The Relationship Between Postoperative Variation of Adjacent Segment Three‐Dimensional Kinematics and Postoperative PI‐LL in Patients With Degenerative Lumbar Scoliosis After Short Segment Fixation

ABSTRACT

Objective

The etiology of adjacent segment diseases and proximal junctional kyphosis has been related to biomechanical alterations after spinal operation. This study investigated the variation in pre‐ and postoperative range of motion in adjacent segments in patients with degenerative lumbar scoliosis (DLS) following posterior lumbar interbody fusion.

Patients and Methods

Eight patients with symptomatic DLS were analyzed using a biplane radiographic imaging system while adopting different postures. Synchronized biplane radiographs from L1 to S1 and the motions of each vertebra were acquired pre‐ and postoperatively. Six degrees of freedom (DOF) of kinematic data were compared between different postoperative pelvic incidence‐lumbar lordosis (PI‐LL) groups (group A: PI‐LL = −10°–~10°; group B: PI‐LL = 10°–~20°).

Results

After surgery, the axial rotational movement (primary rotation) during flexion–extension, bending, and torsion in the first adjacent segment at L3‐4 decreased significantly in three patients (8.14 ± 2.78 vs. 15.13 ± 6.71; 8.36 ± 5.59 vs. 9.08 ± 3.57; 5.07 ± 0.56 vs. 9.25 ± 5.06). At this level, the torsion around the crania‐caudal (CC) axis during bending (1.48 ± 1.01 vs. 7.05 ± 5.84, p < 0.05) and flexion around the mediolateral (ML) axis and bending rotation around the anterioposterior (AP) axis during torsion decreased postoperatively (6.37 ± 6.01 vs. 13.83 ± 8.12, 4.53 ± 1.97 vs. 13.06 ± 6.65; p < 0.05, p < 0.05). After surgery, in the L1‐2 segment, translation along the ML direction decreased during bending (3.69 ± 2.12 vs. 14.76 ± 7.99, p < 0.05). In the adjacent L5‐S1 segment, primary flexed rotation around the ML axis increased in group B postoperatively during flexion–extension, but decreased in group A (6.08 ± 1.17 vs. −13.41 ± 2.99, p < 0.05). Coupled flexed rotation around the ML axis decreased and showed the opposite trend during bending (−10.76 ± 5.51 vs. 18.12 ± 39.83, p < 0.05).

Conclusions

Postoperative coupled motion at the adjacent segment decreased, which indicates an improved balance of the spinal order compared to before the surgery. However, primary motion changed according to the location of the upper instrumented vertebrae. Our results indicate that postoperative PI‐LL values between 10° and 20° were associated with lower coupled motion and higher primary motion at L5‐S1.

In summary, we collected the postoperative variation of motion pattern at the adjacent segment in patients with DLS over short‐term follow‐up. As a result, the reduction of coupled motion in adjacent segments indicates an improved balance of the spine compared to before the surgery. Patients were divided into two groups (−10° < PI‐LL < 10°; PI‐LL:10°‐20°). Postoperative PI‐LL values between 10° and 20° were associated with lower coupled motion and higher primary motion at L5‐S1, indicating that postoperative PI‐LL values between 10° and 20° were associated with lower coupled motion and higher primary motion at L5‐S1.

1. Introduction

It was reported that the incidence of degenerative lumbar scoliosis (DLS) reached as high as 58.6% between 40 and 77 years old [1], which was defined as a developed spinal deformity with a coronal Cobb angle of > 10° [2]. These patients always need surgical treatment such as lumbar interbody fusion. However, after surgery, many patients with DLS experience symptom recurrence because of the development of adjacent segment disease (ASD) with an incidence rate of 27.4%–30.3% during more than 5‐year follow‐up [3, 4]. The success rate of revision surgery is lower than that of primary surgery [5]. In addition, other postoperative complications, such as proximal junctional kyphosis, are as high as 39% in patients with DLS [6]. The pathogenesis of postoperative ASD and PJK remains poorly understood. Besides, the etiology of ASD has been related to biomechanical alterations after spinal operation [7, 8]. Kim et al. further identified construct stiffness as a risk factor for PJK [9, 10]. In addition, during surgery in patients with DLS, the recovery of sagittal balance has been closely related to the clinical outcome, and the ideal postoperative pelvic incidence‐lumbar lordosis (PI‐LL) is regarded as an important factor [11].

