Bronchopulmonary dysplasia (BPD), chronic oxygen dependency, is a common adverse outcome of premature birth, particularly for those born before 29 weeks of gestation.1 Unfortunately, affected infants may require supplemental oxygen at home for many months and frequent readmissions to hospital in the first 2 years. Recurrent, troublesome respiratory symptoms and lung function abnormalities are common even at school age and early adulthood in those who had severe BPD.

BPD has a multifactorial aetiology; infants may be predisposed by very premature birth, a family history, the development of respiratory distress syndrome and/or abnormal antenatal lung growth and then exposed to such insults as sepsis and prolonged mechanical ventilation.2 It is, therefore, desirable to wean low birthweight (LBW) infants from ventilation and extubate them as soon as possible. It is important, however, not to undertake inappropriate weaning which may result in an increased work of breathing, or to extubate prematurely as this may result in an acute deterioration and need for reintubation.

Extubation in its most simplistic sense is “pulling out the endotracheal tube” but clearly there is much more to achieving successful extubation! Indeed, key to reducing difficult extubation is identification of optimal respiratory support strategies before and after extubation, efficacious adjunctive treatments to hasten weaning and facilitate extubation and accurate predictors of weaning/extubation. In addition, it is important to determine the most appropriate level of expertise necessary to oversee the process. The aim of this review is to assess the literature critically to determine if any of those crucial goals have been achieved and are successful specifically in LBW infants at high risk for failure of weaning and extubation.

RESPIRATORY SUPPORT STRATEGIES BEFORE AND AFTER EXTUBATION

Patient-triggered ventilation

Patient-triggered ventilation delivered as assist control ventilation (ACV) or synchronised intermittent mandatory ventilation (SIMV) shortens the duration of ventilation in infants recovering from respiratory distress, but not in those with acute respiratory distress.3 In randomised trials, ACV compared with SIMV was associated with a significantly shorter weaning time when the SIMV rate was reduced below 20 bpm; the explanation being that if the number of spontaneous breaths supported by mechanical inflations is less than a critical level, the work of breathing is increased. In a recent randomised trial,4 pressure support with SIMV rather than SIMV alone in infants who required mechanical ventilation during the first postnatal week was associated with a shorter duration of supplemental oxygen requirement in infants with birthweight between 700 and 1000 g. These results emphasise that weaning is best accomplished by triggered techniques which support every spontaneous breath; whether ACV or pressure-support ventilation is better requires testing, particularly in high-risk groups. During proportional assist ventilation the level of elastic and unloading can be altered to influence the infant’s work of breathing, whether this may facilitate weaning of chronically ventilated LBW infants with disuse muscle atrophy merits testing.

Volume-targeted ventilation

During volume-targeted ventilation (VTV) a preset tidal volume is delivered; meta-analysis of the results of four randomised trials5 demonstrated that VTV was associated with a significant reduction in the duration of ventilation. The trials, however, were of small sample size, and different ventilators, which varied in performance,6 were used. Randomised trials performed subsequently have demonstrated short-term advantages, but both studies were of small sample size and inadequately powered to detect whether VTV facilitated weaning and extubation.7 8

Nasally delivered modes

Continuous positive airways pressure (CPAP)

Nasal CPAP after extubation reduces the incidence of adverse clinical incidents (apnoea, respiratory acidosis and increased oxygen requirements) which lead to the need for additional respiratory support (relative risk (RR) = 0.62, 95% confidence interval (CI) 0.49 to 0.77).9 The efficacy of nasal CPAP, however, may depend on the delivery technique as, in a randomised trial of 162 extremely LBW infants, CPAP delivered by the Infant Flow device rather than by a ventilator and nasal prongs was associated with a shorter requirement for supplementary oxygen, although not significantly greater extubation success.10 In contrast, in another study (n = 87), 24% of those randomised to binasal prongs reached predetermined failure criteria in the 7 days after extubation compared with 57% of those randomised to a single nasal prong.11

