Advanced heart failure (HF) is a multifaceted disease process that affects multiple organ systems. Often, patients with HF have concomitant glucose intolerance manifested as diabetes mellitus in conjunction with chronic renal insufficiency, associated with their diabetes and underlying renal disease processes. Compared to non-diabetic patients, those with type 2 diabetes (T2D) have an increased risk of developing HF. HF is the leading cause of hospitalizations in T2D-related cardiovascular disease (1). There remains an increased risk for morbidity and mortality in patients with HF despite the use of optimal medical therapy (OMT) such as beta-blockade, renin angiotensin aldosterone system (RAAS) inhibition, angiotensin-converting enzyme inhibitors, neprilysin inhibitors or angiotensin receptor blockers (2).

Outcomes with the addition of sodium-glucose cotransporter 2 inhibitors (SGLT2i) in the management of both systolic and diastolic HF patients have demonstrated decreased cardiovascular mortality and readmission rate (3). We felt it important to understand the effect on patients with advanced HF with or without T2D and determine if there was a change in functional capacity after SGLT2i initiation compared to those on OMT alone.

In order to create a streamlined understanding of the testing we evaluated, we provide a detailed background below that highlights the role of cardiopulmonary function testing and SGLT2i in the advanced heart failure population.

After IRB approval, we performed a retrospective review of patient records between August 2020 and August 2021. All patient data was extracted from the electronic medical record. We included all patients with a confirmed diagnosis by ICD-10 code of either systolic (I50.2), diastolic (I50.3), or combined systolic and diastolic heart failure (I50.42). We then verified that a CPET result was available at baseline and one year after either the initiation of SGLT2i or maximally tolerated OMT. Our institution's advanced heart failure team judged OMT tolerance based on provider documentation in the electronic medical record. 13,213 patients were found to be on SGLT2i during our review period. Of these, 2,063 fulfilled one of the three ICD-10 codes. Out of these, 94 patients were identified with a baseline and follow-up CPET within one year of optimization of medical therapy or initiation of SGLT2i therapy.

Our control group, the OMT group, were the patients found to be taking guideline directed OMT (20). Our study group, OMT+SGLT2i group, had an additional SGLT2i (dapagliflozin) in their regimen. Of the 94 patients with CPET results available, 50 patients with OMT while 44 had OMT with the addition of an SGLT2i (dapagliflozin).

The primary endpoint was the effect of SGLT2i on Peak VO₂ recorded during CPET. We also report outcomes in terms of progression of therapy to LVAD/transplant and mortality. We report demographics, baseline medical therapies, follow-up duration, and relevant patient-specific parameters such as body mass index, renal function, and HbA1c levels.

We report median values and interquartile range for these parameters. Where appropriate, descriptive, and comparative statistical analyses were performed using the Mann-Whitney or chi-square analysis to determine statistical significance. IBM SPSS version 27 was used for statistical analysis. P value of ≤0.05 was considered statistically significant for all outcomes.

94 patients were identified with a baseline and follow-up CPET within one year of optimization of medical therapy or initiation of SGLT2i therapy. 50 patients undergoing OMT while 44 had OMT with the addition of an SGLT2i (dapagliflozin). Baseline patient characteristics, relevant parameters and outcomes are outlined in Table 1. CPET results are reported in Table 2.

This is one of the first manuscripts to propose that OMT+SGLT2i use alone does not improve functional capacity in advanced heart failure patients (NYHA 3+) as measured by peak VO₂ during CPET. In this retrospective study, we evaluated all patients with heart failure (preserved or reduced ejection fraction) undergoing serial CPET over 12 months. After initiation of SGLT2i or while on maximally tolerated Optimal Medical Therapy (OMT), pre-CPET and post-CPET were compared within each group as well (Figure 3). For those patients receiving OMT alone 12(24%) patients received heart transplantation or durable left ventricular assist device (LVAD), and 3(6%) died. In the OMT+SGLT2i group, 8(18%) patients underwent heart transplantation or durable LVAD placement and 1(3%) died.

As we begin to understand the effects of novel therapies crossing disease borders and impacting the entire patient, we have learned that the pluripotent effects of medications in the modern era have unforeseen benefits in some cases. Reduction of the heart failure burden in the context of readmissions and improvement in functional capacity is important, however, to better understand the trajectory of heart failure and therapies such as transplant or LVAD, we need to better define risk categories for those patients with the highest risk, but potentially the greatest reward. As we look to traditional literature from Mancini, Packer, Weber, and others who created the platforms for modern medical optimization, it is imperative to not only place patients in the right risk group but work to optimize the individual potential of each medicine (3, 5).

As a result of such innovation, we present our data from a large transplant and advanced heart failure center that helps answer the next question, beyond physical improvements and readmissions, but that of true transformation of patient outcomes in those with advanced HF.