In vitro testing in cadaveric specimens and finite element (FE) analyses have both been widely used to investigate adjacent segment motion and mechanical behavior [12] before and after various fusion surgical treatments [13, 14, 15]. In general, these previous studies predicted that spinal fusion causes an increase in the ranges of motion (ROMs) in adjacent segments. Previous in vivo studies have also investigated the changes in overall lumbar and adjacent segment ROMs after fusion [16], mainly during flexion–extension [17, 18, 19] and upright standing conditions [20] by a dual fluoroscopic imaging system (DFIS) [21], with an average accuracy of ±0.3 mm in translation and ±0.7° in orientation [22]. However, the mechanisms underlying the alteration of biomechanics in the adjacent segments by spinal fusion have not yet been thoroughly investigated.

Hayes et al. evaluated the motion of adjacent segments after fusion for idiopathic scoliosis using lateral flexion and extension radiographs [23]. Computational models of patients with DLS have been created to examine lumbar kinematics and the effects of surgical interventions. Previous studies using FE analysis showed that spinal fusion surgery increased left and right bending intersegmental rotation at the adjacent level [24, 25]. Despite being implicated in the development or acceleration of ASD, direct in vivo evidence to support alteration in the biomechanics and kinematics post‐lumbar fusion in patients with DLS is lacking [26, 27]. This reminded us of the necessity of in vivo postoperative data in patients with DLS to provide a reference for modeling and simulating in vitro kinematics after lumbar fusion.

This study aimed to: (i) characterize the in vivo intervertebral kinematics of the adjacent motion segment relative to the fused motion segment before and after fusion in patients with DLS. (ii) This study compared vertebral kinematics across different postoperative PI‐LL values, given that the optimal clinical value of this parameter in patients with DLS remains under investigation. (iii) We hypothesized that the coupled motion pattern of the adjacent segment would increase postoperatively to compensate for the loss of motion of the fused segment in patients with DLS.

2. Methods

2.1. Patients

In this study, we collected data of eight patients with DLS (males, two; females, six), aged 68.63 ± 5.21 years, ranging from 62 to 77 years old before surgery. All patients experienced low back pain due to DLS and were scheduled to undergo posterior lumbar interbody fusion (PLIF). Approval for the experimental design was obtained from the Institutional Review Board of our hospital before beginning the study. The inclusion criteria were as follows: (1) Age > 60 years; (2) Coronal Cobb angle > 10°; (3) Primary degenerative adult lumbar scoliosis; (4) With symptoms of severe spinal stenosis. The exclusion criteria were as follows: (1) History of adolescent scoliosis; (2) History of vertebral infection and vertebral malignancy; (3) Severe hip and knee joint pain; (4) History of spinal surgery. Informed consents were obtained from all participants prior to study commencement.

2.2. Surgical Management and Measurement of the Patients

Following diagnosis, all patients underwent PLIF performed by two full‐time fellowship‐trained spine surgeons at our hospital within 2 weeks. Three of the symptomatic patients underwent single‐level fusion, four of the symptomatic patients underwent double‐level fusion, and the final patient underwent four‐level fusion. Patients were divided into two groups (group A with three patients: PI‐LL = −10°–10°; group B, five patients; PI‐LL, 10°–20°) according to the postoperative PI‐LL measured on postoperative whole‐spine anteroposterior and lateral standing radiographs. The following parameters were measured on pre‐ and postoperative sagittal radiographs by two senior attending doctors independently (Figure 1): (1) Lumbar lordosis (LL) was defined as the angle between the upper endplate of the L1 and sacral plate; (2)Sacral slope (SS) was defined as the angle between the horizontal line and sacral plate; (3) Pelvic incidence (PI) was defined as the angle between the perpendicular from the midpoint of the upper endplate of S1 and a line connecting the center of the femoral head to the center of the upper endplate of S1; (4) Pelvic tilt (PT) was defined as the angle between the vertical and the line through the midpoint of the sacral plate to the femoral head axis. An interclass correlation coefficient (ICC) was calculated to determine intra‐observer reliability, performed in SPSS (version 23.0 for Windows, IBM Corp., Armonk, NY, USA).

FIGURE 1.

The measurements of the radiographic parameters, including lumbar lordosis (LL), sacral slope (SS), pelvic incidence (PI), and pelvic tilt (PT).

