Henrik Verder, Department of Paediatrics, Holbaek University Hospital, University of Copenhagen, Smedelundsgade 60, DK-4300 Holbaek, Denmark. Tel.: +45 59484200 | Fax: +45 59484209 | Email: firstname.lastname@example.org
The Scandinavian approach is an effective combined treatment for respiratory distress syndrome (RDS) and prevention of bronchopulmonary dysplasia (BPD). It is composed of many individual parts. Of significant importance is the early treatment with nasal continuous positive airway pressure (nCPAP) and surfactant treatment. The approach may be supplemented with caffeine citrate and non-invasive positive pressure ventilation for apnoea. The low incidence of BPD seen as a consequence of the treatment strategy is mainly due to a reduced need for mechanical ventilation (MV).
Conclusion: Early-postnatal treatment with nCPAP and surfactant decreases the severity and mortality of RDS and BPD. This is mainly due to a diminished use of MV in the first days of life.
Respiratory distress syndrome (RDS) is the single most important cause of mortality and morbidity in preterm infants. During the last 50 years, significant and conclusive progress has been made in the understanding of the aetiology and pathophysiology of the syndrome as well as the treatment, while the optimal treatment regimen is still under discussion. Oxygen should be restricted (1) and the oxygen tension should be carefully monitored (2). The introduction of nasal continuous positive airway pressure (nCPAP) more than halved the mortality rate due to RDS (3,4) and the therapeutic possibilities were further enhanced by the development of mechanical ventilation (MV) (5). The introduction of prenatal steroid prophylaxis (6) has significantly improved outcome. Since 1980, surfactant has been used with success in the treatment (7) and the effective combined treatment of nCPAP and surfactant, the Scandinavian INtubation SURfactant Extubation procedure (INSURE) (8,9), in a retrospective analysis was found to reduce the mortality significantly in comparison with nCPAP only (10).
In many parts of Scandinavia, the common practice for more than 20 years has been to commence treatment of RDS and other types of respiratory distress with nCPAP. For many years, this was in sharp contrast with practice in most other western countries. The battle between the neonatologists mainly in North America and the UK which prefer ‘safe’ MV and the Danes, Swedes and Colombia neonatologists which prefer nCPAP is described by Lagercrantz (11). When Kamper et al. (12) published his paper in 1993 about treatment of very low birth weight (VLBW) infants with nCPAP, it was attacked by Roberton (13) who asked: ‘Does CPAP work when it really matters’? Today there is no doubt that nCPAP has become an indispensable part of the primary treatment of RDS (14). The focus has changed and the main question now is: ‘How can we minimize the use of MV’? Thus, in the last years, MV has shown to be potentially harmful to the lung tissue due to the risk of developing bronchopulmonary dysplasia (BPD) (15–17). Furthermore, Björklund et al. (18) have showed that it may be better for immature infants to open up their lungs themselves at birth, if necessary supported by mask CPAP. Besides stabilization of the respiration, establishment of nCPAP soon after birth also gives the infant possibility to remain awake and improves the opportunities for attachment between mother and infant, for breast feeding, care and nursing and observation of the newborn (14).
Nasal CPAP is effective but which type of CPAP should be used? Which surfactant should be used? What about the timing of the surfactant treatment and how should the surfactant be administered to be most effective? When should the treatment be combined with assisted ventilation and can the risk of BPD be reduced? These questions will be dealt with in this article.
Gregory et al. (19) introduced CPAP into modern neonatology, and this soon replaced the more complicated ventilator therapy in a significant number of infants with RDS. Originally, greeted as a significant leap forward, CPAP was overtaken in the late 1970s by the development in ventilator technology combined with a wish to secure ‘cardio-respiratory control’ (13) using early MV with surfactant prophylaxis from the late 1980s. However, after 20 years of oblivion, with the exception of a few ‘dissenting’ centres in Scandinavia (20,21) and the United States (22,23), early nCPAP has been experiencing a veritable renaissance in recent years (23,24). CPAP works in several ways: CPAP increases the transpulmonary pressure which results in a greater thoracic gas volume and functional residual capacity (FRC) (25). It is believed that the increase in FRC is caused by the recruitment of collapsed alveoli, thus increasing the surface area for gas exchange and decreasing intrapulmonary shunt. Improved gas exchange after recruitment of alveoli with surfactant and CPAP can allow for lower FiO2 thereby reducing oxygen toxicity (26). Maintaining an adequate FRC as soon after birth as possible will result in stabilization of air spaces, prevent the formation of atelectasis and promote the release of surfactant stores (27).
