Risk Factors and Three Radiological Predictor Models for the Progression of Proximal Junctional Kyphosis in Adult Degenerative Scoliosis Following Posterior Corrective Surgery: 113 Cases With 2-years Minimum Follow-Up

Abstract

Study design

Retrospective cohort study.

Objective

To identify risk factors and predictive models for proximal junctional kyphosis (PJK) in a long-term follow-up of patients with adult degenerative scoliosis (ADS) following posterior corrective surgeries.

Materials and Methods

A consecutive 113 ADS patients undergoing posterior corrective surgery between January 2008 and April 2019 with minimum 2-year follow-up were included. All patients underwent preoperative, postoperative, and final follow-up by X-ray imaging. Multivariate logistic analysis was performed on various risk factors and radiological predictor models.

Results

PJK was identified radiographically in 46.9% of patients. Potential risk factors for PJK included postoperative thoracic kyphosis (TK) (P < .05), final follow-up Pelvic Tilt (PT) (P < .05), PT changes at final follow-up (P < .05), age over 55 years old at the surgery (P < .05), theoretical thoracic kyphosis–actual thoracic kyphosis mismatch (TK mismatch) (P < .05) and theoretical lumbar lordosis–acutal lumbar lordosis mismatch (LL mismatch) (P < .05). As for the predictive models, PJK was predictive by the following indicators: preoperative global sagittal alignment ≥45° (Model 1), postoperative pelvic incidence–lumbar lordosis mismatch (PI–LL)≤10° and postoperative PI–LL overcorrection (Model 2), and TK+LL≥0° (Model 3) (P < .05). Postoperative TK mismatch (OR = 1.064) was independent as risk factors for PJK, with the cut-off values respectively set at −28.56° to predict occurrence of PJK.

Conclusion

The risk of radiographic PJK increases with an age over 55 years old and higher postoperative TK. In addition, postoperative TK mismatch is an independent risk factor for developing PJK. All three predictive models could effectively indicate the occurrence of PJK.

Keywords

radiology sagittal alignment TK mismatch predictive model proximal junctional kyphosis adult degenerative scoliosis

Introduction

Adult degenerative scoliosis (ADS) refers to a spectrum of disabling curves that results in sagittal imbalance and occurs specifically after skeletal maturity.¹ The main pathogenesis of ADS is the degeneration of the intervertebral discs and facet joints. On top of compromising patient’s life quality, ADS inflicts a significant health burden on elderly patients in current aging society and has become a focus of clinical concern.² For patients with severe ADS, optimal surgical correction is essential in improving clinical outcomes and preventing sagittal decompensation.

Proximal junctional kyphosis (PJK) is a common complication after posterior instrumentation corrective surgery for ADS correction. The proximal junction is between the caudal endplate of the upper instrumented vertebra (UIV) and the cephalad endplate of the vertebra that is two levels above the UIV.³ Lowe et al defined PJK as proximal junctional sagittal Cobb angle >10° and at least 10° greater than the preoperative measurement.^4,5 According to previous literatures, the radiological incidence of PJK is varied and ranges from 6% to 50%.^6,7

Varied risk factors have been considered to be associated with the occurrence of PJK, including age at surgery >55 years, low bone mineral density, obesity, preoperative comorbidities, fusion to sacrum, high preoperative lumbar lordosis (LL), and thoracic kyphosis (TK).⁶ The Scoliosis Research Society (SRS) Schwab classification also proposed that coronal curve types and spinopelvic modifiers (SVA, PT, PI–LL) contributed to a strong correlation with postoperative prognosis in ADS patients, including the occurrence of PJK.⁸ Besides, several radiological predictive models for PJK have been found in previous studies, including global sagittal alignment (GSA), TK+LL, and pelvic incidence—lumbar lordosis (PI–LL).^9-11

However, to the best of our knowledge, the risk factors and predictive radiographical models of PJK are still yet to be defined, when one takes in to account the patient, radiographical, and surgical factors together. The present study was thus designed to identify risk factors associated with development of PJK and to evaluate three proposed predictive models for PJK. Based on our findings, we aimed to further provide clinical recommendations on current operational strategy for preventing PJK.

Materials and Methods

The present study was approved by the Institutional review board of the authors’ affiliated institution (IRB number: M2021239). A retrospective review of the medical records and radiographs was conducted for 113 ADS patients from a single institution. All patients received posterior corrective surgeries performed by the senior author between 2008 and 2019. Patients were eligible for inclusion if they (a) were diagnosed with ADS; (b) were treated with posterior corrective surgery; (c) were at least 50 years old; (d) completed radiographical follow-up at a minimum of three time points (preoperatively, immediately after the operation, and at the last follow-up). Patients with less than 2-years follow-up and incomplete radiographical records were excluded from the study.

