Rod Contour Angle and Postoperative Thoracic Kyphosis: Key Predictors of Proximal Junctional Kyphosis in Pediatric Neuromuscular Scoliosis after Spinopelvic Fusion

Abstract

Study Design

Retrospective cohort study.

Objectives

Proximal junctional kyphosis (PJK) is a postoperative complication in spinal deformity surgery causing pain, functional deterioration, and potential revision surgery, with challenges in the neuromuscular scoliosis (NMS) population. We investigated the incidence, risk factors, and clinical impact of PJK in nonambulatory pediatric NMS patients following spinopelvic fusion.

Methods

Seventy-two NMS patients who underwent spinopelvic fusion from the upper thoracic vertebra to the pelvis were reviewed, with a minimum 2-year follow-up. Radiographic measurements and clinical data were analyzed to identify PJK predictors and evaluate outcomes. PJK was defined as a proximal junctional angle (PJA) ≥10° and an increase of ≥10° from preoperative measurements. Statistical analyses included t-tests, ROC curve analysis, logistic regression, and linear regression.

Results

The incidence of PJK was 25%. Significant predictors included rod contour angle (RCA) (odds ratio [OR]: 1.11, P = 0.04), postoperative T2-T12 kyphosis (OR: 1.04, P = 0.043), and the difference between postoperative T2-T12 kyphosis and RCA (OR: 1.04, P = 0.021). A bidirectional relationship between pre- and postoperative PJA was observed. Lower preoperative PJA was associated with an increased risk of PJK (area under the curve [AUC]: 0.78, P < 0.001). Conversely, high preoperative PJA tended to decrease after surgery. No revision surgeries were performed for PJK.

Conclusion

This study reveals a 25% incidence of PJK in NMS patients undergoing spinopelvic fusion. Preoperative radiographic evaluation warrants particular attention in cases exhibiting decreased PJA values. Furthermore, meticulous intraoperative optimization of PJA, RCA, and thoracic kyphosis appears crucial for minimizing PJK occurrence.

Keywords

proximal junctional kyphosis proximal junctional angle rod contour angle neuromuscular scoliosis spinopelvic fusion

Introduction

Proximal junctional kyphosis (PJK) is characterized by kyphotic alteration above the upper instrumented vertebra (UIV) in spinal deformity surgery.^1,2 It poses significant clinical challenges, with presentations ranging from asymptomatic radiographic changes to pain, functional deficits, and neurological symptoms.^3,4

The etiology of PJK is multifactorial. Its prevalence varies with instrumentation type and level, patient population, and diagnostic criteria.² Neuromuscular scoliosis (NMS) patients frequently sustain scoliosis, pelvic obliquity, and hip deformity.⁵ Spinopelvic fusion is recommended to maintain sitting balance, prevent respiratory complications, and improve the quality of life.^6,7 Long fusion levels and pelvic fixation are known risk factors for PJK.^1,8 Patients with NMS exhibit an elevated risk for PJK due to concurrent osteoporosis, muscle weakness, and neck control issues.⁹ Despite this increased risk, research on PJK in NMS populations remains comparatively scarce, suggesting a significant knowledge gap.^2,10,11

This study investigated PJK in nonambulatory pediatric NMS patients following spinopelvic fusion. Our objectives are to determine the incidence, identify potential risk factors, and evaluate the clinical impact of PJK in this cohort. We aim to develop better strategies to prevent PJK and improve long-term outcomes for these high-risk patients.

Materials and Methods

Study Participants and Surgical Procedures

Following institutional review board approval, we retrospectively reviewed NMS patients who underwent spinal deformity surgery between January 2013 and December 2022. The inclusion criteria included (1) diagnosis of NMS; (2) age at surgery under 18 years old; (3) posterior instrumented fusion from the thoracic vertebra (T1-T4) to the pelvis; and (4) a minimum 2-year follow-up. Exclusion criteria comprised incomplete medical and radiographic data.

At our institution, spinopelvic fusion is specifically indicated for nonambulatory NMS patients. The surgical procedure included posterior spinal fusion from the upper thoracic to the pelvis. We used sacral-alar-iliac screws for spinopelvic fusion.¹² Demographic data, intraoperative reports, and medical records were collected through the electronic health medical record (Epic Systems Corporation).

Radiographic Measurement

Radiographic assessments were conducted preoperatively, postoperatively, and at final follow-up. Major curve size, T2-T12 kyphosis, T5-T12 kyphosis, and T12-S1 lordosis were evaluated using the Cobb method. The Osebold method measured pelvic obliquity (PO).¹³ We also measured the sacral slope (SS), pelvic tilt (PT), pelvic incidence (PI), and sagittal vertical axis (SVA) as previously described.^14-16 Rod contour angle (RCA) was measured as the Cobb angle between the UIV and one vertebra caudal to the UIV (UIV-1) Figure 1.¹⁷

Figure 1.

Proximal Junctional Angle (PJA) and Rod Contour Angle (RCA) Measurements in Neuromuscular Scoliosis (A) Preoperative and (B) postoperative radiographs of a cerebral palsy patient with scoliosis after spinopelvic fusion. PJA is measured between the lower endplate of the upper instrumented vertebra (UIV) and the upper endplate of UIV+2. RCA is the Cobb angle between UIV and UIV-1.

The proximal junctional angle (PJA) was defined as the angle between the lower endplate of the UIV and the upper endplate two levels above the UIV (UIV+2).¹⁶ PJK was defined as a PJA ≥10° and an increase of ≥10° compared to the preoperative measurement.^10,16

Statistical Analysis

We analyzed data using IBM SPSS Statistics (Version 25, SPSS Inc., USA) and R (version 4.3.2; R Core Team (2023), Austria). Categorical variables were analyzed using Chi-square or Fisher’s exact tests. Paired t-tests evaluated differences in paired continuous measures, while independent t-tests determined pairwise group differences when appropriate. Receiver operating characteristic (ROC) curve analysis evaluated the accuracy of factors predicting PJK after spine surgery. The optimal cutoff point was established by calculating the maximum Youden index. Logistic regression analysis examined the association between risk factors and PJK. The logistic regression table provided odds ratios and 95% confidence intervals. Linear regression analysis examined the relationship between postoperative T2-T12 kyphosis and reciprocal changes in RCA. The linear regression model results were presented via the ANOVA table and the reported slope estimate. A P-value less than 0.05 was considered statistically significant.

