Validity of behavioral and physiologic parameters for acute pain assessment of term newborn infants Neonatal

Context: The subjectivity of pain causes enormous difficulties in evaluating neonatal pain with a single, practical and easy-to-apply tool. Pain evaluation in the neonatal period should be performed by valid, safe, useful and feasible methods. Objective: To evaluate the validity of the Neonatal Facial Coding System (NFCS), Neonatal Infant Pain Scale (NIPS), heart rate (HR) and O2 saturation (O2 sat) for neonatal pain assessment. Design: Prospective, double-blind randomized trial. Setting: A secondary level maternity hospital. Participants: 70 healthy neonates requiring bilirubin dosage were randomly assigned to receive a venous puncture (P: n=33, BW 3.2kg, SD 0.6; GA 39wk, SD 1; 59h of life, SD 25) or an alcohol swab friction (F: n=37; BW 3.1kg, SD 0.5; GA 39wk, SD 1; 52h of life, SD 17). Intervention: All measurements were taken prior to (PRE), during (T0), and 1(T1), 3(T3), 5(T5) and 10(T10) minutes after the procedure. Measurements: A neonatologist evaluated NFCS, NIPS, HR and O2 sat by pulse oxymetry. Results: Median NFCS and NIPS results at T0, T1 and T3 were higher in P group, compared to F. More P neonates presented NFCS >2 and/or NIPS >3 at T0, T1 and T3. HR was lower in P group at T1. Average O2 sat was above 90% during the whole study period in both groups. Conclusion: NFCS and NIPS are suitable instruments for neonatal pain evaluation. Heart rate and O2 saturation can be used only as auxiliary methods.

the pain.Moreover, most of the studies that relate pain to modifications in physiologic variables evaluate these changes after an acute and short pain stimulus, which is not the only or the main pain source for patients admitted to NICU. 8 Study of neonatal behavior seems a promising way of evaluating pain in pre-verbal patients, [9][10][11][12][13] and such behavior includes crying, motor activity, and the facial expression of pain.
The cry is considered to be the primary way of communication for newborn infants. 14,15The message of distress sent by the infant through the cry sensitizes the adult, either the child's mother or another adult who is taking care of it. 16,17Several studies have tried to relate pain to different characteristics of the cry. 18,19But the main problem in using the cry as a measure of pain is that approximately 50% of newborn infants do not cry during or after a painful procedure. 1,10,20Moreover, crying is not a specific pain assessment tool and can be elicited by other non-painful stimuli, such as discomfort and hunger. 21Therefore, the cry is useful in evaluating pain in the infant's environmental context associated with other pain assessment methods.

