Feasibility of Virtual Optimization of Guideline Directed Medical Therapy in Hospitalized Patients with HFrEF During the Covid-19 Pandemic: The IMPLEMENT-HF Pilot Study

INTRODUCTION: Implementation of GDMT for HFrEF remains low. We assessed the feasibility of a virtual GDMT Team for optimization of GDMT during hospitalization for non-CV conditions. HYPOTHESIS: A GDMT Team will improve GDMT optimization compared with usual care. METHODS: Consecutive hospitalized patients with HFrEF≤40% were prospectively identified. Patients with critical illness, cardiology consult, de-novo HF, COVID-19 & SBP ≤90mmHg were excluded. February 3 to March 1, 2020 served as a pre-intervention period during which patients were screened, but did not receive GDMT Team interventions. From March 2 to June 21, 2020, a pharmacist-physician team provided up to 1 suggestion daily for GDMT optimization (evidence-based ß-blockers, ACEi/ARB/ARNI, & MRA) to treating teams based on an evidence-based algorithm. The primary outcome of a composite GDMT optimization score, the net of positive therapeutic changes (+1 for new initiations/uptitrations) & negative therapeutic changes (-1 for discontinuations/downtitrations) during hospitalization, was compared between the pre- vs. post-intervention periods. Multivariable linear regression models were built adjusting associations for clinical factors. Safety outcomes requiring intervention or GDMT downtitration were identified. RESULTS: Of 187 encounters, 84 (45%) met eligibility criteria: 28 pre-intervention, 56 post-intervention. Mean age was 68±11 yrs, 70% men, and 61% White. Of 88 GDMT Team suggestions, 49 (56%) were followed by discharge. During the intervention, cumulative COVID-19 hospitalizations rose from 0 to 11085 in MA. Mean GDMT optimization score was -0.14 (95% CI: -0.58 to +0.30) pre-intervention & +0.64 (95% CI: +0.35 to +0.93) post-intervention (P=0.004). In a model inclusive of demographics, comorbidities, vital signs, potassium levels, eGFR, & LVEF, the intervention was the only factor associated with higher GDMT optimization score (β coeff 0.89; P=0.008). Safety events included 1 instance each of AKI, hyperkalemia, bradycardia, & hypotension. CONCLUSION: Admission for non-CV conditions is a feasible setting for GDMT optimization. A virtual GDMT Team was associated with improved GDMT; this implementation strategy warrants testing in a prospective RCT.