IV GLP-1 does not reduce death, stroke, organ damage after heart surgery
Click Here to Manage Email Alerts
Key takeaways:
- Exenatide did not reduce death, stroke, HF or renal failure after cardiopulmonary bypass-assisted heart surgery vs. placebo.
- Restrictive and liberal oxygenation strategies conferred similar results.
CHICAGO — Exenatide, an IV GLP-1 receptor agonist, did not improve clinical outcomes compared with placebo in patients undergoing heart surgery, according to the results of the GLORIOUS trial.
The researchers also found no differences between a restrictive and a liberal oxygenation strategy in this population.
For GLORIOUS, presented at the American Heart Association Scientific Sessions, Sebastian Wiberg, MD, PhD, clinical associate professor of anesthesiology at The Heart Centre, Copenhagen University Hospital Rigshospitalet in Copenhagen, Denmark, and colleagues enrolled 1,389 patients (mean age, 68 years; 17% women; 15% with type 2 diabetes; < 2% with type 1 diabetes) requiring cardiopulmonary bypass (CPB)-assisted CABG and/or surgical aortic valve replacement. Patients were randomly assigned to an infusion of 17.4 µg exenatide (Byetta, Eli Lilly) or placebo after anesthesia but before surgery and to restrictive oxygenation (fraction of inspired oxygen = 50%) or liberal oxygenation (fraction of inspired oxygen = 100%).
“Application of CPB is known to cause organ injury, which is associated with morbidity and death,” Wiberg said during a presentation. “The etiology of this organ injury is rather complex, but involves multiple pathways, including cellular damage caused by oxidative stress and ischemia reperfusion injury caused by flow disturbances. In addition, we know that inflammatory cascades induced by mechanical stress, exposure of blood to foreign surfaces and blood-gas interfaces in the oxygenator are thought to be involved. Back in 2014, when this trial was designed, a number of animal studies had suggested GLP-1 agonists to ameliorate several neurodegenerative diseases. Furthermore, in animal models of stroke and myocardial infarcion, the GLP-1 agonists reduced final infarct size. In humans, administration of the GLP-1 analog agonist exenatide to patients with ST-elevation myocardial infarction had been shown to reduce final infarct size in a subset of patients with a short duration of ischemia, and importantly, it had been shown that exenatide was safe to administer to critically ill patients. Accordingly, we hypothesized that in patients undergoing CPB-assisted CABG and/or AVR, a perioperative 6 hour and 15 minute infusion of 17.4 µg of exenatide vs. placebo ... would reduce mortality and significant organ injury.”
He said if the trial were designed today, the researchers would have used a newer GLP-1 receptor agonist such as semaglutide (Ozempic/Wegovy, Novo Nordisk).
Exenatide vs. placebo
During the procedure, patients in the exenatide group had lower blood glucose levels, Wiberg said during the presentation.
At a median follow-up of 5.9 years, there was no difference between the exenatide and placebo groups in the composite primary endpoint of death, stroke, renal failure and HF (HR = 1; 95% CI, 0.83-1.3; P = .8), he said, noting that there were also no differences in any of the individual endpoints of the primary outcome.
There were no differences in adverse events, and rates of hypoglycemia within 12 hours of the procedure were extremely low (two patients in the exenatide group and none in the placebo group), Wiberg said.
In an analysis of prespecified subgroups, exenatide was favored in patients with known stroke, “but this result should be viewed as exploratory,” he said.
In a prespecified sensitivity analysis, there was no difference between the groups for the first occurrence of a primary endpoint event within 180 days (HR = 0.99; 95% CI, 0.72-1.4).
“In adult patients undergoing CPB-assisted elective or subacute CPB-assisted CABG and/or AVR, a perioperative infusion with the GLP-1 agonist exenatide did not reduce mortality or morbidity from renal failure, stroke or heart failure,” Wiberg said during the presentation.
Oxygenation strategies
In the oxygenation arm, partial pressure of oxygen was greater in the liberal-strategy group (P < .001), but there was no difference between the groups in the primary composite outcome (HR = 1; 95% CI, 0.81-1.3; P = .92), he said.
There was no interaction between the exenatide and oxygenation arms (P for interaction = .4), Wiberg said.
The GLORIOUS II trial of four interventions in patients with CPB-assisted CABG and/or AVR is ongoing and expected to be completed by 2027, he said.
In a discussant presentation, Sigrid Sandner, MD, MS, cardiac surgeon at Medical University of Vienna, said “the neutral results of GLORIOUS may reflect the challenges of defining appropriate outcomes and the optimal follow-up, rather than lack of a potential treatment effect.”
Perspective
Back to Top
I am not impressed by this trial. There is a common critical issue with both analyses, which is that when you design your trial, it’s very important that your primary outcome measure is as closely related as possible to your intervention. If you don’t do that, then what happens is the potential treatment effect of the intervention gets diluted, and even if there is one, you may not see it.
In this case, the researchers tested two interventions where, if they have an effect, they should have a very early effect, in the early postoperative period. Instead, they decided to observe patients for a median follow-up of 5 years. What happens when you do that is the natural history of the patient is mostly determined by their disease, not by an intervention that is supposed to work mostly in the perioperative period. I think that is why they found no difference.
The choice of the endpoint and the timing of it was not ideal. A more mechanistic endpoint, for example, for the GLP-1 outcome, something related to cardiac function and/or renal function rather than clinical events would have been more sensitive. There are surgical scores that measure postoperative end-organ function. I think that would have been much more meaningful than the major adverse CV events outcome the researchers used, which is mostly determined by the natural history of the disease. In the oxygenation factorial, there is actually a signal for a difference early on, but then the signal gets diluted, particularly for death and renal failure.
In general, when your trial design is not ideal, your result is neutral, because it produces a regression to the mean. I think this is what has happened here. I am not sure if the interventions would have worked or not, but if I had designed this trial, I would have used a more sensitive and specific mechanistic outcome rather than a MACE outcome, and then I would have tested it only very early. Other recent cardiac surgery trials testing perioperative interventions have done this. They all had a 30-day or in-hospital primary outcome. I think that is what is right for a perioperative intervention trial. We know that interventions that improve early postoperative outcomes don’t necessarily improve longer-term outcomes. Interventions that reduce postoperative renal failure, postoperative atrial fibrillation, postoperative mechanical ventilation, etc, have not been associated with differences in 5-year outcomes. A follow-up of 5 years dilutes the outcomes and opens the door for the baseline characteristics of the patients and the natural history of the disease to drive the result.
Collapse
Wiberg S, et al. LBS.04: From intervention to prevention: Advances in coronary and valvular heart disease. Presented at: American Heart Association Scientific Sessions; Nov. 16-18, 2024; Chicago.
Click Here to Manage Email Alerts