Portal de Periódicos da CAPES

Historical Perspectives

2003; American Academy of Pediatrics; Volume: 4; Issue: 8 Linguagem: Inglês

10.1542/neo.4-8-e199

1526-9906

Alistair G.S. Philip,

Neurology and Historical Studies

The article chosen for this month’s Historical Perspectives is “Neurological Evaluation of the Maturity of Newborn Infants” by Claudine Amiel-Tison. (1) This paper previously was chosen for inclusion in the Ross series “Landmarks in Neonatology/Perinatology,” at which time (August, 1982) Dr Amiel-Tison provided a commentary, which also is reproduced here.Some of the maneuvers Dr Amiel-Tison described were incorporated into scoring systems for assessing gestational age, to which she raised several objections, but she described newer findings that have formed the basis for her continuing development of neurologic evaluation of the neonate, infant, and young child.When I asked her to provide a personal reminiscence of how the paper came into being, I also asked her to put it into today’s perspective. She prepared an overview of the evolution of her work over the past 40 years. Dr Amiel-Tison continues to emphasize the need for caution in interpreting the results of scoring systems for gestational age. She has maintained the great tradition of careful neurologic evaluation she inherited from Dr André Thomas and Dr Suzanne Saint-Anne-Dargassies. Because she has both trained and collaborated with many international investigators, her approach has been widely disseminated. This review provides some special insights that should be of interest not only for the neonate, but also the infant and young child.The following commentary originally appeared in 1982 as one of a series published by Ross Laboratories entitled “Landmarks in Neonatology/Perinatology—Current Comment” and is reproduced with permission of Ross Products Division, Abbott Laboratories, Inc, Columbus, Ohio.While attending the clinic of Dr André Thomas at the Baudelocque Maternity Hospital in the early sixties, I came to understand Thomas’ passion for the neurology of the newborn.André Thomas was an old man by then, nearly blind and deaf. He concentrated all his strength on the examination of the infant before him, as he explained the reflexes and tone reactions he was eliciting to his few silent disciples. (1)Dr Saint-Anne-Dargassies shared Thomas’ passion, inherited his method of examination, and described the stages of development which characterize maturation between 28 and 40 weeks gestational age. (2) This neurologic approach has been invaluable help in determining gestational age and in determining and describing specific abnormal neurologic signs for a given gestational age.This type of evaluation, however, was difficult to learn other than by direct observation and repeated practice. Having been taught by Saint-Anne-Dargassies, I tried to transmit her neurologic evaluation method to pediatricians. I selected the most important items from the examination, described precisely each maneuver involved, and illustrated, in table form, the normal responses according to six age groups. (3) As I gained personal experience, I came to feel that the maturational pattern of active tone in neck flexors and neck extensors could be better described and illustrated. (4)The need to determine gestational age became increasingly evident with the growth of perinatology as a subspecialty. However, the extensive use of artificial ventilation and other accoutrements of newborn intensive care made it more and more difficult to perform the neurologic examination properly. The logical solution was to find other tools, such as external criteria. (5)(6) Another tendency was to simplify neurologic testing and to use it in conjunction with external criteria to create an even faster and easier evaluation. (7)(8)I have three objections to such simplification of neurologic assessment: Error in estimating gestational age from the mother’s last menstrual period amounts to at least ±1 week, even if the dates are carefully verified.Then, nice linear correlations showing that one score is better than another are perfectly all right as long as one decides to forget about the errors hidden behind the score.Aside from these somewhat irritating distortions, and despite the difficulties and limitations described above, this neurologic approach to the newborn has been the basis of new research. These observations are of considerable theoretical interest, offering one way to show that upper structures are already functioning when lower structures are functionally suppressed temporarily.Thus, the study of the evolution of tone and reflexes from fetal life through the first few months postnatally is one of continuing practical as well as theoretical value.When, as a newcomer, I joined the clinical and research group of Alexandre Minkowski in 1962 in Paris, brain maturation in the fetus and the neonate was the central theme. Jupiter was well set indeed, with