Digitoxin for heart failure with reduced ejection fraction?

A recent study found that patients taking digitoxin who had heart failure with reduced ejection fraction had a decrease in the composite of death from any cause or hospital admission for worsening heart failure, in the DIGIT-HF study (which apparently is short for the Digitoxin to Improve Outcomes in Patients with Advanced Chronic Heart Failure trial): see chf HFrEF digitoxin helps NEJM2025 in dropbox, or DOI: 10.1056/NEJMoa2415471)

 

Details

-- 1212 patients in this double-blind, placebo-controlled trial from 65 sites in Austria, Germany, and Serbia

-- patients with chronic heart failure who had a left ventricular ejection fraction of 40% or less and a New York Heart Association (NYHA) functional class of III or IV, or a left ventricular ejection fraction of 30% or less and an NYHA functional class of II, were randomized to receive digitoxin or matching placebo in addition to guideline-directed medical therapy (all patients had received evidence-based therapy for heart failure for a period of at least 6 months)

-- baseline characteristics:

    -- age 66, 20% female, BMI 29

    -- NYHA functional class, in both the digitoxin and placebo groups:

        -- class II in 180 individuals in (30%)

        -- class III in 400 individuals (67%)

        -- class IV in 23 individuals (4%)

    -- left ventricular ejection fraction (LVEF): 29%; LVEF <30 in 49%

    -- main cause of heart failure: ischemic in 53%, non-ischemic or unknown in 47%

    -- heart rate 74, systolic blood pressure 121, atrial fibrillation 27%

    -- eGFR (creatinine-based) mean 65, <60 in 43%

    -- device therapy: implantable cardioverter-defibrillator therapy in 65%, cardiac re-synchronization therapy in 25%

    -- heart failure medications at baseline (though this might have changed over time as newer medications became available):

        -- beta-blocker: 96%

        -- ACE-inhibitor: 36%

        -- angiotensin receptor blocker: 18%

        -- angiotensin receptor–neprilysin inhibitor: 40%

        -- mineralocorticoid receptor antagonist: 76%

        -- SGLT-2: 20%

        -- cardiac glycoside: 3 individuals in those assigned to digitoxin, 6 if on placebo

-- those in the digitoxin group started at a dose of 0.07 mg daily versus placebo

    -- dose adjustments were made if outside the predefined target range of 8-18 ng/mL (10.5-23.6 nmol/L), with a decrease to 0.05mg daily if too high and an increase to 0.1 mg if too low (those on placebo had random changes their drug levels to mask that they were in the placebo group)

-- primary outcome: a composite of death from any cause or hospital admission for worsening heart failure, whichever occurred first.

-- key secondary outcomes: testing for non-inferiority for digitoxin, as well as a composite of death from any cause, any hospitalizations due to heart failure, death from cardiovascular causes, death from heart failure, sudden death from cardiac causes, hospitalization due to cardiovascular causes, hospitalization due to noncardiovascular cause, hospitalization due to any cause, and a composite of death from cardiovascular causes or first hospitalization for heart failure

-- safety outcomes: serum digitoxin concentrations, adverse and serious adverse events, and serious adverse events that appeared to be unexpected adverse reactions to digitoxin or placebo

-- an intention-to-treat analysis was performed excluding patients who did not take any of the medications

-- median follow-up 36 months with a range of 0-110 months

-- mean duration of treatment: 0-107 months

 

Results:

-- 1212 individuals fulfilled the criteria for inclusion in the modified intention-to-treat population: 613 patients in the digitoxin group and 599 in the placebo group.

