Interventions in type 2 diabetes mellitus and cardiovascular mortality–An overview of clinical trials

Elevated blood glucose levels have been associated with higher cardiovascular risk [], thus it was believed that managing blood glucose levels should lead to a decrease in cardiovascular events. Many trials have been performed, and results have been controversial.

In 1970, the University Group Diabetes Program (UGDP), was the first cardiovascular intervention trial and patients were randomized between either a variable-dose of insulin, standard-dose insulin, tolbutamide, phenformin or a placebo. The tolbutamide arm was prematurely stopped due to excess mortality (12.7% versus 4.9% in the placebo arm) []. These results created major controversy, among others due to problems in the execution and design of the study. It would take over 20 years before other trials attempted to reduce cardiovascular risk in diabetes by improving glycaemia.

The United Kingdom Prospective Diabetes Study (UKDPS), published in 1998, randomly assigned patients with newly diagnosed T2DM to either intensive treatment (sulphonylurea or insulin) or conventional glucose lowering strategy with diet restriction. Additionally, 342 overweight patients were allocated to metformin. After a median follow-up of 10 years, only a significant relative risk reduction was found for all-cause mortality and myocardial infarction in the metformin group [, ]. However, 10 years post-trial follow-up showed significant reductions in both groups for all-cause mortality and myocardial infarctions [].

Trials studying metformin only found significant benefits on composite macrovascular events [, ]. Discussion remains whether the beneficial effects were due to glucose lowering or to other effects of metformin such as weight loss or improvement of endothelial function [].

The peroxisome proliferator-activated receptor (PPAR)-γ agonists, or the so-called thiazolidinediones, have created major controversy and impact on the design of clinical trials. The Diabetes Reduction Assessment with Ramipril and Rosiglitazone Medication (DREAM) trial with rosiglitazone in subjects with impaired fasting glucose or glucose tolerance did not show cardiovascular benefit []. The Prospective Pioglitazone Clinical Trial in Macrovascular Events (PROactive) study resulted in a significant reduction of the composite of mortality and macrovascular events versus placebo, but no significant reduction in mortality []. After additional follow-up all cardiovascular benefits disappeared [].

In 2007, Nissen et al. published a meta-analysis reviewing rosiglitazone where an increased risk for myocardial infarction was found [] resulting in a warning on the use of rosiglitazone the Food and Drug Authority (FDA) []. As a result, studies had to include more patients with a longer follow-up period resulting in many studies aiming at non-inferiority compared to current therapies.

The Rosiglitazone Evaluated for Cardiovascular Outcome in oral agent combination therapy for type 2 Diabetes (RECORD) trial did not show an increased cardiovascular risk nor a cardiovascular benefit, but there was a higher incidence of heart failure (HR 2.10) []. Another trial by Dargie et al. showed similar results in patients with preexisting heart failure []. The reevaluation of the rosiglitazone data resulted in ending of the restrictions in 2013 by the FDA []. The Insulin Resistance Intervention after Stroke (IRIS) trial studied pioglitazone in patients with insulin resistance (but without diabetes) and a recent TIA or ischemic stroke. Again no difference was found in all-cause mortality, but a significant reduction infatal and nonfatal myocardial infarctions and stroke was found []. In light of the beneficial results of the PROactive and the recent IRIS trial, a new interest for thiazolidinediones has emerged and it has been suggested that a combination with an SGLT2 inhibitor might offset the fluid retention [].

Only a few trials compared different insulin regimens. The UGDP compared a variable dose of insulin versus standard dose and no difference in mortality and cardiovascular events was found []. A comparison of a three-time daily regimen with a two-time daily insulin strategy, after an acute myocardial infarction, was stopped prematurely due to a lack of efficacy []. The Outcome Reduction with an Initial Glargine Intervention (ORIGIN) trial, comparing insulin glargine versus standard glucose control in patients with increased cardiovascular risk, found no difference in major cardiovascular disease outcome []. Also no difference in survival was found in a trial comparing insulin-sensitization therapy with insulin-provision (either insulin or sulfonylureas) [].

