Glucose-lowering therapies in patients with type 2 diabetes and cardiovascular diseases

Abstract

Type 2 diabetes mellitus is a major risk factor for developing cardiovascular disease, and many patients with diabetes have prevalent cardiovascular complications. Recent cardiovascular outcome clinical trials suggest that certain new glucose-lowering drugs are accompanied by additional cardioprotective properties. Indeed, selected glucagon-like peptide-1 receptor agonists have a proved cardiovascular benefit in terms of a reduced incidence of ischaemic events, while sodium/glucose co-transporter-2 inhibitors have also shown significant protection, with a striking effect on heart failure and renal endpoints. These findings have been integrated in recent guidelines which now recommend prescribing (when initial metformin monotherapy fails) a glucagon-like peptide-1 receptor agonist or a sodium/glucose co-transporter-2 inhibitor with clinical trial-confirmed benefit in patients with diabetes and atherosclerotic cardiovascular disease, and a sodium/glucose co-transporter-2 inhibitor in such patients with heart failure or chronic kidney disease at initial stages. Furthermore, the new 2019 European Society of Cardiology guidelines in collaboration with the European Association for the Study of Diabetes recommend a glucagon-like peptide-1 receptor agonist or a sodium/glucose co-transporter-2 inhibitor in treatment-naive patients with type 2 diabetes mellitus with pre-existing cardiovascular disease or at high cardiovascular risk. Future research will disentangle the mechanisms underpinning these beneficial effects and will also establish to what extent these results are generalisable to the whole diabetes population. In the meantime, available evidence should prompt a wide diffusion of these two classes of drugs among patients with diabetes and cardiovascular disease. Here, we briefly summarise recent findings emerging from cardiovascular outcome clinical trials, discuss their impact on treatment algorithms and propose new possible approaches to improve our knowledge further regarding the cardiovascular effect of glucose-lowering medications.

Introduction

Patients with type 2 diabetes mellitus (T2DM) have a higher incidence of cardiovascular disease (CVD), with a still striking cardiovascular mortality despite continuous advances targeting multiple risk factors.¹ Initial evidence from the UKPDS cohort clearly indicated that, at least in the absence of modern lipid-lowering and blood pressure-lowering drugs, an early and intensive glycaemic control lowers the incidence of late cardiovascular complications.² However, successive analysis suggested that ‘the means’ by which the glycaemic target is reached is of great importance. Indeed, data from the same cohort showed that metformin was superior to sulphonylureas in terms of long-term cardiovascular protection, at least in obese subjects.³ On the other hand, subsequently introduced drugs, i.e. thiazolidinediones, have been accompanied by cardiovascular safety concerns despite a tangible glucose-lowering effect.⁴ In particular, certain meta-analyses have shown that rosiglitazone may increase the risk of myocardial infarction (MI) and heart failure (HF), while this effect was not observed with another member of this class.⁵ This and other considerations prompted the introduction (in 2008) of US Food and Drug Administration (FDA)/European Medicines Agency (EMA)-requested cardiovascular outcome trials (CVOTs) to demonstrate non-inferiority against placebo in terms of cardiovascular safety for newly introduced glucose-lowering drugs (GLDs). In CVOTs, the primary outcome is usually a combination of cardiovascular endpoints, including cardiovascular mortality, non-fatal MI, non-fatal stroke and, in some cases, the hospitalisation rate for unstable angina pectoris (3 or 4 point major adverse cardiovascular events; MACEs). Non-inferiority is usually defined as a two-sided 95% confidence interval (CI) upper boundary of 1.8 risk ratio (pre-approval) and/or 1.3 risk ratio (post-approval) for MACE versus the control group, and the monitoring must cover at least 2 years with event-driven stop after the occurrence of more than 600 primary events. Secondary outcomes often include hospitalisation for HF but also renal endpoints.^5,6 Since then, a number of trials have been conducted, mainly encompassing four classes of drugs: (a) dipeptidyl-peptidase-4 inhibitors (DPP-4is); (b) glucagon-like glucagon-like peptide-1 receptor agonists (GLP-1RAs); (c) sodium/glucose co-transporter-2 inhibitors (SGLT-2is); and (d) insulins.^5,6 The main findings of these trials and how they have been integrated into newly released therapeutic guidelines are briefly summarised in this article. The importance of determining the underpinnings of the observed cardioprotective effects is underlined, in order to infer which population of non-complicated patients with diabetes will benefit most from their use. Finally, some potential improvements in the design of CVOTs are suggested to fill the still existing gaps in knowledge regarding the effect of glucose-lowering medications on the cardiovascular system.

