Author + information
- Received December 31, 2000
- Revision received May 14, 2001
- Accepted May 23, 2001
- Published online September 1, 2001.
- Michael W Rich, MD, FACC∗,* (, )
- Frances McSherry, MS†,
- William O Williford, PhD†,
- Salim Yusuf, MD, FACC‡,
- for the Digitalis Investigation Group
- ↵*Reprint requests and correspondence: Dr. Michael W. Rich, Associate Professor of Medicine, Washington University School of Medicine, Cardiovascular Division, 660 South Euclid Avenue, Box 8086, St. Louis, Missouri 63110
This study was designed to determine the effect of increasing age on mortality, hospitalizations and digoxin side effects in patients with heart failure (HF), and to determine whether the effect of digoxin on clinical outcomes varies as a function of age.
The incidence and prevalence of HF increase with advancing age, but there are limited data on the clinical course and response to specific therapeutic interventions in elderly patients with HF.
The Digitalis Investigation Group (DIG) study was a prospective, randomized clinical trial involving 7,788 patients with HF randomized to digoxin or placebo and followed for an average of 37 months. In the present analysis, patients were stratified into five age categories: <50 years (n = 841), 50 to 59 years (n = 1,545), 60 to 69 years (n = 2,885), 70 to 79 years (n = 2,092) and ≥80 years (n = 425). Interactions between age and the following clinical outcomes were examined: total mortality, all-cause hospitalizations, HF hospitalizations, the composite of HF death or HF hospitalization, hospitalization for suspected digoxin toxicity and withdrawal from therapy because of side effects.
Increasing age was an independent risk factor for total mortality, all-cause hospitalization, HF hospitalization, HF death or hospital admission, hospitalization for suspected digoxin toxicity and withdrawal from digoxin therapy (all p < 0.001). However, there were no significant interactions between age and digoxin treatment with respect to any of the major clinical end points.
Increasing age is associated with progressively worse clinical outcomes in patients with HF. However, the beneficial effects of digoxin in reducing all-cause admissions, HF admissions, and HF death or hospitalization are independent of age. Thus, digoxin remains a useful agent for the adjunctive treatment of HF due to impaired left ventricular systolic function in patients of all ages.
Both the incidence and prevalence of heart failure (HF) increase progressively with advancing age (1), and individuals over 65 years of age account for 79% of all hospitalizations and 88% of all deaths attributable to HF (2,3). Moreover, it is anticipated that the societal burden of HF will continue to increase as the population ages. However, despite the fact that HF is predominantly a disorder of the elderly, there are limited data on the clinical course, prognosis and response to specific therapeutic interventions in older patients, particularly those over the age of 80 years.
The Digitalis Investigation Group (DIG) study was a prospective, randomized, double-blind placebo-controlled clinical trial designed to evaluate the effects of digoxin on mortality and hospitalizations in patients with chronic HF and sinus rhythm (4). Treatment was administered on a background of diuretic and angiotensin-converting enzyme (ACE) inhibitor therapy. A total of 7,788 patients were enrolled in the trial, including 6,800 with a left ventricular ejection fraction (LVEF) of ≤0.45 (main trial), and 988 with an LVEF >0.45 (ancillary trial). The mean age of patients in the study was 63.9 years, and 32% were 70 years of age or older. Mean follow-up was 37 months and the primary trial end point was all-cause mortality. The principal finding of the DIG study was that although digoxin had no discernible effect on total mortality, the composite end point of HF death or hospitalization was reduced 24% (95% confidence interval 18% to 30%; p < 0.001) in patients randomized to digoxin treatment (4).
The purpose of the present analysis was to examine the effect of age on clinical outcomes, including mortality, hospitalizations and adverse drug reactions, and to determine whether the response to digoxin varied with age.
The rationale, design and main results of the DIG study have been previously published (4,5). Briefly, patients with chronic HF based on clinical symptoms, signs and radiographic findings, and who were in sinus rhythm, were randomized to digoxin or matching placebo at 302 centers in the U.S. and Canada. Patients were stratified according to LVEF (≤0.45 or >0.45), as assessed by radionuclide angiography, contrast ventriculography or echocardiography. Angiotensin-converting enzyme inhibitor therapy was recommended for all patients with an LVEF ≤0.45, and more than 93% of all DIG participants received such treatment. The digoxin dose was determined using a standardized algorithm based on age, gender, weight and renal function. The median dose of digoxin was 0.25 mg, the mean steady-state digoxin level at one-month follow-up was 0.89 ± 0.50 ng/ml and 87% of patients in the digoxin group had digoxin levels within the therapeutic range at the one-month follow-up visit. The DIG study was approved by the institutional review boards at all participating centers, and all patients provided written informed consent.
