This research consisted of two phases. In the first, researchers used data from the UK Clinical Practice Research Datalink (CPRD) to examine factors associated with one-year glycaemic response in patients treated with a sulfonylurea (SU) or thiazolidinedione (TZD). Sex and BMI showed the greatest differential response to therapy, with greater response to TZD (as compared to SU) in women and in those with higher BMI, and vice versa. Based on this, four subgroups with respect to sex and BMI (>30 versus ≤30) were defined. Subsequent analysis found that non-obese males had a greater 1-year response with SU than with TZD (baseline adjusted change in HbA1c: -13.2 vs. -9.7 mmol/mol, P<0.001), whereas obese females had a greater 1-year response with TZD than SU (-13.8 vs. -9.4 mmol/mol, P<0.001).
In the second phase, data from randomised trials comparing SU to TZD (ADOPT; RECORD) was used to determine whether 5-year glycaemic responses differed between these four patient subgroups. Non-obese males had a greater overall HbA1c reduction with SU than with TZD (P < 0.001); in contrast, obese females had a greater HbA1c reduction with TZD than with SU (P<0.001).
In terms of adverse effects, there was an increased risk of moderate/ severe hypoglycaemia with SU versus TZDs in all groups. For TZD, weight was increased versus SU for all subgroups, although this was more marked in obese women. Fracture was more common with TZD compared with SU but only for females.
The authors say their results provide the first example of stratification of therapy in type 2 diabetes based on simple patient characteristics, and that this will allow for much more informed discussion of the benefits and risks of these therapies than the present “one size fits all” approach. They do however acknowledge the study’s limitations, including the fact that the RCT data was for rosiglitazone (trial data were not made available to allow them to repeat their analysis for pioglitazone), and that their findings do not allow prediction at an individual level. They also comment that the ideal stratified approach would be based on cardiovascular endpoints rather than the intermediary measure of glycaemic response.