Our results suggest that venipuncture is a procedure that causes moderate pain in infants. A FLACC score increase of 4.84 (placebo group), post venipuncture falls into the rating of moderate pain as per the authors of the FLACC scale [17–19].
Currently the standard practice during venipuncture in young infants in the PED is not to administer any analgesia. Even though neonatal studies have previously demonstrated the effectiveness of sucrose and/or pacifiers, this practice has not been adopted in general in emergency departments as well as other pediatric departments . We hope that this study will demonstrate the ease of use of sucrose and/or pacifier and we hope that this will inspire practice change in this area.
Our choices of outcome measures were a result of a review of the literature. It should be noted that without direct verbal corroboration from the infants we cannot be entirely sure that any of the above outcome measures actually reflect degree of pain. Previous studies have relied on assessments of behavioral and physiological changes as indirect indicators of pain. We felt that the most comprehensive approach was to use a combination of a validated pain scale, total crying time and change in heart rate. The FLACC scale uses parameters similar to many of the neonatal pain scales, is highly reliable, has been validated, is very easy to use and teach and was best suited to the age group we wished to study. Although crying is associated with pain, it is not exclusive to pain, and thus must be interpreted with caution. In this study, none of the infants were crying prior to the procedure and all cried after it, so it is likely that the pain of this procedure induced this behavioral response. Thus we feel that in this study, crying time is a reasonable measure of pain or discomfort and have interpreted the results in such a light.
Our results show that pacifier appears to be an effective analgesic for the procedural pain of venipuncture in infants. Even though statistical significance was only narrowly missed for the primary outcome measure (p= 0.06), a change in average FLACC score from 4.3 (no pacifier) to 2.5 (pacifier) would be considered by most to be of clinical significance. Pacifier use significantly reduced crying time (statistically and clinically), particularly in the 0–3 month age group, despite small sample sizes of sub-group analysis. It is promising to see that this analgesic effect seems to extend beyond the neonatal period, perhaps up to three months of age. It appears that the effect wanes with age beyond three months. Further trials with larger sample sizes in this age group would be helpful to clarify this matter however.
One caution in the interpretation of results surrounding pacifier use was the fact that the observer was not blinded. This presents potential bias that was unavoidable for the primary outcome measure assessment, as it was necessary to look closely at the infants' faces to give a rating to this parameter on the FLACC scale. The addition of a second observer for the outcome measure of crying time would nearly have doubled the budget of our study and was thus impractical in our setting.
Also, heart rate measurements were assessed at the minute marks only and it is possible that these data points do not accurately represent interim variabilities. This may explain why differences in heart rate were not found. Another possibility is that heart rate monitoring may not be a reliable indicator of the amount of pain experienced. One adult study observed a decrease in heart rate in some patients, likely due to vagal stimulation, on insertion of an IV . Two adult observational studies have noted lack of correlation of heart rate with pain or changes in pain intensity [21, 22]. Review of neonatal studies reveals that heart rate data collection methodology is highly variable and there often does seem to be dissociation between pain scale findings and physiological responses such as heart rate [5, 23]. Pereira et al evaluated the validity of heart rate measurements for neonatal pain assessment in an RCT and concluded that heart rate variations are an inconsistent and insensitive way to evaluate pain in that population . Further clarification as to the reliability of this outcome measure as an indicator of pain across the pediatric spectrum may be warranted.
For sucrose as analgesia, the results are less clear. T-test results demonstrated no significant benefit; however age adjusted regression analysis showed significant reduction in crying time. Trends seem to show greater reduction in the younger age sub-groups. Sucrose appears to be less effective with increasing age at the dosage studied. Further study with larger sample sizes and perhaps using stronger concentrations of sucrose would be required to determine the upper age limit for the effectiveness of sucrose. It seems that sucrose and pacifier have an additive beneficial effect when used together and perhaps this is where the best use for sucrose as analgesia lies- to be used in conjunction with pacifier.
One must consider the dose of sucrose used. We chose 0.88 g (2 ml of a 44% solution) as this was easily prepared by our pharmacy, which uses an 88% sucrose solution to mix oral pediatric medications, and diluted this solution for the purposes of our study. Doses up to 0.5 g have been studied and determined to be safe for use in the neonatal period  and immunization studies have used doses as high as 2.5 g for older infants [14, 16] without adverse events. Future studies could look closely at the issue of optimal doses, especially with older infants.
There are several limitations to our study. One limitation to our study was that the study population was a convenience sample of patients and a few potentially eligible patients were not enrolled. The research nurses were available for 8–16 hours during the day so some children arriving overnight may have been missed. It is unlikely, however that these children would have been different from our study population.
Another limitation of this study was that despite accurate randomization, our randomization produced somewhat of an "unlucky sample" in that there were imbalances in some of the baseline statistics particularly age, NPO, and rate admitted. NPO was found in our adjusted analyses to not have an effect on our outcomes, while admission rates were not too unbalanced, and would be unlikely to have an effect on our final outcomes. Due to our determination in the adjusted analysis that, older children tend to experience more pain, the lower age in the pacifier/placebo group could lead to slight overestimation of pain relief in pacifiers and an underestimation of pain relief in the sucrose. These results may need to be interpreted with caution.
The intention of this study was to recruit infants between the ages of 0 and 6 months. Although infants across this entire age spectrum were recruited, numbers at the upper end of this range were less than had been desired, reflecting the visit and illness spectrum of this group and also chance (Tables 4 and 5). For the 84 infants recruited, the median age was 48 days, the mean age was 30 days and only 20 infants fell into the 3–6 month age range. Thus younger infants were represented strongly and older infants were underrepresented in this study. Therefore we could not draw valid conclusions about the effectiveness of our interventions on infants older than 3 months of age.
We also observed a higher standard deviation than we had originally anticipated. As a result, statistical significance was not achieved when examining age related effects although intriguing trends towards significance were seen which warrant further examination.