Download Sudden Cardiac Death in End-Stage Renal Disease

Sudden Cardiac Death in End-Stage Renal Disease Patients
A 5-Year Prospective Analysis
Angela Yee-Moon Wang, Christopher Wai-Kei Lam, Iris Hiu-Shuen Chan, Mei Wang,
Siu-Fai Lui, John E. Sanderson
Downloaded from http://hyper.ahajournals.org/ by guest on May 6, 2017
Abstract—End-stage renal disease patients experience a high incidence of sudden cardiac death. We performed a 5-year
prospective study in 230 end-stage renal disease patients, aiming to determine the role of echocardiography and the
additional value of serum biomarkers in predicting sudden cardiac death. During follow-up, 24% of all deaths were
attributed to sudden cardiac death. In the multivariable Cox regression analysis considering clinical, biochemical,
dialysis, and echocardiographic parameters, left ventricular systolic dysfunction emerged as the most significant
predictor of sudden cardiac death, followed by a high systolic and a low diastolic blood pressure. An ejection fraction
cutoff ⱕ48.0% is associated with a specificity of 78.6% and a sensitivity of 57.7% in predicting sudden cardiac death.
In biomarker-based multivariable Cox regression analysis, N-terminal probrain natriuretic peptide displays an
independent association with sudden cardiac death and is more significantly associated with sudden cardiac death than
cardiac troponin T. In the combined echocardiography and biomarker-based multivariable Cox regression model,
N-terminal probrain natriuretic peptide loses significance to left ventricular ejection fraction, whereas cardiac troponin
T retains a significant association with sudden cardiac death independent of echocardiographic parameters. In
conclusion, systolic dysfunction is the most significant predictor of sudden cardiac death followed by a high systolic and
a low diastolic blood pressure. Our data suggest additional value in measuring cardiac troponin T for sudden cardiac
death risk stratification. N-terminal probrain natriuretic peptide may be used in place of echocardiography to identify
patients at risk of sudden cardiac death but had no added value over echocardiography in predicting sudden
cardiac death. (Hypertension. 2010;56:210-216.)
Key Words: sudden cardiac death 䡲 cardiac troponin T 䡲 N-terminal probrain natriuretic peptide 䡲 end-stage renal
disease 䡲 peritoneal dialysis 䡲 echocardiography 䡲 systolic dysfunction
D
ialysis patients are at a heightened risk of developing
cardiovascular mortality. According to recent data from
the US Renal Data System, cardiac disease is the leading
cause of mortality, accounting for 43% of all-cause mortality
in patients receiving hemodialysis or peritoneal dialysis
(PD).1 The incidence of death attributable to cardiac arrest/
cause unknown or arrhythmia is high in end-stage renal
disease (ESRD) patients, accounting for ⬇25% of all-cause
mortality. These rates are similar in hemodialysis and PD
patients.1 However, there have been very few prospective
data on risk factors associated with sudden cardiac death
(SCD) in ESRD patients and virtually none in patients
receiving long-term PD.
Data from the general population cannot simply be extrapolated to the ESRD population, because ESRD patients are
subjected not only to traditional Framingham risk factors but
also, more importantly, kidney disease-related risk factors,
such as inflammation, increased calcium⫻phosphorus product, uremic toxins, anemia, and fluid overload. The risk of
cardiovascular mortality is ⱖ10- to 100-fold higher in the
ESRD population compared with the age-, sex-, and racematched general population.2 Other than being at risk for
accelerated atherosclerosis, ESRD patients show a very high
prevalence of vascular3,4 and valvular calcifications,5 and
these have been shown to be associated with increased
arterial stiffening6 and adverse outcomes.7–9 ESRD patients
also have very high prevalence of left ventricular (LV)
hypertrophy10,11 with associated fibrosis, and that may predispose to inadequate coronary reserve and cardiac ischemia,
leading to an increased risk of heart failure.12 Inflammation
and sympathetic overactivity may also act as potential triggers for arrhythmias in patients with diseased myocardium,
and they may result in SCD.13
Given this background, we first aimed to test the hypothesis that specific echocardiographic parameters, including LV
Received February 3, 2010; first decision February 9, 2010; revision accepted February 24, 2010.
From the Departments of Medicine and Therapeutics (A.Y.-M.W., M.W., S.-F.L., J.E.S.) and Chemical Pathology (C.W.-K.L., I.H.-S.C.), Chinese
University of Hong Kong, Prince of Wales Hospital, Shatin, New Territories, Hong Kong; current address (C.W.-K.L.): Macau Institute for Applied
Research in Medicine and Health, Macau University of Science and Technology, Taipa, Macau; current address (A.Y.-M.W., M.W.): Department of
Medicine, Queen Mary Hospital, University of Hong Kong, Hong Kong.
Correspondence to Angela Yee-Moon Wang, University Department of Medicine, University of Hong Kong, Queen Mary Hospital, 102 Pokfulam Rd,
Pokfulam, Hong Kong. E-mail [email protected]
© 2010 American Heart Association, Inc.
Hypertension is available at http://hyper.ahajournals.org
DOI: 10.1161/HYPERTENSIONAHA.110.151167
210
Wang et al
mass index and ejection fraction, may be useful in predicting
SCD in ESRD patients receiving PD treatment and, if so, the
best cutoff of these parameters in predicting SCD. Second,
given that echocardiography is not always readily available in
most dialysis centers, we aimed to test whether serum
biomarkers may be used in place of echocardiography to
identify ESRD patients at risk of SCD. Third, we aimed to
examine the additional role, if any, of a panel of serum
biomarkers, including C-reactive protein (CRP), interleukin 6
(IL-6), fetuin-A, cardiac troponin T (cTnT), N-terminal
probrain natriuretic peptide (NT-pro-BNP), and myeloperoxidase (MPO), to echocardiography in predicting SCD in these
patients.
Methods
Study Design
Downloaded from http://hyper.ahajournals.org/ by guest on May 6, 2017
This is a 5-year prospective observational cohort study performed
at a single regional dialysis center in Hong Kong between 1999
and 2005. The study protocol was approved by the institutional
research ethics committee of the Chinese University of Hong
Kong. All of the patients provided informed consent before
enrollment into the study.
