European Journal of Heart Failure Advance Access published May 12, 2013

European Journal of Heart Failure

doi:10.1093/eurjhf/hft082

Renin–angiotensin system blockade in

heart failure patients on long-term
haemodialysis in Taiwan
Chao-Hsiun Tang 1, Tso-Hsiao Chen 2,3, Chia-Chen Wang 4, Chuang-Ye Hong 2,3,
Kuan-Chih Huang 1, and Yuh-Mou Sue 2,3*
1
School of Health Care Administration, Taipei Medical University, Taipei, Taiwan; 2Department of Internal Medicine, Wan Fang Hospital, Taipei Medical University, Taipei, Taiwan;

Downloaded from http://eurjhf.oxfordjournals.org/ at Jichi Medical School on October 23, 2013

3
Department of Internal Medicine, School of Medicine, College of Medicine, Taipei Medical University, Taipei, Taiwan; and 4School of Medicine, Fu-Jen Catholic University, New Taipei
City, Taiwan

Received 13 February 2013; revised 15 April 2013; accepted 19 April 2013

Aims Heart failure is among the most frequent complications of patients on long-term haemodialysis. The benefits of renin –
angiotensin system (RAS) blockade on the outcomes of these patients have yet to be determined.
.....................................................................................................................................................................................
Methods We conducted a nationwide observational study using data from the Taiwan National Health Insurance claims database,
and results between 1999 and 2010. We enrolled patients aged ≥35 years with new-onset heart failure [diagnosed by International
Classification of Diseases, 9th revision, clinical modification (ICD-9-CM) codes] under treatment with medications. New
users of a RAS blocker (RASB; i.e. an ACE inhibitor or an ARB used as monotherapy or dual therapy) were selected to
compare with non-RASB users. We used Cox proportional hazards regression with and without propensity score ad-
justment to compare the risk of 3-year all-cause and cardiovascular mortality. Stratified analyses and RASB therapy dur-
ation as a time-dependent covariate were also performed. In all, 4771 were treated with an RASB (n ¼ 3024) or without
an RASB (n ¼ 1747). RASB users had a higher prevalence of hypertension and diabetes, and a higher number of hospi-
talization. Among RASB users, 1148 deaths (38.0%) occurred during 5272 person-years of follow-up compared with 734
deaths (42.0%) among non-RASB users during 2683 person-years of follow-up. Three-year mortality rates were 45.4%
and 49.1% for patients receiving and those not receiving an RASB, respectively (log-rank test, P , 0.001). Adjusted hazard
analysis revealed that RASB therapeutic effects remained significant on all-cause [hazard ratio (HR) 0.8; 95% confidence
interval (CI) 0.72–0.89; P , 0.001] and cardiovascular mortality (HR 0.76; 95% CI 0.64–0.90; P , 0.01).
.....................................................................................................................................................................................
Conclusions RASB therapy reduced all-cause and cardiovascular mortality in heart failure patients on long-term haemodialysis.
-----------------------------------------------------------------------------------------------------------------------------------------------------------
Keywords Haemodialysis † Heart failure † Renin –angiotensin system † ACE inhibitor † ARB † Mortality

HF is 36%, while its incidence is  71/1000 person-years, which is

Introduction
The worldwide incidence and prevalence of populations requiring
dialysis is increasing. The incidence and prevalence of patients on dia-
lysis in Taiwan increased rapidly after the launch of National Health
Insurance (NHI) in 1995.1 Compared with patients with normal
renal function, patients on long-term dialysis are at increased risk of
cardiovascular events, which comprise a major cause of morbidity
and mortality.2,3 Chronic heart failure (HF) appears to be among
the most frequent cardiovascular complications. The prevalence of
higher than that of acute coronary syndrome (29/1000 person-
years).4
Multiple large clinical trials of the general population revealed that
the use of renin–angiotensin system blockers (RASBs) (including
ACE inhibitors and ARBs) improves survival in various clinical condi-
tions, including post-myocardial infarction and chronic HF.5,6 The use
of ARBs for patients intolerant to ACE inhibitors is recommended as
an alternative approach in patients with chronic HF and preserved or
low LVEF.5 – 9 However, patients on long-term dialysis were excluded

* Corresponding author. Department of Internal Medicine, Wan Fang Hospital, Taipei Medical University, 5th Floor, 111 Xing-Long Road, Section 3, Wen Shan District, Taipei City 116,
Taiwan. Tel: +886 2 29307930 ext. 8103, Fax: +886 2 29302448, Email: sueym@tmu.edu.tw
Published on behalf of the European Society of Cardiology. All rights reserved. & The Author 2013. For permissions please email: journals.permissions@oup.com.
Page 2 of 9 C.-H. Tang et al.

from these trials because they greatly differed from individuals not
undergoing dialysis. These patients experienced daily volume fluctua-
tions, heightened inflammatory states, and altered drug pharmaco-
kinetics. Therefore, it is not known whether the results of those
studies also apply to a large population of patients on long-term dia-
lysis who are at high risk of cardiovascular events.
Several studies of patients on dialysis but not specific to chronic HF
demonstrated that blockade of the RAS is associated with reductions
in LV mass, mortality, and cardiovascular events.10 – 14 In a retrospect-
ive analysis, Efrati et al. 11 studied the effects of ACE inhiboitors on
mortality in patients undergoing long-term haemodialysis (HD)
therapy. The results showed a survival benefit for patients on HD

