JOURNAL OF THE AMERICAN COLLEGE OF CARDIOLOGY VOL. 74, NO.

5, 2019

PUBLISHED BY ELSEVIER ON BEHALF OF THE AMERICAN

COLLEGE OF CARDIOLOGY FOUNDATION

Digoxin Discontinuation and

Outcomes in Patients With Heart Failure
With Reduced Ejection Fraction
Awais Malik, MD,a,b Ravi Masson, MD,a,b Steven Singh, MD,a,b Wen-Chih Wu, MD,c,d Milton Packer, MD,e
Bertram Pitt, MD,f Finn Waagstein, MD, PHD,g Charity J. Morgan, PHD,a,h Richard M. Allman, MD,h,i
Gregg C. Fonarow, MD,j Ali Ahmed, MD, MPHa,b,i

ABSTRACT

BACKGROUND The deleterious effects of discontinuation of digoxin on outcomes in ambulatory patients with chronic
heart failure (HF) with reduced ejection fraction (HFrEF) receiving angiotensin-converting enzyme inhibitors are
well-documented.

OBJECTIVES The authors sought to determine the relationship between digoxin discontinuation and outcomes in
hospitalized patients with HFrEF receiving more contemporary guideline-directed medical therapies including
beta-blockers and mineralocorticoid receptor antagonists.

METHODS Of the 11,900 hospitalized patients with HFrEF (EF #45%) in the Medicare-linked OPTIMIZE-HF (Organized
Program to Initiate Lifesaving Treatment in Hospitalized Patients with Heart Failure) registry, 3,499 received
pre-admission digoxin, which was discontinued in 721 patients. Using propensity scores for digoxin discontinuation,
estimated for each of the 3,499 patients, a matched cohort of 698 pairs of patients, balanced on 50 baseline charac-
teristics (mean age 76 years; mean EF 28%; 41% women; 13% African American; 65% on beta-blockers) was assembled.

RESULTS Four-year post-discharge, digoxin discontinuation was associated with significantly higher risks of HF read-
mission (hazard ratio [HR]: 1.21; 95% confidence interval [CI]: 1.05 to 1.39; p ¼ 0.007), all-cause readmission (HR: 1.16;
95% CI: 1.04 to 1.31; p ¼ 0.010), and the combined endpoint of HF readmission or all-cause mortality (HR: 1.20; 95% CI:
1.07 to 1.34; p ¼ 0.002), but not all-cause mortality (HR: 1.09; 95% CI: 0.97 to 1.24; p ¼ 0.163). Discontinuation of
digoxin was associated with a significantly higher risk of all 4 outcomes at 6 months and 1 year post-discharge. At
30 days, digoxin discontinuation was associated with higher risks of all-cause mortality (HR: 1.80; 95% CI: 1.26 to 2.57;
p ¼ 0.001) and the combined endpoint (HR: 1.36; 95% CI: 1.09 to 1.71; p ¼ 0.007), but not of HF readmission (HR: 1.19;
95% CI: 0.90 to 1.59; p ¼ 0.226) or all-cause readmission (HR: 1.03; 95% CI: 0.84 to 1.26; p ¼ 0.778).

CONCLUSIONS Among hospitalized older patients with HFrEF on more contemporary guideline-directed medical
therapies, discontinuation of pre-admission digoxin therapy was associated with poor outcomes.
(J Am Coll Cardiol 2019;74:617–27) Published by Elsevier on behalf of the American College of Cardiology Foundation.

From the aVeterans Affairs Medical Center, Washington, DC; bGeorgetown University, Washington, DC; cVeterans Affairs Medical
Center, Providence, Rhode Island; dBrown University, Providence, Rhode Island; eBaylor University Medical Center, Dallas, Texas;
f
University of Michigan, Ann Arbor, Michigan; gUniversity of Gothenburg, Gothenburg, Sweden; hUniversity of Alabama at
Birmingham, Birmingham, Alabama; iGeorge Washington University, Washington, DC; and the jUniversity of California, Los
Angeles, Los Angeles, California. The content is solely the responsibility of the authors and does not necessarily represent the
official views of the Department of Veterans Affairs. Dr. Ahmed was supported in part by the National Institutes of Health through
Listen to this manuscript’s grants R01-HL085561, R01-HL085561-S, and R01-HL097047 from the National Heart, Lung, and Blood Institute. The OPTIMIZE-HF
audio summary by registry was sponsored by GlaxoSmithKline, but played no role in the design, conduct, analyses, or interpretation of the current
Editor-in-Chief study. Dr. Packer has served as a consultant for Abbvie, Akcea, Amgen, AstraZeneca, Boehringer Ingelheim, Bayer, Cardiorentis,
Dr. Valentin Fuster on Daiichi-Sankyo, Gilead, NovoNordisk, Pfizer, Relypsa, Sanofi, and Theravance. Dr. Pitt has been a consultant to Bayer, Sanofi,
JACC.org. AstraZeneca, KBP Pharmaceuticals, scPharmaceuticals, Sarfez, Relypsa/Vifor, and Cereno Scientific; has stock options in KBP
Pharmaceuticals, Relypsda, and Sarfez; and holds U.S. patent # 9931412 for site-specific delivery of eplerenone to the myocardium.
Dr. Fonarow has been a consultant to Abbott, Amgen, Bayer, Janssen, Medtronic, and Novartis; has received research funding
from Novartis; and was the principal investigator of the OPTIMIZE-HF registry. All other authors have reported that they have no
relationships relevant to the contents of this paper to disclose.

Manuscript received April 7, 2019; revised manuscript received May 23, 2019, accepted May 28, 2019.

