for the main comparison.

Sucrose (20% to 33%) compared with water for pain associated with heel lance
Patient or population: neonates with heel lance‐associated pain Settings: hospital Intervention: sucrose (20% to 33%) Comparison: water
Outcomes	Illustrative comparative risks (mean and range)		No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Water	Sucrose (20% to 33%)
PIPP at 30 s after heel lance Range of scale 0‐21 for infants < 28 weeks PMA and 0‐18 for infants > 36 weeks PMA. A lower score = less pain (Stevens 1996)	The mean for PIPP ranged across control groups from 8.5 to 9.62	The WMD for PIPP in the intervention groups was lower: ‐1.42 (95% CI ‐2.86 to 0.01)	105 (2)	⊕⊕⊝⊝ low	Bias: there were some concerns about bias in both studies (see RoB tables) Consistency: there was moderate inconsistency between the study point estimates (12 = 51 %) Precision: there was low precision in the point estimate with wide 95% CIs) Indirectness: all trials were conducted in the target population (no concern about indirectness)
PIPP at 60 s after heel lance Range of scale 0‐21 for infants < 28 weeks PMA and 0‐18 for infants >3 6 weeks PMA A lower score = less pain (Stevens 1996; Stevens 2014a)	The mean for PIPP in the control group was 8.59	The mean for PIPP in the intervention groups was lower: ‐1.80 (95% CI ‐3.81 to 0.21)	31 (1)	⊕⊕⊝⊝ low	Bias: there were concerns about allocation concealment and performance bias in this single study Consistency: N/A as there was only one study Precision: there was lack of precision due to small sample size Indirectness: the study was conducted in the target population
PIPP score during heel lance (1st heel lance) Range of scale 0‐21 for infants < 28 weeks PMA and 0‐18 for infants > 36 weeks PMA A lower score = less pain (Stevens 1996; Stevens 2014a)	The mean for PIPP in the control group was 7.3	The mean for PIPP in the intervention group was the same as in the control group: 0.00 (95% CI ‐1.52 to 1.52)	107 (1)	⊕⊕⊕⊝ moderate	Bias: there were no concerns about bias in this study Consistency: as there was only one study in this analysis concerns about consistency were N/A Precision: this was a relatively large single study (no lack of precision) Indirectness: the study was conducted in the target population
DAN score at 30 s after heel lance Range of scale 0‐10 A lower score = less pain (Carbajal 1997)	The mean DAN score in the control group was 9.5	The mean DAN score in the intervention group was lower: ‐1.90 (95% CI ‐8.58 to 4.78)	32 (1)	⊕⊕⊝⊝ low	Bias: concerns about blinding of performance and detection bias Consistency: as there was only one study in this analysis concerns about consistency were not N/A Precision: small sample size Indirectness: the study was conducted in the target population
NIPS during heel lance Range of scale 0‐7 A lower score = less pain Lawrence 1993	The mean NIPS score in the control group was 3	The mean NIPS score in the intervention group was lower: ‐2.00 (95% CI ‐2.42 to ‐1.58)	56 (1)	⊕⊕⊕⊝ moderate	Bias: no concerns about bias Consistency: as there was only one study in this analysis concerns about consistency were N/A Precision: small sample size Indirectness: the study was conducted in the target population (no concern about indirectness)
*The basis for the assumed risk was 'The mean PIPP, DAN and NIPS scores across control groups according to the values reported in the Assumed risk column. The corresponding risk was the mean in the intervention groups for the PIPP, DAN and NIPS scores with their 95% CI'. CI: confidence interval; DAN: Douleur Aiguë du Nouveau‐né Scale; PIPP: Premature Infant Pain Profile; PMA: postmenstrual age; N/A: not applicable; NIPS: Neonatal Infant Pain Scale; WMD: weighted mean difference
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.