Infant motor milestones: poor predictive value for outcome of healthy children

Abstract

The acquisition of motor skills is among the most remarkable achievements in the first years of life. Motor milestones such as the emergence of sitting without support or the first independent steps provide a framework for developmental monitoring of children in health supervision visits because these milestones belong to the most salient and best demarcated markers of developmental processes that parents and healthcare professionals can observe. In clinical practice, general paediatricians, child neurologists and developmental paediatricians are often asked to predict future outcomes on the basis of early developmental milestones. While there is some evidence from the study of preterms or children who have experienced early brain injury that a delay in motor development is a predictor of subsequent motor and cognitive problems [see e.g. 1], the association between infant motor milestones and later developmental outcome in normal healthy children without risks is less clear. Several papers have been recently published assessing this relationship in large-scale birth cohorts [e.g. 1946 National Survey of Health and Development in the United Kingdom 2 or Northern Finland 1966 Birth Cohort 3, 4]. Although these studies described an association between early motor milestones and later developmental outcome, the magnitude of the reported effects was generally small and the findings were questioned by others 5. In fact, infant milestone information in these studies was acquired either at age 12 months (before all children were able to walk independently) or at 24 months (past the time most children learned to sit without support or made their first steps), and later cognitive functions were not measured with existing standards 5. In our longitudinal study of normally developing children, milestones were recorded at a time point contemporaneous with the emergence of motor milestones and later outcomes were repeatedly assessed up to 18 years using standardized intelligence and neuromotor tests. Our cohort consisted of 222 healthy, term-born individuals [119 boys, 103 girls, median gestational age 40 weeks, MAD (median absolute deviation) 1.2 weeks, range 37–44 weeks; median birth weight 3.26 kg, MAD 0.4 kg, range 2.37–4.57 kg] of the Zurich Generational Study [community sample of children born between 1978 and 1993 6]. Participants did not have mental retardation [full-scale intelligence quotient (IQ) < 70] or a motor disorder that required therapeutic interventions (physiotherapy, occupational therapy, etc.). The mean socioeconomic status (SES) according to Largo et al. 7 was 8 (range 3–12). Informed consent was obtained at birth from the parents. The institutional review board of the Canton Zurich confirmed that the study was performed according to the Declaration of Helsinki and conformed to legal and ethical norms. All children were followed with comprehensive neurodevelopmental assessments from birth to age 18 years by experienced developmental paediatricians (visits at the centre at ages 1, 3, 6, 9, 12, 18 and 24 months, and then in yearly intervals). Depending on the outcome variable considered, between 79% and 85% of children with a milestone record participated in the endpoint visit at age 18 years, and these children were representative of those included at baseline (i.e. same age distribution for milestones). Parents were asked to record the date when their child achieved motor milestones. At each visit, the paediatrician obtained the date from the parent. For the current analysis, the motor milestones ‘sitting without support’ (child sits up and does not use arm or hands to support position) and ‘independent walking’ (at least 5 steps in upright position without support) were used. The ZNA was performed at ages 7, 8, 10, 12, 15 and 18 years. The ZNA is a standardized motor test for assessing timed performance of 12 motor tasks and the quality of movements (i.e. intensity of associated movements) 6, 8. Here, we used motor scores measuring contralateral associated movements (CAM) and timed performance of repetitive, alternating and sequential tasks of fingers, hands and feet (HFT), the pegboard task, dynamic balance (DB, side and forward jumping) and static balance [for more details, see 6, 8]. Intelligence quotient was assessed at ages 7, 9, 11, 13, 14, 16 and 18 years by three standardized instruments: the German version of the Wechsler Preschool and Primary Scale of Intelligence WIPPSI at age 7 years; the German version of the Wechsler Intelligence Test for Children – revised version WISC-R at ages 9, 14 and 18 years; and the German-based Adaptives Intelligenz Diagnostikum (AID) at ages 11, 13 and 16 years. Neither motor nor cognitive performances showed evidence for a selection bias (informative dropout) that would manifest as a consistent shift or trend in the data across age. To remove possible IQ test and random age effects, we restandardized IQ scores to get a median of 100 and a MAD of 15 at each age. For comparability, we did the same for the motor scores of the ZNA. The distributions of ages at which children started to sit without support and walk independently were skewed to the right. Therefore, a logarithmic transformation was used to obtain more symmetric distributions. Transformed milestone data were then standardized using their respective means and standard deviations, and their sign was inverted so that a positive value referred to a milestone achieved earlier than the median age. Standardized milestone data were used as predictors in linear regression