Motor impairment in very preterm-born children: links with other developmental deficits at 5 years of age

Authors

  • Janeline W Van Hus,

    Corresponding author
    1. Department of Rehabilitation, Academic Medical Centre, University of Amsterdam, Amsterdam, the Netherlands
    • Correspondence to Janeline W P Van Hus, Department of Rehabilitation A-01, Academic Medical Centre, University of Amsterdam, Post Box 22660, 1100 DD Amsterdam, the Netherlands. E-mail: j.w.vanhus@amc.uva.nl

    Search for more papers by this author
  • Eva S Potharst,

    1. Psychosocial Department, Academic Medical Centre, Emma's Children's Hospital, University of Amsterdam, Amsterdam, the Netherlands
    2. Department of Developmental Psychology and EMGO, Institute for Health and Care Research, Free (VU) University of Amsterdam, Amsterdam, the Netherlands
    Search for more papers by this author
  • Martine Jeukens-Visser,

    1. Department of Rehabilitation, Academic Medical Centre, University of Amsterdam, Amsterdam, the Netherlands
    Search for more papers by this author
  • Joke H Kok,

    1. Department of Neonatology, Academic Medical Centre, Emma's Children's Hospital, University of Amsterdam, Amsterdam, the Netherlands
    Search for more papers by this author
  • Aleid G Van Wassenaer-Leemhuis

    1. Department of Neonatology, Academic Medical Centre, Emma's Children's Hospital, University of Amsterdam, Amsterdam, the Netherlands
    Search for more papers by this author

Abstract

Aim

To elucidate the relation between motor impairment and other developmental deficits in very preterm-born children without disabling cerebral palsy and term-born comparison children at 5 years of (corrected) age.

Method

In a prospective cohort study, 165 children (81 very preterm-born and 84 term-born) were assessed with the Movement Assessment Battery for Children - 2nd edition, Touwen's neurological examination, the Wechsler Preschool and Primary Scale of Intelligence, processing speed and visuomotor coordination tasks of the Amsterdam Neuropsychological Tasks, and the Strengths and Difficulties Questionnaire.

Results

Motor impairment (≤15th centile) occurred in 32% of the very preterm-born children compared with 11% of their term-born peers (p=0.001). Of the very preterm-born children with motor impairment, 58% had complex minor neurological dysfunctions, 54% had low IQ, 69% had slow processing speed, 58% had visuomotor coordination problems, and 27%, 50%, and 46% had conduct, emotional, and hyperactivity problems respectively. Neurological outcome (odds ratio [OR]=41.7, 95% confidence intervals [CI] 7.5–232.5) and Full-scale IQ (OR=7.3, 95% CI 1.9–27.3) were significantly and independently associated with motor impairment. Processing speed (OR=4.6, 95% CI 1.8–11.6) and attention (OR=3.2, 95% CI 1.3–7.9) were additional variables associated with impaired manual dexterity. These four developmental deficits mediated the relation between preterm birth and motor impairment.

Interpretation

Complex minor neurological dysfunctions, low IQ, slow processing speed, and hyperactivity/inattention should be taken into account when very preterm-born children are referred for motor impairment.

Abbreviations
MND

Minor neurological dysfunction

MABC-2

Movement Assessment Battery for Children - 2nd edition

Deficits on multiple developmental domains occur more often in children born very preterm children (<30wks' gestation and/or birthweight <1000g) than in term-born children.[1] Owing to improved neonatal intensive care, the rate of cerebral palsy (CP) in very preterm-born children has dropped to approximately 5%.[2] However, preterm birth is still associated with significant motor impairment persisting throughout childhood.[3] In several studies, motor impairment has been found in about 30% to 40% of very preterm-born children at 5 years of age.[1, 4-6]

Impaired motor development early in life may affect children's ability to explore their environment and gain experiences, which may in turn result in later cognitive delay, intellectual disability, or behavioural problems.[7-9] In an earlier study on very preterm-born children at age 5, we described that, in addition to motor impairment occurring in 30%, minor neurological dysfunction (MND) occurred in 45%, cognitive problems in 39%, and behaviour problems in 27%.[1]

Spittle et al.[10] found that white matter abnormalities in very preterm-born children predict motor impairment at 5 years of age. Because white matter injury occurs in various areas of the brain,[11] and is often accompanied by damage of the grey matter, corpus callosum, and cerebellum,[12] it is not surprising that, apart from motor impairment, there is also a high frequency of deficits in other domains. Furthermore, Diamond's[13] overview of studies indicated a close functional relation between motor and cognitive development.

