Extended Supplementary Materials

Manuscript ID: ICME-D-16-00739 R1

Extended supplementary materials

Extended methods

Data source and study population

Outcomerea (www.outcomerea.org) is an ongoing collaborative study group involving several French ICUs. In each participating ICU detailed clinical and outcome data of ICU admissions are prospectively entered by trained senior physicians or clinical research monitors into the Outcomerea database [1]. This study did not require individual patient consent because it involved research on a previously developed and approved (declared to the Comission Nationale de l’Informatique et des Libertés) database. In our study, eligible patients were consecutive adults (≥18 years) who were admitted in two ICUs from the Outcomerea database between 1997 and 2012 and who developed a first episode of GNB-related ventilator-associated pneumonia (VAP) and received appropriate empirical antimicrobial therapy.

Exclusion criteria were i) a time interval between the microbiological sample and the initiation antimicrobial therapy higher than 48 hours; ii) no beta-lactam as pivotal agent of the VAP-antimicrobial regimen; iii) a follow-up (after antibiotic treatment initiation) shorter than 5-days; and iv) polymicrobial VAP due to GNB and Gram positive cocci.

Definitions

As previously reported [1], VAP was defined as a persistent radiological pulmonary infiltrates combined with purulent tracheal aspirates and body temperature ≥38.5°C or ≤36.5°C and peripheral blood leukocyte count ≥10x109/L or ≤4x109/L. VAP diagnosis required systematic microbiological confirmation using quantitative culture from protected specimen brush (≥103 colony-forming unit (CFU)/mL), plugged telescopic catheter (≥103 CFU/mL), bronchoalveolar lavage fluid specimen (≥104 CFU/mL), or endotracheal aspirate (≥105 CFU/mL).

The intensity of DE was calculated as the difference between the ranking of the pivotal beta-lactams before and after DE. All VAP episodes were treated for 7 days.

Data Collection

The following data were collected at admission: demographic data admission category, chronic illness and comorbidities (using Knaus definition). Severity of illness was evaluated at ICU admission, and on a daily basis, using the following scores: Simplified Acute Physiology Score 2 (SAPS 2) and Sequential Organ Failure Assessment (SOFA) [2]. Multidrug-resistant bacteria carriage (including extended spectrum beta-lactamase-producing Enterobacteriaceae (ESBL-PE), methicillin-resistant Staphylococcus aureus but not carbapenemase-producing Enterobacteriaceae) was systematically screened on once weekly nasal and rectal samples. Pseudomonas aeruginosa resistant to ticarcillin or carbapenems and AmpC-hyperproducing Enterobacteriaceae obtained from clinical samples, whatever it was, were recorded.

Clinical outcome data such as VAP recurrence rate, length of mechanical ventilation (after initiation of an appropriate antimicrobial treatment of VAP) and mortality rate were also recorded.

This work aimed to evaluate the impact of DE on patient outcome and, at a more general level, the effect of DE on antibiotic consumption and on multiresistant strains carriage acquisition. Secondary end-points were the VAP relapse rate, length of mechanical ventilation, length of ICU stay, and 28 day-mortality rate.

The median (IQR) length of follow-up of our population was 43 [24 ; 71] days after VAP antimicrobial treatment initiation. As we wanted to maximize chances to detect multiresistant strain carriage acquisition following antimicrobial therapy while minimizing the lost-of follow-up patients, we chose a 21-day time period after VAP treatment initiation. Antibiotic (global, carbapenems, rank-4 molecules and fluoroquinolones)-free days were also evaluated during the VAP treatment period.

Statistical analysis

Categorical variables were described as number (%) and continuous variables as median (interquartile range (IQR)). For comparisons, we used the chi-square test for categorical data and the Mann-Whitney U test for continuous data.

Factors associated with DE by univariate analysis were entered into a multivariate logistic regression model using a stepwise selection procedure. Factors yielding values of P < 0.2 by univariate analysis were included in the initial multivariate model into a multiple logistic regression model and retained in the final model when associated with values of P < 0.05.

We used linear regression and logistic regression adjusted on factors associated with DE (medical type, Enterobacteriaceae VAP and empirical treatment including ciprofloxacin) to assess the DE effect on continuous and categorical variables.