Study Subjects
ESRD patients who have been started on long-term continuous PD
therapy for ⱖ3 months and did not fulfil any of the exclusion criteria
were considered eligible for study inclusion. Exclusion criteria
included patients with underlying malignancy, chronic obstructive
airways disease, chronic rheumatic heart disease, or congenital heart
disease, as well as patients who refused to provide study consent or
patients with incomplete data. On the basis of the inclusion and
exclusion criteria, 117 men and 113 women were recruited, and they
represented 85% of the entire PD population in the center. All of the
patients were dialyzed using conventional lactate-buffered, glucosebased PD solutions.
Assessments
Echocardiography
At study baseline, 2D echocardiography was carried out in all of the
patients lying in the left decubitus position using a GE-VingMed
System 5 echocardiographic machine (GE-VingMed Sound AB)
with a 3.3-mHz multiphase array probe by a single experienced
cardiologist blinded to all of the clinical details of patients. All of the
echocardiographic data were recorded according to the guidelines of
the American Society of Echocardiography.14,15
Biochemical Measurements
A 20-mL fasting venous blood sample was collected at study baseline
for the measurement of cTnT, NT-pro-BNP, high-sensitivity CRP
(hs-CRP), IL-6, fetuin-A, MPO, albumin, fasting lipids, renal function,
bone profile, intact parathyroid hormone, and hemoglobin. Detailed
methods for the different biochemical assays are provided in the online
Data Supplement (please see http://hyper.ahajournals.org).
Measurements of Indices of Dialysis Adequacy
At study entry, 24-hour urine sample and dialysate were also
collected for measurement of total weekly urea and creatinine
clearances using standard methods.16 Residual glomerular filtration
rate was estimated as the average of 24-hour urine urea and
creatinine clearances.17
Data Collection
At study entry, a thorough medical history was taken from all of the
study patients. Details on data collection are provided in the online
Data Supplement.
Sudden Cardiac Death in Peritoneal Dialysis
211
Follow-Up and Outcome Measure
The cohort was assembled purposely to study the outcome of SCD,
which was defined “a priori” as sudden unexpected natural death
within 1 hour from the symptom onset and without any immediate
previous condition that would appear fatal.18,19 Given the sudden and
unexpected nature of death, there is usually not enough time to allow
extensive investigations to be performed before death. If investigations were performed and the cause of death identified, the deaths
were not considered as SCD. For instance, in patients who developed
acute myocardial infarction, aortic dissection, acute pulmonary
embolism, or some other acute infections and then developed SCD
shortly or immediately after the acute event, these deaths were not
considered as SCDs because there were attributable causes. Death
preceded by an arrhythmic event was classified as attributed to
malignant arrhythmia and not as SCD. Deaths occurring after
termination of dialysis were not regarded as unexpected and, thus,
not as SCD. All of the deaths were accurately recorded with the exact
cause of death provided by the attending physician. For out-ofhospital deaths, family members were interviewed by telephone to
ascertain the circumstances surrounding the death. All of the patients
were followed up prospectively for 5 years from the day of baseline
assessment at study entry, until death, or until they underwent kidney
transplant. No patient was lost to follow-up.
Statistical Analysis
Baseline characteristics were presented as mean⫾SD or median
(interquartile range) for continuous data where appropriate or number (%) for categorical data. Kaplan-Meier survival analysis was
used to estimate the cumulative SCD event-free survival probability
in relation to tertiles of cTnT and NT-pro-BNP. Patients who
underwent kidney transplantation during follow-up were censored at
the time of transplantation. Patients who died from other cardiac
causes or noncardiovascular causes were also censored at the time of
death. Relative risks and 95% CI from the Cox proportional-hazards
regression models were used to evaluate the associations of different
risk factors to SCD. To test the 3 hypotheses, separate multivariable
Cox regression analyses for SCD were performed using a backward
elimination strategy considering variables with P⬍0.2 on univariate
Cox regression analysis. Details on multivariable Cox regression
analysis are provided in the online Data Supplement. Although the
study cohort is relatively large, the number of SCDs remained fairly
low. We adjusted each regression coefficient in the final multivariable Cox regression model with a shrinkage factor to correct for
overfitting.20 Details on how the shrinkage factor was derived are
shown in the online Data Supplement. A P value of ⬍0.05 was
considered to be statistically significant. Statistical analysis was
performed using SPSS version 16.0 (SPSS, Inc).
Results
The baseline characteristics of the study population are
shown in Table 1. Background causes of ESRD included 74
cases of chronic glomerulonephritis (32.2%), 55 cases of
diabetic nephropathy (23.9%), 31 cases of hypertensive
nephrosclerosis (13.5%), 12 cases of polycystic kidney disease (5.2%), 13 cases of obstructive uropathy (5.7%), 6 cases
of tubulointerstitial nephritis (2.6%), and no cause identified
in 39 cases (17.0%). During the 5-year follow-up, a total of
115 deaths occurred, of which 28 deaths were attributed to
SCD, and 87 deaths were attributed to non-SCD causes.
Table 2 presents the univariate Cox regression analysis of
factors predicting SCD. Of the serum biomarkers, only cTnT
and NT-pro-BNP are significantly associated with SCD, and
fetuin-A shows marginal association (P⬍0.2) on univariate
analysis. The associations of hs-CRP, IL-6, and MPO with
SCD are all highly insignificant on univariate analysis
(P⬎0.2), and, thus, they are not further considered in multivariable Cox regression analysis. We stratified patients by
212
Hypertension
Table 1.
