37th International Symposium on Intensive Care and Emergency Medicine (part 1 of 3)

P1 Inhibition of jak/stat3 signaling pathway by small molecule r548 prevents inflammation in experimental murine lung injury model

V. Karavana¹, I. Smith², G. Kanellis³, I. Sigala¹, T. Kinsella², S. Zakynthinos¹

¹George P. Livanos and Marianthi Simou Laboratories, Athens, Greece; ²Rigel Pharmaceuticals, Inc, South San Francisco, CA, USA; ³Evangelismos Hospital, Athens, Greece

Introduction: Aberrant inflammation is a hallmark of acute respiratory distress syndrome (ARDS) pathophysiology. JAK/STAT3 pathway is critical for macrophages and neutrophils activation and persistent inflammation. This study aims to investigate the therapeutic potential of inhibiting JAK1/3 activity using the small-molecule R548 inhibitor in LPS induce lung injury model.

Methods: Lung injury was induced in adult male C57BL/6 mice, by intratracheal LPS administration followed by post subcutaneous injection of R548 inhibitor R548 inhibitor, prodrug for the active compound, R507 (Rigel Pharmaceuticals Inc.). Mice sacrificed at 6 h and 24 h after LPS administration. Lung inflammation was examined by protein content, number and type of inflammatory cells in bronchoalveolar lavage fluid (BALF). Protein expression levels of JAK1, p-STAT3, ERK1/2 were analysed by Western blotting.

Results: LPS administration increased BALF cellularity, total protein content, and neutrophils cells number at both 6 h and 24 h. Elevated levels of JAK1, p-STAT3 and ERK1/2 protein expression were observed. In addition, post LPS treatment with the JAK 1/3 inhibitor significantly reduces BALF protein content (P < 0.05), total cells number (P < 0.01), neutrophils cells number (P < 0.01) as early as 6 h. Moreover, R548 treatment decreased JAK1 protein expression by 2 fold (P < 0.01) and p-STAT3 levels by 2.7 fold (P < 0.001) below the LPS group.

Conclusions: These data suggest the JAK/STAT3 signaling pathway plays a critical role in ARDS mediated lung inflammation and injury. Additional studies are warranted to further investigate JAK/STAT3 inhibition as a therapeutic treatment for this serious and life threatening disease.

P2 Stable genetic alterations of CXCR7 regulate the CXCL12/CXCR7 axis, a new "passport" ± for the homing of mscs

L. Liu, J. Chen, X. Zhang, A. Liu, F. Guo, S. Liu, Y. Yang, H. Qiu

Department of Critical Care Medicine, Nanjing Zhongda Hospital, School of Medicine, Southeast University, Nanjing, China

Introduction: Mesenchymal stem cells (MSCs) have several properties that make them attractive therapeutic candidates for treatment of acute disease, but in vivo homing, MSCs does not appear to be highly efficient. CXCL12/CXCR7 axis not only improves the motility of stem cells, but also regulates many essential biological processes. The aim of this study is to evaluate the effects of overexpressing or suppressing CXCR7 on proliferation and migration abilities of mice mesenchymal stem cells.

Methods: The lentivirus vector overexpressing and suppressing the murine CXCR7 gene was transducted into mMSCs. The transfection efficiency of LV in passage 20 transduced-mMSCs was identified using fluorescence microscopy, and the percentage of ZsGreen positive cells was determined by flow cytometry analysis (FCM). CXCR7 mRNA expression in mMSCs was verified by quantitative real-time PCR, and CXCR7 protein expression was analyzed by FCM. The effect of CXCR7 on the migration of mMSCs was evaluated using the scratch assay and the transwell migration assay. In transwell assay, 50 ng/ml CXCL12 was added in two groups to simulation inflammation microenvironments. The ELISA assay was used to detect the concentration of VCAM-1, CXCL12 the supernatant.

Results: The efficiencies of the lentiviral vector transduction of MSC-OE-CXCR7 (overexpression of CXCR7), MSC-OENC-CXCR7 (normal control of overexpression of CXCR7), MSC-Sh-CXCR7 (suppression of CXCR7) and MSC-ShNC-CXCR7 (normal control of suppression of CXCR7) after 20 passages in mMSCs were 91.29%, 91.39%, 91.69% and 91.28% respectively. The CXCR7 mRNA and protein expression were significantly higher in the MSC-OE-CXCR7 cells than in the MSC-OENC-CXCR7 cells, and were significantly lower in the MSC-Sh-CXCR7 when compared with the MSC-ShNC-CXCR7. Moreover, CXCR7 gene overexpression promoted mMSCs migration. In contrast, the suppression of CXCR7 inhibited mMSCs migration when compared with the MSC-ShNC-CXCR7 group. Overexpression CXCR7 increased MSC-secreted CXCL12 and VCAM-1, which contributed to the improvement of mMSCs homing.

Conclusions: Overexpression CXCR7 improved the homing abilities of mMSCs, which might attribute mainly to increasement of CXCL12 and VCAM-1 secreted by mMSCs.

Fig. 1 (abstract P2).

Legend 1: Figure 1 Long-term transgene expression efficiency in mMSCs after lentiviral vector transduction

Fig. 2 (abstract P2).

Legend 2: Figure 2 The effect of CXCR7 on the migration of mMSCs. The concentration of VCAM-1 and CXCL12 in the supernatant of Transwell Chambers was examined using ELISA assay.

P3 Withdrawn

P4 Spectrum study severe hypoxemia: prevalence, treatment and outcome in 2016

D. G. Grimaldi¹, SR The SRLF Trial Group²

¹Hôpital Erasme, Brussels, Belgium; ²SRLF, Paris, France

Introduction: Limited information exists about the prevalence, subsequent management, and outcomes of hypoxemia among ICU patients.

The aim of the study was to assess the prevalence of hypoxemia among ICU patients and to stratify them according to the severity of hypoxemia (PaO2/FIO2 between 300 and 201 in mild, between 200 and 101 in moderate, and <101 mm Hg in severe). Management and outcomes of hypoxemic patients and the proportion of them who met the criteria for acute respiratory distress syndrome (ARDS) were analyzed

Methods: This study was an international, multicenter, 1-day point prevalence study conducted in 117 units during the spring of 2016. All patients already hospitalized or newly admitted the day of the study were susceptible to be enrolled.

Hypoxemia was defined as a P/F ratio of 300 mmHg or less

Results: Of 1604 patients included, 859 (54%) were hypoxemic, 440 (51%) were mildly, 345 (40%) moderately and 74 (9%) severely hypoxemic. Among hypoxemic patients, 183 (21%) had ARDS (37% mild, 46% moderate and 16% severe). Characteristics of patients according to hypoxemia severity are reported in the Table 1. Pneumonia was the main cause of hypoxemia (53%) and of ARDS (79%). Modalities of oxygen treatment vari.

