Lead Investigator: David Cherney, University Health Network
Title of Proposal Research: Modelling Studies to Assess Sodium-Glucose co-Transporter-2 (SGLT2) Inhibition Effects and Kidney Protection in Type 1 Diabetes
Vivli Data Request: 8582
Funding Source: Canadian Institutes of Health Research (CIHR) Team Grant
Potential Conflicts of Interest: None
Summary of the Proposed Research:
Diabetes is a growing public health issue that is estimated to affect over 463 million adults worldwide. Patients with diabetes are characterized by chronically elevated blood sugar (glucose) and are categorized into three main subtypes – type 1 diabetes (T1D), type 2 diabetes (T2D), and gestational diabetes. T1D occurs when the pancreas is unable to produce insulin whereas T2D occurs when the body does not respond to insulin, gestational diabetes is high blood glucose that develops during pregnancy and usually disappears after giving birth. Type 1 diabetes (T1D) is associated with an increased risk in mortality compared to the general population, with premature cardiovascular disease (CVD) and chronic kidney disease (CKD) being leading causes of reduced life expectancy. Although less common compared to type 2 diabetes (T2D), T1D afflicts approximately 9 million people worldwide. Given that the coexistence of CVD and kidney disease is more prevalent within the T1D population, it is imperative to proactively develop tools to prevent diabetes-associated complications.
This project aims to provide new insights into how sodium-glucose cotransporter 2 (SGLT2) inhibitors reduce cardiorenal risk as well as societal costs due to type 1 diabetes (T1D). SGLT2 inhibitors are a class of prescription medicines that are for use with diet and exercise to lower blood sugar in adults with type 2 diabetes. They work by causing the kidneys to remove sugar from the body through urine.
We will use data analysis tools to evaluate potential its effects on markers of long-term diabetic kidney disease (DKD) outcomes and physiological factors associated with kidney health and disease in T1D. The overarching aim of this project is to determine if the mechanisms associated with kidney-heart protection in T2D are also present in people with T1D. We will use previous insights from people with T2D and non-diabetic chronic kidney disease (CKD) to assess long-term clinical benefits on kidney and cardiovascular risk with SGLT2 inhibition in T1D.
Regardless of the magnitude of anticipated benefits, these analyses will be informative for our ultimate goal of designing a larger clinically meaningful trial in collaboration with the Preventing Early Renal Function Loss (PERL) consortium to assess kidney function loss. The PERL consortium is a group of high-quality academic centres and investigators focused on preventing kidney dysfunction. Results of the study will support ongoing efforts to design a clinical trial examining the use of SGLT2 inhibitors in patients with T1D. This may potentially extend the use of these agents with associated heart and kidney protective benefits to a patient population with limited therapeutic options. The ultimate goal of this project is to provide better personalized care to patients with T1D by understanding the mechanisms that are associated with protection against diabetic complications. This will add to our understanding on SGLT2 inhibitors and may provide useful for use in other patient groups as well as the T1D population.
Statistical Analysis Plan:
We will be doing two identical but separate analyses on the DEPICT and EASE clinical trial cohorts. We will first assess the longitudinal effects of SGLT2 inhibitor (dapagliflozin or empagliflozin) versus placebo on urinary albumin-creatinine ratio (UACR), kidney function loss, clinical efficacy measures (blood pressure [BP], metabolic control) and safety. In Aim 1 we will use parameter response efficacy or “PRE” analysis techniques to identify clinical factors that impact longer-term sodium glucose cotransporter-2 (SGLT2) inhibitor effects in type 1 diabetes (T1D), including kidney function loss, need for dialysis, and cardiovascular disease (CVD). We will apply data analytical techniques including regression analysis to process trial data to understand how clinical determinants of cardiorenal outcomes influence DKD and CVD in response to SGLT2 inhibitor use in T1D. In Aim 1, we hypothesize that by using risk markers identified in people with type 2 diabetes (T2D), we can develop a similar PRE model to estimate the real-world long-term impact of SGLT2 inhibition on cardiorenal risk in T1D, thereby guiding personalized T1D care. Risk markers of interest include hemoglobin A1C (HbA1C), body weight, BP, UACR, low density lipoprotein cholesterol (LDL-C), high density lipoprotein cholesterol (HDL-C), uric acid, hemoglobin, potassium, albumin, calcium phosphate, and biological sex. The economic evaluation will be conducted using a Markov health state cost-utility model to assess the incremental direct and indirect medical costs to produce incremental benefits in quality-adjusted life year (cost/“QALY”), based on completed Patient Reported Outcome Measures (PROM) questionnaires for T1D patients treated with SGLT2 inhibition in the DEPICT and EASE trials, as well as incremental cost-effectiveness ratio (ICER) to compare these different treatment options. This tool has been used to assess diabetes-related complications and treatments. The economic evaluation will be reported using the consolidated health economic evaluation reporting standards.
