Secondary Logo

Radiographic Assessment in Total Knee Arthroplasty

Bach, Christian, Michael; Steingruber, Iris, Eva; Peer, Siegfried; Nogler, Michael; Wimmer, Cornelius; Ogon, Michael

Clinical Orthopaedics and Related Research: April 2001 - Volume 385 - Issue - p 144-150
SECTION II ORIGINAL ARTICLES: Knee
Free
SDC

Sixty-five total knee arthroplasties were evaluated by the Knee Society Radiological Evaluation System which was developed to encourage uniform reporting of the results of total knee arthroplasty. All patients were examined by three independent experienced radiologists 8.9 years after surgery (range, 3–16 years) to analyze the interobserver variability. For measurement of angles, high interobserver correlation was calculated for the prosthetic component angles and the femorotibial shaft angle. The comparison of the means indicated no significant differences except for the femorotibial shaft angle. For measurement of radiolucent lines, interobserver correlation was low for all components. The differences of the means were significantly different for all components. The results of interobserver variability of the patellar evaluation revealed high interobserver correlation for the patellar angle and for patellar subluxation and dislocation evaluation. For assessment of patellar mediolateral and superoinferior displacement, a low interobserver correlation was found. For radiographic assessment of total knee arthroplasty, the measurement of angles, including alpha, beta, femorotibial shaft angle, sagittal femoral and tibial component angle, patellar angle, and patellar subluxation and dislocation evaluation are recommended. The method of assessing radiolucent lines should be reconsidered.

From the *Department of Orthopedic Surgery, and the **Department of Radiology, University of Innsbruck, Austria.

Reprint requests to Christian Michael Bach, MD, Department of Orthopedic Surgery, University of Innsbruck, Anichstrasse 35, A-6020 Innsbruck, Austria.

Received: March 28, 2000.

Revised: June 19, 2000; September 8, 2000.

Accepted: September 22, 2000.

Knee replacement has become a common procedure. Several studies have been conducted on the outcome of total knee arthroplasty. 2,3,7,10,14,15,18 Radiographic assessment is important in postoperative evaluation of total knee arthroplasty. In some studies, surgeons or orthopaedic units are providing their own assessment forms without defining their criteria. Most studies include measurements of knee alignment, positioning of components, and assessment of radiolucent lines. The lack of uniformity among the systems and the use of different terminology make it difficult to compare radiographic outcomes after total knee arthroplasty. In 1989 the Knee Society developed, by consensus, a new rating system with separate scales for clinical 13 and radiologic 6 evaluation to encourage uniform reporting of results of total knee arthroplasty. Since its introduction, the radiologic evaluation system of the Knee Society has been used in studies by Duffy et al 4 and Gill et al, 10 however, the reliability of this score has never been evaluated properly.

The current study analyzed interobserver variability of the radiographic variables used by the Knee Society for radiographic assessment of total knee arthroplasty.

Back to Top | Article Outline

MATERIALS AND METHODS

The Knee Society Radiological Evaluation System was used to evaluate a series of 65 total knee arthroplasties. The population consisted of 45 patients, 26 women and 19 men, with a mean age of 72 years (range, 57–89 years) who had total knee replacement 8.9 years (range, 3–16 years) previously.

All patients were examined by three independent experienced radiologists. All investigators were well acquainted with the assessment of total knee arthroplasty radiographs. The investigators also were familiar with the scoring system. Each observer completed the score independently for each patient. Submission of the results to a personal computer and statistical calculations started after data acquisition were complete. There was no discussion of individual results before patient examinations were completed.

Standard radiographs of the knee included one anteroposterior (AP) view and one lateral view of the knee and one axial view of the patella. For the AP view, the patient was positioned supine, and the xray beam was directed perpendicular to the joint line, which was determined by palpation. For the lateral view, the xray beam was directed laterally, again perpendicular to the joint line. The patella was evaluated using 30° axial views.

As it was the aim to test a clinical setting, all lines were drawn on printed copies of the radiographs. Three copies were made from all radiographs to provide each observer with one complete set. No templating guides were used.

