Expandable cages increase the risk of intraoperative subsidence but do not improve perioperative outcomes in single level transforaminal lumbar interbody fusion

BACKGROUND CONTEXT: Expandable cages (EXP) are being more frequently utilized in transforaminal lumbar interbody fusions (TLIF). EXP were designed to reduce complications related to neurological retraction, enable better lordosis restoration, and improve ease of insertion, particularly in the advent of minimally invasive surgical (MIS) techniques, however they are exponentially more expensive than the nonexpandable (NE) alternative. PURPOSE: To investigate the clinical results of expandable cages in single level TLIF. STUDY DESIGN/SETTING: Retrospective review at a single institution. PATIENT SAMPLE: Two hundred and fifty-two single level TLIFs from 2012 to 2018 were included. OUTCOME MEASURES: Clinical characteristics, perioperative and neurologic complication rates, and radiographic measures. METHODS: Patients ≥18 years of age who underwent single level TLIF with minimum 1 year follow-up were included. Outcome measures: clinical characteristics, perioperative and neurologic complications. Radiographic analysis included pelvic incidence-lumbar lordosis (PI-LL) mismatch, segmental lumbar lordosis (LL) mismatch, disc height restoration, and subsidence ≥2 mm. Statistical analysis included independent t tests and chi-square analysis. For nonparametric variables, Mann-Whitney U test and Spearman partial correlation were utilized. Multivariate regression was performed to assess relationships between surgical variables and recorded outcomes. For univariate analysis significance was set at p<.05. Due to the multiple comparisons being made, significance for regressions was set at p<.025 utilizing Bonferroni correction. RESULTS: Two hundred and fifty-two TLIFs between 2012 and 2018 were included, with 152 NE (54.6% female, mean age 59.28±14.19, mean body mass index (BMI) 28.65±5.38, mean Charlson Comorbidity Index (CCI) 2.20±1.89) and 100 EXP (48% female, mean age 58.81±11.70, mean BMI 28.68±6.06, mean CCI 1.99±1.66) with no significant differences in demographics. Patients instrumented with EXP cages had a shorter length of stay (3.11±2.06 days EXP vs. 4.01±2.64 days NE; Z=-4.189, p<.001) and a lower estimated blood loss (201.31±189.41 mL EXP vs. 377.82±364.06 mL NE; Z=-6.449, p<.001). There were significantly more MIS-TLIF cases and bone morphogenic protein (BMP) use in the EXP group (88% MIS, p<.001 and 60% BMP, p<.001) as illustrated in Table 1. There were no significant differences between the EXP and NE groups in rates of radiculitis and neuropraxia. In multivariate regression analysis, EXP were not associated with a difference in perioperative outcomes or complications. Radiographic analyses demonstrated that the EXP group had a lower PI-LL mismatch than the NE cage group at baseline (3.75±13.81° EXP vs. 12.75±15.81° NE; p=.001) and at 1 year follow-up (3.81±12.84° EXP vs. 8.23±12.73° NE; p=.046), but change in regional and segmental alignment was not significantly different between groups. Multivariate regression demonstrated that EXP use was a risk factor for intraoperative subsidence (2.729[1.185–6.281]; p=.018). CONCLUSIONS: Once technique was controlled for, TLIFs utilizing EXP do not have significantly improved neurologic or radiographic outcomes compared with NE. EXP increase risk of intraoperative subsidence. These results question the value of the EXP given the higher cost.