Weaning from ventilation and extubation of children in critical care

Learning objectives.

By reading this article, you should be able to:

• •Identify the predictors of successful extubation in paediatric ICU (PICU) patients.
• •Describe the factors affecting mechanical ventilator weaning in PICU patients.
• •Distinguish the tests used to determine readiness for extubation in PICU patients.
• •Detail the factors associated with failure of extubation.
• •Anticipate and manage common complications after extubation in children.

Key points.

• •No standardised protocols exist to guide weaning from mechanical ventilation in critically ill children.
• •A systematic approach to identifying predictors of successful weaning and risk factors for failed extubation is recommended to guide weaning from ventilation and extubation.
• •Paediatric intensivists must balance the risk of premature extubation with the risks of prolonged mechanical ventilation.
• •Extubation readiness testing (ERT) and spontaneous breathing trials (SBTs) can be used to identify patients who are at risk of failed extubation.
• •The causes of failed extubation are multifactorial; upper airway obstruction is the most common single cause.

With the increase in paediatric bone marrow and solid organ transplantation, renal replacement therapy and extracorporeal membrane oxygenation, the use of mechanical ventilation in paediatric critical care continues to increase.¹ Children who require mechanical ventilation often have underlying respiratory disorders as the main cause of their respiratory failure.^2,3 Weaning from mechanical ventilation and extubation are vital steps in their recovery; however, these steps are associated with significant risks. Balancing the hazards of premature extubation such as extubation failure and emergent reintubation with those of prolonged mechanical ventilation is a challenge. There is currently a paucity of published evidence-based protocols or guidelines to guide this process; therefore, predictors for successful extubation in critically ill children are not always clear cut. This has resulted in wide variation in practice amongst paediatric intensive care unit (PICU) providers that is typically based on institutional norms. This paper summarises published predictors for successful extubation, ventilator weaning strategies, the use of extubation readiness testing (ERT) and spontaneous breathing trials (SBTs) to predict successful extubation, risk factors for extubation failure and complications associated with extubation in paediatric critical care. Tracheal extubation in children and adults after anaesthesia for surgery have been covered recently in this journal.^4,5

Predictors of successful extubation in critically ill children

There are two conditions associated with successful extubation: the ability to maintain a patent airway upon removal of the tracheal tube and tolerance of spontaneous breathing without invasive mechanical ventilatory support. There are several predictive indices that that can help to guide extubation in critically ill children (Table 1).³

Table 1.

Predictive indexes for successful extubation in PICU.^2,3,6,7 Cdyn: dynamic compliance; NIF, negative inspiratory force; V_D, dead space ventilation; V_T, tidal volume.

	Definition	Predictive value
Spontaneous ventilatory frequency	Number of independent breaths per minute	RR <45
Spontaneous tidal volume	The amount of air that moves in and out of the lungs with each respiratory cycle	VT >5.5 ml kg−1
Rapid Shallow Breathing Index (RSBI)	RR/VT	RSBI ≤8 breaths min−1 ml−1 kg−1
Compliance, Resistance, Oxygenation, Pressure (CROP) index	Cdyn × NIF × (Pao2/PAo2)/RR	CROP index ≥0.15 ml kg−1 breaths−1 min−1
Volumetric capnography	VD/VT	VD/VT <0.50
Crying vital capacity (CVC)	Maximum volume of air expired during a single cry	CVC 17 ml kg−1
Peak inspiratory flow rate (PIFR)	Maximum amount of air inhaled (as fast as possible) over one deep breath after maximal exhalation to residual volume	PIFR 3.5 ml s−1 cm−1
Maximum inspiratory pressure (MIP)	Maximum pressure generated with forceful inspiration after an expiration to residual volume	MIP 50 cmH2O

Patients in the PICU who are extubated successfully often have ventilatory frequencies that approximate normal age-appropriate rates, larger tidal volumes and higher negative inspiratory force (NIF).⁶ The Rapid Shallow Breathing Index (RSBI) assesses ventilatory frequency against tidal volume. The Compliance, Rate, Oxygenation and Pressure index (CROP index) assesses lung compliance, alveolar and arterial oxygen pressure against ventilatory frequency. Both of these have been shown to be very useful in predicting successful extubation.^2,3,6 An RSBI value ≤8 bpm ml⁻¹ kg⁻¹ has a sensitivity of 74% and specificity of 73% for predicting a successful extubation and a CROP index value ≥0.15 ml kg⁻¹ breaths⁻¹ min⁻¹ has a sensitivity of 83% and specificity of 53%.⁶ Of note, in adult studies, the predictive value of RSBI has been shown to be affected by factors such as duration of mechanical ventilation, timing and technique of the measurements, sex, tracheal tube size, presence of agitation and the use of opioid medications. These should be considered as they may affect the use of RSBI to predict successful extubation in children.^2,3,6,8