Results

103 patients were initially reviewed, but 31 were excluded due to insufficient data or follow-up duration. The final cohort comprised 72 patients, with PJK occurring in 18 patients (25%). (Table 1) All patients were nonambulatory preoperatively and remained nonambulatory at final follow-up. No revision surgeries were performed for PJK. In the non-PJK group, seven patients experienced neck pain. There were no significant differences between the PJK and non-PJK groups in sex (P = 0.577), age at surgery (P = 0.583), follow-up duration (P = 0.868), or BMI (P = 0.284). The UIV was mostly located at T2 (47.2 %) or T3 (40.3 %). Hooks were the UIV fixation method in 41 patients (56.9%), while pedicle screws were used in 29 patients (40.3%). No statistically significant differences were found between groups regarding UIV instrumentation use (P = 0.331).

Table 1.

Demographic Result and Patient Characteristics.

Variables	Overall	Non-PJK Group	PJK Group	P-Value^a
Patient number	72	54	18
Male	28 (38.9%)	22 (40.7%)	6 (33.3%)	0.577
Neuromuscular diagnosis
Cerebral palsy	40 (55.6%)	27 (50.0%)	13 (72.2%)	0.273^b
Spinal muscular atrophy	3 (4.2%)	2 (3.7%)	1 (5.6%)
Duchenne muscular dystrophy	5 (6.9%)	5 (9.3%)	0 (0%)
Spinal bifida	2 (2.8%)	2 (3.7%)	0 (0%)
Spinal cord injury	3 (4.2%)	3 (5.6%)	0 (0%)
Rett syndrome	5 (6.9%)	4 (7.4%)	1 (5.6%)
Others	14 (19.4%)	11 (20.4%)	3 (16.7%)
Age at surgery (years)	12.5 (2.30)	12.6 (2.21)	12.2 (2.62)	0.583
Triradiate cartilage closure	43 (59.7%)	33 (61.1%)	10 (55.6%)	0.677
Body height (cm)	138 (14.3)	138 (13.8)	137 (16.1)	0.672
Body weight (kg)	34.7 (12.5)	35.7 (13.5)	31.7 (8.64)	0.239
Body Mass index (kg/m²)	18.8 (9.13)	19.5 (10.4)	16.8 (2.91)	0.284
Follow-up duration (Year)	4.21 (1.93)	4.23 (1.95)	4.14 (1.93)	0.868
Previous growing rod used	10 (13.9%)	8 (14.8%)	2 (11.1%)	0.520
UIV Level
T1	2 (2.8%)	2 (3.7%)	0 (0%)
T2	34 (47.2%)	28 (51.9%)	6 (33.3%)	0.043^c
T3	29 (40.3%)	18 (33.3%)	11 (61.1%)
T4	7 (9.7%)	6 (11.1%)	1 (5.6%)
UIV instrumentation
All pedicle screws	29 (40.3%)	20 (37.0%)	9 (50.0%)	0.331^d
Combine screw and hook	2 (2.8%)	2 (3.7%)	0 (0%)
All hooks	41 (56.9%)	32 (59.3%)	9 (50.0%)
Postoperative neck pain
None	65 (90.3%)	47 (87.0%)	18 (100%)
Mild	6 (8.3%)	6 (11.1%)	0 (0%)
Severe	1 (1.4%)	1 (1.9%)	0 (0%)

The values are given as patient numbers (percentage) or mean (standard deviation).

PJK, proximal junctional kyphosis; UIV, upper instrumented vertebra (UIV).

^aChi-square test for categorical variables; Student's t test for continuous variables.

^bChi-square test: Cerebral Palsy (Yes or No) vs PJK occurrence.

^cChi-square test: UIV level (T2 or T3) vs PJK occurrence.

^dChi-square test: All pedicle screws/instruments (Yes or No) vs PJK occurrence.

Table 2 summarizes the preoperative and postoperative radiographic results. Postoperative radiographs revealed significant deformities correction following spinopelvic fusion. The PJK group demonstrated a significantly higher RCA (P < 0.001) and postoperative PJA (P < 0.001). PJA increased significantly in the PJK group, from 3.00° preoperatively to 25.7° postoperatively (P < 0.001). Greater preoperative (P = 0.002) and postoperative (P < 0.001) T2-T12 kyphosis were observed in the PJK group. The non-PJK group significantly increased T12-S1 lumbar lordosis following surgery (P < 0.001). SVA reduction was observed in both groups postoperatively. The non-PJK group demonstrated a statistically significant decrease in SVA (P = 0.001), while the reduction in the PJK group did not reach statistical significance (P = 0.071).

Table 2.

Result of Radiographic Measurement before and Initial after Surgery.