Evidence emerging from CVOTs

Patients at high cardiovascular risk

Results have been variable and are the subject of intense research.^5,6 However, it is increasingly evident that: (a) DPP-4is are accompanied by a neutral effect on short-term cardiovascular outcomes;^7–9 (b) four out of six GLP-1RAs have a confirmed cardiovascular benefit against MACEs, especially in terms of the reduced incidence of ischaemic events;^10–15 (c) empagliflozin and canagliflozin have shown protection against MACEs, while both plus dapagliflozin (all SGLT-2is) have also demonstrated a tangible effect on HF and renal endpoints.^16–18

In detail, analysis of secondary endpoints revealed an increased risk (or a trend towards) for HF for selected DPP-4is, namely saxagliptin and alogliptin.^7,8 A later meta-analysis did, however, question these results.⁶ However, the FDA issued a safety warning on these two drugs increasing the risk of HF, especially in patients with prevalent cardiovascular morbidity.¹⁹ The other DPP-4is, i.e. sitagliptin and linagliptin, have not been accompanied by similar concerns and can safely be used in patients with diabetes and HF.²⁰ Of note, the recent CARMELINA trial suggested that linagliptin may decrease the incidence of microvascular outcomes, as shown in an exploratory analysis.²¹ It is worth mentioning that linagliptin has also been tested in comparison with glimepiride in the only active-comparator CVOT, i.e. the CAROLINA trial, which showed long-term cardiovascular safety for this drug.²²

Regarding GLP-1RAs, liraglutide,¹¹ semaglutide,¹² albiglutide,¹⁴ and dulaglutide¹⁵ showed superiority against placebo in the composite MACE primary outcome, while lixisenatide²³ and exenatide¹³ did not. Liraglutide was also accompanied by a consistent (but not statistically significant) decrease in the incidence of MI and stroke,¹¹ an effect also observed with semaglutide, dulaglutide (especially stroke) and albiglutide (especially MI).^12,14,15 This evidence, coupled with the observation that the beneficial effect becomes evident after one year of treatment, argues in favour of an ameliorating effect of GLP-1RAs against the progress of the atherosclerotic process and/or stabilisation of existing plaques. On the other hand, two GLP-1RAs have shown neutral effects in CVOTs. While for exenatide a non-significant, protective trend was observed,¹³ neutral results regarding lixisenatide may suggest that short-acting GLP-1RAs are not as effective as their long-acting counterparts, or that drug structure may influence their cardiovascular effects.¹⁴ Of note, the REWIND trial (dulaglutide) also included a consistent proportion of T2DM patients in primary prevention, with an overlapping protective effect against MACEs when compared to patients with pre-existing CVDs.¹⁵

An unresolved question is why the SGLT-2is empagliflozin and canagliflozin have shown protection against 3 point MACEs^16,17 while dapagliflozin did not, despite a significantly lower rate of cardiovascular death or hospitalisation for HF than placebo.¹⁸ Importantly, all the SGLT-1is tested so far have been accompanied by protective effects on HF hospitalisation and renal outcomes in secondary endpoint analysis.²⁴ Worth mentioning is the fact that these results appear also to apply to patients without previous HF, as shown in DECLARE-TIMI,¹⁸ but also in real world studies, i.e. observational data obtained outside the context of randomised controlled trials and generated during routine clinical practice.²⁵ Of note, these are characterised by a higher risk of confounding variables but a wider and a more heterogenous population of patients. Of note, the CREDENCE trial showed a decreased risk of cardiovascular events and kidney failure in diabetes patients with chronic kidney disease (CKD) treated with canagliflozin vs placebo.²⁶ Finally, ‘ischaemic’ endpoints were not significantly affected by SGLT-2i treatment in CVOTs. However, real-world studies suggest a beneficial effect of this class also against MI and stroke.²⁷

Thiazolidinediones increase insulin sensitivity, which may lead to increased sodium and water retention. In clinical trials, they have been associated with an increased risk of HF.^4,28 On the other hand, pioglitazone showed evidence of cardiovascular benefit (against fatal or non-fatal stroke or MI) in the PROactive²⁹ and IRIS trials,³⁰ which enrolled patients with previous ischaemic events and, in the latter, increased insulin resistance. Later, in a head-to-head comparison trial, pioglitazone was found not to be superior to sulfonylureas for cardiovascular event prevention in a lower-risk diabetes population.³¹ However, the results of that study have been questioned, mainly due to the small sample size and the high rate of treatment discontinuation.³²

In the DEVOTE trial, insulin degludec was not superior compared to insulin glargine in terms of MACEs, despite a significantly decreased risk of hypoglycaemia.^33,34

