Journal http://jcn.sagepub.

com/
of Child Neurology

Predictors of Independent Walking in Children With Spastic Diplegia

Ermellina Fedrizzi, Paola Facchin, Michela Marzaroli, Emanuela Pagliano, Gabriella Botteon, Luciana Percivalle and
Elisa Fazzi
J Child Neurol 2000 15: 228
DOI: 10.1177/088307380001500405

The online version of this article can be found at:

http://jcn.sagepub.com/content/15/4/228

Published by:

http://www.sagepublications.com

Additional services and information for Journal of Child Neurology can be found at:

Email Alerts: http://jcn.sagepub.com/cgi/alerts

Subscriptions: http://jcn.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Citations: http://jcn.sagepub.com/content/15/4/228.refs.html

>> Version of Record - Apr 1, 2000

What is This?

Downloaded from jcn.sagepub.com at UNIV OF UTAH SALT LAKE CITY on November 28, 2014
Original Article

Predictors of Independent Walking

in Children With Spastic Diplegia
Ermellina Fedrizzi, MD; Paola Facchin, MD; Michela Marzaroli, MD; Emanuela Pagliano, MD;
Gabriella Botteon, MD; Luciana Percivalle, PhT; Elisa Fazzi, MD

ABSTRACT

A prospective study was carried out to identify predictors of independent walking in 31 children with either spastic diple-
gia or triplegia, observed from the age of 9 to 18 months (mean, 11months) and followed for a mean period of 30 months
(range, 24 to 36 months). Mean age at most recent examination was 41 months (range, 36 to 54 months). We used an 18-item
scheme to chart the acquisition, from the prone position, of prelocomotor, sitting, and locomotor skills. Examinations
were conducted every 6 months and videotaped according to a standardized procedure. At latest assessment 18 (58%) of

the 31 children had achieved walking, 7 (23%) independently and 11 (35%) with assistance; 13 (42%) did not achieve walk-
ing. Ambulatory status was related to developmental quotient and visual acuity: all the children who became independent
walkers had normal visual acuity and in 86% of cases a normal general development quotient. Moreover, we found a sig-
nificant correlation between the number of gross motor skills achieved and the rate of achievement before 2 years of age
and ambulatory status at 3 to 5 years of age. Ability to put weight on the hands while prone and to roll from supine to prone
position by 18 months of age were significantly related to independent walking, while ability to sit without support was
predictive only at around 24 months of age. ( J Child Neurol 2000;15:228-234).

It is important to be able to predict whether children with
cerebral palsy will walk in order to plan appropriate ther-
apeutic and rehabilitation goals, and to inform the parents.
Spastic diplegia and triplegia, the clinical forms of cere-
bral palsy related to preterm birth, are increasing in fre-
quency in relation to the increase in such births and the
decrease in perinatal mortality. 1,2 However, the prognosis of
ambulatory outcome in these conditions is more difficult
than in other forms of cerebral palsy because of the variety
of factors that can influence ambulatory status during devel-
opment. Thus, while children with spastic hemiplegia almost
always learn to walk3-5 and those with spastic quadriplegia
Received March 1, 1999. Received revised April 5, 1999. Accepted for pub-
lication April 7, 1999.
From the Department of Developmental Neurology, C. Besta Neurological
Institute (Drs Fedrizzi, Marzaroli, Pagliano, Botteon, and Percivalle), Milan,
the Child Neuropsychiatry Division of the C. Mondino Neurological Insti-
tute (Dr Fazzi), Pavia, and the Department of Child Epidemiology of the
University of Padua (Dr Facchin), Padua, Italy.
Address correspondence to Dr Ermellina Fedrizzi, Department of Devel-
opmental Neurology, C. Besta Neurological Institute, via Celoria, 11, 20133
Milan, Italy. Tel: 39-022-394295; fax: 39-027-0635350; e-mail:
efedrizzi.besta@interbusiness.it.
rarely achieve ambulation,6 children with spastic diplegia
eventually achieve ambulation in 86% to 91% of cases,~9
and a similar percentage of spastic triplegia cases (87%)
acquired walking in the series of Campos da Paz.9
Most studies on predictive factors for walking have
been retrospective and involved children with all clinical
forms of cerebral palsy. Furthermore, no studies have
addressed the relationship between ambulatory status, cog-
nitive development, and visuoperceptual disorders in cere-
bral palsy. In a review of the literature spanning the past 50
years, Sala and Grant6 reported that primitive reflexes, pos-
tural reactions, gross motor skills, and the type of cerebral
palsy seemed to be the main factors in predicting the achieve-
ment of ambulation. More specifically, most studies indicate
that age at the achievement of sitting (defined as the abil-
ity to maintain the sitting position without support) is a
predictor of ambulatory status. Thus, in Badell-Ribera’s
prospective study7 of 50 children with spastic diplegia, all
of whom became independent walkers, sitting and crawl-
ing were achieved at between 18 and 30 months of age, and
Watt and coworkers5 reported a positive correlation between
sitting by the age of 2 years and walking at 8 years. Campos
da Paz et all investigated the role of age at attainment of
major motor milestones in 272 children with cerebral palsy

