Indomethacin for symptomatic patent ductus arteriosus in preterm infants

Abstract

Background

Symptomatic patent ductus arteriosus (PDA) is associated with mortality and morbidity in preterm infants. In these infants, prophylactic use of indomethacin, a non‐selective cyclooxygenase inhibitor, has demonstrated short‐term clinical benefits. The effect of indomethacin in preterm infants with a symptomatic PDA remains unexplored.

Objectives

To determine the effectiveness and safety of indomethacin (given by any route) compared to placebo or no treatment in reducing mortality and morbidity in preterm infants with a symptomatic PDA.

Search methods

We used the standard search strategy of Cochrane Neonatal to search Cochrane Central Register of Controlled Trials (CENTRAL; 2020, Issue 7), in the Cochrane Library; Ovid MEDLINE(R) and Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, Daily and Versions(R); and Cumulative Index to Nursing and Allied Health Literature (CINAHL), on 31 July 2020. We also searched clinical trials databases and the reference lists of retrieved articles for randomized controlled trials (RCTs) and quasi‐RCTs.

Selection criteria

We included RCTs and quasi‐RCTs that compared indomethacin (any dose, any route) versus placebo or no treatment in preterm infants.

Data collection and analysis

We used the standard methods of Cochrane Neonatal, with separate evaluation of trial quality and data extraction by at least two review authors. We used the GRADE approach to assess the certainty of evidence for the following outcomes: failure of PDA closure within one week of administration of the first dose of indomethacin; bronchopulmonary dysplasia (BPD) at 28 days' postnatal age and at 36 weeks' postmenstrual age; proportion of infants requiring surgical ligation or transcatheter occlusion; all‐cause neonatal mortality; necrotizing enterocolitis (NEC) (≥ Bell stage 2); and mucocutaneous or gastrointestinal bleeding.

Main results

We included 14 RCTs (880 preterm infants). Four out of the 14 included studies were judged to have high risk of bias in one or more domains. Indomethacin administration was associated with a large reduction in failure of PDA closure within one week of administration of the first dose (risk ratio (RR) 0.30, 95% confidence interval (CI) 0.23 to 0.38; risk difference (RD) ‐0.52, 95% CI ‐0.58 to ‐0.45; 10 studies, 654 infants; high‐certainty evidence). There may be little to no difference in the incidence of BPD (BPD defined as supplemental oxygen need at 28 days' postnatal age: RR 1.45, 95% CI 0.60 to 3.51; 1 study, 55 infants; low‐certainty evidence; BPD defined as supplemental oxygen need at 36 weeks' postmenstrual age: RR 0.80, 95% CI 0.41 to 1.55; 1 study, 92 infants; low‐certainty evidence) and probably little to no difference in mortality (RR 0.78, 95% CI 0.46 to 1.33; 8 studies, 314 infants; moderate‐certainty evidence) with use of indomethacin for symptomatic PDA. No differences were demonstrated in the need for surgical PDA ligation (RR 0.66, 95% CI 0.33 to 1.29; 7 studies, 275 infants; moderate‐certainty evidence), in NEC (RR 1.27, 95% CI 0.36 to 4.55; 2 studies, 147 infants; low‐certainty evidence), or in mucocutaneous or gastrointestinal bleeding (RR 0.33, 95% CI 0.01 to 7.58; 2 studies, 119 infants; low‐certainty evidence) with use of indomethacin compared to placebo or no treatment. Certainty of evidence for BPD, surgical PDA ligation, NEC, and mucocutaneous or gastrointestinal bleeding was downgraded for very serious or serious imprecision.

Authors' conclusions

High‐certainty evidence shows that indomethacin is effective in closing a symptomatic PDA compared to placebo or no treatment in preterm infants. Evidence is insufficient regarding effects of indomethacin on other clinically relevant outcomes and medication‐related adverse effects.

Plain language summary

Indomethacin for treatment of symptomatic patent ductus arteriosus (PDA) in preterm infants

Review question

We wanted to see how safe and effective indomethacin is when compared with no treatment or placebo (a substance with no therapeutic effect) in treating symptomatic patent ductus arteriosus (PDA) in a preterm infant.

Background

PDA is a common complication among preterm and low birth weight infants. PDA is an open vascular channel between the lungs and the heart, which usually closes shortly after birth. In preterm infants, the PDA remains open and may contribute to life‐threatening complications. We wanted to see if a medicine commonly used to close PDA, indomethacin, was better or worse than no treatment or placebo for preterm infants who were experiencing symptoms from their PDA.

Study characteristics

We searched scientific databases for randomized controlled trials (clinical studies where people are randomly put into one of two or more treatment groups) in preterm (born at less than 37 weeks into pregnancy) or low birth weight (weighing less than 2500 grams) infants with a symptomatic PDA diagnosed by a combination of specific clinical features and/or ultrasound of the heart. Included studies compared indomethacin versus no treatment or placebo. We searched for studies that had been published up to 31 July 2020.

Key results

This review of 14 clinical trials (880 infants) found that indomethacin is very effective in closing a PDA, reducing the risk of PDA being open at one week by 70%, when compared to no treatment or placebo. Of the 14 total studies, eight reported on death, which was not different between groups. Only one study reported on the number of infants needing oxygen at 28 days of life, and only one study reported on the number of infants using oxygen at 36 weeks' postmenstrual age (how many weeks pregnancy plus how long the infant has been alive). Data are not adequate to permit conclusions on these outcomes. Both studies showed no difference between those who received indomethacin or no treatment or placebo. Medication side effects or other adverse effects (such as gut bleeding, kidney injury, and bowel infection/lack of blood flow called necrotizing enterocolitis) were not different between groups.

Certainty of evidence

According to GRADE (a method used to score the certainty of trials supporting each outcome), high‐certainty evidence shows that indomethacin is effective in closing a PDA in a preterm infant with symptoms. However, there is not enough evidence to say if indomethacin helps to reduce death or lung injury. Also, there is not enough evidence to say if indomethacin causes increased harm compared to no treatment or placebo. More studies are needed to see if indomethacin can reduce death or lung injury in infants with symptomatic PDA.

Summary of findings

Summary of findings 1. Indomethacin compared to placebo or control for symptomatic patent ductus arteriosus in preterm infants.

Indomethacin compared to placebo or control for symptomatic patent ductus arteriosus in preterm infants
Patient or population: symptomatic patent ductus arteriosus in preterm infants Setting: neonatal intensive care setting Intervention: indomethacin Comparison: placebo or control
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
	Risk with placebo or control	Risk with indomethacin
Failure of PDA closure within 1 week of administration of the first dose of indomethacin	Study population		RR 0.30 (0.23 to 0.38)	654 (10 RCTs)	⊕⊕⊕⊕ HIGH
	732 per 1000	220 per 1000 (168 to 278)
Bronchopulmonary dysplasia at 28 days	Study population		RR 1.45 (0.60 to 3.51)	55 (1 RCT)	⊕⊕⊝⊝ LOWa
	222 per 1000	322 per 1000 (133 to 780)
Bronchopulmonary dysplasia at 36 weeks' gestation	Study population		RR 0.80 (0.41 to 1.55)	92 (1 RCT)	⊕⊕⊝⊝ LOWa
	313 per 1000	250 per 1000 (128 to 484)
All‐cause neonatal mortality before hospital discharge	Study population		RR 0.78 (0.46 to 1.33)	314 (8 RCTs)	⊕⊕⊕⊝ MODERATEb
	164 per 1000	128 per 1000 (75 to 217)
Proportion of infants requiring surgical ligation of PDA	Study population		RR 0.66 (0.33 to 1.29)	275 (7 RCTs)	⊕⊕⊕⊝ MODERATEb
	113 per 1000	75 per 1000 (37 to 146)
NEC (≥ Bell stage 2)	Study population		RR 1.27 (0.36 to 4.55)	147 (2 RCTs)	⊕⊕⊝⊝ LOWc
	53 per 1000	68 per 1000 (19 to 243)
Mucocutaneous or gastrointestinal bleeding	Study population		RR 0.33 (0.01 to 7.58)	119 (2 RCTs)	⊕⊕⊝⊝ LOWc
	16 per 1000	0 per 1000 (0 to 0)
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; NEC: necrotizing enterocolitis; OR: odds ratio; PDA: patent ductus arteriosus; RCT: randomized controlled trial; RR: risk ratio.
GRADE Working Group grades of evidence. High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aDowngraded two levels for very serious imprecision (there are few events in a small sample single RCT, and the CI includes appreciable benefit and harm).

^bDowngraded one level for serious imprecision (the CI includes appreciable benefit and harm).

^cDowngraded two levels for very serious imprecision (there are few events from two small sample RCTs, and the CI includes appreciable benefit and harm).

Background

Description of the condition

Fetal circulation relies on the placenta, as well as on a patent ductus arteriosus (PDA) (Matthew 1998). The ductus arteriosus connects the main pulmonary artery to the proximal descending aorta, allowing the vast majority of right ventricular output to bypass the pulmonary circulation (Clyman 2000). Shortly after birth, with initiation of breathing and separation of the low‐resistance placenta, functional closure of the ductus arteriosus begins. Physiological mechanisms for closure include increased oxygen tension and decreased circulating prostaglandin (PGE₂) and prostacyclin (PGI₂). This generally occurs within 24 to 72 hours of birth in the term infant (Clyman 2000). However, the ductus arteriosus frequently fails to close in the preterm infant, with an inverse relationship between gestational age and ductal patency (Clyman 2000; Hammerman 1995). Seventy per cent of infants born before 28 weeks' gestational age have historically received medical or surgical closure of the PDA (Clyman 2000). Infants with respiratory distress syndrome (RDS) (Thibeault 1975), as well as those with perinatal asphyxia (Cotton 1981), are more likely to have a significant PDA present, as are infants who receive large volumes of fluid early in their life (Bell 1980). The likelihood of spontaneous closure of a PDA in infants less than or equal to 1500 grams remains high, reaching 85% in one cohort before discharge (Semberova 2017).

Failure of the ductus arteriosus to close in preterm infants is partially related to altered physiological mechanisms, including increased ductal sensitivity to the vasodilatory effects of prostaglandins and nitric oxide (Dice 2007). Additionally, tone and muscle fibers are decreased in the ductus of preterm infants compared to term infants (Hammerman 1995). RDS can further exacerbate these mechanisms, increasing the likelihood of ductal patency (Hammerman 1995).

Although several clinical findings may be noted, none is entirely specific to a hemodynamically significant PDA. Although often associated with a continuous, 'machine‐like' murmur best heard in the left infraclavicular region, a murmur can be purely systolic or absent, particularly in the first few days of life (Ellison 1983). In cases of hemodynamically significant PDA with an absent murmur, the infant often presents with respiratory distress or cardiac enlargement on chest radiograph (Hammerman 1986). Other common clinical findings include a hyperdynamic precordial impulse, tachycardia, bounding pulses, widened pulse pressure, worsening respiratory status, and apnea (Ellison 1983).

In the late 1970s, it was recognized that preterm infants without clinical findings of a PDA had lower risk for RDS, bronchopulmonary dysplasia (BPD), and death when compared to similar infants with signs or symptoms of PDA (Brown 1979). Both surgical and pharmacological interventions were shown to be possible and potentially beneficial for preterm infants with a significant left‐to‐right ductal shunt (Cotton 1978; Friedman 1976; Heymann 1976; Kitterman 1972). Although systolic blood pressure may be sustained despite a left‐to‐right ductal shunt, the drop in diastolic blood pressure and localized vasoconstriction are believed to contribute to the clinical consequences of PDA (Clyman 2000). A significant left‐to‐right shunt secondary to the ductus has been associated with complications that include increased rates of BPD (Brown 1979), intraventricular hemorrhage (IVH) (Ballabh 2010), and necrotizing enterocolitis (NEC) (Dollberg 2005), as well as decreased middle cerebral artery blood flow (Weir 1999), worsening RDS (Jacob 1980), and death (Dice 2007). However, a precise causal link between these associations has not been demonstrated (Benitz 2010).

Description of the intervention

For a hemodynamically significant PDA that does not spontaneously close, a medical or surgical intervention may be chosen to achieve ductal closure. Procedural interventions include surgical ligation and transcatheter occlusion. Pharmacological agents include cyclooxygenase inhibitors, such as ibuprofen or indomethacin, and acetaminophen (paracetamol), which is a derivative of acetanilide with weak anti‐inflammatory properties.

Current data are inconclusive regarding the comparative efficacy of surgery or medical management as initial treatment for a PDA in preterm infants (Malviya 2013). In clinical practice, surgical ligation or occlusion is frequently used following failure of medical management. Surgical ligation of the ductus is associated with reduced mortality, but surviving infants are at increased risk of neurodevelopmental impairment. Studies addressing survival bias and confounding by indication are lacking (Weisz 2014).

Cyclooxygenase inhibitors (including indomethacin and ibuprofen) have been used to treat preterm infants with symptomatic PDA following the finding in both observational and randomized controlled trials (RCTs) that indomethacin increases the PDA closure rate (Friedman 1976; Gersony 1983; Heymann 1976). Although numerous studies have indicated that prophylactic closure of the ductus with indomethacin reduces the incidence of severe IVH, hemodynamically significant PDA, and surgical ligation, complications of indomethacin therapy have been noted, including increased risk of bleeding and transient renal insufficiency (Davis 1990; Fowlie 2010; Gersony 1983; Stavel 2017; Yeh 1981a). Several studies report NEC as an adverse outcome of indomethacin treatment (Fujii 2002; Grosfeld 1996). However, this association was not found in a systematic review of prophylactic intravenous indomethacin (Fowlie 2010). Systematic reviews have indicated that ibuprofen may offer similar efficacy to indomethacin, with lower rates of transient renal insufficiency and NEC, although optimal dosing, duration, and timing of both therapies are uncertain (Noori 2009; Ohlsson 2020).

Regarding cyclooxygenase inhibitors, several approaches have been taken for PDA treatment, including prophylactic medical therapy within the first 24 hours of life for at‐risk infants (Fowlie 2010; Ohlsson 2020), treatment following PDA diagnosis, and symptomatic treatment based on clinical and/or echocardiographic criteria to define hemodynamic significance (Clyman 1996).

This Cochrane Review will evaluate treatment of symptomatic PDA with indomethacin in the preterm infant. A previously published review has evaluated the role of ibuprofen in PDA treatment (Ohlsson 2020).

How the intervention might work

Prostaglandins play a key role in maintaining ductal muscle relaxation (Clyman 1977). Indomethacin is a non‐selective cyclooxygenase inhibitor that prevents the enzymatic process leading to production of prostaglandins. Inhibition of prostaglandin production results in arteriolar vasoconstriction, which aids ductal closure. Prophylactic treatment aims to achieve ductal closure before development of significant left‐to‐right shunting and the clinical consequences of hemodynamic instability. PDA closure rates with prophylactic use of ibuprofen and indomethacin approach 58% and 57%, respectively (Fowlie 2010; Ohlsson 2020). However, exposure to these drugs has revealed potential side effects and unclear benefits (Benitz 2010). Although benefits are associated with indomethacin use, treatment with indomethacin may reduce perfusion to certain organs. This includes reduced cerebral blood flow, which may oppose neurodevelopmental outcomes associated with decreased IVH (Edwards 1990), decreased intestinal perfusion potentially increasing the rate of NEC (Coombs 1990), and decreased renal perfusion (Cifuentes 1979). Indomethacin may also lead to bleeding via inhibition of platelet function (Friedman 1976). Rather than medically treating all preterm infants at risk or with trivial symptoms attributed to a PDA, targeting treatment only to infants with hemodynamically significant PDA (by clinical and/or echocardiographic/Doppler criteria) may improve the risk‐to‐benefit ratio and limit known adverse exposure to those with the greatest potential to benefit from therapy.

