Causes of death and rehospitalization in patients
hospitalized with worsening heart failure and reduced
left ventricular ejection fraction: Results from efficacy of
vasopressin antagonism in heart failure outcome study
with tolvaptan (EVEREST) program
Christopher M. O’Connor, MD, a Alan B. Miller, MD, b John E. A. Blair, MD, c Marvin A. Konstam, MD, d
Patricia Wedge, RN, CCRC, e Maria C. Bahit, MD, f Peter Carson, MD, g Markus Haass, MD, h Paul J. Hauptman, MD, i
Marco Metra, MD, j Ron M. Oren, MD, k Richard Patten, MD, d Ileana Piña, MD, l Sherryn Roth, MD, m
Jonathan D. Sackner-Bernstein, MD, n Brian Traver, MS, o Thomas Cook, PhD, o and Mihai Gheorghiade, MD p
for the Efficacy of Vasopressin Antagonism in heart Failure Outcome Study with Tolvaptan (EVEREST)
investigators Durham, NC; Jacksonville, FL; San Antonio, TX; Boston, MA; Buenos Aires, Argentina; Mannheim,
Germany; St Louis, MO; Brescia, Italy; Iowa City, IA; Cleveland, OH; Ontario, Canada; Madison, WI; and Chicago, IL
Background The postdischarge rehospitalization and death rates are high in patients with acute heart failure (HF)
syndromes despite optimization of standard therapy for chronic HF. To the best of our knowledge, there has been no systematic
analysis of the causes of death and rehospitalization in this patient population.
Methods
This was a prespecified analysis of adjudicated cause-specific all-cause mortality and cardiovascular (CV)
hospitalization in the Efficacy of Vasopressin Antagonism in Heart Failure Outcome Study with Tolvaptan (EVEREST) trial, a
randomized, double-blind, placebo-controlled study in patients hospitalized with worsening HF and left ventricular ejection
fraction ≤40% comparing tolvaptan, an oral vasopressin receptor antagonist to placebo, in addition to standard care.
Results Of the 4,133 randomized, there were 5,239 rehospitalizations and 1,080 deaths during a median of 9.9
months. Of all deaths, 41.0% were due to HF, 26.0% due to sudden cardiac death (SCD), 2.6% due to acute myocardial
infarction (MI), 2.2% due to stroke, and 13.2% due to non-CV causes. Of all hospitalizations, 39.2% were non-CV, whereas
46.3% were for HF, and a minority of hospitalizations was due to stroke, MI, arrhythmia, or other CV causes.
Conclusions Despite close follow-up and evidence-based therapy within a clinical trial, rehospitalization and death
remain high. Although most deaths were from HF, one quarter of patients had SCD. In addition, there were almost as many
non-CV hospitalizations as HF hospitalizations. Knowledge of the causes of death and rehospitalization may be essential for
proper management and early initiation of therapy. (Am Heart J 2010;159:841-849.e1.)
From the
a
Duke University Medical Center, Durham, NC,
c
b
University of Florida,
d
Jacksonville, FL, Wilford Hall Medical Center, San Antonio, TX, Tufts University-New
England Medical Center, Boston, MA, eCardiovascular Clinical Studies, Boston, MA,
f
Hospital Italiano de Buenos Aires, Buenos Aires, Argentina, gWashington VAMC,
Washington DC, hTheresienkrankenhaus, Mannheim, Germany, iSt Louis University School
of Medicine, St Louis, MO, jUniversity of Brescia, Brescia, Italy, kIowa City Heart Center PC,
Iowa City, IA, lCase Western Reserve University, Cleveland, OH, mUniversity of Toronto,
Toronto, Ontario, Canada, nDobbs Ferry, New York, oUniversity of Wisconsin, Madison,
WI, and pNorthwestern University, Chicago, IL.
Clinicaltrials.gov Identifier: NCT00071331.
Submitted December 29, 2009; accepted February 23, 2010.
Reprint requests: Mihai Gheorghiade, MD, Center for Cardiovascular Quality and
Outcomes, Northwestern University, Feinberg School of Medicine, 645 N. Michigan
Avenue, Suite 1006, Chicago, IL 60611.
E-mail: m-gheorghiade@northwestern.edu
0002-8703/$ - see front matter
© 2010, Mosby, Inc. All rights reserved.
doi:10.1016/j.ahj.2010.02.023
Death and hospitalization are common in chronic
heart failure (HF), with the annual death rate approaching 20% and the number of hospital admissions for HF as
a primary diagnosis N1 million or 390 per 100,000 in the
United States.1,2 The available data regarding causes of
hospitalization3,4 and mode of death5-9 are in patients
with chronic HF. Acute heart failure syndromes (AHFS),
defined as rapid development of HF signs and symptoms
requiring urgent therapy and hospitalization, have been
identified as a group with an extremely high death and
rehospitalization rate, especially within 60 to 90 days
after discharge from the hospital, despite most patients
have symptomatic improvement in response to therapy
received during hospitalization.10,11 Recently, there has
been interest from policymakers in reducing the rates of
842 O'Connor et al
rehospitalization to reduce costs and improve quality of
care.12 To achieve this goal, it is essential to understand
the causes of rehospitalization and mode of death so that
therapies and treatment strategies can be developed to
prevent these events. There are limited data on
frequency and patterns of rehospitalization and practically no data on the mode of death and cause for
rehospitalization after an episode of AHFS.13 The
purpose of this analysis was to explore the causes of
rehospitalization and mode of death for patients
discharged after hospitalization for worsening HF and
reduced left ventricular ejection fraction in the Efficacy
of Vasopressin Antagonism in Heart Failure Outcome
Study with Tolvaptan (EVEREST) study, a large contemporary study in AHFS patients.
Methods
Study design and population
The design and results of the EVEREST program has been
reported previously.14-16 The trial was an event-driven, randomized, double-blind, placebo-controlled study designed to evaluate both the short-term and long-term efficacy of tolvaptan (an
oral vasopressin V2 receptor blocker) supplemented to standard
HF therapy in patients hospitalized with worsened HF. This
patient population was enrolled at 359 sites and 20 countries
throughout North America, Europe, and South America
between October 7, 2003, and February 3, 2006. Two identical
studies (trials A and B) were designed to evaluate short-term
outcomes during the inpatient period and embedded into a longterm outcome study.
