Abstract
Respiratory distress syndrome is a disease of prematurity and is caused by a relative deficiency of endogenous surfactant production. Respiratory distress syndrome is the most common cause of mortality and morbidity in the newborn population and the standard of care is to provide exogenous surfactant therapy. This saves lives and reduces respiratory complications but, despite treatment, a significant proportion of these infants go onto develop chronic lung disease, the severest form of which is bronchopulmonary dysplasia. Once developed, this is a multisystem disease and treatment is mostly supportive by using various therapeutic adjuncts. Some of these have been proven to be safe and effective in large randomized, controlled trials but similar evidence for other drugs is lacking. The aim of this paper is to provide an overview and critically appraise the available scientific evidence for or against their use in routine practice.
Introduction
Respiratory distress syndrome (RDS) is a specific clinical entity occurring mostly, but not exclusively, in preterm infants. It occurs predominantly due to a lack of endogenous surfactant (a surface tension lowering agent) in the alveoli, but overall anatomical and structural immaturity also plays an important part. The key treatment, therefore, is replacement exogenous surfactant therapy; although some infants will still go on to have respiratory failure of sufficient severity to require mechanical respiratory support. Surfactant therapy saves the lives of premature infants, including those who are born at borderline viability (23–25 weeks gestation), but despite this improvement in survival, the majority of such infants then experience difficulties with the onset of chronic lung disease (CLD), the severest form of which is bronchopulmonary dysplasia (BPD). Although a disease of prematurity, BPD is multifactorial in origin (Table 1) and results from a sequence of events which increases infants’ dependency on ventilation and oxygen. The incidence of BPD remains approximately 30–40% in very preterm infants. Once developed, there are many health consequences of BPD, including prolonged hospital stay requiring artificial respiratory support, asthma, pulmonary hypertension, failure to thrive, cognitive impairment, and neurodevelopmental delay. There is also a high incidence of postnatal mortality and frequent rehospitalization among infants diagnosed with BPD. Once developed, the treatment for BPD is only supportive but can be optimized with the use of therapeutic adjuncts aimed to ameliorate the underlying pathogenesis [Donn and Sinha, 2009].
RDS, respiratory distress syndrome; BPD, bronchopulmonary dysplasia.
Some of these therapeutic approaches are based on sound physiologic rationales and their efficacy and safety are well tested in large controlled trials, whilst others may still be experimental and lacking scientific credibility but are still being used on individual neonatal units [De Luca et al. 2011]. The purpose of this paper is to provide an overview of currently used therapeutic options for the treatment of respiratory disease in newborns and critically appraise the available scientific evidence for or against their use in routine practice. The list, however, is not exhaustive and beyond the limits of this review. Readers are therefore advised to keep abreast of the developments of individual drugs or therapies relevant to the neonatal population.
Pharmacotherapy to induce antenatal lung maturation and endogenous surfactant production
As RDS is a direct effect of surfactant deficiency resulting from decreased production due to prematurity and corticosteroids induce surfactant production, it makes sense to induce antenatal lung maturation and enhance endogenous surfactant production by administering antenatal steroids to mothers whose pregnancies are likely to culminate in premature delivery. Antenatal steroids reduce the incidence of RDS by almost 50% and, even in those infants who develop RDS, it tends to result in a less severe form. Evidence from a Cochrane systematic review supports the use of a single course of antenatal corticosteroids (two doses of betamethasone) to accelerate foetal lung maturation in women at risk of preterm birth [Roberts and Dalziel, 2006].
Post hoc analysis of studies included in this systematic review, however, does not show consistency in the benefits for infants of less than 26 weeks gestation, both with regard to neonatal death and RDS. Moreover, the data in the Cochrane review relates to very few foetuses of less than 26 weeks at exposure (n = 49, of which 22 were stillborn), which makes it difficult to interpret the data in relation to its effect in this subgroup of pregnancies. Similarly, in the EPICure study, which recorded the outcome of a UK national cohort between 22 and 25 weeks gestation, there was a significant overall reduction in deaths before discharge with the use of antenatal corticosteroids [odds ratio (OR) 0.38, 95% confidence interval (CI) 0.28–0.53] but there was no comment made for infants under 24 weeks gestation [Costeloe et al. 2000].