Data Collection and Radiographic Measurements

We reviewed patient charts and radiographs, and recorded the following demographic and clinical data for each patient: age, sex, and body mass index (BMI). Surgical data such as surgical approach and UIV were also recorded. Preoperative, postoperative, and final follow-up full-length standing spine radiographs were measured and evaluated by a single observer. Spinal measurements were performed through the Picture Archiving and Communication System. They included sagittal vertical axis (SVA), sacral slope (SS), lumbar lordosis (LL), thoracic kyphosis (TK), pelvic incidence (PI), and pelvic tilt (PT). Theoretical values of TK and LL were calculated according to the formulas proposed by Sebaaly et al: LL = .67PI+23.7 and TK = .15PI+43¹¹. In addition, we calculated the mismatch values of theoretical TK and LL minus their actual values. Coronal Cobb angles were also measured.

We focused on 3 predictive models which combined both spinal and pelvic parameters to assess sagittal alignment. The predictive models were defined and calculated respectively:

Model 1: GSA was calculated as a predictor. Normal GSA was defined as less TK+PI+LL than 45°, as suggested by the study of Rose et al.¹²

Model 2: Pelvic incidence minus lumbar lordosis (PI–LL) was calculated according to the SRS-Schwab classification.¹³ We defined ideal sagittal realignment as |PI

-

LL|≤10°.¹⁴ Meanwhile, according to the study by Lafage,¹⁵ if the value of PI

-

LL after corrective operation was smaller than the normal values specific to the patient’s age group, the case would be categorized as overcorrected (OVER); otherwise, the case would be categorized as aligned (ALIGN). Ideal age-specific alignment criteria provided by Lafage et al are reported in Table 1¹⁵.

Table 1.

Ideal age-specific alignment criteria for PI–LL.

Age group	PI–LL ( $°$ )
<35	−10.5
35–44	−4.6
45–54	0.5
55–64	5.8
65–74	10.5
$\geq$ 74	17

Abbreviations: PI, pelvic incidence; LL, lumbar lordosis.

Model 3: the predictor was the sum of thoracic kyphosis and lumbar lordosis (TK+LL), according to the approach proposed by Mendoza–Lattes.¹⁰ Normal TK+LL was defined as less than $0 °$ .

Proximal junctional kyphosis was defined at the final follow-up radiograph as described by Glattes: proximal junctional Cobb angle greater than 10° and at least 10° greater than the corresponding preoperative measurement. The presence of both criteria was necessary to be considered abnormal.¹⁶

Statistical Analysis

Data was analyzed with SPSS 23.0. Continuous variables with normal distribution were presented as mean ± standard deviation, continuous variables with skewed distribution were presented as median (lower quartile, upper quartile), and categorical variables were expressed as frequency or percentages. Independent t test and Mann–Whitney U-test were used to analyze the differences in continuous variables between groups. Fisher’s exact test and $χ^{2}$ analyses were used to examine the differences among categorical variables. Variables with P values <.05 in the univariate analyses were entered into a multivariate logistic regression model. We used receiver operating characteristic (ROC) curve to find out the optimal cut-off points which presented the largest Youden index. For each variable, we computed the odds ratio (OR) with 95% confidence interval.

Results

Prevalence and Characteristics of PJK Patients

There were 94 female and 19 male patients. The average age at the time of surgery was 60.8 $\pm$ 9.39 years. Radiographical PJK was developed in 53 of the 113 patients (46.9%) by the final follow-up. Accordingly, patients were divided into two groups, PJK group (n = 53) and non-PJK group (n = 60). For the PJK group, the mean age was 61.72 $\pm 9.57$ years, and combined anterior-posterior surgery was performed in 22 patients. Of the 53 subjects with PJK, the UIV of 17 patients was at T10 or above, 16 at the thoracolumbar junction (T11–L1), and 20 at the L2 or below, while the lower instrumented vertebrae (LIV) of 20 patients was at L5 or above, and 33 at S1. For the non-PJK group, the mean age was 59.98 $\pm 9.23$ years, and combined anterior–posterior surgery was performed in 20 patients. Of the 60 subjects without PJK, the UIV of 18 patients was at T10 or above, 17 at the thoracolumbar junction(T11–L1), and 23 at L2 or below, while the lower instrumented vertebrae (LIV) of 25 patients was at L5 or above, and 30 at S1.

As shown in Table 2, the baseline characteristics and surgical parameters in both groups were similar (all P > .05), except for a significantly higher number of older patients in the PJK group (>55 years; P = .023).

Table 2.

Comparison of patient characteristics and surgical factors between PJK and non-PJK group.

	PJK group(N = 53)	Non-PJK group(N = 60)	Statistic^a	P value
Gender			χ² = .002	.964
Male	9(17.0%)	10(16.7%)
Female	44(83.0%)	50(83.3%)
BMI (kg/m2)	25.92±4.70	25.14±3.49	t = −.639	.526
Age at surgery (years)	63.00(58.50,67.50)	62.00(54.25,65.75)	U = −1.204	.229
Age at surgery ≥55(years)	45(86.5%)	41(68.3%)	χ² = 5.180	.023*
Surgical approach			χ² = .806	.369
Combined anterior-Posterior approach	22(41.5%)	20(33.3%)
Otherwise	31(56.4%)	40(66.7%)
UIV			χ² = .036	.982
Above T10	17(32.1%)	18(30.0%)
T11–L1	16(30.2%)	17(28.3%)
Below L2	20(37.7%)	23(38.3%)
LIV			χ² = 3.645	.302
Above L5	20(37.7%)	25(41.7%)
S1	33(62.35)	30(50.0%)

Abbreviations: PJK, proximal junction kyphosis; BMI, body mass index; yr, years; UIV, upper instrumented vertebrae; LIV, lower instrumented vertebrae.