Many trials have been performed evaluating intensive glucose control on cardiovascular end-points. The Diabetes Mellitus Insulin-Glucose Infusion after Myocardial Infarction (DIGAMI) study, published in 1997, compared intensive glucose control with usual metabolic control after an acute myocardial infarction. The overall mortality was significantly [] and a post-trial follow-up showed that this effect attenuated slightly []. However, a next trial with twice the number of patients did not show a significant reduction, neither did the post-trial follow up [, ]. There are some differences between the studies, the second included patients with lower glucose levels at baseline and not all subjects reached their target. Furthermore, due to slow patient recruitment the study was stopped prematurely with only a third of the planned subjects included.

In 2008 and 2009, three large randomized controlled trials were published assessing intensive versus standard glucose-lowering therapy. The Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial was stopped due to an increase in all-cause and cardiovascular mortality in the intensive group (respectively HR 1.22 and HR 1.35) []. A post-trial follow-up showed that the increased mortality had attenuated and there was no effect on macrovascular events []. The Action in Diabetes and Vascular Disease: Pretarax and Diamicron Controlled Evaluation (ADVANCE) trial did not report an increased mortality; neither did it show benefits with respect to mortality and macrovascular events. The post-trial follow up also failed to show benefits [, ]. The ACCORD trial aimed for a stricter glucose regulation with a target HbA1c below 6% versus <6.5%. Finally, the Veterans Administration Diabetes Trial (VADT) assigned veterans to either a 1.5% reduction in HbA1c or standard therapy. Again, no reduction in all-cause or cardiovascular mortality was found []. Interestingly, the post-trial follow-up, in which participants of VADT were followed for an additional 9.8 years revealed a significant reduction in time to first major cardiovascular event preventing 8.6 major cardiovascular events per 1000 person years. However no difference in mortality was found []. Looking at these three trials and with the UKDPS in mind, intensive glucose treatment does not seem to reduce cardiovascular mortality. However, after extended follow-up, there is evidence for reduction of cardiovascular events. This was also found in a meta-analysis from 2016, concluding that intensive treatment resulted in a significant relative risk reduction in time to first major cardiovascular events. There was no reduction in all-cause or cardiovascular mortality [].

The last decade, several new therapeutic options have become available. GLP-1 is an incretin mainly synthesized in the intestine and secreted postprandially. The cardiovascular benefit of GLP-1 agonists is not only by stimulation of insulin release but also due to effects on inflammation, slowed gastric emptying, enhanced satiety and improved endothelial function []. Three major trials have been conducted with GLP-1 agonists. The Evaluation of Lixisenatide in Acute Coronary Syndrome (ELIXA) trial showed no differences in mortality or the composite of cardiovascular events []. The Liraglutidin Effect and Action in Diabetes: Evaluation of Cardiovascular Outcome Results (LEADER) trial, with patients at high cardiovascular risk, resulted in small but significant reductions in all-cause and cardiovascular mortality and in composite cardiovascular events []. Furthermore, the recently published Trial to Evaluate Cardiovascular and Other Long-term Outcomes with Semaglutide in Subjects with Type 2 Diabetes (SUSTAIN-6) trial, with patients on a standard regimen, also showed benefit in composite cardiovascular mortality and events [].

Dipeptidyl-peptidase 4 (DDP4) inhibitors prevent breakdown of GLP-1. In 3 trials with different drugs, no reductions in cardiovascular events or mortality compared to placebo were found [, , , ]. Of note, there was a significant, unexplained increase in hospitalization of patients on saxagliptin in the Saxagliptin and Cardiovascular Outcomes in Patients with Type 2 Diabetes Mellitus (SAVOR-TIMI) trial due to heart failure []. It needs to be underlined that all three studies were designed to show non-inferiority and with algorithms to minimalize differences in HbA1c between the groups.

