Driving Pressure and Survival in ARDS-Amato-ESM-NEJM 2015

Supplementary Appendix

This appendix has been provided by the authors to give readers additional information about their work.

Supplement to: Amato MBP, Meade MO, Slutsky AS, et al. Driving pressure and survival in the acute respiratory distress syndrome. N Engl J Med 2015;372:747-55. DOI: 10.1056/NEJMsa1410639

1

Driving Pressure and Survival in Acute Respiratory Distress Syndrome

SUPPLEMENTARY WEB MATERIAL

Marcelo B. P. Amato, MD 1

Maureen O. Meade, MD, MSc 2

Arthur S. Slutsky, MD 3,4

Laurent Brochard, MD 3,4

Eduardo L.V. Costa, MD 1,5

David A. Schoenfeld, PhD 6

Thomas E. Stewart, MD 2

Matthias Briel, MD, MSc 2,7

Daniel Talmor, MD, MPH 8

Alain Mercat, MD 9

Jean-Christophe M. Richard, MD 10

Carlos R.R. Carvalho 1

Roy G. Brower, MD 11

1

Cardio-Pulmonary Department, Pulmonary Divison, Heart Institute (Incor), University of So Paulo, So Paulo, Brazil;

2 Departments of Clinical Epidemiology, Biostatistics and Medicine, McMaster University, Hamilton, Ontario, Canada;

3 Keenan Research Centre for Biomedical Science, St. Michaels Hospital, Toronto, Ontario, Canada;

4 Interdepartmental Division of Critical Care Medicine, and Department of Medicine, University of Toronto, Ontario, Canada;

5 Research and Education Institute, Hospital Sirio-Libans, So Paulo, Brazil

6 Massachusetts General Hospital Biostatistics Center, Harvard Medical School, Boston, MA;

7 Basel Institute for Clinical Epidemiology and Biostatistics, University Hospital Basel, Switzerland;

8 Department of Anesthesia, Critical Care, and Pain Medicine, Beth Israel Deaconess Medical Center, Harvard Medical School,

Boston, MA;

9 Department of Intensive Care and Hyperbaric Medicine, Angers University Hospital, Angers, France;

10 Emergency Department, General Hospital of Annecy, Annecy, France and INSERM UMR 955, Creteil, France;

11 Pulmonary and Critical Care Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland.

2

SUPPLEMENTARY WEB MATERIAL

This supplement has additional information on methods and results, organized as:

I) DESCRIPTION OF THE STUDIED POPULATION: (Tables S1-S2; Figure S1)

II) ADDITIONAL RESULTS / ANALYSIS REFERRED TO IN THE MAIN MANUSCRIPT:

1. Accounting for residual (intrinsic) heterogeneity across the trials (Figure S2)

2. Univariate analysis (Table S3)

3. Length of risk exposure and test of proportional hazards assumption (Tables S4-S6)

4. Sensitivity analysis for different estimates of baseline elastance (Figure S3)

5. Homogeneous P-risks across the trials (Figure S4)

6. Consistency of higher P-risks in the validation cohorts (Tables S7-S8)

7. Tidal volume predicts survival only if normalized to compliance (CRS) (Figure S5)

8. Survival in patients under protective ventilator settings (Figure S6)

9. P (but not VT) predicts Barotrauma after randomization (Figure S7)

10. Mediation analysis: more than grading the severity of lung disease

P strongly mediates survival, independently of baseline elastance of

respiratory system (Figures S8-S9)

VT and PEEP were not independent mediators (Figures S10-S11)

11. P consistently mediates survival benefits across/within trials (Table S9)

III) DETAILS ON STATISTICS AND METHODS:

1. Screening the dataset and compatibility analysis

2. Missing data

3. Double-stratification (used for the analysis shown in Figure 1, main text)

3

I) DESCRIPTION OF THE STUDIED POPULATION:

Please, refer to the tables and figures within the next pages

Tables S1-S2:

Studied cohorts and baseline patient characteristics recalculated from individual

patient data

Figure S1:

Overview of the results of randomization in each of the trials

4

Table S1(website): Studied cohorts and baseline patient characteristics recalculated from individual patient data. The trials in the first four rows were pooled and formed our hypothesis generation sample, used to elect a multivariate model for survival (Model-1, Table 1). The ARDSnetVT

study was used as a first validation sample. The studies in the last four rows (testing a higher vs. lower-PEEP strategy) were pooled and used as a second validation sample.

Lower vs. Higher

VT-trials :

Years of

recruitment

Patients

(N)

Randomization

Cont. / Treat.

Age

mean (SD)

Sepsis at

Entry (%)

Pneumonia/

Aspiration*

MV.Days

at entry

Interventions

(within treatment-arm)

Outcome Treatment-arm

(RR; 95%CI)

Amato et al.1 1991-1995 53 24 / 29 34 (13) 83% 28% 1 VT 6mL/kg; P 20cmH2O PPLAT 40cmH2O;

0.38 (0.180.79)

Stewart et al.2 1995-1996 118 59 / 59 59 (18) 40% 58% 0 VT 8mL/kg; PPEAK 30cmH2O

0.99 (0.601.70)

Brochard et al.3 1994-1996 113 57 / 56 57 (15) n.a. n.a. 2 VT < 10mL/kg; PPLAT 25cmH2O

1.28 (0.732.25)

Brower et al.4 1994-1996 52 26 / 26 48 (16) 23% 54% n.a. VT 8mL/kg; PPLAT 30cmH2O

1.11 (0.482.57)

ARDSnetVT5 1996-1999 861 429 / 432 51 (17) 27% 49% 1 VT 6mL/kg;

PPLAT 30cmH2O; 0.74 (0.580.93)

Higher vs. Lower

PEEP-trials :

ARDSnetPEEP6 1999-2002 545 271 / 274 51 (17) 38% 55% 1 Higher PEEP guided by

higher PEEP/ FIO2 table; VT = 6.00.9 mL/kg/pbw

1.11 (0.801.54) stopped for futility

EXPRESS7 2002-2005 767 382 / 385 60 (15) 61% 72% 1.5 Highest PEEP keeping PPLAT < 30cmH2O;

VT = 6.10.3 mL/kg/pbw

0.87 (0.691.09) vent. free days

stopped for futility

LOVS8 2000-2006 983 508 / 475 56 (17) 47% 64% 2 Higher PEEP guided by

higher PEEP/ FIO2 table; VT = 7.01.5 mL/kg/pbw

0.88 (0.711.08) refract. hypoxemia

Talmor et al.9 2004-2007 61 31 / 30 53 (20) 48% 20% n.a. Higher PEEP guided by esophageal-pressure;

VT = 7.61.5 mL/kg/pbw

0.49 (0.201.24) oxygenation

compliance, rs

5

LEGEND FOR TABLE S1:

n.a. not available information.

ARDSNetTV: First ARDSNet study5 comparing high versus low tidal volume strategies

ARDSNetPEEP: Second ARDSNet study6 comparing high versus low PEEP strategies

*: P < 0.001 - Chi-squared test comparing differences in prevalence of primary ARDS across the trials.

RR: non-adjusted relative-risk (mortality-rate) associated with the treatment arm - calculated by Cox Proportional Hazards regression.

: In the supplement (Figure S1) we present the results of the adjusted relative-risk according to Model-1 (Table 1), where we noticed two important findings:

- the intervention arm in the EXPRESS trial presented a significant reduction in the relative-risk: 0.75 (95%CI: 0.590.96; P=0.02).

- the intervention arm in the ARDSnetPEEP trial presented an inversion of the trend for the relative-risk:0.82 (95%CI: 0.561.12; P=0.29).

(This was related to an important imbalance in covariates at baseline, as reported in the original publication of this trial).

95% C.I. 95% confidence interval;

P: driving-pressure defined as the difference between plateau pressure and PEEP

PPLAT: plateau pressure at airways; PPEAK: peak-inspiratory airway pressure

: means a significant improvement in the variable within the treatment arm

: means a significant decrease in the number of patients suffering refractory hypoxemia during the hospital stay.

: Median of days under mechanical ventilation (intubation) before entering the study.

: The study of Amato et al. was the only one where a target for maximum P was explicit in the lung-protective protocol.

This trial also used a higher PEEP strategy in the treatment arm.

: Except for the study of Amato et al. the VT-trials used the same PEEP strategy in the control and intervention arms

: The PEEP-trials (forming our validation cohort) used the same tidal volume strategy (< 8 mL/kg) in the control and intervention arms.

6

Table S2 (website): Baseline patient characteristics: recalculated from individual patient data. The trials in the first four rows were pooled and formed our hypothesis generation sample, used to elect a multivariate model for survival (Model-1, Table 1 - main text).

The ARDSnetVT study (testing a high versus low VT strategy) was used as a first validation sample. The studies in the last four rows (testing a high versus low PEEP

strategy) were pooled and used as a second validation sample.

Values represent the median, with the interquartile range inside parenthesis

*Risk of Death was calculated according to the equations of APACHE II, APACHE III and SAPS II, depending on the individual scores available in each of the trials.

