Safety
Four included trials reported the AEs during the trial phase14,15,19,23, and 307 out of a total of 605 patients had at least one AE. As shown in Figure 5 , GKAs did not significantly increase the incidence of at least one AE compared with placebo (RR 1.11, 95% CI: 0.95 to 1.3, P = 0.19). Of note, GKAs increased the incidence of hypoglycemic events(RR 1.181, 95% CI: 1.35 to 2.42, P <0.0001), but they did not increase the relative risk of headache(RR 1.3, 95% CI: 0.49 to 3.43, P = 0.60), diarrhea(RR 1.64, 95% CI: 0.69 to 3.87, P = 0.26), and nausea(RR 2.65,95%CI: 0.53 to 13.26,P=0.24).
Then, we did further subgroup meta-analysis according to the type of GKAs. As shown in Figure 6 , we found that there was no significant difference in the risk of hypoglycemia between Liver-selective glucokinase activator and placebo (RR 1.59, 95% CI: 0.58 to 4.39, P = 0.37), but GKAs (pancreas liver double activator) did increase the risk(RR 1.84, 95% CI: 1.36 to 2.49, P <0.0001). And in GKAs (pancreas liver double activator) subgroup analysis, there was heterogeneity (I2 > 50%). It seems that there are significant differences in the effectiveness and incidence of hypoglycemia among different types of GKAs.
Quality of Evidence
Most of the findings were supported by high certainty evidence.Table 2 shows the GRADE summary of findings for FPG, HbA1c, Total AEs, Hypoglycemia, Headache, Diarrhea and Nausea.