Limited comparisons of these two newer drug classes
“Real-world evidence comparing these two therapeutic classes based on CV outcomes is limited,” Poonawalla said at the start of her presentation, and relative treatment persistence, utilization, and cost data are even less well studied.
To investigate this, the researchers identified patients aged 19 to 89 years who were newly prescribed one of these two types of antidiabetic agents during January 1, 2015 through June 30, 2017.
Poonawalla and senior study author Phil Schwab, PhD, research lead, Humana Healthcare Research, Louisville, Kentucky, clarified the study design and findings in an email to this new organization.
The team matched 5507 patients initiated on a GLP-1 agonist with 5507 patients newly prescribed an SGLT2 inhibitor.
Patients were a mean age of 65 years and 53% were women.
More than a third (37%) had established ASCVD, including myocardial infarction (MI) (7.9%) and stroke (9.8%), and 11.5% had heart failure.
About two thirds were receiving metformin and about a third were receiving insulin.
In the GLP-1 agonist group, more than half of patients were prescribed liraglutide (57%), followed by dulaglutide (33%), exenatide, and lixisenatide (two patients).
In the SGLT2 inhibitor group, close to 70% received canagliflozin, about a quarter received empagliflozin, and the rest received dapagliflozin.
During up to 3.5 years of follow-up, a similar percentage of patients in each group had either an MI, stroke, or died (the primary composite CV outcome) (hazard ratio [HR], 0.98; 95% CI, 0.89 - 1.07).
However, more patients in the GLP-1 agonist group had heart failure or died (the secondary composite CV outcome), driven by a higher rate of heart failure in this group.
But after adjusting for time to events there was no significant between-group difference in the secondary composite CV outcome (HR, 1.09; 95% CI, 0.99 - 1.21).
During the 12-months after the initial prescription, patients who were started on a GLP-1 agonist versus an SGLT2 inhibitor had higher mean monthly medical costs, which included hospitalizations, emergency department (ED) visits, and outpatient visits ($904 vs $834; P < .001).
They also had higher pharmacy costs, which covered all drugs ($891 vs $783; P < .001).
And they were more likely to discontinue treatment (HR, 1.15; 95% CI, 1.10 - 1.21), be hospitalized (14.4% vs 11.9%; P < .001), or visit the ED (27.4% vs 23.5%; P < .001).
“Not too surprising” and “somewhat reassuring”
Overall, Kosiborod did not find the results surprising.
Given the sample size and follow-up time, event rates were probably quite low and insufficient to draw firm conclusions about the composite CV outcomes, he reiterated.
However, given the comparable effects of these two drug types on major adverse cardiac events (MACE) in similar patient populations with type 2 diabetes, it is not too surprising that there were no significant differences in these outcomes.
It was also “somewhat reassuring” to see that heart failure rates were lower with SGLT2 inhibitors, “as one would expect,” he said, because these agents “have been shown to significantly reduce the risk of hospitalization for heart failure in multiple outcome trials, whereas GLP-1 receptor agonists’ beneficial CV effects appear to be more limited to MACE reduction.”
The higher rates of discontinuation with GLP-1 receptor agonists “is also not a surprise, since patients experience more gastrointestinal tolerability issues with these agents (mainly nausea),” which can be mitigated in the majority of patients with appropriate education and close follow up — but is not done consistently.
Similarly, “the cost differences are also expected, since GLP-1 receptor agonists tend to be more expensive.”
On the other hand, the higher rates of hospitalizations with GLP-1 agonists compared to SGLT2 inhibitors “requires further exploration and confirmation,” Kosiborod said.
But he suspects this may be due to residual confounding, “since GLP-1 agonists are typically initiated later in the type 2 diabetes treatment algorithm,” so these patients could have lengthier, more difficult-to-manage type 2 diabetes with more comorbidities despite the propensity matching.
Poonawalla and Schwab are employed by Humana. Kosiborod has disclosed research support from AstraZeneca and Boehringer Ingelheim; honoraria from AstraZeneca, Boehringer Ingelheim, and Novo Nordisk; and consulting fees from Amarin, Amgen, Applied Therapeutics, AstraZeneca, Bayer, Boehringer Ingelheim, Eisai, GlaxoSmithKline, Glytec, Intarcia, Janssen, Merck, Novartis, Novo Nordisk, and Sanofi Aventis .
This article first appeared on Medscape.com