Tidal compliance is represented in milliliters per centimeter of water / ideal body weight. In the hypothesis generation sample (first four rows) as well as the ARDSnetVT

trial, the value was obtained from the first measurement after randomization, taken a few hours after entry;

ARDSnetVT: First ARDSNet study 5 comparing lower versus higher tidal-volume strategies

ARDSnetPEEP: Second ARDSNet study 6 comparing higher versus lower PEEP strategies

Trial:

Risk of Death*

arterial pH at entry PaO2/FIO2 at entry Tidal-compliance at entry or

at first day

Amato et al.1 48(35-71) 7.32 (7.24-7.40) 113 (74 - 165) 0.40 (0.32-0.53)

Stewart et al.2 40(26-64) 7.39 (7.33-7.43) 182 (135 246) 0.53 (0.39-0.66)

Brochard et al.3 23(15-36) 7.36 (7.30-7.42) 137 (110 177) 0.52 (0.38-0.64)

Brower et al.4 44(19-64) 7.42 (7.36-7.45) 119 (97 - 147) 0.45 (0.36-0.56)

ARDSnetVT5 39(24-64) 7.41 (7.36-7.45) 123 (89 175) 0.46 (0.36-0.60)

ARDSnetPEEP6 49(29-64) 7.39 (7.32-7.43) 142 (104 200) 0.47 (0.36-0.60)

EXPRESS7 42(23-68) 7.37 (7.31-7.42) 138 (98 180) 0.47 (0.39-0.60)

LOVS8 54(35-74) 7.37 (7.30-7.42) 141 (106 180) 0.45 (0.36-0.58)

Talmor et al.9 60(46-72) 7.38 (7.33-7.42) 135 (108 178) 0.47 (0.39-0.59)

7

ARDSnet PEEP

Amato et al. Stewart et al.

EXPRESS LOVS EPVENT

Brochard et al. Brower et al.

ARDSnetVT

Figure-S1: Overview of the results of randomization in each of the study-trials. (survival curves were pre-adjusted according to covariates 1-5 of model-1 ; Cox proportional-hazards)

Days after randomization Days after randomization Days after randomization Days after randomization

Days after randomization

Days after randomization Days after randomization Days after randomization Days after randomization

Ad

just

ed s

urv

ival

Ad

just

ed s

urv

ival

Ad

just

ed s

urv

ival

Hypothesis-generation Cohort ( N = 336)

First validation Cohort ( N = 861)

Higher vs. Lower VTstudy -trial

Higher vs. Lower PEEPstudy -trials

Second validation Cohort ( N = 2360)

Higher vs. Lower VTstudy -trials

*

*

* : significant survival differences in the original report(the study of Amato et al. tested a combined strategy of higher PEEP and lower VT)

: significant survival differences after multivariate adjustment (model-1)

Treatment arm

Control arm

8

II) ADDITIONAL RESULTS AND ANALYSIS REFERRED TO

IN THE MAIN MANUSCRIPT:

II.1. Accounting for residual (intrinsic) heterogeneity across the trials: Please, refer to the figure on the next page

Figure S2: Accounting for residual heterogeneity across the trials

Additional details:

As shown in Figure S2, in spite of covariate adjustments according to Model-1, there was some

residual, unexplained heterogeneity in the pooled mortality (both arms considered together)

across trials (P < 0.001). Overall the mortality in the derivation cohort (45.3%) was higher than

that observed in the validation cohorts (33.6% and 34.2%, for ARDSnetVT cohort and second

validation cohort, respectively). This higher mortality was a general trend observed in both arms

of each of the studies (from derivation cohort) and could not be fully explained by their baseline

disease (expressed by the covariates Age, APACHE III, arterial pH, PaO2/FIO2ratio) or by P

applied. The source of this heterogeneity is beyond the scope of this study and might relate, for

instance, to improvements in general patient support, independent of ventilation strategies.

We must stress that this residual heterogeneity did not cause any bias in Table 1 or Figures 1-2

presented in the main manuscript, since we pre-adjusted our survival models according to a

categorical variable "Trial". The reported effects of P on mortality were, therefore, necessarily

calculated in proportion to this intrinsic pooled mortality of each trial.

9

Figure S2: Accounting for residual heterogeneities across the trials (control and treatment are pooled together for each trial)

Figure S2: Accounting for residual heterogeneities across the trials.

After accounting for differences in baseline covariates, the intrinsic mortality of the derivation cohort remained higher than in the validation cohort. In order to minimize

such intrinsic differences (not fully expressed by baseline covariates), our analysis (Table 1 and Figures 1-2, main manuscript) always included a categorical dummy

variable that balanced the different risks among the trials (the panel on the right shows the respective survival curves after such adjustment).

1

2

3

Adjusted for individual patient covariates

+ trial dummy covariate

High vs. Low PEEP trials second validation cohort

High vs. Low VT trial ARDSnet first validation cohort

High vs. Low VT trials preARDSnet derivation cohort

Figure-S8: Accounting for residual heterogeneities across the trials( control and treatment survival are pooled together for each trial )

Adjusted for individual patient

covariates

study-trial:

Days after randomization Days after randomization

Adjustment to balance intrinsic differences in baseline mortality across the trials(not fully expressed by baseline covariates)

1

3

2

10

II.2. Univariate analysis: Please, refer to the table on the next page

Table S3: Univariate Cox Regression Model 60-Day Mortality.

11

Table S3 (website): Univariate Cox Regression Model 60-Day Mortality

Hypothesis generation cohort

- Univariate -

First Validation cohort

- Univariate -

Second Validation cohort

- Univariate -

(N = 336) (N = 861) (N = 2360)

VARIABLES: RR (95% C.I.) P-value RR (95% C.I.) P-value RR (95% C.I.) P-value

Trial * 0.27 --- < 0.001

Randomized arm 0.93 (0.68 1.28) 0.67 0.74 (0.58 0.93) 0.01 0.90 (0.78 1.03) 0.13

Days on MV before 1.12 (0.97 1.27) 0.16 --- --- --- ---

Age 1.03(0.88 1.22) 0.68 1.73 (1.52 1.97) < 0.001 1.70 (1.57 1.83) < 0.001

APACHE/SAPS risk 1.59 (1.34 1.89) < 0.001 1.51 (1.34 1.69) < 0.001 1.83 (1.70 1.98) < 0.001

Organ Failures --- --- 1.40 (1.25 1.57) < 0.001 1.48 (1.37 1.59) < 0.001

Arterial pH at entry 0.69 (0.61 0.79) < 0.001 0.66 (0.58 0.77) < 0.001 0.59 (0.55 0.63) < 0.001

PaO2/FIO2 at entry 0.73 (0.65 0.83) < 0.001 0.84 (0.74 0.96) 0.01 0.70 (0.64 0.76) < 0.001

Tidal compl. at entry --- --- --- --- 0.76 (0.67 0.87) < 0.001

P at entry --- --- --- --- 1.27 (1.15 1.40) < 0.001

Tidal compl. 1st day 0.80 (0.66 0.97) 0.02 0.90 (0.74 1.09) 0.29 0.91 (0.87 0.94) < 0.001

PaCO2 - 1st day 1.08 (0.95 1.23) 0.22 0.85 (0.72 1.00) 0.05 1.14 (1.06 1.22) < 0.001

FIO2 - 1st day 1.51 (1.28 1.77) < 0.001 1.39 (1.22 1.57) < 0.001 1.54 (1.45 1.65) < 0.001

VT - 1st day 1.08 (0.91 1.30) 0.37 1.06 (0.98 1.15) 0.16 0.99 (0.88 1.12) 0.92

Respir. rate - 1st day 1.18 (0.88 1.88) 0.12 1.17 (1.07 1.28) < 0.001 1.30 (1.21 1.41) < 0.001

PPLAT - 1st day 1.50 (1.26 1.77) < 0.001 1.32 (1.20 1.45) < 0.001 1.39 (1.28 1.51) < 0.001

PEEP - 1st day 1.15 (0.98 1.36) 0.09 1.62 (1.38 1.89) < 0.001 1.13 (1.04 1.22) 0.003

P - 1st day 1.35 (1.16 1.58) < 0.001 1.19 (1.07 1.33) 0.001 1.50 (1.36 1.67) < 0.001

Mean PAW 1st day 1.42 (1.19 1.70) < 0.001 1.48 (1.33 1.65) < 0.001 1.44 (1.24 1.67) < 0.001

12

LEGEND FOR TABLE S3:

* Categorical variable with four classes in the hypothesis generation, plus 4 classes in the second validation sample. The first validation sample had only one single

study (ARDS Network tidal volume study).

Methods for defining organ-failures differed within hypothesis generation sample and pooled relative risks could not be calculated.

A random variable (mean = 57; std = 15; as reported in the original publication) was imputed to all patients in the study of Brochard et al.

Abbreviations: RR: relative risk associated to 1 standard-deviation (STD) increment in the respective variable;

By normalizing RR according to STD, the strength of the association of different variables with survival can be grossly compared as the RR per se

(using 1/RR when RR < 1). For instance, in the second validation sample, P showed stronger association with survival (1.50) than tidal-compliance. (1/0.91 = 1.10).

95% C.I.: 95% confidence interval; VT = tidal-volume; PPLAT = plateau-pressure; P = driving-pressure; Mean PAW= mean airway pressure; FIO2= fraction of

inspired oxygen; PEEP = positive end-expiratory pressure; tidal compl. = tidal-compliance.

13

II.3. Length of risk exposure and test of proportional hazards

assumption:

Please, refer to the tables on the next pages

Table S4:

Non-parametric Correlation Between Individual Values Observed During the First Day of

Mechanical Ventilation (Ventilation-Variables), and the Individual Values Observed in

the Following Days

Table S5:

Multivariate Cox Regression Model (60-day Hospital Mortality) comparing the

performance of Ventilation-Variables on Day 1 versus Days 1 to 3.

