In general, critically ill patients in need of invasive ventilation receive mechanical ventilation via an endotracheal tube. It is a common practice to replace the initial translaryngeal tube with a tracheostomy to facilitate oral care and weaning from the ventilator. However, currently, there are no clear guidelines describing indications and contraindications to convert an endotracheal tube to a tracheostomy in critically ill patients. Such a guideline could reduce the current practice variation and inform decisions that currently may be somewhat arbitrary.

Summary of Findings

Click here to see this table in a document

_{^1. Risk of Bias: serious. Inadequate/lack of blinding of participants and personnel, resulting in potential for performance bias; Inconsistency: serious. The direction of the effect is not consistent between the included studies; Imprecision: serious. Wide confidence intervals.}

_{^2. Risk of Bias: serious. Inadequate/lack of blinding of participants and personnel, resulting in potential for performance bias, due to recruitment bias; Imprecision: very serious. Wide confidence intervals.}

_{^3. Risk of Bias: serious. Inadequate/lack of blinding of participants and personnel, resulting in potential for performance bias; Inconsistency: serious. The direction of the effect is not consistent between the included studies; Indirectness: no serious. Differences between the outcomes of interest and those reported (e.g short-term/surrogate, not patient-important); Imprecision: serious. Wide confidence intervals.}

_{^4. Risk of Bias: serious. Inadequate/lack of blinding of participants and personnel, resulting in potential for performance bias; Inconsistency: serious. The direction of the effect is not consistent between the included studies; Imprecision: very serious. Wide confidence intervals.}

_{^5. Risk of Bias: very serious. Inadequate/lack of blinding of participants and personnel, resulting in potential for performance bias, due to recruitment bias; Imprecision: serious. Wide confidence intervals.}

_{^6. Risk of Bias: very serious. Inadequate/lack of blinding of participants and personnel, resulting in potential for performance bias, due to recruitment bias; Imprecision: serious. Wide confidence intervals.}

_{^7. Risk of Bias: very serious. Inadequate/lack of blinding of participants and personnel, resulting in potential for performance bias, due to recruitment bias; Imprecision: serious. Wide confidence intervals.}

Description of studies

A total of five studies (one systematic review describing eight different studies, and four randomized controlled trials) were included in the analysis of the literature. Important study characteristics and results are summarized in table 2. In the analysis of the literature we included multiple definitions for the control group, e.g., either prolonged intubation or late tracheostomy. The exact intervention and control, and the definition of the comparison, can also be found in table 2. 

Andriolo (2015) performed a systematic review to evaluate the effectiveness and safety of early (≤ 10 days after tracheal intubation) versus late tracheostomy (> 10 days after tracheal intubation) in critically ill adults predicted to be on prolonged mechanical ventilation with different clinical conditions. The Cochrane Central Register of Controlled Trials (CENTRAL) (2013, Issue 8); MEDLINE (via PubMed) (1966 to August 2013); EMBASE (via Ovid) (1974 to August 2013); LILACS (1986 to August 2013); PEDro (Physiotherapy Evidence Database) at www.pedro.fhs.usyd.edu.au (1999 to August 2013) and CINAHL (1982 to August 2013) were searched. The review included all (quasi-) RCTs that met the following inclusion criteria. For types of patients, the inclusion criteria were 1) Critically ill patients (for whom death is possible or imminent); 2) Patients expected to be on prolonged mechanical ventilation; and 3) Adults (≥ 18 years).  Inclusion criteria for the type of intervention was 1. Early tracheostomy, if no serious attempt was made to wean the patient from the ventilator (tracheostomy based only on clinical or laboratory results and performed from two days to 10 days after intubation); and 2) late tracheostomy, if weaning had not been successful; performed later than 10 days after intubation. Eight studies were included.

Mohamed (2014) performed a single center prospective randomized study on 40 patients who underwent bedside percutaneous dilatational tracheostomy (PDT). Patients included were older than 18 years, were mechanically ventilated for respiratory failure>24 h and have had no previous pulmonary infection by chest X-ray.

Blot (2008) performed a multicenter prospective randomized trial in twenty-five medical and surgical ICUs in France. Patients with ‘‘main underlying pathology’’ coded ‘‘neurological’’ were included; the study database did not record if this underlying neurological pathology was acute or not (personal communication with study authors). Patients were excluded if they had an irreversible neurological disease or post-trauma intracranial hypertension (intracranial pressure >20 cmH2O and/or cerebral edema on head CT scan).

