Extended Data Fig. 4: Causal mediation pathway analysis between hyperuricemia and in-hospital mortality in cohort 2.
From: Elevated uric acid levels, mortality and cognitive impairment in children with severe malaria
a, Causal pathway diagrams with adjusted odds ratios (OR) from binary logistic regression, adjusted for age, sex and site. Two-sided p-values shown in diagram and are not adjusted for multiple comparisons. b, The indirect and direct effects of hyperuricemia on in-hospital mortality with potential mediators, showing point estimates with 95% confidence intervals (error bars), obtained from causal mediation analysis. Sample sizes (independent biological samples from individual patients) were n = 595 for AKI and hyperuricemia; n = 594 for intestinal injury, metabolic acidosis and coma; and impaired n = 593 for perfusion. The percentage mediated (indirect effect divided by total effect) is presented to the right of the diagram in the diagram. All mediation analyses are adjusted for age, sex and site, with p-values from two-sided tests. All mediators were statistically significant (p < 0.05). For the relationship between AKI and mortality, mediation analysis was also performed for hyperuricemia as the potential mediator. AKI, acute kidney injury.
