Background: The evidence about the effects of trace elements on overall survival(OS) of patients with esophageal squamous cell carcinoma(ESCC) is limited. This study aims to evaluate mixed effects of plasma trace elements on OS of ESCC.
Methods: This prospective cohort analysis included 497 ESCC patients with a median follow-up of 52.3 months. The concentrations of 17 trace elements were measured. We fitted Cox's proportional hazards regression, factor analysis and Bayesian kernel machine regression (BKMR) models to estimate the association between trace elements and OS.
Results: Our analysis found that in the single-element model, Co, Ni, and Cd were associated with an increased risk of death, while Ga, Rb, and Ba were associated with a decreased risk. Cd had the strongest risk effect among all elements. As many elements were found to be mutually correlated, we conducted a factor analysis to identify common factors and investigate their associations with survival time. The factor analysis indicated that the factor with high factor loadings in Ga, Ba and B was linked to a decreased risk of death, while the factor with high factor loadings in Co, Ti, Cd and Pb was associated with a borderline significantly increased risk. Using BKMR analysis to disentangle the interaction between elements in significant factors, we discovered that Ga interacted with Ba and both elements had U-shaped effects with OS. Cd, on the other hand, had no interaction with other elements and independently increased the risk of death.
Conclusions: Our analysis revealed that Ga, Ba and Cd were associated with ESCC outcome, with Ga and Ba demonstrating an interaction. These findings provide new insights into the impact of trace elements on the survival of patients with ESCC.
Keywords: Bayesian kernel machine regression; Esophageal squamous cell carcinoma; Outcome; Trace elements.
© 2023. The Author(s).