Milk-induced Hypoalgesia in Human Newborns

Subjects

A total of 72 infants born in the Tompkins Community Hospital, Ithaca, New York, were randomly assigned to nine experimental groups. All infants were delivered vaginally. Sixty-one were white, 8 Asian, and 3 black. Forty-two of the infants were female. Birth weights ranged between 2976 and 3697 kg, with a mean of 3321 kg. All infants had Apgar scores of 8 or above 1 minute after delivery, and 9 or above at 5 minutes. Body weights were equivalent among the experimental groups as were gender and racial distributions. Infants were between 22 and 40 hours old when their heels were lanced for the standard blood collection for phenylketonuria evaluation. Data were collected between 7 am and 8 am with the assistance of a single nurse to minimize variability. Infants had not eaten for 2.5 to 3.0 hours and were almost always asleep at the time of the experiment.

No mother had received meperidine or other morphine-like substances during delivery or during the postdelivery period if nursing. Fifty-eight of the 72 infants were breastfed and were essentially equally distributed among the nine groups. No infants were on special diets. An infant was enrolled in the study only after a parent, generally the mother, provided informed written consent; 68% of the parents approached did so. The study and consent forms were approved by the human subjects committees of both Cornell University and the Tompkins Community Hospital.

Eight infants served in each condition, namely, sucrose, water, protein, lactose, dilute fat, concentrated fat, fat/lactose, Similac, and RSF. Each infant was studied as shown by the time line in Fig1. Seven minutes before blood collection, the infant's foot was placed in a heated matrix, thereby facilitating circulation at the heel, the source of blood collection. Three minutes later, the infant drank exactly 2 mL of its designated solution through a sterile plastic syringe for 2 minutes. An additional 2 minutes elapsed after delivery, during which time the nurse prepared the surgical procedure and removed the matrix from the infant's foot. The procedure itself took place with the infant lying supine in his or her bassinet. For scoring purposes the procedure was considered to start when the nurse squeezed the infants heel immediately before heel-stick, and to end when a Band-Aid was applied to the wound and the foot released. Three to 6 minutes were necessary for blood collection. Mean collection times did not differ significantly among groups, nor did the number of heel-sticks. The entire session from solution administration to 3 minutes after Band-Aid application to the heel was recorded via videotape and later scored for crying. Neither the nurse, who was not present during fluid delivery, nor the person scoring the videotape were informed about experimental condition. Except for Similac, which is opaque, syringe contents were clear and, therefore, unknown to the experimenter. Fluids were delivered at body temperature and prepared as follows:

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Time line depicting the experimental protocol followed in this experiment, starting 7 minutes before blood collection, when a heated matrix was placed on the infant's foot, and terminating 3 minutes after the collection procedure.

Sucrose, 12%, was prepared on a weight/volume basis by adding sterile water to 6 g of sucrose until a volume of 50 mL was achieved.

Protein was prepared by adding 35 mL of water to 7.5 mL of protein (Provimin), a gift from Ross Laboratories, and vigorously agitating the mixture.

Lactose was of a 7% concentration, ie, the physiological range, and was prepared by adding water to 3.5 g of lactose until 50 mL was achieved. We did not use a more concentrated lactose solution. In our experience human newborns have found more concentrated solutions aversive, as judged by increased crying and grimacing.¹⁸

Dilute fat was composed of 1.85 mL of a coconut and soy oil blend mixed with 48.15 mL of water and agitated until homogeneity.

Concentrated fat was composed of 3.7 mL fat and 46.3 mL water. This concentration was apparently slightly aversive to the infants because 5 of the 8 infants in this group cried during the 2-minute interval between consuming the fat and the onset of the procedure. Of the remaining eight groups, only 1 infant cried in each of five groups, 3 each in the protein and Similac groups, and 2 infants cried after receiving the fat/lactose mixture.

Fat/lactose solution was made by combining 3.5 g lactose and 1.85 mL fat, adding water until 50 mL was achieved, and vigorously agitating the solution.

RSF was fabricated according to the percentage of each component that is found in human milk. Accordingly, the recipe of Ross Laboratories was followed, namely, adding water to 7.5 mL protein (Provimin), 3.24 g lactose, and 1.85 mL fat until 50 mL were reached and the entire mixture was agitated vigorously by sonication for 5 minutes.

Similac, a gift from Ross Laboratories, was also warmed to body temperature and presented to the infants as above.

Crying was declared to occur when a cry sound was detected. We did not distinguish among different crying intensities. In actuality crying was intense, almost always commencing when the heel was squeezed by the nurse in preparation for collection. Intensity increased when the heel was lanced.

Statistical Evaluations

Because the primary goal of these studies was to evaluate the effectiveness of milk and its individual components in reducing crying during and after blood collection, the sequence of planned comparisons was sucrose vs water, Similac vs water, and RSF vs water. Individual component comparisons were to be made against water if the milk or special formula evaluation was statistically reliable. Because of the relatively small number of subjects in each group and the lack of normal distribution of scores, the Mann-Whitney U statistic was chosen to determine statistical reliability during blood-harvesting. Also, because the collection procedure was of a somewhat variable duration, the percentage of time spent crying during the procedure was used for statistical evaluation. All probability values are presented after Bonferoni corrections for multiple tests.¹⁹

The effectiveness of the different substances in relieving crying during the recovery period was evaluated statistically via ttest for difference scores in crying between the first and third minutes after treatment. This is justified by the protracted effects of chemosensory stimulation in both rat^5,13,14 and human infants.^6,7,18 This comparison also acknowledges the possibility that substances could become effective against diminished stress due to the passsage of time since blood collection.