PEDIATRICS Vol. 120 No. 4 October 2007, pp. e1051-e1058 (doi:10.1542/10.1542/peds.2006-2295)
ARTICLE |
Weight Faltering in Infancy and IQ Levels at 8 Years in the Avon Longitudinal Study of Parents and Children
a Centre for Child and Adolescent Health, Department of Community-Based Medicine
b Departments of Clinical Sciences (South)
c Social Medicine, University of Bristol, Bristol, United Kingdom
d Department of Psychology, University of Durham, Durham, United Kingdom
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
OBJECTIVE. Our goal was to investigate the association between failure to thrive (defined as weight faltering in the first 9 months of life) and IQ levels 8 years later.
METHODS. Weight gain (conditional on initial weight) from birth to 8 weeks, 8 weeks to 9 months, and birth to 9 months was measured on term infants from the Avon Longitudinal Study of Parents and Children. Cases of weight faltering were defined as those infants with a conditional weight gain below the 5th centile who were compared with the rest of the cohort as the control group. At the age of 8 years, 5771 infants born at term with no major congenital abnormalities had IQ measured by using the Wechsler Intelligence Scale for Children, Third Revision.
RESULTS. Mean (SD) IQ scores were 104.7 (16.3) (total), 107.6 (16.5) (verbal), and 100.2 (16.9) (performance). Children whose weight faltered from birth to 9 months had a total IQ that was significantly lower by an average of –2.71 points at 8 years, equivalent to 0.17 SD. Weight gain from birth to 8 weeks had a positive linear association with child IQ at 8 years. This remained significant in a multivariate regression despite controlling for correlates of both infant growth and child IQ; 1 SD of weight gain was associated with a difference of 0.84 points in the total IQ score. In contrast to early weight faltering, weight gain from 8 weeks to 9 months was not related to IQ at 8 years.
CONCLUSIONS. Failure to thrive in infancy was associated with persisting deficits in IQ at 8 years; the critical period for growth faltering was birth to 8 weeks. The relationship between infant growth from birth to 8 weeks and later intellectual development was approximately linear over the whole range of weight velocities.
Key Words: infant growth • weight faltering • growth faltering • failure to thrive • intellectual performance • IQ
Abbreviations: FTT—failure to thrive • CI—confidence interval • ALSPAC—Avon Longitudinal Study of Parents and Children
Failure to thrive (FTT) is a term used to describe young children who are not growing as well as expected.1,2 Because the most objective clinical finding associated with the condition of FTT is poor weight gain,3 in many ways the term "weight faltering" is preferable for these infants,4 because it avoids the pejorative use of the word "failure."
There is strong evidence that weight faltering is associated with developmental delay in young children,5–7 but the evidence that it is associated with later intellectual deficits is less consistent.8–10 A recent meta-analysis of controlled studies11 found a weighted average deficit of 
4 IQ points (95% confidence interval [CI]: –2 to –6). This is sufficiently large to be of importance at a population level, but many of the studies reviewed were small, and in observational studies it is difficult to be sure that all relevant covariates have been appropriately controlled for.
However, it is not clear whether there is a threshold for the association of poor growth with intellectual impairment, nor whether there is a sensitive period when weight faltering has a particular impact on cognitive development. Low birth weight in term infants is also associated with IQ deficits, but birth weight is in fact related to IQ across its whole range,12 with no obvious discontinuity at 2500 g. Although it is widely believed that malnutrition early in life might have particularly adverse impacts on intellectual development, the empirical evidence for this is mixed.12–15 It is not clear whether the most important factor for long-term cognition is the timing of the insult (ie, is there a "sensitive period"), or the severity of postnatal undernutrition.
