Abstract
Background
Given the different biomechanical properties and surgical techniques between L5-S1 and at ≥ L4-5 levels, different factors might contribute to rod fracture (RF) development between the L5-S1 and ≥ L4-5 levels. Recognizing segment-specific risk factors can be helpful when planning surgery. Therefore, this study aims to analyze the risk factors for RF separately for L5-S1 and ≥ L4-5 levels after adult spinal deformity surgery.
Methods
This study analyzed consecutive patients who underwent ≥ 5-level fusion including the sacrum or pelvis and were followed up for ≥ 2 years. RF was assessed at the segment level between the L1-2 and L5-S1 levels. Multivariate logistic regression analyses were performed to identify independent risk factors for RF separately for L5-S1 and ≥ L4-5 levels.
Results
A total of 318 patients were included in the study (mean age, 69.3 years; female, 88.4%; mean fusion length, 7.3 levels). A total of 1,329 segments were evaluated (247 for L5-S1 and 1,082 for ≥ L4-5 levels). The mean follow-up duration was 47.4 months. Multivariate analysis revealed that increased fusion length (odds ratio [OR], 1.251; p = 0.021) and inadequate pelvic tilt (PT) correction (OR, 4.781; p = 0.005) were significant risk factors for RF at L5-S1, while non-use of teriparatide (OR, 3.460; p = 0.021), anterior column realignment (ACR) (vs. posterior lumbar interbody fusion [PLIF]: OR, 5.424; p = 0.003), posterior fusion (PF) (vs. PLIF: OR, 3.276; p = 0.029), pedicle subtraction osteotomy (PSO; OR, 3.803; p = 0.002), and dual rod construct (vs. 4-rod configuration: OR, 2.873; p = 0.044) were significant risk factors for RF at or above L4-5.
Conclusions
This study demonstrated that the risk factors for RF are different between L5-S1 and ≥ L4-5 levels. Increased fusion length and poor postoperative correction of PT were significant risk factors for RF at the L5-S1 level, while non-use of teriparatide, ACR, and PF, performance of PSO, and dual rod construct significantly increased the risk of RF at ≥ L4-5 levels. Understanding the differing risk factors can help establish suitable preventive strategies tailored to each patient.
Keywords: Risk factors, Pseudoarthrosis, Rod fracture, Adult spinal deformity
Adult spinal deformity (ASD) is a complex condition, which leads to debilitating effects on both the physical and mental health of patients.1) Previous research has reported the effectiveness of surgical treatment for ASD and its superiority over conservative treatment.2) However, since patients are typically older and long-segment fixation is commonly required, implant-related complications such as rod fracture (RF) can develop. Despite extensive research regarding RF, it remains a challenging issue in ASD surgery with its rates reported up to 30%.3,4,5,6,7,8,9,10,11,12,13) Therefore, a precise understanding of the risk factors for RF is essential for improving surgical outcomes.
To date, numerous risk factors have been demonstrated in the literature.6,7,8,9,10,11,12) However, it remains unclear whether these factors are equally applicable to both the L5-S1 and ≥ L4-5 levels. Given the different biomechanical properties and surgical techniques between the L5-S1 and ≥ L4-5 levels,14,15,16) different factors might contribute to RF development between the L5-S1 and ≥ L4-5 levels. Recognizing segment-specific risk factors can be helpful when planning surgery. For instance, when the L5-S1 level is already solidly fused due to previous surgery, attention should be directed toward the risk factors relevant to ≥ L4-5 levels.
Given that surgical procedures for each operated segment are highly variable depending on the patient’s deformity status and previous fusion surgery, identifying distinct risk factors between the L5-S1 and ≥ L4-5 levels could enable the implementation of tailored strategies to prevent RF more effectively during surgical planning. In this study, we aimed to analyze the risk factors for RF separately for L5-S1 and ≥ L4-5 levels in patients undergoing long-segment fusion for ASD. Considering that fusion techniques may vary by segment, RF was assessed at the segment level.
METHODS
This study was approved by Institutional Review Board of Samsung Medical Center (IRB No. 2024-07-144). The requirement for informed consent was waived due to its retrospective nature.
Study Cohort
We retrospectively analyzed patient records from a prospective ASD database at our academic institution. Consecutive patients who underwent corrective surgery for degenerative ASD between 2012 and 2022 were enrolled in this study. Patients were included if they met the following criteria: (1) ASD, defined as pelvic incidence minus lumbar lordosis (PI – LL) mismatch ≥ 10°, pelvic tilt (PT) ≥ 25°, thoracic kyphosis ≥ 60°, sagittal vertical axis (SVA) ≥ 5 cm, or a coronal Cobb angle ≥ 30°; (2) ≥ 5-level fusion including the sacrum or pelvis; and (3) a minimum follow-up period of 2 years. RF cases that developed within 2 years were included in the study, even if the follow-up period had not been completed. Patients were excluded if they had neuromuscular, inflammatory, or other non-degenerative pathological conditions; or if they had incomplete radiographic data. Surgery was performed by 3 attending surgeons (DHK, JSP, and SJP). For interbody fusion, polyetheretherketone cages were used for anterior lumbar interbody fusion (ALIF), lateral lumbar interbody fusion (LLIF), and anterior column realignment (ACR), while titanium cages were used for posterior lumbar interbody fusion (PLIF). Pelvic fixation was routinely performed using iliac screws except for cases with solid lumbosacral fusion due to previous fusion surgery. Titanium rods were used in all cases.
