Close X

Incretin therapies for microvascular diabetes complications

9th November 2023

A number of drugs that mimic or enhance incretin agonism are now being used for glucose lowering in diabetes. They have proven cardiovascular safety and may even reduce cardiovascular complications. A new review in Diabetologia looks at the impact of incretin therapies on microvascular outcomes, such as nephropathy and retinopathy. Dr Susan Aldridge reports.

In the last 20 years, glucagon-like peptide-1 receptor agonists (GLP-1RAs, incretin mimics) and dipeptidyl peptidase-4 inhibitors (DPP-4is, incretin enhancers) have been developed as glucose-lowering agents for people with type 2 diabetes. New glucose-lowering agents now need to go through cardiovascular outcome trials (CVOTs) to prove safety before regulatory approval. These have shown cardiovascular safety for DPP-4is, while GLP-1RAs actually reduce major cardiovascular events. 

What is less clear is whether these incretin-based therapies also reduce microvascular complications because CVOTs do not report microvascular outcomes in detail, nor were they designed to do so. Clearly, though, the impacts – positive or negative – of GLP-1RAs and DPP-4is on microvascular complications are highly relevant for people with type 2 diabetes who may be using these therapies long-term.  

Both therapies increase GLP-1 agonism, which is associated with the attenuation of various pathological processes involved in microvascular disease. These include reduction of reactive oxygen species production, decreased inflammation and improved vascular function. DPP-4is also increase stromal cell-derived factor-1 (SDF-1), which is associated with neovascularisation and tissue repair.  

Rodent studies have already demonstrated several benefits of these agents in prevention or reversal of nephropathy, retinopathy and neuropathy. However, evidence from human studies is, so far, limited. What we do know about the impact of incretin therapies on various microvascular outcomes in people with type 2 diabetes is described in a review by Jonathan Goldney from the University of Leicester and colleagues.  


Individual CVOTs have not shown a reduction in the risk of hard renal endpoints with GLP-1RAs, but the trials do show a consistent reduction in albuminuria compared with placebo. And meta-analyses have now confirmed a risk reduction for a renal composite outcome too. The ongoing FLOW trial of semaglutide versus placebo, which is powered for a composite renal outcome excluding albuminuria, is underway to clarify the situation.

The mechanism behind the renoprotective effect of GLP-1 RAs is probably multifactorial, with GLP-1 receptors in the kidney likely playing an important role. One meta-analysis has suggested that reduction in systolic blood pressure by GLP-1 RAs explains away most of the reduction in nephropathy, while another found that HbA1c reduction was responsible. However, weight reduction from GLP-1 RAs appears not to play a role in reducing the risk of nephropathy. 

When it comes to DPP-4is, two of the CVOTs (CARMELINA and SAVOR-TIMI) showed a reduction in albuminuria, although to a lesser extent than with GLP-1RAs. However, a meta-analysis showed that this did not translate into a decrease in composite renal outcomes. So when it comes to nephropathy, both GLP-1RAs and DPP-4is appear to improve albuminuria, but only GLP-1RAs improve the more clinically meaningful composite renal outcomes. 

There are two possible reasons for this difference. GLP-1RAs are incretin mimics and therefore have a much higher degree of GLP-1 agonism than DPP-4is, which are only incretin enhancers. Furthermore, some of the benefits of GLP-1 agonism could be offset via the increase in SDF-1 with DPP-4is, as SDF-1 is associated with a number of pathological changes, such as natriuresis and renal hyperfiltration. 


The SUSTAIN-6 trial has shown that semaglutide increases the risk of retinopathy complications. However, other CVOTs and meta-analyses have shown no difference in ocular outcomes between GLP-1RAs and placebo. There has been much debate over why retinopathy risk was greater in SUSTAIN-6. Participants in this trial had a high prevalence of background retinopathy and post-hoc analysis has suggested that those without pre-existing retinopathy were at no greater risk of ocular events. So adverse outcomes on semaglutide might be limited to those who already have retinopathy. Furthermore, compared with other CVOTs, SUSTAIN-6 participants did have higher HbA1c, as the trial did not specify an upper limit. When on semaglutide, therefore, their fall in HbA1c was relatively large. It has been seen previously that a rapid fall in HbA1c on GLP-1RAs is associated with worsening retinopathy and this is backed by meta-analysis. 

The Diabetes Control and Complications Trial (DCCT) also showed that intensive glucose control was associated with a larger fall in HbA1c from baseline and worsening retinopathy at six and/or 12 months. However, over a mean follow-up of 6.5 years in the DCCT, the overall risk of retinopathy was actually reduced with intensive therapy compared with conventional treatment. Given that SUSTAIN-6 had an observation period of approximately two years, it is not clear whether a longer run would have seen the reversal of the negative relationship between semaglutide and retinopathy. The ongoing FOCUS trial, which is comparing semaglutide and placebo, will investigate further, reporting retinopathy progression at five years as its primary outcome.

