Table 3:

Results from meta-analyses (where possible) or individual studies for intake of nonnutritive sweeteners and long-term cardiometabolic health outcomes in randomized controlled trials and cohort studies

Outcome: change or incidenceNo. of studies* (participants)ComparisonEstimate of NNS effect (95% CI) from meta-analysis or individual studiesAssoc.Citation(s)*Figure
Randomized controlled trials
BMI3 (242)NNS v. controlMD −0.37 kg/m2 (−1.10 to 0.36), I2 9%NS20, 36, 372
Weight5 (791)NNS v. controlSMD −0.17 (−0.54 to 0.21), I2 81%NS19, 20, 34, 35, 382
Percentage of fat mass1 (25)NNS v. controlMD −1.01% (−3.01 to 0.99)NS35–
Waist circumference3 (683)NNS v. controlSMD −0.16 (−0.56 to 0.25), I2 83%NS19, 20, 34S1‡
Insulin resistance: HOMA-IR3 (99)NNS v. controlSMD +0.10 (−0.57 to 0.76), I2 55%NS20, 35, 37S3‡
HbA1c1 (62)NNS v. controlMD +0.07% (−0.00 to 0.14)NS20–
Cohort studies
BMI2 (21 256)Continuous correlationWMC +0.05 (0.03 to 0.06), I2 0%↑ Gain14, 152
1 (3371)Highest NNS intake quantile v. noneMD +0.77 kg/m2 (0.47 to 1.07)↑ Gain50–
Weight4 (32 405)Continuous correlationWMC +0.06 (0.05 to 0.07), I2 46%↑ Gain21, 572
Gestational weight gain1 (347)Highest v. lowest NNS intake quantileMD +2.5 kg (0.5 to 4.5)↑ Gain23–
Weight gain > 1 kg1 (7,194)Highest v. lowest NNS intake quantileOR 1.05 (0.93 to 1.19)NS41–
Waist circumference1 (384)Daily v. no NNS consumptionMD +2.27 cm (0.96 to 3.58)↑ Gain18–
Incident abdominal obesity1 (5011)Highest v. lowest NNS intake quantileHR 1.59 (1.23 to 2.07)↑ Gain60
Incident overweight/obesity3 (7917)Highest v. lowest NNS intake quantileOR 1.84 (1.28 to 2.66), I2 0%↑ Risk22, 50, 59S1‡
Metabolic syndrome5 (27 914)Highest v. lowest NNS intake quantileRR 1.31 (1.23 to 1.40), I2 0%↑ Risk39, 47, 48, 54, 60S2‡
Type 2 diabetes4 (221 363)Per daily serving of NNSRR 1.03 (1.01 to 1.05), I2 0%↑ Risk24, 42, 56S2‡
9 (400 571)Highest v. lowest NNS intake quantileRR 1.14 (1.05 to 1.25), I2 52%↑ Risk16, 24, 42, 49, 55, 56, 58, 602
Gestational diabetes1 (13 475)Highest v. lowest NNS intake quantileRR 0.87 (0.71 to 1.02)NS44–
Impaired glucose tolerance1 (3728)No v. any NNS consumptionHR 1.07 (0.91 to 1.26)NS48–
Hypertension5 (232 630)Highest v. lowest NNS intake quantileHR 1.12 (1.08 to 1.13), I2 53%↑ Risk45, 48, 60S4‡
Stroke2 (128 176)Highest v. lowest NNS intake quantileRR 1.14 (1.04 to 1.26), I2 0%↑ Risk40S4‡
Cardiovascular events†2 (62 178)Highest v. lowest NNS intake quantileRR 1.32 (1.15 to 1.52), I2 0%↑ Risk17, 52S4‡
Coronary heart disease2 (131 403)Highest v. lowest NNS intake quantileRR 0.98 (0.90 to 1.07), I2 0%NS46, 51S4‡
Chronic kidney disease1 (14 002)Highest v. lowest NNS intake quantileOR 0.80 (0.64 to 1.00)NS43–

  • Note: BMI = body mass index, CI = confidence interval, HbA1c = glycosylated hemoglobin, HOMA-IR = homeostatic model assessment for insulin resistance, HR = hazard ratio, MD = mean difference, NNS = nonnutrititve sweetener, NS = not significant, OR = odds ratio, RR = risk ratio, SMD = standardized mean difference, WMC = weighted mean group correlation (unitless).

  • ↵* Number of studies does not always equal the number of citations, because some citations report results from multiple studies.

  • ↵† Defined by the study authors as coronary heart disease, heart failure, myocardial infarction, coronary revascularization procedure, ischemic stroke, peripheral arterial disease and cardiovascular death;17 or stroke, myocardial infarction and vascular death.52

  • ↵‡ Appendix 1, available at www.cmaj.ca/lookup/suppl/doi:10.1503/cmaj.161390/-/DC1.