Elsewhere, the 1994 consensus guideline of the National Institutes of Health and the revised 2004 guideline of the Royal College of Obstetricians and Gynaecologists recommend the use of antenatal corticosteroids in women at risk of delivery between 24 and 34 weeks gestation [National Institutes of Health, 1994; Royal College of Obstetricians and Gynaecologists, 2004] but they do not endorse its use before 24 weeks gestation. This is on the basis that (a) infants born before 24 weeks were not always resuscitated, (b) during the canalicular stage of lung development, type II pneumocytes (the target cells for corticosteroids) are lacking and the gas exchange surface may be nonexistent, and (c) that use of antenatal steroids at this stage might have deleterious effects on subsequent lung and brain development. Since then, however, clinical practice has changed and more infants born at borderline viability are being offered initial resuscitation and intensive care treatment [Janvier and Barrington, 2005; Wapner, 2004].
Given the uncertainty about the role of antenatal steroids in such very early gestational age infants, we have analysed our own data available from a well-defined population based cohort of premature infants to see if there was any difference in the effect of antenatal corticosteroids on neonatal outcomes according to different gestational ages at birth. In this study, the overall mortality amongst infants born between 24 and 29 weeks, who were administered antenatal steroids, was lower (n = 850/4370; 19.4%) compared with their counterparts whose mothers did not receive steroids (n = 323/920; 35.1%) The gestation specific mortality in the steroid treated group between 24 and 29 weeks gestation was 61.5%, 36.9%, 28.5%, 17.5%, 10.2% and 5.1%, respectively, and this was significantly lower than the group without steroid treatment. There was a 9.9% reduction in mortality amongst infants born at 23 weeks gestation in the steroid treated group (n = 8/102; 79.4%) compared with the nonsteroid group (n = 75/84; 89.3%) but this did not reach statistical significance (p = 0.068). Interestingly, there was no significant effect of antenatal steroid treatment seen in this study on length of hospital stay, duration of respiratory support and CLD among infants who survived until discharge [Manktelow et al. 2010].
These observations are interesting and need further exploration in larger cohort studies but use of antenatal steroids clearly seems to have a favourable effect on survival at birth in all gestational age groups below 29 weeks. Whether this beneficial effect is maintained amongst higher gestational ages, such as late preterm gestation above 34 weeks, is not clear.
The only other lung maturation strategy that has been tried is the combination of corticosteroids and thyrotropin-releasing hormone, as they can act synergistically. Unfortunately, when compared in a large randomized, controlled trial, this not only showed little benefit, but demonstrated possible serious side effects. Therefore, the use of thyrotropin-releasing hormone to supplement antenatal corticosteroid therapy is not recommended [ACTOBAT Study Group, 1995].
Exogenous surfactant replacement therapy
Considering that preterm infants have a significantly lower pool of surfactant (10 mg/kg) compared to a surfactant pool size of around 100 mg/kg in full-term infants, surfactant replacement therapy makes sense. Administration of exogenous surfactant to a surfactant deficient preterm human newborn decreases the minimum pressure required to open the lungs, increases the functional residual capacity and maximal lung volumes, and prevents lung collapse (atelectasis) at low pressure [Moya and Javier, 2012].
There are many preparations of surfactant available for use in preterm infants. Although these surfactant preparations are not similar, they are generally grouped into two categories depending on whether they are derived from animal lungs or are of synthetic origin. Since the availability of newer preparations, the synthetic surfactants are now further subdivided depending upon whether they do or do not contain peptides that mimic surfactant protein (SP)-B [such as lucinactant (Surfaxin™)] or SP-C. Lucinactant is now approved for clinical use in the USA and has been the subject of study in a number of clinical trials but there is no published experience of SP-C involving newborns.