Why it is important to do this review

A Cochrane Review on prophylactic use of indomethacin for medical treatment of PDA in the preterm infant examined studies of asymptomatic preterm infants with PDA and reported only short‐term benefit, with no benefit or harm for long‐term neurodevelopment identified (Fowlie 2010). Benefit may be increased relative to adverse consequences for treatment of symptomatic preterm infants with PDA when compared to prophylactic use. For this Cochrane Review, we will include only studies of indomethacin for preterm infants with a symptomatic PDA.

A Cochrane Review assessed the duration of indomethacin therapy for PDA treatment in preterm infants (Herrera 2007). A prolonged course of indomethacin did not seem to significantly impact clinical outcomes. For this Cochrane Review, we will not assess duration of therapy but will examine differences in dosing.

A non‐Cochrane review by Nehgme and colleagues assessed six small randomized trials of intravenous indomethacin for symptomatic PDA (Nehgme 1992). The meta‐analysis suggests that indomethacin significantly increased rates of PDA closure, without differences in mortality, IVH, NEC, or chronic lung disease (Nehgme 1992). In this Cochrane Review, we will perform an updated examination of the literature.

Objectives

To determine the effectiveness and safety of indomethacin (given by any route) compared to placebo or no treatment in reducing mortality and morbidity in preterm infants with a symptomatic patent ductus arteriosus (PDA).

Methods

Criteria for considering studies for this review

Types of studies

We included all published and unpublished RCTs, quasi‐RCTs, and cluster‐RCTs, and randomized cross‐over trials. Both superiority trials and non‐inferiority trials were eligible for inclusion.

Types of participants

We included preterm infants born at less than 37 weeks' gestation and low birth weight infants (< 2500 grams) treated for symptomatic PDA enrolled within the first 28 days of life.

Symptomatic PDA was defined by clinical and/or echocardiographic criteria. Clinical criteria included one or more of a characteristic heart murmur, hyperdynamic precordial impulse, tachycardia, bounding pulses, widened pulse pressure, or worsening respiratory status (Davis 1995). Echocardiographic criteria included one or more of a transductal diameter greater than 1.5 mm, left atrial aortic root ratio (LA:Ao) greater than 1.3, a left‐to‐right shunt, and “disturbed diastolic flow in the main pulmonary artery with diastolic backflow in the aorta immediately below the ductus arteriosus and forward flow above the ductal insertion” (Lago 2002).

We excluded infants with other known congenital heart disease and infants who received prior treatment with a cyclooxygenase inhibitor (indomethacin, ibuprofen).

Types of interventions

Indomethacin (any dose, any route) versus placebo or no treatment.

Types of outcome measures

Primary outcomes

1. Failure of PDA closure within one week of administration of the first dose of indomethacin

2. Bronchopulmonary dysplasia (defined as supplemental oxygen need at 28 days' postnatal age (Bancalari 1979) or supplemental oxygen at 36 weeks' postmenstrual age with or without compatible clinical and radiographic findings (Shennan 1988))

3. All‐cause neonatal mortality at 28 days and before hospital discharge

Secondary outcomes

PDA‐related outcomes

1. Proportion of infants receiving rescue medical treatment (repeated cyclooxygenase or paracetamol/acetaminophen dosing, or both). This outcome was revised post‐hoc to include any 'open‐label' treatment and was defined as any cyclooxygenase treatment provided after the initial assigned treatment

2. Proportion of infants requiring surgical ligation or transcatheter occlusion

Other outcomes

1. Pneumothorax

2. Pulmonary hemorrhage

3. Late‐onset bacterial sepsis

4. Necrotizing enterocolitis (NEC) (≥ Bell stage 2) (Bell 1978)

5. IVH (intraventricular hemorrhage) (any grade) (Papile 1978)

6. Severe IVH (grade III to IV) (Papile 1978)

7. Periventricular leukomalacia

8. Retinopathy of prematurity (ROP) (any stage) (ICROP 2005)

9. Severe retinopathy of prematurity (severe ROP) (≥ stage III)

10. Surgery for severe ROP (added post hoc)

11. Infant mortality (first year of life)

12. Use of inotropic agents

13. Duration of assisted ventilation (days)

14. Duration of oxygen dependence (days to last discontinuation of any supplemental oxygen)

15. Duration of hospital stay (days)

16. Time to full enteral feeds (days)

17. Cerebral palsy at approximately 2 years' corrected age (as defined by study authors)

18. Neurodevelopmental outcome at approximately 2 years' corrected age (acceptable range 18 months to 28 months) including cerebral palsy, delayed neurodevelopment (Bayley Scales of Infant Development Mental Developmental Index < 70) (Bayley 2006), legal blindness (< 20/200 visual acuity), and hearing deficit (aided or < 60 decibels on audiometric testing). The composite outcome 'neurodevelopmental impairment' was defined as having any one of the aforementioned deficits

Safety outcomes (harms reported within one week of completion of the intervention)

1. Intestinal perforation

2. Renal function

   1. Oliguria (< 1 mL/kg/hour)

   2. Serum/plasma creatinine (µmol/L) levels during treatment

   3. Serum/plasma creatinine (µmol/L) after treatment

   4. Post‐hoc: creatinine > 150 micromole during treatment (post‐hoc)

3. Hemostasis

   1. Mucocutaneous or gastrointestinal bleeding

   2. Platelet count (< 50,000 platelets/µL blood)

4. Pulmonary hypertension (diagnosed by echocardiographic or Doppler criteria)

Search methods for identification of studies

We used criteria and standard methods of Cochrane and Cochrane Neonatal (see the Cochrane Neonatal search strategy for specialized register). We searched for errata or retractions from included studies published in full text on PubMed (www.ncbi.nlm.nih.gov/pubmed), and, if noted, we reported the date this was done.

Electronic searches

We conducted a comprehensive search including Cochrane Central Register of Controlled Trials (CENTRAL; 2020, Issue 7), in the Cochrane Library; Ovid MEDLINE(R) and Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, Daily and Versions(R) (1946 to 31 July 2020); and Cumulative Index to Nursing and Allied Health Literature (CINAHL; 1981 to 31 July 2020). We included the search strategies for each database in Appendix 1. We did not apply language restrictions.

We searched clinical trial registries for ongoing and recently completed trials. We searched the World Health Organization International Clinical Trials Registry Platform (ICTRP) (www.who.int/ictrp/search/en/), as well as the US National Library of Medicine ClinicalTrials.gov (clinicaltrials.gov), via Cochrane CENTRAL. Additionally, we searched the ISCRTN Registry (i.e. International Standard Randomized Controlled Trials Number) for any unique trials not found through the Cochrane CENTRAL search.

Searching other resources

We also searched the reference lists of any articles selected for inclusion in this review to identify additional relevant articles.

Data collection and analysis

Selection of studies

Working in pairs, review authors independently screened search results by title and abstract for studies that potentially met the inclusion criteria. We obtained the full text of any articles that were potentially eligible, and at least two review authors independently performed full‐text assessments. We resolved any disagreements by discussion between the review author team. We recorded the selection process in sufficient detail to complete a PRISMA flow diagram (Moher 2009), and we listed in a Characteristics of excluded studies table all studies excluded after full‐text assessment.

Data extraction and management

At least two review authors (of PE, JF, DO, and SM) extracted, assessed, and coded all data for each study, using a form designed specifically for this review. We replaced any standard error of the mean by the corresponding standard deviation. We resolved any disagreements by discussion. For each study, one review author (PE) entered extracted data into Review Manager 5 (RevMan 5) (Review Manager 2020); a second review author (SM) checked data entry. If there was a conflict regarding interpretation of data, another review author (RS) was brought in to adjudicate. All review authors assessed the protocol, analysis, and draft manuscript.

We collected information regarding the method of randomization, blinding, drug intervention, stratification, and whether the trial was single or multi‐center for each included study. We noted information regarding trial participants including gestational age criteria, birth weight criteria, and other inclusion or exclusion criteria. We analyzed the information on clinical outcomes of the primary and secondary outcomes.

Assessment of risk of bias in included studies

Two review authors (PE and SM) independently assessed risk of bias (low, high, or unclear) of all included trials using the Cochrane ‘Risk of bias’ tool for the following domains (Higgins 2011).

1. Sequence generation (selection bias).

2. Allocation concealment (selection bias).

3. Blinding of participants and personnel (performance bias).

4. Blinding of outcome assessment (detection bias).

5. Incomplete outcome data (attrition bias).

6. Selective reporting (reporting bias).

7. Any other bias.

We resolved any disagreements by discussion or by consultation with a third review author. See Appendix 2 for a more detailed description of risk of bias for each domain.

Measures of treatment effect

We performed statistical analyses using RevMan 5 (Review Manager 2020). We analyzed categorical data using the risk ratio (RR) and the risk difference (RD). For statistically significant outcomes, we calculated the number needed to treat for an additional beneficial outcome (NNTB) or the number needed to treat for an additional harmful outcome (NNTH). We analyzed continuous data using the mean difference (MD) and the standardized mean difference (SMD). We reported the 95% confidence interval (CI) for all estimates.

Unit of analysis issues

The unit of analysis was the individual infant. We considered an infant only once in an analysis. We excluded infants with multiple enrollments unless we could obtain data from the report or the investigators related to the first episode of randomization. If we could not separate data from the first randomization, we excluded the study, as we were not be able to address unit of analysis issues that arise from multiple enrollments of the same infant.

We conducted intention‐to‐treat analyses.

We intended to use the participating neonatal unit or section of a neonatal unit as the unit of analysis in cluster‐randomized trials. We planned to analyze these using an estimate of the intracluster correlation coefficient derived from the trial (if possible) or from another source, as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2019).

If we identified both cluster‐RCTs and individual RCTs, we planned to combine results from both only if there was little heterogeneity between study designs, and if the interaction between the effect of an intervention and the choice of a randomization unit was considered unlikely. We identified no cluster trials in our review.

Dealing with missing data

We attempted to obtain data from the primary author if published data provided inadequate information for the review.

Assessment of heterogeneity

We estimated treatment effects of individual trials and examined heterogeneity among trials by inspecting the forest plots and quantifying the impact of heterogeneity using the I² statistic. We assessed the degree of heterogeneity as:

1. < 25% = no heterogeneity;

2. 25% to 49% = low heterogeneity;

3. 50% to 75% = moderate heterogeneity; and

4. > 75% = substantial heterogeneity.

If statistical heterogeneity was present (I² statistic value > 50%), we explored possible causes (e.g. differences in study quality, participants, intervention regimens, or outcome assessments).

Assessment of reporting biases

If we suspected reporting bias, we contacted trial investigators to request missing outcome data. If this was not possible, and we considered the missing data to have introduced serious bias, we planned to explore the impact of including such trials in the overall assessment of results by conducting a sensitivity analysis.

Data synthesis

In studies judged to be sufficiently similar, we performed meta‐analysis using RevMan 5 (Review Manager 2020). For categorical outcomes, we calculated typical estimates of RR and RD, each with its 95% CI; and for continuous outcomes, we determined the MD or a summary estimate for the SMD, each with its 95% CI. We used a fixed‐effect model for meta‐analysis. If we considered meta‐analysis to be inappropriate, we analyzed and interpreted individual trials separately.

Subgroup analysis and investigation of heterogeneity

Subgroup analyses

1. Gestational age at birth (< 28 weeks' gestation, 28 to 32 completed weeks' gestation, 33 to 36 completed weeks' gestation)

2. Birth weight (< 1000 grams, 1000 to 1500 grams, 1501 to 2500 grams)

3. Chronological age (< 3 days, 3 to 7 days; > 7 days)

4. Route of administration (intravenous indomethacin, oral indomethacin)

5. Higher‐dose versus lower‐dose indomethacin: low dose (≤ 0.4 mg/kg/dose) versus high dose (> 0.4 mg/kg/dose)

6. Cumulative indomethacin dose: standard or lower cumulative dose (≤ 0.6 mg/kg) versus higher cumulative dose (> 0.6 mg/kg)

7. Method of PDA diagnosis (clinical diagnosis, echocardiographic diagnosis, Doppler diagnosis)

8. Trial methods (low bias, high bias)

9. Co‐interventions (antenatal steroids, postnatal surfactant therapy)

Sensitivity analysis

We conducted sensitivity analyses to determine whether or not findings were affected by inclusion of only studies that reported adequate methods (low risk of bias), defined as adequate randomization and allocation concealment, blinding of intervention and measurement, and up to and including 10% loss to follow‐up.

Summary of findings and assessment of the certainty of the evidence

We used the GRADE approach, as outlined in the GRADE Handbook (Schünemann 2013), to assess the certainty of the evidence for the following (clinically relevant) outcomes.

Efficacy

1. Failure of PDA closure within one week of administration of the first dose of indomethacin

2. Proportion of infants requiring surgical ligation or transcatheter occlusion

3. Bronchopulmonary dysplasia defined as supplemental oxygen need at 28 days postnatal age (Bancalari 1979)

4. Bronchopulmonary dysplasia defined as supplemental oxygen need at 36 weeks' postmenstrual age with or without compatible clinical and radiographic findings (Shennan 1988))

5. All‐cause neonatal mortality at 28 days and prior to hospital discharge

6. NEC (≥ Bell stage 2) (Bell 1978)

Safety outcomes (harms)

1. Mucocutaneous or gastrointestinal bleeding

Two review authors (SM, RS) independently assessed the certainty of the evidence for each of the outcomes above. We considered evidence from RCTs as high certainty but downgraded the certainty of the evidence by one level for serious (or two levels for very serious) limitations based upon the following: design (risk of bias), consistency across studies, directness of the evidence, precision of estimates, and presence of publication bias. We used the GRADEpro Guideline Development Tool to create Table 1 to report the certainty of the evidence (GRADEpro GDT).

The GRADE approach results in an assessment of the certainty of a body of evidence as one of four grades.

1. High certainty: further research is very unlikely to change our confidence in the estimate of effect.

2. Moderate certainty: further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.

3. Low certainty: further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.

4. Very low certainty: we are very uncertain about the estimate.

Results

Description of studies

Results of the search

The literature search on 31 July 2020 yielded 1085 studies. An ongoing clinical trial was identified in the search of clinical trial registries (NCT03456336), along with one study awaiting classification (Clyman 2019) (Figure 1). See Characteristics of ongoing studies and Characteristics of studies awaiting classification .

1.

Study flow diagram.

Included studies

We included in this review 14 clinical trials that enrolled 880 infants. Individual study characteristics, inclusion criteria, treatment details, outcome details, and study author information can be found in the Characteristics of included studies table.

Cotton 1980 conducted a single‐center RCT to compare the efficacy of intravenous indomethacin versus standard medical management. Ventilator‐dependent preterm infants with symptomatic PDA were randomized to receive intravenous indomethacin without digoxin (n = 9) or a standard protocol that included anti‐congestive measures and digoxin (n = 9). Indomethacin treatment was allowed 6 to 19 days after study enrollment if infants remained ventilator‐dependent. Demographic characteristics did not differ between groups. Infants who had received indomethacin experienced high rates of ductal closure within 24 hours of administration (100% versus 37.5%) and achieved extubation within 72 hours (100% versus 37.5%).

Gersony 1983 conducted a multi‐center RCT to compare the efficacy of indomethacin with standard medical management versus medical management alone for treatment of symptomatic PDA. The study was conducted at 13 tertiary care centers as part of the National Collaborative Study on Patent Ductus Arteriosus. Infants weighing less than 1750 grams with evidence of a hemodynamically significant PDA were enrolled (n = 421). Infants were randomly assigned to receive three doses of intravenous indomethacin while the control group received standard medical therapy, which included fluid restriction, diuretic use, and occasional digoxin. Infants receiving standard medical therapy were given indomethacin or underwent surgical ligation if they failed to improve within 36 to 48 hours after initiation of therapy. Permanent PDA closure occurred at a higher rate among infants who received indomethacin (79% versus 28.1%). As most patients assigned to standard medical therapy received indomethacin after failure of ductal closure, other outcomes were not considered in this analysis.

