The Journal of Nutrition

Community and International Nutrition

Children Who Recover from Early Stunting and

Children Who Are Not Stunted Demonstrate
Similar Levels of Cognition1,2
Benjamin T. Crookston,3* Mary E. Penny,4 Stephen C. Alder,3 Ty T. Dickerson,5 Ray M. Merrill,6
Joseph B. Stanford,3 Christina A. Porucznik,3 and Kirk A. Dearden7
3

Department of Family and Preventive Medicine, University of Utah, Salt Lake City, UT 84108; 4Instituto de Investigacion Nutricional,
La Molina, Lima-12, Peru; 5Department of Pediatrics, University of Utah, Primary Childrens Medical Center, Salt Lake City, UT 84113;
6
Department of Health Sciences, Brigham Young University, Provo, UT 84602; and 7Department of International Health and Center for
Global Health and Development, Boston University, Boston, MA 02118

Abstract
decreased productivity, and reduced adult stature. Recovery from early stunting is possible; however, few studies explore
whether those who demonstrate linear catch-up growth experience long-term cognitive deficits. Using longitudinal data
on 1674 Peruvian children from the Young Lives study, we identified factors associated with catch-up growth and
assessed whether children who displayed catch-up growth have significantly lower cognition than children who were not
stunted during infancy and childhood. Based on anthropometric data for children 618 mo of age and again for the same
children when they were 4.56 y of age, we categorized participants as not stunted, stunted in infancy but not childhood
(catch-up), stunted in childhood, and stunted in infancy and childhood. Children who had grandparents in the home, had
less severe stunting in infancy, and had taller mothers were more likely to demonstrate catch-up growth by round 2.
Children who experienced catch-up growth had verbal vocabulary and quantitative test scores that did not differ from
children who were not stunted (P = 0.6 and P = 0.7, respectively). Those stunted in childhood as well as those stunted in
infancy and childhood scored significantly lower on both assessments than children who were not stunted. Based on
findings from this study, policy makers and program planners should consider redoubling efforts to prevent stunting and
promote catch-up growth over the first few years of life as a way of improving childrens physical and intellectual
development. J. Nutr. 140: 19962001, 2010.

Introduction
Undernutrition contributes to more than one-half of the nearly
10 million childhood deaths that occur annually (1,2). Stunting
affects 150 million (24%) of all children worldwide (3). Stunting
is defined as a height-for-age Z-score (HAZ)8 , 22 SD from the

Supported by the UK Department for International Development (DFID) for the

benefit of developing countries (core-funding for Young Lives). Substudies are
funded by the Bernard van Leer Foundation, the Inter-American Development
Bank (in Peru), the International Development Research Centre (in Ethiopia), and
the Oak Foundation. The views expressed here are those of the authors. They
are not necessarily those of the Young Lives project, the University of Oxford,
DFID, or other funders. Additional funding for this research came from the
University of Utah, the Instituto de Investigacion Nutricional, Brigham Young
University, and Boston University.
2
Author disclosures: B. T. Crookston, M. E. Penny, S. C. Alder, T. T. Dickerson,
R. M. Merrill, J. B. Stanford, C. A. Porucznik, and K. A. Dearden, no conflicts of
interest.
8
Abbreviations used: CDA, Cognitive Development Assessment; HAZ, heightfor-age Z-scores; PPVT, Peabody Picture Vocabulary Test.
* To whom correspondence should be addressed. E-mail: benjamin.crookston@
utah.edu.

1996

mean and is often categorized as mild (22 to 21 SD), moderate

(22 to 23 SD), and severe (, 23 SD) (4).
For many countries in resource-poor settings, mean HAZ at
birth is close to the National Center for Health Statistics
reference score but begins faltering immediately after birth and
continues into the 3rd y of life (5). A majority of stunted children
become stunted adults (6).
Stunting is associated with adverse cognitive development in
childhood and adolescence, delayed entrance into school, fewer
years of schooling, decreased productivity, and reduced adult
stature (4,711). Stunting during childhood is associated with
delays in motor development and lower IQ (12). Stunted
children are less likely than their nonstunted counterparts to
enroll in school; those who do enroll have lower grades and
poorer cognition (13,14).
Subsequent to growth faltering during the first 3 y, children
from resource-poor countries grow at rates similar to children
from more affluent nations, neither losing additional ground nor
experiencing substantial recovery (5,12). There is no established
definition for catch-up growth. (15). It can refer to acceleration of growth or partial or complete recovery from stunting

2010 American Society for Nutrition.

Manuscript received November 16, 2009. Initial review completed December 28, 2009. Revision accepted August 11, 2010.
First published online September 15, 2010; doi:10.3945/jn.109.118927.