To account for interobserver variability of angle and radiolucent line measurements, the pairwise Pearson’s correlation coefficient was calculated between Observer 1 and Observer 2, Observer 1 and Observer 3, and Observer 2 and Observer 3. For all angle and radiolucent measurements, analysis of variance (ANOVA) for repeated measurements was used to evaluate mean differences of the scores among the three observers. Cohen’s kappa measure was used to determine the interobserver reliability of position measurements of the patella referred to the femoral condyles (patellar subluxation and dislocation) and the measurements of the position of the prosthesis in the patella (patellar mediolateral and superoinferior displacement).

Back to Top | Article Outline

Radiologic Variables

Angles

The amount of tibial and femoral component deformity in the frontal plane was measured by the alpha (femoral) and beta (tibial) joint line angles. For assessment of the alpha angle, one line was drawn parallel to the femoral condyles and a second line was drawn in the femoral shaft axis. For calculation of the beta angle, one line was drawn parallel to the plateau of the metal tibial component, and the second line was drawn in the tibial shaft axis. Femoral component flexion and extension angles were measured between a line perpendicular to the distal metal-cement interface of the femoral component and a line parallel to the femoral shaft axis. The tibial component sagittal angle was measured between a line parallel to the metal tibial plateau and the tibial shaft axis. The femorotibial angle was measured by the intersection of femoral and tibial shaft axis (Fig 1). The patellar angle was measured between a line drawn parallel to the surface of the patellar prosthesis and a line parallel to the femoral condyles (Fig 2).

Fig 1.

Fig 1.

Fig 2.

Fig 2.

Back to Top | Article Outline

Radiolucent Lines

Measurement of radiolucent lines at the cement-bone interface was done for all three prosthetic components. Figure 3 shows a radiolucent line of the lateral part of the plateau of the tibial component prosthesis. Radiolucent lines were measured in millimeters.

Fig 3.

Fig 3.

The total widths of the radiolucent lines at a component of the prosthesis were determined by adding up the widths of the radiolucent lines as measured at the specific component sites defined by Ewald et al. 7 The total widths were added for each zone and for each component. Therefore, the total width of the radiolucent lines is a calculated value and not one measurement. The total widths were recorded independently for the tibial component in the AP and sagittal views, for the femoral component in the sagittal view, and for the patellar prosthesis in the axial view.

Back to Top | Article Outline

Patella

For assessment of patellar position, evaluation of patellar subluxation and dislocation was used. For assessment of the patellar prosthesis, the patellar angle, and mediolateral and superoinferior displacement were used. Assessment of patellar displacement included measurement of displacement of the center of the patellar prosthesis relative to the patella. Mediolateral displacement of the patellar prosthesis was assessed on axial views. If the center of the prosthesis was medial or lateral to the center of the patella, medial or lateral displacement was recorded. The relative position of the patellar prosthesis in the sagittal plane was evaluated by the vertical distance between the center of the patellar prosthesis and the center of the patellar articular surface. If the center of the patellar prosthesis was superior or inferior to the center of the patellar articular surface, superior or inferior displacement was seen (Fig 2).

Back to Top | Article Outline

RESULTS

For measurement of component angles (alpha angle, sagittal femoral angle, beta angle, sagittal tibial angle, patellar angle), the interobserver correlation was good. Comparison of the mean differences revealed no significant differences between the measures of the three observers (p > 0.05 each). For the femorotibial shaft angle, the interobserver correlation was good but the comparison of the means indicated significant differences between the single measures (p < 0.02). Results are summarized in Table 1.

TABLE 1

TABLE 1

Correlation coefficients for the radiolucent lines were generally low except for measurements at the tibial side in the AP view. Calculation of the mean differences indicated significant differences for all components (p < 0.05 each).

The results of interobserver variability of the patellar position determination revealed good correlation for patellar subluxation and dislocation evaluation (Kappa > 0.65). For assessment of patellar mediolateral and superoinferior displacement, however, a low correlation was found (Kappa < 0.35).

