American Journal of Respiratory and Critical Care Medicine

Rationale: The weaning process concerns all patients receiving mechanical ventilation. A previous classification into simple, prolonged, and difficult weaning ignored weaning failure and presupposed the use of spontaneous breathing trials.

Objectives: To describe the weaning process, defined as starting with any attempt at separation from mechanical ventilation and its prognosis, according to a new operational classification working for all patients under ventilation.

Methods: This was a multinational prospective multicenter observational study over 3 months of all patients receiving mechanical ventilation in 36 intensive care units, with daily collection of ventilation and weaning modalities. Pragmatic definitions of separation attempt and weaning success allowed us to allocate patients in four groups.

Measurements and Main Results: A total of 2,729 patients were enrolled. Although half of them could not be classified using the previous definition, 99% entered the groups on the basis of our new definition as follows: 24% never started a weaning process, 57% had a weaning process of less than 24 hours (group 1), 10% had a difficult weaning of more than 1 day and less than 1 week (group 2), and 9% had a prolonged weaning duration of 1 week or more (group 3). Duration of ventilation, intensive care unit stay, and mortality (6, 17, and 29% for the three groups, respectively) all significantly increased from one group to the next. The unadjusted risk of dying was 19% after the first separation attempt and increased to 37% after 10 days.

Conclusions: A new classification allows us to categorize all weaning situations. Every additional day without a weaning success after the first separation attempt increases the risk of dying.

Scientific Knowledge on the Subject

Despite the importance of the weaning period during mechanical ventilation, our knowledge of the separation process from mechanical ventilation as it is currently performed and the complications associated with its prolongation are poorly described.

What This Study Adds to the Field

The WIND (Weaning according to a New Definition) classification is based on the duration of ventilation after the first separation attempt. This classification shows that prolongation of weaning has a direct and immediate impact on morbidity and mortality, by contrast with previous reports using the International Consensus Conference classification. The first separation attempt is a major milestone, and each additional day without a weaning success after this first attempt is associated with an increased crude mortality.

Mechanical ventilation is used daily as a life-saving technique in intensive care units (ICUs). Its application, however, is associated with serious complications and costs, often directly linked to the duration of ventilation (1, 2). Likewise, extubation failure contributes to impaired outcomes: longer length of ventilation, longer length of stay in the ICU and in the hospital, and association with higher mortality (3, 4). Therefore, weaning from mechanical ventilation represents a crucial step for every patient (57), and aiming to shorten duration of ventilation is fundamental (810). Despite the importance of this period, the weaning process is not rigorously defined, with wide variations in definitions and practices, thus making epidemiological studies difficult to conduct and interpret. In addition, the specific impact of weaning difficulties is still poorly understood. This is an important problem, because general recommendations regarding the entire weaning process may encompass completely different causes and consequences of its prolongation and therefore may be totally inappropriate for individual patients. A relatively simple classification may help to differentiate the various situations; in 2007, an International Consensus Conference (ICC) on weaning from mechanical ventilation (5) proposed a classification in three different groups (ICC groups) on the basis of number, timing, and results of spontaneous breathing trials (SBTs, defined as a T-piece trial or a low-level pressure support ≤ 8 cm H2O) as well as extubation outcomes. This approach, though useful to illustrate the different problems associated with these three groups, was not initially proposed as an operational tool for prognostication. In this regard, its major shortcoming was that it only considered patients who were ultimately weaned. In addition, the use of well-identified SBTs is not a universal practice, and some patients are weaned without an SBT. These patients cannot be classified in any of these three ICC groups. To our knowledge, six studies applied this classification using different adaptations of the initial description (1116), but mostly on a small number of patients or in single center (1216) or trying to retrospectively classify patients from an ICU database (11, 12). The different designs and the limitations in the definitions may have explained differences among the studies and some surprising findings. For instance, previous studies have been unable to find a different prognosis associated with simple and difficult weaning (11, 12). Prospective data concerning the weaning process as it is currently performed and its associated mortality in a large population of ICU patients are therefore lacking.

We thus designed a multicenter multinational prospective observational study, the WIND (Weaning according to a New Definition) study, in a large population of patients receiving invasive mechanical ventilation to obtain a comprehensive view of weaning trajectories, including weaning prolongation and its associated prognosis. We defined the start of weaning as the first attempt at separating the patient from the ventilator, whatever its modality, and did not try to describe the possible gradual reduction of support that may have occurred before. We propose a modification of the ICC classification (5) that could better fit daily ICU practice and could be operational for every patient under invasive mechanical ventilation in various environments. Such a classification may be useful for epidemiological studies, quality-improvement projects and benchmarking, health economics calculations, conducting trials, and guidelines and recommendations. Some of the results of this study have been previously reported in the form of abstracts (17, 18).

A full version of the methods is in the online supplement.

This prospective multicenter observational study was endorsed by the Réseau Européen de Recherche en Ventilation Artificielle (REVA) Network and included patients admitted in 36 ICUs in France (N = 29), Spain (N = 6), and Switzerland (N = 1) over a 12-week period (April 2013 to June 2013). All patients newly admitted during this period and requiring intubation for ventilation were enrolled at the date of intubation and followed until ICU discharge or Day 60, whichever came first.

Ethical and Legal Aspects

This study was approved by the French Intensive Care Society (Société de Reanimation de Langue Française) ethics committee with a waiver of consent, and as of April 9, 2013 by the Commission Nationale de l’Informatique et des Libertés, the French independent administrative authority that operates in accordance with the data-protection legislation. For the centers in Switzerland and Spain, local ethics committee approval was obtained.

Data Collection

Investigators first had to answer a questionnaire about their centers, including the use of sedation and weaning protocols. Participating investigators collected daily ventilation parameters and weaning strategies. Investigators recorded the weaning modalities, including performance of SBTs, and, if there were SBTs, the technique used. Data collection was continued until ICU discharge or Day 60, whichever occurred first.

