Background Malnutrition among HIV-infected children in resource-limited settings is common, but rapid weight gain usually follows ART initiation. The effect of ready-to-use therapeutic foods (RUTF) in this context is unclear. The nutritional status of a child is likely to influence both the use and effect of RUTF, so causal models are required to establish benefit.
Methods ARROW is an ongoing randomised trial of monitoring and treatment strategies in ART-naïve children in Uganda/Zimbabwe. Enrolment was from May 2007 to November 2008; nutritional supplement data was collected from December 2007. Three of four centres gave Plumpy'nut, a peanut-based RUTF, using criteria that included weight-for-height, mid-upper arm circumference, clinical status and knowledge of food insecurity. Marginal structural models with stabilized time-dependent inverse probability treatment weights were used to estimate the causal effect of Plumpy’nut on weight-for-age Z-scores (WAZ) over the first 48 weeks on ART among children aged less than 9 years who were enrolled after December 2007, stratifying by centre to compute weights.
Results 387 children aged less than 9years at ART initiation were enrolled from the 3 centres after Dec 2007; 207 (53%) were female, median(IQR) age was 3(1,6) years, weight-for-age Z-score -1.8(-2.9,-1.0), height-for-age Zscore -2.4(-3.4,-1.3) and weight-for-height Z-score -0.6(-1.4,0.2). Plumpy’nut was given to 66 (17%) children for median(IQR) 17(6,30) weeks, starting 2(2,8) weeks after enrolment. Time-dependent predictors of Plumpy’nut use were age, WAZ, weight less than 80% expected, mid-upper arm circumference and previous hospital admission. Children given Plumpy’nut had lower WAZ at baseline (-4.0(-4.7,-3.0) vs -1.5(-2.2,-0.8)). In unadjusted analysis, median improvement in WAZ at 48 weeks was greater among those receiving Plumpy’nut (+2.0 vs +0.5); however this reflected greater weight gain in children starting from very low WAZ regardless of Plumpy’nut use. In adjusted analysis, the causal effect of Plumpy’nut was small and not statistically significant (+0.02 per month 95%CI (-0.03,0.08); p=0.45). Conclusion Malnourished HIV-infected children initiating ART plus RUTF generally experience rapid weight gain. However, a marginal structural model applied to observational data with time-varying confounders suggests that the weight gain was not attributable to the use of RUTF. Routine use of RUTF for stable, malnourished children entering ART programs may therefore not be clinically beneficial.
Bwakura-Dangarembizi, M.; Cook, A.; Karungi, C.; Kasirye, P.; Mapinge, F.; Musinguzi, M.; Nahirya-Ntege, P.; Babirye, A.; Najjuko, J.; Prendergast, A. Weight gain in children initiating ART is not increased by nutritional supplements. Presented at the 19th Conference on retroviruses and opportunistic infections, 5-8 March 2012, Seattle, WA, USA. [Poster 718]. (2012) 1 pp.