The long-term outcome of a total knee arthroplasty (TKA) depends on patient selection, implant design, and surgical technique.32 The most common reasons for early failure are component malalignment and extensor mechanism complications4,34,43 resulting in abnormal wear patterns and component loosening.5,10,15,32-34,37,39,43,44,46 Varus and valgus malalignments greater than 3° accelerate polyethylene wear and early TKA failure.2,11,13,26,40 Malalignments greater than 3° occur in ⅓ of all TKAs even when performed by experienced surgeons with modern techniques.40 Several surgical navigation systems for TKA have been introduced to reduce early failure caused by suboptimal component positioning.8,9,20,36,40
There are three types of computer-based navigation systems. The first is an image-based system using preoperative computed tomography (CT) scans to guide component positioning. This approach requires additional preoperative imaging and considerable preoperative planning to process the radiographic data. The second is image-free navigation based on real-time data gathered perioperatively using kinematic analysis. The third is transferring intraoperative fluoroscopic images to a workstation to virtually plan component positioning.19,20
Navigation provides more precise component positioning with fewer outliers.1,4,6,7,22,25,38,42 Its main disadvantages are the additional time and cost.1,6,7,14,35,42 It is unknown whether the advantage of improved alignment is from more precise femoral component positioning,1,25,35,38 tibial component positioning,23 or both.6,7,14,22,24,42
We hypothesized that the less time consuming image-free component positioning using femoral navigation alone would be as accurate as any other navigated positioning method. We also raised the following questions: (1) Were any additional costs or time requirements associated with navigation?; (2) How accurate was femoral and tibial component positioning with or without navigation?;(3) Was there a correlation between component malalignment and mechanical axis malalignment between groups?; and (4) Was there a decrease in total navigation time without tibial navigation, and did it affect component positioning accuracy?
MATERIALS AND METHODS
We prospectively enrolled 88 patients (90 consecutive knees) with osteoarthritis (OA) having TKAs. Informed consent was obtained from all patients. Patients with rheumatoid arthritis (RA) or other forms of inflammatory arthritis were excluded. Thirty patients had image-based navigation TKAs (image-based group), 30 patients had image-free navigation TKAs (image-free group), and 30 patients had conventional TKAs (conventional group). They were grouped sequentially based on the availability of the technologies studied. Two experienced orthopaedic surgeons (PMB and RH) performed the TKAs. We used a medial parapatellar approach for patients with varus and valgus mal-alignment less than 10°. Patients with a preoperative valgus mal-alignment greater than 10° were operated on using a lateral approach with an osteotomy of the tibial tuberosity.
Four patients who had surgery with computer navigation were excluded. Surgical navigation was stopped in three patients in the image-based group and in one patient in the image-free group because of hardware problems beyond the surgeons' control that necessitated conversion to standard TKAs. One patient in the image-based group and one patient in the conventional group were lost to followup. We recorded age, side, gender, varus/valgus malalignment, and preoperative Knee Society score (Table 1).12,16 There were no differences among the groups.
We used the Navitrack® surgical navigation system (Zimmer, Warsaw, IN ). The system consists of an infrared-based reflex camera system, a portable workstation, a monitor, a keypad, and a foot control. Additional equipment includes two rigid bodies (trackers) containing reflex elements (spheres), and 2.5-mm stainless steel bicortical K wires designed to fix the rigid trackers to the bone. The optical localizer consists of an infrared lamp sending out an infrared signal, which is reflected by the spheres and detected by cameras. The rigid trackers are fixed to the femur and tibia with bicortical K wires. Leg and bone positioning are observed on the monitor during the operation. The surgeon guides the procedure using a pointer, a keypad, and a foot pedal.
We performed the CT-based image navigation technique with a computer program (Navitrack TKR 1.2, Orthosoft, Montreal, Canada ) that used CT scans to preoperatively create a three-dimensional (3-D) model with bony landmarks.26 We calibrated the instruments preoperatively and confirmed the intraoperative position of the sensors on the femur and the tibia. The patient's bony anatomy and bony landmarks then were matched to the virtual model. We also confirmed the instruments (a positioning block and a pointer) to observe displacement of the pointer and the positioning block in relation to the defined bony landmarks and anatomy. The computer showed the lines targeting the center of the femur and the center of the ankle. Rotation was based on the transepicondylar axis, the line tangential to the posterior condyles, and the anteroposterior (AP) line.3,45
We performed the image-free navigation technique without preoperative CT reconstruction. We calibrated the instruments preoperatively, and intraoperatively fixed trackers on the tibia and the femur. We defined the bony landmarks using the calibrated pointer. The 3-D model of the surgically treated knee was created by a kinematic analysis by moving the leg in different positions.
