Determinants of linear growth and predictors of severe stunting during infancy in rural Malawi


Per Ashorn, University of Tampere Medical School, PO Box 607, FI-33101 Tampere, Finland (Tel. +358 3 215 8410, fax. +358 3 215 8420, e-mail.


Stunting is common among children under 5 y of age in sub-Saharan Africa. Several risk factors have been associated with poor growth but few studies have prospectively addressed the development of linear growth faltering and stunting during the first year of life. The present study was designed to analyse typical growth among rural Malawian infants, focusing particularly on the impact of birth size, adherence to feeding guidelines and morbidity in the development of severe stunting during infancy. A community-based cohort of 613 singleton newborns was prospectively followed by monthly home visits. Data were collected on the children's socioeconomic background, maternal size and weight gain during pregnancy, birth events, morbidity, breastfeeding and complementary feeding, growth and mortality. Univariate and multivariate analyses were used to determine associations between predictor variables and poor linear growth. The proportions of stunted infants (Height-for-age Z-score > –2) at 3, 6 and 9 mo of age were 27%, 51%, and 63%, respectively. At 1 y of age, over two-thirds (71%) of the infants were at least moderately (HAZ > –2) and 31% severely stunted (HAZ > –3).

Conclusion: The strongest predictor of severe stunting at 12 mo of age was small birth size. Other variables independently associated with this outcome included inappropriate complementary feeding, high morbidity, maternal short stature, male gender, and home delivery. Faltering of linear growth started soon after birth and continued throughout infancy. Interventions increasing birth size could have a significant role in the prevention of early childhood stunting. The ideal strategy should also emphasize the importance of appropriate infant feeding and decreasing the number of illness episodes amongst the infants.