Variables	Time Points	Overall	Non-PJK Group	PJK Group	P-Value^b
Major curve cobb angle (°)	Preoperative	77.7 (21.2)	78.6 (20.9)	74.8 (22.2)	0.512
	Postoperative	25.7 (15.6)	25.4 (15.6)	26.6 (15.7)	0.778
	P-value^a	<0.001	<0.001	<0.001
Pelvic obliquity (°)	Preoperative	16.4 (11.1)	17.4 (11.8)	13.5 (8.5)	0.206
	Postoperative	5.2 (4.5)	5.6 (4.8)	4.0 (3.6)	0.135
	P-value^a	<0.001	<0.001	<0.001
Coronal T1 pelvic angle (°)	Preoperative	21.1 (13.9)	22.0 (14.4)	18.6 (12.3)	0.370
	Postoperative	6.9 (6.0)	7.5 (6.4)	4.8 (3.9)	0.039
	P-value^a	<0.001	<0.001	<0.001
Rod contour ange (°)	Postoperative	11.0 (7.7)	9.2 (6.3)	16.3 (9.0)	<0.001
Proximal junctional angle (°)	Preoperative	6.7 (7.4)	7.9 (8.1)	3.0 (2.6)	<0.001
	Postoperative	12.9 (11.6)	8.6 (7.2)	25.7 (12.8)	<0.001
	P-value^a	<0.001	0.190	<0.001
T2-T12 thoracic kyphosis (°)	Preoperative	37.4 (21.7)	32.9 (21.1)	50.8 (18.1)	0.002
	Postoperative	41.4 (20.3)	36.8 (18.6)	55.4 (19.1)	<0.001
	P-value^a	0.068	0.135	0.308
T5-T12 thoracic kyphosis (°)	Preoperative	30.8 (20.0)	27.4 (19.4)	40.9 (18.8)	0.011
	Postoperative	27.1 (15.8)	26.2 (16.7)	29.8 (12.4)	0.399
	P-value^a	0.097	0.646	0.01
T12-S1 Lumbar Lordosis (°)	Preoperative	34.2 (28.2)	31.7 (29.2)	41.3 (24.2)	0.214
	Postoperative	48.2 (18.1)	48.4 (17.9)	47.4 (19.0)	0.846
	P-value^a	<0.001	<0.001	0.154
Sacral slope (°)	Preoperative	32.1 (23.0)	31.5 (24.8)	33.7 (17.1)	0.730
	Postoperative	37.2 (13.7)	38.2 (13.2)	34.3 (15.0)	0.303
	P-value^a	0.032	0.025	0.856
Pelvic tilt (°)	Preoperative	20.7 (23.6)	22.3 (23.7)	15.4 (23.2)	0.304
	Postoperative	14.8 (18.3)	16.1 (18.8)	11.4 (16.7)	0.355
	P-value^a	0.005	0.006	0.427
Pelvic incidence (°)	Preoperative	52.1 (22.9)	52.4 (23.2)	51.2 (22.5)	0.850
	Postoperative	51.6 (17.6)	53.4 (17.3)	46.4 (17.8)	0.150
	P-value^a	0.532	0.738	0.521
Sagittal vertical axis (mm)	Preoperative	55.3 (54.4)	58.5 (56.0)	46.3 (49.7)	0.416
	Postoperative	27.9 (40.9)	27.6 (41.0)	28.7 (41.7)	0.921
	P-value^a	<0.001	0.001	0.071

The values are given as mean (standard deviation).

^aPair t test: preoperative vs postoperative value comparison.

^bStudent t test: non-PJK group vs PJK group data comparison.

Table 3 revealed the result of the radiographic assessment at the final follow-up and the difference between the initial postoperative measurements (∆Final-Postop). The ∆Final-Postop for major curve Cobb angles increased slightly in both groups but were not statistically significant. In the non-PJK group, ∆Final-Postop exhibited significant increases in PO (P = 0.005) and decreases in PJA (P = 0.001) and T2-T12 kyphosis (P = 0.01). The PJK group’s ∆Final-Postop for these parameters remained stable (P > 0.05). Both groups significantly decreased ∆Final-Postop T5-T12 kyphosis (non-PJK: P = 0.003; PJK: P = 0.024). Notably, the PJK group showed a significant increase in ∆Final-Postop lumbar lordosis (P = 0.028), while the non-PJK group remained stable (P = 0.961). The ∆Final-Postop for pelvic parameters and SVA remained stable in both groups.

Table 3.

Result of Radiographic Measurement at Final Follow-Up.

Variables	Time Points	Overall	Non-PJK Group	PJK Group	P-Value^b
Major curve cobb angle (°)	Final	27.8 (16.9)	27.3 (17.2)	29.2 (16.4)	0.685
	∆Final-postop	2.2 (9.8)	2.0 (10.2)	2.7 (8.5)	0.798
	P-value^a	0.065	0.159	0.202
Pelvic obliquity (°)	Final	6.8 (6.0)	7.7 (6.4)	4.2 (3.5)	0.006
	∆Final-postop	1.6 (4.8)	2.0 (5.0)	0.2 (4.0)	0.162
	P-value^a	0.008	0.005	0.862
Coronal T1 pelvic angle (°)	Final	8.3 (7.3)	9.4 (8.0)	4.9 (3.0)	0.001
	∆Final-postop	1.5 (4.5)	1.9 (4.9)	0.1 (2.5)	0.141
	P-value^a	0.007	0.006	0.851
Proximal junctional angle (°)	Final	11.8 (16.0)	6.1 (6.9)	28.7 (22.7)	0.001
	∆Final-postop	−1.1 (15.0)	−2.5 (8.5)	3.1 (26.3)	0.393
	P-value^a	0.542	0.001	0.628
T2-T12 thoracic kyphosis (°)	Final	38.3 (21.7)	33.1 (18.4)	54.2 (23.5)	<0.001
	∆Final-postop	−3.1 (11.7)	−3.7 (10.2)	−1.3 (15.5)	0.449
	P-value^a	0.027	0.01	0.731
T5-T12 thoracic kyphosis (°)	Final	22.1 (12.9)	21.0 (13.8)	25.4 (9.3)	0.207
	∆Final-postop	−5.0 (11.3)	−5.2 (12.4)	−4.4 (7.5)	0.746
	P-value^a	0.027	0.003	0.024
T12-S1 Lumbar Lordosis (°)	Final	50.0 (16.4)	48.1 (16.2)	55.6 (16.3)	0.098
	∆Final-postop	2.0 (14.6)	−0.1 (14.2)	8.1 (14.3)	0.039
	P-value^a	0.252	0.961	0.028
Sacral slope (°)	Final	38.5 (12.7)	38.5 (12.1)	38.7 (14.8)	0.939
	∆Final-postop	1.5 (11.5)	0.4 (10.9)	4.4 (12.9)	0.213
	P-value^a	0.296	0.780	0.168
Pelvic tilt (°)	Final	15.2 (19.6)	15.7 (20.0)	13.8 (19.1)	0.724
	∆Final-postop	1.2 (15.4)	0.5 (15.5)	3.4 (15.5)	0.513
	P-value^a	0.526	0.843	0.386
Pelvic incidence (°)	Final	53.1 (18.0)	53.1 (18.0)	52.9 (18.3)	0.972
	∆Final-postop	1.9 (16.4)	0.3 (16.4)	6.2 (16.2)	0.211
	P-value^a	0.365	0.906	0.136
Sagittal vertical axis (mm)	Final	21.8 (57.0)	24.2 (59.3)	14.9 (50.8)	0.579
	∆Final-postop	−11.5 (62.5)	−9.2 (63.5)	−18.1 (61.2)	0.628
	P-value^a	0.151	0.325	0.256

