**3. Results**

#### *3.1. Demographic Data*

All demographic, clinical, and operative data, including sex, age, body mass index (BMI), bone mineral density (BMD), symptom duration, main diagnosis of LSDs, spinal stenosis grade on MRI, fusion levels, and Kellgren-Lawrence grade were summarized in Table 1. In preoperative TKA group, mean interval between TKA and spinal fusion was 1.2 years. The mean age in the non-TKA and preoperative TKA groups was 68.4 years and 72.1 years, respectively (*p* = 0.110). Mean BMI in the non-TKA and preoperative TKA groups was 26 and 25.5, respectively (*p* = 0.602). Mean BMD in non-TKA and preoperative TKA groups was −0.7 and −1.1 at the spine as well as −1.1 and −1.4 at the femur. There were no significant differences in BMD of the spine and femur between the two groups (*p* = 0.696, *p* = 0.284). In total, 58% and 59.2 of patients had a symptom duration of more than 5 years in the non-TKA and preoperative TKA groups, respectively. A severe grade of spinal stenosis was presented in 52% and 54.5% of the non-TKA and preoperative TKA groups, respectively. The fusion levels in non-TKA and preoperative TKA group were not significant different (*p* = 0.409). Spondylolisthesis was presented in 26% of the non-TKA group and 45% of the preoperative TKA group for the main diagnosis of LSDs. All KOA were bilateral, which showed more than Kellgren-Lawrence grade III. There were no significant differences in demographic and operative data between the two groups (Table 1).

**Table 1.** Demographic and operative data for spinal fusion only and preoperative TKA with spinal fusion groups.



**Table 1.** *Cont.*

*p* < 0.05 is significant. \* All values are expressed as mean ± standard deviation. *p* values were calculated by independent t-test for parametric data and Mann Whitney U test for non-parametric data. † *p*-values were calculated by chi-square test for parametric data and Fisher's exact test for non-parametric data. *n* = number; TKA = Total knee arthroplasty; M = Male; F = Female; BMI = Body mass index; BMD = Bone mineral density; LSDs = Lumbar spinal diseases; MRI = Magnetic resonance imaging.

#### *3.2. Radiological Outcomes*

Regarding the regional and global parameters of radiological outcomes, preoperative LL was significantly lower in the preoperative TKA group (32◦) than the non-TKA group (23◦) (*p* = 0.045). The 2-year follow-up LL was lower in the non-TKA group (35.3◦) than the preoperative TKA group (27.1◦) with statistical significance (*p* = 0.041). Preoperative SVA was 51.6 mm in the non-TKA group and 72.5 mm in the preoperative TKA group, with no significance (*p* = 0.066). Immediate postoperative (40 mm, 47.2 mm) and 2-year follow-up (41.2 mm, 47 mm) SVA in non-TKA and preoperative TKA groups was distributed within an age-adjusted target (about 54.5 mm from 65 to 74 years) with no significance (*p* = 0.455, 0.561) [3]. All TPAs were greater than 20◦ and those in the preoperative TKA group were higher than non-TKA group, but statistical difference was not significant. Regional and global parameters demonstrated worse outcomes in the preoperative TKA group than the non-TKA group. Only the preoperative and 2-year follow-up LL showed statistically significant differences (Table 2).

Regarding the coronal parameters, Cobb's angle preoperatively, immediate postoperative, and at 2-year follow-up was within 10◦ in both groups (all *p* > 0.05). All coronal balance values preoperatively, immediate postoperatively and at 2-year follow-up evaluations were within 20 mm and showed statistical insignificance between the two groups (all *p* > 0.05). For sagittal spinopelvic parameters, preoperative PI was significantly higher in the preoperative TKA group (62.8◦) than the non-TKA group (53.5◦) (*p* = 0.041). However, after spinal fusion, there were no significance differences between immediate postoperative (*p* = 0.398) and 2-year follow-up (*p* = 0.729) PI. All values of PI/LL mismatch were more than 11◦. Preoperative PI/LL mismatch was significantly higher in the preoperative TKA group (39.8◦) than the non-TKA group (21.5◦) with statistical significance (*p* = 0.013). However, there were no significant difference observed in immediate postoperative (*p* = 0.286) and 2-year follow-up (*p* = 0.265) PI/LL mismatch. PT was greater at

more than 22◦ and was higher in the preoperative TKA group (30.7◦) than the non-TKA group (24◦). Only preoperative PT showed a statistically difference (*p* = 0.011). All sacral slopes were greater in the preoperative TKA group than in the non-TKA group but without statistical significance (all *p* > 0.05) (Table 3).

