Background
Although it is a life-saving intervention, mechanical ventilation is associated with many complications ( Box 24-1 ), making timely and safe weaning an important imperative. However, the process of discontinuing mechanical ventilation in newborn infants with significant pulmonary morbidity and those who are extremely premature remains a major challenge, made more complex by the variety of modes of respiratory support currently in use. There is no consensus about the most appropriate way to wean babies from mechanical ventilation, and their management remains largely subjective, depending on institutional practices and individuals’ training or preferences. Unfortunately, many important gaps in knowledge remain in the science of weaning from mechanical ventilation and assessment of extubation readiness in neonates, which results in significant variations in periextubation practices worldwide. In this chapter we provide a comprehensive review of this subject and some evidence-based recommendations to guide clinical practice for weaning from mechanical ventilation, assessment of extubation readiness, and postextubation management.
| Ventilator-Induced Lung Injury (Volutrauma) |
| Atelectasis |
| Overdistention |
| Bronchopulmonary dysplasia |
| Air-Leak Syndromes |
| Pulmonary interstitial emphysema |
| Pneumothorax |
| Pneumomediastinum |
| Pneumopericardium |
| Airway Trauma |
| Vocal cord injury |
| Subglottic stenosis |
| Subglottic cysts |
| Granulomas |
| Tracheobronchomalacia |
| Palatal deformities |
| Nasal septal defects |
| Endotracheal Tube Complications |
| Obstruction |
| Displacement |
| Accidental extubation |
| Infection |
| Ventilator-associated pneumonia |
| Late-onset sepsis |
| Cardiovascular |
| Decreased cardiac output |
| Neurologic |
| Hypocarbia (cerebral vasoconstriction) |
| Neurodevelopmental impairment |
Weaning from Ventilatory Support
Weaning from mechanical ventilation is the process of decreasing the amount of ventilatory support, with the patient gradually assuming a greater proportion of the overall work of ventilation. As mentioned in Chapter 15 , weaning and extubation at the earliest possible time are among the a priori goals of mechanical respiratory support. In addition to the obvious goal of reducing ventilator-induced lung injury, early weaning will reduce the risk of nosocomial sepsis, reduce patient discomfort and need for sedation, minimize the development of oral aversion with subsequent feeding difficulties, and facilitate parental bonding and developmentally appropriate care. Therefore, as soon as the patient’s condition stabilizes and the underlying respiratory disorder that led to the initiation of ventilation begins to improve, weaning should be initiated. This approach differs from that employed in the past (and persisting in some centers) when patients were heavily sedated or even paralyzed during the acute phase of the illness and weaning from mechanical ventilation did not begin until some arbitrary weaning criteria were met. In the early days of mechanical ventilation of newborn infants, there was a widespread practice of keeping infants on mechanical ventilation until they reached the arbitrary weight of 1 kg. This practice, long ago abandoned, was based on the assumption that small preterm infants would expend too much energy to breathe and were unlikely to remain extubated. This concept was in large part related to the practice of extubating preterm infants to an oxygen hood rather than to continuous positive airway pressure (CPAP). Today we know that with appropriate noninvasive support many extremely low birth-weight infants are able to be extubated within a few days of birth and thrive. Those who remain ventilator dependent beyond the first week of life have a much higher risk of bronchopulmonary dysplasia (BPD). A retrospective study of infants of ≤1000 g and ≤28 weeks demonstrated a seventeenfold increase in the risk of any BPD in infants ventilated for >7 days, compared to those extubated on days 1 to 3, with a 62% incidence of moderate or severe BPD in the babies extubated for the first time beyond 7 days of age.
