What to do with diabetes therapies when HbA1c lowering is inadequate: add, switch, or continue? A MASTERMIND study

BACKGROUND: It is unclear what to do when people with type 2 diabetes have had no or a limited glycemic response to a recently introduced medication. Intra-individual HbA1c variability can obscure true response. Some guidelines suggest stopping apparently ineffective therapy, but no studies have addressed this issue. METHODS: In a retrospective cohort analysis using the UK Clinical Practice Research Datalink (CPRD), we assessed the outcome of 55,530 patients with type 2 diabetes starting their second or third non-insulin glucose-lowering medication, with a baseline HbA1c > 58 mmol/mol (7.5%). For those with no HbA1c improvement or a limited response at 6 months (HbA1c fall < 5.5 mmol/mol [0.5%]), we compared HbA1c 12 months later in those who continued their treatment unchanged, switched to new treatment, or added new treatment. RESULTS: An increase or a limited reduction in HbA1c was common, occurring in 21.9% (12,168/55,230), who had a mean HbA1c increase of 2.5 mmol/mol (0.2%). After this limited response, continuing therapy was more frequent (n = 9308; 74%) than switching (n = 1177; 9%) or adding (n = 2163; 17%). Twelve months later, in those who switched medication, HbA1c fell (− 6.8 mmol/mol [− 0.6%], 95%CI − 7.7, − 6.0) only slightly more than those who continued unchanged (− 5.1 mmol/mol [− 0.5%], 95%CI − 5.5, − 4.8). Adding another new therapy was associated with a substantially better reduction (− 12.4 mmol/mol [− 1.1%], 95%CI − 13.1, − 11.7). Propensity score-matched subgroups demonstrated similar results. CONCLUSIONS: Where glucose-lowering therapy does not appear effective on initial HbA1c testing, changing agents does not improve glycemic control. The initial agent should be continued with another therapy added. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (10.1186/s12916-019-1307-8) contains supplementary material, which is available to authorized users.