In addition, any improvement in insulin resistance will diminish the allostatic load on the glucoregulatory system captured by the ''hyperbolic'' relationship between insulin sensitivity and beta-cell function. The key question is whether any therapy goes beyond these two general mechanisms and can specifically protect the beta-cell. Among the commonly used oral antidiabetic agents metformin is beta-cell neutral and sulphonylureas/meglitinides appear to accelerate beta-cell deterioration.

In contrast, GLP-1 based therapies (DPP IV inhibitors and GLP-analogues) and thiazolidinediones have been shown to improve beta-cell function. GLP-1 has been shown to stimulate beta-cell proliferation and inhibit beta-cell apoptosis in vitro but it is unclear how this translates into long-term clinical scenarios.

Thiazolidinediones have also been shown to preserve beta-cell mass through a number of mechanisms. For example, rosiglitazone reduces beta-cell death in ZDF rats and pioglitazone protects beta-cells from glucose and IL1-b induced apoptosis. In humans, a significant improvement in the acute insulin response to IV glucose was observed after 6 months of therapy with rosiglitazone which was independent of reduced glucose toxicity because the control group (insulin treatment) had a similar reduction in HbA1c. Moreover, a decrease in the proinsulin:insulin ratio seen after 6 months of rosiglitazone therapy in patients with type 2 diabetes indicates an improvement in insulin granule maturation. Furthermore, a key component of thiazolidinedione action is redistribution of fatty acid fluxes which may alleviate beta-cell lipotoxicity. In the American Diabetes Prevention Program in patients with IGT, evaluation of the 1-year troglitazone arm demonstrated an even greater effect on reduction of diabetes risk than metformin or lifestyle intervention, a finding well compatible with a beta-cell protective effect. Independent evidence for a beta-cell protective effect of rosiglitazone comes from studies in renal transplant patients treated with beta-cell-toxic immunosuppressants.

In 18 out of 22 patients, adequate glycaemic control was achieved with rosiglitazone suggesting a dramatic amelioration of beta-cell toxicity. In conclusion, both GLP-1 based therapies and thiazolidinediones have potential for chronic beta-cell protection but clinical data in support of this concept are only available for thiazolidinediones. Both direct and indirect mechanisms are likely to be involved.