ISSN 0735-1097/$36.00 https://doi.org/10.1016/j.jacc.2019.05.064

618 Malik et al. JACC VOL. 74, NO. 5, 2019

Digoxin Discontinuation and Outcomes in HFrEF AUGUST 6, 2019:617–27

H METHODS
ABBREVIATIONS eart failure (HF) is a leading cause
AND ACRONYMS of hospital admission and readmis-
sion (1). Digoxin is approved by the DATA SOURCE AND STUDY POPULATION. The
ACE = angiotensin-converting
enzyme
U.S. Food and Drug Administration for the OPTIMIZE-HF (Organized Program to Initiate Life-
treatment of mild-to-moderate HF to reduce saving Treatment in Hospitalized Patients With
AKI = acute kidney injury
the risk of HF-related hospitalizations and Heart Failure) registry is a national web-based regis-
ARB = angiotensin receptor
blocker emergency care. According to the American try of acute HF based on 48,612 HF hospitalizations
CI = confidence interval
College of Cardiology Foundation/American occurring in 259 hospitals across 48 states between
Heart Association HF guideline, digoxin March 1, 2003, and December 31, 2004 (14–18). The
GDMT = guideline-directed
medical therapies may be used, unless contraindicated, to registry contains extensive data on demographics,
HF = heart failure decrease hospitalizations due to worsening patient and hospital characteristics, quality of care,
HFrEF = heart failure with
HF in patients with HF with reduced ejection and short-term outcomes in a small subset of pa-
reduced ejection fraction fraction (HFrEF) (2). In the DIG (Digitalis tients. We obtained information about outcomes
HR = hazard ratio Investigation Group) trial, the largest ran- through probabilistic linking of OPTIMIZE-HF registry
MRA = mineralocorticoid domized controlled trial of digoxin in HF, with the Medicare data (19). The Medicare-linked
receptor antagonist digoxin reduced the risk of all-cause and HF OPTIMIZE-HF data included 25,345 unique patients
OR = odds ratio hospitalizations in patients with HFrEF but who were discharged alive, of whom 11,900 had HF
had no effect on all-cause mortality (3–5). with left ventricular EF #45% (Figure 1). We used this
The lack of a mortality benefit of digoxin in the DIG cutoff for EF because this was used in the DIG main
trial, taken together with the emergence of other trial to define HFrEF (3). We excluded 8,401 patients
evidence-based guideline-directed medical therapies not receiving digoxin before hospital admission.
(GDMT) with proven efficacy and effectiveness in Thus, our study cohort consisted of 3,499 patients
lowering the risks of both all-cause mortality and hos- who were receiving digoxin before hospital admis-
pitalization, has led to a dramatic decline in the use of sion. In 721 of these patients, digoxin was dis-
digoxin in patients with HFrEF (6). continued before hospital discharge.
ASSEMBLY OF A BALANCED COHORT. We used
SEE PAGE 628
propensity score matching to assemble a cohort in
Both a lower rate of initiation of digoxin therapy which patients with continuation and discontinua-
and a higher rate of discontinuation of pre-admission tion of pre-admission digoxin therapy would be
digoxin therapy may lead to the underutilization of balanced on key measured baseline characteristics. A
digoxin. The efficacy and effectiveness of digoxin in notable advantage of using the propensity score
lowering the risk of admission and readmission in approach is that as in a randomized controlled trial,
patients with HFrEF are well established (3–5,7–10). the process of cohort assembly is outcome-blinded
However, less is known about the effect of discon- and the balance in measured baseline characteristics
tinuation of digoxin in patients with HFrEF. Findings can be displayed in a tabular form (20,21). To achieve
from the RADIANCE (Randomized Assessment of the this goal, we first calculated propensity scores for
effect of Digoxin on Inhibitors of the Angiotensin- each of the 3,499 patients using a nonparsimonious
Converting Enzyme) and PROVED (Prospective Ran- multivariable logistic regression model (22–24). In the
domized Study of Ventricular Failure and the Efficacy model, discontinuation of digoxin was the dependent
of Digoxin) trials suggest that discontinuation of variable and 50 baseline characteristics listed in
digoxin therapy increased the risk of adverse out- Online Figure 1 were used as covariates. Using a
comes in ambulatory patients with chronic HFrEF matching algorithm described elsewhere (25), we
(11,12). A post hoc analysis of the DIG trial demon- matched 698, or 97%, of the 721 patients whose
strated that discontinuation of digoxin is associated digoxin was discontinued, with another 698 patients
with poor outcomes in ambulatory patients with whose digoxin was continued, but had the same
chronic HFrEF (13). Patients in these studies were probability or propensity scores for discontinuation.
receiving angiotensin-converting enzyme (ACE) in- Propensity score models are sample-specific ad-
hibitors, but not beta-blockers or mineralocorticoid justers and are not intended to be used for out-of-
receptor antagonists (MRAs). The objective of the sample prediction or estimation of coefficients, and
current study was to examine the relationship be- measures of fitness and discrimination are not
tween discontinuation of pre-admission digoxin important for the assessment of the model’s effec-
therapy and outcomes in a propensity score–matched tiveness (25,26). As such, we assessed our propensity
cohort of hospitalized patients with HFrEF receiving score model by estimating pre- and post-match be-
more contemporary GDMT. tween-group absolute standardized differences for
JACC VOL. 74, NO. 5, 2019 Malik et al. 619
AUGUST 6, 2019:617–27 Digoxin Discontinuation and Outcomes in HFrEF

the 50 measured baseline characteristics used in our

F I G U R E 1 Assembly of the Study Cohort
propensity score model (22–24). Absolute standard
difference values <10% suggest inconsequential re-
sidual imbalance and a value of 0% indicates no re- OPTIMIZE-HF hospitalizations (n = 48,612)
sidual imbalance. We also used the same logistic
regression model used for the estimation of pro-
pensity scores to estimate multivariable-adjusted Linked to Medicare fee-for-service data (n = 26,376)
odds ratios (ORs) for digoxin discontinuation associ-
ated with each of the 50 baseline characteristics.
ASSEMBLY OF A SENSITIVITY COHORT. Because
Discharged alive (n = 25,345)
worsening kidney function during hospitalization is
one of the common reasons why digoxin may be
Excluded patients with ejection
fraction >45% (n = 13,445)
discontinued in the inpatient setting, we assembled a
sensitivity cohort that accounted for acute kidney
Restricted to ejection fraction ≤45% (n = 11,900)
injury (AKI). Data on both admission and discharge
serum creatinine levels were available on 2,971 of the
Excluded patients not receiving
3,499 patients, of whom 418 (14%) had AKI defined as
digoxin upon admission (n = 8,401)
a rise of serum creatinine from admission to discharge
by $0.3 mg/dl. Digoxin was discontinued in 21% of
Receiving digoxin upon admission (n = 3,499)
the patients (622 of 2,971). We then calculated the
probability (propensity scores) of digoxin discontin-
uation for each of the 2,971 patients using the same
Digoxin continued on Digoxin discontinued on
regression model described in the preceding text,
discharge (n = 2,778) discharge (n = 721)
thus assembling a matched cohort of 602 pairs of
patients. In the model, we replaced discharge serum
creatinine with admission serum creatinine and AKI
Estimated propensity scores for digoxin discontinuation
so that patients would be balanced on AKI in the
and matched patients 1:1 by propensity scores
matched cohort.
OUTCOMES DATA. We examined the following out-
comes: HF readmission, all-cause readmission, all-
Digoxin continued on Digoxin discontinued on
cause mortality, and the combined endpoint of HF
discharge (n = 698) discharge (n = 698)
readmission or all-cause mortality. All outcomes were
examined at 30 days, 6 months, 1 year, and 4 years
after hospital discharge. Data on all outcome events
The Matched Cohort (n = 1,396)
and times to events were collected from the Medicare
data (19).
STATISTICAL ANALYSES. Descriptive analyses be- Flow chart displaying assembly of a matched cohort of patients with heart failure
with reduced ejection fraction, by discontinuation of digoxin therapy. OPTIMIZE-
tween group baseline characteristics were compared
HF ¼ Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with
using the Pearson chi-square and Wilcoxon rank sum Heart Failure registry.
tests, as appropriate. All outcome analyses were
conducted using matched data. Cox regression
models were used to estimate hazard ratios (HRs) and
95% confidence intervals (CIs) associated with 30 days for 30-day readmission), whichever occurred
digoxin discontinuation and time to events. We first. For mortality events, we used time to death or
assessed the assumption of the proportional hazard time to censoring for that follow-up period. Subgroup
by visual examinations of the log (minus log) curves. analyses were conducted to determine the homoge-
Kaplan-Meier survival analysis was used to generate neity of the association of digoxin discontinuation
survival plots by digoxin discontinuation. For read- and the combined endpoint of HF readmission or all-
mission events within any given time, we used time cause mortality in the matched cohort.
to event within that time (e.g., within 30 days for To examine whether significant associations
30-day readmission) for patients who had an event, observed in our matched data could be explained
and for those without a readmission, we censored by away by an unmeasured baseline characteristic,
time to death or time to end of follow-up (e.g., at we conducted formal sensitivity analyses using
620 Malik et al. JACC VOL. 74, NO. 5, 2019