Among hypoxemic patients under invasive mechanical ventilation, 77% received a tidal volume of 8 mL/kg or less of predicted body weight with a median PEEP level of 6 [5–10] cmH2O. Among ARDS patients, 145 (83%) received a tidal volume of 8 mL/kg or less with a median PEEP of 8 [6–12] cmH2O and a median plateau pressure of 23 [19–27] cmH2O.

In the whole population, median ICU length of stay was 12 days [5–28] and the ICU mortality was 20%. In the hypoxemic population, median ICU length of stay appeared to be higher (16 [7–32] days) and further increased in ARDS patients (20 [11–38] days, P < 0.001). ICU mortality was 27% in hypoxemic patients (21% in mild hypoxemic patients, 26% in moderate hypoxemic patients, and 50% in severe hypoxemic patients, P < 0.001)

Conclusions: Hypoxemia affected more than half of the patients in ICU, whereas only 21% of hypoxemic patients met ARDS criteria. Mortality was significantly associated with severity of hypoxemia.

Table 1 (abstract P4). See text for description

P5 Correlation of horowitz ratio with oxygen saturation index in ARDS

E. Kaya, O. Acicbe, I. Kayaalp, S. Asar, M. Dogan, G. Eren, O. Hergunsel

Bakirkoy Dr. Sadi Konuk Training and Research Hospital, Istanbul, Turkey

Introduction: In definition of ARDS PaO2 and PaO2/FiO2 (Horowitz) ratio are the most reliable parameters used. In this study it is aimed to investigate the correlation of Horowitz ratio with oxygenation saturation index (OSI) derived from SpO2 instead of PaO2.

Methods: Demographical data of 307 patients treated with mechanical ventilation in our ICU between 01.01.2014–31.12.2015 due to respiratory failure were revised from iMDsoft (Metavision) system together with FiO2, SpO2, PaO2, Paw parameters. Oxygenation saturation index was calculated with the formula using [FiO2 x Paw]/SpO2. ROC (receiver operator characteristic) analysis was used to search for the relation of Horowitz ratio and OSI.

Results: The median age of the patients were 58,5 years (IQR 18–96). For the patients with Horowitz ratio < 100 cut-off value for OSI was found to be 7,0768 (sensitivity 94.4% and specificity 86.9%). OSI cut-off value was 3,7178 for Horowitz ratio < 200 (sensitivity 90.2% and specificity 79%). For those with Horowitz ratio < 300, OSI cut-off value was calculated as 2,561 (sensitivity 83.3% and specificity 72%).

Conclusions: OSI was defined for ARDS diagnosis in pediatric population which includes the parameter of mean airway pressure in its formula rendering this index a more objective parameter for ventilatory support instead of PaO2/FiO2 ratio. It was a big controversy before Berlin definition that ARDS definition was relying on cirteria which do not directly consider the extent of ventiatory support on oxygenation, however there is still a debate that Berlin definition did not bring a solution in that aspect. Therefore, we think that this study, which we aim to reveal the OSI values that correlate with Horowitz values of mild, moderate and severe ARDS, should be supplemented with other studies and validated by further prospective studies, considering OSI as a reflection of positive pressure changes affecting oxygenation in ARDS.

References

1. Thomas NJ, et al. Pediatr Crit Care Med. 2010; 11(1): 12–17

2. Rotta AT, et al. Rev Bras Ter Intensiva. 2015;27(3):266–273

P6 Ards features of the severe burn patients from colective tragedy

D. Pavelescu, I. Grintescu, L. Mirea

Emergency Hospital Floreasca, Bucharest, Romania

Introduction: ARDS is an independent risk factor for death in burn patients, with appreciable mortality of 26–58%. More than 30% of thermal injured patients have a concomitant smoke inhalation injury

Methods: A retrospective observational study which include 14 young patients 21–36 y.o. with severe burns to the face, neck, chest, extremities, admitted in the ICU of level I trauma Center after a deadly fire in Bucharest in an enclosed space with air temperature in 60 seconds between 900–1500 degrees C, in which 64 people were killed and 147 severely injured.

All of them had signs and symptoms of acute hypoxia requiring mechanical ventilation.

We asses the incidence of smoke inhalation injury, the CO toxicity, the etiology, the development time, the relation between severity of ARDS and % of BSA, the mortality

Results: All of them have smoke inhalation injury and CO toxicity, the etiology was multifactorial (trauma, multiple transfusion, sepsis, smoke injury, resuscitated cardio-respiratory arrest), there was a strong correlation between the severity of ARDS and the % of Burn Surface Area, the mortality was extremely high

Conclusions: The severity of burn associated with inhalatory injury dramatically increase the incidence of ARDS and despite new promising therapeutic interventions, ARDS still remain a devastating disease in severe burn patients.

Fig. 3: (abstract P6).

See text for description

P7 Relationship between energy load and lung regional inflation status in ARDS patients: a CT scan study

M. Guanziroli¹, M. Gotti², A. Marino¹, M. Cressoni¹, G. Vergani¹, C. Chiurazzi³, D. Chiumello¹, L. Gattinoni⁴

¹Università degli Studi di Milano, Milano, Italy; ²ASST Santi Paolo e Carlo, Milano, Italy; ³Humanitas, Rozzano, Italy; ⁴University of Gottingen, Gottingen, Germany

Introduction: Tidal volume, pressure and flow are components of the energy load (EL). We investigated the relationship between intra-tidal lung inflation status variation and EL in mechanically ventilated ARDS patients.

Methods: Twenty-eight ARDS patients underwent end-inspiratory and end-expiratory low dose CT scans at PEEP 5 and 15 cmH2O, maintaining the same tidal volume 7 ± 1.8 mL/kg and respiratory rate 15 ± 4.2 breaths/min. Quantitative CT scan analysis was performed to obtain the amount of not, poorly, well and over inflated tissue (g). EL was computed as the area between the inspiratory limb of pressure-volume curve and the y axis, summed to the energy needed to inflate the PEEP volume. EL at PEEP 5 and 15 cmH2O was then normalized by the End-Expiratory Lung Volume at 5 cmH2O (EELV5).

Results: EL/EELV5 (mJ/mL) is lower at PEEP 5 than 15 cmH2O (1.3 ± 0.73 and 2.4 ± 0.95, p < 0.001). Higher EL/EELV5, respectively for PEEP 5 and 15 cmH2O, is associated with a decrease in not inflated tissue (Δ not inflated tissue = −22.87–61.38*EL/EELV5, r2 = 0.28, p < 0.0001 and Δ not inflated tissue = −4.35–24.59*EL/EELV5, r2 = 0.15, p < 0.01) and with an increase in well inflated tissue (Δ well inflated tissue = 67.72 + 55.91*EL/EELV5, r2 = 0.27, p < 0.0001 and Δ well inflated tissue = −25.13 + 51.91*EL/EELV5, r2 = 0.46, p < 0.0001) at the end of inspiration (Fig. 4). No relationships were found between poorly or over inflated tissue and total energy load.

Conclusions: Higher PEEP is associated to higher EL. Higher EL, increasing lung stress and intra-tidal opening and closing, could be unprotective.