August 2010
Baseline Characteristics
Characteristic
Data
Male sex, n (%)
117 (50.9)
Age, y
56⫾11
Table 2. Univariate Cox Regression Analysis of Factors
Predicting SCD in ESRD Patients
Parameters
Hazard Ratios (95% CI)
P
Clinical and biochemical parameters
86 (37.4)
Age, y
1.06 (1.04 to 1.08)
⬍0.001
69 (30)
Male sex, %
1.28 (0.61 to 2.69)
0.5
52 (22.6)
Positive smoking history, %
1.90 (0.90 to 3.98)
0.09
Background AVD, %
3.22 (1.50 to 6.92)
0.003
23.1⫾3.4
Background diabetes mellitus, %
1.42 (0.66 to 3.08)
0.4
Systolic blood pressure, mm Hg
147⫾17
Duration of dialysis, mo
1.00 (0.99 to 1.01)
0.7
Diastolic blood pressure, mm Hg
83⫾10
Systolic blood pressure, mm Hg
1.02 (1.00 to 1.04)
0.1
0.63 (0, 1.94)
Diastolic blood pressure, mm Hg
0.96 (0.92 to 1.00)
0.038
Hemoglobin, g/dL
1.08 (0.87 to 1.35)
0.5
0.95 (0.88 to 1.02)
0.1
1.003 (0.996 to 1.010)
0.4
Smoking status, n (%)
Background diabetes mellitus, n (%)
Background atherosclerotic vascular disease, n (%)
Duration of dialysis, mo*
2
Body mass index, kg/m
Residual glomerular filtration rate, mL/min per
1.73 m2*
26.0 (14.5, 50.5)
Total weekly urea clearance
1.81⫾0.43
Serum albumin, g/L
Total weekly creatinine clearance, L/wk per 1.73 m2
56.5⫾21.3
Intact parathyroid hormone, pmol/L
2
2
Downloaded from http://hyper.ahajournals.org/ by guest on May 6, 2017
Hemoglobin, g/dL
9.17⫾1.68
Calcium⫻phosphorus, mmol /L
1.10 (0.86 to 1.40)
0.5
Serum albumin, g/L
28.6⫾5.1
LDL cholesterol, mmol/L
0.92 (0.62 to 1.37)
0.7
Calcium⫻phosphorus, mmol2/L2
4.30⫾1.33
Triglyceride, mmol/L
0.86 (0.62 to 1.21)
0.4
Erythropoietin use
0.50 (0.21 to 1.18)
0.1
0.031
Intact parathyroid hormone, pmol/L*
40.7 (17.6, 74.0)
LDL cholesterol, mmol/L
3.29⫾0.98
Dialysis parameters
Triglyceride, mmol/L
2.07⫾1.46
Total weekly Kt/V
0.35 (0.13 to 0.91)
Peritoneal dialysis Kt/V
0.60 (0.23 to 1.56)
0.3
Residual GFR, mL/min per 1.73 m2
0.83 (0.64 to 1.09)
0.2
hs-CRP, mg/L*
2.66 (0.92, 8.04)
IL-6, pg/mL*
9.7 (5.2, 17.8)
Serum MPO, ␮g/L*
31.7 (24.4, 42.5)
Serum fetuin-A, g/L
cTnT, ␮g/L*
NT-pro-BNP, pg/mL*
LV mass index, g/m
2.7
LV ejection fraction, %
Echocardiographic parameters
LV mass index, g/m2.7
1.008 (1.000 to 1.016)
0.06 (0.01, 0.15)
LV ejection fraction, %
0.93 (0.90 to 0.95)
⬍0.001
5698 (1944, 16711)
Cardiac valvular calcification, %
2.51 (1.13 to 5.55)
0.023
0.31⫾0.06
0.051
103.5⫾39.9
Serum biomarkers
52.5⫾8.3
hs-CRP, mg/L
1.01 (0.98 to 1.04)
0.6
IL-6, pg/mL
1.00 (0.98 to 1.02)
0.9
Continuous data are shown as mean⫾SD unless specified otherwise. LDL
indicates low-density lipoprotein.
*Data are shown as median (interquartile range).
tertiles of NT-pro-BNP and cTnT. Figure A and B show the
SCD event-free survival probability of patients stratified by
tertiles of NT-pro-BNP and cTnT, respectively. A significant
graded increase in the risk of developing SCD is observed
across the 3 tertiles of increasing NT-pro-BNP (P⬍0.001)
and cTnT (P⫽0.001). We compared the strength of association between cTnT and NT-pro-BNP with SCD in Table S1
(please see the online Data Supplement). NT-pro-BNP shows
a stronger association with SCD than cTnT. NT-pro-BNP
retains significant association with SCD when adjusting for
cTnT, whereas cTnT loses significance when adjusting for
NT-pro-BNP.
Table 3 presents 3 separate multivariable Cox regression
models for SCD. Model 1 considers clinical, biochemical,
and dialysis parameters, as well echocardiographic parameters with P⬍0.2 on a univariate analysis (echocardiographybased model). Model 2 includes the same clinical, biochemical, and dialysis parameters, as well as serum biomarkers
with P⬍0.2 on univariate analysis but not echocardiographic
parameters (biomarker-based model). Because hs-CRP, IL-6,
and MPO are all highly insignificant in the univariate
analysis, only cTnT, NT-pro-BNP, and fetuin-A are consid-
Fetuin-A, g/L
0.007 (0 to 3.07)
0.1
MPO, ␮g/L
0.99 (0.97 to 1.02)
0.5
cTnT, 0.1 ␮g/L
1.25 (1.13 to 1.38)
⬍0.001
NT-pro-BNP, 1000 pg/mL
1.05 (1.03–1.08)
⬍0.001
LDL indicates low density lipoprotein; Kt/V, urea clearance; GFR, glomerular
filtration rate; AVD, atherosclerotic vascular disease.
ered in the multivariable Cox regression analysis. Model 3
includes both echocardiographic parameters and serum biomarkers, in addition to clinical, biochemical, and dialysis
parameters, with P⬍0.2 on univariate analysis (combined
echocardiography and biomarker-based model). In the echocardiography-based Cox model, LV systolic dysfunction is
the most significant predictor of SCD, followed by a high
systolic blood pressure and a low diastolic blood pressure.
Using the receiver-operator characteristics curve analysis, the
best cutoff of LV ejection fraction in predicting SCD is
ⱕ48% and is associated with a specificity of 78.6% (95% CI:
72.2% to 84.1%) and a sensitivity of 57.7% (95% CI: 36.9%
to 76.6%). In the biomarker-based Cox model, NT-pro-BNP
is independently predictive of SCD and is more significant
than cTnT in predicting risk of SCD. In the combined
echocardiography and biomarker-based model, NT-pro-BNP
does not retain independent significance in the model,
Wang et al
A
lower tertile of NT-pro-BNP
1.0
Cumulative sudden cardiac
death event-free survival
middle tertile of NT-pro-BNP
0.8
upper tertile of NT-pro-BNP
0.6
0.4
0.2
P<0.0001
0.0
0
B
12
24
36
48
Follow-up Time (months)
60
lower tertile of cTnT
1.0
Cumulative sudden cardiac
death event-free survival
Downloaded from http://hyper.ahajournals.org/ by guest on May 6, 2017
middle tertile of cTnT
0.8
upper tertile of cTnT
0.6
0.4
0.2
0.0
P=0.0001
0
12
24
36
48
Follow-up Time (months)
60
Figure. A, Cumulative SCD event-free survival probability of
patients stratified by tertiles of NT-pro-BNP. Lower tertile
(n⫽77): NT-pro-BNP ⬍2665 pg/mL; middle tertile (n⫽77):
NT-pro-BNP ⱖ2666 to 11 893 pg/mL; upper tertile (n⫽76):
NT-pro-BNP ⱖ11 894 pg/mL. B, Cumulative SCD event-free
survival probability of patients stratified by tertiles of cTnT. Lower
tertile (n⫽81): cTnT ⱕ0.010 ␮g/L; middle tertile (n⫽70): cTnT
⬎0.010 to 0.099 ␮g/L; upper tertile (n⫽79): cTnT ⱖ0.100 ␮g/L.
whereas LV systolic dysfunction and cTnT retain independent, significant predictive values for SCD.