Volumetric capnography is another useful technique that helps to predict successful extubation in critically ill children. It involves measurement of physiological and alveolar dead space at the bedside by plotting the concentration of carbon dioxide (CO₂) in airway gas against expired volume. Physiologic dead space (V_D)/tidal volume (V_T) ratio <0.50 has been found to reliably predict successful extubation with 75% sensitivity and 92% specificity, whereas V_D/V_T >0.65 predicts failure of extubation.^2,3

Crying vital capacity (CVC) 17 ml kg⁻¹, peak inspiratory flow fate (PIFR) 3.5 ml s⁻¹ cm⁻¹ and maximum inspiratory pressure (MIP) 50 cmH₂O are other good predictors of successful extubation in the PICU (Table 1). For patients who meet all three of these respiratory criteria, successful extubation rate has been reported to be 97.9%. For those who pass one to two of them, the successful extubation rate is 88.8%, and for those who fail all three it is only 66.7%. In a retrospective study by Toida and colleagues, the few patients who passed all three assessments yet failed extubation were noted to have postextubation respiratory muscle fatigue, altered mental status, or upper airway obstruction (UAO) from glottic oedema.⁷ The authors recommend using these predictive indices in combination with clinical assessment to guide the patient's readiness to wean and improve the chances of successful extubation.

Mechanical ventilator weaning

Historically it was assumed that critically ill children with respiratory failure behave similarly to their adult ICU counterparts and should be gradually weaned from mechanical ventilation. This has been found to be false. In a multicentre randomised trial, Randolph and colleagues demonstrated that the majority of children were able to wean from ventilator support in ≤2 days. The use of weaning protocols did not significantly shorten this already brief duration of weaning.⁹ They also found that more than a third of the patients, whose physicians determined that they were ready to wean, had already met bedside criteria for being ready for extubation. Most of these patients were successfully extubated within 24 h. The study also found that although 50% of patients receiving mechanical ventilation in PICU were extubated by day 2 of admission; the remainder typically required prolonged ventilator support. PICU patients who require prolonged ventilator support may benefit from optimisation of weaning strategies.⁹ Table 2 highlights the key factors that have been shown to affect ventilator weaning in patients in the PICU.

Table 2.

Factors affecting weaning in patients in PICU.

Factors	Study data
Fluid status	Cumulative fluid balance (ins and outs) is not associated with ventilator weaning duration or extubation outcomes in critically ill children.10 Studies of children with multiorgan failure showed survival may be associated with less fluid overload in the setting of continuous renal replacement therapy and that children placed on extracorporeal life support for respiratory failure showed increased survival with fluid removal and return to dry body weight.2,3
Sedative regimen	Choice of sedative regimen and sedation level can affect ventilator weaning and readiness for extubation. Excessive sedation may depress respiratory drive while inadequate sedation can lead to agitation and thrashing movements, which can result in trauma from the tracheal tube.2,9 Increased use of sedative drugs in the first 24 h of weaning predicts failure of extubation.2,9 Targeting a state behavioural scale score of 0, where a patient is awake and able to be calm helps optimise sedation level.11
Pulmonary hypertension	Pulmonary hypertension affects patient oxygenation. Because oxygen therapy and mechanical ventilation are cornerstones of the management of pulmonary hypertension, hesitance to wean such support can significantly affect readiness for extubation especially when direct measures of pulmonary arterial pressure or pulmonary vascular resistance are unavailable.2,12
Diaphragmatic function	Infants and young children have less well developed accessory respiratory muscles and weaker diaphragmatic function at baseline, which may relate to their longer weaning times.2,13 Significant diaphragm atrophy and a decreased diaphragmatic thickening fraction (DTF) measured by point-of-care ultrasound have been observed within 24 h of mechanical ventilation. DTF of <10% and the recovery of diaphragmatic thickness predicts successful weaning from mechanical ventilation.14,15
Airway steroids	Using corticosteroids to prevent (or treat) postextubation stridor has not proven effective for neonates or children; however, there have been findings suggesting benefit particularly for high-risk children or neonates.16 Of note, multiple doses of corticosteroids begun 12–24 h before extubation do appear beneficial for adult patients with a high likelihood of postextubation stridor.16
Disease reversibility and chronicity	Patients with rapidly reversible respiratory disease (like RSV bronchiolitis) as the cause of their respiratory failure and need for mechanical ventilation wean quicker and extubate faster than patients with slowly reversible disease (such as pneumonia, acute respiratory distress syndrome [ARDS]) or patients with chronic issues (such as genetic disorders/syndromes, chronic respiratory failure, chronic neuromuscular disease, chronic noninvasive ventilation, and prolonged steroid exposure).2,3,15,17
Cardiac function	Patients requiring prolonged ventilation in cardiac intensive care units (CICUs) are more likely to suffer from severe cardiovascular dysfunction compared with patients in a general PICU.18
Nutritional status	A study of ventilator-induced diaphragmatic atrophy in mechanically ventilated PICU patients found that optimal nutritional intake during mechanical ventilation averaged 54% and 53% of recommended calories and proteins. There may be an association between nutritional status and diaphragmatic atrophy and dysfunction in critically ill children who are mechanically ventilated.15