Bouderka (2004) performed a single-center prospective randomized trial in one ICU in Morrocco. Patients were included in the study if they met the following criteria: isolated severe head injury (admission Glasgow coma scale (GCS) score ≤8); cerebral contusion on CT scan; GCS score <8 on the fifth day without any sedation.

Saffle (2002) performed a single-center prospective randomized trial in the Intermountain Burn Center in the United States. Patients were eligible if they were older than 18 years of age, hospitalized within 24 hours of acute burn injury, and if they required ongoing mechanical ventilatory support on postburn day (PBD)2 (ie, 48–72 hours postinjury). All intubated patients were evaluated for the presence of inhalation injury using fiberoptic bronchoscopy. The following patients were excluded from participation in the study: 1) pregnant women; 2) patients with preexisting significant renal or hepatic disease; 3) patients taking corticosteroids prior to admission; and 4) patients who did not have cutaneous burn injuries.

Table 2. Characteristics of included studies

Results

Duration of mechanical ventilation (critical)

Two of the eight studies in the review of Andriolo (2015) reported the outcome mean days of mechanical ventilation during 1 to 60 days (Rumbak, 2004; Trouillet, 2011). The other four studies also reported the outcome mechanical ventilation duration (Blot, 2008; Bouderka, 2004; Saffle; 2002; Mohamed; 2004).

Since Blot (2008) reported the median (range) duration of mechanical ventilation during the first 28 days, these results could not be used in the pooled analysis. Blot (2008) reported a median (range) duration in severely ill ICU patients of 14 (2 – 18) days for the early tracheostomy group (n=61) and 16 (3-28) days for the prolonged intubation group (n=62).

The combined mean difference in the duration of mechanical ventilation in days was 4.14 (95% CI: -10.79 to 2.51) in favour of the early tracheostomy group. This difference was considered to be clinically relevant.

_{Figure 1. The Mean Difference (MD) in the duration of mechanical ventilation (in days) between ICU-patients receiving early of late tracheostomy}

_{Z: p-value of the pooled effect; df: degrees of freedom; I2: statistical heterogeneity; CI: confidence interval}

Use of sedatives (critical)

The outcome use of sedatives was reported in one study (Blot, 2008).

Blot (2008) reported on the outcome sedation and measured the number of sedation-free days within the first 28 days. The median (range) of sedation-free days were 18 days (0 to 27) in de early tracheostomy group (n=61) and 15 days (0 to 27) in the prolonged intubation group (n=62). This difference was considered to be clinically relevant.

Mobilization (critical)

The outcome mobilization was reported in one study (Blot, 2008).

Blot (2008) reported the outcome chair positioning on day 28 and day 60, which was measured as the time of the first transfer from bed to chair. In the early tracheostomy group (n=61) the chair positioning rate on day 28/day 60 was 80%/100% with a median time (95% CI) of 20 days (15-23), compared to 65%/88%, and a median time of 22 days (16-28), in the prolonged intubation group (n=61). Since the mean was not reported, the mean difference could not be calculated. Therefore, it could not be determined whether this difference is clinically relevant.

Complications (important)

Pneumonia  

Five of the eight studies in the review of Andriolo (2015) reported the incidence rate of pneumonia, defined as the incidence of ICU-acquired pneumonia within 28 days. (Dunham 1984; Rumbak 2004; Terragni 2010; Trouillet 2011; Zheng 2012). In addition, the four separate studies also reported on the incidence rate of pneumonia (Blot, 2008; Bouderka, 2004; Saffle; 2002; Mohamed; 2004).

The combined percentage of pneumonia events in the early tracheostomy group is 32.0% (193 out of 603 patients), versus 38.4% (236 out of 614 patients) in the late tracheostomy group, with a risk ratio (RR) of 0.86 (95% CI: 0.68 to 1.08).  The risk difference (RD) was calculated using the baseline risk (BR) from the PRECISe Trial (Bels et al., 2024), which included Dutch and Belgium patients. Based on the incidence of ventilator-acquired pneumonia in this population, which was 115 of the 465 (25%), the RD is -0.035 (95%CI: -0.08 to 0.02), in favour of the early tracheostomy group. The RD of -0.035 represents a difference of 3.5%, which was not considered to be clinically relevant.