Patients in primary prevention

In contrast to the more common CVOT design, the EXSCEL, REWIND, CANVAS and the DECLARE-TIMI trials also enrolled T2DM patients in primary prevention and not only patients with pre-existing CVD or at high risk of CVD.^5,6 While these findings may still be taken with caution, also considering that the same definition of high cardiovascular risk and/or CVD has been heterogeneous in different trials,⁶ these results should encourage further exploration of the cardiovascular effects of new GLDs in the general diabetes population. Furthermore, the putative independence of the population enrolled and the homogeneity of the results in terms of specific endpoints support the theory that SGLT-2is are characterised by a protective class effect against HF and renal failure, while GLP-1RAs bear consistent anti-atherosclerotic properties. The results of the CVOTs completed at present are summarised in Table 1.³⁵

Despite these results showing a tangible impact of SGLT-1is and GLP-1RAs on cardiovascular outcomes in both primary and secondary prevention, the patients receiving these medications in clinical settings still represent a small proportion of the T2DM population. Indeed, applying the inclusion criteria of the EMPA-REG and LEADER trials (secondary prevention) in a large outpatient registry, a highly significant number of patients meeting these criteria were not treated with any of these two classes.³⁶ Similar or lower percentages of new GLD usage were found in populations in primary prevention,³⁷ suggesting that a broader and better targeted use of these medications should be considered.

Mechanisms of action

GLP-1RAs are degradation-resistant GLP-1 analogues presenting a plethora of pleiotropic effects both dependent and independent of the GLP-1-R.³⁸ The main effects explaining the glucose-lowering properties of GLP-1 include: (a) glucose-dependent increase in insulin synthesis and secretion; (b) glucose-dependent inhibition of glucagon release; (c) slowing of gastric emptying; (d) induction of satiety; (e) mild insulin sensitisation to promote glucose uptake in muscle and adipose tissue while halting hepatic glucose production.³⁸ DPP-4is are held to have almost overlapping effects, given that GLP-1 is cleaved by the DPP-4 enzyme.³⁹ Indeed, comparable HbA1c reductions have been observed among these two classes in clinical trials.⁵ However, as described above, the cardiovascular outcomes of DPP-4is and GLP-1RAs are different, implying that additional, probably divergent, cardiovascular mechanisms characterise these drugs.³⁹

SGLT-2 is the key co-transporter promoting the reabsorption of glucose from the glomerular filtration back into circulation, being responsible of almost 90% of the kidney’s glucose reabsorption.⁴⁰ SGLT-2 is mainly expressed on the epithelial cells lining the first segment of the proximal convoluted tubule. SGLT-2is prevent the kidney’s reuptake of glucose from the glomerular filtrate, thus lowering glycaemia by fostering glycosuria (accompanied by natriuresis).⁴¹

It is very unlikely that the reduction in glycated haemoglobin observed in CVOTs is the only or even the major reason for the beneficial cardiovascular and renal effect of GLP-1RAs and SGLT-1is. Similar or indeed more pronounced HbA1c reductions have been observed in trials of GLDs without cardiovascular benefits. The new GLDs are accompanied by a plethora of effects, e.g. blood pressure lowering, weight reducing, metabolic and haemodynamic adjustments, hormonal alterations, anti-oxidant and anti-inflammatory effects. Which and how many (if any) of these mechanisms underlie cardioprotection is matter of basic and clinical investigations.^24,42,43 The possible mechanisms of GLP-1RAs and SGLT-2is in reducing the risk of CVD or kidney disease are shown in Figure 1. Understanding the underpinnings of the beneficial effect of new GLDs is paramount for at least three reasons: (a) they could help to identify which kind of non-complicated patient with diabetes may benefit most from their use; (b) it can eventually prompt further drug designs with specific additional, beneficial effects; and (c) their use could eventually also be extended to some patients without diabetes, e.g. those with HF with regard to SGLT-2is.

ADA/EASD treatment algorithm

New information derived from CVOTs has been integrated in the most recent expert consensus document for the management of T2DM released by the American Diabetes Association (ADA) and EASD.⁴⁴ In contrast to previous indications, more flexibility and a personalised approach is advised instead of a fixed glycaemic target. Lifestyle interventions and metformin are still the first-line approach to manage hyperglycaemia. However, when these measures are insufficient, the add-on therapies should be personalised according to patient characteristics and comorbidities. In particular: (a) in the case of patients with atherosclerotic CVD, a GLP-1RA or a SGLT-2i (when the estimated glomerular filtration rate (eGFR) is acceptable) with confirmed cardiovascular benefit are recommended; (b) in patients with HF or CKD, SGLT-2is with evidence of reducing HF and/or CKD progression are recommended (with proper adjustments according to eGFR cut-offs);⁴⁵ (c) in patients with any of these conditions the risk of hypoglycaemia, body weight, side effects and cost should be considered to tailor the appropriate therapy.⁴⁴ Thiazolidinediones, saxagliptin and alogliptin should not be used in patients with diabetes with HF.⁴⁴

What if a cardiac patient develops diabetes? 2019 ESC guidelines in collaboration with EASD