three ladies familiarly designated as “the three graces”: Saint-Anne-Dargassies was the clinician, Dreyfus-Brisac the electrophysiologist, and Larroche the pathologist. They described typical maturation patterns on the one hand and specific types of brain damage in correlation with gestational age (GA) at birth on the other. Their pioneering work was rapidly and widely recognized.As a clinician, I was taught by Saint-Anne-Dargassies. It was a permanent delight to observe her manipulating a “premie,” but it took me years to determine how each maneuver could be analyzed and both typical and abnormal responses defined. I had to go from magic to analysis and synthesis. In this long process, I benefited from various effective supports. When I spent a few months as a fellow in Stanford in 1966, Marshall Klaus, with his usual communicative enthusiasm, strongly encouraged me to publish a simplified method that would be accessible to neonatologists. I was offered the help of the illustration department of Stanford University to elaborate the graphic representation, which resulted in the three tables included in my first paper. (1)During the following years, due to the chronic lack of space in the research unit, my desk was in the pathology lab of Jeanne-Claudie Larroche. I was, therefore, participating in the lab life, with brains all over the place at various stages of pathologic processing from macroscopic inspection to serial slicing, coloration, microscopic analysis, and finally clinicopathologic conference. Each time Dr Larroche had an interesting slide under her microscope, she invited me to have a look (“Claudine viens voir”). This was such amazing training for a clinician that I was very well prepared for the imaging revolution that came in the early 1980s; anatomoclinical correlations were transposed with no surprise into clinicoradiologic correlations.When Saint-Anne-Dargassies applied André Thomas’ assessment of the term neonate (2) to 100 preterm infants born from 28 to 37 weeks’ gestation, the approach was descriptive. (3) From these observations, organized at 2-week intervals, she brilliantly identified the upward (caudocephalic) progression of maturation during the last 3 months of fetal life. However, she did not elaborate on anatomic and physiologic correlates to understand the meaning of this caudocephalic progression. It was clear that most of the criteria used were dependent on brainstem activity, particularly primary reflexes that exerted such a fascination for adult neurologists such as André Thomas and many others.At the same time, Peiper, a German pediatrician, described the downward (cephalocaudal) progression of maturation during the first 2 years after birth, depending on the cerebral hemispheres. (4) To illustrate the waxing and waning pattern, I proposed a graphic representation for both passive and active tone (Fig. 1 ). (5) Anecdotally, Heinz Prechtl did not like at all the upward progression of passive tone in flexion in the last 3 months of fetal life and sent me, with his kind regards, a reprint of his study in full contradiction of this upward progression. (6) In 1998, however, some support came from a fetal ultrasonographic study on prenatal development of arm posture that demonstrated a clear developmental trend toward increased flexion from 12 to 38 weeks’ gestation. (7)It was only in the 1980s, with the general review of Harvey Sarnat (8), that I began to feel comfortable about anatomophysiologic correlates (9) with two distinct neuromotor systems: the subcorticospinal (or “lower”) system and the corticospinal (or “upper”) system, including specific physiologic correlates and specific timing and direction of maturation. This new step resulted in a didactic representation (Fig. 2 ). Antigravity activity is dependent on the lower system. In contrast, the head passing forward in the raise-to-sit maneuver is the first neuromotor achievement under the control of the upper system, which can be demonstrated clinically from 34 weeks’ gestation onward.These correlates shed light on clinical observations I made long ago on head straightening reactions (10) without fully understanding their meaning at the time: 1) When a term newborn or a preterm infant reaching 40 weeks’ gestation is capable of showing identical activity in the neck flexor muscles (going forward in the raise-to-sit maneuver) and the neck extensor muscles (going backward in the sitting-to-lying maneuver), this is the best demonstration of an intact upper system. 2) In contrast, flexor muscles showing poor or no activity associated with a powerful response in the extensor muscles is one of the best early signs of damage in the upper system. 