Bulletin of the World Health Organization

Use of new World Health Organization child growth standards to assess how infant malnutrition relates to breastfeeding and mortality

Linda Vesel a, Rajiv Bahl a, Jose Martines a, Mary Penny b, Nita Bhandari c, Betty R Kirkwood d & the WHO Immunization-linked Vitamin A Supplementation Study Group

a. Department of Child and Adolescent Health and Development, World Health Organization, 20 avenue Appia, 1211 Geneva 27, Switzerland.
b. Instituto de Investigacion Nutricional, Lima, Peru.
c. Society for Applied Studies, New Delhi, India.
d. Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London, England.
e. Division of Child Health and Development Immunization-linked Vitamin A Group, World Health Organization, Geneva, Switzerland.

Correspondence to Rajiv Bahl (e-mail: bahlr@who.int).

(Submitted: 18 August 2008 – Revised version received: 27 April 2009 – Accepted: 03 June 2009 – Published online: 13 October 2009.)

Bulletin of the World Health Organization 2010;88:39-48. doi: 10.2471/BLT.08.057901

Introduction

Malnutrition contributes to about one-third of the 9.7 million child deaths that occur each year.1,2 Recently, the World Health Organization (WHO) introduced new child growth standards for use in deriving indicators of nutritional status, such as stunting, wasting and underweight. These standards are based on the growth of infants from six different regions of the world who were fed according to WHO and United Nations Children’s Fund (UNICEF) feeding recommendations, had a non-smoking mother, had access to primary health care and did not have any serious constraints on health during infancy or early childhood.36 It is recommended that these new growth standards replace the previously recommended international growth reference devised by the National Center for Health Statistics (NCHS) in the United States.7

The prevalence of malnutrition estimated using WHO standards is expected to differ from that based on the NCHS growth reference because there are differences in median weight-for-age, height-for-age and weight-for-height between the two.8 Recent studies have investigated the direction and magnitude of these differences.912 In children aged 6–59 months, the prevalence of stunting (i.e. low height-for-age) and wasting (i.e. low weight-for-height) were higher when WHO standards were used but that of underweight (i.e. low weight-for-age) was lower.9,11,12 In the first half of infancy (i.e. the period from birth up to the end of the 6th month), the prevalence of stunting, wasting and underweight has been reported to be higher with WHO growth standards.10,12 It is important that the magnitude of these apparent changes in the prevalence of malnutrition are investigated in different settings in order to gain a better understanding of their implications, particularly for child health and nutrition programmes whose progress is monitored through large household surveys.

Another important question connected with growth in the first 6 months of life is its relationship with feeding practices. Exclusive breastfeeding is recommended for infants up to 6 months of age because of its benefits in reducing morbidity and mortality.13 In their systematic review of the optimal duration of exclusive breastfeeding, Kramer and Kakuma14 combined the results of two studies conducted in Honduras and found that exclusively breastfed infants had a lower prevalence of stunting, wasting and underweight, although not significantly so.

Malnourished children are known to be at an increased risk of death.1520 Nutritional status indicators can be used to identify those infants and children at a higher risk of dying so they can be provided with special care both at a population level in emergency settings and individually following screening. For example, low weight-for-age is used in the UNICEF–WHO Integrated Management of Childhood Illness (IMCI) programme to identify infants whose feeding practices should be assessed and who would benefit from additional counselling on infant feeding.21 It is not yet known whether nutritional status assessed using WHO growth standards or the NCHS growth reference would be a better predictor of death, and answering this question has been proposed as a research priority.12,22

We carried out a secondary analysis of a large data set obtained from a randomized controlled trial of vitamin A supplementation conducted in Ghana, India and Peru23 to determine how using the NCHS growth reference or WHO growth standards influences the calculated prevalence of malnutrition, the relationship between exclusive breastfeeding and malnutrition, and the sensitivity and specificity of nutritional status indicators for predicting the risk of death during infancy.