 

-- primary outcome (a composite of death from any cause or hospital admission for worsening heart failure, whichever occurred first):

    -- digitoxin group: 242 patients (39.5%); 12.8 events per 100 patient-years

    -- placebo group: 264 patients (44.1%); 15.7 events per 100 patient-years

        -- 18% improvement with digitoxin, hazard ratio (HR) 0.82 (0.69 to 0.98), P = 0.03

        -- this difference translates to the number-needed-to-treat to prevent one primary outcome: 22

-- death from any cause:

    -- digitoxin group: 167 patients (27.2%), 7.8 events per 100 patient-years

    -- placebo group: 177 patients (29.5%), 8.9 events per 100 patient-years

        -- 14% better with digitoxin, HR 0.86 (0.69 to 1.07), which exceeded their threshold for noninferiority, p<0.001

-- first hospital admission for worsening heart failure

        -- digitoxin group: 172 patients (28.1%), 9.1 events per 100 patient-years

        -- placebo group: 182 patients(30.4%), 10.8 events per 100 patient-years

            -- 15% better with digitoxin, HR 0.85 (0.69 to 1.05), a strong trend to being statistically significant

subgroup analyses:

-- males did 16% better with digitoxin, HR 0.84 (0.69-1.02), very close to being statistically significant

-- those with NYHA class III or IV did 17% better with digitoxin, HR 0.83 (0.68-1.02), very close to being statistically significant

-- atrial fibrillation when present did 28% better, HR 0.72 (0.50-1.03), very close to being statistically significant

-- less than 70 years old did 23% better, HR 0.77 (0.60-1.00)

-- left ventricular ejection fraction less than 30% did 23% better, HR 0.77 (0.59-0.99)

-- nonischemic heart failure did 30% better, HR 0.70 (0.53-0.93)

-- heart rate greater than 75 bpm did 37% better, HR 0.63 (0.48-0.83)

-- systolic blood pressure less than 120 mmHg did 39% better, HR 0.61 (0.48-0.79)

-- BMI<30 did 26% better, HR 0.74 (0.58-0.93)

-- those without hypertension did 43% better, HR 0.57 (0.34-0.93)

-- those without diabetes did 24% better, HR 0.76 (0.60-0.96)

-- those without  implantable cardioverter-defibrillator did 43% better, HR 0.57 (0.41-0.80)

-- those without cardiac-resynchronization therapy did 28% better, HR 0.72 (0.58-0.90)

-- those not on an SGLT2 inhibitor did 41% better, HR 0.59 (0.34-1.04), very close to being statistically significant

-- those on triple heart failure therapy did 19% better, HR 0.81 (0.65-1.01), very close to being statistically significant

    -- overall, there was no relationship with being on or off mineralocorticoid receptor antagonists or angiotensin receptor–neprilysin therapies

-- other secondary outcomes:

    -- none reached statistical significance, though all were skewed to a strong trend to being significantly better with digitoxin except for sudden death from cardiac causes, which had very few people (only 12 in each group)

-- Safety

    -- serum digitoxin levels at 6 weeks 17.0+/- 5.9 ng/ml in 550 patients, target range of 8-18 ng/mL (ie, near to the top of the targeted range)

        -- at 12 months (assessed for safety reasons): 13.5+/-5.1 ng/mL

    --serious adverse events:

        -- digitoxin group: at least 1 in 29 patients (4.7%)

        -- placebo group: at least 1 in 17 patients (2.8%)

            -- these events included cardiac disorders in 3.4% on digitoxin and 1.8% in placebo

                -- the major ones (buried in the supplement) were ventricular fibrillation in 10 patients (1.6%) on digitoxin and 3 (0.5%) on placebo and ventricular tachycardia in 7 (1.1%) vs 3 (0.5%) respectively. both not reaching statistical significance

            -- adverse events leading to discontinuation of drugs: 9.1% on digitoxin vs 10.2% on placebo

        -- total numbers of adverse events:

            -- total: 369 (60.2%) on digitoxin and 599 (56.9%) on placebo

                 -- for patients with the largest numbers of events:

                    -- anemia: 8 (1.3%) on digitoxin vs 3 (0.5%) on placebo

                    -- angina: 12 (2.0%) vs 9 (1.5%)

                    -- arrhythmia: 6 (1.0%) vs 2 (0.3%)

                    -- atrial fibrillation: 33 (5.4%) vs 31 (5.2%)

                    -- bradycardia: 10 (1.6%) vs 3 (0.5%)

                    -- cardiac failure: 29 (4.7%) vs 26 (3.3%)

                    -- ventricular tachycardia: 36 (5.9%) vs 38 (6.3%)

                    -- vertigo: 12 (2.0%) vs 8 (1.3%)

