Cognitive status,
language attainment,
and prereading skills
of 6-year-old very
preterm children
and their peers: the
Bavarian Longitudinal
Study
Dieter Wolke* PhD Dipl-Psych;
Renate Meyer Dipl-Psych; University of Munich Children’s
Hospital, Bavarian Longitudinal Study, Munich, Germany.
*Correspondence to first author at University of Hertfordshire,
Faculty of Health and Human Sciences, Department of
Psychology, College Lane, Hatfield AL10 9AB, UK.
The prevalence of intellectual-, language-, and prereadingskill deficits was investigated in a geographically defined
whole-population sample of very preterm children at 6 years
of age in southern Germany. The sample consisted of the
following: 264 very preterm children (75.6% of Germanspeaking survivors); 264 term controls (matched for sex,
socioeconomic status [SES], marital status and age of
mother); and a representative normative sample for Bavaria
(N=311). Compared with term peers, very preterm children
scored significantly lower (approximately –1 SD) on the
measures of cognitive and language skills and had major
cognitive deficits (<–2 SD) 10 to 35 times more often than the
controls. Deficits in speech articulation and prereading skills
(<10th centile) were three to five times more frequent in very
preterm children. More than 18% of very preterm children
had cognitive deficits in more than five areas of functioning,
compared with no control children. The differences between
very preterm children and controls remained highly
significant when only very preterm children (N= 229) and
controls (N= 261) without major neurosensory impairment
were considered. Little evidence for specific cognitive deficits
was found once mental processing measured in the Kaufman
Assessment Battery for Children (K-ABC) was controlled for.
The effect of preterm birth on cognitive abilities was found to
be larger than the influence of SES. In conclusion, there was a
high prevalence of long-term multiple cognitive problems in
very preterm children. These persistent cognitive problems
appear to be of pre- or neonatal (treatment) rather than
postnatal social origin.
94 Developmental Medicine & Child Neurology 1999, 41: 94–109
Meta-analyses of reports on the intellectual development of
very preterm or very-low-birthweight infants and children
have concluded that their IQs are usually within the normal
range but approximately 0.5 SD below those of term infants
(Aylward and Pfeiffer 1989, Aylward et al. 1989, Escobar et al.
1991, Ornstein et al. 1991, Wolke 1993, Hall et al. 1995).
Analysis of IQ scores have found significant developmental
delay or learning difficulties are more common in very-lowbirthweight children (Aylward et al. 1989; Wolke 1991,
1998a). A greater variability of test scores or discrepancies
between Verbal and Performance IQs have also been noted
by some researchers (Vohr and Garcia Coll 1985, Breslau et
al. 1988, Ornstein et al. 1991) but not others (Li et al. 1990).
Other developmental disorders such as language delays
and deficits (Rocissano and Yatchmink 1983, Wright et al.
1983, Byers et al. 1986, Vohr et al. 1988, Lewis and Bendersky
1989, Largo et al. 1990), articulation problems (Largo et al.
1990), and specific learning disorders such as poor reading,
writing, and numerical skills and mathematics problems
(Klein et al. 1989, Ross et al. 1991) have also been noted to be
more common in very-low-birthweight children (Cohen et
al. 1988, McCormick 1989, Friedman and Sigman 1992,
Wolke 1993). It has been speculated that specific deficits may
suggest either damage to, or inhibition of, normal development in specific areas of the brain (Mutch et al. 1993).
However, only a few previous studies have evaluated
whether these developmental or learning problems represent specific developmental deficits or could be accounted
for by general IQ deficits in very-low-birthweight children
(Lindgren et al. 1986, Breslau et al. 1988, Hunt et al. 1988,
Klein et al. 1989, Saigal 1991, Saigal et al. 1991).
There is thus still considerable uncertainty regarding the
cognitive outcome of very-low-birthweight children. Definite
conclusions are difficult to draw owing to shortcomings of earlier studies: many previous investigations included no control
groups (Saigal 1991, Saigal et al. 1991, Gross et al. 1992); often
employed variable and arbitrary criteria to define significant
cognitive impairment (Wolke et al. 1994); were single-centre
studies, with epidemiological studies being the exception (Bax
1983, Saigal et al. 1991, Veen et al. 1991, Robertson et al. 1992,
Scottish Low Birthweight Study Group 1992, Johnson et al.
1993, Wolke et al. 1994); had no or poor documentation of
those lost to follow-up (Aylward et al. 1986, Wariyar and
Richmond 1989, Leonard et al. 1990, Halsey et al. 1993, Wolke
et al. 1995b); and generally included only short follow-up periods ( Aylward and Pfeiffer 1989; Aylward et al. 1989; Collin et al.
1991; Wolke 1991, 1993; McCormick 1992; McCormick et al.
1992; Lukeman and Melvin 1993). Lack of attention to study
design and methodological factors may have resulted in large
underestimations of the adverse cognitive outcome of verylow-birthweight infants (Escobar et al. 1991, Ens-Dokkum et al.
1992, Gross et al. 1992, Wolke et al. 1994, Wolke 1997).
This study investigated the intellectual abilities, language
comprehension and expression, and prereading skills in a
population of very preterm children, matched term controls,
and a representative sample of children of the same cohort
living in a geographically defined area in the south of
Germany. We report the incidence of cognitive impairment
of very preterm children in comparison to a matched control
group. It was hypothesized that very preterm children usually have multiple cognitive problems and that specific developmental deficits are rare in very preterm children.
Method
THE BAVARIAN LONGITUDINAL STUDY: COHORTS AND DESIGN
Details of the design of the Bavarian Longitudinal Study
(BLS) have been described elsewhere (Wolke et al. 1994,
1995a,b; Riegel et al. 1995; Wolke 1997a,b) and are only
briefly outlined here. All infants born alive in a geographically defined region in south Bavaria between 1st February 1985
and 31st March 1986 and who required admission to one of
17 children’s hospitals in this region within the first 10 days
of birth comprised the target sample (index children).
During the study period, 70600 births were registered in the
region. The inclusion criterion was met by 7505 children
(10.6% of all births). The population ranged from very ill
preterm infants to term infants who required only brief inpatient observation for diagnostic purposes in the special
care units. In addition, 916 healthy infants receiving care on
the normal postnatal wards in 12 obstetric units in the same
hospital centre or adjacent to the children’s hospital were
recruited as control infants during the same period.
POPULATIONS
The following populations were assessed at 6 years 3 months
of age and are the subjects of this report.
Very preterm children
Of the 7505 children, 560 were very preterm infants with a gestation of <32 weeks. This comprised virtually all (>99%) very
preterm infants born during this period in this region
(Bayerische Landesärztekammer 1985). Of the 560 very
preterm children, 158 died during the initial hospitalization,
seven died between discharge from hospital and 6 years 3
months of age, four parents gave no written informed consent
for participation in the study, and 42 parents and children were
non-German speaking. The latter group of children had been
followed-up until the age of 4 years 8 months but were excluded
at 6 years 3 months because the language differences precluded
accurate testing of verbal abilities, language and prereading
skills. The verbal and language assessments at 4 years 8 months
had indicated that 80% of these children scored <–2 SD (Riegel
et al. 1995). The potential sample of survivors thus comprised
349 very preterm children. Of these, 264 very preterm children
(75.6%) were assessed at 6 years 3 months of age. Of the 85 children not assessed at 6 years 3 months, three could not be traced
and therefore no information on their development was available. Seven families could only ever be interviewed on the telephone; their physicians’ reports were obtained but they did not
attend any of the follow-up assessments, thus limiting the information on their development. Twenty-three children had only
one previous assessment, 26 had two previous assessments,
and 26 had all previous neurodevelopmental follow-up assessments at 5, 20, and 56 months of age.
Control group
Seven-hundred and eighteen of the 916 originally enrolled
non-referred children (79% survivors; five infants had died
between birth and 4 years 8 months) attended all previous
follow-up assessments at 5, 20, and 56 months (Riegel et al.
1995, Wolke et al. 1995a). Of these 718 children, 689 were
term (gestational age >36 weeks at birth). From this sample
pool of term children, a comparison group of 264 children
was group-matched to the 264 very preterm children (stratified random sampling) according to sex of child, family
socioeconomic status (SES), marital status of the parents,
and maternal age.
The characteristics of the very preterm children and their
controls are shown in Table I.
Normative sample
A normative sample (NC), representative of the total population of Bavarian infants born in 1985, was drawn from the
complete BLS sample. Five stratification variables for constituting a representative sample were available from the
Statistical Yearbook 1985 for Bavaria (Bayerisches Landesamt
für Statistik und Datenerhebung 1986) and the Bavarian
Perinatal and Neonatal Survey 1985 (Bayerische
Landesärztekammer 1985). These were sex distribution of
newborn infants (51% male); size of the community the parents were inhabiting (>50000 inhabitants: 27%); educational
level of mothers giving birth in 1985 (basic education: 9 years
or less of schooling or no basic educational qualification,
14%; moderate education:10 to 12 years’ schooling with
passed final exams/job related training, 76%; completed high
school or university education: successfully completed education of 13 years or more, 10%); gestation at birth (<32
weeks’: 0.9%, 32 to 36 weeks’, 6%); and whether the infant
had been admitted to a children’s hospital within the first 10
days of life (10.6%). There were 311 children in the normative
sample. The normative sample (see below) was solely studied
to determine cohort-specific and regional norms (i.e. means,
SDs, cut-off points for determining cognitive impairment).
MEDICAL AND PREVIOUS FOLLOW- UP DATA
Prenatal data were obtained from the medical histories in the
obstetric units, while peri- and neonatal data were collected
prospectively. The information was summarized into four
scales: prepregnancy (eight items), pregnancy (14 items),
birth (15 items), and neonatal complications (20 items). This
is similar to previously proposed optimality scoring systems
(e.g. Prechtl 1967, Kyllerman and Hagberg 1983, Molfese
and Thomson 1985, Michaelis and Haas 1987, St JamesRoberts and Wolke 1989) (see Riegel et al. 1995, Wolke et al.
1995a). Daily assessments of level of care, respiratory support, and feeding dependency, and neurological status such
as mobility, muscle tone, and neurological excitability were
conducted from the first day after birth. Each of the six variables was scored daily on 4-point rating scales (0 to 3)
according to Casaer and Eggermont (1985). The ‘Duration of
Treatment Index’ was computed as the number of days until
the infant reached a stable clinical state (total daily scores <3
for 3 consecutive days), and the ‘Intensity of Neonatal
Treatment Index’ was computed as the mean of daily ratings
during the first 10 days of life or until a stable clinical state
was reached, depending on which occurred sooner, multiplied by 100. Gestational age was determined from maternal
dates of the last menstrual period and serial ultrasounds during pregnancy. Additionally, all infants had a Dubowitz examination of gestational age (Dubowitz and Dubowitz 1979).