All patients in the study population were adults ≥18 years of age
with reduced left ventricular ejection fraction (≤40%), hospitalized with ≥2 signs of HF at the time of randomization. Patients
were randomized within 48 hours of admission to receive either
30 mg/d of tolvaptan or matching placebo in addition to standard
chronic HF therapy. Significant exclusion criteria included cardiac
surgery within 60 days of enrollment, comorbidities with a life
expectancy b6 months, hemodynamically significant valvular
disease, end-stage HF, acute myocardial infarction (MI), serum
potassium level N5.5 mEq/L, serum creatinine level N3.5 mEq/dL,
hemoglobin level b9 g/dL, and those patients on dialysis or
hemofiltration. Those patients with planned revascularization,
electrophysiologic device implantation, cardiac mechanical support implantation, cardiac transplantation, or other cardiac
surgery within 30 days of enrollment were also excluded.
Adjudication and definition of end points
The clinical events committee (see Appendix online) cochairs
were chosen by the steering committee, who together chose the
remaining 10 committee members based on their expertise in
HF.15 All members of the committee were blinded to treatment
arms to eliminate the possibility of bias. Once an end point was
reached, the site coordinator collected source documentation
for end point validation according to a predefined protocol. The
sponsor reviewed all documentation for completeness, contacted the site if additional documentation was required, and
ensured proper blinding before submission to the clinical events
committee. One of 3 clinical events committee coordinators
American Heart Journal
May 2010
then assigned the event to 2 reviewers using a random
assignment mechanism. Each reviewer completed an electronic
case report form for each end point. If both reviewers per event
agreed, the event was considered adjudicated. If there was
discrepancy, the information was forwarded to 1 of the 2
committee chairmen using a randomized mechanism, who
either reviewed the data and made a final adjudication decision
if the discrepancy could be resolved or discussed the event
during a quarterly meeting for committee decision. Meetings
were held at least quarterly or more frequently if deemed
necessary by the committee chairmen.
Rehospitalization was defined as a nonelective hospital
admission for medical therapy with a duration that extended
over a change in calendar date and based on the primary
reason for admission. Noncardiovascular (CV) hospitalizations
were those hospitalizations without a CV classification. The
CV hospitalizations included HF, arrhythmia, acute MI, stroke,
and other hospitalizations and defined as follows: (1) HF
hospitalization—hospitalization that included substantive
worsening of HF symptoms and/or signs resulting in
augmentation of oral medications or new administration of
intravenous HF therapies including ultrafiltration; (2) arrhythmia hospitalization—typical electrocardiographic evidence or
a clear indication in the narrative; (3) acute MI hospitalization
—evidence of myocardial necrosis plus either ischemic
symptoms, or ST-T wave changes, as defined below; (3)
stroke hospitalization—N24 hours of stroke symptoms defined
below; and (5) other hospitalization—a CV event that does
not meet above criteria.
Mode of death was classified as CV, non-CV, or unknown. NonCV death was defined as a death due to a specific non-CV event,
whereas unknown death was defined as a death for which no
information surrounding the event was available. The CV death
comprised sudden cardiac death (SCD), HF death, acute MI,
stroke, or other and defined as follows: (1) SCD—an unexpected
death in a previously stable patient with recent human contact.
If the subject was out of contact for 24 hours to 1 week, the
event was classified as “presumed SCD.” (2) HF death—a death
involving a substantive worsening of HF symptoms and/or signs
resulting in augmentation or addition of HF therapies. (3) Acute
MI—death occurring after a hospital-verified, definite acute MI,
or if the death was outside the hospital, with autopsy findings
showing a recent MI or recent occluding coronary thrombus.
If autopsy findings were not available, 2 of the following must
be present: symptoms compatible with an acute coronary
event, evidence of myocardial necrosis (creatinine kinase level
N1.5 times the upper limit of normal, creatinine kinase-MB
isoenzyme greater than the upper limits of normal, or troponin
level N2 times the upper limits of normal), or ST-T wave changes
(N2 mm ST-segment elevation in 2 contiguous leads, new
Q waves, or ischemic ST-segment elevation or depression or
T-wave inversion). (4) Stroke death—persistent disturbance of
focal neurologic function resulting in symptoms thought to be
due to thrombotic cerebral infarction, embolus, or hemorrhage
based on history, physical, imaging techniques, and/or autopsy
data. (5) Other CV death—any other CV death that does not fit
into the above criteria.
Defined study end points and statistical analysis
The EVEREST patient population was observed for a median
period of 9.9 months (interquartile range [IQR] 5.1-16.1 months).
American Heart Journal
Volume 159, Number 5
The primary long-term efficacy end points for the EVEREST trial
were all-cause mortality and the combined end point of CV
mortality and HF hospitalization. A cause-specific analysis of the
cause of rehospitalization and death, based on adjudicated end
points, was prespecified in the original trial design.
Baseline clinical characteristics and medication use at
randomization and discharge were tabulated among the major
causes of first postdischarge hospitalization and of death. For
administrative reasons, B-type natriuretic peptide (BNP) collection was performed on only 2,941 patients, and the distribution
of BNP measurements was highly skewed, so we summarized
available BNP values using the medians and IQRs. In addition,
the QRS interval was not reported in 142 patients, and 1,029
patients were excluded due to presence of a pacemaker and/or
implantable cardioverter-defibrillator, leaving 2,962 analyzed
patients. Pearson χ2 analysis and Kruskal-Wallis statistical tests
were used to compare categorical and continuous variables
respectively. The timing of major causes of first rehospitalization
and death was tabulated using post-randomization days 1 to
30, days 31 to 60, and N60 days.
The EVEREST program was funded by the Otsuka Pharmaceutical Group, Tokyo, Japan, under the direction of the
Steering Committee. The authors are solely responsible for the
design and conduct of the study, all study analyses, the drafting
and editing of the article, and its final contents.
