statistics Homeworke for clinical trails

Ibtihal abdullah

nejm_effect_of_digoxin_feb1997.pdf

Home >Applied Sciences homework help >statistics Homeworke for clinical trails

Vo l u m e 3 3 6 N u m b e r 8

�

525

The New England

Journal

Medicine

V O L U M E 3 3 6

E B R U A R Y

2 0 , 1 9 9 7

N U M B E R 8

THE EFFECT OF DIGOXIN ON MORTALITY AND MORBIDITY IN PATIENTS WITH HEART FAILURE

IGITALIS

NVESTIGATION

ROUP

BSTRACT

Background

The role of cardiac glycosides in treat- ing patients with chronic heart failure and normal si- nus rhythm remains controversial. We studied the effect of digoxin on mortality and hospitalization in a randomized, double-blind clinical trial.

Methods

In the main trial, patients with left ven- tricular ejection fractions of 0.45 or less were random- ly assigned to digoxin (3397 patients) or placebo (3403 patients) in addition to diuretics and angioten- sin-converting–enzyme inhibitors (median dose of digoxin, 0.25 mg per day; average follow-up, 37 months). In an ancillary trial of patients with ejection fractions greater than 0.45, 492 patients were random- ly assigned to digoxin and 496 to placebo.

Results

In the main trial, mortality was unaffect- ed. There were 1181 deaths (34.8 percent) with dig- oxin and 1194 deaths (35.1 percent) with placebo (risk ratio when digoxin was compared with placebo, 0.99; 95 percent confidence interval, 0.91 to 1.07; P

�

0.80). In the digoxin group, there was a trend to- ward a decrease in the risk of death attributed to wor- sening heart failure (risk ratio, 0.88; 95 percent con- fidence interval, 0.77 to 1.01; P

�

0.06). There were 6 percent fewer hospitalizations overall in that group than in the placebo group, and fewer patients were hospitalized for worsening heart failure (26.8 percent vs. 34.7 percent; risk ratio, 0.72; 95 percent confi- dence interval, 0.66 to 0.79; P

�

0.001). In the ancillary trial, the findings regarding the primary combined outcome of death or hospitalization due to worsen- ing heart failure were consistent with the results of the main trial.

Conclusions

Digoxin did not reduce overall mor- tality, but it reduced the rate of hospitalization both overall and for worsening heart failure. These find- ings define more precisely the role of digoxin in the management of chronic heart failure. (N Engl J Med 1997;336:525-33.)

Rekha Garg, M.D., Richard Gorlin, M.D., Thomas Smith, M.D., and Salim Yusuf, M.D., assume responsibility for the contents of this article on behalf of the Digitalis Investigation Group.

Address reprint requests to Dr. Gorlin at Mount Sinai Medical Center, Box 1018, 1 Gustave L. Levy Pl., New York, NY 10029-6574.

*The participants in the Digitalis Investigation Group are listed in the Appendix.

LTHOUGH digoxin is one of the most commonly prescribed drugs for the treat- ment of heart failure, there is uncertainty about its long-term efficacy and safety.

1-3

Several recent short-term, randomized trials indi- cated that withdrawing digoxin worsens functional status, exercise capacity, and the left ventricular ejec- tion fraction in patients with heart failure.

4,5

However, the long-term effect of digoxin on mortality and hospitalization for heart failure or other causes is un- known. We conducted a randomized, double-blind, placebo-controlled trial to evaluate the effects of dig- oxin (Lanoxin, Glaxo Wellcome) on mortality from any cause (the primary end point) and on hospital- ization for heart failure (the secondary end point) over a three-to-five-year period in patients with heart failure and normal sinus rhythm.

METHODS

Design

The rationale and design of the study and the base-line char- acteristics of the patients have been reported previously.

Patients were enrolled at 302 clinical centers in the United States and Canada. The study was organized and conducted by a Steering Committee representing the National Heart, Lung, and Blood Institute; the Department of Veterans Affairs Cooperative Studies Program; and cardiologists from the United States and Canada. An independent Data and Safety Monitoring Board monitored the progress of the study. The study was approved by the institu- tional review board at each participating center. All the patients gave written informed consent.

Eligibility

Patients were eligible for the main trial if they had heart failure and a left ventricular ejection fraction of 0.45 or less (6800 pa-

526

�

Fe b r u a r y 2 0 , 1 9 9 7

T h e New E n g l a n d Jo u r n a l o f Me d i c i n e

tients) and were in normal sinus rhythm. Patients with heart fail- ure and a left ventricular ejection fraction of more than 0.45 (988 patients) were enrolled in an ancillary trial conducted parallel to the main study. The diagnosis of heart failure was based on cur- rent or past clinical symptoms (limitation of activity, fatigue, and dyspnea or orthopnea), signs (edema, elevated jugular venous pressure, rales, or S

gallop), or radiologic evidence of pulmonary congestion. Patients were eligible for the study whether or not they were already being treated with digoxin. The left ventricular ejection fraction was assessed by radionuclide left ventriculogra- phy, left ventricular contrast angiography, or two-dimensional echocardiography. The criteria for exclusion from the study have been published previously.

Randomization, Dose Titration, and Follow-up

In telephone conversations between the clinical centers and the data coordinating center from February 1991 through August 1993, the patients were randomly assigned to receive digoxin or placebo. The randomization was stratified according to center and left ventricular ejection fraction (

�

0.45 or

�

0.45). The rec- ommended initial dose of the study drug was determined with an algorithm that took into account the patient’s age, sex, weight, and renal function.

