This article is commented on by Spittle, p. 698 of this issue.
Hand movements at 3 months predict later hemiplegia in term infants with neonatal cerebral infarction
Article first published online: 23 OCT 2009
© The Authors. Journal compilation © Mac Keith Press 2009
Developmental Medicine & Child Neurology
Volume 52, Issue 8, pages 767–772, August 2010
How to Cite
GUZZETTA, A., PIZZARDI, A., BELMONTI, V., BOLDRINI, A., CAROTENUTO, M., D’ACUNTO, G., FERRARI, F., FIORI, S., GALLO, C., GHIRRI, P., MERCURI, E., ROMEO, D., ROVERSI, M. F. and CIONI, G. (2010), Hand movements at 3 months predict later hemiplegia in term infants with neonatal cerebral infarction. Developmental Medicine & Child Neurology, 52: 767–772. doi: 10.1111/j.1469-8749.2009.03497.x
ACKNOWLEDGEMENTS The authors thank the families of the infants participating in the study and the staff of Pisa, Catania, and Modena neonatal intensive care units for their support. AG and GC were supported by the Italian Ministry of Health (grant RC 1/2008).
- Issue published online: 8 JUL 2010
- Article first published online: 23 OCT 2009
- PUBLICATION DATA Accepted for publication 7th August 2009. Published online.
Aim The aim of this study was to explore the predictive value of quantitative assessment of hand movements in 3-month-old infants after neonatal stroke.
Method Thirteen infants born at term (five females, eight males; mean gestational age 39.4wks, SD 1.19, range 37–41wks; mean birthweight 3240g, SD 203, range 2900–3570g) with neonatal arterial ischaemic cerebral infarction, and 13 healthy infants (mean gestational age 39.1wks, range 37–41wks, SD 1.26; mean birthweight 3190g, SD 259, range 2680–3490g) were enrolled in the study. The absolute frequency and the asymmetry of global hand opening and closing, wrist segmental movements, and independent digit movements were assessed from videotapes recorded at around 12 weeks. Neurological outcome was assessed when the infants were at least 18 months old using Touwen’s neurological examination.
Results Five of the 13 infants with neonatal stroke had normal neurological development, and eight had hemiplegia. Asymmetry of wrist segmental movements and the absolute frequency of independent digit movements were significantly different between infants with and without hemiplegia (p=0.006 and p=0.008, respectively). No differences were found in global hand movements.
Interpretation We propose that the observed abnormalities of hand movements are the result of two different mechanisms: direct disruption of the corticospinal projection to the spinal cord, and altered modulation of the central pattern generators of general movements.
New evidence suggests that spontaneous hand and digit movements in infants from birth to 5 months are various and abundant, including segmental wrist movements, global opening and closing of the hand, independent finger movements, pre-precision grips (sideways contact between thumb pad and the side of other fingers), and precision grips (contact between thumb pad and other digit pads).1 The emergence and early development of some of these complex movement patterns has been proposed to be related to the maturation and function of the corticospinal tract, based on recent electrophysiological and anatomical studies suggesting that, in humans, direct connections of the pyramidal tract to motor neurons are established before birth.2 Should this be true, development of hand movements could be impaired from the first weeks of life in infants with unilateral brain damage and later hemiplegia.
We recently showed how, in preterm and term infants with unilateral brain damage, asymmetries of segmental distal movements of the upper limb at 3 months predict later development of hemiplegia.3,4 In the present study we assessed a wider range of hand and digit movements in 3-month-old infants with neonatal cerebral infarction, a disorder occurring in about one in 4000 term neonates and leading in about half of the cases to permanent neurological or cognitive abnormalities.5 Our aims were as follows: (1) to confirm in another series of patients, and with stricter methodology, previous findings of the predictive value for later hemiplegia of the presence or asymmetry of segmental movements; (2) to assess the predictive value of the presence or asymmetry of a wider range of hand and digit movements; and (3) to explore the correlation between the pattern of hand and digit movements and the severity of later hemiplegia.