228

Downloaded from jcn.sagepub.com at UNIV OF UTAH SALT LAKE CITY on November 28, 2014

and found that achievement of head balance before
9 months, sitting by 24 months, and crawling by 30 months
were predictors of the achievement of ambulation.
The purpose of this prospective study was first to iden-
tify predictors of independent walking before 2 years of age
in children with spastic diplegia or triplegia using a functional
motor chart that analyzes the developmental uprighting
sequence from prone position and the prelocomotor skills,
and second, to study the relationship between ambulatory
status and other factors, such as cognitive development
and visual acuity.
PATIENTS AND METHODS
Patients
We enrolled patients with spastic diplegia or triplegia presenting
at the Department of Developmental Neurology, C. Besta Neuro-
logical Institute, Milan, or the Child Neuropsychiatry Division of
the C. Mondino Neurological Institute, Pavia, from 1993 to 1995.
Inclusion criteria were (1) clinical diagnosis of spastic diplegia
according to Hagberg and colleaguesl° or spastic triplegia accord-
ing to Michaelis et all (2) enrolment before the age of 2 years, and
(3) likelihood of availability for regular follow-up until at least the
age of 8 years. Forty-five patients conforming to the inclusion cri-
teria presented over the study period, but only 31 had a follow-up
of at least 24 months by December 1997 (mean, 30 months; range,
24 to 36); these 31 form the population of this study. Fifteen were
boys and 16 were girls. Mean gestational age was 29 weeks (range,
24 to 38); mean age at recruitment was 11 months (range, 9 to 18).
Mean age at most recent examination was 41 months (range, 36 to 54).
Ultrasonography revealed early brain damage in 27 patients:
periventricular leukomalacia in 26 cases and intraperiventricular
hemorrhage in one. Nineteen of the children had spastic diplegia
and 12 had spastic triplegia.
Achievement of Gross Motor Skills
The children were examined every 6 months and the development
of their motor skills was recorded using an 18-item protocol (Table
1), developed and validated in a previous sturdy. 12 The 18 items were
chosen to reflect the development sequence of acquisition of motor
skills in the normal child as described by Touwenl° and Hempel and
Touwen.’4 The number of motor skills attained at each examina-
tion was noted. For analysis, skills were grouped into four func-
tional classes (plus outcome class) as shown in Table 2, each class
indicating a progressive stage in the overall acquisition of motor
ability. Within this grouping, the sitting function was divided into
three levels: level 1, maintaining the sitting position using both arms;
level 2, maintaining the sitting position with one arm free, and level
3, maintaining the sitting position without arm support.
Each child was assessed at play in a standardized setting. Activ
ity was recorded with a hand-held video camera over 30 to 60 mint
utes, include frontal, sagittal anterior, and posterior views, noting~
the maximum skills attained. Ambulatory status, as determined a1t
the most recent follow-up examination, was classified as follows9 :
independent ambulation, ambulation only with assistance (sticks
crutches, or walkers), and ambulation not achieved.
Downloaded from jcn.sagepub.com at UNIV OF UTAH SALT LAKE CITY on November 28, 2014
229
All the children underwent physiotherapy according to a neu-
rodevelopmental approach, from the time of diagnosis and during
follow-up, attending the two centers involved in the research pro-
ject, with a mean frequency of three sessions per week.
Assessment of Cognitive Development
and Visual Function
At 2 years of age cognitive development was assessed as the gen-
eral quotient of the Griffiths Mental Developmental Scale, exclud-
ing the locomotor subscale. Each child also underwent an
ophthalmologic evaluation (fundus oculi, ocular motility, refraction)
and visual acuity assessment (Teller acuity cards; normal value,
more than 10th percentile).
Statistical Analysis
The nonparametric Kruskal-Wallis analysis of variance (ANOVA)
was used to determine whether the distribution of the number of
acquired motor functions differed in relation to ambulatory outcome
at the most recent examination. Linear, square, and log-linear
regression analyses were carried out to model the rate of acquisi-
tion of motor functions in relation to age. The best models were
analyzed and compared to each other. It was possible to describe
the typologies of functions related to age. With this aim the 18 func-
tions described in the protocol (Table 1) have been grouped into