Historically, surfactant has been used in either a ‘prophylactic’ or a ‘rescue’ approach. The former involves administration within 30–60 minutes of birth regardless of respiratory status and is usually used in very preterm newborns who are at high risk of RDS such as those born before 28–29 weeks gestation. This means that surfactant is administered to a proportion of infants who would not have ever developed RDS.
Alternatively, rescue treatment is done in infants with established signs of respiratory failure associated with changes of RDS on X-ray. In this approach, infants who are intubated and require FiO2 of more than 30–35% would be deemed eligible for treatment, which often occurs several hours after birth.
Investigators have also looked at the potential of late administration of surfactant in infants with either evolving or established BPD, and who have shown transient improvement in oxygenation and ventilatory support. These have been confirmed in two recent placebo controlled, randomized trials using either lucinactant or calfactant [Laughon et al. 2009]. However, no major impact on the prevention of BPD has been reported to date. Therefore administration of surfactant for infants with evolving or established BPD remains under investigation and cannot be widely recommended.
Many randomized trials have compared the efficacy of animal-derived surfactants to synthetic surfactants. Considering the heterogeneity of the methods used in the earlier surfactant trials, however, a better comparison of data is derived from head to head comparisons which do not demonstrate any overall difference in mortality or BPD as a result of using different surfactants [Moya and Maturana, 2007].
There have been two randomized clinical trials comparing the peptide-containing synthetic surfactant lucinactant with colfosceril palmitate, beractant and poractant. The authors reported more survivors without BPD with lucinactant compared to colfosceril but the latter is no longer available for clinical use. There was also no significant difference in survival without BPD between the three protein-containing surfactants used in the trials [Sinha et al. 2005; Moya et al. 2005]. Further, there were no differences in other common complications generally associated with prematurity.
Although exogenous surfactant therapy is an established treatment for neonatal RDS, it requires endotracheal intubation which is an invasive procedure and likely to be hazardous and cause complications. Moreover, the skill required to place an endotracheal tube, especially in the delivery room, may not be universally available. This has led to a search for alternative strategies of providing exogenous surfactant using noninvasive or minimally invasive techniques which essentially avoid the use of an endotracheal tube; collectively labelled as ‘minimally invasive surfactant therapy’ (MIST) [Gupta and Donn, 2012]. These include pharyngeal instillation of surfactants [Kattwinkel et al. 2004], administration via laryngeal mask airway [Trevisanuto et al. 2005] and administration via a thin endotracheal catheter/feeding tube [Dargaville et al. 2011]. The protocols for using these techniques are described in the individual publications but these are observational studies and their safety and efficacy is still not validated in controlled trials, and hence not generalizable to other units. Moreover, some bench data have shown that MIST may be associated with the loss of approximately 11% of the surfactant dose due to the smaller diameter of the feeding tube [De Luca et al. 2013]. Further studies are required, not only to verify the clinical utility of MIST but also the effect of surfactant activity on different devices and materials for its administration.
Of all available noninvasive techniques, aerosolised administration of surfactant [Mazela et al. 2007] seems to be the most sophisticated and minimally invasive, but this is in the preliminary stages of development and its efficacy and safety still needs to be tested in a large, properly designed controlled trial comparing it with the traditional method of surfactant therapy as the ‘gold standard’.
The typical protocol for aerosolized surfactant administration involves the use of an aerosol generator with surfactant administered by a nasal continuous positive airway pressure (CPAP) system, using either a tight face mask or nasopharyngeal tube. The efficacy of any aerosolized therapy depends on a number of factors including the infant’s weight or size, aerosol flow, aerosolized particle size (large enough to avoid potential exhalation and yet small enough to bypass the oropharynx), type of aerosol generator and retention of the biological property of the surfactant preparation after nebulization [Cole, 2000; Pillow and Minocchieri, 2012].
