The management of patent ductus arteriosus is controversial, and there are diverse approaches to treatment, ranging from very conservative management through to early and aggressive securing of ductus closure, either pharmacologically or surgically. This lack of consensus on best management reflects a paucity of high quality randomised controlled trials, with many published studies focusing on establishing points of treatment, rather than looking for benefits of intervention over more conservative management. Despite this lack of good evidence views on ductus management can be entrenched, with accompanying loss of equipoise. This review looks at our current situation with regard to ductus arteriosus management and the need for good quality trials especially in the light of other published studies, concerning postnatal steroids, caffeine and oxygen which have demonstrated unexpected benefits – or sometimes unexpected harm – from long-familiar drugs.
Statistics from Altmetric.com
There are vastly differing practices among neonatologists when it comes to managing patent ductus arteriosus (PDA). Different approaches include conservative management with no active treatment, prophylactic use of cyclo-oxygenase (COX) inhibitors (indomethacin and ibuprofen), treatment with COX inhibitors when a duct is assessed as being ‘significant’, early surgical closure and surgical ligation of only refractory cases. The reason for this variation is that current evidence does not mandate one treatment over any other. Beliefs regarding the best treatment option are developed individually, and can become entrenched. Such strongly held views are open to criticism; there is no national or international consensus on the best management strategy, and the evidence on which these views are based is often questionable; in many cases, studies of ductus management ask the wrong question, concerning the wrong outcome measure.
PDA is commonly diagnosed in extremely preterm infants and has been associated with numerous pathologies including chronic lung disease, necrotising enterocolitis, intraventricular haemorrhage and with increased mortality.1 ,2 The evidence for some of these associations is conflicting and of course association does not imply causation. Nevertheless, these associations have made closing the duct an attractive prospect to neonatologists, despite the lack of robust evidence as to long-term benefit. This article is not intended to be a comprehensive review of the literature concerning the PDA; there are many such summaries already published. Nevertheless, some reference to what the substantial body of literature concerning the PDA tells us is warranted. Even a cursory examination quickly uncovers a tendency to assume benefit, the focus of many studies being to determine which treatment is most effective. The question asked by any particular study might concern optimal timing of therapy, optimal drug dosage or some aspect of surgical management. Few studies have attempted to address the more thorny issue of whether active duct closure is more beneficial than conservative management.
Treatment decisions should be based on evidence that they reduce important adverse outcomes, especially adverse neurodevelopment, that might result from ongoing ductal patency and, furthermore, that benefits of treatment outweigh the risks. Surprisingly, only a few high-quality multicentre randomised controlled trials have looked at neurodevelopmental and other long-term outcomes following PDA management. The Trial of Indomethacin Prophylaxis in Preterms study, which studied prophylactic indomethacin use in extremely low birthweight infants, did not demonstrate any positive or negative effect on survival or neurosensory impairment despite reduced frequency of PDA and severe periventicular or intraventricular haemorrhage.3 Furthermore, there was no difference in rates of chronic lung disease, necrotising enterocolitis or retinopathy of prematurity.3 Likewise, the Indomethacin Intraventricular Haemorrhage Prevention Trial showed that low-dose indomethacin significantly reduces the presence and severity of intraventricular haemorrhage.4 Nevertheless, long-term follow-up showed no benefit in terms of improved neurodevelopmental outcome at 36 and 54 months corrected.5 ,6 Meta-analysis of published randomised controlled trials concurs with these individual studies; Fowlie's review showed that indomethacin prophylaxis reduces incidence of intraventricular haemorrhage (IVH), symptomatic PDA and the need for surgical ligation, but there were no reductions in other morbidities, including chronic lung disease (CLD), necrotising enterocolitis (NEC) and, most importantly, neurodevelopmental impairment.7 Similarly, although less extensively studied, later treatment of the PDA judged to be ‘haemodynamically significant’ has not been shown to influence 2-year developmental outcome in very low birthweight infants.8
In short, the body of literature concerning the PDA has failed to demonstrate any improvement in mortality or long-term neurodevelopmental outcome in infants who have received COX inhibitors to close the PDA. This lack of benefit should prompt us to re-evaluate our rationale for administering these medicines. If this rationale focuses on the reduction in the incidence of common neonatal pathologies such as CLD or NEC, then the results published to date suggest no benefit, in keeping with the relationships being associations rather than causation. Either that, or there are deleterious effects of the medicines used that are masking any potential benefit. This might be of particular concern when considering the reduction in incidence of IVH, a long-standing feature in the results tables of randomised controlled trials of PDA treatment which has never translated into a convincing benefit in terms of long-term neurodevelopmental outcome. Either way, if there is no evidence of long-term benefit, is it right to continue treating infants with these medicines?