Weaning Classification according to the ICC

We first applied the ICC classification (5). Three weaning groups were defined on the basis of number, timing, and results of SBTs as well as extubation outcomes: simple weaning (ICC group 1) was defined by a successful extubation after the first SBT, difficult weaning (ICC group 2) was defined by a successful extubation after two to three SBTs and taking less than 7 days, and prolonged weaning (ICC group 3) was defined by a successful extubation after more than three SBTs or by more than 7 days.

The New WIND Definition and Classification

After observing the variety of practices and to cover the range of clinical situations encountered, we proposed an evolution of the ICC classification (see online supplement for methods). This “WIND classification” defined the start of weaning as any kind of separation attempt (without accounting for the previous reduction in the ventilation support), computed the duration of this process and its prognosis, and proposed the following definitions:

For intubated patients

  • Separation attempt from mechanical ventilation: an SBT with or without extubation, or an extubation directly performed without identified SBT (whatever the type: planned or unplanned extubation).

  • Successful weaning or separation: extubation without death or reintubation within the next 7 days whether postextubation noninvasive ventilation (NIV) was used or not, or ICU discharge without invasive mechanical ventilation within 7 days, whichever comes first. The date of the successful weaning was counted retrospectively to the actual day of extubation after the patient has completed 7 days without reintubation (or was discharged earlier without reintubation).

For tracheostomized patients

  • Separation attempt from mechanical ventilation: 24 hours or more with spontaneous ventilation through tracheostomy without any mechanical ventilation.

  • Successful weaning or separation: spontaneous ventilation through tracheostomy without any mechanical ventilation during 7 consecutive days or discharged with spontaneous breathing, whichever comes first.

The whole population was then classified into four mutually exclusive groups, on the basis of the duration of the weaning process (i.e., delay between the first separation attempt and weaning termination):

  • Group “no weaning,” comprising patients who never experienced any separation attempt.

  • Group 1 (short weaning): the first attempt resulted in a termination of the weaning process within 1 day (successful separation or early death).

  • Group 2 (difficult weaning): the weaning was completed after more than 1 day but in less than 1 week after the first separation attempt (successful separation or death).

  • Group 3 (prolonged weaning): weaning was still not terminated 7 days after the first separation attempt (by successful separation or death).

This last group was further split in Group 3a (prolonged weaning leading to a successful weaning after 7 days or more after the first attempt) and Group 3b (prolonged weaning without success).

Outcome according to the WIND Definition

Ventilator-free days were calculated as the number of days without invasive ventilation to Day 28. Nonsurvivors were considered to have a ventilator-free days value of 0. We assessed the crude mortality rate of each group. We also calculated the daily risk of dying for patients still not having a successful weaning each day after the first separation attempt.

Use of Sedation and Weaning Protocols and Use of SBTs

We assessed the association between the presence of protocols for sedation or for weaning and the likelihood of having a short phase of separation or short weaning. We also compared patients whose first separation attempt was an SBT to patients who had another type of first separation attempt.

Statistical Analysis

Descriptive statistics included frequency (percentages) for categorical variables and mean and SD or median and interquartile range (IQR) for continuous variables. Comparisons of proportions were made using chi-square or Fisher exact tests, and continuous variables were compared using Student t test or Wilcoxon rank sum test when two groups were compared and analysis of variance or Kruskal-Wallis tests when more than two groups were compared. We performed a multivariable analysis of factors associated with simple weaning by means of a logistic regression, forcing both sedation and weaning protocols in the final model. Last, we performed a multivariable logistic regression to assess factors associated with the use of an SBT before a planned extubation.

ICU and Patient Characteristics

A total of 36 units participated, with a median (IQR) number of beds of 12 (10–17) and 740 (498–901) admissions per year, including a median of 324 (222–462) intubated patients. Most of these centers used a sedation protocol (57%), and 43% had a weaning protocol. Over the 3 studied months, a total of 2,729 patients met the inclusion criteria, which represented 71 (43–106) patients per unit. Among them, 20 patients were excluded because the main reason for mechanical ventilation was a tracheostomy for orotracheal obstruction and thus they were not concerned by the weaning process. The final cohort therefore comprised 2,709 patients, whose main characteristics are detailed in Table 1. A majority (66.9%) were discharged alive from the ICU and weaned from mechanical ventilation; 125 (4.6%) were still alive but not fully weaned at discharge or at Day 60, and the ICU mortality of the total cohort was 28.5%.

Table 1. WIND Study Population Characteristics

Characteristic 
Baseline characteristics 
 Age, yr61 ± 16
 Sex ratio, M/F1.90
 SAPS II at admission, points52 ± 21
 SOFA at admission, points7.9 ± 4.2
 SOFA at Day 3 (n = 2,225), points5.8 ± 4.6
 Admission type 
  Medical2,042 (75.4)
  Planned/unplanned surgery283 (10.5)/384 (14.2)
   Abdominal203 (30.4)
   Cardiac157 (23.5)
   Vascular57 (8.5)
   Neurological54 (8.1)
   Urologic49 (7.3)
   Trauma49 (7.3)
   Thoracic33 (4.9)
   Head and neck24 (3.6)
   Other41 (6.1)
Outcome 
 Total No. of days of invasive MV4 (2–9)
 Ventilator-free days*21 (0–26)
 Delay from intubation to first SA, d3 (2–7)
 Length of stay in the ICU, d6 (3–13)
 Length of stay in the ICU in survivors, d7 (4–14)
 Status at ICU discharge or Day 60 
  Dead771 (28.5)
  Alive and weaned 
   Spontaneous breathing1,700 (62.8)
   NIV65 (2.4)
   Tracheostomy (permanent spontaneous breathing)48 (1.8)
  Alive and not weaned 
   Intubation88 (3.3)
   Tracheostomy with MV37 (1.4)

Definition of abbreviations: ICU = intensive care unit; IQR = interquartile range; MV = mechanical ventilation; NIV = noninvasive ventilation; SA = separation attempt; SAPS II = Simplified Acute Physiology Score II; SOFA = Sequential Organ Failure Assessment; WIND = Weaning according to a New Definition.

