ORIGINAL
E n d o c r i n e
ARTICLE
R e s e a r c h
Childhood Stunting and Mortality Between 36 and 64
Years: The British 1946 Birth Cohort Study
Ken K. Ong, Rebecca Hardy, Imran Shah, and Diana Kuh on behalf of the
National Survey of Health and Development Scientific and Data Collection Teams
Medical Research Council Unit for Lifelong Health and Ageing (K.K.O., R.H., I.S., D.K.), London WC1B
5JU, United Kingdom; and Medical Research Council Epidemiology Unit (K.K.O.), Institute of Metabolic
Science, Addenbrooke’s Hospital, Cambridge, CB2 0QQ, United Kingdom
Background: Our aim was to examine the associations between childhood or adult height and adult
mortality.
Methods: In the prospective British 1946 Birth Cohort Study, childhood height was measured at 2,
4, 6, 7, 11, and 15 years, and adult height was measured at 36 years. Deaths were reported from
the national health service register.
Results: A total of 3877 study members (1963 male) contributed 106,333 person-years of follow-up;
391 deaths (228 male) were reported between the ages of 36 and 64 years. The strongest sexadjusted association between height and mortality between ages 36 and 64 years was seen for
height at age 6 years. The association was nonlinear; only study members in the shortest quintile
at 6 years had a higher relative risk of adult mortality compared with those in the tallest quintile.
By contemporary growth standards, 5.7% (n ⫽ 188) had heights at 6 years less than the second
percentile, and a further 15.0% (n ⫽ 490) had heights between the second to ninth percentiles;
these groups had higher adult mortality than all other study members (hazard ratio, 2.18; 95%
confidence interval, 1.52–3.13; P ⬍ .001; and hazard ratio, 1.42; 95% confidence interval, 1.08 –1.88;
P ⫽ .01, respectively). Several determinants of childhood stunting (height at 6 years less than the
second percentile) were directly associated with adult mortality; these included shorter parental
heights and adverse early life nutrition and housing.
Conclusions: British men and women born in 1946 were relatively stunted as children by contemporary standards. Those who were short at age 6 years had substantially higher mortality 30 to 60
years later. Furthermore, they accounted for the well-recognized inverse association between
adult height and mortality. (J Clin Endocrinol Metab 98: 2070 –2077, 2013)
inear growth retardation (“stunting”) during early
childhood is a marker of chronic undernutrition
caused by poor nutrition and compounded by infectious
disease (1). In deprived settings, stunting is associated with
poor childhood outcomes, permanent deficits in adult
height, poor adult health, and other adverse adult outcomes (1, 2). Furthermore, even across a wide range of
settings, shorter adult height is consistently associated
with higher risks of cardiovascular disease (CVD) and allcause mortality in men and women (3–7). Understanding
L
the association of poor childhood growth with mortality
in contemporary aging populations is key to the developmental origins of disease hypothesis (8) and adds an important perspective to studies in contemporary birth cohorts, who are increasingly prone to rapid childhood
weight gain and obesity (9, 10).
We hypothesized that short height in early childhood
would be adversely associated with adult survival even in
a high-income setting, that it would explain the association between shorter adult height and mortality, and that
ISSN Print 0021-972X ISSN Online 1945-7197
Printed in U.S.A.
Copyright © 2013 by The Endocrine Society
Received October 11, 2012. Accepted February 22, 2013.
First Published Online March 26, 2013
Abbreviations: BMI, body mass index; BP, blood pressure; CI, confidence interval; CVD,
cardiovascular disease; HR, hazard ratio; ICD, International Classification of Diseases;
NSHD, National Survey of Health and Development; OR, odds ratio; PEFR, peak expiratory
flow rate; SEP, socioeconomic position; WHO, World Health Organization.
2070
jcem.endojournals.org
J Clin Endocrinol Metab, May 2013, 98(5):2070 –2077
doi: 10.1210/jc.2012-3595
doi: 10.1210/jc.2012-3595
it would be independent of the association of weight status
with mortality. We compared the associations between
childhood or adult height and adult mortality in the British
1946 Birth Cohort Study. We also investigated whether
adverse early childhood environments, poor nutrition,
and physical illness, which are risk factors for poor childhood growth (11), might be related to adult mortality.
Finally, we investigated whether adult risk factors might
confound or mediate the relationship between height and
mortality.
Materials and Methods
Study design
The Medical Research Council’s National Survey of Health
and Development (NSHD) is a socially stratified birth cohort of
2547 men and 2815 women of white European descent born
during 1 week in 1946 who have been followed with repeated
data collections since then (12). The study received Multi-Centre
Research Ethics Committee approval, and informed consent was
given by cohort participants.