four performance classes plus an outcome class. Each class includes
a series of functions related to ability and motor control in increas-
ing steps (Table 2). For each function class the progressive matu-
ration achieved at the different control ages by each subject,
grouped according to the outcome, has been described (Table 3).
Finally, two series of stepwise logistic analyses were per-
formed to determine the predictive values of the variables for two
outcomes (walking or not walking). In the first analysis the vari-
ables initially included were age, general development quotient, gen-
eral quotient subscale scores, the achievement of uprighting
sequences from the prone position, prelocomotor skills, sitting
without support, and the first two items of kneeling. In the second
logistic analysis, five motor items only were considered (rolling,
prone weight on hands, crawling, sitting without support, and
kneeling upright with support), and these were dichotomized in
terms of achievement before or after 18 months of age (for sitting,
before or after 24 months).
The aim of the logistic analyses was to obtain models that
allowed us to attribute an a priori probability of walking for each
patient and for each assessment. By attributing a negative prognosis
to each case with an a priori probability lower than 0.5 and a pos-
itive prognosis for each case with a probability equal to or greater
than 0.5, it was possible to formulate an a priori prognostic judg-
ment about the outcome. Comparison between the prognosis pro-
_
duced from these models and the observed outcome allowed
evaluation of the two models. These performances were described
-
in terms of efficiency,sensibility, and specificity of the models. To
-
evaluate the models’ extendability to new cases, and then to the
possibility of using them as a tool of clinical practice, a double val-
idation has been carried out. With this aim, 21 new patients, with
clinical characteristics (spastic diplegia, recruiting age, and age at
,
the end of follow-up) similar to the patients of the previous study
were evaluated and a prognosis was assigned based on our mod-
230
Table 1.
els by physicians blinded to the instrument we wanted to vali-
date. We therefore calculated the models’ efficiency in the prog-
nostic classification of new cases.
RESULTS
At latest assessment, 18 (58%) of the 31 children had achieved
walking, 7 (23%) independently and 11 (35%) with assis-
tance ; the other 13 (42%) did not learn to walk. The age at
walking was 2 years in 2 cases, 3 years in six cases, 4 years
in seven cases, and 5 years in three cases. Visual function
was severely compromised (less than 10th percentile) in 9
children and normal in the remaining 22. Cognitive devel-
opment was assessed (as Griffiths general quotient) in 29
of 31 children; in 15 (52%) the general quotient was normal
(greater than 85), in 6 (21%) it was borderline (71 to 84), and
in 8 (28%) it was below normal (less than 70).
Ambulation in Relation toCognitive
Development and Visual Impairment
As shown in Figure lA, most (86%) of the children who
achieved independent walking had normal general quo-
tients on the Griffiths Scale, while in children walking with
assistance and in those not walking the general quotient was
normal in 55% and in 27% of cases, respectively. All children
Downloaded from jcn.sagepub.com at UNIV OF UTAH SALT LAKE CITY on November 28, 2014
Functional Assessment Chart
who achieved autonomous walking had normal visual acu-
ity, while 46% (6 of 13) of children who did not achieve
walking had defective visual acuity (Figure 1B).
Ambulation in Relation to Achievement
of Early Motor Skills
The relation between acquisition of motor skills and achieve-
ment of walking was analyzed both in terms of the total num-
ber and types of motor skills acquired.
Number of Motor Skills
Figure 2 shows the number of motor skills acquired with
advancing age for the three outcome groups (independent
walking, walking with assistance, and no walking). The
children who did not walk achieved fewer skills than those
who did. This difference was evident from the earliest
assessment, and among those who did not walk, the num-
ber of skills learned increased little with the passage of
time to reach a plateau of approximately five skills. Among
the children who achieved walking, those who walked inde-
pendently achieved more motor skills at a faster rate than
those who learned to walk with assistance. However, the dif-
ference between these two latter groups was rather small
compared to the difference between these groups and the
nonwalkers. Kruskal-Wallis ANOVA showed a significant dif-