Data are presented as n (%), mean ± SD, or median (IQR). N = 2,709 included patients.

*Ventilator-free days were defined as 28 − the total number of days with invasive MV. Nonsurvivors were considered as having 0 ventilator-free days.

Classifications and Characteristics of the Groups

Applying the ICC classification to our cohort (n = 2,709), only 1,379 (51.0%) patients could be classified: 1,016 (37.5%) did not have an SBT, 735 (27.1%) were never extubated, 337 (12.4%) had an extubation without an SBT, 1,070 (39.5%) never met successful weaning criteria of the ICC (comprising 170 patients still under NIV 48 h after extubation); 786 (29.0%) had no SBT or weaning success. Among the classified patients, 962 (69.8%) belonged to ICC group 1, 308 (22.3%) to ICC group 2, and 109 (7.9%) to ICC group 3.

The new WIND classification allowed description of the distribution and mortality of the whole cohort as follows (Figures 1 and 2):

  • Group no weaning comprised 658 (24.3%) patients who never had any separation attempt.

  • Group 1 comprised 1,543 patients (57.0%) whose weaning was terminated within 24 hours after the first separation attempt. Most had a successful weaning (94.5%), but 82 patients (5.3%) died and 2 patients (0.2%) were transferred before weaning success in the 24 hours after the first separation attempt.

  • Group 2 comprised 273 patients (10.1%) whose weaning was terminated between 2 and 6 days after the first separation attempt. Among these patients, 83.5% had a successful weaning, 15.4% died, and 1.1% were transferred before weaning success.

  • Group 3 comprised 235 (8.7%) patients who did not have a separation attempt leading to weaning success 1 week after their first separation attempt. This group had a death rate of 29.8%. Of this group, 145 (61.7%) patients finally had a successful weaning (group 3a), and 90 (38.3%) patients were never weaned (group 3b) because they died (74.4%), were transferred under mechanical ventilation (13.3%), or were still under mechanical ventilation at Day 60 (13.3%). Patients’ baseline characteristics and outcomes in the groups are described in Table 2 and Table E1 in the online supplement, and differences between the ICC and the WIND classifications are described in Table E2. Patients from group 1 significantly differed from patients of groups 2 and 3 by being younger and with lower admission Sequential Organ Failure Assessment (SOFA) score (P < 0.001 for each comparison, chi-square tests). No evidence of any differences in baseline characteristic between patients of groups 2 and 3 was observed. Outcomes were significantly different across the three groups: length of ICU stay, ventilator-free days (Figure 3 and Figure E1), and mortality were significantly different between all groups (P < 0.001 for each comparison, chi-square tests) (Table 2).

Table 2. Characteristics of the Population according to the Weaning Group (WIND Classification)

 Group 1 (N = 1,543)Group 2 (n = 273)Group 3 (n = 235)P Value*Group NW (n = 658)
Age, yr59 ± 1765 ± 1565 ± 13<0.00164 ± 15
Sex ratio, M/F1.92.11.90.911.9
SAPS II at admission, points45 ± 1751 ± 1853 ± 18<0.00168 ± 23
SOFA at admission, points6.6 ± 3.68.0 ± 3.38.1 ± 3.7<0.00111 ± 4.3
SOFA at Day 3, points4.3 ± 3.86.7 ± 4.06.8 ± 3.7<0.00110.4 ± 5.1
Admission type   <0.001 
 Medical1,084 (70.3)220 (80.6)185 (78.7)553 (84)
 Planned surgery234 (15.2)10 (3.7)18 (7.7)21 (3.2)
 Emergent surgery225 (14.6)43 (15.8)32 (13.6)84 (13)
Type of surgery    
  Cardiac129 (8.4)8 (2.9)10 (4.3)10 (1.5)
  Abdominal125 (8.1)24 (8.8)14 (6.0)40 (6)
  Vascular37 (2.4)2 (0.7)9 (3.8)9 (1.4)
  Neurological23 (1.5)7 (2.6)5 (2.1)19 (2.9)
  Urologic40 (2.6)1 (0.4)1 (0.4)7 (1)
  Thoracic17 (1.1)3 (1.1)6 (2.6)7 (1)
  Trauma37 (2.4)6 (2.2)2 (0.9)4 (0.6)
  Head and neck17 (1.1)2 (0.7)0 (0.0)5 (0.8)
  Other34 (2.2)0 (0.0)3 (1.3)4 (0.6)
Total No. of days of invasive MV3 (2–6)9 (6–13)19 (15–31)<0.0013 (2–7)
Ventilator-free days25 (22–26)18 (9–21)0 (0–12)<0.0010 (0–0)
Delay from intubation to first SA, d3 (2–5)6 (3–10)6 (3–10)<0.001
Length of stay in the ICU, d5 (3–9)14 (8–25)31 (20–46)<0.0013 (2–8)
Length of stay in the ICU in survivors, d5 (3–9)15 (9–25)37 (23–52)<0.0018 (2–31)
Status at ICU discharge (or Day 60)   <0.001 
 Dead90 (5.8)45 (16.5)70 (29.8)566 (86)
 Alive and weaned1,449 (93.9)224 (82.0)140 (59.6)0 (0)
  Spontaneous breathing1,394 (90.3)210 (76.9)96 (40.9)0 (0)
  NIV44 (2.9)11 (4.0)10 (4.3)0 (0)
  Tracheostomy11 (0.7)3 (1.1)34 (14.5)0 (0)
 Alive and under invasive mechanical ventilation4 (0.3)4 (1.5)25 (10.6)92 (14)
  Tube2 (0.1)2 (0.7)5 (2.1)79 (12.0)
  Tracheostomy2 (0.1)2 (0.7)20 (8.5)13 (2.0)
Decision of withholding or withdrawing invasive MV     
 Total130 (8.4)46 (16.9)63 (26.8)<0.001117 (17.8)
 Among deceased patients81 (90.0)30 (66.7)42 (60.0)<0.001113 (20.0)
 Among survivors49 (3.4)16 (7.0)21 (12.7)<0.0014 (4.4)

Definition of abbreviations: ICU = intensive care unit; IQR = interquartile range; MV = mechanical ventilation; NIV = noninvasive ventilation; NW = no weaning; SA = separation attempt; SAPS II = Simplified Acute Physiology Score II; SOFA = Sequential Organ Failure Assessment; WIND = Weaning according to a New Definition.