21,22
After a pinprick in one foot, the neonate withdraws the opposite leg in 0.3 seconds, the affected leg in 0.4 seconds, and he cries in 1.8 seconds. 23Term and preterm newborn infants have an organized repertoire of movements in response to a painful stimulus. 24,25Motor activity is a sensitive method of pain assessment and its specificity is enhanced by the concurrent use of other physiologic and behavioral pain evaluation tools.21   Several studies have indicated that the observation of facial expression seems to be a non-invasive, sensitive, specific and useful method of pain evaluation in term and preterm newborn infants. 9,10,24The facial activity of pre-verbal infants is expressive and can inform the observer about the emotional status of the neonate, beyond the pain, fulfilling the pain-subjective criteria of the International Association for the Study of Pain's definition. 28There are several ways of evaluating pain from facial features, but the most studied is the Neonatal Facial Activity Coding System (NFCS).NFCS evaluates eight parameters: brow bulge, squeezed eyes, nasolabial furrow, open lips, stretched mouth, lip purse, taut tongue, and chin quiver.Studies show that brow bulge, squeezed eyes, deepening of naso-labial furrow and open lips are present in more than 90% of the neonates exposed to a painful stimulus. 10Analysis of facial expression allows an effective communication between the neonates and the professionals that take care of them.
Multidimensional pain assessment is considered ideal because it gives information about individual responses to pain and their interaction with the environment. 8Among several published neonatal pain scales, the most studied are the Neonatal Infant Pain Scale -NIPS 12 and the Premature Infant Pain Profile -PIPP. 29The NIPS assesses five behavioral pain parameters (facial expression, cry, position of arms and legs and state of arousal) and one physiologic (breathing pattern).It is a neonatal adaptation of the Children's Hospital of Eastern Ontario Pain Scale -CHEOPS. 30Evaluations are performed at oneminute intervals prior to, during and after a painful procedure.NIPS seems to be a valid pain assessment method because it is based on known behavioral responses to pain, widely reported in the literature. 12The scale is able to differentiate term and preterm newborn infants that received a painful stimulus from those that had a distressing non-painful stimulus.Apgar scores at 1 and 5 minutes, post-natal age in hours, and minutes after the last feeding.
Neonates were randomly assigned to receive one of the following procedures: venous puncture in the back of the hand (P), considered as a painful stimulus; or alcohol swab friction on the back of the hand (F), considered as a distressing but not painful stimulus.A neonatologist blind to the procedure (P or F) evaluated physiologic and behavioral pain parameters.
After allocation, the patient was placed under a radiant warmer and a shield was positioned between the hand to be punctured or frictioned and the rest of the body, in order to obstruct the vision of the neonatologist in charge of pain assessment.At this time, a pulse oxymeter probe was located in the foot and asepsis of the hand was performed.Then, the neonates rested for five minutes prior to puncture or friction.Only one attempt at venous puncture was made, for all patients, and one milliliter of blood was collected, followed by a quick compression of the hand with dry gauze for hemostasis.F neonates received an alcohol swab friction in the back of the hand followed by a quick compression with dry gauze, simulating the hemostatic phase.After puncture or friction, the newborn infants were observed for 10 minutes.During this period, no pain relief attempts were performed.
The following pain parameters were evaluated: 1) Neonatal Facial Activity Coding System (NFCS) score, defined by the presence or absence of the following facial features: brow bulge, eye squeeze, naso-labial furrow, open lips, stretched mouth, lip purse, taut tongue and chin quiver.One point was given to each feature present (total score = 8).Pain was considered present when at least three features were observed.2) Neonatal Infant Pain Scale (NIPS) score, defined by the following parameters: facial expression (0/1 point), cry (0/1/2 points), breathing pattern (0/1 point), position of arms (0/1 point), position of legs (0/1 point), and state of arousal (0/1 point).The possible total score was seven, and pain was considered present when the score was greater than three.
12 3) Heart rate was evaluated by pulse oxymetry, and considered only when the quality of the wave registered was adequate.Bradycardia was defined by heart rate less than 80 bpm and tachycardia by heart rate above 160 bpm; 41 4) Oxygen saturation was also evaluated by pulse oximetry, taking into account the quality of the wave registered.Hypoxia was defined by oxygen saturation below 90%. 42 These physiologic and behavioral pain parameters were evaluated at six different times: immediately prior to the procedure (PRE), during hand puncture or friction (T0), one (T1), three (T3), five (T5) and ten (T10) minutes after the procedure.
Statistical Methods.Statistical analysis of data included the Student t test or Mann-Whitney test to compare quantitative variables between the P and F groups at each study time; chi-square or Fisher tests to compare qualitative variables between the P and F groups at each study time; and the Friedman test to compare the results obtained at the different study times for the P and F groups.The results were considered significant when p < 0.05.

