Bouwmeester 2003a.

Study characteristics
Methods	Study design: Double‐blind, randomized clinical trial Study grouping: Continous morphine versus three‐hourly placebo or intermittent morphine with placebo infusion. Additional morphine was administered on guidance of pain outcomes.
Participants	Baseline Characteristics: Neonates aging from 0 to 4 weeks (gestational age between 35 to 42 weeks; body weight 1500 g or more) Inclusion criteria: Included neonates were admitted to the pediatric surgical intensive care unit following non‐cardiac thoracic and abdominal surgery. Exclusion criteria: Patients were excluded if they had received morphine < 6 h before surgery, or suffered from hepatic, renal or neurological disorders. Pretreatment: Anesthesia was induced intravenously with 3 to 5 mg/kg thiopentone or by inhalation with halothane in oxygen. Fentanyl at 5 mcg/kg was given before orotracheal intubation, which was facilitated with 0.5 to 1 mg/kg atracurium or 2 mg/kg suxamethonium.
Interventions	Intervention Characteristics Dose regimen: Continous morphine infusion (10 mcg/kg/h) versus intermittent morphine (10 mcg/kg every 3 hours) Description of procedure: All neonates received a dose of morphine hydrochloride 100 mcg/kg at the end of surgery and were randomly allocated to equivalent intravenous doses of continuous morphine infusions (10 mcg/kg/hour) or intermittent morphine boluses (10 mcg/kg/3 hours).
Outcomes	Pain assessed with validated methods during the administration of selected drugs Outcome type: Continuous outcome Pain was assessed by nurses trained in the use of the behavioral part of the COMFORT scale (CS), the total score of which can range from 6 to 30, and a 0 ± 10 visual analogue scale (VAS).
Identification	Sponsorship source: The study was supported by the Netherlands Research Council and the Sophia Foundation for Medical Research. Country: The Netherlands Setting: Pediatric Intensive Care Unit Comments: None Authors names: Nancy J. Bouwmeester, Wim C. J. Hop, Monique van Dijk, K.J.S. Anand, John N. van den Anker, and Dick Tibboel Institution: Department of Anesthesiology and Pediatric Surgery, Sophia Children's Hospital Email: j.bouwmeester.1@erasmusmc.nl Address: Sophia Children's Hospital, University Hospital Rotterdam, Dr Molewaterplein 60, 3015 GJ Rotterdam, The Netherlands
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"They were randomly assigned to receive either i.v. continuous morphine (CM) or intermittent morphine (IM)." Computer‐generated (stratification)
Allocation concealment (selection bias)	Low risk	"The pharmacist prepared all study drugs and strata‐specific schedules for randomization." "For each age group, the boxes containing the study drugs were numbered consecutively and used in sequence. Each patient received a study number consisting of a number of the age group (I ± IV) and a sequence number (1 ± 68)." Concealed
Blinding of participants and personnel (performance bias) All outcomes	Low risk	"For children in the CM group this was followed by a morphine infusion 10 mg kg‐1 h‐1, combined with 3‐hourly i.v. placebo (saline) boluses." "Pharmacist prepared all study drugs and strata‐specific schedules for randomization." "The amount of glucose and the volume of fluid was the same in both treatment groups." Blinded
Blinding of outcome assessment (detection bias) All outcomes	Low risk	"The clinical staff were blinded to the study group allocation until data collection was complete."
Incomplete outcome data (attrition bias) All outcomes	Low risk	"Logistic and laboratory problems resulted in missing data for several of the 197 included patients." "All 204 patients were included in an intention‐to‐treat analysis." Limited missing data
Selective reporting (reporting bias)	Unclear risk	No registry reported so unable to compare to protocol
Other bias	Low risk	None