A Comparative
Study
of the Behavior
of Greek
Neonates
T. Berry Brazelton, M.D., Yolanda Tryphonopoulou, M.D., and Barry M. Lester, Ph.D.
Froum time Harvard Unicersity .%fedical Seimool an(l Cimildren s Hospital .fedical Center, Boston, 011(1 tIme
institute of Cimild Health, Athens, Greece
ABSTRACT. Three groups of 30 Greek neonates each (an adoptive group from the Metera orphanage, a lower-class group, and a middle-class group) were evaluated at days 1, 5, and 10 after birth using a behavioral scale of 21 items and a neurologic evaluation of 16 items. Behaviors were examined for group differences and group-by-age recovery curves were determined during the first ten days. Significant differences were found in the separate items and items grouped to reflect interactive, motor, and state behavioral dimensions. The adoptive 1)abies at the Metera orphanage generally performed the most poorly. This difference seems not only to reflect intrauterine differences, especially in regard to
nutri-tion, 1)ut to point to the likelihood of eliciting less than optimal responses from future environments. The middle-class group had the worst scores on physiologic items and were similar to the Metera babies in having initially depressed interactive, motor, and state behavior. Improve-ment in these areas over ten days suggested that temporary effects of maternal medication caused the poor scores. The
recovery curves of the infants pointed to the important effects of such l)erinatal variables as maternal medication on earl neonatal l)ehaviOr. Pediatrics 63:279-285, 1979,iieona-t#{252}lbelmarior, lmc’on(ztal behacioral assessment scale, socioeco-7lOHl U (llff()r();l(.(..s Greek ;zeonat(’.s.
Neonatal behavior in the human infant results from interaction between the genotypic behavior with which the infant is endowed and the shaping of it by intrauterine and perinatal experience. Recent animal and human research points to some of the factors that affect genotypic behavior before birth. Undernutrition in the mother,’4 closely repeated pregnancies,3 intrauterine infec-tion,2 and drugs given to the mother at critical periods of fetal development6 are some of the factors, and there probably are many others.
The Brazelton Neonatal Behavioral Assessment Scale has been used to reflect neonatal behavioral
differences related to full-term but underweight infants,7TM poor nutrition,9 obstetric risk factors,” and maternal obstetric medication.” #{149}‘ 2 This scale
tests behavioral and neurologic adequacy in the immediate extrauterine period before the neonate is shaped by his new environment. It is assimed that the infant’s behavioral responses as well as
his physical condition reflect the powerful influences of intrauterine and perinatal experi-ences as they interact with genetic endowment. These influences must be seen as partially respon-sible for individual and group differences at birth. In studies where differences in neonatal behavior are observed in groups of babies within a culture,”’4 the influence of these variables on genetic endowment can be accounted for more easily.
In this study we evaluated a group of 30 Greek neonates in an Athens orphanage, the Metera, on days 1, 5, and 10 after birth. We were interested in this group because the early placement of infants in adoptive homes demands assessment of the integrity of the CNS and the capacity to adjust in a nurturing environment.
Two other groups of 30 Greek neonates each were studied to contrast the behavior of the Metera infants with infants from the same culture (and possibly similar genetic endowment) who varied in obstetric history and socioeconomic status (SES). A group of babies whose mothers’ SES (lower class) was comparable to that of the Metera mothers and a middle-class group formed the comparison groups. We used the 21 behavior-al items from the scale to document group differ-ences in behavior in the neonatal period, to estimate the effects of different intrauterine expe-rience, and to determine the effects of different environments on these behaviors during the first ten days after birth.
PROCEDURE Subjects
The three groups of Greek infants were
evalu-ated by two of us (T.B.B. and Y.T.) on days 1, 5,
Received Febniary 14; revision accepted for publication June 2, 1978.