Methods

Data collection

Between 1995 and 1997, 9424 mother–infant pairs were enrolled in a randomized controlled trial of vitamin A supplementation linked to the WHO Expanded Programme on Immunization (EPI): 2919 mother–infant pairs were from 37 villages in the Kintampo district of Ghana, 4000 were from two urban slums in New Delhi, India, and 2505 were from a periurban shanty town in Lima, Peru. Child morbidity, mainly diarrhoea and respiratory infection, was high at all study sites. Breastfeeding was almost universal and more than 94% of infants were still consuming breast milk after 9 months. Full details of the original study are described elsewhere.23 In this paper, we present only the information essential for the secondary analysis.

Mothers and infants were enrolled 21–42 days after childbirth in Ghana and 18–28 days after childbirth in India and Peru. Information on each infant’s age, sex and breastfeeding status and on several family characteristics was collected by fieldworkers, who visited participants’ homes after enrolment.23

During the first follow-up visit at 6 weeks in Ghana and India and at 10 weeks in Peru, the first doses of diphtheria, tetanus and pertussis vaccine and oral poliomyelitis vaccine were administered and the infants’ weight and length were measured. Weights and lengths were also measured when the infants were 6, 9 and 12 months of age. However, of the 9424 enrolled infants, 983 from either Ghana or Peru were followed up only to 6 months of age to enable the study to be completed in the intended time period.23 Standard procedures were followed for measuring weight and length. Field workers were extensively trained and standardization exercises were conducted before data collection to ensure high levels of reproducibility and validity. Length measurements were taken three times and the median was used to calculate the z-score, which is the number of standard deviations an observation is above or below the mean. Weight was measured only once. A length board with a sliding foot scale and a precision of 0.1 cm was used to measure length and a hanging spring scale accurate to 100 g and calibrated daily was used to measure weight.

Information on each infant’s vital status and feeding mode was collected at follow-up visits which took place every 4 weeks until the infant was 12 months of age. Infants were divided into subgroups by feeding history using data collected at 6 (Ghana and India only), 10, 14, 18 and 22 weeks. At each visit, mothers were asked what they had offered their child to eat or drink during the past week. After the mother’s unprompted response was documented, she was asked whether she had offered her own breast milk, breast milk from a wet nurse, animal milk, infant formula, other fluids or solid food at any time during the week. An exclusively breastfed infant was one who received only breast milk from his mother or a wet nurse and who took no other liquids or solids, except vitamins, mineral supplements or medication.

The study was approved by the ethics review committees of all the participating institutions and by WHO’s ethics review board.23

Secondary data analysis

The nutritional survey option in WHO Anthro 2005 software (WHO, Geneva, Switzerland) for assessing growth and development was used to calculate weight-for-age, length-for-age and weight-for-length z-scores on the basis of both WHO standards and the NCHS growth reference. An infant was defined as stunted, wasted or underweight if his or her length-for-age, weight-for-length or weight-for-age z-score, respectively, was less than –2. Severe stunting, severe wasting or severe underweight corresponded to a length-for-age, weight-for-length or weight-for-age z-score, respectively, less than –3.

Weight and length measurements were considered valid at 6 or 10 weeks if measured within 2 weeks of the target date and valid at 6, 9 and 12 months if measured within 4 weeks of the target date. Data were checked to exclude extreme outliers (only one outlier from Ghana was excluded from the analysis) and missing or unknown values. The length of a small number of infants could not be measured, for example, because the parents refused permission or because of deformity. Consequently their length-for-age and weight-for-length could not be calculated. In addition, the weight-for-length of infants whose length was under a certain value (i.e. 49 cm for the NCHS growth reference and 45 cm for WHO standards) was not calculated. The analysis only included data on infants whose z-scores could be calculated using both the NCHS growth reference and WHO standards. Weight could not be adjusted for oedema as its presence was not recorded.