The Avon Longitudinal Study of Parents and Children (ALSPAC),16 a large contemporaneous British birth cohort, provides an opportunity to clarify some of these questions. We previously described the familial, social, and prenatal factors,17 and the postnatal and feeding factors18 associated with FTT in this cohort. The specific aims of this study were to use the ALSPAC cohort to investigate the association between poor weight gain in infancy and later intellectual deficits, with careful controlling for confounding variables. The hypotheses to be tested were first that there is a sensitive period in early infancy when poor weight gain is associated with subsequent cognitive impairments, and second that there is a threshold of weight gain in infancy above which there is no impact on subsequent IQ.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
All births to women resident in the former Avon Health Authority area, with an expected date of delivery between April 1, 1991, and December 31, 1992, were eligible for enrollment in the ALSPAC, and 85% of the known births from the geographically defined catchment area were included, resulting in a total cohort of 14062 live births. Avon is a mixture of rural and urban areas, and the study population has social and demographic characteristics similar to United Kingdom national census surveys.19
Weight data were extracted from the Avon Child Health Computer system, using measurements made as part of the local preschool-aged child health surveillance program. Measurements were taken at birth, at 8 weeks (range: 1–3 months), and at 9 months (range: 6–12 months). All weights were standardized to z scores adjusting for differences in gender and age (gestational age in weeks for weight at birth and infant age in weeks for subsequent weights). Growth was assessed by calculating the difference in z scores between 2 time points and adjusting for regression toward the mean using correlates provided by the British 1990 growth reference.20 The initial weight is used to determine the rate of growth rather than where the infant lies on the growth chart. Weight gain, therefore, was "conditional" on gender, age, and initial weight. This increasingly used technique more accurately reflects infant growth as it accounts for the smaller infants who tend to grow faster and the larger infants who tend to grow slower.21
The "cases" of weight faltering were defined as those infants below the 5th centile for weight gain, which corresponded with a conditional growth score of –1.645. The controls were all the other infants in the cohort above the 5th centile.
Background socioeconomic and health data were collected by using a series of postal questionnaires completed by both the mother and partner in pregnancy and at 4 weeks, 6 months, and 15 months postdelivery. We previously tested >60 factors that were potentially associated with failure to thrive,17,18 including prenatal factors, family and infant characteristics, postnatal events, and problems with infant feeding. Factors that were statistically significant in these multivariate analyses were used in the following analyses.
At the age of 8 years, all surviving children were invited to a research clinic where their IQ was measured by using a validated age-adjusted shortened form of the Wechsler Intelligence Scale for Children, Third Revision, United Kingdom (Psychological Corporation, London, United Kingdom), using alternate items from each subtest.22 The IQ tests were conducted over a 2-year period by using 31 qualified testers who were blinded to whether the child was a case of growth faltering.
Normal distributions were described by using the mean and SD and other distributions using medians and interquartile ranges. If a standard cutoff point for a continuous variable was not available, we examined the approximate lowest quartile of the cohort (or quintile if both ends of the distribution were of interest). The difference in means was tested by using the unpaired t test for 2 means and analysis of variance for several means. The test of linearity was conducted by using the F statistic on 1 degree of freedom. The difference in medians was tested using the Mann Whitney U test. Confidence intervals for single proportions were calculated by using Wilson's method.22 Multivariate analysis was conducted by using a linear regression model with SAS (SAS Institute, Inc, Cary, NC), the dependent variable being the total Wechsler Intelligence Scale for Children IQ score.
Ethical permission for the ALSPAC was granted by the research ethics committees of United Bristol Healthcare Trust and the Frenchay and Southmead health care trusts, and the study is monitored by the ALSPAC Ethics and Law Advisory Committee.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Ascertainment
Of the 13970 children still in the ALSPAC at 1 year, 89 infants (0.6%) with major congenital abnormality likely to affect growth (eg, cleft palate, cerebral palsy, Down syndrome, congenital heart disease) were excluded. At 8 years, all surviving children in the study were invited to a research clinic: 7188 children (51% of original birth cohort) had their IQ measured, and 5771 were term infants with weight measurements available at birth and at 8 weeks and 9 months of age. There was no difference in the birth weight or subsequent growth patterns of those who had IQ measured and those who did not.