Results
Study population
Of the 4,133 patients randomized in the EVEREST trial
observed for a median of 9.9 months, there were 5,239
rehospitalizations and 1,080 deaths. As previously
reported, there were no differences in the primary end
points between the tolvaptan and placebo groups, and
the Kaplan-Meier estimated 1-year mortality rate was
25.5%, and the hospitalization rate was 57.6%.16
Mode of death
Most deaths were CV deaths (86.8% of all deaths). The
leading mode of death was for HF (47.2%), followed by
SCD (30.0%), non-CV causes (13.2%), unknown (11.4%),
or other CV death (5.8%). Acute MI (3.0%) and stroke
(2.6%) represented a small percentage of the deaths
(Table I).
Baseline characteristics were compared across groups
of patients who died of HF, SCD, all other causes, and
those still alive during the follow-up period (Table II).
Compared to patients who died of SCD, patients who
died of HF had lower systolic blood pressure (111.0 ±
17.2 vs 115.2 ± 18.2 mm Hg), higher New York Heart
Association (NYHA) functional class (54.9% vs 44.6%
NYHA class IV), higher serum blood urea nitrogen (43.4 ±
24.5 vs 32.0 ± 16.1 mg/dL), higher serum creatinine level
(1.7 ± 0.8 vs 1.4 ± 0.5 mg/dL), lower serum sodium
(137.1 ± 5.5 vs 139 ± 4.8 mg/dL), higher arginine
vasopressin (6.8 ± 8.2 vs 5.8 ± 5.8 pg/mL), higher serum
BNP (1,519 [IQR 2,470] vs 966 [IQR 1,793] pg/mL),
O'Connor et al 843
Table I. Leading modes of death in EVEREST
Mode of death
No. of Deaths
% of total deaths
Total death
Non-CV death
CV death
HF
SCD
Unknown
Other CV death
Acute MI
Stroke
1080
143
937
443
281
107
54
28
24
100.0
13.2
86.8
41.0 (47.2)⁎
26.0 (30.0)⁎
9.9 (11.4)⁎
5.0 (5.8)⁎
2.6 (3.0)⁎
2.2 (2.6)⁎
⁎ Percentage of CV death.
longer QRS duration (68.2% vs 62.7% with QRS ≥120
milliseconds), higher percentage of diabetes (43.8% vs
39.2%), and higher use of implantable cardioverterdefibrillator implantation (23.9% vs 12.5%). In general,
patients were well treated with evidence-based therapies
for chronic HF, which included angiotensin-converting
enzyme (ACE) inhibitors, angiotensin receptor blockers,
β-blocking agents, spironolactone, and diuretics. Compared to patients with SCD, patients who died of HF had
slightly lower use of ACE inhibitors, angiotensin receptor
blockers, β-blockers, and spironolactone, and a higher
use of inotropic agents and amiodarone. The HF death
was approximately twice the rate of SCD among all
regions studied, with the exception of Eastern Europe,
where there were equal HF and SCD events.
Mode of hospitalization
Of the total number of hospitalizations, 60.8% were CV
hospitalizations and 39.2% were non-CV hospitalizations.
Of the CV hospitalizations, most (46.5% of total hospitalizations) were for HF, whereas arrhythmia, stroke, MI,
and other CV hospitalizations made up a minority of CV
hospitalizations (Table III). Because a single patient could
be hospitalized multiple times during the follow-up
period, further analysis of the 2,159 patients (52.2%)
with at least one type of hospitalization was performed.
There were 1,225 (29.6% of all patients) who had non-CV
hospitalizations and 1,625 (39.3% of all patients) who had
CV hospitalizations. The HF hospitalization was the most
common cause of CV hospitalization (31.1% of all
patients), with arrhythmia, stroke, MI, and other CV
rehospitalizations representing a minority of CV hospitalizations. Of all the causes of first hospitalization after
discharge for worsened HF, only 813 (19.7% of all
patients) were for non-CV reasons, whereas 1,346
(32.6% of all patients) were for CV causes, again with
HF hospitalization dominating the first hospitalization
(23.8% of all patients).
Compared to patients with non-CV hospitalization,
patients hospitalized for HF had higher NYHA class
(45.1% vs 39.4% NYHA class IV), higher BNP (970 [IQR
American Heart Journal
May 2010
844 O'Connor et al
Table II. Patient characteristics for mode of death
Demographic
Total (%)
Age, mean (SD)
Male, n (%)
Race
White, n (%)
Black, n (%)
Other, n (%)
Region
North America, n (%)
South America, n (%)
Western Europe, n (%)
Eastern Europe, n (%)
Physical and laboratory findings
Weight, mean (SD), kg
Dyspnea, n (%)
Systolic BP, mean (SD), mm Hg
Diastolic BP, mean (SD), mm Hg
Heart rate, mean (SD), beat/min
NYHA class, n (%)
III
IV
JVD ≥10 cm, n (%)