Weaning techniques

Although weaning protocols have been shown to result in faster weaning in adults, this has not been found to be true for most children.^2,9 For children who require ventilatory support for >24–48 h, weaning protocols may be beneficial. The most common method of weaning is a gradual decrease of ventilator support. This is performed by decreasing the ventilatory frequency, decreasing the inspiratory pressure or tidal volume depending on the mode of ventilation, or both. Pressure support (PS) is often combined with intermittent mandatory ventilation during weaning. The patient's trachea is extubated when the individual is noted to be performing a higher (and adequate) work of breathing compared with the ventilator support.³

‘Sprint’ weaning is another approach to weaning that involves alternating between total ventilator support and spontaneous breathing with graded levels of assistance. The premise behind this approach is to gradually train the respiratory muscles to take on the full work of breathing; however, there is no evidence to support this theory or approach.³ Sprinting is not commonly used in the PICU.

The mode of ventilation has not been found to affect the speed of weaning. An RCT to evaluate whether a volume support weaning protocol using continuous automated adjustment of PS by the ventilator (VSV) is superior to manual adjustment of PS by clinicians (PSV) found no significant difference in duration of weaning among PSV, VSV and no protocol.^2,9

Extubation readiness testing

An extubation readiness test (ERT) is a bundle of assessments (oxygen saturation, exhaled tidal volume and ventilatory frequency) used to assess a patient's readiness to be weaned from the ventilator. The patient is placed on minimal ventilator support and then observed for signs of respiratory distress or poor gas exchange.^2,3,7 Randolph and colleagues evaluated an ERT on patients who were receiving minimal sedation, breathing spontaneously and requiring minimal ventilatory support.⁹ They were placed on PS ventilation with fractional inspired oxygen (Fio₂) of 0.50 and PEEP of 5 cmH₂O for 2 h. The positive predictive value of this ERT for predicting successful extubation was found to be 87%.^8,9

The RESTORE (Randomized Evaluation of Sedation Titration for Respiratory Failure) study used a modified version of the Randolph ERT – they added oxygenation index (OI) ≤6 as a way to assess lower respiratory tract function.⁸ Oxygenation index measures arterial oxygenation (Pao₂) based on varying levels of ventilatory support (MAP and Fio₂; OI=[MAP × Fio₂]/Pao₂). The lower the oxygenation index (<25), the healthier the lungs, because it indicates that a higher Pao₂ is being achieved with lower ventilatory support. The RESTORE study found that passing the modified ERT increased the positive predictive value for successful extubation to 92–93% and was associated with extubation earlier in the day. The study also found that children likely to fail were younger, had underlying obstructive lung disease, more severe lung disease on the day of the ERT and were less likely to have received early neuromuscular blocking agents.⁸

Pressure support of 10 cmH₂O with PEEP of 5 cmH₂O is typically used in the PICU during ERT to overcome the increased resistance of breathing through a small tracheal tube during spontaneous ventilation. However, Khemani and colleagues¹⁹ demonstrated that, regardless of tube size (down to an ID of 3.0 mm), PS during ERT significantly underestimates the postextubation effort of breathing and creates a false sense of security about the patient's anticipated postextubation work of breathing. This happened even when patients' tracheas were extubated to noninvasive respiratory support. They concluded that PS should not be added to continuous positive airway pressure (CPAP) to overcome additional work of breathing imposed by the tracheal tube during ERT in children.¹⁹ These findings challenge the routine practice of using PS during ERT in children. The authors recommend performing ERT with CPAP alone for patients at high risk of failed extubation and high risk of difficult reintubation, to prevent underestimation of the postextubation work of breathing. A proposed algorithm is provided in Table 3.

