Effect of complementary feeding with lipid‐based nutrient supplements and corn–soy blend on the incidence of stunting and linear growth among 6‐ to 18‐month‐old infants and children in rural Malawi

Introduction

Almost one‐third of all under 5‐year‐old children in low‐ or middle‐income countries and over 40% of those living in Africa are estimated to be stunted, i.e. they have suffered from linear growth failure that has made them shorter than expected for their age (Black et al. 2008). Stunting is not only associated with reduced final height, but also with increased morbidity, mortality, developmental delay or deficits, poor school performance and lower cognitive function in childhood, and less income as an adult (Pelletier et al. 1995; Grantham‐McGregor et al. 2007; Victora et al. 2008). Given these adverse outcomes and the frequency of the condition, it is not surprising that prevention of childhood stunting has been identified as a major global health priority (UNICEF 2009).

On a population level, stunting is associated with a number of harmful exposures, including inadequate diet and frequent infections (Martorell et al. 1994). However, most interventions addressing these risk factors have so far failed to markedly improve linear growth or prevent stunting among infants and young children (Dewey & Adu‐Afarwuah 2008). Recently results from studies in Ghana and Malawi suggested that consumption of lipid‐based nutrient supplements (LNS), a novel class of micronutrient–fortified, ready‐to‐use products, might boost length gain and reduce growth failure and the incidence of severe stunting among 6–18‐month‐old infants (Adu‐Afarwuah et al. 2007; Phuka et al. 2008; Phuka et al. 2009). While these findings were encouraging, they are also limited; the Malawi study did not include a parallel no‐intervention control group and the non‐intervention group in Ghana study had no weekly follow‐up visits compared with the intervention groups.

The main aim of the present study was to test a hypothesis that provision of LNS would reduce the incidence of severe stunting or other forms of linear growth faltering among 6–18‐month‐old infants in Malawi, southern Africa. Two separate LNS products were tested, one that contained dried skim milk and thus resembled products used in earlier prevention trials with LNS (Adu‐Afarwuah et al. 2007; Phuka et al. 2008), and another one in which the milk powder was substituted with soy flour. The latter choice was motivated by the use of soy in nationally recommended complementary foods (Malawi MoHP 2010), its cheaper price and also its acceptability to vegetarians. To improve on the earlier trials, we also designed the study with two control groups: one of these included children who received no intervention during the period of interest while participants in the other control group were given rations of micronutrient‐fortified corn–soy blend (CSB), a nationally recommended complementary food for infants and young children (Malawi MoHP 2010). Because of our earlier finding that LNS provision might have the strongest impact on the most disadvantaged infants (Phuka et al. 2009), we selected the incidence of severe stunting as the primary outcome measure.

Methods

Results

Of the 1385 infants who were identified through community census, 490 were either ineligible or not brought to an enrolment session. A further 55 were found ineligible in a more detailed assessment. The remaining 840 infants were randomised into four intervention groups as shown in Fig. 1. None of the infants was allergic to a 5‐g test dose of milk–LNS.

Table 2 shows the baseline demographics, anthropometrics and parental characteristics of the participants by intervention group. The summary statistics were largely comparable in the four groups although the CSB group had slightly fewer boys and lower mean initial weight and LAZ. At enrolment, the prevalence of severe and very severe stunting in the combined sample was approximately 8.5% and 3.5%, respectively. All except one child were breastfeeding at enrolment, but none of them exclusively. The proportion of children receiving breast milk remained very high throughout the study (99.7% at 12 months and 96.8% at 18 months of age).

During the 12 month follow up, 25 children (3.0%) died and 68 (8.1%) dropped out (Fig. 1). Hence, final measurements were obtained from 747/840 (88.9%) participants. The success of follow‐up was not significantly different between intervention groups (P = 0.852). A total of 45 participants had no anthropometric data at all after enrolment (13 deaths, 32 drop‐outs) and there was no difference in this proportion between the intervention groups (P = 0.847).

All mothers reported that their infants readily ate the provided supplement and diversion of any portion to someone other than the intended beneficiary was reported at only 69 of 18 906 (0.36%) supplement delivery interviews: 29 in milk–LNS, 19 in soy–LNS, and 21 in CSB groups (P = 0.383). From the 2‐weekly home visits during which leftover trial products were checked, the percentage of visits with leftovers found were 1.3%, 1.3%, and 0.6% in the milk–LNS, soy–LNS, and CSB groups respectively (P < 0.001).