Table S6:

Comparison of the Original Model 1 (with constant hazards) versus Alternative Models

with Time-Dependent Covariates Included..

14

Table S4 (website): Non-parametric Correlation Between Individual Values (Ventilation-Variables Observed During the First Day of Mechanical

Ventilation) and the Individual Values Observed in the Following Days

(data from the ARDSNetPEEP trial)

Spearman correlation coefficient

Ventilator-variables: Mean value

1stday

Mean value

2nd

day

Mean value

3rd

day

Mean value

4th

day

Mean value

7th

Day

FIO2 - 1

stday 1 0.64* 0.51* 0.47* 0.29*

VT - 1stday 1 0.87* 0.79* 0.75* 0.68*

Respir. rate - 1stday 1 0.70* 0.59* 0.54* 0.36*

Plateau Press. - 1st day 1 0.66* 0.56* 0.56* 0.51*

PEEP - 1st day 1 0.73* 0.62* 0.58* 0.48*

Driving Press. - 1st day 1 0.64* 0.51* 0.52* 0.52*

Mean PAW 1stday 1 0.69* 0.63* 0.57* 0.50*

* :P < 0.001 ; P-value of the two-tailed test of significance for the Spearmans-Rho correlation-coefficient. The correlation was

calculated between individual data collected at day one (first 24 hs. after randomization) and data at each of the following days,

for the same respective patients.

15

Table S5 (website): Multivariate Cox Regression Model (60-day Hospital Mortality) comparing the performance of Ventilation-

Variables on Day 1 versus Days 1 to 3.

(data from the ARDSNetPEEP trial)

Considering Ventilation-Variables

to 1st day.

Considering Ventilation-Variables to

3rd

day

- Multivariate - - Multivariate -

(Valid cases = 483) (Valid cases = 328)

RR (95% C.I.) P-value RR (95% C.I.) P-value

Model: (1) Age 1.88 (1.54 2.29) < 0.001 1.84 (1.46 2.33) < 0.001

(2) APACHE III 1.79 (1.51 2.12) < 0.001 1.73 (1.38 2.18) < 0.001

(3) Organ Failures 1.09 (0.89 1.34) 0.39 1.21 (0.95 1.53) 0.12

(4) arterial pH at entry 0.59 (0.47 0.75) < 0.001 0.64 (0.48 0.85) < 0.001

(5) PaO2/FIO2 at entry 1.00 (0.78 1.28) 0.81 0.94 (0.73 1.20) 0.72

(6) FIO2 - 1st day 0.99 (0.78 1.25) 0.77 0.86 (0.65 1.13) 0.27

(7) Driving-pressure 1.59 (1.22 2.07) 0.001 1.70 (1.23 2.35) 0.001

Model Chi-Square

(change after including all covariates) 139.9 (P =2 x10-26) 85.1 (P = 5 x10-15)

The average values in time were used for ventilator-variables collected from days 1 to 3.

Only patients surviving longer than 1or 3 days were respectively included in the Cox survival model. This explains why the overall Chi-Square decreased,

despite a preserved association between individual covariates and survival.

RR: adjusted relative risk associated to 1 standard-deviation increment in the respective variable.

95% C.I. 95% confidence interval

16

Table S6 (website): Comparison of the Original Model 1 (with constant hazards) versus

Alternative Models with Time-Dependent Covariates Included.

Original

model

Alternative models after addition of time-dependent covariates

( RR and P-values below refer to the long-term hazard observed during the 60-day period)

RR P-value RR P-value RR P-value RR P-value RR P-value RR P-value

Model 1 (constant hazards)

(2) Age 1.59 < 0.001 1.71 < 0.001 1.59 < 0.001 1.58 < 0.001 1.59 < 0.001 1.58 < 0.001

(3) APACHE/SAPS-risk 1.38 < 0.001 1.38 < 0.001 1.24 < 0.001 1.39 < 0.001 1.38 < 0.001 1.38 < 0.001

(4) arterial pH at entry 0.68 < 0.001 0.68 < 0.001 0.69 < 0.001 0.80 < 0.001 0.68 < 0.001 0.68 < 0.001

(5) PaO2/FIO2 at entry 0.87 < 0.001 0.87 < 0.001 0.87 < 0.001 0.87 < 0.001 0.92 0.33 0.87 < 0.001

(6) Driving Press. - 1st day 1.41 < 0.001 1.41 < 0.001 1.40 < 0.001 1.40 < 0.001 1.40 < 0.001 1.35 < 0.001

Time-dependent covariate added (exerting a multiplicative hazard

during the first week)

Age APACHE III arterial pH

at entry

PaO2/FIO2

at entry

Driving Press. -

1st day

Transient hazards observed during the first week *

1.38 1.74 0.56 0.77 1.49

P-value (RR for the first week versus

RR for the rest of the period)

(0.001) (

17

Additional details: (related to Tables S4-S6 above)

As we excluded patients with early death or weaning (i.e. death or weaning within the first 24

hours following randomization), we necessarily included patients exposed to ventilation risk

factors for at least 24 hours. We also assumed that a fixed, ongoing hazard should be related

to the average value of the variable observed during the first 24 hours of mechanical

ventilation, despite possible fluctuations of the variable in the next few days. To test the validity

of this assumption, we performed three additional series of analysis described below. One

important observation here is our censoring procedure: to avoid competing risks, we censored

all patients discharged to home before day-60 as alive at day-60 (instead of censoring them as

alive at the date of discharge)10. Thus, our analysis basically represents the estimation of risks

during hospital stay (i.e. focusing on hospital mortality), avoiding biases caused by unknown

risk exposition at home.

First, we assessed the relationship of a ventilation variable to its respective value in the next

few days. Such analysis is illustrated in Table S4. There was a high degree of correlation,

especially for tidal volume, in which the relationship remained significant for several days. Thus,

a value measured in the first 24 hours was generally representative of values for the

subsequent several days.

Second, we checked if the inclusion of any additional information on days 2 and 3 of

mechanical ventilation could improve our survival model. Variables representing either the

ventilator parameters observed once during days 2 or 3, or the average values during the first 2

or 3 days, were included stepwise in model 1.This analysis, however, required that patients

have survived and have remained on mechanical ventilation for at least 2 or 3 days, decreasing

substantially the number of valid cases. One example of such an analysis is presented in

Table S5, performed with the data from the ARDS Network PEEP trial6. We chose this single

18

trial because of the lowest number of early deaths and missing cases (until day 7) of

mechanical ventilation. As shown, the effect-size of most ventilation variables was either

maintained or increased after considering longer periods of time exposure, suggesting an

ongoing, cumulative effect. However, the power of the analysis decreased and confidence

intervals widened due to the smaller sample size. After also considering the potential survival

bias introduced by such procedure -the selection of healthier patients, able to survive the first 3

days11 - we preferred to use the simpler and more powerful analysis, only considering risk

exposure to 24 hours of mechanical ventilation.

Finally, we tested the possibility that ventilation covariates exerted only transient effects that

diminished after a few days. This would represent a violation to the proportional hazard

hypothesis, which assumes an ongoing hazard till the end of the 60-day period. Thus, using the

approach suggested by Kasal et al.12, we assigned, for each covariate, a respective time-

varying covariate that allows a different hazard (higher or lower) to be applied over days 0 7,

in addition to the fixed hazard (constant during the 60days) already included in the model.

Whenever this new time-dependent covariate brought additional information to the model (at a

significance level of P < 0.01), we considered that there was a non-proportional hazard (higher

or lower) during the first week of mechanical ventilation. Such analysis is presented in

Table S6. The most relevant non-proportional hazards were observed for lower values of

baseline arterial-pH and higher values APACHE-III/SAPS-risk, both conditions associated with

higher mortality during the first week(in addition to their long term hazard). High driving-

pressures, however, did not impose additional risks during the first week.

We concluded, therefore, that high driving-pressure exposure during the first days of

mechanical ventilation was strongly associated to a fixed, ongoing hazard during the first 60-

day after randomization. There was no need of more sophisticated models, with time-

dependent variables, to describe such relationship.

19

II.4. Sensitivity analysis for different estimates of

baseline-elastanceRS

As a surrogate of the severity of underlying lung disease, baseline-elastanceRS should be

ideally calculated from baseline data, providing independent measurements that are not

affected by treatment. For instance, a higher-PEEP might decrease lung elastance after

randomization due to an immediate promotion of lung recruitment, but not because of an actual

change in the underlying lung disease. Thus, baseline-elastanceRS might be underestimated in

in the treatment arm, if measured after randomization (even if a few minutes later).

To circumvent this potential bias, whenever we had to use post-randomization data to calculate

baseline-elastanceRS (necessary for the earlier VT-trials), we calculated it as stratified

elastanceRS-ranks, performed within each treatment arm, for each trial. These ranks were

subsequently scaled within the [-0.5 to 0.5] interval. This procedure was performed under the

reasonable assumption that the systematic changes in ventilation parameters due to

randomization might affect the absolute values of elastanceRS, but could hardly affect the

ranking of individual elastancesRS within the respective study-arm. In the next few paragraphs,

we will demonstrate the good plausibility of this assumption.