Many different modes of invasive respiratory support are used in neonates, and the specific mechanics of weaning from mechanical ventilation are to a large extent a function of the mode of support in use. The basic types of mechanical ventilation are (1) pressure-controlled ventilation, (2) volume-controlled or volume-targeted ventilation, and (3) high-frequency ventilation (HFV). The basic process of weaning for the various pressure-controlled synchronized modes is described in Chapter 18 , and the basic steps for pressure-controlled ventilation and HFV are summarized in Box 24-2 . A detailed discussion of weaning using volume-targeted ventilation is available in Chapter 20 . Available data support the preferential use of (1) any mode of synchronized ventilation over unsynchronized intermittent mandatory ventilation (IMV), (2) modes that support each spontaneous breath over synchronized IMV (SIMV), and (3) volume-targeted over pressure-controlled ventilation. Some important concepts that should be kept in mind to facilitate weaning include the following: (1) Weaning too slowly may be more dangerous than weaning too fast, as it may result in excessive lung injury and hypocarbia. (2) Weaning should be attempted throughout the day, not just during rounds. (3) When gas exchange is satisfactory and the work of breathing is not excessive, weaning should be attempted. (4) Volume-targeted ventilation effectively addresses concepts 1 to 3 by lowering inflation pressure in real time in response to improving lung mechanics and patient effort. Ventilatory settings at which to consider extubation readiness in infants 2 weeks of age or younger are provided in Box 24-3 . These reflect general consensus and are based on values used as extubation criteria in many randomized controlled trials (RCTs). However, they are not based on prospectively obtained trial data.
| Conventional Ventilation (AC, SIMV, PSV) |
|
| Volume Ventilation |
|
| High-Frequency Oscillatory Ventilation |
|
| High-Frequency Jet Ventilation |
|
Weaning from Pressure-Limited Ventilation
With SIMV, weaning is accomplished by reducing both peak inflation pressure (PIP) and the ventilator rate. The rate should not be reduced much until PIP has been reduced to relatively low values (<20 cm H 2 O) that indicate improved lung compliance. Lowering the set rate while the lungs are still quite stiff is likely to impose a high work of breathing and result in rapid shallow spontaneous breathing requiring an excessively large tidal volume (V T ) for the low-rate SIMV inflations to maintain adequate alveolar minute ventilation. In small preterm infants, it is advisable to add pressure support (PS) when SIMV rate is reduced below 30/minute. The combination of SIMV + PS resulted in more rapid weaning from mechanical respiratory support than SIMV alone in extremely low birth-weight (ELBW) infants and significantly lower work of breathing. If SIMV is used without PS, the rate should not be reduced below 15 inflations/minute. There are no studies to inform the best method of weaning from SIMV + PS. It seems reasonable to reduce the SIMV rate gradually to 10 while also reducing the PIP as necessary to avoid excessive V T and maintain PS at a level sufficient to achieve acceptable V T for the spontaneous breaths, typically 4-5 ml/kg.
With assist control and PS ventilation as a stand-alone mode, the infant controls the ventilator rate; therefore, lowering the set rate, which acts as a backup only in the case of apnea, has no real impact on reducing ventilator support. Weaning is accomplished by lowering the PIP and gradually transferring the work of breathing to the infant. For a brief period of several hours just prior to extubation, it may be reasonable to reduce the backup rate to 20 inflations/minute to better recognize any inconsistent respiratory effort that may have been masked by a higher backup rate.
Weaning from High-Frequency Ventilation
Many clinicians are more comfortable changing from HFV to conventional modes prior to extubation, but extubation directly from both jet and oscillatory ventilation is not only possible, it may even be desirable. Clark et al. reported that infants who remained on high-frequency oscillatory ventilation (HFOV) until extubation had a lower incidence of BPD than those ventilated conventionally, but infants who were changed to SIMV after 72 hours of HFOV did not seem to benefit equally. Similarly, the large HFOV trial by Courtney et al., which required infants to remain on HFOV for 14 days or until extubation, reported a lower rate of BPD and shorter duration of ventilation, compared with conventional ventilation, whereas a similar study published in the same issue of the New England Journal of Medicine , which allowed early crossover from HFOV to conventional ventilation, showed no such benefits. The way HFOV support is reduced is based on empiric data and experience, with few experimental data to guide the clinician. In general, both pressure amplitude and mean airway pressure are reduced progressively as tolerated, the former to reduce minute ventilation, the latter to avoid overexpansion as lung compliance and oxygenation improve. There is no consensus regarding “extubatable” settings during HFV, but in general, extubation is considered when mean airway pressure is around 8 cm H 2 O with FiO 2 less than 0.30. Frequency is not reduced as a means of reducing support. With most HFOV devices, delivered V T increases as frequency decreases, so that reducing ventilator frequency has the opposite effect compared to that on conventional ventilation. Some clinicians increase the HFOV frequency as an indirect means of reducing V T , but this approach of making ventilation more inefficient seems counterintuitive and not specifically supported by evidence.
