Summary: In adults with new-onset type 2 diabetes who maintained HbA1c below 6.5% for 6 years in the EDICT trial, initial triple therapy with metformin, pioglitazone and exenatide (n=29) was associated with slower carotid intima-media thickness progression and lower liver fat content and fibrosis than sequential add-on therapy with metformin, glipizide and insulin glargine (n=26). Mean HbA1c was 5.7% versus 6.0% (not significant), so the differences in surrogate markers were observed despite comparable glycaemic control. This is a small secondary analysis of surrogate endpoints, not a clinical-outcome trial.
PICO Summary
| Element | Detail |
|---|---|
| Population | Adults with new-onset type 2 diabetes who maintained HbA1c below 6.5% for 6 years; secondary analysis of the single-centre EDICT randomised controlled trial (UT Health San Antonio, USA; NCT01107717). Total analysed n=55. |
| Intervention | Initial fixed triple therapy with metformin, pioglitazone and exenatide (n=29). |
| Comparison | Sequential add-on therapy: metformin first, then stepwise addition of glipizide and then insulin glargine (n=26). |
| Outcome | Mean HbA1c at 6 years 5.7% (triple) versus 6.0% (sequential), p=NS. The sequential group showed greater carotid intima-media thickness progression and greater liver fat content and fibrosis than the triple-therapy group. The abstract reports no effect sizes, 95% confidence intervals or p-values for these imaging outcomes, and no ARR or NNT; results are described in associational terms. |
Expert Commentary
This secondary analysis of the EDICT trial is best read as hypothesis-generating rather than practice-changing. The signal is biologically coherent: a regimen containing pioglitazone and a GLP-1 receptor agonist was associated with slower carotid intima-media thickness progression and lower hepatic fat and fibrosis than a sulfonylurea-and-insulin pathway, and these differences emerged at matched HbA1c, which is the analysis’s most interesting feature. Several limitations temper enthusiasm. The cohort is very small (29 versus 26) and is restricted to participants who sustained tight control for 6 years, so the groups are a selected, better-prognosis subset rather than the full randomised population, weakening causal inference. The comparison is confounded by drug class, the trial cannot be blinded given the divergent regimens and insulin titration, and the abstract reports the imaging outcomes without effect sizes, confidence intervals or p-values, so the precision of the surrogate-marker differences cannot be judged. These are imaging surrogates, not cardiovascular events or biopsy-confirmed liver outcomes. Can I use this with my patients? Not yet as a reason to start triple therapy upfront; it is a fair talking point when weighing pioglitazone or a GLP-1 agonist in someone with new-onset diabetes and fatty liver. Adequately powered trials with hard endpoints are needed.
References
Abdul-Ghani M, Puckett C, Abdelgani S, Merovci A, Lavrynenko O, Adams J, Triplitt C, DeFronzo RA. Glycemic and non-glycemic benefits of initial triple therapy versus sequential add-on therapy in patients with new-onset diabetes: results from the EDICT study. BMJ Open Diabetes Res Care. 2025;13(2):e004981. doi:10.1136/bmjdrc-2025-004981