Discussion
In this prospective cohort study of Chinese ESRD patients
receiving chronic PD treatment, we observed a high incidence
of SCD over a 5-year longitudinal follow-up. Twenty-four
percent of all-cause mortality was attributed to SCD. This
incidence is almost identical to that reported in the US Renal
Data System database, in which 25% of all-cause death
among PD patients and 27% of all-cause death among
hemodialysis patients in the United States are attributed to
cardiac arrest/cause unknown.1 This observation is not surprising, given that PD patients are similarly subjected to
kidney disease or uremia-related risk factors as in hemodialysis patients.
Our results show that systolic dysfunction is the most
significant predictor of SCD, followed by a high systolic and
a low diastolic blood pressure. This is in keeping with data
from the general population showing that reduced ventricular
function was the strongest predictor for SCD.21,22 Our study
Sudden Cardiac Death in Peritoneal Dialysis
213
shows for the first time that an ejection fraction ⱕ48% is the
best cutoff threshold in predicting an increased risk of SCD in
ESRD patients with ⬇80% specificity. Poor systolic function
may predispose to heart failure and increase the risk of
electric instability and ventricular arrhythmia via neurohumoral/sympathetic activation.23,24
The association among background atherosclerotic vascular disease, valvular calcification, and SCD in our ESRD
patients is consistent with observations in the general population that obstructive coronary artery disease with resulting
myocardial ischemia was the most common triggering factor
for SCD.25 Previous studies show that valvular calcification is
a marker of atherosclerosis26 and predicts cardiovascular
death9 in ESRD patients. There is also evidence that valvular
calcification is associated with vascular calcification27 and
increased arterial stiffening and may, thus, increase LV
hypertrophy.28,29 The diseased myocardium secondary to
coronary ischemia, together with myocardial fibrosis and/or
hypertrophy, provide additional substrate for an increased
electric instability13 and may, thus, contribute to an increased
risk of SCD in uremic patients. In this study, systolic
hypertension and a low diastolic blood pressure, reflecting
increased arterial stiffening or reduced arterial distensibility,
both independently predict an increased risk of SCD. Previous studies also reported the importance of blood pressure
control in dialysis patients.30 –32 It is well recognized that
systolic hypertension and increased arterial stiffness contribute to LV hypertrophy in ESRD patients.29 Our results
showed that LV hypertrophy was a significant predictor of
SCD but lost its significance to systolic hypertension and low
diastolic blood pressure in multivariable Cox regression
models. Of importance, systolic hypertension and low diastolic blood pressure consistently retained independent significance in all 3 of the multivariable Cox regression models
for SCD. These observations clearly confirm the importance
of blood pressure control in ESRD patients and suggest that
systolic hypertension and arterial stiffening may act as
potential triggers in the presence of an abnormal myocardium
and predispose to arrhythmias and, consequently, SCD.
Previous studies demonstrated the prognostic importance
of cTnT in ESRD patients.33–35 cTnT is a highly sensitive
marker of myocardial damage. There is pathological evidence
that elevated cTnT is associated with subclinical myocardial
necrosis or microinfarct.36,37 In uremic cardiac hypertrophy,
myocardial capillary growth does not keep pace with cardiomyocyte hypertrophy, resulting in cardiomyocyte/capillary
mismatch, and reduces ischemic tolerance of the heart.38 This
may further increase the risk of subclinical ischemic insults to
the myocardium and amplify the leakage of cardiac troponins
across the plasma membrane of myocardial cells into the
circulation. In this study, cTnT retained a significant association with SCD independent of echocardiographic parameters. This suggests that measurement of cTnT to stratify SCD
risk in ESRD patients may give additional value. It is likely
that cTnT reflects subclinical myocardial necrosis or injury
that is not captured by standard echocardiographic parameters
LV mass and ejection fraction, although they are correlated
with cTnT.35 Our results are in keeping with a study in
hemodialysis patients showing that cTnT strongly correlated
214
Hypertension
Table 3.
August 2010
Multivariable Cox Regression Models for SCD*
Model and Variable
Unit
Increase
Shrinkage
Corrected HR
Shrinkage
Corrected 95% CI
P
0.0057
Echocardiography-based model
Systolic blood pressure
1 mm Hg
1.04
1.01 to 1.07
Diastolic blood pressure
1 mm Hg
0.92
0.87 to 0.97
0.0044
LV ejection fraction
1%
0.93
0.89 to 0.97
0.0012
Erythropoietin use
…
0.45
0.19 to 1.11
0.082
Diabetes mellitus
…
0.54
0.22 to 1.28
0.16
Valvular calcification
…
2.09
0.94 to 4.66
0.071
0.023
Biomarker-based model
Systolic blood pressure
1 mm Hg
1.04
1.01 to 1.07
Diastolic blood pressure
1 mm Hg
0.92
0.87 to 0.97
0.0023
…
0.38
0.16 to 0.91
0.03
1000 pg/mL
1.05
1.01 to 1.08
0.0044
0.1 ␮g/L
1.14
0.99 to 1.30
0.061
Systolic blood pressure
1 mm Hg
1.05
1.02 to 1.08
0.0031
Diastolic blood pressure
1 mm Hg
0.92
0.87 to 0.97
0.0033
Erythropoietin use
NT-pro-BNP
cTnT
Downloaded from http://hyper.ahajournals.org/ by guest on May 6, 2017
Combined echocardiography- and biomarker-based model
Erythropoietin use
…
0.44
0.18 to 1.06
0.066
Diabetes mellitus
…
0.48
0.20 to 1.15
0.10
1.14
1.01 to 1.31
0.031
cTnT
0.1 ␮g/L
Valvular calcification
…
1.90
0.83 to 4.32
0.13
LV ejection fraction
1%
0.94
0.89 to 0.98
0.004
HR indicates hazard ratio.