Study members were flagged for death on the National Health
Service Central Register from 1972. At that time, 2325 men and
2136 women were alive and residing in Britain; of the remaining
cohort, 288 had already died, 586 had emigrated, and 27 were
not flagged on the register. The underlying cause of death was
coded according to the International Classification of Diseases
(ICD)-9 or ICD-10 and were categorized as follows: cardiovascular diseases (ICD-9 codes 401– 454 and ICD-10 codes I10 –
I89); cancers (ICD-9 codes 140 –239 and ICD-10 codes C00 –
C97); externalizing disorders (including violent accidental and
suicidal deaths; ICD-9 codes 291–292, 295–305, 307–309, 311–
316, 570 –571.3, 800 –994, 1800 –1869, and 1880 –1999 and
ICD-10 codes F61–F69, K70 –K71, S00 –X99, and Y85–Y98)
and other causes. Cause of death was missing in 10 cases.
Anthropometry
Birth weight to the nearest quarter pound (113 g) was extracted from medical records within a few weeks of birth and was
converted to kilograms. Heights and weights were measured by
trained personnel using standard protocols at ages 2, 4, 6, 7, 11,
15, 36, 43, 53, and 60 to 64 years and were self-reported at ages
20 and 26 years. Adult height was taken as the measured height
at age 36 years; for 604 study members with missing data, we
used self-reported height at age 26 years. Body mass index (BMI)
was calculated as weight in kilograms divided by height in meters
squared. Measures of BMI, weight, and height were converted
into Z-scores using internally generated growth charts, constructed using the LMS method (13). To relate childhood height
to current growth references, we also calculated height Z-scores
by comparison with the World Health Organization (WHO)
2007 growth reference and defined stunting as height less than
the second percentile (⬍⫺2 Z-scores) (14).
Mothers were, when possible, measured at the examination
of their cohort child at age 6 years, and they reported their husband’s height. Midparental height was calculated as an indicator
of genetic height potential.
jcem.endojournals.org
2071
Indicators of early childhood environment,
nutrition, and illness
Various early life factors, namely, female sex, shorter midparental height, lower birth weight, higher birth order, higher
number of young siblings, lower childhood (father’s) social class,
lower father’s education, lower mother’s education, and higher
household crowding, have been examined previously as being
potential determinants of adult height in this study (11). We
examined those and other available nutritional factors as potential determinants linking childhood height to adult mortality.
Childhood socioeconomic position (SEP) was based on father’s
occupation at age 4 years; for 159 study members with missing
data, we used father’s occupation at age 11 or 15 years. Fathers’
and mothers’ education levels were classified as at least secondary level or primary level with or without further education or
qualifications (11). Information on breastfeeding (classified as
never or ever breastfed) and number of lower respiratory tract
infections were obtained from mother’s reports to health visitors
when children were aged 2 years. Family size (study member plus
surviving siblings at age 2 years) and number of young children
(aged ⬍5 years) in the house were also recorded at age 2 years.
Parity was based on the reported number of live births. At age 4
years, health visitors were asked to assess the mother’s general
management and cleanliness of the child and home. These responses were stratified in 3 groups: best; intermediate; and
worst. Although these responses are subjective, they discriminated very poor environments that other measures failed to distinguish (15). Household crowding was calculated as the total
number of people divided by the number of bedrooms or living
rooms at age 4 years (11). Chronic illness between ages 0 and 4
years was defined as an illness that necessitated a hospital admission for at least 28 days. Information on energy and fat intake
at 4 years was derived from maternal reports of what the child
had eaten in the last 24 hours (16).
Potential adult mediators and confounders
Height in childhood and adulthood is strongly related to adult
SEP, smoking status, and lung function, which are themselves
strong risk factors for mortality (17). Adult SEP was based on the
head of household’s occupation at age 36 years and classified as
professional/intermediate, skilled nonmanual, skilled manual,
or partly skilled/unskilled; for 768 study members with missing
data, we used head of household’s occupation at age 26 years.
Smoking status (current smoker, ex-smoker, or never smoker)
was available at 36 years; for 376 study members with missing
data, we used smoking status at age 31 years. Systolic and diastolic blood pressure (BP) (millimeters of mercury), resting pulse
rate, and peak expiratory flow rate (PEFR) (liters per minute) at
36 years were measured by nurses at home visits. BP was measured using a Hawksley random zero sphygmomanometer. The
second of 2 measures was used unless the second recording was
missing or invalid; then the first recording was used. PEFR was
measured using a mini-Wright peak flow meter and was taken as
the mean of the last 3 of up to 5 recordings.
