Patients who develop oedema on initiating thiazolidinedione therapy have an improved glycaemic response: a MASTERMIND study

Abstract

Background/aim: Oedema is a common and serious side effect ofthiazolidinedione therapy. A stratified medicine approach wouldaim to give thiazolidinediones to patients likely to have a goodglycaemic response but not to develop oedema. We investigatedwhether oedema was associated with glycaemic response tothiazolidinedione therapy.Methods: We studied 10,486 patients initiating a thiazolidinedionefrom Clinical Practice Research Datalink (CPRD), and identifiedmedical records of oedema in the subsequent 12 months. Responsewas defined as change in HbA1c at 12 months and was adjusted forbaseline HbA1c, baseline body mass index, gender and adherence(medication possession ratio). In secondary analyses we restrictedoedema classification to patients with concomitant weight gain. As acomparison the same analysis was performed in 13,089 patientsinitiating a sulfonylurea.Results: The 3% of patients with recorded oedema onthiazolidinediones had a mean (confidence interval) 3 (1.7–4.3)mmol/mol greater fall in HbA1c (p < 0.001) compared to thosewithout oedema. This improved response increased when oedemawas associated with weight gain, with a 4.1 (1.9–6.2)mmol/molgreater HbA1c fall for weight gain >3kg (p < 0.001) and a 5.2 (2.1–8.4)mmol/mol greater fall for weight gain >8kg (p < 0.001). Insulfonylurea patients oedema was not associated with response(HbA1c fall difference 1 (0.5 to 2.5)mmol/mol, p=0.2), evenwhen associated with weight gain >3kg (p=0.19).Conclusion: Patients with Type 2 diabetes who develop oedemaon initiating thiazolidinediones have an improved glycaemicresponse, and more severe oedema is associated with greaterHbA1c reduction. This supports glycaemic lowering andfluid retention being mediated by a common pathway ofthiazolidinedione drug action.