The values are given as mean (standard deviation).

∆Final-Postop = Final follow-up value – Postoperative value.

^aPair t test: postoperative vs final follow-up value comparison.

^bStudent t test: non-PJK group vs PJK group data comparison.

Logistic regression analysis estimated the probability of PJK at the final follow-up (Table 4). The variables for the multivariate models were chosen based on multicollinearity and clinical relevance. Multivariate models incorporated sex, age at surgery, and statistically significant variables from univariate analysis. Not all variables with a P-value <0.05 in the univariate analysis were included to avoid potential confounding effects. For example, the relationship between RCA and ΔT2K-RCA postop shows inherent correlations between variables, which could lead to redundancy and obscure the true effects of other predictors. In the multivariate model adjusted for age and sex, RCA (OR: 1.113, P = 0.040) and postoperative T2-T12 kyphosis (OR: 1.039, P = 0.043) were significant predictors for PJK.

Table 4.

Logistic Regression Analysis for Predicting Proximal Junctional Kyphosis at Final Follow-Up.

Variable	Univariate Model		Multivariate Model*
Variable	Odds Ratio (95% CI)	P-Value	Odds Ratio (95% CI)	P-Value
Age at surgery (years)	0.935 (0.739-1.184)	0.578	1.002 (0.773-1.297)	0.991
Triradiate cartilage closure	0.795 (0.270-2.339)	0.678	-	-
Sex (male vs Female)	1.375 (0.449-4.215)	0.577	1.644 (0.438-6.163)	0.461
Body Mass index (kg/m2)	0.909 (0.785-1.054)	0.208	-	-
Preoperative major curve cobb angle (°)	0.991 (0.966-1.071)	0.506	-	-
Preoperative pelvic obliquity (°)	0.961 (0.904-1.022)	0.204	-	-
Preoperative coronal T1 pelvic angle (°)	0.980 (0.939-1.023)	0.367	-	-
Preoperative T2-T12 kyphosis (°)	1.044 (1.013-1.075)	0.005	-	-
Preoperative T5-T12 kyphosis (°)	1.036 (1.006-1.067)	0.017	-	-
Rod contour angle (°)	1.149 (1.048-1.260)	0.003	1.113 (1.005-1.232)	0.040
Postoperative T2-T12 kyphosis (°)	1.052 (1.019-1.086)	0.002	1.039 (1.001-1.078)	0.043
Postoperative T5-T12 kyphosis (°)	1.015 (0.981-1.049)	0.394	-	-
∆T2K-RCA postop (°)	1.040 (1.007-1.075)	0.018	-	-
∆T2K preop-postop (°)	1.003 (0.974-1.032)	0.864	-	-
∆T5K preop-postop (°)	0.970 (0.941-1.000)	0.052	-	-

∆T2K-RCA postop, the difference between postoperative T2-T12 kyphosis and rod contour angle.

∆T2K preop-postop, the difference between preoperative and postoperative T2-T12 kyphosis.

∆T5K preop-postop, the difference between preoperative and postoperative T5-T12 kyphosis.

*Multivariate models included sex, age at surgery, and select variables significant in univariate analysis. Selection criteria considered multicollinearity and clinical relevance, not solely P-values <0.05, to avoid confounding effect.

Table 5 and Figure 2 revealed the radiographic parameters demonstrating significant predictive value for PJK, with AUC values exceeding 0.7. ROC curve analysis revealed that RCA (AUC: 0.753, P = 0.001) and postoperative T2-T12 kyphosis (AUC: 0.767, P = 0.001) demonstrated fair predictive ability for PJK, with optimal cutoff points of 13.5° and 42.5°, respectively. The difference between postoperative T2-T12 kyphosis and RCA (∆T2K-RCA postop) also demonstrated fair predictive ability (AUC: 0.731, P = 0.003), with an optimal cutoff point of 30°. Preoperative PJA exhibited the highest predictive ability among the parameters (P < 0.001), with lower preoperative PJA values associated with increased PJK risk. To enhance interpretability, the original AUC of 0.223 for preoperative PJA was inverted to 0.777. Other effective predictors included postoperative PJA and the difference between postoperative PJA and RCA (∆PJA-RCA postop) and preoperative and postoperative PJA (∆PJA preop-postop). Caution is warranted in interpreting the strong predictive ability of PJA-related parameters, as the definition of PJK was inherently based on changes in PJA.

Table 5.

ROC Curve Analysis of Predicting Proximal Junctional Kyphosis at Final Follow-Up.