_{Figure 2. The Risk Ratio (RR) for the number of pneumonia events in ICU-patients receiving either early of late tracheostomy}

_{Z: p-value of the pooled effect; df: degrees of freedom; I2: statistical heterogeneity; CI: confidence interval}

Peri-procedural mortality  

The outcome ‘peri-procedural mortality’ was not reported in the included randomized controlled trial of Andriolo (2014) or the four separate randomized controlled trials (Blot, 2008; Bouderka, 2004; Saffle; 2002; Mohamed; 2004).

Bleeding

The outcome ‘bleeding’ was not reported in the included randomized controlled trial of Andriolo (2014) or the four separate randomized controlled trials (Blot, 2008; Bouderka, 2004; Saffle; 2002; Mohamed; 2004).

Accidental extubation/loss of airway

The outcome ‘accidental extubation/loss of airway’ was not reported in the included randomized controlled trial of Andriolo (2014) or the four separate randomized controlled trials (Blot, 2008; Bouderka, 2004; Saffle; 2002; Mohamed; 2004).

Hospital mortality (important)

Three of the eight studies in the review of Andriolo (2015) reported mortality on day 28 (Terragni 2010; Trouillet 2011; Zheng 2012), in addition to one of the separate studies (Blot, 2008). The other three separate studies also reported hospital mortality, although not explicitly mentioned on which day the outcome was measured (Bouderka, 2004; Saffle; 2002; Mohamed; 2004).

In total, 116 of the 509 patients died in the early tracheostomy group, compared to 131 of the 504 patients in the late tracheostomy/prolonged intubation group. The incidence in the late tracheostomy group does not represent the Dutch population. Unfortunately, due to the absence of a Dutch baseline risk we could not adapt the RD to the Dutch population. Therefore, the RD might not represent the Dutch situation. The outcome mortality had a RR of 0.88 (95%CI: 0.71 to 1.10) and a RD of -0.03 (95%CI: -0.08 to 0.02), in favour of the early tracheostomy group. The RD of 3% was not considered to be clinically relevant.

_{Figure 3. The Risk Ratio (RR) for mortality in ICU-patients receiving either early of late tracheostomy}

_{Z: p-value of the pooled effect; df: degrees of freedom; I2: statistical heterogeneity; CI: confidence interval}

Comfort (important)

One study reported on the outcome comfort (Blot, 2008).

Blot (2008) reported on self-evaluated comfort using a questionnaire and on objective comfort criteria evaluated on day 28 and day 60, including the time of recovery time of oral feeding and speech. Results are expressed according to a numerical scale from 1 (acceptable) to 10 (unbearable). Most of the criteria evaluated in the self-evaluation questionnaire were in favour of tracheostomy. The differences in the following criteria, all in favour of early tracheostomy, were considered clinically relevant: difficulty to move (early tracheostomy: 2 (1-7); prolonged intubation: 5 (1-10)), mouth discomfort (early tracheostomy: 2 (1-10); prolonged intubation: 5 (1-10)), feeling of mouth cleanliness (early tracheostomy: 3 (1-8); prolonged intubation: 5 (2-10)), feeling of overall safety (early tracheostomy: 1 (1-10); prolonged intubation: 4 (1-9)), perception of change in body image (early tracheostomy: 3 (1-7); prolonged intubation: 5 (1-9)), and overall feeling of comfort (early tracheostomy: 3 (1-9); prolonged intubation: 5 (1-8)). In addition, all patients who had completed the self-evaluation questionnaire and had undergone both trans-laryngeal intubation and early or late tracheostomy (n=13) considered that the tracheostomy, and not prolonged intubation, was the most comfortable technique.

There were no differences regarding the recovery time of oral feeding (early tracheostomy group (n=61): recovery on d28/d60: 64%/91%; median time (95%) 18 days (13-28); prolonged intubation group (n=62); recovery on d28/d60: 64%/86%; median time (95%): 23 days (20-27)), and speech recovery (early tracheostomy group (n=61): recovery on d28/d60: 57%/83%; median time (95%) 26 days (18-32); prolonged intubation group (n=62); recovery on d28/d60: 63%/85%; median time (95%): 22 days (16-30)). Since the median, and not the mean, was reported, the mean difference could not be calculated. Therefore, it could not be determined whether this difference is clinically relevant.