Aerosurf® (Discovery Laboratories, Warrington, PA, USA) is an aerosolized formulation of lucinactant developed for administration by nasal CPAP. The only human neonatal clinical evaluation of Aerosurf® to date was the pilot trial of Finer and colleagues [Finer et al. 2010]. After a single administration, 64% (11/17 infants) were successfully treated. Only one infant required the maximal four doses. All of the infants survived, but six required subsequent endotracheal intubation and intratracheal surfactant administration. The reduction in the FiO2 was comparable to that seen in most intratracheal surfactant trials. The rate of RDS at 24 hours was 24% and the rate of BPD at 28 days was 11.7%. Although the number of infants treated in the pilot trial is small, the study establishes a proof of concept and justifies a larger phase III trial.
Diuretics
Infants with RDS and BPD have a tendency to accumulate excessive interstitial fluid in the lungs. This excess of fluid can lead to a deterioration of their pulmonary function causing a delay in the resolution of lung disease and requiring longer ventilator dependency. When increased lung water persists despite fluid restriction, use of diuretics might be beneficial. However, this benefit in lung mechanics is often short lived and there is no convincing evidence that such therapy causes any beneficial effect in the long term. On the contrary, indiscriminate use of diuretics is often associated with undesirable side effects such as potential ototoxicity, hypokalaemia, hyponatraemia, metabolic acidosis, hypercalciuria with nephrocalcinosis, hypochloraemia and osteopenic bone disease, especially when used in conjunction with postnatal steroids. Therefore, infants who are on chronic diuretic therapy should be closely monitored for these side effects and adjustments made to the type and dosage of the diuretic being used, particularly as these side effects are reversible.
The diuretics often used in infants with RDS and BPD include the following.
Combination of benzthiazide and spironolactone. This is probably the most commonly used combination and thought to be safe for long-term control of fluid retention in congestive cardiac failure and BPD but it is not proven and in fact can cause considerable urinary calcium loss causing nephrocalcinosis.
Furosemide. This is a loop diuretic and has faster action but also causes significant urinary losses of electrolytes and calcium. Its ototoxic effect is enhanced by concomitant use of amnioglycosides. Chronic use of furosemide may also cause nephrolithiasis or nephrocalcinosis. Some studies have shown beneficial effects when given in nebulized form, at least in the short term. However, a systematic review of the use of furosemide for RDS in preterm infants showed that early use had no effect on meaningful outcomes, including duration of mechanical ventilation and oxygen supplementation, length of hospitalization and, more relevantly, on mortality and CLD [Stewart et al. 2011].
Bumetanide. This is also a loop diuretic but more potent than furosemide and with similar mechanism of action. Half life in the newborn is short (2–6 hours) and can cause significant urinary loss of electrolytes. Overuse can also cause significant alkalosis. Evidence to support its routine use in CLD or BPD is lacking.
Postnatal dexamethasone
Because of the importance of inflammation in the pathogenesis of CLD, there has been interest in the use of postnatal corticosteroids during the earlier stages of the disease in order to reduce its progression. Several reports showed rapid improvement in lung function after the administration of steroids, facilitating weaning from the ventilator when compared with controls who received placebo. However, the optimal age of treatment, dose schedule, and duration of therapy have not been established. Although use of corticosteroids does seem to provide some short-term gains, there is no data to suggest any improvement in long-term respiratory or neurodevelopmental outcome. In fact, three Cochrane reviews of randomized, controlled trials, originally published in 1999 and subsequently updated, raise a very worrying concern that ‘early’ use of postnatal corticosteroids may be associated with increased incidence of cerebral palsy (CP). These reviews classified the trials by age at start of corticosteroid treatment. In 19 trials this was early (less than 96 hours), in seven trials moderately early (7–14 days) and in nine trials delayed (more than 3 weeks). These reviews showed significant reduction in mortality at 28 days as well as reduction in CLD, but the most important finding of these reviews was evidence of an increased risk of abnormal neurological outcome [Halliday et al. 2003]. Since then, the approach to postnatal corticosteroid therapy in infants with RDS has changed and suggestions made that the use of dexamethasone should be restricted only to those infants who become ventilator dependent and are deemed unlikely to survive without systemic corticosteroids. In such cases, dexamethasone can be used at the lowest dose for the shortest course possible with beneficial effect.