Table 1 Baseline characteristics of heart failure patients on long-term haemodialysisa

Characteristic Treatment Control group P-value Propensity score

group (n 5 3024) (n 5 1747) adjusted P-value
...................... ......................
n (%) n (%)
...............................................................................................................................................................................
Men 1500 (49.6) 796 (45.6) ,0.01 0.98
Age at cohort entry ,0.001 0.99
35–44 107 (3.5) 61 (3.5)
45–54 444 (14.7) 201 (11.5)
55–64 742 (24.5) 345 (19.8)
65–74 986 (32.6) 561 (32.1)
≥75 745 (24.6) 579 (33.1)
No. of hospitalizationsb ,0.001 0.99
0 448 (14.8) 326 (18.7)
1 789 (26.1) 547 (31.3)
2 542 (17.9) 358 (20.5)
≥3 1245 (41.2) 516 (29.5)
Duration of dialysis at enrolment, mean (months, SD) 31.3 (24.1) 29.1 (25.8) ,0.01 ,0.001
Charlson co-morbid indexb ,0.001 0.99
0 –2 235 (7.8) 196 (11.2)
3 –5 1654 (54.7) 1001 (57.3)
≥6 1135 (37.5) 550 (31.5)
Co-morbiditiesb (ICD-9-CM codes)
Diabetes mellitus (250) 1526 (50.5) 823 (47.1) 0.03 0.99
Hypertension (401–405) 2322 (76.8) 1275 (73.0) ,0.01 0.96
Cancer (140–208) 295 (9.8) 192 (11.0) 0.17 0.07
COPD (491– 493, 495– 496) 368 (12.2) 251 (14.4) 0.03 0.98
Gastric ulcer (531– 534) 585 (19.4) 339 (19.4) 0.96 0.69
Cirrhosis of liver (571) 122 (4.0) 64 (3.7) 0.52 0.92
Dementia (290) 99 (3.3) 67 (3.8) 0.31 0.98
Cerebrovascular disease (430–438) 583 (19.3) 377 (21.6) 0.06 0.98
Peripheral vascular disease (440.2, 443) 109 (3.6) 82 (4.7) 0.06 0.99
Cardiac dysrhythmia (426, 427) 396 (13.1) 226 (12.9) 0.88 0.98
Ischaemic heart disease (411, 413, 414) 1168 (38.6) 606 (34.7) ,0.01 0.94
Myocardial infarction (410, 412) 133 (4.4) 68 (3.9) 0.40 0.98
Tests or proceduresb
24 h ECG 342 (11.3) 179 (10.3) 0.26 0.25
Echocardiography 1981 (65.5) 1092 (62.5) 0.04 0.45
Myocardial perfusion scan 411 (13.6) 213 (12.2) 0.17 0.96
Treadmill exercise test 116 (3.8) 56 (3.2) 0.26 0.76
Coronary angiography 293 (9.7) 131 (7.5) 0.01 0.52
Percutaneous coronary intervention 160 (5.3) 74 (4.2) 0.10 0.99
CABG surgery 24 (0.8) 12 (0.7) 0.68 0.89
Permanent pacemaker implantation 22 (0.7) 25 (1.4) 0.02 0.01

CABG, coronary artery bypass graft; ICD-9-CM, International classification of diseases, 9th revision, clinical modification.
a
ICD-9-CM codes for heart failure: 401.91, 402.01, 402.11, 404.01, 404.03, 404.11, 404.91, 404.93, and 428.
b
Within 2 years before the index date.
RAS blockade in heart failure patients on HD
who were treated with ACE inhibitors. Takahashi et al. 12 also
reported that candesartan therapy significantly reduced cardiovascu-
lar events and mortality in patients on long-term HD. However, some
investigators did not demonstrate the same benefit.15 – 19 In a rando-
mized trial, Zannad et al. 16 reported that fosinopril had no significant
benefit on fatal or non-fatal cardiovascular events in patients on long-
term dialysis. Chang et al. 17 also showed that ACE inhibitor use was
associated with a higher risk of HF-related hospitalization. Bajaj
et al. 18 and Chan et al. 19 recently conducted population-based
studies in Canada and the USA, and showed that ACE inhibitor or
ARB therapy in patients on long-term HD did not favourably affect
cardiovascular outcomes. Only 15–49% of the enrolled subjects of
these studies had chronic HF. Therefore, their conclusions could
not be applied to long-term HD patients with HF. In fact, there are
no large-scale clinical trials of an ACE inhibitor or ARB alone in HF
patients on long-term HD. Cice et al. 20 delineated the association
between combined ACE inhibitor and ARB therapy and the occur-
rence of cardiovascular events or death. However, that study’s
sample size was relatively small (332 patients) and the population
was a highly selected group of dialysis patients with HF in advanced
NYHA functional class.
Figure 1 Enrolment of study participants.
Page 3 of 9
Previous studies have not yet provided sufficient evidence on the
effects of ACE inhibitors and ARBs in patients on dialysis. Therefore,
we performed this population-based retrospective cohort study of
patients on long-term HD based on the real-world clinical practice
information within the Taiwan NHI Research Database (NHIRD)
to investigate the associations between ACE inhibitor and ARB use
and the occurrence of all-cause mortality and cardiovascular mortal-
ity in HF patients on long-term HD.
Methods
Data sources
In this study, we obtained data from the NHIRD, which was established by
the National Health Research Institute with the aim of conducting re-
search into current and emerging issues in Taiwan. The NHIRD contains
healthcare data for .99% of the entire population of Taiwan (23.74
million people). We used the International Classification of Diseases,
9th revision, clinical modification (ICD-9-CM) codes to define the dis-
eases. The following files from the NHIRD between 1999 and 2010
were used in this study: (i) the inpatient expenditure by admission and
the inpatient medical order files; (ii) the outpatient expenditure by visit
Page 4 of 9 C.-H. Tang et al.

