the inliers of the EM group, although no such difference was seen in the IM group. However, BMI showed no signi cant difference between the outliers and inliers of each group. We found no relevant data in the previous studies comparing both techniques. Nevertheless, a recent study by Compton et al. showed that BMI did not in uence the postoperative MPTA following TKA using EM guides. They found no difference between the MPTA outlier (de ned as outside ± 5°) ratio of patients with a BMI < 35 (2/100) and ≥ 35 (2/62) (20). It was also not signi cantly different between those with a BMI < 35 (9/34) and ≥ 35 (1/8) in EM group of our study, which is consistent with the ndings of Compton et al. However, the signi cantly lower weights of the patients, who underwent TKA using an EM guide and have a nal MPTA outside the normal range, might indicate that thinner legs are associated with a higher chance of error in the appropriate mounting of an EM jig and doing tibial cuts. It might be due to the underestimation of a thinner leg by the surgeon, as he usually expects the obese legs to be a challenge. We think further comparative studies between both techniques are needed, which focus particularly on the length and girth of the leg to determine the effect of these parameters on the nal lower extremity alignment. Oxford Knee Score (OKS) between 107 patients undergoing TKA using IM, EM, and computer-assisted techniques at a median follow-up of 46 months. None of these measures differed signi cantly between IM and EM techniques, although OKS adjusted for sex and age showed a difference, close to the signi cance level, between computer-assisted and conventional techniques (15). In our study, the postoperative values of KSS, fKSS, pain VAS, and the increase in ROM were not signi cantly different between IM and EM groups at a mean follow-up of 27 ± 2.9 months. However, a postoperative mechanical axis within ± 3° of neutral was associated with a more increase in knee ROM than that outside this range (Fig. 2).

Conclusions
In this randomized clinical trial, no signi cant difference was found between IM and EM techniques for TKA in terms of the mean postoperative lower extremity alignment angles, including MPTA, JLCA, and VA. However, the IM technique was associated with fewer postoperative MPTA outliers. Both techniques were found equal regarding the good functional outcomes seen with both. In general, a postoperative mechanical axis within ± 3° of neutral was associated with a more increase in ROM at mean 27-month follow-up. Ethics approval and consent to participate: All the patients participated voluntarily in this study and had signed the informed consent formula. The Ethics Committee Board of Tehran university of medical sciences approved of the methodology and declared no ethical concern in this study. All the methods were performed in accordance with the guidelines and regulations of the institutional review board of our university.
Consent for publication: All the patients were informed of the fact that the results of this study are to be published in scienti c medical journals and gave their informed consent. Moreover, the Ethics Committee Board allowed the publishing of study results.
Availability of data and materials: The datasets of the current study are available from the corresponding author on reasonable request.
Competing interests: The authors declare that they have no con ict of interest.
Funding: This research did not receive any speci c grant from funding agencies in the public, commercial, or not-for-pro t sectors. The CONSORT owchart of the study