Downloaded from jn.nutrition.org by guest on November 29, 2012

Stunting is associated with adverse cognitive development in childhood and adolescence, fewer years of schooling,

(15,16). We use HAZ for this study, and we define catch-up

growth as HAZ , 22 during infancy (;1 y of age) and $ 22
during childhood (;5 y of age) (9,15,17).
Some consider the likelihood of catch-up growth to be
limited, because children remain in environments that contribute
to growth restriction (1720). A few studies show that partial
catch-up is possible among stunted children who remain in the
same environment (15,2123). Identifying factors that allow
children to catch up is important for developing programs to
improve nutrition and health.
We tested the hypothesis that children who experience catch-up
growth have significantly lower scores on 2 cognitive assessments
than children who were not stunted in infancy and childhood.
We also identified factors associated with catch-up growth.

Methods

Study participants. A detailed description of methods for this study has

been given elsewhere (24). Briefly, a total of 2052 children from 74
communities representing 20 districts in Peru were enrolled in 2002. As
an initial step, we assembled a complete list of all districts in the country,
excluding the 5% with the highest socioeconomic levels. To determine
socioeconomic levels, we ranked the 1818 districts in Peru according to a
composite index of poverty (25). Using national census data, the size of
the population for all districts was noted. The total population was
divided by 2000 to establish the sampling interval and a random number
within the sampling interval was selected to determine the initial district.
To identify the subsequent district, we added the total number of
individuals in the sampling interval to the initial starting point and
continued in this fashion until all 20 districts were identified. Once the 20
districts were chosen, we randomly selected a community and a specific
house within the community. In many cases, there were ,100 children
aged 617.9 mo in a given community. In such cases, the closest
community was selected in systematic fashion until 100 families per
district were enrolled. Urban areas were subdivided into city blocks. The
sample represents ~95% of Peruvian children, includes participants from
urban and rural areas, and represents coastal, highland, and jungle regions.
Institutional review boards from London South Bank University,
London School of Hygiene and Tropical Medicine, and the University of
Reading approved of this research. Within Peru, the Ministry of Health
approved of the study. Ethical reviews were conducted by the Institutional
Review Boards at the Instituto de Investigacion Nutricional and the
Peruvian Instituto Nacional de Salud.
Data collection. Interviews consisted of a core questionnaire for
primary caregivers and included questions regarding household composition, child health, caregiver characteristics, livelihoods, socioeconomic
status, social capital, childcare, and cognitive development. Fieldworkers collected anthropometric data using digital scales (Soehnle)
and locally made wooden boards and measurement techniques were
standardized per WHO guidelines (26). Fieldworkers were trained for 3
mo in data collection protocols and instruments. Based on a pilot study,
instruments were revised. Three teams of 6 interviewers per team
collected 2 rounds of data in 2002 when children were 617.9 mo of age
(round 1) and 20062007 when they were 4.56 y old (round 2). There
are 2052 children in the cohort. Only 4% of children were lost to follow-