We performed the conventional technique using femoral intramedullary and tibial extramedullary instruments. The cuts were planned preoperatively on long, standing radiographs to measure the difference between the mechanical and anatomic axes. The intramedullary guide rod was inserted slightly anteromedial to the AP45 and transepicondylar axes. The length of the standardized intramedullary guide rod was 25 cm, with an 8-mm diameter.
We recorded the preoperative planning time and total operative time. The patients operated on using the lateral approach, including an osteotomy of the tibial tuberosity, were not included in the surgical time analysis. Therefore, operative time was analyzed for only for 22 of the 26 patients in the image-based group, 24 of the 29 patients in the image-free group, and 22 of the 29 patients in the conventional group. All patients received a rotating platform component (Innex® UCOR, Zimmer). The femur was cemented or press-fitted after intraoperatively assessing the distal femur bone stock. The tibial implant always was cemented. None of the knees received patellar resurfacing. Postoperative care included a femoral catheter for pain control and a standard rehabilitation program with full weightbearing started 1 day postoperatively.
We used the Knee Society outcome measurement tools12,16 to assess preoperative and postoperative ranges of motion (ROM), function, pain relief, and complications. Postoperative limb alignment was determined from full-length, standing AP and lateral radiographs of the knee. To obtain representative radio-graphs, we used a standardized protocol positioning the feet slightly rotated internally with the patella toward the AP beam. Two observers (MZ and DW) evaluated the radiographs in consensus and were blinded for the procedures. There were no differences in the means and the variances for the mechanical axis of the leg and the femoral and tibial components in the frontal plane.
Statistical analyses were performed using SPSS statistical software (Version 11, SPSS Inc, Chicago, IL). Differences between the demographics were analyzed with a Mann-Whitney U test and a chi square test. Differences between the groups were determined using the F Levene test for homogeneity of variance (constant variance)21 and the Mann-Whitney U test for comparing postoperative between-group alignment. Preoperative and postoperative Knee Society scores were compared by Wilcoxon rank tests. Post hoc power analysis was performed using a Fisher's exact binominal distribution test. A two-tailed value of p <0.05 was considered significant. We performed a linear regression analysis using decreased operative times as a function of the number of procedures to distinguish the learning curve and to obtain predicted values. We performed Spearman rank correlations to analyze the correlation of component malalignment.
The image-free navigation method excluding tibial navigation was as accurate as any other navigation system.
The additional costs consist of the price of the preoperative CT scan in the image-based group. The overall mean surgical time including the learning curve was longer in the image-based group (132 ± 26 minutes) compared with the image-free (114 ± 22 minutes; p = 0.017) and conventional (91 ± 21 minutes; p = 0.001) groups. The mean time requirement for preoperative planning for the image-based navigation system was 117 ± 51 minutes. The mean preoperative calibration procedure lasted 9 minutes for the image-based system and 5 minutes for the image-free system. Additional intraoperative time was necessary for navigation because of fixing the tracking device to the femur and tibia, and defining and confirming the axes, bony landmarks, and reference surfaces. This accounted for a mean time of 38 minutes in the image-based group and 25 minutes in the image-free group.