Where gestational age estimates from these methods differed by less than 2 weeks, maternal dates were used; where
the difference between the three methods was greater than 2
weeks (e.g. with uncertain dates), then the Dubowitz examination results alone were used to determine gestational age.
Parents were approached within 48 hours of the infant’s
hospital admission. The aims of the study were explained to
Cognitive Status of 6-Year-Old Very Preterm Children Dieter Wolke and Renate Meyer
95
them and they were asked to give written informed consent for
their child to participate. Socioeconomic data were obtained
by standard interviews with the infants’ parents in the first 10
days of life and at each follow-up visit. SES was computed as a
weighted composite score of maternal highest educational
qualification, paternal highest educational qualification, and
occupation of the head of family according to Bauer (1988). As
part of a full neurological and psychological assessment programme, the cognitive development was assessed at 5 and 20
months of age corrected for prematurity with the Griffiths
Scales of Babies Abilities (Brandt 1983) and the Columbia
Mental Maturity Scales (CMM; Burgemeister et al. 1954, Bondy
et al. 1975) at 4 years 8 months chronological age.
SIX-YEAR ASSESSMENTS
The aim was to assess all children at 6 years 3 months (before
beginning primary school in Germany). As shown in Table I,
more than 94% of children were preschool at the time of
assessment and those who had entered school had less than 3
months of schooling. Somatic and neurological assessments
(with outcome presented below) were conducted by specially
employed and trained paediatricians. Relevant information
for current purposes were diagnosis of severe impairment,
resulting in severe disability (Riegel et al. 1995): unchanged
cerebral palsy (CP) diagnosis between 56 and 75 months of
age (CP grade 3, no unaided walking but rolling or crawling
possible; CP grade 4, no active mobility), major malformations
Table I: Background information on the very preterm children, very preterm children who dropped out, and term control children
Control
children
N=264
Very preterm
children
N=264
Very preterm
drop-out children
N=85
Control
children
versus very
preterm
children
Very
preterm
versus
dropout
children
3407 (3351–3463)
–
–
1288 (1247–1330)
47.7
23.9
1350 (1282–1419)
52.9
12.9
<0.001
–
–
–
–
–
39.6 (39.5–39.7)
–
29.5 (29.3–29.7)
22.7
29.8(29.5–30.1)
16.5
<0.001
<0.001
–
–
Small for dates (<10th centile) (%)
9.8
28.8
25.9
<0.001
–
Multiple births (twins and more) (%)
4.2
23.5
21.2
<0.001
–
–
–
–
Obstetric/neonatal variables
Mean birthweight (g) (95% CI)
<1000–1499 g (VLBW) (%)
<1000 g (ELBW) (%)
Mean gestation (wk) (95% CI)
≤ 28 wk (%)
Parity
1
2
≥3
59.1
31.8
9.1
51.7
34.2
14.1
49.4
30.6
20.0
–
–
–
–
–
56.8
56.5
–
–
1.1 (1.0–1.1)
0.9 (0.8–1.0)
2.2 (2.0–2.4)
0.4 (0.3–0.5)
1.4 (1.3–1.5)
2.3 (2.2–2.5)
4.6 (4.4–4.7)
9.8 (9.5–10.1)
1.2 (1.0–1.4)
2.2 (1.9–2.4)
4.1 (3.8–4.4)
9.5 (9.0–10.0)
<0.001
<0.001
<0.001
<0.001
–
–
<0.01
–
Mean duration of treatment index (95% CI)
–
63.5 (58.8–68.2)
60.9 (52.0–69.7)
–
–
Mean intensity of treatment index (95% CI)
– 1268.7 (1224.6–1312.7)
1175.5 (1099.4–1251.6)
–
<0.05
Mean duration of initial hospitalization (95% CI)
–
84.0 (79.3–88.8)
74.7 (66.6– 82.7)
–
–
Outborn
Complication scores (Mean 95% CI)
Prepregnancy
Pregnancy
Birth
Neonatal
Demographic variables
Child’s sex
Male (%)
56.4
56.4
49.4
–
–
Socioeconomic status (%)
High
Middle
Low
22.3
42.4
35.2
22.3
42.4
35.2
13.3
41.0
45.8
–
–
–
–
–
–
Maternal educational qualification (%)
Basic (≤10 y)
Moderate (work qualification ≤12 y)
High school/university (≤13 y)
10.2
72.0
17.8
9.1
73.9
17.0
14.5
72.3
13.3
–
–
–
–
–
–
Marital status
Single (%)
11.7
16.5
18.3
–
–
Mean maternal age (at birth ) (95% CI)
28.5 (27.9–29.1)
28.8 (28.2–29.4)
27.9 (26.8–29.0)
–
–
Mean child’s age (mo) (95% CI)
74.2 (73.9–74.5)
75.0 (74.7–75.4)
–
<0.001
–
5.7
6.1
–
–
–
Started school (%)
VLBW, very low birthweight; ELBW, extremely low birthweight.
96 Developmental Medicine & Child Neurology 1999, 41: 94–109
or genetic syndromes (e.g. Down syndrome), blindness (noncorrectable loss of sight in at least one eye), or deafness.
All cognitive assessments were carried out by postgraduate clinical psychologists and their psychometric assistants in
specially equipped assessment rooms attached to the children’s hospitals at six sites in south Bavaria (Munich,
Augsburg, Rosenheim, Regensburg, Landshut, and
Deggendorf). All investigators were trained in the various
assessment procedures during a 3-month period before the
main study. Additionally, a pilot study was conducted under
study conditions with 74 children. The investigators were
unaware of the infants’ neonatal course and previous test
results. All assessments during the study period were videotaped and regularly checked by a specially employed supervisor (experienced clinical psychologist) and the first author.
COGNITIVE STATUS
The children were assessed with the German version of the
Kaufman Assessment Battery for Children (K-ABC; Kaufman
and Kaufman 1983, Melchers and Preuss 1991). This cognitive assessment battery is based on neuropsychological and
information-processing theories of intellectual functioning.
Intelligence is measured with the Mental Processing
Composite (MPC; eight subtests), designed to test fundamental mental processes. The MPC is subdivided in two further subscores (also separately standardized). These are the
simultaneous information-processing score (SGD; five subtests requiring the processing of several stimuli at the same
time, such as arranging complex patterns from several pieces)
and the sequential information-processing score (SED; three
subtests requiring the processing of individual stimuli one by
one, such as repeating numbers or repeating hand movements). The Achievement Score (AS) included three subtests designed to measure what has been learned by the
child. Each of the subscores (MPC, SGD, SED, AS) are standardized at 100±15 (Melchers and Preuss 1991). The K-ABC
is being used in an increasing number of very-low-birthweight studies (Li et al. 1990, Teplin et al. 1991, Achenbach et
al. 1993, Weisglas-Kuperus et al. 1993).
LANGUAGE DEVELOPMENT
Four subtests were used from the Heidelberger Sprachentwicklungstest (HSET; Grimm and Schöler 1991), a German
test battery for language development. These included the following subscales: plural-singular-rules (PS, test of grammatical
rules), correction of semantically inconsistent sentences (KS,
test of language comprehension), sentence production (SB,
test of language comprehension and production), and understanding of grammatical structures (VS, test of grammatical
rules). Each subtest is standardized as T-scores (50±10; Grimm
and Schöler 1991), and a total score of the four subtests was
computed standardized as T-scores (HSET total).
A new articulation test was developed with a total of 34
items. The phonemes (letters) previously found to be most
difficult for children at 6 years of age (Largo et al. 1990) were
tested. The child was presented with drawings of objects and
required to name them. The child’s formation of the target
sound was noted as correct or incorrect. There were three
categories for incorrect sound production: sound replaced
by another sound, left-out sound, or no reply. If a child did
not recognize the picture, the word was spoken by the investigator and the child required to repeat it (with the response
noted as correct or incorrect). Thirteen items had the target
letter (sound) at the beginning of the word (e.g. Tisch [table])
and 13 in the middle of the word (e.g. Leiter [ladder]). A further nine items tested specifically for ‘Sch’/ ‘St’/ ‘Sp’phonemes which are most difficult for 6-year-old German
speakers (Largo et al. 1990). The number of correct answers
was totalled to gain three subscores (word beginning, word
middle, ‘Sch’ phonemes) and a total score across all subtests.
Quality of speech and grammatical correctness (which
also included interviews with the children) was judged at the
end of the assessment day by the research team (consensus
ratings). The quality of speech index consisted of five items:
intelligibility of speech, voice quality, melody of speech, volume modulation of speech, and tempo. The items had one
normal category and one to five alternative ‘deviant’ categories. The number of non-deviant codings was totalled.
Grammatical correctness was judged according to seven
items: articles omitted, errors in article use, declension or
conjugation omitted, wrong declension or conjugation,
wrong sentence structure, prepositions omitted, wrong
prepositions. Each of the items was rated according to three
categories: never or rarely, occasionally, and often or always.
The number of never or rarely ratings was totalled. The
research teams were required to judge the reliability (very
good, good, poor, completely inaccurate) of their ratings for
each child and to give reasons for poor reliability (no talking
during assessments, no articulated language). Furthermore,
each research team recorded whether the child and the parents were speaking High German (received pronunciation,
‘Hochdeutsch’) or local dialect.
PREREADING SKILLS
The ability to recognize and categorize sounds has been
found to be predictive of later reading skills and difficulties
(Bradley and Bryant 1978, 1983; Velluntino and Scanlon
1987; Goswami 1990; Bowey et al. 1992; Schneider 1993),
even if controlled for social and environmental factors (Raz
and Bryant 1990). The phone-oddity task (Bradley and
Bryant 1983) involves rhymes in which three of four words
share a common phoneme and the child has to detect the
odd word. This can be problematic, however, because the
child must retain four words in order to make a decision; a
certain level of short-term memory is required. Indeed,
Schneider (Schneider 1993, personal communication)
found that only 26.1% of 6-year-olds have a mean memory
digit span of four, thus making the task problematic for use
with healthy and, in particular, at-risk children. We thus
adapted a measure of phonological awareness which makes
fewer requirements on memory (Jansen et al. 1986,
Skovronek and Marx 1989).
The rhyming task consisted of 18 word pairs, half of which
rhymed. At the very beginning, four practice items were
given to ascertain that the children understood the concept
of rhyme and could principally distinguish between word
pairs that rhymed and those that did not. The 18 word pairs
were presented next (e.g. Bauch – Schlauch [rhyme]; Laub –
Lauf [no rhyme]). The number of correctly identified
rhyming and non-rhyming pairs was counted.