In Aim 2, to be a “mediator” of cardiorenal outcomes, a variable must change significantly in response to SGLT2 inhibition, and the variable change must have an effect on the outcome of interest. In our previous work, we used Cox regression models to evaluate covariates (i.e. “mediators”) that were impacted by SGLT2 inhibition. Mediators are identified as covariates that, when added to an unadjusted model of treatment assignment, yield an attenuated hazard ratio, reflecting a statistical reduction in the clinical benefit when the mediators are accounted for. The percentage of mediation is determined by the proportional increase in the hazard ratio between the unadjusted and adjusted models, calculated as the weighted average of change from baseline from all post-baseline measurements. Each potential mediator will be tested individually, so across analyses, the % mediation attributable to a specific factor may sum to >100%. Potential mediators include estimated glomerular filtration rate (eGFR) dip, UACR reduction, hematocrit, hemoglobin, red blood cell count, uric acid, HbA1c, body weight, albumin, systolic BP, and HDL-C.
In Aim 3, we will use models that simulate kidney function in a heterogeneous population of patients with T1D. Analyses will be developed that simulate male and female patients with T1D, with or without hypertension, taking different medications (e.g. renin-angiotensin-aldosterone system [RAAS] inhibitors) in addition to SGLT2 inhibitors. We will conduct analyses to assess the effects of SGLT2 inhibition in patients with T1D at different stages of chronic kidney disease (CKD) (eGFR 60-89 ml/min/1.73m2, eGFR ≥90 ml/min/1.73m2) characterized by different degrees of nephron loss. Model parameters will be fitted using DEPICT data and EASE data, including eGFR, BP, UACR, hematocrit, serum albumin and electrolytes.
In a second step, we will expand the model patient population by varying “hidden” physiological parameters not observable in existing data (e.g. sensitivity of tubuloglomerular feedback response, degree of nephron hypertrophy, reduction in metabolic efficiency due to diabetes-induced mitochondrial dysfunction). With this virtual patient cohort, we will simulate SGLT2 inhibition and assess kidney protective effects. Simulated patients will be classified as a “responder” in terms of SGLT2 inhibitor-induced protection if a minimum, clinically relevant >20% reduction in UACR is predicted, and as a “non-responder” otherwise. With this “responder” definition, we will seek to identify the mechanisms that explain different kidney responses. We anticipate that GFR and metabolic processes hold the key to kidney protection. Specifically, we will analyze model outputs to assess the extent to which non-responders are those who fail to respond to SGLT2 inhibition with sufficient intraglomerular pressure reduction and/or improvements in renal oxygenation.
We are selecting the DEPICT1 and 2 trials and the EASE 2 and 3 because they are T1D cohorts that we want to analyze to assess the impact of SGLT2 inhibition. We will assign each cohort a study code to maintain independence of the two trials. Missing data will be considered missing completely at random and we anticipate using a multiple imputation method to handle missing values.
Requested Studies:
Dapagliflozin Evaluation in Patients With Inadequately Controlled Type 1 Diabetes (DEPICT 2)
Data Contributor: AstraZeneca
Study ID: NCT02460978
Dapagliflozin Evaluation in Patients With Inadequately Controlled Type 1 Diabetes (DEPICT 1)
Data Contributor: AstraZeneca
Study ID: NCT02268214
A Phase III, Randomised, Double Blind, Placebo-controlled, Parallel Group, Efficacy, Safety and Tolerability Trial of Once Daily, Oral Doses of Empagliflozin as Adjunctive to inSulin thErapy Over 52 Weeks in Patients With Type 1 Diabetes Mellitus (EASE-2)
Data Contributor: Boehringer Ingelheim
Study ID: NCT02414958
Sponsor ID: 1245.69
A Phase III, Randomised, Double Blind, Placebo-controlled, Parallel Group, Efficacy, Safety and Tolerability Trial of Once Daily, Oral Doses of Empagliflozin as Adjunctive to Insulin Therapy Over 26 Weeks in Patients With Type 1 Diabetes Mellitus (EASE-3)
Data Contributor: Boehringer Ingelheim
Study ID: NCT02580591
Sponsor ID: 1245.72
Public Disclosures:
- Nardone, M., Kugathasan, L., Sridhar, V., Dutta, P., Campbell, D., Layton, A.T., Perkins, B.A., Barbour, S., Lam, T.K., Levin, A. and Lovblom, L.E., 2024. Modeling Cardiorenal Protection with SGLT2 Inhibition in Type 1 Diabetes: An Analysis of DEPICT-1 and-2: FR-PO312. Journal of the American Society of Nephrology, 35(10S), pp.10-1681. Doi : 10.1681/ASN.2024s5r4rxwp
- Kugathasan, L., Dutta, P., Nardone, M., Sridhar, V., Campbell, D., Layton, A.T., Perkins, B.A., Barbour, S., Lam, T.K., Levin, A. and Lovblom, L.E., 2024. Modeling Cardiorenal Protection with SGLT2 Inhibition in Type 1 Diabetes: An Analysis of EASE-2 and-3: TH-OR53. Journal of the American Society of Nephrology, 35(10S), pp.10-1681. Doi : 10.1681/ASN.2024w22c75k8