Electronic supplement Material: icm_2009-01128.R2
Do Hypooncotic Fluids for Shock Increase the Risk of late-onset Acute Respiratory Distress Syndrome?
Frédérique Schortgen MD, Emmanuelle Girou PharmD PhD, Nicolas DeyeMD, Laurent Brochard MD, for the CRYCO Study Group
Development of different models of logistic regression exploring the impact of the time for fluid volume collection on the association between late-onset ARDS occurrence and the volume and type of fluids administered
The association between fluids’ volumes and ARDS occurrence might vary according to the time for fluid collection. We therefore tested different ways for fluid volume calculation. In this analysis, types and volumes of fluids were recorded up to 8 days following inclusion (or up to earlier ICU discharge) for patients who did not developed ARDS and up to ARDS occurrence (or 8 days max in case of later occurrence of ARDS) in other wise. For each patient the cumulative volume of each fluid was calculated and divided by the number of days in the ICU (maximum 8 days) for the control group or dived by the number of days in the ICU before ARDS development (maximum 8 days). Volumes are therefore indicated as Liter or mL/patient/day in tables E1-E4. This different time point for fluid collection did not modify the results on the association between fluid volumes and ARDS development.
Table E1: Risk factors for developing late-onsetARDS: univariate analysis
Free oflate-onsetARDS
(N=824) / late-onsetARDS occurrence
(N=81) / p Value
Patient characteristics
Age, yr / 60±18 / 55±18 / 0.008
Male gender, n (%) / 516 (63) / 49 (60) / 0.706
History of heart failure NYHA III-IV, n (%) / 110 (13) / 6 (7) / 0.162
Ultimately or rapidly fatal underlying disease, n (%) / 302 (37) / 33 (41) / 0.430
Immunosuppression, n (%) / 102 (12) / 8 (10) / 0.596
Reason for admission, n (%) / 0.002
Medical / 378 (48) / 50 (66)
Surgical / 417 (52) / 26 (34)
Sepsis as indication for fluid resuscitation, n (%) / 286 (35) / 45 (55) / <0.0001
Patients receiving fluid therapy 24 hours before T0, n (%) / 463 (56) / 40 (50) / 0.212
Severity at initiation of fluid for shock resuscitation (T0)
SAPS II / 46±20 / 53±18 / <0.001
Number of organ dysfunctions# / 3.1±1.2 / 3.4±1.1 / 0.010
Renal dysfunction, n (%) # / 619 (75) / 66 (81) / 0.203
Hepatic dysfunction, n (%) # / 174 (21) / 22 (27) / 0.208
Need for high-dose vasoactive drugs, n (%) § / 161 (20) / 24 (30) / 0.032
Need for mechanical ventilation, n (%) / 581 (71) / 57 (70) / 0.915
Hypoxaemia (Pa O2/Fi O2<200 or O2>5 L/min), n (%) / 385/805 (48) / 53/79 (67) / 0.001
Chest X-Ray score, points (max 4)* / 0.65±1.01 / 1.38±1.40 / <0.0001
Type of fluid administered for volume expansion
Colloids
Yes
No / 713 (87)
111 (13) / 70 (86)
11 (14) / 0.978
Hypooncotic fluids only $
Yes
No / 362 (44)
462 (56) / 44 (54)
37 (46) / 0.07
20-25 % Albumin, n (%)
Yes
No / 69 (8)
755 (92) / 9 (11)
72 (89) / 0.402
Fluid management, median IQR
Proportion of hypooncotic fluids used for volume expansion, %/patient/day $$ / 28 (11-52) / 33 (17-49) / 0.199
Volume of crystalloids, L/patient/day / 0.6 (0.2-1.5) / 1.0 (0.5-2.4) / 0.001
Volume of colloids, L/patient/day / 0.3 (0.1-0.7) / 0.8 (0.4-1.3) / <0.0001
Mean daily fluid balance, L/patient/day / 1.7 (1.1-2.8) / 3.5 (1.8-6.1) / <0.0001
Lowest CVP, mmHg ** / 6 (3-9) / 7 (4-11) / 0.0001
Transfusions,
Any transfusion, n (%)
Yes
No
Volumes, median IQR: / 453 (55)
371 (45) / 46 (57)
35 (43) / 0.754
Red blood cells, mL/patient/day / 0 (0-150) / 80 (0-400) / 0.011
Plasma , mL/patient/day / 0 (0-80) / 0 (0-450) / 0.040
Total transfusions, mL/patient/day / 60 (0-260) / 200 (0-820) / 0.003
Variables were expressed as mean±SD, unless otherwise indicated.
ARDS, acute respiratory distress syndrome; NYHA, New York Heart Association; SAPS II, Simplified Acute Physiology Score II;; CVP, central venous pressure
#Organ dysfunction was defined as a value higher than 0 for each organ among the six organs included in the LOD system (i.e., neurologic, hemodynamic, pulmonary, renal, hematologic and hepatic). The number of organ dysfunctions ranged from 0 to 6 [26].
§High-dose vasoactive drugs was defined as dopamine >15 µg/kg/min or either epinephrine or norepinephrine >0.1 µg/kg/min at the time of fluid resuscitation initiation (T0) [27].
