Intensive monitoring of adverse drug events associated with the use of new glucose-lowering drugs: results from an inception cohort study in Portugal

Aims To determine the frequency and the time-course profile of adverse drug events associated with new glucose-lowering drugs in daily practice and to explore factors potentially associated to these events. Methods An inception cohort study was implemented. Adults with type 2 diabetes mellitus initiating a dipeptidyl peptidase-4 inhibitor, a glucagon-like peptide-1 receptor agonist or a sodium-glucose co-transporter-2 inhibitor were eligible for inclusion. Data were collected through baseline and follow-up telephone questionnaires, administered at 2 weeks, 3 months and 6 months. Kaplan-Meier curves and log-rank were computed to compare the time to adverse drug event onset. Cox models were used to explore potential factors associated with adverse drug events. Results A total of 1328 participants were recruited to the study. In all, 1118 adverse drug events were reported (of which 36% were not listed in the summary of product characteristics) by 41% of participants. The median latency time of adverse drug events reported in >= 1% of participants ranged from 0 to 2 days. Glucagon-like peptide-1 receptor agonist and sodium-glucose co-transporter-2 inhibitor subgroups were associated with an increased likelihood of adverse drug event reporting when compared with the dipeptidyl peptidase-4 inhibitor subgroup. A total of 328 glucose-lowering drugs were withdrawn, more than half as a result of an adverse drug event. Conclusions More than two-fifths of participants reported an adverse drug event; dipeptidyl peptidase-4 inhibitors led to the highest proportion of unlabelled adverse drug events. Adverse drug event latency time data show that counselling and adverse drug event management should be proactively addressed from treatment initiation. There should be greater focus on prevalent new users of glucose-lowering drugs, who were more complex participants in this study in terms of type 2 diabetes disease, as they were more likely to report an adverse drug event than the incident new users.