Diffuse white matter loss may also be responsible for processing speed decrements in very preterm-born adolescents.[14] Processing speed, the basic speed at which the brain processes information, is thought to underlie academic attainments, executive function, and behaviour,[15] but not much is known about the relation between processing speed and motor function in very preterm-born children. The extent to which motor impairment in very preterm-born children is associated with other deficits is unclear. This information would be of interest from the causal point of view and for treatment.

Assuming that very preterm-born children have more motor impairment and other developmental deficits than term-born children, we hypothesized that (1) especially in very preterm-born children with motor impairment, deficits in neurological outcome, cognition, visuomotor coordination, processing speed, and behaviour are frequently found, and (2) other developmental deficits may, in part, mediate the higher occurrence of motor deficits in very preterm-born children.

Therefore we studied the following: (1) the frequency of motor impairment in very preterm-born children without disabling CP and in term-born children at 5 years of (corrected) age; (2) the frequency of abnormal neurological outcome and deficits in cognition, processing speed, visuomotor coordination, and behaviour in very preterm-born and term-born children and in very preterm-born children with and without motor impairment at 5 years of (corrected) age; (3) the association of motor impairment with these developmental deficits; and (4) whether these developmental deficits mediate the relation between preterm birth and motor impairment.

Method

Participants

Two groups of children participated in the study: the very preterm-born group (children born <30wks' gestation and/or with birthweight <1000g) and the comparison group (children born after 37wks' gestation and with birthweight >2500g). Participants in the very preterm-born group were recruited from a single-centre prospective cohort study as part of the follow-up programme of the Emma's Children's Hospital/Academic Medical Centre, Amsterdam, the Netherlands.[1] Children in the comparison group were recruited from the school or social network of the very preterm-born group or from mainstream schools in the neighbourhood of our hospital. The very preterm-born group reached the corrected age of 5 years between December 2007 and June 2009. The children in the comparison group reached the age of 5 years in the same period. Inclusion criteria for the very preterm-born group were (1) hospitalization in our neonatal intensive care unit, (2) participation at least once in our neonatal follow-up programme, and (3) resident in the Netherlands. Exclusion criteria were (1) participation in other studies (because of the use of different instruments and different timing of follow-up), (2) a genetic syndrome, or (3) unable to participate in an intelligence test because of the extent of their disability. The exclusion criterion of the comparison group was a planned or current referral for learning or behavioural problems.

Procedures

The assessment protocol was similar for both groups. Two appointments were made at the (corrected) age of 5 years at the Academic Medical Centre in Amsterdam. At the first appointment, the child's intelligence and visuomotor coordination were assessed by a trained psychologist. At the second visit, within 3 months of the first visit, motor and neurological tests were performed by a trained paediatrician or paediatric physical therapist, and focused attention and processing speed were assessed by the trained psychologist. The investigators were not blind for birth status. Parents and teachers were asked to fill in a questionnaire about the behaviour of the child. Informed consent to participate in the study was obtained from the parents, and the medical ethics committee of the hospital approved the study. Perinatal and sociodemographic characteristics were taken from the medical records, including education level and country of birth of the mother as variables reflecting socio-economic status because of a possible association between low socio-economic status and cognition.

