Patients selection
After approval by the institutional ethics committee, a retrospective evaluation was performed prospectively on the lateral radiographs of 369 consecutive knees of patients who had undergone a primary TKA (NexGen LPS-flex, Zimmer, Warsaw, Indiana) in our hospital between January 2018 and January 2019. All surgeries were performed by a single senior joint surgeon under subarachnoid anesthesia. And the surgical approach, patellar eversion, osteotomy and prosthetic placement were carried out in the same way. We only included patients who underwent TKA without patellar resurfacing for primary osteoarthritis. Importantly, the lateral radiographs of the involved knee were taken twice in our hospital for each patient, and the typical examination times were within one month before TKA and two weeks after TKA. In view of the fact that the following factors may have an impact on the measurement of patellar height, the exclusion criteria were determined: (1) patients who presented preoperative or postoperative radiographs with the knee in less than 30° of flexion; (2) patients who had undergone a unicompartmental knee arthroplasty (UKA), high tibial osteotomy (HTO) or a revision procedure; (3) patients with an excessively rotated knee radiographs due to poor technique, impairing an accurate identification for bone landmarks during measurement of patellar height.
Of a total of 369 knees, one hundred and seventeen (31.7%) were excluded because the preoperative lateral radiographs were taken in other hospitals; fifty-eight (15.7%) were excluded because the radiographs showing the knee in less than 30° of flexion; nine (2.4%) were excluded due to previous UKA or HTO; ten (2.7%) were excluded because the radiographic examination time did not accord with schedule. Eventually, 175 knees (corresponding to 350 lateral radiographs) of 167 patients were included in this study. Of these, eight patients had undergone bilateral TKA at different times.
Study protocol
Using the ST-PACS CD-Medical software (version 1.5.3, Crealife, Beijing, China), the preoperative and postoperative patellar height were respectively measured on the lateral knee radiographs for each case. The IS, mIS, CD, BP indices, and the new method were applied to assessing patellar height. The measurement processes were performed independently by two trained observers in a double-blinded manner. All of the radiographs were measured by observer 1. To assess the reliability and variation in measuring patellar height, intra-observer and inter-observer studies were carried out. Two hundred radiographs of 100 knees were measured by observer 1 with an interval of 3 months, which were randomly selected from the total of 175 knees. Moreover, two hundred and twenty radiographs of 110 knees were independently measured by observer 2, which were also randomly selected from all cases.
Description of the methods for assessing patellar height
According to the description of the original publication [13-16], four classical indices for measuring patellar height were performed (Fig. 1A and C). Each index was drived from two measurements (a/b), “a” was the distance from the lower end of the patella to the tibial landmark and “b” was the length of the patella. For IS index, “a” was the length of the patellar tendon, which was measured from the lower pole of the patella to the insertion site of the patellar tendon on tibia (a notch superior tibial tubercle), and “b” was measured between the upper and lower pole of patella. For mIS, CD and BP indices, the length of patellar articular surface was shared as the measurement “b”. The “a” was the distance between the inferior pole of the patellar articular surface to the tibial landmarks of three indices, including the insertion site of the patellar tendon on tibia, the antero-superior margin of the tibia, and the perpendicular intersection of the inferior pole of the patellar articular surface with the tibial plateau line, respectively. Osteophytes were excluded during the measurements. After TKA, the tibial plateau was defined as the tangent line of the distal femoral component parallel to tibial osteotomy plane, as conducted by Rogers [6] and Cabral et al [3] (Fig. 1B and D).
Description of the new method
In current study, a new method was proposed to evaluate the patellar height and quantitatively measure the joint line position on the lateral radiograph of the knee before and after TKA. Instead of taking marginal landmarks as the tibial references in classical methods, the tibial shaft axis was used to construct the new measurement method, which comprises the axis-patella (AP), joint axis-patella (jAP) indices and joint line height (JLH).
Firstly, the AP index is a ratio between two measurements (a1/b) (Fig. 2). Measurement “b” is the length of the patellar articular surface. Measurement “a1” is the distance from the lower pole of the patellar articular surface to the intersection point (T1) between the tibial shaft axis and its perpendicular line passing through the tip of the fibular head. The tibial shaft axis is the mid-diaphyseal line, as described by previous studies [22, 23]. Secondly, the jAP index is another relative index with a2/b (Fig. 2A). Measurement “a2” is the distance from the lower pole of the patellar articular surface to the intersection point (T2) between the tibial shaft axis and the joint line. The joint line was defined as the tibial plateau line preoperatively and the tangent line of the distal femoral component parallel to tibial osteotomy plane postoperatively (Fig. 2B). Thirdly, the JLH is determined by measuring the distance between the T1 and T2 on the tibial shaft axis (Fig. 2).
Statistical analysis
All data were collected and collated through Excel (version 2016, Microsoft, Redmond, Washington), and data analysis is performed through SPSS (version 25.0, SPSS Inc., Chicago, IL, USA). The analysis results were expressed as mean ± standard deviation. A descriptive analysis and a paired sample T test were performed to depict the distribution and difference of measurements for both pre- and post-operative knee radiographs. To verify the validity of the new method, we conducted a Pearson’s correlation analysis on the values of various measurement methods included in this study. In addition, the intraclass correlation coefficient (ICC) and Bland-Altman plots, constructed by the Medcalc software (version 11.4, MedCalc software Ltd., Acacialaan 22, 8400 Ostend, Belgium), were used to analyze the intra- and inter-observer reliability and variability [24]. P < 0.05 was regarded as statistically significant.