* Statistical significance.

^aCategorical variables were presented as number (percentage) and tested by chi-square test; continuous data with skewed distribution were presented as median (interquartile range, IQR) and tested by Mann-=–Whitney U-test, continuous data with normal distribution were presented as mean ± standard deviation (SD) and tested by t test.

Table 3 shows the comparisons of radiographical parameters between PJK group and non-PJK group at each measurement (preoperation, postoperation, and final follow-up). The PJK group demonstrated higher postoperative TK (P = .009) and higher PT at final follow-up (P = .001) compared to those of the non-PJK group. Meanwhile, the average postoperative PI

-

LL in the PJK group was significantly lower than that in the non-PJK group (P = .047).

Table 3.

Comparison of radiographical parameters between PJK and non-PJK group.

	PJK group(N = 53)	Non-PJK group(N = 60)	Statistic^a	P value
SVA (C7 plumb line to S1) (mm)
Preoperation	42.95±5.49	49.96±5.13	t = .932	.353
Postoperation	18.19±3.84	25.42±4.30	t = 1.240	.217
Final follow-up	27.84±4.27	26.91±2.72	t = −.188	.852
Thoracic kyphosis (T5–T12) (°)
Preoperation	19.50 (13.55, 27.70)	18.30 (11.75, 22.35)	U = −1.116	.264
Postoperation	21.67 (21.67, 29.05)	21.67 (14.73, 22.35)	U = −2.616	.009*
Final follow-up	32.00 (23.90, 40.70)	28.31 (17.55, 35.65)	U = −1.899	.058
Lumbar lordosis (T12–S1) (°)
Preoperation	19.30 (−1.75, 35.70)	11.22 (−25.28, 39.20)	U = −1.427	.154
Postoperation	34.80 (−2.00, 44.65)	24.10 (−31.95, 40.30)	U = −1.602	.109
Final follow-up	28.10 (19.25, 33.10)	21.15 (−24.63, 34.15)	U = −1.459	.145
Sacral slope (°)
Preoperation	22.86±10.89	25.74±12.86	t = 1.275	.205
Postoperation	28.37±7.14	29.48±7.75	t = .790	.431
Final follow-up	22.90±11.70	24.58±10.28	t = .813	.418
Pelvic tilt $(°)$
Preoperation	26.55±12.00	23.20±12.65	t = −1.439	.153
Po-operation	19.63±9.92	19.55±9.64	t = −.039	.969
Final follow-up	44.18±21.78	32.23±15.89	t = −3.359	.001*
Pelvic incidence (°)
Preoperation	48.93±15.74	48.93±10.99	t = .16	.873
Postoperation	47.60±10.48	48.33±10.99	t = −.361	.719
Final follow-up	52.11±8.98	50.24±11.28	t = −.966	.336
Global sagittal alignment (TK+PI+LL) (°)
Preoperation	84.51±33.65	73.26±41.20	t = −1.576	.118
Postoperation	92.22±38.11	79.58±40.89	t = −1.694	.093
Final follow-up	76.58±35.60	77.47±33.50	t = .948	.891
PI $-$ LL (°)
Preoperation	27.40 (12.75,5 1.70)	35.53 (15.38, 68.73)	U = −1.341	.18
Postoperation	11.90 (2.70, 49.55)	19.35 (9.83, 68.40)	U = −1.985	.047*
Final follow-up	24.02 (17.47, 37.15)	26.82 (16.28, 68.20)	U = −.794	.427
TK mismatch (°)
Po-operation	−26.95 (−29.04, −22.44)	−29.22 (−34.98, −26.47)	U =− 2.769	.006*
Final follow-up	−18.67 (−27.50, −9.68)	−22.49 (−32.74, −14.70)	U = −.898	.369
LL mismatch (°)
Po-operation	−19.95 (−56.54, −9.90)	−28.89 (−79.59, −15.36)	U = −2.019	.043*
Final follow-up	−30.85 (−39.75, −24.30)	−33.65 (−76.69, −23.35)	U =− 1.668	.095

Abbreviations: PJK indicates proximal junction kyphosis; SVA, sagittal vertical axis; PI, pelvic incidence; LL, lumbar lordosis; TK mismatch, mismatch values in theoretical TK minus actual TK; LL mismatch, mismatch values in theoretical LL minus actual LL.

* Statistical significance.

^aContinuous data with skewed distribution were presented as median (interquartile range, IQR) and tested by Mann–Whitney U-test, continuous data with normal distribution were presented as mean ± standard deviation (SD) and tested by t test.

