A 661-patient ISSG prospective multicenter study finds that current L1PA and T4-L1PA alignment guidelines have important limitations — overcorrection increases failure risk in long fusions, and neither framework reliably predicts outcomes in shorter constructs.
OPENING SUMMARY
The prevention of proximal junctional failure — the collapse or progressive kyphosis that can develop at the upper end of a spinal fusion — has been one of adult spinal deformity surgery’s most active research priorities. Over the past decade, the field has developed increasingly sophisticated alignment frameworks, from the L1 pelvic angle to the more recently proposed T4-L1 pelvic angle, aimed at defining postoperative targets that predict durable outcomes. This study subjects both frameworks to rigorous prospective multicenter testing — and the results challenge some of the field’s current assumptions. Overcorrecting L1PA in long fusions may actually increase failure risk. The T4-L1PA framework shows limited predictive value once early proximal junctional angle effects are accounted for. And for shorter fusion constructs, neither current framework provides significant predictive guidance. The study’s conclusion is not that alignment-based planning is wrong — it is that the frameworks need refinement, and that spinal shape must be considered alongside individual alignment values.
STUDY SNAPSHOT
Study type: Prospective multicenter ASD database analysis
Number of patients: 661
Female: 78%
Mean age: 65 ± 9 years
Mean BMI: 28.3 ± 5.7
Surgical procedure: Fusion from ilium to thoracic spine
Minimum follow-up: 2 years
Overall PJF rate at 2 years: 17.9%
Revision surgery (proximal extension): 8.3%
Severe radiographic PJK: 10.4%
UIV stratification: Upper thoracic (UT, T2–4) vs. lower thoracic (T9–11)
L1PA guideline: 0.5 × PI − 19 ± 2°
T4-L1PA guideline: −3° to 1°
Published in: Journal of Neurosurgery: Spine, online ahead of print, October 9, 2026
PMID: 42854238
DOI: 10.3171/2026.4.SPINE251528
Lead author: Renaud Lafage
WHY THIS MATTERS
Proximal junctional failure affects a meaningful proportion of adult spinal deformity patients — in this cohort, 17.9% within two years, split between those requiring revision surgery (8.3%) and those experiencing severe radiographic proximal junctional kyphosis (10.4%). It is not only a radiographic phenomenon: revision surgery for PJF is a major undertaking that carries its own complications, recovery demands, and risks. Preventing PJF through better surgical planning is one of the field’s most important ongoing goals.
The L1 pelvic angle was established as an alignment target based on evidence linking it to lower reoperation rates in open ASD correction — a finding extended to MIS surgery in other ISSG research. The T4-L1 pelvic angle was more recently proposed as a complementary target that incorporates the thoracic spine’s orientation into the alignment assessment, recognizing that the upper end of a lumbar or thoracolumbar fusion must negotiate the transition to the thoracic spine in a way that minimizes junctional stress.
What has been lacking is a rigorous head-to-head evaluation of both frameworks against actual failure outcomes — stratified by fusion length, adjusted for confounders, and accounting for the influence of early postoperative proximal junctional angle on apparent alignment measurements. This study provides that evaluation, and its findings have direct implications for how surgeons plan ASD correction.
UNDERSTANDING THE KEY MEASUREMENTS
L1 Pelvic Angle (L1PA)
The L1 pelvic angle captures the sagittal orientation of the L1 vertebra relative to the pelvis — a measure of how the lumbar correction is positioned relative to the pelvic foundation. The ideal L1PA target used in this study is individualized to each patient’s pelvic anatomy using the formula: 0.5 × pelvic incidence − 19 ± 2°. Achieving this target has been associated in prior ISSG research with reduced reoperation rates in MIS-treated ASD patients. Patients in this study were categorized as undercorrected, matched, or overcorrected relative to this guideline.
