Age-related variation in the association between intraoperative dynamic cerebral autoregulation impairment proportion...

Dynamic cerebral autoregulation (dCA) impairment during surgery has been associated with adverse neurologic outcomes. Whether the burden of impaired dCA is related to the postoperative peripheral inflammatory response, and whether this association varies with age, remains uncertain.

To examine whether age modifies the association between intraoperative dCA impairment proportion and the neutrophil-to-lymphocyte ratio on postoperative day 3 (NLR3).

This single-center exploratory observational study included 57 adults undergoing elective major abdominal surgery. dCA impairment proportion was calculated as cumulative impaired dCA duration divided by valid dCA monitoring duration; impaired dCA was defined as a cerebral oximetry index (COx) ≥ 0.3 in at least one hemisphere. NLR3 was modeled using a Gamma generalized linear model with a log link. The model included continuous age, dCA impairment proportion, their interaction, baseline neutrophil-to-lymphocyte ratio (NLR0), operation duration, and sex. Robust variance estimation, bootstrap resampling, leave-one-out analysis, and alternative outcome and exposure definitions were used to assess model stability.

The age × dCA impairment proportion interaction was positive (Exp(B) = 1.193 per 10-year × 10-percentage-point increment; model-based 95% confidence interval [CI], 1.051-1.354; P = 0.006). The point estimate was unchanged with heteroskedasticity-consistent type 3 (HC3) robust standard errors, although the CI widened (95% CI, 0.975-1.459; P = 0.086). The 2,000-replicate bootstrap percentile interval for the interaction coefficient was 0.051-0.370. A log-transformed NLR3 sensitivity model yielded a similar interaction (HC3 P = 0.040), whereas no interaction was observed when cumulative impaired dCA duration was used (HC3 P = 0.915). The age-conditional slopes were imprecise. Longer operation duration was associated with higher NLR3 (Exp(B) = 1.214 per hour; 95% CI, 1.041-1.414; P = 0.013).

In this cohort, the modeled association between dCA impairment proportion and NLR3 became more positive with increasing age. The consistency of the interaction depended on the method used to estimate uncertainty and on whether dCA impairment was expressed as a proportion or as cumulative impaired dCA duration. These findings identify an age-related pattern that merits prospective validation using standardized dCA monitoring and serial inflammatory measurements.