                    -- dizziness: 47 (7.7%) vs 39 (6.5%)

                    -- syncope: 22 (3.6%) vs 13 (2.2%)

                    -- diarrhea: 37 (6.9%) vs 18 (3.0%)

                    -- nausea: 29 (4.7%) vs 22 (3.7%)

                    -- fatigue: 21 (3.4%) vs 12 (2.0%)

                    -- peripheral edema: 16 (2.6%) vs 8 (1.3%)

                    -- hypotension: 18 (2.9%) vs 10 (1.7%)

     

Commentary:

-- there does seem to be an interesting cycle in clinical medicine, with some old studies and treatments being revitalized:

    -- this current study is opening the door for much more use of cardiac glycosides (digitoxin in this case) after these were largely dismissed, promoting a potential recurrence of cardiac glycoside use

    -- lipoprotein (a) has had evident importance in studies performed 50 to 60 years ago, and now sees its importance emerging again: https://gmodestmedblogs.blogspot.com/2025/06/high-lpa-increases-risk-of-recurrent.html

    -- apolipoprotein B has also had long-standing relevance, finding it to have a higher correlation with atherosclerotic events than LDL levels; apoB has a strong association with oxidized LDL, which plays a major role in the initiation of the atherosclerotic process: https://gmodestmedblogs.blogspot.com/2023/05/cad-presumed-mechanism-and-apob-was.html

    -- cystatin-C has been a better test for renal function than creatinine, also known for 40 to 50 years, with added justification more recently: https://gmodestmedblogs.blogspot.com/2023/12/cystatin-c-better-predictor-of-bad.html

        -- this study, as well as most in the medical literature, still use the clearly inferior creatinine-based estimated GFR (eGFR)...

 

-- This digitoxin study does follow a 1997 study on digoxin (DIG trial: Effect of Digoxin on Mortality and Morbidity in Patients with Heart Failure; see https://www.nejm.org/doi/full/10.1056/NEJM199702203360801)

    -- the DIG trial enrolled patients with heart failure and reduced ejection fraction, finding that the primary outcome of death from any cause, a typical primary outcome in studies of that time, was not superior to placebo; this study led to dramatic decreases in digoxin use for heart failure (this was also corroborated indirectly by several studies on new positive inotropic agents more recently that were associated with worse clinical outcomes)

    -- of note, in a secondary outcome that was not widely publicized, digoxin did show a lower incidence of hospitalizations for worsening heart failure

    -- and, patients with dramatically reduced ejection fractions of less than 25% or with advanced heart symptoms (NYHA functional class III or IV) seemed to have received some benefit from digoxin, and those with lower serum digoxin concentrations in the 0.50-0.09 nanograms/mL had better outcomes vs those with concentrations >1.0 ng/mL who had worse outcomes

        -- there were studies at that time finding that digoxin improved exercise capacity and quality of life, and a study showing that taking people off their digoxin had worsening heart failure symptoms and more hospitalizations, though digoxin use still remained low

-- both digoxin and digitoxin are similar cardiac glycosides, although digitoxin is more lipophilic, has higher levels of intestinal absorption, and the serum protein binding is higher. In addition, digitoxin, unlike digoxin, is effectively eliminated by entero-hepatic excretion when renal function is markedly impaired; digitoxin (vs digoxin) therefore does not need dose reductions if there is renal insufficiency (i.e. patients can remain on stable digitoxin doses without dose adjustments if they have progressive renal dysfunction)

   -- the digoxin study was much larger than the current digitoxin one with 6710 participants, increasing the chance for statistically significant outcomes for digoxin (very likely that many of the subgroup analyses that were close to being statistically significant for digitoxin in the Results section above would have been statistically significant if there were 5-fold the number of participants as in the DIG study)

   --  this digoxin study was conducted at a time prior to studies on heart failure that confirmed the benefit of beta blockers, mineralocorticoid inhibitors, SGLT-2 inhibitors and cardiac device therapies. it was notable in the digitoxin study above that there was benefit in patients on many of the newer drugs being used 

-- this trial with digitoxin, with patients having much more advanced symptomatic heart failure than in the DIG study, found a few things:

    -- the combo of death from any cause and or worsening of heart failure was significantly lower in those on digitoxin

    -- unlike DIG, the benefit of digitoxin was for both males and females (though more clearly so for males, though likely related to the low number of women in this study); the DIG study had increased mortality for women at higher doses of digoxin

    -- digitoxin was associated with fewer hospitalizations for worsening heart failure, as was the DIG trial with digoxin

    -- the results of this digitoxin study was similar to several other studies in terms of the number of people who need to be treated to avoid one primary-outcome event: there were similar numbers to the findings in PARADIGM-HF and the EMPEROR-Reduced trials

    -- there were many adverse events with digitoxin, as noted above. many of these are not so dissimilar to those in the placebo group. several of them, however, are related to the mechanism of action of digitoxin: vagal nerve/parasympathetic tone increase leading to decreased heart rate, prolonged SA and AV node conduction, and lowering blood pressure by enhancing baroreceptors

    -- this current digitoxin trial enrolled extremely sick patients, as opposed to the DIG trial, which brings up a few issues:

        -- these were quite sick heart failure patients (>70% had NYHA class III or IV symptomatic heart failure)

        -- these patients did quite well with low doses of digitoxin. this is important since digitoxin has a quite long half-life, so having adverse cardiovascular effects from digitoxin toxicity is a real concern: digitoxin's half-life is high in younger people at 6.7 +/- 1.7 days, but more in patients in 8^th and 9^th decade, the half-life having a mean of 25 +/- 9 days: https://pubmed.ncbi.nlm.nih.gov/16296683/

        -- there was a Norwegian study in 1998 finding that even a dose of digitoxin 0.05 mg/d can lead to digitoxin toxicity (this study had only 6 patients aged 77-93yo who had high digitoxin levels with symptoms of nausea, diarrhea, anorexia and weakness. all did well with withholding the drug; this study found a drug half-life of 25.2 days: https://pubmed.ncbi.nlm.nih.gov/16296683/

        -- a 2024 Canadian study did review the literature on cardiac glycoside toxicity finding 58 studies. most receiving pacemakers and/or Digoxin-Immune Fab fragments for those on digoxin: https://pmc.ncbi.nlm.nih.gov/articles/PMC11246064/ . in this study the small numbers of patients were given magnesium and seemed to benefit a lot. some of the studies were on patients taking digitoxin, but the overall study review did not have much information on how those patients did/what therapies were best

        -- it is not clear in this current digitoxin study what the quality of the "guideline-directed medical therapy" provided for the heart failure on an individual basis. we do know that at baseline almost all were on beta-blockers, and 3/4 on spironolactone, but very low numbers on ACE-inhibitors (18%) or angiotensin receptor-neprilysin inhibitor (40%), or SGLT-2s (20%). 

            -- how did the heart failure meds change during the course of the study? how did the outcomes change with added digitoxin when the baseline heart failure regimen changed?

                -- were there cutpoints in left ventricular ejection fraction that divided the response to digitoxin as the therapy changed?

                -- we really need this granular data in determining the role of digitoxin as a therapeutic intervention in patients on relevant current guideline-directed medical therapy

Limitations:

-- as above, we do not have granular, individual patient-level information on what the baseline cardiac meds were, nor how they changed during the course of the study. were the patients aggressively treated by guideline-directed therapy as it changed? was there a difference in outcomes based on that? how did important lab tests/clinical function change as treatment changed over time? 

    -- oddly enough, there is no clear indication of when the patients were recruited for this study, nor when the last patient evaluation was. Buried in the supplemental materials, this clinical study protocol was approved (with subsequent modifications) in 2014. and we know that this was a 3-year study (though some patients were followed up to 107 months)

    -- BUT, if the first patients were recruited in the study about 8 years ago (ie about 2017-8), it is hard to argue, as per the article, that all patients had received baseline evidence-based therapy for heart failure for a period of at least 6 months before being recruited into the study, given that some known evidence-based treatments were not apparently applied (eg, sacubitril/valsartan became available in 2015 and included in guideline-directed therapy in 2016; empagliflozin in 2014 but not in guideline-based therapy until 2021, though that should still be part of therapy by the time this study was published). at the earliest baseline patient recruitment of presumably 2017-8, only 36% were on ACE-i, 19% on ARB, 40% on ARB-neprilysin inhibitor, 76% on MRA, and 19% on SGLT2s.

