Suggested Electronic Appendix

SUGGESTED ELECTRONIC APPENDIX

Table S1. Baseline characteristics of the study sample and the ESA treated population

Table S2: The association of RDW and all-cause mortality in 723 ESA untreated kidney transplant recipients estimated with Fine-Gray competing-risks analysis, where the death with a functioning graft was the event of interest and graft failure before death was considered as a competing event

Table S3: Association of RDW and all-cause mortality in the total sample (723 ESA-untreated + 83 ESA-treated kidney transplant recipients) estimated with Cox-regression models

Table S4: Detailed report of sensitivity, specificity, positive predictive and negative predictive values at different cut-off points of RDW in 723 ESA untreated kidney transplant recipients

Table S5: Tests of discriminatory value, Goodness of Fit, and Calibration in 723 ESA untreated kidney transplant recipients.

Table S6: Classification tables of predicted risk for model1 and model1 + RDW in 723 ESA untreated kidney transplant recipients

Legend of figures

Figure S1: Flow chart of patients’ selection

Figure S1:

Table S1. Baseline characteristics of the study sample and the ESA treated population

patients without ESA treatment (n=723) / patients with ESA treatment
(n=82) / *p value
RDW (%) / 13.8±1.0 / 15.1±1.5 / <0.001
Mortality (%) / 11 / 26 / <0.001
Age (years) / 51±13 / 51±14 / 0.407
Gender, female (%) / 41 / 70 / <0.001
BMI (kg/m²) / 27.4±4.8 / 25.2±5.4 / <0.001
Abdominal circumference (cm) / 99±14 / 92±15 / <0.001
Time since last transplant (month) / 84 (75) / 92 (71) / 0.461
Previous time on dialysis (month) / 19 (29) / 20 (26) / 0.655
Total ESRD time (month) / 119 (87) / 132 (103) / 0.046
eGFR (CKD-EPI) (ml/min/ 1.73 m²) / 54.3±20.0 / 31.0±15.8 / <0.001
Charlson Comorbidity index / 2 (1) / 2 (2) / 0.114
Presence of hypertension (%) / 93 / 98 / 0.085
Presence of diabetes (%) / 21 / 22 / 0.813
Presence of coronary hearth disease (%) / 8 / 2 / 0.085
Smoking (%) / 19 / 15 / 0.305
Hgb (g/L) / 138±15 / 117±13 / <0.001
Ferritin (ng/mL) / 120 (195) / 185 (334) / 0.011
Transferrin (g/L) / 2.37±0.44 / 2.08±0.52 / <0.001
Sol. transferrin receptor (mg/L) / 3.42±1.41 / 4.09±2.73 / <0.001
Serum iron (µmol/L) / 17.4±9.2 / 15.1±12.4 / <0.001
Total iron binding capacity (µmol/L) / 51.8±12.2 / 44.5±13.2 / <0.001
Serum albumin (g/L) / 40.9±3.7 / 38.2±4.4 / <0.001
CRP (mg/L) / 3.0 (4.9) / 2.7 (6.5) / 0.250
Use of ACE inhibitor or ARB (%) / 31 / 50 / <0.001
Iron or folic acid supplementation (%) / 23 / 70 / <0.001
Primary cause of ESRD (%) / 0.635
chronic GN / 24 / 18
chronic TIN / 13 / 17
PKD / 17 / 18
diabetic nephropathy / 5 / 6
hypertensive nephropathy / 6 / 2
others or unknown / 35 / 39
Cold ischemic time (minute) / 1235±356 / 1272±238 / 0.816
History of delayed graft function (%) / 25 / 36 / 0.035
History of acute rejection (%) / 32 / 53 / <0.001
HLA mismatches (%) / 0.539
0 / 1 / 0
1 / 6 / 1
2 / 22 / 23
3 / 45 / 50
4 / 21 / 22
5 / 4 / 4
6 / 1 / 0
Immunsuppression
Steroid use (%) / 79 / 82 / 0.498
Cyclosporine use (%) / 43 / 30 / 0.027
Tacrolimus use (%) / 47 / 56 / 0.114
Azathioprine use (%) / 3 / 0 / 0.128
MMF use (%) / 82 / 73 / 0.055
Sirolimus use (%) / 7 / 11 / 0.179
Everolimus use (%) / 2 / 5 / 0.065
Abbreviations: ESA: erythropoiesis stimulating agent; RDW, red blood cell distribution width; ESRD, end stage renal disease; eGFR, estimated GFR; CKD-EPI, Chronic Kidney Disease – Epidemiology Collaboration; Hgb, hemoglobin; CRP, C-reactive protein; ACE, angiotensin-converting enzyme; ARB, angiotensin-receptor blocker; GN, glomerulonephritis; TIN, tubulo-interstitial nephritis; PKD, polycystic kidney disease; HLA, human leukocyte antigen; MMF, mycophenolat mofetil.
Values are in median with interquartile range or in mean±SD.
*p value: comparing the subgroups of patients treated or not treated with ESA.