According to the formulas LL = .67PI+23.7 and TK = .15PI + 43, as proposed by Sebaaly, we also calculated the theoretical values of LL and TK and their gaps with actually their observed values. The mismatch of LL, that is, the difference between average theoretical LL and observed postoperative LL, in the non-PJK group was significantly higher than that in the PJK group (P = .043). Likewise, the mismatch of TK, that is, the difference between average theoretical TK and observed postoperative TK, in the non-PJK group was also found higher than that in the PJK group (P = .006, Table 3) (Figure 1 and 2).

Figure 1.

64-year-old woman presenting with degenerative lumbar scoliosis and sagittal imbalance (Cobb angle 19.9°, TK 3.2°, LL −1.7°, SS 21.9°, PT13.9°, PI 36.1°) (A, B). We performed posterior iliac screw fixation on T10–S2. Postoperative Thoracic Kyphosis was 17.4° while the theoretical TK was 47.8°, indicating a large mismatch (C).postoperative 3-year whole spine lateral radiograph showing normal sagittal balance without occurence of proximal junctional kyphosis (proximal junctional kyphosis angle 5.6°) at final follow-up (D).

Figure 2.

63-year-old woman presenting with degenerative lumbar scoliosis and major sagittal imbalance (Cobb angle 31.2°, TK 27.0°, LL −23.1°, SS 25.5°, PT30.2°, PI 54.8°) (A, B). She was operated with posterior instrumentation on T10–S1. Postoperative Thoracic Kyphosis was 40.0° while the theoretical thoracic kyphosis was 50.65°(C). 3 years postoperatively (final follow-up), she presented a proximal junctional kyphosis with a proximal junctional kyphosis angle of 24.7°(D). Compared with the non-proximal junctional kyphosis subject presented in Figure 1, she had a higher postoperative thoracic kyphosis which was closer to the theoretical value. Our findings suggested a necessity for developing personalized orthopedic plan against adult degenerative scoliosis patients specifically.

Table 4 presents the intergroup comparisons in the changes of radiographical parameters from postoperation to the final follow-up. It was observed that the average PT change at the final follow-up in the PJK group was 17.63°, which was larger than the 9.02° in the non-PJK (P = .015).

Table 4.

Comparison of radiographical parameter changes between PJK and non-PJK groups.

	PJK group(N = 53)	Non-PJK group(N = 60)	Statistic^a	P value
SVA (C7 plumb line to S1) (mm)
changes in SVA at postoperation	−24.76±37.04	−24.53±48.29	t = −.925	.357
changes in SVA at final follow-up	−15.12±48.33	−23.05±42.77	t = .027	.978
Thoracic kyphosis (T5–T12) (°)
changes in TK at postoperation	13.70 (.850,20.30)	10.75 (−.93,19.88)	U = −.466	.641
changes in TK at final follow-up	3.46±11.28	2.42±11.33	t = −.485	.629
Lumbar lordosis (T12–S1) (°)
changes in LL at postoperation	5.18±17.66	4.49±19.68	t = −.195	.846
changes in LL at final follow-up	7.40 (−1.30, 27.35)	13.95 (−8.10, 51.02)	U = −.952	.341
Sacral slope (°)
changes in SS at po-operation	4.40 (−2.25, 11.50)	3.20 (−4.30, 8.55)	U = −.656	.512
changes in SS at final follow-up	3.30 (−6.70, 9.30)	.85 (−8.38, 7.13)	U = −.883	.377
Pelvic tilt (°)
changes in PT at po-operation	−6.93±11.64	−3.65±12.34	t = 1.448	.151
changes in PT at final follow-up	17.63±19.43	9.02±17.75	t = −2.446	.015*
Pelvic incidence (°)
changes in PI at po-operation	−.80 (−6.20, 4.55)	1.30 (−6.18, 7.10)	U = −.877	.380
changes in PI at final follow-up	3.52 (−4.69, 10.95)	2.61 (−5.45, 9.69)	U = −.460	.645
Global sagittal alignment (°)
changes in GSA at po-operation	7.72±21.22	6.32±28.15	t = −.295	.769
changes in GSA at final follow-up	17.90±26.31	17.24±33.54	t = −.116	.908
PI $-$ LL (°)
changes in PI-LL at po-operation	−6.11±21.64	−5.09±22.71	t = .242	.809
changes in PI-LL at final follow-up	−1.24±27.96	−4.45±30.16	t = −.584	.560

Abbreviations: PJK indicates proximal junction kyphosis; SVA, sagittal vertical axis; PI, pelvic incidence; LL, lumbar lordosis.

* Statistical significance.

^aContinuous data with skewed distribution were presented as median (interquartile range, IQR) and tested by Mann-Whitney U-test, continuous data with normal distribution were presented as mean ± standard deviation (SD) and tested by a t test.

To further examine the predictive models, the actual values of specific parameters were calculated and compared to their theoretical values in terms of each measurement:

Model 1: For 77.9% of the subjects, the preoperative GSA was greater than 45°, whereas for the rest 22.1%, it was less than 45°. Among the PJK group, the incidence of GSA≥45° was 86.8%, which was significantly higher than the 70.0% in the non-PJK group (P = .032) (Table 5).

Table 5.

