A systematic review and meta-analyses of nonsucrose sweet solutions for pain relief in neonates

1School of Nursing, Federal University of Minas Gerais, Belo Horizonte, Brazil; 2Lawrence S Bloomberg Faculty of Nursing, University of Toronto, 3The Hospital for Sick Children, Toronto; 4Faculty of Health Sciences, University of Ottawa; 5Children’s Hospital of Eastern Ontario, Ottawa, Ontario; 6Critical Care and Neurosciences, Murdoch Children’s Research Institute, Royal Children’s Hospital, Victoria; 7Melbourne School of Health Sciences, The University of Melbourne, Melbourne, Australia; 8Department of Pediatrics, Obstetrics and Gynecology, Faculty of Medicine; 9Department of Health Policy, Management and Evaluation, Faculty of Medicine; 10University of Toronto Centre for the Study of Pain, University of Toronto, Toronto, Ontario Correspondence: Dr Mariana Bueno, Avenida Alfredo Balena, 190, 30130-100, Belo Horizonte, Brazil. Telephone 55-31-3409-9859, fax 55-31-3409-9860, e-mail maribueno@hotmail.com, buenom@enf.ufmg.br Sweet solution analgesia for neonates undergoing painful procedures has been extensively investigated over the past two decades. Sweet taste is believed to trigger the release of endogenous opioids (1-3). The analgesic efficacy of a solution may be dependent on its degree of sweetness, with the order from highest to lowest degree of sweetness being sucrose, fructose, glucose and lactose (4). Several guidelines recommend the use of sucrose for neonatal analgesia during single minor procedures (5-8). Sucrose is the most widely studied sweet solution, and its short-term safety and effectiveness for analgesia during minor procedural pain in neonates have been demonstrated in a systematic review and meta-analysis (9). It is, however, important to establish the effectiveness of alternative interventions for pain relief in neonates, such as sweet solutions other than sucrose. Alternative sweet solutions may be more accessible than sucrose in clinical settings. Randomized controlled trials (RCTs) have been conducted on sweet solutions other than sucrose for several years; however, the results of these trials have not yet been synthesized. The aim of the present study was to review the efficacy and safety of sweet-tasting solutions other than sucrose during acute procedural pain in neonates. review


MetHoDs search strategy
Electronic searches were conducted in MEDLINE, Embase, and PsycINFO on Ovid, and in the Cumulative Index to Nursing and Allied Health Literature on EBSCO, from the start date of each database to July 31, 2011.Search terms included: pain*, infant*, neonat*, newborn*, lactose, glucose, fructose, glycerine, dextrose, aspartame, polycose, saccharose, saccharide and alternative names for each type of sugar.Additional studies were identified from reference lists, manual searches of key literature, key informants and scientific meeting proceedings.No restrictions on language, year of publication or publication status were imposed.

study selection
RCTs studying term and preterm neonates (from birth to one month of age) undergoing any painful procedures and receiving nonsucrose sweet solutions as analgesia were considered for inclusion.Interventions in the control and/or comparison groups included no intervention, oral administration of water, sucrose, breast milk or formula, and breastfeeding.Studies reporting the use of pacifier, positioning, swaddling, cuddling, skin to skin contact and topical anaesthetic agents in combination with or in comparison with sweet solutions were included.
Two independent reviewers (MB and DH) screened abstracts to identify eligible studies.Full-text articles of potentially relevant abstracts were retrieved and independently assessed for inclusion.Discrepancies were resolved through a consensus process between the reviewers or with a third reviewer (JY) if necessary.

Assessment of bias risk
Risk of bias of RCTs were assessed using the Cochrane Collaboration criteria (10), which included the following: adequate randomization sequence generation; adequate allocation concealment; blinding of intervention; incomplete outcome data addressed; free of selective outcome reporting; free of other potential sources of bias.Possible responses were low risk, high risk and unclear risk.Two reviewers (MB and SK) independently assessed each study for risk of bias; differences were resolved by discussion.A third reviewer (JY) was consulted if discrepancies remained.