Analyses
Cox regression models were used to investigate the relationships between childhood or adult height Z-score and adult mortality, adjusted for sex. Follow-up time was from the cohort
members’ 36th birthday in 1982 until the first of death, emigra-
2072
Ong et al
Childhood Stunting and Adult Mortality
J Clin Endocrinol Metab, May 2013, 98(5):2070 –2077
Table 1. Description of the Population by Sex
Adult height, cm
Mother’s height, cm
Father’s height, cm
Follow-up from age 36, y
Adult social class
I (professional)
II (intermediate)
IIIa (skilled nonmanual)
IIIb (skilled manual)
IV (partly skilled)
V (unskilled)
Cigarette smoking
Never smoker
Current smoker
Ex-smoker
Systolic BP, mm Hg
Diastolic BP, mm Hg
Resting pulse rate, beats/min
PEFR, liters/min
Male
Female
P Value*
175.7 ⫾ 6.6 (1963)
161.2 ⫾ 6.4 (1732)
173.2 ⫾ 8.0 (1963)
27.3 ⫾ 5.2 (1963)
162.4 ⫾ 6.1 (1914)
161.2 ⫾ 6.6 (1668)
173.4 ⫾ 8.0 (1914)
27.6 ⫾ 4.7 (1914)
⬍.001
.9
.6
.04
225 (11.6)
597 (30.8)
207 (10.7)
632 (32.6)
229 (11.8)
50 (2.6)
173 (9.1)
544 (28.7)
322 (17.0)
538 (28.3)
269 (14.2)
52 (2.7)
⬍.001
413 (25.0)
569 (34.4)
670 (40.6)
122.8 ⫾ 15.4
78.4 ⫾ 13.0 (1634)
70.9 ⫾ 10.3 (1626)
554 ⫾ 79 (1621)
558 (33.5)
558 (33.5)
550 (33.0)
114.7 ⫾ 15.6
73.3 ⫾ 12.4 (1646)
73.0 ⫾ 9.6 (1646)
412 ⫾ 68 (1638)
⬍.001
⬍.001
⬍.001
⬍.001
⬍.001
Summaries are based on available data at age 36 years; where data are missing additional data on height and social class at 26 years and smoking
at 31 years were used. Data are means ⫾ SD (n) or n (%).
a
t test for continuous variables; ␹2 test for categorical variables.
tion, or the end of February 2011 (just before the cohort’s 65
birthday). If death had not occurred, follow-up was treated as
censored. The proportional hazards assumption was checked
graphically.
To explore the shape of the association with mortality, height
Z-score was also entered as a quadratic term, and, if this term was
significant, mortality rates were calculated by quintiles of height
Z-score. We explored the interactions between childhood height
and adult body size (height and BMI) by adding these additional
interaction terms to the Cox regression models. Additional analyses of the association between childhood stunting and mortality
during specific age periods from childhood were performed by
logistic regression, with adjustment for sex, SEP, and smoking.
We tested which indicators of the early environment, childhood nutrition, and physical illness were independently associated with childhood stunting and whether these factors were also
associated with adult mortality. Finally, further models for adult
height Z-score and adult mortality were adjusted separately for
potential adult confounders or mediators, including adult SEP,
smoking status, weight, arterial BP, resting pulse rate, and PEFR.
All analyses were performed using SPSS (version 15.0).
Results
Details of the 3877 study members (1963 male and 1914
female) with data on adult height (at age 36 or 26 years)
who were flagged on the national death register are summarized in Table 1. Of these, 391 deaths (228 male and
163 female) were recorded between ages 36 and 64 years
(total 106,333 person-years follow-up; mean ⫾ SD,
27.4 ⫾ 5.0 years). The overall mortality rate was 3.7
deaths per 1000 person-years and was higher in men than
in women (hazard ratio [HR], 1.39; 95% confidence interval, [CI], 1.13–1.70; P ⫽ .001).
Childhood or adult height and all-cause mortality
Heights at all ages from 4 years onward were inversely
associated with all-cause mortality between ages 36 and
64 years, after adjustment for sex (Table 2). In linear models, the strongest associations with mortality were seen
with height at age 6 years (HR per ⫹1 SD, 0.80; 95% CI,
Table 2. HRs for All-Cause Mortality Between 36 and 64 Years per ⫹1 SD Increase in Childhood or Adult Height
Age at Measurement
HR
95% CI
Total
Deaths
2y
4y
6y
7y
11 y
14 y
Adult height
0.96
0.87
0.80
0.82
0.84
0.82
0.80
0.86 –1.07
0.79 – 0.97
0.72– 0.89
0.74 – 0.91
0.75– 0.94
0.73– 0.92
0.73– 0.89
3133
3410
3276
3354
3282
2991
3877
319
345
328
336
330
298
391
Values are adjusted for sex.
doi: 10.1210/jc.2012-3595
jcem.endojournals.org
2073
Figure 2. Kaplan-Maier plot of cumulative all-cause mortality
between ages 36 and 64 years old in study members in the shortest
height quintile at age 6 years (dotted line) vs those in any of the 4
taller height quintiles (solid line).
Figure 1. All-cause mortality during follow-up between 36 and 64
years by quintiles of adult height (A) and childhood height at age 6
years (B). The horizontal line indicates the overall mortality rate 3.7
deaths per 1000 person-years.