ADDRESS FOR REPRINTS (T.B.B.) 23 Hawthorn Street,
and 10 after birth. The testing was done in Athens and the babies were selected to fit recognized SES criteria for the middle- and lower-class groups.’3 Although the tester knew the group membership of the infants, which may have introduced bias, she was not aware of the hypotheses of the study. The data analysis was performed without knowl-edge of the group membership of the infants. Criteria for acceptance in each group were (1) the mother had a healthy, uneventful pregnancy; (2) the labor was classified as normal; (3) no congen-ital abnormalities were noted at birth; (4) Apgar scores were 7, 8, and 8 or more at 1, 5, and 15 minutes, respectively; (5) anesthesia was local rather than general (eg, spinal); (6) medication was routine and could be weighted according to criteria listed under “Method”; (7) the infant appeared by clinical evaluation to be mature and the birth weight was greater than 3.0 kg; (8) no imusual jaundice or neurologic signs developed in the neonatal period; and (9) babies were uniform-ly fed in each group; ie, all were bottle-fed except those in the lower-class Greek group who were all breast-fed.
The Metera group was composed of 30 babies whose mothers were cared for in late pregnancy by this Athens orphanage and who decided to leave their infants at the Metera for adoption. Women are admitted to the Metera in the seventh month of pregnancy and from then on they are well fed and cared for. Case histories indicated that many of them had attempted to abort themselves early in pregnancy, and all of them reported that they had tried to starve themselves in the first few months of pregnancy in an effort to hide their condition. As a result, intrauterine conditions changed markedly for the fetus during the seventh month when the mother was admitted to the Metera and received appro-priate prenatal care and nutrition. No anesthesia was administered to the mothers at the time of delivery, and little or no premedication was given
(
two mothers had one dose of promazine hydro-chloride, Sparine, of 25 mg each). The babies were seen in the maternity hospital on day 1 and at the orphanage on days 5 and 10.The second group consisted of 30 babies from middle-class homes. The fathers were all at home and working and earned an annual income of more than $3,000 (United States). The mothers had eaten well during pregnancy and there was no reported attempt to abort any of the infants. The average gravida index was 2.0. In accordance with the practice of private Greek obstetrics, these mothers had premedication (average 2.1 by weighing, see “Method”). Spinal or local
anesthe-sia was administered routinely. There were no complications in labor and labor was not prolonged for any of these mothers. The infants were seen in the maternity hospital on day 1 and at home on days 5 and 10.
The third group consisted of 30 infants of a lower SES group of laborers’ families. They, too, were from intact families. The mothers were gravida 1 except for six who were gravida 2. They all wanted their babies and had made no attempt to abort them. No medication or anesthesia was administered at the time of delivery. Nutritional histories indicated that malnutrition had not been a problem for these families. Infants were evalu-ated in the hospital on day 1 and at home on days 5 and 10.
Method
The three groups were evaluated with an earlier version of the Brazelton Neonatal Behav-ioral Assessment Scale” than that presently in publication.’7 The scale consisted of 21 behavioral items to assess behavioral adequacy at birth and 16 reflex items to assess neurologic adequacy. The latter were drawn from and scored according to the method of Prechtl and Beintema.’5 Interscor-er reliability defined according to presently accepted criteria’2 for three consecutive tests was .90,
.94,
and .91, respectively. For the 21 behav-ioral items this represents the ratio of agreements to agreements plus disagreements; a disagreement is defined as a discrepancy of more than one point on a nine-point scale for each item.Maternal medication given in the one half to eight hours before delivery was weighted as follows: (0) no premedication; (1) alphaprodine hydrochloride (Nisentil), not more than 30 mg; promazine hydrochloride (Sparine), not more than 25 mg; meperidine hydrochloride (Demerol), not more than 50 mg, and only two of these given in the eight hours before delivery; and (3) a combination of three premedicant drugs in doses in excess of 60 mg of alphaprodine, 50 mg of promazine, and 100 mg of meperidine in the eight-hour period before delivery. Table I shows the average drug scores and other relevant obstet-nc history information for the three groups.
Scale
Table I
OBsmrRIc HISTORIES OF THREE GROUPS OF GREEK INFANTS
Obstetric Iiistonj Metera Middle Class Lower Class
Gravida (average) 1.1 2.0 1.2
Average birth weight (kg) 3.2 3.6 3.3
Apgar scores (averages at 1, 5, 15 mm) 8.6,9,9.5 9,9,9.5 9,9.5,9.5
Labor complication 0.0 0.0 0.0
Anesthesia (spinal or epidural) 0.0 Yes 0.0
Niedication (weighted average) 0.6 2.1 0.0
responses, but more importantly, for the infant’s al)ility to respond to stimuli that represent the kinds of stiniulation that caregivers present in a fostering situation.