In our reviewed literature search, the current major clinical trials studying the effects of SGLT2i on heart failure patients fail to include the utilization or assessment of CPET and peak VO₂ within the framework of their outcomes, with most favoring the 6-minute walk (6MW). For trials such as EMPA-REG, EMPEROR Preserved, CANVAS and DAPA-HF, CPET testing was not available (17, 18, 22, 23). For papers like EMPA-Tropism (2020), for example, CPET was considered novel (10). However, this was a placebo-controlled study and therefore did not compare each participant to OMT or to their own baseline. The suggestion proposed based on this trial is that the SGLT2i improves the functional testing, however, this clear correlation cannot be taken at face value based on their data. Volume optimization, an effect well understood due to osmotic diuresis in SGLT2i and the anti-inflammatory effect of less myocardial stress, is known to benefit myocardial contraction. As such, optimization of patients with salt restriction, fluid and diuretic therapy can also conceivably show similar results. In fact, He and colleagues in 2021 showed that aggressive diuresis can augment the response of patients with both HF reduced or preserved ejection fraction (24). Although the study was terminated prematurely due to new guideline recommendations regarding the use of SGLT2 inhibitors, the CANA-HF study (2020) reported similar findings to ours in a 12-week study that canagliflozin did not improve the peak VO₂ and VE/VCO₂ slope (25).

The latest 2021 ACC guidelines for the treatment of advanced heart failure saw the addition of new medications representing a significant advancement in the field, particularly SGLT2i and ARNI (20). Currently, the only approved ARNI in use is sacubitril/valsartan (S/A). ARNI and other medications affecting the RAAS system demonstrate favorable effects in the pathophysiology and symptomatology of heart failure and could potentially lead to improvements in peak VO₂ (26, 27). In a prospective study with 37 patients, Cacciatore et al. showed increased VO2 max and improved 6MWT at 12 months follow up and compared to their baseline (28). Meanwhile, the PARALLAX trial showed no significant impact on exercise tolerance as shown by no improvement in the 6MW test at 24 weeks with ARNI use (29). Santos et al. used 6MW test and CPET to assess functional outcomes with ARNI therapy. The study reported no significant improvement in in 6MW or the peak VO₂ when compared to their enalapril control group (30).

While parameters such as BMI may affect the peak VO₂, studies have shown inconsistent results. Some studies show a negative correlation between BMI and peak VO₂, attributing to the to the increase in type 2 muscle fibers and a decrease in type 1 muscle fibers in obese individuals and increased fat mass which may reduce oxygen uptake (31). Other studies demonstrate that BMI or fat percentage do not affect the peak VO₂ but rather the exercise capacity of individuals (32). This may also be attributed to the lack of BMI to differentiate between fat and muscle mass. Diabetes is another such comorbidity known to affect CPET performance, although the evidence is inconclusive. One study found no significant difference in peak VO2 in diabetics when compared to well matched control group without diabetes (33). While the OMT+SGLT2i group in our study shows higher HbA1c levels and is comprised of 45% diabetics, there was no difference by true value (raw VO2) or in assessing a statistical difference in the baseline peak VO2 in the OMT or the OMT+SGLT2i group.

An observation worth noting is that in our population, we found a trend that showed a patient with higher baseline peak VO₂ may have a significant drop but remain above the cutoff for advanced HF evaluation. This canary in the coal mine perhaps can, with larger data and multi-center support, help us understand the true delta needed to be concerned with poor response to OMT earlier, and consider interventions – exercise programs, medical therapy, dietary/lifestyle changes, or advanced HF workup – to prevent acute decompensation and urgent need for transplantation or inotrope/LVAD therapy as salvage.

Our data, despite being single center, does highlight a potential area for further opportunity. Understanding the profiles of our patients and stratifying low risk or high risk to progress is increasingly important in our field, underscoring the need that benefits of performing CPET may not be only for patients undergoing transplant assessments.

Our study is thought provoking and hypothesis generating, as are all retrospective assessments. The main limitation is our sample size; however, we remain the largest single center study and at present the only to compare VO₂ data in the same patient before and after SGLT2i therapy, not only to the OMT or placebo group. As a retrospective study, we were not able to include KCCQ12, 6MW or NT-Pro-BNP assessment, which may provide further details into quality of life in the setting of unchanged CPET studies.

Our single-center data shows that functional testing in advanced HF patients is not significantly impacted by SGLT2i as assessed by the peak VO₂ through serial CPET testing. Given the pluripotent nature of the effects of SGLT2i, volume status and reduction in intracardiac filling pressures may improve symptoms and reduce hospitalizations but this does not impact long term likelihood of needing an advanced therapy evaluation. Further studies are needed to fully evaluate the effects of SGLT2i in HF particularly with respect to the standardization of study endpoints and the inclusion of late-stage HF (NYHA class 3+) patients.