The investigators were permitted to modify the dose on the basis of other factors, such as the previous dose of digoxin and the use of concomitant drugs that might alter dig- oxin pharmacokinetics, and they were strongly encouraged to give the patients angiotensin-converting–enzyme inhibitors. Patients who were using digoxin before entry into the study were random- ly assigned to receive either digoxin or placebo without a washout period. All the patients returned for follow-up visits 4 weeks and 16 weeks after randomization and every 4 months thereafter. At each follow-up visit, data were recorded on changes in clinical and functional status, the use of selected nonstudy drugs, hospitaliza- tion, adherence to the study regimen, and side effects. When pa- tients had worsening symptoms of heart failure, it was recom- mended that other therapy for heart failure be used in an optimal fashion. If the patients remained symptomatic despite efforts to optimize other forms of treatment, open-label treatment with digoxin was allowed and the study drug was discontinued.

Outcomes

The primary outcome studied in the main trial was mortality. The secondary outcomes were mortality from cardiovascular causes, death from worsening heart failure, hospitalization for worsening heart failure, and hospitalization for other causes, in particular suspected digoxin toxicity. In the ancillary trial, the oc- currence of death or hospitalization due to worsening heart fail- ure was studied as a combined primary outcome. At this writing, the vital status of 47 patients assigned to digoxin and 46 patients assigned to placebo in the main trial (1.4 percent of the total) re- mains unknown.

Classification of Outcomes

Each investigator classified the cause of death or the primary diagnosis leading to hospitalization without knowing the pa- tient’s study-drug assignment and after reviewing the patient’s hospital chart or interviewing relatives. Deaths due to worsening heart failure were classified as such even if the final event was an arrhythmia. Suspected digoxin toxicity was diagnosed according to the judgment of the investigator.

Statistical Analysis

All the analyses were performed on an intention-to-treat basis with two-sided P values. A stratified log-rank statistic was used to compare the survival distributions in the two study groups. Kap- lan–Meier analysis was used to construct life-table plots. In com- paring the digoxin and placebo groups, we estimated the risk ra- tio associated with an event and calculated the 95 percent confidence interval from the Cox proportional-hazards model.

*CHF denotes congestive heart failure, NYHA New York Heart Associ- ation, and ACE angiotensin-converting enzyme. Because of rounding, not all percentages total 100.

†The clinical signs or symptoms studied included rales, elevated jugular venous pressure, peripheral edema, dyspnea at rest or on exertion, orthop- nea, limitation of activity, S

gallop, and radiologic evidence of pulmonary congestion.

‡This category included valvular and alcohol-related causes of congestive heart failure.

§These drugs included clonidine hydrochloride, doxazosin mesylate, flose- quinan, labetalol hydrochloride, minoxidil, prazosin hydrochloride, and ter- azosin hydrochloride.

ABLE

ASE

-L

INE

HARACTERISTICS

THE

TUDY

ATIENTS

CCORDING

REATMENT

ROUP

HARACTERISTIC

IGOXIN

�

3397) P

LACEBO

�

3403)

Age (yr) — mean

�

SD 63.4

�

11.0 63.5

�

10.8

Ejection fraction — mean

�

SD 28.6

�

8.9 28.4

�

8.9

Median duration of CHF — mo 17 16

% of patients

Female sex 22.2 22.5

Nonwhite race 14.4 14.8

Age

�

70 yr 26.7 27.4

Method of assessing ejection fraction Radionuclide ventriculography Two-dimensional echocardiography Contrast angiography

65.0 29.5 5.5

64.2 30.0 5.8

Cardiothoracic ratio

�

0.55 34.6 34.4

NYHA class I II III IV

13.7 53.3 30.7 2.2

13.0 54.5 30.5 1.9

No. of signs or symptoms of CHF† 0 1 2 3

�

1.1 2.4 7.1 9.3

80.1

1.1 2.0 7.1 8.6

81.2

Medical history Previous myocardial infarction Current angina Diabetes Hypertension

64.7 27.1 28.3 45.0

65.3 26.4 28.6 45.8

Previous digoxin use 44.1 44.6

Primary cause of CHF Ischemic Nonischemic

Idiopathic Hypertensive Other‡

70.8 29.0 15.5 8.0 5.4

70.4 29.3 14.1 9.2 6.0

Concomitant medications Diuretics ACE inhibitors Nitrates Other vasodilators§

81.2 94.1 42.1 0.9

82.2 94.8 43.1 1.5

Daily dose of study medication prescribed 0.125 mg 0.250 mg 0.375 mg 0.500 mg

17.5 70.6 10.3 1.1

17.4 70.0 11.3 0.9

E F F E C T O F D I G OX I N O N M O R TA L I T Y A N D M O R B I D I T Y I N PAT I E N T S W I T H H E A R T FA I LU R E

Vo l u m e 3 3 6 N u m b e r 8

�

527

The Wilcoxon rank-sum test was used to determine any differenc- es between groups in the number of hospitalizations. The Cox proportional-hazards model was also used to test for interactions between the study assignments and predefined variables that included the left ventricular ejection fraction (

�

0.25 vs. 0.25 to 0.45), heart size (cardiothoracic ratio,

�

0.55 vs.

�

0.55), the cause of heart failure (ischemic vs. nonischemic), digoxin use be- fore the trial (any vs. none), and New York Heart Association (NYHA) functional class (I or II vs. III or IV). In this article we report primarily the results of the main trial.

The data were reviewed every six months by the Data and Safe- ty Monitoring Board. December 31, 1995, was chosen in advance as the date on which identification of events would terminate.

Data on the 93 patients whose vital status was unknown on De- cember 31, 1995, were censored as of the date of their most recent follow-up visit. A sensitivity analysis in which we assumed either that all the placebo patients had died or that all the digoxin patients had died did not change the overall results with respect to mortality.