During the one year intervention period, children who received milk–LNS had the most length gain both in terms of cm and z‐scores whereas children who received CSB had the least. None of the differences between the intervention groups and the control children was, however, statistically significant (Table 3). Looking into the consecutive measurements within the one year period, the biggest difference was seen in the age period between 9 to 12 months (Fig. 2). During this time children who received milk–LNS had a negligible change in their mean LAZ (–0.02 z‐score units) whereas children in the Control, soy–LNS and CSB group dropped their mean LAZ by −0.15, −0.12 and −0.18 z‐score units, respectively (P = 0.045). (Sample size varied from 520 to 747 across age intervals of measurements.) Pairwise comparisons showed statistically significant differences in change in LAZ between milk–LNS and control groups (P = 0.029) and CSB group (P = 0.014), but not soy–LNS group (P = 0.079) in this age interval.

A total of 86 infants developed severe stunting during the intervention. The number of those who developed very severe stunting, moderate‐to‐severe stunting, severe underweight, and severe wasting were 50, 177, 36, and 11, respectively. Children in the milk–LNS group had less of these events whereas those in the CSB group had more (except very severe stunting), but none of the differences between the groups was statistically significant (Table 4). At the end of the 12‐month intervention, the prevalence of severe stunting and other forms of undernutrition was slightly lower in the milk–LNS group than the other groups but again the differences were not statistically significant (Table 5).

All interventions were well tolerated by the participants. Besides the 25 deaths (8 in control, 4 in milk–LNS, 5 in soy–LNS, 8 in CSB; Fig. 1), 34 other participants were hospitalised (12 in control, 7 in milk–LNS, 6 in soy–LNS, 9 in CSB) and 3 were recorded as having experienced another SAE during the follow‐up. Fifty‐nine (95.2%) of the SAEs were considered unrelated and the rest probably unrelated to the trial interventions.

Discussion

This trial was carried out to test a hypothesis that 12‐month‐long dietary supplementation with LNS would promote linear growth and reduce the incidence of severe growth faltering among rural Malawian infants and young children. Although statistically not significant, the sample findings were consistent with the hypothesis, as participants receiving dietary complementation with milk‐containing LNS gained on average approximately 0.2 cm more in length and 110 g more in weight and had a lower incidence of severe stunting during the follow‐up than control children. The internal validity of the trial was high because of broad inclusion criteria, random group allocation, similarity of the intervention groups at enrolment, comprehensive follow‐up, and blinding of the outcome assessors. The between‐group differences were, however, smaller than anticipated and in many analyses the probability of the observed differences being due to random variation was above the traditionally accepted 5% threshold. Because of this, the primary null‐hypothesis of no difference between the groups could not be conclusively rejected. Nevertheless, the inherent consistency of the findings on various aspects of growth, as well as their biological plausibility and coherence with earlier studies (Adu‐Afarwuah et al. 2007; Phuka et al. 2008; Phuka et al. 2009) suggest a causal relationship between the milk–LNS intervention and moderately improved growth outcomes.

Age–interval‐specific analyses from our sample indicated that the mean length‐for‐age remained constant from 6 to 12 months among participants who received milk–LNS (i.e. no growth faltering occurred), whereas it decreased by 0.1–0.2 z‐score units between 9 and 12 months of age among other infants. From the age of 12–18 months, the mean length‐for‐age fell at approximately equal rate in all groups. One possible explanation for these findings is a specific effect of milk powder on linear growth in early childhood. According to the Infancy–Childhood–Puberty model of growth, the interval from 6 to 12 months of age is the period when an infant changes from the infancy phase of growth to the childhood phase (Karlberg 1989; Liu et al. 1998a, 2000). A delay in this shift [called the infancy‐childhood (IC) spurt] has been associated with the high incidence of growth retardation in some low‐income settings (Liu et al. 1998a,b). Factors leading to the IC spurt remain unclear, but insulin‐like growth factor and numerous other biologically active substances that are present in cow's milk – and hence also in milk–LNS – have been suggested to play a role (Low et al. 2001; Wiley 2005; Wiley 2012). This hypothesis is supported by several studies suggesting that animal source foods, especially bovine milk promote length or height gain in human children (Hoppe et al. 2006; de Beer 2012).