This has opened up debate about the possible beneficial effect of low-dose steroids specifically for infants at high risk of BPD. This argument has been supported by a meta-regression analysis of 28 randomized, controlled trials, with available follow-up data from 1721 randomized infants, showing that the impact of postnatal corticosteroids on the combined outcome of death or CP was modified by the risk of BPD in the control group of infants; postnatal steroid therapy significantly increased the chance of death or CP if used in infants with a risk of BPD <35%, whereas the same therapy reduced the chance of death or CP if used in infants with risk of BPD >65% [Doyle et al. 2005].
The optimal corticosteroid type and the optimal dose are unknown, but it appears that the avoidance of steroids may in fact be detrimental to a subset of infants at high risk of BPD. In an attempt to improve the efficacy of steroids and reduce the associated systemic side effects, studies have looked into the safety and efficacy of giving steroids in a nebulized form with encouraging results but a meta-analysis of seven trials of early inhaled steroids did not confirm any improvement either in survival or incidence of BPD [Shah et al. 2007].
Pulmonary vasodilators
Pulmonary hypertension of varying degrees is often associated with CLD and has a cause and effect relationship. There is an intricate relationship between developing air space and adjacent pulmonary vasculature. It has been postulated that modulating pulmonary vasculature by prophylactic use of low-dose nitric oxide may prevent this condition. Earlier studies did show some short-term improvement in terms of a reduced period of oxygen dependency and ventilatory support [Ballard et al. 2006]. However, a large, multicentric trial testing the efficacy of low-dose nitric oxide as a preventative measure did not show any improvement in pulmonary outcome at 2 years of age [Mercier et al. 2010]. This is also supported by the findings of a recent systematic review [Donohue et al. 2011]. Thus, while nitric oxide may have the potential to prevent BPD, most likely by modifying alveolarization and angiogenesis, its clinical usefulness in preterm infants has not been proven. Yet there is a tendency in individual neonatal units to use nitric oxide in the preterm population with respiratory illnesses without fully understanding the natural history of the disease and use of nitric oxide in this situation should be avoided. There are European Consensus Guidelines on the use of nitric oxide in term and late preterm infants [Macrae et al. 2004]. Although somewhat old, these Guidelines still remain valid for this population and should be followed.
Other agents which have been tried and reported to have efficacy in observational case studies, but which are still investigational, include prostacyclin and sildenafil.
Treatment for closure of patent ductus arteriosus
In a foetus, the ductus arteriosus is a blood vessel that allows blood from the right ventricle to bypass the foetal lungs and reach the placenta. At birth, this foetal circulation changes to adult circulation and the ductus arteriosus closes. In preterm infants, the ductus arteriosus may not close spontaneously, and this is known as patent ductus arteriosus (PDA). It is a common problem amongst extremely low birth weight infants, especially those who are receiving mechanical respiratory support, and alters the mechanics of the lungs leading to prolonged ventilatory dependence and development of CLD. Although practice varies among neonatal units, there is a tendency to close the PDA, especially if deemed to be haemodynamically significant, in the hope that this will improve the lung mechanics and facilitate early weaning from the ventilator. PDA can be closed by surgical ligation but most clinicians rely on medical treatment with either indomethacin or ibuprofen.
There are two recognized approaches for the medical management of PDA in preterm infants:
Prophylactic treatment (commencing treatment in all infants within 24 hours of birth).