Data are presented as n (%), mean ± SD, or median (IQR).

*Overall comparison between group 1, group 2, and group 3. Chi-square test, analysis of variance, or Kruskal-Wallis tests were used.

Ventilator-free days were defined by 28 minus the total number of days with invasive MV. Nonsurvivors were considered as having 0 ventilator-free days.

No statistical test to compare status at ICU discharge between groups was performed, as weaning entered in the group’s definition.

Factors Associated with a Short Weaning Duration

A total of 1,812 patients had at least one separation attempt and did not have any withholding/withdrawing decisions, comprising 1,413 who had a short weaning (group 1) and 399 patient who had a longer weaning (groups 2 and 3) (Table 3). Among these patients, 995 (54.9%) were managed in a center using a sedation protocol, and 653 (36.0%) were managed in a center using a weaning protocol. In a multivariable analysis, younger age, lower SOFA score at admission, shorter duration of mechanical ventilation before the first separation attempt, and admission for planned surgery were associated with a short weaning, whereas being in a unit using a protocol for sedation or for weaning did not influence this proportion (Table 3).

Table 3. Factors Associated with a Short Weaning in Patients with No Decision of Withholding/Withdrawal, Bivariate and Multivariable Analysis

 Patients with a Short Weaning (n = 1,413)Patients with a Longer Weaning (n = 399)Bivariate Analysis P Value*Multivariable Analysis
OR, 95% CIP Value
Age, yr58 ± 1765 ± 14<0.0010.98, 0.97–0.99<0.001
SAPS II at admission, points44 ± 1650 ± 18<0.001  
SOFA at admission, points6.4 ± 3.57.9 ± 3.5<0.0010.94, 0.91–0.970.003
Admission   2.27, 1.12–4.780.024
 Medical968 (68.5)312 (78.2)<0.001
 Planned surgery231 (16.4)25 (6.3)<0.001
 Unplanned surgery214 (15.2)62 (15.5)0.91
Reintubation19 (1.3)181 (45.4)<0.001  
Sedation protocol735 (52.0)260 (65.2)<0.0010.72, 0.39–1.320.576
Weaning protocol513 (36.3)140 (35.0)0.760.84, 0.65–1.090.198
Total No. of days of invasive MV3 (2–5)12 (8–19)<0.001  
Ventilator-free days25 (23–26)15 (2–20)<0.001
Delay from intubation to first SA, d3 (1–5)6 (3–10)<0.0010.92, 0.91–0.94<0.001
Length of stay in the ICU, d5 (3–9)21 (12–38)<0.001
Length of stay in the ICU in survivors, d5 (3–9)22 (12–40)<0.001
Death9 (0.6)43 (10.8)<0.001  

Definition of abbreviations: CI = confidence interval; ICU = intensive care unit; IQR = interquartile range; MV = mechanical ventilation; OR = odds ratio; SA = separation attempt; SAPS II = Simplified Acute Physiology Score II; SOFA = Sequential Organ Failure Assessment.

Data are presented as n (%), mean ± SD, or median (IQR).

*Comparison between patients with a short weaning versus patients with a longer weaning (Student test, Wilcoxon test, or chi-square test).

Ventilator-free days are defined by 28 minus the total number of days with invasive MV. Nonsurvivors were considered as having 0 ventilator-free days.

Key Terms
International Consensus Conference

  • International Consensus Conference weaning groups:

    • Group 1 (simple weaning): successful extubation after the first spontaneous breathing trial (SBT).

    • Group 2 (difficult weaning): successful extubation after two to three SBTs and taking less than 7 days.

    • Group 3 (prolonged weaning): successful extubation after more than three SBTs or taking more than 7 days.

  • International Consensus Conference weaning success and failure definitions:

    • Weaning success is defined as extubation not requiring reinstitution of ventilatory support in the 48 hours after extubation.

    • Weaning failure is defined as one of the following: (1) failed SBT; (2) reintubation and/or resumption of ventilator support in the 48 hours after extubation; or (3) death within 48 hours after extubation.

WIND Study

  • Weaning: the start of weaning is the first attempt at separating the patient from the ventilator, whatever its modality. Weaning is terminated when the patient is successfully separated from the ventilator or if the patient dies (see below). Weaning does not include the possible gradual reduction of support that may have occurred earlier.

  • SBT: test of spontaneous ventilation without or with “minimal” level of support (pressure support, positive end-expiratory pressure, and/or continuous positive airway pressure).

  • Separation attempt

    • For intubated patients: SBT with or without extubation, or an extubation directly performed without identified SBT (whatever the type: planned or unplanned extubation).

    • For tracheostomized patients: 24 hours or more with spontaneous ventilation through tracheostomy without any mechanical ventilation.

  • Successful weaning (or successful separation)

    • Intubated patients: extubation without death or reintubation within 7 days after extubation (whether postextubation noninvasive ventilation was used or not), or ICU discharge without invasive mechanical ventilation within 7 days, whichever comes first.

    • Tracheostomized patients: spontaneous ventilation through tracheostomy without any mechanical ventilation during 7 consecutive days or ICU discharge with spontaneous breathing, whichever comes first.

  • WIND Groups

    • Group no weaning: patients never experienced any separation attempt.

    • Group 1 (short weaning): the first separation attempt resulted in a termination of the weaning process within 24 hours (successful separation or early death).

    • Group 2 (difficult weaning): weaning was terminated after more than 1 day but in less than 1 week after the first separation attempt (successful separation or death).

    • Group 3 (prolonged weaning): weaning was still not terminated 7 days after the first separation attempt (by success or death).