In the sound-to-word-matching task, children had to repeat
each given word, and then indicate if a special phoneme pronounced by the experimenter was included in the presented
word. For example, the experimenter presented the word
Cognitive Status of 6-Year-Old Very Preterm Children Dieter Wolke and Renate Meyer
97
‘Auge’ and asked subjects whether it contained an ‘au’
phoneme. Four practice items were given and then 16 test
items. The number of correct responses was recorded.
Both the rhyming and sound-to-word-matching task items
were presented from a standard prerecorded tape (female
voice) to all children to control for possible differences in
pronunciation by different investigators.
The naming of numbers and letters assessed children’s
knowledge of the alphabet and the number sequence 0 to 9.
Children were instructed to read a row of letters (random
order) and to indicate those they knew. Answers were judged
correct if the child could name, pronounce, or say a word
that began with the letter. Similarly, the digits were presented
and the child asked to name them. The total number of correct answers was calculated separately for the naming of letters and numbers.
STATISTICAL ANALYSIS
Comparisons were made between the very preterm children
and matched control groups. They included parametric
analyses using sequential analysis of variance for interval
data assessing mean differences between very preterm children and controls, after allowing for sex and SES.
For categorical analyses, outcome variables were categorized into three groups: normal performance (≥ –1SD), mild
impairment (<–1SD to ≥ –2 SD), and severe impairment
(<–2SD). Where data distribution deviated significantly from
the normal distribution (Kolomogorov-Smirnov test, P<0.05
in the normative sample), cut-off points were set at the 10th
centile (impairment score <10th centile). All cut-off points
(SD classifications; 10th centile) were determined according
to the distributions of the dependent variable scores in the
normative sample. Area-transformed cut-off points for SD
classifications were used (e.g. –1 SD at 15.9%; –2 SD at 2.3% of
the score distribution). Comparisons between very preterm
children and controls were carried out using the χ2 test.
A number of measures were taken to determine whether
very preterm children had specific cognitive and language
deficits more frequently than controls. Firstly, categorical
comparisons of the frequency of differences in performances
between the test scores MPC and AS, SGD and SED, and AS of
the K-ABC, set at >1 SD, were made to determine specific
cognitive deficits. Secondly, the parametric analyses were
repeated for the HSET using IQ scores (MPC) as covariate to
determine whether possible differences between groups disappeared once IQ was controlled for. Thirdly, odds ratios
adjusted for K-ABC IQ (MPC) were computed for the not
normally distributed variables (articulation test, prereading
scales, speech ratings).
All children underwent repeated analysis except those
children with major impairment according to medical examination. By convention, the P values reported are <0.05,
<0.01, and <0.001.
Results
SAMPLE CHARACTERISTICS AND DROP- OUTS
The characteristics of the very preterm children and control
children studied at 6 years of age are shown in Table I. The
very preterm children and controls differed on all medical
complication and biological background variables except for
parity. The very preterm children had severe medical complications more frequently, and were more commonly twin or
98 Developmental Medicine & Child Neurology 1999, 41: 94–109
higher-order births and small for gestational age. None of the
very preterm children was deaf. Nineteen very preterm children than controls had severe CP (grade three, 13; grade
four, six), compared with no controls. More very preterm
children had severe malformations or genetic aberrations
than controls (very preterm children 12; controls, three). In
the very preterm children these were microcephaly and
severe growth retardation (one); hypothyroidism (four);
multiple malformations (two); heart failure (aorta-valve
stenosis, one); lysteriosis (one); cytomegalovirus embryopathy (one); hypoplastic kidney (one); hypoplastic-multicystic
kidney and hypothyroidism (one). Among the control children, three had major congenital anomalies: aorta-valve
stenosis (one), glycogenosis (one), cytomegalovirus embryopathy (one). There were more blind very preterm children
than controls (one-sided, two children; both eyes, one child;
controls, none). As one very preterm child had both CP grade
four and was blind, there were 33 very preterm children and
three control children with severe medical impairment and
disability (MI). There were no differences between the very
preterm children and controls on any of the sociodemographic variables (Table I). However, the very preterm children were, on average, assessed at a slightly later age (0.8
months). This was due to more frequent cancellations of
appointments and journeys to the families in the very
preterm children group. The difference was of little clinical
significance which was confirmed by analyses using assessment age as covariate. No alterations of findings were found
and thus the results not corrected for the little difference in
assessment age are reported.
The universality and interpretation of the results subsequently reported depends on whether those lost to follow
up from birth until 6 years 3 months of age were different
from those who stayed. The medical and biological background data are thus also shown for the 85 children who did
not attend the 6-year-3-month assessment (Table I). Of the 14
background variables where complete data were available,
there were two significant differences (Table I): the very
preterm children lost to follow-up had fewer birth complications (P< 0.01) and a less intense course of neonatal treatment (Intensity of Treatment Index, P<0.05). Seventy-five of
the 85 children who dropped-out had at least one previous
assessment with the Griffiths Scales or CMM. The number of
children assessed at each of the previous assessment points
and the subsequent results are shown in Table II. There were
no differences between the groups. In particular, 73 of the 85
had Griffiths assessments at 5 or 20 months (and these were
combined) and no differences to those who stayed in the
study were found. The group of children who dropped-out
did not differ significantly in their early cognitive development from those who remained in the study.
COGNITIVE STATUS
The results of the K-ABC assessment are shown in Table III,
first for all children and, secondly, for the group of children
after those with major neurological impairment had been
excluded. The findings from the normative sample are shown
for descriptive purposes. The very preterm children had much
poorer performance in all of the K-ABC subscale composites,
with scores >1 SD below those of the controls in the MPC and
AS. The lowest mean scores were found in the simultaneous
information-processing (SGD) subscales (approximately 1.4
SD below the controls) while the smallest differences were
found for the sequential information-processing scales (SED)
(approximately 0.7 SD below the controls). The mean differences between the two groups were all very significantly different (P<0.001) according to the sequential ANOVAs
controlling for sex and social class of the children (Table III).
Of particular clinical interest is the relative frequency of
children with mild (<–1 SD to ≥ –2 SD) or serious impairment (<–2SD) in intellectual abilities (Table III). Mild impairment was between 2 and 2.5 times more apparent in very
preterm children than control children but up to 34 times
more very preterm children had serious cognitive impairment. The relative risk (RR) of having serious impairment relative to no or mild impairment was 34.5 (95% CI, 8.5–139.3;
P<0.001) for the MPC and 12.4 (95% CI, 5.1–30.3; P<0.001)
on the AS. The exclusion of children with major impairment
caused a reduction in the RR but the very preterm children
still had a much higher risk for serious cognitive impairment
(MPC, RR=24.9 [95% CI, 6.1–101.4], P<0.001; AS, RR=9.7
[95% CI, 3.9–24.1], P<0.001). Considering the SGD and SED
components of the MPC separately, both showed highly significant differences between the very preterm children and
the controls. However, while 12.1% of the very preterm children and 1.1% of the controls (RR=10.7 [95% CI, 3.3–34.4],
P<0.001) had problems in sequential information processing, 34.8% of the very preterm children (1.1% of controls;
RR=30.7 [95% CI, 9.8 to 95.6], P<0.001) had problems in
tasks requiring the child to process stimuli simultaneously.
Similar differences between SED and SGD were found when
only the children without major impairment were considered (Table III; SED, RR=6.0 [95% CI, 1.8–20.4], P<0.001;
SGD, RR=24.1 [95% CI, 7.7–75.7], P <0.001). The findings
indicate that very preterm children have particular problems
processing complex information requiring logical reasoning
and spatial orientation abilities.
To test for specific intellectual deficits, differences between
the individual K-ABC scores (MPC, AS, SED, SGD) were computed, and it was determined whether the very preterm children group had differences between scales exceeding 1 SD
more frequently than the controls (Hunt et al. 1988). The
results are shown in Figure 1. No differences in the frequency
of discrepancies exceeding 1 SD between AS and MPC or SGD
were found between very preterm children and controls.
However, the very preterm children had SGD scores which
were more than 1 SD below the SED scores (35.2%; 33.8%
without MI) more often than the controls (15.2%; 5.3% without MI , P<0.001). The very preterm group had lower SED
scores than AS scores (6.4%; 6.5% without MI) less often than
the controls (14.4%; 14.2% without MI, P<0.01).
The intellectual performance scores grouped according
to the SES of the subjects are shown in Figure 2. Analysis of
variance entering the factors simultaneously showed clearly
the effect of social class (P<0.001) and no interaction
between social class, group status, or sex of child was found.
No overlap of scores between the very preterm children and
controls could be detected. Indeed, not even the control
children with low SES scored, on average, lower than the
very preterm children of high SES. Figure 2 also illustrates, as
already demonstrated in the analyses above, the different
patterns in performances in SED and SGD between the very
preterm children and controls in all social groups. The very
preterm children performed particular poorly when infor-
mation presented simultaneously had to be processed.
We further analysed how SES affected the incidence of
normal intellectual performance, mild, and severe cognitive
impairment in the very preterm children and control group.
The results are shown in Figure 3 for the two groups excluding those children with major medical impairment. In the
control group, the incidence of mild impairment more than
twice the rate for the children of low SES when compared
with those of upper or middle SES. Only control children of
low SES were found to have severe impairment in the MPC
(N=2) and AS (N=5) of the K-ABC. The SES effects were highly significant for both the MPC and AS (Fig. 3).
The effect of SES on cognitive performance differed for
the MPC and AS in the very preterm children group. No
changes in the number of children with mild impairment
were found on the AS according to SES. However, the number of those with serious impairment almost doubled with
each drop in level of SES (e.g. upper SES, 7.8%; middle SES,
15.2%; low SES, 29.6%). The χ2 test (df 2) of the overall
changes was significant at P<0.05 and testing how many very
preterm children had serious impairment compared with
others (≥ 2 SD), the SES effects were highly significant (df 1;
P<0.01). The SES effects were less evident for the MPC. The
Table II: Cognitive development status of the very preterm
children who stayed and of the children who dropped out at 5,
20, and 56 months of age.