RESULTS
The study population consisted of 70 newborn infants divided into two groups with the following characteristics (Table 1).The two groups were similar in relation to all these characteristics, except for time after last feeding (t test -P > F: p = 0.02).
Comparison of median NFCS scores (Table 2) revealed significant differences between the various study times for the P and F groups.P group median NFCS scores were higher than F during the procedure, one minute, and three minutes after the procedure.When presence of pain was considered as NFCS > 2 (Table 3), significantly more P patients showed signs of pain during the procedure, one minute, and three minutes afterwards.
Comparison of median NIPS scores (Table 2) revealed significant differences between the various study times for the P and F groups.P group median NIPS scores were higher than F during the procedure, one minute, and three minutes after the procedure.When presence of pain was considered as NIPS > 3 (Table 3), significantly more P patients showed signs of pain during the procedure, one minute, and three minutes afterwards.
Comparison of average heart rate values (Table 4) revealed significant differences between the various study times for the P and F groups.The heart rate of P neonates was lower than the heart rate of F patients only at one minute after the procedure.Three minutes after the procedure, 21% of P neonates presented bradycardia or tachycardia versus 3% of F neonates (Table 5).
Comparison of average oxygen arterial saturation values (Table 4) revealed significant differences between the various study times for the P and F groups.Oxygen saturation values were lower in the P group prior to the procedure, one minute, three minutes, and 5 minutes after the procedure.More P patients showed hypoxia (Table 5) one minute, and three minutes after the procedure.