models to predict intellectual and motor performances for each age separately, controlling for the effects of gender (possibly interacting with milestone) and SES. However, we observed no consistent trend over age; the milestones did not predict intellectual or motor functions better during childhood than during adolescence, nor was the converse true. Therefore, we pooled the data from all ages and used a mixed model to account for the repeated IQ/motor measurements available for each child. A simple random intercept model was fitted to the data in which a given IQ or motor outcome was used as the response and a given milestone as a predictor, controlling for gender and SES as described above. Under this model, the difference in IQ/motor performances between two assessments on the same child is assumed purely random. The model describes the association between milestones and average IQ/motor performances over the period of 7 to 18 years. The milestone coefficient measures the number of IQ/motor points gained (if positive) or lost (if negative) that corresponds to a shift of one SD on the transformed milestone scale after adjusting for gender and SES. The children achieved the milestones ‘sitting without support’ at a median age of 6.5 months (interquartile range 1.5 months; 10th percentile at 5 months; 90th at 9 months, range 3.75–13 months) and ‘independent walking’ at a median age of 12 months (interquartile range 2.5 months, 10th percentile at 10.5 months, 90th at 16 months, range 8.5–20 months). The age of achievement of a motor milestone was not related to the overall motor and cognitive performance between 7 and 18 years (Table 1), except for the age of independent walking, which was weakly associated with performance in the dynamic balance tasks (Fig. 1). After adjusting for gender and SES, children who could walk earlier by one standard deviation (SD) on the standardized milestone scale scored on average 1.7 points (or roughly 1/10 of a SD) higher in the DB task (p = 0.023). Alternatively, these 1.7 points represent the expected gain in DB performance for a child who started to walk 1.27 month earlier than the median age of 12 months. As expected, SES had a large effect on verbal and full-scale IQ while only a small effect (most of the time non significant) on performance IQ and motor scores was observed. No consistent gender differences were found. A subsequent analysis (not shown) comparing extreme groups (earliest 20% and latest 20%) on the milestone scale did not reveal any association between the age of milestone achievement and developmental outcomes. Similarly, the distribution of the age of milestone achievement was not different between extreme groups when these were defined on the basis of average cognitive or motor performances (best 20% and poorest 20%). We found that 80 per cent of the children of our cohort were able to sit without support between 5 and 9 months of age and walk independently between 10.5 and 16 months, which is in line with previous investigations 9. The age of reaching a milestone was not associated with later developmental outcome, except for the milestone of independent walking, which was weakly related to later gross motor functions in the dynamic balance tasks of the ZNA. This finding is corroborated by an earlier study reporting that gross motor milestones are weakly associated with a higher frequency of sports participation during school age 10. In contrast, we did not find an association of milestones with later cognitive performance, which has been reported by others 2-4, 11. While previous authors have hypothesized about the involvement of common underlying neural systems that govern motor development in infancy and also contribute to later motor and cognitive control 2, our analyses rather indicate that early motor development is not essentially related to later outcomes in healthy children. This view is in agreement with a recent study concluding that the stability of motor performance from birth to school age is generally low, which leads to poor prognostic validity of early motor milestones 12. In fact, several comprehensive theories have articulated that motor development is primarily driven by self-organized and dynamic processes that are shaped through a complex series of interactions between genetic programmes and experience-based selections of movement patterns instead of predetermined genetic sequences that occur as children mature to adulthood 13-15. These developmental theories may explain the poor prediction of outcomes obtained from early motor development. Our study shows that the range of attaining early motor milestones in a normal cohort of children is large. While infant motor development may predict the outcome of children with developmental or neurological risks [e.g. premature born children 1], in healthy children with an uneventful medical history and no other clinical findings the age of achievement of motor milestones has only limited clinical significance and practical relevance. Our findings, however, should not lead to the conclusion that developmental surveillance in the early years is not important. We emphasize that the first achievement of motor milestones beyond a certain age (e.g. walking after 18 or 20 months) has to be further explored by a thorough clinical and developmental examination. We thank Prof. Dr. Remo Largo, PD Dr. Tanja Kakebeeke and PD Dr. Bea Latal for helpful comments and Dr. Leila Tarokh for proofreading the manuscript. This study was supported by the Swiss National Science Foundation, grant no 32473B_129956. The authors have no conflicts of interest to declare.