Fig. 4 (abstract P7).

a and b: PEEP 5 cmh₂O; c and d 15 cmh₂O

P8 Relationship between total or lung superimposed pressure and absolute esophageal pressure in ARDS: a CT scan study

M. Guanziroli¹, M. Gotti², G. Vergani¹, M. Cressoni¹, C. Chiurazzi³, A. Marino¹, S. Spano¹, D. Chiumello¹, L Gattinoni⁴

¹Università degli Studi di Milano, Milano, Italy; ²ASST Santi Paolo e Carlo, Milano, Italy; ³Humanitas, Rozzano, Italy; ⁴University of Gottingen, Gottingen, Germany

Introduction: In ARDS patients lung collapse is due to lung superimposed pressure (SP) [1].

SP seems to change as a function of pleural pressure, and both these pressures change as a function of the sterno-vertebral level [2].

Absolute esophageal pressure (Pes) reflects mid-lung pleural pressure, according to the level where the esophageal balloon is placed [1]. However, Pes is unrelated to all morphological indexes of disease severity (not inflated tissue, lung weight, functional residual capacity) and is unrelated to maximal SP [1].

Our aim is to verify if Pes is related to the total and/or the lung SP at the esophageal catheter balloon level.

Methods: Ninety-two ARDS patients underwent an end-expiratory CT scan at PEEP 5 cmH2O. Quantitative CT scan analysis was performed to compute the total and the lung SP at the esophageal catheter balloon level in the slice above the diaphragm.

Pes was recorded at PEEP 5 cmH2O.

Results: The total and the lung SP and Pes at PEEP 5 cmH2O are weakly related, respectively (Fig. 5):

• - total SP (cmH2O) = 11.97 + 0.17 * Pes (cmH2O), r2 = 0.08, p < 0.01

• - lung SP (cmH2O) = 3.47 + 0.13 * Pes (cmH2O), r2 = 0.07, p = 0.01

As shown, Pes is greater than either the total and the lung SP at the level of the esophageal balloon.

Conclusions: If the increased total or lung SP is the main mechanism for lung collapse in ARDS patients and PEEP keeps the lung open mainly by counter-balancing the SP, Pes could not be used to select the best PEEP.

References

[1] Pelosi P. Am J Respir Crit Care Med 2001; 164:122–130

[2] Bottino N. Crit Care 2000; 4:P115

Fig. 5 (abstract P8).

See text for description

P9 Relationship between energy load and lung inhomogeneity in ARDS patients: a CT scan study

M. Guanziroli¹, M. Gotti², G. Vergani¹, A. Marino¹, M. Cressoni¹, C. Chiurazzi³, D. Chiumello¹, L. Gattinoni⁴

¹Università degli Studi di Milano, Milano, Italy; ²ASST Santi Paolo e Carlo, Milano, Italy; ³Humanitas, Rozzano, Italy; ⁴University of Gottingen, Gottingen, Germany

Introduction: In ARDS patients, lung is inhomogeneous [1]. Energy load (EL) describes the interaction between the ventilator and the respiratory system [2]. We investigated the relationship between intra-tidal lung inhomogeneity variation and EL in ARDS patients.

Methods: Twenty-eight ARDS patients underwent a series of end-inspiratory and end-expiratory low dose CT scans at 2 different levels of PEEP, 5 and 15 cmH2O, maintaining the same tidal volume 7 ± 1.8 mL/kg and respiratory rate 15 ± 4.2 breaths/min.

Lung inhomogeneity extent is defined as the fraction of lung volume presenting an inflation ratio greater than 1.61 (95th percentile of homogeneous lungs) [1].

The airway pressure-volume curve at the 2 ventilatory settings was obtained. EL was computed as the area between the inspiratory limb of that curve and the volume axis, summed to the energy needed to inflate the PEEP volume [2].

Results: At both PEEP levels, lung inhomogeneity variation (end inspiration - end expiration) is related to total EL (Fig. 6):

• PEEP 5: Δ inhomogeneity(%) = 4.38–0.01*EL(mJ), r2 = 0.31, p < 0.001

• PEEP 15: Δ inhomogeneity(%) = 2.78–0.002*EL(mJ), r2 = 0.26, p < 0.0001.

Conclusions: During mechanical ventilation, the EL applied to the respiratory system is spent to increase homogeneity of the lung parenchyma. We could speculated that EL is spent at the interfaces between regions with different inflation status.

References

[1] Cressoni M. Am J Respir Crit Care Med 2014; 189:149

[2] Gattinoni L. Intensive Care Med 2016; 42:1567

Fig. 6 (abstract P9).

a PEEP 5 cmH₂O; a PEEP 15 cmH₂O

P10 Does galunisertib reduce edema in experimental ARDS? An exploratory study

F. Massaro¹, A. Moustakas², S. Johansson², A. Larsson³, G. Perchiazzi³

¹Policlinico di Bari, Bari, Italy; ²Uppsala University Biomedical Center, Uppsala, Sweden; ³Akademiska Sjukhuset, Uppsala University, Uppsala, Sweden

Introduction: The transforming growth factor β (TGF-β) pathway is activated in experimental ARDS and is associated with increased pulmonary edema (1) because it reduces transepithelial sodium transport by inducing endocytosis of ENaC (the epithelial sodium channel). In this way TGF-β decreases the osmotic gradient for re-absorption of water from the alveolar and interstitial spaces (2). We hypothesized that blocking TGF-β signaling with Galunisertib (LY2157299 monohydrate, a TGFβ R1 kinase inhibitor) would reduce edema and tested this in a porcine model of ARDS using aerosolized Galunisertib.

Methods: Five piglets (25–30 kg) were randomized to receive (n = 3) or not (n = 2) Galunisertib 50 μM by aerosol. The animals were anesthetized, tracheotomized and ventilated in volume controlled mode (VT: 8 ml/kg, FiO2: 70%, RR:30, PEEP:5 cmH2O, I:E = 1:2). Cardiac output and extravascular lung water (EVLW) were continuously measured (PiCCO, Pulsion Medical System). We measured pressure/volume curves and arterial pH, paO2, paCO2, HCO3- at baseline, immediately after induction of ARDS and then hourly for 6 h. ARDS was induced by repeated lung lavages and injurious ventilation. After the establishment of the ARDS model, Galunisertib was administered and the experiment was continued for 6 h, after which the animals were sacrificed. Histological samples from lungs, heart, kidney and liver were taken. Activation of the TGF-β pathway was immunohistochemically evaluated using an antibody against phosphorylated Smad2 (Small Mother Against Decapentaplegic) and activation of mechanical signaling in the lungs was evaluated using an antibody against phosphorylated FAK (Focal Adhesion Kinase). Moreover, we measured the wet/dry weight ratio of the lung samples.

Results: Two animals did not survive the experiment. The sample size is too small to draw any strong conclusion. However, we could not find any major difference in EVLW, wet/dry ratio, blood gas parameters or in lung mechanics between the two groups during the tested time period. However, the lung samples from the treated group showed a lower degree of TGF-β pathway activation, while the control samples from liver or heart showed no influence by Galunisertib.

