The DIG trial provides important data concerning the efficacy of digoxin in patients with HF from reduced EF.⁵⁸ In the main part of this trial, 6800 patients with LVEF ≤45% were randomized to digoxin or placebo in addition to diuretics and ACE inhibitors. The primary end point of all-cause mortality was not significantly different between the placebo and the digoxin groups. The need for hospitalization and cointervention (defined as increasing the dose of diuretics and ACE inhibitors or adding new therapies for worsening HF) was significantly lower in the digoxin group, even in those patients who were not previously taking digoxin. Twenty-eight percent fewer patients on digoxin compared with placebo were hospitalized for worsening HF.

Results from the DIG study showed a neutral effect on the primary study endpoint, mortality from any cause, during an average follow-up of approximately 3 years. This differs from other oral agents with inotropic properties, which have been associated with an adverse effect on mortality. These long-term data are consistent with recent results obtained from an analysis of the combined PROVED and RADIANCE databases.⁶¹ In this analysis, patients who continued digoxin as part of triple therapy with diuretics and an ACE inhibitor were much less likely to develop worsening HF (4.7%) than those treated with a diuretic alone (39%, P < .001), diuretic plus digoxin (19%, P = .009), or diuretic plus an ACE inhibitor (25%, P = .001).

Although the number of patients in the DIG trial with NYHA functional class IV HF was limited, retrospective analysis of this subgroup found clear evidence of clinical benefit of digoxin.⁶⁹ Other results from this trial confirm that digoxin works across the spectrum of LV systolic dysfunction. A prespecified subgroup analysis of patients with evidence of severe HF, as manifested by LVEF <25% or cardiothoracic ratio (CTR) >0.55, showed the benefit of digoxin.^66,70 The following reductions in the combined endpoint of all-cause mortality or hospitalization were seen on digoxin compared with placebo: 16% reduction (95% CI 7-24%) in patients with an LVEF <25%, and a 15% reduction (95% CI 6-23%) in patients with a CTR >0.55.⁷⁰ Reductions in the risk of the combined endpoint of HF-related mortality or hospitalization were even more striking: 39% for patients with LVEF <25% and 35% for patients with a CTR >0.55.

Evidence for the efficacy of digoxin in patients with mild symptoms of HF has been provided by a second retrospective cohort analysis of the combined PROVED and RADIANCE databases.⁷¹ The outcome of patients in these trials randomized to digoxin withdrawal or continuation was categorized using a prospectively obtained HF score based on clinical signs and symptoms. Patients in the mild HF group who were randomized to digoxin withdrawal were at increased risk of treatment failure and had deterioration of exercise capacity and LVEF compared with patients who continued digoxin (all P < .01).

Recent data suggest that the target dose (and serum concentration) of digoxin therapy should be lower than traditionally assumed. Although higher doses may be necessary for maximal hemodynamic effects,⁶⁴ beneficial neurohormonal and functional effects appear to be achieved at relatively low serum digoxin concentrations (SDC) typically associated with daily doses of 0.125 to 0.25 mg.^64,72,73 A retrospective analysis of the relationship of serum digoxin concentration to outcomes in the DIG trial demonstrated a strong direct relationship between the risk of death and serum digoxin concentration, with concentrations > 1.2 ng/mL being associated with harm, whereas concentrations < 1.0 ng/mL were associated with favorable outcomes.⁷⁴ These findings supporting the efficacy of low SDC are reinforced by a retrospective cohort analysis of the combined PROVED and RADIANCE databases indicating that patients with a low SDC (<0.9 ng/mL) were no more likely to experience worsening symptoms of HF on maintenance digoxin than those with a moderate (0.9-1.2 ng/mL) or high (> 1.2 ng/mL) SDC.^74,75 All SDC groups were significantly less likely to deteriorate during follow-up compared with patients withdrawn from digoxin.

Therefore, patients with LV systolic dysfunction and normal sinus rhythm should be started on a maintenance dose of digoxin (no loading dose) of 0.125 or 0.25 mg once daily based on ideal body weight, age, and renal function For patients with normal renal function, a dose of 0.25 mg/day will be typical. Many patients with HF have reduced renal function and should begin at 0.125 mg daily. Patients with a baseline conduction abnormality, or who are small in stature or elderly, should be started at 0.125 mg/day, which can be up-titrated if necessary. After dosing has continued for a sufficient period for serum concentration to reach steady state (typically 2 to 3 weeks), some clinicians consider the measurement of a SDC, especially in elderly patients or those with impaired renal function where the digoxin dose often is not predictive of SDC. SDC measurements may be considered when (1) a significant change in renal function occurs; (2) a potentially interacting drug (amiodarone, quinidine, verapamil, itraconazole, erythromycin, clarithromycin, ritonavir, propafenone, or cyclosporine, and others) is added or discontinued; or (3) confirmation of suspected digoxin toxicity is necessary in a patient with signs/ symptoms or ECG changes consistent with this diagnosis. Samples for trough SDC should be drawn more than 6 hours after dosing; otherwise, the result is difficult to interpret because the drug may not be fully distributed into tissues.

Digoxin alone is often inadequate to control ventricular response in patients with atrial fibrillation. Digoxin slows ventricular response to atrial fibrillation through enhancement of vagal tone. However, with exertion or other increases in sympathetic activity, vagal tone may diminish and ventricular rate accelerate. Addition of a β-blocker complements the pharmacologic action of digoxin and provides more optimal rate control. For patients with a contraindication to β-blockers, amiodarone is a reasonable alternative, although chronic amiodarone use is associated with both thyroid disease and lung toxicity. If amiodarone is added, the dose of digoxin should be reduced and the SDC should be monitored to maintain the serum concentration in the desired range. Some clinicians advocate the short-term, intravenous administration of diltiazem for the acute treatment of patients with very rapid ventricular response, especially those with hemodynamic compromise. In the acute and chronic treatment of atrial fibrillation with rapid ventricular response the clinician must consider the benefits of rate control versus the negative inotropic effects of the agent.

Although digoxin continues to play a role in some patients with HF and atrial fibrillation, the traditional practice of arbitrarily increasing the dose and SDC of digoxin until ventricular response is controlled should be abandoned, because the risk of digoxin toxicity increases as well.

AV node ablation is a consideration in patients who remain symptomatic with atrial fibrillation despite adequate rate control or in those who cannot tolerate drug therapy for rate control. Although there are studies that determine adequate rate control in atrial fibrillation, the recommendations followed in the Atrial Fibrillation Follow-up Investigation of Sinus Rhythm Management (AFFIRM) trial are a reasonable starting point.⁷⁶ These recommendations include: a resting heart rate ≤80 bpm, an average heart rate by Holler monitor of ≤100 bpm, and no heart rate >110% of the age-predicated maximum or a heart rate ≤110 bpm in a 6-minute walk test.