Respiratory distress syndrome is a disease of surfactant deficiency. The link between the surface properties of lung extracts and hyaline membrane disease was drawn by Avery and Mead (1959), and two decades later Fujiwara et al. (1980) showed that instilling surfactant into the airways of affected infants improved gas exchange. For the next quarter of a century surfactant and the endotracheal tube travelled together: the infant was intubated, given surfactant and ventilated. Each part of that package has since been questioned. Trials of nasal continuous positive airway pressure (CPAP) from birth showed that many very preterm infants could be managed without routine intubation (Morley et al., 2008; SUPPORT Study Group, 2010), and professional guidance now treats CPAP as the starting point for respiratory support (Committee on Fetus and Newborn, 2014).

Starting on CPAP creates a new problem. An infant with genuine surfactant deficiency who is left on CPAP alone may deteriorate, and intubating that infant late forfeits the advantage of early treatment. Less invasive surfactant administration (LISA), also described as minimally invasive surfactant therapy (MIST), is an attempt to resolve the dilemma by delivering surfactant through a thin catheter while the infant breathes spontaneously on CPAP. The technique has been studied chiefly in infants born before 32 weeks. This review is concerned with those born a little later, at 32 and 33 completed weeks, who are conventionally described as moderately preterm. They are far more numerous than extremely preterm infants, they are often cared for outside tertiary centres, and decisions about their respiratory care are made with less direct evidence than is commonly assumed.

This is a narrative review of the published literature. It draws on the randomised trials of thin-catheter surfactant, the principal meta-analyses, observational reports that bear on technique and safety, and the current European consensus guideline. No formal search protocol was registered and no pooled estimates are calculated here; where a quantitative synthesis is referred to, it is the published one. The trials are discussed by the gestational age of the infants they enrolled, because the central question is one of applicability. Findings are described in words and readers are referred to the original reports for effect sizes. Figure 1 places the main developments in sequence.

The first step away from prolonged ventilation was the approach reported by Verder et al. (1994), in which infants on nasal CPAP were briefly intubated, given surfactant and extubated again. Known as INSURE (intubate, surfactant, extubate), it reduced the need for subsequent mechanical ventilation when compared with later selective surfactant and continued ventilation (Stevens et al., 2007). INSURE still requires laryngoscopy, an endotracheal tube, usually premedication, and a period of positive pressure inflations, and a proportion of infants cannot be extubated promptly.

Kribs et al. (2007) in Cologne described a different idea: a fine feeding tube, guided through the vocal cords with Magill forceps under direct laryngoscopy, through which surfactant is given slowly while CPAP continues and the infant breathes. Dargaville et al. (2011) in Hobart developed a variant using a semi-rigid vascular catheter that can be passed without forceps, and reported its use in infants of 25 to 32 weeks (Dargaville et al., 2013). The two methods differ in equipment and in habits of premedication, but share the essential features: spontaneous breathing is preserved, the glottis is not held open by a tube, and no positive pressure breaths are delivered. Purpose-made catheters have since become available, and the practical aspects of the procedure have been reviewed by Herting et al. (2019).

Table 1 summarises the main randomised trials and the Cochrane review. They fall into three groups according to the infants enrolled, and the pattern of results differs between them in ways that matter for the moderately preterm infant.

Three features of the moderately preterm infant change how the trial results should be read. The first is baseline risk. Respiratory distress syndrome becomes less frequent as gestation advances, and antenatal corticosteroids, given to women at risk of preterm birth, reduce it further (Roberts et al., 2017). Most infants of 32 to 33 weeks who need support settle on CPAP alone. Bronchopulmonary dysplasia, defined by continuing need for oxygen or respiratory support at 36 weeks of postmenstrual age (Jobe and Bancalari, 2001), is uncommon in this group. A treatment whose headline benefit in smaller infants is a reduction in bronchopulmonary dysplasia cannot be expected to deliver the same benefit where the outcome seldom occurs.

The second feature is that the consequences of CPAP failure are different in kind. In a large cohort of infants managed initially on CPAP, Dargaville et al. (2016) found that failure in those of 29 to 32 weeks was associated with a higher risk of pneumothorax and a longer course of respiratory support, even though mortality was low. For a moderately preterm infant in a unit without facilities for prolonged ventilation, failure may also mean transfer to another hospital and separation from the mother. These are the outcomes on which thin-catheter surfactant could plausibly act, and they are the ones least often chosen as primary endpoints.

The third feature is the infant. A baby of 33 weeks is larger, more vigorous and more likely to resist laryngoscopy than one of 27 weeks. The procedure may be technically easier in terms of anatomy and harder in terms of tolerance. This bears directly on the question of sedation, discussed below, and on how many attempts are needed, which trial reports record inconsistently.

The evidence base has an uneven shape. It is deepest between 25 and 28 weeks, thinner below that, and thinnest above 32 weeks, where the number of infants who could be affected by a change in practice is greatest. A trial confined to moderately preterm infants would need outcomes suited to them: failure of non-invasive support, pneumothorax, days of respiratory support, time to full enteral feeding, transfer to a higher level of care and length of stay. Masking with a sham procedure, as in OPTIMIST-A, is feasible and would guard against the bias that affects decisions to intubate in open trials.

Several secondary questions follow. The oxygen threshold for treatment should be tested and not assumed. Sedation regimens should be compared with procedural success and the need for rescue ventilation as co-primary outcomes. Training deserves study in its own right, since the technique was developed by teams who perform it often, and its safety in hospitals that treat a handful of eligible infants each year is unknown. Video laryngoscopy may shorten the learning curve, and simulation-based training programmes have been described, but their effect on first-attempt success in practice has yet to be measured. Long-term respiratory and neurodevelopmental follow-up is available for very preterm cohorts and absent for moderately preterm ones.

Surfactant given through a thin catheter to a spontaneously breathing infant reduces exposure to mechanical ventilation, and in very preterm infants it is associated with less bronchopulmonary dysplasia among survivors. These findings justify its place in current guidance. For infants born at 32 and 33 weeks the case rests on one small trial, on subgroups and on extrapolation. The likely benefits are fewer intubations, fewer air leaks and fewer transfers, and the likely costs are a laryngoscopy that some infants would not have needed and a sedation question without an agreed answer. Units that extend the technique to moderately preterm infants are acting reasonably on indirect evidence, and should audit their thresholds, their procedural success and their failures so that the direct evidence can be built.

To be declared by the author(s) before publication.

To be declared by the author(s) before publication.