Computed tomography (CT) of the lumbar spine was collected by a Siemens CT scanner (Siemens, Erlangen, Germany). Each three‐dimensional (3D) bone model of the vertebrae was constructed from the reconstructed CT images using Mimics (21.0; Materialize, Leuven, Belgium), and the contours of the vertebrae were digitized (Figure 2A,B). Anatomic coordinate systems were created at the geometric center of the vertebrae (Figure 2C). The x‐axis was set in the frontal plane to represent the mediolateral (ML) direction and was then pointed to the left. The y‐axis was set in the sagittal plane and pointed in the posterior direction, representing the anteroposterior (AP) direction. The z‐axis was set perpendicular to the X–Y plane, representing the cephalad‐caudad (CC) direction, and pointed to the cranial direction. We selected the center of the vertebral body as the origin, defined the direction toward the spinous process as posterior, and the orientation toward the superior endplate as the craniocaudal direction. A line perpendicular to the plane formed by these two axes was established as the mediolateral axis. In our study, the anatomical planes (frontal, sagittal, and coronal) were defined following the standard definitions provided in Terminologia Anatomica (TA) and commonly used in biomechanical studies. The sagittal plane divides the body into left and right sections, the coronal plane divides it into anterior and posterior sections, and the frontal plane divides it into superior and inferior sections. Synchronized biplane radiographs of the lumbar spine (L1‐S1) were acquired using two fluoroscopes (BV Pulsera, Phillips, Bothell, WA, USA) placed perpendicular to each other. It was performed 1 month before and 1 year after spinal decompression and fusion surgery. Then, the 3D anatomical vertebral models were subsequently imported into a DFIS, Rhinoceros 3D (Robert McNeel & Associates, Seattle, WA), and FluoMotion (Innomotion Inc., Shanghai, China). The 3D models of the vertebrae were then matched to two orthogonal fluoroscopic images (Figure 3) [22]. During imaging, we were careful to decrease the radiation quantity exposure to patients, with about 6.5 mSv per person. The patients were asked to move their lumbar spines to different postures as follows: upright standing position, 45° flexion of the trunk, maximal extension, maximal left–right bending, and maximal left–right twisting. The relative motion, including rotation and translation, of the proximal vertebrae with respect to that of the distal vertebrae was subsequently calculated at four vertebral levels: L1‐2, L2‐3, L3‐4, and L5‐S1. During the bending of the trunk, the translation along the ML direction and the rotation around the AP axis were defined as the primary translation and rotation. The translation along the AP direction and the rotation around the ML axis during the flexion–extension were also defined as primary translation and rotation. The rotation around the CC axis during torsion was defined as primary rotation. The ROM data included the primary and coupled motions in all the six degrees of freedom (DOFs).

FIGURE 2.

Three‐dimensional anatomic vertebral model (b) was constructed from the computed tomography (a) with DLS patient who underwent PLIF. Anatomic coordinate system was created to measure kinematics of the vertebra (c).

FIGURE 3.

Reproduction of vertebra in the dual fluoroscopic system.

Patient‐reported outcomes included the visual analog scale (VAS) for back pain and leg pain. All patients completed the Modems version of the Oswestry Disability Index (ODI) 1 month before and 1 year after surgery.

3. Statistical Analysis

Two‐way repeated measures ANOVA was used to compare the ROMs of the vertebrae at the adjacent segments. Kinematics was the dependent variable, and the vertebral level and activity were set as the independent variables. Statistical significance was set at p < 0.05. Another multiway analysis of variance was used to compare the kinematics between the different groups. The participant group was the categorical factor, and the levels and activities were set as the independent variables. A Newman–Keuls post hoc test was performed when a statistically significant difference was detected. Statistical analyses were performed using SPSS version 23 (IBM Corp.) and Prism 9 software (version 5.01; GraphPad Software Inc., CA, USA).

4. Results

4.1. Patients Demographics

Overall, the scoliosis apical vertebra was located at L2 (two patients), L3 (three patients), and L4 (three patients). The pre‐ and postoperative sagittal spinopelvic parameters were shown in Table 1 and the ODI and VAS were improved after surgery (Table 1). The ICC of radiological parameters for the interobserver agreement was 0.985 (95% CI, 0.968–0.993; p < 0.001).

TABLE 1.

Patient demographics and radiographic parameters before and after fusion.