and the outpatient medical order files; (iii) the registry of contracted
medical facilities provided data on each medical institution’s accreditation
level and geographical location; (iv) the registry for medical personnel
provided data on each medical professional’s date of birth, gender, pro-
fession, and specialty; and (v) the registry for catastrophic illness patients
provided data of 30 illness and injury categories. In Taiwan, patients who
need long-term dialysis are recognized as having a catastrophic illness
according to the NHI Bureau and they are thus exempt from
co-payments. Anonymous identifiers of the medical institutions and phy-
sicians were used to link the inpatient expenditure file and outpatient ex-
penditure file to the physician and hospital registries. This study was
approved by the National Health Research Institute, and all the data
are delinked information. Confidentiality was ensured by abiding by the
NHI Bureau data regulations.
Study subjects
Our study subjects were HF patients on long-term HD. ‘Long-term HD’
in this context means that each patient received at least 26 HD sessions
within 3 months after starting HD during 2001– 2009, and the patients
were identified through the catastrophic illness registry in the NHIRD.
We used the NHIRD between 1999 and 2010 for co-morbidity and
follow-up analysis purposes. We included patients with at least three out-
patient visit claims with a major/minor diagnosis of HF within 365 days or
one claim for incident hospitalization with a major/minor HF diagnosis.
We also defined co-morbidities by the same criteria. Diagnostic
ICD-9-CM codes for HF and other co-morbidities are listed in Table 1.
The exclusion criteria are shown in Figure 1. We also excluded patients
who had taken an RASB in the 3 months before the HF diagnosis (i.e.
washout period) and who were not taking any drugs for HF. In this
study, we defined drugs for HF as ACE inhibitors, ARBs, beta-blockers,
calcium channel blockers, hydralazine, nitrates, loop diuretics, and
digoxin. Prescription data were classified according to the Anatomical
Therapeutic and Chemical classification. We further classified the
study sample into two groups: the treatment group was defined as
‘new users’ of RASBs (ACEI and ARB used as monotherapy or dual
therapy) after HF diagnosis, whereas the control group was defined as
patients not taking any RASB for HF after the HF diagnosis. Medications
used at enrolment are listed in Table 2. Figure 1 contains a schematic illus-
tration of the study sample.
Possible confounding factors
We also considered other covariates in this study, including cumulative
days taking an RASB, age, gender, number of hospitalizations, urbaniza-
tion level, medications, co-morbidities, and tests/procedures within the
2 years before the first HF prescription. We used cumulative days
having taken an RASB as the time-dependent covariate to reflect its
nature of changing over time since it might otherwise be censored at
death.
Propensity score computation
We used propensity score analyses to adjust for the differences between
the treatment and control groups in terms of baseline patient character-
istics and to reduce the biases in the estimation process that may be
caused by these differences.21 We computed propensity scores by mod-
elling a logistic regression with the variables of patient age and sex, long-
term HD year, the duration of HD at enrolment, insured’s income,

Table 2 Medication use at enrolmenta of heart failure patients on long-term haemodialysis

Characteristic Treatment group Control group P-value Propensity score

(n 5 3024) (n 5 1747) adjusted P-value
........................... ..........................
n (%) n (%)
...............................................................................................................................................................................
Beta-blockers 1092 (36.1) 457 (26.2) ,0.001 0.85
Calcium channel blockers 1540 (50.9) 649 (37.2) ,0.001 0.88
Other hypertension drugsb 348 (11.5) 141 (8.1) ,0.001 0.88
Loop diuretics 707 (23.4) 398 (22.8) 0.67 0.99
Nitrates 1097 (36.3) 566 (32.4) ,0.01 0.91
Digoxin 160 (5.3) 92 (5.3) 0.97 0.99
Antiarrhythmics 320 (10.6) 205 (11.7) 0.22 0.99
Platelet inhibitors 804 (26.6) 426 (24.4) 0.09 0.96
Warfarin 65 (2.2) 49 (2.8) 0.15 0.99
Statins 338 (11.8) 195 (11.2) 0.99 0.99
Fibrates 180 (6.0) 95 (5.4) 0.46 0.99
H2-antagonists 574 (19.0) 320 (18.3) 0.57 0.99
Proton pump inhibitors 448 (14.8) 279 (16.0) 0.28 0.99
Benzodiazepinec 1077 (35.6) 563 (32.2) 0.02 0.97
Antidepressants 407 (13.5) 258 (14.8) 0.21 0.98
NSAIDs 1436 (47.5) 800 (45.8) 0.26 0.97
Oral hypoglycaemic drugs 608 (20.1) 337 (19.3) 0.50 0.29
Insulin 699 (23.1) 374 (21.4) 0.17 0.69