*All of the models considered the same clinical, biochemical, and dialysis parameters with P⬍0.2 on univariate analysis.
with all-cause mortality even after adjusting for LV systolic
function.39
On the other hand, NT-pro-BNP independently predicts
SCD in a biomarker-based model but not in a combined
echocardiography and biomarker-based model. Brain natriuretic peptide is a peptide hormone released primarily by the
ventricular myocytes in response to myocyte stretch, such as
increased cardiac filling pressure.40 Population studies have
suggested that plasma brain natriuretic peptide and NT-proBNP levels are useful screening tests for heart failure41 and
systolic dysfunction.42 NT-pro-BNP has been shown to
strongly correlate with LV systolic dysfunction and is more
strongly associated with mortality than cTnT, suggesting that
volume overload may play a role in predicting outcome.39
This is in keeping with our current observation that NT-proBNP had a stronger association with SCD than cTnT. There
is increasing evidence that brain natriuretic peptide and
NT-pro-BNP are powerful prognostic tools in the ESRD
population.43,44 In PD patients, NT-pro-BNP is associated
with LV hypertrophy and systolic dysfunction independent of
residual renal function.44 In a more recent study, elevated
NT-pro-BNP has been demonstrated to be a useful biomarker
in predicting severe LV hypertrophy and ruling out systolic
dysfunction in the PD patients.45 The finding that NT-proBNP predicts SCD in a biomarker-based model but loses
significance to LV ejection fraction in a combined echocardiography and biomarker-based model suggests that its association with SCD is likely mediated via its close relations
with systolic dysfunction. These findings on cTnT and
NT-pro-BNP have several important novel clinical implications in that NT-pro-BNP measurement may be used in place
of echocardiography to identify ESRD patients at risk of SCD
but has no added value over echocardiography in predicting
risk of SCD. If echocardiography is not available, then
NT-pro-BNP appears to be the most useful serum biomarker
in predicting SCD in ESRD patients compared with cTnT,
CRP, IL-6, fetuin-A, and MPO. On the other hand, our data
suggest additional value in measuring cTnT in ESRD patients, because it displays independent predictive value for
SCD beyond the standard echocardiographic parameters.
Contrary to the Choices for Healthy Outcomes in Caring
for ESRD Study showing how inflammation, such as CRP
and IL-6, increased the risk of SCD,46 we found no significant
association between a single time point of measured serum
inflammatory protein, including CRP, IL-6, MPO, or
fetuin-A, and SCD in our patients. It is not clear whether this
difference may be explained by the fact that our study
includes prevalent PD patients in contrast to the Choices for
Healthy Outcomes in Caring for ESRD Study, which included incident and predominantly hemodialysis patients.
There is some suggestion that hemodialysis may induce
CRP more than PD,47,48 but this will require confirmation
in a larger prospective controlled study. Unlike our study,
the Choices for Healthy Outcomes in Caring for ESRD
Study46 did not include any cTnT, NT-pro-BNP, or echocardiographic data.
The strengths of our study include that it is so far the
longest prospective follow-up (5 years) study that evaluated
Wang et al
Downloaded from http://hyper.ahajournals.org/ by guest on May 6, 2017
SCD in ESRD patients. The outcome of SCD was defined
a priori using strict standard criteria. All of the patients had
detailed comorbid data, echocardiographic examination together with a comprehensive panel of serum inflammatory
proteins, cardiac biomarkers including cTnT and NT-proBNP, as well as indices of dialysis adequacy at baseline
before entering a 5-year prospective follow-up period. There
was no patient loss to follow-up. The incidence of SCD in our
population is very similar to that reported in the US Renal
Data System, and, thus, the results should be representative of
the dialysis population at large. However, there are also
limitations. First, the study was performed in prevalent PD
patients and may introduce survival bias. Second, although
the study cohort was followed for a fairly long period, the
total number of SCDs remains quite low, because SCD
accounts for 24% of total mortality. We, therefore, adjusted
the analysis by a “shrinkage factor,” making the analysis
more conservative. Our observations will require confirmation in a larger study. Third, we evaluated the association of
a single time point measurement of different parameters with
SCD and did not take into account changes with time on
dialysis. However, this reproduces the typical situation of
daily clinical practice and is also the primary aim of our
study, that is, to determine how the different parameters
measured at a 1-off time point may be used to identify
patients at risk of SCD years later.
Sudden Cardiac Death in Peritoneal Dialysis
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Perspectives
There are only scant data on risk factors predicting SCD in
the ESRD population. In this 5-year prospective follow-up
study, we evaluated how echocardiography, serum biomarkers, and other clinical parameters assessed at a 1-off time
point may be used to identify ESRD patients at risk of SCD
some years later. Our results demonstrate the importance of
systolic dysfunction (with the best ejection fraction cutoff
ⱕ48%), a high systolic blood pressure, and a low diastolic
blood pressure in predicting SCD in ESRD patients. Our data
suggest additional value in measuring cTnT, because it gives
independent predictive value for SCD beyond echocardiographic parameters. NT-pro-BNP may be used in place of
echocardiography to predict SCD but has no additional role if
echocardiography is available. These observations require
confirmation in larger ESRD cohorts.
13.
14.
15.
16.
17.
Sources of Funding
18.
This study was supported by the Hong Kong Health Service
Research grant (project 6901023), of which A.Y.-M.W. is the
principal investigator.
19.
Disclosures
None.
References
1. US Renal Data System. USRDS 2006 Annual Data Report. Bethesda,
MD: National Institutes of Health, National Institute of Diabetes and
Digestive and Kidney Diseases; 2006.
2. Levey AS, Beto JA, Coronado BE, Eknoyan G, Foley RN, Kasiske BL,
Klag MJ, Mailloux LU, Manske CL, Meyer KB, Parfrey PS, Pfeffer MA,
Wenger NK, Wilson PW, Wright JT Jr. Controlling the epidemic of
cardiovascular disease in chronic renal disease: what do we know? What
do we need to learn? Where do we go from here? National Kidney
20.
21.
22.
23.
215
Foundation Task Force on Cardiovascular Disease. Am J Kidney Dis.
1998;32:853–906.
Braun J, Oldendorf M, Moshage W, Heidler R, Zeitler E, Luft FC.
Electron beam computed tomography in the evaluation of cardiac calcification in chronic dialysis patients. Am J Kidney Dis. 1996;27:394 – 401.
