The prevalence of cardiovascular disease and antidiabetes treatment characteristics among a large type 2 diabetes population in the United States

Abstract Objectives The purpose of this study was to assess atherosclerotic cardiovascular disease (ASCVD) prevalence, antidiabetes medication usage and physician specialty encounters among individuals with type 2 diabetes mellitus (T2DM) in the United States during 2015. Design Retrospective, cross‐sectional analysis. Patients Adults with T2DM in a large US administrative claims database. Patients were divided into ASCVD and non‐ASCVD groups. Subgroup analyses were conducted for three age groups (18‐44, 45‐64 and 65+ years). Results Of 1 202 596 patients with T2DM, 45.2% had established ASCVD. About 40% of T2DM patients with ASCVD had visited a cardiologist during 2015, compared to 11% in the non‐ASCVD group. The use of glucagon‐like peptide‐1 receptor agonists (GLP‐1RAs) and sodium‐glucose co‐transporter 2 inhibitors (SGLT‐2is) was low overall (<12%), and even lower in the ASCVD group (<9%). The prevalence of ASCVD was 15%, 36% and 71% in the 18‐44, 45‐64 and 65+ year age groups, respectively. GLP‐1RA and SGLT‐2i use was ≤5% in the 65+ subgroup, regardless of ASCVD status. Conclusions These real‐world data showed a high prevalence of ASCVD among T2DM patients, and confirmed, as a baseline assessment, low use of GLP‐1RAs and SGLT‐2is in these at‐risk patients prior to the 2017 American Diabetes Association guidelines recommending use of agents with proven cardiovascular benefits.

In recent years, large cardiovascular outcomes trials (CVOTs) have demonstrated CV benefits with glucagon-like peptide-1 receptor agonists (GLP-1RA) and sodium-glucose co-transporter-2 inhibitor (SGLT2is). The first such studies to be published were the EMPA-REG-OUTCOME trial in 2015 8 and the Liraglutide Effect and Action in Diabetes: Evaluation of Cardiovascular Outcome Results (LEADER) trial in 2016. 9 In the EMPA-REG-OUTCOME study, patients with established, stable CV disease treated with the SGLT2i empagliflozin had a lower rate of the primary composite outcome (death from cardiovascular causes, nonfatal myocardial infarction or nonfatal stroke) than patients receiving placebo, as well as significantly lower rates of death from cardiovascular causes, hospitalization for heart failure and death from any cause. 8 In the LEADER trial, patients with T2DM and concomitant CV disease or at high CV risk treated with the GLP-1RA liraglutide had a lower rate of the primary composite outcome (first occurrence of cardiovascular death, nonfatal myocardial infarction or nonfatal stroke in the time-to-event analysis), and lower risks of death from cardiovascular causes, death from any cause and microvascular events than did those receiving placebo. 9 Based on the results of these two studies, the American Diabetes The impact of the updated diabetes guidelines regarding GLP-1RA and SGLT2i use in at-risk patients in real-world clinical practice will be of interest in the coming years. A recent study used the electronic health record system at Cleveland Clinic (Ohio and Florida) to create a crosssectional summary of patients with T2DM and CVD in 2016 (ie the year prior to release of the 2017 ADA guidelines) to establish a baseline of real-world treatment patterns in these patients. 12 Utilization rates of GLP-1RA and SGLT2i agents were found to be low (<10%) in patients with T2DM, whether with or without established CVD.
The current study used real-world claims data to determine the prevalence of ASCVD among patients with T2DM and to assess antidiabetes medication usage and healthcare specialty utilization in these high-risk patients prior to availability of the 2017 ADA guidelines.

| Data source and study population
This was a retrospective, cross-sectional analysis of a large, nationwide Eligible individuals were aged ≥18 years on 1 January 2015 and had an established diagnosis of T2DM before 1 January 2015, defined as ≥2 diagnoses for T2DM, based on international classification of diseases, ninth revision (ICD-9) codes of 250.x0 or 250.

| S TUDY COHORTS-PATIENTS WITH AND WITHOUT E S TAB LIS HED A SC VD
Eligible patients with T2DM were divided into two groups based on the presence (ASCVD group) or absence (non-ASCVD group) of established ASCVD. To be included in the ASCVD group, ASCVD must have been present prior to 1 January 2015 and was defined based on the ICD-9/-10 codes (Table S1)

| VARIAB LE S OF INTERE S T
Patient demographics were determined as of 1 January 2015 and included age, sex, geographic region and insurance type. Comorbidities were identified from claims from 2014 and 2015 and were used to determine the Diabetes Complications Severity Index (DCSI) score, 13 and Charlson Comorbidity Index (CCI) score. 14 Individual comorbidities of hypertension and dyslipidemia were also recorded, as these were not captured in the definition of ASCVD nor were they components of DCSI or CCI scoring.
Study endpoints during 2015 included claims for any anti-diabetes medications, use of GLP-1RA and SGLT2i agents specifically and visits with endocrine or cardiovascular specialists.