Very preterm
children
(compiler)
N=264
Very preterm
children
(drop-outs)
N=8
Griffiths General Development
Quotient
5 mo
Number assessed
262
Mean (95% CI)
95.5 (92.8–98.2)
Mild delaya (%)
13.0
Serious delayb (%)
23.3
69
93.8 (88.0–99.5)
14.5
24.6
20 mo
Number assessed
Mean (95% CI)
Mild delay a (%)
Serious delayb (%)
261
90.7 (87.9 – 93.6)
19.2
29.5
51
90.0 (83.8 – 96.1)
25.5
31.5
Griffiths DQ at 5 or 20 moc
Number assessed
Mean(95% CI)
Mild delay a (%)
Serious delayb (%)
261
90.7 (87.9 – 93.5)
19.2
29.5
73
88.9 (83.2 – 94.6)
24.7
31.5
Columbia Mental Maturity Scale
Number assessed
262
Mean(95% CI)
82.7 (79.6 – 85.8)
Mild Impairmenta (%)
18.3
Serious Impairment (%)
22.5
32
87.2 (77.7 – 96.6)
21.9
15.6
a <–1SD to >–2SD categorization according to prospectively
studied normative sample; normal, mild, and serious delay
expressed as percentage of children
b <–2SD
c two further drop-outs without Griffiths Tests had a CMM at 56
months within the normal range (CMM-IQ: 94, 96).
Cognitive Status of 6-Year-Old Very Preterm Children Dieter Wolke and Renate Meyer
99
number of cognitively impaired children increased in those
with middle SES and the number of children with serious
impairment increased largely in the low SES group (df 2;
P<0.05). However, testing whether SES predicted serious
impairment only (<–2 SD compared with others) just failed
significance (df 1; P<0.06). Categorical analyses of SGD and
SED, the two composites of the MPC, showed no significant
association between degree of cognitive impairment and SES
in the very preterm children but significant associations in
the control group (SED, P<0.05; SGD, P<0.01).
LANGUAGE DEVELOPMENT
The results of the HSET language test are shown in Table IV.
Seventeen children in the very preterm children group could
not be assessed because of severe cognitive impairment.
These children had a mean MPC intelligence score of 46.8.
Excluding these children would have biased the HSET
results and therefore a score below the lowest test score
reached by a tested child was assigned. One child in the very
preterm children group with intellectual performance within the normal range (MPC≥ –2SD) could not be tested
because of severe behaviour problems (test refuser); and
one child each in the very preterm children and control
group (both with normal IQ) could not take the test because
of time constraints. There were thus test results missing for
two very preterm children and one control child.
The very preterm children had significantly poorer language abilities in all subtests. In particular, they had problems
with grammatical rules and detecting semantically incorrect
sentences. More very preterm children had language impairments than the controls according to the HSET total score
(Table IV). While there were few differences between the
groups regarding mild impairment, many more very preterm
children (All, RR=18.1 [95% CI: 4.4–74.3], P<0.001; without
MI, RR=13.1 [95% CI: 3.1–54.8], P<0.001) had serious language impairment than their term peers.
None of the tests reported in the following was normally
distributed and thus the frequency of children <10th centile
according to the normative sample was computed. Group
differences were tested with χ2 (df 1). The very preterm children had more problems articulating words (total score: very
preterm children, 24.1%; controls, 4.9%, P<0.001) (Table V).
Table III: Mean performance and classifications of normal IQ, mild and severe impairment in intellectual abilities according to K-ABC
Normative
sample
N=311
Control group
All
N=264
Mental Processing Composite (MPC)
Mean (95% CI)
99.5 (98.2 – 100.8)
99.7 (98.4 – 101.1)
Mild impairment (<–1 to ≥ –2SD)
Cut-off point
87/88
Frequency (%)
13.2
12.5
Serious impairment (<–2SD)
Cut-off point
76/77
Frequency (%)
1.6
0.8
Simultaneous Processing Scale (SGD)
Mean (95% CI)
103.0 (101.4 – 104.6) 103.2 (101.6 – 104.9)
Mild impairment
Cut-off point
87/88
Frequency (%)
13.2
11.4
Serious impairment
Cut-off point
77/78
Frequency (%)
2.3
1.1
Sequential Processing Scale (SED)
Mean (95% CI)
96.1 (94.6 – 97.5)
96.0 (94.5 – 97.6)
Mild impairment
Cut-off point
81/82
Frequency (%)
10.9
10.6
Serious impairment
Cut-off point
68/69
Frequency (%)
1.9
1.1
Achievement Scale (AS)
Mean (95% CI)
100.5 (99.0 – 102.1) 100.9 (99.2 – 102.5)
Mild impairment
Cut-off point
86/87
Frequency (%)
10.6
10.6
Serious impairment
Cut-off point
72/73
Frequency (%)
2.9
1.9
Very preterm children
All
Without MIb
N=264
N=231
All
P
χ2
Without MI
P
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
a
84.8 (82.7 – 86.9) 87.5 (85.5 – 89.5)
23.5
26.0
26.1
19.0
83.7 (81.3 – 86.7)
87.0 (84.6– 89.4)
22.3
24.2
34.8
27.7
86.9 (84.7 – 89.1) 89.0 (86.9 – 91.0)
24.6
26.0
12.1
6.9
84.6 (82.2 – 87.0) 87.2 (84.9 – 89.6)
25.0
25.1
23.5
18.6
a χ2 comparisons of normal, mild, and severe impairment between VPI and controls (df=2) are reported. The comparisons of severe
impairment (<–2 SD) vs others (≥ –2SD) were all also highly significant (P<0.001; df=1)
b All mean differences (VPI versus controls) after controlling for sex and SES P<0.001.
100
Developmental Medicine & Child Neurology 1999, 41: 94–109
They had particular problems with the ‘Sch’ sounds (<10th
centile: very preterm children, 25.7%; controls, 7.6%,
P<0.001) and significantly more, although not so marked,
problems with sounds placed at the beginning (<10th centile: very preterm children, 18.0%; controls, 7.6%, P<0.01)
or in the middle of word <10th centile: very preterm children, 18.0%; controls, 4.2, P<0.001). More articulation
problems were still found when those with major impairment (MI) were excluded (Table V).
Eleven very preterm children had no articulated language at
all (with MPC-IQs well below –2 SD; mean, 45.3), one very
preterm child did not speak at all during the assessments (elective mutism). For these 12 children a score of 0 was assigned for
the quality of speech and grammatical correctness index. One
very preterm child (MPC-IQ=82) had no completed form and
was excluded. The reliability of ratings for the very preterm
children was judged by the investigators to be poor for 16.7%
of very preterm children and 9.1% of the controls (P<0.01).
However, once the children with MI were excluded, the examiners did not judge the reliability of their consensus ratings as
poorer for the very preterm children and controls (13.9% versus 9.2%, ns). Of the very preterm children group, 30.7% of the
children and 28.6% of the accompanying parents spoke High
German; in the control group the figures were 25.7% of children and 24.7% of parents (not significantly different). The
speech quality (P<0.001) and the grammatical correctness of
expressive language (P<0.001) was much poorer in the very
preterm children (Table V) independent of whether all or only
those without MI were considered (P<0.001).
PREREADING SKILLS
Many more very preterm children, regardless whether all or
only those without MI were considered, had significant
deficits (<10th centile) in all the prereading tests (Table V).
As previously discussed (Table I), 16 very preterm children
and 15 controls had already started school at time of testing.
The analyses were thus repeated excluding these school children, resulting only in very minor alterations. All differences
between the groups remained very significant (P<0.001).
MULTIPLE COGNITIVE DEFICITS
It was evaluated whether very preterm children had problems in more areas of cognitive functioning than their controls. For this purpose, the number of children in each group
with scores below the 10th centile (according to the normative sample) was calculated in the following seven assessments: K-ABC MPC, K-ABC AS, HSET total score, rhyming
task, sound-to-word-matching task, quality of speech index,
and grammatical correctness index. Children could thus
have a potential multiple cognitive problem score between 0
and 7. Figure 4 shows that very preterm children had deficits
in more than one of the cognitive assessments more frequently. Only 31.9% of very preterm children (34.8% without
MI) had no cognitive problems at all (controls, 67.8%) and
18.3% (without MI, 14.0%) had cognitive deficits in five or
more cognitive tests (controls, 0%). The Mann–Whitney U
test was highly significant (P<0.001), independent of
whether all very preterm children or those excluding the
children with MI were considered. Applying a <10th centile
cut off for this multiple cognitive problem index itself, 38.0%
of very preterm children (without MI, 32.6%) had cognitive
deficits in three or more areas, compared with only 7.6% of
the control children (χ2 test, df 1; P<0.001).
SPECIFIC DEFICITS : CONTROLLING FOR IQ
To test whether the differences between very preterm children and their controls in the HSET language were maintained if controlled for general intelligence, all analyses of
variance were repeated entering the MPC of the K-ABC as a
covariate. The analyses were made for all children and
repeated excluding those with MI. Furthermore, the analyses
were repeated a third time, but including only those very
preterm children and control children with an IQ according
to the MPC ≥ –2SD. This was done to exclude any exaggeration of the regression of IQ on the dependent variable
because we had dealt with missing language or prereading
skills score for 17 children who had severe cognitive impairment (IQs well below –2 SD) by setting their value at or
below the lowest score of any tested child.
40
Figure 1: Discrepancies between
the four K-ABC subscores: very
preterm children versus control
children.
❏ Very preterm children
■ Control children
a P<0.001
b P<0.01
Percentage
30
20
10
0
All
Without MI
AS>MPC
Alla
Without MIa
SED>SGD
Allb
Without MIb
AS>SED
All
Without MR
AS>SGD
Cognitive Status of 6-Year-Old Very Preterm Children Dieter Wolke and Renate Meyer 101
Secondly, the effect of very preterm birth on cognitive abilities was found to be much larger than the effect of socioeconomic factors. Thirdly, there was seldom an isolated
problem: very preterm children had multiple cognitive problems much more frequently than controls. Fourthly, very
preterm children were found to have a specific deficit in the
processing of simultaneous information (SGD) to solve tasks
(i.e. visual spatial recognition, pattern building and memory,
logical reasoning). The risk of a simultaneous information
processing deficit was 31 times greater in very preterm children compared with term controls. Finally, most differences
in achievement and language abilities between very preterm
children and controls disappeared when we controlled for
mental processing differences. Specific developmental
deficits were rare and only found for speech articulation,
quality of speech, and numeracy.
The findings of this study, in indicating that very preterm
birth is associated with deficits in a wide range of abilities, are
consistent with previously reported findings (Portnoy et al.
1988, Klein et al. 1989, Saigal 1991, Saigal et al. 1991, Halsey
et al. 1993, Weisglas-Kuperus et al. 1993, Riegel et al. 1995).
Even when children with major neurosensory impairment or
IQ scores below normal were excluded from the analyses,
the IQ, language, and prereading test scores of the very
preterm children were lower, and more very preterm children had mild to serious deficits than term children matched
for age, sex, SES, maternal education, and marital status.