**Table 2.** Comparison of regional and global parameters between spinal fusion only and preoperative TKA with spinal fusion groups.


Data represent mean ± standard deviation values for each group. In the case of the sagittal vertical axis, the statistical analysis between groups was performed by converting negative numbers to positive numbers to analyze how the difference from the reference point. *p*-values were calculated by independent t-test for parametric data and Mann Whitney U test for non-parametric data. Significant differences were accepted for *p* < 0.05. *n* = number; TKA = Total knee arthroplasty.

#### *3.3. Clinical Outcomes*

ODI and VAS were used for assessing clinical outcomes preoperatively, immediate postoperatively, and at 6-month follow-up. The mean preoperative ODI was significantly worse in the preoperative TKA group (62.4) than the non-TKA group (50.4) (*p* = 0.001). However, after spinal fusion, the mean immediate postoperative ODI was 45.4 in the non-TKA group and 37.6 in the preoperative TKA group (*p* = 0.008). Mean 6-month follow-up ODI was 45.8 in the non-TKA group and 34. 1 in the preoperative TKA group (*p* < 0.001). Mean preoperative VAS of the back was 7.57 in the non-TKA group and 8.44 in the preoperative TKA group. Mean immediate postoperative VAS of the back was 4.00 in the non-TKA group and 4.44 in the preoperative TKA group. Mean 6-month followup VAS of the back was 3.19 in the non-TKA group and 3.33 in the preoperative TKA group. None of these back VAS values were significantly different between groups (all *p* > 0.05). Preoperative VAS of the leg was close to 7.2 in the non-TKA group and 7.3 in the preoperative TKA (*p* = 0.965). Mean immediate postoperative VAS of the leg was 6.1 in the non-TKA group and 3 in the preoperative TKA group (*p* < 0.001). Six-month follow-up VAS of the leg was 6 in the non-TKA group and 2.7 in the preoperative TKA group, a significant difference (*p* < 0.001) (Table 4).


**Table 3.** Comparison of coronal and sagittal spinopelvic parameters between spinal fusion only and preoperative TKA with spinal fusion groups.

Data represent mean ± standard deviation values for each group. In the case of coronal balance, the statistical analysis between groups was performed by converting negative numbers to positive numbers to analyze how the difference from the reference point. *p* values were calculated by independent t-test for parametric data and Mann Whitney U test for non-parametric data. Significant differences were accepted for *p* < 0.05. *n* = number; TKA = Total Knee Arthroplasty; PI/LL mismatch = Pelvic incidence minus lumbar lordosis.

**Table 4.** Comparison for clinical outcomes between spinal fusion only and preoperative TKA with spinal fusion.


Data represent mean ± standard deviation values for each group. *p*-values were calculated by independent t-test for parametric data and Mann–Whitney U test for non-parametric data. Significant differences were accepted for *p* < 0.05. *n* = number; TKA = Total knee arthroplasty; ODI = Oswestry Disability Index; VAS = Visual Analog Scale.

> The ODI differences between preoperative and immediate postoperative was 5.0 ± 4.7 in non-TKA and 24.9 ± 6.2 in preoperative TKA with statistical significance (*p* < 0.001). VAS

leg differences between preoperative and immediate postoperative was 1.0 ± 0.9 in non-TKA and 4.3 ± 1.9 in preoperative TKA with statistical significance (*p* < 0.001). However, ODI differences and VAS leg differences between immediate postoperative and 6-month follow-up showed not statistical insignificance (*p* = 0.780).
