Table 2.

Bivariate and Multivariable Predictors of Predialysis Systolic Blood Pressure in Prevalent Hemodialysis Patients

Variable	Bivariate Parameter Estimate of Slope (mm Hg)	P	Multivariable Parameter Estimate of Slope (mm Hg)	P
%IDWG (/1% increase)	1.00 ± 0.24	<0.0001	1.00 ± 0.24	<0.0001
Dry weight (/1 kg increase)	0.29 ± 0.12	<0.0001	0.26 ± 0.12	<0.0001
Left ventricular hypertrophy	8.2 ± 3.4	<0.0001	8.9 ± 1.8	<0.0001
Hypertension	11.7 ± 6.6	0.0005	7.2 ± 6.0	0.02
Diabetes as cause of ESRD	6.9 ± 3.7	0.0003	2.7 ± 3.7	0.2
Erythropoietin use	10.1 ± 6.1	0.001	7.7 ± 5.5	0.006
DM	5.1 ± 3.5	0.004	*	Not significant
Black race (v nonblack)	5.2 ± 3.4	0.004	3.7 ± 4.0	0.08
Hispanic ethnicity	11.1 ± 8.8	0.01	11.5 ± 6.6	0.0007
Arrhythmia	−7.9 ± 4.9	0.002	−6.9 ± 5.0	0.008
Creatinine (/1 mg/dL increase)	−0.4 ± 0.41	0.06	−0.46 ± 0.43	0.04
Age (/1 y increase)	−0.05 ± 0.12	0.5	0.03 ± 0.12	0.6
Treatment group	2.0 ± 3.5	0.3	1.4 ± 3.4	0.4
Interaction terms
%IDWG * DM as cause of ESRD				0.04
%IDWG * creatinine				0.0002
%IDWG * age				0.04
%IDWG * dry weight				0.6
%IDWG * black race				0.3

Note: Values expressed as number ± 95% confidence interval. Nonsignificant predictors of systolic blood pressure in bivariate analyses included sex; duration of dialysis therapy; tobacco use; antihypertensive class (angiotensin-converting enzyme inhibitor, β-blocker, calcium channel blocker); history of bilateral nephrectomy, coronary artery disease, congestive heart disease, cerebrovascular disease, chronic obstructive pulmonary disease, or peripheral vascular disease; serum albumin level; or a history of medical noncompliance. Because of missing data, 408 patients and 28,803 dialysis sessions are included in the multivariate analysis.

Abbreviations: %IDWG, percentage of interdialytic weight gain; DM, diabetes mellitus; ESRD, end-stage renal disease.

^*Removed from the final multivariate model with P > 0.10. Treatment group, DM as cause of ESRD, and age were forced into all final models.