Conclusions: We cannot prove or disprove that TGF-β blockade reduces pulmonary edema associated with experimental ARDS. However, the TGF-β blocker seemed to reach its target as indicated by reduced activation of the TGF-β pathway in the lungs, suggesting that aerosol administration is feasible.

References

1. Acta Anaesthesiol Scand. 2016;60(1):79–92,

2. Proc Natl Acad Sci USA. 2014;111(3):E374-83

P11 Stable over expression of p130 and E2F4 makes a difference on multipotential differentiation in bone marrow derived MSCs

X. W. Zhang, F. M. Guo, J. X. Chen, M. Xue, Y. Yang, H. B. Qiu

Department of Critical Care Medicine, Nanjing Zhongda Hospital, School of Medicine, Southeast University, Nanjing, China

Introduction: Bone marrow derived mesenchymal stem cells (BMSCs) are proved to have potential therapeutic effects in ARDS models. However, the mechanism of differentiation in MSCs is still unknown. There is growing evidence suggesting p130 or E2F4 plays an important role in regulating cellular differentiation. Our aim is to evaluate the role of p130/E2F4 in regulating the differentiation of mouse MSCs (mMSCs).

Methods: mMSCs with p130 or E2F4 overexpression were constructed using lentiviral vectors. The transfection efficiency of mMSCs was identified using fluorescence microscopy, and the percentage of GFP positive cells was determined by flow cytometry analysis. The mRNA and protein levels in MSC-p130 (overexpressing p130) and MSC-E2F4 (overexpressing E2F4) were detected by qRT-PCR and Western blotting assay, respectively. The effect of p130/E2F4 on the multipotential differentiation abilities of mMSCs were evaluated by adipogenesis, osteoblastic and chondrogenic differentiation medium respectively. The expression of the osteogenic gene OSX, adipogenisis gene PPAR-¦Ã and chondrogenic gene Sox9 measured by qRT-PCR were used to evaluate the differentiation of each group of mMSCs treated with differentiation induction media.

Results: The transduction efficiencies mediated by the lentiviral vectors were 80.3–84.4% and were maintained over 20 passages of mMSCs. The p130 or E2F4 mRNA expression was significantly higher in the MSC-p130 and MSC-E2F4 cells than in the MSC-NC cells. Similar results for p130 or E2F4 protein expression were also observed in Western blotting. Moreover, p130/E2F4 gene overexpression promoted the differentiation of mMSCs into osteoblasts, while inhibiting adipogenesis and chondrogenic differentiation of mMSCs.

Conclusions: The successful construction of stable and long-term mMSCs lines with overexpressing p130/E2F4 change the multipotential differentiation of mMSCs.

References

1. Liu AR et al.: J Cell Physiol 2013; 228: 1270.

Fig. 7 (abstract P11).

Legend 1: Long-term transgene expression efficiency in mMSCs

Fig. 8 (abstract P11).

Legend 2: The effect of p130 or E2F4 on multipotential differentiation of mMSCs

P12 The homing and protective effects of mesenchymal stem cells overexpressing CXCR7 in LPS-induced acute respiratory distress syndrome mice

J. X. Chen, L. Liu, L. Yang, X. W. Zhang, F. M. Guo, Y. Yang, H. B. Qiu

Department of Critical Care Medicine, Nanjing Zhongda Hospital, School of Medicine, Southeast University, Nanjing, China

Introduction: The low efficiency of homing to injured organ in mesenchymal stem cells (MSCs) has become the bottleneck for the treatment of acute respiratory distress syndrome (ARDS). In a previous in vitro study, the results showed that the CXCL12/CXCR7 axis promoted the migration of MSCs. It is hypothesized that MSCs overexpressing CXCR7 could further benefits LPS-induced ARDS mice.

Methods: mMSCs were transfected with CXCR7 by a lentiviral vector and were transplanted into mice following LPS-induced intratracheal lung injury. Histopathology with haematoxylin and eosin staining and lung injury scoring was performed to evaluate lung tissue injury. Homing of mMSCs were assayed by NIR815 fluorescence imaging and immunofluorescent staining.

Results: The administration of MSC, MSC-GFP and MSC-CXCR7 all improved both lung histopathology (Fig. 9) and lung injury score. Moreover, the MSC-CXCR7 showed lower lung/body weight and lung injury score especially at 24 h compared with MSC-GFP. NIR fluorescence imaging and immunofluorescent staining showed that overexpressing CXCR7 significantly increased the accumulation of mMSCs in the lung especially at 24 h compared to MSC (Fig. 10).

Conclusion: It is demonstrated that overexpressing CXCR7 in mMSCs could further promote their homing to injured lung and improve the pathological damage in ARDS mice.

Fig. 9 (abstract P12).

Legend: Histopathology of the effects of MSC-GFP or MSC-CXCR7 in LPS-induced lung injury

Fig. 10 (abstract P12).

Legend: Effect of overexpressing CXCR7 on the retention of mMSC in the lung after LPS change.

P13 Independent lung ventilation (ilv) in ICU-forgotten, neglected or useless?

M. Fister, R. Knafelj

Rihard Knafelj, Ljubljana, Slovenia

Introduction: ILV remains valuable rescue therapy for refractory hypoxemia in patients with predominantly single sided lung injury (contusion, aspiration, hemorrhage, bronchopleural fistula) especially in centers with no extracorporeal membrane oxygenation (ECMO). Special intubation technique, double lumen endotracehal tube and 2 ventilators are needed.

Methods: Single center prospective observational study. Characteristics of consecutive patients assigned to ILV from 2010–2016 are reported.

Results: Total 8 patients underwent ILV during observed period (0,9% of all ventilated patients). All patients had paO2/FiO2 < 200, all were ventilated in pressure modes (P-AC, BiLEVEL), ventilators were not synchronized - PEEP, RR, FiO2 and Vt were set individually for each ventilator following blood gas results. Recruitment manuvers and NO were used at physician discretion. In 5 patients oxygenation improved significantly within 2 hours of ILV. 2 patients were upgraded to V-V ECMO (Wegener’s, pneumonia), in 4 contraindication to ECMO was present. 3 patients died (lung carcinoma, emphysema, Wegener’s). All tube placements were confirmed radiologically, 1 intubation with double lumen tube was performed via bougie and proper placement confirmed bronhcoscopically. In all patients left sided double lumen tubes were used only. In one patient carinal decubitus was observed due to double lumen tube’s hook. Patients underwent ILV for median 4 days (range 2–10 days).

Conclusions: In severe single sided lung injury necessitating MV, ILV should be considered as it represents valuable adjunct in treating refractory hypoxemia. With lung separation both lungs can be ventilated according to protective ventilation strategy. Special care is needed to prevent tube dislocation, cuffs overinflation and trauma to larynx, trachea and bronchi. Since tubes’ lumens are narrower special attention to possible tube blockage must be paid. Daily reevaluation and conventional tube placement is warranted as soon as possible.

