FormalPara Key Summary Points

Cardiovascular outcomes trial (CVOT) results have led to changes in diabetes position statements and indications for some glucose-lowering agents

This study aimed to understand internal medicine (IM) and family medicine (FM) physicians’ knowledge of CVOTs and beliefs about type 2 diabetes mellitus (T2DM) medications for CV disease risk reduction

Overall, 39% of respondents were slightly/not at all familiar with the 2018 American Diabetes Association treatment recommendations for those with T2DM and atherosclerotic CV disease

Most respondents were not familiar with the findings of recent CVOTs

Findings from this survey suggest that there are knowledge gaps among IM and FM physicians regarding the results from CVOTs

These gaps have implications for the treatment of patients with T2DM and CV disease


Cardiovascular (CV) disease is the leading cause of mortality and morbidity for patients with type 2 diabetes mellitus (T2DM) [1, 2]. Internal medicine (IM) and family medicine (FM) physicians may appreciate the need to consider CV risk when treating patients with T2DM, but they may not be aware of the plethora of CV outcomes trials (CVOTs) that have been published over the last decade. In these CVOTs, the primary endpoint was most often a composite of time to first occurrence of CV death, non-fatal myocardial infarction, or non-fatal stroke, referred to as a 3-point major adverse CV event (3P-MACE) [3]; hospitalization for heart failure (HHF) was usually evaluated as a secondary outcome. To date, all completed CVOTs have confirmed the CV safety of the glucose-lowering agents tested and results from several CVOTs have shown that specific agents improve CV outcomes in patients with T2DM and CV disease [4, 5].

The first CVOT results (Table 1) published were for dipeptidyl peptidase 4 (DPP4) inhibitors (SAVOR-TIMI 53 [6], TECOS [7], and EXAMINE [8, 9]). These trials showed that compared with placebo, the addition of saxagliptin, sitagliptin, or alogliptin, respectively, to standard care diabetes therapy did not increase or decrease the event rate for the primary composite endpoint. More recently, CARMELINA [10] demonstrated the CV and renal safety of linagliptin compared with placebo in patients with T2DM and at high risk of CV and/or renal events. The first DPP4 inhibitor CVOT with an active comparator, CAROLINA, showed neither an increase nor a decrease in the risk of 3P-MACE for linagliptin compared with the sulfonylurea [11].

Table 1 Results of CVOTs for non-insulin glucose-lowering agents included in survey

Among the sodium-glucose co-transporter 2 (SGLT2) inhibitors, EMPA-REG OUTCOME was the first CVOT to demonstrate a reduced risk of 3P-MACE as well as reduced risk of CV mortality for empagliflozin vs. placebo added to standard care [12, 13]. The trial also showed that empagliflozin reduced the risk of HHF. This was followed by the CANVAS program, which also demonstrated a reduced risk of MACE compared with placebo [14]. More recently, the DECLARE-TIMI 58 study [15] showed that dapagliflozin reduced the risk of the combined endpoint of CV death or HHF, but showed non-inferiority for 3P-MACE. Results for the glucagon-like peptide 1 receptor agonist (GLP-1 RA) CVOTs have been mixed. LEADER [16], SUSTAIN-6 [17], Harmony Outcomes [18], and REWIND [19] demonstrated superiority for the composite MACE outcome vs. standard care. In contrast, other GLP-1 RA CVOTs demonstrated non-inferiority, including ELIXA [20], EXSCEL [21], and PIONEER 6 [22].

Results from these trials have led to changes in treatment guidelines and therapeutic indications for some glucose-lowering agents [23, 24]. Little is known about primary care providers’ knowledge of recent CVOTs and practice guidelines for the management of patients with T2DM. This survey was aimed at US IM and FM licensed physicians to examine their knowledge and beliefs about CVOTs for non-insulin glucose-lowering agents.