Mortality over Time

Crude mortality of the patients remaining on mechanical ventilation was calculated daily after the first separation attempt for patients still present but not successfully weaned at each specific day. Each additional day was associated with a progressively increasing mortality, from 19.0% in patients not weaned 1 day after the first attempt to 36.8% in patients still present in the ICU and not successfully weaned 10 days after the first attempt (Figure 4).

Extubation Modalities

The 2,051 patients in the three weaning groups underwent a total of 2,172 extubations: most of them occurred after an SBT success (77.9%), but 13.7% were performed without any SBT, 7.6% were unplanned extubations, and 0.8% were performed after an SBT failure. Physicians performed a total of 2,904 SBTs, almost equally distributed between T-piece ventilation (49.9%) and low pressure support ventilation trials (47.5%); a few (2.7%) used other types of trials with low support or low positive end-expiratory pressure. All these results are detailed in Table 4 and Table E3.

Table 4. Characteristics of the Weaning Process for the Whole Cohort of the 2,709 Included Patients

Patient characteristics (N = 2,709) 
 No. of SAs* 
  0658 (24.3)
  11,398 (51.6)
  ≥2653 (24.1)
 No. of SBTs 
  01,016 (37.5)
  11,112 (41)
  ≥2581 (21.4)
 Patients with extubation after an SBT success1,565 (57.8)
 Patients with planned extubation without SBT290 (10.7)
 Patients with unplanned extubation161 (5.9)
 Delay from inclusion to first SA, d3 (2–7)
 Patients tracheostomized112 (4.1)
 Patients reintubated240 (8.8)
 Patients receiving postextubation NIV308 (11.4)
 Patient with Do Not Reintubate order356 (13.1)
SA characteristics (N = 3,471) 
 Type of SA 
 T-tube1,448 (41.7)
 LPSV1,379 (39.7)
 Other SBT77 (2.2)
 Planned extubation without SBT298 (8.6)
 Unplanned extubation166 (4.8)
 Planned extubation despite SBT failure17 (0.5)
 Continuous period of spontaneous breathing during tracheostomy86 (2.5)
SBT characteristics (N = 2,904) 
 SBT success followed by planned extubation (within 24 h)1,691 (58.2)
 T-tube success followed by planned extubation812 (56.1)
 Low PSV success followed by planned extubation827 (60.0)
 “Other types” of SBT success followed by planned extubation52 (67.5)
 SBT success followed by successful extubation (within 24 h)1,457 (50.1)
 T-tube success followed by successful weaning§688 (47.5)
 Low PSV success followed by successful weaning§729 (52.9)
 “Other types” success followed by successful weaning§40 (51.9)
Extubation characteristics (N = 2,172) 
 Extubation with SBT success1,691 (77.9)
 Extubation despite SBT failure17 (0.8)
 Planned extubation without SBT298 (13.7)
 Unplanned extubation166 (7.6)
 Delay from inclusion to first extubation, d (N = 1,974)4 (2–7)
 Successful extubation1,772 (81.6)
 Delay from inclusion to first successful extubation, d (N = 1,772)3 (2–8)
 Delay from inclusion to first unplanned extubation, d4 (2–8)
 Successful unplanned extubation113 (5.2) [68% of the unplanned extubations]

Definition of abbreviations: ICU = intensive care unit; IQR = interquartile range; MV = mechanical ventilation; LPSV = low-level pressure support ventilation; NIV = noninvasive ventilation; PSV = pressure support ventilation; SA = separation attempt; SBT = spontaneous breathing trial.

Data are presented as n (%) or median (IQR). The number of patients having one separation attempt is lower than the number of patients in group 1 because, as defined in the manuscript, patients having an SBT but extubated the day after (within 24 h of the first SBT) were classified in group 1 despite having two separation attempts (the SBT at Day 1 and the extubation at Day 2). On the other hand, some patients in group 2 had only one separation attempt: if a patient had an extubation as first separation attempt and then died between Day 2 and Day 7 without being reintubated, he was classified in group 2 (weaning terminated between 2 and 7 d after the first separation attempt).

*As defined in the manuscript: For an intubated patient, a separation attempt was defined as an SBT with or without extubation or an extubation directly performed without identified SBT (whatever the type: planned, unplanned extubation). For a tracheostomized patient, a separation attempt was defined as a period of consecutive days (at least 1 d) with complete spontaneous ventilation through tracheostomy without any mechanical ventilation.

An SBT was defined as a T-piece trial, a low-level pressure support ≤ 8 cm H2O and a positive end-expiratory pressure ≤ 5 cm H2O, or another type trial considered by the treating physician as an SBT.

Five patients had two unplanned extubations.

§As defined in the manuscript: For an intubated patient, successful weaning was defined as an extubation without death or reintubation within the next 7 days, whether postextubation NIV was used or not, or ICU discharge without MV within 7 d, whichever comes first. For a tracheostomized patient, a successful weaning was defined as spontaneous ventilation through tracheostomy without any mechanical ventilation during 7 consecutive days or discharged with spontaneous breathing, whichever comes first.

The 2,051 patients who had at least one separation attempt comprised 1,669 patients whose first separation attempt was an SBT and 382 who had another type of first separation attempt (Table E4). We identified three types of patients who were extubated without an SBT: (1) patients with unplanned extubation, (2) patients with withholding/withdrawal who were accordingly extubated without probability of success assessment, and (3) patients probably assumed by their physician to have a high likelihood of weaning success. After excluding patients with a decision of withholding/withdrawal (n = 95) and patients with an unplanned extubation as a first separation attempt (n = 124), we compared patients who had a planned extubation with or without SBT as first separation attempt (Table 5). In a multivariable analysis restricted to these patients, younger age, lower SOFA score at admission, and admission for planned or emergent surgery were associated with performing a first separation attempt without an SBT (Table 5). Use of a weaning protocol was not associated with use of an SBT.