This study is part of a multicentre project to evaluate early development of spontaneous motor activity in infants with and without brain damage. The patients selected for this study were born at or after 37 weeks’ gestation in neonatal units in Pisa, Catania, and Modena, Italy, between 2004 and 2007. Inclusion criteria were a diagnosis of neonatal arterial ischaemic cerebral infarction and at least one video-recorded assessment of general movements at 12 weeks (SD 3wks). As we wanted to report on a new sample of patients with neonatal stroke, participants in our earlier study4 were excluded.
All infants presented at birth with neonatal seizures or with signs of mild to moderate encephalopathy (e.g. weak muscle tone and reflexes, weak cry, or reduced responsiveness). They had serial cranial ultrasounds or neonatal brain magnetic resonance imaging (MRI) showing signs of acute focal damage (transient unilateral focal echogenicity or abnormal focal signal on conventional or diffusion-weighted MRI) suggestive of arterial infarction.6 There were no signs of antenatal lesions on the early scans performed in the first few days after birth.7 The arterial distribution of the lesion was further confirmed on follow-up brain MRI performed after the neonatal period. Lesions were classified according to previously proposed methods.8,9
All 13 infants diagnosed with neonatal stroke during the period of this study had at least one video-recorded assessment at around 12 weeks, and were thus recruited into the study (mean gestational age 39.4wks, SD 1.19wks, range 37–41wks; mean birthweight 3240g, SD 203g, range 2900–3570g). Thirteen healthy infants matched for gestational age were selected from a large sample of healthy children from the three centres as a comparison group (mean gestational age 39.1wks, range 37–41wks; SD 1.26wks, mean birthweight 3190g, SD 259g, range 2680–3490g).
The study was approved by the ethics committees of the three institutions. Informed, written consent for participation was obtained from the parents of all participants.
As part of a long-standing multicentre project to evaluate general movements in infants at neurological risk, the follow-up programmes of the three centres involved in this study are uniform in terms of timing and type of clinical assessments. All infants with brain damage undergo video recordings during their hospital stay and every 3 to 4 weeks in the outpatient clinic, with a fixed scheduled appointment at around 12 weeks after birth. The video camera is positioned high above the infant, at an angle of 45° from the vertical on the sagittal plane, or slightly on a side. The infants are recorded between feeding times, in the supine position.
In the comparison group of healthy infants, recordings were performed in the respective outpatient clinics so as to be indistinguishable from those of the infants with brain damage.
Assessment of spontaneous motility
Two types of analysis, global and detailed, were performed on the video recordings by two evaluators (VB and MGD). Interobserver agreement, calculated using Cohen’s kappa10 after categorization of the continuous variables, was 0.74.
Global analysis consisted of the assessment of fidgety movements according to standard methodological principles for qualitative assessment of general movements (Prechtl’s method).11 Normal fidgety movements were defined as rounded movements of small amplitude, moderate speed, and variable acceleration (jerkiness) in all directions, involving the neck, trunk, and limbs. Fidgety movements were classified as absent when normal fidgety movements were never observed in the recording or abnormal when fidgety movements could be detected but their amplitude, speed, and jerkiness were moderately or greatly exaggerated.11 When more than one assessment was available during the fidgety age, defined as 12 weeks (SD 3) after birth, the one closest to 12 weeks was selected.
Detailed analysis consisted of the assessment of (1) global hand movements, (2) segmental movements at the wrist, and (3) independent digit movements. Global hand movements were defined as the simultaneous flexion (closing) or extension (opening) of all of the digits towards or away from the palm (Fig. S1, Supporting information published online). Segmental movements at the wrist were defined as movements of moderate speed at the level of the wrist joint, including rotation, palmar flexion and extension, and ulnar or radial flexion (see Fig. S1); this type of movement is typical of the fidgety pattern of general movements and may involve both the upper limbs and the lower limbs, trunk, and the neck.11 Independent digit movements include all of the digit movements that do not fulfil the criteria for the previous patterns, such as isolated movement of one finger, simultaneous movement of two or three fingers, and sequential movement of two or more fingers (see Fig. S1); this category also includes two patterns described by Wallace and Whishaw1: the pre-precision grip (the meeting of the thumb pad with the side of either the index or the middle finger) and the precision grip (the meeting of the thumb pad with the pad of either the index or the middle finger). Overall the analyses took 30 to 45 minutes for each infant.