Weaning from high-frequency jet ventilation more closely parallels conventional ventilation, with stepwise reduction in peak and mean pressures. The primary means of reducing support is reduction in pressure amplitude, which is the difference between PIP and positive end-expiratory pressure (PEEP). With both jet and oscillatory ventilation, when support is reduced enough to allow mild respiratory acidosis, spontaneous breathing will be observed as the infant begins to take over more of the respiratory effort. When there is good spontaneous effort and the settings are judged to be sufficiently low, extubation should be attempted. There are no studies of extubation readiness for infants on HFV.
General Strategies to Facilitate Weaning
Permissive Hypercarbia
Permissive hypercarbia is a ventilatory strategy that accepts higher than normal PaCO 2 levels (between 45 and 65 mm Hg), as long as the pH is ≥7.20, while using lower rates and/or V T s. This strategy may reduce injury to the developing lung through a variety of mechanisms, which include more efficient CO 2 removal, better ventilation–perfusion matching, stabilization of or increase in respiratory drive facilitating weaning, and improvement in cardiac output. Permissive hypercarbia appears to be an effective strategy to allow continued use of noninvasive support and avoidance of mechanical ventilation. This approach, though widely practiced, has not actually been directly evaluated as an isolated intervention in randomized trials. Nonetheless, it is generally accepted as appropriate and was an important part of several large trials comparing delivery room management. It remains unclear if permissive hypercarbia is effective in facilitating weaning. Mariani et al. showed shorter duration of mechanical ventilation with a trend to less BPD in a single-center randomized pilot trial comparing target PaCO 2 of 35 to 45 mm Hg to 45 to 55 mm Hg for the first 96 hours of life. A subsequent multicenter trial targeting an even higher level of hypercarbia failed to replicate these findings, possibly because unlike in the pilot study, a clear separation in between the two arms could not be achieved. Furthermore, the trial was stopped early after enrollment of 220 infants because of increased complications associated with the other intervention involved in this 2 × 2 factorial design trial. There was a nonsignificant trend to less BPD at 36 weeks (63% vs 68%), and fewer infants required mechanical ventilation at 36 weeks’ postmenstrual age.
The largest trial that incorporated permissive hypercarbia as part of a strategy to avoid mechanical ventilation and hasten extubation before and during the first14 days of mechanical ventilation was the Surfactant, Positive Pressure, and Oxygenation Randomized Trial (SUPPORT). The study compared a strategy of routine intubation and surfactant administration in the control arm to primary use of noninvasive support in the delivery room and permissive hypercarbia. The target P co 2 was >65 mm Hg with a pH >7.20 in the group assigned to noninvasive support and <50 mm Hg and a pH >7.30 in the routine intubation group. There was no difference in the primary outcome of BPD/death (47.8% vs 51.0%), but subgroup and secondary analysis showed decreased mortality in infants with gestational age (GA) between 24 and 25 weeks (23.9% vs 32.1%, p = 0.03), lower rates of mechanical ventilation and surfactant supplementation, shorter duration of mechanical ventilation, and less use of postnatal corticosteroids for BPD. However, a secondary analysis of data from SUPPORT reported increased risk of adverse outcomes, such as severe intraventricular hemorrhage, and death, with both high PaCO 2 and fluctuating PaCO 2 . A single-center trial focusing solely on permissive hypercarbia enrolled 65 infants of between 23 and 28 weeks’ gestation who received mechanical ventilation within 6 hours of birth. Infants were randomized to a PaCO 2 target of either 55 to 65 or 35 to 45 mm Hg for the first week of life. The trial was stopped early after enrolling about one-third of the projected sample size. BPD or death occurred in 64% of the infants in the hypercarbia group and 59% of control infants ( p = NS). A concerning finding was that permissive hypercarbia was associated with trends toward higher mortality and higher incidence of neurodevelopmental impairment with a significantly increased combined outcome of mental impairment or death ( p < 0.05). The authors concluded that permissive hypercarbia, as performed in that study, did not improve clinical outcome and may actually be associated with a worse neurodevelopmental outcome.