Digoxin Discontinuation and Outcomes in HFrEF AUGUST 6, 2019:617–27

T A B L E 1 Baseline Characteristics of Patients With HFrEF

Before Propensity Score Matching (n ¼ 3,499) After Propensity Score Matching (n ¼ 1,396)

Digoxin Discontinuation Digoxin Discontinuation

No (n ¼ 2,778) Yes (n ¼ 721) p Value No (n ¼ 698) Yes (n ¼ 698) p Value

Age, yrs 75  11 76  11 0.047 76  11 76  11 0.827

Women 1,086 (39) 290 (40) 0.580 287 (41) 279 (40) 0.663
African American 389 (14) 105 (15) 0.700 84 (12) 103 (15) 0.135
Left ventricular ejection fraction, % 26  9 28  10 <0.001 28  9 28  10 0.973
Smoker in past 1 yr 359 (13) 81 (11) 0.223 83 (12) 80 (11) 0.803
Past medical history
Prior heart failure 2,621 (94) 683 (95) 0.691 658 (94) 662 (95) 0.637
Prior HF hospitalization in 6 months 602 (22) 161 (22) 0.702 154 (22) 153 (22) 0.948
Hypertension 1,805 (65) 465 (64) 0.809 459 (66) 449 (64) 0.575
Myocardial infarction 872 (31) 220 (31) 0.651 228 (33) 215 (31) 0.455
Coronary revascularization 1,097 (39) 281 (39) 0.801 278 (40) 268 (38) 0.583
Diabetes mellitus 1,135 (41) 304 (42) 0.525 302 (43) 292 (42) 0.588
Stroke/transient ischemic attack 430 (15) 101 (14) 0.327 103 (15) 100 (14) 0.820
Peripheral vascular disease 438 (16) 107 (15) 0.541 106 (15) 104 (15) 0.881
Atrial fibrillation 1,271 (46) 324 (45) 0.696 313 (45) 314 (45) 0.957
Ventricular arrhythmia 299 (11) 69 (10) 0.352 60 (9) 67 (10) 0.515
Implantable cardioverter defibrillator 341 (12) 86 (12) 0.800 88 (13) 82 (12) 0.623
Biventricular pacemaker 230 (8) 50 (7) 0.236 54 (8) 49 (7) 0.609
Chronic obstructive pulmonary disease 813 (29) 179 (25) 0.018 165 (24) 176 (25) 0.493
Anemia 407 (15) 94 (13) 0.270 93 (13) 92 (13) 0.937
Depression 251 (9) 70 (10) 0.577 64 (9) 65 (9) 0.926
Admission findings
Dyspnea at rest 1,183 (43) 291 (40) 0.281 300 (43) 284 (41) 0.385
Dyspnea on exertion 1,751 (63) 433 (60) 0.142 417 (60) 421 (60) 0.827
Orthopnea 838 (30) 165 (23) <0.001 156 (22) 165 (24) 0.567
Paroxysmal nocturnal dyspnea 503 (18) 95 (13) 0.002 87 (12) 94 (13) 0.577
Jugular venous pressure elevation 948 (34) 250 (35) 0.782 226 (32) 242 (35) 0.364
Third heart sound 378 (14) 92 (13) 0.552 83 (12) 88 (13) 0.683
Pulmonary rales 1,709 (62) 423 (59) 0.162 415 (59) 411 (59) 0.828
Peripheral edema 1,687 (61) 467 (65) 0.047 459 (66) 452 (65) 0.694
Pulse, beat/min 85  20 83  20 0.014 82  19 83  20 0.538
Systolic blood pressure, mm Hg 133  28 134  30 0.537 133  29 134  30 0.706
Diastolic blood pressure, mm Hg 74  17 72  18 0.006 71  17 72  18 0.553
Laboratory findings
Admission serum sodium, mEq/l 137  10 136  13 0.349 136  10 136  13 0.941
Admission hemoglobin, g/dl 13  3 12  2 0.040 12  2 12  2 0.795
Admission serum BNP, pg/ml 1,208 (792–1,527) 1,222 (830–1,553) 0.435 1,186 (772–1,557) 1,223 (784–1,561) 0.363
Discharge serum creatinine, mg/dl 1.6  0.9 1.7  1.0 0.008 1.7  1.2 1.7  1.0 0.729
Discharge medications
ACE inhibitors or ARBs 2,001 (72) 436 (60) <0.001 449 (64) 429 (61) 0.268
Beta-blockers 2,027 (73) 450 (62) <0.001 457 (65) 444 (64) 0.467
Aldosterone antagonists 559 (20) 106 (15) 0.001 105 (15) 106 (15) 0.940
Loop diuretic agents 2,354 (85) 513 (71) <0.001 507 (73) 510 (73) 0.857
Nitrates 795 (29) 170 (24) 0.007 157 (22) 170 (24) 0.411
Amlodipine 118 (4) 43 (6) 0.050 42 (6) 40 (6) 0.820
Other calcium-channel blockers 189 (7) 39 (5) 0.176 43 (6) 39 (6) 0.649
Antiarrhythmic agents 483 (17) 127 (18) 0.886 135 (19) 123 (18) 0.408
Hospital length of stay, days 4 (3–7) 5 (3–8) <0.001 5 (3–8) 5 (3–8) 0.332
Continued on the next page

Rosenbaum’s approach (27). From the 698 pairs of clearly had a longer survival or event-free survival
matched patients, we identified pairs in which we time than the other member (28). We then tested
could directly compare survival times within each whether, in the absence of a hidden bias, patients in
pair to determine whether one member of the pair the digoxin discontinuation group had shorter
JACC VOL. 74, NO. 5, 2019 Malik et al. 621
AUGUST 6, 2019:617–27 Digoxin Discontinuation and Outcomes in HFrEF