Peckham 1984 conducted one‐year follow‐up of patients enrolled in Gersony 1983. Two hundred seventy‐one infants were examined at one year, allowing comparisons between three treatment strategies: (1) immediate administration of IV indomethacin in addition to standard medical therapy with surgery as a backup measure, (2) standard medical therapy alone initially, with indomethacin as the backup followed by surgery, and (3) standard medical management with surgery alone as backup. Cumulative mortality from trial entry to one year was 21% and did not vary between treatment strategies. No significant differences in Bayley scores or rates of BPD were noted between groups. The rate of ROP was lower in those who received indomethacin (2% and 4%) compared to those who could receive surgery alone (10%).

Kluckow 2014 completed a multi‐center, double‐blind RCT to compare the efficacy of indomethacin versus placebo for infants diagnosed with a large PDA. This study was conducted at three Australian neonatal intensive care units (NICUs). Infants born at less than 29 weeks who had a screening cardiac ultrasound demonstrating a large PDA were included (n = 92). Infants were randomly given three doses of indomethacin (n = 44) or three doses of placebo (n = 48). No differences in death, BPD, NEC, or sepsis were noted between groups. A trend towards less IVH in the indomethacin group was evident (4.5% versus 12.5%), although this did not reach statistical significance. The study abstract was published in 2012 as a conference proceeding that showed identical results as mentioned above (Kluckow 2012). Varghese and colleagues reported two‐ to three‐year neurodevelopmental outcomes of randomized infants as a conference proceeding. Follow‐up data were available for 137/149 (92%) surviving infants at mean corrected age of 25 months. Neurodevelopmental outcomes were assessed using the Bayley III Scale for all but five infants who had Griffiths scores (Griffiths Mental Development Scales ‐ Extended Revised (GMDS‐ER)). Moderate/severe cognitive delay was seen in 2.6% of infants randomized to the indomethacin group versus 9.7% in the placebo group. Moderate/severe motor delay was noted in 2.6% of infants in the indomethacin group versus 4.8% in the placebo group, and moderate/severe language delay was found in 5.2% versus 9.7% of infants, respectively (Varghese 2016).

Knight 2011 conducted a feasibility study comparing the efficacy of IV indomethacin versus placebo for infants with a hemodynamically significant PDA. Infants were enrolled if they were at less than 28 weeks' gestational age or weighed less than 1250 grams with a PDA ≥ 1.5 mm and reversed descending aortic diastolic flow. The IV indomethacin group (n = 13) received indomethacin 200/100 mcg/kg IV every 24 hours for six days, and the control group (n = 13) received placebo. No major differences in patient demographics were noted, nor were any differences in peak inspiratory pressure and inspired oxygen for either group during therapy. This study contributed no data to the meta‐analysis.

Krauss 1989 conducted a randomized trial to compare the efficacy of intravenous indomethacin for PDA treatment versus no intervention. Newborn infants (n = 27) weighing less than 1500 grams, ventilator‐dependent at 24 hours of age, with a physiologically significant PDA were enrolled. Infants were randomly assigned to receive three doses of intravenous indomethacin between 72 and 96 hours of life or no treatment. Infants who received indomethacin had a tendency for higher PDA closure rates within one week of administration of indomethacin (58% versus 13%). No differences were reported in all‐cause mortality at discharge, the proportion of infants requiring surgical ligation of PDA, or the duration of oxygen dependence.

Merritt 1981 conducted an RCT in the NICU of the University of Rochester to compare the efficacy of intravenous indomethacin and standard medical management to achieve closure of a hemodynamically significant PDA. Newborn infants weighing 1350 grams or less who had severe respiratory distress and ventilatory failure requiring intermittent mandatory ventilation (IMV) during the first hours of life were considered eligible for inclusion. Those with clinical and radiographic signs of PDA were enrolled, and infants were randomized to receive 0.2 mg/kg indomethacin (n = 12) every 24 hours for up to three days or medical management (n = 13), including fluid restriction and/or furosemide 1 to 2 mg/kg every 12 hours. The medical management group was not excluded from receiving indomethacin nor from undergoing surgical ligation after therapies failed to eliminate pulmonary hyperperfusion through the PDA. Although rates of mortality were higher in the control group (30.8% versus 9.1%), they did not reach the level of statistical significance. A difference in BPD was reported between control and indomethacin groups (66.7 versus 18.2%). In the control group, 11/13 infants developed symptoms of congestive heart failure (CHF) and required indomethacin rescue at a mean age of 167.4 hours of life. No significant difference in duration of assisted ventilation was evident between control and intervention groups (7.3 ± 2.5 days versus 7.6 ± 6.05 days).

Monset‐Couchard 1983 conducted a randomized trial comparing intravenous indomethacin versus no intervention for treatment of symptomatic PDA in preterm infants. In this study, 24 preterm infants with symptomatic PDA were randomized to receive either 0.2 mg IV indomethacin just after diagnosis (n = 12) or no treatment for 48 hours (n = 12). Among 12 infants in the intervention group, the PDA closed within 4 to 26 hours of the first dose in eight cases. In all 12 untreated infants, the PDA persisted for at least 48 hours. Ten of the 12 untreated infants subsequently received IV (n = 5) or oral (n = 5) indomethacin.

Nestrud 1980 conducted a single‐center RCT to compare the efficacy of indomethacin versus placebo for treatment of hemodynamically significant PDA. The study was conducted in the NICU at University of Arkansas for Medical Sciences or at St. Vincent’s Infirmary, Little Rock, Arizona, USA. Newborn infants (n = 23) at less than 35 weeks' gestational age (GA) with large left‐to‐right PDA determined by clinical signs, radiography, and echo or cardiac catheterization were enrolled. Infants were randomized to receive oral indomethacin (n = 12) or placebo (n = 11). The indomethacin group received 0.2 mg/kg every 12 hours up to three doses, and the placebo group received the same volume of saline at the same interval. PDA closure occurred at a higher rate among infants receiving indomethacin compared to those receiving placebo (58% versus 18%).

Neu 1981 conducted a single‐center, randomized cross‐over study to compare the efficacy of oral indomethacin versus placebo for closure of clinically significant PDA. Infants (n = 21) with RDS and a clinically significantly PDA who had failed to respond to fluid restriction were enrolled. Infants were randomly assigned to up to two oral doses of 0.25 mg/kg indomethacin or placebo. If infants demonstrated signs of a clinically significant PDA at 48 hours, they received the alternative treatment one time. Infants receiving indomethacin had a PDA closure rate higher than those receiving placebo before cross‐over (80% versus 27.2%).

Osborn 2003 conducted a multi‐center RCT to compare the efficacy of intravenous indomethacin and placebo in maintaining superior vena cava (SVC) flow rates in preterm infants with significant PDA. Newborn infants (n = 70) born at gestational age < 30 weeks, < 12 hours of life, with a PDA Doppler diameter > 1.6 mm were eligible. Infants randomly received a single dose of 0.2 mg/kg intravenous indomethacin (n = 35) or saline (n = 35). Infants were evaluated by echocardiography one hour following infusion, and were crossed over to the other treatment if PDA remained > 1.6 mm and if < 30% constriction from initial diameter occurred. No significant differences in mean per cent change in DA diameter (‐20% versus –15%), SVC flows (‐3.2 to 10.5 mL/kg/min), or change in mean blood pressure (+1.4 mmHg versus +0.1 mmHg) were noted one hour after infusion.

Rudd 1983 conducted a single‐center RCT to evaluate the efficacy of oral indomethacin compared to placebo to achieve closure of hemodynamically significant PDA. Newborn infants (n = 15) weighing < 1500 grams with a clinically significant PDA were fluid‐restricted for 24 hours and were randomized to receive up to three doses (0.2 mg/kg) of oral indomethacin (n = 15) or placebo (n = 15). PDA closure occurred in 13/15 (87%) of those receiving indomethacin compared to 3/15 (20%) of those receiving placebo. Six infants receiving indomethacin had relapses at a mean of eight days, tended to weigh less and have a lower GA, and were younger at time of treatment. Infants who failed placebo or relapsed could be given indomethacin on an open basis. 6/11 (54.5%) of infants in the control group achieved ductal closure following open indomethacin administration. Of the 15 infants who did not respond to indomethacin, nine achieved spontaneous ductal closure, and three required surgical closure. Three infants died before PDA closure.

Valaes 1980 conducted a single‐center RCT to compare the efficacy of oral indomethacin and placebo for closure of a symptomatic PDA. Preterm infants (n = 13) with evidence of a symptomatic PDA were randomized to receive up to three doses (0.2 mg/kg) oral indomethacin (n = 7) or placebo (n = 6). Among patients receiving indomethacin, 3/7 achieved PDA closure and the remainder became asymptomatic; 2/6 patients receiving placebo had ductal closure and the other 4 remained symptomatic. The four patients were given indomethacin at an unspecified time; three patients achieved ductal closure, and one became asymptomatic.

Yanagi 1981 conducted an RCT to compare the efficacy of oral indomethacin versus placebo for treatment of symptomatic PDA. Thirty‐nine preterm infants with RDS requiring ventilator support with symptomatic PDA (defined as left atrial‐to‐aortic root ratio ≥ 1.3) who had failed a 24‐hour course of standardized medical management were randomized to indomethacin (0.2 mg/kg of enteral indomethacin) or placebo in a double‐blind manner. Second and third indomethacin doses were administered at 24‐hour intervals in phase 1 (n = 17), and at eight‐hour intervals in phase 2 (n = 22). The PDA closure rate in phase 1 was 75% in the indomethacin group versus 44% in the placebo group. In phase 2, the PDA closure rate in the indomethacin group was 85% versus only 11% in the placebo group (P < 0.01). In this review, we have synthesized study results by combining results from the two phases.

Yeh 1981a conducted an RCT to compare the effects of intravenous indomethacin versus placebo for closure of significant PDA. Infants (n = 55) weighing < 2040 grams or transferred to the study site with evidence of cardiovascular dysfunction due to a PDA were included after failure of medical management. Ductal closure was achieved at a higher rate among those receiving indomethacin than in those receiving placebo (92.9% versus 37.0%), and surgical ligation rates were lower (3.6% versus 29.5%). Rates of mortality before discharge, NEC, and late‐onset sepsis were similar between groups.

Yeh 1982b conducted one‐year follow‐up of infants enrolled in Yeh 1981a. Three further infants in both indomethacin and control groups died at one year (32.1% versus 29.6%). Of the 38 surviving infants, 30 (indomethacin = 13, control = 17) presented for one‐year follow‐up. It was noted that there had been further surgical ligation, including one case in the indomethacin group and four in the control group. Two infants in the control group had Mental Developmental Index (MDI) and Psychomotor Developmental Index (PDI) scores < 50. None in the indomethacin group had Bayley scores < 70. Study authors categorized survivors as having a composite major and minor neurological deficit; 5/13 (38%) of those receiving indomethacin and 7/17 (41%) of those receiving control had this composite outcome. Additionally, three (23%) infants receiving indomethacin had abnormal electroencephalogram (EEG) findings compared to seven (41%) in the control group. The mean MDI and PDI scores for the indomethacin group were 100.8 ± 24.6 and 92.4 ± 16.9. The group receiving placebo had an MDI of 97.7 ± 29.6 and a PDI of 103.4 ± 29.2.

Yeh 1983 reported on the incidence of ROP in 47 infants from the Yeh 1981a trial. 2/24 (8%) infants receiving indomethacin and 6/23 (26%) receiving placebo developed ROP. Severe ROP developed in 1/24 (4%) receiving indomethacin and in 1/23 (4%) of those receiving saline.

Yeh 1982c reported on plasma glucose levels following indomethacin administration in 47 infants of the Yeh 1981a cohort. No statistical differences in baseline plasma glucose or daily caloric or IV glucose intake before and during the study were noted. Infants receiving indomethacin (n = 25) had plasma glucose of 70.8 ± 3.9 at 24 hours and 70.9 ± 5.9 at 49 hours, and those receiving placebo (n = 22) had plasma glucose of 94.2 ± 3.4 at 24 hours and 90.4 ± 4.1 at 48 hours.

Yeh 1981b reported on left atrium (LA): aorta at the level of the aortic valve (Ao), left ventricular end‐diastolic diameter (LVEDD) on echocardiogram, change in tidal volume, and lung compliance in 11 infants of the Yeh 1981a cohort who did not require mechanical ventilation. Significant pretreatment and post‐treatment changes in tidal volume, lung compliance, LA:Ao, and LVEDD were evident among those who received indomethacin but not in those receiving placebo. No changes in fraction of inspired oxygen (FiO₂) requirements, pH, or base deficit pretreatment and post‐treatment were reported for either group.

Excluded studies

We excluded six studies (see Characteristics of excluded studies).

Carmo 2009 conducted a single‐center, non‐blinded RCT at a perinatal center in Australia. Infants at < 30 weeks' GA, requiring respiratory support and found to have a PDA > 2 mm on echocardiogram performed between 3 and 12 hours of age, were included. All infants were initially given an indomethacin dose of 0.1 mg/kg. Infants were than randomized to receive 0.1 mg/kg indomethacin every 24 hours only if their ductal diameter remained > 1.6 mm, or to receive 0.1 mg/kg every 24 hours up to three total doses. The primary outcome was failure of PDA closure, defined as ductal patency 24 hours after the final dose of indomethacin. No differences between groups in failure of PDA closure, PDA reopening, or need for rescue therapy (medical or ligation) were noted. Fewer doses of indomethacin were given in the targeted therapy group. This study was excluded because infants in both groups received indomethacin.

Mardoum 1991 conducted a single‐center RCT at a perinatal center in the USA. Preterm infants with clinically significant PDA were enrolled if they had adequate arterial catheterization and continuous intra‐arterial monitoring of blood pressure, as well as transcutaneous monitoring of oxygen and carbon dioxide pressure. Infants were randomized to receive intravenous indomethacin 0.2 mg/kg or placebo. Infants were crossed‐over to receive the alternative therapy after 30 minutes. The primary outcome was cerebral blood flow velocity. We excluded this study as infants were given indomethacin regardless of their PDA status after initial treatment.

Mullett 1982 conducted a randomized, double‐blind trial of preterm infants with birth weight < 1750 grams and a systolic murmur consistent with a PDA. Infants received either 0.2 mg/kg/dose of oral indomethacin two doses 24 hours apart or placebo. The primary outcome was PDA closure. We excluded this study as infants with PDA, regardless of symptomatic status, were included.

Nair 1986 conducted a randomized trial at a perinatal center in India. All infants with a PDA diagnosis were included. Infants were randomized to receive either 0.2 mg/kg three times at 12‐hour intervals or to no treatment. The primary outcome was PDA closure. The rate of ductal closure was higher among those receiving indomethacin (60%) compared to those given no treatment (16.7%). We excluded this study as it did not specify whether infants had a clinically significant PDA.

Nuntnarumit 2011 conducted a trial evaluating use of NT‐proBNP levels to guide indomethacin therapy at a single perinatal center in Thailand. Infants at < 33 weeks' GA were enrolled. Infants with N‐terminal pro hormone B‐type natriuretic peptide (NT‐proBNP) > 10,180 pg ⁄ mL on day of life 2 were given three doses of 0.1 mg/kg IV indomethacin at 24‐hour intervals. Of the remaining infants, those who developed clinical symptoms of hemodynamically significant PDA, defined as the presence of left‐to‐right ductal shunting plus at least two clinical signs (heart murmur, heart rate (HR) > 160 beats per minutes (BPM), pulse pressure > 25 mmHg, hyperactive precordium, bounding pulse, hepatomegaly, increasing respiratory support by 20% increase in oxygen supplementation or in pressure support, chest radiographic evidence of cardiomegaly or pulmonary congestion) received three doses of 0.1 mg/kg IV indomethacin at 12‐hour intervals. This course could be repeated if clinical signs persisted. If infants in the NT‐proBNP guided group developed clinical signs attributable to a hemodynamically significant PDA, the treatment interval was moved to every 12 hours. Primary outcomes were incidence of hemodynamically significant PDA and rate of exposure to indomethacin. 19 (38%) infants had NT‐proBNP greater than threshold and were treated with indomethacin. Only 2/19 (10.5%) had no echocardiographic evidence of PDA at the time of NT‐proBNP collection. None of the early treatment group developed significant PDA; 1/31 infants with NT‐proBNP below cutoff developed a significant PDA requiring treatment. This study was not included in this review as it was not randomized.