DISCUSSION
The focus of this prospective study was to evaluate the use of physiologic and behavioral parameters as valid assessment tools for acute pain in healthy term newborn infants.In order to achieve this goal, the two study groups were comparable in terms of their main demographic characteristics.A statistical difference in the interval between feeding and observation was detected between the groups: the fact that the P group was last fed on average 33 minutes prior to the study and the F group 31 minutes does not, however, seem large enough to interfere with the clinical status of the neonates and the results here obtained.
The facial response of neonates to pain was well-documented in the present research.The Neonatal Facial Activity Coding System was able to differentiate term neonates who received the painful stimulus from those who had the non-painful stimulus.Not only were the median NFCS scores higher during puncture, compared to friction, but also there were significantly more P patients showing signs of pain during the procedure, and at one and three minutes afterwards.Therefore, evaluation of facial movements seems to be a valid and specific tool for acute pain assessment in term neonates and these findings are consistent with the literature.9,11,24,27 However, the modifications of facial mimic in response to pain were transient, and after three minutes they could not be seen anymore in their full expression.This finding raises difficulties for the clinical application of NFCS to pain evaluation at the bedside.New studies should clarify how often the facial movements should be monitored and for how long they should remain altered before analgesic therapeutic intervention is initiated or prior to analgesic dose adjustments.Further studies should analyze the frequency of each facial movement recorded during NFCS assessment, and verify whether there is any association between a set of facial movements, namely brow bulge, eye squeeze, deepening of naso-labial furrow and open lips, with the painful procedure, as  especially in premature infants.It would be interesting to compare the performance of both scales, the NIPS and the PIPP, in a neonatal intensive care unit context.[12]24,26 More studies should be done to evaluate these scales in the presence of various factors that could modulate the pain sensation in term and preterm newborn infants, for instance, the sleep/alert status, the neonatal clinical status and the patients' previous experience of pain, among others.
In both groups the heart rate values decreased during and one minute after the puncture or friction and increased afterwards.previously described. 10Furthermore, the adoption of a higher cutoff to define the presence of pain by NFCS could increase its sensibility and specificity for neonatal pain assessment.
Among the several multidimensional pain scales described when this study was started, the Neonatal Infant Pain Scale was chosen because it was a simple and easy-to-apply tool, with great potential for clinical use.NIPS was able to differentiate term neonates who received the painful stimulus from those who had a non-painful stimulus.Similarly to NFCS, not only were median NIPS scores higher during puncture, compared to friction, but also there were significantly more P patients showing NIPS greater than three during the procedure, and at one and three minutes afterwards.Therefore, this multidimensional pain scale seems to be a valid and specific tool for acute pain assessment in term neonates and these results are consistent with the literature. 11,12It should be noted again that after three minutes NIPS scores returned to baseline.Therefore the relationship between the duration of pain responses and the necessity for therapeutic interventions is not clear.Moreover, further studies should establish whether this cutoff (NIPS > 3 = pain) is the ideal one, in order to achieve the maximum sensitivity and specificity for the tool.In relation to its clinical use, the Neonatal Infant Pain Scale is an easier method than NFCS, and requires less personnel training.Also, NIPS scores the facial expression as a whole and values other behavioral parameters and one physiologic pain parameter, i.e.NIPS takes into account the notion that the best tool for evaluating pain is the one that considers its multiple dimensions. 8 The Premature Infant Pain Profile (PIPP) is another structured and validated multidimensional pain scale recently described. 29However, PIPP was not available when this study was completed.Perhaps the use of the Premature Infant Pain Profile could give more consistent information about pain, Only at one minute after the procedure was the average heart rate lower in P neonates compared to F infants.Three minutes after puncture more neonates had bradycardia or tachycardia, compared to the friction group.These findings show that the variation of heart rate is an inconsistent and insensitive way to evaluate pain in term newborn infants.Nor is the heart rate variation specific in relation to pain assessment, as occurs with other physiologic variables.It is important to note that in most infants the increase or decrease in heart rate during the puncture was not significant enough to alert the clinician to any need for analgesia or interruption of procedure.
In this study most neonates decreased their heart rate after the puncture instead of showing the usual response, i.e. tachycardia after a painful stimulus.No specific association between a distressing or painful stimulus on the back of the hand and bradycardia could be found.Possibly, technical artifacts interfered with the correct evaluation of heart rate.
Nevertheless, it appears important to repeat the study with heart rate evaluation performed simultaneously by pulse oxymetr y and electrocardiography, in order to understand better the heart rate response to painful stimulus in healthy term newborn infants.With the methodology here applied, heart rate assessment was not useful in detecting the neonatal pain.Since heart rate monitoring is routine in almost all neonatal intensive care units, care must be taken in associating its variations with the presence of pain.At best, together with other behavioral and multidimensional pain scales, and interpreted within the environmental context of the patient, the heart rate can only give a secondary support to the clinician who is evaluating the presence of pain in the neonate.
The oxygen saturation presented a more persistent decrease after the painful procedure, compared to the heart rate, and the decay was more pronounced in the neonates that received the puncture, compared to those that received the friction.Therefore O2 saturation values seem to be a valid tool for evaluating the newborn infants' pain.However, just as for any physiologic pain parameter, it has a low specificity, and it can be altered by other non-painful stimuli.
Although statistical differences were noted between P and F groups at 1 and 3 minutes after the procedure, the average O 2 saturation values were above 90% during the whole study period, and only a third of P neonates had hypoxia after the procedure.These findings mean that the clinical importance of P and F statistical differences is questionable, and that O2 saturation also has low sensitivity as a pain assessment tool.That is, it would be difficult to use O2 saturation values as the only indicator of pain in neonates.Although the monitoring of this physiologic parameter is generally available in neonatal intensive care units, it should be considered as an accessory tool for pain assessment.
In the present study the various pain parameters were analyzed separately, and their assessment of pain presence often did not agree in the same patient.This finding indicates that the use of multiple information can help to establish the presence of neonatal pain, and confirms the utility of composite scales for pain

Table 1 -Demographic data
a a: t test; b: chi-square test; SD = Standard Desviation

Table 5 -Number (percentage) of patients with HR < 80 bpm and/or HR > 160 bpm or O2 sat. < 90% at each study time Cut-off point
In conclusion, this study helps to validate the Neonatal Facial Activity Coding System and the Neonatal Infant Pain Scale as valid and suitable instruments for neonatal pain evaluation.Heart rate and O2 saturation have low sensitivity and specificity for evaluating neonatal pain, and can be used only as auxiliary methods.New studies should be performed in order to validate the NFCS and the NIPS in critically ill term and preterm patients.The availability of adequate pain assessment tools is critical for reducing the undertreatment of neonatal pain.