Data extraction
Data extracted from each study included age, number of infants allocated, comparison groups including type of painful procedure, type of sweet solution(s), concentration and volume of solution(s) administered, study outcomes and adverse events.Where possible, authors were contacted for missing information/data or data clarification.Two reviewers (MB and SK) extracted data from the studies using a standardized, pretested form.Differences were resolved by discussion or by a third reviewer (JY) when necessary.

Data synthesis
The primary outcome of interest was pain scores assessed by a validated unidimensional, multidimensional and/or composite pain measurement tool.Physiological pain indicators (ie, heart rate [HR], HR variability, respiratory rate [RR], oxygen saturation [O 2 Sat], skin conductance), behavioural pain indicators (eg, crying features, facial or body actions) and adverse events were reported.
When data were available from at least two studies using the same intervention and outcome measure, a meta-analysis was performed using a fixed-effects model.A mean difference (MD) was reported for continuous variables and typical relative risk for dichotomous data with 95% CIs.Risk difference (RD) was calculated, and number needed to treat to benefit or harm were calculated if the RD was significant.Between-study heterogeneity was determined using the I 2 statistic.RevMan 5.1 software (Cochrane Collaboration, Denmark) was used to analyze the data (11).

study organization
Reporting for the present systematic review was conducted in accordance with the recommendations of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement (12).

included studies
Of 1998 identified citations, 901 were unique studies.Forty-two studies met the inclusion criteria.On further evaluation of full-text articles, it was unclear whether four studies (13)(14)(15)(16) were RCTs.Authors were contacted but did not respond before the completion of the present review; therefore, these studies are currently awaiting classification.Thirty-eight studies (enrolling 3785 infants) were included (Figure 1).
Solutions were administered to the anterior portion of the tongue using a syringe in the majority of the trials.Akcam and Ormeci (21) compared the use of spray and syringe to administer 30% glucose to infants.Non-nutritive sucking was offered in combination with the sweet solution in five trials (22)(23)(24)(25)(26).In four studies, neonates were stimulated to suck the syringe during the administration of the solutions (27)(28)(29)(30).Gradin et al (31) described the use of a pacifier or a finger for providing sucking after offering the solution as optional.Finally, Kass and Holman (32), Sajedi et al (33) and Gharehbaghi and Ali (34) did not provide sufficient information regarding the methods used to administer sweet solutions before the procedure.
Control and/or comparison groups received water (with or without sucking), pacifier, swaddling, skin to skin contact, sensorial saturation, facilitated tucking, sucrose solution (with or without sucking), breastfeeding, expressed breast milk, 2.5% lidocaine/2.5% prilocaine cream (EMLA; AstraZeneca, United Kingdom), dorsal penile nerve block, acetaminophen, oxycodone or inhaled sevoflurane.No intervention groups were included in eight trials and water groups were included in 19 trials.In six trials, both no intervention and water groups were evaluated (Table 1).
Painful procedures investigated were heel lance (19 studies), venipuncture (10 studies), intramuscular injection (three studies), subcutaneous injection (one study), peripherally inserted central catheter (PICC) placement (one study), eye examination for retinopathy of prematurity (one study) and circumcision (one study).Two studies analyzed pain during different procedures: heel lance and venipuncture (one study), and heel lance and pharyngeal suctioning (one study).

