THIAMINE VERSUS PLACEBO IN OLDER HEART FAILURE PATIENTS: A PILOT RANDOMIZED CONTROLLED CROSSOVER TRIAL (THIAMINE-HF)

Thiamine supplementation is hypothesized to improve cardiac function in older adults with heart failure (HF). This pilot randomized trial was conducted to determine the feasibility of a larger trial and to explore the effects of thiamine on clinical outcomes.

We conducted a randomized placebo-controlled two-period crossover feasibility study of thiamine supplementation in patients with HF from June 2018 to April 2021. Inclusion criteria were adults age ≥60 years with symptomatic HF (New York Heart Association, NYHA, class II-IV), reduced LVEF (≤45%), and either a recent HF hospitalization in the past 12 months or NT-proBNP ≥600ng/L in the past 60 days. Participants were randomized to receive either thiamine mononitrate 500mg or placebo once daily for 90 days, followed by a 6-week washout period. Participants then switched to the opposite treatment (placebo or thiamine) for 90 days. The primary feasibility outcome was the recruitment of 24 participants in 11 months. Secondary clinical outcomes included change in peak global longitudinal strain (GLS), LVEF, NYHA functional class, NT-proBNP, and quality of life using the Kansas City Cardiomyopathy Questionnaire (KCCQ). We also collected information on rates of death, emergency department visits, and hospitalization. We screened 330 patients over 21 months to recruit 24 patients. Recruitment was challenging due to illness severity and comorbidities. The refusal rate was 37.0% (feasibility threshold < 40%), retention rate was 91.6% (feasibility threshold >80%) and study medication adherence rate was 95.0% (feasibility threshold >90%). Participants' mean age was 73.4 (standard deviation, SD, 7.4) and 7 were female (29.2%). Eighteen patients (75.0%) had NYHA class II symptoms and 6 (25.0%) had class III symptoms. All participants were on a beta-blocker, 19 (79.2%) were on a loop diuretic, 17 (70.8%) were on an angiotensin-blocking medication (ACE inhibitor, angiotensin-receptor blocker, or neprilysin inhibitor), and 12 (50.0%) were on aldosterone antagonist. The mean baseline LVEF was 33.1% (SD 10.5) and mean GLS was -8.1% (SD 3.1). Exploratory outcomes (Table 1) were not statistically significantly different between groups. There were four deaths in the study, all related to HF and unrelated to study medication. There were 13 serious adverse events in seven patients; none related to the study drug. Thiamine was well tolerated.

Although we were unable to recruit our target number of patients within the pre-specified time period, our other feasibility thresholds were met. It is feasible to conduct a larger trial of thiamine in heart failure if recruitment can be improved.