Values of P < 0.05 were considered significant. Analyses were performed using SAS 9.4 software (SAS Institute, Cary, NC).

Extended results

Table S1. 6-rank classification of pivotal beta-lactams in ascending order of ecological consequences (according to[3])

Rank / Drug(s)
6 / Imipenem
Meropenem
Doripenem
5 / Ertapenem
4 / Piperacilin+Tazobactam
Ticarcilin+ClavulanicAcid
4th generation cephalosporin,
Antipseudomonal 3rd generation cephalosporin
3 / 3rd generation cephalosporin
Ureido/carboxy-penicillin
2 / Amoxicillin+ClavulanicAcid
1 / Amoxicillin

Table S2. Antimicrobial molecules used as empirical therapy during the first 72 hours

No-de-escalation (n=61) / De-escalation (n=70) / P value*
Pivotal therapya
- Carbapenem (%) / 10 (16) / 23 (32) / 0.03
- Group 4 molecule (%)b / 41 (67) / 54 (77) / 0.2
- 3rd generation cephalosporin (%) / 14 (23) / 26 (37) / 0.08
- Amoxicillin-clavulanic acid (%) / 6 (10) / 6 (9) / 0.8
Combination therapy (%) / 45 (74) / 53 (76) / 0.8
- Combination therapy including fluoroquinolones (%) / 18 (29) / 6 (9) / <.01
- Combination therapy including aminoglycosides (%) / 27 (44) / 47 (67) / <.01
Variables are described as number (percentage). The chi-square test was used for comparison. aOne patient can have more than one pivotal antimicrobial during the first 3 days. b Group 4 molecules included: Piperacilin+Tazobactam, Ticarcilin+Clavulanic acid, 4th generation cephalosporin and antipseudomonal 3rd generation cephalosporin (see table S1)

Table S3. Characteristics of patients

No-de-escalation (n=61) / De-escalation (n=70) / P value
Male / 42 (68.9) / 57 (81.4) / 0.09
Chronic illness
- Liver cirrhosis / 7 (11.5) / 4 (5.7) / 0.24
- NYHA IV / 3 (4.9) / 9 (12.9) / 0.12
- Respiratory disease / 12 (19.7) / 13 (18.6) / 0.87
- Terminal Kidney disease / 2 (3.3) / 0 (0) / 0.13
- Immunocompromised patient / 7 (11.5) / 5 (7.1) / 0.39
Category at admission
- Medical / 48 (78.7) / 39 (55.7) / <.01
- Surgical / 39 (55.7) / 48 (78.7)
SAPS at admission / 50 [34 ; 68] / 43 [33 ; 54] / 0.14
SAPS at initiation of VAP antimicrobial treatmenta / 38 [33 ; 48] / 39.5 [29 ; 46] / 0.81
SAPS on day 2 of VAP antimicrobial treatment / 37 [29 ; 45] / 35 [26 ; 44] / 0.20
Delta SAPS between initiation and day 2 of VAP antimicrobial treatment / -1 [-8 ; 4] / -3 [-8 ; 3] / 0.45
SOFA on day 1 after admission / 8 [5 ; 11] / 8 [5 ; 10] / 0.82
SOFA at initiation of VAP antimicrobial treatment / 5 [3 ; 8] / 6 [3 ; 9] / 0.46
SOFA on day 2 of VAP antimicrobial treatment / 4 [3 ; 7] / 4.5 [3 ; 7] / 0.94
Delta SOFA between initiation and day 2 of VAP antimicrobial treatment / 0 [-1 ; 1] / 0 [-3 ; 0] / 0.35
Enterobacteriaceae-related VAP / 24 (39.3) / 39 (55.7) / 0.06
Pseudomonas aeruginosa-related VAP / 29 (47.5) / 24 (34.3) / 0.12
Haemophilus influenzae-related VAP / 6 (9.8) / 7 (10) / 0.99

Table S4. Predictors of de-escalation in multivariate analysis

Parameter / OR / 95% CI / P
Medical category at admission / 0.388 / 0.168 / 0.896 / 0.0266
Enterobacteriaceae-related VAP / 2.368 / 1.109 / 5.054 / 0.0259
Empirical combination therapy including fluoroquinolones (%) / 0.284 / 0.099 / 0.815 / 0.0193

Factors associated with DE by univariate analysis (yielding values of P < 0.2) that were entered into a multivariate logistic regression model using a stepwise selection procedure: medical category at admission, Enterobacteriaceae-related VAP, Empirical combination therapy including fluoroquinolones, Carbapenem as empirical therapy and SAPS at admission.