Risk of bias
Six trials were considered to have a low risk of bias because authors provided sufficient information to make a judgement (18,21,24,31,35,36).
In fewer trials, a risk for selective outcome reporting was identified (17,19,27,29,32,45).There was not enough information on incomplete outcome data in one trial (19) and risk of other sources of bias in four studies (19,26,46,53).
Results from two of the six RCTs that used the PIPP to assess pain during heel lancing were included in a meta-analysis.A 3.6-point reduction in PIPP scores during and/or after heel lances was observed for 20% to 30% glucose (1 mL to 2 mL) compared with no intervention (49,50) (124 neonates; MD −3.61 [95% CI −4.58 to −2.63]; P<0.001;I 2 =54%) (Figure 2).
Two trials used the NIPS and the PIPP to assess the same heel lance (47,51).Bonetto et al (47) reported no differences in PIPP scores, while NIPS scores significantly favoured 25% glucose (1 mL) versus 2.5% lidocaine/2.5% prilocaine cream, acetaminophen or water.Axelin et al (51) reported significantly lower PIPP scores for neonates receiving 24% glucose (0.2 mL) versus water (0.2 mL) but reported no differences in NIPS scores.In the one trial that compared 20% glucose and sucrose (2 mL) solutions (36), differences in NFCS scores were not significant.These results suggest similar effectiveness of both glucose and sucrose.Physiological outcomes: HR was assessed at various intervals of time in 11 studies.Three trials reported significantly lower increases in HR for infants receiving 10% to 50% glucose (1 mL to 2 mL) compared with water or no intervention during and after heel lance (36,40,48).Ramenghi et al (20) reported significantly lower increases in HR for maltitol compared with the water group.
A significant increase in HR was reported by Eriksson et al (49) among infants receiving 30% glucose versus no intervention and by Guala et al (28) when comparing 5% glucose with 30% glucose and water.
No significant differences were observed between glucose and control groups regarding O 2 Sat and RR (36,43), with the exception of one study that reported a lower O 2 Sat reduction from baseline for infants receiving 10% to 50% glucose compared with water (40).Okan et al (36) described no differences in O 2 Sat and RR for 20% glucose and sucrose groups.Cry behaviour: Glucose (1 mL to 2 mL of 10% to 50% solution) significantly decreased cry duration when compared with water or no intervention in four trials (30,40,48,49); however, no significant differences between glucose and no intervention or water groups were reported in three trials (36,38,43).
A longer duration of crying within 3 min after heel lance was described for neonates receiving 10% or 30% glucose compared with 30% sucrose (38), although no significant differences in crying characteristics were observed by Okan et al (36) between 20% glucose and sucrose groups.
All 11 studies in which 25% to 50% glucose (1 mL to 2 mL) was administered demonstrated significant reduction in pain scores compared with 10% glucose, water, no intervention or 2.5% lidocaine/2.5% prilocaine cream.However, Liu et al ( 60) described lower NIPS scores for infants receiving sucking compared with 25% glucose.
Carbajal et al (22) described similar DAN scores for 2 mL of 30% glucose and sucrose groups after venipuncture.Physiological outcomes: HR was assessed in five trials involving venipuncture.No significant differences in HR responses were observed in infants receiving 10% to 30% glucose compared with water (34,35,39) or 2.5% lidocaine/2.5% prilocaine cream (31).An increased HR in infants receiving 30% glucose was reported in one trial (49).
Incidence of cry after venipuncture was reduced among infants receiving 25% to 30% glucose (1 mL to 2 mL) versus water after combining results of three studies (35,39,42) (relative risk 0.80 [95% CI 0.66 to 0.96], P=0.02 and RD −0.18 [95% CI −0.31 to −0.05]; P=0.008, number needed to treat = 6 [95% CI 3 to 20]; I 2 =63%) (Figure 3).intramuscular injections: Significantly lower NIPS scores were reported in one study for neonates receiving 30% glucose (2 mL) compared with water at 3 min after injection (33), but Chermont et al (52) reported no differences in pain scores (NFCS, NIPS or PIPP) for 25% glucose versus no intervention.There were no significant differences for changes in HR among 30% to 50% glucose, water and no intervention groups (33,45).subcutaneous injections: Carbajal et al (23) described significantly lower DAN scores for infants who received 30% glucose (0.3 mL) compared with water during subcutaneous injections.Picc placement: A single trial investigated the effects of 0.2 mL of 30% glucose and sucking compared with sevoflurane during PICC placement (26).All infants included in the trial required assisted ventilation and received 2.5% lidocaine/2.5% prilocaine cream before puncture.Results indicate that the glucose and sucking group presented higher incidence of bradycardia, tachycardia and hypertension than the sevoflurane group.Body movements were increased for the glucose group.eye examination for retinopathy of prematurity: No significant differences were reported in PIPP scores, cry, HR and O 2 Sat between infants who received 30% glucose or water (1 mL) before eye examination for retinopathy of prematurity (53).circumcision: No differences in Modified Behavioural Pain Scale (MBPS) (62) scores, HR, O 2 Sat and cry were observed for infants receiving 50% glucose and water before circumcision (32).Figure 3) Incidence of crying after venipuncture for infants receiving 25% to 30% glucose (1 mL to 2 mL) compared with water Pharyngeal suctioning: Axelin et al (51) reported significantly lower PIPP scores for neonates receiving 24% glucose (0.2 mL) versus water, but no significant differences for NIPS scores.

