Associations between sepsis occurrence, hemoglobin level and mortality in patients with non-trauma hemorrhagic brain injuries: trajectory-based analysis

Table 2 Estimated associations between sepsis and mortality in multivariate logistic, IPWRA and PSM models

Multivariate logistic model
Variables	OR (95%CI)	p
Sepsis	2.31 (1.77—3.01)	< 0.001
SOFA on ICU admission	1.12 (1.03—1.21)	0.010
Hypertension	1.31 (1.001—1.713)	0.049
Cirrhosis	2.78 (1.56—4.95)	0.001
Initial WBC count	1.08 (1.05—1.11)	< 0.001
Initial Serum creatinine level	1.39 (1.24—1.55)	< 0.001
Initial Platelet count	0.996 (0.994—0.997)	< 0.001
Vasopressor usage in 3 days	2.28 (1.63—3.19)	< 0.001

Propensity score matching
Outcomes	HBI with sepsis (n = 942) vs. HBI without sepsis (n = 942)	p
Poor neurologic outcome [n (%)]	148/942 vs. 102/942	< 0.001
28-day mortality [n (%)]	212/942 vs. 130/942	< 0.001

IPWRA model (ATE on 28-day mortality)
Variables	ATE (95% CI)	p
No sepsis	Ref
Sepsis	0.073 (0.048 to 0.099)	< 0.001

In the PSM model, the imbalanced variables including Age over 65, initial white blood cell count, creatinine level, SOFA on ICU admission, Hypertension, Diabetes and vasopressor use in first 3 days were included in the sepsis assignment model (propensity score model), and confounders including Age over 65, initial white blood cell count, creatinine level, SOFA on ICU admission, Hypertension, Diabetes and vasopressor use in first 3 days were included in the inverse probability weighted regression model.
IPWRA inverse probability weighted regression model, PSM Propensity score matching, ATE estimate average treatment effect, OR odds ratio, 95% CI 95% confidence interval

ISSN: 2047-783X