Abstract
We examined the association between low-calorie sweeteners (LCS) consumption during preconception, pregnancy, and breastfeeding and child health outcomes. A systematic search of electronic databases in PubMed, Embase, Cumulated Index to Nursing and Allied Health Literature, the Cochrane Library, Scopus, Web of Science, PsycINFO, ProQuest Health and Medical, ClinicalTrials.gov, and Google Scholar was conducted up to 21 September 2023. A random effects model with restricted maximum likelihood estimation was used for the meta-analysis. Seventeen eligible studies were included. The standardised mean difference (SMD) and 95% confidence interval (CI) in birth weight between those who frequently consumed LCS (≥1 serve/day) during pregnancy and those who did not consume LCS was 0.04 (0.00, 0.08) (four cohort studies). Any LCS consumption during pregnancy compared with no consumption was not associated with birth weight [SMD (95% CI) = 0.03 (-0.03, 0.08)] (four cohort studies). Any LCS consumption during pregnancy was not associated with body mass index z-scores. The weighted mean difference (95% CI) was 0.00 (-0.05, 0.06) at birth, 0.06 (-0.29, 0.40) at 6 months, -0.04 (-0.19, 0.10) at 1 year, 0.00 (-0.16, 0.17) at 3 years, and 0.10 (-0.15, 0.34) at 7 years of the child age, compared with no intake (five cohort studies). The odds of being overweight at 1 year among children exposed to LCS during pregnancy was 1.19 (OR [95% CI]: 1.19 [0.81, 1.58]) compared with unexposed children (two cohort studies). The effect sizes were not precise for all the outcomes as the 95% CI indicated the effect estimates could range from small protective to a higher risk. The effect of LCS consumption on child behaviour and cognition was inconsistent. There is not enough evidence to confirm LCS consumption during pregnancy affects birth weight and risk of overweight in children. However, frequent consumption increased birth weight and the risk of overweight at different ages, though the effects were imprecise. More robust research evidence is required as the quality of evidence is low.