Avoiding intubation will promote normal airway mucociliary transport, improved humidification of inspired air, decreased risk of airway damage and secondary infection (28). CPAP also decreases the respiratory frequency in proportion to the applied pressure level (29) and elicits pulmonary reflexes, influencing apnoea rates (30). In experimental situations using baboons, CPAP has been shown to preserve alveolar counts and surface area (31), improve cerebral outcome (32) and decrease the pulmonary inflammatory response compared with MV (33).
In the pioneering study by Gregory et al. (19), the indications for CPAP treatment were similar in principle to the indications for ventilator treatment (except for total apnoea) and the air–oxygen mixture generating the CPAP was administered, as in ventilator treatment, via a tracheal tube with all its inherent complications. Obviously, the way forward to earlier treatment and presumably improved results would necessitate less invasive systems.
Caliumi-Pellegrini et al. (34) solved the problem by simply replacing the tracheal tube with a 2 cm long ‘twin nasal cannula’. nCPAP elegantly takes advantage of the fact that most neonates are nose breathers and spontaneously form a seal between the palate and the tongue – a mechanism which is enhanced by CPAP. In addition, the mouth can act as a natural safety valve in cases of excessive pressures. Within a short span of years, an outburst of observational, experimentally controlled studies, as surveyed by Bancalari and Sinclair (5), documented the feasibility as well as the effect of nCPAP in infants with RDS – improving oxygenation and decreasing the oxygen requirement and mortality as well as the need for assisted ventilation. In these early controlled studies, CPAP appeared to increase the risk of pulmonary air leaks without having any certain effect on the incidence of BPD. The best effect on MV was obtained when nCPAP was applied early rather than later during the course of RDS. Side-effects of nCPAP proved to be pneumothorax, air distension of the stomach, pressure necrosis of the nose, especially the nasal septum and noise – complications which are, to a large extent, avoidable with proper technique, experience and, last but not least, skilled nursing care.
The perfect CPAP
The perfect CPAP should be able to keep airway pressure as constant as possible to reduce work of breathing (WOB), as any deviations in pressure add to WOB performed by the patient, while FRC should rise proportionately to the increase in CPAP pressure (35). This requirement means that the device must have highly variable flow characteristics to be able to accelerate flow on inspiration without time lag. On expiration, pressures should not overshoot, thus maintaining a low WOB. In reality this is not so. Even the best systems on the market demonstrate a pressure variance and cannot compensate fully for leaks. Differences in failure rates between studies can possibly be attributed to the performance of the devices used (36).
A variety of devices and strategies for administering CPAP have been developed (37). Many of these are complicated and difficult to handle. Therefore, the use of head box, face mask and endotracheal tube to administer CPAP is rare today, with CPAP applied via nasal prongs being the most common mode. Short bi-nasal prongs are to be preferred in any CPAP system for preterm infants. The diameter and length of the prongs affects the pressure and flow significantly (36). The development of more efficient nasal prong systems has coincided with reports in the early 1990s from studies that report an increased CPAP success rate in neonates (38,39). While facilitating patient care, these systems are easy to apply. (40,41). However, generally, there is a lack of studies that have investigated the physical properties and performance of and between different CPAP systems.
Variable flow systems
The variable flow systems use jet entrainment, i.e. use of a fast moving jet of compressed gas to accelerate an air mass. The CPAP pressure is built at the nasal orifices via a jet stream passing through an opening in the nosepiece. The Benveniste jet (Fig. 1) was the first jet system to be introduced (40). The principle is used in some commercially available systems (Medijet®; Medin, Puchheim, Germany, Arabella®; Hamilton Medical, Bonaduz, Switzerland). The variable flow jet (Infant Flow®; CardinalHealth, Dublin, OH, USA) (Fig. 2) was developed a decade later by Moa et al. (41) and achieves its variable flow by enabling air entrainment via the exhaust limb of the system. Positive pressure on inspiration and expiration is controlled by using the jet as a fluidistor achieving a ‘fluidic flip’ of the jet stream by attaching the jet to different surfaces – the Choanda effect, which is unique. The silicone nasal prongs developed for this system have the lowest flow resistance of all available prongs (36).