Back to Top | Article Outline

DISCUSSION

The results of radiographic evaluation of total knee arthroplasty of this study are similar to the results described in the literature. For component angles, the results of this study were compared with the data of Mont et al 15 (Table 2). For assessment of radiolucent lines, Ewald 6 suggested that a total sum in millimeters for each component of 4 or less and nonprogressive is probably not significant; 5 to 9 should be closely observed for progression; and 10 or greater signifies possible or impending failure. In the current study, the mean total sum for each component was below the 4 mm limit.

TABLE 2

TABLE 2

For assessment of the patellar prosthesis, results were similar to the results of Gomes et al. 11 In this series 40 patients were assessed. For the patellar angle, a lateral tilt of 1.8° to 8.4° was found. In all patients the patellar prosthesis was displaced superiorly, and in seven cases there was medial displacement of the patellar prosthesis.

Schneider et al 17 reported that a shift in the position of a prosthetic component is a definite sign of loosening. Ewald 6 stated that a migrating or shifting prosthesis with or without disappearance of radiolucent lines should be considered possible or impending failure. For measurement of component angles, low interobserver variability and low mean score differences were found. Correlation was somewhat higher for the femoral than for the tibial component. The method is shown to be reliable in determining component position and this may be correlated with component loosening.

For the femorotibial shaft angle a good interobserver correlation was found, but the differences of the means between the observers indicated that the method of assessment might be unreliable (p < 0.002). The absolute mean values of the angle were 86°, 88°, and 87° for Observers I, II, and III. Even though the means differed significantly, this is not clinically relevant and the angle can be recommended for radiographic assessment.

Loosening of a prosthetic component is the most frequent causes of failure of a joint. As the importance of a plain radiograph for assessment of total knee arthroplasty loosening is evident, the importance of radiolucent lines has been discussed. 1 Ecker et al 5 reported that radiolucent lines could not be visualized if the central beam was angulated either cephalad or caudad more than 6°. If the central beam was more than 2.5 cm from the tibial plateau, the lines were not seen.

Reckling et al 16 reported that a radiolucent zone at the bone-cement interface was seen under the tibial component in 42 of 59 knees. Flynn 8 reported that, in a series of 200 total knee arthroplasties, the presence of a radiolucent zone occurred in 80% on the tibial side and 0% on the femoral side after a minimum followup of 2 years. The presence of radiolucent lines does not necessarily indicate loosening or surgical failure. 19 Insall et al 12 reported that the design of a prosthetic device affected the incidence of a radiolucent zone and proposed that the zone was related to movement between the implant and bone. Other authors suggested that generation of heat by the bone cement 19 or the biomechanical effect of the cement on the bone 9 are important in development of a radiolucent zone. However, there is no general agreement regarding the mechanism of development of the radiolucent lines. 19

Nevertheless, measurement of radiolucent zones in total knee arthroplasty has been described. 1,16,17 The current results indicate a low interobserver correlation. The means of the measures were significantly different for assessment of radiolucent lines.

Although major radiolucencies are probably the main indicator for implant loosening, the clinical significance of small radiolucent lines for diagnosis of total knee arthroplasty loosening is questionable. Because in the current study measurement of radiolucencies was proven to be unreliable, this method of assessment appears debatable. A more simple system, although less sensitive, probably would be more appropriate. The system of assessment of radiolucent lines is very detailed and numerous, different zones of each component are assessed in two planes. It was suggested that the more complex a system is constructed, the more it is likely that errors occur. Ecker et al 5 stated that the presence of thin radiolucent lines under the tibial plateau did not appear to have clinical significance. Thick lines or lines about all portions of the tibial component indicated significant loosening which usually was associated with symptoms and possible failure. A simple system may only differ between, for example, no or minor radiolucent lines and extensive radiolucent lines.