Adverse events
Thirteen trials recorded possible adverse events related to the solutions (18,21,23,26,28,31,34,37,39,42,46,51,52).Carbajal et al (23) reported one death from necrotizing enterocolitis in an infant who received water and seven episodes of oxygen desaturation (five in neonates receiving 0.3 mL of 30% glucose and two in neonates receiving glucose plus pacifier).No adverse events were reported for infants receiving water.Axelin et al (51) reported short-term desaturation and/or bradycardia episodes in 21% of infants receiving 0.2 mL of 24% glucose, and in 12.5% of the infants receiving water.No adverse events related to the administration of sweet solutions were recorded in the other 11 studies.

Discussion
The increasing interest in exploring the analgesic properties of orally administered glucose for pain relief over other sugars may be explained by its availability in clinical settings for intravenous use (22,32,38,40,42,48).Manufactured sucrose may not always be readily available or institutions may not be equipped with pharmacies to prepare sucrose solutions; therefore, the efficacy of alternative sweet solutions, such as glucose, as analgesic strategies for painful procedures in neonates needs to be determined.
To our knowledge, the present systematic review and meta-analyses are the first to evaluate the efficacy of nonsucrose sweet-tasting solutions during painful procedures in neonates.The large number of trials included (38 RCTs; 3785 neonates) and the results of the meta-analyses contribute to strengthen the present review.A recently published systematic review by Kassab et al (63) investigated the analgesic effects of glucose during needle-related pain in infants younger than 12 months of age.The authors included 18 studies and no meta-analyses were conducted.Differences may be explained by search strategies, assessment of study quality adopted and by the age of included infants because Kassab et al (63) excluded studies involving preterm infants.
The assessment of risk of bias is dependent on the authors' report of methodological details of their study.Efforts were made to contact the authors and, in some cases, additional information on randomization processes, allocation concealment techniques, blinding and losses, among others, was provided.In addition, methodological discrepancies were observed across the included studies, especially regarding the variety of volumes and concentration of glucose solutions and the outcome measures.
The results of the present systematic review and meta-analyses indicate lower pain scores for neonates receiving glucose compared with water or no intervention for heel lances and venipuncture.These findings are similar to those reported by Stevens et al (9) in a systematic review and meta-analysis on the effectiveness of sucrose solution for pain relief in neonates, emphasizing that glucose is an effective alternative to sucrose.
The results of single trials demonstrated the analgesic effects of glucose during other minor skin-breaking procedures (subcutaneous injections [23], intramuscular injections [33] and pharyngeal suctioning [51]).Additional studies are warranted to confirm the efficacy of glucose during these procedures.Conversely, glucose did not provide adequate analgesia during circumcision (32) and eye examination (53).Procedures of longer duration may require additional pharmacological and nonpharmacological interventions to be effective (9) and further investigation is needed.
Pain scores were the most commonly reported outcome measure.A meta-analysis including 124 infants demonstrated that 1 mL to 2 mL of a 20% to 30% glucose solution compared with no intervention effectively reduced PIPP scores during the first 30 s after heel lancing.Eriksson et al (49) included healthy term infants, whereas Freire et al (50) enrolled healthy preterm neonates.Differences among the facial expression of term and preterm infants undergoing painful procedures are well described in the literature (58,64,65).These differences may have been minimized between these trials due to the use of the PIPP to assess pain.The PIPP is a validated, composite multidimensional pain score that considers gestational age as a modifying factor on pain expression in infants.Outcome measurement procedures and procedure technique may have contributed to high between-study heterogeneity.These findings are consistent with the meta-analyses reported in a systematic review on the effectiveness of sucrose on PIPP scores of infants undergoing heel lancing (9).
Although PIPP scores were considered the primary outcome in four other studies (27,44,47,51), differences in the intervention, especially volume and concentration of glucose solution, and comparison interventions used, precluded their inclusion in meta-analysis.