Cognitive outcomes. This study used the Peabody Picture Vocabulary

Test (PPVT) to evaluate vocabulary skills and listening comprehension
and the Cognitive Development Assessment (CDA) to judge the
quantitative reasoning of children 4.56 y of age. The PPVT is a
commonly used measure for evaluating cognitive development in both
industrialized and resource-poor countries (2730). The PPVT is highly
correlated with the Wechsler and McCarthy Scales, which are validated
measures of intelligence (3133). A validated, Spanish version of the
PPVT consisting of 125 questions was given to children participating in
the study (34,35).
The CDA was developed by the International Association for the
Evaluation of Educational Achievement to study the effect of preschool
attendance on cognitive development in children 4 y of age (36). The
CDA has 3 main components: quantity, time, and spatial relations. Due
to the great amount of time it took to administer (spatial relations
subtest) and low reliability among the Young Lives sample in Peru (time
subtest), only the quantity subtest was used in the study. The quantity
subtest relies on 15 items to measure a childs perception of amount. For
each item, interviewers show children images and ask a question such
as Look at the cats and dogs and point to the picture where the dog
has less food than the cat.
It is possible that because the CDA was developed for children 4 y of
age, it might not be an appropriate measure of cognition for children
older than 4 y. We compared the mean raw and Rasch scores for the CDA
from children without preschool, with preschool, and currently attending first grade. Rasch scores are used extensively in psychometrics and
increasingly in public health. Researchers use them to determine the
quality of tests and to build true interval-scale measures based on raw
scores (37). With Rasch scores, one can judge the degree to which
measurement has been successful. Rasch scores are based on a
probability model that when a group of individuals is asked multiple
questions, e.g. as part of a cognitive assessment, individual j makes a
mistake in answering question k.
As expected, data from Young Lives suggest that children who had
formal schooling performed better on the test than children who never
attended preschool. Similarly, the longer children spent in preschool, the
higher the score. While test scores improved with increased levels of
education, we conclude that the CDA was an appropriate measure for
children 4.56 y of age, as noted below.
Items in the PPVT and the CDA are arranged in order of increasing
difficulty. A given child received only those items within his or her critical
range based on chronological age and ceiling. We reviewed all observations to verify that basal and ceiling sets were correctly established for
each child (36). Not all distributions of raw and Rasch scores were
normal. In rare cases, children achieved close to the highest possible
score. Thus, test scores may have underestimated childrens abilities for
the construct measured. This appears to be a very infrequent occurrence.
For example, only 1 child obtained a perfect score on the CDA.
Validity and reliability for both tests were established by assessing the
degree to which evidence and theory supported the interpretations of test
scores. Reliability was established according to Classical Test theory and
Item Response theory. Details about the establishment of validity and
reliability can be found elsewhere (36). For children who were native
Spanish speakers, both tests were found to have appropriate psychometric properties. However, each test proved unreliable for assessing
cognition among native Quechua speakers. Thus, only Spanish-speaking
children (n = 1706) were included in the analysis. We used standardized
scores for both the PPVT and the CDA.
Catch-up growth. Catch-up growth was examined both as an outcome
variable and as a determinant of cognitive scores. We defined stunted in
infancy but not childhood (catch-up) as having a HAZ , 22 during
round 1 and a HAZ $ 22 at round 2 (i.e. during infancy but not
childhood), stunted in infancy and childhood as having a HAZ , 22
at rounds 1 and 2, stunted in childhood as having a HAZ $ 22 during
round 1 and a HAZ , 22 at round 2, and not stunted as having
HAZ $ 22 at both rounds (9,15,17). We used information from all
children when evaluating catch-up growth as a predictor of cognitive
Catch-up growth and improved cognition

1997

Downloaded from jn.nutrition.org by guest on November 29, 2012

Study design and background. This study assessed a prospective

cohort of children who are part of a multicountry study called Young
Lives. The Young Lives study, which is funded by the United Kingdoms
Department for International Development, explores childhood poverty,
intergenerational poverty, and poverty reduction policies. Two cohorts
of children (1 and 8 y old) from Ethiopia, India, Peru, and Vietnam have
been followed since 2002 with the intent to track the children for 15 y.
Our study examines catch-up growth following early stunting among the
younger Peruvian cohort. Researchers from the Instituto de Investigacion
Nutricional and Grupo de Analisis para el Desarrollo manage the
Peruvian cohort.

up. Data from both rounds were entered for analyses using Delphi
software.

scores. When considering determinants of catch-up growth, we used data

only from children with a HAZ , 22 at round 1. Our assessment of
changes in nutritional status was based on data collection at 2 points in
time. We do not know whether children were stunted before round 1 or
between rounds 1 and 2.

were tested for interaction and compliance with logistic and linear
regression model assumptions. No interaction terms were retained based
on P , 0.1. Significance was declared when P , 0.05.

Covariates. Covariates were chosen based on results from an earlier

study using Young Lives data, which used a well-known conceptual
framework for the determinants and consequences of undernutrition
(24,38). Covariates included: urban/rural setting; geographic region
(coast/highland/jungle); childs age, sex, and preschool attendance; and
maternal age, height, education, and ethnicity. Severity of stunting was
used when considering determinants of catch-up growth, where only
children with a HAZ , 22 at round 1 were examined. In this instance,
severity of stunting was the HAZ for a given child. Thus, the lower the
HAZ, the more severe the stunting experienced by the child. A composite
indicator of wealth ranging from 01 reflected consumer durables,
services, and housing quality (39).

Childrens mean age at round 1 was 12 (6 3.5) mo. Slightly less

than one-half (49.5%) were female and only a few (1.0%) had
begun formal schooling by round 2. A majority (74.1%) lived in
urban areas. Highland and coastal regions had the greatest
proportion of study participants, 43.0 and 40.2%, respectively.
The remaining respondents lived in jungle regions (16.8%).
Most mothers described themselves as mestizo (91.9%), with
the principal religion being Catholicism (80.2%). Two-thirds
(63.8%) of children were not stunted in either round 1 or round
2, whereas 8.4% were stunted at round 1 but not round 2 and
were considered to have experienced catch-up growth, 13.4%
were stunted at round 2 only (childhood), and 14.4% were
stunted in both infancy and in childhood.
Children classified into these 4 categories differed significantly with respect to sex, region, area of residence, preschool
attendance, and sociodemographic conditions (Table 1). More
males than females experienced catch-up growth. Children from
rural areas made up one-half of those stunted in infancy and
childhood but only one-quarter of the overall sample. Mean
wealth index was highest among those who were not stunted,
followed by children experiencing catch-up growth.
Change in HAZ between rounds 1 and 2 differed by stunting
classification as follows: catch-up growth, 1.13 6 0.66; stunted