As of this writing, the most recent randomized trial was the Permissive Hypercapnia in Extremely Low Birthweight Infants (PHELBI) trial, which compared permissive hypercarbia and more traditional P co 2 targets. The high-target group aimed at PaCO 2 values of 55 to 65 mm Hg on postnatal days 1 to 3, 60 to 70 mm Hg on days 4 to 6, and 65 to 75 mm Hg on days 7 to 14; the control group target was P co 2 40 to 50 mm Hg on days 1 to 3, 45 to 55 mm Hg on days 4 to 6, and 50 to 60 mm Hg on days 7 to 14. The trial was stopped after interim analysis when it became evident that the probability of showing a benefit of the intervention became vanishingly small. Ultimately, 359 infants of 400- to 1000-g birth weight were analyzed. The rate of the combined outcome of BPD or death in the permissive hypercarbia group versus control (36% vs 30%; p = 0.18), death (14% vs 11%; p = 0.32), and grade III to IV intraventricular hemorrhage (15% vs 12%; p = 0.30) showed nonsignificant trends favoring the control group. As with the previous randomized trials, the PaCO 2 in the hypercarbia group, although higher than in the control group, was generally below the target range, despite significantly lower ventilator settings. This highlights the practical limitation of targeting a significant respiratory acidosis in infants who are not heavily sedated or paralyzed—the baby’s own respiratory drive will cause increased spontaneous respiratory effort, which may lower the PaCO 2 below the target range but also cause tachypnea, retractions, and agitation, necessitating either heavy sedation, which is associated with substantial adverse effects, or an increase in respiratory support. The above studies, taken together with observational studies that suggested that PaCO 2 values >60 to 65 mm Hg during the first few days of life were associated with increased risk of severe intraventricular hemorrhage (IVH), especially when rapid changes occur, indicate that marked permissive hypercarbia is not beneficial and may in fact be harmful. Mild degrees of permissive hypercarbia similar to the control arm of the PHELBI trial are based on sound physiologic principles and are generally accepted as safe and effective in reducing the need for mechanical ventilation.
Permissive Hypoxemia
Less aggressive oxygenation targets became incorporated into neonatal care without the benefit of large randomized trials in an effort to reduce oxidative stress and adverse effects on the lungs and other organs. Early observational studies suggested that lower oxygen saturation targets were associated with less retinopathy of prematurity, less chronic lung disease, and faster weaning from mechanical ventilation. The safety of this permissive hypoxemia approach was called into question by a series of studies attempting to define the optimal target saturation as measured by pulse oximetry (Sp O 2 ). Five trials with similar designs including a prespecified composite outcome of death/disability at 18 to 24 months’ corrected GA were conducted to compare two different ranges of oxygen saturation (high, 91% to 95%, and low, 85% to 89%). Two systematic analyses of these trials have been published. An increased relative risk for mortality and necrotizing enterocolitis and a reduced risk for severe retinopathy of prematurity was noted in the low, compared to the high, oxygen saturation target group. There were no differences regarding BPD (physiologic definition), brain injury, or patent ductus arteriosus between the groups. Another systematic review found no difference in the prespecified composite outcome of death/disability (risk ratio [RR], 1.02; 95% CI, 0.92 to 1.14) or mortality before 24 months (RR, 1.13; 95% CI, 0.97 to 1.33). A significant increase in hospital mortality was found in the restricted oxygen group (RR, 1.18; 95% CI, 1.03 to 1.36). However, even though many questions remain unanswered, it has been suggested that SpO 2 targets should be raised between 90% and 95% in infants with GA <28 weeks until 36 weeks’ postmenstrual age. Permissive hypoxemia therefore does not appear to be a useful strategy to accelerate weaning from mechanical ventilation.