The primary outcome for the DIG study was all-cause mortality. Secondary outcomes were cardiovascular mortality, death from worsening HF, hospitalization due to worsening HF and hospitalization for other reasons, including suspected digoxin toxicity. Data were reviewed at six-month intervals by an independent Data and Safety Monitoring Board. The trial was terminated as planned on December 31, 1995, after a mean follow-up of 37 months (range 28 to 58 months). When the main results of the trial were reported, the vital status was unknown in 1.4% of patients, and data on these patients were censored at the date of the last known follow-up visit.
In the present analysis, all 7,788 DIG study patients were retrospectively grouped into five age categories: <50 years (n = 841), 50 to 59 years (n = 1,545), 60 to 69 years (n = 2,885), 70 to 79 years (n = 2,092) and ≥80 years (n = 425). Total mortality, all-cause hospitalizations, HF hospitalizations and the combined end point of HF death or HF hospitalization were determined within each age group, both overall and within subgroups according to treatment assignment. Because there was no heterogeneity in the effect of digoxin on any of these end points with respect to LVEF ≤0.45 or >0.45, data from the main and ancillary trials were pooled for the present analysis.
The Mantel-Haenszel statistic was used to test for linear association between age and each of the above outcomes, and for the association between age and adverse drug reactions, defined as drug withdrawal because of side effects or hospitalization for suspected digoxin toxicity. Kaplan-Meier curves were plotted by age category for total mortality and for the combination of HF death or HF hospitalization. Multiple regression of survival data based on the Cox proportional hazards model with backward selection of covariates was used to determine independent correlates of each outcome, both overall and within age groups. Variables considered for inclusion in the model, in addition to age category and treatment assignment, were: cardiothoracic ratio, body mass index, ejection fraction, serum creatinine, heart rate, systolic blood pressure, gender, race, etiology of HF, prior myocardial infarction, diabetes mellitus, hypertension, use of ACE inhibitors and New York Heart Association functional class. All data are reported as means (± standard deviation) or as percentages.
Baseline characteristics of patients in each of the five age groups are summarized in Table 1. Older patients were more likely to be women, to be Caucasian and to take diuretics. They also had higher mean systolic blood pressures, serum creatinine levels, cardiothoracic ratios on chest radiograph, New York Heart Association functional classifications and LVEFs than younger patients. The prevalence of diabetes mellitus and prior myocardial infarction increased up to the seventh decade, but then declined at older age. Use of ACE inhibitors declined slightly at older age, but this could reflect the fact that older patients were more likely to have an LVEF >0.45.
Effect of age on mortality and hospitalizations
As shown in Table 2and Figures 1 and 2, ⇓⇓increasing age was associated with a progressive increase in the incidence of death, hospitalization, HF hospitalization and HF death or hospitalization (p for linear trend <0.001 for each outcome). Compared with patients <50 years old, the adjusted relative risk of death in patients over 80 years of age was 2.05. Similarly, the adjusted relative risk of hospitalization for any reason was 1.55, the adjusted relative risk of hospitalization for HF was 1.94 and the adjusted relative risk for HF death or admission was 2.04.
Effect of digoxin on clinical outcomes by age group
Table 2also shows the effect of digoxin on mortality, all-cause and HF hospitalizations and HF death or hospitalization in each age group. Consistent with the findings from the primary DIG study analysis, digoxin had no effect on total mortality, and there also was no interaction between age and digoxin treatment with respect to this end point. Similarly, whereas digoxin was associated with a modest reduction in all-cause hospitalizations (6%), and more substantial reductions in HF hospitalizations (27%) and in HF death or hospitalization (24%), there was no interaction between age and treatment assignment with respect to any of these outcomes. Thus, digoxin was associated with a risk reduction in HF death or hospitalization within each age category, and the absolute reduction attributable to digoxin therapy was 87 per 1,000 in individuals <50 years of age (p = 0.12), 34 per 1,000 in patients 50 to 59 (p = 0.0001), 62 per 1,000 in patients 60 to 69 (p = 0.0001), 91 per 1,000 in patients 70 to 79 (p = 0.0001) and 67 per 1,000 in patients ≥80 years of age (p = 0.20).
