Table 4:

Sensitivity analysis for change in cognitive scores over time before and after post-enrollment exposure to general anesthesia (GA) with or without N₂O and estimated difference in change in slope for GA without and with N₂O, where post-enrollment exposure is defined only for those without exposures in the 20 years before enrollment.

Cognitive Ssore		GA without N2O*		GA with N2O*		GA without N2O – GA with N2O‡
	Slope est. †	Change in slope est. (95% CI)	P	Change in slope est. (95% CI)	P	Difference in change in slope est. (95% CI)	P
Global z-score	−0.086	−0.071 (−0.110 to −0.031)	<0.001	−0.019 (−0.057 to 0.019)	0.325	−0.052 (−0.104 to −0.001)	0.047
Domain specific z-score
Memory	−0.027	−0.090 (−0.142 to −0.038)	<0.001	−0.040 (−0.077 to −0.004)	0.030	−0.049 (−0.109 to 0.011)	0.105
Language	−0.083	−0.035 (−0.072 to 0.003)	0.068	−0.008 (−0.042 to 0.026)	0.631	−0.027 (−0.073 to 0.019)	0.251
Visuospatial	−0.045	0.008 (−0.030 to 0.047)	0.660	−0.006 (−0.044 to 0.032)	0.749	0.015 (−0.036 to 0.065)	0.567
Attention/executive	−0.104	−0.055 (−0.101 to −0.008)	0.022	−0.028 (−0.079 to 0.024)	0.285	−0.027 (−0.093 to 0.039)	0.415

Analyses presented are from adjusted linear mixed effects models. Six participants were excluded from the analyses because of missing covariates. After exclusions there were 1,813 participants included in the analysis (1,217 had an exposure in the 20 years prior to MCSA enrollment; of 596 without a prior exposure, 83 had a new exposure to GA without N₂O, and 63 had a new exposure to GA with N₂O). All models are adjusted for prior 20-year exposure to GA without N₂O, GA with N₂O, age, sex, cognitive status at enrollment, education, marital status, Charlson comorbidity index, presence of APOE-ε4 genotype, ever-diagnosed alcohol problem, smoking status, mid-life diagnosis of diabetes, hypertension, and dyslipidemia, and history of atrial fibrillation, congestive heart failure, stroke, and coronary artery disease.

^†Slope estimate represents average annual change (slope) in z-score for participants prior to any post-enrollment exposure to surgery/GA.

^*Estimates, 95% confidence intervals, and P-values for the time-dependent change in slope following new exposure to GA without and with N₂O based on a post-enrollment exposure;

^‡P-value for comparison of change in slope estimates between GA without and with N₂O.