The apparent difference in linear growth and the incidence of severe stunting between participants who received milk‐containing as opposed to soy‐containing LNS could be a spurious finding or related to the lack of milk or other differences in the composition of the soy–LNS. Depending on the type and extent of processing, soy products contain antinutrients that might negatively affect linear growth, such as enzyme inhibitors, phytate and lectins (Sarwar 1997; Hoppe et al. 2008). Although soy is cheaper than milk, has a favorable protein profile, and is increasingly used in therapeutic LNS preparations (de Pee & Bloem 2009), our results do not support its use in the formulation for preventive LNS. Similarly, the observation that the growth outcomes were worse among participants who received fortified CSB than among non‐supplemented controls (e.g. incidence of moderate‐to‐severe stunting was 39.7% vs. 29.7%, respectively) does not lend support to the Malawian national recommendation of using CSB as a complementary food (Malawi MoHP 2010). Unlike the energy‐ and nutrient‐dense LNS, the intake of which does not appear to alter breast milk intakes (Galpin et al. 2007), CSB is typically fed as a high‐volume dilute porridge, which may more easily result in displacement of breast milk or other nutritious foods in the child's diet. The relatively poor growth outcomes in the CSB group could also explain why the effects observed in the current trial (in which the LNS groups were compared with a no‐supplement control group) were smaller than those observed in a previous trial conducted in the same area in which the LNS groups were compared with a group given CSB (Phuka et al. 2009). However, these potential explanations are speculative and require confirmation. The current study was designed to compare each intervention group to the control group, but not to make direct comparisons between the intervention groups.

As a whole, the results from the current trial and those of the two earlier ones (Adu‐Afarwuah et al. 2007; Phuka et al. 2008; Phuka et al. 2009) are consistent with a hypothesis that provision of milk‐containing LNS may slows down the process of infant growth faltering around the time of the transition into the childhood phase of growth in environmental conditions similar to those in rural Malawi. Unfortunately the intervention appears insufficient to promote catch‐up growth or even to maintain normal growth after the first year of life in this population, possibly because of increasing nutrient needs, high incidence and prevalence of morbidity and environmental enteropathy (Prentice & Paul 2000; Goto et al. 2009; Humphrey 2009), or programming of postnatal growth rate due to malnutrition in utero. The failure to reverse linear growth faltering is not surprising, because stunting typically involves multiple aetiologies (Waterlow 1994) and very few single‐pronged interventions have ever achieved catch‐up growth in length, either in a trial or programmatic setting (Dewey & Adu‐Afarwuah 2008). Long‐term reduction in stunting most likely requires a health and nutrition package that comprehensively covers the critical ‘1000 days’, i.e. the period from conception to approximately the point when the child turns 2 years old (Piwoz et al. 2012).

Source of funding

This study was supported by Academy of Finland (grants 200720, 108873, 111685 and 109796), Foundation for Pediatric Research in Finland, Medical Research Fund of Tampere University Hospital, and the American people through the support of the Office of Health, Infectious Disease, and Nutrition, Bureau for Global Health, United States Agency for International Development (USAID), under terms of Cooperative Agreement No. GHN‐A‐00‐08‐00001‐00, through the FANTA‐2 Project, and operated by the Academy for Educational Development (AED). The contents are the responsibility of the authors and do not necessarily reflect the views of the funders, USAID or the United States Government. The micronutrient mixture used in the production of LNS was provided free of charge by Nutriset Inc. (Malaunay, France). John Phuka and Chrissie Thakwalakwa received personal stipends from Nestle Foundation. Yin Bun Cheung was supported by the Singapore Ministry of Health's National Medical Research Council under its Clinician‐Scientist Award.

Conflicts of interest

The authors declare that they have no conflicts of interest.

Contributions

All authors participated in the design of the trial. PA, KM coordinated and supervised the research team. CM, CT coordinated the research team at local centres and were responsible for data collection. YBC, CM designed the details of statistical analysis and analysed the data. CM, KM, PA wrote the first draft of the manuscript. All authors commented on the analysis, reviewed and approved the final manuscript.

Mangani, C. , Maleta, K. , Phuka, J. , Cheung, Y. B. , Thakwalakwa, C. , Dewey, K. , Manary, M. , Puumalainen, T. , and Ashorn, P. (2015) Effect of complementary feeding with lipid‐based nutrient supplements and corn–soy blend on the incidence of stunting and linear growth among 6‐ to 18‐month‐old infants and children in rural Malawi. Matern Child Nutr, 11: 132–143. doi: 10.1111/mcn.12068.