Sodium-glucose transporter 2 (SGLT-2) inhibitors prevent the reabsorption of glucose in the proximal tubule, resulting in lower blood glucose levels. The Empagliflozin, Cardiovascular Outcomes, and Mortality in Type 2 Diabetes (EMPA-REG) study published in 2015, with 8020 patients with T2DM at high cardiovascular risk, randomized to either empagliflozin or a matching placebo and was designed as a non-inferiority trial. The results were unexpected and surprising. After a median follow-up of 3.1 years the primary outcome, a composite of death from cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke showed a protective effect by empagliflozin (HR 0.86; 95% confidence interval 0.74–0.99). This reduction was also seen in all-cause and cardiovascular mortality []. Interestingly, most patients did not reach their glycemic target, therefore, other mechanisms must have been responsible for these protective effects. Since its publication, EMPA-REG has been the subject of many speculations and several mechanisms have been proposed. At this moment, no clear explanation can be given for these positive results and future trials with SGLT-2 inhibitors must show reproducible and solid results before this type of drugs can be used widely for the reduction of cardiovascular disease in T2DM. The Canagliflozin Cardiovascular Assessment Study (CANVAS) is expected to appear in 2017 and the Dapagliflozin Effect on Cardiovascular Events (DECLARE) trial is expected in 2019.

Diabetic dyslipidemia is characterized by increased triglycerides (TG), decreased HDL-C, increased small density LDL particles, reflected by elevated plasma apolipoprotein B levels and postprandial hyperlipidemia []. In clinical guidelines, LDL-C lowering is the main target for lipid-lowering therapy and cardiovascular risk reduction for patients with and without diabetes. Only few studies have been performed in T2DM and most results are from studies with a subgroup analysis.

In primary prevention, treatment with statins resulted in conflicting results. Treatment of patients with hypercholesterolemia and hypertension with pravastatin or atorvastatin did not show significant reductions in both all-cause and cardiovascular mortality [, ]. Only one primary prevention study specifically aimed at diabetic patients has been performed. The Collaborative Atorvastatin Diabetes Study (CARDS) randomized patients, with either hypertension, retinopathy, microalbuminuria or smoking, between atorvastatin or placebo. Significant reductions in major cardiovascular events and cardiovascular mortality were found, but not in all-cause mortality []. The Heart Protection Study (HPS) included a group of diabetic patients without coronary artery disease. In this subgroup a significant reduction in coronary mortality and nonfatal myocardial infaction was found [].

In secondary prevention, the landmark Scandinavian Simvastatin Survival Study (4S), published in 1994, randomized patients with angina pectoris or a previous myocardial infarction to either simvastatin or placebo. A significant reduction in major cardiovascular events was found in the T2DM subgroup, but the benefit of all-cause mortality, found in the total study group, was not reproduced in the diabetes subgroup [, ]. However, the diabetes subgroup only consisted of 4.5% of subjects from the total study population.

Concerning diabetes subgroups of patients with coronary heart disease, three trials evaluated treatment with pravastatin and only the Cholesterol And Recurrent Event (CARE) trial showed significant reductions in a composite of fatal coronary events or nonfatal myocardial infarctions [, , ]. The Pravastatin in Elderly Individuals at risk of Cardiovascular Disease (PROSPER) trial even found a non-significant higher incidence of macrovascular events after treatment with pravastatin, however the diabetes subgroup was small (5.5%) and the study was performed with elderly patients []. In the Greek Atorvastatin and Coronary-heart-disease Evaluation (GREACE) study, with patients with coronary heart disease, atorvastatin treatment showed a significant reduction of all-cause mortality, coronary mortality and coronary morbidity, including the diabetes subgroup [, ]. Both, simvastatin and fluvastatin, in both the total study groups as well as in the diabetes subgroups, reduced major cardiovascular events [, ]. In contrast, the “Deutche Diabetes Dialysis Studie” (4D) trial with T2DM patients on hemodialysis did not show any benefit from atorvastatin, neither did the 2016 post-trial follow-up [, ]. The Atorvastatin Study for Prevention of Coronary Heart Disease Endpoints in Non-insulin-dependent Diabetes Mellitus (ASPEN) initially started as secondary prevention in diabetic patients but due to changes in treatment and prevention guidelines also patients without a prior history of myocardial infarction were included. Treatment with atorvastatin failed to demonstrate a significant reduction in a composite of cardiovascular events and also no significant benefit in mortality was found [].

A meta-analysis from the Cholesterol Treatment Trialists Collaboration in 2008, which included 14 randomized trials and 18,686 individuals with diabetes, focused on the reduction in cardiovascular risk with statin therapy. After a mean follow-up of 4.3 years a significant reduction in major vascular events per 1 mmol/L LDL-C reduction (RR 0.79) was found, irrespective of the baseline characteristics or prior history of vascular disease. Only a proportional reduction in all-cause mortality was found [].

