Abstract
Glucagon-like peptide-1 receptor agonists (GLP-1 RAs) were originally developed for the treatment of type 2 diabetes but have recently been approved for chronic weight management and reducing cardiovascular risk in individuals with overweight and obesity. Despite this approval, significant heterogeneity in the cardioprotective benefits and less desirable increases in resting heart rate (RHR) with GLP-1 RAs have been reported. To better understand cardiovascular responses to GLP-1 RAs and the potential role of health behaviors in influencing these responses, we leveraged wearable technology and causal inference analysis. We tracked RHR, heart rate variability (HRV), physical activity, and sleep in 66 individuals (42 ± 9 yr, body mass index: 30.0 ± 7 kg/m2) from the week before to 12 wk following the initiation of GLP-1 RA medication. Propensity score matching on a larger sample of wearable users identified a control group with similar anthropometric and cardiovascular characteristics (Ps > 0.26). After the 12-wk study period, GLP-1 users showed significant (Ps < 0.05) weight loss (−10.0%, 95% CI: −11.2% to −8.5%) and changes in RHR (3.2 ± 0.8 beats/min) that were mediated (P < 0.01) by changes in HRV (−6.2 ± 1.4 ms) compared with control. Trends (Ps < 0.10) suggested that increases in weekly physical activity were associated with GLP-1 RA medication (31.5 ± 13.2 min) and that higher physical activity levels accompanied an attenuation of RHR increases. Our real-world findings align with clinical trial data in showing rapid and significant weight loss with GLP-1 RAs, coinciding with increases in RHR that are mediated by changes in autonomic function (i.e., HRV). Physical activity may help to offset RHR increases, but further research is needed to confirm these effects.
NEW & NOTEWORTHY These findings are among the first to provide daily insights into cardiovascular and behavioral responses following GLP-1 RA initiation. Substantial weight loss and significant increases in resting heart rate mediated by reductions in heart rate variability during the initial 12 wk of GLP-1 RA therapy were observed. In addition, trends suggest an increase in physical activity with GLP-1 therapy, and that physical activity may help to temper GLP-1 RA-associated increases in resting heart rate.
INTRODUCTION
Glucagon-like peptide-1 receptor agonists (GLP-1 RA) were originally developed for the treatment of type 2 diabetes but have recently been approved for chronic weight management and reducing cardiovascular risk in individuals with overweight and obesity. GLP-1 RAs delay gastric emptying and promote feelings of satiety, leading to weight loss that surpasses what is achieved with lifestyle intervention alone or earlier weight loss medications (1, 2). Intriguingly, GLP-1 RAs provide cardiometabolic benefits beyond weight loss, such as reductions in blood pressure (3) that occur despite less desirable increases in resting heart rate (RHR) (4). However, significant heterogeneity in GLP-1 RA cardiovascular outcome trials has been observed (5).
Wearable technology offers an attractive approach to better understand cardiovascular responses to GLP-1 RAs by enabling continuous monitoring of cardiovascular health outside the clinic or laboratory environment. Moreover, wearable tracking of health behaviors, such as physical activity and sleep, may provide unique insights into moderators of these cardiovascular responses and inform strategies to mitigate RHR increases and optimize the cardioprotective benefits of GLP-1 RAs. Unlike tightly controlled clinical trials, this real-world approach captures data in everyday settings, offering a more practical understanding of interactions between GLP-1 RAs, cardiovascular health, and health behaviors.
In this real-world evidence study, we leveraged longitudinal measurements, wearable technology, and causal inference to examine changes in cardiovascular measures [RHR, heart rate variability (HRV), and respiratory rate (RR)] and health behaviors (physical activity and sleep) during the initial 12 wk of GLP-1 RA therapy. We hypothesized that GLP-1 RAs would increase RHR (4) and reduce HRV and RR (6, 7), with these effects on RHR and HRV attenuated by regular physical activity and healthy sleep patterns.
DISCUSSION
We leveraged longitudinal measurements, wearable technology, and causal inference to characterize cardiovascular and behavioral responses during the initial 12 wk of GLP-1 RA therapy. As hypothesized, GLP-1 RA therapy was associated with substantial weight loss, alongside significant increases in RHR mediated by reductions in HRV. Trends alluded to an increase in physical activity with GLP-1 RAs, which may play a role in attenuating the observed increase in RHR.
The observed ∼10% reduction in total body weight after only 12 wk of GLP-1 RA therapy in our real-world study matches, if not exceeds, the weight loss reported in large clinical trials of the latest GLP-1 RAs (i.e., ∼6%–9%) (1, 2). In addition, the proportion of our participants reporting a weight loss of ≥5% (i.e., 90%)—a clinically meaningful threshold (13)—after just 12 wk was comparable with what was achieved in these clinical trials after 12–18 mo (1, 2). These data underscore the profound weight loss efficacy of GLP-1 RAs in an everyday setting, with effects that appear amplified in our sample. The reasons behind this augmented weight loss are unclear; however, it is notable that Food and Drug Administration guidelines recommend that GLP-1 RAs be used alongside diet and exercise. Greater motivation, potentially associated with participants’ willingness to invest in a wearable technology subscription, may contribute to improved medication adherence and behavior modification (i.e., diet and exercise) that promotes enhanced weight loss. However, it is also possible that wearable technology itself may play a role in facilitating behavior change that supports GLP-1 RA weight loss. Notably, changes in physical activity during the study period were not associated with changes in body weight, as may have been anticipated based on literature questioning the sufficiency of exercise alone for weight loss (14).