Segments for Evaluation
The primary outcome of this study was the development of RF evaluated at the segment level. RFs above T12-L1 (n = 2) were excluded to avoid compromising statistical power, so only the segments from L1-2 to L5-S1 were included in the evaluation. While all naive segments were included in the evaluation, the segments with solid fusion due to previous surgery were excluded from the analysis. However, even if a segment had undergone prior fusion surgery, it was included in the analysis if one of the following procedures had been performed: (1) segments with prior posterior fusion (PF) where the PF mass was broken down and correction was re-performed through the anterior approach and (2) segments with nonunion after prior interbody fusion where cage removal and revisional interbody fusion were performed, either anteriorly or posteriorly.
The presence or absence of RF was confirmed using plain radiographs taken during routine follow-up (i.e., 6 weeks and 3, 6, 12, 18, and 24 months, and annually thereafter). Once RF was detected, fusion status of the affected segment was evaluated using computed tomography (CT). The decision to proceed with close observation or revision surgery was made based on the bilaterality of the RF, CT findings on fusion status, and symptom aggravation. Unilateral RFs were managed conservatively, as they did not aggravate pain. For bilateral RFs, revision surgery was considered when radiographic motion or displacement was observed together with mechanical pain. However, even in bilateral fractures, when fracture sites showed no motion on dynamic radiographs, and CT scans confirmed sufficient fusion mass without any associated pain exacerbation, revision surgery was not performed.
Presumed Risk Factors
Risk factors were analyzed with regard to patient, surgical, and radiographic factors. Patient factors included sex, age, American Society of Anesthesiologists physical status, body mass index (BMI), T-score (lowest values in spine or hip bone densitometry), and perioperative administration of teriparatide. Only patients whose total duration of teriparatide use before and after surgery was 3 months or longer were included. Surgical factors were assessed for each segment, including fusion methods (ALIF, PLIF, LLIF, ACR, and no anterior support [i.e., PF]), pedicle subtraction osteotomy (PSO), corpectomy, and the number of rods. The number of rods was assessed at the segmental levels by determining how many rods spanned at each segment (Fig. 1). Additionally, the performance of pelvic fixation and length of fusion were evaluated. The sagittal parameters, including PI, LL, PI – LL, sacral slope (SS), PT, T1 pelvic angle (TPA), and SVA, were measured preoperatively and at 6 weeks postoperatively for descriptive statistics. The analysis of risk factors related to radiographic parameters was conducted using the following alignment assessment metrics: sagittal modifiers of the Scoliosis Research Society (SRS)-Schwab classification for PI – LL, PT, and SVA (classified as grade 0, +, and ++), Global Alignment and Proportion (GAP) score (classified as proportioned, moderately disproportioned, and severely disproportioned), and restoration relative to Roussouly curve types (classified as restored and non-restored).17,18,19)
Fig. 1. Examples of different multi-rod configurations. (A) A satellite rod was anchored from L1-2 space to pelvis. From the perspective of the segment, the segments from L2-3 to L5-S1 were reinforced with 3 rods. (B) Two accessory rods were placed between T12-L1 and L4-5. Despite the use of 4 rods, the segments from L1-2 to L3-4 were covered by the satellite rods, while other segments were not reinforced by the satellite rods. Segments where a transverse connector was inserted for satellite rod attachment were not classified as being supported by a multi-rod construct.
Statistical Analysis
All data were presented as numbers with percentages for categorical variables and as means with standard deviations for continuous variables. Risk factor analyses were conducted separately for the L5-S1 and ≥ L4-5 levels. Bivariate comparisons between RF and non-RF groups were performed using independent t-tests (or Mann-Whitney U-test, as appropriate) for continuous variables and chi-square tests for categorical variables. Then, multivariate logistic regression analysis was performed using the significant variables on the bivariate comparisons. T-score and teriparatide use were included in the multivariate analysis regardless of their statistical significance. All statistical analyses were performed using IBM SPSS Statistics for Windows version 27 (IBM Corp.). Statistical significance was set at p < 0.05.
RESULTS
Baseline Characteristics
A total of 318 patients were included in this study (female, 88.4%; age, 69.3 years; BMI, 25.9 kg/m2; T-score, −1.4) (Table 1). Teriparatide was administered perioperatively to 52 patients (16.4%). The mean duration of teriparatide use was 3.8 ± 1.2 months and 6.0 ± 2.1 months before and after surgery, respectively. Regarding surgical techniques, anterior-posterior combined approaches were conducted in 74.5% of patients. PSO and corpectomy were performed in 10.1% and 6.3% of patients, respectively. The mean number of interbody fusion levels was 3.6. The mean fusion length was 7 with the most common uppermost instrumented vertebra levels of T10 (42.8%). The numbers of rods were 4, 3, and 2 in 19.2%, 12.3%, and 68.5% of patients, respectively. Pelvic fixation was performed in 73.6% of patients. The mean follow-up duration was 47.4 months.
Table 1. Baseline Data.
| Overview of cohorts | Value (n = 318) | |
|---|---|---|
| Sex (female) | 281 (88.4) | |
| Age (yr) | 69.3 ± 6.7 | |
| ASA physical status | ||
| Grade 1 | 18 (5.7) | |
| Grade 2 | 252 (79.2) | |
| Grade 3 | 48 (15.1) | |
| BMI (kg/m2) | 25.9 ± 4.0 | |
| T-score | −1.4 ± 1.4 | |
| Teriparatide use | 52 (16.4) | |
| Previous fusion | 128 (40.3) | |
| Anterior-posterior combined approach | 237 (74.5) | |
| PSO | 32 (10.1) | |
| Corpectomy | 20 (6.3) | |
| Interbody fusion level | 3.6 ± 1.0 | |
| Fusion length | 7.3 ± 1.9 | |
| Number of rods | ||
| 4 | 61 (19.2) | |
| 3 | 39 (12.3) | |
| 2 | 218 (68.5) | |
| Pelvic fixation | 234 (73.6) | |
| Follow-up (mo) | 47.4 ± 22.1 | |
Values are presented as number (%) or mean ± standard deviation.