Table S2: The association of RDW and all-cause mortality in 723 ESA untreated kidney transplant recipients estimated with Fine-Gray competing-risks analysis, where the death with a functioning graft was the event of interest and graft failure before death was considered as a competing event

unadjusted / +eGFR, age and gender / +iron markers / +inflammatory markers / Final model
SHR / 95% CI / SHR / 95% CI / SHR / 95% CI / SHR / 95% CI / SHR / 95% CI
RDW (1 % increase) / 1.63 / 1.39 to 1.92 / 1.50 / 1.26 to 1.80 / 1.56 / 1.27 to 1.92 / 1.52 / 1.23 to 1.90 / 1.60 / 1.24 to 2.06
RDW (upper median) / 2.73 / 1.67 to 4.47 / 2.02 / 1.20 to 3.42 / 1.88 / 1.09 to 3.23 / 1.55 / 0.89 to 2.73 / 1.33 / 0.73 to 2.41
After a median 35 months of follow-up time the number of death with functioning graft were 66 and the number of graft failure were 56. Independent variables in the final model were age, gender, eGFR, iron markers: (serum iron, soluble transferrin receptor, hemoglobin), inflammatory markers: (serum albumin, C-reactive protein, abdominal circumference), Charlson Comorbidity index, total time in ESRD, steroid use, mammalian target of rapamycin (mTOR) use, ACEi or ARB use, iron and folic acid supplementation.
Abbreviations:
ESA: erythropoiesis stimulating agent; eGFR, estimated GFR; Hb, hemoglobin; ESRD, end stage renal disease; ACEi, angiotensin converting enzyme inhibitor; ARB, angiotensin receptor blocker; SHR, subhazard ratio; CI, confidence interval; RDW, red blood cell distribution width.

Table S3: Association of RDW and all-cause mortality in the total sample (723 ESA-untreated + 83 ESA-treated kidney transplant recipients) estimated with Cox-regression models

unadjusted / +eGFR, age and gender / +iron markers / +inflammatory markers / Final model
HR / 95% CI / HR / 95% CI / HR / 95% CI / HR / 95% CI / HR / 95% CI
RDW (1 % increase) / 1.46 / 1.32 to 1.60 / 1.43 / 1.27 to 1.60 / 1.35 / 1.17 to 1.57 / 1.37 / 1.17 to 1.61 / 1.60 / 1.22 to 1.75
RDW upper median / 3.28 / 2.04 to 5.26 / 2.35 / 1.43 to 3.84 / 1.98 / 1.19 to 3.31 / 1.67 / 1.00 to 2.83 / 1.40 / 0.82 to 2.38
After 35 months of follow-up time the total number of deaths was 101. Independent variables in the final model were age, gender, eGFR, iron markers (serum iron, soluble transferrin receptor, hemoglobin, iron and folic acid supplementation), inflammatory markers (serum albumin, C-reactive protein, abdominal circumference), Charlson Comorbidity index, total time in ESRD, steroid use, mammalian target of rapamycin (mTOR) use, ACEi or ARB use.
Abbreviations:
ESA: erythropoiesis stimulating agent; eGFR, estimated GFR; Hb, hemoglobin; ESRD, end stage renal disease; ACEi, angiotensin converting enzyme inhibitor; ARB, angiotensin receptor blocker; HR, hazard ratio; CI, confidence interval; RDW, red blood cell distribution width.

Table S4: Detailed report of sensitivity, specificity, positive predictive and negative predictive values at different cut-off points of RDW in 723 ESA untreated kidney transplant recipients