The survey included 253 IM and 250 FM physicians from a database of 47,534 Medscape members who were invited to participate in an online survey via e-mail. To be eligible, physicians had to be a licensed MD (Doctor of Medicine) or DO (Doctor of Osteopathic Medicine) in the USA or its territories, board-eligible or board-certified (self-reported), see patients for at least 75% of the working week, and treat at least 10 patients with T2DM per week. Physicians were excluded from participation if they were from Vermont and Maine because of state laws regarding this type of research. The survey was fielded September 18–20, 2018; a quota of 500 responses was set. The questionnaire consisted of 23 questions. The first section assessed how familiar physicians were with 14 CVOTs, and their knowledge of the primary MACE outcome, secondary outcomes, and safety signals. Physicians ranked how familiar they were with the 2018 American Diabetes Association (ADA) Standards of Medical Care in Diabetes treatment recommendations for individuals with atherosclerotic CV disease and T2DM [25]. The survey also asked physicians their opinions about management of their patients with T2DM and CV disease, and to identify the challenges they face when incorporating new recommendations from guidelines into their practice. Survey items were multiple choice and most used a 5-point Likert-type scale. Responses to each item were summarized and analyzed descriptively. A Z-test was used for statistical comparisons. Calculations were performed using Decipher software. The study protocol met the criteria for exemption by the Touro University of California Institutional Review Board (TUC IRB #M-3118). Participants were informed about the intentions of the research and how their personal information and responses would be used and their confidentiality protected; informed consent was provided via a tick box at the beginning of the survey.


A total of 702 physicians responded, of whom 199 did not meet eligibility criteria. Data were collected from 503 physicians (IM, n = 253; FM, n = 250) who met screening criteria and completed the survey.

Familiarity with CVOTs and Knowledge of MACE

Familiarity with CVOTs was low. Respondents were most familiar with the three CVOTs that had positive CV outcomes reported at the time of the survey (LEADER, 39%; EMPA-REG OUTCOME, 33%; CANVAS, 30%) as well as TECOS, 42% (P < 0.05 vs. other CVOTs). Physicians were least familiar with EXAMINE, ELIXA, and VERTIS-CV (P < 0.05 vs. other CVOTs). All of the CVOTs that had been reported at the time of the survey demonstrated at least non-inferiority (i.e., no difference in risk between the active agent and placebo) for the time to first occurrence of CV death, non-fatal myocardial infarction, or non-fatal stroke ± hospitalization for unstable angina (MACE endpoint). At the time of the survey, four trials showed superiority (i.e., decreased risk vs. placebo) for the composite MACE endpoint; fewer than half of respondents identified those as EMPA-REG OUTCOME, LEADER, CANVAS, and SUSTAIN-6 (Fig. 1a). Similarly, 36% and 26% correctly chose EMPA-REG OUTCOME and LEADER, respectively, as showing superiority vs. placebo for CV mortality (Fig. 1b). Only 21% and 16% selected that EMPA-REG OUTCOME and CANVAS, respectively, showed superiority for HHF.

Fig. 1
figure 1

Percentage of physicians who correctly responded that these CVOTs showed superiority for MACE (a) or CV death (b). CV cardiovascular, CVOT cardiovascular outcomes trial, MACE major adverse cardiovascular event

Knowledge of Adverse Events in CVOTs

Knowing the adverse events of the different classes of glucose-lowering agents is essential when choosing the appropriate treatment for patients with T2DM and CV disease. Acute pancreatitis is a labeled warning and precaution in the prescribing information for incretin-based therapies (DPP4 inhibitors and GLP-1 RAs). However, no significant differences in the incidence of acute pancreatitis between DPP4 inhibitor and placebo groups were observed in SAVOR-TIMI 53, TECOS, and EXAMINE [6,7,8]. Respondents were unaware of these findings; 20% thought all three CVOTs showed a significant increase in the risk of acute pancreatitis. Similarly, no significant differences in the incidence of pancreatitis were observed between groups in the GLP-1 RA CVOTs [16, 17, 20, 21]. For this adverse event, 18% thought all of the listed GLP-1 RA CVOTs showed a significant increase in risk. Among the GLP-1 RA CVOTs, the rates of retinopathy complications were significantly higher in the SUSTAIN-6 trial (HR 1.76; 95% CI 1.11, 2.78) [17]. Only 8% of respondents selected that retinopathy occurred more frequently with semaglutide than placebo in SUSTAIN-6.