Yeh 1982a conducted a randomized study at a single center in the USA. Infants with PDA, clinical evidence of cardiovascular dysfunction, and LA:Ao ≥ 1.3 were randomized to receive 0.3 mg/kg IV indomethacin and 1 mg/kg IV furosemide or 0.3 mg/kg IV indomethacin. Primary outcomes included urine output, fractional excretion of sodium and chloride, and glomerular filtration rate (GFR). The same number of infants in each group had PDA closure. Those in the furosemide had higher urine output, fractional excretion of sodium and chloride, and GFR. We excluded this study from this review as both groups received indomethacin.

Risk of bias in included studies

Many studies were conducted over 30 years ago. In subsequent years, methodological expectations have changed, with greater care taken to minimize bias and to report details of trial conduct (Figure 2).

2.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

1. Random sequence generation

Most studies randomized infants using a random numbers generator (Gersony 1983; Krauss 1989; Merritt 1981; Yanagi 1981; Yeh 1981a), or the method of randomization was unclear (Cotton 1980; Monset‐Couchard 1983; Neu 1981; Osborn 2003; Valaes 1980).

Nestrud 1980 randomized by block numbers to receive either oral indomethacin or placebo. Rudd 1983 randomized by number code by batching of 10, and Kluckow 2014 used block randomization by gestational age and by center using a random number‐generating table.

2. Allocation concealment

In most studies, the method of allocation concealment was unclear (Cotton 1980; Krauss 1989; Merritt 1981; Monset‐Couchard 1983: Neu 1981: Osborn 2003; Rudd 1983; Valaes 1980). Gersony 1983 and Yeh 1981a placed allocation inside concealed envelopes. Yanagi 1981 concealed allocation from both physician and pharmacist, and randomization numbers were kept by a pharmacist not involved in patient care in Nestrud 1980.

Blinding

In five studies, staff were not blinded to the drug administered (Cotton 1980; Krauss 1989; Merritt 1981), or blinding was not reported as it is unlikely to have occurred (Monset‐Couchard 1983; Valaes 1980).

Blinding

Other studies were more successful in blinding the intervention using a placebo (Gersony 1983; Nestrud 1980; Neu 1981: Osborn 2003; Rudd 1983; Yanagi 1981; Yeh 1981a). One study stated that staff were blinded to drug administration without communicating details of how this occurred (Kluckow 2014).

Incomplete outcome data

All randomized infants are accounted for in all 12 of the 14 included studies that contributed to the meta‐analysis (Cotton 1980; Gersony 1983; Kluckow 2014; Krauss 1989; Monset‐Couchard 1983; Nestrud 1980; Neu 1981; Rudd 1983; Valaes 1980; Yanagi 1981; Yeh 1981a; Merritt 1981).

Selective reporting

Most trials were conducted before the routine expectation of trials registration (Cotton 1980; Gersony 1983; Krauss 1989; Merritt 1981; Monset‐Couchard 1983; Nestrud 1980; Neu 1981; Osborn 2003; Rudd 1983; Valaes 1980; Yanagi 1981; Yeh 1981a). Therefore, selective bias was difficult to interpret.

Only one study was formally registered, and no deviations from the original protocol were reported (Kluckow 2014).

Other potential sources of bias

None were noted.

Effects of interventions

See: Table 1

Indomethacin versus placebo or control (Comparison 1)

Out of the 14 included trials, 12 trials contributed to the meta‐analysis. Pooled results from these 12 trials are discussed in detail below (Cotton 1980; Gersony 1983; Kluckow 2014; Krauss 1989; Merritt 1981; Monset‐Couchard 1983; Nestrud 1980; Neu 1981; Rudd 1983; Valaes 1980; Yanagi 1981; Yeh 1981a).

Primary outcomes

Failure of PDA closure within one week of administration of the first dose of indomethacin (Outcome 1.1)

Ten studies reported on PDA closure within one week of administration of the first dose of indomethacin (parenteral administration: Cotton 1980; Gersony 1983; Krauss 1989; Monset‐Couchard 1983; Yeh 1981a; enteral administration: Nestrud 1980; Neu 1981; Rudd 1983; Valaes 1980; Yanagi 1981). Many individual studies reported increased rates of PDA closure within one week of indomethacin administration (parenteral administration: Gersony 1983; Krauss 1989; Monset‐Couchard 1983, enteral administration: Neu 1981; Rudd 1983; Yanagi 1981). Overall, indomethacin administration was associated with a large decrease in failure of PDA closure within one week of administration of the first dose of indomethacin (typical risk ratio (RR) 0.30, 95% confidence interval (CI) 0.23 to 0.38; typical risk difference (RD) ‐0.52, 95% CI ‐0.58 to ‐0.45; 10 studies, 654 infants; high‐certainty evidence; Analysis 1.1; Figure 3). This was seen in both studies that utilized enteral (typical RR 0.35, 95% CI 0.23 to 0.54; typical RD ‐0.49, 95% CI ‐0.64 to ‐0.35; 5 studies, 126 infants) or parenteral administration of indomethacin (typical RR 0.28, 95% CI 0.21 to 0.37; typical RD ‐0.52, 95% CI ‐0.60 to ‐0.45; 5 studies, 528 infants).

1.1. Analysis.

Comparison 1: Indomethacin vs placebo or control, Outcome 1: Failure of PDA closure within 1 week of administration of the first dose of indomethacin

3.

Forest plot of comparison: 1 Indomethacin vs placebo or control, outcome: 1.1 Failure of PDA closure within 1 week of administration of the first dose of indomethacin.

Bronchopulmonary dysplasia (defined as supplemental oxygen need at 28 days' postnatal age with or without compatible clinical and radiographic findings) (Outcome 1.2)

Only one study reported on supplemental oxygen need at 28 days' postnatal age with or without compatible clinical and radiographic findings (parenteral administration: Yeh 1981a). In that small study, no impact on the risk of requiring supplemental oxygen at 28 days' postnatal age was reported (RR 1.45, 95% CI 0.60 to 3.51; 1 study, 55 infants; low‐certainty evidence; Analysis 1.2).

1.2. Analysis.

Comparison 1: Indomethacin vs placebo or control, Outcome 2: Bronchopulmonary dysplasia at 28 days

Bronchopulmonary dysplasia (defined as supplemental oxygen at 36 weeks' postmenstrual age with or without compatible clinical and radiographic findings) (Outcome 1.3)

Only one study reported on supplemental oxygen need at 36 weeks' postmenstrual age with or without compatible clinical and radiographic findings (parenteral administration: Kluckow 2014). In that study, no impact on the risk of requiring supplemental oxygen at 36 weeks' postmenstrual age was reported (RR 0.80, 95% CI 0.41 to 1.55; 1 study, 92 infants; low‐certainty evidence; Analysis 1.3).

1.3. Analysis.

Comparison 1: Indomethacin vs placebo or control, Outcome 3: Bronchopulmonary dysplasia at 36 weeks' gestation

All‐cause neonatal mortality at 28 days

This was not reported.

All‐cause neonatal mortality before hospital discharge (Outcome 1.4)

Eight studies reported on all‐cause neonatal mortality before hospital discharge (parenteral administration: Kluckow 2014; Krauss 1989; Merritt 1981; Monset‐Couchard 1983; Yeh 1981a; enteral administration: Nestrud 1980; Rudd 1983; Yanagi 1981). None of the individual studies reported an impact on all‐cause neonatal mortality before hospital discharge, and no effect of either parenteral or enteral indomethacin was demonstrated in the meta‐analysis (all studies typical RR 0.78, 95% CI 0.46 to 1.33; 8 studies, 314 infants; parenteral administration typical RR 1.00, 95% CI 0.52 to 1.91; 5 studies, 222 infants; enteral administration typical RR 0.46, 95% CI 0.17 to 1.22; 3 studies, 92 infants; moderate‐certainty evidence; Analysis 1.4). No heterogeneity was noted.

1.4. Analysis.

Comparison 1: Indomethacin vs placebo or control, Outcome 4: All‐cause neonatal mortality before hospital discharge

Secondary outcomes: PDA‐related outcomes

Proportion of infants receiving rescue medical treatment (repeated cyclooxygenase or paracetamol/acetaminophen dosing, or both) (Outcome 1.5)

Six studies reported on the proportion of infants receiving rescue medical treatment (repeated cyclooxygenase or paracetamol/acetaminophen dosing, or both) (parenteral administration: Cotton 1980; Kluckow 2014; Merritt 1981; Monset‐Couchard 1983; enteral administration: Nestrud 1980; Rudd 1983).

Two of the six studies reported a decrease in the risk of receiving rescue medical treatment (repeated cyclooxygenase or paracetamol/acetaminophen dosing, or both) (Merritt 1981; Monset‐Couchard 1983).

Overall, indomethacin administration was associated with a large decrease in the risk of receiving rescue medical treatment (typical RR 0.35, 95% CI 0.23 to 0.54; typical RD ‐0.34, 95% CI ‐0.45 to ‐0.22; 6 studies, 211 infants; Analysis 1.5).

1.5. Analysis.

Comparison 1: Indomethacin vs placebo or control, Outcome 5: Proportion of infants receiving rescue medical treatment

This was seen in both studies that utilized enteral (typical RR 0.23, 95% CI 0.06 to 0.82; typical RD ‐0.31, 95% CI ‐0.53 to ‐0.09; 2 studies, 53 infants) or parenteral administration of indomethacin (typical RR 0.38, 95% CI 0.24 to 0.61; typical RD ‐0.35, 95% CI ‐0.48 to ‐0.21; 4 studies, 158 infants).

Proportion of infants requiring surgical ligation or transcatheter occlusion (Outcome 1.6)

Seven studies reported on the proportion of infants requiring surgical ligation or transcatheter occlusion (parenteral administration: Kluckow 2014; Krauss 1989; Merritt 1981; Monset‐Couchard 1983; Yeh 1981a; enteral administration: Nestrud 1980; Rudd 1983).

None of the included studies reported a decrease in the risk of requiring surgical ligation or transcatheter occlusion, and no effect of either parenteral or enteral indomethacin was demonstrated in the meta‐analysis (all studies typical RR 0.66, 95% CI 0.33 to 1.29; 7 studies, 275 infants; moderate‐certainty evidence; Analysis 1.6; parenteral administration typical RR 0.50, 95% CI 0.19 to 1.33; 5 studies, 222 infants; enteral administration typical RR 0.94, 95% CI 0.37 to 2.39; 2 studies, 53 infants). No heterogeneity was noted.

1.6. Analysis.

Comparison 1: Indomethacin vs placebo or control, Outcome 6: Proportion of infants requiring surgical ligation of PDA

Other outcomes

Pneumothorax

This was not reported.

Pulmonary hemorrhage (Outcome 1.7)

Only one study reported on pulmonary hemorrhage (parenteral administration: Kluckow 2014). In that study, no impact on the risk of pulmonary hemorrhage was reported (RR 0.40, 95% CI 0.14 to 1.16; 1 study, 92 infants; Analysis 1.7).

1.7. Analysis.

Comparison 1: Indomethacin vs placebo or control, Outcome 7: Pulmonary hemorrhage

Late‐onset bacterial sepsis (Outcome 1.8)

Three studies reported on the risk of late‐onset bacterial sepsis (parenteral administration: Kluckow 2014; Yeh 1981a; enteral administration: Nestrud 1980).

None of the included studies reported a decrease in the risk of late‐onset bacterial sepsis, and no effect of either parenteral or enteral indomethacin was demonstrated in the meta‐analysis (all studies typical RR 0.78, 95% CI 0.46 to 1.33; 3 studies, 170 infants; Analysis 1.8; parenteral administration typical RR 0.82, 95% CI 0.48 to 1.40; 2 studies, 147 infants; enteral administration typical RR 0.31, 95% CI 0.01 to 6.85; 1 study, 23 infants). No heterogeneity was noted.

1.8. Analysis.

Comparison 1: Indomethacin vs placebo or control, Outcome 8: Late‐onset bacterial sepsis

Necrotising enterocolitis (NEC) (≥ Bell stage 2) (Bell 1978) (Outcome 1.9)

Two studies, both using parenteral administration of indomethacin, reported on the risk of necrotizing enterocolitis (NEC) (≥ Bell stage 2) (Kluckow 2014; Yeh 1981a).

Neither of the included studies reported a decrease in the risk of necrotizing enterocolitis, and no effect of parenteral indomethacin was demonstrated in the meta‐analysis (parenteral administration typical RR 1.27, 95% CI 0.36 to 4.55; 2 studies, 147 infants; low‐certainty evidence; Analysis 1.9). No heterogeneity was noted.

1.9. Analysis.

Comparison 1: Indomethacin vs placebo or control, Outcome 9: NEC (≥ Bell stage 2)

Intraventricular hemorrhage (IVH) (any grade) (Papile 1978) (Outcome 1.10)

Three studies reported on the risk of intraventricular hemorrhage (IVH) (any grade) (parenteral administration: Kluckow 2014; Monset‐Couchard 1983; Yeh 1981a).

None of the included studies reported a decrease in the risk of intraventricular hemorrhage (IVH) (any grade), and no effect of parenteral indomethacin was demonstrated in the meta‐analysis (parenteral administration typical RR 0.58, 95% CI 0.20 to 1.65; 3 studies, 171 infants; Analysis 1.10). No heterogeneity was noted.

1.10. Analysis.

Comparison 1: Indomethacin vs placebo or control, Outcome 10: IVH (any grade)

Severe intraventricular hemorrhage (IVH) (grade III to IV) (Papile 1978) (Outcome 1.11)

Only one study reported on severe IVH (grade III to IV) (parenteral administration: Monset‐Couchard 1983). In that small study, no impact on the risk of severe IVH was reported (RR 0.33, 95% CI 0.01 to 7.45; 1 study, 24 infants; Analysis 1.11).

1.11. Analysis.

Comparison 1: Indomethacin vs placebo or control, Outcome 11: Severe IVH (grade III to IV)

Periventricular leukomalacia (PVL) (Outcome 1.12)

Only one study reported on periventricular leukomalacia (PVL) (parenteral administration: Kluckow 2014). In that study, no impact on the risk of PVL was reported (RR 1.09, 95% CI 0.29 to 4.10; 1 study, 92 infants; Analysis 1.12).

1.12. Analysis.

Comparison 1: Indomethacin vs placebo or control, Outcome 12: Periventricular leukomalacia

Retinopathy of prematurity (any stage) (ICROP 2005) (Outcome 1.13)

Only one study reported on retinopathy of prematurity (any stage) (parenteral administration: Yeh 1981a). In that small study, no impact on the risk of requiring supplemental oxygen at 28 days' postnatal age was reported (RR 0.32, 95% CI 0.07 to 1.42; 1 study, 47 infants; Analysis 1.13).

1.13. Analysis.

Comparison 1: Indomethacin vs placebo or control, Outcome 13: Retinopathy of prematurity (any stage)

Severe retinopathy of prematurity (≥ stage III) (Outcome 1.14)

Only one study reported on severe retinopathy of prematurity (ROP) (≥ stage III) (parenteral administration: Yeh 1981a). In that small study, no impact on the risk of severe ROP was reported (RR 0.96, 95% CI 0.06 to 14.43; 1 study, 47 infants; Analysis 1.14).