All video recordings were first digitalized and then edited with Adobe Premier Pro software (version 2.0 for Windows, Adobe Inc., San Jose, CA, USA) to obtain 26 audio–video interleaved (AVI) files of spontaneous motility, one for each infant. Periods lasting from 3 to 5 minutes in which the infant was awake and not crying were selected. Files were numbered in random order using a free tool to generate random integers (RANDOM.ORG available at http://www.random.org/integers). To achieve a high level of blindedness for scoring, right and left arms were coded separately as follows. In a first step, numbered files were randomly ordered, and the right arm was scored for the files with even numbers, while the left arm was scored for the files with odd numbers. In a second step, numbered files were randomly reordered, and the same procedure was followed with inverted number/side relation.
For each of the three hand-movement patterns (global hand movements, segmental movements at the wrist, and independent digit movements) the number of movements observed was counted, and the frequency, expressed as the number of movements per minute, was calculated.
An asymmetry index was also calculated according to the following equation: , where νCM is the movement frequency of the contralesional hand and νIM is the movement frequency of the ipsilesional hand. The asymmetry index could range from −1 to +1, a negative score indicating higher movement frequency in the ipsilesional hand. In the healthy infants the right hand was considered as ipsilesional and the left hand as contralesional, with negative scores indicating a prevalence of right-hand movements.
Neurological outcome was assessed when the infants were 18 months or older using Touwen’s12 neurological examination. Hemiplegia, when present, was graded according to the degree of involvement of the upper limb, as described by Claeys et al.:13 mild, when pincer grasp or isolated finger movements were possible; moderate, when only global use of the hand was possible; severe, in the absence of any use of the hand.
For statistical analysis three groups were considered: infants with neonatal stroke and normal neuromotor outcome, infants with neonatal stroke and hemiplegia, and healthy infants. As the samples were small and the variables were not always normally distributed, non-parametric tests were used in all analyses.
Clinical data from the participants are reported in Table I. All infants had unilateral arterial infarction in the territory of the middle cerebral artery (right side n=4; left side n=9). In five infants the infarction involved the main branch, in six a cortical branch, and in the remaining two a lenticulostriate branch. Eight infants developed hemiplegia, which was mild in four and moderate in four. The remaining five infants showed a normal neuromotor outcome.
|Infant||Type of lesion||Outcome||Fidgety movements||Frequency (movements/min)|
|Contralesional upper limb||Ipsilesional upper limb|
|1||Right MCA, cortical branch||Normal||Normal||22.9||12.0||2.2||19.6||18.5||3.3|
|2||Left MCA, main branch||Normal||Normal||26.5||2.8||2.8||26.3||10.2||2.9|
|3||Left MCA, main branch||Normal||Normal||26.8||19.1||2.6||21.0||10.0||2.0|
|4||Right MCA, lenticulostriate||Normal||Normal||46.1||5.4||1.1||42.9||7.5||4.3|
|5||Left MCA, cortical branch||Normal||Normal||25.3||21.3||0.7||21.8||6.5||2.4|
|6||Right MCA, lenticulostriate||Mild left hemiplegia||Normal||28.6||0.6||2.2||40.2||4.9||5.6|
|7||Left MCA, cortical branch||Mild right hemiplegia||Absent||3.9||0.0||2.6||22.2||11.1||2.2|
|8||Left MCA, cortical branch||Mild right hemiplegia||Absent||3.4||0.8||4.2||8.5||0.8||0.8|
|9||Left MCA, cortical branch||Mild right hemiplegia||Normal||36.0||8.6||3.4||32.6||17.1||3.4|
|10||Left MCA, main branch||Moderate right hemiplegia||Absent||5.6||0.7||2.8||40.4||12.1||6.7|
|11||Left MCA, main branch||Moderate right hemiplegia||Absent||6.1||0.0||1.7||19.1||0.9||1.7|
|12||Right MCA, cortical branch||Moderate left hemiplegia||Absent||0.9||0.9||2.8||41.5||11.1||6.5|
|13||Left MCA, main branch||Moderate right hemiplegia||Absent||2.3||0.8||2.3||23.4||0.7||1.5|
Global quality of general movements and hemiplegia
The results of the global assessment of general movements at the fidgety period are summarized in Table I. In the comparison group, normal fidgety movements were observed in all infants. In the group of infants with neonatal stroke, absence of normal fidgety movements was significantly associated with the presence of hemiplegia (p<0.01, Fisher’s exact test; Table II). Sensitivity was 0.75 (95% confidence interval [CI] 36–96), and specificity was 1 (95% CI 46–100).