Heart murmur, n (%)
Rales, n (%)
Pedal edema, n (%)
Ejection fraction, mean (SD), %
Serum BUN, mean (SD), mg/dL
Serum creatinine level, mean (SD), mg/dL
Serum sodium, mean (SD), mg/dL
Arginine vasopressin, mean (SD), pg/mL
Serum BNP, median (IQR), pg/mL⁎
QRS ≥120 ms, n (%)†
Medical history, n (%)
Hospitalization for HF
CAD
MI
Hypertension
Hypercholesterolemia
Mitral valve disease
Atrial fibrillation on admission
Diabetes
Chronic kidney disease
Severe COPD
Peripheral vascular disease
Revascularization and device use, n (%)
Previous PCI
Previous CABG
No implantable cardioverter-defibrillator or pacemaker
Pacemaker
Implantable cardioverter-defibrillator
Medications at discharge, n (%)‡
ACE I/ARBs
β-Blocking agents
Spironolactone
Digoxin
Diuretics
Lipid lowering
Nitroglycerin
Amiodarone
Inotropic agents
Intravenous nitroglycerin
Death from HF
SCD
Other cause of death
No death
443 (10.7)
68.8 (12.4)
345 (77.9)
281 (6.8)
65.1 (12.6)
229 (81.5)
356 (8.6)
68.9 (11.6)
255 (71.6)
3053 (73.9)
65.0 (11.6)
2246 (73.6)
380 (10.8)
28 (9.0)
35 (12.1)
230 (6.5)
30 (9.7)
21 (7.3)
299 (8.5)
35 (11.3)
22 (7.6)
2624 (74.3)
217 (70.0)
211 (73.3)
186
81
78
98
(14.9)
(11.6)
(13.8)
(6.1)
90
51
34
106
158 (12.6)
52 (7.4)
51 (9.0)
95 (5.9)
817
515
401
1320
79.7
392
111.0
66.9
79.7
(19.3)
(91.0)
(17.2)
(11.2)
(15.9)
83.5
261
115.2
71.1
80.4
P
b.001
.005
.153
b.001
193 (44.7)
237 (54.9)
144 (33.9)
292 (67.9)
353 (81.7)
42 (9.7)
24.5 (8.0)
43.4 (24.5)
1.7 (0.8)
137.1 (5.5)
6.8 (8.2)
1519 (2470)
289 (68.2)
(7.2)
(7.3)
(6.0)
(6.5)
(65.3)
(73.7)
(71.1)
(81.5)
b.001
.133
b.001
b.001
.123
b.001
(18.8)
(94.6)
(18.2)
(12.8)
(15.4)
80.0 (17.7)
308 (89.3)
117.3 (18.5)
69.0 (13.4)
78.7 (16.5)
84.1 (18.8)
2718 (90.9)
122.7 (19.7)
74.0 (12.5)
80.0 (15.6)
153 (55.4)
123 (44.6)
96 (34.9)
168 (61.1)
225 (81.5)
24 (8.7)
25.2 (7.8)
32.0 (16.1)
1.4 (0.5)
139.0 (4.8)
5.8 (4.3)
966 (1793)
168 (62.7)
181 (52.5)
163 (47.3)
111 (32.7)
198 (57.6)
280 (81.2)
28 (8.1)
26.4 (8.2)
35.7 (19.1)
1.5 (0.6)
138.9 (5.2)
5.6 (5.4)
1169 (2183)
225 (66.6)
1877 (62.7)
1099 (36.7)
731 (24.6)
1688 (56.4)
2437 (81.4)
292 (9.8)
28.3 (7.9)
27.4 (13.0)
1.3 (0.5)
140.1 (4.3)
5.5 (5.0)
571 (1264)
1444 (49.7)
b.001
b.001
.997
.005
b.001
b.001
b.001
b.001
.001
b.001
b.001
388
321
243
268
225
180
122
194
212
57
107
(88.0)
(72.5)
(54.9)
(60.5)
(50.9)
(40.6)
(27.6)
(43.8)
(48.0)
(12.9)
(24.2)
229
199
152
202
137
79
74
110
89
31
60
(82.1)
(70.8)
(54.1)
(71.9)
(48.8)
(28.1)
(26.4)
(39.2)
(31.7)
(11.0)
(21.4)
293 (82.5)
258 (72.9)
196 (55.1)
262 (73.6)
173 (48.6)
148 (41.6)
95 (26.8)
162 (45.5)
138 (38.8)
61 (17.1)
86 (24.3)
2340 (77.0)
2133 (69.9)
1496 (49.1)
2200 (72.1)
1468 (48.4)
897 (29.4)
816 (26.9)
1132 (37.1)
668 (21.9)
267 (8.8)
613 (20.1)
b.001
.519
.017
b.001
.808
b.001
.985
.002
b.001
b.001
.088
93
129
262
139
106
(21.0)
(29.1)
(59.1)
(31.4)
(23.9)
52
62
218
50
35
(18.5)
(22.1)
(77.6)
(17.8)
(12.5)
76 (21.4)
100 (28.1)
231 (64.9)
90 (25.3)
70 (19.7)
517 (17.0)
571 (18.7)
2393 (78.4)
436 (14.3)
389 (12.8)
.052
b.001
b.001
b.001
b.001
276
235
210
191
347
132
48
103
24
0
(73.8)
(62.8)
(56.2)
(51.1)
(92.8)
(35.3)
(12.8)
(27.5)
(6.4)
(0.0)
216
189
170
142
254
99
25
54
7
0
(79.7)
(69.7)
(62.7)
(52.4)
(93.7)
(36.5)
(9.2)
(19.9)
(2.6)
(0.0)
251 (74.7)
229 (68.2)
162 (48.2)
180 (53.6)
319 (94.9)
124 (36.9)
55 (16.4)
70 (20.8)
16 (4.8)
2 (0.6)
2674 (87.7)
2367 (77.7)
1813 (59.5)
1396 (45.8)
2832 (92.9)
1196 (39.2)
275 (9.0)
476 (15.6)
19 (0.6)
3 (0.1)
b.001
b.001
b.001
.005
.539
.372
b.001
b.001
b.001
.073
American Heart Journal
Volume 159, Number 5
O'Connor et al 845
Table II (continued )
Death from HF
Nesiritide
12 (3.2)
SCD
5 (1.9)
Other cause of death
9 (2.7)
No death
P
59 (1.9)
.354
Reference laboratory values are as follows: arginine vasopressin, 0 to 8 pg/mL; BNP, 0 to 100 pg/mL. Race and region data are presented as percentages within each race or region.
JVD, Jugular venous distention; BUN, blood urea nitrogen; CAD, coronary artery disease; COPD, chronic obstructive pulmonary disease; PCI, percutaneous coronary intervention;
CABG, coronary artery bypass graft; ACE I, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker.
⁎ Reported on 2941 patients (see text).
† Reported on 2962 patients (see text).
‡ n = 374, 271, 336, and 3048 at discharge for HF deaths, SCD, other deaths, and no deaths, respectively.