ASA: American Society of Anesthesiologists, BMI: body mass index, PSO: pedicle subtraction osteotomy.
Postoperative Radiographic Results
Except for PI, all sagittal parameters such as LL, PI – LL, SS, PT, TPA, and SVA significantly improved after surgery (Table 2). With regard to SRS-Schwab classification, 65.1%, 57.7%, and 72.3% of patients achieved grade 0 sagittal modifier for PI – LL, PT, and SVA, respectively. On the GAP score, 26.7%, 46.5%, and 26.7% of patients were classified as proportioned, moderately disproportioned, and severely disproportioned, respectively. Roussouly types were restored in 40.3% of patients.
Table 2. Postoperative Radiographic Results for Overall Cohort (n = 318).
| Variable | Preoperative | Postoperative at 6 wk | p-value | ||
|---|---|---|---|---|---|
| Sagittal parameter | |||||
| PI – LL (°) | 41.3 ± 19.8 | 6.6 ± 11.6 | < 0.001* | ||
| SS (°) | 20.9 ± 11.2 | 35.0 ± 9.2 | < 0.001* | ||
| PT (°) | 32.9 ± 10.7 | 18.8 ± 9.2 | < 0.001* | ||
| TPA (°) | 32.9 ± 11.9 | 15.6 ± 8.7 | < 0.001* | ||
| SVA (mm) | 80.3 ± 58.0 | 20.4 ± 31.4 | < 0.001* | ||
| Correction status at 6 weeks postoperatively | |||||
| SRS-Schwab PI – LL sagittal modifier | |||||
| Grade 0 (< 10°) | 207 (65.1) | ||||
| Grade + (10°– 20°) | 71 (22.3) | ||||
| Grade ++ (> 20°) | 40 (12.6) | ||||
| SRS-Schwab PT sagittal modifier | |||||
| Grade 0 (< 20°) | 174 (54.7) | ||||
| Grade + (20°–30°) | 114 (35.8) | ||||
| Grade ++ (> 30°) | 30 (9.4) | ||||
| SRS-Schwab SVA sagittal modifier | |||||
| Grade 0 (< 4 cm) | 230 (72.3) | ||||
| Grade + (4.0–9.5 cm) | 84 (26.4) | ||||
| Grade ++ (> 9.5 cm) | 4 (1.3) | ||||
| GAP score | |||||
| Proportioned | 85 (26.7) | ||||
| Moderately disproportioned | 148 (46.5) | ||||
| Severely disproportioned | 85 (26.7) | ||||
| Roussouly curve type | |||||
| Restored | 128 (40.3) | ||||
| Non-restored | 190 (59.7) | ||||
Values are presented as mean ± standard deviation or number (%).
PI – LL: pelvic incidence minus lumbar lordosis, SS: sacral slope, PT: pelvic tilt, TPA: T1 pelvic angle, SVA: sagittal vertical axis, SRS: Scoliosis Research Society, GAP: Global Alignment and Proportion.
*Indicates statistical significance.
Incidence of RF and Revision Surgery
Overall, RF developed in 69 patients (21.7%), among whom 32 patients (46.4%) underwent revision surgery for bilateral RFs. None of the patients with unilateral RF required revision surgery since new-onset pain was not detected. Three patients with bilateral RFs did not undergo revision, as imaging revealed solid fusion without segmental motion and no pain aggravation was observed. At the segment level, a total of 1,329 segments were evaluated (247 and 1,082 segments for L5-S1 and ≥ L4-5 levels, respectively) (Table 3). The incidences of RF and related revision surgery were significantly higher at the L5-S1 level than at the ≥ L4-5 levels (11.7% vs. 4.6%, p < 0.001 and 4.5% vs. 2.2%, p = 0.048).
Table 3. Incidence of Rod Fracture and Revision Surgery According to the Segment Groups.
| Group | No. of segments for evaluation | Rod fracture | Revision surgery | ||
|---|---|---|---|---|---|
| Incidence | p-value | Incidence | p-value | ||
| L5-S1 | 247 | 29 (11.7) | < 0.001* | 11 (4.5) | 0.048* |
| ≥ L4-5 | 1,082 | 50 (4.6) | 24 (2.2) | ||
Values are presented as number (%).
*Indicates statistical significance.
Risk Factors for RF at L5-S1
In the bivariate comparison, fusion length was significantly greater in the RF group than in the non-RF group (8.0 vs. 7.1 levels, p = 0.012) (Table 4). The proportion of grade ++ in the PT sagittal modifier was significantly higher in the RF group than in the non-RF group, while the proportion of grade 0 was significantly lower. Multivariate analysis revealed that increased fusion length (odds ratio [OR], 1.251; 95% CI, 1.035–1.512; p = 0.021) and grade ++ of PT sagittal modifier (vs. grade 0) (OR, 4.781; 95% CI, 1.596–14.323; p = 0.005) were significant risk factors for RF at L5-S1.