RESULTS

Main Trial (Left Ventricular Ejection Fraction,

�

0.45)

Except as otherwise specified, the results reported here are those of the main trial. In that trial, there were no significant differences in base-line character- istics between the 3397 patients assigned to digoxin and the 3403 patients assigned to placebo (Table 1). The mean duration of follow-up was 37 months (range, 28 to 58).

Mortality

There were 1181 deaths in the digoxin group (34.8 percent) and 1194 deaths in the placebo

group (35.1 percent) (risk ratio when digoxin was compared with placebo, 0.99; 95 percent confi- dence interval, 0.91 to 1.07; P

�

0.80) (Fig. 1 and Table 2).

There were 1016 deaths from cardiovascular caus- es in the digoxin group (29.9 percent) and 1004 such deaths in the placebo group (29.5 percent) (risk ratio, 1.01; 95 percent confidence interval, 0.93 to 1.10; P

�

0.78). In the digoxin group as compared with the placebo group, there was a trend toward a lower risk of mortality attributable to worsening heart failure (number of deaths, 394 vs. 449; risk ra- tio, 0.88; 95 percent confidence interval, 0.77 to 1.01; P

�

0.06) (Fig. 2).

Hospitalizations for Heart Failure and Arrhythmia

Fewer patients were hospitalized for worsening heart failure in the digoxin group (910) than in the placebo group (1180) (risk ratio, 0.72; 95 percent confidence interval, 0.66 to 0.79; P

�

0.001) (Table 3). In all, there were 1927 hospitalizations with worsening heart failure as the primary diagnosis in the digoxin group and 2553 such hospitalizations in the placebo group. There was no significant differ- ence between the groups in the number of patients hospitalized for ventricular arrhythmia or cardiac ar- rest (142 vs. 145). The risk of death from any cause or hospitalization for worsening heart failure was lower in the digoxin group (risk ratio, 0.85; 95 per-

Figure 1.

Mortality in the Digoxin and Placebo Groups. The number of patients at risk at each four-month interval is shown below the figure.

0 52

4 8 12 16 20 24 28 32 36 40 44 48

Months

M o

rt a li ty

f ro

m A

n y C

a u

se (

% )

NO. OF PATIENTS AT RISK Placebo Digoxin

3403

Placebo

Digoxin P � 0.80

3239 3105 2976 2868 2758 2652 2551 2205 1881 1506 1168 734 339 3397 3269 3144 3019 2882 2759 2644 2531 2184 1840 1475 1156 737 335

528

�

Fe b r u a r y 2 0 , 1 9 9 7

T h e New E n g l a n d Jo u r n a l o f Me d i c i n e

*Absolute differences were calculated by subtracting the percentage of deaths in the placebo group from the percentage of deaths in the digoxin group (before values were rounded).

†Risk ratios and confidence intervals (CI) were estimated from the Cox proportional-hazards model.

‡This category includes patients who died from worsening heart failure, even if the final event was an arrhythmia.

§This category includes deaths presumed to result from arrhythmia without evidence of worsening heart failure and deaths due to atherosclerotic coronary disease, bradyarrhythmias, low-output states, and cardiac surgery. Although this outcome was not prespecified, P

�

0.04 for the comparison of study groups with respect to death from other cardiac causes.

¶This category includes deaths due to stroke, embolism, peripheral vascular disease, vascular surgery, and carotid end- arterectomy.

ABLE

EATHS

CCORDING

TUDY

ROUP

AND

AUSE

EATH

IGOXIN

�

3397) P

LACEBO

�

3403) A

BSOLUTE

IFFERENCE

* R

ISK

ATIO

(95% CI)† P V

ALUE

no. of patients (%) %

All 1181 (34.8) 1194 (35.1)

�

0.4 0.99 (0.91–1.07) 0.80 Cardiovascular

Worsening heart failure‡ Other cardiac§ Other vascular¶ Unknown

1016 (29.9) 394 (11.6) 508 (15.0) 50 (1.5) 64 (1.9)

1004 (29.5) 449 (13.2) 444 (13.0) 45 (1.3) 66 (1.9)

0.4

�

1.6 1.9 0.1

�

0.1

1.01 (0.93–1.10) 0.88 (0.77–1.01) 1.14 (1.01–1.30) 1.11 (0.74–1.66) 0.97 (0.69–1.37)

0.78 0.06

Noncardiac and nonvascular

165 (4.9) 190 (5.6)

�

0.7 0.87 (0.71–1.07)

Figure 2.

Mortality Due to Worsening Heart Failure in the Digoxin and Placebo Groups. The number of patients at risk at each four-month interval is shown below the figure.

18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1

0 524 8 12 16 20 24 28 32 36 40 44 48

Months

M o

rt a li ty

D u

e t

o W

o rs

e n

in g

H e a rt

F a il

u re

( %

)

NO. OF PATIENTS AT RISK Placebo Digoxin

3403

Placebo

Digoxin

P � 0.06

3239 3105 2976 2868 2758 2652 2551 2205 1881 1506 1168 734 339 3397 3269 3144 3019 2882 2759 2644 2531 2184 1840 1475 1156 737 335

E F F E C T O F D I G OX I N O N M O R TA L I T Y A N D M O R B I D I T Y I N PAT I E N T S W I T H H E A R T FA I LU R E

Vo l u m e 3 3 6 N u m b e r 8

�

529

cent confidence interval, 0.79 to 0.91; P

�

0.001). The risk associated with the combined outcome of death due to worsening heart failure or hospitaliza- tion related to that diagnosis was lower in the dig- oxin group (in which 1041 patients had the out- come) than in the placebo group (1291 patients) (risk ratio, 0.75; 95 percent confidence interval, 0.69 to 0.82; P

�

0.001) (Fig. 3 and Table 4). These benefits were seen soon after randomization, and they persisted throughout the trial.