Raggi P, Boulay A, Chasan-Taber S, Amin N, Dillon M, Burke SK,
Chertow GM. Cardiac calcification in adult hemodialysis patients: a link
between end-stage renal disease and cardiovascular disease? J Am Coll
Cardiol. 2002;39:695–701.
Wang AY, Woo J, Wang M, Sea MM, Ip R, Li PK, Lui SF, Sanderson JE.
Association of inflammation and malnutrition with cardiac valve calcification in continuous ambulatory peritoneal dialysis patients. J Am Soc
Nephrol. 2001;12:1927–1936.
Haydar AA, Covic A, Colhoun H, Rubens M, Goldsmith DJ. Coronary
artery calcification and aortic pulse wave velocity in chronic kidney
disease patients. Kidney Int. 2004;65:1790 –1794.
Blacher J, Guerin AP, Pannier B, Marchais SJ, London GM. Arterial
calcifications, arterial stiffness, and cardiovascular risk in end-stage renal
disease. Hypertension. 2001;38:938 –942.
Block GA, Raggi P, Bellasi A, Kooienga L, Spiegel DM. Mortality effect
of coronary calcification and phosphate binder choice in incident hemodialysis patients. Kidney Int. 2007;71:438 – 441.
Wang AY, Wang M, Woo J, Lam CW, Li PK, Lui SF, Sanderson JE.
Cardiac valve calcification as an important predictor for all-cause mortality and cardiovascular mortality in long-term peritoneal dialysis
patients: a prospective study. J Am Soc Nephrol. 2003;14:159 –168.
Wang AY, Wang M, Woo J, Lam CW, Lui SF, Li PK, Sanderson JE.
Inflammation, residual kidney function, and cardiac hypertrophy are
interrelated and combine adversely to enhance mortality and cardiovascular death risk of peritoneal dialysis patients. J Am Soc Nephrol. 2004;
15:2186 –2194.
Foley RN, Parfrey PS, Harnett JD, Kent GM, Martin CJ, Murray DC,
Barre PE. Clinical and echocardiographic disease in patients starting
end-stage renal disease therapy. Kidney Int. 1995;47:186 –192.
Foley RN, Parfrey PS, Kent GM, Harnett JD, Murray DC, Barre PE.
Serial change in echocardiographic parameters and cardiac failure in
end-stage renal disease. J Am Soc Nephrol. 2000;11:912–916.
Herzog CA, Mangrum M, Passman R. Sudden cardiac death and dialysis
patients. Semin Dialysis. 2008;21:300 –307.
Schiller NB, Shah PM, Crawford M, DeMaria A, Devereux R,
Feigenbaum H, Gutgesell H, Reichek N, Sahn D, Schnittger I. Recommendations for quantitation of the left ventricle by two-dimensional
echocardiography: American Society of Echocardiography Committee on
Standards, Subcommittee on Quantitation of Two-Dimensional Echocardiograms. J Am Soc Echocardiogr. 1989;2:358 –367.
Schiller NB. Two-dimensional echocardiographic determination of left
ventricular volume, systolic function, and mass: summary and discussion
of the 1989 recommendations of the American Society of Echocardiography. Circulation. 1991;84(suppl 3):1280 –1287.
Nolph KD, Moore HL, Twardowski ZJ, Khanna R, Prowant B, Meyer M,
Ponferrada L. Cross-sectional assessment of weekly urea and creatinine
clearances in patients on continuous ambulatory peritoneal dialysis.
ASAIO J. 1992;38:M139 –M142.
Van Olden RW, Krediet RT, Struijk DG, Arisz L. Measurement of
residual renal function in patients treated with continuous ambulatory
peritoneal dialysis. J Am Soc Nephrol. 1996;7:745–748.
Engelstein ED, Zipes DP. Sudden cardiac death. In: Alexander RW,
Schlant RC, Fuster V, eds. The Heart, Arteries and Veins. New York,
NY: McGraw-Hill; 1998:1081–1112.
Myerburg RJ, Castellanos A. Cardiac arrest and sudden death. In:
Braunwald E, ed. Heart Disease: A Textbook of Cardiovascular
Medicine. Philadelphia, PA: WB Saunders; 1997:742–779.
Van Houwelingen JC, Le Cessie R. Predictive value of statistical models.
Stat Med. 1990;9:1303–1325.
Bigger JT, Fleiss JL, Kleiger R, Miller JP, Rolnitzky LM. The relationships among ventricular arrhythmias, left ventricular dysfunction, and
mortality in the 2 years after myocardial infarction. Circulation. 1984;
69:250 –258.
Mukharji J, Rude RE, Poole WK, Gustafson N, Thomas LJ Jr, Strauss
HW, Jaffe AS, Muller JE, Roberts R, Raabe DS Jr. Risk factors for
sudden death after acute myocardial infarction: two-years follow-up.
Am J Cardiol. 1984;54:31–36.
Packer M. Sudden unexpected death in patients with congestive heart
failure; a second frontier. Circulation. 1985;72:681– 685.
216
Hypertension
August 2010
Downloaded from http://hyper.ahajournals.org/ by guest on May 6, 2017
24. Bigger JT Jr. Why patients with congestive heart failure die: arrhythmias
and sudden cardiac death. Circulation. 1987;75:IV28 –IV35.
25. Spaulding CM, Joly LM, Rosenberg A, Monchi M, Weber SN, Dhainaut
JF, Carli P. Immediate coronary angiography in survivors of out-ofhospital cardiac arrest. N Engl J Med. 1997;336:1629 –1633.
26. Wang AY, Ho SS, Wang M, Liu EK, Ho S, Li PK, Lui SF. Cardiac
valvular calcification as a marker of atherosclerosis and arterial calcification in end-stage renal disease. Arch Intern Med. 2005;165:327–332.
27. Bellasi A, Ferramosca E, Muntner P, Ratti C, Wildman RP, Block GA,
Raggi P. Correlation of simple imaging tests and coronary artery calcium
measured by computed tomography in hemodialysis patients. Kidney Int.
2006;70:1623–1628.
28. London GM, Marchais SJ, Guerin AP, Metivier F, Adda H. Arterial
structure and function in end-stage renal disease. Nephrol Dial
Transplant. 2002;17:1713–1724.
29. London GM. Left ventricular hypertrophy: why does it happen? Nephrol
Dial Transplant. 2003;18(suppl 8):viii2–viii6.