A number of our results deviate from reports of cognitive
development of very preterm children in previous work. The
mean scores of the very preterm children in the various tests
are lower, and the incidence of serious impairment found in
our sample is higher than in most previous reports. The very
preterm children performed, on average, around 1 SD below
the term controls and had serious deficits 10 to 35 times more
frequently than the term controls. These findings compare
unfavourably to previous metaanalyses findings of the cognitive outcome of very preterm children or very-low-birthweight
children (e.g. Aylward et al. 1989, Ornstein et al. 1991). They
indicated that very-low-birthweight children have, on average,
IQs approximately 0.5 SD below the mean. These differences
The analysis of covariance had the effect that all differences between the groups disappeared in the four subtests
of the HSET and the HSET-total score (i.e. were all not significant), independent of whether all children, excluding
those with MI or those with an MPC ≥ –2SD were analysed.
The regression of the covariate was in all cases highly significant (P<0.001). For example, while the observed HSET-total
mean score for all very preterm children had been 47.0 and
51.1 for the controls (Table IV) and significantly different,
the IQ-adjusted scores were 49.9 for the very preterm children and 49.5 for the controls and not significantly different.
For the non-normally distributed articulation tests, language ratings, and prereading skills tests, odds ratios between
very preterm children and controls adjusted for K-ABC
MPC were computed. All differences between very preterm
children and controls in specific articulation problems
(word beginning, word middle, ‘Sch’ sounds), grammatical
correctness ratings, and phonological prereading skills
(rhyming task, sound-to-word matching, naming of letters)
disappeared once controlled for mental processing abilities
(MPC). This was found to be independent of whether all, or
only children without MI or IQs in the normal range (>–2
SD) were considered. For example, while the unadjusted
odds ratio for sound-to-word-matching was 5.1 (95% CI,
3.1–8.33) this reduced to an adjusted odds ratio of 1.7 (95%
CI, 0.96 –3.04) in the group of all very preterm children and
was no longer significant. Three significant differences
between very preterm children and controls remained once
adjusted for mental processing: total score differences in
articulation, differences in quality of speech, and in the naming of numbers (Table VI). The significant adjusted odds
ratios were, however, generally small in magnitude.
Discussion
The major findings of this study can be summarized as follows: firstly, compared with term peers of the same age, 6year-old very preterm children scored significantly lower on
all measures of cognitive and language abilities, including
assessments of general intelligence, language comprehension, and expression, articulation and prereading skills.
Table IV: Results of the assessments of language development: very preterm children versus controls
All
P (df)
χ2
Without MI
P (df)
50.3 (48.8–51.7)
46.5 (45.3–47.7)
46.1 (45.1–47.2)
47.5 (46.0–49.0)
47.8 (46.7–48.8)
–
–
–
–
–
–
–
–
–
–
12.2
13.1
–
–
13.7
10.0. <0.001 (2) <0.001 (2)
Normative
sample
N=308
Control group
All
N=263
51.7 (50.6–52.8)
51.2 (50.2–52.3)
48.9 (48.0–49.8)
51.9 (50.8–52.9)
51.1 (50.3–51.9)
52.4 (51.3–53.6)
51.4 (50.3–52.5)
48.8 (47.8–49.7)
52.2 (51.1–53.3)
51.3 (50.5–52.2)
49.4 (48.0–50.9)
45.8 (44.7–46.9)
45.8 (44.8–46.8)
46.5 (44.8–48.0)
47.0 (46.0–48.1)
Mild impairment
Cut-off point
Frequency (%)
42/43
10.1
10.3
Serious impairment
Cut-off point
Frequency (%)
37/38
1.9
0.8
HSET a Total score, mean (95% CI)
PS, plural-singular
KS, semantics
SB, sentence production
VS, grammatical structure
HSET-Total score
a
Very preterm children
All
Without MI
N=262
N=229
Mean differences between VPI and controls were significant at P<0.001 after controlling for sex and SES.
102
Developmental Medicine & Child Neurology 1999, 41: 94–109
year. Where losses occur, they are often selective and therefore
findings on the compliant sample are often biased, i.e. underestimate impairment rates (Wariyar and Richmond 1989,
Halsey et al. 1993, Wolke et al. 1995b). We showed that those
lost to follow-up were not different in most medical characteristics and had the same social characteristics. Most importantly, those lost to follow-up did not differ in their early cognitive
development (5 to 56 months) from those who stayed. The
losses in our sample were random and can thus be generalized
to the whole cohort of very preterm children in Bavaria.
Thirdly, we included a closely matched control group of children born in the same region and during the same period.
Additionally, we studied a large representative sample of children of the same cohort to determine SDs and classifications
of cognitive impairment. These were determined according to
the area under the curve of the score distributions. We found
that the means or SDs deviated from the standard scores cited
in test manuals, even for recently standardized measures (e.g.
may be best explained by methodological shortcomings in
many previous studies (Lloyd et al. 1988, Aylward and Pfeiffer
1989, Aylward et al. 1989, Abel Smith and Knight-Jones 1990,
Escobar et al. 1991, Ens-Dokkum et al. 1992, Wolke et al. 1994,
Wolke 1998a). Firstly, our study followed up children from a
geographically defined area and, unlike many previous studies, included very preterm children from more than one single
hospital centre treating large numbers of very preterm children. It has been demonstrated that treatment (Avery et al.
1987, Hack et al. 1991) and associated outcome between different hospital centres in the same (Wolke and Söhne 1997)
and different regions (Wolke 1997b, 1998b) varies greatly.
Other geographical studies have also reported higher levels of
impairment than single centre studies (Veen et al. 1991, Hille
et al. 1994). Secondly, our drop-out rate was low, with less
than 25% of surviving children lost to full face-to-face assessment at 6 years 3 months. Aylward et al. (1989) noted that
most developmental studies have a 10% drop-out rate per
Ñ
▲ Upper
120
Figure 2: K-ABC subscores
according to SES in very
preterm and control children.
Ñ
● Middle
110
▲
●
▲
100
●
IQ
+
90
80
●
▲
●
+
+
▲
▲
●
+
+
●
+
Lower
▲
+
Control
children
▲
▲
●
+
●
+
Very
preterm
children
70
MPC
Figure 3: Serious and mild
cognitive impairments in the
K-ABC mental processing and
achievement subscores
according to SES.
SED
SGD
AS
■ Serious impairment
❏ Mild impairment
60
50
a P<0.05, b P<0.01, c P<0.001.
Percentage
40
30
20
10
0
Upper Middlea Low
Upper Middleb Low
Very preterm children
Control children
Mental processing composite
Upper MiddleaLow
Upper Middleb Low
Very preterm children
Control children
Achievement
Cognitive Status of 6-Year-Old Very Preterm Children Dieter Wolke and Renate Meyer 103
German version of the K-ABC; Melchers and Preuss 1991). For
example, the standard scores for the K-ABC MPC was
99.5±11.6 in the normative sample rather than 100±15 (test
manual). This had important implications for setting the cutoff point for serious impairment (<–2 SD) which was at 76/77
rather than <70 if done according to the test-manual norms.
Applying our concurrent norms and classifications, 26.1% of
very preterm children had serious impairment, a figure
reduced to only 17.4% if determined according the test-manual norms published in 1991. These underestimations of cognitive impairment become more likely as the time period since
the tests were standardized increases (Flynn 1984, Flynn
1987, Murphy 1987, Fuggle et al. 1992, Gross et al. 1992,
Halsey et al. 1993, Wolke et al. 1994). Fourthly, the assessments of all children were videotaped and subject to constant
supervision by one clinical psychologist, and regular feedback
seminars were conducted. Finally, the follow-up period
extended into the late preschool years, allowing for more certain judgments of long-term problems than assessments in
infancy (Hindley and Owen 1979, McCall 1981, Escobar et al.
1991, Colombo 1993, McCall and Carriger 1993). The study
thus meets and exceeds demands for methodologically sound
follow-up studies (Mutch et al. 1989, Escobar et al. 1991,
Ornstein et al. 1991, Saigal 1991, Saigal et al. 1991, Harrison
1993, Lukeman and Melvin 1993, Siegel 1994, Wolke 1998a).
Our findings cannot be explained by lack of reliability of
assessments or measures.
GENERAL INFORMATION PROCESSING DEFICIT
One result of particular relevance is that very preterm children
have a specific deficit in simultaneous information processing,
and once MPC-IQ was controlled for statistically, most differences in language performance and other cognitive tasks
between very preterm children and controls disappeared. It
suggests that the ability to perceive, process, and integrate different stimuli at the same time may be fundamental to the multiple cognitive problems that very preterm children experience.
We found some, but generally little, evidence that very
preterm children encounter specific developmental disorders
or learning difficulties. Rather, language comprehension,
Table V: Language abilities and prereading skills of very preterm children and their controls
Articulation test (N)
Word beginning (N=13)
Cut-off point
<10th centile (%)
Word middle (N=13)
Cut-off point
<10th centile (%)
Sch-sounds (N=8)
Cut-off point
<10th centile (%)
Total score (N=34)
Cut-off point
<10th centile (%)
All
P (df)
χ2
Without MI
P (df)
14.0
<0.001 (1)
<0.05 (1)
18.0
14.0
<0.001 (1)
<0.001 (1)
25.7
23.1
<0.001 (1)
<0.001 (1)
<0.001 (1)
<0.001 (1)
Normative
sample
Control group
All
Very preterm children
All
Without MI
310
264
261
229
11/12
9.7
7.6
18.0
11/12
7.4
4.2
6/7
9.0
7.6
29/30
8.7
4.9
24.1
21.0
Language ratings a
Quality of speech (N=5)
Cut-off point
<10th centile (%)
Grammatical correctness (N=7)
Cut-off point
<10th centile (%)
N=310
N=264
N=261
N=230
2/3
8.1
6.8
26.2
23.5
<0.001 (1)
<0.001 (1)
2/3
7.4
7.6
25.1
21.7
<0.001 (1)
<0.001 (1)
Prereading skills a
Rhyming task (N=18)
Cut-off point
<10th centile (%)
Sound-to-word-matching (N=16)
Cut-off point
10th centile (%)
Naming of letters (N=26)
Cut-off point
10th centile (%)
Naming of numbers (N=11)
Cut-off point
<10th centile (%)
N=306
N=262
N=261
N=228
9/10
9.5
9.2
30.3
24.6
<0.001 (1)
<0.001 (1)
5/6
11.2
9.2
34.0
30.3
<0.001 (1)
<0.001 (1)
0/1
10.5
11.1
31.9
27.6
<0.001 (1)
<0.001 (1)
2/3
9.2
8.0
35.0
31.1
<0.001 (1)
<0.001 (1)
a
Scores were missing for the following reasons: behaviour problems (test refuser: two very preterm children, one control), time constraints
(one very preterm child, one control).