Variable	Area Under Curve	95% CI	P-Value	Optimal Cutoff Point^a	Sensitivity	Specificity
Rod contour angle (RCA)	0.753	0.625-0.880	0.001	>13.5°	66.7%	74.1%
Postoperative T2-T12 kyphosis	0.767	0.643-0.891	0.001	>42.5°	77.8%	70.4%
∆T2K-RCA postop	0.731	0.601-0.862	0.003	>30°	77.8%	66.7%
Preoperative PJA	0.777*	0.664-0.891	<0.001	<4.5°	66.7%	77.8%
Postoperative PJA	0.872	0.760-0.985	<0.001	>16°	83.3%	85.2%
∆PJA-RCA postop	0.707	0.533-0.882	0.009	>9.5°	50.0%	94.9%
∆PJA preop-postop	0.907	0.822-0.992	<0.001	>14°	83.3%	90.7%

*The original AUC of 0.223 for preoperative PJA was inverted to 0.777 for consistent interpretation.

CI, confidence interval; PJA, proximal junctional kyphosis.

∆T2K-RCA postop, the difference between postoperative T2-T12 kyphosis and RCA.

∆PJA-RCA postop, the difference between postoperative PJA and RCA.

∆PJA preop-postop, the difference between preoperative and postoperative PJA.

^aThe optimal cutoff point was established by calculating the maximum Youden index.

Figure 2.

ROC Curve Analysis of Radiographic Parameters for Assessing Proximal Junctional Kyphosis (PJK) (A) Comparison of individual radiographic parameters, including rod contouring angle (RCA), postoperative proximal junctional angle (PJA postop), preoperative PJA (PJA preop), postoperative T2-T12 kyphosis (T2-T12 kyphosis postop). (B) Comparison of derived radiographic parameters, including the difference between postoperative PJA and RCA (∆PJA-RCA postop), preoperative and postoperative PJA (∆PJA preop-postop), and between postoperative T2-T12 kyphosis and RCA (∆T2K-RCA postop).

Scatter plots illustrated the distinct patterns between PJK-positive and PJK-negative cases. Reference lines indicated potential thresholds for each parameter, as determined by ROC curve analysis (Table 5). Our findings revealed a bidirectional impact of surgery on PJA. (Figure 3) While surgery generally decreased PJA in patients with higher preoperative measurements, patients with lower preoperative PJA, specifically those with values <4.5°, experienced a marked increase postoperatively. Patients with preoperative PJA <4.5° exhibited significantly higher rates of PJK (P = 0.001), greater postoperative T2-T12 kyphosis (P = 0.04), and increased T2-T12 kyphosis (∆T2K preop-postop, P = 0.04) after surgery. (Table 6)

Figure 3.

Relationship between Preoperative (Preop) and Postoperative (Postop) Proximal Junctional Angle (PJA). This scatter plot illustrates preoperative (preop) vs postoperative (postop) PJA for PJK-negative (blue) and PJK-positive (red) cases. The horizontal dashed line at 4.5° represents the significant preop PJA cutoff point determined by ROC curve analysis. The light blue shaded ellipse encompasses patients with large preop PJA who experienced effective PJA reduction after surgery and did not develop PJK at the final follow-up. Conversely, the light red shaded ellipse includes patients with low preop PJA who developed PJK at the final follow-up.

Table 6.

Kyphosis Measurements in Patients With Different Preoperative PJA.

Variables	Preoperative PJA <4.5°	Preoperative PJA ≥4.5°	P Value*
Proximal junction kyphosis (Yes: No)	14:18	4:36	0.001
Preoperative T2-T12 kyphosis	37.9° (22.7°)	37.0° (21.1°)	0.862
Postoperative T2-T12 kyphosis	46.9° (23.3°)	37.1° (16.5°)	0.040
∆T2K preop-postop	9.0° (15.5°)	0.05° (19.7°)	0.040

*Fisher’s exact test for categorical variables; Student's t test for continuous variables.

PJA, Proximal junctional angle.

∆T2K preop-postop, difference between preoperative and postoperative T2-T12 kyphosis.

Figure 4 illustrated the association between RCA and postoperative PJA. The pink shaded area (∆PJA-RCA <9.5°) predominantly contains PJK-positive cases, with 75% (9/12) falling within this region. Cases with postoperative PJA ≥30° were all PJK-positive. Figure 5 showed the association between RCA and postoperative T2-T12 kyphosis. Most PJK-negative cases had RCA <13.5° and postoperative T2-T12 kyphosis <42.5°, possibly suggesting these thresholds may predict PJK.

Figure 4.

Relationship between Rod Contour Angle (RCA) and Postoperative Proximal Junctional Angle (Postop PJA) This scatter plot revealed the relationship between RCA and postop PJA for proximal junctional kyphosis (PJK)-negative (blue) and PJK-positive (red) cases. The pink shaded area indicates where the difference between postop PJA and RCA (ΔPJA-RCA postop) >9.5° was the significant cutoff point from our ROC curve analysis. The vertical dashed line at 16° represents the significant postoperative (postop) PJA cutoff point determined by ROC curve analysis.

Figure 5.

Relationship between Rod Contour Angle (RCA) and Postoperative T2-T12 Kyphosis (Postop T2-T12 Kyphosis) This scatter plot revealed the relationship between RCA and postop T2-T12 kyphosis for proximal junctional kyphosis (PJK)-negative (blue) and PJK-positive (red) cases. Linear trend lines are shown for each group. Linear regression analysis revealed a significant interaction between RCA and postop T2-T12 kyphosis in the PJK-positive group (red line, P = 0.028). Dashed lines indicate ROC-derived critical values: RCA (13.5°, horizontal) and postoperative T2-T12 kyphosis (42.5°, vertical).

The ANOVA table of the linear regression model revealed a significant interaction effect between RCA and postoperative T2-T12 kyphosis in the PJK-positive group (F = 5.034, P = 0.028). (Figure 5, Table 7) In PJK-positive cases, a 1-degree increase in postoperative T2-T12 kyphosis corresponded to an average 0.32-degree increase in RCA (t = 2.24, P = 0.028)

Table 7.

ANOVA Table: Factors Affecting Rod Contour Angle (RCA).