1.14. Analysis.

Comparison 1: Indomethacin vs placebo or control, Outcome 14: Severe retinopathy of prematurity (≥ stage III)

Surgery for severe retinopathy of prematurity (added post hoc) (Outcome 1.15)

Only one study reported on surgery for severe retinopathy of prematurity (ROP) (parenteral administration: Kluckow 2014). In that study, no impact on the risk of surgery for severe ROP was reported (RR 0.16, 95% CI 0.01 to 2.93; 1 study, 92 infants; Analysis 1.15).

1.15. Analysis.

Comparison 1: Indomethacin vs placebo or control, Outcome 15: Surgery for severe retinopathy of prematurity*

Infant mortality (first year of life) (Outcome 1.16)

Only one study reported on infant mortality (first year of life) (parenteral administration: Yeh 1981a). In that small study, no impact on the risk of infant mortality (first year of life) was reported (RR 1.08, 95% CI 0.49 to 2.40; 1 study, 55 infants; Analysis 1.16).

1.16. Analysis.

Comparison 1: Indomethacin vs placebo or control, Outcome 16: Infant mortality (first year of life)

Use of inotropic agents (Outcome 1.17)

Only one study reported on use of inotropic agents (parenteral administration: Kluckow 2014). In that study, no impact on the risk of using inotropic agents was reported (RR 1.09, 95% CI 0.50 to 2.37; 1 study, 92 infants; Analysis 1.17).

1.17. Analysis.

Comparison 1: Indomethacin vs placebo or control, Outcome 17: Use of inotropic agents

Duration of assisted ventilation (days) (Outcome 1.18)

Only one study reported on the duration of assisted ventilation (parenteral administration: Merritt 1981). In that small study, no impact on the duration of assisted ventilation was reported (mean difference (MD) 0.28, 95% CI ‐3.55 to 4.11; 1 study, 24 infants; Analysis 1.18).

1.18. Analysis.

Comparison 1: Indomethacin vs placebo or control, Outcome 18: Duration of assisted ventilation (days)

Duration of oxygen dependence (days to last discontinuation of any supplemental oxygen)

This was not reported.

Duration of hospital stay (days) (Outcome 1.19)

Only one study reported on the duration of hospital stay (days) (parenteral administration: Yeh 1981a). In that small study, no impact on the duration of hospital stay (days) was reported (MD ‐14.30, 95% CI ‐51.36 to 22.76; 1 study, 44 infants; Analysis 1.19).

1.19. Analysis.

Comparison 1: Indomethacin vs placebo or control, Outcome 19: Duration of hospital stay (days)

Time to full enteral feeds (days)

This was not reported.

Cerebral palsy at approximately 2 years' corrected age (as defined by study authors)

This was not reported.

Moderate to severe neurodevelopmental outcome: neurodevelopmental outcome at approximately 2 years' corrected age (acceptable range 18 months to 28 months) including cerebral palsy, delayed neurodevelopment (Bayley Scales of Infant Development Mental Developmental Index < 70), legal blindness (< 20/200 visual acuity), and hearing deficit (aided or < 60 dB on audiometric testing). The composite outcome 'neurodevelopmental impairment' was defined as having any one of the aforementioned deficits

This was not reported.

Safety outcomes (harms reported within one week of completing intervention)

Intestinal perforation (Outcome 1.20)

Only one study reported on Intestinal perforation (parenteral administration: Kluckow 2014). In that study, no cases of intestinal perforation were reported in either group (RR not estimable; RD 0.00, 95% CI ‐0.04 to 0.04; 1 study, 92 infants; Analysis 1.20).

1.20. Analysis.

Comparison 1: Indomethacin vs placebo or control, Outcome 20: Intestinal perforation

Renal function

Oliguria (< 1 mL/kg/hr)

This was not reported.

Serum/plasma creatinine (µmol/L) levels during treatment

This was not reported.

Serum/plasma creatinine (µmol/L) after treatment

This was not reported.

Post hoc: creatinine > 150 µmol/L during treatment (post hoc) (Outcome 1.21)

Only one study reported on creatinine > 150 µmol/L during treatment (parenteral administration: Kluckow 2014). In that study, no impact on the risk of elevation in creatinine > 150 µmol/L during treatment was reported (RR 1.09, 95% CI 0.07 to 16.92; 1 study, 92 infants; Analysis 1.21).

1.21. Analysis.

Comparison 1: Indomethacin vs placebo or control, Outcome 21: Creatinine > 150 micromole during treatment

Hemostasis

Mucocutaneous or gastrointestinal bleeding (Outcome 1.22)

Two studies reported on the risk of mucocutaneous or gastrointestinal bleeding (parenteral administration: Kluckow 2014; enteral administration: Rudd 1983).

None of the included studies reported a decrease in the risk of mucocutaneous or gastrointestinal bleeding. Kluckow 2014 reported no events (RR not estimable; RD 0.00, 95% CI ‐0.04 to 0.04; 1 study, 89 infants), and Rudd 1983 reported one event in the control group (RR 0.33, 95% CI 0.01 to 7.58; RD ‐0.07, 95% CI ‐0.23 to 0.10; 1 study, 30 infants). No effect of indomethacin was demonstrated in the meta‐analysis (RR 0.33, 95% CI 0.01 to 7.58; RD ‐0.02, 95% CI ‐0.07 to 0.04; 2 studies, 119 infants; low‐certainty evidence; Analysis 1.22).

1.22. Analysis.

Comparison 1: Indomethacin vs placebo or control, Outcome 22: Mucocutaneous or gastrointestinal bleeding

Platelet count (< 50,000 platelets/µL blood)

This was not reported.

Pulmonary hypertension (diagnosed by echocardiographic or Doppler criteria)

This was not reported.

Indomethacin versus placebo or control (sensitivity analysis) (Comparison 2)

We conducted a sensitivity analysis for the primary outcomes excluding studies that were judged to have high risk of bias in one or more domains as specified a priori (Cotton 1980; Krauss 1989; Merritt 1981; Monset‐Couchard 1983).

Primary outcomes

Failure of PDA closure within one week of administration of the first dose of indomethacin (Outcome 2.1)

Seven studies reported on PDA closure within one week of administration of the first dose of indomethacin in this analysis. Overall, indomethacin administration was associated with a large decrease in failure of PDA closure within one week of administration of the first dose of indomethacin (typical RR 0.29, 95% CI 0.22 to 0.38; typical RD ‐0.51, 95% CI ‐0.58 to ‐0.44; 7 studies, 586 infants; Analysis 2.1). This was seen in both studies that utilized enteral (typical RR 0.35, 95% CI 0.23 to 0.54; typical RD ‐0.49, 95% CI ‐0.64 to ‐0.35; 5 studies, 126 infants) or parenteral administration of indomethacin (typical RR 0.27, 95% CI 0.19 to 0.37; typical RD ‐0.52, 95% CI ‐0.60 to ‐0.44; 2 studies, 460 infants).

2.1. Analysis.

Comparison 2: Indomethacin vs placebo or control (sensitivity analysis), Outcome 1: Failure of PDA closure within 1 week of administration of the first dose of indomethacin

Bronchopulmonary dysplasia (defined as supplemental oxygen need at 28 days' postnatal age with or without compatible clinical and radiographic findings) (Outcome 2.2)

Only one study reported on supplemental oxygen need at 28 days' postnatal age with or without compatible clinical and radiographic findings (parenteral administration: Yeh 1981a). In that small study, no impact on the risk of requiring supplemental oxygen at 28 days' postnatal age was reported (RR 1.45, 95% CI 0.60 to 3.51; 1 study, 55 infants; low‐certainty evidence; Analysis 2.2).

2.2. Analysis.

Comparison 2: Indomethacin vs placebo or control (sensitivity analysis), Outcome 2: Bronchopulmonary dysplasia at 28 days

Bronchopulmonary dysplasia (defined as supplemental oxygen at 36 weeks' postmenstrual age with or without compatible clinical and radiographic findings) (Outcome 2.3)

Only one study reported on supplemental oxygen need at 36 weeks' postmenstrual age with or without compatible clinical and radiographic findings (parenteral administration: Kluckow 2014). In that study, no impact on the risk of requiring supplemental oxygen at 36 weeks' postmenstrual age was reported (RR 0.80, 95% CI 0.41 to 1.55; 1 study, 92 infants; low‐certainty evidence; Analysis 2.3).

2.3. Analysis.

Comparison 2: Indomethacin vs placebo or control (sensitivity analysis), Outcome 3: Bronchopulmonary dysplasia at 36 weeks' gestation

All‐cause neonatal mortality before hospital discharge (Outcome 2.4)

Five studies reported on all‐cause neonatal mortality before hospital discharge in this analysis. None of the individual studies reported an impact on all‐cause neonatal mortality before hospital discharge, and no effect of either parenteral or enteral indomethacin was demonstrated in the meta‐analysis (all studies typical RR 0.73, 95% CI 0.39 to 1.34; 5 studies, 239 infants; parenteral administration typical RR 1.02, 95% CI 0.45 to 2.29; 2 studies, 147 infants; enteral administration typical RR 0.46, 95% CI 0.17 to 1.22; 3 studies, 92 infants; Analysis 2.4).

2.4. Analysis.

Comparison 2: Indomethacin vs placebo or control (sensitivity analysis), Outcome 4: All‐cause neonatal mortality before hospital discharge

Discussion

Summary of main results

Fourteen studies completed to date have compared indomethacin to placebo or to no treatment. A total of 880 infants have been enrolled in these studies. Nine of the 14 studies compared intravenous indomethacin to placebo or no treatment (Cotton 1980; Gersony 1983; Kluckow 2014; Knight 2011; Krauss 1989; Merritt 1981; Monset‐Couchard 1983; Osborn 2003; Yeh 1981a). Five studies compared oral indomethacin to placebo or no treatment (Nestrud 1980; Neu 1981; Rudd 1983; Valaes 1980; Yanagi 1981).

Indomethacin appears to be significantly more effective than placebo or no treatment for patent ductus arteriosus (PDA) closure in preterm infants (high‐certainty evidence). Indomethacin use is also associated with a significantly reduced need for rescue medical treatment for PDA. The effectiveness of both intravenous and oral administration of indomethacin appears to be similar when compared to placebo or no treatment. There were no differences in the rates of other clinical outcomes nor in adverse effects.

Overall completeness and applicability of evidence

We recommend cautious interpretation and application of these findings. Although it was clearly demonstrated that indomethacin administration, as intravenous or oral formulation, was superior to placebo or no treatment in closing a PDA, evidence is insufficient to allow inferences on the clinical implications of this finding. No statistically significant differences could be demonstrated in rate of surgical PDA ligation, duration of assisted ventilation, bronchopulmonary dysplasia (BPD), necrotizing enterocolitis (NEC), intraventricular hemorrhage (IVH), retinopathy of prematurity (ROP), or mortality. However, several limitations of the included studies prevent us from drawing firm conclusions on the lack of efficacy of indomethacin for clinical outcomes. First, imprecise effect estimates, as evident from their wide confidence intervals, for the above‐mentioned clinical outcomes fail to provide convincing evidence for absence of effect on these outcomes. Second, a significantly higher proportion of infants in the placebo or no treatment group ended up receiving open‐label medical therapy, thereby likely pulling the effect estimates for clinical outcomes towards null. The latter, especially, might be an important reason why significantly improved rates of PDA closure did not translate into lower rates of PDA ligation or longer‐term clinical benefit; this might need to be addressed in future trials.

Evidence was insufficient to confidently establish the safety of indomethacin. No differences were demonstrated with respect to intestinal perforation, NEC, mucocutaneous or gastrointestinal bleeding, and transient renal impairment. However, imprecise effect estimates again fail to provide convincing evidence for safety with respect to these outcomes. Clinicians should consider this uncertainty of evidence on safety while considering indomethacin treatment for preterm infants.

Clinicians should also consider the applicability of this evidence in the current clinical context. Eleven out of the 14 included studies were published before 1990, when neonatal clinical practices were vastly different, such as infrequent use of antenatal corticosteroids, lesser use of non‐invasive ventilatory support, and no availability of surfactant therapy. Therefore, clinicians should exercise caution when applying these results in the current context, especially with regards to extremely low gestational age infants at the limits of viability.

Quality of the evidence

Certainty of evidence for the primary outcome, failure of PDA closure, was high according to GRADE. Evidence for other clinical outcomes was of low to moderate certainty. Certainty of evidence should be considered in light of the fact that most included trials were conducted in the pretrial registration era, and publications were not mandated to follow CONSORT guidelines when reporting a clinical trial (Moher 1998). Therefore, risk of bias was unclear for several domains across trials. Furthermore, most trials were small‐sample pilot trials with insufficient power to detect statistically significant differences in rates of clinically relevant outcomes, thereby leading to serious or very serious imprecision on GRADE assessment. Despite the above limitations, certainty of evidence for ‘failure of PDA closure’ was high, as most of the evidence for this outcome was contributed by studies with low risk of bias with minimal inconsistency and precise effect estimates.

Potential biases in the review process

We are not aware of any biases in the review process. Departures from the protocol were primarily related to clarification of inclusion criteria and relevant outcomes. Review authors were not involved with any of the included trials.

Agreements and disagreements with other studies or reviews

The result for the primary outcome of ‘failure of PDA closure’ is consistent with the two existing network meta‐analyses comparing pharmacotherapeutic options for treatment of a symptomatic PDA (Jones 2011; Mitra 2018). Mitra 2018 showed that both intravenous indomethacin (network odds ratio (OR) 0.15, 95% credible interval (CrI) 0.09 to 0.22) and ‘indomethacin, other types’, which included oral indomethacin (network OR 0.15, 95% CrI 0.09 to 0.24), were associated with significantly lower rates of failure of PDA closure compared to placebo or no treatment. The earlier network meta‐analysis by Jones 2011 also shows that intravenous indomethacin was associated with significantly lower rates of failure of PDA closure compared to placebo (pooled risk ratio (RR) 0.42, 95% confidence interval (CI) 0.36 to 0.49).

Although the network meta‐analysis by Jones 2011 did not show any statistically significant differences in other clinical outcomes, in keeping with the results of this review some differences were noted in the results of the network meta‐analysis by Mitra 2018. Mitra and colleagues demonstrated that the need for surgical PDA ligation was significantly lower with both intravenous indomethacin (network OR 0.29, 95% CrI 0.12 to 0.65) and ‘indomethacin, other types’ (network OR 0.25, 95% CrI 0.11 to 0.57) compared to placebo or no treatment (Mitra 2018). Furthermore, it was noted in Mitra 2018 that the risk of NEC was significantly higher with ‘indomethacin, other types’ (network OR 2.58, 95% CrI 1.06 to 6.65) compared to placebo or no treatment. No such differences were noted in our results. Differences in statistical analytical techniques might contribute to such differences in results. In our analyses, we used fixed‐effect meta‐analysis of direct comparisons only. Conversely, network meta‐analysis uses both direct and indirect comparisons to obtain network effect estimates.

It is important to note key differences between our results and those of the previously published Cochrane Review on prophylactic intravenous indomethacin in preterm infants (Fowlie 2010). Although prophylactic intravenous indomethacin was shown to significantly reduce PDA surgical ligation (typical RR 0.51, 95% CI 0.37 to 0.71) and severe intraventricular hemorrhage (typical RR 0.66, 95% CI 0.53 to 0.82) (Fowlie 2010), we were unable to demonstrate in our review such findings with use of indomethacin for symptomatic PDA. Such differences could largely be attributed to differences in methodological quality of included studies in the two respective reviews.

Authors' conclusions

Implications for practice.

High‐certainty evidence shows that indomethacin is effective in closing a symptomatic patent ductus arteriosus (PDA) in preterm infants compared to placebo or no treatment. Given that we were unable to demonstrate any other clinically important benefit or to confidently rule out any significant adverse effect with this intervention, the decision to use indomethacin for symptomatic PDA will depend on the value that families and neonatal care providers attach to the surrogate outcome of PDA closure.

Implications for research.