3) The association of hypotonic upper limbs with difficulties in oral feedings very likely indicates more severe damage reaching the lower structures (brainstem). These findings make sense, based on our knowledge that periventricular leukomalacia in the preterm infant typically is located in the upper structures; only in more severe cases are hypoxic-ischemic lesions also found at the lower level.In my first paper, (1) passive tone criteria were overrepresented. In the following years, it became apparent that active tone is the most meaningful part of the assessment, both for assessment of maturation and for detection of neurologic signs. Passive tone is more contingent, within the first days, on many types of interference (eg, too much flexion in the limbs at birth due to restricted space in utero or hypotonia due to general condition). Primary reflexes also are not very helpful. At this stage, they simply have to be present to exclude central nervous system depression; therefore, their elicitation may be restricted to a few of them.As a consequence, the next step has resulted in only one table (Fig 3 ) based on 10 criteria, including four passive tone items (popliteal angle, scarf sign, forearm recoil, dorsiflexion angle of the foot), three active tone items (righting reaction, neck flexor tone, neck extensor tone), and three primary reflexes (finger grasp and response to traction, crossed extension, sucking). Changes are described at 2-week intervals from 32 to 40 weeks’ gestation, as originally proposed by Saint-Anne-Dargassies. Periods of rapid modification are highlighted, indicating the most discriminative period for each observation.How to define the neurologic age of one infant from the responses obtained? A definite conclusion can be reached only if most of the responses are within the same age period, which is the case for most neonates in good general condition. However, responses can be scattered in some cases, with about 50% of them in one age period, a few in the “younger” interval, and a few in the “older” interval. For research purposes, the so-called “uniform pattern” is defined as 7 of 10 responses being in the same 2-week gestation period and the so-called “scattered pattern” as more than 3 responses being “out of line.”We attempted to discern the meaning of a “scattered pattern” by studying a few sets of twins who had discordant birthweights. We repeated the clinical assessment every 2 to 3 days until the “thin” (wasted) infant recovered from various degrees of undernutrition and underhydration. The “thin” twin often exhibits a scattered pattern in the first postnatal week, reaching a uniform pattern with improvement in general condition, usually by the beginning of the second postnatal week. In some of these infants, however, the maturation pattern remains scattered. These unpublished findings helped us to determine that: 1) Neurologic performances are not independent of the general condition of the neonate, which seems obvious to clinicians (asking active tone responses, such as response to traction or head straightening, of a very tired neonate is the equivalent of asking a convalescent adult to run a marathon) and 2) A uniform pattern, obtained immediately after birth or after convalescence, is a good marker for an intact brain. In contrast, a scattered pattern can suggest risk about transient or persistent brain dysfunction.These examples illustrate how maturational profiles, derived from the “routine use” of Figure 3 , can be an important part of the evaluation of the newborn brain. This is the opportunity for me to thank once more my sister, Annette Tison, who created most of the illustrations accompanying my didactic writing, particularly Figure 3 . I have found that it is better for me not to ask her to produce a graphic representation of a maneuver still imprecise in my mind; she needs to understand before drawing, and she has no mercy for haziness. It is why her drawings are so sharp and helpful.There was a dogma in Port-Royal that brain maturation in utero is independent of unfavorable gestational circumstances and proceeds at the same speed ex utero. As a consequence of this independence, one can determine GA from the assessment of neurologic age at birth. At a time when GA was unknown in about one third of pregnancies, neurologic age was very helpful to separate small-for-gestational age (SGA) from appropriate-for-gestational age (AGA) babies, when early fetal measurements by ultrasonography (US) were not even a dream. However, being so confident in their abilities, pediatricians became somewhat arrogant and began to “correct” the GA calculated from last menstrual period or assessed on early fetal US. When assessing neonates whose GA was known precisely (based on temperature curve or stimulation of ovulation), I was puzzled by discrepancies of several weeks between GA and neurologic age, but I did not dare to challenge the dogma.In 1972, Lou Gluck visited us in Paris on his way to the international pediatric meeting in Moscow. He was caught up in the excitement of his discovery concerning lecithin/sphingomyelin ratio and lung maturation and observed that in a few abnormal pregnancies, the fetus showed accelerated maturation of