Both navigated groups showed less outlier of the mechanical axis than the conventional group. The mean and the standard deviation of the tibial components were not different in the navigated and the conventional groups. However, in both navigated groups, the femoral component was positioned less aberrantly from the mechanical axis than the conventional group. The mechanical axis of the leg (center of the hip and dome of the talus) was within 3° of neutral alignment in 89% of the patients in the image-based group, 83% in the image free-group, and 73% in the conventional group (Table 2). There were no differences in the means between the image-based group or image-free group and the conventional TKA group. There were no differences in the means or variances in the navigated groups. Compared with the conventional group(10.6°), the image-based (5.8°; p = 0.004) and image-free(6.4°; p = 0.009) groups had lower variances. Both navigated groups were more precise and had fewer outliers (Fig 1). The femoral component in the frontal plane was within 3° of neutral alignment in 92% of patients in the image-based group, 97% in the image-free group, and 80% in the conventional group. There were no differences between the means of the three groups for femoral frontal plane alignment. The conventional group had a greater variance (7.2°) in the frontal plane measurement than the image-based (4.1°; p = 0.005) and image-free groups(2.3°; p = 0.001) (Table 2). The navigated technique was more precise for femoral component alignment in the frontal plane. Tibial component positioning in the frontal plane was within 3° of neutral in 92% of the patients in the image-based group, 90% in the image-free group, and 90% in the conventional group. There were no differences between the means and variances (Table 2).
We found a correlation (r = 0.66) between malalignment of the femoral component and total mechanical axis malalignment in the conventional group. The probability value (p = 0.0001) suggests the two variables were linearly related. The correlation (r = 0.46; p = 0.021) of tibial component malalignment and total mechanical axis was weaker. Correlation coefficients for the entire series were lower for the femoral (r = 0.393) and the tibial components (r = 0.362), respectively. There was no correlation between component malalignment.
Surgical time decreased after we performed more procedures, suggesting a learning curve. Consecutive surgical interventions reduced (regression coefficient β, −0.732; p = 0.001) the total surgical time in the image-based group. Consecutive surgeries did not influence the total surgical time in the image-free group. However, the mean surgical times (excluding the first 15 patients) decreased (p = 0.009) to 109 ± 7 minutes in the image-based group and to 101 ± 17 minutes in the image-free group. The were no differences between the image-free and the conventional groups (Fig 2). After excluding tibial navigation, the total surgical times in the image-based group (106 ± 6 minutes) and in the image-free group (93 ± 19 minutes) decreased (Table 3).
There were technical difficulties during surgical navigation in four patients (three patients in the image-based group and one patient in the image-free group), necessitating conversion from surgical navigation to conventional TKA. Conversion was the result of hardware failure in two patients and tracker fixation-device loosening in two patients. One patient in the image-free group had arthrofibrosis, which was treated with manipulation under anesthesia. This patient also had a patella ligament rupture during manipulation and had reconstruction. There were no wound healing problems, no infections, and no additional extensor mechanism complications.
Computer-assisted surgery has become popular by minimizing component malalignment in TKAs. However, image-based and image-free navigation are time consuming. It is not known whether navigational support is mandatory in both components to achieve accurate prosthesis alignment. We determined the additional time requirements and costs for navigation and the accuracy of femoral and tibial navigation for component alignment. Because of the high positive and linear correlation between the femoral component and malalignment in the conventional group, surgical time can be reduced by tibial navigation only, not affecting accuracy in component positioning.
Our study has several limitations. First, our patients were not prospectively randomized. However, they were assessed prospectively and consecutively, which is helpful in limiting selection bias. Equal means among all three groups and equal differences between the variances of the image-based and image-free groups indicated a lack of a systematic selection bias between techniques. The order of the procedures (image-based before image-free) could have affected the learning curve. The surgeons were familiar with the navigational technology. Therefore, the learning curve may have had less influence on the surgical time of the image-free group. The read-out of the radio-graphs by two independent observers was not tested on reasonable reliability because of the evaluation in consensus. Another limitation is that we used radiographs rather than CT scans to determine axial alignment and malrotation. Precise statements on rotation alignments can only be made using CT scans38 or specific radiographs.41 We do not suspect this substantially affected our results. There were no intraoperative or postoperative patella complications and no insufficiencies of the knee extensor mechanism; therefore, we do not suspect any major femoral mal-rotation.
Surgical navigation increased preoperative and intraoperative times. The image-based group required an extra 2 hours for preoperative planning. The preoperative planning, an additional 38 minutes for the image-based group and an additional 25 minutes for the image-free group, would be too long for a standard procedure. However, it is important to know whether a surgical procedure is performed early in the learning curve or later. Regression analyses indicated the number of surgical interventions was an important factor for the total surgical time. After a learning curve of 15 interventions, the calculated estimated value for operative time with navigation support was 109 minutes for the image-based group and 101 minutes for the image-free group. The software application used previously with the image-based group was similar to that used with the image-free group. This may have decreased the influence of when the surgery was actually performed.