| Data analysis
This was a descriptive analysis. Population characteristics were measured using counts with percentages for the categorical variables and means with standard deviation (SD) for continuous TA B L E 1 Demographic characteristics of a real-world 2015 population with type 2 diabetes (N = 1 202 596), stratified by atherosclerotic cardiovascular disease (ASCVD) status variables. Subgroup analyses were conducted for three age categories (18-44, 45-64 and ≥65 years).

| Antidiabetes drug use patterns
In the total population with T2DM, the majority of patients had claims for OADs only, regardless of ASCVD status Table 2. Among OADonly users, most patients were using 1 (56%-57%) or 2 (30%) OADs.
Overall, the use of GLP-1RAs and SGLT2is was low (<12% of patients) and slightly lower in the ASCVD group compared to the non-ASCVD group. Liraglutide and canagliflozin were the most prevalent GLP-1RA and SGLT2i agents, respectively. Insulin use was more prevalent in the ASCVD cohort vs the non-ASCVD cohort (18.5% vs 13.9%).

| Healthcare specialist visits
A low and similar proportion of patients visited an endocrinologist during 2015, regardless of ASCVD status (8.0% of those non-ASCVD; 8.7% of those with ASCVD). In the ASCVD group, 40% had visited a cardiologist during 2015, compared to 11% in the non-ASCVD group.

| Subgroup analysis by age category
The prevalence of ASCVD increased with increasing age category Figure 1. In each age category, all diabetes-related complications included in the analysis were present at a higher prevalence among the ASCVD cohorts as compared to the non-ASCVD cohorts Table 3.
The proportion of patients that used GLP-1RA or SGLT2i agents was 5% or lower among the ≥65 age subgroup, regardless of ASCVD F I G U R E 1 Prevalence of atherosclerotic cardiovascular disease (ASCVD) among 1 202 596 patients with T2DM within age subgroups TA B L E 3 Prevalence of diabetes-related complications by atherosclerotic cardiovascular disease (ASCVD) status and age category in a real-world 2015 population with type 2 diabetes a Category includes any cardiovascular complication, not limited to those used to define "ASCVD" (acute coronary syndrome, history of myocardial infarction, angina pectoris, peripheral arterial disease presumed to be of atherosclerotic origin, transient ischaemic attack and coronary or other arterial revascularization). b Category includes any peripheral vascular disease, not limited to "peripheral arterial disease presumed to be of atherosclerotic origin" which was part of the "ASCVD" definition. c Category includes ketoacidosis, hyperosmolar and "other coma."

Non-ASCVD ASCVD
is being provided by primary care physicians. Among patients with T2DM and ASCVD, 40% had seen a cardiologist during 2015. A limitation of the current analysis is the reliance on ICD-9/-10 codes alone to document ASCVD and comorbidities, since these codes may be impacted by provider coding practices and subject to coding error. Also, substantial differences in insurance coverage patterns observed between the groups could potentially impact drug and healthcare resource utilization patterns: the non-ASCVD group had a much higher proportion of commercially insured patients, whereas approximately half the ASCVD group were covered by Medicare. Another potential limitation is that patients may have had visits to endocrinologists and/or cardiologists closely adjacent to the 1-year study period window, and therefore not been captured in the assessment of utilization. Further, it was not possible to determine the type of provider who prescribed the antidiabetes therapies. In addition, these data did not include uninsured patients and thus may not be entirely generalizable. Nonetheless, the large number of patients (more than one-half-million in each cohort) that were included, and the nationwide sampling, allow for a certain degree of generalizability of these findings.
Although these cross-sectional data do not capture nuances of clinical practice, they do serve as a useful tool for capturing the prevalence of medication use and overall care patterns in a large cohort of patients with T2DM and ASCVD. Tian was a contracted employee of Novo Nordisk Inc., the funding body for this study, during study conduct. S. X. Kong was an employee of Novo Nordisk, Inc., during study conduct.

AUTH O R S' CO NTR I B UTI O N S
W. Weng, Y. Tian, S. X. Kong, R. Ganguly and T. Hobbs made substantial contributors to conception and design, or acquisition of data, or analysis and interpretation of data. All others were involved in drafting the manuscript or revising it critically for important intellection content, gave final approval of the version to be published and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

M E D I C A L WR ITI N G A SS I S TA N CE
Writing assistance was provided by Kulvinder Katie Singh, PharmD, through Churchill Communications (Maplewood, NJ) and funded by Novo Nordisk, Inc.

E TH I C S S TATEM ENT
This was a retrospective, noninterventional study using data from the IBM MarketScan database which contains de-identified patient claims data. The IBM MarketScan database is fully compliant with the Health Insurance Portability and Accountability Act of 1996.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from IBM/Truven. Restrictions apply to the availability of these data, which were used under licence for this study. Data are available Wayne Weng with the permission of IBM/Truven.