Thank you very much. I thank the organizing committee for their invitation to talk about the Winsafe study. So I was part of the Winsafe executive committee and it was funded by ERS and ESICM. So we should think about extubation and weaning as soon as we intubate a patient. It was previously showed that up to 40% of the duration of total mechanical ventilation is dedicated to weaning and the duration of mechanical ventilation is associated with worse outcome and weaning failure is associated with higher mortality. So unfortunately there is no unanimous definition of weaning. In 2007 experts from Europe and America published a consensus conference trying to address this issue and proposed some definitions of weaning groups based on the spontaneous breathing trial, on the number of spontaneous breathing trial, and on the duration between the first SBT and the weaning success. Unfortunately further research and investigation on the patient's cohorts showed that this classification failed to group almost half of the patients. Trying to address the limitations of these definitions, the Riva Network in France and with some other centers in Europe, in Spain and Switzerland, proposed another definition, the WIND definition, and this was able to classify more patients. The most recent data we have about epidemiology of weaning comes from Karen Burns' group in Toronto and in an international survey followed by an international observational study on intubated patients, they showed that there was huge heterogeneity in the sedation policy, in the practice of spontaneous breathing trial, and in the practice of tracheostomy. So trying to address this, we proposed an international observational study on the weaning of the intubated patients and we had collected data daily in patients who received mechanical ventilation for more than two days. More than 1,000 investigators participated to this observational study and the executive committee was formed by Laurent Brochard, Leo Ernst, Giacomo Bellani, John Laffey, and myself. So each participating ICU screened patients for four weeks between October 2017 and May 2018 and all patients who received invasive mechanical ventilations were screened and patients who received more than two calendar days of mechanical ventilations were included. We defined weaning eligibility criteria based on FIO2, the PEEP, the fact that patients were not receiving vasopressor or low doses of vasopressors and were not receiving paralyzing agents anymore. We defined separation attempt in patients receiving ventilations through a tube as an SBT or an extubation with or without SBT. And we defined also separation attempt in patients with a tracheostomy when this patient had spontaneous breathing without assistance. Weaning success based on the weaned definition was not receiving any assistance for seven days. We defined delayed separation attempt when a patient met criteria for weaning eligibility but did not have a separation attempt in the following day. So if a patient was receiving mechanical ventilation and we defined the first separation attempt at day zero, any patients who were transferred or died before any separation attempt was classified in the no weaning group. If there was successful weaning within 24 hours following the first separation attempt, the patient was in the short wean, then intermediate wean between day two and the end of the first week and successful weaning occurring seven days after the first separation attempt classified patients in the prolonged wean group. If death or transfer happens after the first separation attempt, the patients were classified in the failed wean group. We defined a sedation category based on the previous classifications such as RAS, SAS, or RANSI, but to simplify because every ICU had different types of classification, we grouped these different levels of sedation into three simplified groups. Patients could be awake with moderate sedation or deep sedation. The goals and the questions of the study were to describe the weaning process and outcomes, the trajectories of the patients, and the main question was how many patients were weaned within the first 90 days. We also wanted to identify factors associated with delayed separation attempt and weaning failure and to address if delay in initiating weaning impacts outcome and what would be the factors explaining such a delay. Just to clarify, our multivariable analysis, because it was a multilevel, there were some characteristics based on individual patients, but these patients belonged to different ICUs, so we also clustered these in different ICUs, and these different ICUs came from different levels and we also took this into account in our analysis. So going to the results, more than 10,000 patients were screened all over the world, mainly from Europe, but all the continents were represented. Among these 10,000 intubated patients, 4,000 were excluded mainly because they received less than two days of mechanical ventilation, and 5,800 were included. Looking at the characteristics of the included patients, 38% were female, the mean age was 61, the mean BMI was 27, 23% had chronic lung disease, 22% had diabetes, 14% were smokers, and 22% had criteria for frailty. For the cause of admission in the ICU, it was mainly medical, 22% were surgical, and 9% were trauma. And the three main causes for intubation were hypoxemic or hypercapnic respiratory failure, infection, and non-traumatic brain issues. Among these patients, 4,500 received at least one separation attempt, and 65% of the patients included were weaned. Among the patients who were not weaned, it's mainly because they died, so they had at least one separation attempt but died, and 153 were transferred out of the participating ICU, likely for weaning or expertise such as ECMO in another ICU. This graph shows the daily percentage of the patients, depending on their state. They could be weaned, still receiving invasive mechanical ventilation, being transferred in gray, or deceased. And you can see that in this global population, 65% of the patients are weaned. If we restrict this analysis to the patients who had at least one separation attempt, 84% of the patients are weaned. Then looking at the different weaning groups based on the delay between the first separation attempt and the weaning success, so most of the patients were in the short weaned group, so were successful weaning within 24 hours following the first separation attempt, and then 10% were in the intermediate group and 10% in the prolonged group. As you can see here, they had very different characteristics. The patients in the short group were more admitted for planned surgeries, had less neurological events, were less frail. Of course, they had less reintubation and received less tracheostomy. So, comparing the outcomes of these groups, again, the short weaned group had shorter delay between the first separation attempt and the weaning success, as I already said. But this first separation attempt also occurred earlier in the course of mechanical ventilation, in the median of five days after intubation, as compared to the prolonged group, where it was seven days in median. The first separation attempt to success was shorter. The total duration of invasive mechanical ventilation was five days in median in the short group, 10 days in the intermediate, and 20 in the prolonged group. And of course, the total duration of ICU hospital was shorter in the short group, and hospital mortality was 10%, as compared to 18% in the patient in the prolonged group. This is the multistate, sorry, multilevel multivariate analysis, looking at the factors associated with delayed weaning. So even if we adjusted on many factors, the ones that were actually significantly associated with delayed weaning were frailty, trauma, non-trauma neuro issues, higher SOFA score, receiving paralyzing agents before the first separation attempt, and receiving moderate or deep sedation at the moment of meeting weaning LGBT criteria. This is the same type of analysis, but restricted to the weaning success or failing. So all these factors were also adjusted in the multivariable analysis, and factors associated with failed weaning were increased age, immunocompromission, frailty, cardiac arrest, neuro issues, decisions of limitation previous to ICU admission, higher SOFA score, again, deep sedation, and delayed separation attempt, meaning a longer delay between meeting weaning LGBT criteria and the first separation attempt were associated with increased failure. And all the other factors were severity at the moment of the first separation attempt, so higher respiratory rate, higher PEEP, higher dynamic driving pressure, and lower PF ratio. So to summarize our findings, in this cohort of severe patients receiving at least two days of mechanical ventilations, we have weaning success in 65%. There are some practice variations, but the delay between the meeting weaning LGBT criteria and the first separation attempt seems to impact the weaning outcome. And there is a high importance of early management, sedation management, and the severity at the first separation attempt. With this, I thank you very much.

international observational study

weaning process

intubated patients

mechanical ventilation

weaning outcome