Control of Confounding and Claims of Causality in Observational Studies of Neuraxial Anesthesia and Recurrence in Urologic Cancer: A Systematic Review

Background and objective: Several observational studies have claimed that the use of neuraxial anesthesia reduces the risk of recurrence after urologic cancer surgery. However, its use is associated with age, comorbidity, and cancer stage, raising concerns of confounding. We aimed to assess the published literature to evaluate causal inference reporting. Methods: We performed a systematic review of observational studies evaluating the association between anesthesia type and oncologic outcomes in urologic cancer surgery. We used a three-point scale to describe how causal inference was recorded on seven separate reporting criteria. To explore the issue of confounding directly, we evaluated the oncologic outcomes of patients undergoing radical cystectomy at a major cancer hospital according to the type of anesthesia received. Key findings and limitations: We retrieved 18 studies on anesthesia and recurrence after urologic cancer surgery. Reference to causality was completely absent in more than half of the papers; all but one paper failed to assess the control of confounding in the Results section or addressed causal inference in the Discussion section. An analysis of the cohort at our own institution demonstrated large differences in age, comorbidity, and cancer stage, and inconsistent findings between overall, recurrence-free, and cancer-specific survival. Conclusions and clinical implications: Reporting of causal inference is poor for observational studies of neuraxial anesthesia in urologic cancer surgery. The current literature is beset by residual and unmeasured confounding, and randomized controlled trials appear to be the only solution to this research question. However, these are unlikely to be a research priority given the weak biologic rationale. Reporting of causal inference in the urologic literature needs dramatic improvement.