T A B L E 1 Continued

Before Propensity Score Matching (n ¼ 3,499) After Propensity Score Matching (n ¼ 1,396)

Digoxin Discontinuation Digoxin Discontinuation

No (n ¼ 2,778) Yes (n ¼ 721) p Value No (n ¼ 698) Yes (n ¼ 698) p Value

Hospital characteristics
Region
Midwest 958 (34) 225 (31) 244 (35) 222 (32)
Northeast 439 (16) 96 (13) 79 (11) 94 (13)
<0.001 0.490
South 928 (33) 250 (35) 236 (34) 240 (34)
West 453 (16) 150 (21) 139 (20) 142 (20)
Size, number of beds 375 (267–500) 375 (270–500) 0.136 375 (280–500) 375 (270–500) 0.198
Academic center 1,410 (51) 353 (49) 0.390 365 (52) 345 (49) 0.284

Values are mean  SD, n (%), or median (interquartile range). The p values comparing medians are based on nonparametric independent sample median test.
ACE ¼ angiotensin-converting enzyme; ARB ¼ angiotensin receptor blocker; BNP ¼ B-type natriuretic peptide; HF ¼ heart failure; HFrEF ¼ heart failure with reduced ejection
fraction.

survival time than their matched counterparts. A readmission (HR: 1.21; 95% CI: 1.05 to 1.39;
significant sign-score test provides strong evidence of p ¼ 0.007), all-cause readmission (HR: 1.16; 95% CI:
a relationship between discontinuation of digoxin 1.04 to 1.31; p ¼ 0.010), and the combined
and time to a particular event. The sign-score test is endpoint of HF readmission or all-cause mortality
used to calculate “sensitivity bounds” for a hypo- (HR: 1.20; 95% CI: 1.07 to 1.34; p ¼ 0.002), but
thetical unmeasured confounder to determine how not with all-cause mortality (HR: 1.09; 95% CI: 0.97
much it would need to increase the odds of digoxin to 1.24; p ¼ 0.163) (Table 2, Central Illustration,
discontinuation to explain away its significant asso- Online Figure 2). The association between digoxin
ciations with outcomes. Our sensitivity analysis
assumed that the potential unmeasured confounder
is a binary baseline characteristic that is a near perfect
T A B L E 2 Outcomes in 1,396 Propensity Score–Matched Hospitalized Patients With HFrEF
predictor of the outcomes, which is also not strongly
correlated with any of the 50 baseline characteristics Events by Digoxin
Discontinuation at
used in our propensity score model. However, sensi- Hospital Discharge,
tivity analysis cannot determine whether such an n (%)
Hazard Ratio Associated
unmeasured confounder exists. All statistical tests No Yes With Digoxin Discontinuation,
(n ¼ 698) (n ¼ 698) (95% CI); p Value
were 2-tailed, and a p value <0.05 was considered
30-day outcomes
significant. All statistical analyses were conducted
HF readmission 89 (13) 101 (14) 1.19 (0.90–1.59); 0.226
using IBM SPSS Statistics for Windows software,
All-cause readmission 195 (28) 193 (28) 1.03 (0.84–1.26); 0.778
version 24 (IBM, Armonk, New York). All-cause mortality 47 (7) 82 (12) 1.80 (1.26–2.57); 0.001
HF readmission or all-cause mortality 131 (19) 171 (24) 1.36 (1.09–1.71); 0.007
RESULTS 6-month outcomes
HF readmission 205 (29) 246 (35) 1.31 (1.08–1.57); 0.005

The 1,396 matched patients had a mean age of 76 All-cause readmission 377 (54) 409 (59) 1.18 (1.03–1.36); 0.019
All-cause mortality 181 (26) 215 (31) 1.25 (1.02–1.52); 0.028
 11 years and left ventricular ejection fraction of 28 
HF readmission or all-cause mortality 321 (46) 377 (54) 1.28 (1.10–1.48); 0.001
9%; 41% were female, and 13% were African Amer-
1-yr outcomes
ican. Before matching, patients in the digoxin HF readmission 276 (40) 317 (45) 1.28 (1.09–1.51); 0.003
discontinuation group were older, had higher left All-cause readmission 472 (68) 483 (69) 1.15 (1.02–1.31); 0.028
ventricular ejection fraction, and fewer of these pa- All-cause mortality 251 (36) 287 (41) 1.21 (1.02–1.43); 0.028
tients had received GDMT for HFrEF (Table 1). After HF readmission or all-cause mortality 415 (59) 475 (68) 1.27 (1.11–1.45); <0.001
propensity score matching, absolute standardized 4-yr outcomes
HF readmission 391 (56) 407 (58) 1.21 (1.05–1.39); 0.007
differences for all 50 baseline characteristics
All-cause readmission 595 (85) 579 (83) 1.16 (1.04–1.31); 0.010
were <10%, suggesting inconsequential residual bias
All-cause mortality 498 (71) 510 (73) 1.09 (0.97–1.24); 0.163
(Online Figure 1).
HF readmission or all-cause mortality 607 (87) 624 (89) 1.20 (1.07–1.34); 0.002

4-YEAR OUTCOMES. Digoxin discontinuation was CI ¼ confidence interval; other abbreviations as in Table 1.
associated with significantly higher risks of HF
622 Malik et al. JACC VOL. 74, NO. 5, 2019

Digoxin Discontinuation and Outcomes in HFrEF AUGUST 6, 2019:617–27

C E N T R A L IL L U ST R A T I O N Kaplan-Meier Plots for Outcomes by Digoxin Discontinuation

1.0 HR (95% CI): 1.21 (1.05–1.39); p = 0.007

Heart Failure Readmission

0.8 Digoxin Discontinuation

0.6
Digoxin Continuation
0.4

0.2
6-Month HR (95% CI):
1.31 (1.08–1.57); p = 0.005
0.0
0 1 2 3 4
Follow-Up (Years)
Number at risk
Digoxin Discontinuation 698 223 134 94 74
Digoxin Continuation 698 283 180 129 91

1.0 HR (95% CI): 1.09 (0.97–1.24); p = 0.163

0.8
All-Cause Mortality

Digoxin Discontinuation
0.6

0.4 Digoxin Continuation

0.2
6-Month HR (95% CI):
1.25 (1.02–1.52); p = 0.028
0.0
0 1 2 3 4
Follow-Up (Years)
Number at risk
Digoxin Discontinuation 698 411 308 236 188
Digoxin Continuation 698 447 336 254 200
Malik, A. et al. J Am Coll Cardiol. 2019;74(5):617–27.