Adverse drug reactions
Table 3shows mean digoxin dosages and serum digoxin levels measured one month after initiation of therapy in a randomly selected subset of DIG study patients receiving active treatment. Despite an age-related decline in mean dosage, mean digoxin levels tended to be higher in older patients, although levels were within the therapeutic range in all age groups. Table 4characterizes the frequency of side effects and suspected digoxin toxicity throughout the follow-up period as a function of age and treatment assignment. Data are based on reports filed by study physicians at each local site. Electrocardiographic tracings documenting arrhythmic events were obtained when available, but specific details regarding nonfatal arrhythmias (such as supraventricular tachycardias) were not collected. Overall, 3.4% of patients were withdrawn from study medication because of apparent side effects, and the frequency of drug withdrawal increased progressively with age, from 1.7% in patients <50 years of age to 5.4% in patients over the age of 80 (adjusted risk ratio = 1.36, p = 0.0001). However, because medication withdrawal increased with age in both the digoxin and placebo groups, there was no significant interaction between age and medication withdrawal due to digoxin toxicity. Similarly, hospitalizations for suspected digoxin toxicity increased with age (adjusted risk ratio = 1.31, p = 0.0036), but there was no significant interaction between age and hospitalization for digoxin toxicity. With respect to specific side effects associated with digoxin toxicity, atrioventricular block, supraventricular arrhythmias and nausea or vomiting all occurred more frequently with advancing age, but again there was no interaction between age and treatment assignment. Of note, there was no apparent relationship between age and the occurrence of visual disturbances or serious ventricular arrhythmias.
Clinical outcomes by age and ejection fraction
Table 5displays clinical outcomes based on age group, LVEF category (≤0.45 vs. >0.45) and treatment assignment. The proportion of patients with preserved left ventricular systolic function, as indicated by LVEF >0.45, increased significantly with age, from 8.2% in patients under age 50 to 19.5% among patients over age 80. Total mortality was significantly lower in patients with LVEF >0.45 in all age groups (p = 0.0001 overall; p < 0.05 within each age group), and the absolute mortality difference was similar across age groups. Hospitalizations for HF and the composite end point of death or hospitalization for HF occurred less frequently in patients with LVEF >0.45 (both p = 0.0001), and absolute differences were again similar across age groups. In contrast, all-cause hospitalizations did not differ according to LVEF category (p = 0.12), either overall or within any specific age group.
Table 6lists independent predictors of all-cause mortality based on a Cox proportional hazards model. Increasing age, as defined by the five age categories, was a strong independent predictor of total mortality, and there was a 20% increase in mortality risk per decade. Similarly, increasing age was an independent predictor of all-cause hospitalizations, HF hospitalizations and the composite of death or HF admission (data not shown). Digoxin treatment was independently associated with lower all-cause and HF hospitalization rates (p = 0.0087 and p = 0.0001, respectively), as well as death or HF hospitalization (p = 0.0001), but there was no effect of digoxin on total mortality.
Although the inotropic effect of digoxin is preserved at older age (6), the volume of distribution declines with age, and it has been suggested that older adults are at increased risk for the development of digitalis intoxication and other adverse side effects (7). Moreover, although withdrawal from digoxin has been associated with clinical deterioration in younger patients (8,9), it has been reported that many older adults on chronic digoxin therapy can be safely withdrawn without adverse consequences (10). In light of the fact that the majority of HF patients are over 65 years of age, and that the majority of digoxin prescriptions are written for individuals in the older age groups, the safety and efficacy of digoxin in the elderly are of great clinical relevance.
Although increasing age is associated with a modest but progressive rise in the frequency of hospitalizations for suspected digoxin toxicity and in the number of patients withdrawn from therapy because of side effects, these trends are apparent in placebo-treated as well as digoxin-treated patients. As a result, older age was not associated with a significant excess in the occurrence of digoxin toxicity relative to placebo. Similarly, atrioventricular block, supraventricular arrhythmias and gastrointestinal symptoms occurred more commonly in the elderly, both in the placebo and treatment groups, but there was no age-related excess of these adverse events in patients treated with digoxin. These results are perhaps surprising, as older adults are often considered to be at increased risk for digoxin toxicity (7), and it has also been suggested that the therapeutic range for digoxin may be shifted downward in the very elderly (6). The findings of this study imply that side effects often attributed to digoxin toxicity in older individuals may be due to other causes, including age-related changes in atrioventricular conduction, predisposition to supraventricular arrhythmias and increased prevalence of nonspecific gastrointestinal disturbances. Thus, these findings are reassuring, particularly considering that the incidence of hospitalization for suspected digoxin toxicity was only 4.4% over a three-year period in patients over 80 years of age, despite the fact that digoxin levels were not routinely monitored in the DIG study. This rate is considerably lower than earlier studies indicating a 10% to 20% incidence of digitalis toxicity during long-term therapy in elderly patients (7).