104
Developmental Medicine & Child Neurology 1999, 41: 94–109
grammatical rule application and knowledge, and phonetic
understanding were related to a central information- processing deficit. Differences between very preterm children
and controls remained, even after IQ was controlled in oralmotor-skill-related tasks such as articulation and quality of
speech, suggesting that these may be specific motor-related
deficits of very preterm children (Largo et al. 1990, The
Scottish Low Birthweight Study Group 1992). Naming of
numbers was also only partly explained by MPC-IQ. Klein et
al. (1989) previously found that once controlled for, IQ differences between very-low-birthweight and controls disappeared in auditory mediated skills (e.g. reading) but those
in maths skills remained. They concluded that this may be
due to particular deficits in visually mediated tasks apparent
in very-low-birthweight children. Hunt et al. (1988) found
evidence for specific learning difficulties (e.g. visual motor
and language problems) in very-low-birthweight children.
However, the use of tests standardized at different times
(e.g. the IQ score of the very-low-birthweight was 110.3 in
their sample, 10 points above the norm) and the lack of differences found between the very-low-birthweight children
and controls may have led to an overinterpretation of significant differences between general IQ and specific achievement tests. Further follow-up at school age is necessary to
reassess the issue of specific learning difficulties.
The finding of a specific deficit in simultaneous information processing also requires replication at a later or earlier
age in very preterm children and in other samples. Our
Table VI: Speech problems of very preterm children versus controls after
adjustment for children’s IQ (adjusted odds ratios)
Unadjusted
OR
95% CI
Figure 4: Relative
frequency of
multiple cognitive
deficits of very
preterm and
control children.
OR
Adjusted
95% CI
Articulation test total score
All children
Children without MI
Normal IQ children (>–2SD MPC)
6.14
5.06
3.00
3.29–11.48
2.66–9.61
1.50–6.01
2.20
2.53
2.63
1.08–4.47
1.25–5.12
1.27–5.43
Language rating: quality of speech
All children
Children without MI
Normal IQ children (>–2SD MPC)
4.86
4.14
2.98
2.80–8.44
2.35–7.31
1.63–5.45
2.08
2.16
2.32
1.11–3.88
1.15–4.06
1.23–4.37
Naming of numbers
All children
Children without MI
Normal IQ children (>–2SD MPC)
6.18
5.13
3.34
3.70–10.33
3.03–8.68
1.89–5.90
2.18
2.28
2.36
1.21–3.92
1.26–4.11
1.29–4.30
Control children
All very preterm children
Very preterm children without MI
70
60
Percentage
50
40
30
20
10
0
0
1
2
3
4
<5
Nr of cognitive deÞcits (<10th centile)
Cognitive Status of 6-Year-Old Very Preterm Children Dieter Wolke and Renate Meyer 105
findings of a generalized deficit in integrating complex
information may suggest that they are not related to lesions
in specific areas of the brain but rather to aberrant cortical
development. Atrophy or thinning of white matter, late or
incomplete migration and myelination, and thus poorer
connectivity between different parts of the brain have been
found in very preterm infants (Klein et al. 1989, Cioni et al.
1992, Fujii et al. 1993, Skranes et al. 1993, Goto et al. 1994).
These may be consequences of injuries to the white matter
(subcortical ischaemic/infarctive brain lesions) (Whitaker
et al. 1996, 1997) or a result of impaired brain growth in
those infants who require long neonatal treatment and
show little catch-up growth. The effects of early nutrition
and head growth on IQ development in both preterm and
term children have been documented (Hack et al. 1991,
Lucas et al. 1992, Skuse et al. 1994, Morley and Lucas 1997).
Both imaging techniques such as (functional) MRI or PET
and investigations of the effects of early head (and brain)
growth may provide important clues to the pathogenic
pathways.
SOCIAL FACTORS
In recent years, awareness has increased that the long-term
effects of prematurity should be considered within the context
of sociodemographic and environmental influences (Sameroff
and Chandler 1975, Drillien et al. 1980, Saigal 1991, Saigal et al.
1991, Bendersky and Lewis 1994). Previous studies and
reviews on high-risk infants born in the 1960s (mainly lowbirthweight or term sick infants rarely requiring NICU care
nowadays) suggested that social factors can compensate for
reproductive casualty and that the relation between SES and
reproductive risk is likely to be interactional, that is the risk for
poor cognitive outcome is intensified (disproportionally high)
for those of high biological and social risk (e.g. Sameroff and
Chandler 1975, Escalona 1982, Werner 1993, Bendersky and
Lewis 1994). Similarly, recent research on ex-special care
babies (term or preterm) indicates that sociodemographic factors have a far greater effect on long-term cognitive outcomes
than biological risk factors as the children grow older (Wolke
1993; Censullo 1994; Forfar et al. 1994; Levy-Shiff et al. 1994;
Cohen 1995; Hack et al. 1995; Wolke 1997a, 1998b). Our
results show that SES was related to all cognitive and language
scores. Children from families of higher SES, regardless of
whether they were controls or very preterm children, obtained
higher scores than those of the middle SES, who in turn
obtained higher scores than those of the lower SES. However,
the effect sizes for the influence of prematurity were still much
higher than those of SES at 6 years. Furthermore, the effects of
prematurity and SES were independent from each other and
thus additive (double jeopardy). Thus infants who were very
preterm at birth and came from the lower SES performed most
poorly (additive effect) but they did not perform disproportionally worse than low SES controls (no interactive or potentiating effect). This confirms our findings on the same sample at
5, 20, and 56 months of age (Wolke 1993, Wolke et al. 1994,
Riegel et al. 1995) and is in agreement with recent other longerterm follow-up studies of very preterm children (Eilers et al.
1986, Lloyd et al. 1988, McCormick et al. 1989, McGauhey et al.
1991, Saigal 1991, Saigal et al. 1991, Levy-Shiff et al. 1994). The
evidence thus underlines the importance of social adversity
factors for all children whether at biological risk or not.
However, different from low-birthweight or term infants (low
106
Developmental Medicine & Child Neurology 1999, 41: 94–109
to moderate biological risk), very preterm birth and the associated intensive neonatal treatment received (Wolke 1987, 1997;
Silverman 1992; Als et al. 1994; Wolke 1998a,b; Wolke and
Eldridge1998) may have caused CNS insults which set limits to
the plasticity of the CNS even in very preterm children without
severe neurological deficits. In particular, the ability to take
advantage of environmental stimuli requiring complex information processing may be more often impaired in very
preterm children. Corroborative evidence is available from the
large-scale Infant Health and Development Programme
(Ramey et al. 1992). Brooks-Gunn et al. (1994) and McCarton
et al. (1997) reported that changes in the care-taking environment (intense multimodal intervention in the first 3 years of
life) had no long-term positive effects (at 5 or 8 years of age) on
the cognitive development for infants <2000 g. In short, the
comprehensive program had no effect on those infants at
greatest risk for mental retardation (Baumeister and Bacharah
1996). Hack et al. (1995) also concluded that ‘enrichment programs for low-birthweight children seem to be most effective
for the moderately-low-birthweight child who comes from a
lower socioeconomic group’ (p 176).
Further follow up and comparisons to larger preterm and
term ex-neonatal-special-care children will help to clarify
whether a highly increased number of very preterm children
have suffered specific long-term insult to information processing which can only partly be compensated for by environmental factors (Levy-Shiff et al. 1994). If this is the case, appropriate
prevention has to start before or at birth (Baumeister and
Bacharach 1996). Neonatal care approaches which reduce
treatment invasiveness (Jacobson et al. 1993, Wolke 1997b)
and provide individualized gentle care (Wolke 1991, 1997b,
1998b; Als et al. 1994) may reduce neonatal complications
and help prevent long-term cognitive sequelae.
Acknowledgements
The research was supported by the Bundesministerium für
Forschung und Technik (Federal Government of Germany, Ministry
of Science and Technology) program grants PKE 4 and JUG 14 (FKZs
0706224; 0706564 and 01EP9504) to K Riegel, D Wolke, and B Ohrt.
Thanks to the 17 special-care baby units in south Bavaria, the highly
committed psychologists and psychometric assistants who carried
out the assessments, and the parents and children who continue to
participate in the Bavarian Longitudinal Study.
Accepted for publication 27th July 1998.
References
Abel Smith AE, Knight-Jones EB. (1990) The abilities of very lowbirthweight children and their classroom controls.
Developmental Medicine and Child Neurology 32: 590–601.
Achenbach TM, Howell CT, Aoki MF, Rauh VA. (1993) Nine-year
outcome of the Vermont Intervention Program for low birth
weight infants. Pediatrics 91: 45–55.
Als H, Lawhon G, Duffy F, McAnulty G, Gibes-Grossman R, Blickman J.
(1994) Individualised developmental care for the very low birth
weight preterm infant. Journal of the American Medical
Association 272: 853–932.
Avery ME, Tooley WH, Keller JB, Hurd SS, Bryan MH, Cotton RB,
Epstein MF, Fitzhardinge PM, Hansen CB, Hansen TN et al.
(1987) Is chronic lung disease in low birth weight infants
preventable? A survey of eight centers. Pediatrics 79: 26–30.
Aylward GP, Pfeiffer S. (1989) Follow-up and outcome of low
birthweight infants: conceptual issues and a methodology review.
Australian Paediatric Journal 25: 3–5.
Aylward GP, Pfeiffer SI, Wright A, Verhulst SJ. (1989) Outcome
studies of low birth weight infants published in the last decade: a
meta-analysis. Journal of Pediatrics 115: 515–20.
Bauer A. (1988) Ein Verfahren zur Messung des für das
Bildungsverhalten relevanten Sozial Status (BRSS)
–überarbeitete Fassung. Frankfurt, Germany: Deutsches Institüt
für Internationale Pädagogische Forschung.
Baumeister AA, Benasch VR. (1996) A critical analysis of the Infant
Health and Development Program. Intelligence 23: 79–104.
Bax M. (1983) Following up the small baby. Developmental
Medicine and Child Neurology 25: 415–6. (Editorial.)
Bendersky M, Lewis M. (1994) Environmental risk, biological risk, and
developmental outcome. Developmental Psychology 30: 484–94.
Bondy C, Cohen R, Eggert D, Lüer G. (1975) Testbatterie für Geistig
Behinderte Kinder. TBGB Manual. 3rd edn. Weinheim: Beltz.
Bowey CA, Cain MT, Ryan SM. (1992) A reading level design study of
phonological skills underlying fourth grade children’s word
reading difficulties. Child Development 63: 999–1011.
Bradley L, Bryant PE. (1978) Difficulties in auditory organisation as
a possible cause of reading backwardness. Nature 271: 746–7.
— — (1983) Categorizing sounds and learning to read – a causal
connection. Nature 301: 419–21.
