systematic review and meta-analysis of clinical studies
OBJECTIVE: To investigate glucose-dependent insulinotropic polypeptide (GIP) secretion in patients with type 2 diabetes and nondiabetic control subjects during oral glucose or meal tests. RESEARCH DESIGN AND METHODS: Eligible trials were identified by The Cochrane Library, MEDLINE, Embase, and Web of Science. Data were retrieved and random-effects models for the primary meta-analysis, random-effects meta-regression, and subgroup and regression analyses were applied. RESULTS: Random-effects meta-analysis of GIP responses in 23 trials during 28 different stimulation tests showed that patients with type 2 diabetes (n=363) exhibited no significant differences (P=not significant) in peak plasma GIP, total area under the curve (tAUC), time-corrected tAUC (tAUC×min(-1)), and time-corrected incremental area under the curve (iAUC×min(-1)) in comparison with nondiabetic control subjects (n=325) but had lower GIP responses as evaluated from iAUC (weighted mean difference, -648 pmol/L×min; 95% CI, -1,276 to -21). Fixed-effects models meta-analyses confirmed most of the results of the primary meta-analysis but showed iAUC×min(-1) to be reduced and showed tAUC and tAUC×min(-1) to be higher in diabetic patients. Random-effects meta-regression of the primary meta-analysis showed that age (peak GIP, tAUC, iAUC, and iAUC×min(-1)), BMI (tAUC, iAUC, and iAUC×min(-1)), and HbA1c (iAUC and iAUC×min(-1)) predicted some of the GIP outcomes. Post hoc subgroup analysis showed a negative influence of age and of HbA1c on GIP responses and showed a positive influence of BMI on GIP responses. CONCLUSIONS: Our results suggest that patients with type 2 diabetes are characterized by preserved GIP secretion in response to oral glucose and meal tests. They also suggest that high BMI is associated with increased GIP responses but increasing age and HbA1c are associated with reduced GIP secretion.