Background: Complementing infant diets with lipid-based nutrient supplements (LNSs) has been suggested to improve growth and reduce morbidity, but the daily quantity and the milk content of LNSs affect their cost.

Objective: We tested the hypotheses that the change in mean length-for-age z score (LAZ) for infants provided with 10–40 g LNSs/d from ages 6 to 18 mo would be greater than that for infants receiving no dietary intervention at the same age and that provision of LNSs that did not contain milk would be as good as milk-containing LNSs in promoting linear growth.

Methods: We enrolled in a randomized single-blind trial 6-mo-old infants who were allocated to 1 of 6 groups to receive 10, 20, or 40 g LNSs/d containing milk powder; 20 or 40 g milk-free LNSs/d; or no supplement until 18 mo of age. The primary outcome was change in LAZ.

Results: Of the 1932 enrolled infants, 78 (4.0%) died and 319 (16.5%) dropped out during the trial. The overall reported supplement consumption was 71.6% of days, with no difference between the groups (P = 0.26). The overall mean ± SD length and LAZ changes were 13.0 ± 2.1 cm and −0.45 ± 0.77 z score units, respectively, which did not differ between the groups (P = 0.66 for length and P = 0.74 for LAZ). The difference in mean LAZ change in the no-milk LNS group compared with the milk LNS group was −0.02 (95% CI: −0.10, 0.06; P = 0.72).

Conclusion: Our results do not support the hypothesis that LNS supplementation during infancy and childhood promotes length gain or prevents stunting between 6 and 18 mo of age in Malawi. This trial was registered at clinicaltrials.gov as NCT00945698.