Table 4. Risk Factors for RF at L5-S1 (247 Segments for Evaluation).
| Variable | Bivariate comparison | Multivariate analysis | ||||||
|---|---|---|---|---|---|---|---|---|
| Non-RF group (n = 218) | RF group (n = 29) | p-value | OR | 95% CI | p-value | |||
| Demographic factor | ||||||||
| Sex (female) | 191 (87.6) | 26 (89.7) | 0.752 | |||||
| Age (yr) | 69.1 ± 6.7 | 69.5 ± 5.0 | 0.759 | |||||
| ASA physical status | 0.064 | |||||||
| Grade 1 | 26 (11.9) | 7 (24.1) | ||||||
| Grade 2 | 182 (83.5) | 19 (65.5) | ||||||
| Grade 3 | 10 (4.6) | 3 (10.3) | ||||||
| BMI (kg/m2) | 25.6 ± 3.9 | 26.8 ± 4.0 | 0.139 | |||||
| T-score | −1.5 ± 1.4 | −1.6 ± 1.5 | 0.777 | 1.056 | 0.796–1.400 | 0.706 | ||
| Non-use of teriparatide | 178 (81.7) | 27 (93.1) | 0.123 | 3.164 | 0.699–14.285 | 0.135 | ||
| Surgical factor | ||||||||
| Fusion methods | 0.275 | |||||||
| ALIF | 115 (52.8) | 17 (58.6) | ||||||
| PLIF | 81 (37.2) | 7 (24.1) | ||||||
| PF | 22 (10.1) | 5 (17.2) | ||||||
| Number of rods | 0.410 | |||||||
| 3 | 5 (2.3) | 0 | ||||||
| 2 | 213 (97.7) | 29 (100) | ||||||
| Pelvic fixation | 192 (88.1) | 26 (89.7) | 0.804 | |||||
| Fusion length | 7.1 ± 1.9 | 8.0 ± 1.7 | 0.012* | 1.251 | 1.035–1.512 | 0.021* | ||
| Radiographic factor | ||||||||
| SRS-Schwab PI – LL sagittal modifier | 0.315 | |||||||
| Grade 0 (< 10°) | 147 (67.4) | 18 (62.1) | ||||||
| Grade + (10°– 20°) | 47 (21.6) | 5 (17.2) | ||||||
| Grade ++ (> 20°) | 24 (11.0) | 6 (20.7) | ||||||
| SRS-Schwab PT sagittal modifier | 0.004* | 0.014* | ||||||
| Grade 0 (< 20°) | 125 (57.3) | 13 (44.8) | Reference | |||||
| Grade + (20°–30°) | 76 (34.9) | 8 (27.6) | 1.117 | 0.430–2.903 | 0.821 | |||
| Grade ++ (> 30°) | 17 (7.8) | 8 (27.6) | 4.781 | 1.596–14.323 | 0.005 | |||
| SRS-Schwab SVA sagittal modifier | 0.715 | |||||||
| Grade 0 (< 4 cm) | 165 (75.7) | 21 (72.4) | ||||||
| Grade + (4.0–9.5 cm) | 50 (22.9) | 8 (27.6) | ||||||
| Grade ++ (> 9.5 cm) | 3 (1.4) | 0 | ||||||
| GAP score | 0.199 | |||||||
| Proportioned | 59 (27.1) | 11 (37.9) | ||||||
| Moderately disproportioned | 106 (48.6) | 9 (31.0) | ||||||
| Severely disproportioned | 53 (24.3) | 9 (31.0) | ||||||
| Roussouly curve type | ||||||||
| Restored | 94 (43.1) | 8 (27.6) | 0.110 | |||||
| Non-restored | 124 (56.9) | 21 (72.4) | ||||||
| Follow-up (mo) | 47.7 ± 22.4 | 52.9 ± 23.0 | 0.248 | |||||
Values are presented as number (%) or mean ± standard deviation.
RF: rod fracture, OR: odds ratio, ASA: American Society of Anesthesiologists, BMI: body mass index, ALIF: anterior lumbar interbody fusion, PLIF: posterior lumbar interbody fusion, PF: posterior fusion, SRS: Scoliosis Research Society, PI – LL: pelvic incidence minus lumbar lordosis, PT: pelvic tilt, SVA: sagittal vertical axis, GAP: Global Alignment and Proportion.
*Indicates statistical significance.
Risk Factors for RF at or above L4-5
In the bivariate comparison, the T-score tended to be lower in the RF group than in the non-RF group (−1.8 vs. −1.5, p = 0.084). Teriparatide was administered more frequently in the non-RF group than in the RF group, albeit without significance (17.6% vs. 8.0%, p = 0.078) (Table 5). Regarding the fusion methods, the proportion of ACR and PF was significantly higher in the RF than in the non-RF group (36.0% vs. 19.2% for ACR; 42.0% vs. 29.8% for PF; p < 0.001). PSO was performed more frequently in the RF group than in the non-RF group (14.0% vs. 5.2%, p = 0.009). Regarding the number of rods, multi-rod fixation was used significantly more often in the non-RF group than in the RF group (30.6% vs. 13.7%, p = 0.043). Fusion length was significantly greater in the RF group than in the non-RF group (8.0 vs. 7.3 levels, p = 0.007). Multivariate analysis revealed that non-use of teriparatide (OR, 3.460; 95% CI, 1.204–9.901; p = 0.021), ACR (vs. PLIF) (OR, 5.424; 95% CI, 1.795–16.391; p = 0.003), PF (vs. PLIF) (OR, 3.276; 95% CI, 1.128–9.513; p = 0.029), performance of PSO (OR, 3.803; 95% CI, 1.469–9.848; p = 0.006), and dual rods (vs. 4 rods) (OR, 2.873; 95% CI, 1.028–7.813; p = 0.044) were significant risk factors for RF at ≥ L4-5 levels.
