Predictors of Superior Recovery Kinetics in Adult Cervical Deformity Correction: An Analysis Using a Novel Area Under the Curve Methodology

Abstract

Study design: Retrospective review of a prospective database.

Objective: The aim of this study was to identify demographic, surgical, and radiographic factors that predict superior recovery kinetics following cervical deformity (CD) corrective surgery.

Summary of background data: Analyses of CD corrective surgery use area under the curve (AUC) to assess health-related quality of life (HRQL) metrics throughout recovery.

Methods: Outcome measures were baseline (BL) to 1-year (1Y) health-related quality of life (HRQL) (Neck Disability Index [NDI]). CD criteria were C2-7 Cobb angle >10°, coronal Cobb angle >10°, C2-C7 sagittal vertical axis (cSVA) >4 cm, TS-CL >10°, or chin-brow vertical angle >25°. AUC normalization divided BL and postoperative outcomes by BL. Normalized scores (y axis) were plotted against follow-up (x axis). AUC was calculated and divided by cumulative follow-up length to determine overall, time-adjusted recovery (Integrated Health State [IHS]). IHS NDI was stratified by quartile, uppermost 25% being "Superior" Recovery Kinetics (SRK) versus "Normal" Recovery Kinetics (NRK). BL demographic, clinical, and surgical information predicted SRK using generalized linear modeling.

Results: Ninety-eight patients included (62 ± 10 years, 28 ± 6 kg/m2, 65% females, Charlson Comorbidity Index: 0.95), 6% smokers, 31% smoking history. Surgical approach was: combined (33%), posterior (49%), anterior (18%). Posterior levels fused: 8.7, anterior: 3.6, estimated blood loss: 915.9ccs, operative time: 495 minutes. Ames BL classification: cSVA (53.2% minor deformity, 46.8% moderate), TS-CL (9.8% minor, 4.3% moderate, 85.9% marked), horizontal gaze (27.4% minor, 46.6% moderate, 26% marked). Relative to BL NDI (Mean: 47), normalized NDI decreased at 3 months (0.9 ± 0.5, P = 0.260) and 1Y (0.78 ± 0.41, P < 0.001). NDI IHS correlated with age (P = 0.011), sex (P = 0.042), anterior approach (P = 0.042), posterior approach (P = 0.042). Greater BL pelvic tilt (PT) (SRK: 25.6°, NRK: 17°, P = 0.002), pelvic incidence-lumbar lordosis (PI-LL) (SRK: 8.4°, NRK: -2.8°, P = 0.009), and anterior approach (SRK: 34.8%, NRK: 13.3%; P = 0.020) correlated with SRK. 69.4% met MCID for NDI (<Δ-15) and 63.3% met substantial clinical benefit for NDI (<Δ-10); 100% of SRK met both MCID and substantial clinical benefit. The predictive model for SRK included (AUC = 88.1%): BL visual analog scale (VAS) EuroQol five-dimensional descriptive system (EQ5D) (odds rario [OR] 0.96, 95% confidence interval [CI]: 0.92-0.99), BL swallow sleep score (OR: 1.04, 95% CI: 1.01-1.06), BL PT (OR: 1.12, 95% CI: 1.03-1.22), BL modified Japanese Orthopedic Association scale (mJOA) (OR: 1.5, 95% CI: 1.07-2.16), BL T4-T12, BL T10-L2, BL T12-S1, and BL L1-S1.

Conclusion: Superior recovery kinetics following CD surgery was predicted with high accuracy using BL patient-reported (VAS EQ5D, swallow sleep, mJOA) and radiographic factors (PT, TK, T10-L2, T12-S1, L1-S1). Awareness of these factors can improve decision-making and reduce postoperative neck disability.Level of Evidence: 3.