Table 2.
ference (P =
.004) between the three outcome groups in
terms of the number of skills attained.
Regarding the trend of the number of functions achieved
by each subject in the three different groups of autonomous
walking, walking with assistance, and no walking (Figure
3) we found an extremely different trend in subjects who
could not walk compared to those who achieved walking
(with assistance or independently). Subjects who could not
walk seem to have a very uniform trend: they reached or sur-
passed six items in only two cases. Among those who
achieve walking, the subjects who walk independently have
rapid growth in the achievement of gross motor skills within
the first 24 months. The number of functions achieved tends
to stabilize after 36 months. This trend is apparent in all
cases. Among the subjects who walk with assistance there
is the highest variability both in the rate of increasing func-
tions or in the age at which this occurs, and in the total num-
ber of functions achieved.
Table 3. Typology of Motor Skills and Ambulatory Status
Downloaded from jcn.sagepub.com at UNIV OF UTAH SALT LAKE CITY on November 28, 2014
231
Functional Classes
The regression models that best approximate the empir-
ical data are shown in Table 4. The log linear model provided
the best overall correlation in all three outcome groups, with
Pearson’s r 0.87 in the independent walking group and
=
r = 0.65 in the other two outcome groups.
Types of Motor Skills Achieved
The results obtained with motor skills grouped into four
classes are shown in Table 3. It was found that none of the
nonwalking children were able to perform all uprighting
sequences from a prone position before 18 months of age,
while most (71%) of the children who became independent
walkers had completed the sequence (Item 6) by 18 months
of age and all of them by 24 months. Among the preloco-
motor skills, rolling was acquired by all independent walk-
ers by 18 months, while the acquisition of creeping occurs
later, by 24 months. All except one of the children who
achieved independent walking also achieved level 3 sitting
232

Figure1. Ambulatory status in relation to (A) cognitive develop-

ment and (B) visual acuity.

(no arm support) by 24 months; the exceptional case

achieved level 3 sitting by 36 months. Only 55% of the chil-
dren who achieved ambulation with assistance achieved level
3 sitting by age 24 months. Only one of the nonwalking
children had achieved level 3 sitting by the latest follow-up

Figure 3. Trend of the number of functions achieved by each sub-

ject belonging to the three different groups of outcome.

examination. Similarly, all the children who walked inde-

pendently were able to sit on their heels by 30 months,
while only one of the nonwalking children was able to do
this by 30 months.
Therefore, for the uprighting and prelocomotor
sequences from the prone position, 18 months seems to be
a good cut-off age to distinguish potential walkers (inde-

pendent or with assistance) from those who will not walk.

Figure 2. Ambulatory status
skills
acquisitions and age.
in relation to mean number of motor
As far as the sitting posture and kneeling are concerned, a
better cut-off age seems to be 24 months. Table 5 shows the
two models obtained by logistic regression analysis that best
predicted walking from the independent variables consid-
ered before subjects reach 2 years of age. A model with only
two variables (putting weight on hands while prone and over-

Downloaded from jcn.sagepub.com at UNIV OF UTAH SALT LAKE CITY on November 28, 2014