An element of selection bias was inevitable when trying to conduct intensive measurements with a cohort 8 years after recruitment. In comparison to those tested, children who were not assessed at 8 years were more likely to come from families from social class IV or V (23% vs 14%), with low maternal educational level (27% vs 13%), and insecure housing tenure (30% vs 16%). Thus, the cohort seen at 8 years overrepresents the better-off families; however, a spectrum of social background was obtained, with 900 of our study children coming from households with 
1 of these 3 markers of socioeconomic deprivation. Also, we previously showed17 that although poor weight gain in infancy was more common among the more deprived families, socioeconomic factors did not have independent effects on velocity of growth, even at the univariate level.
Univariate Analysis
Infant Growth
The mean (SD) birth weight of the 5771 infants at birth was 3840 (410) g. Corrected for gestational age and gender, the weight z scores at birth, 8 weeks, and 9 months were normally distributed, with a mean just above 0 and a SD of 1. The weight-gain z scores of the infants were also normally distributed, from birth to 8 weeks (mean weight-gain z score: –0.03; SD: 0.95), from 8 weeks to 9 months (mean weight-gain z score: 0.16; SD: 1.10), and from birth to 9 months (mean weight-gain z score: 0.11; SD: 1.04). Centiles were constructed for weight gain between birth and 8 weeks, 8 weeks and 9 months, and birth and 9 months, and infants below the 5th centile were defined as "cases" of weight faltering (FTT).
Factors found previously17,18 to be significant multivariate predictors of failure to thrive in the larger ALSPAC cohort (see Appendices 1 and 2) were also significant in this reduced cohort.
|
|
Child IQ at 8 Years Old
The total IQ scores were normally distributed, ranging from 51 to 151 with a mean score of 104.7 (SD: 16.3). Verbal scores (mean: 107.6; SD: 16.5) were relatively higher than performance scores (mean: 100.2; SD: 16.9). The majority of children (92.8%) were measured between 8 and 9 years old, and all but a few (0.3%) before their tenth birthday. The age-adjusted scores clearly show a decline in IQ scores with age of testing. Strong predictors of IQ such as maternal education and paternal social class also declined over age of testing, suggesting that recall for testing was easier for children of better educated families. The decline in IQ of time also partially explained the significant differences in the mean IQ scores between testers and for different months of the year. These factors (age at testing, month of year and tester) were adjusted for in the multivariate analysis. Correlates of child IQ scores are shown in Table 1.
|
Infant Growth and Child IQ
There was a linear relationship between early growth and subsequent IQ. The association was more marked for the verbal tests, but still significant for the performance tests. Figure 1 plots the mean total IQ of the children at 8 years old by weight-gain z score from birth to 8 weeks.
|
The increase in child IQ scores with improved weight gain is significant (P < .0001), as is the linear relationship (F1= 36.09; P < .0001). In this univariate analysis, the IQ of children who showed faltering growth early in infancy was
3 points lower than the average IQ measured at 8 years old, whereas in those with good weight gain early in infancy it was 3 points higher. Growth later in infancy did not seem to be related to subsequent IQ. Figure 2 plots the mean total IQ of the children at 8 years old by weight-gain z score from 8 weeks to 9 months: the distribution shows no obvious pattern (P = .52).
|
For the longer period from birth to 9 months, the mean IQ scores were remarkably consistent, not deviating by > 1 point except for those infants with the slowest weight gain. These are infants who would meet traditional criteria for failure to thrive, and they differed significantly (P = .02) from the rest of the cohort by 4 or 5 IQ points. Figure 3 plots the mean total IQ of the children at 8 years old by weight gain from birth to 9 months.
|
Multivariate Analysis
Using a regression model with the child IQ score as the dependent variable, the significance of infant weight gain for the 3 different time periods was analyzed after adjusting for correlates of both infant growth and childhood IQ (Table 2).
|
The unadjusted effect of weight gain between birth and 9 months on IQ was –3.71 points. After adjusting for confounders, the relationship between weight faltering in the first 9 months and subsequent IQ remained significant (P = .008), with a reduction in total IQ score of –2.71 points (95% CI: –0.69 to –4.73). This association was explained by growth during the first 8 weeks, and not by growth between 8 weeks and 9 months. After adjustment, the positive linear relationship between weight gain from birth to 8 weeks and child IQ 8 years later remained significant (P = .0002); the adjusted association remained stronger with verbal IQ (P = .0001) than performance IQ (P = .007). One SD of weight gain equated to a total IQ score adjustment of 0.84 points (95% CI: 0.39 to 1.29).