The 21 scale items are organized into four clusters that represent the conceptual dimensions the scale was designed to reflect (Table II). The first dimension, interactive processes, assesses the capacity to respond to social stimuli, especially during alert states. The second dimension, motor processes, includes the ability to maintain adequate tone, control motor behavior, and perform integrated motor acts. Dimension 3, regulation of state, assesses the capacity to organ-ize and modulate states of consciousness, to respond in alert states, and to shut out disturbing stimuli during sleep. Dimension 4, physiologic organization, looks at the baby’s stability in response to stress as reflected in tremors, startles, and changes in skin color.
The 16 reflex items were scored on a three-point scale in an effort to rule in or out neurologic integrity. The reflexes were scored as follows: 1, low; 2, medium (containing 80% of the “normal” expected responses); and 3, hypertonic. More than three deviant scores (one or three) indicates the need for a complete neurologic examination. The reflexes assessed were plantar foot grasp, hand grasp, ankle clonus, Babinski’s reflex, stand-ing, walking, placing, incurvation, crawling, glabella, vestibiilar responses on rotation, ocular nystagmus, tonic neck, Moro’s reflex, sucking, and rooting responses. Since the reflexes were scored to establish normality, no attempt was made to grade them on a continuum that might establish individual or group differences. The 16 reflex items reflected average, normal performances in all three groups, and there were no differences by group or by age. Because there were no differ-ences in the reflex items among the three groups, they have not been included in this article.
Of the 21 behavior items, 13 are scored along scales that are positive and linear, ie, higher scores indicate better performance. For the other
nine items, the scales are either curvilinear, with midrange performance considered to be optimal, or linear, with higher scores representing poorer performance. Since our interest in this study was to compare mean differences on each item between the three groups, we rescored these nine items to represent positive, linear scales. Thus, for all 21 items, higher scores indicate better performance.
RESULTS
The 21 scale items were analyzed in a series of 3 (group) x 3 (age) repeated measures of analysis of variance. For multiple mean comparisons, the critical difference for a two-tailed t test (alpha equal to .05) was used. Table II shows the mean group scores for age and day of testing for the 21 behavioral items of the Brazelton scale. Table III presents a summary of the analysis of variance effects for the three groups of neonates.
Table III indicates that significant differences among the groups were found for 13 of the 21 items. All but three items showed significant changes in performance during the three exami-nations. Five of the items showed significant interactions.
Age Effects
All four items included in the dimension of interactive processes showed significant age effects. Significant improvement in inanimate visual orientation was found over all three days. These changes in performance were signficant for the middle- and lower-class groups, but not for the Metera infants. Performance in inanimate auditory orientation, animate auditory orienta-tion, and alertness improved from day 1 to day 5 with no significant change from day 5 to day 10 for all three groups.
TABLE II
MEAN SCORES ON BRAZELTON SCALE ITEMS BY GROUP AND DAY OF TESTING FOR THE METERA,
MIDDLE-CLASS, AND LOWER-CLASS GREEK NEONATES
&azelton Scale Items Group
1 Day
5 10
Metera Middle Lower
Class Class
Interactive processes
Orientation inanimate visual 3.33 4.36 4.98 3.18 4.30 5.12
Orientation inanimate auditory 5.39 5.46 5.73 5.24 5.73 5.61
Orientation animate auditory 4.67 5.07 5.62 4.56 5.19 5.61
Alertness animate visual 4.59 5.44 6.09 4.49 5.71 5.92
Motor processes
Motor maturity 5.05 5.42 5.80 4.61 5.47 6.19
General tonus 4.18 4.08 4.44 3.86 4.36 4.49
Defensive movements 3.90 4.57 4.88 3.73 4.55 5.06
Pull-to-sit 4.47 4.56 4.80 3.75 4.67 5.40
Hand-to-mouth 4.32 5.24 5.54 4.23 5.03 5.78
Activity 3.78 3.42 3.70 3.51 3.71 3.69
Vigor of movements 4.60 5.03 5.17 4.54 5.00 5.27
Tempo of movements 3.93 3.88 4.42 3.65 4.32 4.26
Regulation of state
Habituation to light 5.03 6.11 6.44 5.14 5.92 6.53
Rapidity of buildup 3.81 3.63 4.60 3.92 4.02 4.10
Lability of states 2.97 2.88 3.36 3.00 2.91 3.29
Irritability 4.15 4.04 4.47 3.96 4.16 4.54
Consolability 4.56 3.97 4.42 3.34 4.44 5.17
Self-quieting 5.47 6.11 5.99 5.60 5.83 6.13
Physiologic organization
Tremulousness 6.47 3.83 4.68 4.29 4.90 5.79
Amount of startles 5.55 4.71 5.21 4.80 5.14 5.53
Lability of skin color 5.30 4.72 5.65 6.12 4.63 3.61
among groups on days 5 and 10. Significant improvement across all three days was found for defensive movements, motor maturity, pull-to-sit, and hand-to-mouth activity. Vigor of movement changed only from day 1 to day 5 and general tonus from day 1 to day 10.
