Response to the comment on: "Dulaglutide treatment results in effective glycaemic control in latent autoimmune diabetes in adults (LADA): A post-hoc analysis of the AWARD-2, -4 and -5 trials"

To the Editor The paper by Jones et al. in which they studied the response of a cohort of 19 patients with adult-onset diabetes who were glutamic acid decarboxylase autoantibodies (GADA) positive, most with very low levels of C peptide and multiple autoantibodies, is consistent with published results studying GLP-1 agonists in patients with established type 1 diabetes who show little by way of a beneficial response.1,2 Importantly, they note that those with GADA, who were not on insulin therapy from diagnosis, had a similar response to a GLP-1 agonist to those who were GADA negative and by implication had type 2 diabetes. This point is important as the implication is that patients with GADA, the usual diabetes-associated autoantibodies screening test in adults, could benefit from GLP-1 agonists as long as they have limited insulin deficiency, that is, as long as they are not on multiple insulin injections. By inference, GLP-1 agonists are a reasonable option in the management of adult-onset autoimmune diabetes patients who are not on multiple insulin therapy. Jones and McDonald have misread the assay specificity of our GADA assay; their assay specificity is 97.5% and ours, based on the international workshop (DASP), is 99%; therefore, our assay is more specific than theirs.3 In addition, the positive predictive value of an assay increases as the population under test is enriched, as with the adult-onset diabetes cases we studied.4 We found a rate of GADA positive cases in our cohort (7.6%, 188 GADA+ out of 2466 patients) comparable to the rate they found with their assay, despite theirs having only a slightly lower assay specificity: “Eight percent of insulintreated participants” compared with a lower rate and longer disease duration at “0.9% of non–insulin-treated participants.”1 Finally, the high titre cut-off is arbitrary in this cohort in order to make it comparable with other studies in which it was not arbitrary. Those previous studies used inflections in the Quartile-Quartile plot to define a second population with a high GADA titre different from those with a lower GADA titre. We accept that our study has limitations, despite it comprising tenfold the number of cases than their analysis. In addition, this is the first study of a once-weekly as compared to a once-daily GLP-1 agonist regime, which could be relevant to differences in responses. Large case-controlled cases are awaited, but the data we presented on our large cohort is compelling, while the combined results from their study and the study by Jones et al. indicate that GLP-1 agonists are likely to have a limited role in patients with severe insulin deficiency and frank clinical type 1 diabetes.