Variables	Sum of Squares	df	F-Stat	P Value
Postop T2-T12 kyphosis (°)	265.370	1	7.003	0.01
PJK-negative vs PJK-positive group	63.878	1	1.686	0.199
Postop T2-T12 kyphosis (°) of PJK-positive group	190.772	1	5.034	0.028

ANOVA: Analysis of variance; df: Degrees of freedom; Postop: Postoperative PJK: Proximal junctional kyphosis; F stat: F-statistic.

Full Regression Formula: RCA(°) = 4.76 + 0.12 K - 6.4P + 0.20(K × P).

Where.

K = Postop T2-T12 Kyphosis (°).

P = PJK status (0 if Negative, 1 if Positive).

Simplified formulas.

For PJK-Negative group: RCA(°) = 4.76 + 0.12 K.

For PJK-Positive group: RCA(°) = −1.64 + 0.32 K.

Discussion

Our study found that 25% of NMS patients developed PJK following spinopelvic fusion. We observed a bidirectional relationship between pre- and postoperative PJA relationships. Surgery generally decreased PJA in patients with higher preoperative values. However, many PJK patients had preoperative PJA <4.5°. Significant predictors of PJK included RCA, postoperative T2-T12 kyphosis, and ∆T2K-RCA postop. The complex interplay between these parameters offers potential targets for preventing PJK in NMS surgical planning.

Toll et al. reported a 27% incidence of PJK in NMS using the definition of PJK as a >10° increase.² Despite variations in PJK definition, neuromuscular condition distribution, ambulatory status, and pelvic fixation rates, we observed a similar PJK prevalence. Our findings corroborate their observation regarding the PJK incidence and risk factors for PJK, including lower preoperative PJA and increased lumbar lordosis at the final follow-up in the PJK group. Figure 3 demonstrated that patients with higher preoperative PJA typically experienced reduction post-surgery, while those with PJK often presented with preoperative PJA <4.5°. We propose three potential explanations for the association between lower preoperative PJA and increased PJK risk. First, a lower initial PJA could more easily reach the 10° increase criterion for PJK diagnosis. Second, our patients with preoperative PJA <4.5°exhibited significantly higher postoperative T2-T12 kyphosis and greater kyphosis increases. (Table 6) They simultaneously presented two conditions associated with PJK risk. The interplay between these factors warrants further study. Third, poor cervical muscle control might cause unintended neck hyperextension during radiographic positioning, resulting in artificially lower preoperative PJA measurements. The increased lumbar lordosis observed in the PJK group was consistent with Toll et al's findings. The underlying mechanisms for this phenomenon remain unclear, but we hypothesize that it may indicate compromised bone quality, potentially affecting screw fixation. Further studies are needed to elucidate these factors in PJK etiology.

Larger postoperative PJA is a significant risk factor for PJK in adolescent idiopathic scoliosis (AIS) and adult spinal deformity (ASD) surgery.^18-20 Sakuma et al. reported that postoperative PJA >15.5° and changes in postoperative PJA >14.5° predicted PJK in ASD surgery.²¹ Our findings corroborate these results, with significantly higher PJK risk in patients with postoperative PJA >16°. (Table 5) While some patients’ PJA decreased at final follow-up, this was uncommon in those with postoperative PJA >16°. (Table 3) Given the increased risk when postoperative PJA exceeds 16°, intraoperative strategies to mitigate PJK risk are crucial. Beyond extending fusion levels, our findings suggest adjusting parameters like RCA and postoperative thoracic kyphosis may reduce risk, especially with high PJA values.

RCA and its relationship with postoperative PJA are significant predictors of PJK across various patient populations.^{17,19,20,22-25} Previous studies found that ∆PJA-RCA post >5° predicted PJK in AIS and early-onset scoliosis (EOS)^{20, 26}. We observed that RCA >13.5° and ∆PJA-RCA postop >9.5° predict PJK in the NMS population. These observations align with recent studies suggesting that higher RCA and mismatch between postoperative PJA and RCA increased PJK risk.^19,23-25 ROC curve analysis revealed that PJA-RCA postop >9.5° as a predictor for PJK with high specificity. (Table 5) However, the scatter plot revealed a more complex relationship between these radiographic parameters. (Figure 4) Adjusting RCA is primarily effective when postoperative PJA values are low. Specifically, when postoperative PJA is below 30°, maintaining ΔPJA-RCA postop under 9.5° appears to increase the likelihood of PJK-negative outcomes. However, for higher postoperative PJA values (>30°), modifying RCA seems to have limited impact on preventing PJK-positive results. This highlights the limitations of single-factor thresholds and underscores the need for multivariable analysis in assessing PJK risk in NMS patients.

Our findings did not indicate that the preoperative and postoperative thoracic kyphosis change was a significant PJK predictive factor as previous literature described.^19,24,26,27 Instead, we observed that postoperative thoracic kyphosis and ∆T2K-RCA postop were associated with PJK. Figure 5 demonstrates the interaction between RCA and postoperative T2-T12 kyphosis in predicting PJK. Non-PJK patients predominantly clustered in the lower left quadrant (RCA <13.5° and postoperative T2-T12 kyphosis <42.5°). The steeper regression line slope for the PJK group indicated a stronger positive correlation between postoperative T2-T12 kyphosis and RCA in PJK patients compared to non-PJK patients. Future studies must validate these findings and explore optimizing the RCA and postoperative thoracic kyphosis balance.

Spinopelvic fusion and long-level fusion are known as PJK risk factors.^17,28 In our study, while the absence of a control group precluded direct comparisons, we observed that the PJK rate falls between EOS and AIS, two distinct pediatric spinal deformities.^18,26,29,30 Notably, no revision surgery was performed, possibly due to reduced functional demands, complex comorbidities, and wheelchair adjustments compensating for PJK deformities. However, this should not diminish PJK’s significance in NMS. Prevention remains crucial to avoid additional surgeries. Considering these patients’ unique clinical characteristics and functional requirements, reevaluating the PJK definition for this population may be necessary.