TABLE 1

Characteristics of study participants by stunting classification1

Independent
variable2
n
Sex, %
Male
Female
Region, %
Coast
Highland
Jungle
Area, %
Urban
Rural
Attended preschool, %
Yes
No
HAZ score (round 1)
HAZ score (round 2)
Child age, mo
Maternal age, y
Maternal height, cm
Wealth index
Maternal education, y
Verbal vocabulary
Quantitative score

Not
stunted3

Catch-up4

Stunted in
childhood5

Stunted in infancy
and childhood6

1065

140

224

240

49.9
50.1

64.3
35.7

38.8
61.2

56.7
43.3

50.9
34.6
14.6

30.7
53.6
15.7

25.5
52.7
21.9

12.9
64.2
22.9

83.5
16.5

79.3
20.7

52.7
47.3

49.6
50.4

90.4
9.6
20.6 6 0.9
20.9 6 0.8
11.8 6 3.5
26.5 6 6.3
151.3 6 5.2
0.47 6 0.20
9.7 6 3.6
95.9 6 19.9
10.9 6 2.2

90.0
10.0
22.6 6 0.5
21.4 6 0.5
13.8 6 3.2
27.5 6 7.2
150.0 6 4.5
0.44 6 0.20
8.3 6 4.3
90.2 6 20.3
10.9 6 2.1

75.5
24.6
21.1 6 0.8
22.5 6 0.5
10.9 6 3.4
26.0 6 6.4
149.0 6 5.6
0.33 6 0.19
7.3 6 3.7
82.1 6 19.0
9.4 6 2.7

75.4
24.6
22.8 6 0.7
22.8 6 0.6
13.1 6 3.5
27.5 6 7.5
147.1 6 4.9
0.29 6 0.18
6.1 6 4.0
76.9 6 18.3
9.4 6 2.6

P-value7

0.0001

0.0001

0.0001

0.0001

0.0001
0.0001
0.0001
0.0366
0.0001
0.0001
0.0001
0.0001
0.0001

Values are mean 6 SD or %.

Data reported in table are from round 1 (when child was 618 mo of age) except for verbal vocabulary and quantitative score.
3
Not stunted: HAZ $ 22 at rounds 1 and 2.
4
Catch-up: HAZ , 22 at round 1 and HAZ $ 22 at round 2.
5
Stunted in childhood: HAZ $ 22 at round 1 and HAZ , 22 at round 2.
6
Stunted in infancy and childhood: HAZ , 22 at round 1 and at round 2.
7
Chi-square and F tests compare percents and means for each stunting classification for each independent variable.
1
2

1998

Crookston et al.

Downloaded from jn.nutrition.org by guest on November 29, 2012

Data analysis. Anthropometric indicators were calculated using the

latest WHO International Growth Reference standard (26). SAS (version
9.2) was used for statistical analyses. Pearson chi-square tests compared
distributions among stunting classifications; F tests compared means
among groups.
For analyses, we used mixed regression models to account for the
cluster sample design. The MIXED procedure from SAS was used for
linear models (cognitive outcomes) and the GLIMMIX procedure was
used for logistic models (catch-up growth outcomes). Covariates were
retained or dropped based on P-values (,0.1) and conceptual considerations. Regression coefficients (linear models), odds ratios (logistic
models), and 95% CI were calculated for retained variables. All models

Results

TABLE 2

Odds ratios from logistic regression model for catchup growth (1 = yes, 0 = no) among Peruvian children
who experienced stunting in infancy (n = 374)

Independent variable1

Odds ratio

Area population
Rural
Urban
Maternal education in y
Preschool attendance
No
Yes
Child age in mo
Grandparents live in home
No
Yes
Severity of stunting (HAZ at round 1)2
Maternal height Z-score

1.00
1.70
1.06
1.00
2.03
1.14
1.00
1.89
2.10
1.66

P-value

0.08
0.07
0.06
0.0006
0.04
0.0009
0.0002

95% CI

0.94, 3.08
1.00, 1.13

0.98, 4.17
1.05, 1.18

1.05, 3.40
1.36, 3.25
1.28, 2.17

Data reported in table come from round 1 (when child was 618 mo of age) except
preschool attendance and child age.
2
HAZ at round 1 is a continuous variable.