NSAIDs, non-steroidal anti-inflammatory drugs.

a
Within 3 months before the index date.
b
Other hypertension drugs: alpha-blockers, hydralazine, and clonidine.
c
Benzodiazepine used as anxiolytics, hypnotics, and sedatives.
RAS blockade in heart failure patients on HD
Figure 2 Kaplan– Meier estimates of (A) all-cause mortality and
(B) cardiovascular mortality up to 3 years.
hospital ownership, urbanization level 1 (high) to 7 (low) of residence,
Charlson co-morbidity index, medications taking at enrolment, and all
co-morbidities and tests/procedures listed in Table 1.
Outcome measurements
The main outcome variable of interest in this study was all-cause mortal-
ity, while the secondary outcome was cardiovascular mortality within 36
months after the index date. We defined the date of the first RASB pre-
scription as the index date in the treatment group and the date of the first
HF drug prescription as the index date in the control group after the HF
diagnosis. The follow-up period started on the index date and continued
for 36 months or until the date of death, whichever came first. A death
event was identified if (i) patients were coded with ‘death’ as the discharge
status in the inpatient claims data files; (ii) dates of death were claimed by
catastrophic illness registration files from the NHIRD; or (iii) patients
were unenrolled from the NHI programme and had not enrolled in the
NHI beneficiary registry files. Because the NHI is a compulsory plan, it
is rare that a patient, especially one with a disease, would unenrol from
the NHI unless death was to occur. Lien et al. 22 compared the end date
of coverage with the actual date of death for a cohort of stroke patients
and showed that the differences were longer than a month in ,2% of the
patients. Cardiovascular death was defined as diseases of the circulatory
system (myocardial infarction, unstable angina pectoris, coronary athero-
sclerosis, HF, severe arrhythmia, and sudden death) according to the diag-
noses made at the latest hospitalization/emergency department visit
within 2 months before death.
Page 5 of 9
Statistical analysis
We divided the patients of the treatment and control groups into five
strata using quintiles of the estimated propensity scores. To evaluate
the propensity score, we performed logistic regression using the treat-
ment group as the outcome for each of the variables listed in Table 1
with and without adjustment for propensity score strata. We then
charted the survival curves using the Kaplan– Meier method and exam-
ined the treatment effect using the log-rank test. Finally, we performed
several univariate and multivariate proportional hazard regression
models with and without adjustment for demographic and clinically rele-
vant variables, propensity score strata variables, and cumulated days of
taking an RASB to assess the independent treatment effects on the prob-
ability of death. Differences between groups were considered significant
if two-sided P-values were ,0.05. All analyses were performed using SAS
9.1 software (SAS Institute Inc., Cary, NC, USA).
Results
Participant characteristics
In all, 4771 patients were enrolled in the study (Figure 1). Of these,
3024 patients took an RASB for HF and 1746 patients did not take
any RASB for HF. The baseline patient characteristics before and
after the propensity score analyses are shown in Table 1. Those
treated with an RASB were younger (66.3 years vs. 68.7 years, P ,
0.001) and more likely to be male (49.6% vs. 45.6%, P , 0.01).
There were more patients with ≥ 3 hospitalizations in the treatment
group (2.66 vs. 2.08, P , 0.001). The duration of dialysis at enrolment
was also longer in the treatment group. More patients in the treat-
ment group had a Charlson co-morbidity index of ≥ 6. However,
after the propensity score adjustment, there were no significant
differences in baseline demographic or clinical characteristics
between patients in the two groups, except for the differences in the
duration of HD at enrolment. Hypertension was the most frequent
co-morbidity in both groups, followed by diabetes mellitus and is-
chaemic heart disease. More than 60% of patients in each group
had undergone echocardiography.
At the time of enrolment, compared with the patients in the
control group, patients in the treatment group more commonly
used beta-blockers, calcium channel blockers, alpha-blockers, hydra-
lazine, clonidine, and antianxiety drugs (Table 2). However, there was
no significant difference between groups after the propensity score
adjustment.
Primary outcome
The primary outcome was all-cause mortality during a follow-up of
up to 3 years. ACE inhibitors and ARBs were used in 69.7% and
65.9% of the patients in the treatment group, respectively. In the
treatment group, 2379 patients (78.7%) received monotherapy and
645 patients (21.3%) received dual therapy. The mean follow-up
periods were 20.9 + 13.3 and 18.4 + 13.7 months in the treatment
and control groups, respectively (P , 0.001). There were 1148
deaths (38.0%) in the treatment group during 5272 person-years of
follow-up, a lesser proportion than the 734 deaths (42.0%) in the
control group during 2683 person-years of follow-up. The case fatal-
ity rate (per 1000 person-years) in the treatment group was smaller
than that of the control group (1.53 vs. 1.92, P , 0.001). The mean
age at death at the incident year in the treatment group was less
Page 6 of 9 C.-H. Tang et al.