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The results of this study support the hypothesis that the critical period for the effect of undernutrition on cognitive development is early in infancy. Multivariate analysis shows that early growth faltering (slowest gaining 5% of term infants in the first 8 weeks) is associated with an average deficit of
3 IQ points. This is consistent with the estimate (4 IQ points; 95% CI: 2–6) provided by a previous meta-analysis,11 which was based on 502 cases and 523 controls, to which this article now adds a additional 258 cases and 5510 controls. A recent systematic review of the long-term outcome of failure to thrive24 found a similar effect size on IQ: although this effect is not large (0.2 SD), it is comparable in size to effects of other identified environmental variables that affect IQ, such as bottle-feeding, prenatal cocaine use, and low birth weight.11,25 Our results confirm previous studies in the United States26 and the United Kingdom6 suggesting that it is early failure to thrive that is associated with delayed development. The results do not, however, provide evidence for the hypothesis that there is a threshold effect, because the relationship between infant growth from birth to 8 weeks and later intellectual development is approximately linear over the whole range of weight velocities, with no obvious threshold above which no additional increase in IQ is seen. This is similar to the relationship between weight at birth and later intellectual development, which in 6 large independent studies have also been shown to be linearly related over the whole range of birth weights.12
The strengths of our study are that the conclusions are derived from a large, well-documented data set, which has allowed adjustment for the important factors associated with IQ, and enabled us to show the linear relationship between early weight gain and subsequent IQ. The main weakness is the missing data, which have skewed the IQ results, and probably led to an underestimate of the size of effect of growth faltering on cognitive development. We did not have parental IQ, so were not able to estimate the size of genetic effect on the child's IQ: maternal educational level was used as a proxy variable. Also, the study would have been improved by having weight measurements at more ages, but we were constrained by having to use measurements collected at routine surveillance contacts. We acknowledge the limitation of observational studies such as ours that there is the always the possibility of residual confounding by unmeasured variables.
Our finding that the deficit in IQ is associated with early rather than later growth faltering has important implications. Firstly, early growth faltering is associated with symptoms of weak sucking and other feeding difficulties,18,27 which has led to the suggestion that FTT is a feeding skills disorder which is neurophysiological in origin.28 If infants who falter early in weight gain are neurologically impaired, then it is perhaps not surprising that they show deficits in later IQ scores. We have previously shown in the ALSPAC cohort that prolonged breastfeeding is associated with weight faltering. The effect of breastfeeding on IQ may seem to be discrepant with this, but also supports the argument that main effect we have found on IQ might not be because of nutrition but rather be a developmental issue. The alternative explanation is that the first weeks of life are a critical period nutritionally, when relative undernutrition can have significant impacts on brain growth and maturation.
The second major implication of our finding is for the identification of FTT, which has been defined traditionally with reference to the third or 5th centile for weight or weight gain. Our data from a large representative cohort show that these cutoffs are reasonably soundly based. For example, a weight-gain z score from birth to 9 months of –1.75 identifies children in the ALSPAC below the 4th centile, and Figure 3 shows clearly that this group has a lower IQ than children in any other group. Our data imply that to enable timely intervention to prevent any deleterious intellectual consequences, cases of growth faltering will have to be identified early in infancy. Children with poor weight gain are often not referred until much later in infancy, which is illustrated by a report from the Children's Hospital in Denver, Colorado, which showed that the average age of children referred for failure to thrive was 21 months.29 Health professionals monitoring the growth and developments of infants need to be aware that the early onset of growth faltering associated with slow or difficult feeding is both warning of inadequate nutritional intake, and a marker for possible subtle neurologic impairment, which may be associated with long-term cognitive impairments.
|
ACKNOWLEDGMENTS |
|---|
The United Kingdom Medical Research Council, the Wellcome Trust, and the University of Bristol provide core support for the ALSPAC. This publication is the work of the authors, and Dr Emond will serve as guarantor for the contents of this article. This study was funded by Wellcome Trust grant 59579.