TABLE 3

Discussion
Results suggest that one-third of children who were stunted in
infancy (at 1 y of age) recovered by the time they were 5 y old.
Those who demonstrated catch-up growth had cognitive test
scores similar to children who were not stunted at either age.
These results are important because they demonstrate that
children can recover from early nutritional insult, identify
predictors of catch-up growth, and suggest that catch-up growth
contributes to cognition. These findings can guide future
research about the impact of stunting on cognition as well as
policy and program efforts to improve childrens physical and
intellectual well-being.
Potential for catch-up growth. The proportion of stunted
children who experienced catch-up growth (36.8%) is consistent
with findings from previous research (9,15,17,40,41). In a study
of 2131 Filipino children, researchers reported that 34% of
stunted children experienced catch-up growth by 8 y of age (18).
More than one-half of 239 Peruvian children in Lima who were
stunted in the first 2 y exhibited catch-up growth by 9 y (9). These
studies used similar methods for collecting anthropometry and
defining catch-up growth. But the timing of assessment, cognitive
tests administered, and setting differed from our study. The
severity of stunting varied across studies discussed here. Finally,
the magnitude of recovery (1.13 SD) in our catch-up group was
nearly identical to that described by Adair (15) (1.14 SD).
Predictors of catch-up growth. According to our results,
children with higher HAZ at initial assessment (i.e. children who
experienced less severe stunting) were more likely to experience
catch-up growth. Others (15,17,22) also report that higher HAZ
at initial assessment improves the chances of experiencing catchup growth, although variables included in their models differ
somewhat from ours. Additional factors associated with catch-

Unadjusted differences in cognitive scores by stunting status in infancy and early childhood
among Peruvian children (n = 1674)
Verbal vocabulary score

Quantitative score

Characteristics

Estimate

P-value

95% CI

Estimate

P-value

95% CI

Intercept
Not stunted
Catch-up
Stunted in childhood
Stunted in infancy and childhood

92.33

23.15
26.21
210.03

,0.0001

0.04
,0.0001
,0.0001

87.17, 97.49

26.18, 20.12
28.77, 23.67
212.57, 27.49

10.64

0.14
20.91
20.86

,0.0001

0.50
,0.0001
,0.0001

10.20, 11.08

20.26, 0.53
21.24, 20.59
21.19, 20.54

Catch-up growth and improved cognition

1999

Downloaded from jn.nutrition.org by guest on November 29, 2012

in childhood, 21.38 6 0.96; not stunted, 20.28 6 0.85; and

stunted in infancy and childhood, 0.02 6 0.76. Changes in HAZ
for children demonstrating catch-up growth, those stunted in
childhood only, and those not stunted were significantly different from children who were stunted in infancy and childhood.
We conducted logistic regression to identify determinants of
catch-up growth, limiting our analyses to children with HAZ ,
22 at round 1 (Table 2). Predictors of catch-up growth included
maternal height, child age at round 2, grandparents living at
home, and severity of stunting in round 1. Severity of stunting
was the most important predictor of catch-up growth; children
who were less stunted at round 1 (i.e. higher HAZ) were more
likely to experience catch-up growth. It is worth pointing out
that although not significant, area of residence, maternal
education, and preschool attendance were retained in the model
for conceptual reasons (P , 0.10).
Estimates for unadjusted regression models predicting verbal
vocabulary and quantitative scores are reported in Table 3. For
verbal vocabulary scores, the reference group (not stunted)
scored significantly higher than the 3 comparison groups.
However, for quantitative cognition, children who were not
stunted and those who demonstrated catch-up growth did not
differ (P = 0.5).
When adjusted for maternal age and education, area of
residence, preschool attendance, wealth index, and number of
siblings, verbal vocabulary scores did not differ between the not
stunted and catch-up groups (Table 4). However, children who
were stunted in childhood or stunted in infancy and childhood

had significantly lower scores on the verbal vocabulary test.

Children who had older mothers and mothers with more
education, and children who lived in urban areas, attended
preschool, had fewer siblings, and came from better socioeconomic conditions scored higher on the verbal vocabulary
assessment.
Differences between quantitative scores for the not stunted
and catch-up groups did not differ (P = 0.7) (Table 5), but
children who were stunted in childhood or stunted in infancy
and childhood had significantly lower quantitative scores.
Children who were older at round 1, had older mothers or
mothers with more education, lived in urban areas, attended
preschool, had fewer siblings, and came from better socioeconomic conditions scored significantly better on the quantitative
assessment.