By using our second validation cohort, in which we measured individual data of baseline-

elastance (pre-randomization, when the patients received VT = 82 mL/kg/PBW and PEEP =

104 cmH2O), as well as individual data of post-randomization elastanceRS, we could compare

the information provided by the two estimates of elastance (baseline-elastance as actually

measured versus stratified elastance-ranks measured post-randomization, using N = 1656

patients enrolled in three PEEP-trials). We observed that (Figure S3, next page):

20

Figure S3: correlation between ElastanceRS estimated from baseline data

versus

stratified ElastanceRS ranks (estimated from post-randomization data).

Figure S3: Correlation between the two estimates for baseline ElastanceRS

Using the data from 1656 patients participating in the higher-PEEP trials, for whom we had measurements of

elastance performed at baseline (average VT = 82 mL/kg/ibw and PEEP = 104 cmH2O), as well as after

randomization, we could check the correlation between the two estimates. In order to avoid post-treatment

bias in the measurements performed after randomization, we calculated elastance ranks within each of the

arms, and within each of the studies. The ranks were then scaled within the [-0.5 to 0.5] interval. This

procedure was performed under the reasonable assumption that the systematic changes in ventilation

parameters due to randomization might affect the absolute values of elastance, but could hardly affect the

ranking of individual elastances within the respective study-arm. As shown, the relationship between both

variables was reasonably linear, with similar slopes and determination coefficients for both arms. This

suggest that the stratified ranking avoided post-treatment bias.

Rank of Elastance respiratory system

( post randomization data )

- stratified by arm, and by study-trial -

Bas

elin

e E

last

ance

resp

irat

ory

sys

tem

(P

BW

)

( b

asel

ine

dat

a )

21

1. The relationship between both variables, i.e., baseline-elastanceRS (actually measured)

versus stratified elastanceRS-ranks (from post-randomization data), was reasonably

linear (Figure S3), with similar slopes and determination coefficients (R2 = 0.35;

P

22

4. Finally, we observed that after pre-adjusting model-1 for elastanceRS-ranks (post-

treatment), the test of entry of baseline-elastanceRS (pre-treatment) in the model was no

longer significant. This suggest a consistent overlap of information coming from both

methods of estimation of underlying lung disease: there was no further independent

information (correlated with outcomes) in baseline-elastanceRS that was missed in the

prediction models, or in the mediation models.

This sequence of tests shows that, in the hypothesis that baseline-elastanceRS was causing

some confounding effect in our mediation analysis, this bias was equally removed from P

by both adjustments (i.e. using pre or post randomization data).

23

II.5. Homogeneous P-risks across the trials:

Please, refer to the figure in next page

Figure S4:

Relative risk of death associated to increments in P within each of the trials

Additional details:

Model adjustments for unexplained differences in the pooled mortality (both arms together) for

each trial and for differences in effect-size across the trials did not change the relative risk

associated with general increments of one standard deviation in P. The first adjustment was

performed by assigning one dummy variable for each trial in the Cox model (Figure S2); the

effect-size adjustment was performed by assigning one dummy interaction term for each trial

expressing the freedom for each trial to present a peculiar response to P.

After all such adjustments1, the relative risk associated to increments of P was consistently

1.45 (95% CI, 1.28 to 1.64; P < 0.0001), similar to the numbers presented in Table 1, main text.

This analysis suggests a consistency of effects: whatever the individual severity of disease

(expressed by baseline covariates), or whatever the baseline mortality of a specific population

(expressed by the dummy variables representing each trial), increments in P were always

deleterious, and associated with the same risk magnitude across the 9 trials (Figure S4).

1 The inclusion of such dummy interaction terms in the Cox model did not contribute to its predictive power

(likelihood-ratio test: P = 0.13) suggesting that the effect-size of increments of P was homogeneous across the different trial populations.

24

Figure S4: Relative risk of death associated to increments in P within each of the trials

Figure S4: Relative-risks associated to increments in P within each of the trials

The relative-risk of death (Cox survival analysis) associated to 1 standard-deviation increment in P

(7.0 cmH2O) measured after randomization (first 24hs.) was calculated for each trial, and for the

combined sample. We performed multivariate adjustment (at patient-level) for covariates specified in

Model 1 (Table 1, main text) plus dummy variables accounting for residual heterogeneities in baseline

mortality among the trials. Error bars represent 95% confidence intervals. There was no significant

heterogeneity of P effects across the trials (P = 0.13; test of driving-pressure-by-trial interaction term),

despite the different distributions of primary cause of ARDS across the trials (P < 0.001, Table S1).

X Data

0.33 0.5 0.7 1 1.4 2 3

Y D

ata

Amato

Stewart

Brochar

Brower

ARDSnet

ALVEOLI

EXPRESS

LOVS

EPVENT

OVERALL

Figure-S2: Relative risk of death associated to increments in P within each of the trials

HazardBenefit

Amato et al.

Stewart et al.

Brochard et al.

Brower et al.

ARDSnetVT

ARDSnetPEEP

Talmor et al.

EXPRESS

LOVS

COMBINED EFFECTS

Adjusted Relative-risk of death ( for one-standard-deviation increment in P )

Heterogeneity test:P = 0.13

P < 0.0001

n=53

n = 120

n = 116

n = 52

n = 861

n = 549

n = 61

n = 767

n = 983

n = 3562

25

II.6. Consistency of higher P-risks in the validation cohorts:

Please, refer to the tables on next 2 pages

Table S7:

Multivariate Cox Regression Model (60-day Hospital Survival)

Original derivation model and posterior test in the ARDSNetVT population (first validation

cohort)

Table S8:

Multivariate Cox Regression Model (60-Day Hospital Survival)

Posterior test of derivation model in the VT trials (derivation cohort combined with the

ARDSNetVT cohort) and in the PEEP trials (second validation cohort) .

This table complements Table 1, main manuscript.

26

Table S7 (website): Multivariate Cox Regression Model (60-day Hospital Survival) - Original derivation model and posterior test in the ARDSNetVT population (first validation cohort)

Original Derivation model.

Early ventilation trials

Test of Derivation Model

ARDSNETVT trial

Refined model

ARDSNETVT trial

- Multivariate - - Multivariate - - Multivariate -

(Valid cases = 331) (Valid cases = 705) (Valid cases = 704)

RR (95% C.I.) P-value

RR(95% C.I.)

RR (95% C.I.) P-value RR (95% C.I.)

P-value

Model: (1) APACHE - risk * 1.36 (1.17 1.57) < 0.001 1.39 (1.23 1.57) < 0.001 1.25 (1.10 1.42) 0.001

(2) arterial pH at entry 0.73 (0.63 0.83) < 0.001 0.82 (0.70 0.96) 0.013 0.72 (0.61 0.83) < 0.001

(3) P - 1st day 1.42 (1.21 1.66) < 0.001 1.21 (1.08 1.35) 0.001 1.29 (1.16 1.44) < 0.001

(4) FIO2 - 1st day 1.24 (1.05 1.48) 0.014 1.24 (1.09 1.42) 0.001

(5) PaO2/FIO2 at entry N.S. N.T. 0.81 (0.71 0.92) 0.001

(6) Age N.S. N.T. 1.77 (1.55 2.03) < 0.001

Model Chi-Square

(step change after inclusion

of block of covariates) 78.7 (P =3 x10-16) 81.6 (P =1 x10-16) 145.7 (P =1 x10-29)

ARDSnetVT: First ARDSNet study 5 comparing lower versus higher tidal-volume strategies

RR: adjusted relative risk associated to one standard-deviation increment in the respective variable. Values above 1.00 indicate increased mortality-rate. The values used for standard-deviation were: age (17), death-risk (26), arterial pH (0.09), PaO2/FIO2 (60), P (7), FIO2 (0.19). 95% C.I. 95% confidence interval

P - 1st day: average driving-pressure during the first 24 Hs after randomization. *: Risk of Death was calculated according to the equations of APACHE II, APACHE III, depending on the trial.

N.S.: Non significant entry in the backward/forward process of selection of variables

N.T.: Not tested

27

Table S8 (website): Multivariate Cox Regression Model 60-Day Hospital Survival Posterior test of derivation model in the VT trials (derivation cohort combined with the ARDSNetVT cohort)

and in the PEEP trials (second validation cohort) . This table complements Table 1, main manuscript

High vs. Lower-VT trials

- Multivariate -

High vs. Lower-PEEP trials

- Multivariate -

Combined analysis

- Multivariate -

(N = 1020) (N = 2060) (N = 3080)

RR(95% C.I.) P-value RR(95% C.I.) P-value RR(95% C.I.) P-value

Model 1:

(1) TRIAL --- < 0.001 --- 0.83 --- < 0.001

(2) Age 1.51(1.36 1.69) < 0.001 1.64(1.50 1.79) < 0.001 1.59(1.48 1.70) < 0.001

(3) Risk of Death 1.34(1.20 1.49) < 0.001 1.41(1.29 1.54) < 0.001 1.38(1.29 1.48) < 0.001

(4) Arterial pH at entry 0.69(0.63 0.77) < 0.001 0.68(0.63 0.74) < 0.001 0.68(0.64 0.72) < 0.001

(5) PaO2/FIO2 at entry 0.85(0.77 0.95) 0.004 0.88(0.80 0.96) 0.005 0.87(0.81 0.93) < 0.001