This study assessed the relationship of discontinuation of pre-admission digoxin therapy with heart failure readmission (top) and all-cause
mortality (bottom) in 698 pairs of propensity score–matched patients with heart failure with reduced ejection fraction. During 4 years of
follow-up, digoxin discontinuation was associated with a significantly higher risk of heart failure readmission but not of all-cause mortality,
compared with patients whose digoxin therapy was continued. During the first 6 months of follow-up, digoxin discontinuation was asso-
ciated with a significantly higher risk of both outcomes. CI ¼ confidence interval; HR ¼ hazard ratio.

discontinuation and the combined endpoint of HF hospital discharge (Figure 2). HRs (95% CIs) for 4-
readmission or all-cause mortality at 4 years was year combined endpoint associated with digoxin
homogeneous among various clinically relevant discontinuation in the subgroups receiving and not
subgroups of patients, except by beta-blocker use at receiving beta-blockers were 1.11 (0.96 to 1.27;
JACC VOL. 74, NO. 5, 2019 Malik et al. 623
AUGUST 6, 2019:617–27 Digoxin Discontinuation and Outcomes in HFrEF

F I G U R E 2 Forest Plots for Subgroup Analyses of Combined Outcome by Digoxin Discontinuation

Total Patients Digoxin Discontinuation Hazard Ratio P value

(N = 1,396) No (n = 698) Yes (n = 698) (95% CI) Effect Interaction
Discontinuation
of digoxin worse
Age
<80 years (n = 878) 366/435 (84) 384/443 (87) 1.15 (0.99 - 1.32) 0.063
0.276
≥80 years (n = 518) 241/263 (92) 240/255 (94) 1.31 (1.10 - 1.57) 0.003
Sex
Male (n = 830) 361/411 (88) 374/419 (89) 1.19 (1.03 - 1.38) 0.018
0.940
Female (n = 566) 246/287 (86) 250/279 (90) 1.20 (1.01 - 1.43) 0.043
Race
Non-African American (n = 1,209) 536/614 (87) 529/595 (89) 1.18 (1.05 - 1.33) 0.007
0.612
African American (n = 187) 71/84 (85) 95/103 (92) 1.28 (0.94 - 1.74) 0.120
Systolic Blood Pressure
<120 mm Hg (n = 495) 229/255 (90) 224/240 (93) 1.27 (1.05 - 1.52) 0.012
0.466
≥120 mm Hg (n = 901) 378/443 (85) 400/458 (87) 1.18 (1.03 - 1.36) 0.020
Heart Rate
<70 beats/min (n = 439) 185/219 (84) 191/220 (87) 1.17 (0.95 - 1.43) 0.139
0.695
≥70 beats/min (n = 957) 422/479 (88) 433/478 (91) 1.21 (1.06 - 1.39) 0.005
Coronary Artery Disease
No (n = 518) 222/258 (86) 229/260 (88) 1.16 (0.96 - 1.39) 0.118
0.728
Yes (n = 878) 385/440 (88) 395/438 (90) 1.22 (1.06 - 1.40) 0.007
Diabetes Mellitus
No (n = 802) 339/396 (86) 355/406 (87) 1.17 (1.01 - 1.36) 0.042
0.679
Yes (n = 594) 268/302 (89) 269/292 (92) 1.24 (1.05 - 1.47) 0.012
Atrial Fibrillation
No (n = 769) 331/385 (86) 340/384 (89) 1.25 (1.08 - 1.46) 0.004
0.359
Yes (n = 627) 276/313 (88) 284/314 (90) 1.12 (0.95 - 1.33) 0.169
Glomerular Filtration Rate
≥45 ml/min/1.73 m2 (n = 744) 316/383 (83) 307/361 (85) 1.16 (0.99 - 1.36) 0.065
0.626
2
<45 ml/min/1.73 m (n = 652) 291/315 (92) 317/337 (94) 1.22 (1.04 - 1.43) 0.016
Ejection Fraction
≥25% (n = 922) 403/470 (86) 401/452 (89) 1.21 (1.05 - 1.39) 0.007
0.768
<25% (n = 474) 204/228 (89) 223/246 (91) 1.16 (0.96 - 1.40) 0.138
ACE Inhibitor/ARB on Discharge
Yes (n = 878) 379/449 (84) 381/429 (89) 1.23 (1.07 - 1.42) 0.005
0.539
No (n = 518) 228/249 (92) 243/269 (90) 1.13 (0.94 - 1.35) 0.184
Beta-Blocker on Discharge
Yes (n = 901) 397/457 (87) 393/444 (89) 1.11 (0.96 - 1.27) 0.161
0.044
No (n = 495) 210/241 (87) 231/254 (91) 1.38 (1.14 - 1.66) 0.001

Overall (N = 1,396) 607/698 (87) 624/698 (89) 1.20 (1.07 - 1.34) 0.002

Combined endpoint of heart failure

readmission or all-cause mortality, n / N (%) 0.8 1.0 1.2 1.4 1.6 1.8
Hazard Ratio (95% CI)

In all subgroups analyzed, the association between digoxin discontinuation and the combined endpoint of heart failure readmission or all-cause mortality was ho-
mogeneous, except by beta-blocker use at hospital discharge. Note: Results of subgroup analyses need to be interpreted with caution because they may be false
positive due to multiple comparisons and false negative due to inadequate power. ACE ¼ angiotensin-converting enzyme; ARB ¼ angiotensin receptor blocker;
CI ¼ confidence interval.
624 Malik et al. JACC VOL. 74, NO. 5, 2019

