Rethinking sagittal alignment: the value of evaluating spinal shape in conjunction with T4-L1 pelvic angle measurement to prevent proximal junctional failure.

OBJECTIVE: The T4-L1 pelvic angle (T4-L1PA) has been proposed as a novel alignment target associated with mechanical complications in long-segment adult spinal deformity (ASD) surgery. However, its relationship to proximal junctional failure (PJF) and its relevance in shorter thoracolumbar constructs remain unclear. The aim of this study was to evaluate whether T4-L1PA and L1 pelvic angle (L1PA) guidelines predict PJF in ASD surgery and to examine how early proximal reciprocal changes influence this relationship.

METHODS: A prospective ASD database was reviewed for patients who underwent fusion from the ilium to the thoracic spine with a minimum follow-up of 2 years. Postoperative PJF was defined as 1) revision surgery requiring proximal extension or 2) severe radiographic proximal junctional kyphosis (PJK) (proximal junctional angle [PJA] > 28° and ΔPJA > 22°). Patients were categorized based on adherence to L1PA (0.5 × PI -19 ± 2°) and T4-L1PA (from -3° to 1°) guidelines, stratified by upper instrumented vertebra position (upper thoracic [UT], T2-4; lower thoracic, T9-11). Simulated postoperative alignment was used to assess the effect of the PJA on guideline adherence and failure risk.

RESULTS: Among 661 patients (78% female, mean age 65 ± 9 years, mean BMI 28.3 ± 5.7), 17.9% experienced PJF at 2 years (8.3% with revision, 10.4% with severe PJK). Patients with UT UIVs had higher failure rates when L1PA was overcorrected (25.6%) versus matching (12.9%) or undercorrected (11.1%, p = 0.029). In contrast, failure rates across T4-L1PA groups were not different (p = 0.094). The lowest failure rate was identified for either L1PA undercorrected and T4 aligned or L1PA aligned but T4 ventral to L1. For short fusion, no significant differences were found across L1PA guidelines (p = 0.132), and the significant difference in T4-L1PA was not maintained after adjusting for PJA influence on early postoperative alignment (p = 0.148). Similarly, no significant differences were found across the 9 combinations after controlling for PJA influence.

CONCLUSIONS: The T4-L1PA concept demonstrated some association with PJF alone in the setting of long fusions, although adjusting L1PA thresholds toward less correction might further improve outcomes. Unfortunately, current targets seem inappropriate for short fusions, and a different framework might be required.