GLP-1 RAs vs SGLT2 inhibitors: What are the comparative risks of ischaemic stroke and incident AF?

A population-based retrospective study of Hong Kong patients with type 2 diabetes (T2D) finds that use of glucagon-like peptide-1 receptor agonists (GLP-1 RAs) is associated with a lower risk of ischaemic stroke, while use of sodium-glucose cotransporter 2 (SGLT2) inhibitors is associated with a lower risk of atrial fibrillation (AF).

“SGLT2 ihibitors and GLP-1 RAs have shown cardiorenal benefits beyond glycaemic control in landmark trials. However, head-to-head comparisons of SGLT2 inhibitors and GLP-1 RAs, especially with regards to stroke risk, are currently limited,” wrote the researchers from Li Ka Shing Faculty of Medicine, University of Hong Kong. [Lancet 2021;398:262-276] “To facilitate selection of appropriate anti-diabetic agents, we carried out a real-world population-based study of Chinese patients with T2D who were started on a SGLT2 inhibitor or a GLP-1 RA to compare their risks of different types of stroke and incident AF.”

The researchers analyzed electronic medical records of 5,840 propensity score–matched adult patients with T2D (mean age, 55.5 years; male, 56.1 percent) managed within Hong Kong Hospital Authority who received a SGLT2 inhibitor (n=2,920) or a GLP-1 RA (n=2,920) between 1 January 2008 and 31 December 2020. “The starting date was set as 1 January 2008, because the first prescription of a GLP-1 RA in Hong Kong was in 2008, while the first prescription of a SGLT2 inhibitor took place in 2015,” they explained. [Cardiovasc Diabetol 2023;doi:10.1186/s12933-023-01772-0]

The median durations of SGLT2i and GLP-1 RA cumulative use were 15.4 months and 10.1 months, respectively. With a median follow-up of 17 months, the incidence rate of composite stroke in SGLT2i and GLP-1 RA groups were 1.10 cases and 0.80 cases per 100 person-years, respectively.

While no significant difference in composite stroke was observed between the SGLT2 nhibitori and GLP-1 RA groups (hazard ratio [HR], 1.46; 95 percent confidence interval [CI], 0.99–2.17; p=0.058), there was a significantly higher risk of ischaemic stroke among SGLT2 inhibitor vs GLP-1 RA users (HR, 1.53; 95 percent CI, 1.01–2.33; p=0.044). “Our results suggest that GLP-1 RAs may be preferred over a SGLT2 inhibitors for T2D patients at risk of ischaemic stroke,” noted the researchers.

When analyzing all stroke, no significant interaction was observed between subgroups by age, sex, glycaemic control, or history of AF. However, patients with history of stroke appeared to benefit more from GLP-1 RA use (p_interaction=0.043) vs those without, as did those with shorter (<10 years) vs longer duration of diabetes (p_interaction=0.017). When focusing on ischaemic stroke, no interaction was observed across any subgroups, except for patients with shorter duration of diabetes (<10 years), who again appeared to benefit more from GLP-1 RA use than those with longer duration of diabetes (p_interaction=0.009).

The incidence rates of AF in the SGLT2 inhibitor and GLP-1 RA groups were 0.27 cases and 0.58 cases per 100 person-years, respectively. The risk of incident AF was significantly lower among SGLT2 inhibitor users (HR, 0.43; 95 percent CI, 0.23–0.79; p=0.006). Although AF is closely linked to cardioembolic stroke, the risk of cardioembolic or haemorrhagic strokes was comparable between SGLT2 inhibitor and GLP-1 RA users. [Cardiol Clin 2016;34:255-268]

“An earlier meta-analysis of six trials showed that SGLT2 inhibitor use was associated with a lower risk of cardioembolic stroke vs placebo. The lack of a statistically significant difference in cardioembolic stroke events between SGLT2 inhibitor and GLP-1 RA users may be due to the relatively small number of events during our study’s median follow-up of 17 months. Longer follow-up with larger cohorts may be necessary to clarify the comparative risk of cardioembolic stroke between SGLT2 inhibitor and GLP-1 RA users,” suggested the researchers. [Cardiovasc Diabetol 2021;20:100]