Measurements

Movement Assessment Battery for Children - second edition (MABC-2)

The MABC-2 (Henderson et al.[16]) is a standardized and normative referenced test, designed to identify impairment of motor functions in children aged 3 to 16 years. Motor outcomes were calculated using the age band 3 to 6 years. Within this age band, eight tasks (items) are grouped under three components: manual dexterity, aiming and catching, and balance. Raw scores of the three components and the total test (the sum of all eight items) are converted to standard scores. Reference means (SD) for the total test and component standard scores are 10 (3). A standard score of less than 7 reflects performances no greater than the 15th centile and is regarded as motor impairment.

The neurological examination according to Touwen

The neurological examination according to Touwen[17] is a standardized and age-specific examination to assess the neurological condition of a child between the ages of 4 years and 12 years and pays special attention to MND. It addresses eight functional clusters: posture, reflexes, involuntary movements, associated movements, coordination problems, fine manipulation disability, sensory deficits, and cranial nerve dysfunctions. Because children with disabling CP were excluded from this study, neurological outcome was considered abnormal if there was either complex MND or non-disabling CP.

Wechsler Preschool and Primary Scale of Intelligence

Intelligence was assessed using the Dutch translation of the third edition of the Wechsler Preschool and Primary Scale of Intelligence.[18] The seven core subtests were administered. The Full-scale IQ was calculated. The reference mean (SD) is 100 (15). The Full-scale IQ was considered abnormal if it was more than 1SD below the mean (<85 points).

Baseline speed task of the Amsterdam Neuropsychological Tasks

Processing speed and consistency of speed were measured using the baseline speed task of the Amsterdam Neuropsychological Tasks.[19] This task requires children to react as fast and as accurately as possible to simple stimuli by hitting a large button. The dependent measure was processing speed, the time a child needed to process the information and generate a motor response. Scores were considered abnormal if they were more than 1SD above the mean of the comparison group. The outcomes of the left and the right hand were combined.

Tracking and pursuit tasks of the Amsterdam Neuropsychological Tasks

Visuomotor coordination was measured using the tracking and pursuit tasks of the Amsterdam Neuropsychological Tasks assessment program.[19] The tracking task requires the child to trace a circle with a mouse cursor. The pursuit task requires tracking a randomly moving target with a mouse cursor. The dependent measure was the distance between the cursor and the circle or target. Scores were considered abnormal if they were more than 1SD above the mean of the comparison group. The outcomes of the left and the right hand were combined.

Strengths and Difficulties Questionnaire

This behaviour questionnaire has five subscales each consisting of five items, and a total difficulty score.[20] For this study, the ‘conduct’, ‘emotional’, and ‘hyperactivity/inattention’ subscales were assessed, using both the parents and teacher form. According to the test manual, a score was classified as abnormal if it was above the 80th centile of the norm score.

Statistical analysis

Data analyses used the computer program SPSS version 20.0 (IBM SPSS Statistics, IBM Corporation, NY, USA). Differences in sociodemographic and perinatal characteristics and the frequency of (abnormal) test outcomes between the very preterm-born and the comparison groups and between the very preterm-born and term-born children with and without motor impairment were analysed using the independent-samples t-test and the χ[2] test, as appropriate. Two-sided p values of <0.05 were considered statistically significant.

To investigate the association between motor impairment and other developmental deficits, four binary logistic regression analyses were performed with dependent variables normal versus abnormal outcome (>15th centile vs ≤15th centile) on the total score of the MABC-2 and the three components. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated. The model consisted of two blocks. The first block contained the variable birth status (preterm/term) and the variables that differed between the groups. All variables were entered to the model at the same time. The second block contained the dichotomous variables (normal/abnormal): neurological outcome, Full-scale IQ, processing speed, tracking and pursuit, conduct, emotional and hyperactivity\inattention behaviour problems. The variables in the second block were by a forward stepwise procedure added to the model when the variables had a p value of 0.05. To derive the robustness of the 95% confidence intervals vested through stepwise logistic regression, additional bootstrapping analyses for the estimated regression coefficient b were done.

Subsequently a mediation model was tested to analyse the extent to which these developmental deficits mediate the association between preterm birth and total motor impairment, and between preterm birth and impaired manual dexterity. Univariate binary logistic regression was analysed for the different pathways. The potential mediators were entered separately to the mediation model, combined with the variables that differed between the groups. The Sobel technique,[21, 22] was used to address whether the effect of birth status on motor impairment was significantly reduced by including a potential mediator.