Other Hospitalizations

There were 6 percent fewer hospitalizations in the digoxin group than in the placebo group (6356 vs. 6777), and 10 percent fewer hospitalizations for car- diovascular causes (4106 vs. 4570). Furthermore, there were fewer hospitalizations for any reason per patient in the digoxin group than in the placebo group (P

�

0.01 by the Wilcoxon test), and fewer hospitalizations per patient for cardiovascular causes (P

�

0.001). The difference between the groups be- came more apparent when the number of hospital- izations was standardized according to the duration of follow-up. In all, 64.3 percent of the patients in the digoxin group and 67.1 percent of those in the

placebo group were hospitalized for all reasons com- bined, with cardiovascular causes as the primary rea- son in both groups (Table 3). There was no signifi- cant difference between the digoxin group and the placebo group in the number of hospitalizations for myocardial infarction (195 vs. 201) or unstable an- gina (399 vs. 398), but there were fewer hospitaliza- tions for supraventricular arrhythmia in the digoxin group (133 vs. 152). Small proportions of patients in both groups were hospitalized with suspected digoxin toxicity, with significantly more in the dig- oxin group than in the placebo group (2.0 percent vs. 0.9 percent, P

�

0.001). The proportion of pa- tients hospitalized for noncardiovascular reasons was similar in the two groups.

Effects in Subgroups

We examined the influence of digoxin on both the relative and the absolute reduction in the risk of a secondary outcome in prespecified subgroups. With regard to the combined outcome of mortality from worsening heart failure or a hospitalization related to that diagnosis, the benefit of digoxin ap- peared to be greater among patients at high risk — that is, those with lower ejection fractions or en-

*Data shown include the first hospitalization of each patient for each reason.

†Absolute differences were calculated by subtracting the percentage of patients hospitalized in the placebo group from the percentage of patients hospitalized in the digoxin group (before values were rounded).

‡Risk ratios and confidence intervals (CI) were estimated from a Cox proportional-hazards model that used the first hospitalization of each patient for each reason.

§This category includes atrioventricular block and bradyarrhythmia.

¶This category includes coronary-artery bypass grafting and percutaneous transluminal coronary angioplasty.

�

This category includes embolism, venous thrombosis, peripheral vascular disease, hypertension, other vascular surgery, cardiac catheterization, other types of catheterization, pacemaker implantation, installation of automatic implantable car- diac defibrillator, electrophysiologic testing, transplant-related evaluation, nonspecific chest pain, atherosclerotic heart dis- ease, hypotension, orthostatic hypotension, and valve operation.

ABLE

PATIENTS HOSPITALIZED DURING THE STUDY, ACCORDING TO STUDY GROUP AND REASON FOR HOSPITALIZATION.

REASON FOR HOSPITALIZATION* DIGOXIN

(N � 3397) PLACEBO

(N � 3403) ABSOLUTE

DIFFERENCE† RISK RATIO (95% CI)‡ P VALUE

no. of patients (%) %

Cardiovascular Worsening heart failure Ventricular arrhythmia, cardiac arrest Supraventricular arrhythmia§

Atrioventricular block, bradyarrhythmia

1694 (49.9) 910 (26.8) 142 (4.2) 133 (3.9) 14 (0.4)

1850 (54.4) 1180 (34.7) 145 (4.3) 152 (4.5)

9 (0.3)

�4.5 �7.9 �0.1 �0.6

0.1

0.87 (0.81–0.93) 0.72 (0.66–0.79) 0.98 (0.78–1.24) 0.87 (0.69–1.10) 1.56 (0.68–3.61)

�0.001 �0.001

Suspected digoxin toxicity Myocardial infarction Unstable angina Stroke Coronary revascularization¶ Cardiac transplantation Other cardiovascular �

67 (2.0) 195 (5.7) 399 (11.7) 157 (4.6) 83 (2.4) 25 (0.7)

452 (13.3)

31 (0.9) 201 (5.9) 398 (11.7) 164 (4.8) 71 (2.1) 16 (0.5)

381 (11.2)

1.1 �0.2

0.1 �0.2

0.4 0.3 2.1

2.17 (1.42–3.32) 0.97 (0.79–1.18) 1.01 (0.87–1.16) 0.95 (0.77–1.19) 1.17 (0.85–1.61) 1.57 (0.84–2.94) 1.20 (1.05–1.38)

�0.001

Respiratory infection 238 (7.0) 252 (7.4) �0.4 0.94 (0.79–1.12) Other noncardiac and nonvascular 1126 (33.1) 1079 (31.7) 1.4 1.06 (0.98–1.15) Unspecified 20 (0.6) 18 (0.5) 0.1 1.11 (0.59–2.10)

No. of patients hospitalized 2184 (64.3) 2282 (67.1) �2.8 0.92 (0.87–0.98) 0.006

No. of hospitalizations 6356 6777

530 � Fe b r u a r y 2 0 , 1 9 9 7

T h e New E n g l a n d Jo u r n a l o f Me d i c i n e

larged hearts and those in NYHA functional class III or IV (Table 4).

Serum Digoxin Levels

Among the 1485 patients in the digoxin group for whom blood samples obtained more than six hours after the last dose were available, the mean steady- state serum digoxin level was 0.86 ng per milliliter (1.10 nmol per liter) at the 1-month visit and 0.80 ng per milliliter (1.02 nmol per liter) at the 12-month visit. At the one-month visit, the mean serum digox- in levels were 0.76 ng per milliliter (0.97 nmol per liter) in the patients receiving 0.125 mg of digoxin per day, 0.89 ng per milliliter (1.14 nmol per liter) in those receiving 0.250 mg per day, 0.88 ng per milli- liter (1.13 nmol per liter) in those receiving 0.375 mg per day, and 0.88 ng per milliliter in those receiv- ing 0.500 mg per day. At one month, 88.3 percent of the patients in the digoxin group had serum dig- oxin levels within the therapeutic range of 0.5 to 2.0 ng per milliliter (0.6 to 2.6 nmol per liter).

