Supplemental data:

The use of erythropoietin stimulating agents in renal transplant recipients with hemoglobin above 12.5 g/dl is associated with elevated mortality

G. Heinze1, A. Kainz2, W. H. Hörl2, R. Oberbauer2,3,4

1- Core Unit of Medical Statistics and Informatics, Medical University of Vienna, Austria
2- Department of Nephrology, Medical University of Vienna, Austria
3- Department of Nephrology, KH Elisabethinen, Linz, Austria
4- Austrian Dialysis and Transplant Registry, Austria


All supplemental data are available for download as a single PDF document.

FIGURES
Webfigure 1 Webfigure 1

Adjusted hazard ratio for ESA users (top) and non-users (bottom) at various hemoglobin levels vs. a reference level of 12.5 g/dl. Gray: original analysis (excluding deaths up to day 90), black: analysis including deaths up to day 90.
Webfigure 2 Webfigure 2

Adjusted hazard ratios of ESA users vs. non-users at various hemoglobin levels. Gray: original analysis (excluding deaths up to day 90), black: analysis including deaths up to day 90.
Webfigure 3 Webfigure 3

Adjusted hazard ratio for ESA users (top) and non-users (bottom) at various hemoglobin levels vs. a reference level of 12.5 g/dl. Gray: original analysis (multiple imputation; N=1794, 286 events), black: analysis using complete cases only (N=1386, 208 events).
Webfigure 4 Webfigure 4

Adjusted hazard ratios of ESA users vs. non-users at various hemoglobin levels. Gray: original analysis (multiple imputation; N=1794, 286 events), black: analysis using complete cases only (N=1386, 208 events).
Webfigure 5 Webfigure 5

Adjusted hazard ratio for ESA users (top) and non-users (bottom) at various hemoglobin levels vs. a reference level of 12.5 g/dl. Gray: original analysis (multiple imputation; N=1794, 286 events), black: analysis after multiple imputation of artificially generated nonrandomly missing data.
Webfigure 6 Webfigure 6

Adjusted hazard ratios of ESA users vs. non-users at various hemoglobin levels. Gray: original analysis (multiple imputation; N=1794, 286 events), black: analysis after multiple imputation of artificially generated nonrandomly missing data.
Webfigure 7 Webfigure 7

Adjusted hazard ratio for ESA users (top) and non-users (bottom) at various hemoglobin levels vs. a reference level of 12.5 g/dl. Gray: original analysis (multiple imputation; N=1794, 286 events), black: analysis including time-dependent effects for cholesterol level and number of antihypertensive drugs.
Webfigure 8 Webfigure 8

Adjusted hazard ratios of ESA users vs. non-users at various hemoglobin levels. Gray: original analysis (multiple imputation; N=1794, 286 events), black: analysis after including time-dependent effects for cholesterol level and number of antihypertensive drugs.
Webfigure 9 Webfigure 9

Adjusted hazard ratio for ESA users (top) and non-users (bottom) at various hemoglobin levels vs. a reference level of 12.5 g/dl. Gray: original analysis (multiple imputation; N=1794, 286 events), black: analysis including nonlinear effects for cholesterol level.
Webfigure 10 Webfigure 10

Adjusted hazard ratios of ESA users vs. non-users at various hemoglobin levels. Gray: original analysis (multiple imputation; N=1794, 286 events), black: analysis after including nonlinear effects for cholesterol level.
Webfigure 11 Webfigure 11

Adjusted hazard ratio for ESA users (top) and non-users (bottom) at various hemoglobin levels vs. a reference level of 12.5 g/dl. Gray: original analysis (multiple imputation; N=1794, 286 events, follow-up until 31 Dec 2004), black: analysis including follow-up data until 31 Dec 2008 (313 events).
Webfigure 12 Webfigure 12

Adjusted hazard ratios of ESA users vs. non-users at various hemoglobin levels. Gray: original analysis (multiple imputation; N=1794, 286 events, follow-up until 31 Dec 2004), black: similar, but follow-up until 31 Dec 2008 (313 events)
Webfigure 13 Webfigure 13

Adjusted hazard ratio for ESA users (top) and non-users (bottom) at various hemoglobin levels vs. a reference level of 12.5 g/dl. Gray: original analysis (multiple imputation; N=1794, 286 events, follow-up until 31 Dec 2004), black: analysis excluding cardiovascular disease variables from final model.
Webfigure 14 Webfigure 14

Adjusted hazard ratios of ESA users vs. non-users at various hemoglobin levels. Gray: original analysis (multiple imputation; N=1794, 286 events, follow-up until 31 Dec 2004), black: analysis excluding cardiovascular disease variables
Webfigure 15 Webfigure 15

Adjusted hazard ratio for ESA users (top) and non-users (bottom) at various hemoglobin levels vs. a reference level of 12.5 g/dl. Gray: original analysis (multiple imputation; N=1794, 286 events, follow-up until 31 Dec 2004), black: analysis with ESA users left in ESA user group even after the time point where treatment is stopped.
Webfigure 16 Webfigure 16

Adjusted hazard ratios of ESA users vs. non-users at various hemoglobin levels. Gray: original analysis (multiple imputation; N=1794, 286 events, follow-up until 31 Dec 2004), black: analysis with ESA users left in ESA user group even after the time point where treatment is stopped.
Webfigure 17 Webfigure 17

Adjusted hazard ratio for ESA users (top) and non-users (bottom) at various hemoglobin levels vs. a reference level of 12.5 g/dl. Gray: original analysis of patient survival (multiple imputation; N=1794, 286 events, follow-up until 31 Dec 2004), black: analysis of transplant survival (graft loss and death counted as event; N=1794, 367 events).
Webfigure 18 Webfigure 18

Adjusted hazard ratios of ESA users vs. non-users at various hemoglobin levels. Gray: original analysis of patient survival (multiple imputation; N=1794, 286 events, follow-up until 31 Dec 2004), black: analysis of transplant survival (graft loss and death counted as event; N=1794, 367 events).
Webfigure 19 Webfigure 19

Adjusted hazard ratio for ESA users (top) and non-users (bottom) at various hemoglobin levels vs. a reference level of 12.5 g/dl. Gray: original analysis of patient survival (multiple imputation; N=1794, 286 events, follow-up until 31 Dec 2004), black: analysis additionally adjusting for primary indication for transplantation (diabetes, immune mediated, PCKD, or other).
Webfigure 20 Webfigure 20

Adjusted hazard ratios of ESA users vs. non-users at various hemoglobin levels. Gray: original analysis of patient survival (multiple imputation; N=1794, 286 events, follow-up until 31 Dec 2004), black: analysis additionally adjusting for primary indication for transplantation (diabetes, immune mediated, PCKD, or other).