Table 5. Factors Associated with Extubation after or Not a Spontaneous Breathing Trial in Patients with Planned Extubation and with No Decision of Withholding/Withdrawal, Bivariate and Multivariable Analysis

 Patients with Planned Extubation without SBT (n = 177)Patients with Planned Extubation after SBT (n = 1,489)Bivariate Analysis P Value*Multivariable Analysis
OR, 95% CIP Value
Age, yr53 ± 1661 ± 16<0.0010.97, 0.97–0.98<0.001
SAPS II at admission, points41 ± 1646 ± 17<0.001  
SOFA at admission, points5.9 ± 3.26.7 ± 3.5<0.0010.95, 0.90–0.990.044
Admission  0.0041.69, 1.20–2.390.003
 Medical107 (60.5)1,056 (70.9)
 Surgery (planned or unplanned)70 (39.5)433 (29.1)
Reintubations11 (6.3)148 (11.7)0.11  
Total No. of days of invasive MV2 (1–6)4 (2–9)<0.001  
Delay from intubation to first SA, d2 (1–5)3 (2–6)<0.001  
Length of stay in the ICU, d4 (3–12)7 (4–13)<0.001  
Death5 (2.8)41 (2.8)0.85  

Definition of abbreviations: CI = confidence interval; ICU = intensive care unit; IQR = interquartile range; MV = mechanical ventilation; OR = odds ratio; SA = separation attempt; SAPS II = Simplified Acute Physiology Score II; SBT = spontaneous breathing trial; SOFA = Sequential Organ Failure Assessment.

Data are presented as n (%), mean ± SD, or median (IQR).

*Comparison between the patients with a short weaning versus patients with a longer weaning (Student’s test, Wilcoxon test, chi-square test, or Fisher test).

To our knowledge, the WIND study is the largest prospective observational study offering a daily systematic description of the weaning process after the first separation attempt and the outcome associated with the prolongation of the weaning process. We found several new and relevant results: (1) The ICC definition of weaning from mechanical ventilation could not classify almost half of invasively ventilated patients from this mixed medical and surgical ICU population (5). The WIND classification, based on the concept of separation attempts, an evolution of the weaning success definition, and mainly on the duration of the weaning process after a first attempt, allowed classification of all patients. (2) These groups based on weaning duration have gradually increasing morbidity, length of stay, and mortality, by contrast with previous reports using the ICC classification (11, 1316). (3) Each additional day without a weaning success after the first attempt was associated with an increased crude mortality. (4) The use of protocols for sedation and weaning were not associated with a higher rate of short weaning. (5) SBTs were equally distributed between T-piece and low pressure support ventilation, and both modalities led to a successful weaning rate of approximately 50% (Table 3). (6) A small number of patients are intentionally extubated without an SBT and are less severely ill than the patients extubated after an SBT.

The classification proposed in 2007 by the ICC (5), based on expert opinion, failed to allocate one-half of the invasively ventilated patients included in the present cohort, for two main reasons. First, only successfully weaned patients could enter this classification. Many patients of the present cohort were never eventually weaned either because they died before entering the weaning process (21%), had an unsuccessful weaning process, or were discharged to another unit or a weaning center under mechanical ventilation (3%) and/or with a tracheostomy (3%). Moreover, the ICC classification did not specify how to classify tracheostomized patients and gave an unclear definition of patients receiving noninvasive ventilation 48 hours after extubation as a “weaning in progress” group. Second, the ICC classification was based on the performance of SBTs, whereas a substantial number of patients were weaned after a planned extubation time without any SBT or after unplanned extubation (10.7 and 5.9%, respectively). The six published studies that used the 2007 ICC so far also evidenced a very high rate of patients who could not be classified, ranging from 40 to 75% (1116). Even though none of these studies discussed the applicability of the 2007 classification, several authors did not precisely follow ICC group definitions: only two groups strictly applied the classification proposed in 2007 (14, 15).

We propose modification of the ICC classification to overcome these issues. The new WIND classification is operational for daily clinical practice, allows a straightforward epidemiological description, and accounts for all the patients receiving invasive ventilation. Because we considered all types of extubation with or without SBT, the first main change was the definition of “separation attempts” that brought together all the situations leading to extubation: a formal SBT as well as an extubation without an SBT. In a study where patients were extubated without an SBT, the rate of failure with reintubation in the next 48 hours was as high as 39.3% (19). Nevertheless, a recent study found the same rate of successful extubation with or without using an SBT in a mixed medical and surgical population of patients (20). Our data suggest, however, that they may not be the same group of patients. The fact that this was not associated with weaning protocols may also suggest that it is more driven by individual patient characteristics than by ICU policy. Furthermore, we also considered unplanned extubation as a separation attempt (leading to successful weaning in most of the cases). In addition, we also took into account tracheostomized patients in the WIND classification, because almost 5% of the patients included in the cohort had a tracheostomy at any point of their ICU stay.

The second main change was the weaning success definition: in the ICC, a delay of 48 hours without any form of mechanical ventilation (invasive or not) was retained. We chose a 7-day delay whether noninvasive ventilation was used or not. Time after which extubation is considered successful remains a subject of debate and is further complicated by the use of NIV, which may prolong the time to reintubation (4, 21).