The beneficial effect of GLP-1RAs and SGLT-2is on cardiovascular outcomes is emphasised in the most recent guidelines. However, metformin still represents the first approach for newly diagnosed patients. Usually, T2DM precedes the development of CVDs in unhealthy pathological trajectories.¹ However, the opposite can also happen. In case a patient with pre-existing CVD develops T2DM, which should be the first-line drug? Metformin has been safely used for decades in almost any kind of patient with diabetes, as suggested by ever-increasing observational data.⁴⁶ On the other hand and in contrast to GLP-1RAs and SGLT-2is, metformin did not undergo a randomised, placebo controlled clinical trial to show non-inferiority in terms of cardiovascular outcomes in populations with both T2DM and CVDs (for obvious historical reasons). The position of metformin in the clinical management of diabetes remains an open question worthy of discussion.⁴⁷

The guidelines released in 2019 by the American Heart Association (AHA) and the American College of Cardiology (ACC) suggests as a key take-home message that ‘for adults with T2DM and additional ASCVD risk factors who require glucose-lowering therapy despite initial lifestyle modifications and metformin, it may be reasonable to initiate a SGLT-2i or a GLP-1RA to reduce CVD risk’ (COR IIb, LOE B-R).⁴⁸ Noteworthy, and in contrast to all the previously published indications, the new 2019 ESC guidelines in collaboration with EASD have now clearly addressed this point for the first time.⁴⁹ Indeed, in the case of pre-existing CVD or a condition of high to very high cardiovascular risk (defined as long diabetes duration plus one additional risk factor and pre-existing organ damage or three risk factors, respectively), the first line therapy for treatment-naive T2DM patients should be a SGLT-2i or a GLP-1RA with confirmed cardiovascular benefit, while metformin remains the first option for all the other T2DM patients, especially those overweight and without CVD/at low cardiovascular risk.⁴⁹

Conclusion and future prospects

Recent pharmacological research has provided new medications in T2DM patients. Ten years of CVOTs have disclosed a plethora of findings suggesting that GLP-1RA and SGLT-2 inhibitors and, to a certain extent, pioglitazone, are accompanied by beneficial effects against a wide range of cardiovascular outcomes. Thus, the prevailing glucocentricity should be abandoned to the advantage of a holistic view of patients’ treatment with protection from cardiovascular events as an important and presently feasible priority. Both diabetologists and cardiologists should take advantage of these new tools to tailor personalised therapies for their patients. Patients with atherosclerotic CVD should be prescribed a GLP-1RA or a SGLT-2i at an early stage, not only to intensify the glucose-lowering regimen but to reduce the cardiovascular risk as an important part. Patients with T2DM and prone to develop HF will benefit from the use of a SGLT2i to prevent HF and CKD.^5,6,44 Importantly, this also applies to treatment-naive T2DM patients with high cardiovascular risk or pre-existing CVD being prescribed their first GLD.⁴⁹

As recently suggested,^6,33 there is still a chance to improve our knowledge further regarding the cardiovascular effect of glucose-lowering medications through CVOTs. Indeed, enrolling a more variegated and real-world representative population of diabetes patients, coupled by a longer follow-up would undoubtedly strength the findings, possibly amplifying the ‘target audience’ of new drugs and further driving treatment decisions. In addition, in CVOTs the drugs are compared to placebo. Despite the tentative treatment intensification in the placebo group to reach a comparable glycaemic target (often unsuccessful), it is highly likely that drugs used as background therapy themselves have cardiovascular effects (positive or negative). Also, given these results, future CVOTs should not exclude the use of new drugs with confirmed cardiovascular benefit for ethical reasons. The new 2019 ESC guidelines in collaboration with EASD have now split the approach of the treatment algorithm according to the cardiovascular background of the patient, suggesting GLDs with cardiovascular benefit as the first-line approach in T2DM patients with CVD or high cardiovascular risk while proposing metformin for the others.⁴⁹ However, it is still unclear which drug represents the best second line therapy in not-at-target T2DM patients without any co-morbidity, even if the use of a new GLD with cardiovascular benefit is advocated in the latest guidelines.^44,49 The use of active comparator medications, longer follow-ups and the inclusion of patients without pre-existing CVD and short diabetes duration in CVOTs would probably help to disentangle this point. Whatever the case, the best approach to take care of patients is staying up to date, in order to avoid clinical inertia⁵⁰ and always to provide the best option to increase the chances of preventing or delaying the harmful effects of T2DM on the cardiovascular system.

Author contribution

FP and AC contributed to the conception and design of the study and drafted the manuscript. LLS, LR, NM and MF contributed to acquisition and interpretation of literature and critically revised the manuscript. All authors gave final approval and agree to be accountable for all aspects of the work ensuring integrity and accuracy.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship and/or publication of this article: this work has been supported by the Italian Ministry of Health (Ricerca Corrente).

References