We are extremely grateful to all the families who took part, the midwives for help in recruiting them, and the whole ALSPAC team, which includes interviewers, computer and laboratory technicians, clerical workers, research scientists, volunteers, managers, receptionists, and nurses. Sue Bonnell and Colin Steer made particular contributions to handling these data.
|
FOOTNOTES |
|---|
Accepted Mar 14, 2007.
Address correspondence to Alan M. Emond, MB, MD, FRCPCH, Centre for Child and Adolescent Health, Hampton House, Bristol BS6 6JS, United Kingdom. E-mail: alan.emond{at}bristol.ac.uk
The authors have indicated they have no financial relationships relevant to this article to disclose.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
- Kessler DB, Dawson P, eds. Failure to Thrive and Pediatric Undernutrition. Baltimore, MD: Brookes Publishing; 1999
- Skuse D. Epidemiological and definitional issues in failure to thrive. Child Adolesc Psychiatric Clin. 1993;2 :37 –59
- Wright CM, Waterston A, Matthews J, Aynsley-Greene A. What is the normal rate of weight gain in infancy? Acta Paediatr. 1994;83 :351 –356[Web of Science][Medline]
- Wright CM, Parkinson KN, Drewett RF. How does maternal and child feeding behavior relate to weight gain and failure to thrive? Data from a prospective birth cohort.
Pediatrics. 2006;117
:1262
–1269
[Abstract/Free Full Text] - Black M, Hutcheson J, Dubowitz H, Berenson-Howard J. Parenting style and developmental status among children with nonorganic failure to thrive.
J Pediatr Psychol. 1994;19
:689
–707
[Abstract/Free Full Text] - Skuse D, Pickles A, Wolke DF, Reilly S. Postnatal growth and mental development: evidence for a "sensitive period. " J Child Psychol Psychiatry. 1994;35 :521 –545[Web of Science][Medline]
- Wilensky D, Ginsberg G, Altman M, et al. A community based study of failure to thrive in Israel.
Arch Dis Child. 1996;75
:145
–148
[Abstract/Free Full Text] - Dowdney L, Skuse D, Heptinstall F, Puckering C, Zur-Szpiro S. Growth retardation and developmental delay amongst inner city children. J Child Psychol Psychiatry. 1987;28 :529 –540[Web of Science][Medline]
- Drewett RF, Corbett SS, Wright CM. Cognitive and educational attainments at school age of children who failed to thrive in infancy: a population-based study. J Child Psychol Psychiatry. 1999;40 :551 –561[CrossRef][Web of Science][Medline]
- Boddy J, Skuse D, Andrews B. The developmental sequelae of nonorganic failure to thrive. J Child Psychol Psychiatry. 2000;41 :1003 –1014[CrossRef][Web of Science][Medline]
- Corbett SS, Drewett RF. To what extent is failure to thrive in infancy associated with poorer cognitive development? A review and meta-analysis. J Child Psychol Psychiatry. 2004;45 :641 –654[CrossRef][Web of Science][Medline]
- Shenkin SD, Starr JM, Deary IJ. Birth weight and cognitive ability in childhood: a systematic review. Psychol Bull. 2004;130 :989 –1013[CrossRef][Web of Science][Medline]
- Simeon DT, Grantham-McGregor SM. Nutritional deficiencies and children's behavior and mental development. Nutr Res Rev. 1990;3 :1 –24[CrossRef][Medline]
- Mendez M, Adair L. Severity and timing of stunting in the first two years of life affect performance on cognitive tests in late childhood.