Results

One hundred and thirty-eight very preterm-born children fulfilled our inclusion criteria. Twenty-three children were excluded: 16 participated in another study, four were unable to participate because of the extent of their disability, and three had a genetic syndrome. Eleven children were lost to follow-up. Participating children (n=104) differed only from non-participants (n=34) in the proportion who were part of twins (27% and 77% respectively, p<0.001). The 95 participating term-born children were recruited from schools attended by very preterm-born children (n=63), were friends (n=14), or family (n=1) of very preterm-born children, or were from schools in the neighbourhood of our hospital (n=17).

Nineteen very preterm-born children were excluded because motor functioning was assessed using a different version (1st edition) of the MABC, and four very preterm-born children and 11 term-born children were excluded because motor functioning was not addressed owing to no show or logistical problems.

One hundred and sixty-five children (81 very preterm-born [40 males, 41 females]; and 84 term-born [34 males, 50 females]) remained in the study. Non-participating very preterm-born children (n=23) did not differ from the participants (n=104) in sociodemographic and perinatal factors.

The sociodemographic factors did not differ between the very preterm-born and comparison group except for two factors: the very preterm-born group comprised more parents born outside the Netherlands (mothers, p=0.011; fathers, p=0.005) and more low-educated mothers (p=0.002; Table 1). The countries/regions of maternal origin were Surinam, Turkey, Morocco, and West Africa. In our analyses we corrected for mothers with low education (<6y post elementary school) and mothers not born in the Netherlands. No correction for fathers born outside the Netherlands was made because of the high correlation with the mothers (Pearson's r=0.839).

Table 1. Perinatal and social background characteristics of very preterm-born and comparison children
CharacteristicsVery preterm-born group (n=81)Comparison group (n=84) p
  1. Values are n (%) except where otherwise stated. Differences in mean scores and proportions between the groups were analysed using t-tests or χ2 tests. aIntraventricular haemorrhage defined according to Papile. bPeriventricular leukomalacia 1 defined according to de Vries et al.[32] There were no cases of periventricular leukomalacia 2–4. cHydrocephalus defined as 4mm>p97 of Levene curves.[33] dLow education defined as <6years post elementary schooling. –, not applicable.

Perinatal characteristics
Mean (SD) gestational age, wk28.7 (1.5)40.0 (1.7)<0.001
Mean (SD) birthweight, g1078.7 (264.2)3448.1 (511.9)<0.001
Sex: male/female, n40/4134/500.250
Part of a multiplet24 (29.6)3 (3.6)<0.001
Small for gestational age25 (30.9) 
Surfactant38 (46.9)  
Inotropics20 (24.7)  
Postnatal dexamethasone3 (3.7) 
Indomethacin for patent ductus arteriosus25 (30.9) 
Requiring ventilation47 (58.0) 
Continuous positive airway pressure (d)17.0 (14.1) 
Oxygen support ≥28d postmenstrual age27 (33.3) 
Oxygen support at 36wks postmenstrual age13 (16.0) 
Sepsis22 (27.2) 
Necrotizing enterocolitis, stage 22 (2.5) 
Subependymal haemorrhage18 (22.2) 
Intraventricular haemorrhagea6 (7.4) 
Grade 23 (3.7) 
Grade 33 (3.7) 
Periventricular leukomalacia 1b3 (3.7) 
Post haemorrhagic hydrocephalusc4 (4.9) 
Social background characteristics
Mean (SD) age of infant at test date, y:m5.2 (0.2)5.2 (0.1)0.614
Mean (SD) maternal age at date birth31 (6.0)31 (4.0)0.880
Mean (SD) paternal age at birth date, y34 (7.0)34 (5.0)0.990
Mother not born in the Netherlands21 (25.9)9 (10.7)0.011
Father not born in the Netherlands

81/81

19 (25.0)

82/84

7 (8.4)