Fig. 11 (abstract P13).

See text for description

P14 Initiation of mechanical ventilation with the lower threshold of tidal volume in ARDS

M. A. Suzer¹, M. E. Kavlak², H. K. Atalan³, B. Gucyetmez⁴, N. Cakar⁴

¹Cankaya Hospital, Ankara, Turkey; ²Atasehir Memorial Hospital, Istanbul, Turkey; ³Acibadem Fulya Hospital, Istanbul, Turkey; ⁴Acibadem University School of Medicine, Istanbul, Turkey

Introduction: We hypothesized that in patients with ARDS, a pH level >7.20 can be achieved by using a tidal volume (TV) of 6 ml/kg. It was demonstrated that low TV (6–8 ml/kg) strategy was related with decreased volutrauma in patients with ARDS (1–3). According to ARDSNet (ARMA) trial, the initial mechanical ventilation (MV) settings should be 8 ml/kg TV, 7–9 L/min minute volume and a respiratory rate up to 35/min to provide pH > 7.20 and it is suggested to reduce the TV to 7 ml/kg and then gradually to 6 ml/kg (4).

Methods: The present study was designed in 2015 as prospective observesional. Patients with ARDS who were >18 years old were included in this study. The initial MV was set with TV:6 ml/kg, MV:7–9 L/min as recommended by ARDSNet (ARMA) trial. Blood gas samples were taken at the 1st and 2nd hours of MV. In patients with pH < 7.20, TV was increased by 1 ml/kg. Demographic data, 1st and 2nd hours blood gas values and ventilation parameters were recorded.

Results: One hundred and ten patients were included in this study. At the 1st hour, pH values in 86.4% (95) of all patients were above 7.20. In patients with pH < 7.20, while pH-corrected-APACHE II and lactate values were significantly higher; HCO3 and SBE were significantly lower than patients with pH > =7.20 (P = 0.006 and P < 0.001 for others). In patients with pH < 7.20, pH values were increased above 7.20 by increasing TV to 7 ml/kg.

Conclusions: In a significant percentage of patients with ARDS, targeted pH value can be achieved by using the lower threshold of recommended initial TV value. In patients with severe hyperlactatemic acidosis, higher initial TV values may be required.

References

1. Petrucci N et al. Cochrane Database Syst Rev 2004; 2:CD003844

2. Petrucci N et al. Cochrane Database Syst Rev 2013; 28:CD003844

3. Needham DM et al. Am J Crit Care Med 2015; 191:177–85

4. ARDSNet Study I, ARMA trial 1998.

P15 Lung protective ventilation: what is the effect of education on tidal volume and p0.1 values?

D. Weller, A. F. Grootendorst, A. Dijkstra, T. M. Kuijper, B. I. Cleffken

Maasstad Hospital, Rotterdam, Netherlands

Introduction: Lung protective ventilation may improve patient outcome on ICU’s and clear guidelines are available for ventilator settings. It is unknown whether nurses and doctors adhere to these guidelines. The purpose of this study was to determine the effect of predefined lung-protective ventilation regime in pressure support ventilation (PS) introduced via an educational programme on tidal volumes and airway occlusion pressure.

Methods: The study was a case control single-center study. For the intervention group, an education programme was integrated consisting of eight clinical lessons of ±30 min, bed side teaching, four times written instruction by email and twice a dissemination of an educational presentation to the entire ICU team, in the period from May to June 2016. This program set guidelines for mechanical ventilation, with the specific instructions to ventilate patients with a tidal volume of 6–8 ml/kg/ideal body weight (IBW) and airway occlusion pressure (P0.1) of 2.8 to 6.0 cmH2O. Prior to ventilation instruction a control group was obtained from the general ICU population.

Inclusion criteria were: mechanical ventilation (PS) within 7 days after presentation on the ICU, exclusion criteria were: any neurological diagnosis, re-intubation and delirium.

Results: In total 14 patients were included in the intervention group versus 17 in the control group (baseline, Table 3).

Thirteen parameters were monitored during the study, primary endpoints where tidal volume ml/kg/IBW and the P0.1. In the intervention group (Table 3) finale tidal volumes were 1.2 ml/kg/IBW lower (p = 0.04). The P0.1 shows no difference between the two groups (p =0.75).

Conclusions: Conclusion: Adherence to lung-protective ventilation is increased with education and improves ventilation practice on ICU’s. An educational program on specific guidelines of mechanical ventilation decreases tidal volumes to optimal values between 6–8- ml/kg/IBW. P0.1 values could not be optimized by this regimen. It seems likely that higher pressure support levels lead to lower inspiratory work of breathing.

Table 2 (abstract P15). See text for description

Table 3 (abstract P15). See text for description

P16 Intra-abdominal pressure adjusted positive end-expiriatory pressure – a pilot study

A. Regli¹, B. De Keulenaer¹, P. Van Heerden²

¹Fiona Stanley Hospital, Perth, Australia; ²Hadassah University Hospital, Jerusalem, Israel

Introduction: Intra-abdominal hypertension (IAH) is associated with increased morbidity and mortality. IAH reduces lung volumes and oxygenation and increases airway pressures. The optimal level of positive end-expiratory pressure (PEEP) to be applied in such patients remains unknown.

Animal data suggests that high PEEP adjusted for intra-abdominal pressure (IAP) may counteract IAH-induced lung volume and arterial oxygen level reductions.

In this pilot trial, our primary aim was to assess whether high and PEEP, adjusted for IAP, can be safely applied in patients with IAH requiring mechanical ventilation. Our secondary aim was to assess the effect of such PEEP levels on oxygenation.

Methods: Prior to enrolment, the patients next of kin were asked for written informed consent. Patients were excluded if deemed unsuitable or if predefined severe cardiovascular dysfunction or severe hypoxemia were present.

Ventilation was standardized. Following a recruitment manoeuvre, the following PEEP levels in cmH2O were randomly applied: 5 (baseline), PEEP = 50% of IAP, and PEEP = 100% of IAP. At each measurement step, we allowed 30 min for equilibration before measuring arterial blood gases and respiratory parameters. For the safety of our patients, predefined stopping criteria were applied.

Results: 15 patients were enrolled. The protocol was stopped in one patient (excluded from analysis) due to an unexpected hypertensive episode (drug error unrelated to protocol). Three patients did not tolerate PEEP = 100% IAP due to hypoxemia, hypotension, or ETT cuff leak in one patient each. PaO2/FiO2 ratios (maen and SD) were 234 (68), 271 (99), and 329 (107) at each PEEP level respectively. The difference was significant (p = 0.009) only between baseline and PEEP = 100% IAP.

Conclusions: PEEP = 100% was commonly not tolerated and only marginally improved oxygenation in ventilated patients with IAH. Such high PEEP pressures can therefore not be recommended for routine clinical practice.