Constant flow systems
Constant flow systems are either ventilator derived or use the classic underwater bubble CPAP system (19). In its basic design, the bubble system is cheap and relatively easy to monitor due to the absence or presence of bubbling, indicating insufficient vs. sufficient achievement of the set pressure in the circuit. Even though 6 L/min is sufficient to cover minute volume, it leads to larger pressure swings as it is unable to cope with leaks and peak flow, increasing WOB, especially in low compliant lungs, and subsequently will achieve higher pressures than the set value in the absence of leaks. A possible explanation for this is that prong pressure during bubble nCPAP delivery in preterm infants is highly variable and dependent on the interaction of submersion depth and flow amplitudes.
In one study comparing it with ventilator derived CPAP in preterm infants, the bubble system has been speculated to facilitate CO2 removal due to oscillations (42). A small cross-over study was unable to confirm this (43). Further, in a lung model, the pressure oscillations during bubble CPAP were found to become progressively attenuated distal to the nasal prongs (44). In contrast, investigators found that bubble CPAP improved gas exchange and lung morphology in a preterm lamb model (45), and recently in a single centre extubation study bubble CPAP vs. infant flow CPAP favoured bubble CPAP (46).
Comparing different CPAP systems
Very little is known about system performance in trials comparing invasive (intubated positive pressure) ventilation and nCPAP (5).
Variable flow systems tend to show better lung recruitment compared with ventilator derived or bubble CPAP (47). Indeed, comparisons between different delivery systems tend to show an advantage for variable flow over bubble systems that in turn perform better than constant flow systems regarding pressure stability, resulting in lower WOB and improved oxygenation. A variable flow device was more effective than constant flow systems (ventilator derived and bubble) in preventing apnoea episodes in preterms (48).
When trying to determine CPAP efficacy drawing on the present body of clinical studies, the inclusion of trials where the design has been extubation to nCPAP poses a problem. Little or no information is supplied regarding whether infants have been intubated and bagged during resuscitation (18) with or without multiple surfactant doses and normally extubation has taken place after undefined time on ventilator, set at unknown pressures and FiO2.
CPAP and NIPPV
The use of non-invasive positive pressure ventilation (NIPPV) is well established (49). Considering the increased use of CPAP in newborns during the last decade and the realization that MV is one of the culprits in the development of BPD, it is natural that interest in this modality has increased in the neonatal community (22,50).
The addition of NIPPV will increase patency of the upper airways, activate the respiratory drive, decrease inspiratory effort, reduce chest wall distortion and further reduce WOB compared with CPAP only (51,52). Further, NIPPV has been shown to increase the efficacy of breaths by obtaining larger tidal volumes and minute ventilation, compared with nCPAP only.
Two systematic reviews exist comparing NIPPV with nCPAP only. One examines the effect on treatment of apnoea in preterm infants and the other the effect on reduction in the need for re-intubation in preterms (53,54). The two trials investigating the effect on apnoea of prematurity (n = 54 patients) only reported short-term results of the intervention (4–6 h) on apnoea episodes. Only Lin et al. found a significant difference between the groups in the incidence of apnoeas, favouring NIPPV. The systematic review of CPAP vs. NIPPV following extubation indicates that NIPPV may be superior to nCPAP only (55). The three trials included (n = 159 patients) compared short-term outcomes and showed a significant reduction in the need for re-intubation, favouring NIPPV. The meta-analysis showed a relative risk reduction of 0.21 (CI: 0.10–0.45) with a number needed to treat of three to avoid one extubation failure. Further, in two of the trials, a non-significant trend towards reduction in BPD was seen.
Three recent trials have investigated the use of NIPPV as the primary early respiratory support in infants greater than 28 weeks gestation with respiratory distress, or early after surfactant delivery (56,57). Kugelman et al. (58) recently conducted a randomized trial (n = 84) where infants supported with NIPPV required less intubation and had lower BPD rates. All trials concluded that NIPPV can be used safely in these conditions.
