Fifty-three trials (1,126 participants) were included in the review. Sample sizes ranged from 10 to 37 participants (median 17). Three trials had more than 30 participants; all of these were randomised cross-over trials. Thirteen trials (10 trials compared saccharides and three compared non-caloric sweeteners to sucrose) reported one week or longer of follow-up; ten of these trials scored 1 out of 5 on Jadad and none adequately concealed treatment allocation.
In diabetic participants, fructose reduced two-hour blood glucose concentrations by 4.8mmol/L (95% CI -6.34 to -3.29) compared with glucose (five trials, 52 diabetic participants). Results for the comparison of hypocaloric sweeteners with sucrose or high fructose corn syrup in non-diabetic patients were inconclusive.
Based on two trials of 10 weeks or longer (174 participants), non-caloric sweeteners reduced energy intake compared with sucrose groups by approximately 250kcal to 500kcal per day (95% CI 153 to 806). One trial (41 participants over 10 weeks) reported a decrease in body mass index for participants using non-caloric sweetener (BMI change -0.40kg/m²; weight change -1.00kg) compared with participants using sucrose (BMI change 0.50kg/m²; weight change 1.60kg) whose body mass index increased.
Two trials found that high fructose corn syrup or fructose did not increase levels of cholesterol relative to other sweeteners (12 healthy and 10 diabetic participants).
Back-calculation methods did not indicate an inconsistency in the network meta-analysis.