Some authors suggest that evidence of new radiolucent zone formation after more than 1 postoperative year, or enlargement, even partially, of an existing radiolucent zone could be regarded as loosening. 1,16,17

For assessment of the patella, measurement of the patellar angle and patellar position relative to the femoral condyles was more reliable than assessment of positioning of the patellar prosthesis relative to the patella. This possibly was because of the difficulty of identifying the circumference of the patellar prosthesis.

No rating system is ideal, but if centers use the same reporting system, then relative comparisons will be possible. 6 According to the results of the current study, for radiographic assessment of total knee arthroplasty, the system introduced by the Knee Society in 1989 can be recommended for assessment of angles, including alpha, beta, femorotibial shaft angle, sagittal femoral and tibial component angles, patellar angle, and patellar subluxation and dislocation evaluation. The proposed assessment of radiolucent lines, however, is unreliable and should be reconsidered.

Back to Top | Article Outline

Acknowledgment

The authors thank Dr. Benedikt Czermak, Department of Radiology, University of Innsbruck, Austria, for his assistance.

Back to Top | Article Outline

References

1. Ahlberg A, Linden B: The radiolucent zone in arthroplasty of the knee. Acta Orthop Scand 48:687–690, 1977.
2. Barck AL: Patient’s memory or repeated pain and function scores as index for major clinical change caused by knee replacement? Arch Orthop Trauma Surg 116:484–485, 1997.
3. Ducheyne P, Kagan A, Lacey JA: Failure of total knee arthroplasty due to loosening and deformation of tibial component. J Bone Joint Surg 60A:384–391, 1978.
4. Duffy GP, Trousdale RT, Stuart MJ: Total knee arthroplasty in patients 55 years old or younger. Clin Orthop 356:22–27, 1998.
5. Ecker ML, Lotke PA, Windsor RE, et al: Long-term results after total condylar knee arthroplasty: Significance of radiolucent lines. Clin Orthop 216:151–158, 1987.
6. Ewald FC: The Knee Society total knee arthroplasty roentgenographic evaluation and scoring system. Clin Orthop 248:9–12, 1989.
7. Ewald FC, Jacobs MA, Miegel RE, et al: Kinematic total knee replacement. J Bone Joint Surg 66A: 1032–1040, 1984.
8. Flynn LM: Experiences with UCI total knee. Clin Orthop 135:188–191, 1978.
9. Freeman MA, Bradley GW, Revell PA: Observations upon the interface between bone and polymethylmethacrylate cement. J Bone Joint Surg 64B:489–493, 1982.
10. Gill GS, Joshi AB, Mills DM: Total condylar knee arthroplasty. Clin Orthop 367:210–215, 1999.
11. Gomes LSM, Bechtold JE, Gustilo RB: Patellar prosthesis positioning in total knee arthroplasty. Clin Orthop 236:72–81, 1988.
12. Insall J, Scott WN, Ranawat CS: The total condylar knee prosthesis. J Bone Joint Surg 61A:173–180, 1979.
13. Insall JN, Dorr LD, Scott R, et al: Rationale of the Knee Society clinical rating system. Clin Orthop 248:13–14, 1989.
14. Kettelkamp DB, Thompson C: Development of a knee scoring scale. Clin Orthop 107:93–99, 1975.
15. Mont MM, Fairbank AC, Yammamoto V, et al: Radiographic characterization of aseptically loosened cementless total knee replacement. Clin Orthop 321:73–78, 1995.
16. Reckling FW, Asher MA, Dilion W: Longitudinal study of the radiolucent line at the bone cement interface following total joint-replacement prostheses. J Bone Joint Surg 59A:355–358, 1977.
17. Schneider R, Hood RW, Ranawat CS: Radiologic evaluation of knee arthroplasty. Orthop Clin North Am 13:225–244, 1982.
18. Schurman DJ, Parker JN, Orstein D: Total condylar knee replacement. J Bone Joint Surg 67A:1006–1010, 1985.
19. Torisu T, Morita H: Roentgenographic evaluation of geometric total knee arthroplasty with a six-year average follow-up period. Clin Orthop 202:125–133, 1986.
© 2001 Lippincott Williams & Wilkins, Inc.