TABLE 4

Estimates from mixed linear regression model for

predictors of verbal vocabulary score among Peruvian
children (n = 1649)

Independent variable1
Intercept
Stunting
Not stunted
Catch-up
Stunted in childhood
Stunted in infancy and childhood
Maternal age in y
Maternal education in y
Area population
Rural
Urban
Preschool attendance
No
Yes
Wealth index
Number of siblings

Estimate

P-value

95% CI

59.0

,0.0001

53.2, 64.9

20.7
22.5
24.6
0.3
1.4

0.62
0.04
0.0002
0.0001
,0.0001

23.5, 2.1
24.8, 20.1
27.0, 22.2
0.1, 0.4
1.2, 1.7

0.002

1.7, 7.3

0.002
,0.0001
0.0002

1.5, 6.3
14.0, 28.2
21.9, 20.6

4.5

3.9
21.1
21.2

up growth included childs age at second assessment, maternal

height, and grandparents living in the home. Although childs
age and presence of grandparents in the home cannot be changed
by program planners and implementers, maternal height is
amenable to improvement and should be considered as one
option among several that can improve the potential for catchup growth.
Catch-up growth and cognition. Our results are different
from previous research that suggests that stunting has long-term
consequences from which children do not recover (20). Similar
to our findings, 2 previous studies found no significant cognitive

TABLE 5

Estimates from mixed linear regression model for

predictors of quantitative score among Peruvian
children (n = 1649)

Independent variable1
Intercept
Stunting
Not stunted
Catch-up
Stunted in childhood
Stunted in infancy and childhood
Child age in mo
Maternal age in y
Maternal education in y
Area population
Rural
Urban
Preschool attendance
No
Yes
Wealth index
Number of siblings
1

P-value

95% CI

6.63

,0.0001

5.83, 7.42

0.08
20.49
20.60
0.11
0.02
0.09

0.69
0.003
0.0003
,0.0001
0.01
,0.0001

20.30, 0.46
20.81, 20.17
20.93, 20.28
0.08, 0.14
0.00, 0.04
0.06, 0.12

0.52

0.005

0.16, 0.89

0.52
1.6
20.10

0.002
0.001
0.02

0.19, 0.85
0.66, 2.59
20.19, 20.02

Estimate

Data reported in table come from round 1 (when child was 618 mo of age) except
preschool attendance, wealth index, and number of siblings.

2000

Crookston et al.

Acknowledgments
B.T.C., M.E.P., S.C.A., T.T.D., R.M.M., J.B.S., C.A.P., and
K.A.D. designed the research; B.T.C. and K.A.D. conducted the
research and analyzed data; B.T.C., M.E.P., S.C.A., T.T.D.,
R.M.M., J.B.S., C.A.P., and K.A.D. wrote the paper; and B.T.C.
and K.A.D. had primary responsibility for its final content. All
authors read and approved the final manuscript.

Literature Cited
1.

Caulfield LE, Richard SA, Black RE. Undernutrition as an underlying

cause of malaria morbidity and mortality in children less than five years
old. Am J Trop Med Hyg. 2004;71 Suppl 2:5563.

Downloaded from jn.nutrition.org by guest on November 29, 2012

1
Data reported in table come from round 1 (when child was 618 mo of age) except
preschool attendance, wealth index, and number of siblings.

deficit between children who were not stunted and children who
experienced catch-up growth (9,17). Although Berkman (9)
documented differences in cognition between children who were
stunted in infancy and childhood and children who were not
stunted during either period, he found no differences in
cognition between children who were not stunted and children
who demonstrated catch-up growth, even after adjusting for
covariates. Mendez and Adair (17) found moderate differences
in cognition between children who had catch-up growth and
those who were not stunted but found no difference when
adjusting for many of the same confounding factors used in this
study, such as number of siblings, area population (urban vs.
rural), maternal education, and wealth. Given the apparent lack
of difference in cognitive scores between those who experienced
catch-up growth and those who were never stunted, children
who were at greatest risk of cognitive deficits were those who did
not experience catch-up growth after stunting in infancy (41).
This study has limitations. First, there is no standard
definition for catch-up growth; thus, our interpretations may
be different from those in other studies. However, we used a
definition that was employed by 3 other authors (15,17,22) and
our results are consistent with findings where the same definition
was used. Second, our sample is not nationally representative;
we were unable to include non-Spanish speakers, because each
test proved unreliable for assessing cognition among nonSpanish speakers. Third, we have measures at only 2 points in
time and therefore do not have measures of the frequency or
duration of stunting. Thus, it is not possible to determine
whether lack of cognitive impact in childhood (round 2) is due to
catch-up growth alone or whether differences in the severity,
frequency, and duration of stunting as well as access to
interventions designed to improve nutritional status account
for this effect. For example, some infants at round 1 may have
been moderately stunted only once (i.e. only at the time of data
collection).
Our results suggest that children can recover from early
nutritional insult, the severity of stunting influences potential for
recovery, and children who demonstrate catch-up growth
display levels of cognition (as measured by the PPVT and the
CDA) that are similar to those of children who are not stunted in
infancy nor childhood. They also suggest that preschool is
important in improving the cognition of children who were
stunted. Reducing the severity of early stunting is critical to
improving childrens chances of recovery. Additionally, improvements in maternal education can increase the possibility of
catch-up growth. Programs and policies designed to improve
physical growth and cognition should focus first and foremost
on children , 2 y of age. However, interventions that include
older children may also improve their growth and intellectual
development.