Weaning Protocols
The availability of a variety of ventilators and ventilatory modes and the diverse backgrounds and training of neonatal practitioners contribute to highly variable approaches to mechanical ventilation practices with significant intra- and intercenter differences in outcome. Such variability can be harmful. The use of ventilation protocols driven by health care providers is an effective way of decreasing unnecessary variations in practice. Indeed, in adults the use of ventilation protocols has been demonstrated to improve weaning from mechanical ventilation, with no complications and decreasing costs. Several studies involving patients in the pediatric intensive care setting similarly indicated benefits of protocol-driven weaning. In neonates, especially in the preterm population, the highest level of evidence is lacking and weaning practices remain very physician-dependent, but some evidence in favor of weaning protocols in newborn infants is available. In a retrospective study, Hermeto et al. demonstrated a significant improvement in short-term respiratory outcomes in preterm infants after a respiratory therapy-driven protocol was implemented. Despite the lack of solid evidence, a Canadian survey showed that almost 40% of the neonatal intensive care units (NICUs) have already adopted the use of protocols for mechanical ventilation. A 2002 paper suggested that experienced NICU nurses are more effective at weaning infants from mechanical ventilation than physicians in training, adding another potential avenue to facilitate weaning.
Adjunctive Therapies
Adjunctive therapies are interventions designed to help preterm infants maintain adequate gas exchange and respiratory effort during the weaning and postextubation periods. Some of these therapies have been investigated by large RCTs, whereas for others there is still a glaring lack of evidence to support or refute their use. A more complete discussion of pharmacologic therapies can be found in Chapter 34 .
Caffeine
When given to infants receiving mechanical ventilation, caffeine was associated with faster weaning, within 2 to 7 days after initiation of the treatment. Caffeine also improved important outcomes such as extubation failure, number of days on invasive and noninvasive ventilation, and neurodevelopment at 18 to 22 months of age. The most appropriate dose is still not well defined but caffeine is commonly administered at 10 mg/kg/day of caffeine citrate once a day given 24 hours after a loading dose of 20 mg/kg. Higher doses have been shown to be beneficial in single-center studies. The most appropriate age at which caffeine should be initiated has also not been established. Caffeine is usually started soon after birth, but some centers prefer to initiate caffeine only around the time of extubation, either prophylactically or in infants with significant episodes of apnea after extubation. Furthermore, additional evidence is needed to better define when to safely discontinue caffeine for extremely preterm infants and possible long-term side effects.
Diuretics
High fluid intake and low weight loss during the first 10 days of life is associated with an increased risk of death or BPD in preterm infants. A meta-analysis on this subject concluded that restricted water intake reduces the risk of patent ductus arteriosus (PDA) and necrotizing enterocolitis with trends toward reduced risks of BPD, intracranial hemorrhage, and death. These findings naturally led clinicians to consider therapies to address fluid overload as a means of improving lung function and reducing the need for respiratory support. However, diuretics have been shown to be ineffective in the treatment of the acute phase of respiratory distress syndrome, even though it is clearly associated with fluid retention. Furosemide is the most commonly used diuretic in the newborn period. Furosemide decreases interstitial lung water and improves lung mechanics in the short term, and thus it is commonly used in infants who are difficult to wean from mechanical ventilation, especially when there is evidence of fluid overload. However, it is important to recognize the lack of objective evidence for its effectiveness in the long term. It is also important to note that furosemide stimulates the renal synthesis of prostaglandin E2, a potent dilator of the ductus arteriosus, and therefore should be used with caution within the first days of life.
Closure of Patent Ductus Arteriosus
Left-to-right shunting of blood through a widely patent ductus arteriosus can cause pulmonary edema and decreased lung compliance. PDA is associated with increased mortality and morbidity and BPD, and difficult weaning from mechanical ventilation in preterm infants is commonly attributed to a “hemodynamically significant PDA.” Treatment (pharmacologic and/or surgical) was routinely instituted to facilitate weaning. However, aggressive PDA treatment is a practice that has been questioned based on analyses of individual RCTs and meta-analyses. Pharmacologic ductal closure is effective in reducing the need for ductal ligation, but no long-term benefits of PDA treatment were documented. Both medical and surgical PDA closure have significant associated risks, and spontaneous PDA closure commonly occurs, supporting a more conservative approach. Thus, the role of treatment of PDA in facilitating weaning from mechanical ventilation remains unclear.
