Association of type 2 diabetes mellitus with postoperative outcomes following single-level endoscopic lumbar interbody fusion through a posterior approach: A global propensity score-matched cohort study.

Abstract

Type 2 diabetes mellitus (T2DM) has been associated with impaired bone healing and may increase the risk of pseudarthrosis after lumbar fusion. Its impact after endoscopic lumbar interbody fusion (Endo-LIF) performed through a posterior approach remains unclear. In this retrospective cohort study using the TriNetX Global Collaborative Network, adults undergoing single-level Endo-LIF through a posterior approach between 2015 and 2025 were identified from ICD-10-PCS procedure codes, which record a percutaneous endoscopic interbody fusion approached from behind but do not distinguish uniportal from biportal technique or posterior from transforaminal variants. Patients with and without T2DM were propensity score-matched 1:1 on demographic and clinical variables. The primary outcome was code-defined pseudarthrosis (ICD-10-CM M96.0) within three years; secondary outcomes were 90-day complications and one-year healthcare utilization. Time-to-event outcomes were analyzed with Kaplan-Meier estimates, log-rank tests and Cox regression. After matching, 755 patients remained per cohort. Early complications and healthcare utilization were comparable (all p > 0.05). Code-defined pseudarthrosis was recorded more often with T2DM at every interval (observed proportions 11.5% vs 6.2% by 36 months; all p < 0.01). Kaplan-Meier cumulative incidence at 36 months was 12.0% (95% CI 9.8-14.6) with T2DM versus 6.4% (95% CI 4.9-8.5) without (log-rank p = 0.0003), an absolute risk difference of 5.5% (95% CI 2.6% to 8.5%); the hazard ratio was 1.90 (95% CI 1.33-2.71). In this database cohort, T2DM was associated with a higher rate of code-defined pseudarthrosis after single-level Endo-LIF, despite similar short-term outcomes. Because the outcome was code-defined and the procedure codes do not characterize operative technique, these findings describe an association only: they do not establish that T2DM independently causes radiographic nonunion, and are consistent with, but do not demonstrate, a contribution of systemic metabolic factors to long-term fusion.