Cutpoint / Sensitivity / Specificity / Correctly classified / Positive predictive value / Negative predictive value
( >= 13.5 ) / 77.78% / 41.90% / 45.92% / 14.45% / 93.73%
( >= 13.6 ) / 75.31% / 47.51% / 50.62% / 15.33% / 93.85%
( >= 13.7 ) / 72.84% / 53.43% / 55.60% / 16.48% / 93.97%
( >= 13.8 ) / 69.14% / 58.10% / 59.34% / 17.23% / 93.72%
( >= 13.9 ) / 66.67% / 60.75% / 61.41% / 17.65% / 93.53%
( >= 14.0 ) / 62.96% / 64.80% / 64.59% / 18.41% / 93.27%
( >= 14.1 ) / 59.26% / 68.22% / 67.22% / 19.05% / 92.99%
( >= 14.2 ) / 51.85% / 73.36% / 70.95% / 19.72% / 92.35%
( >= 14.3 ) / 49.38% / 76.48% / 73.44% / 20.94% / 92.29%
( >= 14.4 ) / 44.44% / 78.35% / 74.55% / 20.57% / 91,79%
( >= 14.5 ) / 37.04% / 81.93% / 76.90% / 20.55% / 91,16%
( >= 14.6 ) / 35.80% / 83.96% / 78.56% / 21.97% / 91.20%
( >= 14.7 ) / 34.57% / 85.51% / 79.81% / 23.14% / 91.20%
( >= 14.8 ) / 33.33% / 86.60% / 80.64% / 23.89% / 91.15%
( >= 14.9 ) / 32.10% / 88.63% / 82.30% / 26.26% / 91.19%
( >= 15.0 ) / 28.40% / 89.88% / 82.99% / 26.14% / 90.87%
Abbreviations: RDW, red blood cell distribution width; ESA: erythropoiesis stimulating agent

Table S5: Tests of discriminatory value, Goodness of Fit, and Calibration in 723 ESA untreated kidney transplant recipients.

Estimate / 95% CI / p value
Adjusted Hazard ratio of RDW (1 % increase) in Cox proportional hazards model / 1.46 / 1.17 / 1.82 / <0.001
Adjusted Odds ratio of RDW (1% increase) in logistic regression model / 1.54 / 1.18 / 2.01 / <0.001
Goodness of Fit - Cox regression
Model1 / C-statistic / 0.789
AIC / 930
LR chi2 / 91.7 / <0.001
Model1 + RDW / C-statistic / 0.794
AIC / 922
LR chi2 / 101.7 / <0.001
LR-test / 10.01 / <0.001
Goodness of Fit - Logistic regression
Model1 / AIC / 427
LR chi2 / 94.3 / <0.001
Model1 + RDW / AIC / 419
LR chi2 / 104.3 / <0.001
LR-test / 9.96 / <0.001
Test of discriminatory value
IDI / 0.021 / 0.028
NRI (cut-off 5%, 15% ) / 0.117 / 0.002
NRI (cut-off 5%, 30% ) / 0.118 / 0.003
NRI (cut-off 5%, 15%, 30% ) / 0.189 / <0.001
Variables in model1 were age, gender, eGFR, soluble transferrin receptor, hemoglobin, serum albumin, C-reactive protein, abdominal circumference, Charlson Comorbidity index, total time in ESRD.
Abbreviations: ESA: erythropoiesis stimulating agent; RDW, red blood cell distribution width; ESRD, end stage renal disease; eGFR, estimated GFR; Hosmer-Lemeshow tests. CI, confidence interval; AIC, Akaike Information Criterion; LR, likelihood ratio; IDI, integrated discrimination improvement; NRI, net reclassification improvement (risk classification tables corresponding to NRI are presented in Table S6).

Table S6. Classification tables of predicted risk for model1 and model1 + RDW in 723 ESA untreated kidney transplant recipients

Non-cases / Predicted risks from model A + RDW
Low / Mid-low / Mid-high / High / Total
Predicted risks from model A / Low / 248 / 16 / 267
Mid-low / 34 / 199 / 13 / 1 / 247
Mid-high / 22 / 61 / 4 / 87
High / 6 / 22 / 28
Total / 282 / 237 / 80 / 27 / 626
Cases / Predicted risks from model A + RDW
Low / Mid-low / Mid-high / High / Total
Predicted risks from model A / Low / 6 / 2 / 8
Mid-low / 17 / 5 / 22
Mid-high / 1 / 18 / 6 / 25
High / 1 / 24 / 25
Total / 6 / 20 / 24 / 30 / 80
Data in the table represent numbers of persons in classes of predicted risk from model A and model A + RDW, separately for non-cases and cases. Variables in model1 were age, gender, eGFR, soluble transferrin receptor, hemoglobin, serum albumin, C-reactive protein, abdominal circumference, Charlson Comorbidity index, total time in ESRD.
Probability Cut-off points were 0.05, 0.15 and 0.30. Low: < 5%, Mid-low: 5-14%, Mid-high: 15-29%, High: ≥30%
Abbreviations: ESA, erythropoiesis stimulating agent; RDW, red blood cell distribution width.