The duration of exclusive breastfeeding was defined as the age of the infant at the follow-up visit when exclusive breastfeeding was last reported. Detailed feeding information was available only after the first immunization visit, at 6 weeks in Ghana and India and at 10 weeks in Peru. We categorized the duration of exclusive breastfeeding as < 10 weeks, 10 to < 18 weeks or ≥ 18 weeks. The prevalences of malnutrition in infants aged 6 months in these three categories were compared using Cuzick’s nonparametric statistical test for the trend across ordered groups. Logistic regression analysis was used to adjust for potential confounding factors, which included the infant’s weight, age at enrolment, sex and birth order, the mother’s education, whether there was a multiple or single birth, and the death of a previous child from the same mother.

The risk of death among malnourished and adequately nourished infants was compared. The sensitivity and specificity of nutritional status indicators (i.e. stunting, wasting and underweight) at 6 weeks and 6 months for predicting death in the subsequent periods of 6 weeks to 6 months and 6 months to 12 months, respectively, were calculated. If a malnutrition indicator had a high sensitivity but a low specificity, we also calculated the sensitivity and specificity of the corresponding indicator of severe malnutrition. For example, as underweight had a high sensitivity and low specificity in India, we additionally calculated the sensitivity and specificity of severe underweight in the country.

The area under the receiver operating characteristic curve (AUC) and its 95% confidence interval (CI) were used to assess the performance of nutritional status indicators for identifying those at risk of death. An AUC of 0.5 means the indicator is no better than chance; the closer the AUC is to 1, the better the performance of the indicator.

The statistical analysis was performed using Intercooled Stata 8.0 (Stata Corp LP, College Station, TX, United States of America). Baseline characteristics and results are summarized by country.

Results

Baseline characteristics

For this secondary analysis, 8787 of the 9424 infants (93.2%) enrolled in the study were included at the first immunization visit, 7964 (84.5%) were included at 6 months and 5890 (62.5%), at 12 months. Fig. 1 summarizes the reasons for infants being lost to follow-up.

Fig. 1. Number of infants included at different times in the study of infant malnutrition in Ghana, India and Peru, 1995–1997
Fig. 1. Number of infants included at different times in the study of infant malnutrition in Ghana, India and Peru, 1995–1997
a Of the 9424 infants overall, 983 from Ghana or Peru were followed up only to 6 months of age to enable the study to be completed in the intended time period.

The baseline characteristics of the mothers and infants enrolled in the randomized controlled trial and aspects of their living environment are summarized in Table 1. The sites in the three countries were similar with regard to the infants’ age at enrolment, sex and breastfeeding status. The proportion of twins was much higher in Ghana. Peru had the highest parental educational level and the lowest proportion of mothers who reported a previous child death, followed sequentially by India and Ghana.

Prevalence of malnutrition

The prevalence of malnutrition in infancy was high, particularly in India and Ghana (Fig. 2). As expected, the prevalence of stunting, wasting and underweight in infants in the first half of infancy (i.e. before 6 months of age) was lower at all study sites when the NCHS growth reference was used. When infants were 1 year of age, the prevalence of stunting and wasting continued to be lower with the NCHS growth reference, while that of underweight was much higher. Consequently, there appeared to be a sharp increase in the prevalence of malnutrition during the second half of infancy when the NCHS growth reference was used. In contrast, a more gradual increase in the prevalence of malnutrition was observed throughout infancy when WHO standards were used (Fig. 2).

Fig. 2. Prevalence of stunting, wasting and underweight in infants during the first year of life in Ghana, India and Peru as determined using WHO child growth standards and the NCHS growth reference, 1995–1997
Fig. 2. Prevalence of stunting, wasting and underweight in infants during the first year of life in Ghana, India and Peru as determined using WHO child growth standards and the NCHS growth reference, 1995–1997
NCHS, National Center for Health Statistics; WHO, World Health Organization.