The DIG study demonstrated that digoxin reduced all-cause hospitalizations by 6%, HF hospitalizations by 27% and the combined end point of HF death or hospitalization by 24%, but there was no effect on total mortality (4). The present analysis reveals that these findings are consistent across all age groups, including the very elderly. These findings support the view that digoxin is an effective agent for the treatment of HF, even at very elderly age.
HF with preserved systolic function
Heart failure with preserved systolic function accounts for 30% to 50% of all cases of HF, and the prevalence of this disorder increases with age (11,12). In the DIG study, patients with an LVEF >0.45 comprised only 12.7% of the total population, most likely reflecting the fact that many physicians consider preserved LVEF to be a contraindication to the use of digoxin for the treatment of HF. Compared to patients with reduced LVEF, the etiology of HF in DIG study patients with preserved LVEF was less likely to be ischemic (71 vs. 57%, p < 0.001). Despite the relatively low enrollment of patients with preserved systolic function, the proportion of patients with this syndrome increased progressively with age, from 8.2% in patients <50 years to 19.5% in patients over 80 years. Notably, the incidence of all major outcomes increased with advancing age, regardless of whether the LVEF was < or >0.45. In addition, although the prognosis with respect to both mortality and admission for HF was somewhat better in patients at all ages with preserved LVEF, there was no difference in all-cause hospitalizations as a function of LVEF. These findings are consistent with other studies indicating that LVEF predicts mortality, but that it correlates poorly with either symptom severity (such as New York Heart Association functional class) (13)or functional capacity (such as Vo2max) (14).
The results of these analyses indicate that age should not be a major factor in determining when to use digoxin for the treatment of HF. However, because side effects potentially due to digoxin therapy occur more commonly in older individuals, it is important to monitor therapy more closely. In addition, lower dosages of digoxin may be prudent in the very elderly, especially because the therapeutic range may be lower, and recent data suggest that there is no additional benefit from higher doses (15). With respect to HF with preserved LVEF, results from the DIG ancillary study indicate that digoxin may be useful in selected patients with this condition (4). However, additional data are needed, particularly in the very elderly, before digoxin can be recommended as routine therapy in patients with HF and preserved LVEF.
Randomization in the DIG study was not stratified by age, and there were relatively few patients over 80 years of age enrolled in the trial. In addition, although up to two-thirds of all HF patients in the general population are over age 70, only 32% of DIG study patients were in this age group, indicating a significant selection bias in patient recruitment. Importantly, the exclusion of patients with atrial fibrillation, which increases in prevalence with advancing age, likely contributed to this selection bias. The DIG study also enrolled an insufficient number of elderly racial and ethnic minority group subjects to permit firm conclusions about the safety and efficacy of digoxin in these subgroups. In light of these considerations, the study’s findings may not be generalizable to elderly HF patients in the community.
Increasing age is a strong independent predictor of total mortality, hospital admissions, HF death or hospitalization and suspected adverse reactions to digoxin in patients with HF. However, the effects of digoxin on clinical outcomes are independent of age, and digoxin remains a useful adjunctive agent for the treatment of HF associated with reduced left ventricular systolic function in patients of all ages. The role of digoxin in HF patients with preserved systolic function requires further study.
☆ The DIG Study was sponsored by the National Heart, Lung and Blood Institute and the Department of Veterans Affairs Cooperative Study Program. Digoxin and placebo were provided by Glaxo Wellcome.
- angiotensin-converting enzyme
- Digitalis Investigation Group
- heart failure
- left ventricular ejection fraction
- Received December 31, 2000.
- Revision received May 14, 2001.
- Accepted May 23, 2001.
- American College of Cardiology
- ↵Owings MF, Lawrence L. Detailed diagnoses and procedures. National Hospital Discharge Survey, 1997. National Center for Health Statistics. Vital and Health Statistics 1999; Series 13, No. 145, p. 14.
- Graves EJ. Detailed diagnoses and procedures. National Hospital Discharge Survey, 1990. Vital and health statistics, series 13, data from the National Health Survey. Hyattsville, MD: National Center for Health Statistics, 1992;113:1–225.
- Schneeweiss A,
- Schettler G
- Uretsky B.F,
- Young J.B,
- Shahidi F.E,
- Yellen L.G,
- Harrison M.C,
- Jolly M.K
- Gottdiener J.S,
- Arnold A.M,
- Aurigemma G.P,
- et al.
- Vasan R.S,
- Larson M.G,
- Benjamin E.J,
- Evans J.C,
- Reiss C.K,
- Levy D
- Slatton M.L,
- Irani W.N,
- Hall S.A,
- et al.