Predicting malnutrition in South African preterm infants in kangaroo mother care

Preterm infants often exhibit poor short- and long-term growth. The World Health Organization recommends kangaroo mother care for all preterm and low birth weight infants, emphasising skin-to-skin contact, exclusive human milk feeding, and continued skin-to-skin care post-discharge. While kangaroo mother care supports short-term growth in weight, length, and head circumference, long-term outcomes are less understood.

The study reported on in this paper evaluated one-year anthropometric outcomes in a South African cohort of preterm infants (<37 weeks gestation) who received early-life kangaroo mother care by analysing longitudinally collected routine clinical data. The study compared outcomes of small-for-gestational-age infants with those of appropriate-for-gestational-age infants and analysed various predictors of one-year anthropometric outcomes.

At one year, small-for-gestational-age infants (n=111) had statistically significantly lower weight-for-age (WAZ), length-for-age, weight-for-length, and body mass index-for-age z-scores compared to appropriate-for-gestational-age infants (n=210). They also had higher rates of stunting (34.2% vs. 9.1%; p<0.001), underweight (31.2% vs. 7.2%; p<0.001), and wasting (12.6% vs. 4.3%, p=0.012), with no significant difference in overweight (4.5% vs. 7.7%, p=0.397).

The odds of underweight were increased by maternal pregnancy conditions and decreased by higher birth weight z-scores (BWZ) and early WAZ gains. Stunting odds were increased by being small for gestational age and decreased by higher BWZ and early WAZ gains. Wasting odds were increased by congenital heart conditions and decreased by higher BWZ. Overweight odds were only increased by greater early WAZ gains. However, the regression models had low R² values, indicating poor predictive ability for these outcomes.

A comparison of infants who had been followed up for a full year (‘included infants’) and a random sample of 489 infants with less than one year follow-up data (‘non-included infants’) revealed significant differences. Non-included infants had a statistically significantly higher gestational age and BWZ. However, the study sample had a higher proportion of small-for-gestational-age infants due to deliberate over-sampling; BWZ differences were not significant when appropriate-for-gestational-age and small-for-gestational-age infants were separated, though the gestational age difference remained. Non-included infants had higher rates of maternal HIV infection (29.7% vs. 21.8%, p=0.013) and fewer congenital heart defects (24.7% vs. 31.2%, p=0.045).

A key limitation of the study was potential selection bias due to early discharge, missed clinic visits, or loss to follow-up. The routine clinical data used in the study is likely to have been of a lower quality than data in research settings, though an experienced dietitian likely improved measurement reliability. Incomplete records, especially on maternal comorbidities, might have underestimated some conditions. Future research should follow infants beyond two years to fully assess catch-up growth and changes in stunting and overweight rates over time.

The study concluded that preterm-born small-for-gestational-age infants remain more underweight, stunted, and wasted than their preterm-born appropriate-for-gestational-age peers at one year, despite greater WAZ gains. Targeted interventions are needed to promote appropriate catch-up growth, especially for small-for-gestational-age preterm infants.