Assessing the Comparative Ability of Early Postoperative T4-L1PA Mismatch versus L1PA or T4PA Alone for Predicting Mechanical Complications or Spinal Reoperations

STUDY DESIGN: Retrospective single-center study. OBJECTIVE: We aim to compare the prognostic capacity of postoperative T4-L1PA mismatch versus T4PA and L1PA alone. SUMMARY OF BACKGROUND DATA: The assessment of T4-L1PA mismatch, the difference between traditional radiographic parameters T4PA and L1PA, is a newer radiographic measurement that has generated interest due to the potential ability to predict complications. METHODS: We performed a retrospective analysis of a prospective single-center database and included surgical patients for ASD with pelvic fixation with complete baseline and two-year follow-up. Primary outcomes included mechanical complications and reoperation rates. Secondary outcomes included proximal junctional kyphosis (PJK) and proximal junctional failure (PJF). Predictors included were six-week T4-L1PA mismatch, T4PA, L1PA, corrected T4-L1PA mismatch defined by pelvic incidence, SAAS score, and SRS-Schwab Score. We utilized receiver operating curve bivariable analyses and multivariable regression analyses for adjustment of confounding. RESULTS: A total of 486 patients (mean age: 63.7 y; 75.5% female; mean preoperative T4-L1PA mismatch: 9.7 degrees; mean number of levels fused: 11.9 levels) had 88 mechanical complications (18.1%) and 125 reoperations (25.7%). In a multivariable regression model for mechanical complications, adding six-week T4-L1PA mismatch did not improve the model (P=0.698) in contrast to six-week T4PA (Δ-2LL: 6.9), L1PA (Δ-2LL: 7.5), and SAAS score (Δ-2LL: 6.4) (P<0.05). For spinal reoperations, adding six-week T4PA (Δ-2LL: 4.4; P=0.036) or T4-L1PA mismatch (Δ-2LL: 3.9; P=0.049) improved the regression model; however, six-week L1PA did not improve the model (P>0.05). Among patients with inadequate preoperative T4-L1PA mismatch (n=401), adding six-week L1PA improved model fit for mechanical complications (Δ-2LL: 4.7; P=0.031); however, six-week T4PA or T4-L1PA mismatch did not improve the model (P>0.05). Six-week L1PA consistently outperformed T4-L1PA mismatch on bivariable analysis. CONCLUSION: Early postoperative T4-L1PA mismatch may not serve as a greatly enhanced radiographic predictor of all-type mechanical complications or spinal reoperations, especially in comparison to either six-week T4PA or L1PA alone. LEVEL OF EVIDENCE: III.