Comparison of predictive models between PJK and non-PJK group.

	PJK group(N = 53)	Non-PJK group(N = 60)	Statistic	P value
\|PI $-$ LL\|≤10˚n(%)
Preoperation	8 (15.1%)	10 (16.7%)	χ² = .052	.82
Postoperation	22 (41.5%)	11 (18.3%)	χ² = 7.311	.007*
Final follow-up	9 (17.0%)	9 (15.0%)	χ² = .082	.774
PI $-$ LL overcorrection n (%)
Postoperation	24 (45.3%)	16 (26.7%)	χ² = 4.265	.039*
Final follow-up	47 (88.7%)	50 (83.3%)	χ² = .662	.416
GSA≥45° n(%)
Preoperation	46 (86.8%)	42 (70.0%)	χ² = 4.606	.032*
Postoperation	43 (81.1%)	44 (73.3%)	χ² = .966	.326
Final follow-up	45 (84.9%)	50 (83.3%)	χ² = .52	.82
TK+LL≥0° n(%)
Preoperation	45 (84.9%)	39 (65.0%)	χ² = 5.845	.016*
Postoperation	41 (77.4%)	41 (68.3%)	χ² = 1.151	.283
Final follow-up	46 (86.8%)	48 (80.0%)	χ² = .928	.335

Abbreviations: PJK indicates proximal junction kyphosis; PI, pelvic incidence; LL, lumbar lordosis; GSA, global sagittal alignment; TK, thoracic kyphosis.

* Statistical significance.

Model 2: The average postoperative PI $-$ LL was 32.14°. The PI $-$ LL within $\pm 10 °$ was more common in the PJK group (41.5%) than that in the non-PJK group (18.3%) (P = .007). When compared to the ideal age-specific alignment thresholds (Table 1), 35.4% of the patients had postoperative PI $-$ LL overcorrected, as opposed to the 64.6% who did not. In addition, more subjects in PJK group (45.3%) showed postoperative PI $-$ LL overcorrection than those in non-PJK group (26.7%) (P = .039) (Table 5).

Model 3: The average preoperative TK+LL was 29.79°. Moreover, 84.9% of PJK group was associated with positive TK+LL preoperation, compared to the 65.0% of non-PJK group (P = .016) (Table 5).

Risk Indicators

Table 6 shows the results of multiple logistic regression. Among the radiographic parameters, pelvic tilt (PT) at the final follow-up was significantly associated with PJK (OR = 1.042, P = .001). As for the predictive models, postoperative TK mismatch was identified as an independent risk indicator of PJK (OR = 1.064, P = .016).

Table 6.

Risk factors for PJK in ADS patients identified through multivariate analysis.

Risk factors	Odds ratio	95% CI for OR	P value
Final follow-up PT	1.042	1.017–1.068	.001
Postoperative TK mismatch	1.064	1.012–1.120	.016

Abbreviations: PT, pelvic tilt; TK mismatch, mismatch values in theoretical TK minus actual TK; CI, confidence interval; OR, odds ratio.

To further evaluate the effectiveness of the two parameters in indicating PJK, we performed ROC curves and calculated the Youden index to figure out the best cut-off points. The ROC curve was painted based on the division of PJK group and non-PJK group (Figure 3). We found that the best cut-off values for indicating PJK were final follow-up PT = 39.05° (AUC = .676, sensitivity = 64%, specificity = 68%) and postoperative TK mismatch = −28.56° (AUC = .651, sensitivity = 74%, specificity = 58%).

Figure 3.

Receiver operating characteristic curve to find the optimal cut-off points of final follow-up Pelvic Tilt and postoperative thoracic kyphosis mismatch to predict proximal junctional kyphosis.

Discussion

Proximal junctional kyphosis has been identified as a common complication in patients undergoing corrective surgeries for adult spinal deformities.¹⁷ As reported by Bridwell, this condition may progress with deterioration of sagittal alignment with an apex, and may lead to compromised balance in the sagittal plane, vertebral collapse, and, in some cases, neurological symptoms.¹⁸

This study reviewed 113 ADS patients with a 46.9% prevalence of PJK, which was similar to the incidence described in previous studies.^19,20 Existing literatures suggested that the development of PJK is associated with a variety of baseline characteristics, surgical parameters, and radiographical parameters.^6,7 Nonetheless, related indicators for predicting the occurrence of PJK are still controversial without a comprehensive examination. Therefore, the present study systematically evaluated potential risk factors for PJK as well as 3 radiological predictive models.