Factors were retained in the final model when associated with values of P < 0.05.

References

1. Planquette B, Timsit JF, Misset BY, Schwebel C, Azoulay E, Adrie C, Vesin A, Jamali S, Zahar JR, Allaouchiche B, Souweine B, Darmon M, Dumenil AS, Goldgran-Toledano D, Mourvillier BH, Bedos JP, Group OS. Pseudomonas aeruginosa ventilator-associated pneumonia. predictive factors of treatment failure. American journal of respiratory and critical care medicine 2013: 188(1): 69-76.

2. Timsit JF, Fosse JP, Troche G, De Lassence A, Alberti C, Garrouste-Orgeas M, Bornstain C, Adrie C, Cheval C, Chevret S, Outcomerea Study Group F. Calibration and discrimination by daily Logistic Organ Dysfunction scoring comparatively with daily Sequential Organ Failure Assessment scoring for predicting hospital mortality in critically ill patients. Crit Care Med 2002: 30(9): 2003-2013.

3. Weiss E, Zahar JR, Lesprit P, Ruppe E, Leone M, Chastre J, Lucet JC, Paugam-Burtz C, Brun-Buisson C, Timsit JF, De-escalation' Study G, De-escalation' Study G. Elaboration of a consensual definition of de-escalation allowing a ranking of beta-lactams. Clin Microbiol Infect 2015: 21(7): 649 e641-649 e610.

Acknowledgements: the authors thank Celine Feger, MD, (EMIBiotech), for her support in editing the manuscript.

Author Contributions: Conception and design, E.W., J.-R.Z. and J.-F.T. Analysis and interpretation: S.R., W.E., E.W., J.-R.Z. and J.-F.T. Drafting the manuscript for important intellectual content: E.W., J.-R.Z. and J.-F.T, S.R., M.G.-O., A.B., C.S., B.M. Guarantor: J.-F.T.

Collaborator list:

MEMBERS OF THE OUTCOMEREA STUDY GROUP

Scientific Committee: Jean-François Timsit (Medical and Infectious Diseases ICU, Bichat-Claude Bernard Hospital, Paris, France; UMR 1137 Inserm –Paris Diderot university IAME, F75018, Paris); Elie Azoulay (Medical ICU, Saint Louis Hospital, Paris, France); Yves Cohen (ICU, Avicenne Hospital, Bobigny, France); Maïté Garrouste-Orgeas (ICU, Saint-Joseph Hospital, Paris, France); Lilia Soufir (ICU, Saint-Joseph Hospital, Paris, France); Jean-Ralph Zahar (Infection Control Unit, Angers Hospital, Angers, France); Christophe Adrie (ICU, Delafontaine Hospital, Saint Denis, and Physiology, Cochin Hospital, Paris, France); Michaël Darmon (Medical ICU, Saint Etienne University Hospital, St Etienne, France); and Christophe Clec’h (ICU, Avicenne Hospital, Bobigny, and U823, Grenoble, France).

Biostatistical and Information System Expertise: Jean-Francois Timsit (Medical and Infectious Diseases ICU, Bichat-Claude Bernard Hospital, Paris, France; UMR 1137 Inserm –Paris Diderot university IAME, F75018, Paris); Corinne Alberti (Medical Computer Sciences and Biostatistics Department, Robert Debré Hospital, Paris, France); Adrien Français (Integrated Research Center U823, Grenoble, France); Aurélien Vesin (OUTCOMEREA organization and Integrated Research Center U823, Grenoble, France); Stephane Ruckly (OUTCOMEREA organization and Inserm UMR 1137 IAME, Paris); Christophe Clec’h (ICU, Avicenne Hospital, Bobigny, and Integrated Research Center U823, Grenoble, France); Frederik Lecorre (Supelec, France); Didier Nakache (Conservatoire National des Arts et Métiers, Paris, France); and Aurélien Vannieuwenhuyze (Tourcoing, France).

