Treatment of HF in an Era of Multiple Therapies: Statement From the HF Collaboratory

Statement From the HF Collaboratory

Ankeet S. Bhatt, MD, MBA; William T. Abraham, MD; JoAnn Lindenfeld, MD; Michael Bristow, MD; Peter E. Carson, MD; G. Michael Felker, MD, MHS; Gregg C. Fonarow, MD; Stephen J. Greene, MD; Mitchell A. Psotka, MD, PHD; Scott D. Solomon, MD; Norman Stockbridge, MD, PHD; John R. Teerlink, MD; Muthiah Vaduganathan, MD, MPH; Janet Wittes, PHD; Mona Fiuzat, PHARMD; Christopher M. O'Connor, MD; Javed Butler, MD, MPH, MBA

Disclosures

JACC Heart Fail. 2021;9(1):1-12. 

In This Article

Potential Approaches to Guide use and Sequencing of Therapies

Several solutions and concepts with their advantages and disadvantages were proposed to integrate the growing body of trial data into routine clinical care.

Option 1: Trial Eligibility–Based Approach

One option is strict application of trial eligibility criteria to guide initiation of new therapies. Current trials have used varying left ventricular ejection fraction cutoffs (≤40% for SGLT2 inhibitors, ≤45% for vericiguat, and ≤35% for omecamtiv mecarbil). The trials also differ with respect to baseline natriuretic peptide levels, systolic blood pressure, estimated glomerular filtration rate, and worsening symptom status (Table 2). Leveraging these differences may allow niche indications for drugs with fewer proportions of overlapping patients. However, eligibility criteria are generally chosen for trial feasibility and event rate projection, rather than for biologic reasons. Given its complexity, such an approach would likely only be feasible in specialized HF clinics, and the changing evidence-based and trial nuances may be too burdensome for widespread use across primary care settings. Development of clinical decision support systems could be used to help select patients who meet trial enrollment criteria for a given agent; however, symptoms, blood pressure, renal function, natriuretic peptide levels, and other factors typically all vary considerably, making a patient potentially eligible at one time point and not at another if strict criteria are used. Most importantly, limiting patients to strict definitions of eligibility criteria, although "evidence based," is not grounded in biology and may deprive treatment to subsets of HF patients who could benefit. Given that nuances of trial design might influence ultimate regulatory indications, patients may further be deprived access to possibly beneficial therapies. Lastly, there will remain a large proportion of patients who will remain eligible for multiple agents.

Option 2: Comorbidity/Clinical Status–Based Approach

Certain therapies may provide greater potential to target comorbidities simultaneously. SGLT2 inhibitors might be preferable to other therapy in patients with type 2 diabetes, given the modest antihyperglycemic effect of these agents. Patientswith predilection for congestion or those with refractory hypertension may benefit from the use of ARNI[24] and SGLT2 inhibitors[25] given their natriuretic effects and improved outcomes in this high-risk cohort. Those with continued congestion or recurrent HF hospitalizations may further benefit from vericiguat or ivabradine. Patients with lower range ejection fraction and those limited by hypotension may preferentially benefit from omecamtiv mecarbil given its hemodyanamic properties. However, such an approach is complicated by the varying HF severity and distribution of comorbid conditions, and by the uncertain effects of conventional HF treatments on comorbidities. If drug therapy needs to be discontinued because of hypotension or other reasons, treating physicians may opt to discontinue the therapy with a lessermagnitude of benefit first. However, there is a lack of direct efficacy comparisons. Overall, such an approach may be important for prioritizing therapies, but simply focusing on therapies with comorbid condition indications may unnecessarily disenfranchise patients in whom the therapy was proven to be effective, despite not having a given specific comorbid condition or clinical scenario of dual benefit (e.g., the similar benefit of dapagliflozin in HFrEF patients with and without diabetes).

Option 3: Subgroup Analysis–Based Approach

Subgroup analyses, based on clinical factors or a risk-response score, may suggest that some patients might derive greater relative or absolute benefit from a specific treatment or sequence of treatments. Although subgroup analyses may identify possible groups with differential responses to a therapy, these groups should be biologically plausible. Ideally, when performing these analyses, the subgroup of interest should be prespecified and the trial large enough to have adequate power to test for interactions (i.e., to test whether the effect of treatment in a specific subgroup differs from the effect in others). Given the very large sample sizes required for meaningful subgroup analyses, a more practical approach might be to create statistical models that explore the relationship of predictor variables to treatment effect. In addition, statistical methods for adjusting for multiple testing should be used to control false-positive (i.e., type I error) rates. Valid analysis of subgroups does not require stratification of enrollment, but if such stratification is desired, the logistics of the trial become complicated. Informative subgroup analyses will require willingness from all sponsors to provide such analyses from their own data or to fund larger well-powered trials based on signals seen in subgroup analyses of predecessor trials. Subgroup analyses require very cautious interpretation because failure to show an effect within a given subgroup may simply reflect the inadequate sample size. For example, in the VICTORIA (Vericiguat Global Study in Subjects With Heart Failure With Reduced Ejection Fraction) trial evaluating vericiguat, patients in the highest quartile of N-terminal pro–B-type natriuretic peptide showed no benefit with vericiguat therapy, despite reductions in major endpoints across all other quartiles of N-terminal pro–B-type natriuretic peptide.[20] Establishing biomarker cutoff values to guide use in contemporary practice would likely be arbitrary and arduous, and it could lead to patients being eligible during one visit and ineligible at subsequent visits. Importantly, recent analyses with ARNI and dapagliflozin have not identified heterogeneity in treatment benefit among prespecified and post hoc subgroup analyses, suggesting a population-wide benefit of these agents for patients with HFrEF.[26–28]

Option 4: Biologic Improvement–Based Approach

An alternative option is to use biologic markers or surrogate outcomes to help guide treatment. For example, tumor markers and viral load are used to target incremental therapies or guide change in therapy in oncology and infectious diseases. Although the GUIDE-IT (Guiding Evidence-Based Therapy Using Biomarker-Intensified Treatment in Heart Failure) study[29] found no difference in cardiovascular death/HF hospitalization with a natriuretic peptide–driven approach to uptitration, both study arms showed similar reductions in natriuretic peptides. In fact, the GUIDE-IT trial suggests that clinical inertia and the guise of the stable HF patient may be important drivers for the lack of enhanced disease-modifying therapy use.[30] Meta-analyses have generally favored a biomarker-based approach;[31,32] monitoring of blood or imaging-based markers could be used to guide when more aggressive therapy is needed. However, this approach may potentially deprive patients who would have otherwise met inclusion criteria for a trial in which benefits were confirmed. In addition, trials have differed in baseline biomarker profiles, and the data are inconsistent regarding the prognostic implications or relative versus absolute changes in biomarker profiles. For example, despite modest reductions in N-terminal pro–B-type natriuretic peptide levels,[33] dapagliflozin has shown consistent, important improvements in clinical outcomes.[17] This approach would run the risk of only titrating therapies with more robust effects on surrogate measures and failing to escalate or de-escalate doses of therapies with little effect on surrogates (e.g., MRA, cardiac remodeling), despite demonstrated improvement in clinical outcomes. Ultimately, the lack of ideal surrogates makes such an approach more difficult in HF compared with other medical disciplines such as oncology or infectious disease.

processing....