Table 3.

Readiness for extubation algorithm.

Spontaneous breathing trials

Spontaneous breathing trials are used to identify patients who are likely to fail extubation. Unlike in the operating theatre where patients are routinely switched to pure spontaneous ventilation before extubation (without PS), patients in the PICU do not necessarily undergo a similar pure spontaneous ventilation trial as part of ERT. Spontaneous breathing trials in PICU are performed while a patient is still connected to the ventilator but at minimal settings (PS 5–10 cmH₂O, PEEP 1–5 cmH₂O) or, rarely, the patient can be disconnected from the ventilator to an independent oxygen source (T-piece). Spontaneous breathing trials can last from 15 min to 2 h or more depending on how well the patient tolerates the trial. Furthermore, SBT has been shown to have a high sensitivity (95%) and high positive predictive value (92%) for predicting extubation success in PICU patients.²⁰

There is a long-held belief that having infants and young children breathe spontaneously through small tracheal tubes is comparable with having them breathe through a straw, but this has not been supported by the literature.^21,22 Again, it is important to note that PS during SBT, like in ERT, can underestimate the work of breathing after extubation; therefore, SBT should be performed with a T-piece alone. A spontaneous breathing trial with a T-piece alone has been shown to be well tolerated by PICU patients – even infants and young children.²²

It is important to note that patients typically undergo SBTs only when their primary physician deems them ready. The critical care literature has clearly demonstrated that most patients who are ventilated mechanically (paediatric and adult) would be likely to pass an SBT well before weaning is even initiated.^2,22 It is therefore crucial to use predictive indices of successful extubation in combination with PICU provider clinical assessment of the patient's readiness to wean and extubate to determine the optimal time for ERT and SBT and avoid unnecessary ventilator days.

Leak test

A leak test involves listening for air escaping around the tracheal tube at low pressures (less than 20 cmH₂O) and is typically used to predict the presence of UAO from airway oedema. Leak tests are typically performed when the intubated patient is first admitted to the PICU; they are then repeated before weaning and before extubation. A leak test of <20 mmHg has been shown to be better at predicting stridor in children older than 7 yrs than those younger. Most PICU providers would recommend delaying extubation and prescribing steroids if there is no leak detected under 30 cmH₂O.²

Non-invasive ventilation after extubation in PICU

Non-invasive ventilation (NIV) is an important tool for improving respiratory function and preventing intubation in patients with respiratory failure.^23,24 NIV includes high-flow nasal cannula oxygen therapy, CPAP and bilevel positive airway pressure (BiPAP). The safety and efficacy of NIV in preventing intubation in PICU patients with respiratory failure has led providers to consider the role of NIV in the prevention and management of failed extubation. Two key strategies have been to extubate high-risk patients directly to NIV (also known as elective NIV), or to use NIV when extubation failure is detected (also known as rescue NIV).²³ Non-invasive ventilation has been shown to be effective in reducing the work of breathing in such patients (decreasing their accessory muscle use), and improving heart rate, ventilatory frequency, oxygen saturation, pH and Pco₂.^23,24 In a multicentre study, Bonora and colleagues demonstrated that success rates of rescue compared with elective NIV were 68.8% and 72.7%, respectively, and that the mortality rate was higher among patients in whom rescue NIV failed compared with those who were successful with NIV.²³ Patients with failed NIV were found to have higher paediatric risk of mortality (PRISM) scores, longer durations of PICU and hospital stay, and a higher prevalence of chronic underlying disease.²⁴ Studies have shown that NIV is a safe therapy. Complication rates have been found to be about 18% and include, mostly, mild discomfort, superficial skin lesions, and, rarely, skin necrosis, pneumothorax and aspiration.24, 25, 26