Digoxin Discontinuation and Outcomes in HFrEF AUGUST 6, 2019:617–27

p ¼ 0.161) and 1.38 (1.14 to 1.66; p ¼ 0.001), respec- FINDINGS FROM THE SENSITIVITY COHORT. The
tively (p for interaction ¼ 0.044) (Figure 2). 1,204 matched patients had a mean age of 76  11
Findings from our sensitivity analyses demon- years and left ventricular ejection fraction of
strate that the significant 4-year associations of 28  10%; 40% were female, and 15% were African
digoxin discontinuation with all-cause readmission American; and were balanced on 51 baseline charac-
and the combined endpoint were insensitive to un- teristics including admission and discharge serum
measured confounders. Of the 698 matched pairs, in creatinine and in-hospital AKI (Online Table 1).
580 pairs, we were able to determine which patient Digoxin discontinuation was associated with signifi-
within a matched pair had a shorter 4-year total cantly higher risks of HF readmission (HR: 1.26;
readmission-free survival, and in 55% of those pairs 95% CI: 1.08 to 1.46; p ¼ 0.003), all-cause readmission
(318 of 580), they belonged to the digoxin discontin- (HR: 1.15; 95% CI: 1.02 to 1.30; p ¼ 0.026), and the
uation group (sign-score test p ¼ 0.020). A hidden combined endpoint of HF readmission or all-cause
baseline characteristic would need to increase the mortality (HR: 1.22; 95% CI: 1.08 to 1.38; p ¼ 0.001),
odds of digoxin discontinuation by 3% to explain but not with all-cause mortality (HR: 1.12; 95% CI:
away this association. For the combined endpoint, in 0.98 to 1.28; p ¼ 0.098) (Online Table 2).
682 pairs we were able to determine which member of PREDICTORS OF DISCONTINUATION OF DIGOXIN.
the pair had a shorter time to event and in 55% of Among the 3,499 pre-match patients with HFrEF who
those pairs (378 of 682), they belonged to the digoxin were on digoxin at the time of hospital admission,
discontinuation group (sign-score test p ¼ 0.005). A discharge prescriptions of ACE inhibitors or angio-
hidden baseline characteristic would need to increase tensin receptor blocker (ARBs) (OR: 0.64; 95% CI: 0.53
the odds of digoxin discontinuation by 7% to explain to 0.77; p < 0.001), beta-blockers (OR: 0.68; 95% CI:
away this association. 0.57 to 0.82; p < 0.001), loop diuretics (OR: 0.52;
6-MONTH AND 1-YEAR OUTCOMES. Digoxin discon- 95% CI: 0.43 to 0.64; p < 0.001), and nitrates
tinuation was associated with higher risks of HF (OR: 0.76; 95% CI: 0.62 to 0.94; p ¼ 0.011) were
readmission, all-cause readmission, all-cause mor- associated with lower odds of digoxin discontinua-
tality, and the combined endpoint of HF readmission tion. The odds of digoxin discontinuation were also
or all-cause mortality at 6 months and 1 year after lower among patients with orthopnea (OR: 0.77;
hospital discharge (Table 2). Of the 698 matched 95% CI: 0.61 to 0.97; p ¼ 0.026), anemia (OR: 0.72;
pairs, in 325 pairs, we were able to determine which 95% CI: 0.56 to 0.94; p ¼ 0.016), and chronic
patient within a matched pair had a shorter 6-month obstructive pulmonary disease (OR: 0.81; 95% CI,
HF readmission-free survival, and in 56% of those 0.66–0.99; p ¼ 0.038), but higher among those with
pairs (182 of 325), they belonged to the digoxin lower extremity edema (OR: 1.32; 95% CI: 1.10 to 1.60;
discontinuation group (sign-score test p ¼ 0.031). A p ¼ 0.003). Interestingly, in-hospital AKI was not
hidden baseline characteristic would need to increase associated with digoxin discontinuation (OR: 0.89;
the odds of digoxin discontinuation by 2% to explain 95% CI: 0.68 to 1.17; p ¼ 0.420).
away this association.
DISCUSSION
30-DAY OUTCOMES. Digoxin discontinuation had no
association with HF or all-cause readmissions at 30- Findings from our study demonstrate that discon-
day post-discharge (Table 2). Digoxin discontinua- tinuation of pre-admission digoxin therapy in hospi-
tion was associated with a higher risk of all-cause talized older patients with HFrEF is associated with a
mortality (HR: 1.80; 95% CI: 1.26 to 2.57; p ¼ 0.001) significantly higher risk of poor outcomes. The asso-
(Table 2). In 120 of the 698 matched pairs, we were ciations with readmissions became significant at
able to determine which patient within a matched 6 months and lasted for 4 years but were not signifi-
pair had a shorter survival, and in 65% of those pairs cant during the first 30 days after hospital discharge.
(78 of 120), they belonged to the digoxin discontinu- By contrast, the association with mortality was sig-
ation group (sign-score test p ¼ 0.001). A hidden nificant at 30 days but disappeared after the first year.
baseline characteristic would need to increase the The risk of the combined endpoint of HF readmission
odds of digoxin discontinuation by 22% to explain or all-cause mortality was significantly higher in the
away this association. Digoxin discontinuation was group with digoxin discontinuation throughout the
also associated with a higher risk of the combined entire follow-up. To the best of our knowledge, this is
endpoint (HR: 1.36; 95% CI: 1.09 to 1.71; the first report of adverse outcomes associated with
p ¼ 0.007) (Table 2). discontinuation of pre-admission digoxin therapy in
JACC VOL. 74, NO. 5, 2019 Malik et al. 625
AUGUST 6, 2019:617–27 Digoxin Discontinuation and Outcomes in HFrEF

a propensity score–matched cohort of hospitalized Although we were able to balance the use of these
older patients with HFrEF receiving contemporary drugs in our propensity score–matched cohort,
GDMT including ACE inhibitors/ARBs, beta-blockers, matching may not balance the underlying reasons for
and MRAs. their underuse in the first place. Persistence of those
The elucidation of mechanistic explanations of the reasons during follow-up may result in subsequent
association between discontinuation of digoxin and underuse of those drugs, which in turn may
poor outcomes is outside the scope of the current contribute to poor outcomes. These drugs are known
study. Digoxin is known for its efficacy and effec- for their proven efficacy and effectiveness in
tiveness in lowering the risk of HF and all-cause lowering the risks of both all-cause mortality and
readmissions (3–5,7–10). It has been suggested that hospitalization. If the underuse of ACE inhibitors or
the positive inotropic effect of digoxin on cardiac ARBs, beta-blockers, and MRAs in the digoxin
performance may be attenuated during long-term discontinuation group confounded the worse out-
therapy and may not be related to clinical outcomes comes in that group, then they would be expected to
(11,29,30). The clinical benefit of digoxin is often increase the risk of mortality. However, in our study,
attributed to its ability to suppress neurohormones, digoxin discontinuation had no relationship with
specifically those in the sympathetic nervous system 4-year mortality, suggesting that results of our study
(31–37). In the RADIANCE and PROVED trials, the 2 are unlikely to be explained by the underuse of other
major randomized controlled trials of digoxin with- neurohormonal antagonists. A higher comorbidity
drawal, significantly more patients in the digoxin burden may also explain worse outcomes. However,
discontinuation group had worsening functional ca- patients in our matched cohort were balanced on
pacity and symptoms, and a reduction in ejection 50 baseline characteristics including prior HF
fraction (11,12). Although neurohormonal parameters hospitalizations.
were not measured in these trials, indirect evidence HF remains a leading cause for hospital admission
from a significant increase in heart rate and body and readmission for older adults, and digoxin has
weight in the digoxin discontinuation groups sug- been shown to lower both risks without adversely
gested that digoxin discontinuation was associated affecting mortality (1,3–5,7–10). Digoxin is an inex-
with activation of the sympathetic neurohormonal pensive and relatively safe drug at low doses and is
system (11,12). Plasma noradrenaline concentration recommended for patients with HFrEF who remain
has been shown to be significantly higher among symptomatic despite optimal GDMT (2). The use of
patients with HF with higher New York Heart Asso- digoxin has declined substantially in recent years (6),
ciation functional class symptoms (38). The neuro- which is likely due to both lower rates of initiation as
hormonal activation would be expected to be even well as higher rates of discontinuation. The effect of
higher in hospitalized patients with decompensated discontinuation of digoxin therapy on outcomes is
HF, which may in turn explain a more pronounced well documented in the RADIANCE and PROVED
relationship between digoxin discontinuation and trials (11,12). However, both the RADIANCE and the
adverse outcomes in these patients. It is also possible PROVED trials included ambulatory patients with
that the higher risk of readmission in the digoxin HFrEF from an earlier era of HF therapy limited to use
discontinuation group may in part be explained by of ACE inhibitors, had small sample sizes, and used
the negative rebound effect from digoxin withdrawal. intermediate and softer endpoints. To the best of our
However, unlike other cardiovascular drugs such as knowledge, the current study is the first to examine
beta-blockers, the rebound syndrome phenomenon the relationship between the discontinuation of
is less well documented for digoxin (39,40). digoxin therapy and harder endpoints in a relatively
Findings from the studies of initiation of digoxin large sample of hospitalized patients with decom-
therapy suggest evidence for clinical effectiveness pensated HFrEF. More importantly, the use of GDMT
of digoxin in lowering the risk of hospitalization in our study was similar to that in more contemporary
(8–10). HFrEF populations (6). Considering the recent decline
Considering the observational nature of the cur- in the use of digoxin, a drug known for its efficacy and
rent study, we also explored the roles of confounding effectiveness in reducing the risk of hospitalization
due to bias, such as indication bias. Findings from the (3,4,6–10), the findings of the current study are
American Heart Association’s Get With The Guide- important because they suggest that in hospitalized
lines HF registry (6) and the pre-match data from the patients with HFrEF receiving contemporary GDMT
current study suggest that fewer patients in the including beta-blockers and MRAs, the discontinua-
digoxin discontinuation group were receiving ACE tion of digoxin therapy may be associated with worse
inhibitors or ARBs, beta-blockers, and MRAs. post-discharge outcomes.
626 Malik et al. JACC VOL. 74, NO. 5, 2019