Subject ID	Age (years)	Gender	BMI (kg/m2)	Scoliotic apex vertebrae	Preoperative							Postoperative
					PI	PT	SS	LL	PI‐LL	ODI	VAS	PI	PT	SS	LL	PI‐LL	ODI	VAS
1	75	Female	29.59	L4	38.4	28.9	9.5	17.4	21	22	8	40.9	21.8	19.1	23.8	17.1	2	0
2	71	Female	25.35	L3	62.3	31.4	30.9	36.9	25.4	32	2	62.8	32.1	30.7	27.6	19.6	3	1
3	66	Female	22.03	L3	73.4	27.1	46.3	50.4	23	24	7	59	28.7	30.3	50.2	8.8	1	0
4	67	Female	26.93	L4	76.6	28.1	48.5	38.2	38.4	24	7	65.9	25.1	40.8	47	18.9	2	1
5	66	Male	28.41	L4	58.8	21.7	37.1	30.9	27.9	19	8	44.2	12.9	31.3	41.3	2.9	1	1
6	77	Male	27.01	L3	30.2	6.9	23.3	32.5	−2.3	24	6	33	8	25	26.1	6.9	3	0
7	65	Female	29.69	L2	71.3	31.8	39.5	22.2	49.1	22	3	59.3	25.6	33.7	40.6	18.7	4	1
8	62	Female	28.04	L2	41.8	21.7	20.1	10.1	31.7	23	7	42	16.3	25.7	24.0	18.0	2	0

Abbreviations: BMI, body mass index; PI, pelvic incidence; PT, pelvic tilt; SS, sacral slope; LL lumbar lordosis.

4.2. Primary Motion of the Adjacent Vertebrae Postoperatively

We found no consistent trend in terms of changes in adjacent segments after PLIF. Regarding the first adjacent segment at L3‐4, the primary flexion, bending, and torsion rotation were decreased in three patients (8.14 ± 2.78 vs. 15.13 ± 6.71; 8.36 ± 5.59 vs. 9.08 ± 3.57; 5.07 ± 0.56 vs. 9.25 ± 5.06), but increased in one patient, even though there was no significant difference. In addition, the primary translation along the ML direction increased during bending. For the L2‐3 motion, the primary rotation increased in all three postures. Conversely, when the L2‐3 segment was set as the second adjacent segment, the primary rotation decreased. When the L1‐2 was the second or third adjacent level, the primary flexion and bending decreased, but the primary torsion increased, although there was no significant difference (6.61 ± 3.67 vs. 5.8 ± 2.01). Additionally, when the L5‐S1 was set as the adjacent segment, the primary flexion, bending, and torsion increased in four patients, but decreased in the two remaining patients.

4.3. Coupled Motion of Adjacent Vertebrae Postoperatively (Figure 4)

FIGURE 4.

Ranges of vertebral postoperative rotations and translations of DLS patients during (a and d) flexion–extension, (b and e) bending, and (c and f) torsion of the trunk are shown in the figure. The symbol (*) represents statistical significance on between‐level comparison (p < 0.05).

In the L3‐4 segment, the coupled rotation of torsion decreased postoperatively during bending (1.48 ± 1.01 vs. 7.05 ± 5.84, p < 0.05), while the coupled flexion and bending rotation during the torsion posture decreased (6.37 ± 6.01 vs. 13.83 ± 8.12, 4.53 ± 1.97 vs. 13.06 ± 6.65; p < 0.05, p < 0.05). In the L1‐2 segment, the postoperative coupled rotation around the ML direction decreased during bending (3.69 ± 2.12 vs. 14.76 ± 7.99, p < 0.05), and the coupled flexed and bending rotation during the torsion posture also decreased (3.22 ± 2.11 vs. 9.6 ± 3.79, 3.97 ± 2.26 vs. 8.57 ± 6.73; p < 0.05, p < 0.05). At the L5‐S1 segment, most coupled rotational motions decreased or were not significantly different. However, the rotational coupled motion increased after surgery during the flexion–extension posture (4.43 ± 2.93 vs. 1.20 ± 1.05, p < 0.05). Furthermore, the coupled translational motion along AP and CC direction decreased during the torsion posture (4.46 ± 4.27 vs. 11.42 ± 3.83, 3.87 ± 3.50 vs. 8.00 ± 3.32; p < 0.05, p < 0.05).