Besides statins, other options are also available for LDL-c lowering. The Improved Reduction of Outcomes: Vytorin Efficacy International Trial (IMPROVE-IT) showed a reduction in the composite of cardiovascular events by adding ezetimibe on top of simvastatin in patients hospitalized after an acute coronary syndrome []. More importantly, treatment with the novel proprotein convertase subtilisin/kexin type 9 (PCSK9) inhibitors seems promising. Recently, the first cardiovascular outcome trial was published, which included 36.6% of subjects with diabetes. In the Further Cardiovascular Outcome Research with PCSK9 Inhibition in Subjects with Elevated Risk (FOURIER) patients with atherosclerotic cardiovascular disease who were on statins were treated either with evolucumab or placebo. After a follow-up of 48 weeks, a significant reduction in the composite cardiovascular endpoint (HR 0.85) and cardiovascular events was found []. A second trial, with a different PCSK9 inhibitor, alirocumab is expected in the course of 2017 [].

Other options are also available for improving the lipemic spectrum. Two HDL-C raising trials, with either niacin or dalcetrapid, were stopped prematurely due to futility [, ]. Three trials have been published with cardiovascular endpoints with fibrates, of which two trials were specifically designed for patients with diabetes, however no cardiovascular benefit was found [, ]. However, treatment with fibrates reduces TG and improves HDL-c levels [] and both studies included subjects with normal TG levels were included. Within subgroups with high TG and low HDL-c levels significant cardiovascular benefit were found [, ]. The Veterans Affair High-density lipoprotein cholesterol Intervention Trial (VA-HIT), in men with a cardiovascular history (and low HDL-c and high TGs) randomized between either gemfibrozil or a matching placebo, showed a significant reduction in cardiovascular risk in the T2DM subgroup [, ]. Within diabetic subjects with high TG and low HDL-cholesterol levels, treatment with fibrates seem to reduce cardiovascular risk but also no effect on mortality was found.

The Hypertension Optimal Treatment (HOT) was the first trial to compare different blood pressure (BP) regiments and showed that reduction of the diastolic BP ≤80 mmHg led to a significant reduction in cardiovascular events in diabetic patients with hypertension []. The UKDPS showed that an increase in BP was related to a higher prevalence of micro- and macrovascular disease, but also suggested that tight BP control in patients with T2DM can be beneficial []. However, the 10-year post trial follow-up showed that all these effects did not sustain probably because the differences in BP were lost []. The ACCORD trial, also comparing intensive versus regular BP therapy in diabetic patients at high cardiovascular risk, did not result in a significant reduction in the composite of cardiovascular endpoints and in the intensive BP lowering arm a higher number of serious adverse events occurred []. However, the follow-up showed a significant reduction in the composite of cardiovascular endpoints [].

The ADVANCE trial also studied high risk diabetic patients. The fixed combination of perindopril and indapamide versus placebo resulted in a significant reduction in all-cause and cardiovascular mortality, but no significant reduction in macrovascular events []. In the post-trial follow-up these benefits attenuated but remained [].