Existing evidence has pointed to age as a risk factor for PJK.²¹ As reported by Liu et al, patients who were older than 55 years at the surgery were more liable to developing PJK compared to younger patients.⁶ In addition, Kim et al suggested that advanced age correlated with a high prevalence of revision surgeries for PJK.²² Consistent with previous findings, our study revealed a higher prevalence of PJK among patients over 55 years old (52.3%), as contrasted to the 28% among the younger patients, which indicated advanced age as a risk factor for PJK. Age dependents disc and facet joint degeneration also led to the development of spinal extension musculature in the older adult patients. These degenerative factors might affect the sagittal balance of patients, which may eventually lead to PJK.¹⁷

There were also some radiographical variables identified as potential risk factors for PJK. According to Buell et al, postoperative TK was significantly associated with developing PJK.²³ The present study also demonstrated that greater postoperative TK may serve as risk factors for PJK. This could be explained by the increased junctional stress at the proximal end of UIV caused from large thoracic kyphosis. In patients with a hyperkyphotic thoracic spine, the compensatory mechanism often fails with weak back muscle strength caused by surgery.²⁴ Besides, we postulated that a relative higher postoperative TK implied a substantial correction after surgery, which may cause a reciprocal kyphosis at the proximal junction.²⁵ We further suggested a necessity of monitoring TK intraoperatively to prevent PJK.

According to previous study, a large TK was often accompanied with a large LL, which implies that the latter was also a risk factor for PJK.²² Non-ideal LL might lead to an acute change in the sagittal alignment and eventually results in PJK as a compensatory response to restore sagittal balance.²⁶ Considering the essential role of TK and LL in keeping sagittal spinal balance, Sebaaly proposed to restore LL and TK to their theoretical ideal values based on PI, as calculated through the formulas LL = .67PI+23.7 and TK = .15PI+43¹¹. After applying the two formulas to our subjects, we noted that observed postoperative TK and LL were closer to their theoretical values in the PJK group compared with non-PJK group. One possible explanation is that the construction of the formula was based on symptomatic subjects, whose sagittal parameters might have a different range from those of ADS patients.²⁷ For example, normal TK values were universally higher than observed TK values in the present study. However, a relative higher postoperative TK was found to be a potential risk factor for PJK among ADS patients. It is possible that excessive amount of sagittal balance correction against normal value would lead to a higher prevalence of PJK. Multivariable analysis further indicated that postoperative TK mismatch was an independent risk factor for PJK. Therefore, one should note that merely restoring TK and LL to their theoretical values based on normal population is insufficient to prevent the occurrence of PJK among ADS patients, and it is vital to develop a personalized orthopedic plan for a moderate correction.

Besides sagittal thoracolumbar parameters, pelvic parameters also have important implications for the postoperative development of PJK. In line with previous study, we found that a higher final follow-up PT was significantly associated with developing PJK.¹⁴ Increased PT may be the primary cause of sagittal malalignment, it could also be explained as compensatory mechanisms for the forward incline of sagittal alignment caused by PJK.^28,29 In the present study, the early occurrence of PJK in some cases precluded the ability to obtain a follow-up radiograph without the presence of PJK and the accompanying compensatory changes. Therefore, whether final follow-up PT is a risk factor for PJK or is a compensatory consequence of PJK needs further study. We suggested PT as an indicator for diagnosing PJK postoperatively with the cut-off value of 39.05°.

Finally, we calculated the three radiographical models to assess their efficacy for predicting PJK in ADS patients, based on preoperative, postoperative and final follow-up radiographs. All three radiographical models were found statistically predictive for the occurrence of PJK.

The first predictive model suggests that GSA (PI+LL+TK) greater than 45° implies the occurrence of PJK, as proposed by Rose et al.¹² The model was also supported by the study of Yagi et al, who found that 84% of patients in the PJK group were subject to non-ideal GSA (PI+LL+TK>45°) in the preoperative period. In line with these studies, our results suggested that preoperative GSA>45° was a risk factor for PJK. A possible account is that positive GSA may induce positive SVA and increase the stress for the proximal and distal junctions during the instrumentation, which may result in PJK.⁷

The second predictive model was PI $-$ LL should be less than ± 10° as described by Maruo.¹⁴ PI $-$ LL reflects the proportionality of the pelvic parameters relative to the sagittal parameters.³⁰ Sentler suggested that a higher mismatch between PI and LL indicated higher shear stresses, which are responsible for accelerated degenerations of adjacent segments.³¹ Gomez et al found that postoperative PI $-$ LL >20° was associated with the occurrence of PJK.³² In addition, Maruo et al reported that postoperative PI $-$ LL >10° was associated with the occurrence of PJK.¹⁴ However, this study found that there was a higher prevalence of PJK among patients with ideal PI–LL mismatch (41.5%) compared to those with non-ideal PI–LL mismatch (8.3%). This inconsistency with previous studies should be interpreted with caution.

It is possible that the postoperative PI–LL mismatch in the PJK group was generally overcorrected, which ultimately resulted in sagittal overcorrection and therefore PJK. Using the age-adjusted ideal PI–LL mismatch criteria proposed by Lafage,¹⁵ we found that patients who developed PJK underwent significantly greater correction. Therefore, our data revealed that global alignment overcorrection was associated with a higher prevalence of PJK. Overcorrection of PI–LL mismatch have been proposed as potential risk factors leading to the development of PJK. Yang demonstrated that larger PI–LL mismatch could lower the risk of PJK.³³ More specifically, Kim et al suggested that excessive correction of the sagittal plane might lead to increased stress concentrations at the proximal end of the fusion.¹⁶ Furthermore, they pointed out that patients with PI–LL mismatch <11° were more likely to have late-onset PJK that would require further surgeries.³⁴ Patients with an overcorrected PI-LL mismatch would unconsciously generate compensatory mechanisms to realign and correct the sagittal balance, which might lead to PJK.³⁵ Surgical plans should incorporate age-adjusted ideals to minimize the risk of PJK and increase the likelihood of sustainable realignment success.