Distribution of the Weaning Groups

The WIND classification strongly discriminated three weaning groups with very different weaning trajectories and prognosis using the weaning time span. Defining groups by the delay between the first separation attempt and the outcome (weaning success or death) makes the classification easy and appears to be of paramount importance for the prognosis. The short weaning group gathered patients who could be successfully weaned early or died immediately after extubation without reintubation (90% of them with Do Not Reintubate orders; Table 2). This group 1 accounted for most of the patients included (57.0% of the whole cohort, 75.2% of the weaning cohort). Almost 25% of the patients entering the weaning process were allocated to group 2 (difficult weaning) and 3 (prolonged weaning) with an almost equal distribution. This result could justify future studies to better understand the pathophysiological process leading to difficult weaning to apply strategies to reduce the number of patients transitioning to the prolonged weaning group. By contrast with previous reports, we found a very different outcome for these two groups (11, 1315), thus bringing an important predictive validity of the classification. Indeed, duration of ventilation, ICU stay, and mortality all differ among the groups. As found in previous reports, the total length of mechanical ventilation increased from group 1 to 3 (1116); nevertheless, in this cohort, the difference between group 2 and 3 was mainly due to the length of mechanical ventilation after the first separation attempt: surprisingly, the length of ventilation before the first separation attempt did not differ between group 2 and group 3, emphasizing the importance of what happens after the first separation attempt and shining a new light on the mortality related to the duration of mechanical ventilation. This supports the finding that after a first separation attempt, each additional day before the eventual weaning increased the risk of death. Our results suggest that the three groups are very different and have different outcomes. Including the three groups in randomized controlled trials, however, has often produced results heavily driven by group 1 and leads to recommendations potentially not applicable to other patients. Patients in group 3 represent a population with special needs, prolonged hospitalization, and high resource consumption, whose specific organizational consequences need to be better analyzed (22).

Mortality over Time

Because patients who died early after extubation or who failed weaning were excluded from the ICC definition, ICU mortality reported in previous weaning cohort studies was very low, especially in ICC groups 1 and 2 (from 0 to 7% in a general population [11, 12, 14, 15] to 12% in a cohort with chronic obstructive pulmonary disease [13]), and did not match with clinical reality. Including those patients in our cohort resulted in a more usual case mix and highlighted differences between the three weaning groups. The daily data collection allowed description of the risk of death associated with the duration of the weaning process. We have shown that each additional day of mechanical ventilation after failure of the first attempt is associated with an increasing risk of death, with a kind of plateau between Day 2 and Day 6. We believe this represents useful information for clinicians. In contrast, Peñuelas and colleagues, on the basis of a different population and classification definitions, found higher mortality only for weaning lasting more than 7 days (11).

Strengths and Weaknesses

This study comprises the largest prospective cohort with daily collection of weaning assessment, extubation outcome, and was especially designed for this purpose. The simple definitions of “separation attempt” and “successful weaning” allowed classification of the whole cohort and overcame most of the ICC issues, meeting daily clinical practice needs. Beyond limitations inherent to the observational design of the WIND study, it should be noted that only European ICUs participated (half of them located in university hospitals), and our conclusions may not apply worldwide. We believe, however, that the type of classification proposed is operational to be adopted worldwide. Apart from mechanical ventilation settings, only a few clinical data were collected, and patients were not followed up after ICU discharge to reduce investigators’ workload. Although this limited the possibilities of performing additional analyses and adjustments, it also resulted in a high-quality dataset without missing data.

Conclusions

The first separation attempt is a major milestone in the ventilated patients’ ICU course. The proposed new definitions of separation attempts and successful weaning allow the classification of all patients in meaningful groups. This is relevant to optimize the weaning process and to design and conduct future research in this field. The three weaning groups defined by the delay between the first separation attempt and the weaning termination generate groups with very different morbidity and mortality. Mortality starts to increase after the first unsuccessful separation attempt.

The authors thank Sylvie Chevret (Paris 7 University) for her edit of the manuscript and her statistical advice.

WIND Study Group Collaborators: Website designer: B. Sanchiz. Amiens (Respiratory ICU): C. Andrejak and D. Basille. Amiens (Cardiothoracic and vascular ICU): H. Dupont and M. Miclo. Angers: M. Royer. Barcelone: F. Roche-Campo and F. Parrilla. Caen: B. Sauneuf and C. Daubin. Créteil: E. Akachian. Dieppe: P. L. Declercq and S. Gelinotte. Dijon: J. P. Quenot and A. Pavon. Dreux: P. Dubosq. HEGP: E. Guerot and J. L. Diehl. La Chaux de Fond: H. Zender. La Pitié (ICU and respiratory unit): J. Mayaux and A. Demoule. Le Mans: B. Derrien. Lille (ICU 1): E. Jaillette and S. Nseir. Lille (ICU 2): E. Parmentier-Decrucq and D. Mathieu. Mayotte: F. Lion. Melun: S. Jochmans and J. Chelly. Mostoles: B. Quesada and M. Muñoz-Varea. Mulhouse: A. Poidevin and M. Egard. Nancy: M. Conrad and P. E. Bollaert. Nantes: C. Guitton and C. Bretonniere. Nice: J. Dellamonica and P.E. Danin. Orléans: A. Bretagnol. Poitiers: R. Coudroy and A. W. Thille. Reus: H. Aguirre and I. Vallverdú. Roanne: P. Beuret. Rodez: S. Ena. Rouen (surgical ICU): P. Gouin and B. Veber. Rouen (medical ICU): D. Carpentier and F. Tamion. Saragosse: B. Zalba-Etayo and N. Raventos-Irigoyen. St. Joseph: C. Grégoire and J. Fournier. St. Louis: V. Lemiale and D. Reuter. Tahiti: V. Lassalle and C. Huet. Tenon: D. Contou and M. Fartoukh. Troyes: G. Simon and B. Beilouny. Valence: Q. Blanc and P. Fernandez.