Symptomatic treatment (commencing treatment only when PDA is symptomatic, usually after 72 hours).
A recent meta-analysis of prophylactic treatment with ibuprofen reported a significant decrease in the incidence of PDA on day 3 as compared to controls [typical relative risk (RR) 0.36 (95% CI 0.28, 0.46); number needed to treat 4 (95% CI 3, 5)]. But in this cohort, the PDA had already closed spontaneously by day 3 in 58% of the neonates in the control group and there were no differences in complications of prematurity between the two groups. Moreover, this meta-analysis also included infants up to 34 weeks gestation who may not be as prone to PDA as smaller infants. The major criticism of the prophylactic approach came from the fact that a large group of preterm infants with PDA, which would have closed spontaneously anyway, were unnecessarily exposed to the side effects of a toxic medication. In another large trial of prophylactic treatment with indomethacin, there was a significant reduction in intraventricular haemorrhage and PDA in extreme low birth weight infants but this advantage in short-term outcome failed to show any improvement in long-term neurodevelopmental outcome [Schmidt et al. 2001]. Currently, most neonatal units do not use a prophylactic approach to PDA.
With regard to treatment of a symptomatic PDA (usually after 72 hours) which is now widely practised, the Cochrane meta-analysis reported a significant decrease in the composite outcome of deaths or those requiring rescue treatment [RR 0.58 (95% CI 0.38–0.89); number needed to benefit 5 (95% CI 3–17)] with the use of ibuprofen compared with placebo. There was no difference observed in the failure rates of treatment comparing ibuprofen and indomethacin. Ibuprofen, however, significantly reduced the incidence of necrotizing enterocolitis (NEC), a serious complication in newborns [RR 0.68 (95% CI 0.47–0.99)] [Thomas et al. 2005]. Thus, both approaches seem to have the problem of either giving treatment when it is not needed or unnecessarily exposing the infant to the harmful effect of drugs which are potentially toxic, or delaying the treatment for too long when the damage to the lungs has already occurred.
A third approach of selective early-targeted treatment of PDA before it is symptomatic, should overcome the limitations of prophylactic as well as rescue treatment approaches. With this approach, infants could be screened for PDA using echocardiography and given treatment only if showing signs of abnormal haemodynamics. This targeted approach could have the advantages of treatment similar to prophylactic trials or early asymptomatic periods, by attempting closure of the PDA before it is symptomatic, yet avoiding potentially toxic drug exposure to infants who do not have a ‘significant’ PDA. One study has utilized this approach and compared indomethacin with ibuprofen, commencing targeted treatment within 24 hours of birth in very preterm infants. It reported no difference in the efficacy of the two study drugs in closing the PDA and observed no difference in the complications between the two study drugs [Sellmer et al. 2013].
Vitamin A
Vitamin A is involved in the regulation and promotion of growth and differentiation of multiple cells. It also maintains the integrity of the epithelial cells of the respiratory tract [Hey, 2007]. Preterm infants are relatively deficient in vitamin A, and this has been shown to be associated with BPD. A large randomized, controlled trial involving 807 infants with a birth weight of less than 1000 g, showed that a large dose of intramuscular vitamin A, given three times a week for 4 weeks from birth, reduced the risk of BPD (RR 0.89, 95% CI 0.8–0.99) [Tyson et al. 1999]. However, this therapy has not been adopted readily into practice probably because it requires intramuscular injections. A trial of oral vitamin A therapy daily for 4 weeks in a similar population of infants failed to detect any benefit [Wardle et al. 2001]. Meta-analysis of 8 trials that compared vitamin A supplementation with control (1291 infants) suggested that vitamin A supplementation reduced the risk of BPD at 36 weeks post menstrual age (RR 0.87, 95% CI 0.77–0.98) [Darlow et al. 2011].