4.4. Effects of Different Postoperative PI‐LL (Group A: PI‐LL = −10°~10°; Group B: PI‐LL = 10°~20°)

In the adjacent L5‐S1 segment, the primary flexed rotation increased in group B postoperatively during the flexion–extension posture but decreased in group A (6.08 ± 1.17 vs. −13.41 ± 2.99, p < 0.05) (Table 2). The coupled flexed rotation decreased in group B postoperatively during bending but increased in group A (−10.76 ± 5.51 vs. 18.12 ± 39.83, p < 0.05) (Table 2). The same trend, with a decrease in group B and an increase in group A, was observed for coupled translation along the ML and CC directions (−2.34 ± 3.05 vs. 4.78 ± 6.05, −4.64 ± 3.45 vs. 6.90 ± 5.18; p < 0.05, p < 0.05) (Table 3).

TABLE 2.

Comparison of vertebral rotation ranges (°) between different postoperative PI‐LL (group A: PI‐LL = −10°–~10°; group B: PI‐LL = 10°–~20°).

	L1‐L2			L2‐L3			L3‐L4			L5‐S1
	ML	AP	CC	ML	AP	CC	ML	AP	CC	ML	AP	CC
Flexion–extension
Group A	−2.94 ± 8.75	−4.70 ± 7.36	−0.64 ± 0.69	−3.17 ± 6.07	−0.42 ± 1.54	−0.53 ± 3.98	−7.62 ± 3.73	−2.22 ± 5.78	−3.40 ± 2.75	−13.41 ± 2.99	2.11 ± 6.19	1.95 ± 3.65
Group B	−3.75 ± 3.65	−2.04 ± 4.56	−4.09 ± 4.85	−5.99 ± 10.76	−6.00 ± 4.21	−2.62 ± 2.37	−6.36 ± 13.70	−2.65 ± 0.45	−1.19 ± 1.55	6.08 ± 1.17	2.08 ± 2.13	3.88 ± 3.63
P	0.882	0.503	0.296	0.606	0.214	0.393	0.832	0.943	0.542	0.006*	0.993	0.454
Left–right bend
Group A	−8.12 ± 6.29	−1.70 ± 4.43	−1.57 ± 0.19	−7.26 ± 4.32	−7.89 ± 6.91	−0.17 ± 2.70	−4.90 ± 1.90	1.74 ± 9.73	−3.52 ± 3.97	18.12 ± 39.83	1.28 ± 1.56	0.18 ± 1.32
Group B	−12.85 ± 9.73	−0.34 ± 4.66	−2.80 ± 6.06	2.15 ± 6.45	−0.21 ± 5.60	−1.10 ± 1.97	0.95 ± 5.61	−3.18 ± 1.23	−7.63 ± 7.49	−10.76 ± 5.51	1.28 ± 1.56	−0.59 ± 5.68
P	0.553	0.771	0.762	0.249	0.141	0.704	0.530	0.527	0.243	0.008*	0.765	0.799
Left–right torsion
Group A	−3.53 ± 4.06	−6.38 ± 10.32	1.35 ± 2.53	−3.74 ± 5.26	−0.35 ± 2.16	−1.00 ± 1.41	−3.25 ± 21.04	−5.77 ± 9.32	−0.19 ± 0.24	−0.83 ± 5.63	−11.14 ± 17.85	0.35 ± 5.80
Group B	−8.09 ± 3.40	−3.54 ± 6.61	0.48 ± 4.35	−3.70 ± 6.89	−1.76 ± 6.44	−1.95 ± 5.20	−11.67 ± 4.12	−11.28 ± 4.70	−8.16 ± 0.33	−4.38 ± 14.99	−0.66 ± 7.23	−1.57 ± 8.93
P	0.448	0.584	0.819	0.994	0.811	0.814	0.357	0.353	0.215	0.619	0.142	0.699

Note: Rotation around axis: ML, AP, and CC. The values were presented as mean ± SD in degree. The values were presented as mean ± SD in degree. *p < 0.05.

Abbreviations: AP, indicates anterioposterior; CC, craniocaudal; ML, mediolateral.

TABLE 3.

Comparison of Vertebral Translation Ranges (mm) between different postoperative PI‐LL (group A: PI‐LL = −10°~10°; group B: PI‐LL = 10°~20°).

