Table 4 |.

Univariate and multivariable logistic regression model predicting AKI while on DAA therapies

	Univariate model		Multivariable model
Baseline Predictors of AKI on DAAs	Odds ratio (95% CI)	P	Odds ratio (95% CI)	P
Age, per 10 yr	1.14 (0.81–1.6)	0.46
Sex (female vs. male)	0.70 (0.31–1.6)	0.38
Race (white vs. nonwhite)	0.85 (0.39–1.9)	0.69
Diabetes (vs. nondiabetic)	1.50 (0.69–3.3)	0.31
Cirrhosis (vs. noncirrhotic)	2.93 (1.3–6.5)	0.009	2.32 (1.0–5.4)	0.051
Hypertension (vs. nonhypertensive)	2.19 (1.01–4.8)	0.048	1.62 (0.73–3.7)	0.24
HIV co-infection	0.38 (0.07–2.0)	0.26	0.43 (0.08–2.3)	0.33
Ribavirin use	1.01 (0.46–2.2)	0.99		—
Prior treatment failure	1.42 (0.68–3.0)	0.35
DAA course length (>12 wk vs. ≤12 wk)	2.03 (0.94–4.4)	0.07	1.61 (0.73–3.6)	0.24
Baseline eGFR < 60 ml/min per 1.73 m2	3.25 (1.4–7.6)	0.007	3.02 (1.3–7.3)	0.014

AKI, acute kidney injury; CI, confidence interval; DAA, direct-acting antiviral; eGFR, estimated glomerular filtration rate.

The multivariable model includes demographics and predictors with P < 0.3 in the univariate model. Only presence of cirrhosis and baseline eGFR < 60 ml/min per 1.73 m² predicted the development of AKI while on DAAs in multivariable models. Prolonged DAA course was defined as >12 weeks of treatment.