T4-L1 Pelvic Angle (T4-L1PA)
The T4-L1 pelvic angle extends the alignment assessment upward, measuring the sagittal orientation of the T4 vertebra relative to L1 within the context of the overall spinal construct. The target range used in this study was −3° to 1°. The conceptual rationale for T4-L1PA is that achieving the correct lumbar alignment (captured by L1PA) may not be sufficient if the thoracic spine’s orientation creates unfavorable mechanical loading at the proximal junction — a shape-related factor that L1PA alone does not capture.
Proximal Junctional Angle (PJA)
The proximal junctional angle measures the kyphotic angle at the junction between the most proximal instrumented vertebra and the two vertebrae immediately above it. The study’s definition of severe radiographic PJK required a PJA exceeding 28° and a change from baseline exceeding 22°. PJA also plays an important methodological role in this study: the authors note that early postoperative PJA values can influence the apparent T4-L1PA measurement, and they tested whether the T4-L1PA associations survived adjustment for this influence.
Proximal Junctional Failure (PJF)
PJF in this study was defined as either revision surgery requiring proximal extension of the fusion or severe radiographic PJK meeting both the absolute and delta PJA thresholds described above. This composite definition captures both the clinical consequence (reoperation) and the radiographic severity threshold that is generally considered to reflect clinically meaningful failure.
KEY FINDINGS
1. Nearly 1 in 5 patients experienced proximal junctional failure within two years.
The overall PJF rate of 17.9% at two years — comprising 8.3% revisions and 10.4% severe radiographic PJK — establishes the scale of the problem this study is trying to address. This rate is consistent with prior literature and underscores why better predictive frameworks are needed. It also means that any alignment guideline that does not meaningfully stratify failure risk represents a missed opportunity for prevention.
2. In long fusions with upper thoracic UIVs, overcorrecting L1PA was associated with higher failure rates — not lower.
This is the study’s most counterintuitive and clinically important finding. Among patients with upper thoracic fusion endpoints, the failure rate was:
Overcorrected L1PA: 25.6%
Matched L1PA: 12.9%
Undercorrected L1PA: 11.1%
The gradient runs in the opposite direction from what clinicians might expect — the drive to maximize correction beyond the L1PA guideline was associated with more than double the failure rate compared to undercorrection. This finding aligns with a broader body of evidence suggesting that overcorrection of sagittal alignment in ASD surgery carries mechanical penalties, and it argues for recalibrating L1PA targets toward less aggressive correction rather than more.
3. T4-L1PA showed some association with PJF in long fusions alone — but not after adjusting for early PJA influence.
In isolation, T4-L1PA groupings showed a trend toward different failure rates in long fusions (p=0.094), but this difference was not statistically significant and was further attenuated when the analysis adjusted for the influence of early postoperative PJA on T4-L1PA measurements. The authors interpret this as an important methodological caution: apparent T4-L1PA relationships may partly reflect the confounding influence of early junctional angle changes on the measurement itself, rather than a true independent association between T4-L1PA and failure risk.
4. The combination of L1PA and T4 position — not either alone — identified the lowest failure rates.
When the analysis examined the relationship between L1PA status and T4 position simultaneously, two combinations emerged with the lowest failure rates: either L1PA undercorrected with T4 well-aligned, or L1PA matched with T4 positioned ventral to L1. Neither target alone was sufficient to identify the optimal outcome group. This finding supports the study’s central argument: spinal shape — specifically the orientation of the upper thoracic spine relative to the lumbar correction — provides predictive information beyond what any single alignment parameter captures.
5. For shorter fusions to the lower thoracic spine, neither framework predicted PJF significantly.
Among patients with lower thoracic UIVs, neither L1PA (p=0.132) nor T4-L1PA guidelines showed significant differences in failure rates. After adjusting for the PJA influence on T4-L1PA measurements, and after examining all nine combinations of the two guidelines, no significant pattern emerged. The authors’ conclusion is direct: current alignment targets appear inappropriate for shorter fusion constructs, and a different planning framework may be required for this patient group.