	L1‐L2			L2‐L3			L3‐L4			L5‐S1
	ML	AP	CC	ML	AP	CC	ML	AP	CC	ML	AP	CC
Flexion–extension
Group A	−1.40 ± 3.27	−6.35 ± 4.76	−0.69 ± 2.78	4.44 ± 1.93	−4.29 ± 6.21	0.24 ± 2.84	−0.63 ± 4.31	−3.94 ± 1.07	0.73 ± 0.47	4.78 ± 6.05	−2.98 ± 2.33	6.90 ± 5.18
Group B	0.61 ± 3.03	−0.99 ± 6.58	−0.10 ± 4.39	−0.94 ± 1.91	1.48 ± 7.06	−1.33 ± 0.88	−2.03 ± 0.95	−1.21 ± 8.23	0.61 ± 1.82	−2.34 ± 3.05	−2.65 ± 8.02	−4.64 ± 3.45
P	0.415	0.236	0.862	0.046	0.205	0.595	0.528	0.665	0.972	0.029*	0.954	0.003*
Left–right bend
Group A	−1.88 ± 4.66	−0.30 ± 2.97	0.13 ± 4.93	1.17 ± 4.38	−2.47 ± 7.76	0.70 ± 2.65	4.29 ± 5.64	3.61 ± 0.73	0.37 ± 2.92	−4.27 ± 11.77	6.02 ± 15.51	5.14 ± 0.66
Group B	−0.32 ± 2.11	−0.50 ± 5.18	0.33 ± 1.68	0.74 ± 2.69	−4.22 ± 6.75	−3.57 ± 2.82	5.23 ± 0.97	−3.81 ± 0.04	2.27 ± 5.41	0.27 ± 2.29	−4.68 ± 11.36	0.77 ± 1.85
P	0.652	0.967	0.830	0.910	0.744	0.078	0.799	0.407	0.571	0.301	0.117	0.110
Left–right torsion
Group A	−3.64 ± 6.02	−3.05 ± 0.65	−1.39 ± 2.55	3.66 ± 3.07	−0.59 ± 6.84	−2.58 ± 3.03	4.38 ± 7.13	−1.35 ± 1.22	2.41 ± 0.13	−0.04 ± 2.67	−7.75 ± 0.30	−2.73 ± 0.37
Group B	−3.17 ± 1.47	−0.96 ± 5.21	−1.01 ± 1.98	−2.80 ± 3.08	−0.91 ± 8.64	−1.21 ± 4.45	0.30 ± 1.55	−10.83 ± 10.71	−5.46 ± 1.02	−6.24 ± 6.36	−6.57 ± 8.89	−4.84 ± 5.62
P	0.865	0.655	0.905	0.043	0.948	0.639	0.247	0.247	0.070	0.104	0.829	0.534

Note: Rotation around axis: ML, AP, and CC. The values were presented as mean ± SD in mm. *p < 0.05.

Abbreviations: AP, indicates anterioposterior; CC, craniocaudal; ML, mediolateral.

5. Discussion

To the best of our knowledge, in vivo study to measure the variations in intervertebral kinematics in patients with DLS before [28] and after undergoing PLIF is scarce. In this study, we found that the adjacent segment primary motion appeared to be subject specific, but the coupled motion decreased postoperatively in all participants, while the adjacent L5‐S1 segment coupled motion increased in six patients. Importantly, we found that different postoperative PI‐LL values had different effects on vertebral kinematics, and postoperative PI‐LL at 10° to 20° may be more appropriate in patients with DLS in the Chinese population.

Previous studies have investigated biochemical changes at the adjacent level postoperatively in vitro and using FE. For example, Ha et al. found that adjacent segmental motion increased when the lumbar spine reproduced the same ROM after lumbosacral immobilization using canine spines [29]. Similarly, Bastian et al. reported that the adjacent segment above fixation became more mobile in 10 cadaveric human spines, which may lead to accelerated degeneration of the facet joints due to increased stress at this point [7]. Huang et al. reported that the values of intradiscal pressure and intersegmental rotation at the adjacent level in the operative model were higher than those in the normal model [15]. On the contrary, Ambati et al. observed no significant difference at the adjacent level postoperatively, according to the FE model [30].