Table 4. Trend Analysis of Number of
Functions by Age and Outcome*
*Regression models analysis.
all general quotient) correctly predicted walking status in
96% of cases with 93% sensitivity and 100% specificity. A sim-
ilar result was obtained using a model including rolling
before 18 months, putting weight on hands when prone, and
the oculomotor subscale score of the general development
quotient. This model predicted walking status in 94% of the
children with a sensitivity of 96% and a specificity of 91%.
DISCUSSION
Many professionals in clinical practice develop their own
ad hoc criteria for predicting ambulatory status in children
with cerebral palsy. These criteria might provide reasonable
prognostic accuracy, but are not necessarily transferable to
other contexts, and their a priori errors cannot be assessed.
It is therefore desirable to develop assessment instruments
of known validity, transferability, reliability, and limitation,
in order to provide a uniform approach to prognosis and to
give the families of these children a more reliable prediction
of future development. This was the purpose of the present
study. We used a simple 18-item instrument, developed at
our center, 12 for the regular assessment of achieved gross
motor skills in relation to age. More complex instruments,
such as the Gross Motor Functional Measure or Gross Motor
Performances Measure 15 give too much information and
are used more as tools for evaluation than for prediction.
We found a significant correlation between the rate of
attainment of motor milestones and ambulatory status. Pre-
Table 5. Logistic Models to Predict Walking
Downloaded from jcn.sagepub.com at UNIV OF UTAH SALT LAKE CITY on November 28, 2014
233
vious studies7.9 have reported that the earlier the achieve-
ment of milestones, the better the prognosis for ambulation,
either independently or with assistance. We found that the
rate of acquisition of gross motor skills in our diplegic and
triplegic children differed significantly by Kruskal-Wallis
ANOVA, in the three outcome groups. Those who became
independent walkers were characterized, as a group, by
rapid and uniform accumulation of early milestones before
30 months of age (Figure 3A), while those who did not
achieve walking were characterized by very slow acquisi-
tion of these skills between the ages of 18 and 30 months
(Figure 3C). This indicates that the period up to 30 months
is critical for determining ambulatory potential in these
children, and it is therefore important clinically for defin-
ing, within this age, the walking prognosis and consequently
for planning therapeutic goals. By contrast, in children who
will achieve walking, the prognosis of the future level of
ambulation-independent or with assistance-appeared
more uncertain in patients less than 5 years of age.
In fact, the children who walked with assistance
(between 3 and 5 years of age) showed variable trends: we
found children who acquired motor milestones at a rate sim-
ilar to that of the independent walkers, and children who
showed a rate of acquisition similar to that of the nonam-
bulatory group. We expect that by the time these children
reach 8 years of age they will have provided more infor-
mation to enable prediction, at between 3 and 5 years, of the
type of ambulation they achieve.
As also reported in previous studies,5,7,9 we found that
the acquisition of specific motor skills at an early age was
more predictive of future ambulatory status than the total
number of functions acquired at a given age. Thus, acqui-
sition of head and chest raising while prone, with the aid of
arm support (Item 6) and rolling from supine to prone posi-
tion (Item 4) before 18 months of age predicted independent
walking in a high percentage (71% and 100%, respectively)
of our children at between 3 and 5 years of age. Achievement
of Item 6 before 18 months was therefore an earlier indicator
of independent walking than achievement of the sitting
position, which was acquired by 24 months only in the inde-
pendent walkers, and is in accordance with the findings of
Campos da Paz and coworkers,9 who reported achieve-
ment of head balance by 9 months of age as the first mile-
stone prognostic for autonomous walking. This finding is
important for prognosis as well as treatment. Early ability
to hold the head and chest upright against gravity, to put
weight on the hands with elbows extended, and to rotate the
body from supine to prone are indicative, in diplegic and
triplegic patients, of a variety of motor behavior involving
the trunk and shoulders, which are less compromised than
the pelvis and legs, and provide balance and stability when
shifting weight from one side to the other. These compe-
tencies in uprighting sequences from the prone position
and in rolling are prerequisites for independent sitting and
independent walking: if they are acquired before 18 months
they can be defined as earlier predictors of ambulation than
sitting by 2 years of age. We also found a relationship
234
between ambulatory status at age 3 to 5 years, general
development quotient, and visual acuity, in agreement with
the findings of Watt et all in their prospective study. The asso-
ciation between severity of motor involvement, delay in
cognitive development, and visuoperceptual disorders could
be related to periventricular leukomalacia, the typical lesion
of spastic diplegia. Furthermore, the development of motor,
cognitive, and visual skills is very closely linked at an early