Additional meta-analyses showed significant mean reductions in crying behaviours during and after heel lancing (reduction of 36.4 s) or venipuncture (reduction of 14.6 s) for neonates receiving 25% to 50% glucose solutions compared with water or no intervention.Incidence of crying after venipuncture was reduced by 20% for infants receiving 25% to 30% glucose after venipuncture compared with water.Methodological discrepancies may contribute to between-study heterogeneity, especially regarding definitions of crying behaviour, outcome measurement procedures and procedural technique.
A statistically significant reduction in the total duration of crying was reported when comparing sucrose solution with water for infants undergoing heel lances, although no differences were observed for duration of first cry when comparing neonates receiving sucrose and water during lancing (9).Crying behaviour was significantly reduced by the administration of glucose or sucrose before immunization in infants between one and 12 months of age (66).
A relatively small number of infants were included in each metaanalysis, and moderate to high between-study heterogeneity can be observed in some cases.However, these results can be considered to be clinically significant in favour of glucose for heel lances and venipunctures.
Physiological indicators were not significantly affected by the administration of 25% to 50% glucose, which is similar to the results of sucrose administration as reported in a systematic review (9).Although physiological indexes are important indicators of neonatal pain, lack of specificity may explain these results.
In one trial, an increase in HR was observed for infants receiving 30% glucose compared with no intervention (49).The effects of 30% glucose on increasing HR were later confirmed by Gradin (67) in a trial that involved healthy term infants with no painful intervention.A possible explanation provided by the author is an increased awareness caused by the sweet taste as well as an increase in sympaticus activation.
The administration of sweet solutions in combination with nonnutritive sucking were commonly reported.The analgesic and calming effects of sucking are mainly explained by the stimulation of orogustatory and mechanoreceptors due to the sucking action (68).However, it remains unclear whether there are additive or synergistic effects when sweet solutions and sucking are combined (69).Overall, most included studies identified positive effects of combining both interventions for term and preterm infants undergoing heel lances and venipuncture.
Glucose combined with sensorial saturation was described as more effective than glucose and sucking for analgesia during heel lances.However, sensorial saturation without glucose was not effective and the authors reported that sensorial saturation caused increased awareness and irritation in the infants (29).Further investigation regarding the efficacy and safety of sensorial saturation is required before clinical practice recommendations can be made.Breastfeeding (41) and skin to skin contact (52) combined with glucose were reported to be more effective as pain-reducing interventions than glucose alone.The effects of combining these interventions with sweet solutions should be further investigated.
Several trials included placebo and/or no intervention groups.In some cases, both water and no intervention were compared with an experimental intervention (22,24,28,37,43,45).The results of the present review, combined with results reported by Stevens et al (9) and Harrison et al (66), provide strong and consistent evidence for the analgesic efficacy of sweet-tasting solutions during single procedures in healthy term and preterm infants.Therefore, researchers should consider including sweet-tasting solution as the control intervention in future studies involving this population.
Short-term effects of glucose administration (bradycardia and desaturation) were reported in two studies that enrolled very preterm neonates.There were no differences in adverse event rates between the glucose and water groups, and these events were explained as being more related to the neonate's ability to swallow rather than the type of solution (51).
Overall, results indicate no significant differences in the effects of glucose and sucrose solutions, although the limited number of studies comparing the two solutions precluded meta-analysis.Further research is required to establish the comparative efficacy of these solutions.
With the exception of Michel et al (26), studies included in the present review did not investigate the effects of nonsucrose sweet solutions in critically ill infants, although this population undergoes multiple painful procedures during prolonged hospitalizations (70,71).No studies measured the effects of repeated doses of glucose for procedural pain.Further investigation is needed to establish the efficacy and safety of nonsucrose solutions in both situations.