Considering the skeletal forces transduce force, the true lever arm is difficult to calculate [31]; the collection of postoperative ROM data in vivo is necessary to clarify this situation. However, only a few in vivo studies have discussed the postoperative change in kinematics in patients with low back pain [20] or lumbar degenerative spondylolisthesis [32], with these studies lacking data on postoperative kinematics in patients with DLS. Hayes et al., using lateral flexion and extension radiographs, concluded that slippage and increased translational motion played important roles in low back pain [23]. Richard suggested that although adjacent segment flexion and slippage changes were inconsistent after surgery, all patients demonstrated either increased adjacent slip or increased adjacent flexion during the flexion movement [32], which was similar to our results. Wiltse et al. found some increased adjacent segment mobility and other decreased mobility [33]. Axelsson et al. used roentgen stereophotogrammetric analysis to measure the change of lumbar position from supine to standing in patients with spondylolisthesis; they reported increased adjacent segment translation in two patients, decreased translation in two patients, and no motion change in two patients, concluding that fusion may alter adjacent segment kinematics and lead to hypermobility of the adjacent segment. Mobility was measured before and after the posterolateral fusion. Zhou et al. further investigated the changes in anatomical dimensions and biomechanical responses of the adjacent segment under weight‐bearing, upright‐standing conditions in five patients after fusion using a DFIS, and reported increased intradiscal pressure at the adjacent disc [20].

5.1. Characterization of Intervertebral Kinematics of the Adjacent Motion Segment

In Christine's cadaveric study [34], based on cadaveric spines, the ROM of flexion–extension, bending, and torsion was 11.1° ± 5.1°, 12.0° ± 8.1°, and 7.9° ± 5.5°, respectively. These were smaller than those in our study; the reason might be that the cadaveric study ignored the motion of soft tissue and ligaments. In our study, the postoperative vertebral ROMs increased in some patients at the first adjacent segment, which was similar to some previous findings [23, 32]. In Haddas's FE study, the motion of the adjacent level in left bending increased (4.18 ± 0.54 vs. 4.53 ± 0.68) and right bending increased (4.01 ± 0.84 vs. 4.27 ± 0.85) after fusion surgery [24]. However, in some patients, the postoperative ROMs either did not change significantly or decreased [33]. The change in the motion pattern was subject‐specific and dependent on the location of the adjacent segment. In this study, the postoperative ROM of primary rotation and translation increased, while coupled motion decreased at the adjacent level when the fusion exceeded the scoliotic apex vertebrae, including in the L2‐3 and L3‐4 in some patients. However, when the upper instrumented vertebra was lower than the scoliosis apex vertebra, the postoperative ROMs of the primary and coupled motions decreased. This may be because the adjacent segment is located between the scoliotic apex vertebrae and the fusion level, decreasing vertebral motion. Regarding the second and third adjacent levels, the postoperative coupled motion decreased, while the primary motion did not change significantly. The spine was subject to diverse loading conditions, with coupled motions being different [35].

5.2. Change of Coupled Motion of Adjacent Segments

We believe that the reduction in coupled motion of adjacent segments after surgery may be an indication of postoperative recovery in patients. By surgically reducing excess motion, it alleviates the symptoms and allows patients' motion patterns to recover as closely as possible to patterns revolving around the main axis of rotation. This helps in relieving pain and avoiding mechanical complications. This is also why the symptoms of this group of patients have recovered relatively well after surgery, and it explains why there were no occurrences of ASD during the short‐term follow‐up period after spinal surgery. Cunningham et al. previously demonstrated adjacent lumbar intradiscal pressures increase 45% [36]. Stress on the adjacent cephalad levels to fusion was increased because of a larger lever arm and non‐physiological center of motion. The increase in adjacent motion was caused by the effective force experienced by a level adjacent to a fused motion segment [37]. In other words, arthrodesis of the motion segment caused overactivity at the adjacent segment, which may lead to ASD over time. This finding indicates that ASD does not occur solely because of natural disc degeneration. Moreover, when the increased stress is associated with a decreased primary ROM and increased coupled motion pattern, the development of ASD may be accelerated; further FE studies in the future are required to prove this conclusion.