Investigators of the OUTCOMEREA Database: Christophe Adrie (ICU, Delafontaine Hospital, Saint Denis, and Physiology, Cochin Hospital, Paris, France); Bernard Allaouchiche (ICU, Pierre Benite Hospital, Lyon, France); Laurent Argaut (Medical ICU, Hospices Civils de Lyon, Lyon, France); Claire Ara-Somohano (Medical ICU, University Hospital, Grenoble, France); Elie Azoulay (Medical ICU, Saint Louis Hospital, Paris, France); François Barbier (Medical ICU, Orleans, France), Jean-Pierre Bedos (ICU, Versailles Hospital, Versailles, France); Julien Bohé (ICU, Hôpital Pierre Benite, Lyon France), Lila Bouadma (ICU, Bichat Hospital, Paris, France); Alexandre Boyer (ICU, Pellegrin Hospital, Bordeaux, France); Christine Cheval (ICU, Hyeres Hospital, Hyeres, France); Christophe Clec’h (ICU, Avicenne Hospital, Bobigny, France); Jean-Pierre Colin (ICU, Dourdan Hospital, Dourdan, France); Michael Darmon (ICU, Saint Etienne Hospital, Saint Etienne, France); Anne-Sylvie Dumenil (Antoine Béclère Hospital, Clamart, France); Adrien Descorps-Declere (ICU, Antoine Béclère Hospital, Clamart, France); Jean-Philippe Fosse (ICU, Avicenne Hospital, Bobigny, France); Marc Gainier (hôpital la Timone, Marseille, France), Akim Haouache (Surgical ICU, H Mondor Hospital, Creteil, France); Samir Jamali (ICU, Dourdan, Dourdan Hospital, Dourdan, France); Hatem Khallel (ICU, Cayenne General Hospital, Cayenne, France); Alexandre Lautrette (ICU, G Montpied Hospital, Clermont-Ferrand, France); Guillaume Marcotte (Surgical ICU, Hospices Civils de Lyon, Lyon, France); Eric Le Miere (ICU, Louis Mourier Hospital, Colombes, France); Maxime Lugosi (Medical ICU, University Hospital Grenoble, Grenoble, France); Laurent Montesino (ICU, Bichat Hospital, Paris, France); Bruno Mourvillier (ICU, Bichat Hospital, Paris, France); Benoît Misset (ICU, Saint-Joseph Hospital, Paris, France); Delphine Moreau (ICU, Saint-Louis Hospital, Paris, France); Bruno Mourvillier (ICU, Bichat Hospital, Paris, France); Laurent Papazian – Hopital Nord, Marseille, France), Benjamin Planquette (ICU, Versailles Hospital, Versailles, France); Etienne Pigné (ICU, Louis Mourier Hospital, Colombes, France); Bertrand Souweine (ICU, G Montpied Hospital, Clermont-Ferrand, France); Carole Schwebel (ICU, A Michallon Hospital, Grenoble, France); Gilles Troché (ICU, Antoine Béclère Hospital, Clamart, France); Marie Thuong (ICU, Delafontaine Hospital, Saint Denis, France); Guillaume Thierry (ICU, Saint-Louis Hospital, Paris, France); Dany Toledano (ICU, Gonesse Hospital, Gonesse, France); and Eric Vantalon (SICU, Saint-Joseph Hospital, Paris, France); Emmanuel Weiss (Beaujon Hospital, Clichy France).

Study Monitors: Julien Fournier, Caroline Tournegros, Stéphanie Bagur, Vanessa Vindrieux, Loic Ferrand, Nadira Kaddour, Boris Berthe, Samir Bekkhouche, Sylvain Anselme, Kaouttar Mellouk, Sylvie Conrozier, Igor Theodose, Veronique Deiler, and Sophie Letrou.

Funding information: The research leading to these results has received partial support from the Innovative Medicines Initiative Joint Undertaking under grant agreement n° [115523] COMBACTE, resources of which are composed of financial contribution from the European Union's Seventh Framework Programme (FP7/2007-2013) and EFPIA companies, in kind contribution.