0.005
Maternal low educationd

n=79/81

21 (26.6)

n=84/84

7 (8.3)

0.002
Paternal low education

n=73/81

22 (30.1)

n=81/84

16 (19.8)

0.136

Test outcomes

A significant difference in motor impairment was found between the very preterm-born (32.1%) and the comparison group (10.7%; p=0.001; Table 2). On the MABC-2 components, significant differences between the groups were found for manual dexterity (very preterm-born 69.4% vs comparison group 30.6%, p=0.006). No significant differences were found on balance (very preterm-born 18.5% vs comparison group 9.5%, p=0.095) and aiming and catching (very preterm-born 25.9% vs comparison group 23.8%, p=0.753). Significant differences between both groups were found on all other tests to the disadvantage of the very preterm-born group (Table 2). Within the very preterm-born group, children with motor impairment had significantly more complex MND, low IQ, slow processing speed, and abnormal visuomotor coordination than children with normal motor outcome. Behavioural problems were comparable between very preterm-born children with and without motor impairment (Table 2). Within the comparison group, no significant differences between children with and without motor impairment were found (Table 2).

Table 2. Comparison of multi-domain outcomes between very preterm-born (VPT) and comparison groups, and between VPT children with and without motor impairment
 VP group (n=81)Comparison group (n=84) p a VPT group MABC-2 ≤15th centile n=26VPT group MABC-2 >15th centile n=55 p b Comparison group MABC-2 ≤15th centile n=9Comparison group MABC-2 >15th centile n=75 p c
  1. Differences in proportions and mean scores between the groups were analysed using χ2 test, t-test, or independent sample t-test. ap value between very preterm born (VPT) and comparison group. bp value, VPT group with and without motor impairment. cp value, comparison group with and without motor impairment: MABC-2 ≤15th centile = abnormal. dIn the 18 VPT children with abnormal neurological outcome, 16 had complex minor neurological dysfunction and two non-disabling cerebral palsy. Both term-born children had complex minor neurological dysfunction. MABC-2, Movement Assessment Battery for Children - 2nd edition; WPPSI, Wechsler Preschool and Primary Scale of Intelligence; ANT, Amsterdam Neuropsychological Tasks; SDQ, Strengths and Difficulties Questionnaire.

Motor development (MABC-2)
Mean total test (SD)8.37 (3.3)10.04 (2.6)<0.001
Mean manual dexterity (SD)8.02 (3.1)9.77 (2.4)<0.001
Mean aiming and catching (SD)8.35 (3.5)9.17 (3.3)0.120
Mean balance (SD)9.59 (3.6)11.32 (3.2)0.001
Abnormal MABC-2, n (%)26 (32.1)9 (10.7)0.001
Neurological examination (Touwen)
Normal, n (%)42 (51.9)71 (84.5)<0.001
Complex minor neurological dysfunction/non-disabling CP,d n (%)18 (22.2)2 (2.4)<0.00115 (57.7)3 (5.5)<0.0012 (22.2)0 (00.0)0.246
Intelligence (WPPSI)
Mean Full-scale IQ (SD)92.09 (17.5)103.39 (11.4)<0.001   
Full-scale IQ<85 points, n (%)21 (25.9)2 (2.4)<0.00114 (53.8)7 (12.7)<0.0011 (11.1)1 (1.3)0.069
Processing speed (ANT)         
Mean baseline speed reaction time (SD)677.28 (191.8)575.54 (108.7)<0.001   
Abnormal, n (%)29 (35.8)12 (14.3)0.00218 (69.2)11 (20.0)<0.0011 (11.1)11 (14.9)0.762
Visuomotor coordination (ANT)
Mean tracking distance (SD)13.3 (5.9)10.65 (3.3)0.001   
Abnormal tracking, n (%)27 (33.3)13 (15.5)0.00617 (65.4)10 (18.2)<0.0010 (00.0)13 (17.3)0.171
Mean pursuit distance (SD)7.91 (4.9)5.64 (2.2)<0.001   
Abnormal pursuit, n (%)24 (29.6)11 (13.1)0.00713 (50.0)11 (20.0)0.0021 (11.1)10 (13.3)0.852
Behaviour (SDQ)
Conduct problems, n (%)18 (22.2)10 (11.9)0.0787 (26.9)11 (20.0)0.4841 (11.1)9 (12.0)0.938
Emotional symptoms, n (%)30 (37.0)13 (15.5)0.00213 (50.0)17 (30.9)0.0970 (00.0)13 (17.3)0.174
Hyperactivity/inattention problems, n (%)32 (39.5)13 (15.5)0.00112 (46.2)20 (36.4)0.4003 (33.3)10 (13.3)0.117

