The diagnosis of a patent ductus arteriosus (PDA) in preterm neonates is unreliable based on clinical and radiographic findings and requires the use of echocardiography (1). The incidence of a PDA diagnosed by echocardiography in neonates with a birth weight of less than 1750 g is 43% (2). The presence of a PDA is associated with reduced middle cerebral artery blood flow velocity (3). A PDA is associated with an increased risk of intraventricular hemorrhage, necrotizing enterocolitis, chronic lung disease, and death (4).

Historically, progress in the management of the preterm infant with a PDA includes surgical closure introduced in the early 1970s by Kitterman et al. (5) and the use of i.v. indomethacin, a prostaglandin synthesis inhibitor introduced by Heyman et al. (6) and Friedman et al. (7) in the mid-1970s. Later it was shown in randomized controlled trials that the prophylactic use of indomethacin in infants at risk of a clinically significant PDA reduces the risk of a symptomatic PDA and severe intraventricular hemorrhage, but the effectiveness with regard to long-term neurodevelopment outcomes was poorly ascertained (2). As indomethacin reduces cerebral blood flow velocity (CBFV) (8), treatment of a PDA with indomethacin may contribute to the development of periventricular leukomalacia, a condition associated with a very high risk for adverse neurodevelopmental outcome. The proven effectiveness of indomethacin in reducing the incidence of intraventricular hemorrhage might therefore be counteracted by an increase in periventricular leukomalacia, and as a result there would be no net gain in the number of neonates with normal neurodevelopmental outcome. The recruitment phase of an ongoing international multicenter trial of prophylactic treatment with indomethacin to assess its effectiveness in reducing long-term adverse outcomes has been completed [Trial of Indomethacin Prophylaxis in Preterm Infants (TIPP)]. Results on the primary outcome of death or impairment at 18–21 mo corrected gestational age will be available in the year 2000 (B. Schmidt, Mc Master University Hamilton, Ontario, Canada, 1999, personal communication).

In the mid-1970s Coceani and Olley were concerned about the toxic effects of indomethacin and conducted experimental work on the ductus with another prostaglandin inhibitor, ibuprofen (9, 10). In an abstract on the contractile response of the developing lamb ductus arteriosus to ibuprofen, they concluded: “The relatively greater effectiveness of ibuprofen in younger fetuses suggests its usefulness in the management of the premature human infant with patent d.a.” (9). Coceani recalls “At the time industry had difficulties to prepare an intravenous formulation, and consequently the researchers resorted to using ibuprofen suppositories. However, for some reason the drug was not well absorbed by this route and the results were inconclusive. Meanwhile, other investigators used indomethacin, a treatment that became standard care” (F. Coceani, Scuola Superiore S. Anna, Pisa, Italy, 1999, personal communication). Ironically 20 y later, in Toronto, we experienced the same difficulties in obtaining an i.v. preparation of ibuprofen to study, in a randomized controlled trial, its effect on CBFV (using color Doppler techniques) in comparison to indomethacin.

In 1996 Varvarigou et al. demonstrated in a randomized controlled trial that three doses of ibuprofen administered within 3 h after birth in preterm neonates reduced the incidence of PDA without notable adverse drug reactions (11). The authors did not study the effect of ibuprofen on CBFV or cerebral blood flow (CBF).

In this issue of Pediatric Research, Patel et al. (12) report on the results of a well-designed and conducted double-blind randomized controlled trial using near infrared spectroscopy to assess the effects of ibuprofen and indomethacin on cerebral hemodynamics in preterm infants with a PDA. The primary endpoints were the effects on CBF and cerebral blood volume (CBV) in response to the first dose of ibuprofen or indomethacin. The study confirms previous reports that ibuprofen does not reduce CBF or CBFV in contrast to treatment with indomethacin (13, 14). Contrary to a previous study (13), the authors found no change in CBV. There was no statistically significant difference in the success rates for PDA closure between the two drugs. The results of this study adds to the increasing evidence that ibuprofen might become an alternative to indomethacin, as a drug to prevent a PDA from becoming clinically significant or to treat a PDA diagnosed by echocardiographic criteria in preterm infants.

Table 1 summarizes the findings of studies on ibuprofen in relation to PDA closure published to date. Using meta-analytic techniques (Review Manager, May 1998, Version 3.1.1, The Cochrane Collaboration), a more precise estimate of the effectiveness of ibuprofen versus indomethacin in closing a PDA can be calculated. Combining the data from the three trials (12, 15, 16) that used the same dosing regimen results in a relative risk (RR) of 1.0 (95% confidence interval 0.85, 1.17) and a risk difference (RD) of −0.3% (95% confidence interval −12.7, 12.0%). Although only 176 neonates have been enrolled in such studies to date, the RR, RD, and the relatively narrow confidence intervals makes one quite confident that the drugs are equally effective in closing a PDA. As ibuprofen does not reduce CBF, it should not influence the incidence of periventricular leukomalacia, but whether it reduces the incidence of intraventricular hemorrhage or retinopathy of prematurity (ROP) is not known. Therefore comparative studies of indomethacin versus ibuprofen with regard to long-term outcomes are needed before ibuprofen replaces indomethacin as the standard of practice.

Table 1 Randomized controlled studies on the effectiveness of intravenous ibuprofen vs intravenous indomethacin to close a patent ductus arteriosus in preterm infants
Full size table

The important next step would be to conduct randomized controlled trials of adequate power to study the long-term effects of ibuprofen versus indomethacin for the closure of a PDA. Both the use of ibuprofen versus indomethacin for the treatment of a PDA diagnosed by echocardiography and the prophylactic use of ibuprofen versus indomethacin at birth to prevent long-term adverse sequelae should be studied. For both such studies the primary outcome should be rate of impairment at 2 y of age. Secondary outcomes would be short-term outcomes such as the rate of ductal closure, intraventricular hemorrhage, periventricular leukomalacia, necrotizing enterocolitis, ROP, renal impairment etc. Assessment of the cost effectiveness of one drug versus the other should be part of the study design. The most efficient study design may be a trial with four arms; where infants of a certain gestational age or birth weight are randomized and allocated at birth either to a prophylaxis arm starting immediately or to an arm in which treatment starts when a PDA is diagnosed by echocardiography within the first 7 d of life. Within the prophylaxis and treatment arms, the infants would be randomized to either ibuprofen or indomethacin. The required sample size would be based on the final results of the on-going TIPP-study.