0.72– 0.89; P ⬍ .001) and adult height (HR per ⫹1 SD,
0.80; 95% CI, 0.73– 0.89; P ⬍ .001). In a combined multivariable regression model, height at 6 years (HR per ⫹1
SD, 0.85; 95% CI, 0.73– 0.99; P ⬍ .05) but not adult
height (HR per ⫹1 SD, 0.91; 95% CI, 0.78 –1.06; P ⫽ .2)
was associated with all-cause mortality between ages 36
and 64 years.
The shape of the associations between height at age 6
years or adult height and all-cause mortality were nonlinear, as indicated by the highly significant height-squared
terms in the regression models (P ⬍ .001). Adult height
showed a reverse J-shape association with the lowest mortality rate being in the second tallest quintile (Figure 1A).
At age 6 years, higher mortality rates were only seen in the
shortest quintile compared with the tallest (HR, 1.81;
95% CI, 1.30 –2.54; P ⬍ .001) (Figures 1B and 2). Exclusion of those in the shortest quintile at 6 years nullified the
association between adult height and mortality (HR per
⫹1 SD, 0.94; 95% CI, 0.81–1.09; P ⫽ .4).
There was a significant interaction between height at 6
years and adult height with mortality (P ⫽ .01); taller adult
height was protective against mortality only among those
in the shortest height quintile at 6 years (HR per ⫹1 SD,
0.72; 95% CI, 0.55– 0.94; P ⫽ .01), but not among those
in any of the taller childhood quintiles (HR per ⫹1 SD,
0.94; 95% CI, 0.81–1.09; P ⫽ .4) (Supplemental Figure 1
published on The Endocrine Society’s Journals Online
web site at http://jcem.endojournals.org). There was weak
evidence of interaction between short height at 6 years and
adult BMI status (underweight, normal weight, overweight, or obese) and mortality (P for interaction ⫽ .08);
the highest mortality rates were seen in those in the shortest height quintile at 6 years and subsequently either underweight (7.4 deaths per 1000 person-years; 15 deaths/
n ⫽ 77 at baseline) or obese as adults (12.9 deaths per 1000
person-years; 14 deaths/n ⫽ 41 at baseline) (Supplemental
Figure 2).
Cause-specific mortality
Adult height was not associated with death from cancer, either in linear models (HR per ⫹1 SD, 0.89; 95% CI,
0.77–1.04; P ⫽ .1; 171 deaths) or nonlinear models (not
shown) but was inversely associated with deaths from
CVD (HR per ⫹1 SD, 0.78; 95% CI, 0.63– 0.97; P ⫽ .02;
88 deaths), externalizing disorders (HR per ⫹1 SD, 0.71;
95% CI, 0.54 – 0.94; P ⫽ .02; 50 deaths), and all other
causes (HR per ⫹1 SD, 0.72; 95% CI, 0.57– 0.92; P ⫽
.007; 72 deaths). Similarly, being in the shortest vs all
other quintiles at 6 years was not associated with a higher
risk of death from cancer (HR per ⫹1 SD, 0.72; 95% CI,
0.50 –1.04; P ⫽ .08; 149 deaths), but showed similar sizes
of associations with deaths from CVD (HR per ⫹1 SD,
2074
Ong et al
Childhood Stunting and Adult Mortality
J Clin Endocrinol Metab, May 2013, 98(5):2070 –2077
Table 3. Independent Risk Factors for Childhood Stunting: ORs for Childhood Stunting and HRs for All-Cause
Mortality Between 36 and 64 Years
Midparental height per SD
Breastfed: never vs ever)
3⫹ young children aged ⬍5 y: yes vs noc
Birth weight per kg
Chronic illness ⬍4 y: yes vs nod
Care of house and childe
Best
Intermediate
Worst
OR (95% CI) for Stuntinga
HR (95% CI) for Adult Mortalityb
0.35 (0.27– 0.46)
1.75 (1.15–2.65)
2.70 (1.55– 4.71)
0.58 (0.39 – 0.85)
2.59 (1.39 – 4.83)
0.85 (0.73– 0.99)
1.37 (1.08 –1.73)
1.53 (1.09 –2.15)
0.95 (0.77–1.17)
1.16 (0.77–1.74)
Reference
1.44 (0.85–2.44)
1.79 (1.11–2.88)
Reference
1.13 (0.85–1.50)
1.04 (0.79 –1.37)
OR (95% CI) from the final multivariable model for childhood stunting, defined as height Z-score at age 6 years ⬍⫺2 (equivalent to less than the
second percentile) according to the WHO 2007 growth standard. Only those determinants that remained in the final model, with significant
independent associations with stunting, are shown. The full list of potential determinants that were entered into the initial model is shown in
Supplemental Table 1.
a
b
HR (95% CI) for mortality between age 36 and 64 years, adjusted for sex, adult social class, and smoking. Significant (P ⬍ .05) associations are
highlighted in bold.
c
Number of young children (aged ⬍5 years) in the house was recorded at age 2 years.
d
Chronic illness was defined as an illness that necessitated a hospital admission for at least 28 days between ages 0 to 4 years.
e
Based on assessments by health visitors at age 4 years.