The cardiovascular changes observed in this study, specifically the increase in RHR and decrease in HRV following GLP-1 RA initiation, are consistent with the documented chronotropic (4) and autonomic (6) effects of GLP-1 RAs. However, the mechanism(s) by which GLP-1 RAs increase RHR remains a topic of debate. A recently published study used in vivo and ex vivo methods to demonstrate a direct chronotropic effect of GLP-1 on the heart that is mediated by GLP-1 receptors in pacemaker cells of the sinus node (4). In addition, a systematic review and meta-analysis on the effects of GLP-1 RAs on autonomic function in participants with diabetes concluded that increases in RHR with GLP-1 RAs seem to occur without an alteration of sympatho-vagal balance (i.e., autonomic function) (15). However, as the included studies exclusively involved chronic GLP-1 RA users, the possibility remains that the autonomic alterations seen with GLP-1 RAs, such as the decrease in HRV observed in the present study, may be short-lived. Interestingly, RR remained unchanged in our study, contrary to our hypothesis that it might decrease due to reductions in body mass (7). Longer observation periods are needed for a more comprehensive understanding of associations between RR and GLP-1 RA use. Nonetheless, it is noteworthy that our mediation analysis showed that HRV changes accounted for nearly three-quarters of the total effect of GLP-1 RAs on RHR, alluding to the autonomic nervous system’s significant contribution to initial heart rate responses to GLP-1 RA medication. This finding aligns with prior evidence of autonomic involvement in RHR increases following acute GLP-1 RA administration in patients with type 2 diabetes (T2D) (16). Collectively, these findings indicate that the acute effects of GLP-1 RA therapy on RHR are likely at least partially mediated by alterations in autonomic function. The duration of this autonomic impact and its role in sustaining longer-term RHR increases observed in other studies (17, 18) is unclear, but worthy of future investigation.
Considering interactions between GLP-1 and dopamine signaling (19), it has been hypothesized that these medications may facilitate adoption and adherence to positive lifestyle changes, such as physical activity (20). Moreover, initial weight loss among individuals with overweight and obesity who begin GLP-1 RAs may alleviate discomfort associated with physical activity, thereby increasing the likelihood of exercise adoption and persistence. Consistent with this premise, trends in our study suggested an increase in physical activity among GLP-1 RA users, though statistical significance was not achieved. Similarly, a tendency toward attenuation of the RHR increase among GLP-1 RA users with higher versus lower levels of physical activity was observed. Although this result failed to meet the conventional significance threshold, the potential for even a modest level of physical activity (i.e., median split was at <10 min/day) to temper increases in RHR warrants consideration. Indeed, higher resting heart rate is a well-established independent risk factor for all-cause and cardiovascular mortality (21). Moreover, exercise participation is known to temper increases in RHR with advancing aging (22), providing evidence for the efficacy of physical activity to attenuate undesirable increases in RHR, albeit in a different context. Notably, post hoc sample size analysis for analysis of covariance with two predictors (baseline RHR and physical activity group) indicates that a total sample of 163 participants would be required to detect a statistically significant effect of the physical activity group on RHR, assuming a small-to-medium effect size observed in the current study (f = 0.22) with 80% power and an α = 0.05. In sum, our real-world data suggest that a portion of individuals beginning GLP-1 RAs seem to increase physical activity participation, and, in doing so, may offset the RHR elevations associated with this medication.
Our findings should be interpreted in the context of our study’s strengths and limitations. First, the study population was a convenience sample, which may have limited statistical power for certain analyses, such as those examining associations between GLP-1 RAs and behavior modification. Relatedly, though our sample consisted of a relatively balanced proportion of males and females (∼60% male and ∼40% female), we were unable to account for the potential influence of biological and sociocultural factors that may influence physiological and behavioral responses (23). Second, although participants self-reported GLP-1 RA use, we lacked specific detail on medication type (e.g., semaglutide and tirzepatide) and dosing strategies. Outside of overweight and obesity, GLP-1 RA medications are also prescribed for T2D and even off-label for cosmetic weight loss (24), but we lack a detailed medical history of our sample. Third, although participants wore the device almost daily and weekday versus weekend wear time was relatively similar between groups, our study design did not account for potential differences in cardiovascular metrics and health behaviors between weekdays and weekends (25). Last, body weight was self-reported among GLP-1 users and unavailable in our comparative group. In addition, over a third of participants in the GLP-1 group did not report their weight at week 12, possibly biasing our analysis. Objectively measured body weight and composition values in both cohorts would have enhanced our analysis. Furthermore, while we performed mediation analysis using HRV as a mediator and RHR as an outcome, we could not account for their bidirectional influence, limiting causal inference. Notable strengths of this study include our use of propensity matching for causal inference on real-world data, being among the first to report serial (i.e., daily) measurements of cardiovascular measures and health behaviors among GLP-1 RA users, and our use of wearable technology that has been validated against gold-standard electrocardiogram and polysomnography measures with a low degree of bias and low precision of errors (10, 11).
In conclusion, we demonstrate substantial weight loss and significant increases in RHR mediated by reductions in HRV during the initial 12 wk of GLP-1 RA therapy. In addition, we observed trends portending to an increase in physical activity with GLP-1 therapy in conjunction with evidence that physical activity may help to temper GLP-1-associated RHR increases. These findings are among the first to provide daily insights into cardiovascular and behavioral responses following GLP-1 initiation, offering valuable data for understanding the real-world impact of this medication.