The third predictive model, proposed by Mendoza-lattes et al, suggested that the sum of LL and TK is positively associated with the incidence of PJK. Mendoza-lattes reported that, for every 10 degrees of increase in preoperative TK+LL, the risk of developing PJK would increase by 140%. Our study also found that the PJK group had a higher prevalence of positive preoperative TK+LL, indicating that preoperative TK+LL>0°was a risk factor for PJK. The relationship between TK and LL plays a critical role in keeping sagittal balance, as TK largely depends on LL and the number of vertebras involved in the LL curve.³⁶ Our findings implied that the risk of PJK could be minimized by including sufficient correction of LL and TK in the surgical plans.

There are some limitations of this study. Firstly, as described above, causal relationship between pelvic tilt and PJK cannot be warranted with the cross-sectional data at final follow-up. Secondly, since our sample included more female patients, the results might be subject to gender-bias. Finally, it is a retrospective study with some inaccessible variables (such as bone density to study the effect of osteoporosis) and missing values. Therefore, we recommend further prospective studies that are adequately designed to examine our present conclusions.

Conclusion

For ADS patients who receive corrective surgeries, our study found that postoperative TK mismatch was an independent risk factor for PJK. In addition, the risk of radiographic PJK increases with an age over 55 years old and higher postoperative TK. All three predictive models can satisfactorily predict the occurrence of PJK. Respectively, the three models suggest three conditions to prevent postoperative PJK: 1) GSA less than 45°; (2) surgical overcorrection as indicated by PI–LL, and (3) positive sum of TK and LL. To summarize, our results stress the importance of restoring the global alignment balance to its ideal range and avoiding the overcorrection of sagittal balance in preventing postoperative PJK among ADS patients.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author Note

The present study was approved by the Institutional review board of the authors’ affiliated institution (IRB number: M2021239)

ORCID iDs

Han Xiao

Zhongqiang Chen

Xiaoguang Liu

Feng Wei

WeiShi Li

Miao Yu

References

Graham

Sugrue

Koski

. Adult degenerative scoliosis. Clin Spine Surg. 2016;29(3):95-107.

Jiang

Wang

. Related classification and selection strategy of degenerative scoliosis. Journal of Shandong University (Health Science).2019,57(05):3-6.

Glattes

Bridwell

Lenke

Kim

Rinella

Edwards

Proximal junctional kyphosis in adult spinal deformity following long instrumented posterior spinal fusion: incidence, outcomes, and risk factor analysis. Spine. 2005, 30(14):1643-1649.

Helgeson

Shah

Newton

, et al. Evaluation of proximal junctional kyphosis in adolescent idiopathic scoliosis following pedicle screw, hook, or hybrid instrumentation. Spine. 2010, 35(2):177-181.

Bridwell

Lenke

Cho

, et al. Proximal junctional kyphosis in primary adult deformity surgery: evaluation of 20 degrees as a critical angle. Neurosurgery. 2013, 72(6):899-906.

Liu

Wang

Yang

Wang

Yang

Ding

. Incidence and risk factors for proximal junctional kyphosis: a meta-analysis. Eur Spine J. 2016; 25: 2376-2383.

Yagi

Akilah

Boachie-Adjei

. Incidence, risk factors and classification of proximal junctional kyphosis: Surgical outcomes review of adult idiopathic scoliosis. Spine. 2011;36: E60-E68.

Slattery

Verma

Classification in brief: SRS-schwab classification of adult spinal deformity. Clin Orthop Relat Res. 2018;476(9):1890-1894.

Yagi

King

Boachie-Adjei

. Incidence, risk factors, and natural course of proximal junctional kyphosis: surgical outcomes review of adult idiopathic scoliosis. minimum 5 years of follow-up. Spine. 2012;37:1479-1489.

10.

Mendoza-Lattes

Ries

Gao

Weinstein

. Proximal junctional kyphosis in adult reconstructive spine surgery results from incomplete restoration of the lumbar lordosis relative to the magnitude of the thoracic kyphosis. Iowa Orthop J. 2011; 31:199-206.

11.

Sebaaly

Riouallon

Obeid

, et al. Proximal junctional kyphosis in adult scoliosis: comparison of four radiological predictor models. Eur Spine J. 2018;27:613-621.

12.

Rose

Bridwell

Lenke

, et al. Role of pelvic incidence, thoracic kyphosis, and patient factors on sagittal plane correction following pedicle subtraction osteotomy. Spine. 2009;34:785-791.

13.

Lee

Kang

, et al.

Proximal junctional kyphosis in degenerative sagittal deformity after under- and overcorrection of lumbar lordosis: does overcorrection of lumbar lordosis instigate PJK?

Spine. 2020;45(15):E933-E942.

14.