Brandt I. (1983) Griffiths Entwicklungsskalen. Weinheim: Beltz.
Breslau N, Klein N, Allen L. (1988) Very low birthweight:
behavioural sequelae at nine years of age. Journal of the
American Academy of Child and Adolescent Psychiatry
27: 605–12.
Brooks-Gunn J, McCarton C, Casey P, McCormick M, Bauer C,
Bernbaum J. (1994) Early intervention in low birth weight
premature infants. Journal of the American Medical Association
272: 1257–62.
Burgemeister B, Blum L, Lorge I. (1954) The Columbia Mental
Maturity Scale Manual. Yonkers-on-Hudson, NY: World.
Byers B, Bendersky M, Chapman T. (1986) The early utterances of
preterm infants. British Journal of Disorders of Communication
21: 307–19.
Casaer P, Eggermont E. (1985) Neonatal clinical neurological
assessment. In: S Harel, NJ Nicholas, editors. The At-Risk Infant:
Psycho/socio/medical Aspects. Baltimore: Brookes. p 197-220
Censullo M. (1994) Developmental delay in healthy premature infants
at age two years: implications for early intervention. Journal of
Developmental and Behavioural Pediatrics 15: 99–104.
Cioni G, Bartalena L, Biagioni E, Boldrini A, Canapicchi R. (1992)
Neuroimaging and functional outcome of neonatal leukomacia.
Behavioural Brain Research 49: 7–19.
Cohen SE. (1995) Biosocial factors in early infancy as predictors of
competence in adolescents who were born prematurely. Journal
of Developmental and Behavioural Pediatrics 16: 36–41.
—Parmelee AH, Sigman M, Beckwith L. (1988) Antecedents of
school problems in children born preterm. Journal of Pediatric
Psychology 13: 493–508.
Collin MF, Halsey CL, Anderson CL. (1991) Emerging
developmental sequelae in the ‘normal’ extremely low birth
weight infant. Pediatrics 88: 115–20.
Colombo J. (1993) Infant Cognition –Predicting Later Intellectual
Functioning. Individual Differences and Development. London:
Sage.
Drillien CM, Thomson AJM, Burgoyne K. (1980) Low-birthweight
children at early school-age: a longitudinal study. Developmental
Medicine and Child Neurology 22: 26–47.
Dubowitz L, Dubowitz V. (1979) Clinical Assessment of Gestational
Age in the Newborn Infant. London: Cow and Gate.
Eilers BL, Desai NS, Wilson MA, Cunningham MD. (1986)
Classroom performance and social factors of children with birth
weights of 1,250 grams or less: follow-up at 5 to 8 years of age.
Pediatrics 77: 203–8.
Ens-Dokkum M, Schreuder AM, Veen S, Verloove-Vanhorick SP,
Brand R, Ruys JH. (1992) Evaluation of care for the preterm
infant: review of literature on follow-up of preterm and low
birthweight infants. Report from the collaborative project on
preterm and small for gestational age infants (POPS) in The
Netherlands. Paediatric and Perinatal Epidemiology 6: 434–59.
Escalona S. (1982) Babies at double hazard: early development of
infants at biologic and social risk. Pediatrics 70: 670–6.
Escobar GJ, Littenberg B, Petitti DB. (1991) Outcome among
surviving very low birthweight infants: a meta-analysis. Archives
of Disease in Childhood 66: 204–11.
Flynn J. (1987) Massive IQ gains in 14 nations: what IQ tests really
measure. Psychological Bulletin 101: 171–91.
Flynn JR. (1984) The mean IQ of Americans: massive gains 1932 to
1978. Psychological Bulletin 95: 29–51.
Forfar JO, Hume R, McPhail FM, Maxwell SM, Wilkinson EM, Lin J-P,
Brown JK. (1994) Low birthweight: a 10-year outcome study of
the continuum of reproductive casualty. Developmental
Medicine and Child Neurology 36: 1037–48.
Friedman SL, Sigman MD. (1992) Past, present and future directions
in research on the development of low birthweight children. In:
SL Friedman, MD Sigman, editors. The Psychological
Development of Low Birthweight Infants. New Jersey: Ablex
Publishing Corporation.
Fuggle PW, Tokar S, Grant DB, Smith I. (1992) Rising IQ scores in
British children: recent evidence. Journal of Child Psychology
and Psychiatry 33: 1241–7.
Fujii Y, Konishi Y, Kuriyama M, Maeda M, Saito M, Ishii Y, Sudo M.
(1993) MRI assessment of myelination patterns in high risk
infants. Pediatric Neurology 9:194–7.
Goswami U. (1990) A special link between rhyming skill and the use
of orthographic analogies by beginning readers. Journal of Child
Psychology and Psychiatry 31: 301–11.
Goto M, Ota R, Iai M, Sugita K, Tanabe Y. (1994) MRI changes and
deficits of higher brain functions in preterm diplegia. Acta
Paediatrica 83: 506–11.
Gross SJ, Slagle TA, D’Eugenio, Mettelman BB. (1992) Impact of a
matched term control group on interpretation of
developmental performance in preterm infants. Pediatrics
90: 681–7.
Hack M, Horbar JD, Malloy MH, Tyson KJE, Wright E, Wright L.
(1991a) Very low birth weight outcomes of the National Institute
of Child Health and human development neonatal network.
Pediatrics 87: 587–97.
— Breslau N, Weissman B, Aram D, Klein N, Borawski E. (1991b)
Effect of very low birth weight and subnormal head size on
cognitive abilities at school age. The New England Journal of
Medicine 325: 231–7.
— Klein N, Taylor H. (1995) Long term developmental outcomes of
low birth weight infants. Future Child 5: 176–96.
Hall AM, Counsell C, Thomson L, Mutch L. (1995) School
attainment, cognitive ability, and motor function in a total
Scottish low birthweight population at eight years: A controlled
study. Developmental Medicine and Child Neurology
37: 1037–50.
Halsey CL, Collin MF, Anderson CL. (1993) Extremely low birth
weight children and their peers: a comparison of preschool
performance. Pediatrics 91: 807–11.
Harrison H. (1993) The principles for family-centered neonatal
care. Pediatrics 92: 643–50.
Hille E, Den Ouden AL, Bauer L, Brand R, Verloove-Vanhorick SP.
(1994) School performance at nine years of age in very
premature and very low birth weight infants: perinatal risk
factors and predictors at five years of age. Journal of Pediatrics
125: 426–34.
Hindley CB, Owen CF. (1979) An analysis of individual patterns of
DQ and IQ curves from 6 months to 17 years. British Journal of
Psychology 70: 273–93.
Hunt J, Cooper B, Tooley W. (1988) Very low birth weight infants at 8
and 11 years of age: role of neonatal illness and family status.
Pediatrics 82: 596–603.
Jacobson SW, Jacobson JL, Sokol RJ, Martier SS, Ager JW. (1993)
Prenatal alcohol exposure and infant information processing
ability. Child Development 64: 1706–21.
Jansen H, Knorn P, Mannhaupt G, Marx H, Beck M, Skowronek H.
(1986) Bielefelder Screening zur Vorhersage von LeseRechtschreibschwierigkeiten.
Johnson A, Townshend P, Bull D, Wilkinson AR. (1993) Functional
abilities at age 4 years of children born before 29 weeks of
gestation. British Medical Journal 306: 1715–8.
Kaufman A, Kaufman N. (1983) Kaufman Assessment Battery for
Children. Minnesota,USA: Circle Pines.
Klein NK, Hack M, Breslau N. (1989) Children who were very low
birth weight: development and academic achievement at nine
years of age. Journal of Developmental and Behavioural
Pediatrics 10: 32–7.
Kyllerman M, Hagberg G. (1983) Reduced optimality in pre- and
perinatal conditions in a Swedish newborn population.
Neuropediatrics 14: 37–42.
Cognitive Status of 6-Year-Old Very Preterm Children Dieter Wolke and Renate Meyer 107
Largo RH, Molinari L, Kundu S, Duc G. (1990) Intellectual outcome,
speech and school performance in high risk preterm children
with birth weight appropriate for gestational age. European
Journal of Pediatrics 149: 845–50.
Leonard CH, Clyman RI, Piecuch RE, Juster RE, Ballard RA, Behle
MB. (1990) Effect of medical and social risk factors on outcome of
prematurity and very low birth weight. Journal of Pediatrics
116: 620–6.
Levy-Shiff R, Einat G, Mogilner MB, Lerman MKR. (1994) Biological
and environmental correlates of developmental outcome of
prematurely born infants in early adolescence. Journal of
Pediatric Psychology 19: 63–78.
Lewis M, Bendersky M. (1989) Cognitive and motor differences
among low birth weight infants: impact of intraventricular
hemorrhage, medical risk, and social class. Pediatrics
83: 187–92.
Li AKF, Sauve RS, Creighton DE. (1990) Early indicators of learning
problems in high-risk children. Journal of Developmental and
Behavioural Pediatrics 11: 1–6.
Lindgren SD, Harper DC, Blackman JA. (1986) Environmental
influences and perinatal risk factors in high-risk children. Journal
of Pediatric Psychology 11: 531–47.
Lloyd BW, Wheldall K, Perks D. (1988) Controlled study of
intelligence and school performance of very low-birthweight
children from a defined geographical area. Developmental
Medicine and Child Neurology 30: 36–42.
Lucas A, Morley R, Cole T, Lister G, Leeson-Payne C. (1992) Breast
milk and subsequent intelligence quotient in children born
preterm. The Lancet 339: 261–4.
Lukeman D, Melvin D. (1993) The pre-term infant: psychological
issues in childhood. Journal of Child Psychology and Psychiatry
34: 837–50. (Annotation.)
McCall RB. (1981) Early predictors of later IQ: the search continues.
Intelligence 5:141–7.
McCall RB, Carriger MS. (1993) A meta-analysis of infant habituation
and recognition memory performance as predictors of later IQ.
Child Development 64: 57–79.
McCarton CM, Brooks-Gunn J, Meinhert CL. (1997) Results at age 8
years of early intervention for low-birth-weight premature
infants: the infant health and development program. Journal of
the American Medical Association 277: 126–32.
McCormick MC. (1989) Long term follow up of infants discharged
from neonatal intensive care units. Journal of the American
Medical Association 261: 1767–72.
— (1992) Advances in neonatal intensive care technology and their
possible impact on the development of low-birthweight infants.
In: Friedman S, editor. The Psychological Development of LowBirthweight Children. New Jersey, USA: Ablex.
— Brooks-Gunn J, Workman-Daniels K, Turner JA, Peckham GJ.
(1992) The health and developmental status of very low-birthweight children at school age. Journal of the American Medical
Association 267: 2204–8.