On dimension 3, regulation of state, habit-uation to light and consolability improved signif-icantly over the three test days. For the latter, a significant interaction revealed that the Metera infants did not improve from day 5 to day 10. Improvement in lability of states was seen from day 5 to day 10 and irritability improved from day 1 to day 10.
In the last dimension, physiologic organization, significant age effects were demonstrated in all three items. For startles, the change from day 1 to day 10 was significant. Lability of skin color and tremulousness showed significant improvement across all three days. Tremulousness, however, improved significantly only in the lower-class group. The Metera infants showed consistently higher scores whereas the middle-class infants scored consistently lower.
Group Effects
Three of the items in the interactive process showed significant group differences. The Metera infants had poorer performances than the lower-class group on alertness and animate auditory orientation. For inanimate visual orientation, all three groups demonstrated significant differ-ences, with the Metera infants showing the poor-est performance and the lower-class infants having the best scores.
Five items showed significant group differences
in the dimension of motor processes. Lower-class
infants scored higher than the Metera infants on
motor maturity and defensive movements. For tempo of movement, the lower-class group had better scores than both Metera and middle-class infants. The Metera group showed poorer
performance than both lower- and middle-class
infants on hand-to-mouth activity. The interac-lion for general tonus was significant due to
poorer performance by the Metera group than by
the lower-class infants on days 1 and 10 (not on day 5).
differ-TABLE III
SUMMARY OF ANALYSIS OF VARIANCE EFFECTS FOR THREE GROUPS OF GREEK NEONATES
Brazelton Scale Items Group
F(2, 86)
Age F(2, 172)
Group X Age
F(4, 172)
Interactive processes
Orientatiom3 inanimate visual 14.24#{176} 84.92#{176} 6.510
Orientation inanimate auditory 0.62 5.29t 1.61
Orientation animate auditory 5.51t 28.53#{176} 1.50
Alertness animate visual 12.43#{176} 62.42#{176} 2.0
Motor processes
Motor maturity 3.22 54.96#{176} 0.30
General tonus . 1.25 5.80f 3.7
Defensive movements 5.70t 38.03#{176} 1.10
Pull-to-sit 0.56 84.51#{176} 0.58
Hand-to-mouth 4.35* 24.81 #{176} 1.56
Activity 2.80 2.35 4.70t
Vigor of movements 1.56 15.70#{176} 2.20
Tempo of movements 3.64 9.43#{176} 0.38
Regulation of state
Habituation to light 8.89t 31.36#{176} 0.82
Rapidity of buildup 11.43#{176} 0.48 0.31
Lability of states 3.57 3.33 0.31
Irritability 1.89 6.52t 0.99
Consolability 1.39 35.04#{176} 4.05t
Self-quieting 1.62 2.61 0.75
Physiologic organization
Treirnilousness 29.47#{176} 36.45 #{176} 6.010
Amount of startles 4.92t 10.45#{176} 0.63
Lability of skin color 3.37 67.46#{176} 0.43
OP< .001
tP < .01. :t:P < .05.
ences on the dimension of regulation of state. The Metera infants scored lower than both the lower-class and middle-class infants on
habitua-tion to light. Performance by the lower-class infants was superior to that of both the Metera and middle-class infants for rapidity of buildup and lability of state.