than that in the control group (69.5 + 11.0 vs. 72.2 + 10.9, P ,
0.001). All-cause mortality rates at 12, 24, and 36 months for the
treatment group were 22.9, 35.1, and 45.4%, respectively, while the
mortality rates for the control group were 30.7, 42.3, and 49.1%, re-
spectively. Log-rank tests of unadjusted survival showed a significant-
ly lower mortality rate in the treatment group than in the control
group (P , 0.001, Figure 2A). Men had higher mortality rates than
women in both groups. Both men and women had lower mortality
rates in the treatment group than in the control group after 3 years
of follow-up (men, P , 0.001; women, P ¼ 0.04).
Secondary outcome
The secondary outcome was cardiovascular mortality during a
follow-up of up to 3 years. There were 420 cardiovascular deaths
(13.9%) in the treatment group, a lesser proportion than the 302 car-
diovascular deaths (17.3%) in the control group. Cardiovascular mor-
tality rates at 12, 24, and 36 months for the treatment group were
10.0, 14.5, and 18.8%, respectively. The respective mortality rates
of the control group were 14.8, 18.9, and 22.5%. Log-rank tests of un-
adjusted survival showed a significantly lower cardiovascular mortal-
ity rate in the treatment group than in the control group (P , 0.001,
Figure 2B).
Table 3 shows the hazard ratios (HRs) of all-cause and cardiovas-
cular mortality before and after adjustment for demographic and clin-
ically relevant variables. The treatment group had a 15% lower risk of
all-cause death than the control group in the final model [HR 0.85;
95% confidence interval (CI) 0.77–0.94; P , 0.001]. Including the
propensity score strata in the regression did not change the results
(HR 0.84; 95% CI 0.76–0.92; P ,0.001). After time-dependent cov-
ariate adjustment, the magnitude of the treatment effect on all-cause
mortality increased (HR 0.8; 95% CI 0.72–0.89; P , 0.001). The
treatment group also had a 23% lower risk of cardiovascular death
than the control group in the final model (HR 0.77; 95% CI 0.66–
0.90; P ,0.01), as shown in Table 3. The treatment effect was
similar after time-dependent covariate adjustment (HR 0.76; 95%
CI 0.64–0.90; P , 0.01). Interactions between RASB therapy and
clinical parameters were estimated. These estimations are shown
in Figure 3. Subgroup analysis showed that both monotherapy and
dual therapy had significant treatment effects on all-cause mortality
(P ¼ 0.02 and P , 0.001, respectively) and dual therapy had a
better effect than monotherapy. Significant interactions were also
observed with the treatment with calcium channel blockers and
digoxin (P ¼ 0.02 and P , 0.001, respectively).
Discussion
The question addressed by the present study was whether RASB
therapy prolonged the survival of HF patients on long-term HD. As
shown in Table 3, the main finding of this population-based study is
that RASB therapy in HF patients on long-term HD significantly
decreased all-cause mortality risk compared with that of the
control group (P , 0.001). To diminish bias, we used propensity
scores and time-dependent covariates to adjust for the selection
bias. The difference was still significant even after adjustment for clin-
ical variables.

Table 3 Cox proportional hazard regression for all-cause and cardiovascular mortality before and after propensity score
adjustment

Treatment group vs. control group Before propensity 95% CI P-value After propensity 95% CI P-value
score adjusted HR score adjusted HR
...............................................................................................................................................................................
Univariate model, all-cause 0.81 0.74– 0.89 ,0.001 0.87 0.79–0.96 ,0.01
Multivariate model, all-cause
Adjusted for age and sex 0.86 0.78– 0.94 ,0.01 0.87 0.79–0.95 ,0.01
Adjusted for hypertension 0.80 0.73– 0.88 ,0.001 0.87 0.79–0.95 ,0.01
Adjusted for diabetes 0.80 0.73– 0.87 ,0.001 0.87 0.79–0.95 ,0.01
Adjusted for ischaemic heart disease 0.80 0.73– 0.88 ,0.001 0.87 0.79–0.95 ,0.01
Adjusted for duration of dialysis at enrolment 0.81 0.74– 0.89 ,0.001 0.87 0.79–0.96 ,0.01
Adjusted for no. of hospitalizations 0.77 0.70– 0.84 ,0.001 0.87 0.79–0.96 ,0.01
Adjusted for Charlson co-morbid index 0.81 0.74– 0.89 ,0.001 0.87 0.79–0.95 ,0.01
Adjusted for medicationa 0.80 0.73– 0.88 ,0.001 0.86 0.78–0.94 ,0.01
Multivariate final model, all-cause
Final 0.83 0.75– 0.91 ,0.001 0.85 0.77–0.94 ,0.001
With propensity scores strata 0.84 0.76–0.92 ,0.001
With time-dependent covariateb 0.78 0.70– 0.86 ,0.001 0.80 0.72–0.89 ,0.001
Multivariate final model, cardiovascular
Final 0.76 0.65– 0.88 ,0.001 0.77 0.66–0.90 ,0.01
With propensity scores strata 0.76 0.66–0.89 ,0.001
With time-dependent covariateb 0.74 0.63– 0.88 ,0.001 0.76 0.64–0.90 ,0.01