Table 5. Risk Factors for RF at or above L4-5 (1,082 Segments for Evaluation).
| Variable | Bivariate comparison | Multivariate analysis | ||||||
|---|---|---|---|---|---|---|---|---|
| Non-RF group (n =1,032) | RF group (n = 50) | p-value | OR | 95% CI | p-value | |||
| Demographic factor | ||||||||
| Sex (female) | 921 (89.2) | 41 (82.0) | 0.111 | |||||
| Age (yr) | 69.4 ± 6.4 | 68.6 ±7.6 | 0.409 | |||||
| ASA physical status | 0.301 | |||||||
| Grade 3 | 140 (13.6) | 3 (6.0) | ||||||
| Grade 2 | 830 (80.4) | 44 (88.0) | ||||||
| Grade 1 | 62 (6.0) | 3 (6.0) | ||||||
| BMI (kg/m2) | 25.8 ± 3.8 | 25.1 ± 3.5 | 0.171 | |||||
| T-score | −1.5 ± 1.4 | −1.8 ± 1.4 | 0.084 | 0.845 | 0.683–1.044 | 0.119 | ||
| Non-use of teriparatide use | 850 (82.4) | 46 (92.0) | 0.078 | 3.460 | 1.204–9.901 | 0.021* | ||
| Surgical factor | ||||||||
| Fusion methods | < 0.001* | |||||||
| PLIF | 175 (17.0) | 5 (10.0) | Reference | |||||
| LLIF | 351 (34.0) | 6 (12.0) | 0.994 | 0.278–3.554 | 0.993 | |||
| ACR | 198 (19.2) | 18 (36.0) | 5.424 | 1.795–16.391 | 0.003* | |||
| PF | 308 (29.8) | 21 (42.0) | 3.276 | 1.128–9.513 | 0.029* | |||
| PSO | 54 (5.2) | 7 (14.0) | 0.009* | 3.803 | 1.469–9.848 | 0.006* | ||
| Corpectomy | 32 (3.1) | 3 (6.0) | 0.258 | |||||
| Number of rods | 0.043* | |||||||
| 4 | 185 (17.9) | 3 (5.9) | Reference | |||||
| 3 | 131 (12.7) | 4 (7.8) | 1.849 | 0.867–9.091 | 0.085 | |||
| 2 | 715 (69.4) | 44 (86.3) | 2.873 | 1.028–7.813 | 0.044* | |||
| Pelvic fixation | 791 (76.6) | 39 (78.0) | 0.825 | |||||
| Fusion length | 7.3 ± 1.9 | 8.0 ± 1.9 | 0.007* | 1.145 | 0.996–1.316 | 0.061 | ||
| Radiographic factor | ||||||||
| SRS-Schwab PI – LL sagittal modifier | 0.538 | |||||||
| Grade 0 (< 10°) | 699 (67.7) | 31 (62.0) | ||||||
| Grade + (10°– 20°) | 218 (21.1) | 11 (22.0) | ||||||
| Grade ++ (> 20°) | 115 (11.1) | 8 (16.0) | ||||||
| SRS-Schwab PT sagittal modifier | 0.292 | |||||||
| Grade 0 (< 20°) | 587 (56.9) | 23 (46.0) | ||||||
| Grade + (20°–30°) | 359 (34.8) | 21 (42.0) | ||||||
| Grade ++ (> 30°) | 86 (8.3) | 6 (12.0) | ||||||
| SRS-Schwab SVA sagittal modifier | 0.534 | |||||||
| Grade 0 (< 4 cm) | 767 (74.3) | 40 (80.0) | ||||||
| Grade + (4.0–9.5 cm) | 251 (24.3) | 10 (20.0) | ||||||
| Grade ++ (> 9.5 cm) | 14 (1.4) | 0 | ||||||
| GAP score | 0.125 | |||||||
| Proportioned | 297 (28.8) | 8 (16.0) | ||||||
| Moderately disproportioned | 484 (46.9) | 26 (52.0) | ||||||
| Severely disproportioned | 251 (24.3) | 16 (32.0) | ||||||
| Roussouly curve type | 0.282 | |||||||
| Restored | 430 (41.7) | 17 (34.0) | ||||||
| Non-restored | 602 (58.3) | 33 (66.0) | ||||||
| Follow-up (mo) | 47.6 ± 22.3 | 51.2 ± 22.2 | 0.271 | |||||
Values are presented as number (%) or mean ± standard deviation.
RF: rod fracture, OR: odds ratio, ASA: American Society of Anesthesiologists, BMI: body mass index, PLIF: posterior lumbar interbody fusion, LLIF: lateral lumbar interbody fusion, ACR: anterior column realignment, PF: posterior fusion, PSO: pedicle subtraction osteotomy, SRS: Scoliosis Research Society, PI – LL: pelvic incidence minus lumbar lordosis, PT: pelvic tilt, SVA: sagittal vertical axis, GAP: Global Alignment and Proportion.
*Indicates statistical significance.