Adherence to the Study Regimen

At randomization, the median daily dose of the assigned study drug was 0.250 mg in both treat- ment groups. At one year, 85.6 percent of the pa-

tients in the digoxin group were taking the study drug, and 82.9 percent of the patients in the place- bo group were taking placebo. At the final study vis- it, 70.8 percent of the surviving patients in the dig- oxin group were taking the study drug, and an additional 10.3 percent were taking open-label dig- oxin. In the placebo group, 67.9 percent of the sur- viving patients were taking placebo and 15.6 percent were taking open-label digoxin. Open-label digoxin was used at some time during the trial by 14.2 per- cent of patients in the digoxin group as compared with 22.0 percent of those in the placebo group (P�0.001). The primary reasons for discontinuing the study drug were the use of open-label digoxin to treat worsening heart failure (rate of discontinuation at one year, 3.0 percent in the digoxin group vs. 6.4 percent in the placebo group; by the end of the trial, 6.7 percent vs. 11.0 percent) and atrial fibrillation (rate of discontinuation at one year, 1.5 percent vs. 1.6 percent; by the end of the trial, 2.8 percent vs. 3.4 percent).

Digoxin Toxicity

More patients had suspected digoxin toxicity in the digoxin group than in the placebo group (11.9 percent vs. 7.9 percent). Among these patients, 16.5

Figure 3. Incidence of Death or Hospitalization Due to Worsening Heart Failure in the Digoxin and Placebo Groups. The number of patients at risk at each four-month interval is shown below the figure.

0 52

4 8 12 16 20 24 28 32 36 40 44 48

Months

D e

a th

o r

H o

sp it

a li

za ti

o n

D u

e t

o W

o rs

e n

in g

H e

a rt

F a

il u

re (

% )

NO. OF PATIENTS AT RISK Placebo Digoxin

3403

Placebo

Digoxin P � 0.001

2915 2674 2473 2328 2197 2071 1954 1659 1397 1111 859 546 250 3397 3120 2888 2696 2544 2392 2241 2115 1825 1521 1188 916 578 255

E F F E C T O F D I G OX I N O N M O R TA L I T Y A N D M O R B I D I T Y I N PAT I E N T S W I T H H E A R T FA I LU R E

Vo l u m e 3 3 6 N u m b e r 8 � 531

percent of those in the digoxin group were hospital- ized, as compared with 11.4 percent of those in the placebo group. The most common reasons for sus- pected digoxin toxicity were ventricular fibrillation or tachycardia (1.1 percent in the digoxin group vs. 0.8 percent in the placebo group; risk ratio, 1.40; 95 percent confidence interval, 0.84 to 2.30; P � 0.20), supraventricular arrhythmia (2.5 percent vs. 1.2 per- cent; risk ratio, 2.10; 95 percent confidence interval, 1.45 to 3.07; P�0.001), and second- or third-degree atrioventricular block (1.2 percent vs. 0.4 percent; risk ratio, 2.87; 95 percent confidence interval, 1.56 to 5.28; P�0.001). The increase in the risk of dig- oxin toxicity in the digoxin group was similar for all subgroups.

Ancillary Trial (Left Ventricular Ejection Fraction, >0.45)

In the ancillary trial, there were no significant dif- ferences in base-line characteristics between the 492 patients assigned to digoxin and the 496 patients as- signed to placebo. There were 115 deaths in the dig- oxin group (23.4 percent) and 116 deaths in the placebo group (23.4 percent; risk ratio, 0.99; 95 percent confidence interval, 0.76 to 1.28). With re- gard to the combined outcome of death or hospital- ization due to worsening heart failure, the results in

the ancillary trial (risk ratio, 0.82; 95 percent confi- dence interval, 0.63 to 1.07) were consistent with the findings of the main trial.

DISCUSSION

In the main trial, in which patients with left ven- tricular ejection fractions of 0.45 or lower were studied, we found that digoxin had no effect on overall mortality when it was added to diuretics and angiotensin-converting–enzyme inhibitors to treat heart failure. There were fewer deaths due to wor- sening heart failure in the digoxin group. Although death attributable to other cardiac causes was not specified in advance as a study outcome, there was a statistically significant difference between the study groups (P � 0.04) with regard to that outcome. The risk of hospitalization, especially for worsening heart failure, was reduced with digoxin treatment. When the combined outcome was analyzed, the incidence of death from worsening heart failure or hospitaliza- tion for that diagnosis was markedly reduced. Small- er studies, such as the Randomized Assessment of Digoxin on Inhibitors of the Angiotensin-Convert- ing Enzyme (RADIANCE) and Prospective Ran- domized Study of Ventricular Failure and the Effica- cy of Digoxin (PROVED) trials, have suggested that

*Numbers of patients shown for the subgroups do not all add up to the total number in the group because of missing data for some patients.

†Absolute differences were calculated by subtracting the percentage of patients with one or more events in the placebo group from the corresponding percentage of patients in the digoxin group (before values were rounded). P values for the interaction of the variables shown with the study assignments were as follows: ejection fraction, P� 0.02; previous digoxin use, P � 0.54; cause of heart failure, P� 0.11; cardiothoracic ratio, P� 0.02; and NYHA class, P� 0.02.