|No fidgety movements||Fidgety movements|
Asymmetry of movement patterns and hemiplegia
The degree of asymmetry of wrist movements, expressed by the asymmetry index, was significantly different in the three groups (Fig. 1; p<0.002, Kruskal–Wallis non-parametric test for n independent samples). Infants who developed hemiplegia showed negative asymmetry indexes, indicating a higher frequency of wrist movements in the ipsilesional (unaffected) arm. The asymmetry index for wrist movements was significantly lower than in infants with neonatal stroke and normal neurological outcome (mean difference 0.68; 95% CI 0.20–1.16; p=0.006, Mann–Whitney non-parametric test for two independent samples; Cohen’s d=4.4) and healthy infants (mean difference 0.70; 95% CI 0.26–1.15; p<0.001, Mann–Whitney test; Cohen’s d=3.5).
The degree of asymmetry of digit movements was significantly different in the three groups (Fig. 1; p<0.05, Kruskal–Wallis test). Infants who developed hemiplegia tended to show negative asymmetry indexes, indicating a higher frequency of digit movements in the ipsilesional (unaffected) arm. The asymmetry index for digit movements was significantly lower than in healthy infants (mean difference 0.69; 95% CI 0.01–1.37; p<0.09, Mann–Whitney test; Cohen’s d=1.5), whereas the difference from infants with neonatal stroke and normal outcome only approached significance (p=0.065, Mann–Whitney test). The degree of asymmetry of global hand movements did not differ in the three groups.
Frequency of movement patterns and hemiplegia
Table I shows the frequency per minute of the three hand movement patterns for each infant with neonatal stroke. The frequency of wrist movements in the contralesional hand was significantly different in the three groups (a mean value of the two hands was considered for healthy infants; Fig. 2; p<0.003, Kruskal–Wallis test). No difference was found for the ipsilesional hand (p=0.33, Kruskal–Wallis test). Infants who developed hemiplegia showed significantly lower frequency of contralesional wrist movements than healthy infants (mean difference 31.1 wrist movements/min; 95% CI 27.9–34.3; p=0.001, Mann–Whitney non-parametric test for two independent samples; Cohen’s d=3.1), but the frequency was not significantly different from that of infants who had a normal neurological outcome after neonatal stroke (p=0.79, Mann–Whitney test).
The frequency of digit movements was significantly different in the three groups for the contralesional hand (a mean value of the two hands was considered for healthy infants; Fig. 3; p<0.001, Kruskal–Wallis test) but not for the ipsilesional hand (Fig. 3; p<0.093, Kruskal–Wallis test). Infants who developed hemiplegia showed significantly lower frequencies of contralesional digit movements than infants with normal neurological outcome after neonatal stroke (mean difference 11 contralesional digit movements/min; 95% CI 08.6–13.4; p=0.008, Mann–Whitney non-parametric test for two independent samples; Cohen’s d=2.8) and healthy infants (mean difference 16.2 digit movements/min; 95% CI 14–18.4; p<0.001, Mann–Whitney test; Cohen’s d=3.5). A significant difference was also found between ipsilesional digit movements in infants with hemiplegia and healthy infants (mean difference 8.6 digit movements/min; 95% CI 6.1–11.2; p=0.04, Mann–Whitney test; Cohen’s d=1.3). The frequency of global hand movements did not differ in the three groups (Fig. S2, Supporting information published online).