In the 26 very preterm-born children with motor impairment, motor impairment always co-occurred with one or more other abnormal test outcomes, whereas in four of the nine term-born children with motor impairment all other test outcomes were normal. Moreover, 18 very preterm-born children with motor impairment had abnormal outcomes on three to five different developmental tests, whereas this did not occur in term-born children with motor impairment.

Multivariate associations and mediation effects

We investigated which of the other deficits were most associated with motor impairment, taking preterm birth status into account and correcting for mothers with low education and mothers not born in the Netherlands. Motor impairment (total score) and an abnormal score on the component manual dexterity were significantly associated with preterm birth. For impaired aiming and catching, and balance, no significant association with preterm birth was found (Table 3). The significant association between preterm birth status and motor impairment (total score) and impaired manual dexterity disappeared when other developmental deficits were entered into the model. Motor impairment was significantly associated with complex MND (OR=41.7, 95% CI 7.5–232.5) and low IQ (OR=7.3, 95% CI 1.9–27.3). Impaired manual dexterity was significantly associated with low IQ (OR=4.5, 95% CI 1.4–14.8), slow processing speed (OR=4.6, 95% CI 1.8–11.6), and hyperactivity/inattention (OR=3.2, 95% CI 1.3–7.9).

Table 3. Associations with motor impairment in total group
Dependent variablesIndependent variables b 95% CI for b p Exp (B) (OR)95% CI for exp (B)
  1. Associations between motor impairment and other developmental deficits were analysed using four binary logistic regressions. The variable preterm/term was included in the model as a constant factor, the variables complex minor neurological dysfunction (MND), IQ<85 points, slow processing speed, hyperactivity/inattention, conduct problems, and emotional symptoms were entered by a forward stepwise procedure. Data are presented as b (regression coefficient), bootstrapped confidence intervals for b and p values, and exp (B) (odds ratio [OR]) and confidence intervals for exp (B). Outcomes were adjusted for low-educated mothers (<6y post elementary schooling) and mothers not born in the Netherlands. MABC-2, Movement Assessment Battery for Children - 2nd edition.

MABC-2: total score ≤15th centileBlock 1 
Preterm/term1.3210.6–2.40.0023.71.6–8.9
Block 2 
Preterm/term0.171−1.2 to 1.50.7561.20.4–3.6
Complex MND3.7302.5–23.80.00141.77.5–232.5
IQ<85 points1.9870.6–4.00.0017.31.9–27.3
MABC-2: manual dexterityBlock 1 
Preterm/term0.9960.2–2.00.0152.71.2–6.2
Block 2 
Preterm/term0.178−0.8 to 1.30.7271.20.5–3.2
IQ<85 points1.5080.3–3.30.0134.51.4–14.8
Slow processing speed1.5230.6–2.70.0014.61.8–11.6
Hyperactivity/inattention1.1510.2–2.20.0143.21.3–7.9
MABC-2: aim and catchBlock 1 
Preterm/term0.178−0.6 to 0.90.6181.20.6–2.5
Block 2
Preterm/term−0.227−1.2 to 0.60.5790.80.4–1.8
Complex MND1.7130.5–3.20.0025.51.8–16.9
MABC-2: balanceBlock 1 
Preterm/term0.662−0.3 to 1.80.1871.90.7–4.9
Block 2 
Preterm/term−0.641−2.5 to 0.80.3330.60.2–2.1
Complex MND2.4531.1–4.90.0019.62.6–35.4
IQ<85 points1.5810.0–3.40.0115.01.4–17.5