‡Risk ratios and confidence intervals (CI) were estimated from the Cox proportional-hazards model that used the date of the first event. P values for the interaction of the variables shown with the study assignments were as follows: ejection fraction, P � 0.05; previous digoxin use, P� 0.60; cause of heart failure, P� 0.06; cardiothoracic ratio, P� 0.10; and NYHA class, P � 0.15.

TABLE 4. EFFECT OF THE STUDY DRUG ON THE OCCURRENCE OF DEATH OR HOSPITALIZATION DUE TO WORSENING HEART FAILURE.

VARIABLE DIGOXIN* PLACEBO*

ABSOLUTE DIFFERENCE (95% CI)†

RISK RATIO (95% CI)‡

no. of patients with �1 event/ no. randomized (%) %

Ejection fraction 0.25–0.45 �0.25

613/2270 (27.0) 428/1127 (38.0)

735/2273 (32.3) 556/1130 (49.2)

�5.3 (�8.0 to �2.7) �11.2 (�15.3 to �7.2)

0.80 (0.72 to 0.89) 0.68 (0.60 to 0.77)

Previous use of digoxin Yes No

550/1498 (36.7) 491/1899 (25.9)

688/1519 (45.3) 603/1884 (32.0)

�8.6 (�12.1 to �5.1) �6.2 (�9.0 to �3.3)

0.74 (0.66 to 0.83) 0.77 (0.68 to 0.86)

Cause of heart failure Ischemic Nonischemic

731/2405 (30.4) 306/983 (31.1)

873/2398 (36.4) 413/996 (41.5)

�6.0 (�8.7 to �3.3) �10.3 (�14.5 to �6.1)

0.79 (0.72 to 0.88) 0.67 (0.58 to 0.77)

Cardiothoracic ratio �0.55 �0.55

600/2220 (27.0) 441/1176 (37.5)

724/2233 (32.4) 567/1170 (48.5)

�5.4 (�8.1 to �2.7) �11.0 (�14.9 to �7.0)

0.79 (0.71 to 0.88) 0.69 (0.61 to 0.78)

NYHA class I or II III or IV

601/2275 (26.4) 438/1118 (39.2)

739/2296 (32.2) 552/1105 (50.0)

�5.8 (�8.4 to �3.1) �10.8 (�14.9 to �6.7)

0.78 (0.70 to 0.87) 0.70 (0.61 to 0.79)

Overall study population 1041/3397 (30.6) 1291/3403 (37.9) �7.3 (�9.5 to �5.0) 0.75 (0.69 to 0.82)

532 � Fe b r u a r y 2 0 , 1 9 9 7

T h e New E n g l a n d Jo u r n a l o f Me d i c i n e

worsening heart failure and hospitalization occurred less often in patients treated with digoxin.4,5 Al- though our trial showed that digoxin had no effect on mortality, studies of other inotropic agents not related to glycosides, such as dobutamine, beta-ago- nists, milrinone, and enoximone, have demonstrated excess mortality.6,9-11

The population studied in the main trial rep- resented a wide spectrum of patients with heart failure. A large proportion of the study patients received background therapy with angiotensin-con- verting–enzyme inhibitors (94.4 percent), diuretics (81.7 percent), one or the other of these classes of drugs (98.4 percent), or both (77.7 percent). Among the patients in the study, 22.3 percent were women, and patients with diverse causes of heart failure and a broad range of symptoms were includ- ed. Although only 2 percent of patients were in NYHA functional class IV, 30.6 percent were in class III. Inclusion of patients irrespective of base-line ejection fraction distinguishes this trial from prior studies.

Although there were more patients with suspect- ed digoxin toxicity in the digoxin group (11.9 per- cent, as compared with 7.9 percent in the placebo group), the proportion of patients actually hospital- ized was low (2.0 percent vs. 0.9 percent over a pe- riod of 3.5 years). This excess of suspected cases took the form of new episodes of ventricular tachy- cardia or fibrillation, supraventricular arrhythmia, and advanced atrioventricular block. The vast major- ity of the study patients, however (88.3 percent), had serum digoxin levels in the therapeutic range at the one-month visit, and only 2 percent had levels exceeding 2.0 ng per milliliter.

Subgroups were assessed for differences in the benefits and risks of digoxin. The reduction in the occurrence of either death or hospitalization due to worsening heart failure was seen at all levels of the left ventricular ejection fraction, but it was greatest in patients with ejection fractions of 0.25 or lower, those who had enlarged hearts, and those in NYHA functional class III or IV.

In conclusion, digoxin had no effect on overall mortality in patients receiving diuretics and angio- tensin-converting–enzyme inhibitors, but it did re- duce the overall number of hospitalizations and the combined outcome of death or hospitalization at- tributable to worsening heart failure. In clinical prac- tice, digoxin therapy is likely to affect the frequency of hospitalization, but not survival.

We are indebted to the study patients for their willingness to vol- unteer for the trial; to the several hundred physicians who referred patients, for their cooperation; and to Glaxo Wellcome for supplying the drug and placebo.