The image-free system does not need a confirmation process after the bony landmarks are defined. Therefore, additional total surgical time for the image-free group after the learning curve was 14 minutes, consistent with previous findings.4,26,36,40
Other reports indicate a more precise component position and good functional short-term outcome with surgical navigation for TKA.24,35,36,40 Investigators also suggest component malalignment greater than 3° in the frontal plane is associated with less satisfactory clinical and functional results.5,10,11,13,15,29,31,33,37,44
Our data suggest tibial component positioning was accurate regardless of surgical navigation. In contrast to a multicenter study,17 where better orientation was achieved in the frontal plane of the tibial component in the navigated group, we found no differences in the frontal plane. Matsuda et al23 reported the tibial axis can be determined precisely at the time of surgery according to bony landmarks by showing satisfactory postoperative mechanical tibial alignment. Intramedullary setting of the cutting block of the tibia is strongly influenced by bowing of the guide rod. Using an intramedullary system, even small deviations of the entry point of the guide rod lead to substantial bowing.28 We suspect accurate tibial component positioning associated with the conventional procedure occurs because of the good visible bony landmarks in the extramedullary alignment system. Based on our data, we do not think computer-assisted navigation in the tibia is essential.
The femoral component was positioned more accurately with navigation than with conventional implantation techniques. Regarding the accuracy of femoral component positioning, 12% more knees of patients in the image-based group and 17% in the image-free group were positioned within 3° malalignment. More than 88 patients in each group would have been necessary to reach sufficient power. The difference between better femoral alignment in the image-free group (97%) compared with the image-based group (92%) might be because the image-free navigation software allows observation of intraoperative ligament balancing and the mechanical axes.
In the conventional group, there was a stronger correlation between femoral component malalignment and total mechanical axis than between tibial component malalignment and mechanical axis. Therefore, malalignment in the total mechanical axis was predicted more by inaccurate femoral component positioning than by malpositioning of the tibial component. Although navigation did not improve tibial component positioning, correct tibial alignment is important. Some authors reported other errors with the conventional technique for component positioning, including anatomic variation of the femur, the entry point for the guide rod, and the diameter, length, and intramedullary bow of the guide rod.27,28,30 Nuno-Siebrecht et al28 reported minor deviations from the intramedullary instrumentation insertion point can result in malalignment of several degrees. Any inaccuracy of the cuts can be observed and detected before surgical intervention in the navigated groups, and can be corrected at any time during surgery.
Computer-assisted surgical navigation is based on an accurate anatomic definition of the bony landmarks and kinematically calculated and calibrated structures. For the image-based group, we identified bony landmarks preoperatively and confirmed their locations intraoperatively. We assumed image-free surgical navigation was less accurate than image-based systems. The points of reference were determined at the time of surgery, which may have substantial variability, and there is no confirmation with preoperative radiographic studies. Jerosch et al18 described substantial interobserver variability during definition of the bony landmarks. However, we were unable to find differences between the image-based and the image-free groups. Our results did not support that intraoperative calibration errors and lack of confirmation with preoperative CT scans cause substantial errors in component positioning. We think the additional radiation, cost of preoperative CT scans, and additional time for preoperative planning do not improve accuracy, and therefore, image-based navigation is no longer used at our institution.
Errors in total mechanical axis alignment with the conventional implantation technique occurred on the femoral side. Femoral component placement was more accurate with navigation, but there was no difference for the tibial component. There were no advantages in using tibial computer navigation, and eventually we may stop using tibial navigation. After a learning curve of 15 interventions and excluding the tibial navigation, the mean additional surgical time was 19 minutes for the image-based group and 6 minutes for the image-free group. This amount of time is considerable for a procedure performed so frequently.1,6,7,42 Femoral restricted navigated image-free systems and adapted instruments should be developed to reduce operative time. Longer followups and larger patient cohorts are needed to determine the influence of these slight variances on accuracy and survival rates of TKAs.
We thank M. Piot, PhD, for support in the statistical analysis, and B. Fuchs, MD, PhD, C. Gerber, MD, FRCS, and T. O'Brien, MD, for editorial support.