Our investigation of risk factors was informed by both AIS and ASD literature, as these populations share important characteristics with our NMS cohort. While AIS patients are similar in age to our population, ASD patients often share compromised bone quality issues. Regarding instrumentation, while transverse process hooks at UIV have been associated with decreased PJK risk in AIS, we found no significant differences between pedicle screws and hooks in our NMS cohort (P = 0.331).³¹ This warrants careful interpretation due to selection bias in our surgical approach, where pedicle screws were preferentially used when anatomically feasible. While other fixation techniques such as transitional rods, hybrid constructs, and sublaminar bands have shown promise but were not evaluated in our cohort, their effectiveness in NMS patients requires further investigation due to several patient-specific factors.^32,33 Age-related effects differed from AIS literature, where younger age is a known risk factor for PJK development. In our NMS cohort, neither age nor skeletal maturity (triradiate cartilage closure) showed significant association with PJK development (P = 0.583 and P = 0.746, respectively). Additionally, disease-specific growth patterns in our heterogeneous cohort may have confounded our ability to detect age-related effects on PJK development, as these conditions can have markedly different impacts on skeletal growth and maturation.

Osteoporosis, a well-established risk factor in ASD, that warrants further investigation in NMS patients. We did not routinely perform DXA scans for patients without fracture history, as there is no consensus on bone quality assessment protocols in NMS patients undergoing spinal surgery.³⁴ Future prospective studies are needed to elucidate the unique considerations in NMS patients. These should systematically assess bone quality, spasticity levels, and detailed postoperative protocols - factors that may significantly influence surgical outcomes in this population.

Our study had several limitations. (1) The generalizability may be limited as a single-center retrospective study. (2) Our radiographic analysis was confined to preoperative, initial postoperative, and final follow-up. Intraoperative and longitudinal postoperative assessments were lacking. Comprehensive temporal evaluations could provide critical insights into PJK’s dynamic progression and improve surgical decision-making. (3) We did not assess patients’ spasticity or motor control capabilities of the head, neck, and upper extremities—crucial neuromuscular parameters influencing postoperative functional outcomes. (4) Some patients were excluded due to inadequate imaging quality in the cervicothoracic junction region, attributed to poor bone quality, positioning difficulties during radiographic examination, and severe spinal deformities. Advanced imaging techniques and patient-specific protocols are essential to improve future investigations' reliability. Despite these limitations, our strengths include a relatively large sample size for NMS, extended follow-up duration, and a homogenous patient cohort with similar surgical procedures and fusion levels.

This study reveals a 25% incidence of PJK in NMS patients undergoing spinopelvic fusion, highlighting the need for effective preventive strategies. Our findings highlight the importance of careful preoperative assessment, particularly for patients with lower preoperative PJA values. Intraoperative optimization of PJA, RCA, and thoracic kyphosis emerged as key factors in mitigating PJK risk. The complex interplay between these parameters necessitates further research to refine surgical goals, ultimately reducing PJK incidence and enhancing outcomes in this vulnerable population.

Footnotes

Author Contributions

PCS: Writing – original draft, Investigation, Data curation, Conceptualization. CK: Statistical analysis, Visualization, Software. TS: Project administration, Conceptualization. NHM: Supervision, Conceptualization. MAE: Supervision, Conceptualization. Review and editing

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.

Ethical Statement

ORCID iD

Pochih Shen

References

Dubousset

Diebo

. Proximal junctional kyphosis in modern spine surgery: why is it so common? Spine. Surg Relat Res. 2023;7:120-128. doi:10.22603/ssrr.2022-0100

Toll

Gandhi

Amanullah

, et al. Risk factors for proximal junctional kyphosis following surgical deformity correction in pediatric neuromuscular scoliosis. Spine. 2021;46:169-174. doi:10.1097/brs.0000000000003755

Smith

Shaffrey

Ames

Lenke

. Treatment of adult thoracolumbar spinal deformity: past, present, and future. J Neurosurg Spine. 2019;30:551-567. doi:10.3171/2019.1.SPINE181494

Lee

Park

. Proximal junctional kyphosis: diagnosis, pathogenesis, and treatment. Asian Spine J. 2016;10:593-600. doi:10.4184/asj.2016.10.3.593

Patel

Shapiro

. Simultaneous progression patterns of scoliosis, pelvic obliquity, and hip subluxation/dislocation in non-ambulatory neuromuscular patients: an approach to deformity documentation. J Child Orthop. 2015;9:345-356. doi:10.1007/s11832-015-0683-7

Geiger

Eberl

Wirries

Forth

Hammad

. The indication of fusion to the pelvis in neuromuscular scoliosis is based on the underlying disease rather than on pelvic obliquity. Eur Spine J. 2023;32:4063-4072. doi:10.1007/s00586-023-07943-7

Loughenbury

Tsirikos

. Current concepts in the treatment of neuromuscular scoliosis: clinical assessment, treatment options, and surgical outcomes. Bone Jt Open. 2022;3:85-92. doi:10.1302/2633-1462.31.BJO-2021-0178.R1

Murata

Fujibayashi

Otsuki

Shimizu

Matsuda

. Risk factors for proximal junctional kyphosis in fusions from the sacrum to L1 or L2 for adult spinal deformity. Spine Surg Relat Res. 2022;6:395-401. doi:10.22603/ssrr.2021-0157

Rezende

Cardoso

Leonel

, et al. Bone mineral density evaluation among patients with neuromuscular scoliosis secondary to cerebral palsy. Rev Bras Ortop. 2015;50:68-71. doi:10.1016/j.rboe.2014.12.003

10.

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. doi:10.1097/01.brs.0000179084.45839.ad. (Phila Pa 1976).

11.

Desai

Sayama

Vener

, et al. The feasibility and safety of using sublaminar polyester bands in hybrid spinal constructs in children and transitional adults for neuromuscular scoliosis. J Neurosurg Pediatr. 2015;15:328-337. doi:10.3171/2014.9.PEDS1468

12.