1. Fagon JY, Chastre J, Vuagnat A, Trouillet JL, Novara A, Gibert C. Nosocomial pneumonia and mortality among patients in intensive care units. JAMA 1996;275:866869.
2. Provost KA, El-Solh AA. Complications associated with mechanical ventilation. In: Tobin MJ, editor. Principles and practices of mechanical ventilation, 3rd ed. New York: McGraw-Hill Education; 2012. pp. 973994.
3. Epstein SK, Ciubotaru RL. Independent effects of etiology of failure and time to reintubation on outcome for patients failing extubation. Am J Respir Crit Care Med 1998;158:489493.
4. Thille AW, Richard J-CM, Brochard L. The decision to extubate in the intensive care unit. Am J Respir Crit Care Med 2013;187:12941302.
5. Boles J-M, Bion J, Connors A, Herridge M, Marsh B, Melot C, Pearl R, Silverman H, Stanchina M, Vieillard-Baron A, et al. Weaning from mechanical ventilation. Eur Respir J 2007;29:10331056.
6. Esteban A, Ferguson ND, Meade MO, Frutos-Vivar F, Apezteguia C, Brochard L, Raymondos K, Nin N, Hurtado J, Tomicic V, et al.; VENTILA Group. Evolution of mechanical ventilation in response to clinical research. Am J Respir Crit Care Med 2008;177:170177.
7. McConville JF, Kress JP. Weaning patients from the ventilator. N Engl J Med 2012;367:22332239.
8. Ely EW, Baker AM, Dunagan DP, Burke HL, Smith AC, Kelly PT, Johnson MM, Browder RW, Bowton DL, Haponik EF. Effect on the duration of mechanical ventilation of identifying patients capable of breathing spontaneously. N Engl J Med 1996;335:18641869.
9. Girard TD, Kress JP, Fuchs BD, Thomason JWW, Schweickert WD, Pun BT, Taichman DB, Dunn JG, Pohlman AS, Kinniry PA, et al. Efficacy and safety of a paired sedation and ventilator weaning protocol for mechanically ventilated patients in intensive care (Awakening and Breathing Controlled trial): a randomised controlled trial. Lancet 2008;371:126134.
10. Blackwood B, Alderdice F, Burns K, Cardwell C, Lavery G, O’Halloran P. Use of weaning protocols for reducing duration of mechanical ventilation in critically ill adult patients: Cochrane systematic review and meta-analysis. BMJ 2011;342:c7237.
11. Peñuelas O, Frutos-Vivar F, Fernández C, Anzueto A, Epstein SK, Apezteguía C, González M, Nin N, Raymondos K, Tomicic V, et al.; Ventila Group. Characteristics and outcomes of ventilated patients according to time to liberation from mechanical ventilation. Am J Respir Crit Care Med 2011;184:430437.
12. Tonnelier A, Tonnelier J-M, Nowak E, Gut-Gobert C, Prat G, Renault A, Boles J-M, L’Her E. Clinical relevance of classification according to weaning difficulty. Respir Care 2011;56:583590.
13. Sellares J, Ferrer M, Cano E, Loureiro H, Valencia M, Torres A. Predictors of prolonged weaning and survival during ventilator weaning in a respiratory ICU. Intensive Care Med 2011;37:775784.
14. Funk G-C, Anders S, Breyer M-K, Burghuber OC, Edelmann G, Heindl W, Hinterholzer G, Kohansal R, Schuster R, Schwarzmaier-D’Assie A, et al. Incidence and outcome of weaning from mechanical ventilation according to new categories. Eur Respir J 2010;35:8894.
15. Pu L, Zhu B, Jiang L, Du B, Zhu X, Li A, Li G, He Z, Chen W, Ma P, et al. Weaning critically ill patients from mechanical ventilation: a prospective cohort study. J Crit Care 2015;30:862.e7862.e13.
16. Jeong BH, Ko MG, Nam J, Yoo H, Chung CR, Suh GY, Jeon K. Differences in clinical outcomes according to weaning classifications in medical intensive care units. Plos One 2015;10:e0122810.
17. Beduneau G, Pham T, Schortgen F, Chretien J-M, Richard J-CM, Mercat A, Mancebo J, Brochard LJ. REVA network: a new classification for patients weaning from mechanical ventilation [abstract]. Intensive Care Med 2014;40:A13.
18. Beduneau G, Pham T, Richard J, Schortgen F, Chretien J, Mercat A, Mancebo J, Brochard L. WIND Study Group on behalf of REVA Network. Prognosis of mechanically ventilated patients entering the weaning process [abstract]. Intensive Care Med Exp 2015;3:A426.
19. Zeggwagh AA, Abouqal R, Madani N, Zekraoui A, Kerkeb O. Weaning from mechanical ventilation: a model for extubation. Intensive Care Med 1999;25:10771083.
20. Wang J, Ma Y, Fang Q. Extubation with or without spontaneous breathing trial. Crit Care Nurse 2013;33:5055.
21. Girault C, Bubenheim M, Abroug F, Diehl JL, Elatrous S, Beuret P, Richecoeur J, L’Her E, Hilbert G, Capellier G, et al.; VENISE Trial Group. Noninvasive ventilation and weaning in patients with chronic hypercapnic respiratory failure: a randomized multicenter trial. Am J Respir Crit Care Med 2011;184:672679.
22. Beduneau G, Richard J-CM, Brochard L. Prolonged respiratory insufficiency and ventilator dependence in the ICU. In: Stevens RD, Hart N, Herridge MS, editors. Textbook of post-ICU medicine: the legacy of critical care. Oxford: Oxford University Press; 2014. pp. 150161.
Correspondence and requests for reprints should be addressed to Gaëtan Beduneau, M.D., Rouen University Hospital, Medical Intensive Care, 1 Germont Street, Charles Nicolle Hospital, 76000 Rouen, France. E-mail:

*These authors contributed equally to this work.

A complete list of members may be found before the beginning of the References.

Supported by a grant from the nonprofit Association Départementale des Insuffisants Respiratoires (ADIR) of the Haute Normandie area in France.

Author Contributions: Conception, design, analysis, and interpretation: G. Béduneau, T.P., F.S., J.-M.C., J.M., A.M., J.-C.M.R., and L.B. Drafting the manuscript for important intellectual content: all authors.

This article has an online supplement, which is accessible from this issue’s table of contents at www.atsjournals.org

Originally Published in Press as DOI: 10.1164/rccm.201602-0320OC on September 14, 2016

Author disclosures are available with the text of this article at www.atsjournals.org.

Comments Post a Comment




New User Registration

Not Yet Registered?
Benefits of Registration Include:
 •  A Unique User Profile that will allow you to manage your current subscriptions (including online access)
 •  The ability to create favorites lists down to the article level
 •  The ability to customize email alerts to receive specific notifications about the topics you care most about and special offers
American Journal of Respiratory and Critical Care Medicine
195
6

Click to see any corrections or updates and to confirm this is the authentic version of record