Patient satisfaction (important)

The outcome ‘patient satisfaction’ was not reported in the included randomized controlled trial of Andriolo (2014) or the four separate randomized controlled trials (Blot, 2008; Bouderka, 2004; Saffle; 2002; Mohamed; 2004).

Family satisfaction (important)

The outcome ‘Family satisfaction’ was not reported in the included randomized controlled trial of Andriolo (2014) or the four separate randomized controlled trials (Blot, 2008; Bouderka, 2004; Saffle; 2002; Mohamed; 2004).

Dysphagia (important)

One study reported on the outcome dysphagia (Blot, 2008).

Blot (2008) reported on the outcome swallowing disorders, which belonged to the laryngeal symptoms. In the early tracheostomy group, 4 of the 61 (6.6%) patients experienced a swallowing disorder compared to 8 of the 62 (12.9%) patients in the prolonged intubation group. The incidence in the late tracheostomy group does not represent the Dutch population. Unfortunately, due to the absence of a Dutch baseline risk we could not adapt the RD to the Dutch population. Therefore, the RD might not represent the Dutch situation. The outcome dysphagia had a RR of 0.51 (95% CI: 0.16 to 1.60) and a RD of 0.06 (95% CI: -0.17 to 0.04), in favour of the early tracheostomy group. The RD of 6% in the incidence of swallowing disorders was not considered to be clinically relevant.

Reintubation (important)

The outcome ‘reintubation’ was not reported in the included randomized controlled trial of Andriolo (2014) or the four separate randomized controlled trials (Blot, 2008; Bouderka, 2004; Saffle; 2002; Mohamed; 2004).

Weaning duration (important)

The outcome ‘weaning duration’ was not reported in the included randomized controlled trial of Andriolo (2014) or the four separate randomized controlled trials (Blot, 2008; Bouderka, 2004; Saffle; 2002; Mohamed; 2004).

A systematic review of the literature was performed to answer the following question(s):

What is the effectiveness of tracheostomy versus no tracheostomy and early tracheostomy versus prolonged intubation or late tracheostomy in critically ill adult patients who are intubated? 

Table 1. PICO

Relevant outcome measures

The guideline panel considered duration of mechanical ventilation, mobilization, and use of sedatives as critical outcome measures for decision making; and complications, comfort, patient satisfaction, family satisfaction, dysphagia, reintubation, weaning duration, and hospital mortality as an important outcome measure for decision making.

A priori, the guideline panel did not define the outcome measures duration of mechanical ventilation, patient satisfaction, family satisfaction, dysphagia, reintubation, and weaning duration, but used the definitions used in the studies. The guideline panel defined the other outcomes outcome measures as follows:

• Mobilization: active motion, in or out of bed.
• Use of sedatives: midazolam, propofol, dexmedetomidine or other benzodiazepines.
• Complications: Peri-procedural mortality, bleeding, accidental extubation/loss of airway, pneumonia.
• Comfort: either self-reported or as an objective measure such as resumption of oral feeding or speech recovery.

The guideline panel defined the following thresholds as minimal clinically (patient) important differences:

• Duration of mechanical ventilation: ≥1 day
• Mobilization: 0.5 SD
• Use of sedatives (sedation-free days): ≥1 day
• Complications: Risk difference of ≥5%, except for peri-procedural mortality, then risk difference of ≥1%
• Hospital mortality: Risk difference of ≥3.5%
• Comfort: 0.5 SD, and on the 1-10 scale ≥ 2 points
• Patient satisfaction: 0.5 SD
• Family satisfaction: 0.5 SD
• Dysphagia: Risk difference of ≥10%
• Reintubation: Risk difference of ≥10%
• Weaning duration: ≥1 day

Search and select (Methods)

The databases Medline (via OVID) and Embase (via Embase.com) were searched with relevant search terms until 27-06-2024. The detailed search strategy is listed under the tab ‘Literature search strategy’. The systematic literature search resulted in 698 hits. Studies were selected based on the following criteria ‘intubated ICU patients’ AND ‘tracheostomy’ AND adults. Initially, 48 studies were selected based on title and abstract screening. After reading the full text, 43 studies were excluded (see the exclusion table under the tab ‘Evidence tabellen’), and five studies were included.