P - 1st day 1.35(1.24 1.48) < 0.001 1.50(1.35 1.68) < 0.001 1.41(1.31 1.51) < 0.001

Model 2 (including variables 1-5 as above):

P - 1st day 1.41(1.26 1.59) < 0.001 1.48(1.28 1.71) < 0.001 1.41(1.30 1.53) < 0.001

Compliance,RS 1.18(0.96 1.44) 0.12* 0.98(0.88 1.10) 0.75* 1.01(0.92 1.10) 0.90*


P - 1st day 1.32(1.19 1.47) < 0.001 1.51(1.35 1.68) < 0.001 1.40(1.30 1.51) < 0.001

Tidal Volume - 1st day 1.04(0.95 1.14) 0.42* 1.05(0.90 1.23) 0.52* 1.02(0.95 1.10) 0.58*


P - 1st day 1.44(1.10 1.88) 0.008 1.51(1.31 1.75) < 0.001 1.37(1.22 1.53) < 0.001

Plateau Press. - 1st day 0.94(0.72 1.23) 0.65* 0.99(0.87 1.13) 0.90* 1.04(0.93 1.15) 0.53*


P - 1st day 1.36(1.24 1.49) < 0.001 1.50(1.34 1.68) < 0.001 1.41(1.32 1.52) < 0.001

PEEP - 1st day 0.97(0.80 1.18) 0.78* 0.99(0.91 1.09) 0.90* 1.03(0.95 1.11) 0.51*

28 LEGEND FOR TABLE S8:

RR: adjusted relative-risk associated to 1 standard-deviation increment in the respective variable. Values above 1 indicate increased mortality-rate. The values used for

standard-deviation were: age (17), death-risk (26), arterial pH (0.09), PaO2/FIO2 (60), P (7), PEEP (5), Plateau pressure (7), Tidal volume (2), Compliance,RS (0.3). By

normalizing RR in this way, the strength of the association of different variables with survival can be grossly compared as the RR per se (using 1/RR when RR < 1). For

instance, in the combined analysis, P showed stronger association with survival (1.4) than the PaO2/FIO2 (1/0.87 = 1.15)

95% C.I. 95% confidence interval; P - 1st day: average driving-pressure during the first 24 Hs after randomization.

* Test of variable inclusion in the model (net contribution to predictive power likelihood-ratio test) where variables 1-6 plus driving-pressure were previously included.

Test of variable inclusion in the model where variables 1-5 plus the extra covariate in the line below were previously included.

Risk of Death was calculated according to the equations of APACHE II, APACHE III and SAPS II, depending on the trial.

Although not shown in the table, the variable mean-airway-pressure was tested before and after inclusion of P in model 1, showing no significant association with survival.

29

II.7. Tidal volume predicts survival only if normalized to

respiratory system compliance (CRS):

Please, refer to the figure on the next page

Figure S5: Survival impact of tidal volume, before and after lung-sizing

30

Figure S5: Survival effects of tidal volume, before and after lung-sizing

Figure S5: Survival impact of tidal volume, before and after lung-sizing

Using double stratification procedures (like in Figure 1, main manuscript), we partitioned our dataset into five distinct sub-samples (each one

with approximately 600 patients with ARDS), and calculated the relative risk for each sub-sample in comparison with the average risk of the

combined population. Patients are the same as those included in the combined analysis of Table 1.

In the upper scatter/error-bar diagrams (open triangles) we show the average values for plateau pressures across quintiles of tidal volumes

(left) or P (right). In the middle scatter/error-bar diagrams (black squares) we show the average values for tidal volume (left panel, VT scaled

to predicted body-weight, PBW) and for driving-pressure (right panel, VT scaled in proportion to respiratory-system compliance, so

representing P), found in each quintile. The error bars represent one standard-deviation. Note that each resampling (D and E) produced sub-

samples with comparable mean values for plateau pressures, but very distinct values for tidal volume or P.

At the bottom, we show the respective relative-risk calculated for each sub-sample after multivariate adjustment (at patient-level) for covariates

1-5 specified in Model 1 (Table 1). Error bars represent 95% confidence intervals. A relative-risk of 1 represents the average risk of the pooled

population, which presented an adjusted survival at 60-day of 68%.

Note that reductions in tidal volume per se had no impact on mortality risks, whereas reductions of a re-scaled tidal volume (so representing

P) were associated with a marked risk reduction. Note the mathematical equivalence: P = ( VT / CRS) = VT normalized to CRS

DR

IVIN

G -

PR

ES

SU

RE

5

10

15

20

25

30

35

PLA

TE

AU

-PR

ES

SU

RE

0

5

10

15

20

25

30

35

RESAMPLING : MATCHED PPLAT , QUINTILES OF P

0 1 2 3 4 5 6

RE

LA

TIV

E -

RIS

K (

adju

ste

d )

0.6

0.8

1.0

1.2

1.4

1.6

P < 0.0001

Figure 1B:

P

( cm

H2O

)

5

Resampling D:

- matched PPLAT, - decreasing ranks of VT / PBW

1.6

1.4

1.2

0.6

1.0

0.8

TID

AL

- V

OL

UM

E

4

6

8

10

12

14

PL

AT

EA

U-P

RE

SS

UR

E

0

5

10

15

20

25

30

35

RESAMPLING : MATCHED PPLAT , QUINTILES OF VT

0 1 2 3 4 5 6

RE

LA

TIV

E -

RIS

K (

ad

juste

d )

0.6

0.8

1.0

1.2

1.4

1.6

P = 0.92

1.6

1.4

1.2

0.6

1.0

0.8

20

15

10

VT / PBW( mL / kg )

PPLAT ( cmH2O )

35

30

25

10

8

6

4

Resampling E:

- matched PPLAT, - decreasing ranks of VT / CRS (=P)

611 620 611 623 607

*: mortality-rate adjusted for age, APACHE/SAPS risk, arterial-pH, P/F ratio, and study-trial (Cox Proportional Hazards Regression)

Scaling VT to CRS

Instead of PBW

VT / CRS(P, cmH2O )

PPLAT ( cmH2O )

35

30

25

5

600 624 644 598 614( Sub-sample N ):

S1 S2 S3 S4 S5 S1 S2 S3 S4 S5

Rel

ativ

e r

isk

( ad

just

ed

mo

rtal

ity

rat

e*

)

Rel

ativ

e r

isk

( ad

just

ed

mo

rtal

ity

rat

e*

)

P < 0.001

* : mortality rate adjusted for age, APACHE/SAPS risk, arterial-pH, P/F ratio, and Trial (Cox Proportional Hazard Regression)

Resampling D Resampling E

31

II.8. Survival in patients under protective ventilator settings:


Figure S6: Survival in patients under protective ventilator settings

32

Figure S6: Survival in patients under protective ventilator settings

Figure S6: Survival in patients under "protective" ventilator settings

Survival curves were obtained after multivariate adjustment at patient level (Cox Proportional Hazards model)

for covariates 1-5 specified in Model-1 (Table 1, main manuscript, and Table S8). For each survival plot, the

selected sub-sample (N=1745) was stratified according to the median values of P, plateau pressure and VT,

respectively (median values = 13 cmH2O, 26 cmH2O and 6 mL/kg PBW, respectively, from top to bottom

plots), producing two strata with similar number of patients. Treating ventilator variables as continuous variables

did not improve the association of tidal volume or plateau pressure with survival, but did improve the

association of P with survival (P 14 cmH2O

PPLAT > 25 cmH2O

PPLAT 25 cmH2O

VT < 6 mL / kg

VT 6 mL / kg

Days after randomization

0 10 20 30 40 50 60

100

95

90

85

80

75

70

100

95

90

85

80

75

70

100

95

90

85

80

75

70

Cum

m. S

urv

iva

l (

%)

( a

dju

ste

d*

)

Cum

m. S

urv

iva

l (

%)

( a

dju

ste

d*

)

Cum

m. S

urv

iva

l (

%)

( a

dju

ste

d*

)

P = 0.98

P = 0.30

P < 0.001 stratification: ( N )

( 989 )

( 756 )

( 955 )

( 790 )

( 867 )

( 878 )

( N = 1745 )

Subsample of patients under protective settings

MEDIAN

> MEDIAN

MEDIAN

> MEDIAN

> MEDIAN

MEDIAN

*: su

rviv

al a

dju

ste

d fo

r

Ag

e, A

PA

CH

E/S

AP

S r

isk,

Art

eria

l-p

H,

P/F

ra

tio

, a

nd

T

ria

lFigure-S4:

33

II.9. P (but not VT) predicts Barotrauma after randomization:


Figure S7: Odds for barotrauma (Pneumothorax) across quintiles of P or VT

(combined population of ARDS: N = 3,080)

34

Figure S7: Odds for Barotrauma across quintiles of P or VT :

Combined population of ARDS ( N = 3080)

Figure S7: Odds-ratio for Barotrauma during the first 28 days after randomization

Barotrauma was strictly defined as pneumothorax requiring chest tube drainage (with 313 events, or 9% of

the sample). The odds-ratio for each quintile was calculated in relation to the average risk of the combined

population (assumed to be 1.00). The mean odds and 95% confidence intervals (error bars, enclosing the

gray zone) for each quintile were calculated after multivariate adjustment at patient level (Logistic regression

model) for covariates 1-5 specified in Model-1 (Table 1, main manuscript). After including tidal volume and

P in the multivariate model, the adjusted odds-ratio for progressive quintiles of both variables were

calculated (each quintile had approximately 600 patients). The number of percentiles was chosen in order to

have at least 40 events per percentile, guarantying reliable confidence intervals. P-values indicate the overall

differences in risk across quintiles (as categorical variable).