There appeared to be an association between a longer duration of exclusive breastfeeding and a lower prevalence of malnutrition in 6-month-old infants at all study sites, particularly when judged by the proportion of underweight infants (Table 2). This association was no longer statistically significant for most indicators after adjustment for potential confounding factors, except for wasting in Peru determined using WHO standards and underweight in India determined using the NCHS growth reference.

Prediction of infant mortality

The relationship between malnutrition at the first immunization visit (at 6 weeks in Ghana and India and 10 weeks in Peru) and the risk of death between then and the age of 6 months is presented in Table 3. The presence of stunting, wasting or underweight, as determined using WHO child growth standards or the NCHS growth reference, did not identify infants at risk of death in either Ghana or Peru. However, in India, malnutrition at 6 weeks of age was associated with an increased risk of death before 6 months of age, and the indicator that had the best performance was severe underweight determined using the WHO child growth standard (sensitivity: 70.2%; specificity: 85.8%; AUC: 0.78, 95% CI: 0.72–0.84), followed by underweight determined using the NCHS reference chart.

The relationship between infant malnutrition at 6 months of age and the risk of death between 6 and 12 months of age is presented in Table 4. In Ghana, being underweight at 6 months of age as determined using WHO child growth standards had the highest sensitivity and specificity for death: 37.0% and 82.2%, respectively, with an AUC of 0.60 (95% CI: 0.50–0.69). In Peru, the same indicator was the most predictive (sensitivity: 33.3%; specificity: 97.9%; AUC: 0.66, 95% CI: 0.45–0.86). In India, wasting at 6 months as determined using WHO child growth standards gave the best performance (sensitivity: 54.6%; specificity: 85.5%; AUC: 0.70, 95% CI: 0.61–0.79), closely followed by severe underweight as determined using WHO child growth standards (sensitivity: 50.0%; specificity: 86.3%; AUC: 0.68, 95% CI: 0.60–0.77) and underweight determined using the NCHS growth reference (sensitivity: 58.8%; specificity: 76.6%; AUC: 0.68, 95% CI: 0.59–0.76).

Discussion

This analysis shows the prevalence of stunting, wasting and underweight in the first half of infancy was higher when determined using WHO child growth standards rather than the NCHS growth reference. In contrast, the prevalence of underweight was lower at 1 year of age. Nutritional status indicators determined using WHO standards were better predictors of mortality during infancy than those determined using the NCHS growth reference.

Our analysis was based on data collected on large samples taken from sites on three continents, namely Africa, Asia and Latin America. Well-standardized anthropometric measurements were available for a large proportion of enrolled infants at four time points during their first year of life. This analysis is the first to examine the relative merits of WHO child growth standards and the NCHS growth reference for predicting mortality during infancy.

The analysis has a number of potential limitations related to the characteristics of the study population and how they were followed up. Infants were not enrolled at birth and, therefore, birth weights and lengths were not available. While around 85% of the cohort was followed up to 6 months, the proportion fell to 63% at 12 months, largely because a shorter follow-up was planned. However, the baseline characteristics of those with a short follow-up did not differ from that of those who were followed up to 1 year of age (data not shown). In addition, the weight-for-length z-score at 6 weeks of age could not be estimated for infants whose length was too short (i.e. < 49 cm for the NCHS growth reference and < 45 cm for WHO standards). Because information on feeding involved 7-day recall at intervals of 4 weeks, it is possible that the feeding practices recorded may differ from those in the intervals between follow-up visits. Finally, the number of deaths in Peru was small, which makes it difficult to draw conclusions about the performance of the different malnutrition indicators in predicting subsequent mortality.