P17 Withdrawn

P18 Withdrawn

P19 Neurally adjusted ventilatory assist versus pressure support in prolonged mechanical ventilation: a randomised feasibility study

D. Hadfield¹, P. A. Hopkins², B. Penhaligon², F. Reid¹, N. Hart³, G. F. Rafferty¹

¹King’s College London, London, UK; ²King’s College Hospital, London, UK; ³Guy’s and St Thomas’ Hospital, London, UK

Introduction: Neurally adjusted ventilatory assist (NAVA) is a complex intervention involving diaphragmatic monitoring and a proportional, neurally triggered ventilation mode [1]. Clinical effectiveness has not been demonstrated and feasibility data are required prior to larger definitive randomised controlled trials (RCTs). Although there is no agreement on minimum levels in clinical trials, low protocol compliance will impact on statistical power and interpretation of results in a definitive study [2]. The aim of this trial is to assess feasibility by measuring protocol compliance in the setting of a pragmatic RCT.

Methods: A feasibility RCT in 76 ventilated adult ICU patients, currently being conducted in a central London hospital. Patients at high risk of prolonged ventilation are randomly assigned to NAVA or Pressure Support (PSV) in the weaning phase. Feasibility end-points include protocol compliance (time in the NAVA mode), recruitment rate and acceptability. Secondary outcomes include ventilator free days and sedation load. The data from the first 48 patients (n = 25 and 23 in the NAVA and PSV groups respectively) are presented.

Results: At 14/11/2016 480 patients had been screened and 55 recruited with a mean recruitment rate of 1.8 patients per month. 71 patients were approached for inclusion with 16 patients (22.5%) excluded due to personal or professional consultee refusal. 68.2% of patients in the intervention group were compliant [CI 47.3–83.6] with a median adherence to the NAVA mode of 87.5% [38.9, 99]. Ventilator free days to day 28 were 18 [8, 24] days in the NAVA group vs. 19 [0, 23] days in the PSV group. There were no adverse events related to the intervention.

Conclusions: These data suggest an acceptable protocol compliance of > 65%. Further analysis of the reasons for non-compliance will be conducted; initial data suggest that a mix of technological, clinical and human factors underlie reduced adherence and protocol compliance. This is the first trial to compare NAVA to PSV in prolonged ventilation in the context of a pragmatic RCT. The results will provide guidance for the design of future trials.

References

1. Sinderby, C., et al., Neural control of mechanical ventilation in respiratory failure. Nature medicine, 1999. 5(12): p. 1433–6.

2. Chan, A.-W., et al., SPIRIT 2013 Statement: Defining Standard Protocol Items for Clinical Trials. Annals of Internal Medicine, 2013. 158(3): p. 200–207.

Table 4 (abstract P19). See text for description

P20 Effects of prone position and recruitment manoeuvres on gas exchange and regional respiratory mechanics

G. Grasselli¹, T. Mauri², M. Lazzeri³, E. Carlesso², B. Cambiaghi⁴, N. Eronia⁴, E. Maffezzini⁴, A. Bronco⁴, C. Abbruzzese¹, N. Rossi¹, G. Foti⁴, G. Bellani⁴, A. Pesenti²

¹Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico, Milan, Italy; ²University of Milan, Milan, Italy; ³University of Ferrara, Ferrara, Italy; ⁴University of Milan-Bicocca, Monza, Italy

Introduction: Prone position (PP) improves survival and delays the risk of ventilation-induced lung injury in severe ARDS [1]. PP may exploit its benefits in the presence of optimized alveolar recruitment. Thus, we performed prone positioning in combination with recruitment manoeuvres (RMs) to assess the effects on aeration and oxygenation.

Methods: Eleven intubated, volume-controlled, ARDS patients (1 severe, 8 moderate, 2 mild) were studied in supine and PP before and after RMs (40 cmH2O airway pressure for 10 breaths). In each phase electrical impedance tomography (EIT) was recorded for 15 minutes to assess global and regional (dependent and non-dependent) respiratory system compliance (Crs), tidal volume (VT) distribution, end-expiratory lung volume changes (ΔEELV). Gas exchange and respiratory mechanics were monitored at the end of each phase.

Results: PaO[sub]2[/sub]/FiO[sub]2[/sub] was 159 ± 71 mmHg. Clinical PEEP (13 ± 3 cmH[sub]2[/sub]O), VT (6.6 ± 1.2 ml/kg IBW), RR (25 ± 5 bpm) and FiO[sub]2[/sub] (0.6 ± 0.2) were left unchanged. PaO[sub]2[/sub]/FiO[sub]2[/sub] improved after RMs (p = 0.054) and by interaction between PP and RMs (p = 0.046). ΔEELV significantly increased after RM performed in supine (p = 0.032; 40 ± 53 ml vs. pre-RM), in PP (p = 0.002; 95 ± 78 ml vs. pre-RM) but by similar magnitude (p = 0.089) post-RM supine vs. prone. PP didn’t modify Crs, but dependent Crs significantly improved (p = 0.006) and non-dependent Crs (p = 0.009) decreased. VT distribution was more homogenous during PP and lung homogeneity index improved, albeit non-significantly. Neither regional Crs nor homogeneity were affected by RMs.

Conclusions: PP improves regional mechanics and homogeneity; RM performed after proning seems efficacious and might increase PP protective effects.

References

1. Valenza F. et al. Crit Care Med. 2005;33:361–7

Table 5 (abstract P20). See text for description

P21 Feasibility of the lateral-trendelenburg vs semi recumbent body position – the GRAVITY-VAP TRIAL

G. Li Bassi¹, M. Panigada², O. Ranzani¹, T. Kolobow³, A. Zanella ², M. Cressoni⁴, L. Berra⁵, V. Parrini⁶, H. Kandil⁷, G. Salati⁸, S. Livigni⁹, S. Livigni¹⁰, A. Amatu¹⁰, M. Girardis¹¹, M. Barbagallo¹², G. Moise¹³, G. Mercurio¹⁴, A. Costa ¹², A. Vezzani¹², S. Lindau¹⁵, J. Babel¹⁶, M. Cavana¹⁷, A. Torres¹

¹Hospital Clinic, Barcelona, Spain; ²Policlinico di Milano, Milan, Italy; ³National Institutes of Health, Bethesda, USA; ⁴Ospedale San Paolo, Milan, Italy; ⁵Massachusetts General Hospital, Boston, USA; ⁶Ospedale Nuovo del Mugello, Borgo San Lorenzo, Italy; ⁷Gruppo Ospedaliero San Donato, San Donato Milanese, Italy; ⁸Arcispedale S. Maria Nuova, Reggio Emilia, Italy; ⁹Ospedale San Giovanni Bosco, Torino, Italy; ¹⁰Policlinico San Matteo, Pavia, Italy; ¹¹Policlinico di Modena, Modena, Italy; ¹²Azienda Ospedaliero-Universitaria di Parma, Parma, Italy; ¹³Ospedale Citta di Sesto San Giovanni, Sesto San Giovanni, Italy; ¹⁴Policlinico Gemelli, Roma, Italy; ¹⁵University Hospital Frankfurt, Frankfurt, Germany; ¹⁶University Hospital Zagreb, Zagreb, Croatia; ¹⁷Ospedale Santa Chiara, Trento, Italy

Introduction: We recently completed a large randomized clinical trial in critically ill patients to test the LTP vs. SRP in the prevention of ventilator-associated pneumonia (VAP). Here we focus on the feasibility of the study interventions.