Correlation between hand movement frequency and severity of hemiplegia
Within the cohort of infants with hemiplegia after neonatal stroke, those with mild and those with moderate hemiplegia did not show significant differences in frequency of upper-limb distal movement patterns or degree of asymmetry of segmental movements (Mann–Whitney test). Differences in wrist movement asymmetry only approached significance (p=0.057).
The results of the present study confirm our previous findings of a significant correlation between asymmetry of wrist movements at 3 months after birth and the presence of later hemiplegia in infants with neonatal stroke.4 As the methodology applied in the two studies was similar, and the patients previously assessed were not included in the present study, we can affirm that a significant correlation has been demonstrated in 25 infants with neonatal stroke. All 12 infants with asymmetrical wrist movements at fidgety age developed a hemiplegia, whereas 10 of 13 infants with symmetrical wrist movements had normal motor development. All three of the infants with symmetrical wrist movements who had abnormal neurological outcomes developed a very mild form of hemiplegia. A similar correlation between asymmetry of wrist movements at fidgety age and neuromotor outcome was also found in preterm infants with periventricular unilateral brain damage (usually venous infarction), suggesting that an asymmetry of motor function around 3 months after birth might be an early sign of hemiplegia of different causes and timings.3
A more detailed analysis of movements in the present study allowed us to identify new findings. For the first time, we have shown that in infants with neonatal stroke, hemiplegia can be predicted by an early reduction in the frequency of another type of distal upper-limb movement, that of the digits. In particular, none of the infants with abnormal outcome showed more than one digit movement per minute, with the exception of one infant. Some evidence suggests that independent digit movement might be related to the maturation of direct cortical motor control on spinal networks.1 Digit movements appear to follow a developmental pattern. For example, a gradual shift in behaviour has been reported, in which a pattern of movement in which the thumb first meets the side of the index finger (pre-precision grip) is followed by the thumb meeting the tip of the index finger (precision grip), in a phase in which these movements are not goal directed.1 Furthermore, recent data show that, starting from the first few days of life, infants are capable of imitating isolated finger movements that are differentiated from global hand-movement patterns, thus supporting the functional activity of the monosynaptic projections and extensive innervations of the spinal neurons at a very early stage of development.14,15 Our results further support this hypothesis.
When we compared the frequency of movements in the ipsilesional hand in infants with hemiplegia with the frequency in healthy children, no difference was found for wrist movements, whereas significant differences were found for digit movements. In particular, 3 of the 14 infants with abnormal neurological outcomes showed less than one digit movement per minute in the ipsilesional hand (see Fig. 3). Also, the frequency of wrist movements in the contralesional hand tended to correlate negatively with the severity of hemiplegia, whereas the frequency of digit movements did not. These findings suggest that the neurophysiological mechanisms underlying wrist movements are different from those of digit movements. Segmental wrist movements in the present study were defined as the motor behaviour observed as part of the fidgety pattern. This type of general movement is often considered to be the result of the activity of a central pattern generator, as it is endogenously generated, i.e. there is no detectable external stimulus, it is constant in form, and it is easily recognizable with a Gestaltic perception.16
It might be proposed that the abnormalities of hand movements observed in infants with neonatal stroke and later hemiplegia reflect complex neurophysiological mechanisms. The abnormal development of the corticospinal projection to the spinal cord and the subsequent abnormal refinement of connections are likely to be responsible for the impoverishment of complex motor patterns such as isolated finger movements and pre-precision and precision grips. At the same time, brain damage may alter the cortical modulation exerted on the central pattern generators of the fidgety pattern, resulting in a significant reduction of segmental wrist movements in the contralesional hand. It is not surprising that an abnormality of these two types of movement is predictive of later motor development as they are both the result of brain lesions affecting cortical motor control at some level.
The reason that isolated finger movements were also impaired in the ipsilesional hand in three of the 14 infants who developed hemiplegia is not straightforward. It could be related to a general delay in the development of finger movements resulting in a time-shift in the onset of complex motor patterns in the unaffected hand. Alternatively, it could be dependent on the bilateral process of plastic reorganization of cortical control, in terms of activity-dependent competition between the affected and unaffected hemisphere.17 In infants with unilateral brain damage the degree of later motor impairment depends not only on the extent of the acute loss of corticospinal projections, but also on a protracted process of competitive, activity-dependent refinement of the bilateral corticospinal projections from the affected and unaffected hemisphere, a process that at 3 months of life is still at its very beginning.17 In our study, no correlation was found between the severity of hemiplegia and the degree of impoverishment of distal movements of the contralesional upper limb.