APPENDIX

The following investigators participated in the study. United States: G. Perry, E. Brown, R. Thornton, T. Shiva, J. Hubbard, K.R. Reddy, J.E. Doherty III, F.P. Cardello, A. Fast, M.J. Radford, J.S. Folger, G. Bhaskar, R.G. Zoble, V. Sridharan, M.R. Sridharan, R.R. Loungani, M. Gheorghi- ade, A. Hsieh, C. Tommaso, M. Mansuri, M.A. Guess, S. Akhtar, S. Wag- ner, K. Hagan, K.M. McIntyre, P. Ruble, M. Moten, A. Riley, G. Pierpont, I. Anand, G. Patel, B.S. Puram, B.R. Eladasari, J. Karnegis, E. Gillie, M.H. Crawford, W.F. Graettinger, A. Shah, J. Sacco, M.A. Chaudhry, D. Dolen, N. El-Sherif, S. Bekheit-Saad, E.E. Campos, J.G. Greene, V. Khanijo, U. Kumar, G.I. Mallis, S. Mookherjee, M. Dibner-Dunlap, S.C. Gupta, K. Danisa, U. Thadani, A. Tan, M. Rajachar, M. Amidi, M.V. O’Reilly, C.A. Hassapoyannes, M.L. Davies, V.A. Kumar, D. Okerson, K.B. Ra- manathan, B. Putatunda, A. Gollapudi, A. Montero, P.K. Mohanty, N. Jar- mukli, C. Lui, S. Thagirisa, R.W. Lee, T.R. Glatter, K. Bodine, D. Roberts, M. Bertoglio, G.W. Dennish, R.J. Sarma, G. Gregoratos, D.C. Rausch, W.A. Pitt, B.M. Kennelly, D. Fahrenholtz, R. Gordon, L. Horwitz, R. Rothbart, P. Nutting, L. Lutz, D.L. Copen, A. Rashkow, J. Babb, L. Van Voorhees, A. Silverman, M.E. Stillabower, A.B. Miller, H.T. Nguyen- Pho, R.E. Safford, S. Fishman, J.C. Neiman, M. Stein, J.C. Dominguez, G.T. Abernathy, P.H. Nair, L.S. Goodman, T.H. Cook, W.J. Wickemeyer, D.M. Berkson, J. Mathew, H.G. Richman, D.L. Lubell, R. Lang, E.J. Zajac, R. Rosenstein, M.A. Silver, J.G. Shanes, K.J. Kelly, M.H. Pequignot, R. Campbell, P.C. Kirlin, D.B. Ziperman, D.M. Denny, L. Gamble, M. Weiss, P.K. Kaimal, R.W. Dhurandhar, H. Ventura, M.L. Godley, C. Pu, E.C. Schick, Jr., D.J. Barbour, D. Salmon, M. Goldstein, M.B. Ef- fron, J.L. Fleg, K.L. Baughman, R. Weiss, K. Ericson, W.A. Sturrock, J. Heinsimer, G.C. Timmis, S. Smith, D.D. Shrestha, W.F.C. Duvernoy, J. Hickner, B.K. Lewis, T.K. O’Brien, S.A. Yarows, H.J. Willens, D.J. Mast, T.H. Johnson, R.J. Rodeheffer, M. Seifert, L. Swenson, P. Denes, R. Guda- pati, F.R. Charles, M.W. Rich, V. Beckham, W.P. Hamilton, P.B. Abele, D.M. Harper, R.L. O’Kelly, A.D. Forker, F.R. Kahl, B.W. Garrou, Sr., T.E. Klang, K. Popio, S. Mohiuddin, E.M. Pollak, Jr., A. Chiaramida, J.J. Greg- ory, L.A. Papa, J. Abrams, S. Ung, C.K. Jutila, R.P. Croke, S. Chiaramida, R. Lucariello, D.A. Rim, M.H. Goldberger, R.M. Kohn, S. Graham, E.F. Philbin, K. Sheesley, A. Nafziger, G. Macina, J.T. Hsueh, S. Zoneraich, A. Binder, C. Weinstein, J. Morrison, A. Cameron, E. Vanderbush, J. Brown, P.N. Pande, E. Lader, R. Kay, D. Bloomfield, T. Costantino, D.E. Heiselman, S. Radwany, M.R. Smith, R.E. Hobbs, T.D. Fraker, Jr., T.R. Frerking, A.C. De Leon, Jr., L.G. Christie, Jr., M. Toren, J. Grover, F.D. McBarron, D.E. Harris, R.W. McLean, L. Morris, J. Zatuchni, J.P. Boehmer, B.S. Clemson, H. Lipshutz, R.S. Small, R. Ufret, J.E. Lugo- Rodriquez, A.H. Khan, M. Yousefian, G.C. Friesinger, D. Ely, J.A. Farmer, J.B. Young, R.M. Payne, R.E. Fowles, A.B. Lee, Y. Shalev, I. Al-Bitar, G.S. Schroeder, L. Radant, T.L. Hankey, S. Rezkalla, and D.L. Groden. Cana- da: K.K. Teo, D.P. Humen, D. Wong, P.V. Greenwood, T. Talibi, W. Hui, W.P. Klinke, M.P.J. Senaratne, M. Goeres, D.F. Hughes, M.A.R. Sayeed, D.L. Roth, I. Belenkie, D.E. Manyari, K. Borgersen, L. Read, D. Kinloch, R. Reid, N.C. Barber, B.A. Horner, G. Kenefick, R.A. Kuritzky, J.R. Imrie, K.R. Wagner, S.W. Rabkin, H.F. Mizgala, L. Tarry, A. Dodek, J.M. Kornder, T. Ashton, D. Barr, J. Dufton, R. Sweeny, A. Morris, R. Bessoudo, D. Marr, J.R. Milton, B. Thompson, V. Robinson, B.A. Sussex, M. Tobin, D.P. McMahon, D. Folkins, R. Crowell, L.D. Lalonde, C. Koilpillai, R.J. Hatheway, M.G. O’Reilly, T. Machel, I. Hack, J.W. Stewart, P.H. Tanser, B. Sullivan, S. Hagar, B. Quinn, R.A. Davies, M.G. Baird, W.L. Williams, M. Le May, L. Higginson, M. Turek, R.A. Sochowski, A. Weeks, M.H. Ja- cobs, R. Baigrie, M.A. DeVilla, G. Vertes, B. Bozek, J.E. Goode, A.J. Ricci, J. Swan, D.A. Fell, S.N. Levinoff-Roth, J.M.O. Arnold, L. Patrick, R.F. Southern, C. Rinne, D. Brisbin, J.P. DeYoung, T. Baitz, S. Vizel, J. Minkowitz, P. Letarte, Y.K. Chan, K.K. Kwok, S. Nawaz, F. Ganjavi, L. Yao, J. Misterski, D.L. Raco, A.R. Hess, G. Kuruvilla, L. Silverberg, D. Borts, J.C. Fulop, M.E. Weingert, R.P. Carter, B.M. Sahay, J.E. Hickey, G. Lalonde, G. Gosselin, M.G. Bourassa, C. Goulet, M. Joyal, M. Methe, G. Honos, D. Fitchett, A. Serpa, F. Sestier, G. Roberge, Y. Latour, C. Ron- deau, D. Gossard, J.H.F. Lenis, R. Brossoit, L. Simard, F. Delage, P. Auger, D. Saulnier, P. Talbot, J. Beaudoin, J. Campeau, G. Pruneau, M. Tamilia, G. Boutros, B. Comeau, R. Starra, R. St.-Hilaire, J. Veilleux, M. Mercier, J.F. Poulin, K. MacLellan, S. Lepage, J.L. Rouleau, N. Bruinsma, C. Levesque, R. De Larochelliere, M. LeBlanc, S. Kouz, G.S. Kiwan, M. Laforest, J. Brooks, G. Bouchard, P.B. Gervais, J. Brophy, J. Gagnon, N. Habib, A.K. Basu, A.T. Lutterodt, and M. Khouri.