Table III. Hospitalizations in EVEREST
Cause of
hospitalization
Any hospitalization
Patients
1st hospitalization
Total hospitalizations
Non-CV hospitalization
Patients
1st hospitalization
Total hospitalizations
CV hospitalization
Patients
1st hospitalization
Total hospitalizations
HF hospitalization
Patients
1st hospitalization
Total hospitalizations
Arrhythmia
Patients
1st hospitalization
Total hospitalizations
MI
Patients
1st hospitalization
Total hospitalizations
Stroke
Patients
1st hospitalization
Total hospitalizations
Other CV
Patients
1st hospitalization
Total hospitalizations
n
% of
patients
2159
2159
5239
52.2
52.2
1225
813
2053
29.6
19.7
1625
1346
3186
39.3
32.6
1287
982
2426
31.1
23.8
192
100
220
4.7
2.4
67
33
69
1.6
0.8
69
42
70
1.7
1.0
342
189
401
8.3
4.6
% of
hospitalizations
100.0
39.2
60.8
46.3
4.2
1.3
1.3
7.7
1,918] vs 788 [IQR 1,505] pg/mL), lower rates of diabetes
(39.2% vs 43.8%), chronic kidney disease (31.7% vs
48.0%), and severe chronic obstructive pulmonary
disease (11.0% vs 12.9%), and lower rates of revascularization and device use (Table IV). Medication use was
similar between patients hospitalized for HF versus nonCV hospitalization. The HF hospitalizations were the most
common cause of first hospitalization in South America
and Eastern Europe, whereas there was equal numbers of
HF and non-CV hospitalizations in North America and a
higher percentage of patients with non-CV than HF
hospitalizations in Western Europe.
Timing of events
There were 178 deaths (16.5% of all deaths) that
occurred in the first 30 days post-randomization, whereas
112 (10.4%) occurred at 31 to 60 days and 790 (73.1%)
after 60 days over the entire follow-up period. Of all SCD,
41 (14.6%) occurred within 30 days after discharge,
whereas 28 (10.0%) occurred within 31 to 60 days
(Figure 1). Of all HF deaths, 102 (23.0%) occurred in the
first 30 days after discharge, compared to 52 (11.7%)
within 31 to 60 days. There remained a high percentage
of patients with both SCD and HF death continued after
60 days (75.4% and 65.2%, respectively) during the
entire follow-up period. Of all first rehospitalizations,
496 (23.0%) occurred within the first 30 days after
discharge, 413 (19.1%) between 31 and 60 days and 1,250
(57.9%) after 60 days (Figure 2). Of all first rehospitalizations for HF, 237 (24.1%) occurred within the first
30 days, whereas 191 (19.5%) occurred within 31 to
60 days. Of all first non-CV hospitalizations, 188 (23.1%)
occurred within 30 days after discharge, 148 (18.2%)
within 31 to 60 days, and 477 (58.7%) after 60 days. A
similar pattern was noted non-HF CV hospitalizations.
Discussion
To the best of our knowledge, this is the first study to
systematically and prospectively analyze the cause of
death and rehospitalization in a population of patients
hospitalized for HF and on modern medical therapy. We
demonstrated a high rate of both death and rehospitalization. In addition, we found that most patients died of
progressive HF and approximately one quarter of sudden
and unexpected death. Despite a clinical trial population
with relatively few comorbidities, approximately one half
of rehospitalizations not related to HF.
Retrospective analysis of a subset of patients of the
Candesartan in Heart failure: Assessment of Reduction in
Mortality and morbidity (CHARM) trials demonstrated
that in patients with chronic HF and preserved and
reduced ejection fraction, nonfatal hospitalizations for HF
were associated with increased mortality, with the
American Heart Journal
May 2010
846 O'Connor et al
Table IV. Patient characteristics for mode of first hospitalization
Demographic
Total (%)
Age, mean (SD)
Male, n (%)
Race
White, n (%)
Black, n (%)
Other, n (%)
Region
North America, n (%)
South America, n (%)
Western Europe, n (%)
Eastern Europe, n (%)
Physical and laboratory findings
Weight, mean (SD), kg
Dyspnea, n (%)
Systolic BP, mean (SD), mm Hg
Diastolic BP, mean (SD), mm Hg
Heart rate, mean (SD), beat/min
NYHA class, n (%)
III
IV
JVD ≥10 cm, n (%)
Heart murmur, n (%)
Rales, n (%)
Pedal edema, n (%)
Ejection fraction, mean (SD), %
Serum BUN, mean (SD), mg/dL
Serum creatinine level, mean (SD), mg/dL
Serum sodium, mean (SD), mg/dL
Arginine vasopressin, mean (SD), pg/mL
Serum BNP, median (IQR), pg/mL⁎
QRS ≥120 ms, n (%)†
Medical history, n (%)
Hospitalization for HF
CAD
Previous MI
Hypertension
Hypercholesterolemia
Mitral valve disease
Atrial fibrillation on admission
Diabetes
Chronic kidney disease
Severe COPD
Peripheral vascular disease
Revascularization and device use, n (%)
Previous PCI
Previous CABG
No implantable cardioverter-defibrillator
or pacemaker
Pacemaker
Implantable cardioverter-defibrillator