conclusion
A substantial variability in the volumes and concentrations of glucose solutions administered and the large variety in outcome measurements precluded further meta-analysis.However, the large number of included studies contributes to the strength of the present systematic review.
Overall, results indicate that 20% to 30% glucose solutions have analgesic effects on both term and preterm neonates undergoing a single heel lance and venipuncture and can, therefore, be recommended as an alternative to sucrose for procedural pain reduction in this population.

Figure 1 )
Figure 1) Study selection process for systematic review.RCT Randomized controlled trial

Figure 2 )
Figure 2) Mean Premature Infant Pain Profile scores after heel lancing for infants receiving 20% to 30% glucose (1 mL to 2 mL) compared with no intervention

FunDing:
Canadian Institutes of Health Research (CIHR) Systematic Review of Sweet Solutions for Acute Pain Relief in Infants Knowledge Synthesis Team (KRS91774).CIHR Strategic Training in Health Research Pain in Child Health (PICH) Initiative (STP53885).CIHR Team Grant in Children's Pain (CTP-79854 and MOP-86605)

TaBle 1 Characteristics of included studies author (reference), year; n Comparison groups Outcome measures and main results Heel lance
SS and glucose more effective than glucose and sucking at reducing DAN score (P=0.004).No significant difference in HR increase among all groups.HR recovered more quickly in the sweetener group compared with other groups (P=0.04).Reduced crying time (P=0.04)for sweetener group and longer crying time for glycine group (P=0.01)compared with other solutions.Quicker recovery (P=0.01)for sweetener group compared with other solutions.
Lower DAN scores for glucose and sucking; SS, glucose and sucking (P<0.0001 for both); and water and sucking (P=0.001) compared with no intervention.Lower PIPP scores for glucose alone (P<0.01),sucking(P<0.01),glucose and sucking (P<0.001), and SS (P< 0.0001) compared with no intervention.SS and glucose more effective than glucose and sucking (P<0.01) at reducing PIPP scores.No significant differences in PIPP scores between glucose and breastfeeding groups.Breastfeeding more effective than SS and swaddling (P<0.001 for both) at reducing PIPP scores.No significant differences in IBCS and NFCS scores among the groups.Lower combined score (IBCS and NFCS) for the sweetener group (P=0.036).Continued on next page

TaBle 1 -CONTINUeD Characteristics of included studies author (reference), year; n Comparison groups Outcome measures and main results
No significant differences in HR values among the groups.No significant differences in crying time for 10% and 30% glucose groups compared with water.Sucrose superior to 10% glucose (P=0.003) and 30% glucose (P=0.006) for crying time.25% and 50%) more effectively reduced grimacing, crying time, HR increase and O 2 Sat decrease than 10% glucose and EBM (P<0.05 for all, except for grimacing, for which P value is not provided).No significant differences between 25% and 50% glucose groups, or between 10% glucose and EBM groups for all outcomes.Lower HR values in glucose and sucrose groups compared with water at 1 min postlancing (P=0.007).No significant differences between the groups for crying time, respiration rate and O 2 Sat .
Lower PIPP scores for 2 mL glucose group compared with water (P=0.01).Reduced crying time and incidence of cry for 2 mL glucose group compared with water (P<0.05).No significant differences in HR and oxygen consumption between groups.

TaBle 1 -CONTINUeD Characteristics of included studies author (reference), year; n Comparison groups Outcome measures and main results Heel lance and venipuncture
Lower PIPP scores during heel lance (P<0.0001) and venipuncture (P<0.0241) for glucose group compared with water.Higher increase in HR for glucose group compared with water, for both procedures (P=0.0201).Reduced crying time for glucose group versus water for heel lance (P<0.0001)but not venipuncture.