Effect of GLP-1 Agonist Initiation On Remotely Monitored Pulmonary Arterial Pressures in Patients with Heart Failure

Glucagon-like-peptide-1 receptor agonists (GLP-1 RA), initially developed to treat type 2 diabetes mellitus, have significantly expanded their approval in recent years. Since 2014, GLP-1 RAs have been approved for weight loss and, more recently, have been shown to reduce the risk of major adverse cardiovascular events (MACE) in patients with established cardiovascular disease.1 There has been a strong link between obesity and the development of heart failure with preserved ejection fraction (HFpEF), with improved outcomes in patients with HFpEF randomized to semaglutide or tirzepatide versus placebo.1-4 A recent study by Jiang et al. revealed improved hemodynamics in patients with HF treated with GLP-1 RAs. Their single-center retrospective study showed significant reductions in pulmonary artery pressures (PAP) in patients with HF remotely monitored with CardioMEMS devices (Abbott, Abbott Park, IL) along with a modest 5% reduction in weight.5 We aimed to expand on this limited, single-center analysis through the addition of a patient cohort at our institution. In this retrospective, single-center study, we identified patients followed at our institution with a CardioMEMS device in place for remote hemodynamic monitoring who had been on at least 6 months of maximally tolerated GLP-1 RA therapy. This study was approved by the institutional review board. Patients who were initiated on GLP-1 RA therapy prior to the placement of CardioMEMS were excluded. Baseline characteristics as well as systolic PAP, diastolic PAP and mean PAP were collected in the week preceding GLP-1 RA initiation and after 6 months on each patient's maximum dose of therapy. Guideline-directed medical therapy (GDMT) and loop diuretic dosages (measured as furosemide equivalents) were also assessed in this time frame. To maintain consistency across data sets, we modeled our analysis after that of Jiang et al.5 GDMT and loop diuretic utilization were assessed as the proportion of patients on therapy and percentage at target dosing at baseline and 6 months with ANOVA analysis. Systolic, diastolic and mean PAP changes were assessed utilizing Wilcoxon rank-sum testing. A total of 12 patients were included in this study: 92% women with a mean age at the start of GLP-1 RA therapy of 61 (39–78), 83% Black or African American, and a mean BMI of 39.8 ± 4.8 kg/m2. The mean BMI after 6 months of therapy was 38.3 kg/m2 (P = 0.5). Ejection fraction (EF) was >40% in 75% of this cohort, with an average EF of 50 ± 18%. Full baseline characteristics can be seen in Table 1. A total of six patients (50%) received tirzepatide with a median dose of 10 mg/week (range: 2.5–15 mg/week), four patients (33%) received semaglutide with a median dose of 0.625 mg/week (range: 0.25–2 mg/week), and two patients (17%) received liraglutide 1.8 mg/week. There were no significant differences in GDMT or loop diuretic therapy at 0 and 6 months. There were no significant differences in systolic, diastolic or mean PAPs after 6 months of maximum-tolerated dosage of GLP-1 RA therapy (Figure 1). However, there was a numeric decrease in mean (28.7 to 26.7 mmHg, P = 0.93) and diastolic (19.2 to 16.2 mmHg, P = 0.58) PAPs from baseline to 6 months. Given wide ranges of dosing and potential discrepancies between GLP-1 RA therapies, sensitivity analyses were conducted. First, to compare patients on tirzepatide to those on semaglutide and liraglutide, and second, to compare patients on >50% of goal dose (tirzepatide 15 mg, liraglutide 3 mg, semaglutide 2–2.4 mg based on formulation) versus <50% across all GLP-1 RA therapies. Notably, there was a significant reduction in BMI (40 to 37.6 kg/m2, P = 0.01) for the tirzepatide-only sub-group, but otherwise, there were no significant differences in PAPs in these sensitivity analyses. In this single center, retrospective study of patients with active remote PAP monitoring receiving GLP-1 RA therapy, we found no significant differences in PAP reduction over 6 months of therapy, though a numerical reduction in diastolic and mean PAPs was noted. These results are in contrast to a recent, single center retrospective study of similar size (n = 9) in which significant reductions in PAP were noted with GLP-1 RA therapy. We suspect that the discrepancy between these studies reflects a difference in certain key baseline patient characteristics, in addition to the small sample sizes of both studies. Most notably, our patient population was predominantly HF with moderately reduced ejection fraction (HFmrEF) and HFpEF, with 75% of our patients having an EF > 40% and a mean EF of 50%. In contrast, the study by Jiang et al. had 56% of patients with EF > 40%, and the mean EF was 37%. Additionally, our patient population had no significant change in BMI or weight during the period assessed, though we did see a numerical downward trend of mean BMI from 39.8 to 38.3 kg/m2. This is in contrast to Jiang et al., which showed a significant reduction in both BMI and weight over the 6 months of therapy (BMI 41.4 to 39.2 kg/m2, Weight 123.6 to 117.2 kg). In comparison to Jiang et al., the population in this analysis had a higher mean baseline NT-pro brain natriuretic peptide (2,368 vs 783 pg/mL) and lower mean creatinine (1.58 vs 1.80 mg/dL). Notably, similar to Jiang et al., there was no significant difference in GDMT or loop diuretic therapy over the 6 months assessed, providing reassurance that the PAP assessment is representative of the change in GLP-1 RA medication.5 Our findings are, at least directionally, expected given the current GLP-1 RA landscape. In particular, following the STEP-HFpEF trial showing a greater symptom reduction, weight loss and exercise capacity in HFpEF patients on semaglutide, as well as SUMMIT with similar results in those on tirzepatide, we would anticipate a greater impact of these medications on our HFpEF predominant population.1, 4 Though even outside of weight loss, the impact of GLP-1 RAs is present regardless of EF, with reduced HF hospitalizations and cardiovascular death with GLP-1 RA utilization.6 As we continue to see research showing cardiovascular outcome benefits of GLP-1 RA medications in patients with HF, and particularly those with HFmrEF and HFpEF, it is important to understand the hemodynamic impact of these medications. This would advance our understanding of their mechanism of action and efficacy in this population. Our outcome discrepancy to a recent similarly sized, retrospective study comes in the setting of important population differences whose impact is difficult to quantify with such small sample sizes. This limitation could be addressed with expansion to multiple centers and larger numbers of patients included. As more GLP-1 RAs are prescribed to patients with HF, it will be key to leverage remote PAP monitoring data to assess their related treatment effects.