both lung and brain. (11) He liked so much to share his enthusiasm that he announced to me when washing his hands before going into our neonatal unit: “Claudine, I know that maturation rhythm may vary with adverse gestational circumstances.” He “knew” on the basis of five babies, a remarkable intuition, and a precious self-confidence. This was the trigger I needed to switch from the notion of immutability to the notion of flexibility of maturation speed.From 1972 to 1978, I collected 32 cases in which the menstrual history was precise and neurologic assessment indicated a maturational advance of 2 or more weeks. Most of these 32 infants had abnormal pregnancies with causes similar to the cases published by Gluck and his group in 1977: placental insufficiency due to preeclampsia, multiple pregnancy, or uterine malformation. (12) As I gathered these data, the atmosphere was tense in the department because no one in Port-Royal really wanted to consider this flexibility. The electrophysiologists disagreed because they could not confirm an acceleration of maturation with electroencephalographic (EEG) criteria in the presence of clinical acceleration. (This disagreement is not surprising because EEG assesses primarily cortical activity, and clinical criteria at these ages primarily reflect brainstem activity.) Clinical criteria alone did not convince the editorial boards. Therefore, I had to follow the advice of Leonard Strang (on sabbatical in Port-Royal at this time) and sacrifice 50% of my cohort to concentrate the publication on the 16 cases in whom the acceleration was 4 weeks or more (ie, outside the 95% confidence limits). (13) From my initial cohort, however, I demonstrated that the adaptation was not an all-or-nothing phenomenon; rather, it was a progressive response of variable degree.Confirmation of acceleration of maturation in stressed pregnancies came in 1985 from various horizons. Experimental data in intrauterine growth-restricted rats showed that the Na/K-ATPase activity was accelerated in the brainstem and delayed in the forebrain, cerebellum, and hippocampus. (14) At the same time, based on brainstem auditory evoked potentials, Alan Pettigrew found a significant decrease of mean brainstem conduction among SGA infants compared with AGA infants. (15) When Pettigrew came on sabbatical to Paris, we enjoyed preparing together a general review on the topic. (16)During all these years, the suggested physiopathologic mechanism of response to stress was adrenal stimulation, but the mechanism remained vague until the publications of Robert Denver, a biologist in Ann Arbor, Michigan, which came as a fantastic surprise to me. (17)(18) He reported that tadpoles of desert amphibians can accelerate metamorphosis as their pond dries, thereby escaping mortality in their larval habitat. This observation has been replicated in laboratory experiments, with tadpoles subjected to habitat desiccation exhibiting elevated hypothalamic content of corticotrophin-releasing hormone, precocious activation of the thyroid and interrenal axes, and endocrine activation preceding morphologic changes by 3 days. Although endocrine mechanisms are probably not identical in amphibians and humans, sharing these adaptive processes with amphibians is a thrilling concept for those who are interested in the evolutionary aspects of maturation. A joint general review of the concept is in process. (19) Demonstration of fetal adaptation to the environment being shared with others in the animal kingdom at last may make this flexibility politically correct. If a convincing meta-analysis still is wanted, I am afraid it may have to wait for another 30 years!Neonatologists certainly feel frustrated when considering the objectivity of the Gosner Table for staging anuran embryos and larvae based on morphologic changes. (17)(18) However, such frustration is not an excuse for scoring neurologic findings for several reasons. First, scoring systems such as the Dubowitz (20) and the Ballard (21) suggest a level of accuracy that really does not exist in these dynamic systems. Second, nearly exclusive use of passive tone criteria (easier to evaluate than active tone, but often misleading, as discussed previously) decreases the validity of the assessment, which hardly can be qualified as neurologic. Unfortunately, most neonatologists accept the use of scores, even if not accurate, because they want a GA at any price.To me, this leveling is a real deprivation compared with identification of a maturational profile. As an example, it is typical for an infant born at 34 weeks’ gestation to lie in a froglike position, to have a difficult response going forward in the pull-to-sit maneuver (head rolling on the shoulder), and to have a much more powerful response going backward. The same pattern at 40 weeks corrected age is definitely abnormal, indicating impaired control of the lower system by an impaired upper system. In other words, what is physiologic