Treatments used for duct closure are not without potential harm and indeed there are significant risks associated with their use. The effects of indomethacin administration on renal and platelet function, and its association with intestinal perforation, are well described, but of greater concern is the drug's potential to reduce cerebral blood flow increasing the likelihood of cerebral hypoxia.9 Considering maintenance of tissue oxygen delivery is the primary aim of modern neonatal intensive care, it seems paradoxical that we should continue administering a pharmacological agent which is associated with a reduction in cerebral blood flow and cerebral oxygen delivery in the absence of any long-term outcome data favouring its use. Surgical duct ligation is also not without risk; infants undergoing duct ligation experience significant changes in cerebral oxygenation, and there is substantial late mortality and high incidence of morbidity in survivors.10 Ductal ligation has also been associated with several other significant morbidities including hypotension, pneumothorax, chylothorax, infection, vocal cord paralysis, myocardial dysfunction, neurodevelopmental impairment and increased risk of bronchopulmonary dysplasia.11,–,15 A recent Scottish study declared a 30-day mortality of 4.8%, 1-year mortality of 12.8% and a 32% incidence of neurodisability in survivors who had undergone duct ligation.16
Despite the absence of evidence of beneficial effect on survival or long-term neurodevelopmental outcome, some neonatologists maintain a seemingly immovable belief that active duct closure is the correct course of action. This lack of equipoise has in turn meant that good quality, double-blinded, randomised controlled trials comparing active medical or surgical treatment with conservative management, and which may have given definitive answers to this fundamental question, have not received widespread support. Lack of equipoise is warranted when it is based on good evidence. However, as we have already alluded, the current level of trial data does not, in our opinion, merit such strongly held beliefs.
The reluctance to commit to randomised controlled trials of treatments, especially those therapies that have crept into common practice over time, is not a new phenomenon in neonatology. The history of our specialty is littered with therapies introduced without good trial data that have turned out to be harmful, and similarly with therapies which have turned out to have unexpected benefits. The administration of postnatal steroids to treat chronic lung disease and aid extubation is a good example of the former; a widely used therapy introduced without being subject to randomised controlled trial, only later investigation revealed the harm that this therapy did to so many babies.17,–,19 The Caffeine for Apnoea of Prematurity study subjected another medicine long-established on the neonatal unit to an appropriate level of study, prompted by concerns about potential long-term neurodevelopmental harm.20 Such concerns were not merely laid to rest; the study revealed unimagined benefits in terms of improved long-term neurodevelopmental outcome, reduction in the incidence of chronic lung disease, as well as – ironically considering the subject at hand – less ductal patency.20 ,21 More recently, the Second Benefits Of Oxygen Saturation Targeting (BOOST II) study, against expectations, showed that infants <28 weeks gestation who were randomised to oxygen target saturations of 91% to 95% had a higher survival rate than those with the lower target range of 85% to 89%.22 The magnitude of the effect was striking, with a 20% difference in mortality between the two groups, a dramatic difference in survival similar to that brought about by antenatal steroids or endotracheal surfactant administration. That this study was undertaken despite the widely accepted belief that lower oxygen saturation target ranges are preferred for their lower incidence of retinopathy of prematurity is sobering and shows that we must continue to question our beliefs, just as we have done with postnatal steroids and caffeine for apnoea of prematurity. Time and again, in these studies and others, firmly held preconceptions have been challenged and often shown to be wanting. Results have at times been unexpected, sometimes strikingly so, and long-established therapies have been discarded as a result. We should therefore be cognisant that, once COX inhibitors are subjected to appropriate study, our current management of the PDA may be viewed in a similar way – as one more in a list of harmful treatment strategies we took too long to subject to appropriate study. Or, like caffeine, perhaps it will be more firmly embraced.
Our current position on the management of the ductus arteriosus is open to question, and it is only through well-designed multicentre randomised control trials assessing active management of the PDA against placebo that our current beliefs as to the benefit or otherwise of these treatments will be refuted or confirmed. We may find unexpected benefits, or of course unexpected harm. We should therefore collectively take a deep breath and, once more, grasp the nettle of a double-blind randomised controlled trial of a long-familiar drug without further delay.23
Competing interests None.
Provenance and peer review Commissioned; externally peer reviewed
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.