Digoxin Discontinuation and Outcomes in HFrEF AUGUST 6, 2019:617–27

STUDY LIMITATIONS. Despite our use of propensity CONCLUSIONS

score–matched balanced cohorts, bias due to imbal-
ances in unmeasured baseline characteristics is Among hospitalized older patients with HFrEF
possible. Findings from our sensitivity analyses sug- receiving more contemporary GDMT including ACE
gest that the associations observed could be sensitive inhibitors/ARBs, beta-blockers and MRAs, discontin-
to an unmeasured confounder. However, sensitivity uation of pre-admission digoxin therapy was associ-
analysis cannot determine if such an unmeasured ated with poor outcomes. Findings from this study
confounder exists. For an unmeasured binary base- suggest that it may be premature to abandon the use
line characteristic to be a confounder, it will need to of digoxin in patients with HFrEF.
be a near perfect predictor of the outcome being
studied and also could not be strongly related to the ADDRESS FOR CORRESPONDENCE: Dr. Ali Ahmed,
50 baseline characteristics used in our logistic Washington DC VA Medical Center, 50 Irving Street
regression model to estimate propensity scores, NW, Washington, DC 20422. E-mail: [email protected].
which is an unlikely possibility. The medical, surgi- Twitter: @DCVAMC, @gcfmd.
cal, and device management of HFrEF have evolved
since the OPTIMIZE-HF registry, and the current
PERSPECTIVES
study is based on fee-for-service Medicare benefi-
ciaries, which may limit generalizability. We did not
COMPETENCY IN PATIENT CARE AND
have data on digoxin use during follow-up. If digoxin
therapy was restarted in a large proportion of patients PROCEDURAL SKILLS: In hospitalized older patients

in the digoxin discontinuation group or digoxin was with HFrEF receiving guideline-directed medical

discontinued in the digoxin therapy group, that could management, discontinuation of pre-admission

potentially attenuate between-group differences and digoxin therapy is associated with poor outcomes.

underestimate true associations. Similarly, we did not

TRANSLATIONAL OUTLOOK: Prospective studies
have data on the use and uptitration of other HF
are needed to confirm the effect of digoxin discon-
drugs during follow-up, a differential use of which
tinuation on clinical outcomes in patients with HFrEF
may influence outcomes. Results of subgroup ana-
managed with contemporary therapy and determine
lyses need to be interpreted with caution because
responsible mechanisms.
they may be false positive due to multiple compari-
sons and false negative due to inadequate power (41).

REFERENCES

1. Jencks SF, Williams MV, Coleman EA. Rehospi- Association Get With The Guidelines-Heart Fail- 12. Uretsky BF, Young JB, Shahidi FE, Yellen LG,
talizations among patients in the Medicare fee- ure Registry. J Am Coll Cardiol HF 2016;4: Harrison MC, Jolly MK, PROVED Investigative
for-service program. N Engl J Med 2009;360: 348–56. Group. Randomized study assessing the effect of
1418–28. digoxin withdrawal in patients with mild to mod-
7. Bourge RC, Fleg JL, Fonarow GC, et al. Digoxin
erate chronic congestive heart failure: results of
2. Yancy CW, Jessup M, Bozkurt B, et al. 2013 reduces 30-day all-cause hospital admission in
the PROVED trial. J Am Coll Cardiol 1993;22:
ACCF/AHA guideline for the management of heart older patients with chronic systolic heart failure.
955–62.
failure: a report of the American College of Car- Am J Med 2013;126:701–8.
diology Foundation/American Heart Association 13. Ahmed A, Gambassi G, Weaver MT, Young JB,
8. Ahmed A, Bourge RC, Fonarow GC, et al.
Task Force on Practice Guidelines. J Am Coll Car- Wehrmacher WH, Rich MW. Effects of discontin-
Digoxin use and lower 30-day all-cause read-
diol 2013;62:e147–239. uation of digoxin versus continuation at low serum
mission for Medicare beneficiaries hospitalized for
3. The Digitalis Investigation Group Investigators. heart failure. Am J Med 2014;127:61–70. digoxin concentrations in chronic heart failure. Am
The effect of digoxin on mortality and morbidity in J Cardiol 2007;100:280–4.
9. Lam PH, Bhyan P, Arundel C, et al. Digoxin use
patients with heart failure. N Engl J Med 1997; 14. Fonarow GC, Abraham WT, Albert NM, et al.
and lower risk of 30-day all-cause readmission in
336:525–33. Organized Program to Initiate Lifesaving Treat-
older patients with heart failure and reduced
4. Ahmed A, Rich MW, Love TE, et al. Digoxin and ejection fraction receiving beta-blockers. Clin ment in Hospitalized Patients with Heart Failure
reduction in mortality and hospitalization in heart Cardiol 2018;41:406–12. (OPTIMIZE-HF): rationale and design. Am Heart J
failure: a comprehensive post hoc analysis of the 2004;148:43–51.
10. Qamer SZ, Malik A, Bayoumi E, et al. Digoxin
DIG trial. Eur Heart J 2006;27:178–86. 15. Fonarow GC, Stough WG, Abraham WT, et al.
use and outcomes in patients with heart failure
5. Gheorghiade M, Patel K, Filippatos G, et al. Ef- with reduced ejection fraction. Am J Med 2019 Characteristics, treatments, and outcomes of
fect of oral digoxin in high-risk heart failure pa- May 28 [E-pub ahead of print]. patients with preserved systolic function hospi-
tients: a pre-specified subgroup analysis of the talized for heart failure: a report from the
11. Packer M, Gheorghiade M, Young JB, et al.
DIG trial. Eur J Heart Fail 2013;15:551–9. OPTIMIZE-HF registry. J Am Coll Cardiol 2007;50:
Withdrawal of digoxin from patients with chronic
768–77.
6. Patel N, Ju C, Macon C, et al. Temporal trends heart failure treated with angiotensin-converting-
of digoxin use in patients hospitalized with enzyme inhibitors. RADIANCE study. N Engl J Med 16. Bayoumi E, Lam PH, Dooley DJ, et al. Spi-
heart failure: analysis from the American Heart 1993;329:1–7. ronolactone and outcomes in older patients with
JACC VOL. 74, NO. 5, 2019 Malik et al. 627
AUGUST 6, 2019:617–27 Digoxin Discontinuation and Outcomes in HFrEF

