Sex Differences In Near-infrared Spectroscopy Reactive Hyperemia Outcomes: Adjusting For Confounders

Sex differences in near-infrared spectroscopy reactive hyperemia outcomes have been previously reported, with females generally having lower rates of tissue resaturation. This may be due to females having greater adipose tissue thickness affecting the signal, or less desaturation during occlusion compared to males. Various approaches have been taken to account for these confounders, though statistical adjustment via analysis of covariance has not been widely used. PURPOSE: Compare statistically adjusted and unadjusted sex differences in tissue resaturation during reactive hyperemia. METHODS: 23 female and 22 male participants completed forearm and thigh post-occlusive reactive hyperemia tests consisting of a 5-minute arterial occlusion, followed by reperfusion. Adipose tissue thickness (ATT) was measured via B-mode ultrasound. Tissue saturation index (TSI %/s) slopes were calculated during occlusion (Slope 1) and during reperfusion (Slope 2). Variables were compared between sexes using Bayesian t tests, then Bayesian linear models were fit with Slope 1 and ATT as covariates. Results are mean [95% credible interval]. RESULTS: In both limbs, females had greater ATT (arm BF10 = 15.2, leg BF10 = 4594.2) and Slope 1 was lower for the leg (BF10 = 1772.1) but not the arm (BF10 = 0.3). Unadjusted, Slope 2 differed between sexes for the leg (females = 0.84 [0.66, 1.03], males = 1.35 [1.17, 1.53], BF10 = 320.2) but not the arm (females = 1.50 [1.29, 1.71], males = 1.51 [1.29, 1.73], BF10 = 0.3). For Slope 2 in both limbs, Bayesian model comparisons showed that accounting for both Slope 1 and ATT, but not sex, were most likely to explain the data. Adjusted Slope 2 values in the leg (accounting for Slope 1 and ATT) were similar between sexes (females = 1.11 [0.98, 1.22], males = 1.11 [1.0, 1.23]) with an adjusted mean difference of -0.01 [-0.19, 0.17]. In the arm, adjustment made less impact on Slope 2 values which remained similar between sexes (females = 1.52 [1.35, 1.70], males = 1.49 [1.31, 1.66]) with an adjusted mean difference of 0.04 [-0.2, 0.29]. CONCLUSIONS: Sex differences may exist in these outcomes before accounting for confounders but are no longer evident after adjustment. Including desaturation rates and adipose tissue thickness as covariates in between-group comparisons may be a useful strategy for correction. Supported by: Mississippi Center for Clinical and Translational Researc