Methods: We randomized 194 and 201 critically ill patients into LTP and SRP group, respectively. Patients in LTP were placed in semi-lateral (60°) - Trendelenburg position and turned from one side to the other every 6 hours. LTP was encouraged specifically during the first days of mechanical ventilation, in compliance with the patient’s wish. In the SRP group, patients were kept with the head of the bed elevated > = 30°. We assessed interventions feasibility, and nursing feasibility/workload

Results: Patients in the LTP and SRP group were kept in the randomized position for 38% and 90% of the study time, respectively (p = 0.001). Median fraction of time in LTP reached up to 51.8% (interquartile range 20.7–79.2) during the first 2 days; yet, it progressively decreased in subsequent days and was inversely related to the level of sedation. Median lorazepam equivalents was 46.7 mg/day (interquartile range, 20.8–120.0) and 58.0 (28.0–124.2) in the SRP and LTP groups, respectively (p = 0.37); propofol dose was 1393 mg/day (627–2345) and 1661 (826–2700) (p = 0.13). Finally, morphine equivalents were 112 mg/day (30–217) in the SRP and 119 (31–234) in the LTP (p = 0.56). Nurses reported greater difficulties in positioning the patient in LTP and higher workload (p < 0.001) vs SRP. However, approximately 50% of the LTP patients were easily or very easily positioned. Finally, these nursing challenges slightly ameliorated through practice, as more patients were enrolled in each center.

Conclusions: The LTP was specifically applied during the first days of MV. Of note, higher level of sedation/analgesia was not needed to keep the patient in LTP. At times, the nurses encountered difficulties in executing the LTP.

P22 Lateral-trendelenburg vs. semi recumbent body position for the prevention of ventilator-associated pneumonia – the GRAVITY-VAP TRIAL

M. Panigada¹, G. Li Bassi², O. T. Ranzani², T. Kolobow³, A. Zanella¹, M. Cressoni⁴, L. Berra⁵, V. Parrini⁶, H. Kandil⁷, G. Salati⁸, S. Livigni⁹, A. Amatu¹⁰, M. Girardis¹¹, M. Barbagallo¹², G. Moise¹³, G. Mercurio¹⁴, A. Costa¹², A. Vezzani¹², S. Lindau¹⁵, J. Babel¹⁶, M. Cavana¹⁷, A. Torres²

¹Policlinico di Milano, Milan, Italy; ²Hospital Clinic, Barcelona, Spain; ³National Institutes of Health, Bethesda, USA; ⁴Ospedale San Paolo, Milan, Italy; ⁵Massachussets General Hospital, Boston, USA; ⁶Ospedale Nuovo del Mugello, Borgo San Lorenzo, Italy; ⁷Gruppo Ospedaliero San Donato, San Donato Milanese, Italy; ⁸Arcispedale S. Maria Nuova, Reggio Emilia, Italy; ⁹Ospedale San Giovanni Bosco, Torino, Italy; ¹⁰Policlinico San Matteo, Pavia, Italy; ¹¹Policlinico di Modena, Modena, Italy; ¹²Azienda Ospedaliero-Universitaria di Parma, Parma, Italy; ¹³Ospedale di Sesto San Giovanni, Sesto San Giovanni, Italy; ¹⁴Policlinico Gemelli, Rome, Italy; ¹⁵University Hospital Frankfurt, Frankfurt, Germany; ¹⁶University Hospital Center Zagreb, Zagreb, Croatia; ¹⁷Ospedale Santa Chiara, Trento, Italy

Introduction: The semi-recumbent position (SRP) could increase risk of gravity-driven pulmonary aspiration and ventilator-associated pneumonia (VAP) (1–2). We investigated whether the lateral-Trendelenburg position (LTP) vs. the SRP would prevent microbiologically confirmed VAP.

Methods: 194 patients were randomized into the LTP group and 201 in SRP, and analyzed in an intention to treat approach. Patients in LTP were turned from one side to the other every 6 hours. Whereas, in the SRP group, patients were kept with the head of the bed elevated > = 30°. Primary outcome was VAP incidence rate, based on quantitative bronchoalveolar lavage fluid culture. Secondary outcomes were duration of mechanical ventilation, intensive care unit (ICU) and hospital stays, and ICU/hospital/28-day mortality.

Results: The data safety monitoring board recommended stopping the study at the second interim analysis for low incidence of VAP in the control group, lack of benefits in any major secondary outcome and adverse events in the LTP group. Microbiologically confirmed VAP was 0.5% (1/194 patients) in patients positioned in LTP, and 4.0% (8/201 patients) in patients in SRP, risk ratio (RR) between groups 0.13, 95% confidence interval (CI) 0.02–1.03, p = 0.04. Microbiologically confirmed VAP per 1000 ventilator days was 7.19, 95%CI 3.60–14.37 and 0.88, 95%CI 0.12–6.25 in the LTP and SRP, respectively, RR 0.12, 95%CI 0.01–0.91, p = 0.02. Competing risk analysis, which accounted for the concomitant risk of death and discontinuation of MV on VAP, confirmed lower cumulative probability of VAP in the LTP (cause-specific hazard ratio 0.13, 95% CI 0.02–1.00, p = 0.05). No statistically significant between-group differences were observed in secondary outcomes.

Conclusions: Critically ill patients positioned in the LTP had a statistically significant reduction in the incidence of VAP, in comparison with the SRP. Yet, further clinical examinations are mandatory to corroborate our findings, specifically in populations at high risk of VAP.

References

1) Li Bassi G et al. Crit Care Med 2014; 42: e620–7

2) Panigada M et al. Crit Care Med 2003; 31: 729–37

P23 Effects of lateral position and open chest on partitioned respiratory mechanics during thoracic surgery

M. Umbrello, M. Taverna, P. Formenti, G. Mistraletti, F. Vetrone, A. Marino, G. Vergani, A. Baisi, D. Chiumello

Ospedale San Paolo, Università degli Studi di Milano, Milano, Italy

Introduction: Airway pressure (Paw) based respiratory mechanics is used to guide ventilation. As an altered chest wall may influence the interpretation of Paw, aim of the study was to investigate the behavior of lung and chest wall during different phases of thoracic surgery

Methods: 20 patients undergoing pulmonary resection were enrolled. Double-lung ventilation (DLV) setting: PEEP 8 cmH2O, tidal volume (TV) 8 ml/kg; in one-lung ventilation (OLV) TV was reduced to 5 ml/kg and respiratory rate increased accordingly. Esophageal pressure (Pes) was measured and used to assess partitioned respiratory mechanics (ie. elastance of the respiratory system - Ers, chest wall - Ecw and lung - El). Respiratory mechanics was assessed during DLV in supine and lateral decubitus, OLV in lateral decubitus during closed and open chest

Results: Supine Ers (cmH2O/L) in DLV was 15.8 ± 4.6; no change in lateral position and DLV (19.3 ± 4.4, p > 0.05); significant increase in lateral position with OLV, with no difference between closed and open chest (31.2 ± 9.6 vs. 27.2 ± 8.9, p > 0.05). Figure 12 shows changes of El and Ecw: Ecw increased in lateral position and decreased when chest was opened. El was unchanged in lateral position but significantly increased during OLV.