DISCUSSION
Considering the negative clinical impacts, a thorough analysis and recognition of risk factors for RF are essential for its prevention. Previous studies have demonstrated several risk factors, which were assessed at the “patient” level.6,7,8,9,10,11) However, even in the same patient, the surgical procedures may differ across the segments. For instance, even if multi-rod constructs are utilized, they do not uniformly cover all instrumented segments, leading to some segments being reinforced by satellite rods while others are not (Fig. 2). In addition, in patients with segments that have already achieved solid fusion due to prior surgery, RF is unlikely to occur at those levels. Therefore, failing to account for this may lead to underestimation of the true incidence of RF and hinder the accurate identification of risk factors. In this context, a key strength of this study lies in segment-based analysis of risk factors for RF, as opposed to the conventional patient-level approach. We believe that analyzing risk factors for RF separately for L5-S1 and ≥ L4-5 levels is necessary for the following reasons. First, the lumbosacral junction has distinct anatomical characteristics (e.g., caudal inclination of the disc space) and it is subjected to greater mechanical load compared with other segments. Second, the fusion techniques differ between those levels. ALIF is usually performed exclusively for the L5-S1 level, while ACR or LLIF are the procedures indicated at ≥ L4-5 levels. Likewise, analyzing risk factors for RF at the patient level without stratifying the segments may result in either underestimating or overestimating their true impacts.
Fig. 2. (A) A case of a 72-year-old woman with severe thoracolumbar kyphosis. (B) Corrective surgery was performed using anterior column realignment at the L3-4 and L4-5 levels, and pedicle subtraction osteotomy at the L2 vertebra. Two satellite rods were inserted covering from T12-L1 to L4-5 on the left side and from T12-L1 to L5-S1 on the right side. Therefore, the L1-2, L2-3, and L3-4 levels were considered to be covered by 4 rods, while the L4-5 segment was covered by 3 rods. (C) Two years after surgery, bilateral rod fractures developed at the L4-5 segment (white arrows). (D) Despite bilateral rod fractures, revision surgery was not performed because global alignment had remained relatively stable without symptom worsening.
In this study, we found that different factors contribute to RF occurrence between L5-S1 and ≥ L4-5 levels: increased fusion length and remnant high PT for RF at the L5-S1 level, and non-use of teriparatide, ACR, and PF (vs. PLIF), and performance of PSO for RF at ≥ L4-5 levels. These findings suggest that RF at the L5-S1 level is affected by overall mechanical loading (represented by increased fusion length) and global sagittal alignment (represented by inappropriate PT correction), as demonstrated in previous studies.6,10,20,21) ALIF has been traditionally considered to lower the rate of pseudoarthrosis compared to PLIF.21,22) However, recent studies have demonstrated no notable difference in fusion rates between ALIF and PLIF14,15) or even a higher nonunion rate in the ALIF than in the PLIF.23) Moreover, Cavagnaro et al. have reported that PF demonstrated similar overall construct pseudoarthrosis rates as interbody fusion.11) Consistent with previous studies, we also found that fusion methods did not significantly affect occurrence of RF at L5-S1.
In contrast to L5-S1, RF developing at ≥ L4-5 appears to be more affected by local factors such as fusion methods, PSO, and number of rods. In this region, fusion length and sagittal alignment status did not influence RF, suggesting that segments at ≥ L4-5 differ from L5-S1 in their biomechanical circumstances regarding the development of RF. PSO is a well-known risk factor for RF.24) We also found that ACR significantly increases the risk of RF compared with PLIF (OR, 5.424). The ACR procedure involves the release of the anterior longitudinal ligament, allowing for powerful restoration of the segmental angle. However, similar to our results, previous studies have reported that ACR increases the risk of RF, likely due to disrupted anterior tension band integrity.8,25) We found that 4-rod fixation significantly reduced the risk of RF compared with dual rod construct. Our finding about number of rods is consistent with previous clinical studies and meta-analyses advocating its use for prevention of RF.7,9,13,26,27) Furthermore, several biomechanical studies confirmed that 4-rod constructs significantly decrease primary rod strains.28,29) Therefore, in segments with a high risk of RF, such as those treated with PSO or ACR, satellite rod augmentation is recommended. Lastly, we found that the non-use of teriparatide significantly increased the rate of RF (OR, 3.460), consistent with a recent study demonstrating that perioperative administration of teriparatide was associated with a lower rate of pseudoarthrosis, thanks to its promoting effect on fusion.30)
This observation indicates that the biomechanical characteristics differ between the L5-S1 and ≥ L4-5 levels. Because the L5-S1 segment functions as the foundational base of a long fusion construct, it inevitably bears concentrated mechanical load transmitted from the upper instrumented levels. Consequently, RF at this segment is presumed to be influenced by lengthy fusion levels. Furthermore, global sagittal malalignment may amplify this stress concentration by disrupting stable load transmission. In contrast, the segments above L4-5 are less affected by this global environment and more susceptible to local mechanical factors such as the fusion technique or the number of rods, owing to the preferential loading across the L5-S1 junction.
This study has some limitations. First, the retrospective nature of the study is an inherent limitation despite using prospectively collected data. Second, since we analyzed radiographically diagnosed RFs, a considerable number of asymptomatic unilateral RFs were included. Therefore, there might be a concern regarding the clinical implications of our results. However, even a unilateral RF should not be considered clinically insignificant, as it may progress to a symptomatic bilateral fracture over time. In such cases, CT evaluation may be helpful to assess the degree of fusion, and closer follow-up is warranted to monitor potential progression to bilateral RF. A future study with longer follow-up is necessary to determine the fate of these unilateral RFs. Lastly, we acknowledge that the varied experience levels of the surgeons, non-uniform use of multiple rods, and heterogeneous use of bone substitutes or teriparatide may have differently influenced pseudoarthrosis.