6. The PJA’s influence on alignment measurements is a methodological concern that warrants attention.
A notable contribution of this study is its explicit examination of how early postoperative proximal junctional angle values can influence the apparent T4-L1PA measurement. Because T4-L1PA captures the relationship between T4 and L1 within the fusion construct, early junctional settling at the proximal end can alter the measured T4-L1PA value in ways that may make apparent guideline compliance or non-compliance an artifact of early junctional mechanics rather than a reflection of the planned alignment. This methodological point has implications for how future studies should analyze and report T4-L1PA outcomes.
PRACTICAL IMPLICATIONS
Overcorrection of L1PA in long fusions with upper thoracic UIVs should be avoided.
The finding that overcorrected L1PA was associated with a 25.6% failure rate — more than double the rate of undercorrection — is the most immediately actionable finding in the study. For surgeons planning long fusion constructs to the upper thoracic spine, the evidence now suggests that erring toward the undercorrected side of the L1PA target, rather than maximizing correction, may reduce failure risk. This is a meaningful recalibration of the intuition that “more correction is better.”
Thoracic spinal shape should be assessed alongside lumbar alignment targets.
The finding that the lowest failure rates emerged from specific combinations of L1PA status and T4 position — rather than from either target alone — supports integrating thoracic shape assessment into preoperative planning. Surgeons should consider not only whether the planned lumbar correction will achieve the L1PA target, but also where the T4 vertebra will sit relative to L1 after correction, and whether that configuration is mechanically favorable for the proximal junction.
Current alignment guidelines should not be applied uniformly to shorter fusion constructs.
The absence of significant L1PA or T4-L1PA associations in the lower thoracic UIV group is an important practical caution. Surgeons planning fusions that end in the lower thoracic spine should recognize that the alignment targets developed primarily in the context of long fusions may not be validated — or appropriate — for their patients. This finding calls for dedicated research to develop fusion-length-specific alignment frameworks.
T4-L1PA measurements in the early postoperative period should be interpreted cautiously.
The study’s demonstration that early PJA influences apparent T4-L1PA values means that T4-L1PA measurements obtained at early postoperative timepoints — before junctional settling has stabilized — should be interpreted with awareness of this potential confound. Clinical and research applications of T4-L1PA should consider whether early measurements reflect planned alignment or early junctional mechanics.
The field may need fusion-length-specific alignment frameworks.
The study explicitly concludes that a different framework may be required for shorter fusions. This is a call to action for future research: developing and validating alignment targets that are calibrated to fusion length, UIV position, and the specific mechanical environment at each type of proximal junction. The current practice of applying the same alignment targets across the full spectrum of ASD fusion constructs appears insufficient.
A NOTE ON STUDY DESIGN: THE VALUE OF PROSPECTIVE MULTICENTER DATA
This study’s use of a prospective multicenter ASD database — rather than a retrospective single-center series — is an important strength that bears noting. Prospective data collection reduces selection bias and ensures that patient characteristics, alignment measurements, and outcomes are captured systematically rather than reconstructed from chart review. The multicenter design increases the generalizability of findings across different surgical practices and patient populations. With 661 patients and a uniform two-year follow-up definition, the study has sufficient statistical power to detect meaningful differences across alignment guideline subgroups — and, critically, sufficient power to demonstrate the absence of significant differences in the shorter fusion group, which is itself a finding rather than a null result.
CONCLUSION
This study delivers a nuanced but important message to the field: the alignment targets currently used to prevent proximal junctional failure in adult spinal deformity surgery are not uniformly validated, and applying them without accounting for fusion length, thoracic shape, and the direction of correction error may not be achieving their intended purpose. Overcorrecting L1PA in long fusions appears counterproductive. T4-L1PA’s independent predictive value attenuates when early junctional angle effects are controlled. And for shorter fusion constructs, neither framework provides the guidance the field needs. The path forward requires not just refinement of existing targets but a more sophisticated framework that integrates spinal shape — the orientation of the thoracic spine relative to the lumbar correction — as a primary planning consideration alongside individual alignment values.
READ THE FULL STUDY
