4. Example of stunting: Egypt project

4.1. Infant length and its determinants

The mean birth length of infants was normal, with a Z score of 0.07 ±0.54 (SD). As in the Mexican and Kenyan children, the onset of deceleration of linear growth in the Egyptian children also started at 3 to 4 months of age (Table 11) (Rahmanifar et al., 1993). By 6 months of age the mean Z score was -1.01 ±0.03 (SD). There was no significant deficit in weight-for-length (WL). The main determinant of faltering in length was the percent of time ill with diarrheal episodes (R² = 0.09, P < 0.09, n = 34). Egyptian infants are fed sugar water daily, a likely source of pathogens.

4.2. Toddlers' length and its determinants

The Egyptian toddlers, at entry into the study at 18 months of age, were stunted with a mean Z score for LA of -1.94, and a normal WL (Table 11).

With only household and parental size variables available, multiple regression analysis was used to examine the length of the 18 month old child upon entry into the study. The best regression model for predicting toddler's size at 18 months included household sanitation, which accounted for 45% of the variation in height of boys. For girls only maternal height entered the equation. By comparison, the length of Kenyan toddlers at 18 months was predicted mainly by socioeconomic status and maternal and paternal height.

Table 11. Z scores for length-for-age and weight-for-age in 0 to 6-month-old Egyptian infants*, toddlers, and schoolchildren

* Semi-longitudinal data-sexes combined.
** na = Not available.

Linear growth velocity from 18 to 30 months remained at 70% of expected growth velocity. No relationships between initial weight and weight velocity, or initial length and length velocity were found. The most stunted children did not necessarily remain so, nor did they necessarily show catch-up growth.

The best predictor of rate of gain in linear growth over the 18 to 30 month period, was whether the child was still being breast fed after the age of 18 months. This variable accounted for 8% of the variation in length gain after controlling for sex, length at 18 months and for maternal height. None of the expected factors contributed additional predictive power (e.g. food intake, socioeconomic status, morbidity rate). About 60% of Egyptian toddlers had been completely weaned by 18 months, most between 15 to 18 months of age. The 40% still partially breast fed at 18 months were completely weaned between 18 to 28 months.

Partially breast fed toddlers received a mean of about 78% as much energy from foods other than breast milk as fully weaned toddlers. Thus, continued breast feeding at this age contributed, inferentially, less than a quarter of total calories. However, the effect appears to be not only due to the enhanced dietary quality and bioavailability provided by the relatively small amount of breast milk, but also to the fact that the quality of the non-breast-milk food was very different before and after complete weaning (Table 12). Toddler diets were essentially indistinguishable from those of schoolchildren and adults in terms of the percent of dietary energy from the various food groups. However, before complete weaning, toddlers received significantly less traditional bread and vegetables and more nutrient rich dairy products and chicken. Differences in micronutrient intake for weaned and non-weaned toddlers are shown in Table 13.

Weaning is a very significant developmental milestone in the Egyptian culture, and implies an important shift in all aspects of child care. The fully weaned child is essentially on its own with regard to participation in the household diet, while the still breast fed child is more actively fed and cared for by the mother. However, children weaned later do not get any less diarrhea than children weaned earlier.

Table 12. Comparison of Egyptian toddler diets before and after complete weaning (mean % of dietary energy)

4.3. School-age children's height arid its determinants

Moderate stunting was seen in the schoolchildren with a mean Z score for HA of -1.03 ±0.93 (SD). Less stunting was seen than in toddlers, with some catch-up having occurred. However, over 51% of schoolchildren had HA Z scores of -1, and 16.4% had HA Z scores below -2.

Parental height, but not weight or BMI, predicted both initial HA and WA for the 7-year-old children upon entry into the study. For girls, sanitation variables were powerful predictors of body size. A multiple regression model including father's height, a kitchen contained within the household and piped water explained 50% of the variation in height for girls. For boys maternal and paternal height and sanitation predicted 30% of the variability in height.

For height gain in boys from age 7 to 9 years, maternal literacy accounted for 9% of the variability. For the girls, maternal literacy and percent energy from meat (both positive), and phytate:zinc molar ratio and household crowding (both negative) accounted for 23% of the variation in linear growth. Thus diet quality was relatively more important for girls, who may have been more deprived than the boys. Maternal education was an important factor in height gain both in boys and girls. As for birth order, being the oldest child had a negative impact on height for girls but not for boys.

Table 13. Dietary quality for toddlers before and after complete weaning; percent of mineral and vitamin intake from food groups (mean)