This study demonstrated that the risk factors for RF are different between L5-S1 and ≥ L4-5 levels. Increased fusion length and poor postoperative correction of PT were significant risk factors for RF at the L5-S1 level, while non-use of teriparatide, ACR, and PF, performance of PSO, and dual rod construct significantly increased the risk of RF at ≥ L4-5 levels. Understanding the differing risk factors can help establish suitable preventive strategies tailored to each patient.
Footnotes
CONFLICT OF INTEREST: Se-Jun Park is an editorial board member of the journal but was not involved in the peer reviewer selection, evaluation, or decision process of this article. No other potential conflicts of interest relevant to this article were reported.
References
- 1.Kim HJ, Yang JH, Chang DG, et al. Adult spinal deformity: a comprehensive review of current advances and future directions. Asian Spine J. 2022;16(5):776–788. doi: 10.31616/asj.2022.0376. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Smith JS, Kelly MP, Yanik EL, et al. Operative versus nonoperative treatment for adult symptomatic lumbar scoliosis at 5-year follow-up: durability of outcomes and impact of treatment-related serious adverse events. J Neurosurg Spine. 2021;35(1):67–79. doi: 10.3171/2020.9.SPINE201472. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Sardi JP, Lazaro B, Smith JS, et al. Rod fractures in thoracolumbar fusions to the sacrum/pelvis for adult symptomatic lumbar scoliosis: long-term follow-up of a prospective, multicenter cohort of 160 patients. J Neurosurg Spine. 2023;38(2):217–229. doi: 10.3171/2022.8.SPINE22423. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Soroceanu A, Diebo BG, Burton D, et al. Radiographical and implant-related complications in adult spinal deformity surgery: incidence, patient risk factors, and impact on health-related quality of life. Spine (Phila Pa 1976) 2015;40(18):1414–1421. doi: 10.1097/BRS.0000000000001020. [DOI] [PubMed] [Google Scholar]
- 5.Smith JS, Klineberg E, Lafage V, et al. Prospective multicenter assessment of perioperative and minimum 2-year postoperative complication rates associated with adult spinal deformity surgery. J Neurosurg Spine. 2016;25(1):1–14. doi: 10.3171/2015.11.SPINE151036. [DOI] [PubMed] [Google Scholar]
- 6.Marques MF, Fiere V, Obeid I, et al. Pseudarthrosis in adult spine deformity surgery: risk factors and treatment options. Eur Spine J. 2021;30(11):3225–3232. doi: 10.1007/s00586-021-06861-w. [DOI] [PubMed] [Google Scholar]
- 7.Zhao J, Nie Z, Zhang Z, Liao D, Liu D. Multiple-rod constructs in adult spinal deformity surgery: a systematic review and meta-analysis. Asian Spine J. 2023;17(5):985–995. doi: 10.31616/asj.2022.0266. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Park SJ, Park JS, Kang M, Jung K, Lee CS, Kang DH. Incidence and risk factors for mechanical failure after anterior column realignment in adult spinal deformity surgery. Spine (Phila Pa 1976) 2025;50(1):10–18. doi: 10.1097/BRS.0000000000005114. [DOI] [PubMed] [Google Scholar]
- 9.Sardi JP, Berlin C, Buell TJ, et al. Use of supplemental rod constructs in adult spinal deformity surgery: a review. Oper Neurosurg. 2025;29(3):323–334. doi: 10.1227/ons.0000000000001484. [DOI] [PubMed] [Google Scholar]
- 10.Cho W, Mason JR, Smith JS, et al. Failure of lumbopelvic fixation after long construct fusions in patients with adult spinal deformity: clinical and radiographic risk factors: clinical article. J Neurosurg Spine. 2013;19(4):445–453. doi: 10.3171/2013.6.SPINE121129. [DOI] [PubMed] [Google Scholar]
- 11.Cavagnaro MJ, Orenday-Barraza JM, Khan N, et al. Is L5/S1 interbody fusion necessary in long-segment surgery for adult degenerative scoliosis? A systematic review and meta-analysis. J Neurosurg Spine. 2022;36(6):997–1004. doi: 10.3171/2021.9.SPINE21883. [DOI] [PubMed] [Google Scholar]
- 12.Sakuma T, Kotani T, Iijima Y, Akazawa T, Ohtori S, Minami S. Analysis of rod fracture at the lumbosacral junction following surgery for adult spinal deformity. Asian Spine J. 2024;18(1):79–86. doi: 10.31616/asj.2023.0182. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Park JS, Kim HJ, Park SJ, Kang DH, Lee CS. A comprehensive review of risk factors and prevention strategies: how to minimize mechanical complications in corrective surgery for adult spinal deformity. Asian Spine J. 2025;19(3):463–475. doi: 10.31616/asj.2024.0505. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Park SJ, Lee CS, Park JS, et al. L5-S1 nonunion occurrence even after anterior column support combined with iliac screw fixation in long fusion for adult spinal deformity: CT-based analysis at 2-year follow-up. J Neurosurg Spine. 2022;37(3):420–428. doi: 10.3171/2022.1.SPINE211000. [DOI] [PubMed] [Google Scholar]
