Renal effects of synthetic colloids R1 1
Supplementary Table 1: Patients exposed to nephrotoxic drugs
HES group
(n = 118) p value adjusted
p value Gelatin group
(n = 87) p value adjusted
p value Crystalloid group (n
= 141)
NSAID, n (%)a 17 (14) 0.167 0.833 12 (14) 0.267 1.000 12 (9)
Diuretics, n (%)b 99 (84) 0.270 1.000 76 (87) 0.082 0.491 110 (78)
ACE inhibitors, n (%)c 37 (31) 0.584 1.000 40 (46) 0.006 0.045 39 (28)
Vancomycin, n (%) 41 (35) 1.000 1.000 27 (31) 0.565 1.000 50 (36)
Aminoglycosides, n (%) d 21 (18) 0.389 1.000 7 (8) 0.284 1.000 19 (14)
Antimycotics, n (%) e 31 (26) 0.002 0.012 15 (17) 0.163 0.814 15 (11)
Iodinated contrast media, n (%) 59 (50) 0.059 0.352 54 (62) 1.000 1.000 87 (62)
HES hydroxyethyl starch, IQR interquartile range. P values were calculated with Fisher’s exact test. For p value adjustment the Bonferroni-Holm method was used.
a Nonsteroidal anti-inflammatory drugs, including ibuprofen and diclofenac b including mannitol and furosemide
c angiotensin-converting enzyme inhibitors, including ramipril, perindopril, lisinopril and enalapril d including tobramycin and gentamicin
e including amphotericin B and fluconazole
Renal effects of synthetic colloids R1 2
Supplementary Table 2: Serum creatinine at onset of renal replacement therapy HES group
(n = 42)
p value Gelatin group (n = 30)
p value Crystalloid group (n = 27)
Serum creatinine (µmol/l) at onset of renal replacement therapy, median [IQR]
238 (160-277)
0.146 182 (132-241)
0.991 190 (113-273)
HES hydroxyethyl starch, IQR interquartile range. P values were calculated with the Mann–Whitney test
Renal effects of synthetic colloids R1 3
Suppelmentary Table 3: Multiple logistic regression analysis with RRT as dependent binary variable.
n Adjusted OR (95%CI) p value
SAPS II (per point) 333 1.03 (1.01,1.05) < 0.001
Baseline creatinine: (per µmol/l) 333 1.010 (1.005,1.014) < 0.001 Cardiac / thoracic surgery: yes vs. no 333 3.8 (2.04,7.06) < 0.001
Antimycotics: yes vs. no 333 2.78 (1.35,5.72) 0.005
Vancomycin: yes vs. no 333 4.11 (2.19,7.70) < 0.001
Human albumin 20%: yes vs. no 333 2.20 (1.18,4.12) 0.014
Added after model selection:
Period effects:
Period: ref.=Crystalloids
Gelatin 4% 333 2.37 (1.11,5.06) 0.025
HES 6% 130/0.4 333 2.26 (1.10,4.67) 0.027
Dose-effect relationship:
Crystalloids (per ml/kg BW) 136 1.0018 (0.9996,1.0041) 0.102 Crystalloids – quadratic (per (ml/kg BW)²) 136 1.00000 (1.00000,1.00000) 0.310 HES 6% 130/0.4 (per ml/kg BW) 112 0.999 (0.993, 1.004) 0.651 HES 6% 130/0.4 – quadratic (per (ml/kg BW)²) 112 1.00000 (0.99999,1.00001) 0.817
Gelatin 4% (per ml/kg BW) 85 1.001 (0.991,1.011) 0.863
Gelatin 4% - quadratic (per (ml/kg BW)²) 85 1.00000 (0.99998,1.00003) 0.973
Renal effects of synthetic colloids R1 4
BW body weight, HES hydroxyethyl starch.
Variables considered in the analysis: age, liver cirrhosis, diabetes, baseline creatinine, baseline SAPS-II, cardiac / thoracic surgery, non-steroidal anti-inflammatory drugs, ACE inhibitors, aminoglycosides, antimycotics, vancomycin, ionidated contrast media, diuretics, human albumin 20%, period (HES 6% 130/0.4, Gelatin 4%, Crystalloids), crystalloids (linear and quadratic, per ml/kg BW), HES 6% 130/0.4 (linear and quadratic, per ml/kg BW), Gelatin 4% (linear and quadratic, per ml/kg BW).
Forward and backward stepwise multiple logistic regression analysis based on AIC (Akaike Information Criterion) was used to derive a
multivariable model where the Hosmer-Lemeshow goodness of fit test (C-test: p = 0.690, H-test: p = 0.093) and ROC curve analysis (AUC = 0.70) indicate an acceptable fit and discrimination, respectively. P values were obtained by Wald's test. After model selection a period effects of HES 6%
130/0.4 and Gelatin 4% as categorical variables with Crystalloids as reference category was studied; then the model was applied to the sub- cohorts and the variables for synthetic colloids and crystalloids were added correspondingly to investigate the dose-effect relationship.