In the mediation model (Fig. 1) the effect of preterm birth on motor impairment, the direct pathway, became non-significant after controlling for the effects of the indirect pathway (Table 4). The association between preterm birth and motor impairment was mediated by complex MND and low IQ. The effect of preterm birth on impaired manual dexterity was mediated by low IQ, slow processing speed, and hyperactivity/inattention (Table 4). In addition, Sobel's (Z) test achieved significance (p<0.05), indicating a significant mediation between birth status and motor impairment and between birth status and impaired manual dexterity (Table 4).

Figure 1.

Model of direct (path C) and indirect (paths A and B) effects of preterm birth on motor impairment, corrected for low-educated mothers and mothers not born in the Netherlands.

Table 4. Results of mediation analysis for motor impairment in very preterm-born children
Motor impairment (total)Path APath BPath CTest of mediationPath C (adjusted for indirect path)
a (Sa) p b (Sb) p c (Sc) p Z p c (Sc) p
  1. Path A: effect of preterm birth on the mediators complex minor neurological dysfunction (MND),/IQ<85 points, slow processing speed,/hyperactivity/inattention. Path B: effect of the mediators on motor impairment/impaired manual dexterity. Path C: effect of preterm birth on motor impairment/impaired manual dexterity. Path C (adjusted for indirect path): effect of preterm birth on motor impairment/impaired manual dexterity adjusted for indirect path by path A and path B. a, b, c: unstandardized path coefficients. Sa, Sb, Sc: standard errors of the path coefficients. All outcomes are corrected for low-education mothers and mothers not born in the Netherlands.

Complex MND3.067 (1.05)0.0044.061 (0.82)<0.0011.321 (0.44)0.0032.520.0110.591 (0.51)0.248
IQ<85 points2.381 (0.78)0.0022.412 (0.54)<0.0011.321 (0.44)0.0032.520.0120.880 (0.47)0.065
Impaired manual dexterity
IQ<85 points2.381 (0.78)0.0022.305 (0.53)<0.0010.996 (0.42)0.0182.500.0120.529 (0.46)0.252
Slow processing speed0.993 (0.41)0.0151.884 (0.42)<0.0010.996 (0.42)0.0182.130.0330.731 (0.45)0.103
Hyperactivity/inattention1.217 (0.39)0.0021.379 (0.41)0.0010.996 (0.42)0.0182.190.0220.733 (0.44)0.096

Discussion

Our study confirms the high frequency both of motor impairment and of deficits in other developmental domains in very preterm-born children at 5 years' corrected age. The motor impairment rate (32.1%) was comparable to an earlier Dutch study of very preterm-born children,[4] and to studies from other countries.[5, 6] Contrary to the other studies in which problems on all MABC components were described, we found especially worse outcomes on manual dexterity and balance and less on aiming and catching. This might be because our comparison group had a standard score of 9 in aiming and catching, which is lower than the reference mean of 10.

Very preterm-born children with motor impairment had a substantially higher rate of other abnormal test outcomes than very preterm-born children without motor impairment. The frequency of complex MND, low intelligence, slow processing speed, and visuomotor coordination problems occurred in more than 50% of very preterm-born children with motor impairment. We defined motor impairments as not more than the 15th centile on the MABC-2, including children with mild–moderate motor impairment, because research has suggested that very preterm-born children whose scores fall between the 6th and 15th centiles are at significant risk for associated problems in learning, attention, and psychosocial adjustment.[23] Indeed, this mildly to moderately impaired motor group had a worse developmental profile. These impairments may co-occur because of the possible underlying white matter damage, causing multiple impairments.[10, 11]

Although the very preterm-born children had significantly more motor impairments than term-born children, this difference disappeared when other developmental deficits were also taken into account in the analyses. Complex MND and low intelligence were found to mediate between preterm birth and motor impairment. Low intelligence, slow processing speed, and hyperactivity/inattention played a mediating role between preterm birth and manual dexterity problems.