It is important to acknowledge that most of the trials included in this review were designed to examine whether indomethacin was effective in closing a symptomatic PDA. These trials were not designed or powered to explore whether indomethacin treatment for symptomatic PDA improved clinically relevant outcomes. Hence most trials had a backup treatment option had the PDA remained open following the initial intervention, as evident from the significantly higher rate of open‐label treatment in the placebo or no treatment group. Therefore, if future trials are designed to examine effects of indomethacin treatment on clinical outcomes, then any open‐label treatment should be strictly disallowed before outcome assessment. Furthermore, trials should be powered for outcomes relevant to patients and families such as mortality, necrotizing enterocolitis (NEC), and neurodevelopment rather than short‐term surrogates such as PDA closure (Webbe 2020). Finally, with increasing adoption of conservative management even for symptomatic PDA, future trials should focus on enrolling infants at highest risk of PDA‐related morbidity, based on gestational age, degree of hemodynamic significance of the PDA, and other perinatal characteristics (Mitra 2020).

History

Protocol first published: Issue 9, 2018
Review first published: Issue 1, 2021

Appendices

Appendix 1. Search strategy

The RCT filters have been created using Cochrane's highly sensitive search strategies for identifying randomized trials (Higgins 2019). The neonatal filters were created and tested by the Cochrane Neonatal Information Specialist.

Cochrane CENTRAL via CRS Web

1MESH DESCRIPTOR Indomethacin EXPLODE ALL AND CENTRAL:TARGET

2(indomethacin or indometacin or indocid or indocin) AND CENTRAL:TARGET

3#1 OR #2

4MESH DESCRIPTOR Ductus Arteriosus, Patent EXPLODE ALL AND CENTRAL:TARGET

5(patent ductus arteriosus or PDA or (patency ADJ3 ductus arteriosus)) AND CENTRAL:TARGET

6#5 OR #4

7MESH DESCRIPTOR Infant, Newborn EXPLODE ALL AND CENTRAL:TARGET

8infant or infants or infant's or "infant s" or infantile or infancy or newborn* or "new born" or "new borns" or "newly born" or neonat* or baby* or babies or premature or prematures or prematurity or preterm or preterms or "pre term" or premies or "low birth weight" or "low birthweight" or VLBW or LBW or ELBW or NICU AND CENTRAL:TARGET

9#8 OR #7 AND CENTRAL:TARGET

10#3 AND #6 AND #9

OVID MEDLINE Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, Ovid MEDLINE(R) Daily, and Ovid MEDLINE(R) 1946 to Present

1. exp Indomethacin/

2. (indomethacin or indometacin or indocid or indocin).mp.

3. 1 or 2

4. exp Ductus Arteriosus, Patent/

5. (patent ductus arteriosus or PDA or (patency adj3 ductus arteriosus)).mp.

6. 4 or 5

7. exp infant, newborn/

8. (newborn* or new born or new borns or newly born or baby* or babies or premature or prematurity or preterm or pre term or low birth weight or low birthweight or VLBW or LBW or infant or infants or 'infant s' or infant's or infantile or infancy or neonat*).ti,ab.

9. 7 or 8

10. randomized controlled trial.pt.

11. controlled clinical trial.pt.

12. randomized.ab.

13. placebo.ab.

14. drug therapy.fs.

15. randomly.ab.

16. trial.ab.

17. groups.ab.

18. or/10‐17

19. exp animals/ not humans.sh.

20. 18 not 19

21. 9 and 20

22. randomi?ed.ti,ab.

23. randomly.ti,ab.

24. trial.ti,ab.

25. groups.ti,ab.

26. ((single or doubl* or tripl* or treb*) and (blind* or mask*)).ti,ab.

27. placebo*.ti,ab.

28. 22 or 23 or 24 or 25 or 26 or 27

29. 8 and 28

30. limit 29 to yr="2018 ‐Current"

31. 21 or 30

32. 3 and 6 and 31

CINAHL via EBSCOhost

(indomethacin or indometacin or indocid or indocin) AND
(patent ductus arteriosus or PDA or (patency N3 ductus arteriosus)) AND
(infant or infants or infant’s or infantile or infancy or newborn* or "new born" or "new borns" or "newly born" or neonat* or baby* or babies or premature or prematures or prematurity or preterm or preterms or "pre term" or premies or "low birth weight" or "low birthweight" or VLBW or LBW) AND
(randomized controlled trial OR controlled clinical trial OR randomized OR randomised OR placebo OR clinical trials as topic OR randomly OR trial OR PT clinical trial)

Appendix 2. ‘Risk of bias' tool

1. Sequence generation (checking for possible selection bias). Was the allocation sequence adequately generated?

For each included study, we categorized the method used to generate the allocation sequence as:

1. low risk (any truly random process, e.g. random number table; computer random number generator);

2. high risk (any non‐random process, e.g. odd or even date of birth; hospital or clinic record number); or

3. unclear risk.

2. Allocation concealment (checking for possible selection bias). Was allocation adequately concealed?

For each included study, we categorized the method used to conceal the allocation sequence as:

1. low risk (e.g. telephone or central randomization; consecutively numbered sealed opaque envelopes);

2. high risk (open random allocation; unsealed or non‐opaque envelopes, alternation; date of birth); or

3. unclear risk.

3. Blinding of participants and personnel (checking for possible performance bias). Was knowledge of the allocated intervention adequately prevented during the study?

For each included study, we categorized the methods used to blind study participants and personnel from knowledge of which intervention a participant received. We assessed blinding separately for different outcomes or classes of outcomes. We categorized the methods as:

1. low risk, high risk, or unclear risk for participants; and

2. low risk, high risk, or unclear risk for personnel.

4. Blinding of outcome assessment (checking for possible detection bias). Was knowledge of the allocated intervention adequately prevented at the time of outcome assessment?

For each included study, we categorized the methods used to blind outcome assessment. Blinding was assessed separately for different outcomes or classes of outcomes. We categorized the methods as:

1. low risk for outcome assessors;

2. high risk for outcome assessors; or

3. unclear risk for outcome assessors.

5. Incomplete outcome data (checking for possible attrition bias through withdrawals, dropouts, protocol deviations). Were incomplete outcome data adequately addressed?

For each included study and for each outcome, we described the completeness of data including attrition and exclusions from the analysis. We noted whether attrition and exclusions were reported, numbers included in the analysis at each stage (compared with the total randomized participants), reasons for attrition or exclusion when reported, and whether missing data were balanced across groups or were related to outcomes. When sufficient information was reported or supplied by trial authors, we re‐included missing data in the analyses. We categorized the methods as:

1. low risk (< 20% missing data);

2. high risk (≥ 20% missing data); or

3. unclear risk.

6. Selective reporting bias. Are reports of the study free of the suggestion of selective outcome reporting?

For each included study, we described how we investigated the possibility of selective outcome reporting bias and what we found. For studies for which study protocols were published in advance, we compared pre‐specified outcomes versus outcomes eventually reported in the published results. If the study protocol was not published in advance, we contacted study authors to gain access to the study protocol. We assessed the methods as:

1. low risk (when it was clear that all of the study's pre‐specified outcomes and all expected outcomes of interest to the review had been reported);

2. high risk (when not all of the study's pre‐specified outcomes had been reported; one or more reported primary outcomes were not pre‐specified outcomes of interest and were reported incompletely and so could not be used; study failed to include results of a key outcome that would have been expected to have been reported); or

3. unclear risk.

7. Other sources of bias. Was the study apparently free of other problems that could put it at high risk of bias?

For each included study, we described any important concerns we had about other possible sources of bias (e.g. whether there was a potential source of bias related to the specific study design, whether the trial was stopped early due to some data‐dependent process). We assessed whether each study was free of other problems that could put it at risk of bias as:

1. low risk;

2. high risk; or

3. unclear risk.

If needed, we explored the impact of the level of bias through undertaking sensitivity analyses.

Data and analyses

Comparison 1. Indomethacin vs placebo or control.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Failure of PDA closure within 1 week of administration of the first dose of indomethacin	10	654	Risk Ratio (M‐H, Fixed, 95% CI)	0.30 [0.23, 0.38]
1.1.1 Parenteral administration	5	528	Risk Ratio (M‐H, Fixed, 95% CI)	0.28 [0.21, 0.37]
1.1.2 Enteral administration	5	126	Risk Ratio (M‐H, Fixed, 95% CI)	0.35 [0.23, 0.54]
1.2 Bronchopulmonary dysplasia at 28 days	1	55	Risk Ratio (M‐H, Fixed, 95% CI)	1.45 [0.60, 3.51]
1.2.1 Parenteral administration	1	55	Risk Ratio (M‐H, Fixed, 95% CI)	1.45 [0.60, 3.51]
1.3 Bronchopulmonary dysplasia at 36 weeks' gestation	1	92	Risk Ratio (M‐H, Fixed, 95% CI)	0.80 [0.41, 1.55]
1.3.1 Parenteral administration	1	92	Risk Ratio (M‐H, Fixed, 95% CI)	0.80 [0.41, 1.55]
1.4 All‐cause neonatal mortality before hospital discharge	8	314	Risk Ratio (M‐H, Fixed, 95% CI)	0.78 [0.46, 1.33]
1.4.1 Parenteral administration	5	222	Risk Ratio (M‐H, Fixed, 95% CI)	1.00 [0.52, 1.91]
1.4.2 Enteral administration	3	92	Risk Ratio (M‐H, Fixed, 95% CI)	0.46 [0.17, 1.22]
1.5 Proportion of infants receiving rescue medical treatment	6	211	Risk Ratio (M‐H, Fixed, 95% CI)	0.35 [0.23, 0.54]
1.5.1 Parenteral administration	4	158	Risk Ratio (M‐H, Fixed, 95% CI)	0.38 [0.24, 0.61]
1.5.2 Enteral administration	2	53	Risk Ratio (M‐H, Fixed, 95% CI)	0.23 [0.06, 0.82]
1.6 Proportion of infants requiring surgical ligation of PDA	7	275	Risk Ratio (M‐H, Fixed, 95% CI)	0.66 [0.33, 1.29]
1.6.1 Parenteral administration	5	222	Risk Ratio (M‐H, Fixed, 95% CI)	0.50 [0.19, 1.33]
1.6.2 Enteral administration	2	53	Risk Ratio (M‐H, Fixed, 95% CI)	0.94 [0.37, 2.39]
1.7 Pulmonary hemorrhage	1	92	Risk Ratio (M‐H, Fixed, 95% CI)	0.40 [0.14, 1.16]
1.7.1 Parenteral administration	1	92	Risk Ratio (M‐H, Fixed, 95% CI)	0.40 [0.14, 1.16]
1.8 Late‐onset bacterial sepsis	3	170	Risk Ratio (M‐H, Fixed, 95% CI)	0.78 [0.46, 1.33]
1.8.1 Parenteral administration	2	147	Risk Ratio (M‐H, Fixed, 95% CI)	0.82 [0.48, 1.40]
1.8.2 Enteral administration	1	23	Risk Ratio (M‐H, Fixed, 95% CI)	0.31 [0.01, 6.85]
1.9 NEC (≥ Bell stage 2)	2	147	Risk Ratio (M‐H, Fixed, 95% CI)	1.27 [0.36, 4.55]
1.9.1 Parenteral administration	2	147	Risk Ratio (M‐H, Fixed, 95% CI)	1.27 [0.36, 4.55]
1.10 IVH (any grade)	3	171	Risk Ratio (M‐H, Fixed, 95% CI)	0.58 [0.20, 1.65]
1.10.1 Parenteral administration	3	171	Risk Ratio (M‐H, Fixed, 95% CI)	0.58 [0.20, 1.65]
1.11 Severe IVH (grade III to IV)	1	24	Risk Ratio (M‐H, Fixed, 95% CI)	0.33 [0.01, 7.45]
1.11.1 Parenteral administration	1	24	Risk Ratio (M‐H, Fixed, 95% CI)	0.33 [0.01, 7.45]
1.12 Periventricular leukomalacia	1	92	Risk Ratio (M‐H, Fixed, 95% CI)	1.09 [0.29, 4.10]
1.12.1 Parenteral administration	1	92	Risk Ratio (M‐H, Fixed, 95% CI)	1.09 [0.29, 4.10]
1.13 Retinopathy of prematurity (any stage)	1	47	Risk Ratio (M‐H, Fixed, 95% CI)	0.32 [0.07, 1.42]
1.13.1 Parenteral administration	1	47	Risk Ratio (M‐H, Fixed, 95% CI)	0.32 [0.07, 1.42]
1.14 Severe retinopathy of prematurity (≥ stage III)	1	47	Risk Ratio (M‐H, Fixed, 95% CI)	0.96 [0.06, 14.43]
1.14.1 Parenteral administration	1	47	Risk Ratio (M‐H, Fixed, 95% CI)	0.96 [0.06, 14.43]
1.15 Surgery for severe retinopathy of prematurity*	1	92	Risk Ratio (M‐H, Fixed, 95% CI)	0.16 [0.01, 2.93]
1.15.1 Parenteral administration	1	92	Risk Ratio (M‐H, Fixed, 95% CI)	0.16 [0.01, 2.93]
1.16 Infant mortality (first year of life)	1	55	Risk Ratio (M‐H, Fixed, 95% CI)	1.08 [0.49, 2.40]
1.16.1 Parenteral administration	1	55	Risk Ratio (M‐H, Fixed, 95% CI)	1.08 [0.49, 2.40]
1.17 Use of inotropic agents	1	92	Risk Ratio (M‐H, Fixed, 95% CI)	1.09 [0.50, 2.37]
1.17.1 Parenteral administration	1	92	Risk Ratio (M‐H, Fixed, 95% CI)	1.09 [0.50, 2.37]
1.18 Duration of assisted ventilation (days)	1	24	Mean Difference (IV, Fixed, 95% CI)	0.28 [‐3.55, 4.11]
1.18.1 Parenteral administration	1	24	Mean Difference (IV, Fixed, 95% CI)	0.28 [‐3.55, 4.11]
1.19 Duration of hospital stay (days)	1	44	Mean Difference (IV, Fixed, 95% CI)	‐14.30 [‐51.36, 22.76]
1.19.1 Parenteral administration	1	44	Mean Difference (IV, Fixed, 95% CI)	‐14.30 [‐51.36, 22.76]
1.20 Intestinal perforation	1	92	Risk Ratio (M‐H, Fixed, 95% CI)	Not estimable
1.20.1 Parenteral administration	1	92	Risk Ratio (M‐H, Fixed, 95% CI)	Not estimable
1.21 Creatinine > 150 micromole during treatment	1	92	Risk Ratio (M‐H, Fixed, 95% CI)	1.09 [0.07, 16.92]
1.21.1 Parenteral administration	1	92	Risk Ratio (M‐H, Fixed, 95% CI)	1.09 [0.07, 16.92]
1.22 Mucocutaneous or gastrointestinal bleeding	2	119	Risk Ratio (M‐H, Fixed, 95% CI)	0.33 [0.01, 7.58]
1.22.1 Parenteral administration	1	89	Risk Ratio (M‐H, Fixed, 95% CI)	Not estimable
1.22.2 Enteral administration	1	30	Risk Ratio (M‐H, Fixed, 95% CI)	0.33 [0.01, 7.58]

Comparison 2. Indomethacin vs placebo or control (sensitivity analysis).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Failure of PDA closure within 1 week of administration of the first dose of indomethacin	7	586	Risk Ratio (M‐H, Fixed, 95% CI)	0.29 [0.22, 0.38]
2.1.1 Parenteral administration	2	460	Risk Ratio (M‐H, Fixed, 95% CI)	0.27 [0.19, 0.37]
2.1.2 Enteral administration	5	126	Risk Ratio (M‐H, Fixed, 95% CI)	0.35 [0.23, 0.54]
2.2 Bronchopulmonary dysplasia at 28 days	1	55	Risk Ratio (M‐H, Fixed, 95% CI)	1.45 [0.60, 3.51]
2.2.1 Parenteral administration	1	55	Risk Ratio (M‐H, Fixed, 95% CI)	1.45 [0.60, 3.51]
2.3 Bronchopulmonary dysplasia at 36 weeks' gestation	1	92	Risk Ratio (M‐H, Fixed, 95% CI)	0.80 [0.41, 1.55]
2.3.1 Parenteral administration	1	92	Risk Ratio (M‐H, Fixed, 95% CI)	0.80 [0.41, 1.55]
2.4 All‐cause neonatal mortality before hospital discharge	5	239	Risk Ratio (M‐H, Fixed, 95% CI)	0.73 [0.39, 1.34]
2.4.1 Parenteral administration	2	147	Risk Ratio (M‐H, Fixed, 95% CI)	1.02 [0.45, 2.29]
2.4.2 Enteral administration	3	92	Risk Ratio (M‐H, Fixed, 95% CI)	0.46 [0.17, 1.22]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Cotton 1980.

