Stevens 1999.

Methods	Double‐centre randomised controlled trial performed in Canada.
Participants	120 newborn preterm neonates, range 30 to 36 weeks' gestation (56 intervention, 56 placebo) undergoing a heel prick aged 1 to 5 days' postnatal age in 2 NICU.
Interventions	The study was undertaken in 2 phases. For phase 1, 31 infants in the experimental group receive EMLA 0.5 g on an occlusive dressing for 30 minutes. 5 minutes before heel lance, the dressing was removed and any remaining cream was wiped away with a tissue. The heel was then wrapped in a warm moist compress for 5 minutes before heel lance with an automated lancet (Microtainer, 1.4 mm puncture depth. 29 infants in the control group received 0.5 g Glaxal base and the same procedure as the experimental group was used. For phase 2, the method was the same as phase 1, 25 infants in the experimental group receive EMLA 0.5 g on an occlusive dressing for 60 minutes. 21 infants in the control group received 0.5 g placebo cream that was different to that used in phase 1. It contained all ingredients of EMLA except for the lidocaine and prilocaine, which were substituted with MCT oil (different to phase 1 placebo cream) and the creams remained intact for 60 minutes. Also, a larger automated lancet was used in phase 2 than that used in phase 1. This was due to a change in unit practice. A Microtainer, with a 2.2 mm puncture depth was used.
Outcomes	Pain: used the 0 to 21 PIPP score for infants of lesser gestational age at birth (i.e. < 28 weeks' gestation). The PIPP is a 7‐indicator measure, and gave a total range of scores of 0 to 21. Methaemoglobin levels. Adverse events.
Notes	Sample size based on previous research results comparing infants who received EMLA vs a placebo, sample size was calculated to show a 20% reduction (alpha = 0.05, beta = 0.20) in pain response scores. A total of 120 infants, or 30 infants per study group, was required.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Infants were randomly assigned using a computer‐generated table of random numbers.
Allocation concealment (selection bias)	Low risk	Creams were prepared by the pharmacy department, individually packaged the creams in dark glass ointment jars and labelling the jars by code. Those who coded the data were blinded to infant group and vague to the purpose of the study.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Phase 1 creams were not double blinded as the research nurse applying the creams may have been able to detect group allocation by the cream's colour and consistency. Phase 2 creams were identical in colour and consistency.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	In Phase 2, 14 infants were withdrawn after randomisation because they were discharged or transferred, leaving 106 infants available for analysis. Of the 106 infants, 31 were in group 1 (30 minutes' EMLA), 29 in group 2 (30 minutes' Glaxal base), 25 in group 3 (60 minutes' EMLA) and group 4 (60 minutes' placebo). The attrition appeared to be evenly distributed between the intervention and control groups but this is not reported by the authors.
Selective reporting (reporting bias)	Unclear risk	Protocol not obtained.
Other bias	Low risk

EMLA: eutectic mixture of local anaesthetics; NFCS: Neonatal Facial Coding System; NICU: neonatal intensive care unit; NIPS: Neonatal Infant Pain Scale; PIPP: Premature Infant Pain Profile; SD: standard deviation.