There has, however, been concern that use of dual nasal prongs may be associated with a high incidence of nasal trauma. Randomised trials have not demonstrated that facial/nasal trauma differs according to the CPAP delivery technique, but prolonged CPAP is significantly related to the occurrence of trauma, hence LBW infants should be weaned from CPAP in a timely fashion.12 Two methods of weaning CPAP have become popular, one by reducing pressure and the other by decreasing the time spent on CPAP each day (so called “CPAP holidays”). Weaning by pressure is more efficacious, and in a randomised trial it was associated with significantly more weaning success, both in infants of gestational ages between 23 and 31 weeks and in those born before 27 weeks of gestation.13

Ventilatory modes

Synchronised nasal intermittent mandatory ventilation (NIMV) may be helpful in infants who fail to be maintained on nasal CPAP alone after extubation, as in randomised trials NIMV compared with CPAP increased successful extubation (RR = 0.21, 95% CI 0.10 to 0.45).14 NIMV, however, should be used with caution as the studies included in the meta-analysis were of small sample size and it is not possible to conclude confidently that nasal ventilatory modes do not increase gastrointestinal problems, which were reported in early studies.

ADJUNCTIVE TREATMENTS TO HASTEN WEANING AND FACILITATE EXTUBATION

Chest physiotherapy

Four randomised or quasi-randomised trials have been examined in a Cochrane review15 demonstrating that active chest physiotherapy did not significantly reduce the rate of post-extubation lobar collapse. Overall, there was a reduction in reintubation (RR = 0.32, 95% CI 0.13 to 0.82), but this was achieved by giving physiotherapy at 1 to 2-hourly intervals. The reviewers highlighted the lack of consistency in the results across the trials, the small numbers included in the meta-analysis and the lack of safety data. On the currently available evidence, use of such physiotherapy should be individualised.

Methylxanthines

Treatment with methylxanthines facilitates extubation; meta-analysis of six randomised trials demonstrated a reduction in failure of extubation within 1 week (RR = 0.47). The number needed to treat with a methylxanthine to prevent one case of failed extubation was 3.7.16 Caffeine with its wider therapeutic margin is the preferred treatment. High-dose (20 mg/kg) compared with low-dose (5 mg/kg) caffeine appears to be more effective. In 234 infants born at <30 weeks of gestation, a high dose was associated with significantly less failure to extubate (15% vs 30%, p<0.01) and there were no significant differences in major adverse effects, mortality or neurodisability at 1 year between the two groups, although those receiving the higher dose took a longer time to regain their birthweight (14.8 vs 12.9 days, p<0.01).17 Results from any future trials need to be stratified by gestational age and postnatal age to determine if caffeine is equally effective in high-risk infants. There has been concern about any possible adverse effects of methylxanthine administration in prematurely born infants. The short- and long-term efficacy and safety of caffeine in birthweight of 500 to 1250 g has been examined in a multicentre trial.18 Infants randomised to caffeine in the first 10 days after birth had a lower incidence of BPD; the only adverse effect of caffeine administration was a temporary reduction in birthweight. Although, this study did not examine the use of caffeine in infants about to be extubated, its large sample size (n = 2006) and the lack of major adverse effects should help to reassure practitioners.

Doxapram

In the past, doxapram was given to infants who were difficult to wean as it encourages breathing by stimulating both peripheral and central chemoreceptors, and in prematurely born infants already receiving aminophylline it increased ventilation and respiratory drive. No significant differences in failed extubation or duration of ventilation, however, were demonstrated in two randomised trials and there was a trend for hypertension or irritability leading to cessation of treatment to be more common in infants treated with doxapram (RR = 3.21, 95% CI 0.53 to 19.43).19 This treatment cannot be recommended for the LBW newborn.