Medications at discharge, n(%)‡
ACE I/ARBs
β-Blocking agents
Spironolactone
Digoxin
Diuretics
Lipid lowering
Nitroglycerin
Amiodarone
Inotropic agents
Non-CV
hospitalization
HF
hospitalization
Other CV
hospitalization
No
hospitalization
813 (19.7)
66.9 (12.0)
612 (75.3)
982 (23.8)
65.8 (12.4)
76 (77.0)
364 (8.8)
67.2 (11.6)
273 (75.0)
1974 (47.8)
65.0 (11.5)
1434 (72.6)
696 (19.7)
69 (22.3)
48 (16.6)
799 (22.6)
105 (33.9)
77 (26.6)
312 (8.8)
30 (9.7)
22 (7.6)
1726 (48.9)
106 (34.2)
142 (49.1)
373
116
172
152
(29.8)
(16.6)
(30.5)
(9.4)
378
174
145
285
(30.2)
(24.9)
(25.7)
(17.6)
121
59
47
137
(9.7)
(8.4)
(8.3)
(8.5)
379
350
200
1045
(30.3)
(50.1)
(35.5)
(64.5)
84.4
697
118.5
70.0
78.6
(19.5)
(88.9)
(19.1)
(12.5)
(15.3)
83.1 (18.9)
872 (91.1)
116.3 (18.0)
70.1 (11.9)
79.2 (14.6)
82.9
331
121.5
72.0
79.1
(20.0)
(93.0)
(20.4)
(12.3)
(17.2)
82.9
1779
123.2
75.1
80.8
(18.4)
(91.4)
(20.1)
(12.7)
(16.0)
471 (60.0)
309 (39.4)
222 (28.7)
410 (52.4)
631 (80.5)
593 (75.6)
26.9 (8.2)
33.3 (17.8)
1.5 (0.5)
139.1 (4.7)
5.4 (5.1)
788(1505)
289 (68.2)
523 (54.5)
432 (45.1)
277 (29.3)
579 (60.5)
768 (80.1)
744 (77.6)
25.8 (8.0)
33.4 (16.9)
1.5 (0.5)
138.8 (4.7)
5.6 (4.3)
971 (1918)
168 (62.7)
223 (62.3)
132 (36.9)
88 (24.7)
207 (58.0)
294 (82.4)
277 (77.4)
27.9 (8.3)
30.6 (17.6)
1.4 (0.5)
139.9 (4.3)
5.4 (5.3)
584 (1361)
225 (66.6)
388 (88.0)
321 (72.5)
243 (54.9)
268 (60.5)
225 (50.9)
180 (40.6)
122 (27.6)
194 (43.8)
212 (48.0)
57 (12.9)
107 (24.2)
229 (82.1)
199 (70.8)
152 (54.1)
202 (71.9)
137 (48.8)
79 (28.1)
74 (26.4)
110 (39.2)
89 (31.7)
31 (11.0)
60 (21.4)
293
258
196
262
173
148
95
162
138
61
86
93 (21.0)
129 (29.1)
262 (59.1)
P
b.001
.071
b.001
b.001
.188
.096
b.001
b.001
.001
.031
1187 (61.0)
749 (38.5)
495 (25.6)
1150 (59.1)
1602 (82.2)
1595 (82.0)
28.5 (7.9)
27.3 (14.6)
1.3 (0.6)
140.2 (4.6)
6.0 (6.1)
590 (1377)
1444 (49.7)
.083
.004
.449
.001
b.001
b.001
b.001
b.001
.006
b.001
b.001
(82.5)
(72.9)
(55.1)
(73.6)
(48.6)
(41.6)
(26.8)
(45.5)
(38.8)
(17.1)
(24.3)
2340 (77.0)
2133 (69.9)
1496 (49.1)
2200 (72.1)
1468 (48.4)
897 (29.4)
816 (26.9)
1132 (37.1)
668 (21.9)
267 (8.8)
613 (20.1)
b.001
.519
.017
b.001
.808
b.001
.985
.002
b.001
b.001
.088
52 (18.5)
62 (22.1)
218 (77.6)
76 (21.4)
100 (28.1)
231 (64.9)
517 (17.0)
571 (18.7)
2393 (78.4)
.052
b.001
b.001
139 (31.4)
106 (24.0)
50 (17.8)
35 (12.5)
90 (25.3)
70 (19.7)
436 (14.4)
389 (12.8)
b.001
b.001
668 (82.3)
607 (74.8)
428 (52.7)
381 (46.9)
769 (94.7)
377 (46.4)
112 (13.8)
171 (21.1)
20 (2.5)
795 (81.1)
721 (73.6)
558 (56.9)
514 (52.5)
928 (94.7)
383 (39.1)
117 (11.9)
211 (21.5)
18 (1.8)
1639 (87.5)
1424 (76.0)
1178 (62.9)
839 (44.8)
1715 (91.5)
627 (33.5)
119 (6.4)
257 (13.7)
24 (1.3)
b.001
.508
b.001
.002
.002
b.001
b.001
b.001
.122
315
268
191
175
340
164
55
64
4
(86.8)
(73.8)
(52.6)
(48.2)
(93.7)
(45.2)
(15.2)
(17.6)
(1.1)
American Heart Journal
Volume 159, Number 5
O'Connor et al 847
Table IV (continued )
Intravenous nitroglycerin
Nesiritide
Non-CV
hospitalization
HF
hospitalization
Other CV
hospitalization
No
hospitalization
P
4 (0.5)
25 (3.1)
0 (0.0)
30 (3.1)
1 (0.3)
12 (3.3)
0 (0.0)
18 (1.0)
.004
b.001
Race and region data are presented as percentages within each race or region. Refer to Table II for abbreviations.
⁎ Reported on 2941 patients (see text).
† Reported on 2962 patients.
‡ n = 813, 982, 364, and 1974 at discharge for non-CV, HF, other, and no hospitalizations, respectively.
Figure 1
Figure 2
Timing of primary modes of death.
highest risk occurring in the first month postdischarge,
and primary causes due to HF and SCD.17 The current
study is the only trial available that prospectively analyzes
mode of death in patients hospitalized with HF and found
a high rate of HF deaths (41%) as well as SCD (26%). The
rate of HF deaths is approximately twice that of SCD in all
geographic regions studied with the exception of Eastern
Europe, a difference possibly explained by the relatively
low severity of HF observed in prior analysis of this region
in EVEREST.18 We observed that compared to patients
who died of HF, patients who died of SCD had half the
rate of prior implantable cardioverter-defibrillator implantations (12.5% vs 23.9%).
Recently, the high rate of rehospitalization has caught
the attention of policymakers and international scientific
societies who are determined to improve quality of care
and reduce health care costs. We found that nearly 40% of
all hospitalizations were due to non-CV causes, despite the
exclusion of major comorbidities in this trial. Non-CV
hospitalizations were more common in North America
and Western Europe, populations with more major
comorbidities relative to the other 2 regions studied in
Timing of major causes of first hospitalization.