a
Medications including: beta-blockers, calcium channel blockers, alpha-blockers, hydralazine, clonidine, nitrates, and antianxiety drugs.
b
Time-dependent covariate: duration of ACE inhibitor/ARB therapy.
RAS blockade in heart failure patients on HD
Figure 3 Hazard ratios for all-cause mortality from the final multivariate model, including the duration of ACE inhibitor/ARB therapy and
interaction with selected subgroups. aBenzodiazepine used as anxiolytics, hypnotics, and sedatives. C.I., confidence interval.
Several cardioprotective medications, including ACE inhibitors
and ARBs, have been shown to prolong survival in the general popu-
lation of patients with HF.5,6 However, the examination of HF in
patients on long-term HD requires more large population studies.
In this study, we evaluated long-term outcomes among surviving
HD patients with HF. We defined patients in the treatment group
as new users of RASBs by including those who received an RASB
after the HF diagnosis. According to current guidelines, patients
with HF are suggested first to take ACE inhibitors. ARBs remain an
alternative in patients who are intolerant of ACE inhibitors. In most
clinical trials, beta-blockers were further added when the HF was
stable after ACE inhibitors were taken.5,6 However, hyperkalaemia
and hypotension are the two most concerning adverse reactions of
these drugs.23,24 Roy et al. 23 reported that only 36% of HD patients
with LV systolic dysfunction are given an ACE inhibitor due to its
adverse reaction. In this real-world retrospective study, 3024
patients (63.4%) of the total 4771 enrolled patients received an
RASB after the HF diagnosis (Figure 1). ACE inhibitors were pre-
scribed more than ARBs in the 4771 enrolled patients (44.2% vs.
41.8%). Not all patients in either group received beta-blockers
despite guideline suggestions.
With respect to the cardiovascular benefit of concurrent ACE in-
hibitor and ARB therapy in the general population without chronic
Page 7 of 9
renal failure, two large-scale, double-blind, placebo-randomized
trials have provided conflicting results.25,26 McMurray et al. 25
reported that candesartan was associated with a reduced risk of car-
diovascular mortality or hospitalization in patients already taking an
ACE inhibitor. In the valsartan HF trial,26 valsartan worsened morbid-
ity and mortality in patients receiving an ACE inhibitor or a beta-
blocker. Neither study showed a statistically significant effect on
mortality. Only one prospective randomized controlled trial involv-
ing HF patients on long-term HD showed that telmisartan added to
an ACE inhibitor reduced all-cause mortality, cardiovascular death,
and HF-related hospital stays compared with patients treated with
a placebo plus an ACE inhibitor.20 Cice et al. 20 suggested that ARBs
such as telmisartan might be more effective RAS antagonists in
patients with HD because ACE inhibitors are largely removed
during HD.27 The therapeutic effects of neurohormonal antagonism
resulted in reverse remodelling of the left ventricle. In this study,
there were 4771 patients, all of whom were followed for up to 3
years. Meanwhile, although there were greater numbers of hospitali-
zations and higher Charlson co-morbidity index scores in the treat-
ment group, our study still showed that RASBs could reduce
all-cause and cardiovascular mortality in HF patients on long-term
HD after adjustments for propensity scores and time-dependent
covariates (HR 0.8; 95% CI 0.72–0.89; Figure 2 and Table 3). In this
Page 8 of 9
study, 645 of the 3024 patients (21.3%) in the treatment group used
dual therapy with an ACE inhibitor and ARB concurrently. After sub-
group analysis, dual therapy had a better treatment effect than mono-
therapy (Figure 3). In 2005, K/DOQI guidelines suggested the
treatment of HF in patients on long-term HD with conventional ther-
apies according to expert opinion.28 Our study provided more evi-
dence to support the use of RASBs including monotherapy and
dual therapy in this specific population.
In addition to conventional medications for HF, volume control
with ultrafiltration rather than diuretics in patients on HD was the
preferred choice for treating HF symptoms.28,29 In this study, only
23% of the total 4771 enrolled patients were on loop diuretics. Ultra-
filtration during dialysis may provide an effective means of optimizing
fluid volume and improving blood pressure control. Hyperkalaemia,
intradialytic hypotension, polypharmacy, and non-compliance were
also barriers to these patients receiving adequate HF treatment
even with their high prevalence and mortality rates. To prevent
these biases, we excluded 787 patients who were not taking any con-
ventional drugs for HF within 3 years after the HF diagnosis.
This study has several strengths worth mentioning. First, to the