.

P < 0.0001

Od

ds-

rati

o f

or

Bar

otr

aum

a(

adju

sted

* )

*: pre-adjusted for age, APACHE/SAPS risk, arterial-pH, P/F ratio and study-trial(multivariate logistic regression where both P and VT co-participate in Model-1)

Odds for Barotrauma across quintiles of P or VT: - Combined population of ARDS ( N = 3080 )

Figure 2d:

4 8 12 16 20 24 28

0.6

1.0

1.4

1.8

2.2

P < 0.0001

Driving-pressure (P, cmH2O)

Od

ds-

rati

o f

or

Bar

otr

aum

a(

adju

sted

* )

*: pre-adjusted for age, APACHE/SAPS risk, arterial-pH, P/F ratio and study-trial

(VT and P co-participating in model-1 )

4 6 8 10 12

0.6

1.0

1.4

1.8

2.2

P = 0.87

Tidal-Volume (VT , mL/kg.PBW)

Od

ds-

rati

o f

or

Bar

otr

aum

a(

adju

sted

* )

*: pre-adjusted for age, APACHE/SAPS risk, arterial-pH, P/F ratio and study-trial

(VT and P co-participating in model-1)

Driving-pressure (P, cmH2O)

4 8 12 16 20 24 28

Tidal-Volume (VT , mL/kg[PBW] )

4 6 12108

P < 0.0001 P = 0.87

Figure-S5:

P < 0.001

* : adjusted for age, APACHE/SAPS risk, arterial-pH, P/F ratio, and Trial

(multivariate logistic regression where both P and VT co-participate in Model-1)

35

II.10. Mediation Analysis:

More than a marker for the severity of baseline lung disease.

P strongly correlates with mortality, independently of baseline elastance

of respiratory system)

Please, refer to the figures on the next pages

Figure S8:

Mediation in the Lower vs. Higher VT-trials:

Tested mediator: P-changes driven by randomization

Figure S9:

Mediation in the Higher vs. Lower PEEP-trials:


36

Figure S8: Mediation in the Lower vs. Higher VT-trials:


Mediated-proportion:

74%

Randomization Survival effect

Hazard = 0.68 (ACME)

(0.48 0.75; P < 0.001 )

Mediational model: High vs. Low VT trials



lower P

-8.4 cmH2O

P = 0.46 ( N.S. effect )

Hazard = 0.60

P = 0.004

Figure 3a :

Step 1:

Steps 3 & 4:

*: covariates: age, APACHE/SAPS risk, arterial-pH, P/F ratio, Trial

lower P(-1 STD = - 7 cmH2O)

Survival effect

Baseline disease covariates*

( 0.52 0.74 ; P < 0.001 )

Hazard = 0.62Step 2:

ElastanceRSadjustment


P < 0.001



Subsequently, we jointly calculated the influence of the mediator on survival, after accounting for baseline risk factors, baseline-

elastanceRS, and the direct effects of randomization. This last step shows that reductions in P mediated most (75%, P = 0.004) of the

original effect of randomization and, consequently, randomization is no longer associated with survival in an independent manner

(characterizing complete mediation).

Implicitly, this last step with significant ACME (average causal mediation effect) also suggests that variations in P had an independent

impact on survival: i.e. patients exhibiting larger reductions in P obtained a survival benefit that exceeded the average benefits found in

the lower-VT arm.

Solid arrows in the path diagram represent significant association between variables, with left to right direction representing an independent

to dependent relationship. Red dashed arrows represent non-significant effects.

Top: The first step in our mediational

analysis was the demonstration that

randomization (assignment to lower

tidal-volume arm) had a measurable

impact on survival, after accounting

for baseline risk covariates (Model-1,

Table 1).

Middle: Secondly, we checked if

mediator changes, theoretically

assumed as beneficial, correlated

with better survival. At this step, we

corrected for the baseline-elastance

(of respiratory-system).This allowed

us to examine the exclusive impact of

superimposed variations in P driven

by changes in ventilator settings,

which followed the random treatment

assignment.

Bottom: Finally, a multilinear

regression (mixed effects) calculated

the influence of randomization on the

tested mediator, indicating that

randomization caused a significant

mean reduction of -8.4 cmH2O in P.

37

Figure S9: Mediation in the Higher vs. Lower PEEP-trials:


Randomization Survival effect(0.72 0.97; P = 0.02 )

Mediational model: High vs. Low PEEP trials Tested mediator: P-changes driven by randomization


Hazard = 0.83

Figure 3c :

Step 1:

Steps 3 & 4:

lower P(-1 STD = - 7 cmH2O)

Survival effect


( 0.42 0.72; P < 0.001 )





45%Hazard = 0.91 (ACME)

lower P

-1.2 cmH2O

P = 0.001P < 0.001

P = 0.18 ( N.S. effect )




This last step, with significant ACME (average causal mediation effect) also demonstrates that P had a survival effect that exceeds the

main effect of randomization: i.e. patients exhibiting accentuated reductions in P obtained a survival benefit that exceeded the average

benefit found in the respective higher-PEEP arm.



We followed the same steps

described in the mediational analysis

of Figure S8.

Top: We first demonstrate that

randomization (assignment to higher-

PEEP arm) had a measurable impact

on survival, after accounting for

baseline risk covariates (Model-1,

Table 1).

Middle: Secondly, we checked if

mediator changes, theoretically

assumed as beneficial, correlated

with better survival, especially after

pre-adjustment for baseline-

elastanceRS. This step demonstrated

the significant, independent impact of

superimposed (i.e. caused by

ventilator adjustments) variations in

P.

Bottom: we finally show that

reductions in P had a beneficial

impact, explaining 45% of the original

effects of randomization. The no

longer significant effect of

randomization (red arrow) suggest

complete mediation.

38

Additional details:

The reliability of a randomized clinical trial resides in the lack of correlation between treatment

and baseline condition of patients. This is an essential, well-accepted pre-requisite for

accepting causality implication. The lack of correlation arises from an exogenous, random

process for treatment selection. In case we find a significant correlation between treatment and

outcomes, we can suggest that the benefits are directly caused by treatment, and not by the

circumstantial fact that we applied the treatment in patients with better baseline condition.

Similarly, to suggest that P was more than a risk predictor, working as a mediator of survival,

(i.e. to the extra-survival related to randomized treatment) we had to be sure that the variability

of P observed in our samples were not correlated with baseline disease (i.e. independent

from baseline elastance). Ideally, P variations should be mostly related to ventilator strategy,

implemented according to a randomized process. However, since P is mathematically

correlated with baseline elastance of respiratory system (baseline-elastanceRS), this

independency of P was hardly true and the solution was to first remove (filter out) the

component of P correlated with baseline mechanics, later applying the mediation analysis on

the residual P component.

In multivariate regression analysis, this removal or subtraction of the confounding component

(P-component related to baseline disease) is equivalent to the pre-adjustment of a regression

model by baseline-elastanceRS, if adjusting the same model in which P is simultaneously

tested as an independent explanatory variable. A significant correlation between P and

outcomes would indicate the presence of a residual and independent (orthogonal) source of

variability in P, uncorrelated with baseline disease, which is also affecting outcomes.

Accordingly, in the first logical steps of our mediation analysis (step-2: the check if mediator

changes, theoretically assumed as beneficial, correlated with better survival), we explicitly

tested this hypothesis (Figure S8 and S9). After pre-adjustment of survival Model-1 for

39

baseline-elastanceRS, we checked the correlation between residual changes in P (now mostly

correlated with randomization) and outcomes. As shown in the respective figures (step-2),

reductions in P were significantly associated with better survival in both cohorts,

independently from baseline lung disease, and exhibiting similar effect-size in both cohorts (RR

for 1 STD change: 0.62; 95%CI: 0.520.74 for lower-VT trials; RR: 0.57; 95%CI: 0.420.72,

for higher-PEEP trials). Fitting the rationale of our a priori hypothesis, the sign of this correlation

was negative, i.e. a reduction in P caused improved survival.

After the demonstration of a strong correlation between superimposed-P changes (defined as

residual variations in P, mostly driven by the changes in ventilator settings following

randomization) and mortality, we performed a stepwise, complete mediational analysis (a

powerful and innovative statistical approach that investigates mechanisms explaining why, and

to which extent, a randomized treatment works13-23).

In our case, the hypothesis was that randomization (treatment assignment represents an

intention to treat bundle including various recommendations like VT reduction, plateau pressure

limitation, respiratory rate and acidosis management, etc.) had an impact of survival in

proportion to variations in superimposed-P, despite the fact that manipulation of P was not

an explicit target in most protocols.

To be a mediator, the variable-candidate must be strongly affected by the randomized

treatment, and must be an intermediate variable within the temporal pathway between

randomization (treatment group assignment) and the outcomes. To be significant, a mediation

analysis requires the demonstration that a variation in the mediator causes an impact on

outcome, which is independent of the main effect of randomization (i.e. it causes an impact

above and beyond that caused by randomization). This means that other factors, besides

randomization, cause some variability in the mediator (ideally at random), and this extra-

variability must cause an independent impact on outcomes. Sometimes it is possible to show

40

that variations in the mediator explain the whole effect of the randomized treatment (so called

complete mediation), but this is not a pre-requisite to demonstrate mediation.