EVEREST.18 Major clinical trials in chronic HF have
focused the end points on death and CV morbidity,
assuming the chief cause of hospitalization is for HF or
other CV-related causes. Post hoc analysis of the carvedilol
prospective randomized cumulative survival (COPERNICUS) study of carvedilol or placebo in patients with severe
chronic systolic HF demonstrated that 28% of all
hospitalizations were for non-CV causes,19 whereas the
primary analysis of the controlled rosuvastatin multinational trial in heart failure (CORONA) study of rosuvastatin
versus placebo in elderly patients with NYHA class II to IV
chronic systolic HF reported 56% of all hospitalizations
due to non-CV causes.20 Brown and Cleland3 established
that other major medical comorbidities complicate HF
admissions in a Scottish database, a finding also noted in
a Medicare analysis.4 Comorbidities such as chronic
obstructive pulmonary disease, renal failure, diabetes,
depression, and other lower respiratory diseases may be
undertreated in patients with a chronic disease such as
HF, which may lead to preventable hospitalizations.4
Patients with major chronic conditions often have other
comorbidities, and these comorbidities tend to be
American Heart Journal
May 2010
848 O'Connor et al
undertreated when one chronic condition dominates.6 In
the early postdischarge period, patients are on multiple
new and competing therapies, which can contribute to
polypharmacy and reduced medical compliance. Perhaps
refocusing therapies toward these other important
chronic conditions may reduce rehospitalizations in
patients recently discharged after a HF hospitalization.
To prevent rehospitalizations due to HF, it may be
important to closely monitor signs and symptoms of
congestion. Recent analysis of EVEREST demonstrated
that increase in body weight was an independent
predictor of HF hospitalizations21 and was the only
predictor in which a change in the variable independently predicted rehospitalization,22 suggesting that congestion, or retention of fluid leading to increased body
weight, is the central mechanism behind HF hospitalization. Knowledge of the type of rehospitalizations after the
index HF hospitalizations will allow hospital systems to
establish methods for close follow-up during this period
and potentially treating the additional comorbidities to
reduce hospitalization during this period.
initiation of appropriate treatment strategies in the
early postdischarge period after a hospitalization for
worsened HF.
Disclosures
Limitations
This analysis was performed after the trial results were
analyzed and published. However, such a cause-specific
analysis was prespecified in the trial design, and a
separate adjudication committee independently verified
each death and hospitalization event increasing the
validity of the end points. First, the adjudication
committee has proven essential in past trials for acute
MI, as several important disagreements between investigator assessments at each site and the central events
committee has been noted.23,24 Second, we did not
prospectively adjudicate on the type of non-CV hospitalization or death, limiting the analysis in this group. Third,
mode of death analysis is dependent on the adjudication
of events and is dependent on the amount and quality of
data collected for each event, as well as the definitions
used for modes of death and hospitalization. However,
the adjudication process in this study was rigorously
constructed, and the most current definitions of events
were used, such that any uncertainty surrounding each
event was minimized. Finally, because this was a clinical
trial, with specific entrance criteria, the applicability of
these data is limited to that population. However, the
study is the largest study of AHFS to date, with baseline
and follow-up data carefully collected and validated.
Dr O'Connor reports having received funding from
the National Heart Lung and Blood Institute (NHLBI),
Amgen, Astra, Bristol-Meyers Squibb, GlaxoSmithKline,
Guidant, Medtronic, Merck, Nitrox LLC, Novartis,
Otsuka, Pfizer, ArcaBioPharma, Sanofi-Sythelabo, and
MedPace; Dr Miller Reports having received research
grants from Otsuka and Pfizer and honoraria from
Medtronic, Nitromed, Novartis, Pfizer, Sanofi, and Scios
Inc; Dr Blair reports having been a consultant for
SigmaTau; Dr Konstam reports the following companies,
involved in development of drugs or devices for HF, for
which he consulted, performed research, or have had
other financial relationships: Otsuka, Merck, Cardiokine,
Biogen, Orqis Medical, Boston Scientific, Sanofi, Cytokinetics, and Novartis; Ms Wedge reports having received
funding from Otsuka; Dr Hauptman reports having been
a consultant for Otsuka, BioControl Medical, Merck,
Cardiokine, and ArcaBioPharma and having been on the
speakers bureau of GlaxoSmithKline; Dr Metra reports
having received consulting fees from Corthera, Duke
Clinical Research Institute, Merck, Nile therapeutics, and
Servier; Dr Piña reports having received grant support
through Case Western Reserve University and the
NHLBI, Office of Women's Health (Health and Human
Services); having been on the speakers bureau for
AstraZeneca, Merck, Solvay, Novartis, and Innovia; and
having been a consultant for the Food and Drug
Administration and Sanofi-Aventis; Mr Traver and Dr
Cook report having received compensation through a
contract between the University of Wisconsin and
Otsuka; Dr Gheorghiade reports having been a consultant for Otsuka, Solvay Pharma, Novartis, Bayer, Sigma
Tau, Debiopharm, Medtronic, Merck, Astellas, Cytokinetics, CorThera Inc, Pericor Therapeutics, GlaxoSmithKline, Johnson & Johnson, Abbott, Errekappa
Terapeutici, Protein Design Laboratories, AstraZeneca,
Protein Design Laboratories, and Sanofi-Aventis. No
other disclosures were reported.
Funding sources: This study was supported by Otsuka
Pharmaceuticals under the guidance of the EVEREST
steering committee.
Conclusion
References
Future registries for AHFS should adjudicate events to
obtain a more accurate real-world causes of death and
rehospitalization. Eventually, clinical trials should be
designed to target specific modes of death and
rehospitalization to reduce such events. Knowledge
of the causes of these events may allow for early
1. Lloyd-Jones D, Adams R, Carnethon M, et al. Heart Disease and
Stroke Statistics—2009 Update. A Report From the American Heart
Association Statistics Committee and Stroke Statistics Subcommittee.
Circulation 2008.
2. Fang J, Mensah GA, Croft JB, et al. Heart failure-related
hospitalization in the U.S., 1979 to 2004. J Am Coll Cardiol 2008;
52:428-34.