*From 0 to 4 points [28]
$Crystalloids and/or gelatines and/or 4% albumin
$$Proportion of crystalloids and hypooncotic colloids (i.e., gelatines and/or 4% albumin) over all plasma expanders administered for volume expansion
**At least one central venous pressure measurement was available in 611 patients.
Table E2:Results of logistic regression for late-onset ARDS occurrence: impact of the use of hypooncotic fluids for shock resuscitation
Risk Factor / Odds Ratio (95% CI) / p ValueAge, per y / 0.98 (0.97-0 99) / 0.032
SAPS II, per point / 1.00 (0.99-1.02) / 0.669
Medical reason for admission, n / 1.06 (0.56-2.01) / 0.850
Sepsis as reason for fluid resuscitation, n / 2.22 (1.91-4.15) / 0.012
Chest X-Ray score, per point / 1.60 (1.29-1.98) / <0.001
Need for high-dose vasoactive drugs, n / 1.10 (0.55-2.11) / 0.827
Positive fluid balance, per L/day / 1.29 (1.17-1.42) / <0.001
Use of hypooncotic fluids only for volume expansion $ / 1.42 (0.80-2.50) / 0.231
Transfusion volume, per L/day / 1.65 (1.10-2.46) / 0.015
CI, confidence interval
SAPS II, Simplified Acute Physiology Score II;
$Crystalloids and/or gelatines and/or 4% albumin
The goodness-of-fit chi-square test was 4.94 (p=0.76).
Table E3: Results of logistic regression for late-onset ARDS occurrence: impact of the proportion of hypooncotic fluids used for volume expansion
Risk Factor / Odds Ratio (95% CI) / P ValueAge, per y / 0.98 (0.97-0.99) / 0.037
SAPS II, per point / 1.00 (0.99-1.02) / 0.700
Medical reason for admission, n / 1.16 (0.62-2.17) / 0.639
Sepsis as reason for fluid resuscitation, n / 2.25 (1.21-4.19) / 0.011
Chest X-Ray score, per point / 1.60 (1.29-1.97) / <0.0001
Need for high-dose vasoactive drugs, n (%) / 1.07 (0.55-2.09) / 0.844
Positive fluid Balance, per L/day / 1.28 (1.17-1.41) / <0.0001
Proportion of hypooncotic fluids used for volume expansion, per %/day $ / 1.27 (0.40-4.01) / 0.682
Total transfusion volume, per L/day / 1.68 (1.12-2.51) / 0.012
CI, confidence interval
SAPS II, Simplified Acute Physiology Score II;
$proportion of crystalloids and hypooncotic colloids (i.e., gelatines and/or 4% albumin) over all plasma expanders administered for volume expansion
The goodness-of-fit chi-square test was 2.85 (p=0.94).
Table E4: Results of logistic regression for late-onset ARDS occurrence: impact of colloid volume versus crystalloid volume
Risk Factor / Odds Ratio (95% CI) / p ValueAge, per y / 0.99 (0.97-1.00) / 0.066
SAPS II, per point / 1.01 (0.99-1.02) / 0.338
Medical reason for admission, n / 1.53 (0.82-2.83) / 0.178
Sepsis as reason for fluid resuscitation, n / 1.94 (1.05-3.57) / 0.033
Chest X-Ray score, per point / 1.56 (1.26-1.93) / <0.0001
Need for high-dose vasoactive drugs, n (%) / 1.08 (0.57-2.07) / 0.805
Total transfusion volume, per L / 1.88 (1.27-2.79) / 0.002
Volume of crystalloids, per L / 1.31 (1.12-1.53) / 0.001
Volume of colloids, per L / 2.12 (1.45-3.10) / <0.0001
CI, confidence interval
SAPS II, Simplified Acute Physiology Score version II;
The goodness-of-fit chi-square test was 7.83 (p=0.45).
Table E5: Results of logistic regression for late-onset ARDS occurrencein 748 patients: impact of the proportion of hypooncotic fluids used for volume expansion
Risk Factor / Odds Ratio (95% CI) / P ValueAge, per y / 0.98 (0.97-0.99) / 0.008
SAPS II, per point / 1.01 (0.99-1.02) / 0.223
Medical reason for admission, n / 1.31 (0.71-2.40) / 0.381
Sepsis as reason for fluid resuscitation, n / 1.88 (1.05-3.37) / 0.033
Chest X-Ray score, per point / 1.54 (1.26-1.89) / <0.0001
Need for high-dose vasoactive drugs, n (%) / 1.17 (0.62-2.17) / 0.631
Positive fluid Balance, per L / 1.05 (1.02-1.09) / 0.002
Proportion of hypooncotic fluids used for volume expansion, per % $ / 1.01 (0.99-1.01) / 0.103
Total transfusion volume, per L / 1.14 (1.01-1.29) / 0.033
CI, confidence interval
SAPS II, Simplified Acute Physiology Score II;
$proportion of crystalloids and hypooncotic colloids (i.e., gelatines and/or 4% albumin) over all plasma expanders administered for volume expansion
The goodness-of-fit chi-square test was 7.91 (p=0.44). The area under ROC curve was 0.78.
Table E6. Results of logistic regression for ARDS development in 784 patients: impact of colloid volume versus crystalloid volume
Risk Factor / Odds Ratio (95% CI) / p ValueAge, per y / 0.98 (0.97-0.99) / 0.011
SAPS II, per point / 1.02 (1.00-1.03) / 0.038
Medical reason for admission, n / 1.64 (0.96-3.02) / 0.115
Sepsis as reason for fluid resuscitation, n / 1.76 (0.99-3.13) / 0.055
Chest X-Ray score, per point / 1.54 (1.26-1.89) / <0.0001
Need for high-dose vasoactive drugs, n (%) / 1.16 (0.62-2.15) / 0.641
Total transfusion volume, per L / 1.17 (1.03-1.32) / 0.013
Volume of crystalloids, per L / 1.24 (1.09-1.41) / 0.001
Volume of colloids, per L / 1.05 (1.00-1.10) / 0.023
CI, confidence interval
SAPS II, Simplified Acute Physiology Score version II;
The goodness-of-fit chi-square test was 4.85 (p=0.77). The area under the ROC curve was 0.76.
1