1. Anderson KC, Buehler KC, Markel DC. Computer assisted navigation in total knee arthroplasty: comparison with conventional methods. J Arthroplasty
2. Archibeck MJ, White RE Jr. What's new in adult reconstructive knee surgery. J Bone Joint Surg Am
3. Arima J, Whiteside LA, McCarthy DS, White SE. Femoral rotational alignment, based on the anteroposterior axis, in total knee arthroplasty in a valgus knee: a technical note. J Bone Joint Surg Am
4. Bathis H, Perlick L, Tingart M, Luring C, Zurakowski D, Grifka J. Alignment in total knee arthroplasty: a comparison of computer-assisted surgery with the conventional technique. J Bone Joint Surg Br. 2004;86:682-687
5. Chao EY, Neluheni EV, Hsu RW, Paley D. Biomechanics of mal-alignment. Orthop Clin North Am
6. Chin PL, Yang KY, Yeo SJ, Lo NN. Randomized control trial comparing radiographic total knee arthroplasty implant placement using computer navigation versus conventional technique. J Arthroplasty
7. Decking R, Markmann Y, Fuchs J, Puhl W, Scharf HP. Leg axis after computer-navigated total knee arthroplasty: a prospective ran-domized trial comparing computer-navigated and manual implantation. J Arthroplasty
8. Delp SL, Stulberg SD, Davies B, Picard F, Leitner F. Computer assisted knee replacement. Clin Orthop Relat Res
9. DiGioia AM3rd, Jaramaz B, Colgan BD. Computer assisted orthopaedic surgery: image guided and robotic assistive technologies. Clin Orthop Relat Res
10. Dorr LD, Boiardo RA. Technical considerations in total knee arthroplasty. Clin Orthop Relat Res
11. Ecker ML, Lotke PA, Windsor RE, Cella JP. Long-term results after total condylar knee arthroplasty: significance of radiolucent lines. Clin Orthop Relat Res
12. Ewald FC. The Knee Society total knee arthroplasty roentgeno-graphic evaluation and scoring system. Clin Orthop Relat Res
13. Garg A, Walker PS. Prediction of total knee motion using a three-dimensional computer-graphics model. J Biomech
14. Haaker RG, Stockheim M, Kamp M, Proff G, Breitenfelder J, Ottersbach A. Computer-assisted navigation increases precision of component placement in total knee arthroplasty. Clin Orthop Relat Res
15. Hungerford DS. Alignment in total knee replacement. Instr Course Lect
16. Insall JN, Dorr LD, Scott RD, Scott WN. Rationale of the Knee Society clinical rating system. Clin Orthop Relat Res
. 1989;248: 13-14.
17. Jenny JY, Boeri C. [Navigated implantation of total knee endopros-thesis: a comparative study with conventional instrumentation] Z Orthop Ihre Grenzgeb
18. Jerosch J, Peuker E, Philipps B, Filler T. Interindividual reproducibility in perioperative rotational alignment of femoral components in knee prosthetic surgery using the transepicondylar axis. Knee Surg Sports Traumatol Arthrosc
19. Kahler DM. Image guidance: fluoroscopic navigation. Clin Orthop Relat Res
20. Krackow KA, Bayers-Thering M, Phillips MJ, Bayers-Thering M, Mihalko WM. A new technique for determining proper mechanical axis alignment during total knee arthroplasty: progress toward computer-assisted TKA. Orthopedics
21. Levene H. Robust tests for equality of variance. In: Olkin I, ed. Contributions to Probability and Statistics
. Vol I. Palo Alto, CA: Standford University Press; 1960:278-292.
