4.

Oral ibuprofen compared with intravenous ibuprofen for patent ductus arteriosus
Patient or population: preterm infants with patent ductus arteriosus Settings: NICU Intervention: oral ibuprofen Comparison: intravenous ibuprofen
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	intravenous ibuprofen	oral ibuprofen
Failure to close a patent ductus arteriosus (after single or 3 doses)	High risk population		RR 0.38 (0.26 to 0.56)	406 (5)	⊕⊕⊕⊝ moderate	Bias: there was unclear risk of bias for random sequence generation in all 5 studies. There was low risk of bias for allocation concealment in 4 studies, and unclear risk in one study. The blinding of personnel was inadequate in all five studies and blinding of outcome assessments was good in three studies but inadequate in two studies. We downgraded the evidence by one step. Heterogeneity/consistency: we noted no heterogeneity for RR and for RD (0%). Directness of evidence: studies were conducted in the target population. Precision: the confidence intervals around the point estimates for RR and RD were quite narrow. Presence of publication bias: only 5 studies were included in the analysis so a funnel plot was not constructed.
	363 per 1000	139 per 1000 (115 to 156)
Need for surgical closure of the ductus	High risk population		RR 0.41 (0.41 to 1.21)	406 (5)	⊕⊕⊕⊝ moderate	Bias: there was unclear risk of bias for random sequence generation in all 5 studies. There was low risk of bias for allocation concealment in 4 studies, and unclear risk in one study. The blinding of personnel was inadequate in all five studies and blinding of outcome assessments was good in three studies but inadequate in two studies We downgraded the evidence by one step. Heterogeneity/consistency: we noted no heterogeneity for RR and for RD (0%). Directness of evidence: studies were conducted in the target population. Precision: the confidence intervals around the point estimates for RR and RD were quite narrow. Presence of publication bias: only 5 studies were included in the analysis so a funnel plot was not constructed.
	51 per 1000	19 per 1000 (0 to 31)
Duration of ventilatory support	High risk population		MD 0.54 (days) (‐0.01 to 1.10)	134 (2)	⊕⊕⊝⊝ low	Bias: there was unclear risk of bias for random sequence generation in both studies. There was low risk of bias for allocation concealment in one study, and unclear risk in one study. The blinding of personnel was inadequate in both studies and blinding of outcome assessments was good in one study but inadequate in two one study. We downgraded the evidence by one step. Heterogeneity/consistency: we noted no heterogeneity for MD (10%). Directness of evidence: studies were conducted in the target population. Precision: the confidence intervals around the point estimates for RR and RD were quite wide. We downgraded the evidence by one step. Presence of publication bias: only two studies were included in the analysis so a funnel plot was not constructed.
	The mean duration of ventilatory support (days) ranges across control groups from 3 to 5.1 days	The mean duration of ventilatory support was 0.54 (days) higher in the in the oral ibuprofen group (‐0.01 to 1.10)
Serum/plasma creatinine levels (μmol/L) after treatment Normal values for male and female newborns 17.7 to 88.4 µmol/L	The mean serum/plasma creatinine levels ranged across control groups from 69.84 to 76.02 (μmol/L)	The mean serum/plasma creatinine level in the intervention groups was 22.47 (μmol/L) lower (‐32.40 to ‐12.53)	MD 22.47 (μmol/L) (‐32.40 to ‐12.53)	170 (2)	⊕⊕⊝⊝ low	Bias: there was unclear risk of bias for random sequence generation in both studies. There was low risk of bias for allocation concealment in one study, and unclear risk in one study. The blinding of personnel was inadequate in both studies and blinding of outcome assessments was good in one study but inadequate in the other study. We downgraded the evidence by one step. Heterogeneity/consistency: there was high heterogeneity for MD (81%). We downgraded the evidence by one step. Directness of evidence: studies were conducted in the target population. Precision: the confidence intervals around the point estimates for MD were quite narrow. Presence of publication bias: only 2 studies were included in the analysis so a funnel plot was not constructed.
Oliguria (Urine output < 1 mL/kg/hour)	High risk population		RR 0.14 (0.01 to 2.66)	304 (4)	⊕⊕⊝⊝ low	Bias: there was unclear risk of bias for random sequence generation in all 4 studies. There was low risk of bias for allocation concealment in three studies, and unclear risk in one study. The blinding of personnel was inadequate in all four studies and blinding of outcome assessments was good in three studies but inadequate in one study. We downgraded the evidence by one step. Heterogeneity/consistency: tests for heterogeneity were not applicable for RR as there were only outcomes in one group in one trial. We noted no heterogeneity for RD (19%). Directness of evidence: studies were conducted in the target population. Precision: the confidence interval around the point estimate for RR was quite wide. We downgraded the evidence by one step. Presence of publication bias: only 4 studies were included in the analysis so a funnel plot was not constructed.
	2 per 1000	0 per 1000 (0 to 0)
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (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; IV: intravenous; MD: Mean difference; NICU: Neonatal intensive care unit; RR: Risk Ratio
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: 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. Very low quality: We are very uncertain about the estimate.