For DPP-4is, a retrospective cohort study of German electronic records, with 630 patients, showed that vildagliptin use was associated with a lower incidence of retinopathy. Similar observations come from a smaller study of 82 patients in South Korea showing that DPP-4is are associated with reduced progression of retinopathy. However, a meta-analysis of interventional studies suggests that DPP-4is may actually increase the risk of retinopathy. Similar to the use of GLP-1 RAs, this increased risk may be related to the reduction in HbA1c.  

Together, these findings suggest caution in the use of semaglutide and DPP-4is with respect to retinopathy risk. There is debate over why the protective benefits of incretin therapies for the eye seen in pre-clinical studies were not reflected in CVOTs. Maybe this is due to physiological differences between rodents and humans or the lack of GLP-1 receptors within the retina, in contrast to the kidneys and nerves. For DPP-4is, the increase in SDF-1 may be harmful due to its neovascular effect, which could cause proliferation and damage resembling the pathophysiology of retinopathy.  

Peripheral neuropathy

Only a few studies on peripheral neuropathy have been done in humans, most of them with fewer than 100 participants. In the two larger studies – an observational study from the US national claims database OptumLabs Data Warehouse with 8252 patients and the GRADE trial, with 5047 – no difference in the incidence of neuropathy was seen in comparison with other glucose-lowering therapies. 

With DPP-4is, two large observational studies of electronic medical records have suggested a decreased incidence of neuropathy with both vildagliptin and sitagliptin compared with sulphonylureas. However, findings from clinical trials are mixed, with only one showing a 19.5% reduction in small fibre peripheral neuropathy with linagliptin compared with placebo. It is not clear why these findings differ, but it could be related to differences in comparator, different DPP-4is, different measures of neuropathy and varying statistical power. 

So, although the evidence is inconclusive, it looks as if DPP-4is are more promising than GLP-1 RAs when it comes to reducing the incidence of peripheral neuropathy. This is backed by an observational study showing that liraglutide appears to be associated with a higher incidence of neuropathy than sitagliptin. This may be linked to the increase in SDF-1 that occurs with DPP-4is. 

Other microvascular complications

Given that incretin therapies may decrease peripheral neuropathy risk, could they also protect against autonomic nervous system disorders such as cardiac autonomic neuropathy (CAN), gastroparesis and erectile dysfunction? It is not clear how GLP-1 agonism affects the autonomic nervous system, but the underlying mechanisms may be more complex than those related to peripheral neuropathy. For instance, there are GLP-1 receptors in the carotid body, whose activation decreases the sympathetic response to high plasma glucose and/or insulin.

Conversely, GLP-1RAs cause an increased heart rate, while other studies report decreased vagal tone and heart rate variability. There are inconsistencies in the literature, but evidence does suggest that GLP-1RAs could lead to autonomic imbalance and CAN. 

It is also difficult to determine whether incretins may help reduce autonomic neuropathy that contributes to gastroparesis as GLP-1 receptor activation in the stomach delays gastric emptying. However, long-term glucose lowering might protect against neuropathy-induced gastroparesis, as suggested by the findings of the DCCT/EPIC follow-up.

Finally, erectile function depends upon healthy vascular and nervous function, both of which may be impaired in diabetes and which incretin therapies may be able to modify. There are few data, although the REWIND trial of dulaglutide did find a small reduction in erectile dysfunction incidence compared with placebo. 

Finally, vascular damage may cause other microvascular pathologies, such as microvascular cardiac angina and cerebral small vessel disease. It may also reduce microvascular flow reserve in skeletal muscle. All three complications have been associated with type 2 diabetes. However, the impact of incretin therapies on these outcomes has not been well studied in humans so far.

In summary

Incretin therapies differ in their effect on microvascular disease by both drug type and outcome. There is now high-quality evidence showing that GLP-1RAs reduce the risk of adverse renal outcomes, while DPP-4is do not. It is not clear whether GLP-1RAs cause or worsen diabetic retinopathy, but the risk may be higher among those with background retinopathy or higher HbA1c, particularly with semaglutide. Adequately powered, long-term clinical trials are now needed to clarify the retinopathy risk. Clinical caution with DPP-4is and retinopathy is also advised. 

Meanwhile, there is little clinical evidence that GLP-1RAs reduce the incidence of neuropathic complications – namely peripheral neuropathy, CAN, gastroparesis and erectile dysfunction. Similarly, the relationship between DPP-4is and neuropathy remains unclear, though it is more promising for peripheral neuropathy. 

Going forward, all trials on incretin therapies should report the incidence of microvascular complications in order to enable meta-analysis and improve understanding. Finally, the mechanisms by which incretins exert their positive or negative effects on microvascular outcomes are likely to be multifactorial. Since they are poorly understood, this would also be a fruitful avenue for further research so that these therapies can achieve their full potential in preventing microvascular complications in type 2 diabetes. 

To read this paper, go to: Goldney J, Sargeant JA, Davies MJ. Incretins and microvascular complications of diabetes: neuropathy, nephropathy and microangiopathy. Diabetologia 19 August 2023.

To learn more, enrol on the EASD e-Learning course ‘GLP-1 receptor agonists’.

Any opinions expressed in this article are the responsibility of the EASD e-Learning Programme Director, Dr Eleanor D Kennedy.