22. Martin A, von Strempel A. CT-based and CT-free navigation in total knee arthroplasty: a prospective comparative study with respects to clinical and radiological results. Z Orthop Ihre Grenzgeb
23. Matsuda S, Mizu-uchi H, Miura H, Nagamine R, Urabe K, IwamotoY. Tibial shaft axis does not always serve as a correct coronal landmark in total knee arthroplasty for varus knees. J Arthroplasty
24. Mielke RK, Clemens U, Jens JH, Kershally S. Navigation in knee endoprosthesis implantation: preliminary experiences and prospective comparative study with conventional implantation technique. Z Orthop Ihre Grenzgeb
25. Nabeyama R, Matsuda S, Miura H, Mawatari T, Kawano T, Iwamoto Y. The accuracy of image-guided knee replacement based on computed tomography. J Bone Joint Surg Br. 2004;86:366-371
26. Nizard RS, Porcher R,Ravaud P. Vangaver E, Hannouche D, Bizot P, Sedel L. Use of the Cusum technique for evaluation of a CT-based navigation system for total knee replacement. Clin Orthop Relat Res
27. Novotny J, Gonzalez MH, Amirouche FM, Li YC. Geometric analysis of potential error in using femoral intramedullary guides in total knee arthroplasty. J Arthroplasty
28. Nuno-Siebrecht N, Tanzer M, Bobyn JD. Potential errors in axial alignment using intramedullary instrumentation for total knee arthroplasty. J Arthroplasty
29. Oswald MH, Jakob RP, Schneider E, Hoogewoud HM. Radiological analysis of normal axial alignment of femur and tibia in view of total knee arthroplasty. J Arthroplasty
30. Paley D, Tetsworth K. Mechanical axis deviation of the lower limbs: preoperative planning of multiapical frontal plane angular and bowing deformities of the femur and tibia. Clin Orthop Relat Res
31. Piazza SJ, Delp SL, Stulberg SD, Stern SH. Posterior tilting of the tibial component decreases femoral rollback in posterior-substituting knee replacement: a computer simulation study. J Orthop Res
32. Rand JA, Trousdale RT, Ilstrup DM, Harmsen WS. Factors affecting the durability of primary total knee prostheses. J Bone Joint Surg Am
33. Ritter MA, Faris PM, Keating EM, Meding JB. Postoperative alignment of total knee replacement: its effect on survival. Clin Orthop Relat Res
34. Robertsson O, Knutson K, Lewold S, Lidgren L. The Swedish Knee Arthroplasty Register 1975-1997: an update with special emphasis on 41,223 knees operated on in 1988-1997. Acta Orthop Scand
35. Saragaglia D, Picard F, Chaussard C, Montbarbon E, Leitner F, Cinquin P. Computer-assisted knee arthroplasty: comparison with a conventional procedure: results of 50 cases in a prospective randomized study. Rev Chir Orthop Reparatrice Appar Mot
. 2001;87: 18-28.
36. Sparmann M, Wolke B, Czupalla H, Banzer D, Zink A. Positioning of total knee arthroplasty with and without navigation support: a prospective, randomised study. J Bone Joint Surg Br. 2003;85:830-835
37. Stern SH, Insall JN. Posterior stabilized prosthesis: results after follow-up of nine to twelve years. J Bone Joint Surg Am
. 1992;74: 980-986.
38. Stockl B, Nogler M, Rosiek R, Fischer M, Krismer M, Kessler O. Navigation improves accuracy of rotational alignment in total knee arthroplasty. Clin Orthop Relat Res
39. Stulberg SD. How accurate is current TKR instrumentation? Clin Orthop Relat Res
40. Stulberg SD, Loan P, Sarin V. Computer-assisted navigation in total knee replacement: results of an initial experience in thirty-five patients. J Bone Joint Surg Am
41. Takai S, Yoshino N, Isshiki T, Hirasawa Y. Kneeling view: a new roentgenographic technique to assess rotational deformity and alignment of the distal femur. J Arthroplasty
42. Victor J, Hoste D. Image-based computer-assisted total knee arthroplasty leads to lower variability in coronal alignment. Clin Orthop Relat Res
43. Vince KG. Why knees fail. J Arthroplasty
44. Wasielewski RC, Galante JO, Leighty RM, Natarajan RN, Rosenberg AG. Wear patterns on retrieved polyethylene tibial inserts and their relationship to technical considerations during total knee arthroplasty. Clin Orthop Relat Res
45. Whiteside LA, Arima J. The anteroposterior axis for femoral rotational alignment in valgus total knee arthroplasty. Clin Orthop Relat Res
46. Windsor RE, Scuderi GR, Moran MC, Insall JN. Mechanisms of failure of the femoral and tibial components in total knee arthroplasty. Clin Orthop Relat Res