age, and the development of behavior reflects the com-
plexity of all these interacting factors.
These considerations could explain the results obtained
from our logistic regression analysis, which indicated that
ability to achieve walking can be predicted before the age
of 2 years from two variables only: ability to lift the head and
chest against gravity, using the hands as support, and the ocu-
lomotor subscale score of the Griffiths Developmental
Scale. Evidently the acquisition of these skills is a very
early indicator of the eventual acquisition of other skills,
which come to bear on the achievement of independent
walking. In that case the predictive value of later motor skills,
such as sitting and crawling on hands and knees, becomes
superfluous, because they are already encompassed by the
previously attained skills.
APPENDIX
The logistic analysis studies the trend of the dependent
variable on the basis of a series of explanatory or predic-
tor variables. The response variable is dichotomic (present
or absent), while predictors can be dichotomic or qualita-
tive, with several modalities, or quantitative. The estimate
of parameters or selected predictors is obtained by the
&dquo;maximum likelihood&dquo; method, which provides, through
successive computation iterations, the estimates with the
maximum probability of approaching observed data. The
model resulting from the analysis is a linear combination of
the same explanatory variables, selected using a coeffi-
cient ; this linear combination yields the a priori probabil-
ity of the response variable for each subject.
Since the explanatory variables introduced in the analy-
sis, and then the selected predictors, can be dichotomic,
quantitative, or qualitative, the weight of each predictor
changes depending on values assumed by the predictor
itself. If the predictor is dichotomic (0 or 1), its maximum
contribution to the algorithm will be described by its coef-
ficient (or parameter estimated) because it will be multiplied
by 0 or 1; if the variable is quantitative its maximum con-
tribution will depend on its measurement scale. A more
explanatory comparison of the weight assumed by each
variable in the model can be obtained by the standardiza-
tion of coefficients.
In addition to the significance of every parameter in the
model, it is also possible to evaluate the overall signifi-
cance of the model. With the intent of selecting the most effi-
cient and economic model (which allowed the use of as little
Downloaded from jcn.sagepub.com at UNIV OF UTAH SALT LAKE CITY on November 28, 2014
information or as few predictors as possible), the stepwise
method has been used. It consists of analyzing one explana-
tory variable at a time, selected because it explains most of
the variability compared to the previous step (or to the use
of the intercept alone if it is the first variable taken into con-
sideration). At every new step of the analysis the estimates
of the previous parameters are reassessed and their signif-
icance is redefined. There can be strengthening, or divi-
sion, of the predicting power among different parameters.
The analyst can define, through a series of evaluations, a cut-
off of significance that allows the inclusion or exclusion of
the predictors. The analysis will end when none of the new
variables reaches the cut-off.
Acknowledgment
This work was supported by grants from the Fondazione Pierfranco e Luisa Mar-
iani, Milan, Italy.
References
1. Hagberg B: Nosology and classification of cerebral palsy. G Neu-
ropsich Evol 1989;4(Suppl):S12-S17.
2. Veelken N, Schopf M, Damman O, Schulte FJ: Etiological classi-
fication of cerebral palsy in very low birthweight infants. Neuro-
pediatrics 1993;24:74-76.
3. Crothers B, Paine RS: The Natural History of Cerebral Palsy, 2nd
ed. London, McKeith Press, 1988.
4. Bleck EE: Locomotor prognosis in cerebral palsy. Dev Med Child
Neurol 1975;17:18-25.
5. Watt JM, Robertson CMT, Grace MGA: Early prognosis for ambu-
lation of neonatal intensive care survivors with cerebral palsy. Dev
Med Child Neurol 1989;31:766-773.
6. Sala D, Grant AD: Prognosis for ambulation in cerebral palsy. Dev
Med Child Neurol 1995;37:1020-1026.
7. Badell-Ribera A: Cerebral palsy: Postural locomotor prognosis in
spastic diplegia. Arch Phys Med Rehabil 1995;66:614-619.
8. Trahan J, Marcoux S: Factors associated with the inability of
children with cerebral palsy to walk at six years: A retrospective
study. Dev Med Child Neurol 1994;36:787-795.
9. Campos da Paz A, Burnett SM, Braga LW: Walking prognosis in
cerebral palsy: A 22 year retrospective analysis. Dev Med Child
1994;36:130-134.
Neurol
10. Hagberg B, Hagberg G, Olow I: The changing panorama of cere-
bral palsy in Sweden. II. Analysis of various syndromes. Acta
Paediatr Scand 1975;64:193-200.
11. Michaelis R, Edebol-Tysk K: New aetiopathological and noso-
logical aspects of cerebral palsy syndromes. G Neuropsich Evol
1989;4(Suppl):S25-S30.
12. Fedrizzi E, Oleari G, Botteon G, et al: Motor performance assess-
ment in children with cerebral palsy, in Fedrizzi E, Avanzini G,
Crenna P: Motor Development in Children. London, John Libbey,
1993, pp 51-58.
13. Touwen BCL: Neurological development in infancy. Clin Dev
Med 1976;58:183-189.
14. Hempel MS, Touwen BCL: The Neurological Examination for Tod-
dler Age, doctoral diss. Groningen, Netherlands, Rijksuniversiteit
Groningen, 1993.
15. Boyce WF, Gowland C, Rosenbaum PL, et al: Measuring quality
of movement in cerebral palsy: A review of instruments. Phys Ther
1991;71:813-819.