The dimension of physiologic organization revealed significant group differences for all three items. The middle-class infants received the lowest scores for startles, lability of skin color, and tremulousness. The score for tremulousness was higher for the Metera infants, but scores for
startles and skin color were not significantly different from those of the lower-class infants.
DISCUSSION
The Metera infants exhibited the poorest
performance in most areas. On nine items (visual
orientation to inanimate objects and to the human voice, alertness in general, motor maturity, defen-sive movements, activity, hand-to-mouth motion, tempo of movement, and habituation to light), the Metera group had overall poorer scores than the other two groups. The lower-class group, which
was to be a control for SES, was in no way comparable to the Metera group; it performed significantly better than the Metera infants on all
nine items. Also, on two items, inanimate visual orientation and consolability, the Metera infants did not show the significant improvements with age shown by the other two groups.
Pregnancy out of wedlock is still a serious offense in most of Greece and this orphanage has been the only refuge for pregnant girls. In a practice still observed in the mountains and on some of the Greek islands, the father and brothers of the girl who becomes pregnant out of wedlock are expected to destroy her before the baby can ruin the family.’ Hence, it is common practice for unwed girls to attempt to abort themselves, to eat poorly in early pregnancy in order to hide the pregnancy, and to deny the pregnancy. Not until the seventh month, when they accept the preg-nancy and come to the Metera for care, does the fetus receive an approriate caloric intake. This situation may well account for the poorer performance of these infants in the neonatal period.
Metera infants performed most poorly in overall social responsiveness. The environment of the infants after hospital discharge differed signifi-cantly among the three groups and might well be reflected by the scores. The Metera infants were isolated in brightly lit white cubicles, similar to the hospital. They were fed on a four-hour routine, and little extra stimulation was provided as a routine. The middle- and lower-class infants were fed on demand and were offered the person-al interaction of parents and home. Since this interactive dimension reflects the infant’s impor-tant ability to elicit responses from and to give feedback to the environment, it becomes a signif-icant predictor of both the infant’s capacity to elicit necessary nurturing and the parents’ perception of the infant.
The Metera and the middle-class Greek babies performed significantly poorer than the lower-class group on the items in the dimensions of motor processes and state regulation. Since the middle-class babies were not depleted in utero, their depressed performance on these dimensions may reflect the effects of medication.
Future Studies
The data presented here suggest areas for future investigation to understand differences in genetic endowment, intrauterine depletion, ma-ternal medication, and the effects of postuterine environment on feeding and early stimulation. Certainly, we cannot test all the many variables that are influencing each of these groups of neonates. At this point, one must simply make conjectures as to the cause of each of the differ-ences.
The most encouraging aspect of this study is that it demonstrates that a behavioral assessment repeated several times in the neonatal period can be sensitive to group and group-by-age differ-ences. It calls for more studies of neonatal behav-ior in groups of neonates who have different genetic endowment and are subjected to different intrauterine and perinatal experiences.
The most striking factor in these data is the capacity of the human neonate to recover from intrauterine and postuterine influences that may not always be to his best interest. In all of these groups, the infant’s strengths and ability to recov-er after birth create a powerful impression of developing organization in the early neonatal period.
REFERENCES
1. Cravioto J, deLicardie ER, Birch HG: Nutrition, growth, and neurointegrative development-an
experimental and ecologic study. Pediatrics 38:3 19, 1966.
2. Scrimshaw NS, Taylor CE, Gordon JE: Interactions of nutrition and infection. WHO Monograph Ser
57:1-68, 1968.
3. Winick M: Malnutrition and brain development. I
Pediatr 74:667, 1969.
4. Zamenhof S, VanMarthens E, Margolis FL: DNA (cell number) and protein in neonatal brain. Science
160:322, 1968.
5. Pollitt E: Behavior of infant in causation of nutritional
marasmus. Am I Clin Nutr 26:264, 1973. 6. Brazelton TB, Freedman DG: Manual to accompany
Cambridge newborn behavioral scales, in Normal
and Abnormal Development of Brain and Behacior,
Boerhoave Series. Leiden, Leiden University Press,
1971, pp 104-132.
