Chapter 4 118 Table 4.3. Association between CAR and sleep outcomes per group. The reported estimates reflect the estimated probability of a difference in sleep outcome per 100 nmol/L*minutes increase in the CAR, and a negative estimate implicates that an increase in CAR is accompanied by a possible decrease in the sleep outcome score. The association between sleep quality scores and CAR in OC users and NC women was analyzed using mediation analysis. Sleep quality (PSQI sum score) NC women OC users Men OC users vs. NC women (ref) Men vs. NC women (ref) CAR AUCi (per 100 nmol/L* minutes) Marginal model 0.002 (95% CI: -0.142 to 0.147), p=0.97 -0.241 (95% CI: -0.401 to -0.082), p<0.01 -0.244 (95% CI: -0.454 to -0.034) p=0.023 Covariate-controlled model a. -0.009 (95% CI: -0.161 to 0.144), p=0.91 -0.233 (95% CI: -0.399 to -0.066), p<0.01 -0.233 (95% CI: -0.453 to -0.015) p=0.037 Sleep disruptions (PSQI5B) NC women OC users Men OC users vs. NC women (ref) Men vs. NC women (ref) CAR AUCi (per 100 nmol/L* minutes) Marginal model -0.008 (95% CI: 0.052 to 0.035), p=0.68 -0.138 (95% CI: -0.238 to -0.039) , p<0.01 -0.055 (95% CI: -0.110 to -0.001), p=0.044 -0.129 (95% CI: -0.239 to -0.019) p=0.021 -0.047 (95% CI: -0.117 to 0.022) p=0.18 Covariate-controlled model a. -0.004 (95% CI: -0.048 to -0.042), p=0.87 -0.127 (95% CI: -0.235 to -0.019), p=0.022 -0.049 (95% CI: -0.107 to 0.008), p=0.09 -0.123 (95% CI: -0.241 to -0.005) p=0.042 padjusted=0.084 b. 0.051 (95% CI: 0.122 to 0.020) p=0.16 a. Adjusted for age, smoking and CAR sampling day. b. P-values were adjusted for the primary analysis aims (i.e., the comparison of NC women vs. OC users) using Bonferroni-Holm adjustments on four tests. Confidence intervals were not adjusted for multiple comparisons.
RkJQdWJsaXNoZXIy MTk4NDMw