1.69; 95% CI, 1.05–2.74; P ⫽ .03; 79 deaths), externalizing disorders (HR per ⫹1 SD, 1.90; 95% CI, 1.00 –3.61;
P ⫽ .05; 42 deaths), and all other causes (HR per ⫹1 SD,
1.98; 95% CI, 1.14 –3.44; P ⫽ .01; 56 deaths).
Childhood stunting and its determinants
associated with mortality
When height at 6 years was assessed against the contemporary WHO 2007 growth reference (14), all-cause
mortality between 36 and 64 years was higher in those less
than the second percentile, which is the height threshold
currently used to define childhood stunting (6.8 deaths per
1000 person-years; HR, 2.18; 95% CI, 1.52–3.13; P ⬍
.001), and also in those between the second and ninth
percentiles (4.7 deaths per 1000 person-years; HR, 1.42;
95% CI, 1.08 –1.88; P ⫽ .01) compared with those with
heights at 6 years that were taller than the ninth percentile
(3.2 deaths per 1000 person-years) (Supplemental Figure
3). The association of stunting at age 6 years with earlier
mortality (between ages 6 and 43 years: odds ratio (OR),
2.45; 95% CI, 0.92– 6.55; P ⫽ .07; 105 deaths) was
broadly similar to that with later mortality (between ages
43 and 63 years: OR, 1.69, 95% CI, 1.09 –2.64; P ⫽ .02;
371 deaths).
Several indicators of the early environment, childhood
nutrition, and physical illness were associated with risk of
childhood stunting at age 6 years (Supplemental Table 1).
Because these indicators were interrelated, we performed
a forward stepwise regression model and identified 6 factors that were independently associated with being in the
shortest height quintile at age 6 years. These factors were
shorter midparental height, never breastfed, the presence
of 3 or more young children (age ⬍5 years) in the house,
lower birth weight, a chronic illness requiring prolonged
hospital admission before age 4 years, and intermediate or
poor care of the house and child as assessed by the health
visitor at age 4 years (Table 3).
Of these 6 determinants of stunting, midparental height
(HR per ⫹1 SD, 0.85; 95% CI, 0.73– 0.99; P ⫽ .04), never
breastfed (HR, 1.37; 95% CI, 1.08 –1.73; P ⫽ .009), and
3 or more young children in the house (HR, 1.53; 95% CI,
1.09 –2.15; P ⫽ .01) were also associated with all-cause
mortality at age 36 to 64 years after adjustment for adult
SEP and smoking (Table 3).
Potential confounders and mediators
The association between adult height and all-cause
mortality was similar in men (HR per ⫹1 SD, 0.80; 95%
CI, 0.70 – 0.92) and women separately (HR per ⫹1 SD,
0.81; 95% CI, 0.69 – 0.94) and was only modestly attenuated when adjusted for potential confounders, adult SEP
and smoking (Table 4). Of the possible mediators, the
association between adult height and all-cause mortality showed no attenuation when adjusted for adult
weight, systolic and diastolic BP, or resting pulse rate
but was partially attenuated with adjustment for adult
PEFR (Table 4).
Discussion
In keeping with several previous studies (5, 18 –21), we
found that taller adult height was robustly related to lower
risk of all-cause mortality, and this association was not
doi: 10.1210/jc.2012-3595
jcem.endojournals.org
2075
Table 4. HRs for All-Cause Mortality Between 36 and 64 Years per ⫹1 SD Increase in Adult Height: Basic Models
and Models Adjusted for Potential Confounders or Mediators
Model
HR (95% CI)
Covariables in Models
Total
Deaths
Basic model
Men only
Women only
0.80 (0.73– 0.89)
0.80 (0.70 – 0.92)
0.81 (0.69 – 0.94)
Sex
3877
1963
1892
391
228
163
Basic model
Adjusted model
0.83 (0.75– 0.92)
0.86 (0.78 – 0.96)
Sex
⫹ Adult social class
3838
381
Basic model
Adjusted model
0.80 (0.72– 0.89)
0.82 (0.74 – 0.91)
Sex
⫹ Smoking
3618
357
Basic model
Adjusted model
0.79 (0.71– 0.89)
0.80 (0.70 – 0.92)
Sex
⫹ Parental heights
2754
286
Basic model
Adjusted model
0.81 (0.72– 0.91)
0.78 (0.69 – 0.88)
Sex
⫹ Weight at 36 y
3262
315
Basic model
Adjusted model
0.81 (0.72– 0.90)
0.82 (0.73– 0.91)
Sex
⫹ Diastolic and systolic BP at 36 y
3280
321
Basic model
Adjusted model
0.80 (0.72– 0.90)
0.82 (0.73– 0.92)
Sex
⫹ Resting pulse rate at 36 y
3272
319
Basic model
Adjusted model
0.84 (0.75– 0.94)
0.89 (0.79 –1.00)
Sex
⫹ PEFR at 36 y
3259
312
A
B
C
D
E
F
G
H
confounded by SEP or smoking. Uniquely, compared with
previous studies, we were able to show that the association
with adult mortality was strongest with height in early
childhood. A recent study similarly reported that the inverse association between height and risk of adult coronary heart disease was strongest at age 7 years; however,
they only had data on heights between ages 7 and 13 years
(22). Notably, the associations between height and mortality in our study were nonlinear and appeared to be
largely explained by short height at age 6 years (corresponding to height less than the ninth percentile on current
growth charts). Others have also reported nonlinear associations; for example, in a large Norwegian cohort only
adult height up to 165 cm was related to higher all-cause
mortality (23), but they did not have information on childhood heights.