at 34 weeks must be interpreted as pathologic 6 weeks later. This identification of early neurologic signs is not a useless sophistication; it is a good reason to include a preterm newborn in a follow-up clinic, even if repeated head US studies within the first postnatal month did not show any sign of cerebral atrophy. (22)The 10 criteria of Figure 3 test only neuromotor function and rely nearly exclusively on the lower (brainstem) structures. As discussed previously, only changes observed in neck flexor muscle reaction to the raise-to-sit maneuver are dependent on the upper system (motor cortex, corticospinal tracts, and basal ganglia). Because the brain damage due to perinatal hypoxia-ischemia generally is localized in the upper structures, absence of predictive value of the proposed maturational assessment is not surprising. For this reason, at around the corrected age of 40 weeks’ gestation, we are using a more extensive assessment to define optimality of the central nervous system (Table). (9)(23) Other criteria are added to Figure 3 , primarily in the domains of alertness and interaction (relying mostly on visual fix and track) and cephalic growth. Passive tone in the axis, independent movements of fingers, and abduction of thumbs are included as valuable criteria to define the integrity of cerebral hemispheres.From my experience, typical responses at 40 weeks’ gestation have a high predictive value for normal outcome; very deviant responses have a high predictive value for a poor outcome. Mild and moderate deviations do not have a good predictive value; they simply indicate the need for follow-up, even when imaging results are normal. Particularly in the “new survivors” group of neonates (22 to 26 wk), extreme immaturity at birth makes any prediction unwise. When tested at 40 weeks corrected gestational age, many of these infants are considered typical, based on tone and reflexes, but most demonstrate a poor visual ability to fix and track due to a short attention span. Generally, most will be able to walk independently by 18 months corrected age, but disorganization in many aspects of cerebral function will make their lives difficult. For this group of children, the clinical approach described previously is disappointing for prediction. New technologies to evaluate cerebral function are more promising, including cortical mapping or diffusion tensor magnetic resonance imaging.Who should assess brain maturation of the neonate? The neonatologist can do it and is in the best situation for interpreting the results according to the general status of the neonate. Moreover, this clinical approach is indispensable to developmental pediatrics, and I strongly believe that any pediatrician has to be a “developmentalist.”Which neonates should be assessed? Everyone. It is clear, however, that active tone in particular cannot be assessed in a sick newborn or in a very immature newborn. In these neonates, the assessment will be completed later, when full stabilization is achieved, which may take months when the GA is 27 weeks or less. Using corrected age and assuming that maturation has been progressing at about the same speed as in utero, the type of assessment discussed previously will provide a valuable baseline for evaluation of further maturational steps and possible brain damage.Can a single instrument be used from birth through childhood? Yes, it seems highly desirable because any deviation in clinical methodology makes conclusions uncertain (or epidemiologists unhappy). It is for the sake of a collaborative research project that I recently have tried with Julie Gosselin to adapt the same methodology through the maturational flow from birth to 6 years of age. With criteria similar to those described for assessing maturation up to 4 weeks, we developed a semi-quantitative assessment scored 0, 1, and 2 from 40 weeks (23) to 6 years. (24) We used the same concept of “neurologic age,” but as maturation slows, the 2-week assessment periods up to 40 weeks become 3-month periods up to 1 year and subsequently 6-month periods and finally 12-month periods up to 6 years. We persist in our refusal to assign a total score, using instead clusters to define various types of deviations, just as in the neonatal period.The fluidity of maturation is comparable to the fluidity of water, and as the Greek philosopher Heraclite asserted, we cannot swim twice in the same river, and similarly we cannot assess the same neonate twice. One of Heraclite’s disciples, more extremist than his mentor, claimed that we cannot even swim once in the same river because water flows constantly. In any case, brain maturation proceeds very rapidly, and we have to try hard not to get lost in the stream.[For the original article by Amiel-Tison regarding neurologic evaluation of maturity in newborns, which was published in 1968 in Archives of Disease in Childhood, and which is reprinted with permission, please see the data supplement .]