Study characteristics
Methods	Randomized controlled single‐center study conducted in the Neonatal Intensive Care Unit of Vanderbilt, Nashville, TN, USA
Participants	Inclusion criteria: ventilator‐dependent preterm infants with symptomatic PDA Exclusion criteria: unclear
Interventions	IV indomethacin group (n = 9) received 0.2 mg/kg/dose of indomethacin with no digoxin given Medical management group (n = 8) received customary anti‐congestive measures including digoxin
Outcomes	Primary outcomes Primary PDA closure Secondary outcomes Duration of ventilator dependence
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Unclear method of randomization
Allocation concealment (selection bias)	Unclear risk	Unclear attempts at allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	High risk	No placebo was given. Staff were not blinded to the drugs administered
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	It is not stated whether the person assessing ductal closure was blinded to treatment
Incomplete outcome data (attrition bias) All outcomes	Low risk	All randomized infants are accounted for
Selective reporting (reporting bias)	Unclear risk	Cannot judge if there were any deviations from the original protocol, as study was conducted in pre‐registration era
Other bias	Unclear risk	Given this was an unblinded study, it is unclear if there were other sources of bias

Gersony 1983.

Study characteristics
Methods	Multi‐center randomized cross‐over study conducted at 13 clinical centers in the USA
Participants	Inclusion criteria: infants with birth weight ≤ 1750 grams with hemodynamically significant PDA. First‐order clinical criteria for a hemodynamically significant PDA included continuous murmur, systolic murmur, or no murmur with need for ventilator support for 48 hours. If met, evaluated for second‐order criteria, including hyperactive precordium, increased pulse pressure, tachycardia, tachypnea, hepatomegaly, LA/Ao ≥ 1.15, and cardiomegaly with pulmonary plethora Exclusion criteria: birth weight ≤ 500 grams, congenital anomalies, chromosomal abnormalities, death in first 24 hours of life, admission to study facility after 14 days of age, BUN ≥ 30 mg/dL, serum creatinine clearance ≥ 1.8 mg/dL, total urine output ≤ 0.6 mL/kg/hr over past 8 hours, platelet count < 60,000/mm³, stool Hematest ≥ 3+, bleeding diathesis, NEC, intracranial hemorrhage in past 7 days
Interventions	Initial indomethacin group (n = 135) received 0.2 mg/kg for initial dose, followed by 2 doses at 12‐hour intervals for 3 total doses. If less than 48 hours old, infants 0.1 mg/kg for second and third doses. If 2 to 7 days old, received 0.2 mg/kg for second and third doses. If 8 days of age or older, received 0.25 mg/kg for second and third doses. Those with failure to improve were referred for ligation or surgical closure. Additionally, infants were given standard medical therapy, consisting of fluid restriction, diuretic use, and occasionally, digoxin. The comparison group received standard medical therapy, which included fluid restriction, diuretic use, and occasionally, digoxin. If there was failure to improve within 36 to 48 hours, infants were randomized to receive indomethacin as described above or surgical treatment
Outcomes	Primary outcome Primary PDA closure Secondary outcomes Death Surgical ligation BPD Pneumothorax Intracranial hemorrhage Other bleeding NEC Sepsis Elevated creatinine clearance Retrolental fibroplasia grades 3 and 4 Duration of mechanical ventilation Age at home discharge
Notes	Study period: April 1979 to April 1981
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomized by random number generator to receive indomethacin and standard medical treatment or only standard medical treatment
Allocation concealment (selection bias)	Low risk	Randomization numbers were placed inside opaque envelopes
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Opaque vials containing indomethacin or placebo were given; both were colorless solutions
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Although not explicitly stated, blinding of treatment seems adequate to ensure that outcome assessment was accomplished in a blinded fashion
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Appropriate for primary outcome measure of PDA closure. Other longer‐term outcomes difficult to account for due to cross‐over treatment
Selective reporting (reporting bias)	Unclear risk	We cannot judge if there were any deviations from the original protocol
Other bias	Low risk	Appears free of other bias

Kluckow 2014.

Study characteristics
Methods	Multi‐center double‐blind placebo‐controlled randomized study conducted in the Neonatal Intensive Care Unit of 3 Australian tertiary neonatal intensive care units
Participants	Inclusion criteria: infants < 29 weeks' GA and < 12 hours with a large PDA (based on relevant 50th percentile postnatal age cut points): > 1.8 mm at postnatal age 3 to 5 hours, > 1.6 mm at 6 to 8 hours, and > 1.3 mm at 9 to 12 hours Exclusion criteria: grade 2 to 4 PIVH, PH, clinically unstable and not expected to survive, significant congenital anomaly, structural heart disease, contraindications to indomethacin (thrombocytopenia < 50,000/mm³, serum creatinine > 120 µmol/L, or severe growth restriction with abnormal umbilical Doppler arterial flows) If infants had ductal right‐to‐left shunting for more than one‐third of the cardiac cycle, randomization occurred only if resolved by 12 hours
Interventions	IV indomethacin group (n = 44) received a loading dose of 0.2 mg/kg followed by 0.1 mg/kg for 2 further doses at 24‐hour intervals administered by 30‐minute infusion Placebo group (n = 48) received a dose of 0.2 mg/kg followed by 0.1 mg/kg at 24‐hour intervals administered by 30‐minute infusion
Outcomes	Primary outcome Combined death and/or abnormal cranial ultrasound (IVH ≥ 2 or PVL) at discharge Secondary outcomes Mortality at hospital discharge IVH PVL Total doses of indomethacin Side effects of indomethacin Renal impairment (cr > 150 μmol/L, oliguria < 1.0 mL/kg/hr for > 12 hours) GI bleeding Intestinal perforation NEC ≥ stage 2 Time to full oral feeds (150 mL/kg/d) PH (frank blood in secretions) CXR changes Increase in mechanical ventilation ROP, sepsis (positive blood cultures and antibiotics > 5 days) Duration of mechanical ventilation CPAP Oxygen therapy BPD (oxygen requirement at 36 weeks' GA) Diuretics Postnatal steroid for BPD
Notes	Study was registered: ACTRN12608000295347 Study period: January 2007 to December 2010
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Block randomized by gestational age and center via random number generated tables
Allocation concealment (selection bias)	Low risk	Unclear whether randomization numbers were placed inside sequentially numbered opaque envelopes
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Staff were blinded to drugs administered
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Although not explicitly stated, blinding of treatment seems adequate to ensure that outcome assessment was accomplished in a blinded fashion
Incomplete outcome data (attrition bias) All outcomes	Low risk	All randomized infants are accounted for
Selective reporting (reporting bias)	Unclear risk	Study was registered: ACTRN12608000295347 Study was registered prospectively and there appear to be no deviations from the original protocol
Other bias	Low risk	Appears free of other bias

Knight 2011.

Study characteristics
Methods	Randomized study
Participants	Inclusion criteria: infants < 28 weeks' GA or < 1250 grams between 2 and 6 chronological days with PDA ≥ 1.5 mm and reversed descending aortic diastolic flow Exclusion criteria: not explicitly stated
Interventions	IV indomethacin group (n = 13) received indomethacin 200/100 mcg/kg IV every 24 hours for 6 days Control group (n = 13) received placebo
Outcomes	Primary outcomes Open‐label indomethacin treatment, peak inspiratory pressure, and fraction of inspired oxygen during treatment
Notes	Published only as an abstract. Study authors unavailable for further clarification
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	It is not stated how allocation was completed
Allocation concealment (selection bias)	Unclear risk	Unclear whether randomization numbers were placed inside sequentially numbered opaque envelopes
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	It is not clear if staff were blinded to drugs administered
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	It is not stated whether the person conducting outcome evaluation was blinded to intervention
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	It is not clear if all randomized infants are accounted for
Selective reporting (reporting bias)	Unclear risk	It is unclear if the study’s pre‐specified outcomes and all expected outcomes of interest to the review have been reported
Other bias	Unclear risk	Limited report

Krauss 1989.

Study characteristics
Methods	Single‐center randomized study from Cornell University Medical College, New York, USA
Participants	Inclusion criteria: infants < 1500 grams who were ventilator‐dependent at 24 hours of age Exclusion criteria: serum creatinine level > 110 μmol/L, serum urea nitrogen level > 10.5 mmol/L, platelet count < 60,000/mm³, urinary output < 0.6 mL/kg/hr over the 8 hours preceding evaluation, laboratory evidence of infection or of active bleeding, necrotizing enterocolitis, grade 4 intraventricular hemorrhage, congenital anomalies
Interventions	Intervention group (n = 12) received 3 doses of 0.2 mg/kg of IV indomethacin at between 72 and 96 hours of age Control group (n = 15) received standard medical management
Outcomes	Primary outcomes Arterial carbon dioxide pressure Arterial oxygen pressure pH Total pulmonary compliance Respiratory rate Measurement of inspired oxygen Concentration and calculated a‐A gradient at 1, 3, 5, and 7 days of age Secondary outcomes Mortality at hospital discharge
Notes	Study period: not described
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were assigned by random numbers
Allocation concealment (selection bias)	Unclear risk	Unclear how randomization numbers were concealed
Blinding of participants and personnel (performance bias) All outcomes	High risk	Staff was not blinded to intervention
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	It is not stated whether the person conducting outcome evaluation was blinded to intervention
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	All randomized infants are accounted for
Selective reporting (reporting bias)	Unclear risk	It is unclear if there were deviations from the original protocol
Other bias	Low risk	Appears free of other bias

Merritt 1981.

Study characteristics
Methods	Randomized controlled study conducted in the Neonatal Intensive Care Unit at the University of Rochester, New York, USA
Participants	Inclusion criteria: neonates weighing ≤ 1.35 kg with severe respiratory distress syndrome with clinical signs of hemodynamically significant PDA and cardiothoracic ratio > 0.6. If all signs not present, LA:Ao ratio had to exceed 1.2 Exclusion criteria: infants allocated to treatment group were excluded from further study if they developed total serum bilirubin > 10 mg/dL, blood urea nitrogen > 25 mg/dL, serum creatinine > 1.2 mg/dL, platelet count < 50,000/mm³, signs of necrotizing enterocolitis, clinical evidence of bleeding from the umbilicus, needle pricks, or finding of occult blood in the stool
Interventions	IV indomethacin group (n = 11) received 0.2 mg/kg every 24 hours for up to 3 days if signs of PDA persisted Medical management group (n = 13) received fluid restriction and/or furosemide 1 to 2 mg/kg every 12 hours. This group was not excluded from receiving indomethacin or surgical ligation
Outcomes	Primary outcomes Mortality BPD Secondary outcomes Surgical ligation Duration of assisted ventilation
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomized by random number generator
Allocation concealment (selection bias)	Unclear risk	It is unclear how randomization numbers were distributed
Blinding of participants and personnel (performance bias) All outcomes	High risk	Medical management group did not receive placebo, so staff was not blinded to administration
Blinding of outcome assessment (detection bias) All outcomes	High risk	Staff was not blinded to treatment arm
Incomplete outcome data (attrition bias) All outcomes	High risk	All randomized infants are not accounted for
Selective reporting (reporting bias)	Unclear risk	It is unclear if there were deviations from the original protocol
Other bias	Unclear risk	Given this was an unblinded study, it is unclear if there were other sources of bias

Monset‐Couchard 1983.

Study characteristics
Methods	Single‐center randomized trial conducted in Paris, France
Participants	24 preterm infants with symptomatic PDA (defined as presence of clinical signs of hemodynamically significant PDA and increased LA:Ao on echocardiogram)
Interventions	Intervention group (n = 12) received a single dose of 0.2 mg IV indomethacin just after diagnosis Control group (n = 12) received no treatment for 48 hours
Outcomes	Primary outcome PDA closure Secondary outcomes Receipt of repeat pharmacotherapy Surgical PDA ligation IVH (all grades) Severe IVH
Notes	French language study
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No method described
Allocation concealment (selection bias)	High risk	Reported as open random allocation
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not reported; however no treatment in the control arm suggests that personnel were likely not blinded
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported; unclear if outcome assessors were blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	All randomized infants are accounted for
Selective reporting (reporting bias)	Unclear risk	It is unclear if there were deviations from the original protocol
Other bias	Unclear risk	Given this was an unblinded study, it is unclear if there were other sources of bias

Nestrud 1980.

Study characteristics
Methods	Randomized controlled study conducted in the Neonatal Intensive Care Unit of University of Arkansas for Medical Sciences or St. Vincent’s Infirmary, Little Rock, AR, USA
Participants	Inclusion criteria: < 35 weeks' GA, large left‐to‐right PDA, determined with clinical signs, radiography, and echo or cardiac catheterization. LA:Ao < 1.3 did not exclude patients from study if there were overwhelming signs of cardiac congestion Exclusion criteria: serum creatinine ≥ 1.5 mg/dL and inadequate urine output, serum indirect bilirubin ≥ 10 mg/dL, evidence of bleeding from lungs or GI tract, platelet count < 100,000/mm³ Note: all patients had failed to respond to standard medical management, including fluid restriction, diuretics, and maintenance of hematocrit > 40%. All but 1 received lanoxin before study placement
Interventions	Oral indomethacin group (n = 12) received 0.2 mg/kg diluted to 1 mL with saline every 12 hours up to 3 doses Placebo group (n = 11) received 1 mL of saline at the same frequency
Outcomes	Response to treatment BPD Mortality Requirement of repeat dosing Surgical ligation Late‐onset sepsis Creatinine levels after treatment Increase in creatinine levels after treatment
Notes	Study period: March 1977 to July 1978
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomized by random numbers to receive either oral indomethacin or placebo
Allocation concealment (selection bias)	Low risk	Randomization numbers were kept by a pharmacist not involved in patient care
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Indomethacin was diluted to 1 mL with saline; thus both groups received the same volume
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Although likely, it is not stated if individuals assessing response were blinded to treatment
Incomplete outcome data (attrition bias) All outcomes	Low risk	All randomized infants are accounted for
Selective reporting (reporting bias)	Unclear risk	It is unclear if there were deviations from the original protocol
Other bias	Low risk	Study appears free of other bias

Neu 1981.

Study characteristics
Methods	Single‐center randomized cross‐over study of infants with respiratory distress syndrome diagnosed with a clinically significant PDA conducted at Stanford, CA, USA Study duration: 2 years
Participants	Preterm infants were included if they had clinical signs of a significant PDA and failed to respond to fluid restriction
Interventions	Oral indomethacin group (n = 11) received 0.25 mg/kg of indomethacin. If clinical signs persisted and LA:Ao ratio had not decreased, the same drug was given 12 to 24 hours later. If clinical signs persisted at 48 hours, the alternative drug was given 1 time Placebo group (n = 10) received 0.25 mg/kg of saline. If clinical signs persisted and LA:Ao ratio had not decreased, the same drug was given 12 to 24 hours later. If clinical signs persisted at 48 hours, the alternative drug was given 1 time
Outcomes	Primary outcome Evidence of positive response to therapy based on clinical criteria (resolved murmur, diminished bounding arterial pulse, improved pulmonary vascularity or CT ratio on CXR, decreased LA:Ao ratio)
Notes	The only relevant outcome before cross‐over was response to initial pharmacotherapy. All outcomes post cross‐over were not considered in the review
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not mentioned
Allocation concealment (selection bias)	Unclear risk	Not mentioned
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Placebo was prepared by pharmacy such that indomethacin and placebo were indistinguishable
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Study authors mention (quote): "the study code was not broken until the entire study was completed and the clinical data were evaluated to determine whether, and after which dose, a response occurred" This suggests that outcome assessors were likely blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	All randomized infants were accounted for with respect to the pre‐crossover outcome under consideration
Selective reporting (reporting bias)	Unclear risk	It is unclear if there were deviations from the original protocol
Other bias	Low risk	Study appears free of other bias

Osborn 2003.

