Dipeptidyl peptidase-4 (DPP-4) inhibitors, glucagon-like peptide-1 receptor agonists (GLP-1 RAs), and sodium-glucose cotransporter 2 (SGLT-2) inhibitors are newer glucose-lowering medication classes that are commonly used for the management of type 2 diabetes. These medications are second or third-line agents, meaning they are prescribed as additions to metformin, the standard first-line treatment. The choice of such medications in routine clinical care is not guided by research due to the lack of randomized trials directly comparing second-line glucose-lowering medication classes. While real-world evidence based on electronic healthcare databases helps fill in this gap, few studies have addressed the potential for methodological issues when comparing the newer medication classes relative to older ones such as sulfonylureas (SUs). Since SUs are a cheaper medication class, the characteristics of patients using SUs would be systematically different from those using newer classes, especially in the US, where drug costs were covered by the insurance plans and not paid for by the singer-payer system as in the UK. Thus, we will compare the cardiovascular effectiveness of newer glucose-lowering medication classes vs. SUs, among patients with type 2 diabetes on metformin. We will use data from the US and the UK to compare the methodological issues, particularly with respect to unmeasured confounding (i.e., variables that are unaccounted for) due to socioeconomic status and duration of diabetes, among other patient characteristics in estimating drug effectiveness.
While newer drug classes such as dipeptidyl peptidase-4 (DPP-4) inhibitors, glucagon-like peptide-1 receptor agonists (GLP-1 RAs), and sodium-glucose cotransporter 2 (SGLT-2) inhibitors, are recommended for their relatively safe adverse event profile or their cardiovascular effectiveness, the selection of such medication classes in routine care is not guided by empirical evidence due to a paucity of trial evidence directly comparing second-line glucose-lowering medication classes. Real-world studies using healthcare databases, while commonly used, are susceptible to unmeasured confounding. To date, few studies have evaluated the potential for unmeasured confounding, particularly with respect to socioeconomic status, when comparing newer vs older (and cheaper) medication classes, such as sulfonylureas (SUs).
Thus, we will conduct a multi-site comparative effectiveness study using electronic healthcare databases in both the United Kingdom (UK) and the United States (US) CPRD GOLD and Aurum with linkages to HES Admitted Patient Care (hospitalization records) and ONS Death Registration Data, along with two databases from the US: Optum Clinformatics Data Mart and Medicare fee-for-service. We will construct cohorts consisting of new users of SGLT-2 inhibitors, GLP-1 RAs, DPP-4 inhibitors, and SUs (the comparator), within each data source, with follow-up from date of drug initiation until death, end of follow-up, or an incident cardiovascular outcome event. Comparative cardiovascular effectiveness of each drug class vs. SUs will be estimated using hazard ratios (HR) and rate differences (RD), after propensity score (PS) matching (and disease risk score matching in the US databases) to control for measured confounding by a range of characteristics. The potential for confounding due to socioeconomic status and duration of diabetes will be compared between US and UK databases. Specifically, using the UK CPRD database, we will compare the PS matched with and without adjusting for socioeconomic measures (patient-level index of multiple deprivation), and/or duration of diabetes.
The outcomes of interest in this study are the major adverse cardiovascular events (MACE, i.e., hospitalization for myocardial infarction, hospitalization for ischemic stroke, and cardiovascular-specific mortality), modified MACE (hospitalization for myocardial infarction, hospitalization for ischemic stroke, and all-cause mortality), and hospitalization for heart failure (HHF). We will also evaluate individual components of the MACE and all-cause mortality as secondary outcomes. The diagnoses within MACE, the primary outcome, are recorded in the HES APC database and will be defined using ICD-10 codes (myocardial infarction [ICD 10: I21.x] and ischemic stroke [ICD 10: I63.x and I64.x] in primary position). Cardiovascular mortality is recorded in the ONS database and will be defined using ICD-10 codes [I0-99] where a cardiovascular event was recorded as the underlying cause of death. Hospitalization for heart failure is also recorded in HES APC and will be defined using ICD-10 code I50.x (Appendix 1).
Samy Suissa - Chief Investigator - Sir Mortimer B Davis Jewish General Hospital
Laurent Azoulay - Corresponding Applicant - McGill University
Elisabetta Patorno - Collaborator - Brigham & Women's Hospital
Hui Yin - Collaborator - Sir Mortimer B Davis Jewish General Hospital
PHYO HTOO - Collaborator - Brigham & Women's Hospital
HES Admitted Patient Care;ONS Death Registration Data;Patient Level Index of Multiple Deprivation