The birth cohort study design allowed us to identify
early life factors that were related to both childhood short
height and adult mortality. Midparental height is commonly interpreted as a marker of the child’s genetic height
potential. Most of the variation in growth after age 2 years
is estimated to be genetically determined (24, 25). Alternatively, it is possible that the associations with parental
heights reflected some influence of the parents’ own childhood environment; however, the association with mortality was independent of SEP. Lack of breastfeeding in the
1940s would have increased the risk of gastrointestinal
and respiratory infections, as is seen in contemporary birth
cohorts in lower income settings (26). Furthermore, the
quality and availability of infant formula milk would have
been poor, being largely based on evaporated milk (27). In
contrast, in contemporary Western birth cohorts, formula
milk–fed infants show faster growth rates, presumably
due to increased nutritional intakes in settings of low infection risk (28). Having many young children in the
household was the third factor associated with both childhood short height and adult mortality. Whereas this again
could reflect the higher risks of early life infection and poor
nutrition, the markers of environmental adversity clustered together, and it may be inappropriate to single out
any specific factor in a birth cohort whose members were
relatively deprived and short as children by today’s standards (11).
In the British 1946 Birth Cohort Study, 38% of children
grew up in poor environmental circumstances, defined as
a combination of overcrowding, mothers who had only
primary school education, and fathers in manual occupations, and many such adverse early life factors were associated with shorter adult height (11). Environmental conditions improved substantially in later-born UK cohorts
who have taller childhood heights and higher childhood
and adult BMI (29) and in whom the health consequences
of childhood overnutrition and rapid growth may be more
pertinent to the developmental origins of disease than undernutrition (9, 10). Our findings may, therefore, have
more relevance to current birth cohorts in settings in which
2076
Ong et al
Childhood Stunting and Adult Mortality
childhood stunting remains common (30) and has adverse
short-term and long-term consequences (2). NSHD study
members experienced transition to the obesogenic environment during their third and fourth decades (29), and
previous analyses identified U-shaped associations between childhood and adult BMI and mortality (31). The
effects of height and BMI on adult mortality appeared to
be additive and probably represent distinct mechanisms.
In those who were short as children, the avoidance of
later underweight and obesity may be particularly beneficial for survival in terms of absolute risk reduction.
Our findings may therefore be relevant to settings that
are undergoing transition from undernutrition to overnutrition, resulting in a mismatch between early and
later life environments (32).
Our study has a number of limitations. Adult and childhood height, their determinants, such as breastfeeding,
and mortality are strongly socially patterned. Although we
performed statistical adjustments for SEP, it is possible
that residual effects remained. SEP in childhood has a
strong influence on adult mortality (33), through many
mechanisms that span disease prevention, detection, and
management. Tall height may improve confidence and
self-efficacy and increase “life chances,” with regard to
workplace and social capital. Height is also positively correlated with cognition, and, notably, in the NSHD greater
height growth during early childhood between the ages 2
to 4 years was particularly associated with cognitive development (34). However, the association between cognitive ability and mortality appears to be largely explained
by adult SEP (33). Strong social patterning and other clustering of adverse early life factors also limits our ability to
fully distinguish between the many potential determinants
of short childhood height that lead to subsequent mortality or even to speculate whether infectious, nutritional, or
genetic causes are responsible.
Rather than adult height per se, it is likely that some
other physiological mechanism related to childhood or
adult height explains the link to later mortality. Similar to
some previous studies (5), we found that lower respiratory
function partially mediated the association between
height and mortality. There were too few adult respiratory
deaths (n ⫽ 20) to allow a robust separate analysis of this
outcome as has been studied by others (35); however, our
findings suggest potential links between childhood and
adult respiratory function and deaths from nonrespiratory
disease. Whereas shorter adult height is associated with all
the major CVD risk factors, such as serum cholesterol, BP,
and smoking (36, 37), such markers do not appear to
explain the associations between height and CVD or allcause mortality (5, 35, 38). Causal modeling using genetic
variants related to adult height may give insights into the
J Clin Endocrinol Metab, May 2013, 98(5):2070 –2077
putative biological processes that link short height to later
mortality. Finally, our findings have important relevance
for the design of long-term surveillance studies of adult
disease and mortality after recombinant growth hormone
treatment for childhood short height (39, 40). Our results
show that it is inappropriate to simply compare mortality
in these patients with that in the general population.