This analysis confirms the findings of previous studies10,12 which also showed that the prevalence of malnutrition during the first half of infancy was higher when WHO child growth standards rather than the NCHS growth reference were used and that the difference became less marked thereafter. At 1 year of age, the prevalence of wasting continued to be higher when WHO growth standards were used and the prevalence of stunting was similar, but the prevalence of underweight was lower. Thus, a gradual increase in the prevalence of malnutrition was observed throughout infancy when WHO standards were used, which contrasts with the steep rise in the prevalence observed after 6 months when the NCHS growth reference was used. The high prevalence of malnutrition in the first half of infancy observed using WHO standards is more consistent with the prevailing prevalence of low birth weight for the three study sites as calculated using available data:24 30% in India and 11% in Ghana and Peru.

The differences found in the prevalence of malnutrition are due to differences between the median growth curves derived using WHO standards and the NCHS growth reference. This can be explained by two characteristics of the WHO standards. First, they are based on the growth of breastfed infants only while the NCHS growth reference is based predominantly on the growth of formula-fed infants. Second, the WHO standards are based on more frequent measurements (i.e. every 2 weeks in the first 2 months and monthly thereafter) than the NCHS growth reference (i.e. every 3 months).8

The apparent association observed between the duration of exclusive breastfeeding and most nutritional status indicators at 6 months was not significant after adjustment for confounding factors. Although the association remained significant for two indicators, residual confounding cannot be ruled out. While exclusive breastfeeding has been associated with lower mortality and morbidity in previous studies,13 no clear association with a growth benefit has been reported.25

Anthropometric status, particularly underweight and wasting, has been shown to be a good predictor of subsequent mortality in previous studies.15,1720 Our findings for the second half of infancy are consistent with these previous results. However, no such association was observed between the first immunization visit and 6 months of age in Ghana or Peru, which indicates that adequately nourished infants are exposed to other significant causes of death during this period in these countries. While there are clear guidelines for providing additional feeding counselling for underweight infants beyond 6 months of life, recommendations for underweight infants aged under 6 months are the same as those for the general population of infants, namely to promote exclusive breastfeeding.21

Malnutrition indicators calculated using WHO growth standards had higher AUCs, indicating that their performance may be better for predicting subsequent mortality than those calculated using the NCHS growth reference.

These findings have important implications. The higher prevalence of malnutrition between birth and 6 months of age found using the new WHO child growth standards challenges the traditionally held belief that malnutrition largely begins during the period of complementary feeding between 6 and 24 months of age. Our results add to the limited body of evidence on the relationship between the duration of exclusive breastfeeding and growth during the first 6 months of life. There appears to be little, if any, evidence in this study that exclusive breastfeeding is associated with better growth. However, previous studies provide clear evidence that exclusive breastfeeding reduces the risk of death and illness. Finally, our analysis shows that being judged underweight on the basis of WHO child growth standards may be a better predictor of the risk of death than other indicators or being judged underweight on the basis of the NCHS growth reference. This finding provides an additional reason for countries to implement WHO child growth standards as soon as possible. ■


Acknowledgements

We thank Dr Mercedes de Onis, Department of Nutrition for Health and Development, WHO, Geneva, Switzerland, for reviewing this paper and providing very helpful comments. We also thank members of the WHO Division of Child Health and Development Immunization-Linked Vitamin A Group: P Arthur, Kintampo Health Research Centre, Ghana and London School of Hygiene and Tropical Medicine, London, United Kingdom; S Morris, London School of Hygiene and Tropical Medicine, London, United Kingdom; S Amenga–Etego, C Zandoh and O Boahen, Kintampo Health Research Centre, Ghana; N Bhandari and MK Bhan, All India Institute of Medical Sciences, New Delhi, India; MA Wahed, International Center for Diarrhoeal Disease Research, Bangladesh; CF Lanata, B Butron, AR Huapaya and KB Rivera, Instituto de Investigacion Nutricional, Lima, Peru; for data management, LH Moulton, M Ram, CL Kjolhede and L Propper, Department of International Health, Johns Hopkins University, Baltimore, MA, USA; and for coordination, J Martines and B Underwood, WHO, Geneva, Switzerland.

Competing interests: None declared.

References

Share