Little is known about the efficacy and risk of different serum digoxin concentrations (SDC). The Digitalis Investigation Group (DIG) trial, a randomized, placebo-controlled study, has concluded that digoxin therapy had no effect on mortality, but that it modestly reduced hospitalizations caused by worsening heart failure. Rathore and colleagues conducted a post-hoc evaluation of this trial to determine the relationship between SDC, mortality risk, and risk of hospitalization in patients with heart failure and left ventricular dysfunction.

Study subjects in the DIG trial included patients with stable heart failure and a left ventricular ejection fraction of 45 percent or less. Rathore and colleagues restricted their analysis to men who were randomly assigned to digoxin or placebo and who had a valid SDC level measured at least six hours after their previous digoxin dose. The study sample consisted of 2,611 subjects assigned to placebo and 1,171 assigned to digoxin (N = 3,782). Patients in the digoxin group were divided further into three groups: 0.5 to 0.8 ng per mL, 0.9 to 1.1 ng per mL, and 1.2 ng per mL and higher. The authors used the DIG trial's primary end point of all-cause mortality, and secondary outcomes of death caused by cardiovascular causes and heart failure, and all-cause hospitalization and hospitalization for suspected digoxin toxicity. Of the 1,171 men with SDC levels assessed at one month, 572 (49 percent) had an SDC of 0.5 to 0.8 ng per mL, 322 (27 percent) had an SDC of 0.9 to 1.1 ng per mL, and 277 (24 percent) had an SDC of 1.2 ng per mL and higher. There was no difference in all-cause mortality between placebo patients and patients taking digoxin. However, higher SDC levels were associated with higher crude all-cause and cardiovascular mortality, but not mortality caused by heart failure. Patients with the lowest SDC levels had lower rates of mortality in all categories: 6.3 percent lower for all-cause mortality, 3.7 percent for cardiovascular mortality, and 4.7 percent for mortality caused by worsening heart failure. Patients with the highest SDC levels had higher rates of mortality except with worsening heart failure. Multivariate adjustment did not change the association of lower SDC levels with lower mortality but did attenuate the association between higher SDC levels and mortality. Higher SDC levels were also associated with higher rates of hospitalization, including hospitalization for digoxin toxicity.

The authors conclude that the effectiveness of digoxin varies with SDC, with higher SDC levels associated with higher rates of mortality and hospitalization. The optimal SDC appears to be between 0.5 and 0.8 ng per mL for men with stable heart failure and left ventricular dysfunction. The authors attribute the mechanism of these findings to the neurohormonal benefits of digoxin at lower concentrations, as opposed to the harmful inotropic effects that occur at higher dosages. The authors' analysis was insufficiently powered to determine whether these effects hold for women.