In conclusion, in a long-running British birth cohort
study, the well-recognized association between shorter
adult height and higher all-cause mortality was explained
by short height during early childhood, which in turn appeared to reflect both genetic factors and early life
adversity.
Acknowledgments
We thank the NSHD study members and the staff involved in
data collection for this cohort over the last 65 years.
Address all correspondence and requests for reprints to: Dr
Ken Ong, MRC Epidemiology Unit, Institute of Metabolic Science, Addenbrooke’s Hospital, Box 285, Cambridge CB2 0QQ,
United Kingdom. E-mail: [email protected].
This work was supported by the Medical Research Council
(U120063239, U123092720, and U105960371). The sponsor
and funder had no role in the design and conduct of the study;
collection, management, analysis, and interpretation of the data;
and preparation, review, or approval of the manuscript.
Disclosure Summary: K.K.O. has received honoraria from
Pfizer Ltd and Novo Nordisk Ltd for speaking at and chairing
scientific meetings and from S. Karger AG for editorial work. The
other authors have nothing to disclose.
References
1. Walker SP, Wachs TD, Gardner JM, et al. Child development: risk
factors for adverse outcomes in developing countries. Lancet. 2007;
369:145–157.
2. Victora CG, Adair L, Fall C, et al. Maternal and child undernutrition: consequences for adult health and human capital. Lancet.
2008;371:340 –357.
3. Batty GD, Shipley MJ, Gunnell D, et al. Height, wealth, and health:
an overview with new data from three longitudinal studies. Econ
Hum Biol. 2009;7:137–152.
4. Paajanen TA, Oksala NK, Kuukasjarvi P, Karhunen PJ. Short stature is associated with coronary heart disease: a systematic review of
the literature and a meta-analysis. Eur Heart J. 2010;31:1802–1809.
5. Davey Smith G, Hart C, Upton M, et al. Height and risk of death
among men and women: aetiological implications of associations
with cardiorespiratory disease and cancer mortality. J Epidemiol
Community Health. 2000;54:97–103.
6. Wang N, Zhang X, Xiang YB, et al. Associations of adult height and
its components with mortality: a report from cohort studies of
135,000 Chinese women and men. Int J Epidemiol. 2011;40:1715–
1726.
7. Lee CM, Barzi F, Woodward M, et al. Adult height and the risks of
cardiovascular disease and major causes of death in the Asia-Pacific
doi: 10.1210/jc.2012-3595
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
region: 21,000 deaths in 510,000 men and women. Int J Epidemiol.
2009;38:1060 –1071.
Barker DJ, Winter PD, Osmond C, Margetts B, Simmonds SJ.
Weight in infancy and death from ischaemic heart disease. Lancet.
1989;2:577–580.
Ekelund U, Ong KK, Linne Y, et al. Association of weight gain in
infancy and early childhood with metabolic risk in young adults. J
Clin Endocrinol Metab. 2007;92:98 –103.
Leunissen RW, Kerkhof GF, Stijnen T, Hokken-Koelega A. Timing
and tempo of first-year rapid growth in relation to cardiovascular
and metabolic risk profile in early adulthood. JAMA. 2009;301:
2234 –2242.
Kuh D, Wadsworth M. Parental height: childhood environment and
subsequent adult height in a national birth cohort. Int J Epidemiol.
1989;18:663– 668.
Wadsworth M, Kuh D, Richards M, Hardy R. Cohort profile: the
1946 National Birth Cohort (MRC National Survey of Health and
Development). Int J Epidemiol. 2006;35:49 –54.
Cole TJ, Green PJ. Smoothing reference centile curves: the LMS
method and penalized likelihood. Stat Med. 1992;11:1305–1319.
de Onis M, Onyango AW, Borghi E, Siyam A, Nishida C, Siekmann
J. Development of a WHO growth reference for school-aged children and adolescents. Bull World Health Organ. 2007;85:660 –
667.
Douglas JWB. Children Under Five. London: George Allen and
Unwin; 1958.
Prynne CJ, Paul AA, Mishra GD, Hardy RJ, Bolton-Smith C, Wadsworth ME. Sociodemographic inequalities in the diet of young children in the 1946 British birth cohort. Public Health Nutr. 2002;5:
733–745.
Kuh D, Hardy R, Hotopf M, et al. A review of lifetime risk factors
for mortality. Br Actuar J. 2009;15(suppl):17– 64.
Allebeck P, Bergh C. Height, body mass index and mortality: do
social factors explain the association? Public Health. 1992;106:
375–382.