CPAP and surfactant – the INSURE strategy
The Scandinavian strategy is to start all spontaneously breathing preterm infants on CPAP in the delivery room immediately after birth. The starting time of CPAP is crucial because the non-compliant lung will otherwise collapse and positive pressure ventilation is more likely to be required to open the lung. With the practice of early CPAP rather than intubation in the delivery room, a strategy that allows administration of surfactant as early rescue treatment is necessary. This may be particularly important in the extremely preterm infant. In the continuous positive airway pressure or intubation at birth (COIN) trial of 610 infants born at 25–28 weeks of gestation and assigned to either CPAP or intubation at 5 min after birth, the threshold for intubation, and thereby also surfactant treatment, was set to a FiO2 requirement of >0.6 (24). This translates to late rescue surfactant treatment in a population that can be assumed to be surfactant-deficient and may therefore have negatively influenced the outcome, such as the higher rate of air leaks in the CPAP group. Another recent retrospective report from the Netherlands studied the change in care practice from elective delivery room intubation to early nCPAP combined with early rescue surfactant treatment at FiO2 0.4 and, in contrast to the COIN study, the incidence of pneumothorax was lower in the nCPAP group (59). In a Stockholm survey of all infants less than 1500 g born 1998–1993, failure of CPAP and the need for MV were found to be significantly associated with the presence of RDS and a gestational age <27 weeks (38). This again illustrates the importance of surfactant supplementation and the need to incorporate a strategy for surfactant administration in a programme for CPAP care of VLBW infants.
The Scandinavian model is based on primary nCPAP in combination with early rescue surfactant followed by rapid extubation to nCPAP, the INSURE procedure. The approach was first reported by Victorin in 1990 in spontaneously breathing infants without CPAP (60), then further developed by Verder et al. in 1992 (61) with use of nCPAP and INSURE in severe RDS. In 1994, the Danish–Swedish group published the first randomized controlled trial of surfactant instillation during nCPAP, showing that a single dose of surfactant reduced the need for MV by half, from 85% without surfactant to 43% (8). The effect was even more pronounced if surfactant was given as early rescue treatment, which was reported in a subsequent randomized study of 60 VLBW infants with gestational age < 30 weeks (9). A retrospective follow-up study from Stockholm confirmed that the implementation of INSURE resulted in a reduction in MV rates by 50%, but reported an increase in the overall use of surfactant in the 5-year period after the introduction of INSURE (62). This is consistent with the most recent meta-analysis comparing early surfactant administration with brief MV to later, selective surfactant treatment followed by continued MV (63). Keeping in mind the INSURE procedure is a means to provide surfactant for a selected population of surfactant-deficient infants and making surfactant treatment available to more infants is a desirable effect associated with INSURE, which may contribute to the reduced MV rate.
As observed in previous studies, the Stockholm follow-up indicated that repeated doses of surfactant were rarely needed after INSURE. The improvement in oxygenation was better sustained compared with infants who were kept intubated and mechanically ventilated after surfactant administration (Fig. 3), which might have explained the reduced need for a second dose after INSURE. This observation is supported by experimental data; in rabbits surfactant followed by MV compared with spontaneous breathing resulted in increased surfactant inactivation (15), and in lambs, only a few large manual inflations impaired the treatment response to surfactant (18). Hence, INSURE is a strategy not only to reduce the need for MV, but it may also increase the efficiency of surfactant treatment.
Prevention of BPD
Normally, the first clinical manifestation of BPD begins 8–14 days after birth, and in the first days and weeks of life vulnerable preterm lungs are exposed to several hits. Thus, the mode of ventilation in the delivery room (18), MV (15–17), oxygen treatment, prenatal (64) and neonatal infections (16,65) may all (66) influence the development and severity of the BPD. Nutrition and heritability also play a role.
nCPAP vs. MV and BPD
Conventionally, MV may damage the lungs (15), and the incidence of BPD was low in centres with a long tradition of primary treatment of RDS via nCPAP compared with MV (22,23,39,67). In a multivariate analysis comparing the outcome in VLBW weight infants from centres using primary nCPAP and MV, BPD was explained simply by the incidence of MV (50). In the COIN study (24) with contributions from centres with a relatively recent tradition of using nCPAP, a trend towards less BPD was found in the nCPAP group compared with the MV group. Many observational studies with the use of primary nCPAP vs. MV support the findings of diminished development of BPD when primary nCPAP was used (14,23,39,67). All these studies support the findings of a tight link between the use of MV and the development of BPD.