After approval by the institutional ethics committee, a retrospective evaluation was performed prospectively on the lateral radiographs of 369 consecutive knees of patients who had undergone a primary TKA (NexGen LPS-flex, Zimmer, Warsaw, Indiana) in our hospital between January 2018 and January 2019. All surgeries were performed by a single senior joint surgeon under subarachnoid anesthesia. And the surgical approach, patellar eversion, osteotomy and prosthetic placement were carried out in the same way. We only included patients who underwent TKA without patellar resurfacing for primary osteoarthritis. Importantly, the lateral radiographs of the involved knee were taken twice in our hospital for each patient, and the typical examination times were within one month before TKA and two weeks after TKA. In view of the fact that the following factors may have an impact on the measurement of patellar height, the exclusion criteria were determined: (1) patients who presented preoperative or postoperative radiographs with the knee in less than 30° of flexion; (2) patients who had undergone a unicompartmental knee arthroplasty (UKA), high tibial osteotomy (HTO) or a revision procedure; (3) patients with an excessively rotated knee radiographs due to poor technique, impairing an accurate identification for bone landmarks during measurement of patellar height.
Of a total of 369 knees, one hundred and seventeen (31.7%) were excluded because the preoperative lateral radiographs were taken in other hospitals; fifty-eight (15.7%) were excluded because the radiographs showing the knee in less than 30° of flexion; nine (2.4%) were excluded due to previous UKA or HTO; ten (2.7%) were excluded because the radiographic examination time did not accord with schedule. Eventually, 175 knees (corresponding to 350 lateral radiographs) of 167 patients were included in this study. Of these, eight patients had undergone bilateral TKA at different times.
Study protocol
Using the ST-PACS CD-Medical software (version 1.5.3, Crealife, Beijing, China), the preoperative and postoperative patellar height were respectively measured on the lateral knee radiographs for each case. The IS, mIS, CD, BP indices, and the new method were applied to assessing patellar height. The measurement processes were performed independently by two trained observers in a double-blinded manner. All of the radiographs were measured by observer 1. To assess the reliability and variation in measuring patellar height, intra-observer and inter-observer studies were carried out. Two hundred radiographs of 100 knees were measured by observer 1 with an interval of 3 months, which were randomly selected from the total of 175 knees. Moreover, two hundred and twenty radiographs of 110 knees were independently measured by observer 2, which were also randomly selected from all cases.
Description of the methods for assessing patellar height
According to the description of the original publication [13-16], four classical indices for measuring patellar height were performed (Fig. 1A and C). Each index was drived from two measurements (a/b), “a” was the distance from the lower end of the patella to the tibial landmark and “b” was the length of the patella. For IS index, “a” was the length of the patellar tendon, which was measured from the lower pole of the patella to the insertion site of the patellar tendon on tibia (a notch superior tibial tubercle), and “b” was measured between the upper and lower pole of patella. For mIS, CD and BP indices, the length of patellar articular surface was shared as the measurement “b”. The “a” was the distance between the inferior pole of the patellar articular surface to the tibial landmarks of three indices, including the insertion site of the patellar tendon on tibia, the antero-superior margin of the tibia, and the perpendicular intersection of the inferior pole of the patellar articular surface with the tibial plateau line, respectively. Osteophytes were excluded during the measurements. After TKA, the tibial plateau was defined as the tangent line of the distal femoral component parallel to tibial osteotomy plane, as conducted by Rogers [6] and Cabral et al [3] (Fig. 1B and D).
Description of the new method
In current study, a new method was proposed to evaluate the patellar height and quantitatively measure the joint line position on the lateral radiograph of the knee before and after TKA. Instead of taking marginal landmarks as the tibial references in classical methods, the tibial shaft axis was used to construct the new measurement method, which comprises the axis-patella (AP), joint axis-patella (jAP) indices and joint line height (JLH).
Firstly, the AP index is a ratio between two measurements (a1/b) (Fig. 2). Measurement “b” is the length of the patellar articular surface. Measurement “a1” is the distance from the lower pole of the patellar articular surface to the intersection point (T1) between the tibial shaft axis and its perpendicular line passing through the tip of the fibular head. The tibial shaft axis is the mid-diaphyseal line, as described by previous studies [22, 23]. Secondly, the jAP index is another relative index with a2/b (Fig. 2A). Measurement “a2” is the distance from the lower pole of the patellar articular surface to the intersection point (T2) between the tibial shaft axis and the joint line. The joint line was defined as the tibial plateau line preoperatively and the tangent line of the distal femoral component parallel to tibial osteotomy plane postoperatively (Fig. 2B). Thirdly, the JLH is determined by measuring the distance between the T1 and T2 on the tibial shaft axis (Fig. 2).
Statistical analysis
All data were collected and collated through Excel (version 2016, Microsoft, Redmond, Washington), and data analysis is performed through SPSS (version 25.0, SPSS Inc., Chicago, IL, USA). The analysis results were expressed as mean ± standard deviation. A descriptive analysis and a paired sample T test were performed to depict the distribution and difference of measurements for both pre- and post-operative knee radiographs. To verify the validity of the new method, we conducted a Pearson’s correlation analysis on the values of various measurement methods included in this study. In addition, the intraclass correlation coefficient (ICC) and Bland-Altman plots, constructed by the Medcalc software (version 11.4, MedCalc software Ltd., Acacialaan 22, 8400 Ostend, Belgium), were used to analyze the intra- and inter-observer reliability and variability [24]. P < 0.05 was regarded as statistically significant.