Steering Committee: R. Gorlin, D. Egan, R. Garg, S. Yusuf, T. Mon- tague, T.W. Smith, J.N. Cohn, G.R. Dagenais, R. Davies, D.E. Johnstone, C. Fye, M.R. Sather, D. Deykin, G. Francis, J.F. Collins, W.O. Williford, and B.N. Singh. Data and Safety Monitoring Board: D. Bristow, H.T. Engelhardt, M. Gent, W.B. Hood, S. Jones, P. Meier, B. Pitt, and D. Wa- ters. Canadian Regional Centers: S. Barnhill, R. Miles, E. Kent, M.J. Sayles, C. Liuni, A. Baker, S. Hagar, D. LaForge, L. Harris, and S. Martin.

E F F E C T O F D I G OX I N O N M O R TA L I T Y A N D M O R B I D I T Y I N PAT I E N T S W I T H H E A R T FA I LU R E

Vo l u m e 3 3 6 N u m b e r 8 � 533

Human Rights Committee: S. Jones, L.J. Appel, M.M. Arthur, D.A. Highfield, J.P. Libonati, M.S. Moore, E. Perez, and J.D. Rubin. National Heart, Lung, and Blood Institute Project Office: N. Geller, E. Harris, S. Hunsberger, and P. Mills. Department of Veterans Affairs Cooperative Studies Program: S. Baldwin, A. Burns, H. Caleb, F. Chacon, D.R. Cline, V. Duvall, W. Gagne, J. Gold, S. Harris, R. Hockenbrock, R.A. Horney, C. Howell, P.C. Huang, L.M. Jadwin, J. King, S. Maple, G. Martinez, J. O’Quinn, P. Sexton, and M.E. Spence. Central Laboratory: Smith- Kline Beecham, Van Nuys, Calif.

REFERENCES

1. Kelly RA, Smith TW. Digoxin in heart failure: implications of recent tri- als. J Am Coll Cardiol 1993;22:Suppl A:107A-112A. 2. Yusuf S, Wittes J, Bailey K, Furberg C. Digitalis — a new controversy regarding an old drug: the pitfalls of inappropriate methods. Circulation 1986;73:14-8. 3. Jaeschke R, Oxman AD, Guyatt GH. To what extent do congestive heart failure patients in sinus rhythm benefit from digoxin therapy? A sys- tematic overview and meta-analysis. Am J Med 1990;88:279-86. 4. Packer M, Gheorghiade M, Young JB, et al. Withdrawal of digoxin from patients with chronic heart failure treated with angiotensin-converting– enzyme inhibitors. N Engl J Med 1993;329:1-7.

5. Uretsky BF, Young JB, Shahidi FE, Yellen LG, Harrison MC, Jolly MK. Randomized study assessing the effect of digoxin withdrawal in patients with mild to moderate chronic congestive heart failure: results of the PROVED Trial. J Am Coll Cardiol 1993;22:955-62. 6. Yusuf S, Garg R, Held P, Gorlin R. Need for a large randomized trial to evaluate the effects of digitalis on morbidity and mortality in congestive heart failure. Am J Cardiol 1992;69:64G-70G. 7. The Digitalis Investigation Group. Rationale, design, implementation, and baseline characteristics of patients in the DIG trial: a large, simple, long-term trial to evaluate the effect of digitalis on mortality in heart fail- ure. Control Clin Trials 1996;17:77-97. 8. Jelliffe RW, Brooker GA. A nomogram for digoxin therapy. Am J Med 1974;57:63-8. 9. Yusuf S. Obtaining reliable information from randomized controlled tri- als in congestive heart failure and left ventricular dysfunction. In: Brach- man W, Dietz R, Kubler W, eds. Heart failure and arrhythmias. Berlin, Ger- many: Springer-Verlag, 1990:147-60. 10. Yusuf S, Teo KK. Inotropic agents increase mortality in patients with congestive heart failure. Circulation 1990;82:Suppl III:III-673. abstract. 11. Packer M, Carver JR, Rodeheffer RJ, et al. Effect of oral milrinone on mortality in severe chronic heart failure. N Engl J Med 1991;325: 1468-75.