Failure of extubation in paediatric intensive care

The rate of extubation failure in the PICU ranges from 2.7% to 30% and usually occurs within 24–96 h after extubation.^1,3,7,8,27 Characteristics found to be associated with extubation failure include: age ≤24 months, genetic disorders, syndromic conditions, chronic respiratory disease, chronic neurological conditions, airway disorders (medical or surgical), chronic NIV, immediate postoperative reintubation on admission to PICU, and use of racemic adrenaline (epinephrine), helium–oxygen therapy (heliox) or NIV within 24 h of extubation.¹⁷ Baisch and colleagues found that PICU patients with failed extubation had longer hospital, PICU and ventilator courses and higher rates of tracheostomy placement. However, differences in hospital mortality between those with failed extubation and those whose tracheas were successfully extubated (11% vs 9%) was not statistically significant.²⁸

Negative inspiratory force has been shown to be an important factor in determining the risk of reintubation after a failed extubation. Khemani and colleagues found that risk factors for reintubation included lower pre-extubation maximum airway pressure during airway occlusion (NIF), longer duration of ventilation, UAO after extubation and high respiratory effort after extubation. They also found that nearly 35% of children had diminished respiratory muscle strength (NIF ≤30 cmH₂O) at the time of extubation, and were nearly three times more likely to be reintubated than those with preserved strength (NIF >30 cmH₂O; 14% vs 5.5%; P=0.006).²⁷

Complications after extubation

Upper airway obstruction

Although the cause of extubation failure tends to be multifactorial, UAO, especially from laryngeal oedema, is the most common reason for extubation failure in PICU patients. Although the majority of PICU patients receive systemic steroids around the time of extubation, about 6–13% of these patients will require reintubation. This can lead to increased duration of mechanical ventilation, morbidity and mortality, length of hospital stays and healthcare-related costs.^2,17,29 The most common symptoms of UAO are stridor and increased work of breathing.

Stridor

Given their smaller airway diameter, laryngeal oedema and postextubation stridor (PES) are more common in children as compared with adults.^16,19 The risk factors associated with developing PES include a young age (≤24 months old), weight <5 kg, prolonged intubation (more than 36–48 h), reintubation and underlying respiratory or neurological disease.^4,16 Systemic steroids, in particular dexamethasone, are often used to treat PES as they are powerful anti-inflammatory agents that may reduce the risk of laryngeal oedema. However, the effectiveness of steroids in preventing PES in children has not been adequately demonstrated.¹⁶ In addition to steroids, racemic adrenaline can be used to treat PES, especially when there is PES at rest that is associated with increased work of breathing.

Historically, the use of cuffed tracheal tubes in children was not recommended because of the increased risk of PES. However, with the advent of high-volume/low-pressure cuffed tracheal tubes, which have not been found to lead to clinically significant PES, the use of cuffed tubes in children has increased. Practically, the use of cuffed tubes leads to much less chance of requiring tube exchange or repeat airway manoeuvres that would put the patient at greater risk of laryngeal trauma, oedema, or both.³⁰

Tracheal stenosis

Tracheal or subglottic stenosis (SGS) develops after prolonged intubation from inflammation and subsequent scarring around the tracheal tube. The incidence of SGS has been reported to range from 0.2% to 11%.³¹ Duration of intubation and the need for additional doses of sedative medications, indicating a patient that is difficult to sedate, have been found to be key factors for the development of SGS. Children who developed SGS were noted to be undersedated more often than children who did not develop SGS.³¹

Conclusions

Few evidence-based published protocols exist to guide ventilator weaning and extubation in critically ill children. Using predictive indices in combination with ERTs and clinical assessment of the patient's readiness to wean and extubate is likely to improve the chances of successful extubation. Despite objective criteria that predict successful extubation, the risk of extubation failure is not eliminated. It is important to understand the complications associated with extubation in PICU patients in order to try to minimise their occurrence.

Declaration of interests

The authors declare that they have no conflicts of interest.

Biographies

Chinyere Egbuta MD is a paediatric anaesthetist and paediatric critical care physician with a passion for clinical teaching and medical education. She is the airway management team lead for the MSICU at Boston Children's Hospital and the co-director of the Pediatric Difficult Airway Management Course at Harvard Medical School.

Faye Evans MD is a paediatric anaesthetist passionately involved in paediatric anaesthesia education to improve surgical care for children both in the USA and in low- and middle-income countries. She helped develop the Safer Anaesthesia From Education (SAFE) paediatrics course, is the paediatric anaesthesia subsection editor for Anaesthesia tutorial of the week, and is the current chair of education for the World Federation of Societies of Anaesthesiologists.

Matrix codes: 1C02, 2C01, 3D00

MCQs

The associated MCQs (to support CME/CPD activity) are accessible at www.bjaed.org/cme/home for subscribers to BJA Education.