In the CANVAS program CVOT, an increased risk of lower limb amputations was observed [14]. This was not well recognized by respondents, with 36% selecting CANVAS alone for this finding. Empagliflozin use in EMPA-REG OUTCOME was not associated with an increased risk of lower limb amputations compared with placebo [26]. The majority of respondents correctly recognized this with only 8% selecting EMPA-REG OUTCOME.

Beliefs and Challenges Regarding Treatment Priorities

When provided with a list of seven treatment priorities ranked from most (1) to least (7) important, 33% of respondents ranked using T2DM medications with CV benefits as least important, and 31% ranked managing CV risk in general as most important (P < 0.05 vs. all other categories). The main challenges primary care providers experienced using T2DM agents with CV benefits included insurance coverage (75%), cost (73%), and formulary changes (61%).


Results for some CVOTs were not available at the time the survey was conducted in September 2018 (e.g., CAROLINA, CARMELINA, REWIND, DECLARE-TIMI 58). Nevertheless, this survey showed that in 2018 most IM and FM physicians were not familiar with the findings of recent CVOTs, despite the impact these trials have had on clinical practice guidelines for patients with T2DM at high CV risk. These findings are important as about one third of respondents made reducing CV risk their top priority but did not prioritize agents that have been shown to reduce risk.

Most physicians were unaware of the reduced risk of MACE events demonstrated for active treatment vs. placebo in EMPA-REG OUTCOME, LEADER, CANVAS, and SUSTAIN-6, and the reduced risk of HHF demonstrated in EMPA-REG OUTCOME and CANVAS. As these medications can impact the number one cause of death in people with diabetes, it is important to reconsider how results from CVOTs are disseminated to primary care providers.

Diabetes guidelines, such as the 2020 ADA Standards of Care [27] and the 2019 Consensus Statement from the American Association of Clinical Endocrinologists [28], explicitly recommend SGLT2 inhibitors and GLP-1 RAs with CV benefits for patients with these conditions. Nevertheless, a recent nationwide cohort study that retrospectively analyzed medical and pharmacy claims from 1,054,727 adults with pharmacologically treated diabetes revealed that patients with prior myocardial infarction or heart failure were less likely to receive SGLT2 inhibitors (OR 0.94 and 0.93, respectively, P < 0.001 for both), despite evidence supporting their preferential use in these patients [29].

With respect to adverse events, physicians tended to overestimate the risk of pancreatitis with DPP4 inhibitors and GLP-1 RAs. In contrast, there was under-recognition of the increased risk of retinopathy complications associated with semaglutide in SUSTAIN-6, and the increased rates of lower limb amputations associated with canagliflozin in the CANVAS program.

While it takes time for new scientific findings to be implemented in daily practice, physicians who do not use guideline-recommended approaches to CV risk reduction in patients with T2DM at increased CV risk are likely to miss the opportunity to achieve reductions in CV-related morbidity and mortality for their patients.

There are several limitations in this US study, such as the relatively small sample size that may not be representative of all IM and FM physicians. Because participants were self-selected Medscape users who responded to an email invitation, it is not possible to exclude the potential for bias by over-representation of individuals with similar characteristics. However, the findings of this survey can provide information for hypothesis generation and stimulus for further research. The survey also was conducted at a single point in time in 2018. As additional studies have been published that confirm the CV benefits of SGLT2 inhibitors and GLP-1 RAs, IM, and FM physicians likely have increased awareness of CVOTs compared with when the survey was originally conducted.

In conclusion, the findings from this survey suggest knowledge gaps among IM and FM physicians about CVOTs, which have implications for the treatment of patients with T2DM and CV disease. There is a need for enhanced training and dissemination of clinical practice guidelines to ensure that patients with T2DM and CV risk can benefit from evidence-based therapies.