    -- about 94% had an implantable device (however, they did find that those not on devices actually did 43% better on the outcomes)

    -- so, it seems that the guideline-directed approach to heart failure with reduced ejection fraction that we have been using for many years now, these patients did not seem to really be on optimal background therapy. And this makes it hard to interpret the applicability of the study's results today

-- one other issue is that we have zero information about non-cardiac meds taken, which could also affect the generalizability of the results

-- this trial had lots of prespecified subgroups, as noted above in the Results section. Even prespecified subgroup results are not considered as statistically rigorous as the primary outcome

-- we have only short-term results, though this study is continuing into the future. however, a 3-year time frame as in the current study is still quite helpful given how advanced the heart failure was.

-- we still do not know the best starting dose of digitoxin, nor the best titration dose to achieve the lowish plasma level that seems to be even more effective and with fewer associated adverse events 

so, this study makes the case that digitoxin can be a useful med in patients with severe heart failure with reduced ejection fraction who are already on standard heart failure regimens

-- however, the FDA eliminated digitoxin as an approved medication in 2008, citing its prolonged half-life, lack of available antidote, and potential for severe adverse events

    -- i will admit that i did this blog thinking that digitoxin was available in the US. Wrong. but i will send this blog out since it does raise some significant issues 

-- digitoxin does have some significant benefits over digoxin, including its better absorption, more stable blood levels (higher protein binding, longer half-life), and lack of need for renal dosing.

    -- the toxicity concern about digitoxin is certainly a real issue, though there is an antidote for digoxin (Digifab, digoxin immune fab). After trying difference searches for digitoxin toxicity and treatment, i could only find old foreign studies. so, i cannot say that this digoxin antidote would work for digitoxin, though it is pretty similar structurally (digoxin has an extra hydroxyl group at the C-3 position of the B-ring). the best i could find is https://www.uniba.it/it/ricerca/dipartimenti/scienzebiomediche/didattica1/archivio-didattica/farmacologia/materiale-didattico-bemc/3a-inotropes.pdf

    -- and it is clear from the DIG (digoxin study) and the current study with digitoxin, that lower doses work better (ie, we should probably have a lower range of acceptable blood levels for both drugs)

-- so, is this a useful blog?  i still sort of think so, since:

    -- the approach of this study appropriately started at a low dose of digitoxin with careful, prescribed dose modifications as needed

        -- the Norwegian study mentioned above found that starting elderly on 0.5mg of digitoxin was too much for some. Dosing should be age-specific and probably in elderly the initial dose should be less than 0.5mg/d, with pretty aggressive plasma concentration monitoring

    -- we should monitor drug levels regularly, especially if there could be a drug-drug interaction with any new med

    -- digitoxin dosing could be used and monitored, even with required clinician education prior to the license to prescribe (as happens with some other meds, as with oral retinoic acid and clozapine)

    -- we need information as to whether the digoxin antidote (Digifab, digoxin immune fab) works with digitoxin

    -- and, even with digoxin, the history that some of us lived through was that digoxin toxicity was pretty common

        -- one likely valid conclusion from this study is that, even in the absence of digitoxin, it is very likely (but should be tested) that using digoxin, the close cousin of digitoxin, would have benefits in the group of very sick patients with heart failure with reduced ejection fraction (though it seems that we should target a lower blood level for digoxin than the upper limit of the "thjerapeutic range")

    -- but, digitoxin does provide some pretty clear advantages over digoxin in terms of those with decreasing renal function and dosing that more reliably leads to pretty constant blood levels over time

    -- so, in terms of digitoxin, we are left with an FDA that is largely anti-drug to evaluate whether this phoenix should rise from the ashes...

geoff

-----------------------------------

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@bidmc.harvard.edu

to get access to all of the blogs:  go to http://gmodestmedblogs.blogspot.com/ to see the blogs in reverse chronological order

or you can just click on the magnifying glass on top right, then type in a name in the search box and get all the blogs with that name in them