Complex MND can be considered as a distinct form of perinatally acquired brain dysfunction, which is probably associated with the cortico-striato-thalamo-cortical and cerebello-thalamo-cortical pathways.[24] According to Volpe,[11] these are the circuitries that are often damaged in very preterm-born children. They play a role in sensorimotor aspects of motor programming, movement planning, programme selection and motor memory, and in cognitive tasks. In line with this, Diamond[13] described the interrelation between cognition and motor performance, and the brain areas involved with both functions simultaneously, namely the dorsolateral prefrontal cortex, the cerebellum, and the caudate nucleus. Damage to these circuitries can result in both cognitive and motor impairment.

In addition to the simultaneous involvement of brain structures in cognition and motor performance, the assessment of motor abilities motor abilities also includes functions other than motor performance and these are tested in part as well. Although the MABC-2 is an effective test in identifying motor impairment in the high-risk population of very preterm-born children,[25] cognition, speed, motor planning, spatial precision, and behavioural adaptation to the test situation are also required for normal outcome. In general, it is almost impossible to measure one developmental domain without tapping into other integrated functions.

Although all very preterm-born children with motor impairment also had deficits in other developmental domains, this was not true for the term-born children with motor impairment. In the comparison group, 45% of children with motor impairment had no additional deficits. In term-born children, the cause of motor impairment might be restricted or more isolated.

We found a high frequency of behavioural problems in the very preterm-born group irrespective of motor impairment. Nevertheless, hyperactivity/inattention was associated with impairments in manual dexterity. The relation between manual dexterity and attention has been described in children with attention-deficit–hyperactivity disorder.[26] Furthermore, it is known that very preterm-born children have an increased risk for this disorder.[27]

Of the very preterm-born children with motor impairment, 69% had problems with processing speed. Several studies have shown that processing speed is a basic key ability underlying deficits in intelligence in very preterm-born children,[14, 15, 28, 29] linking reduction of white matter integrity to slow processing speed. We hypothesize that processing speed also plays an important role in motor performance, especially when motor skills become more complex at age 5. Indeed in our study, processing speed was particularly associated with manual dexterity.

Problems in visuomotor coordination were found in 65% (tracking task) and 50% (pursuit task) respectively of the very preterm-born children with motor impairment. This may have been caused by cerebellar damage, by attention difficulties arising from damage to the prefrontal cortex, or by an impaired ability to process and comprehend the visual input.[30] In our multivariate model, however, these visuomotor skills were no longer associated with motor impairment.

A limitation of the study was that the investigators were not blind to birth status. However, highly standardized testing rules were followed in all children to reduce this shortcoming. Further, we included term-born children free from planned or current referral for learning or behavioural problems. A non-selected group of term-born peers might have led to different results. We did not study the effect of neonatal morbidities further, because our focus was the comparison between term and very preterm-born children. Bonifacio et al.[31] showed that brain injury and neonatal comorbidities, and not preterm birth per se, are associated with abnormal brain development of brain microstructure, suggesting that preterm infants with these morbidities are especially at risk of motor problems.

Because motor outcome was the main topic of our paper, we did not study how it might mediate other deficits. Thus, our results focus on interrelations, which of course can be bidirectional.

Using a broad assessment on different developmental domains is a strength of this study. It is also important from the clinical point of view as those treating very preterm-born children with motor impairment should be aware of the interrelation of problems in other developmental domains.

Conclusion

In the absence of disabling CP, motor impairment occurs more frequently in very preterm-born than in term-born children at 5 years of (corrected) age. Very preterm-born children with motor impairment more often have complex MND and impairments in cognition, processing speed, and visuomotor coordination than very preterm-born children without motor impairment; however, behavioural problems are comparable for both groups. Complex MND, low intelligence, slow processing speed, and hyperactivity/inattention mediate the association between preterm birth and motor impairment. These deficits should be taken into account when very preterm-born children with motor impairment are referred for intervention.

Ancillary