OBJECTIVES: Low glycaemic index (GI) foods are recommended to improve glycaemic control in diabetes; however, Health Canada\nconsiders that GI food labeling would be misleading and unhelpful, in part, because selected studies suggest that GI values are\ninaccurate due to an effect of ethnicity. Therefore, we conducted a systematic review and meta-analysis to compare the GI of foods\nwhen measured in Caucasians versus non-Caucasians.\nMETHODS: We searched MEDLINE, EMBASE and Cochrane databases for relevant articles. GI differences were aggregated using the\ngeneric inverse variance method (random effects model) and expressed as mean difference (MD) with 95% confidence intervals\n(95% CI). Study quality was assessed based on how well studies complied with official international GI methodology.\nRESULTS: Review of 1288 trials revealed eight eligible studies, including 28 comparisons of GI among 585 non-Caucasians and 971\nCaucasians. Overall, there was borderline significant evidence of higher GI in non-Caucasians than Caucasians (MD, 3.3 (95% CI,\n? 0.1, 6.8); P = 0.06) with significant heterogeneity (I2, 46%; P = 0.005). The GI of eight types of rice was higher in non-Caucasians\nthan Caucasians (MD, 9.5 (95% CI, 3.7, 23.1); P = 0.001), but there was no significant difference for the other 20 foods (MD, 1.0 (95%\nCI, ? 2.5, 4.6); P = 0.57). MD was significantly greater in the four low-quality studies (nine comparisons) than the four high-quality\nstudies (19 comparisons; 7.8 vs 0.7, P = 0.047).\nCONCLUSIONS: With the possible exception of rice, existing evidence suggests that GI values do not differ when measured in\nCaucasians versus non-Caucasians. To confirm these findings high-quality studies using a wide range of foods are required.
Loading....