Avoidance of Routine Sedation
In preterm infants, a systematic review has evaluated the routine use of opioids in neonates receiving mechanical ventilation and found insufficient evidence to recommend it. Indeed, the use of morphine in this population has been reported as potentially harmful. In a multicenter randomized trial the use of additional analgesia with morphine was associated independently with increased rates of IVH and air leaks and longer duration of mechanical ventilation, nasal CPAP, and oxygen therapy. In an observational study the use of morphine given for preemptive analgesia or without ongoing, patient-specific measures of pain was associated with prolonged hospitalization. In this study, the implementation of a nursing-driven comfort protocol significantly reduced this type of morphine use. Morphine administration to ventilated preterm infants has also been associated with subtle neurobehavioral differences during childhood. Therefore, routine use of morphine analgesia/sedation cannot be recommended as it clearly impairs weaning from respiratory support.
Nutritional Support
Adequate nutritional support is critical to any patient receiving intensive care. Nutrition and lung growth and development are interdependent, with the former playing a critical role in the prevention and management of BPD. Most preterm infants with moderate or severe BPD experience growth failure predominantly due to suboptimal nutritional intake, which in turn can worsen BPD by further compromising lung repair and growth. Indeed, high rates of extrauterine growth restriction have been described in critically ill preterm infants. Therefore, adequate nutritional management of these infants should start immediately after birth to enhance ventilation weaning and lung repair and growth and ultimately minimize respiratory morbidity.
Chest Physiotherapy
Chest physiotherapy, such as percussion and vibration followed by suction, in infants receiving mechanical ventilation was assessed in a systematic review updated in 2009. A total of three trials involving 106 infants were included, and analysis identified insufficient evidence to adequately assess the efficacy and/or adverse effects of this therapy.
Systemic Corticosteroids
The updated Cochrane meta-analysis concluded that the use of early postnatal systemic corticosteroids (<8 days of age) at various doses facilitates weaning and decreases the risk of BPD in preterm infants but also increases the risk of short-term complications and long-term neurologic sequelae. Later use (beyond 7 days of age) also reduced BPD at 36 weeks, facilitated earlier weaning, and reduced mortality at 28 days and at 36 weeks. There was a trend toward a higher rate of cerebral palsy (CP), but this was offset by lower mortality. Following the initial reports of steroid association with increased rates of CP the use of corticosteroids dropped sharply, with a subsequent increase in the rates of BPD. In 2005, a meta-regression analysis of several RCTs demonstrated that the risk of death or CP varies with the a priori level of risk for BPD, with a negative relationship between these variables in the control group. An update of this analysis showed that the results were unchanged by the addition of more studies, but a slightly narrower confidence interval with a slightly greater statistical significance was observed. Based on this evidence, clinicians should use prediction equations or their own local data to identify the highest risk infants who are likely to have a net benefit from postnatal corticosteroid treatment with respect to survival free of CP. Most of the concerns about neurologic sequelae focus on dexamethasone. As of this writing, whether hydrocortisone administered between 14 and 28 days of age is effective in facilitating weaning and reducing the risk of BPD is being investigated in a large randomized trial conducted by the National Institute of Child Health and Human Development (NICHD) Neonatal Research Network.
Inhaled Corticosteroids
Inhaled steroids have long been used in the management of preterm infants on mechanical ventilation. Published surveys from 2014 and 2015 indicate a relatively high and variable use of inhaled steroids, despite the lack of evidence to support the routine use of inhaled steroids in infants receiving mechanical ventilation. A large multicenter trial in Europe (NEUROSIS) has finished patient enrollment, but final results have not been published as of this writing. While the apparent lack of efficacy of inhaled steroids may be a function of ineffective delivery systems, the use of inhaled steroids to facilitate weaning cannot be recommended, unless data from the European trial provide convincing evidence to the contrary.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