Caffeine therapy
Caffeine (usually in the form of caffeine citrate) has been the methylxanthine therapy of choice because of its wider margin between the therapeutic and toxic levels [Hey, 2007]. The use of caffeine for apnoea of prematurity was first reported in 1977 [Aranda et al. 1977] and despite the widespread use of this therapy over the past three decades, there are no reliable data on its long-term efficacy and safety until recently [Schmidt, 1999]. Some small observational studies with follow-up information did not show any long-term adverse effect of methylxanthine therapy [Gunn et al. 1979; Nelson and Resnick, 1981] but one large follow-up study of over 400 very low birth weight infants reported that theophylline administration was significantly associated with an increased risk of CP [Kitchen et al. 1987]. Furthermore, reports from some observational studies suggested that caffeine reduced cerebral and intestinal blood flow velocity in preterm infants [Hoecker et al. 2002], resulting in the perception that caffeine therapy may increase the risk of NEC.
The CAP (Caffeine for Apnoea of Prematurity) Trial was conducted to determine whether survival at 18 months (corrected age) without disability is improved if apnoea of prematurity is managed without methylxanthines in infants with birth weights of 500–1250 g. The trial design was pragmatic, and infants were enrolled in the first 10 days of life, if they were considered suitable candidates for methylxanthine therapy, to receive either caffeine citrate (20 mg/kg of initial loading dose, followed by 5–10 mg/kg/day) or the equivalent volume of placebo until the therapy was deemed necessary. The trial enrolled 2006 infants between 1999 and 2004, and collaborating clinicians managed to keep the use of open-labelled methylxanthines to less than 10%.
This large international trial showed caffeine therapy significantly reduced the risk of BPD, defined as a need for supplemental oxygen at 36 weeks post-menstrual age: (36% in caffeine group versus 47% in placebo group, OR 0.63, 95%CI 0.52–0.76; p < 0.001). Supplemental oxygen therapy, CPAP and mechanical ventilation were discontinued 1 week earlier with the use of caffeine compared with placebo. Unexpectedly, caffeine therapy was also found to be associated with a significant reduction in the risk of PDA deemed to need pharmacological and surgical closure. There was a temporary reduction in weight gain during the first 3 weeks of caffeine therapy, but there was no adverse effect on growth at the time of hospital discharge. This large study also provides reassurance that caffeine therapy did not increase the risk of death, or neonatal complications including NEC and ultrasonographic evidence of brain injuries [Schmidt et al. 2006]. Adequate data for an analysis of the primary outcome data were available for 93% of the enrolments, and the high-quality follow-up data from the CAP trial also showed that caffeine therapy improves the rate of survival without neurodevelopmental disability at corrected age of 18–21 months (death or disability 40.2% in caffeine group versus 46.2% in placebo group, OR 0.77, 95%CI 0.64–0.93; p = 0.008). Treatment with caffeine compared with placebo also significantly reduced the incidence of CP (4.4% versus 7.3%, p = 0.009) and of cognitive delay (33.8% versus 38.3%, p = 0.04) [Schmidt et al. 2007]. The post hoc analysis to explore the likely mechanism as to why caffeine therapy resulted in better neurodevelopmental outcome suggested that earlier discontinuation of positive airway pressure in infants who received caffeine, compared with placebo, was the most important intermediate variable. This mechanism was supported by further analyses of the subgroups that, in terms of death or major disability, infants who were receiving CPAP or mechanical ventilation by endotracheal tube appeared to have gained more benefit from caffeine therapy compared to those infants who did not require any positive pressure ventilation [Davis et al. 2010].
There is a growing list of drugs that can be used in the treatment of respiratory diseases in the newborn, including bronchodilators, proton pump inhibitors, antibiotics, analgesics and sedatives, but there are no clear guidelines for their routine use and the benefit/risk ratio of using such medications should be evaluated on an individual basis.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Conflict of interest statement
The authors declare no conflicts of interest in preparing this article.