J Nutr. 1999;129
:1555
–1562
[Abstract/Free Full Text] - Drewett R, Wolke D, Asefa M, Kaba M, Tessema F. Malnutrition and mental development: is a there a sensitive period? A nested case-control study. J Child Psychol Psychiatry. 2001;42 :181 –187[CrossRef][Web of Science][Medline]
- Avon Longitudinal Study of Parents and Children. ALSPAC home page. Available at: www.alspac.ac.uk. Accessed August 1, 2007
- Blair PS, Drewett RF, Emmett PM, Ness A, Emond AM; ALSPAC Study Team. Family, socio-economic and prenatal factors associated with failure to thrive in the Avon Longitudinal Study of Parents and Children (ALSPAC).
Int J Epidemiol. 2004;33
:1
–9
[Free Full Text] - Emond AM, Blair PS, Drewett RF, Emmett PM. Postnatal factors associated with failure to thrive in term infants in the Avon Longitudinal Study of Parents and Children.
Arch Dis Child. 2007;92
:115
–119
[Abstract/Free Full Text] - Golding J, Pembrey M, Jones R, ALSPAC Study Team. ALSPAC: the Avon Longitudinal Study of Parents and Children. I. Study methodology. Paediatr Perinat Epidemiol. 2001;15 :74 –87[CrossRef][Web of Science][Medline]
- Freeman JV, Cole TJ, Chinn S, Jones PR, White EM, Preece MA. Cross sectional stature and weight reference curves for the UK, 1990.
Arch Dis Child. 1995;73
:17
–24
[Abstract/Free Full Text] - Cole TJ. Conditional reference charts to assess weight gain in British infants.
Arch Dis Child. 1995;73
:8
–16
[Abstract/Free Full Text] - Connery S, Katz D, Kaufman AS, Kaufman NL. Correlations between two short cognitive tests and a WISC-III short form using a sample of adolescent inpatients. Psychol Rep. 1996;78 :1373 –1378[Web of Science][Medline]
- Altman DG, Machin D, Bryant TN, Gardner MJ. Statistics With Confidence. 2nd ed. London, United Kingdom: BMJ Books; 2000
- Rudolf MCJ, Logan S. What is the long term outcome for children who fail to thrive? A systematic review.
Arch Dis Child. 2005;90
:925
–931
[Abstract/Free Full Text] - Anderson JW, Johnstone BM, Remley DT. Breast-feeding and cognitive development: a meta-analysis.
Am J Clin Nutr. 1999;70
:525
–535
[Abstract/Free Full Text] - Drotar D, Nowak M, Devost L, et al. Early psychological outcome in failure to thrive: predictions from an interactional model. J Clin Child Psychol. 1985;14 :105 –111[CrossRef][Web of Science]
- Reilly SM, Skuse DH, Wolke D, Stevenson J. Oral-motor dysfunction in children who fail to thrive: organic or non-organic? Dev Med Child Neurol. 1999;41 :115 –122[CrossRef][Web of Science][Medline]
- Ramsay M, Gisel EG, Boutry M. Non-organic failure to thrive: growth failure secondary to feeding-skills disorder. Dev Med Child Neurol. 1993;35 :285 –297[Web of Science][Medline]
- Miller LA, Grunwald GK, Johnson SL, Krebs NF. Disease severity at time of referral for pediatric failure to thrive and obesity: time for a paradigm shift? J Pediatr. 2002;141 :121 –124[CrossRef][Web of Science][Medline]
PEDIATRICS (ISSN 1098-4275). ©2007 by the American Academy of Pediatrics
CiteULike 
Connotea 
Del.icio.us 
Digg 
Facebook 
Reddit 
Technorati 
Twitter What's this?
This article has been cited by other articles:
|
|
 |
|
  P McDougall, R F Drewett, A P S Hungin, and C M Wright The detection of early weight faltering at the 6-8-week check and its association with family factors, feeding and behavioural development Arch. Dis. Child., July 1, 2009; 94(7): 549 - 552. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C J Kistin and H Bauchner A picture is worth a thousand words Arch. Dis. Child. Ed. Pract., December 1, 2008; 93(6): 177 - 189. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