5.3. Relationship of Vertebral Kinematics and Postoperative PI‐LL

Ha et al. noted a significant correlation between the clinical relevance of the SRS‐Schwab classification for DLS and PI‐LL [38]. From a biomechanical point of view, differences in PI‐LL were expected to be associated with the load patterns in the lumbar spine, with this being even more important in the adjacent segment after spinal fusions, which has been supported by the findings of previous studies [39, 40]. Clinically, the optimum value for postoperative PI‐LL remains under discussion. To avoid disability, Schwab et al. suggested that PI‐LL should be controlled to < 11° [41] and that the postoperative PI‐LL could be corrected to ≤ 9° [42]. However, Lafage demonstrated that operative PI‐LL should be associated with age, with cut‐off limits of −10.5° for patients younger than 35 years and 16.7° for those older than 75 years [43]. Lafage combined the SRS‐Schwab classification, GAP score, and age‐adjusted thresholds to create the sagittal age‐adjusted score (SAAS) and demonstrated its ability to predict postoperative PJK and PJF in patients with ASD [44]. Liu et al. and Srikanth et al. further reported that PI‐LL mismatch did not play an effect on health‐related quality of life or postoperative outcome [45]. Furthermore, Bai et al. found that patients with mismatched PI‐LL under long segment fusion could suffer more low back pain [46]. Zhang et al. and Sun et al. further suggested that the ideal PI‐LL may be between 10° and 20° in adult degenerative scoliosis patients after long posterior instrumentation and fusion [47, 48]. These differences may be related to variations in spinopelvic parameters between populations of different ethnic backgrounds [40].

In the present study, we found that the postoperative coupled motion, including the flexed rotation during bending posture and anteroposterior translation during flexion–extension posture, at the L5‐S1 adjacent segment increased in patients with postoperative PI‐LL < 10°. In contrast, the coupled motion decreased, while the primary motion compensatory increased at the L5‐S1 adjacent segment postoperatively in patients with postoperative PI‐LL between 10° and 20°. This compensatory increase in primary motion may be more suitable for the natural spinal motion pattern, which indicates that controlling the postoperative PI‐LL between 10° and 20° may be easier to return to the normal motion patter.

This study had some limitations. First, the follow‐up time was relatively short about—1 year—which may have influenced the evaluation of the possible relationship between the vertebral ROM and radiographic parameters. Second, the sample size was limited due to the single‐center study design. Nevertheless, this sample size was small, and we would like to collect more postoperative kinematics in the future. Third, this study lacked an analysis of the change in the mechanical factors; more FE studies should be performed in the future to clarify this point. Finally, the laboratory environment collecting our data was more closely related to the real situation of our human bodies with the surrounding tissues (including ligament, muscle and soft tissue) and spinal degeneration (including the degeneration of the disc, osteophytes, and osteoporosis).

6. Conclusions

In summary, we concluded that postoperative coupled motion at the adjacent segment decreased to avoid the development of ASD in patients with DLS over short‐term follow‐up. However, the postoperative primary motion increased when the fusion exceeded the scoliotic apex vertebra and decreased when the upper instrumented vertebrae were lower than the scoliotic apex vertebra. After undergoing correction of spinal order, the overall deformity in patients disappears. There is no need for additional coupled motion in adjacent segments to make compensatory adjustments. As a result, the reduction of coupled motion in adjacent segments indicates an improved balance of the spine compared to before the surgery. Finally, postoperative PI‐LL values between 10° and 20° were associated with lower coupled motion and higher primary motion at L5‐S1.

Author Contributions

Zhuoran Sun: contributed to the study conception and design, and performed material preparation. Yudong Zhao and Siyu Zhou: contributed to the study conception and design, and performed data collection. Zhuofu Li and Weishi Li: contributed to the study conception and design, and performed data analysis. Fei Xu and Shuai Jiang: contributed to the study conception and design, and wrote the first draft of the manuscript. Zheyu Fan: investigation, resources, software. All authors commented on previous versions of the manuscript, and all authors read and approved the final manuscript.

Disclosure

The authors have nothing to report.

Ethics Statement

Ethics approval and informed consent: This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Ethics Committee of Peking University Third Hospital (Date 2019.5.30/No M2019239).

Conflicts of Interest

The authors declare no conflicts of interest.

Xu F., Jiang S., Sun Z., et al., “The Relationship Between Postoperative Variation of Adjacent Segment Three‐Dimensional Kinematics and Postoperative PI‐LL in Patients With Degenerative Lumbar Scoliosis After Short Segment Fixation,” Orthopaedic Surgery 17, no. 12 (2025): 3342–3353, 10.1111/os.70178.