The mediation analysis must be performed according to a sequence of statistical tests and

logical steps. We here used the sequence described by Kraemer and Shrout, in accordance

with the MacArthur approach 20-22.

The initial step was the analysis of the main effect of randomization, traditionally called as the

direct effect. This was calculated through a Cox survival analysis, using trial as a random effect.

In the original reports, only two1,5 out of the nine trials in our combined sample reported a

statistically significant impact of protective strategies. In a recent meta-analysis24 using a

population that had large overlap with the population of our derivation cohort, investigators

have found a marginal benefit of lower VT strategies. And another patient-level meta-analysis25,

considering only the higher versus lower PEEP trials, showed a benefit of the higher PEEP

strategy, although restricted to a subgroup of patients with more severe hypoxemia.

Differently, our combined analysis showed a more consistent efficacy of protective strategies,

either considering the lower-VT trials separately, or the higher-PEEP trials separately. This

difference was essentially caused by the multivariate adjustment at the patient level, according

to survival Model-1. For instance, the crude relative-risk of lower-VT strategies was 0.82 (95%CI

= 0.671.00) before adjustment, and 0.60 (95%CI = 0.480.75) after adjustment. Similarly,

the crude relative-risk of higher-PEEP strategies was 0.92 (95%CI = 0.791.06) before

adjustment, but 0.83 (95%CI = 0.720.97) after adjustment.

When using the R software for mediation analysis, there is an output called total effect,

estimated by simulation according to the principles described by Imai et al15-18. This output must

provide an equivalent result (for the main effect of randomizarion), although slightly different

41

because of the simulations and different estimation methods. After running 5000 simulations,

the results fitted very well with the estimates provided above:

Relative-risk (Hazard) P-Value

Total effect for VT-trials: 0.60 (0.480.74) P < 0.001

Total effect for PEEP-trials 0.82 (0.710.96) P = 0.01

The second logical step was the check described above (page 32, when discussing the

required adjustment for baseline-elastanceRS), where we tested if mediator changes,

theoretically assumed as beneficial, correlated with better survival. At this step, we added the

pre-adjustment of baseline-elastanceRS in the model, making sure that our analysis was not

biased by confounding effects of the severity of underlying lung disease (potentially inflating the

association between P and survival). Among our tested mediator-candidates, PEEP and

plateau pressures failed at this second step within the higher-PEEP trials. After accounting for

baseline elastance, both variables were not significantly associated with survival.

The next two steps involve the analysis of inclusion of the effects of superimposed-P (i.e.

residual variations in P after subtracting the P variations correlated with baseline lung

disease) in a model where the direct effects of randomization are already taken into account:

42

By using the notation indicated in the diagram above, this analysis involves the demonstration

of 2 separate steps (steps 3 and 4 , Figures S8-S9, above)13,20-22 :

Step-3: to show that superimposed-P was significantly correlated to treatment group

assignment. Accordingly, the coefficient "A" produced by the multivariate regression of the

superimposed-P variable over randomization must be significantly different from zero. The

signal of coefficient A is an important issue in this mediation analysis, since we want to prove

that reductions in P cause improved survival. Thus, it has to be negative. As a

counterexample, plateau-pressures did not pass this logical step for the PEEP-trials, since

randomization caused higher plateau-pressure (,i.e. a positive coefficient A, P < 0.001), but an

improved survival.

Step-4: to show that superimposed-P carried significant survival impact (with a significant

coefficient "B") when included in a multivariate model where treatment group was also included

(as well as baseline-elastanceRS). This procedure is equivalent to the Sobel test, or, when using

the R software for mediation, equivalent to the significance test for the ACME (average causal

mediation effect). The ACME considers the combined effects of steps 3 and 4: both steps have

to present some minimum effect-size to guarantee a significant path through the mediator2.

2 In the R package for Mediation Analysis, two models are fit, one modeling the effects of randomization

on the mediator, and a second one jointly modeling the effects of randomization (directly) and mediator (indirectly) on outcomes. Using Monte-Carlo simulations, a mediation proportion is estimated, indicating

43

Whenever the ACME has an important effect, the coefficient "C" (representing the direct effect

of randomization, after multivariate adjustment) becomes weaker after accounting for the

mediating effect (indirect path "A*B"). Ideally, in order to demonstrate complete mediation, "C"

should be virtually zero (non-significant), with a mediation proportion ~100%.

For testing all the steps above, we pre-adjusted our mediational model with the covariates

elected in Model-1 (Table 1, main text), also including the baseline-elastanceRS variable. We

observed that all those covariates were non-specific predictors of survival, uncorrelated with

treatment group. We repeated the stepwise checks for all mediation-candidates, using only one

mediator each time.

how much of the whole risk-reduction in the treatment arm could be explained by the indirect path in which randomization drives a change in the mediator, which then affects the outcome.

The rationale used by this package is the same described by Imai et al8-10

: suppose that randomized treatment is denoted as a z=0 (control) or z=1 (treatment), and the mediating variable for a patient is denoted as M(0) and M(1) if they got treatment 0 or 1. Further suppose that the outcome for a patient with mediating variable M who got treatment 0 is T(M,0) and if they got treatment 1, T(M,1). Then the average mediated effect is the average of (T(M(1),1) -T(M(0),1) and T(M(1),0) -T(M(0),0). In other words it is the effect we would see if the mediating variable changed but treatment did not. This effects needs to be estimated in a model including the treatment group, the mediating variable and all confounding variables the affect both the mediating variable and the outcome.

44

VT and PEEP were not independent mediators

Please, refer to the figures on the next pages

Figure S10:

Mediation in the Lower vs. Higher VT-trials:

Tested mediator: VT-changes driven by randomization

Figure S11:

Mediation in the Higher vs. Lower PEEP-trials:

Tested mediator: PEEP-changes driven by randomization

Additional details:

Except for P, no other mediation-candidate consistently passed through the stepwise

mediation tests. Of note, tidal volume was not a significant mediator in the lower-VT trials

(P=0.67, ACME, Figure S10), and PEEP was not a significant mediator in the higher-PEEP

trials (P=0.50, ACME; Figure S11).

45

Figure S10: Mediation in the Lower vs. Higher VT-trials:

Tested mediator: VT-changes driven by randomization


0%

Randomization Survival effect(0.48 0.75; P < 0.001 )

Mediational model: High vs. Low VT trials

Tested mediator: VT -changes driven by randomization


lower VT

-5.1 mL/kg

P = 0.009 ( Independent effect )

Hazard = 0.60

P = 0.67 ( N.S. effect )

Figure 3b :

Step 1:

Steps 3 & 4:


lower VT(-1 STD = - 2.8 mL/kg)

Survival effect


( 0.70 0.87 ; P < 0.001 )




P < 0.001



Hazard = 0.59

Note that randomization keeps an independent predictive effect at this last step, whereas VT does not. This suggests that the predictive

information provided by treatment group assignment (representing a bundle of intention to treat interventions) exceeds the information

provided by individual levels of VT.

Thus, tidal-volume cannot be considered as an independent mediator of the effects of randomization.



.



of Figure S8.

Top: The first step was the same as

in Figure S8

Middle: VT passed through the

second step.

Bottom: Finally, VT failed in the last

step required to characterize

mediation.

46

Figure S11: Mediation in the Higher vs. Lower PEEP-trials:

Tested mediator: PEEP-changes driven by randomization

Randomization Survival effect(0.72 0.97; P = 0.02 )

Mediational model: High vs. Low PEEP trials Tested mediator: PEEP-changes driven by randomization


Hazard = 0.83

Figure 3c :

Step 1:

Steps 3 & 4:

higher PEEP(+1 STD = + 4.5 cmH2O)

Survival effect


( 0.87 1.03; P = 0.19 )





0%higher PEEP+6.1 cmH2O

P < 0.001

P = 0.03 ( Independent effect )



Hazard = 0.82

P = 0.50 ( N.S. effect )


Note that randomization keeps an independent predictive effect at this last step, whereas PEEP does not. This suggests that the

predictive information provided by treatment group assignment (representing a bundle of intention to treat interventions) exceeds the

information provided by individual levels of PEEP.

Thus, PEEP per se cannot be considered as an independent mediator of the effects of randomization. Other strategies included in

the bundle, or other intermediate variables affected by high-PEEP strategy are likely more important.





of Figure S8.

Top: The first step was the same as

in Figure S9

Middle: PEEP already failed at this

second step, showing that a higher-

PEEP strategy may be beneficial as a

package (shown in step1), but not in

proportion to PEEP increments.

There was no dose-response to

PEEP.

Bottom: PEEP also failed at this last

step, without any independent effect.

47

II.10. P consistently mediates survival across /within trials

Please, refer to the table on the next page

Table S9:

Mediation effects of P within study-trials (intra-trial mediation effects)

Additional details:

The variable Trial was entered as a random-effect in both mediation models: the mixed-effects

linear regression modeling the effects of randomization on the mediator, and the mixed-effects

Cox regression modeling the effects of both, mediator and randomization on survival. The R

package for mediation was designed to perform multi-level mediation analysis, where

individuals may be correlated within groups (trials), and the different trials may have different

randomization processes. Using a conservative approach, we also considered a moderated

mediation analysis, in which the randomization process could be moderated by the trial

variable, i.e. each study could have a different effect of randomization (direct effect), a different

effect of P (mediating effect), and a different mediation proportion. In fact, this analysis

allowed us to test the intra-trial mediation effect of P, shown in Table S9.