best of our knowledge, this is the first population-based study using
real-world data to confirm the benefit of RASB therapy against HF
in patients on long-term HD. To eliminate bias, we included only
new users of RASBs. Furthermore, we used cumulative days taking
RASBs as the time-dependent covariate to adjust for bias in duration
of RASB therapy.
This study has several limitations that must be considered. First, we
could not identify HF severity by the assigned ICD-9-CM codes, al-
though earlier studies proved that the identification of HF was
valid.30 Secondly, we did not take into account potential confounding
factors such as history of high blood pressure, laboratory data, dialysis
doses, nutritional condition, alcohol consumption, smoking status, or
inflammatory status. Thirdly, patients on long-term HD have various
pre-existing diseases. However, we cannot exclude the possibility of
residual confounding factors even after rigorous adjustment for the
important variables.
In conclusion, our findings expand on the prior knowledge to
provide clinicians with more evidence about the effects of RASBs
in HF patients on long-term HD. If patients on long-term HD have
HF, the use of an RASB should be recommended to prolong their
survival.
Funding
The Wan Fang Hospital, Taipei Medical University, Taipei, Taiwan (grant
no. 99wf-eva-09). The funding source had no role in the design or
conduct of the study; in the collection, analysis, and interpretation of
the data; or in the preparation, review, or approval of the manuscript.
Conflict of interest: none.
References
1. Yang WC, Hwang SJ. Incidence, prevalence and mortality trends of dialysis end-stage
renal disease in Taiwan from 1990 to 2001: the impact of national health insurance.
Nephrol Dial Transplant 2008;23:3977 –3982.
2. Cheung AK, Sarnak MJ, Yan G, Berkoben M, Heyka R, Kaufman A, Lewis J, Rocco M,
Toto R, Windus D, Ornt D, Levey AS. Cardiac diseases in maintenance hemodialysis
patients: results of the HEMO Study. Kidney Int 2004;65:2380 – 2389.
3. Gupta R, Birnbaum Y, Uretsky BF. The renal patient with coronary artery disease:
current concepts and dilemmas. J Am Coll Cardiol 2004;44:1343 –1353.
C.-H. Tang et al.
4. Stack AG, Bloembergen WE. A cross-sectional study of the prevalence and clinical
correlates of congestive heart failure among incident US dialysis patients. Am J Kidney
Dis 2001;38:992 –1000.
5. McMurray JJ, Adamopoulos S, Anker SD, Auricchio A, Bohm M, Dickstein K, Falk V,
Filippatos G, Fonseca C, Gomez-Sanchez MA, Jaarsma T, Kober L, Lip GY,
Maggioni AP, Parkhomenko A, Pieske BM, Popescu BA, Ronnevik PK, Rutten FH,
Schwitter J, Seferovic P, Stepinska J, Trindade PT, Voors AA, Zannad F, Zeiher A,
Bax JJ, Baumgartner H, Ceconi C, Dean V, Deaton C, Fagard R, Funck-Brentano C,
Hasdai D, Hoes A, Kirchhof P, Knuuti J, Kolh P, McDonagh T, Moulin C, Reiner Z,
Sechtem U, Sirnes PA, Tendera M, Torbicki A, Vahanian A, Windecker S,
Bonet LA, Avraamides P, Ben Lamin HA, Brignole M, Coca A, Cowburn P,
Dargie H, Elliott P, Flachskampf FA, Guida GF, Hardman S, Iung B, Merkely B,
Mueller C, Nanas JN, Nielsen OW, Orn S, Parissis JT, Ponikowski P. ESC guidelines
for the diagnosis and treatment of acute and chronic heart failure 2012: the Task
Force for the Diagnosis and Treatment of Acute and Chronic Heart Failure 2012
of the European Society of Cardiology. Developed in collaboration with the
Heart Failure Association (HFA) of the ESC. Eur J Heart Fail 2012;14:803 –869.
6. Hunt SA, Abraham WT, Chin MH, Feldman AM, Francis GS, Ganiats TG, Jessup M,
Konstam MA, Mancini DM, Michl K, Oates JA, Rahko PS, Silver MA, Stevenson LW,
Yancy CW. 2009 Focused update incorporated into the ACC/AHA 2005 Guidelines
for the Diagnosis and Management of Heart Failure in Adults A Report of the Ameri-
can College of Cardiology Foundation/American Heart Association Task Force on
Practice Guidelines Developed in Collaboration With the International Society
for Heart and Lung Transplantation. J Am Coll Cardiol 2009;53:e1 –e90.
7. Granger CB, McMurray JJ, Yusuf S, Held P, Michelson EL, Olofsson B, Ostergren J,
Pfeffer MA, Swedberg K. Effects of candesartan in patients with chronic heart failure
and reduced left-ventricular systolic function intolerant to angiotensin-converting-
enzyme inhibitors: the CHARM-Alternative trial. Lancet 2003;362:772–776.
8. Yusu S, Pfeffer MA, Swedberg K, Granger CB, Held P, McMurray JJ, Michelson EL,
Olofsson B, Ostergren J. Effects of candesartan in patients with chronic heart
failure and preserved left-ventricular ejection fraction: the CHARM-Preserved
Trial. Lancet 2003;362:777 –781.
9. Pfeffer MA, Swedberg K, Granger CB, Held P, McMurray JJ, Michelson EL, Olofsson B,
Ostergren J, Yusuf S, Pocock S. Effects of candesartan on mortality and morbidity in