Importantly, besides reduction of BP, angiotensin converting enzyme inhibitors (ACEi) and angiotensin II receptor blockers (ARB) have another important role in diabetes care by reducing urinary albumin excretion, which is an independent risk factor for cardiovascular disease []. Several trials aimed to improve cardiovascular outcomes by reducing urinary albumin excretion. In the Heart Outcome Prevention Evaluation (HOPE) trial, in which patients were included with a history of cardiovascular events or diabetes with an additional CV risk factor, participants were randomized between either ramipril or placebo. Per study design, no differences in BP reduction between groups were found. However, ramipril resulted in reductions in all-cause and cardiovascular mortality and the composite of macrovascular events. Furthermore, also benefits on overt nephropathy were found []. The irbesartan in patients with nephropathy due to Type 2 Diabetes (IDNT) trial, randomized T2DM patients with nephropathy to either irbesartan, amlodipine or placebo. All groups had the same target BP for which other medication was allowed. Treatment with irbesartan showed a significant benefit on slowing the progression of nephropathy []. However, no differences in mortality and cardiovascular events were found []. In other trials in diabetic patients treated to specific BP targets, the addition of ARBs versus placebo slowed the progression of nephropathy, with no significant effects on mortality and morbidity [, , ]. A meta-analysis from 2006, showed renoprotective benefits in patients with established diabetic nephropathy, however only ACEi also showed a reduction of all-cause mortality (RR 0.79) []. In diabetic patients without nephropathy, ACEi and ARBs did not reduce mortality, but delayed the onset of microalbuminuria [, ]. Importantly, most of these trials did not aimed at BP reduction and especially given the benefits found in the HOPE trial, hypertensive diabetic subjects with microalbuminuria should be treatment with either ACEi or ARBs. Other trials studied ACEi or ARB for cardiovascular outcomes in BP management. In patients with T2DM and mild retinopathy, treatment with ARB versus placebo did not result in cardiovascular benefit [] and in diabetic patients and high cardiovascular risk neither ACEi versus ARB or ACEi versus a combination resulted benefit []. The Losartan Intervention For Endpoint reduction in hypertension (LIFE) studied the effect of losartan versus atenolol in subjects with hypertension and left ventricular hypertrophy. Treatment with losartan resulted in significant reductions on the composite of cardiovascular events and mortality in the diabetic subgroup []. Other combinations have also been studied, the combination of ACEi with a calcium channel blocker led to a reduction in combined cardiovascular outcomes compared to ACEi and a diuretic [, ]. Finally, treatment with aliskiren (a renin antagonist) on top of usual therapy did not result in benefit. This trial was terminated prematurely due to adverse events [].

Two recent meta-analysis, including patients with and without a history of cardiovascular disease, showed benefit due to blood pressure lowering in patients with T2DM. Each 10 mmHg lowering of systolic BP was associated with a significant lower risk of all-cause mortality and cardiovascular events. Patients with a baseline BP of >140 mmHg had significant reduction in cardiovascular events and mortality [, ]. Surprisingly, when baseline BP was <140 mmHg, an increased risk for cardiovascular mortality was found. The authors proposed that this could be due to an impaired blood flow to end-organs.

The Look AHEAD (Action for Health in Diabetes) trial is the only trial aiming at an intensive lifestyle intervention in T2DM designed to evaluate cardiovascular outcomes. Overweight or obese patients with T2DM were randomized between intensive lifestyle intervention (promoting weight loss by decreased caloric intake and increased physical activity) and regular diabetes support. The trial was stopped prematurely due to the lack of effect. Though the weight loss was greater in the intervention group (6.0% versus 3.5%), the primary outcome, a composite of cardiovascular events did not differ nor did mortality []. Importantly, the achieved weight loss attenuated over time and the trial was stopped prematurely due to lack of effect. However, since the median duration of DM was only 5 years, one would expect that macrovascular complications of diabetes could take longer to emerge. Even though the difference in definitive weight loss between the groups was only 2.5%, significant reductions in hospitalizations, number of medications, and health-care costs were found [].

The Swedish Obesity Study (SOS) randomized 4047 obese patients to either bariatric surgery or usual care. In the total study group a 29% reduction in overall mortality was found and in the diabetic subgroup of 607 patients a significant reduction in cardiovascular events was found [, ]. When comparing these results with the Look-AHEAD, the improved cardiovascular outcome found in the SOS trial could be explained by both a greater weight loss (16% versus 6%) and an improved glycaemia (in the Look-AHEAD no reduction of glycated A1c was found). In a study from 2015 by Mingrone et al., with 60 patients and a mean follow-up of 5 years, no cardiovascular endpoints occurred during the study period []. In an observational cohort study by Eliasson et al. from 2015, data from 6132 obese patients with T2DM were compared with matched controls from other databases. After a median follow-up period of 3.5 years, significant reductions in overall and cardiovascular mortality were found in patients who underwent a gastric bypass []. These studies suggest benefits with regard to cardiovascular mortality and events by bariatric surgery in patients with T2DM.

The Steno-2 trial showed that a multifactorial intervention (with either tight glucose regulation, the use of renin-angiotensin system blockers, aspirin, and lipid lowering agents) resulted in lower cardiovascular events compared to conventional therapy []. Furthermore, after an additional observational follow-up the significant reduction in cardiovascular events remained and also the total and cardiovascular mortality was significantly lower [].