Conclusions: Paw-based monitoring of respiratory mechanics did not allow to differentiate the cause of an alteration. Partitioned respiratory mechanics gives a deeper insight into the behavior of lung and chest wall.

Fig. 12 (abstract P23).

See text for description

P24 Manual versus closed loop control of oxygenation parameters during invasive ventilation: effects on hyperoxemia

A. G. Garnero¹, D. N. Novotni², J. A. Arnal¹

¹Hôpital Sainte Musse, Toulon, France; ²Hamilton medical, Bonaduz, Switzerland

Introduction: Hyperoxemia occurs up to 50% of mechanical ventilation days in the ICU [1] and is associated with increased mortality as compared to patients ventilated in normoxemia [2]. INTELLiVENT-ASV is a full closed loop ventilation mode adjusting automatically oxygenation parameters according to SpO2 for passive and spontaneously breathing mechanically ventilated patients. This post-hoc analysis of a monocentric randomized controlled parallel group study compared hyperoxemia (PaO2 > 120 mm Hg or SpO2 > 96%) and hypoxemia (PaO2 < 60 mm Hg or SpO2 < 90%) between INTELLiVENT-ASV and conventional ventilation mode in mechanically ventilated ICU patients.

Methods: This randomized controlled trial was performed in the general ICU of Hôpital Sainte Musse, Toulon, France. Eligible participants: adult invasively ventilated for less than 24 h at the time of inclusion with an expected duration of mechanical ventilation of more than 48 h. Exclusion criteria: broncho-pleural fistula, ventilation drive disorder, moribund patients. Patients were allocated to INTELLiVENT-ASV group or to conventional ventilation group (volume assist control and pressure support) using blocked randomization. The post-hoc analysis was performed by the comparison of all arterial blood gases (ABG) performed during the study period: the number of ABG with hyperoxemia and hypoxemia, the median PaO2 and SpO2 for these ABG and FiO2 associated were compared.

Results: 60 patients were included, 30 patients in each group.

Conclusions: The closed loop control of oxygenation settings provided by INTELLiVENT-ASV decreases significantly the number of blood gas with hyperoxemia as compared to manual oxygenation setting without increasing the risk of hypoxemia.

References

[1] Suzuki et al. J Crit Care 28(5):647–654, 2013.

[2] Girardis et al. JAMA 316(15) :1583–1589, 2016.

Fig. 13 (abstract P24).

Legend 1: Demographic results

Fig. 14 (abstract P24).

Legend 2: Blood gas results

P25 Withdrawn

P26 Predicting the presence of spontaneous breathing in mechanically ventilated patients

M. Urner, E. Fan, M. Dres, S. Vorona, L. Brochard, N. D. Ferguson, E. C. Goligher

University of Toronto, Toronto, Canada

Introduction: Databases from large trials are more frequently used to test new hypotheses about the impact of mechanical ventilation. In absence of direct measurement of respiratory effort, it is challenging to know if patients are breathing spontaneously. We sought to establish if the presence of spontaneous breathing can be predicted from readily available clinical variables.

Methods: Logistic regression, extreme gradient boosting, and random forest models were built from clinical variables to predict a diaphragm thickening fraction > 0.2 (mean value in healthy subjects breathing at rest [1]) as reference standard for inspiratory effort. Data from 195 patients was randomly split into a training (75%) and a test dataset (25%). The models were derived from the training dataset by cross-validation and validated on the test dataset. All models were implemented in R using the caret package.

Results: The AUROC for the logistic regression, extreme gradient boosting, and random forest models was 0.65 (0.59 to 0.70), 0.69 (0.63 to 0.72) and 0.68 (0.61 to 0.78). Diaphragm thickening fraction was most closely related to tidal volume, SAPS II score, and minute ventilation (Fig. 15).

Conclusions: We present an approach to detect spontaneous breathing during mechanical ventilation. Our models may be useful to identify spontaneously breathing patient subsets in cohort studies, but predictive accuracy is insufficient to replace direct monitoring of individual patients.

References

[1] Harper CJ et al., J Orthop Sports Phys Ther. 2013; 43(12):927–31

Fig. 15 (abstract P26).

Legend 1: Relative variable importance according to random forest analysis

P27 Mechanical ventilation training using flipped-classroom with a one-hour face-to-face tutorial

C. Leung, G. Joynt, W. Wong, A. Lee, C. Gomersall

Chinese University of Hong Kong, Sha Tin NT, Hong Kong

Introduction: We studied the effectiveness of a flipped classroom approach using e-learning and a one-hour face-to-face (FTF) practical training in teaching medical students to manage basic mechanical ventilation (MV). Teaching MV is challenging due to teacher time limitations.

Methods: We performed a prospective cohort evaluation of the level of MV competence of final year medical students who participated in this optional training. The courseware consisted of an e-chapter (to introduce basic MV settings and physiology); online interactive cases and practice tests with a MV simulator (to facilitate scenario-based practice); a one-hour FTF tutorial (1 teacher to 6 students).

Two critical care educators, informed only of the syllabus outline, created the knowledge and skills assessments at a level expected of interns. Knowledge was assessed by pre- and post-course multiple choice questions (MCQs) testing 10 domains of MV. For each student, two MCQs of equal difficulty per domain were randomized to pre- and post-course tests. The two educators conducted 10-minute post-course skills testing of students’ practical competence in setting the correct MV mode, FiO2, tidal volume, respiratory rate (RR), PEEP in a case scenario; performing appropriate action in response to desaturation, pressure or RR alarms; reassessing after setting or adjusting MV. Students’ feedback on the course’s usefulness was surveyed.

Results: 179(81%) students consented to participate. Mean MCQ score pre-course was 29% and post-course was 53% - with a mean increment (95% CI) of 32.6(28.7–36.6) in students who completed the course including FTF, 7.0(0.8–13.1) when courseware was used without FTF, and 7.7(−1.6–17.1) in those who did not attempt the course (p < 0.001). MCQ score improved in every domain (McNemar’s test, P-values varied from <0.001 to <0.03) and for all domains combined (McNemar’s test P <0.001). 23 randomly selected students participated in the skills test - median score was 8/10(IQR 6.75–8.5). Students reported the course useful for improving their MV competence. Median score of individual course components was 4–5(