Background: A patent ductus arteriosus (PDA) often complicates the clinical course of preterm infants and increases the risk of intraventricular hemorrhage (IVH), necrotizing enterocolitis (NEC), chronic lung disease (CLD) and death. The standard treatment to close a PDA is indomethacin. Its use is associated with renal, gastrointestinal and cerebral side-effects. Ibuprofen has been shown to be effective in closing a PDA without reducing blood flow velocity to the brain, gut or kidneys.
Objectives: To determine the effectiveness and safety of prophylactic ibuprofen compared to placebo/no intervention or other cyclo-oxygenase inhibitor drugs (indomethacin, mefenamic acid, etc) in the prevention of PDA in preterm infants.
Search strategy: Randomized controlled trials comparing prophylactic ibuprofen use with placebo/no intervention/indomethacin were identified by searching the Cochrane Central Register of Controlled Trial (CENTRAL, The Cochrane Library, Issue 3, 2005), MEDLINE (1966-July 2005), CINAHL (1982-July 2005), EMBASE (1980-July 2005), reference lists of published trials and abstracts published in Pediatric Research (1990-July 2005). No language restrictions were applied.
Selection criteria: Randomized or quasi-randomized controlled trials comparing use of ibuprofen with placebo/no intervention or other cyclo-oxygenase inhibitor drugs (indomethacin, mefenamic acid, etc) for the prevention of PDA in preterm and/or low birth weight infants.
Data collection and analysis: Data regarding the clinical outcomes including presence of PDA on day three and day seven, need for surgical ligation, need for rescue treatment with cyclo-oxygenase inhibitors, IVH, mortality, renal and gastrointestinal complications were extracted. Meta-analyses were performed using RevMan 4.2 and treatment estimates were reported as weighted mean difference (WMD), typical relative risk (RR), typical risk difference (RD) and, if statistically significant, number needed to treat (NNT) or number needed to harm (NNH), along with their 95% confidence intervals (CI).
Main results: Four trials (n = 672) were included in the review. There was a statistically significant decrease in the incidence of PDA on day three in the ibuprofen group [typical RR 0.37 (95% CI 0.29, 0.49); typical RD -0.29 (95% CI -0.35, -0.22); NNT 3 (95% CI 3, 5); 4 trials, n = 672], in the need for rescue treatment with cyclo-oxygenase inhibitors [typical RR 0.17 (95% CI 0.11, 0.27), typical RD -0.27 (95% CI -0.35, -0.22); NNT 4 (95%CI 3, 5), and in the need for surgical ligation [typical RR 0.34 (95% CI 0.14, 0.81), typical RD -0.04 (95% CI -0.07, -0.01); NNT 25 (95% CI 14, 100). The PDA had closed spontaneously by day three in 60% of the neonates in the control group. There was a significant increase in the serum creatinine levels in the ibuprofen group [WMD 0.13 mg/dl (95% CI 0.08, 0.17); 2 trials, n = 495]. Ibuprofen reduces urine output. There were no statistically significant differences in mortality, grade 3/4 intraventricular hemorrhage, chronic lung disease at 28 days or 36 weeks, necrotizing enterocolitis , gastrointestinal hemorrhage, intestinal perforation or time to reach full feeds. One trial (Gournay 2002) (n = 135) reported on three infants in the ibuprofen group who developed pulmonary hypertension responsive to nitric oxide treatment.
Authors' conclusions: Prophylactic use of ibuprofen reduces the incidence of PDA, the need for rescue treatment with cyclo-oxygenase inhibitors and surgical closure. However, in the control group, the PDA had closed spontaneously by day three in 60% of the neonates. Prophylactic treatment therefore exposes a large proportion of infants unnecessarily to a drug that has important side effects (mainly involving the kidneys) without conferring any important short term benefits. Prophylactic treatment with ibuprofen is not recommended. Until long-term follow-up results are published from the trials included in this review, no further trials of prophylactic ibuprofen are recommended.