2.
3.

4.

5.

6.
7.

8.

9.

11.

12.

13.

14.

15.
16.
17.

18.

19.
20.

21.

22. Tanner JM. Catch-up growth in man. Br Med Bull. 1981;37:2338.

23. Kulin HE, Bwibo N, Mutie D, Santner SJ. The effect of chronic
childhood malnutrition on pubertal growth and development. Am J
Clin Nutr. 1982;36:52736.
24. Crookston BT, Dearden KA, Alder SC, Porucznik CA, Stanford JB,
Merrill RM, Dickerson TT, Penny ME. Impact of early and concurrent
stunting on cognition. J Matern Child Nutr. In press 2010.
25. Wilson I, Huttly SRA, Fenn B. A case study of sample design for
longitudinal research. Young Lives. Int. J. Social Research Methodology. 2006;9:3516.
26. WHO Multicentre Growth Reference Study Group. WHO child growth
standards based on length/height, weight and age. Acta Paediatr Suppl.
2006;450:7685.
27. Grantham-McGregor SM, Walker SP, Chang SM, Powell CA. Effects of
early childhood supplementation with and without stimulation on later
development in stunted Jamaican children. Am J Clin Nutr. 1997;66:
24753.
28. Kordas K, Lopez P, Rosado JL, Vargas GG, Rico JA, Cebrian ME,
Stoltzfus RJ. Blood lead, anemia, and short stature are independently
associated with cognitive performance in Mexican school children.
J Nutr. 2004;134:36371.
29. Walker SP, Grantham-McGregor SM, Powell CA, Chang SM. Effects of
growth restriction in early childhood on growth, IQ and cognition at
age 1112 years and the benefits of nutritional supplementation and
psychosocial stimulation. J Pediatr. 2000;137:3641.
30. Walker SP, Chang SM, Powell CA, Grantham-McGregor SM. Effects of
early childhood psychosocial stimulation and nutritional supplementation on cognition and education in growth-stunted Jamaican children:
prospective cohort study. Lancet. 2005;366:18047.
31. Campbell JM. Review of the Peabody Picture Vocabulary Test Third
Edition. J Psychoed Assess. 1998;16:3348.
32. Campbell JM, Bell SK, Keith LK. Concurrent validity of the Peabody
Picture Vocabulary Test Third Edition as an intelligence and achievement screener for low SES African American children. Assessment.
2001;8:8594.
33. Gray S, Plante E, Vance R, Henrichsen M. The diagnostic accuracy of
four vocabulary tests administered to preschool-age children. Lang
Speech Hear Serv Sch. 1999;30:196206.
34. Dunn L, Padilla E, Lugo D, Dunn L. Manual del Examinador para el
Test de Vocabulario en Imagenes Peabody (Peabody Picture Vocabulary
Test): Adaptacion Hispanoamericana (Hispanic-American adaptation).
Circle Pines (MN): AGS; 1986.
35. Dunn L, Dunn L. Examiners manual for the PPVT-III. Form IIIA and
IIIB. (MN): AGS; 1997.
36. Cueto S, Leon J, Guerrero G, Munoz I. Psychometric characteristics of
cognitive development and achievement instruments in round 2 of
Young Lives. Technical Note 15. Oxford: Young Lives. 2009.
37. Rasch G. Probabilistic models for some intelligence and attainment
tests. Copenhagen: Denmarks Paedagogiske Institut; 1960.
38. Black RE, Allen LH, Bhutta ZA, Caulfield LE, de Onis M, Ezzati M,
Mathers C, Rivera J. Maternal and child undernutrition: global and
regional exposures and health consequences. Lancet. 2008;371:24360.
39. Filmer D, Pritchett L. Estimating wealth effects without expenditure
data or tears: an application to educational enrolments in states of
India. Demography. 2001;38:11532.
40. Vella V, Tomkins A, Borghesi A, Miglori GB, Oryem VY. Determinants
of stunting and recovery from stunting in northwest Uganda. Int J
Epidemiol. 1994;23:7826.
41. Cheung YB, Khoo KS, Karlberg J, Machin D. Association between
psychological symptoms in adults and growth in early life: longitudinal
follow up study. BMJ. 2002;325:74952.