Maruo

Inoue

, et al. Predictive factors for proximal junctional kyphosis in long fusions to the sacrum in adult spinal deformity. Spine. 2013;38:E1469-E1476.

15.

Lafage

Schwab

Glassman

, et al. Age-adjusted alignment goals have the potential to reduce PJK. Spine. 2017;42(17):1275-1282.

16.

Glattes

Bridwell

Lenke

Kim

Rinella

Edwards

Proximal junctional kyphosis in adult spinal deformity following long instrumented posterior spinal fusion: incidence, outcomes, and risk factor analysis. Spine.2005, 30 (14): 1643-1649.

17.

Kim

Bridwell

Lenke

Glattes

Rhim

Cheh

. Proximal junctional kyphosis in adult spinal deformity after segmental posterior spinal instrumentation and fusion: Minimum five-year follow-up. Spine. 2008;33:2179-2184.

18.

Bridwell

Lenke

Cho

, et al. Proximal junctional kyphosis in primary adult deformity surgery: evaluation of 20 degrees as a critical angle. Neurosurgery. 2013;72:899-906.

19.

Lee

Kim

Jang

Lee

. Analysis of the incidence and risk factors for the progression of proximal junctional kyphosis following surgical treatment for lumbar degenerative kyphosis: minimum 2-year follow-up. Br J Neurosurg. 2014;28:252-258.

20.

Zhao

. Risk factors and treatment of proximal junctional kyphosis after lumbar degenerative disease. Chinese J Bone Joint Surg. 2020;13(10):793-798.

21.

Kim

Lenke

Shaffrey

Van Alstyne

Skelly

. Proximal junctional kyphosis as a distinct form of adjacent segment pathology after spinal deformity surgery: a systematic review. Spine. 2012;37: S144-S164.

22.

Kim

Bridwell

Lenke

, et al. Patients with proximal junctional kyphosis requiring revision surgery have higher postoperative lumbar lordosis and larger sagittal balance corrections. Spine. 2014;39: E576-E580.

23.

Buell

Chen

Quinn

, et al. Alignment risk factors for proximal junctional kyphosis and the effect of lower thoracic junctional tethers for adult spinal deformity. World Neurosurg. 2019;121:e96-e103.

24.

Togawa

Hasegawa

, et al. The risk of proximal junctional kyphosis decreases in patients with optimal thoracic kyphosis. Spine Deform. 2019;7(5):759-770.

25.

Zhong

Cao

Wang

, et al. Incidence and risk factors for proximal junctional kyphosis in adolescent idiopathic scoliosis after correction surgery: a meta-analysis. World Neurosurg. 2019;125:e326-e335.

26.

McClendon

Smith

Sugrue

, et al. Spinal implant density and postoperative lumbar lordosis as predictors for the development of proximal junctional kyphosis in adult spinal deformity. World Neurosurg. 2016;95:419-424.

27.

Vialle

Levassor

Rillardon

Templier

Skalli

Guigui

. Radiographic analysis of the sagittal alignment and balance of the spine in asymptomatic subjects. J Bone Joint Surg Am. 2005;87(2):260-267.

28.

Lafage

Schwab

Patel

Hawkinson

Farcy

. Pelvic tilt and truncal inclination: two key radiographic parameters in the setting of adults with spinal deformity. Spine. 2009;34: E599-E606.

29.

Zhao

Yang

, et al. Proximal junctional kyphosis in adult spinal deformity: a novel predictive index. Eur Spine J. 2018;27(9):2303-2311.

30.

Moal B

Schwab

Ames

, et al. Radiographic outcomes of adult spinal deformity correction: a critical analysis of variability and failures across deformity patterns. Spine deformity. 2014;2:219-225.

31.

Senteler

Weisse

Snedeker

, et al. Pelvic incidence lumbar lordosis mismatch results in increased segmental joint loads in the unfused and fused lumbar spine. Eur Spine J. 2014;23(7):1384-1393.

32.

Gomez

Kubat

Tovar Castro

, et al. The effect of spinopelvic parameters on the development of proximal junctional kyphosis in early onset: mean 4.5-year follow-up. J Pediatr Orthop. 2020;40:261-266.

33.

Yang

Khalife

Lafage

, et al. What factors predict the risk of proximal junctional failure in the long term, demographic, surgical, or radiographic?: Results from a time-dependent ROC curve. Spine. 2019;44(11):777-784.

34.

Kim

Lenke

Bridwell

, et al. Results of revision surgery for proximal junctional kyphosis following posterior segmental instrumentation: minimum 2-year postrevision follow-up. Spine. 2016;41:E1444-E1452.

35.

Yan

Bao

Qiu

, et al. Mismatch between proximal rod contouring and proximal junctional angle: a predisposed risk fac- tor for proximal junctional kyphosis in degenerative scoliosis. Spine. 2017;42: E280-E287.

36.

Kim

Bridwell

Lenke

Kim

Cho

. Proximal junctional kyphosis in adolescent idiopathic scoliosis following segmental posterior spinal instrumentation and fusion: minimum 5-year follow-up. Spine. 2005;30:2045-2050.