Table 3.

Association between IRRT modality as a time-varying exposure and kidney recovery until 90 days

Variable	Univariable (HR, [95% CI])	P-value	Multivariable (aHR, [95% CI])	P-value
Conventional hemodialysis	Ref		Ref
Online hemodiafiltration	1.19 (0.77; 1.84)	0.44	1.18 (0.76; 1.84)	0.47
Age	0.99 (0.97; 0.999)	0.033	0.99 (0.97; 1.01)	0.51
Female sex	1.07 (0.71; 1.61)	0.75	1.32 (0.84; 2.08)	0.23
Charlson comorbidity score	0.95 (0.87; 1.02)	0.16	1.02 (0.92; 1.15)	0.68
Baseline eGFR	1.01 (1.00; 1.01)	0.007	1.01 (1.00; 1.02)	0.024
Initially started on CRRT	1.00 (0.83; 1.22)	0.98	0.66 (0.43; 1.02)	0.06
SOFA score at IRRT initiation	1.01 (0.96; 1.06)	0.70	1.01 (0.95; 1.07)	0.052
Admission diagnostic
Surgical	Ref		Ref
Medical	1.06 (0.86; 1.32)	0.57	1.27 (0.78; 2.06)	0.34
Context at ICU admission
Sepsis	1.23 (0.78; 1.93)	0.38	1.36 (0.84; 2.20)	0.21
Heart failure	0.93 (0.56; 1.55)	0.78	1.03 (0.60; 1.79)	0.91
Both	1.69 (0.83; 3.46)	0.15	1.98 (0.94; 4.18)	0.07

aHR, adjusted hazard ratio; CI, confidence interval; CRRT, continuous renal replacement; eGFR, estimated glomerular filtration rate; HR, hazard ratio; IRRT, intermittent renal replacement therapy; SOFA, Sequential Organ Failure Assessment.

Cox regression model for count processes with HDF exposure as a time-varying covariate. Two patients were excluded from the analysis for kidney recovery and loss to follow-up after a single IRRT session (N = 180).