Corticosteroids

Systemic corticosteroids facilitate weaning regardless of postnatal age at administration—that is, when given in the first 96 hours significantly more infants were extubated both at 3 days (RR = 0.74, 95% CI 0.62 to 0.78) and 28 days (RR = 0.80, 95% CI 0.67 to 0.96) and when given either between 7 and 14 days or after 21 days extubation at 28 days was improved (RR = 0.62, 95% CI 0.51 to 0.78; RR = 0.54, 95% CI 0.32 to 0.92, respectively).20^–22 There are concerns, however, about the long-term side effects of corticosteroid treatment, and as a consequence the efficacy of low-dose dexamethasone (0.89 mg/kg over 10 days) has been explored in a randomised trial of infants ventilator dependent after the first week after birth and born before 28 weeks’ gestation or with a birthweight of <1 kg. Dexamethasone treatment was associated with more infants being successfully extubated by 10 days (60% vs 12%, p<0.01) and a shorter duration of ventilation (14 vs 21 days, p = 0.03).20 No short-term23 or longer-term24 adverse effects were reported, but only 70 of the 814 planned sample size were recruited; further studies are required to explore these encouraging results. An alternative approach to avoid the side effects of systemic steroids is to use inhaled steroids. Meta-analysis of seven randomised trials, one which included 253 infants, demonstrated that inhaled steroids administered for 1–4 weeks improved extubation rates (RR = 0.12, 95% CI 0.03 to 0.43)25; the optimal dosing schedule and method of delivery, however, need defining.

Meta-analysis of three randomised trials showed that intravenous dexamethasone given peri-extubation significantly reduced reintubation,26 but in the trial in which attempts were made to exclude infants at high risk of airway oedema, the incidence of extubation failure was zero and in all the trials side effects were reported. The authors then reasonably concluded that intravenous dexamethasone peri-extubation should be restricted to infants at increased risk of airway oedema and obstruction, such as those who have had repeated or prolonged intubation.26

PREDICTORS OF SUCCESSFUL WEANING/EXTUBATION

Clinical criteria used to determine whether a very LBW infant is ready to be extubated fail on about 30% of occasions. Successful weaning and extubation are dependent on effective respiratory drive, adequate respiratory muscle strength and the magnitude of the respiratory load. Examples of conditions in which there is abnormal respiratory load, inadequate respiratory drive and inadequate respiratory muscle strength are given in box 1; they are not mutually exclusive. Assessment of the individual components influencing weaning and extubation has yielded variable success and a better approach appears to be to determine the balance between them by determining the effectiveness of respiratory efforts. This can be achieved, for example, by measuring the minute ventilation and in one trial, infants who were randomised to extubation decided by a minute ventilation test rather than clinical judgment were extubated significantly sooner (8 vs 36 hours p = 0.04), although there were no significant differences in the extubation failure rate between the two groups.27 This strategy needs to be tested in other centres.

EXPERTISE

Computer-aided decision support tools have been developed to assist the clinician in the decision-making process for extubation. In the development of one support tool, clinical experts identified 51 potentially predictive variables and the 13 most predictive were used to develop an artificial neural network extubation model.28 It is important to determine how such tools perform in routine practice and who in the multidisciplinary team is most effective at managing weaning and extubation. In a randomised trial, the weaning time was shorter (1200 vs 3015 min, p = 0.049) and there were fewer backward steps (p = 0.019) when the process was managed by experienced nurses rather than specialist registrars.29 This study emphasises that efficacious weaning is best accomplished by staff who can devote sufficient time to supervising weaning.

CONCLUSION

The majority of studies of weaning and extubation have included infants who had respiratory distress syndrome, and future research needs to target specifically the chronically ventilated infant developing, and with, BPD. In such infants, who have a high respiratory load and poor respiratory muscle function and may have abnormal respiratory drive, weaning and extubation predictors may be most helpful. In addition, randomised trials are required to optimise extubation of infants with airway problems, such as subglottic stenosis or Pierre Robin syndrome. Such problems are rare and multicentre collaboration will be required.