Suresh

Ikwuezunma

Margalit

Sponseller

. Spinal fusion with sacral alar iliac pelvic fixation in severe neuromuscular scoliosis. JBJS Essent Surg Tech. 2021;11:00060. doi:10.2106/JBJS.ST.20.00060

13.

Yen

Gartenberg

Cho

. Pelvic obliquity associated with neuromuscular scoliosis in cerebral palsy: cause and treatment. Spine Deform. 2021;9:1259-1265. doi:10.1007/s43390-021-00346-y

14.

Legaye

Duval-Beaupere

Hecquet

Marty

. Pelvic incidence: a fundamental pelvic parameter for three-dimensional regulation of spinal sagittal curves. Eur Spine J. 1998;7:99-103. doi:10.1007/s005860050038

15.

Cho

Suk

Park

, et al. Risk factors of sagittal decompensation after long posterior instrumentation and fusion for degenerative lumbar scoliosis. Spine. 2010;35:1595-1601. doi:10.1097/BRS.0b013e3181bdad89

16.

Chen

Qiu

, et al. Proximal junctional kyphosis after posterior spinal instrumentation and fusion in young children with congenital scoliosis: a preliminary report on its incidence and risk factors. Spine. 2017;42:E1197-e1203. doi:10.1097/brs.0000000000002109

17.

Lonner

Ren

Newton

, et al. Risk factors of proximal junctional kyphosis in adolescent idiopathic scoliosis-the pelvis and other considerations. Spine Deform. 2017;5:181-188. doi:10.1016/j.jspd.2016.10.003

18.

Wei

Xing

Duan

. Incidence and risk factors of proximal junctional kyphosis in adolescent idiopathic scoliosis after correction surgery: a meta-analysis and systematic review. J Orthop Surg Res. 2024;19:217. doi:10.1186/s13018-024-04638-7

19.

Wang

Yang

Luo

Xia

. Large difference between proximal junctional angle and rod contouring angle is a risk factor for proximal junctional kyphosis. World Neurosurg. 2020;136:e683-e689. doi:10.1016/j.wneu.2020.01.136

20.

Boeckenfoerde

Schulze Boevingloh

Gosheger

Bockholt

Lampe

Lange

. Risk factors of proximal junctional kyphosis in adolescent idiopathic scoliosis-the spinous processes and proximal rod contouring. J Clin Med. 2022;11:6098. doi:10.3390/jcm11206098

21.

Sakuma

Kotani

Akazawa

, et al. Incidence, risk factors, and prevention strategy for proximal junctional kyphosis in adult spinal deformity surgery. Spine Surg Relat Res. 2021;5:75-80. doi:10.22603/ssrr.2020-0093

22.

Grelat

Sun

Qiu

. Under-contouring of rods: a potential risk factor for proximal junctional kyphosis after posterior correction of Scheuermann kyphosis. J Neurosurg Spine. 2020;33:830-837. doi:10.3171/2020.5.Spine20229

23.

Ishihara

Taniguchi

Adachi

, et al. Rod contour and overcorrection are risk factors of proximal junctional kyphosis after adult spinal deformity correction surgery. Eur Spine J. 2021;30:1208-1214. doi:10.1007/s00586-021-06761-z

24.

Yan

Bao

Qiu

, et al. Mismatch between proximal rod contouring and proximal junctional angle: a predisposed risk factor for proximal junctional kyphosis in degenerative scoliosis. Spine. 2017;42:E280-e287. doi:10.1097/brs.0000000000001883

25.

Yang

Wang

, et al. Unmatched rod contouring at the proximal end predisposes to occurrence of junctional kyphosis in early-onset scoliosis patients undergoing traditional growing rods treatment. BMC Muscoskelet Disord. 2022;23:624. doi:10.1186/s12891-022-05564-7

26.

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. doi:10.1016/j.wneu.2019.01.072

27.

Lee

Betz

Clements

Huss

. Proximal kyphosis after posterior spinal fusion in patients with idiopathic scoliosis. Spine. 1999;24:795-799. doi:10.1097/00007632-199904150-00011

28.

Han

Ren

. Risk factors for proximal junctional kyphosis in adult spinal deformity after correction surgery: a systematic review and meta-analysis. Acta Orthop Traumatol Turcica. 2022;56:158-165. doi:10.5152/j.aott.2022.21255

29.

Yang

Zhou

, et al.

Relook into the risk factors of proximal junctional kyphosis in early onset scoliosis patients: does the location of upper instrumented vertebra in relation to the sagittal apex matter?

Orthop Surg. 2022;14:1695-1702. doi:10.1111/os.13380

30.

Watanabe

Uno

Suzuki

, et al. Risk factors for proximal junctional kyphosis associated with dual-rod growing-rod surgery for early-onset scoliosis. Clin Spine Surg. 2016;29:E428-433. doi:10.1097/bsd.0000000000000127

31.

Erkilinc

Coathup

Liska

Lovevoy

. Can placement of hook at the upper instrumented level decrease the proximal junctional kyphosis risk in adolescent idiopathic scoliosis? Eur Spine J. 2023;32:3113-3117. doi:10.1007/s00586-023-07803-4

32.

Arora

Sharfman

Clark

Theologis

. Proximal junctional kyphosis and failure: strategies for prevention. Neurosurg Clin. 2023;34:573-584. doi:10.1016/j.nec.2023.06.004

33.

Viswanathan

Minnema

Viljoen

Farhadi

. Sublaminar banding as an adjunct to pedicle screw-rod constructs: a review and technical note on novel hybrid constructs in spinal deformity surgery. J Neurosurg Spine. 2019;30:807-813. doi:10.3171/2018.11.SPINE181154

34.

Antoniou

Masouros

Papadopoulos

, et al. A scoping review of the recent clinical practice regarding the evaluation of bone mineral density in children and adolescents with neuromuscular diseases. Medicina (Kaunas). 2023;59:312. doi:10.3390/medicina59020312