7. Als H, Tronick E, Adamson L, et al: The behavior of the full-term yet underweight newborn infant. Dec Med Child Neurol 18:590, 1976.
8. Lester BM, Zeskind PS: The organization of crying in the infant at-risk, in Field T, Sostek A, Goldberg S et al (eds): Infants Born At-Risk. New York, Spec-trum Publications, to be published.
9. Brazelton TB, Tronick E, Lechtig A, et al: Biomedical variables and neonatal performance of Guatemalan
infants. Dec Med Child Neurol 19:364, 1977. 10. Lester BM, Emory EK, Hoffman 5, et al: A ,rnilti-variate
study of the effects of high-risk factors on perfor-mance on the Brazelton neonatal assessment scale.
Child Dec 47:515, 1976.
11. Brackbill Y, Kane J, Manniello RL, et al: Obstetric meperidine usage and assessment of neonatal status.
Anesthesiology 40:116, 1974.
12. Standley K, Soule AB, Copans SA, et al: Local-regional anesthesia during childbirth: Effects on newborn behaviors. Science 186:634, 1974.
13. Brazelton TB, Robey JS, Collier GA: Infant develop-ment in the Zinacanteco Indians of southern Mexi-co. Pediatrics 44:274, 1969.
14. Grantham-McGregor SM, Hawke M: Developmental
assessment of Jamaican infants. Dec Med Child
Neurol 13:582, 1971.
15. Bayley N: Comparisons of mental and motor test scores for ages 1-15 months by sex, birth order, race, geographical location and education of parents.
Child Dec 36:379, 1965.
16. Brazelton TB: Influence of perinatal drugs on the behavior of the neonate, in Hellmuth J (ed): The
Exceptional Infant. New York, Bruner-Mazel, 1971,
vol 2, p 419.
17. Brazelton TB: Neonatal Behacioral Assessment Scale,
publication 50. National Spastics Society Mono-graph Series, London, William Heinemann and Sons, 1974.
18. Prechtl H, Beintema 0: The Neurological Examination
of the Full Term Newborn Infant. London, William
Heinemann and Sons, 1964.
19. Blum R, Blum B: Health and Healing in Rural Greece.
Stanford, Calif, Stanford University Press, 1965.
ACKNOWLEDGMENT
This study was supported in part by a grant to the, Childrens Hospital of the District of Columbia, by grant MH 09228 from the National Institute of Mental Health, and by
grant 3122 to Dr. Brazelton from the William T. Grant
The authors express their gratitude to Spyros Doxiadis,
M.D., professor of pediatrics and president of the Institute of
Child Health in Athens, who provided us the chance to study these infants and whose support and encouragement made the study possible. Dr. Dorothy Huntington, consultant to
the Metera, presently chief, Child and Family Services,
Community Mental Health Center, Peninsula Hospital, Burlingame, California, furnished the inspiration and the plan for the study and the continued advice necessary to complete it.
CMV AND INFANTILE SPASMS
At the Children’s Hospital, University of Helsinki, 205 children with infantile spasms who were born between 1960 and 1976 were studied in a
search for the factors responsible. In 1 1 children (5%) the infantile spasms were possibly associated with cytomegalovirus (CMV) infection. The number may actually have been considerably higher, as no systematic search was made for CMV, especially in the early years. In four of the 1 1 children, the infection was probably congenital, and was the most likely cause of the spasms. One of the remaining seven children had congenital toxoplasmosis and the simultaneous CMV infection was probably also congenital. The children with congenital CMV infection exhibited severe clinical symptoms in the neonatal period or in early infancy. Two frequent symptoms were persistent tremor and meningoen-cephalitis. Later, all five children were severely mentally retarded and had spastic tetraplegia and small heads; three of them had optic atrophy and were blind. In the other six children, the CMV infection was probably acquired, the clinical symptoms being less severe, and the spasms may have been caused by another factor. In two of the 1 1 cases, immunosuppressive therapy (ACTH treatment generally given for infantile spasms) seems to have caused a fulminant CMV infection. Two children with CMV infection still show signs of a slow virus infection in the central nervous system many years later.
Froni 1)ec Med Child Neurol 20:570, 1978.