INSURE, surfactant treatment and BPD
Although the introduction of surfactant treatment has not reduced the incidence of BPD (68,69), there are strong indications and evidence that surfactant therapy has a positive effect on the development of BPD and on the severity of BPD in the individual patients. The incidence of BPD is less following treatment with prophylactic surfactant than after rescue treatment (70), and early selective surfactant treatment reduces the combined mortality and BPD more effectively than rescue treatment of infants mechanically ventilated (71). To this evidence are added the positive experiences obtained by INSURE. Thus, the incidence of BPD was very low in centres which introduced INSURE compared with centres using MV and rescue surfactant (8,9). At the time when INSURE was introduced in the early 1990s, the incidence of BPD on day 28 in premature infants with moderate-to-severe RDS was 9% vs. 30% in infants treated with MV and surfactant (Table 1). A Cochrane meta-analysis from 2009 including six studies indicates that early nCPAP with early surfactant compared with nCPAP with late surfactant significantly reduces BPD, the need for MV and air leaks (63). Most observational studies have also shown that INSURE may diminish the development of BPD. In a recent randomized trial comparing INSURE with nCPAP, the incidence of MV, air leaks and BPD was less in the INSURE group (74). Another U.S. group reduced the incidence of chronic lung disease in infants < 1500 g by avoiding intubation in the delivery room, the adoption of new pulse oximeter limits and the early use of nCPAP and surfactant (75).Very recently, the reported results from the CURPAP study (76) showed no differences in the need for MV after prophylactic vs. early rescue surfactant in extremely preterm infants on nCPAP. The incidence of BPD and pneumothorax was very low in both treatment groups. As mentioned, the positive clinical results with INSURE are supported by animal experiences (15,31).
Table 1. Gestational age, birth weight and outcome from three randomized clinical trials in the early 1990s including infants with severe RDS with entry criteria: a/APO2 < 0.22
Characteristics and outcome
Primary nCPAP + Curosurf (INSURE)† n = 68
MV + Exosurf‡ n = 6757
MV + Curosurf§ n = 2168
Figures are mean values from both treatment groups.
*p (χ2) < 0.001; **p < 0.01; other differences were not significant; (***p = 0.08).
Probably the main effect of surfactant is, indirectly, improving the ability of the infant to breathe supported by nCPAP and reducing the need for MV, thereby lessening barotraumas. The efficacy of this effect is dependent on the type of CPAP system, on the surfactant preparation and on the timing and dosing of the surfactant treatment.
A large database analysis of more than 24 000 preterm infants has demonstrated a lower mortality rate and reduced length of hospital stay with Curosurf compared with Survanta and Infasurf, associated with significant cost benefits (77). Infants treated with Curosurf show faster weaning of oxygen and less need for additional surfactant doses and 200 mg/kg is more effective than 100 mg/kg (78). The new synthetic surfactants are not yet licensed for RDS treatment. Comparative trials of lucinactant (KL4) with natural surfactants suggest equivalence, but further studies are needed to clearly demonstrate superiority.
Surfactant is administered intratracheally via an endotracheal tube or laryngeal mask. Bolus or rapid instillation gives optimal distribution; bagging is probably not necessary. Despite encouraging results with aerosolized surfactant in ventilated animal, nebulization of surfactant in the clinical setting has not yet been successful, although remains an attractive route of administration for further investigation.
Analgesia and sedation for elective intubation are recommended. Morphine is commonly used as premedication, but has limitations, such as delayed onset and a risk for apnoea in association with early extubation. In the INSURE protocol, the opioid effect is reversed with naloxone. Newer, short-acting drugs such as remifentanil are currently being evaluated in newborns with promising results (79).