Song YM, Sung J. Adult height and the risk of mortality in South
Korean women. Am J Epidemiol. 2008;168:497–505.
Notkola V, Punsar S, Karvonen MJ, Haapakoski J. Socio-economic
conditions in childhood and mortality and morbidity caused by coronary heart disease in adulthood in rural Finland. Soc Sci Med.
1985;21:517–523.
Peck AM, Vagero DH. Adult body height, self perceived health and
mortality in the Swedish population. J Epidemiol Community
Health. 1989;43:380 –384.
Silventoinen K, Baker JL, Sorensen TI. Growth in height in childhood and risk of coronary heart disease in adult men and women.
PLoS One. 2012;7:e30476.
Engeland A, Bjorge T, Selmer RM, Tverdal A. Height and body mass
index in relation to total mortality. Epidemiology. 2003;14:293–
299.
Silventoinen K, Magnusson PK, Tynelius P, Kaprio J, Rasmussen F.
Heritability of body size and muscle strength in young adulthood: a
study of one million Swedish men. Genet Epidemiol. 2008;32:341–
349.
Silventoinen K, Pietilainen KH, Tynelius P, Sorensen TI, Kaprio J,
Rasmussen F. Genetic regulation of growth from birth to 18 years
jcem.endojournals.org
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
2077
of age: the Swedish young male twins study. Am J Hum Biol. 2008;
20:292–298.
Oddy WH. Breastfeeding protects against illness and infection in
infants and children: a review of the evidence. Breastfeed Rev. 2001;
9:11–18.
Anderson SA, Chinn HI, Fisher KD. History and current status of
infant formulas. Am J Clin Nutr. 1982;35:381–397.
Ong KK, Preece MA, Emmett PM, Ahmed ML, Dunger DB. Size at
birth and early childhood growth in relation to maternal smoking,
parity and infant breast-feeding: longitudinal birth cohort study and
analysis. Pediatr Res. 2002;52:863– 867.
Li L, Hardy R, Kuh D, Lo Conte R, Power C. Child-to-adult body
mass index and height trajectories: a comparison of 2 British birth
cohorts. Am J Epidemiol. 2008;168:1008 –1015.
Stevens GA, Finucane MM, Paciorek CJ, et al. Trends in mild, moderate, and severe stunting and underweight, and progress towards
MDG 1 in 141 developing countries: a systematic analysis of population representative data. Lancet. 2012;380:824 – 834.
Strand BH, Kuh D, Shah I, Guralnik J, Hardy R. Childhood, adolescent and early adult body mass index in relation to adult mortality: results from the British 1946 birth cohort. J Epidemiol Community Health. 2012;66:225–232.
Low, FM, Gluckman PD, Hanson MA. Developmental plasticity
and epigenetic mechanisms underpinning metabolic and cardiovascular diseases. Epigenomics. 2011;3:279 –294.
Kuh D, Shah I, Richards M, Mishra G, Wadsworth M, Hardy R. Do
childhood cognitive ability or smoking behaviour explain the influence of lifetime socio-economic conditions on premature adult mortality in a British post war birth cohort? Soc Sci Med. 2009;68:
1565–1573.
Richards M, Hardy R, Kuh D, Wadsworth ME. Birthweight, postnatal growth and cognitive function in a national UK birth cohort.
Int J Epidemiol. 2002;31:342–348.
Jousilahti P, Tuomilehto J, Vartiainen E, Eriksson J, Puska P. Relation of adult height to cause-specific and total mortality: a prospective follow-up study of 31,199 middle-aged men and women in
Finland. Am J Epidemiol. 2000;151:1112–1120.
Skidmore PM, Hardy RJ, Kuh DJ, Langenberg C, Wadsworth ME.
Life course body size and lipid levels at 53 years in a British birth
cohort. J Epidemiol Community Health. 2007;61:215–220.
Langenberg C, Hardy R, Kuh D, Wadsworth ME. Influence of
height, leg and trunk length on pulse pressure, systolic and diastolic
blood pressure. J Hypertens. 2003;21:537–543.
Forsen T, Eriksson J, Qiao Q, Tervahauta M, Nissinen A,
Tuomilehto J. Short stature and coronary heart disease: a 35-year
follow-up of the Finnish cohorts of The Seven Countries Study.
J Intern Med. 2000;248:326 –332.
Carel JC, Ecosse E, Landier F, et al. Long-term mortality after recombinant growth hormone treatment for isolated growth hormone
deficiency or childhood short stature: preliminary report of the
French SAGhE study. J Clin Endocrinol Metab. 2012;97:416 – 425.
Savendahl L, Maes M, Albertsson-Wikland K, et al. Long-term mortality and causes of death in isolated GHD, ISS, and SGA patients
treated with recombinant growth hormone during childhood in Belgium, The Netherlands, and Sweden: preliminary report of 3 countries participating in the EU SAGhE study. J Clin Endocrinol Metab.
2012;97:E213–E217.