Effectiveness of a locally produced ready-to-use supplementary food in preventing growth faltering for children under 2 years in Cambodia: a cluster randomised controlled trial.

This cluster randomised controlled trial tested the effectiveness of a locally produced, fish-based, ready-to-use supplementary food (RUSF) to prevent growth faltering (decline in z-scores). Cambodian infants (n= 485), aged 6 to 11 months, were randomised by site to receive the RUSF, Corn-Soy Blend++ (CSB++), micronutrient powders (MNP), or no supplement (control). The intervention was for 6 months. In unadjusted analysis, the control group had statistically significantly decreased weight-for-age z-scores (WAZ; -0.02, 95%CI = -0.03 - -0.01, P= 0.001) and height-for-age z-scores (HAZ; -0.07, 95%CI = -0.09 - -0.05, P < 0.001), and increased mid-upper arm-circumference (MUAC; 0.02cm, 95%CI = 0.01 - 0.04, P = 0.010), but no statistically significant change in weight-for-height z-scores (WHZ). The RUSF group did not differ significantly from the control for WAZ, HAZ or WHZ (in other words, WAZ and HAZ decreased and WHZ did not change), but had increased MUAC in comparison to the control (0.04cm, 95%CI = 0.01 - 0.06, P = 0.008). There were no statistically significant differences between the RUSF group and the CSB++ or MNP groups with respect to WAZ, HAZ, WHZ or MUAC. Interestingly, in adjusted analysis, low consumers of RUSF had increased WAZ, WHZ and MUAC (0.03, 95%CI = 0.01-0.06, P = 0.006; 0.04, 95%CI = 0.01-0.08, P = 0.026; and 0.05cm, 95%CI = 0.02-0.09, P = 0.004, respectively) compared with the control. The novel RUSF, particularly in small quantities, protected against ponderal growth faltering, but the improvements were of limited clinical significance.


Comparators
The RUSF was compared with: 1. CSB++: CSB++ was chosen as a comparator because it is the standard supplementary food that WFP provides to children aged six months to two years to prevent undernutrition. 2. MNP: Sprinkles micronutrient powders were chosen since they are a commonly provided supplement in developing countries, such as Cambodia, with low dietary diversity, and complementary foods with low nutrient density (HF-TAG, 2011) 3. Control: A standard, unsupplemented diet, typically borbor and family foods was chosen as a control. Borbor is the traditional food for weanlings (children transitioning from exclusively milk diets to diets that include complementary foods) and is often the only food given until about nine months.
The active comparators complied with WFP and UNICEF standards for supplementary foods, and had been used and tested in Cambodia and elsewhere (Jack et al., 2012;LaGrone et al., 2012;Manary & Chang, 2012. ). They have been found to be safe and to have no unintended side-effects.  All subjects (N=485) attended baseline. Thereafter, subjects attended 60-75% of data collection sessions. Loss to follow up refers to subjects who failed to attend the endline data collection (n = 192), regardless of how many other data collection points they attended. There were significant differences in loss to follow up between the groups. The MNP group had significantly lower odds of being lost to follow up. Comparisons of the odds of dropping out between the other groups were not statistically significant.
In comparison to the MNP group, subjects in the control group had more than twice the odds of dropping out (OR = 2.37; 95% CI = 1.03, 5.44; p = 0.042), while RUSF subjects had almost four times the odds (OR = 3.89; 95% CI = 1.71, 8.88; p = 0.001), and CSB++ subjects had almost five times the odds of dropping out (OR = 4.84; 95% CI = 2.08, 11.29; p <0.001). In comparison to the control group, subjects in the RUSF and CSB++ groups had about twice the odds of dropping out (OR = 1.65; 95% CI = 0.76, 3.55; p = 0.204; and OR = 2.05; 95% CI = 0.92, 4.57; p = 0.081 respectively) although these were not statistically significant. In comparison to the RUSF group, subjects in the CSB++ groups had slightly greater odds of dropping out (OR = 1.24; 95% CI = 0.57, 2.72; p = 0.585) although this was not statistically significant.
For every additional month that a subject stayed in the study, their odds of dropping out decreased by approximately half (OR = 0.58; 95% CI = 0.54, 0.62; p<0.001). Subjects who were older at baseline had slightly higher odds of dropping out (OR = 1.12; 95% CI = 1.04, 1.01; p=0.002). Subjects whose caregiver had had attended high school or higher had lower odds of dropping out (OR = 0.64; 95% CI = 0.41, 0.99; p=0.047). Subjects whose family were poor card holders had half the odds of dropping out (OR = 0.51; 95% CI = 0.34, 0.77; p=0.001). Sex, primary school education and having diarrhoea in the past two weeks did not make a statistically significant difference.
Appendix Figure 1: Mean anthropometric measures and confidence intervals monthly from baseline to endline for children with baseline and endline measurements.

Mean WAZ monthly from baseline to endline Mean HAZ monthly from baseline to endline
Mean WHZ monthly from baseline to endline Mean MUAC monthly from baseline to endline Appendix Figure 1: Mean anthropometric measures and confidence intervals monthly from baseline to endline for children with baseline and endline measurements Appendix Figure 1 depicts the change in monthly mean anthropometric measures from baseline to endline for the 292 children with baseline and endline measurements. The graphs show that WAZ, HAZ and WHZ decline and MUAC increases. The wide and overlapping confidence intervals of the graphs indicate that there may be no significant difference between the groups.