Table 3.

Pre‑specified multivariable Cox regression models for the association between RKF decline rate and PD technique failure

Characteristic	Model 1		Model 2		Model 3
	HR (95% CI)	p-value	HR (95% CI)	p-value	HR (95% CI)	p-value
RKF decline rate (×10)	1.17 (1.03, 1.32)	0.012	1.18 (1.05, 1.33)	0.006	1.19 (1.05, 1.34)	0.006
RKF decline rate (×10) (quartiles)
Q1	Reference		Reference		Reference
Q2	2.26 (1.37, 3.74)	0.001	2.30 (1.38, 3.83)	0.001	2.35 (1.40, 3.93)	0.001
Q3	1.70 (1.02, 2.82)	0.040	1.84 (1.10, 3.07)	0.019	1.82 (1.09, 3.05)	0.022
Q4	2.37 (1.41, 3.97)	0.001	2.61 (1.55, 4.41)	<0.001	2.69 (1.57, 4.60)	<0.001
P for trend	0.007		0.002		0.002

Note: Model 1 adjusted for age and sex. Model 2 additionally adjusted for diabetes, hypertension, and cause of ESKD. Model 3 further adjusted for hemoglobin, albumin, eGFR, and SBP. Abbreviations: CI, confidence interval; HR, hazard ratio; RKF, residual kidney function; ESKD, end-stage kidney disease; eGFR, estimated glomerular filtration rate; SBP, systolic blood pressure. The RKF decline rate (×10) was derived from the absolute RKF slope multiplied by 10; each 1‑unit increase represents a 0.1 ml/min/1.73 m² /month faster decline