- 15.Park SJ, Park JS, Nam Y, Yum TH, Choi YT, Lee CS. Failure types and related factors of spinopelvic fixation after long construct fusion for adult spinal deformity. Neurosurgery. 2021;88(3):603–611. doi: 10.1093/neuros/nyaa469. [DOI] [PubMed] [Google Scholar]
- 16.Park SJ, Park JS, Lee CS, Lee KH. Metal failure and nonunion at L5-S1 after long instrumented fusion distal to pelvis for adult spinal deformity: Anterior versus transforaminal interbody fusion. J Orthop Surg (Hong Kong) 2021;29(3):23094990211054223. doi: 10.1177/23094990211054223. [DOI] [PubMed] [Google Scholar]
- 17.Yilgor C, Sogunmez N, Boissiere L, et al. Global Alignment and Proportion (GAP) Score: development and validation of a new method of analyzing spinopelvic alignment to Predict mechanical complications after adult spinal deformity surgery. J Bone Joint Surg Am. 2017;99(19):1661–1672. doi: 10.2106/JBJS.16.01594. [DOI] [PubMed] [Google Scholar]
- 18.Schwab F, Ungar B, Blondel B, et al. Scoliosis Research Society-Schwab adult spinal deformity classification: a validation study. Spine (Phila Pa 1976) 2012;37(12):1077–1082. doi: 10.1097/BRS.0b013e31823e15e2. [DOI] [PubMed] [Google Scholar]
- 19.Pizones J, Martin MB, Perez-Grueso FJ, et al. Impact of adult scoliosis on Roussouly sagittal shape classification. Spine (Phila Pa 1976) 2019;44(4):270–279. doi: 10.1097/BRS.0000000000002800. [DOI] [PubMed] [Google Scholar]
- 20.Guler UO, Cetin E, Yaman O, et al. Sacropelvic fixation in adult spinal deformity (ASD); a very high rate of mechanical failure. Eur Spine J. 2015;24(5):1085–1091. doi: 10.1007/s00586-014-3615-1. [DOI] [PubMed] [Google Scholar]
- 21.Singh V, Oppermann M, Evaniew N, et al. L5-S1 pseudoarthrosis rate with ALIF versus TLIF in adult spinal deformity surgeries: a retrospective analysis of 100 patients. World Neurosurg. 2023;175:e1265–e1276. doi: 10.1016/j.wneu.2023.04.113. [DOI] [PubMed] [Google Scholar]
- 22.Buell TJ, Shaffrey CI, Bess S, et al. Multicenter assessment of outcomes and complications associated with transforaminal versus anterior lumbar interbody fusion for fractional curve correction. J Neurosurg Spine. 2021;35(6):729–742. doi: 10.3171/2020.11.SPINE201915. [DOI] [PubMed] [Google Scholar]
- 23.Meyers AJ, Wick JB, Rodnoi P, Khan A, Klineberg EO. Does L5-S1 anterior lumbar interbody fusion improve sagittal alignment or fusion rates in long segment fusion for adult spinal deformity? Global Spine J. 2021;11(5):697–703. doi: 10.1177/2192568220921833. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Noh SH, Kim KH, Park JY, et al. Characteristics and risk factors of rod fracture following adult spinal deformity surgery: a systematic review and meta-analysis. Neurospine. 2021;18(3):447–454. doi: 10.14245/ns.2040832.416. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Godzik J, Haglin JM, Alan N, et al. Retrospective multicenter assessment of rod fracture after anterior column realignment in minimally invasive adult spinal deformity correction. World Neurosurg. 2019;130:e400–e405. doi: 10.1016/j.wneu.2019.06.096. [DOI] [PubMed] [Google Scholar]
- 26.Gupta S, Eksi MS, Ames CP, et al. A novel 4-rod technique offers potential to reduce rod breakage and pseudarthrosis in pedicle subtraction osteotomies for adult spinal deformity correction. Oper Neurosurg. 2018;14(4):449–456. doi: 10.1093/ons/opx151. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Lyu Q, Lau D, Haddad AF, Deviren V, Ames CP. Multiple-rod constructs and use of bone morphogenetic protein-2 in relation to lower rod fracture rates in 141 patients with adult spinal deformity who underwent lumbar pedicle subtraction osteotomy. J Neurosurg Spine. 2022;36(2):235–245. doi: 10.3171/2021.3.SPINE201968. [DOI] [PubMed] [Google Scholar]
- 28.Hallager DW, Gehrchen M, Dahl B, et al. Use of supplemental short pre-contoured accessory rods and cobalt chrome alloy posterior rods reduces primary rod strain and range of motion across the pedicle subtraction osteotomy level: an in vitro biomechanical study. Spine (Phila Pa 1976) 2016;41(7):E388–E395. doi: 10.1097/BRS.0000000000001282. [DOI] [PubMed] [Google Scholar]
- 29.Shekouhi N, Vosoughi AS, Zavatsky JM, Goel VK, Theologis AA. Biomechanical comparison of multi-rod constructs by satellite rod configurations (in-line vs. lateral) and screw types (monoaxial vs. polyaxial) spanning a lumbar pedicle subtraction osteotomy (PSO): is there an optimal configuration? Eur Spine J. 2022;31(11):3050–3059. doi: 10.1007/s00586-022-07331-7. [DOI] [PubMed] [Google Scholar]
- 30.Mohanty S, Sardar ZM, Hassan FM, Lombardi JM, Lehman RA, Lenke LG. Impact of teriparatide on complications and patient-reported outcomes of patients undergoing long spinal fusion according to bone density. J Bone Joint Surg Am. 2024;106(3):206–217. doi: 10.2106/JBJS.23.00272. [DOI] [PubMed] [Google Scholar]