heart failure and reduced ejection fraction. Am J 26. Austin PC. An introduction to propensity score patients: direct evidence from sympathetic neural
Med 2019;132:71–80.e1. methods for reducing the effects of confounding recordings. Circulation 1989;80:65–77.
in observational studies. Multivariate Behav Res
17. Lam PH, Gupta N, Dooley DJ, et al. Role of 36. Covit AB, Schaer GL, Sealey JE, Laragh JH,
2011;46:399–424.
high-dose beta-blockers in patients with heart Cody RJ. Suppression of the renin-angiotensin
failure with preserved ejection fraction and 27. Rosenbaum PR. Sensitivity to hidden bias. In: system by intravenous digoxin in chronic
elevated heart rate. Am J Med 2018;131:1473–81. Rosenbaum PR, editor. Observational Studies. congestive heart failure. Am J Med 1983;75:
New York, NY: Springer-Verlag, 2002:105–70. 445–7.
18. Mujib M, Patel K, Fonarow GC, et al. Angio-
tensin-converting enzyme inhibitors and outcomes 28. Arundel C, Lam PH, Gill GS, et al. Systolic 37. Torretti J, Hendler E, Weinstein E,
in heart failure and preserved ejection fraction. Am blood pressure and outcomes in patients with Longnecker RE, Epstein FH. Functional signifi-
J Med 2013;126:401–10. heart failure with reduced ejection fraction. J Am cance of Na- K-ATPase in the kidney: effects of
Coll Cardiol 2019;73:3054–63. ouabain inhibition. Am J Physiol 1972;222:
19. Zhang Y, Kilgore ML, Arora T, et al. Design and
1398–405.
rationale of studies of neurohormonal blockade 29. Taggart AJ, Johnston GD, McDevitt DG.
and outcomes in diastolic heart failure using Digoxin withdrawal after cardiac failure in patients 38. Sigurdsson A, Amtorp O, Gundersen T,
OPTIMIZE-HF registry linked to Medicare data. Int with sinus rhythm. J Cardiovasc Pharmacol 1983;5: Nilsson B, Remes J, Swedberg K, the Ramipril Trial
J Cardiol 2013;166:230–5. 229–34. Study Group. Neurohormonal activation in pa-
tients with mild or moderately severe congestive
20. Rubin DB. Using propensity score to help 30. Fleg JL, Gottlieb SH, Lakatta EG. Is digoxin
heart failure and effects of ramipril. Br Heart J
design observational studies: application to the really important in treatment of compensated
1994;72:422–7.
tobacco litigation. Health Serv Outcomes Res heart failure? A placebo-controlled crossover
Methodol 2001;2:169–88. study in patients with sinus rhythm. Am J Med 39. Nattel S, Rangno RE, Van Loon G. Mechanism
1982;73:244–50. of propranolol withdrawal phenomena. Circulation
21. Rosenbaum PR, Rubin DB. The central role of
1979;59:1158–64.
propensity score in observational studies for 31. Slatton ML, Irani WN, Hall SA, et al. Does
causal effects. Biometrika 1983;70:41–55. digoxin provide additional hemodynamic and 40. Schmidt TA, Holm-Nielsen P, Kjeldsen K. No
autonomic benefit at higher doses in patients with upregulation of digitalis glycoside receptor (Na,K-
22. Ahmed A, Husain A, Love TE, et al. Heart
mild to moderate heart failure and normal sinus ATPase) concentration in human heart left
failure, chronic diuretic use, and increase in mor-
rhythm? J Am Coll Cardiol 1997;29:1206–13. ventricle samples obtained at necropsy after long
tality and hospitalization: an observational study
term digitalisation. Cardiovasc Res 1991;25:
using propensity score methods. Eur Heart J 32. Gheorghiade M, Hall V, Lakier JB, Goldstein S.
684–91.
2006;27:1431–9. Comparative hemodynamic and neurohormonal
effects of intravenous captopril and digoxin and 41. Rothwell PM. Treating individuals 2. Subgroup
23. Ahmed A, Rich MW, Zile M, et al. Renin-
their combinations in patients with severe heart analysis in randomised controlled trials: impor-
angiotensin inhibition in diastolic heart failure and
failure. J Am Coll Cardiol 1989;13:134–42. tance, indications, and interpretation. Lancet
chronic kidney disease. Am J Med 2013;126:
2005;365:176–86.
150–61. 33. Gheorghiade M, Ferguson D. Digoxin. A
neurohormonal modulator in heart failure? Circu-
24. Ahmed A, Fonarow GC, Zhang Y, et al. Renin-
lation 1991;84:2181–6.
angiotensin inhibition in systolic heart failure and KEY WORDS digoxin discontinuation, heart
chronic kidney disease. Am J Med 2012;125: 34. Krum H, Bigger JT Jr., Goldsmith RL, Packer M. failure, mortality, readmission, reduced
399–410. Effect of long-term digoxin therapy on autonomic ejection fraction
function in patients with chronic heart failure.
25. Ahmed MI, White M, Ekundayo OJ, et al.
J Am Coll Cardiol 1995;25:289–94.
A history of atrial fibrillation and outcomes
in chronic advanced systolic heart failure: a 35. Ferguson DW, Berg WJ, Sanders JS, Roach PJ, A PP END IX For supplemental figures and
propensity-matched study. Eur Heart J 2009;30: Kempf JS, Kienzle MG. Sympathoinhibitory re- tables, please see the online version of
2029–37. sponses to digitalis glycosides in heart failure this paper.