patients with chronic heart failure: the CHARM-Overall programme. Lancet 2003;
362:759 – 766.
10. Matsumoto N, Ishimitsu T, Okamura A, Seta H, Takahashi M, Matsuoka H. Effects of
imidapril on left ventricular mass in chronic hemodialysis patients. Hypertens Res
2006;29:253 –260.
11. Efrati S, Zaidenstein R, Dishy V, Beberashvili I, Sharist M, Averbukh Z, Golik A,
Weissgarten J. ACE inhibitors and survival of hemodialysis patients. Am J Kidney Dis
2002;40:1023 –1029.
12. Takahashi A, Takase H, Toriyama T, Sugiura T, Kurita Y, Ueda R, Dohi Y. Candesar-
tan, an angiotensin II type-1 receptor blocker, reduces cardiovascular events in
patients on chronic haemodialysis—a randomized study. Nephrol Dial Transplant
2006;21:2507 –2512.
13. Suzuki H, Kanno Y, Sugahara S, Ikeda N, Shoda J, Takenaka T, Inoue T, Araki R. Effect
of angiotensin receptor blockers on cardiovascular events in patients undergoing
hemodialysis: an open-label randomized controlled trial. Am J Kidney Dis 2008;52:
501 –506.
14. Tai DJ, Lim TW, James MT, Manns BJ, Tonelli M, Hemmelgarn BR. Cardiovascular
effects of angiotensin converting enzyme inhibition or angiotensin receptor block-
ade in hemodialysis: a meta-analysis. Clin J Am Soc Nephrol 2010;5:623 –630.
15. Trespalacios FC, Taylor AJ, Agodoa LY, Bakris GL, Abbott KC. Heart failure as a
cause for hospitalization in chronic dialysis patients. Am J Kidney Dis 2003;41:
1267 –1277.
16. Zannad F, Kessler M, Lehert P, Grunfeld JP, Thuilliez C, Leizorovicz A, Lechat P. Pre-
vention of cardiovascular events in end-stage renal disease: results of a randomized
trial of fosinopril and implications for future studies. Kidney Int 2006;70:1318 – 1324.
17. Chang TI, Shilane D, Brunelli SM, Cheung AK, Chertow GM, Winkelmayer WC.
Angiotensin-converting enzyme inhibitors and cardiovascular outcomes in patients
on maintenance hemodialysis. Am Heart J 2011;162:324 –330.
18. Bajaj RR, Wald R, Hackam DG, Gomes T, Perl J, Juurlink DN, Manno M, Garg AX,
Kitchlu A, Mamdani MM, Yan AT. Use of angiotensin-converting enzyme inhibitors
or angiotensin receptor blockers and cardiovascular outcomes in chronic dialysis
patients: a population-based cohort study. Arch Intern Med 2012;172:591 –593.
19. Chan KE, Ikizler TA, Gamboa JL, Yu C, Hakim RM, Brown NJ. Combined angiotensin-
converting enzyme inhibition and receptor blockade associate with increased risk of
cardiovascular death in hemodialysis patients. Kidney Int 2011;80:978 –985.
20. Cice G, Di Benedetto A, D’Isa S, D’Andrea A, Marcelli D, Gatti E, Calabro R. Effects
of telmisartan added to angiotensin-converting enzyme inhibitors on mortality
and morbidity in hemodialysis patients with chronic heart failure a double-blind,
placebo-controlled trial. J Am Coll Cardiol 2010;56:1701 –1708.
21. Rosenbaum PR, Rubin DB. The central role of the propensity score in observational
studies for causal effects. Biometrika 1983;70:41 –55.
RAS blockade in heart failure patients on HD
22. Lien HM, Chou SY, Liu JT. Hospital ownership and performance: evidence
from stroke and cardiac treatment in Taiwan. J Health Econ 2008;27:
1208– 1223.
23. Roy P, Bouchard J, Amyot R, Madore F. Prescription patterns of pharmacological
agents for left ventricular systolic dysfunction among hemodialysis patients. Am J
Kidney Dis 2006;48:645 –651.
24. Knoll GA, Sahgal A, Nair RC, Graham J, van Walraven C, Burns KD. Renin–angioten-
sin system blockade and the risk of hyperkalemia in chronic hemodialysis patients.
Am J Med 2002;112:110 –114.
25. McMurray JJ, Ostergren J, Swedberg K, Granger CB, Held P, Michelson EL,
Olofsson B, Yusuf S, Pfeffer MA. Effects of candesartan in patients with chronic
heart failure and reduced left-ventricular systolic function taking angiotensin-
converting-enzyme inhibitors: the CHARM-Added trial. Lancet 2003;362:
767–771.
Page 9 of 9
26. Cohn JN, Tognoni G. A randomized trial of the angiotensin-receptor blocker valsar-
tan in chronic heart failure. N Engl J Med 2001;345:1667 –1675.
27. Roithinger FX, Punzengruber C, Wallner M, Ulbrich W, Pachinger O, Kramar R,
Prischl FC. The influence of ACE-inhibition on myocardial mass and diastolic func-
tion in chronic hemodialysis patients with adequate control of blood pressure. Clin
Nephrol 1994;42:309 –314.
28. K/DOQI Workgroup. K/DOQI clinical practice guidelines for cardiovascular disease
in dialysis patients. Am J Kidney Dis 2005;45:S1–153.
29. Costanzo MR, Guglin ME, Saltzberg MT, Jessup ML, Bart BA, Teerlink JR, Jaski BE,
Fang JC, Feller ED, Haas GJ, Anderson AS, Schollmeyer MP, Sobotka PA. Ultrafiltra-
tion versus intravenous diuretics for patients hospitalized for acute decompensated
heart failure. J Am Coll Cardiol 2007;49:675–683.
30. Svanstrom H, Pasternak B, Hviid A. Association of treatment with losartan vs cande-
sartan and mortality among patients with heart failure. JAMA 2012;307:1506 –1512.