American Heart Journal
Volume 159, Number 5
3. Brown AM, Cleland JG. Influence of concomitant disease on patterns
of hospitalization in patients with heart failure discharged from
Scottish hospitals in 1995. Eur Heart J 1998;19:1063-9.
4. Redelmeier DA, Tan SH, Booth GL. The treatment of unrelated
disorders in patients with chronic medical diseases. N Engl J Med
1998;338:1516-20.
5. O'Connor CM, Carson PE, Miller AB, et al. Effect of amlodipine on
mode of death among patients with advanced heart failure in the
PRAISE trial. Prospective Randomized Amlodipine Survival Evaluation. Am J Cardiol 1998;82:881-7.
6. Carson P, Anand I, O'Connor C, et al. Mode of death in advanced
heart failure: the Comparison of Medical, Pacing, and Defibrillation
Therapies in Heart Failure (COMPANION) trial. J Am Coll Cardiol
2005;46:2329-34.
7. Mozaffarian D, Anker SD, Anand I, et al. Prediction of mode of death
in heart failure: the Seattle Heart Failure Model. Circulation 2007;
116:392-8.
8. Poole-Wilson PA, Uretsky BF, Thygesen K, et al. Mode of death
in heart failure: findings from the ATLAS trial. Heart 2003;89:42-8.
9. Remme WJ, Cleland JG, Erhardt L, et al. Effect of carvedilol and
metoprolol on the mode of death in patients with heart failure. Eur J
Heart Fail 2007;9:1128-35.
10. Gheorghiade M, Zannad F, Sopko G, et al. Acute heart failure
syndromes: current state and framework for future research.
Circulation 2005;112:3958-68.
11. Gheorghiade M, Pang PS. Acute heart failure syndromes. J Am Coll
Cardiol 2009;53:557-73.
12. Epstein AM. Revisiting readmissions—changing the incentives for
shared accountability. N Engl J Med 2009;360:1457-9.
13. Jencks SF, Williams MV, Coleman EA. Rehospitalizations among
patients in the Medicare fee-for-service program. N Engl J Med 2009;
360:1418-28.
14. Gheorghiade M, Konstam MA, Burnett Jr JC, et al. Short-term clinical
effects of tolvaptan, an oral vasopressin antagonist, in patients
hospitalized for heart failure: the EVEREST Clinical Status Trials.
JAMA 2007;297:1332-43.
15. Gheorghiade M, Orlandi C, Burnett JC, et al. Rationale and design
of the multicenter, randomized, double-blind, placebo-controlled study
to evaluate the Efficacy of Vasopressin antagonism in Heart Failure:
Outcome Study with Tolvaptan (EVEREST). J Card Fail 2005;11:260-9.
O'Connor et al 849
16. Konstam MA, Gheorghiade M, Burnett Jr JC, et al. Effects of oral
tolvaptan in patients hospitalized for worsening heart failure: the
EVEREST Outcome Trial. JAMA 2007;297:1319-31.
17. Solomon SD, Dobson J, Pocock S, et al. Influence of nonfatal
hospitalization for heart failure on subsequent mortality in patients
with chronic heart failure. Circulation 2007;116:1482-7.
18. Blair JE, Zannad F, Konstam MA, et al. Continental differences in
clinical characteristics, management, and outcomes in patients
hospitalized with worsening heart failure results from the EVEREST
(Efficacy of Vasopressin Antagonism in Heart Failure: Outcome
Study with Tolvaptan) program. J Am Coll Cardiol 2008;52:
1640-8.
19. Packer M, Fowler MB, Roecker EB, et al. Effect of carvedilol on the
morbidity of patients with severe chronic heart failure: results of the
carvedilol prospective randomized cumulative survival (COPERNICUS) study. Circulation 2002;106:2194-9.
20. Kjekshus J, Apetrei E, Barrios V, et al. Rosuvastatin in older
patients with systolic heart failure. N Engl J Med 2007;357:
2248-61.
21. Blair J, Khan S, Konstam M, et al. Weight changes after
hospitalization for worsening heart failure and subsequent rehospitalization and mortality in the EVEREST trial. European Heart
Journal 2009;30:1666-73.
22. Gheorghaide M, Filippatos G, Pang P, et al. Changes in clinical,
neurohormonal, electrolyte, renal, and hepatic profiles during and
after hospitalization for acute decompensated heart failure: analysis
of the EVEREST trial (late-breaking clinical trial). Presented at
European Society of Cardiology—2008 Scientific Sessions. September 2; 2008.
23. Mahaffey KW, Harrington RA, Akkerhuis M, et al. Disagreements
between central clinical events committee and site investigator
assessments of myocardial infarction endpoints in an international
clinical trial: review of the PURSUIT study. Curr Control Trials
Cardiovasc Med 2001;2:187-94.
24. Mahaffey KW, Roe MT, Dyke CK, et al. Misreporting of myocardial
infarction end points: results of adjudication by a central clinical
events committee in the PARAGON-B trial. Second Platelet IIb/IIIa
Antagonist for the Reduction of Acute Coronary Syndrome Events in
a Global Organization Network Trial. Am Heart J 2002;143:
242-8.
American Heart Journal
Volume 159, Number 5
Appendix. Clinical Events Committee
C. O'Connor (cochair), Duke University, Durham, NC; A.
Miller (cochair), University of Florida, Jacksonville, FL; M.
C. Bahit, Hospital Italiano de Buenos Aires, Buenos Aires,
Argentina; P. Carson, Washington VA Medical Center,
Washington, DC; M. Haass, Theresienkrankenhaus, Mannheim, Germany; R. Patten, Tufts-New England Medical
O'Connor et al 849.e1
Center, Boston, MA; P. Hauptman, St Louis University
School of Medicine, St Louis, MO; I. Piña, Case Western
Reserve University, Cleveland, OH; M. Metra, University
of Brescia, Brescia, Ital; R. Oren, Iowa City Heart Canter
PC, Iowa City, IA; S. Roth, The Scarborough Hospital,
Toronto, Ontario, Canada; J. Sackner-Bernstein, Dobbs
Ferry, New York.