1042-P: Long-Term Complications after Add-on Treatment to Metformin—UK Real-World Progression in Diabetes (RAPIDS) Model

Introduction & Objective: Policy makers must take a long view when deciding which treatments to reimburse. Clinical trials may not provide data for long-term complications, nor capture treatment duration or sequences observed in practice. In this study, we predict probabilities of complications for people initiating add-on treatment to metformin monotherapy with sulfonylureas (SU) vs. dipeptidyl peptidase-4 inhibitors (DPP4i) vs. sodium-glucose co-transporter-2 inhibitors (SGLT2i).

Methods: We used data from the UK Clinical Practice Research Datalink (2011 to 2020) to find people with type 2 diabetes prescribed 2nd-line treatment with a SGLTi, DPP4i or SU. We used the US-based RAPIDS microsimulation model updated to reflect trial evidence and calibrated to the UK. We predicted counterfactual outcomes for the alternative second-line treatments for ‘real-world’ durations and subsequent treatment switching. From these predicted outcomes, we estimated between-treatment differences in the mean probabilities of MI, stroke, diabetic eye disease (DED), eGFR, end-stage renal disease (ESRD) or amputation (LEA) over 5 years.

Results: Starting 2nd-line treatment with SGLTi compared to either SU or DPP4i reduced the predicted probability of long-term complications over 5 years. The probability of MI or stroke were similar across the 2nd-line treatments. For SGLTi, the model predicted an increase in the mean eGFR (mL/min/1.73m²) of 0.4 (95% CI 0.1, 0.7) vs SU, and of 0.6 (0.5, 0.7) vs DPP4i. Following starting SGLTi, the model predicted a reduced probability of ESRD of -0.0025 (-0.002,-0.003) vs SU and -0.0025 (-0.03,-0.02) vs DPP4i; for DED the corresponding reductions were -0.011 (-0.012,-0.01) vs. SU and -0.003 (-0.004, -0.002) vs DPP4i. SGLTi and DPP4i lowered LEA risk compared with SU.

Conclusion: The RAPIDS-UK model, based on real-world treatment use, shows that 2nd-line therapy with SGLTi reduces the probability of complications up to 5 years compared to SU or DPP4i.