Study characteristics
Methods	Randomized double‐blind cross‐over study conducted at the Royal Prince Alfred Hospital and Royal North Shore Hospital Neonatal Intensive Care Units, Sydney, Australia
Participants	Inclusion criteria: preterm infants with gestational age < 30 weeks and < 12 hours of age were eligible. Infants were enrolled if PDA diameter > 1.6 mm Exclusion criteria: major cerebral hemorrhage (grade 3 or 4), major congenital abnormality, cardiac abnormality identified, considered non‐viable by attending, inotropic agents had been given, indomethacin had been given
Interventions	IV indomethacin group (n = 35) received a single dose of 0.2 mg/kg Placebo group (n = 35) received saline of uncertain quantity Infants were evaluated by echocardiography 1 hour following infusion and were crossed over to the other treatment if PDA diameter remained > 1.6 mm and < 30% constriction from initial diameter occurred
Outcomes	Primary outcome Change in SVC flow Secondary outcomes Death IVH PVL NEC
Notes	Study period: October 1998 to September 1999 All reported secondary outcomes were post cross‐over; hence not included in the meta‐analysis
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Unclear how randomization occurred
Allocation concealment (selection bias)	Low risk	Treatment and placebo were placed in identical vials
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Interventions were given at the same time period
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	It is not stated whether the person conducting echocardiography was blinded to treatment
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	All randomized infants are accounted for
Selective reporting (reporting bias)	Unclear risk	It is unclear if the study deviated from the original protocol
Other bias	Low risk	Study appears free of other bias

Rudd 1983.

Study characteristics
Methods	Randomized controlled study conducted at Royal Postgraduate Medical School, Hammersmith Hospital, London, UK Study duration: 21 months
Participants	Inclusion criteria: infants < 1500 grams who developed a murmur characteristic of PDA, had full pulses, had at least 2 of the following: active cardiac impulse, loud pulmonary S2 component, cardiothoracic ratio ≥ 0.6, pleonemia, LA:Ao ratio ≥ 1.2 Exclusion criteria: hemoglobin ≤ 10 g/dL, serum bilirubin > 11.7 mg/100 mL, platelet count < 80 000/mm³, sodium < 120 mmol (mEq)/L, potassium > 6 mmol (mEq)/L, NEC
Interventions	Infants were fluid‐restricted after PDA diagnosis for 24 hours. If still meeting criteria, infants were randomized to receive the following Oral indomethacin group (n = 15) received 0.2 mg/kg every 12 hours up to 3 doses via nasogastric tube Placebo group (n = 15) received saline via nasogastric tube in a preparation visually identical to the indomethacin preparation Fluid restriction was continued 48 hours after last dose of drug
Outcomes	Outcomes Early successful response PDA recurrence Response to repeat treatment Indomethacin therapy Surgical closure GI bleeding Hyponatremia Death at discharge
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomized by numbered code by batching of 10
Allocation concealment (selection bias)	Unclear risk	It is unclear how allocation numbers were concealed
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Preparations appeared visually identical and both were given at the same time
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	It is not stated if those assessing outcomes were blinded to treatment
Incomplete outcome data (attrition bias) All outcomes	Low risk	All randomized infants are accounted for
Selective reporting (reporting bias)	Unclear risk	It is unclear if there were deviations from the study protocol
Other bias	Low risk	Study appears free of other bias

Valaes 1980.

Study characteristics
Methods	Randomized controlled study conducted in the Tufts – New England Medical Center Hospital, Boston, USA
Participants	Inclusion criteria: preterm infants with evidence of symptomatic PDA defined by congestive heart failure and/or respiratory distress plus a large left‐to‐right shunt noted on echocardiogram Exclusion criteria: early death (< 3 days old)
Interventions	Oral indomethacin group (n = 7) received 0.2 mg/kg/dose in 1 to 3 doses It is unclear what intervention the placebo group (n = 6) received
Outcomes	Primary outcome PDA closure by echocardiogram and resolution of symptomatic status
Notes	Placebo group could receive indomethacin if these participants remained symptomatic at undefined time intervals
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	It is unclear how randomization occurred
Allocation concealment (selection bias)	Unclear risk	It is unclear how allocation was concealed
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	It is unclear how blinding of administration occurred
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	It is not stated whether persons conducting echocardiogram or post‐treatment evaluations were blinded to treatment
Incomplete outcome data (attrition bias) All outcomes	Low risk	All randomized infants are accounted for
Selective reporting (reporting bias)	Unclear risk	It is unclear if there are any deviations from the original protocol
Other bias	Low risk	Appears free of other bias

Yanagi 1981.

Study characteristics
Methods	Randomized controlled study conducted at Prentice Women’s Hospital and Children’s Memorial Hospital, Chicago, USA
Participants	Inclusion criteria: preterm infants with severe RDS who required ventilator support and a symptomatic PDA that had failed a 24‐hour course of standardized medical management including fluid restriction, digoxin, and furosemide. The PDA was considered to be symptomatic only if the LA:Ao ratio was ≥ 1.3:1 and continuous ventilatory support was required Exclusion criteria: serum creatinine level > 2.0 mg/100 mL, platelet count < 60,000/mm³, indirect bilirubin level > 12.0 mg/100 mL, abnormal PT/PTT
Interventions	In phase 1, oral indomethacin group (n = 8) received 3 doses of 0.2 mg/kg every 24 hours in the first phase of the study. In phase 2, infants (n = 13) received 3 doses of 0.2 mg/kg every 8 hours In phase 1, control group (n = 7) received placebo every 24 hours. In phase 2, control group (n = 9) received saline every 8 hours
Outcomes	Primary outcome Improvement in clinical signs and symptoms 24 hours after last dose of indomethacin or placebo. Infants were considered improved if they could be weaned from the ventilator and/or if LA:Ao ratio decreased to < 1.3:1 Secondary outcomes Death before hospital discharge Duration of ventilator support Duration of supplemental oxygen therapy Urine output as percentage of total daily fluid intake Change in serum sodium levels Change in serum creatinine levels Change in platelet count Change in PT/PTT Upper gastrointestinal bleeding
Notes	Study period: February 1977 to March 1979
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomization occurred by random number generator
Allocation concealment (selection bias)	Low risk	Allocation was concealed from both physician and pharmacist
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Treatment and placebo were given at the same time interval and in the same volume. They were visually similar in appearance
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Assessments of patients were made while treatment allocation was still concealed
Incomplete outcome data (attrition bias) All outcomes	Low risk	All randomized infants are accounted for
Selective reporting (reporting bias)	Unclear risk	It is unclear if there were any deviations from the original protocol
Other bias	Low risk	Appears free of other bias

Yeh 1981a.

Study characteristics
Methods	Randomized controlled study conducted at Cook County Hospital (CCH), Chicago, USA
Participants	Inclusion criteria: all infants born at CCH weighing < 2040 grams, all neonates transferred to CCH for intensive care were considered. Further criteria included evidence of a PDA and evidence of cardiovascular dysfunction after medical therapy Exclusion criteria: total serum bilirubin > 10 mg/dL, BUN > 20 mg/dL, evidence of shock, evidence of PDA closure, cerebral hemorrhage, NEC, hemorrhagic disease
Interventions	Intravenous indomethacin group (n = 28) received 0.2 mg/kg/dose every 24 hours for up to 3 doses unless the PDA murmur disappeared Control group (n = 27) received 0.2 mg/kg/dose of saline every 24 hours for up to 3 doses unless the PDA murmur disappeared All infants considered for inclusion with a presumptive PDA diagnosis received 1 dose of furosemide (1 mg/kg IV) and 24 hours of fluid restriction
Outcomes	Primary outcome Successful PDA closure Secondary outcomes Surgical PDA ligation Successful wean from ventilator support Mortality at discharge Mortality at 1 year Duration of hospitalization Incidence of BPD Incidence of NEC Incidence of sepsis Evidence of ROP Changes in urine output Changes in serum electrolytes Bayley scores at 1 year
Notes	Study period: November 1977 to February 1979
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomization occurred by random number generator in blocks of 10
Allocation concealment (selection bias)	Low risk	Randomization numbers were contained in sealed envelopes under control of designated pharmacist
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Treatment and placebo were given at the same time interval and in the same volume
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Assessments of patients were made while treatment allocation was still concealed
Incomplete outcome data (attrition bias) All outcomes	Low risk	All randomized infants were accounted for
Selective reporting (reporting bias)	Unclear risk	It is unclear if there were any deviations from the original protocol
Other bias	Low risk	Appears free of other bias

BPD: bronchopulmonary dysplasia; BUN: blood urea nitrogen; CPAP: continuous positive airway pressure; CT: cardiothoracic; CXR: chest X‐ray; GA: gestational age; GI: gastrointestinal; IV: intravenous; IVH: intraventricular hemorrhage; LA/Ao: left atrium‐to‐aortic root ratio; NEC: necrotizing enterocolitis; PDA: patent ductus arteriosus; PH: pulmonary hypertension; PIVH: peri/intraventricular hemorrhage; PT/PTT: prothrombin time/partial thromboplastin time; PVL: periventricular leukomalacia; RDS: respiratory distress syndrome; ROP: retinopathy of prematurity; S2: second heart sound; SVC: superior vena cava.

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Carmo 2009	Excluded from this review because infants in both arms of this trial received indomethacin
Mardoum 1991	Excluded from this review because infants in both arms of the study were crossed to receive indomethacin regardless of change in status of a clinically significant PDA
Mullett 1982	Excluded from this review because included participants did not clearly have a clinically significant PDA
Nair 1986	Excluded from this review because included participants did not clearly have a clinically significant PDA
Nuntnarumit 2011	Excluded from this review due to lack of randomization
Yeh 1982a	Excluded from this review as infants in both arms of the study received indomethacin

PDA: patent ductus arteriosus.

Characteristics of studies awaiting classification [ordered by study ID]

Clyman 2019.

Methods	Multi‐center randomized controlled trial (PDA‐TOLERATE) conducted across 17 international centers
Participants	Preterm infants at < 28 weeks' gestation with moderate to large PDA shunt and receiving greater than minimal respiratory support at between 6 and 14 days of age
Interventions	Early routine treatment (n = 104): treatment with indomethacin, ibuprofen, or acetaminophen initiated at 6 to 14 days of age Conservative treatment (n = 98): no pharmacological treatment for PDA
Outcomes	Primary outcomes PDA ligation Presence of a PDA at discharge
Notes	Data on the subgroup of infants randomized to indomethacin vs placebo sought from the contact author

PDA: patent ductus arteriosus.

Characteristics of ongoing studies [ordered by study ID]

NCT03456336.

Study name	Management of the PDA Trial (PDA)
Methods	Multi‐center randomized controlled trial across 16 centers in the USA
Participants	Infants at 22 0/7 to 28 6/7 weeks' gestation at birth with postnatal age of 48 hours to 21 days. Significant PDA defined by clinical and echocardiographic characteristics
Interventions	Intervention group will receive indomethacin or ibuprofen per local site standard dosing and schedule Comparison group will receive expectant management, unless they develop cardiopulmonary compromise. At this point, they are given indomethacin or ibuprofen
Outcomes	Death or BPD at 36 weeks' PMA Mortality at 36 weeks' PMA Mortality at discharge BPD ‐ physiologic test of oxygen therapy BPD ‐ NIH consensus definition NEC at 36 weeks' PMA and at 2 years ROP at 36 weeks' PMA and at 2 years PDA ligation or cardiac catheterization between birth and 120 days NDI at 2 years Anthropomorphic measurements at 36 weeks' PMA and at 2 years
Starting date	3 December 2018
Contact information	Dr M Laughon, Division of Neonatal‐Perinatal Medicine, UNC School of Medicine
Notes

BPD: bronchopulmonary dysplasia; NEC: necrotizing enterocolitis; NDI: neurodevelopmental impairment; NIH: National Institutes of Health; PDA: patent ductus arteriosus; PMA: postmenstrual age; ROP: retinopathy of prematurity.

Differences between protocol and review

1. As of July 2019, Cochrane Neonatal no longer searches Embase for its reviews. RCTs and controlled clinical trials (CCTs) from Embase are added to the Cochrane Central Register of Controlled Trials (CENTRAL) via a robust process (see How CENTRAL is created). Cochrane Neonatal has validated its searches to ensure that relevant Embase records are found while searching CENTRAL

2. Also starting in July 2019, Cochrane Neonatal no longer searches for RCTs and CCTs on the following platforms: ClinicalTrials.gov or the World Health Organization International Clinical Trials Registry Platform (ICTRP), as records from both platforms are added to CENTRAL on a monthly basis (see How CENTRAL is created). Comprehensive search strategies are executed in CENTRAL to retrieve relevant records. The International Standard Randomized Controlled Trials Number (ISRCTN) platform (at http://www.isrctn.com/, formerly Controlled‐trials.com) is searched separately

3. The definition of symptomatic PDA was clarified. We specified that a PDA was defined as symptomatic if the infant met any one of the listed clinical and/or echocardiographic criteria (and the latter included Doppler criteria).

4. The primary outcome was changed from 'PDA closure within one week of administration of the first dose of indomethacin' to 'failure of PDA closure within one week of administration of the first dose of indomethacin', in keeping with similar reviews on PDA management in preterm infants

5. Neurodevelopmental outcome was listed twice in the protocol. We removed one mention and kept the other, which had a more comprehensive definition of neurodevelopmental outcome.

6. We did not perform GRADE assessment of certainty of evidence on the following pre‐specified outcomes due to lack of data: neurodevelopmental outcome at approximately 2 years' corrected age; intestinal perforation; oliguria; increase in serum/plasma creatinine (µmol/L) levels after treatment; and pulmonary hypertension

Contributions of authors

PE and RS conceived the project and drafted the protocol.

PE, DO, JNF, and SM identified eligible trials for inclusion and performed data abstraction.

PE and SM performed data analyses.

PE, SM, and RS drafted the review.
PE, DO, JNF, SM, and RS reviewed all drafts and approved the final version of the protocol and the review.

Sources of support

Internal sources

• No sources of support supplied

External sources

• Vermont Oxford Network, USA

  Cochrane Neonatal Reviews are produced with support from Vermont Oxford Network, a worldwide collaboration of health professionals dedicated to providing evidence‐based care of the highest quality for newborn infants and their families.

• National Institute for Health Research, UK

  This report is independent research funded by a UK NIHR Cochrane Programme Grant (16/114/03). The views expressed in this publication are those of the review authors and are not necessarily those of the National Health Service, the NIHR, or the UK Department of Health.

Declarations of interest

PE is a Pediatric Resident at the Children's Hospital of Philadelphia. As a pediatric resident, he participated in the care of patients with PDA on his neonatology rotations.

DO has no interest to declare.

JNF is a Paediatric Cardiologist at the University of Vermont Medical Center.

RS is a Professor of Neonatology at the University of Vermont Medical Center. He needs to make the decisions detailed in the review on a daily basis.

SM is a neonatologist at a tertiary care neonatal intensive care unit in Halifax, Nova Scotia, Canada, where he attends to preterm infants diagnosed with a PDA. He has published review articles on PDA management in preterm infants.

New