Development and internal validation of a risk score for subsidence of expandable spacers in transforaminal lumbar interbody fusion (TLIF) surgery.

Introduction: Subsidence of transforaminal lumbar interbody fusion (TLIF) cages is a common complication. Little is known about risk factors for cage subsidence (CS) of expandable spacers. Research question: To develop a score for CS risk stratification. Material and methods: We reviewed consecutive patients treated by TLIF with a specific type of expandable interbody implant. CS at the 3-month follow-up was defined as primary endpoint. We identified risk factors, calculated their effect sizes using multivariable logistic regression models and developed the Expandable TLIF Subsidence Index (ETSI). Area Under the Receiver Operating Curve (AUROC) was performed. Results: We identified n = 388 patients (mean age 64.8 years, 55.9 % female), in which we performed TLIF at 482 levels. CS at 3 months was evident in n = 123 patients and on 153 levels (31.7 %). Independent risk factors for CS were high ASA score (3&4; odds ratio (OR) 2.19, 95 % confidence interval (CI) 1.37-3.50, p = 0.008), indication for surgery other than degenerative or deformity (4.79, 1.75-13.1, p = 0.002), lower lumbar spine (L3/4-L5/S1; 4.31, 1.46-12.7, p = 0.008), small cage size (1.94, 1.22-3.07, p = 0.005), and more posterior position of the cage in the intervertebral space (1.68, 1.15-2.43, p = 0.007). The ETSI with a range from -2-6 points obtained an AUROC of 0.6807, indicating a moderate classification result. For each 1-step increase in the ETSI, an OR of 2.02 (1.63-2.50, p < 0.001) to develop CS was calculated and it predicted 3- and 12-month non-union risk. Discussion and conclusions: We identified risk factors for CS, some of which have strong effect sizes and are potentially modifiable.