Fig 1.
Patient screening flowchart.
Table 1.
Clinical characteristics of the study population.
Fig 2.
Relationship between SvO2 and serum UA levels in the whole study population.
A scatterplot showing the relationship between SvO2 and serum UA levels based on the single regression analysis with SvO2 as the independent variable and serum UA level as the dependent variable. As a correlation was found between UA and SvO2 levels, a regression line is drawn. N, number of samples; R2, coefficient of determination; SvO2, mixed venous oxygen saturation; UA, uric acid.
Table 2.
Results of the single regression analysis of SvO2 and UA levels.
Table 3.
Results of the stratified regression analysis.
Fig 3.
Path diagram based on structural equation modeling (for all patients).
The path diagram is devised from structural equations among eight factors affecting SvO2, UA levels, and both SvO2 and UA levels and also shows the effects of SvO2 on UA levels. The paths between variables are represented by one-way arrows extending from the independent to the dependent variable, indicating a positive or negative effect, whereas two-way arrows between the two variables indicate a correlation. Dependent variables are accompanied by error variables (e), one-way arrows by estimates of the standardized coefficient (blue & red), and two-way arrows by estimates of the correlation coefficient (green). Squared multiple correlations are shown in narrow italics. SvO2 shows a significant relationship with UA levels (P<0.01). BMI, body mass index; CI, cardiac index; eGFR, estimated glomerular filtration rate; Hb, hemoglobin; HbA1c, glycated hemoglobin; SvO2, mixed venous oxygen saturation; TG, triglyceride; UA, uric acid; e, extraneous variable.
Table 4.
Structural equation modeling results.
Table 5.
Partial correlation analysis with all possible affecting factors as control factors.