Neither the asymmetry nor the actual amount of global hand movements was significantly different between the groups. The range of asymmetries was wider in the comparison group than in the study group: in four healthy infants global hand movements were observed in only one hand, either the dominant or the non-dominant hand. There is no clear explanation for this finding. It could be simply related to casual occurrence, as this type of movement is relatively infrequent (<3 movements/min on average in our full cohort). Larger samples are needed to clarify the possible nature of this result.
The major limitation of this study is the small number of patients assessed. Neonatal stroke is a rare disease and is often diagnosed after the first months of life, when the signs of motor impairment become obvious.18 Recruiting a reasonable number of individuals with a definite diagnosis of neonatal stroke and an homogeneous clinical follow-up thus requires multicentre collaboration and extended projects. Indeed further studies are needed to investigate the proposed hypotheses. A larger number of infants with a known pattern of sensorimotor reorganization, followed longitudinally at least for the first 2 years of life with standardized methodology, would shed light on the correlations between early hand function and outcomes such as the type of reorganization, the site and extension of the lesion, the response to environmental intervention, and the final motor outcome.
Another important limitation of this study is the use of a single time point for the assessment of hand motor function. This choice was based on our previous knowledge of the predictive value of qualitative assessments of general movements, performed at 3 months after birth in similar cohorts. Also, this time point is commonly included in the follow-up programmes of infants at neurological risk, as it is considered a time at which important developmental changes occur. Although global abnormalities of general movements (the poor repertoire and the cramped-synchronized patterns) are predictive from birth of abnormal motor development in infants with unilateral brain damage, no asymmetries are reliably detected before the onset of the fidgety period.3,4 Future work will tell us whether upper-limb digit movements, which were not specifically explored in the previous studies, can reliably predict motor outcome in infants with neonatal stoke even earlier than 3 months of age. It will also be of interest to compare these clinical findings in future studies with data on the anatomical and functional integrity of the corticospinal system obtainable using newer MRI techniques (e.g. diffusion tensor imaging) and electrophysiological techniques (e.g. motor evoked potentials).
In conclusion, our findings show that, at 3 months of age, infants with neonatal stroke may present with a relative impoverishment of the contralesional upper limb in terms of distal movements, which is always predictive of later hemiplegia. However, the degree of impoverishment or asymmetry is not correlated with the severity of motor impairment, suggesting that other factors play a significant role in the prediction of functional outcome at this age.17
- 11Prechtl’s Method on the Qualitative Assessment of General Movements in Preterm, Term and Young Infants. Clinics in Developmental Medicine No. 167. London: Mac Keith Press, 2005., , , , .
- 12Neurological Development in Infancy. London: Heinemann, 1976..
- 15Individuated finger movements: rejecting the labeled-line hypothesis. In: WingAM, HaggardP, FlanaganJR, editors. Hand and Brain. San Diego: Academic Press, 1996: 81–98..
Figure S1: Hand and digit movements: (a) global hand movements: opening; (b) global hand movements: closing; (c) segmental movements at the wrist: ulnar/radial flexion; (d) segmental movements at the wrist: rotation; (e) independent digit movements: precision grip; (f) independent digit movements: isolated movement of index finger.
Figure S2: Absolute frequency of global hand movements (movements/minute). Crosses indicate healthy infants (comparison group); open circles indicate infants with no hemiplegia after neonatal stroke; half-filled circles indicate infants with mild hemiplegia after neonatal stroke; filled circles indicate infants with moderate hemiplegia after neonatal stroke.
Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.
|DMCN_3497_sm_Suppl_Fig1.jpg||502K||Supporting info item|
|DMCN_3497_sm_Suppl_Fig2.jpg||32K||Supporting info item|
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.