30. Klassen PS, Lowrie EG, Reddan DN, DeLong ER, Coladonato JA,
Szczech LA, Lazarus JM, Owen WF Jr. Association between pulse
pressure and mortality in patients undergoing maintenance hemodialysis.
JAMA. 2002;287:1548 –1555.
31. Lucas MF, Quereda C, Teruel JL, Orte L, Marcen R, Ortuno J. Effect of
hypertension before beginning dialysis on survival of hemodialysis
patients. Am J Kidney Dis. 2003;41:814 – 821.
32. Liu JH, Chen CC, Wang SM, Chou CY, Liu YL, Kuo HL, Lin HH, Wang
IK, Yang YF, Huang CC. Association between pulse pressure and
30-month all-cause mortality in peritoneal dialysis patients. Am J
Hypertens. 2008;21:1318 –1323.
33. Apple FS, Murakami MM, Pearce LA, Herzog CA. Multi-biomarker risk
stratification of N-terminal pro-B-type natriuretic peptide, highsensitivity C-reactive protein, and cardiac troponin T and I in end-stage
renal disease for all-cause death. Clin Chem. 2004;50:2279 –2285.
34. Khan NA, Hemmelgarn BR, Tonelli M, Thompson CR, Levin A. Prognostic value of troponin T and I among asymptomatic patients with
end-stage renal disease: a meta-analysis. Circulation. 2005;112:
3088 –3096.
35. Wang AY, Lam CW, Chan HI, Goggins WB, Yu CM, Lui SF, Sanderson
JE. Prognostic value of cardiac troponin T is independent of inflammation, residual renal function, cardiac hypertrophy and dysfunction in
peritoneal dialysis patients. Clin Chem. 2007;53:882– 889.
36. Ooi DS, Isotalo PA, Veinot JP. Correlation of antemortem serum creatine
kinase, creatine kinase-MB, troponin I, and troponin T with cardiac
pathology. Clin Chem. 2000;46:338 –344.
37. Antman EM, Grudzien C, Mitchell RN, Sacks DB. Detection of unsuspected myocardial necrosis by rapid bedside assay for cardiac troponin T.
Am Heart J. 1997;138:596 –598.
38. Amann K, Breitbach M, Ritz E, Mall G. Myocyte/capillary mismatch in
the heart of uremic patients. J Am Soc Nephrol. 1998;9:1018 –1022.
39. Satyan S, Light RP, Agarwal R. Relationships of N-terminal pro-brain
natriuretic peptide and cardiac troponin T to left ventricular mass and
function and mortality in asymptomatic hemodialysis patients. Am J
Kidney Dis. 2007;50:1009 –1019.
40. Yasue H, Yoshimura M, Sumida H, Kikuta K, Kugiyama K, Jougasaki M,
Ogawa H, Okumura K, Mukoyama M, Nakao K. Localization and
mechanism of secretion of B-type natriuretic peptide in comparison with
those of A-type natriuretic peptide in normal subjects and patients with
heart failure. Circulation. 1994;90:195–203.
41. Cowie MR, Struthers AD, Wood DA, Coats AJ, Thompson SG, PooleWilson PA, Sutton GC. Value of natriuretic peptides in assessment of
patients with possible new heart failure in primary care. Lancet. 1997;
350:1349 –1353.
42. McDonagh TA, Robb SD, Murdoch DR, Morton JJ, Ford I, Morrison CE,
Tunstall-Pedoe H, McMurray JJ, Dargie HJ. Biochemical detection of
left-ventricular systolic dysfunction. Lancet. 1998;351:9 –13.
43. Zoccali C, Mallamaci F, Benedetto FA, Tripepi G, Parlongo S, Cataliotti
A, Cutrupi S, Giacone G, Bellanuova I, Cottini E, Malatino LS. Cardiac
natriuretic peptides are related to left ventricular mass and function and
predict mortality in dialysis patients. J Am Soc Nephrol. 2001;12:
1508 –1515.
44. Wang AY, Lam CW, Yu CM, Wang M, Chan IH, Zhang Y, Lui SF,
Sanderson JE. N-terminal pro-brain natriuretic peptide: an independent
risk predictor of cardiovascular congestion, mortality and adverse cardiovascular outcomes in chronic peritoneal dialysis patients. J Am Soc
Nephrol. 2007;18:321–330.
45. Wang AY, Lam CW, Wang M, Chan IH, Lui SF, Zhang Y, Sanderson JE.
Diagnostic potential of cardiac biomarkers for left ventricular hypertrophy and systolic dysfunction in chronic peritoneal dialysis patients.
Nephrol Dial Transplant. 2009;24:1962–1969.
46. Parekh RS, Plantinga LC, Kao WHL, Meoni LA, Jaar BG, Fink NE,
Powe NR, Coresh J, Klag MJ. The association of sudden cardiac death
with inflammation and other traditional risk factors. Kidney Int. 2008;74:
1335–1342.
47. Haubitz M, Brunkhorst R, Wrenger E, Froese P, Schulze M, Koch KM.
Chronic induction of C-reactive protein by hemodialysis, but not by
peritoneal dialysis therapy. Perit Dial Int. 1996;16:158 –162.
48. Samouilidou EC, Grapsa EJ, Kakavas I, Lagouranis A, Agrogiannis B.
Oxidative stress markers and C-reactive protein in end-stage renal failure
patients on dialysis. Int Urol Nephrol. 2003;35:393–397.
Sudden Cardiac Death in End-Stage Renal Disease Patients: A 5-Year Prospective
Analysis
Angela Yee-Moon Wang, Christopher Wai-Kei Lam, Iris Hiu-Shuen Chan, Mei Wang, Siu-Fai
Lui and John E. Sanderson
Downloaded from http://hyper.ahajournals.org/ by guest on May 6, 2017
Hypertension. 2010;56:210-216; originally published online July 6, 2010;
doi: 10.1161/HYPERTENSIONAHA.110.151167
Hypertension is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2010 American Heart Association, Inc. All rights reserved.
