Digoxin and reduction in mortality and hospitalization in systolic and diastolic heart failure at low serum digoxin concentrations: reply

We appreciate Dr Rahimtoola's comments regarding our recent report1 based on the Digitalis Investigation Group (DIG) trial. Our analysis, which included both men and women with systolic (ejection fraction ≤45%: main trial) or diastolic (ejection fraction >45%; ancillary trial) heart failure (HF), confirmed one of the key findings of the DIG trial that digoxin significantly reduces hospitalizations due to worsening HF in a broad population of HF patients with sinus rhythm.2 Our study also demonstrated that digoxin reduced HF hospitalization regardless of serum digoxin concentrations (SDC). Furthermore, we observed that in patients who achieved low SDC (0.5–0.9 ng/mL), digoxin was associated with reduction in all-cause mortality and all-cause hospitalizations.1 In a subgroup analysis, we found no significant interaction between digoxin and any major patient characteristic, including sex and ejection fraction. Our subgroup analysis based on low SDC and placebo patients included 4843 patients: 982 patients receiving digoxin who had low SDC and 3861 patients receiving placebo. Of these, 602 (12%) had diastolic HF: 108 patients with low SDC and 494 patients receiving placebo. The magnitude of the absolute and relative reductions in total mortality was comparable between patients with systolic and diastolic HF, and there was no significant heterogeneity in the effect of digoxin between these two groups (adjusted P for interaction 0.834).1 Among systolic HF patients, low SDC was associated with a 3% absolute reduction (compared to 34% mortality in the placebo group, 31% of patients with low SDC died during the study) and a 22% relative reduction (adjusted hazard ratio 0.78, 95% CI, 0.68–0.87; P<0.0001) in all-cause mortality. Similarly, among diastolic HF patients, low SDC was associated with a 4% absolute reduction (compared to 23% mortality in the placebo group, 19% of patients with low SDC died during the study) and a 25% relative reduction (adjusted hazard ratio 0.75, 95% CI, 0.43–1.31; P=0.313) in all-cause mortality. Thus, although Dr Rahimtoola is correct that the mortality difference in the diastolic HF group did not achieve conventional statistical significance, we believe that the overall study findings are most consistent with the view that digoxin at low SDC reduces mortality by about 20–25%, independent of ejection fraction. To further clarify this issue, a more detailed analysis of the DIG ancillary trial, including Kaplan–Meier curves, will be the subject of a future report.3