Catch-up growth and improved cognition

2001

Downloaded from jn.nutrition.org by guest on November 29, 2012

10.

UNICEF [Internet]. The state of the worlds children 2008. 2007 [cited
2009 18 Mar]. Available from: http://www.unicef.org/sowc08/.
de Onis M. Child growth and development. In: Semba RD, Bloem MW,
editors. Nutrition and health in developing countries. 2nd ed. Totowa
(NJ): Humana Press; 2008. p. 11337.
Caulfield L, Richard S, Rivera J, Musgrove P, Black R. Stunting, wasting,
and micronutrient deficiency disorders. In: Jamison D, Breman J,
Measham A, Alleyne G, Claeson M, Evans D, Jha P, Mills A, Musgrove
P, et al, editors. Disease control priorities in developing countries. 2nd ed.
Washington, DC: Oxford University Press; 2006. p. 55167.
Shrimpton R, Victora C, de Onis M, Costa Lima R, Blossner M,
Clugston G. Worldwide timing of growth faltering: implications for
nutritional interventions. Pediatrics. 2001;107:E75.
Frongillo EA. Symposium: causes and etiology of stunting. Introduction. J Nutr. 1999;129 Suppl 2S:S52930.
Grantham-McGregor S, Cheung YB, Cueto S, Glewwe P, Richter L,
Strupp B. Developmental potential in the first 5 years for children in
developing countries. Lancet. 2007;369:6070.
Behrman JR, Hoddinott J, Maluccio JA, Soler-Hampejsek E, Behrman
EL, Martorell R, Quisumbing A, Ramirez M, Stein AD. What determines
adult skills? Impacts of preschool school-years and post-school experiences in Guatemala. Philadelphia: University of Pennsylvania; 2006.
Berkman DS, Lescano AG, Gilman RH, Lopez SL, Black MM. Effects of
stunting, diarrhoeal disease, and parasitic infection during infancy on
cognition in late childhood: a follow-up study. Lancet. 2002;359:56471.
Martorell R, Rivera J, Kaplowitz J, Pollitt E. Long term consequences of
growth retardation during early childhood. In: Hernandez M, Argenta
J, editors. Human growth: basic and clinical aspects. Amsterdam:
Elsevier; 1992. p. 1439.
Victora CG, Adair L, Fall C, Hallal PC, Martorell R, Ritcher L, Sachdev
HS. Maternal and child undernutrition: consequences for adult health
and human capital. Lancet. 2008;371:34057.
Martorell R. Undernutrition during pregnancy and early childhood:
consequences for cognitive and behavioral development. In: Young ME,
editor. Early child development: investing in our childrens future.
Amsterdam and New York: Elsevier Science BV; 1997. p. 3983.
Beasley NMR, Hall A, Tomkins AM, Donnelly C, Ntimbwa P, Kivuga J,
Kihamia CM, Lorri W, Bundy DAP. The health of enrolled and nonenrolled children of school age in Tanga, Tanzania. Acta Trop. 2000;
76:2239.
Ivanovic DM, Perez HT, Olivares MD, Diaz NS, Leyton BD, Ivanovic
RM. Scholastic achievement: a multivariate analysis of nutritional,
intellectual, socioeconomic, sociocultural, familial, and demographic
variables in Chilean school-aged children. J Nutr. 2004;20:87889.
Adair LS. Filipino children exhibit catch-up growth from age 2 to 12
years. J Nutr. 1999;129:11408.
Ashworth A, Milward DJ. Catch-up growth in children. Nutr Rev.
1986;44:15763.
Mendez MA, Adair LS. Severity and timing of stunting in the first two
years of life affect performance on cognitive tests in late childhood.
J Nutr. 1999;129:155562.
Checkley W, Epstein LD, Gilman RH, Cabrera L, Black RE. Effects of
acute diarrhoea on linear growth in Peruvian children. Am J Epidemiol.
2003;157:16675.
de Onis M. Commentary: socioeconomic inequalities and child growth.
Int J Epidemiol. 2003;32:5035.
Martorell R, Khan LK, Schroeder DG. Reversibility of stunting:
epidemiologic findings from children in developing countries. Eur J
Clin Nutr. 1994;48 Suppl:S4557.
Golden MH. Is complete catch-up growth possible for stunted
malnourished children? Eur J Clin Nutr. 1994;48 Suppl 1:5870.