McGauhey PJ, Starfield B, Alexander C, Ensminger ME. (1991)
Social environment and vulnerability of low birth weight
children: a social-epidemiological perspective. Pediatrics
88: 943–53.
Melchers P, Preuss U. (1991) K-ABC: Kaufman Assessment Battery
for Children. In: Deutschsprachige Fassung. Frankfurt: Swets and
Zeitlinger.
Michaelis R, Haas G. (1987) The concept of optimality in evaluating
hazards of prenatal, perinatal and postnatal complications. In:
Rauh H and Steinhausen HC, editors. Psychobiology and Early
Development. Amsterdam: North-Holland. p 145–54.
Molfese VJ, Thomson B. (1985) Optimality versus complications:
assessing predictive values of perinatal scales. Child
Development 56: 810–23.
Murphy G. (1987) Are intelligence tests outmoded? Archives of
Disease in Childhood 62: 773–75.
Morley R, Lucas A. (1997) Nutrition and cognitive development.
British Medical Bulletin 53: 123–34.
Mutch L, Johnson M, Morley R. (1989) Follow up studies: design,
organisation, and analysis. Archives of Disease in Childhood
64: 1394–402.
— Leyland A, McGee A. (1993) Patterns of neuropsychological
function in a low-birthweight population. Developmental
Medicine and Child Neurology 35: 943–56.
108
Developmental Medicine & Child Neurology 1999, 41: 94–109
Ornstein M, Ohlsson A, Edmonds J, Esztalos E. (1991) Neonatal
follow-up of very low birthweight/extremely low birthweight
infants to school age: a critical overview. Acta Paediatrica
Scandinavica 80: 741–8.
Portnoy S, Callias M, Wolke D, Gamsu H. (1988) Five-year follow-up
study of extremely low-birthweight infants. Developmental
Medicine and Child Neurology 30: 590–8.
Prechtl HFR. (1967) Neurological sequelae of prenatal and perinatal
complications. British Medical Journal 12: 763–77.
Ramey CT, Bryant DM, Wasik BH, Sparling JJ, Fendt KH, LaVange LM.
(1992) Infant health and development program for low birth
weight, premature infants: program elements, family
participation, and child intelligence. Pediatrics 89: 454–65.
Raz RS, Bryant P. (1990) Social background, phonological
awareness, and children’s reading. British Journal of
Developmental Psychology 8: 209–25.
Riegel K, Ohrt B, Wolke D, Österlund K. (1995) Die Entwicklung
gefährdet geborener Kinder bis zum fünften Lebensjahr. Die
ARVO-YLLPÖ Neugeborenen-Nachfolge Studie in Südbayern und
Südfinnland. Stuttgart, Germany: Enke Verlag.
Robertson CMT, Hrynchyshyn GJ, Etches PC, Pain KS. (1992)
Population-based study of the incidence, complexity, and severity
of neurologic disability among survivors weighing 500 through
1250 grams at birth: a comparison of two birth cohorts.
Pediatrics 90:750–5.
Rocissano L, Yatchmink Y. (1983) Language skill and interactive
patterns in prematurely born toddlers. Child Development
54: 1229–41.
Ross G, Lipper G, Auld AM. (1991) Educational status and schoolrelated abilities of very low birth weight premature children.
Pediatrics 88: 1125–34.
Saigal S. (1991) Follow-up of high-risk infants: methodological
issues, current status, and future trends. In: Kiely M, editor.
Reproductive and Perinatal Epidemiology, Boca Raton FA: CRC
Press.
— Szatmari P, Rosenbaum P, Campbell D, King S. (1991) Cognitive
abilities and school performance of extremely low birth weight
children and matched term control children at age 8 years: a
regional study. Journal of Pediatrics 118: 751–60.
Sameroff A, Chandler M. (1975) Reproductive risk and the
continuum of caretaking casualty. In: Horowitz F, Hetherington J,
Scarr-Salapatek S, Siegel G, editors. Review of Child Development
Research. Chicago: University of Chicago Press. p 187–243.
Schneider W. (1993) Introduction: the early prediction of reading
and spelling. European Journal of Psychology of Education
8: 199–203.
Scottish Low Birth Weight Study Group. (1992) The Scottish Low
Birthweight Study: II. Language attainment, cognitive status, and
behavioural problems. Archives of Disease in Childhood
67: 682–6.
Siegel LS. (1994) The long-term prognosis of pre-term infants.
Human Nature 5: 103–26.
Silverman WA. (1992) Overtreatment of neonates? A personal
retrospective. Pediatrics 90: 971–6.
Skovronek H, Marx H. (1989) The Bielefeld longitudinal study on
early identification of risks in learning to read and write:
theoretical background and first results. In: Bamboing U, Lösel F,
Skowronek H, editors. Children at Risk: Assessment,
Longitudinal Research and Intervention. New York: Walter de
Gruyter.
Skranes JS, Vik T, Nilsen G, Smevik O, Andersson HW, Rinck P,
Brubakk AM. (1993) Cerebral magnetic resonance imaging (MRI)
and mental and motor function of very low birth weight infants at
one year of corrected age. Neuropediatrics 24: 256–62.
Skuse D, Pickles A, Wolke D, Reilly S. (1994) Postnatal growth and
mental development: evidence for a ‘sensitive period’. Journal of
Child Psychology and Psychiatry 35: 521–45.
St James-Roberts I, Wolke D. (1989) Do obstetric factors affect the
mother’s perception of her newborn’s behaviour? British
Journal of Developmental Psychology 7: 141–58.
Tabachnik BG, Fidell LS. (1989) Using Multivariate Statistics. New
York: Harper and Row.
Teplin SW, Burchinal M, Johnson-Martin N, Humphry RA, Kraybill EN.
(1991) Neurodevelopmental, health, and growth status at age 6
years of children with birth weights less than 1001 grams. Journal
of Pediatrics 118: 768–77.
Veen S, Ens-Dokkum M, Schreuder A, Verloove-Vanhorick S, Brand R,
Ruys J. (1991) Impairments, disabilities, and handicaps of very
preterm and very-low-birthweight infants at five years of age.
Lancet 338: 33–6.
Velluntino FR, Scanlon DM. (1987) Phonological coding,
phonological awareness, and reading ability: evidence from a
longitudinal and experimental study. Merril Palmer Quarterly
33: 321–63.
Vohr BR, Garcia Coll CT. (1985) Neurodevelopmental and school
performance of very low-birth-weight infants: a seven-year
longitudinal study. Pediatrics 76: 345–50.
— — Oh W. (1988) Language development of low-birthweight
infants at two years. Developmental Medicine and Child
Neurology 30: 608–15.
Wariyar UK, Richmond S. (1989) Morbidity and preterm delivery:
importance of 100% follow-up. Lancet 18: 387–8.
Weisglas-Kuperus N, Baerts W, Smrkovsky M, Sauer PJJ. (1993) Effects
of biological and social factors on the cognitive development of very
low birth weight children. Pediatrics 92: 658–65.
Werner E. (1993) Risk, resilience, and recovery: perspectives from
the Kauai longitudinal study. Development and Psychopathology
5: 503–15.
Whitaker A, Feldman J, Van Rossem R, Schonfeld IS, Pinto-Martin JA,
Torre C, Blumenthal SR, Paneth NS. (1996) Neonatal cranial
ultrasound abnormalities in low birth weight infants: relation to
cognitive outcomes at six years of age. Pediatrics 98: 719–29.
Whitaker AH, Van Rossem R, Feldman JF, Schonfeld IS, Pinto-Martin JA,
Torre C, Shaffer D, Paneth N. (1997) Psychiatric outcomes in lowbirth-weight children at age 6 years: relation to neonatal cranial
ultrasound abnormalities. Archives of General Psychiatry
54: 847–56.
Wolke D. (1987) Environmental neonatology. Archives of Disease in
Childhood 62: 987–8. (Annotation.)
— (1991) Supporting the development of low birthweight infants.
Journal of Child Psychology and Psychiatry 32: 723–41.
(Annotation.)
— (1993) Langzeitprognose von Frühgeborenen: was wir wissen
und was wir wissen sollten. In: Lischka A, Bernert G, editors.
Aktuelle Neuropädiatrie 1992. Wehr Baden: Verlag Ciba Geigy.
p 99–121.
— (1997a) Entwicklung sehr Frühgeborener bis zum 7. Lebensjahr.
In: Horstmann T, Leyendecker C, editors. Frühförderung und
Frühbehandlung - wissenschaftliche Grundlagen,
praxisorientierte Ansätze und Perspektiven interdisziplinarer
Zusammenarbeit. Heidelberg: Universitätsverlag C Winter.
p 271–88.
—(1997b) The preterm responses to the environment - long term
effects? In: Cockburn F, editor. Advances in Perinatal Medicine.
Carnforth: Parthenon Publishing. p 305–14.
— (1998a) The psychological development of prematurely born
children. Archives of Disease in Childhood 78: 567–70.
— (1998b) Premature babies and the special care baby Unit
(SCBU)/neonatal intensive care unit (NICU): environmental,
medical and developmental considerations. In: Niven K, Walker A,
editors. The Psychology of Reproduction: Current Issues in
Infancy and Early Parenthood. (Vol. 3). Butterworth:
Heinemann. p 255–82.
— Söhne B. (1997) Wenn der Schein trügt: zur kritischen
Interpretation von Entwicklungsstudien. Teil 1: Studienplan,
Stichprobenbeschreibung, Probandenverluste und
Kontrolgruppen. Monatsschrift fur Kinderheilkunde
145: 444–56.
— Eldridge T. (1998) The environment of care. In: Campbell AGM,
McIntosh N, editors. Textbook of Paediatrics, 5th edn.
Edinburgh: Churchill Livingstone. p 96–100.
— Ratschinski G, Ohrt B, Riegel K. (1994) The cognitive outcome of
very preterm infants may be poorer than often reported: an
empirical investigation of how methodological issues make a big
difference. European Journal of Pediatrics 153: 906–15.
— Meyer R, Ohrt B, Riegel K. (1995a) The incidence of sleeping
problems in preterm and term infants discharged from special
neonatal care units: an epidemiological longitudinal study.
Journal of Child Psychology and Psychiatry 36: 203–23.
— Söhne B, Ohrt B, Riegel K. (1995b) Follow-up of preterm
children: important to document dropouts. Lancet 345: 447.
Wright NE, Thislethwaite D, Elton RA, Wilkinson EM, Forfar JO.
(1983) The speech and language developmnent of low birth
weight infants. British Journal of Disorders in Communication
18: 187–96.
Cognitive Status of 6-Year-Old Very Preterm Children Dieter Wolke and Renate Meyer 109