Background: There is growing evidence in animal and in vitro studies showing that thiazolidinediones (TZDs) improve pancreatic beta cell (β-cell) function. The aim of this study was to evaluate the effect of thiazolidinediones on the β-cell function of patients with Type 2 diabetes mellitus (T2DM) in clinical practice. Patients and Methods: This is an observational, nested case-control study. We identified 28 patients (TZD group), with T2DM, who had had a meal-stimulated C-peptide level documented before the addition of troglitazone to a failing double-therapy regimen with metformin (MET) and sulphonylurea (SU). As a control group (CTRL), we identified 26 patients, with T2DM, who also had had a meal-stimulated C-peptide documented before adding MET to a failing SU monotherapy regimen. We then proceeded to prospectively remeasure their meal-stimulated C-peptide levels and compared the changes over time between the two groups. Results: Both groups were well matched for age, body mass index (BMI), and HgbA1c before and after treatment. The C-peptide in the TZD group increased significantly during therapy (from 3.2 ± 0.5 to 4.2 ± 0.5, p = 0.01), whereas it remained unchanged in the CTRL group (from 4.8 ± 0.6 to 5.0 ± 0.5, p = 0.74). The C-peptide/glucose ratio also increased significantly in the TZD group (from 1.9 ± 0.3 to 3.1 ± 0.3, p = 0.0003) whereas it remained unchanged in the CTRL group (from 3.4 ± 0.7 to 3.4 ± 0.3, p = 0.97). Furthermore, the C-peptide/glucose ratio of the TZD group, which was significantly lower at baseline compared with the CTRL group (1.9 ± 0.3 vs. 3.4 ± 0.7, p = 0.04), caught up to its level during treatment (3.1 ± 0.3 vs. 3.4 ± 0.3, p = 0.48). Conclusion: Thiazolidinediones seem to induce recovery of pancreatic beta cell function, independently of the correction of glucose toxicity.