Print ISSN: 0194-911X. Online ISSN: 1524-4563
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://hyper.ahajournals.org/content/56/2/210
Data Supplement (unedited) at:
http://hyper.ahajournals.org/content/suppl/2010/07/02/HYPERTENSIONAHA.110.151167.DC1
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Hypertension can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial
Office. Once the online version of the published article for which permission is being requested is located,
click Request Permissions in the middle column of the Web page under Services. Further information about
this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Hypertension is online at:
http://hyper.ahajournals.org//subscriptions/
Online Supplement
Sudden Cardiac Death in End-Stage Renal Disease Patients: A 5year Prospective Analysis
Authors list: Angela Yee-Moon WANG1*, MD, PhD, FRCP, Christopher Wai-Kei
LAM2, PhD, FACB, Iris Hiu-Shuen CHAN2, PhD, Mei WANG1*, PhD, Siu-Fai LUI1,
FRCP, John E SANDERSON1, MD, FRCP, FACC.
Affiliations: 1Department of Medicine & Therapeutics, 2Department of Chemical
Pathology, The Chinese University of Hong Kong, Prince of Wales Hospital, Shatin.
New Territories, Hong Kong, 3Macau Institute for Applied Research in Medicine and
Health, Macau University of Science and Technology Foundation, Taipa, Macau
Current affiliations: *Department of Medicine, Queen Mary Hospital, University of
Hong Kong, Hong Kong.
Address for correspondence:
*Dr. Angela Yee-Moon Wang,
University Department of Medicine,
The University of Hong Kong,
Queen Mary Hospital,
102 Pokfulam Road,
Pokfulam, Hong Kong.
Ph: 852-28554949
Fax: 852-28555411
E-mail: [email protected]
Method
Assessments
Data Collection
Clinical and demographic data including the presence of diabetes, smoking
status, details of any previous history of angina, myocardial infarction with or without
percutaneous coronary intervention or coronary artery bypass grafting, previous
stroke or transient ischemic attacks, intermittent claudication or other symptoms
suggestive of peripheral vascular disease with or without history of amputation or
revascularization and use of different anti-hypertensive medications and
erythropoietin were recorded. With a mercury sphygmomanometer, systolic and
diastolic blood pressure were measured once on every follow-up visit after patient
was rested for 15 minutes on either arm at 8 week intervals for 12 months preceding
study entry and were then averaged to give the final systolic and diastolic blood
pressure.
Biochemical measurements
cTnT and NT-pro-BNP were measured by electrochemiluminescence
immunoassays on the Elecsys 2010 analyser (Roche Diagnostics Corp, Indianapolis,
IN, USA) with a detection limit of 0.01 µg/L and 5 pg/mL, respectively. For samples
with NT-pro-BNP concentrations above the measuring range 35,000 pg/mL, the final
concentrations would be taken as 35,000 pg/mL. CRP and albumin were measured
using the Tina-quant CRP latex ultra-sensitive assay (detection limit of 0.01mg/L)
and the bromcresol purple method, respectively, on the Roche modular analyzer
(Roche Diagnostics, USA). IL-6 was measured using the enzyme-linked
immunosorbent assay (ELISA) (BioSource International Inc, Camarillo, CA, USA)
(detection limit of 2 pg/mL). Serum fetuin-A was determined using a human fetuin-A
ELISA assay kit (Epitope Diagnostics, San Diego, USA). Plasma MPO was
measured using a sandwich ELISA kit (Immunodiagnostik AG, Bensheim, Germany)
with a detection limit of 1.6 µg/L. Calcium and phosphorus concentrations were
measured using dye-binding methods on the Dimension AR automatic analyzer (Du
Pont Co, Wilmington, DE, USA). iPTH was determined by chemiluminescence
immunoassay on the Immulite analyzer (Diagnostic Products Corp, Los Angeles, CA,
USA). Total cholesterol and triglyceride were assayed enzymatically (Hitachi 911
analyser, Roche Diagnostics GmbH, Mannheim, Germany).
Statistical Analysis
Multivariable Cox regression analysis
In echocardiography-based model, we considered clinical, biochemical,
dialysis and echocardiography parameters but not serum biomarkers. In biomarkerbased model, clinical, biochemical, dialysis parameters and serum biomarkers were
included but not echocardiographic parameters. In combined echocardiography and
biomarker-based model, clinical, biochemical, dialysis, echocardiography parameters
and serum biomarkers were considered.
Shrinkage factor
In brief, the shrinkage factor was derived using the formula: (overall chi
square of the model to the power 2 – the number of parameters in the model) divided
by overall chi square of the model to the power 2. Each regression coefficient is then
multiplied by the shrinkage factor. The standard error (SE) is considered unchanged.
The shrinkage corrected hazard ratios can then be calculated from the formula:
hazard ratios = exp shrinkage corrected coefficient. The corresponding 95% confidence
intervals can be calculated by the formula: 95% CI = exp shrinkage corrected coefficient ±
1.96*SE
. We calculated the P-value from the Wald statistics which produced a ChiSquare distribution with 1 degree of freedom. The exact P value is the P-value
corresponding to a given Chi-Square value.
Results
Table S1. Cox regression analysis of SCD in relation to tertiles of N-terminal probrain natriuretic peptide and cardiac troponin T considered separately and together.
Biomarker
Unadjusted HR
(95% CI)
P
NT-pro-BNP
Lower tertile
Middle tertile
Upper tertile
1
6.37 (1.41 – 28.78)
11.94 (2.72 – 52.43)
0.016
0.001
cTnT
Lower tertile
Middle tertile
Upper tertile
1
2.77 (0.85 – 8.99)
5.67 (1.87 – 17.14)
0.09
0.002
Adjusted HR
(95% CI)*
1
5.06 (1.09 – 23.51)
7.54 (1.55 – 36.69)
Chi-square = 8.93, P=0.012
P
0.038
0.012
1
1.91 (0.58 – 6.32)
0.29
2.58 (0.78 – 8.57)
0.13
Chi-square = 2.71, P=0.26
NT-pro-BNP, N-terminal-pro-brain natriuretic peptide; cTnT, cardiac troponin T; HR,
hazard ratios; CI, confidence intervals.
*The Cox regression model of NT-pro-BNP in relation to SCD was adjusted for cTnT.
The Cox regression model of cTnT in relation to SCD was adjusted for NT-pro-BNP.
The likelihood ratio chi-square statistic was used for model comparisons.
For NT-pro-BNP: Lower tertile NT-pro-BNP < 2665 pg/mL(n=77); middle tertile ≥
2666 – 11893 pg/mL(n=77); upper tertile ≥ 11894 pg/mL (n=76).
For cTnT: Lower tertile cTnT ≤ 0.01 µg/L (n=81); middle tertile > 0.01 – 0.099 µg/L
(n=70); upper tertile ≥ 0.1 µg/L (n=79).