Early INSURE combined with caffeine prophylaxis seems to be ideal for the prevention of BPD. Thus, treatment with caffeine citrate in a meta-analysis has been shown to reduce the development of BPD significantly (80), probably because this treatment reduces the need for MV. Although doxapram may be effective in caffeine resistant apnoea, the treatment is associated with various side-effects (81) and should at present be used only briefly or not at all in preterm infants.
Besides the known additive effect of prenatal steroids and INSURE on development of RDS (9), steroids also influence the development of BPD. Thus, it is likely that prenatal steroid treatment will have some negative effect on the development of BPD (64) in addition to the positive effect on RDS. Both early- and late-postnatal steroids reduce the incidence of BPD, but early-postnatal steroids may have serious side-effects such as cerebral palsy (82). Late-postnatal steroids seem to be safer, but there is no absolute safe window (83). NIPPV may be considered as prophylaxis or treatment to avoid severe apnoea and to diminish the development of BPD as mentioned above (58).
INSURE and modified INSURE for infants < 26–27 weeks gestation
To minimize unnecessary intubations, The European guidelines for treatment of RDS recommend prophylactic surfactant treatment for infants below 27 weeks of gestation (84). Because of that many clinicians prefer to treat these infants with prophylactic MV and surfactant from birth. As described this imply an increased risk of developing BPD compared with infants allowed to breathe via nCPAP from birth.
In the light of the results from the CURPAP study including infants of 25–28 weeks gestation (76), it can be recommended to treat infants in these age groups with nCPAP from birth and to use rescue INSURE. Another possibility is to install surfactant via a feeding catheter, prophylactically or very early. This procedure often gives infants less than 26 weeks of gestation the ability to breathe supported by nCPAP for long periods (85), and they also have low mortality and BPD compared to historical controls in the same neonatal units (85).
Oxygen treatment and BPD
The formation of free radicals and per-oxidation of lipid membranes in connection with oxygen therapy is an important contributable cause of BPD and retinopathy of prematurity (ROP). In studies looking at the avoidance of ROP, it has been shown that infants with and without surfactant treatment kept at low saturations have lesser BPD than infants allowed to have higher saturations (2). Monitoring of the oxygen tension values is also important in newborns receiving oxygen therapy (86).
Nitric oxide (NO) is a simple gaseous endogenous messenger of complex actions as it exists in several redox forms (87), producing a variety of effects in different organs. It is produced in the nose at up to 4 parts per million (ppm) concentrations at term birth and in premature birth at lower levels (88), is thought to influence the vascular tone in the lung and to have anti-inflammatory effects. In the event of intubation, these signals are lost.
The ability of inhaled NO (iNO) to increase oxygenation due to selective vasorelaxation has been the prime indication for most of the studies forming the basis for its registration as a drug for late rescue therapy; to reduce the need of extra-corporal membrane oxygenation. In a few studies, iNO has led to a reduction in death or BPD predominantly in infants with a low oxygenation index, whom in Scandinavia would have received nCPAP (89). The positive effect on oxygenation exerted by NO was around 20% when tried in a cross-over exposure (2 × 30 min) during nCPAP (90).
To lessen inflammatory response, there is probably an early window of opportunity, where iNO could be tried in combination with surfactant and nCPAP. NO lessens the inflammatory response (88), as well as the leukocyte sequestration, leukocyte and platelet adhesion and aggregation. NO also modulates apoptosis and, as a downstream mediator of vascular endothelial growth factor, it increases the vascular proliferation in developing lungs (91). Thus, postnatal iNO 5 ppm in a foetal baboon model of BPD reduces its incidence (92).
Therefore, we speculate that combined early treatment of VLBW infants with nCPAP, surfactant and iNO by reducing MV could be effective in the prophylaxis of BPD.
The combined non-invasive prophylaxis with prenatal steroids and early-postnatal treatment with INSURE has shown to decrease the severity and mortality of RDS and BPD. Optimum efforts should be made to diminish the use of MV in the first days of life to avoid barotraumas and to reduce the severity of BPD. Key factors to be implemented are minimal use of oxygen and ventilation at birth, careful monitoring of oxygen saturation and tension, early establishment of nCPAP and, if necessary, early surfactant treatment. Prophylaxis with caffeine is recommended, and in cases with early apnoea also NIPPV. The optimal type of CPAP system and surfactant preparation, dosing and method of administration may significantly influence the outcome.