Figure 3

Funnel plot of the highest compared to lowest FPG categories from only 10 studies included in the dose-response meta-analysis.
The funnel plot was based on 4 small studies (fewer than 100 cancer cases for each exposure dose) and 6 large studies. Among the 10 published studies, the results of 2 studies were statistically significant (P < 0.05) and the results of 8 studies were not significant (P > 0.05). For the 2 studies with significant results, one study was small, whereas the other came from the group of 6 large studies. No significant publication bias was detected (P = 0.125 for Egger’s test, P = 0.283 for Begg’s test). This funnel plot shows asymmetry, which might be related to reasons other than publication bias.