Kidney function loss and albuminuria progression with GLP-1 receptor agonists versus basal insulin in patients with type 2 diabetes: real-world evidence

Fig. 1. The association between initiation of GLP-1 RAs versus basal insulin and the risk of the composite kidney outcome. The risk of the composite kidney outcome (confirmed ≥ 40% eGFR decline or new end-stage kidney disease) was compared between initiators of GLP-1 RAs versus basal insulin. Presented are cumulating incidence functions. In an intention-to-treat (ITT) analysis, patients were followed until October 2021, the end of data availability, or death (A). In an as-treated (AT) analysis, follow-up was also censored at study drug discontinuation or comparator-initiation (B). In an ITT-48 months (ITT-48mo) analysis, to emulate the LEADER study, the ITT follow-up was also censored at four years when a large proportion of the participants were still on the study drugs (C). Cox proportional hazards regression models were applied to compare between treatment arms. GLP-1 RA Glucagon-like peptide-1 receptor agonists, eGFR estimated glomerular filtration rate

Fig. 2. The association between initiation of GLP-1 RAs versus basal insulin and the risk of the composite kidney outcome by baseline subgroups. The risk of the composite kidney outcome (confirmed ≥ 40% eGFR decline or new end-stage kidney disease) was compared between initiators of GLP-1 RA and initiators of basal insulin across subgroups. In the intention-to-treat (ITT) analysis, patients were followed until October 2021, end of data availability or death. In the as-treated (AT) analysis, follow-up was also censored at study drug discontinuation or comparator-initiation. Cox proportional hazards regression models were applied to compare treatment arms, with an interaction term between subgroups and the treatment arm. Event rates (ER) are per 100 patients-years. BDL below detectable levels, BMI body mass index, CV cardiovascular, GLP-1 RA Glucagon-like peptide-1 receptor agonists, HbA1c glycated hemoglobin A1c, eGFR estimated glomerular filtration rate, RAAS Renin angiotensin aldosterone system, SGLT2 sodium-glucose co-transporter 2 inhibitors, UACR urine albumin-to-creatinine ratio

Fig. 3. The association between initiation of GLP-1 RAs versus basal insulin and the risk of categorical eGFR decline or albuminuria progression in the ITT and AT analyses. The risks of GFR decline (by different thresholds) or albuminuria progression were compared between initiators of GLP-1 RAs and initiators of basal insulin. In the intention-to-treat (ITT) analysis, patients were followed until October 2021, end of data availability or death. In an as-treated (AT) analysis, follow-up was also censored at study-drug discontinuation or comparator-initiation. A categorical increase of UACR was defined for the following categories: < 30, 30- < 300 or ≥ 300 mg/g. New-onset macroalbuminuria was defined as UACR ≥ 300 among those with UACR ≤ 230 mg/g at baseline (resembling a ≥ 30% increase in UACR). Outcomes were assessed as single- or confirmed measurement. Cox proportional hazards regression models were applied to compare between treatment arms. Event rates are per 100 patients-years. c - confirmed measurement, GLP-1 RAs Glucagon-like peptide-1 receptor agonists, eGFR estimated glomerular filtration rate, ER event rate, ESKD End-stage kidney disease, s -single measurement, UACR Urine…

Fig. 4. The change in eGFR overtime in initiators of GLP-1 RA versus basal insulin. At each time point (every 6 months for the first 3 years and then yearly), we considered the eGFR measurement closest to the end of the period for each patient. Mixed models with repeated measurements were applied to measure the change from baseline in eGFR at each time point and to calculate the p-value for the difference between groups (presented near each point). GLP-1 RAs Glucagon-like peptide-1 receptor agonists, eGFR estimated glomerular filtration rate

Fig. 5. The change in eGFR overtime in initiators of GLP-1 RAs versus basal insulin in the ITT and AT analyses, by baseline subgroups. We calculated an eGFR slope per patient by fitting a linear regression model. We then calculated the mean eGFR slope over time for each group and used t-test to compare the treatment groups. eGFR slopes are presented as mL/min/1.73 m2/year. BDL below detectable levels, BMI body mass index, CVD cardiovascular disease, eGFR estimated glomerular filtration rate, GLP-1 RAs glucagon-like peptide 1 receptor agonist, HbA1c glycated hemoglobin A1c, RAAS renin angiotensin aldosterone system, SGLT2i sodium-glucose transporter 2 inhibitors, UACR urine albumin-to-creatinine ratio