As observed, the ACME had the same consistent trend in all of the nine studies, presenting a

significant value in many of the individual studies. These findings support our pooled analysis

presented in Figure S8 and S9. They also demonstrate that P mediates intra-trials and inter-

trials effects, i.e. differences between the two arms in each study, and differences in the net

effects of randomization across the studies, with some studies presenting larger benefits of

randomization than others.

48

Table S9 (website): Mediation effects of P within study-trials (intra-trial mediation effects)

ACME: Average causal mediation effect, calculated according to Imai et al.16,17

Note that the ACME had the same consistent trend in all of the nine studies, presenting a significant value in many of the individual studies. Thus,

P was always responsible for part of the observed effects of randomization (shown as the total-effect). These findings support the pooled

analysis presented in Figures S8-S9. They also suggest that P mediates intra-trials and inter-trials effects, i.e. differences between the two arms

in each study, and differences in the net effects of randomization across the studies, with some studies presenting larger benefits of randomization

than others.

Trial:

Total effect

(Hazard)

ACME

(Hazard)

P-value

(for ACME)

Amato et al.1 0.12 0.26 0.07

Stewart et al.2 0.63 0.72 0.14

Brochard et al.3 0.69 0.60 0.03

Brower et al.4 1.16 0.86 0.28

ARDSnetTV5 0.62 0.78 0.10

VT-trials combined 0.59 0.73 0.01

ARDSnetPEEP6 0.90 0.83 0.001

EXPRESS7 0.76 0.91 0.004

LOVS8 0.89 0.96 0.06

Talmor et al.9 0.52 0.97 0.46

PEEP-trials combined 0.83 0.92 < 0.001

49

III) DETAILS ON STATISTICS AND METHODS:

All statistical tests were two-sided.

III.1. Screening the dataset and compatibility analysis

After pooling the data from the 9 trials 1-9, patients who died or were weaned before the first 24

hours after randomization were excluded from the final analysis (9 patients). This procedure

was necessary because we planned to test the impact of ventilation variables best representing

the first 24 hours of the randomized strategy. Outcomes occurring before this time period would

characterize incomplete or undetermined risk exposure, potentially introducing bias in the Cox

survival model. We also reasoned that a minimum risk exposure of 24 hours was necessary to

affect the outcome, or to minimally override the effects of treatment received before

randomization.

After this first exclusion, we searched for the following incompatibilities in the dataset of each

patient: peak-pressure < plateau-pressure; plateau-pressure < PEEP; minute ventilation

(tidal-volume times respiratory-rate10%); mean-airway pressure > plateau-pressure or mean-

airway pressure < PEEP; weaning date later than death or discharge date; PaO2/FIO2> 600 or

PaO2/FIO2150 mmHg; arterial-pH < 6.7 or arterial-pH > 7.7.

When one value was obviously wrong, it was removed from the dataset. When incompatibilities

as above were found, without an obviously spurious datum, the whole set of two or three

incompatible variables were excluded from the data set. Afterwards, all outliers (> 3 standard

50

deviations from the population-sample mean) within baseline or ventilation variables were

assessed. If they were consistent with values on the previous or following days (i.e. within the

range of the individual mean along the time 3 STD), they were retained; otherwise they were

either within-patient-interpolated (when possible) or discarded. If there was more than one

value per day, we assumed that the non-outlier value was representative for that day.

In Tables S1-S2 and Figure S1 we present some descriptive statistics for these trials, after the

incompatibility analysis. In Table S3, we tested the univariate association with survival for each

of those screened covariates.

Finally, we tested the impact of excluding additional 87 subjects whose plateau-pressures were

below 10 cmH2O or Ps were below 5 cmH2O or PEEP values were below 5 cmH2O or tidal-

compliance was above 1.25 mL/kg/cmH2O, during the first day after randomization. We

reasoned that such values might indicate errors during data collection, or patients with mild

disease or low risk exposure who might decrease the sensitivity of our analysis. The mortality

rate of these excluded patients was lower (25.3%) than the average mortality of the remaining

population (35.3%; P = 0.07 for the comparison of mortality-rates between excluded vs.

remaining patients).

The performance of Models 1-5 was rechecked within this more restricted sample of 3466

patients. The relative-risk associated with one standard-deviation increment in driving-pressure

or other covariates minimally changed (maximum change of 1% in the relative-risks for all

covariates). After observing this lack of benefit of a more restricted sample, we ultimately

decided to present our results without this filter (as shown in Table 1). The slightly lower

number of patients presented in Table 1 was caused by missing data (see next session).

Whereas patients in the validation sample had their tidal-volumes adjusted according to the

ideal body weight (IBW), most patients in the hypothesis-generation sample had their tidal-

51

volumes adjusted according to actual body weight. The physiological bias generated by this

procedure, in case of prevalent obesity, would be that of a systematically lower tidal-

compliance (affecting both arms) in those studies using actual body weight. Since we found

similar mean values for normalized compliance in both samples (P = 0.23), without significant

differences in other markers of baseline lung disease, we concluded that such bias was unlikely

and that the use of actual body weight in these particular patients had the same physiological

consequences as the use of ideal body weight. In sum, 92% of patients in our combined

sample had their tidal-volumes adjusted according to ideal body weight, but we kept the term

IBW for all.

52

III.2. Missing data

In general, the amount of missing data was low (< 10% for individual variables) for first-day

ventilation variables, with only 4 cases of missing data related to main outcome. The same was

not true for plateau-pressure (and consequently P) at baseline, which was only available for

the second validation cohort.

There were no significant differences in the amount of missing data across the trials, except for

age, which was missing in 100% of the cases in one of the trials3, but complete in all other

studies. Thus, we did not consider any special treatment for missing cases of other covariates,

and the incomplete cases were simply excluded from multivariate analysis. To ensure that the

lack of information was at random, not introducing bias in our model, we subsequently tested a

missingness variable for each covariate (coded as 1 if missing, and zero if present) including

them in univariate survival models or in multivariate models where the original covariate was

replaced by its missingness one. From all missingness variables, only missing-age affected the

models, being significantly related to survival (P = 0.002).

Since we could retrieve the mean plus standard deviation of the variable age for all missing

cases (found in the original publication in which individual values were missing3, we replaced

the missing values by an imputed random value obeying similar distribution (mean = 57,

standard deviation = 15, as reported). After this procedure, which allowed us to include an

additional 113 cases in the multivariate analysis, we checked the performance of the

multivariate Model 1 before and after treatment for missing cases. This data handling was

advantageous, increasing the power of our analysis and minimally affecting the relative risk of

other covariates. Thus, this procedure was adopted in our final analysis.

53

III.3. Double-stratification (used for analysis in Figure 1 and Figure S5)

Description of the resampling procedures:

Resampling A: using the combined population shown in Table 1, we first stratified patients into

deciles of PEEP, forming 10 clusters of patients with similar PEEP. Second, we stratified each

cluster into quintiles of progressively higher P (at this point we had 50 classes). And third, for

each strata of P, we merged the 10 clusters of PEEP, finally forming 5 sub-samples with

matched average PEEP, but distinct average P.

Resampling B: Similarly to resampling A, we first formed 10 clusters of patients with similar

P, and then stratified each cluster into quintiles of progressively higher PEEP. Then, for each

strata of PEEP, we merged the 10 clusters of P, finally forming 5 sub-samples with matched

average P, but distinct average PEEP.

Resampling C: Similarly to resampling A, we first formed 10 clusters of patients with similar

plateau-pressure, and then stratified each cluster into quintiles of progressively higher PEEP.

Then, for each strata of PEEP, we merged the 10 clusters of plateau-pressure, forming 5 sub-

samples with matched average plateau-pressure, but distinct average PEEP.

Resampling D: Similarly to resampling C, we first formed 10 clusters of patients with similar

plateau-pressure, and then stratified each cluster into quintiles of progressively lower VT. Then,

for each strata of VT, we merged the 10 clusters of plateau-pressure, forming 5 sub-samples

with matched average plateau-pressure, but distinct average VT.

Resampling E: Similarly to resampling C, we first formed 10 clusters of patients with similar

plateau-pressure, and then stratified each cluster into quintiles of progressively lower P. Then,

for each strata of P, we merged the 10 clusters of plateau-pressure, forming 5 sub-samples

with matched average plateau-pressure, but distinct average P.

54

References

1. Amato MBP, Barbas CSV, Medeiros DM, et al. Effect of a Protective-Ventilation Strategy on

Mortality in the Acute Respiratory Distress Syndrome. N Engl J Med 1998;338:347-54.

2. Stewart TE, Meade MO, Cook DJ, et al. Evaluation of a ventilation strategy to prevent barotrauma

in patients at high risk for acute respiratory distress syndrome. Pressure- and Volume-Limited

Ventilation Strategy Group. N Engl J Med 1998;338:355-61.

3. Brochard L, Roudot-Thoraval F, Roupie E, et al. Tidal volume reduction for prevention of

ventilator-induced lung injury in acute respiratory distress syndrome. The M

Documents

Driving Pressure and Survival in ARDS-Amato-ESM-NEJM 2015