Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Mar 1.
Published in final edited form as: Am J Kidney Dis. 2021 Jul 19;79(3):383–392. doi: 10.1053/j.ajkd.2021.06.017

Upper Reference Limits for High-Sensitivity Cardiac Troponin T and N-Terminal Fragment of the Prohormone Brain Natriuretic Peptide in Patients With CKD

Nisha Bansal 1, Leila Zelnick 1, Christie Ballantyne 2, Paulo Chaves 3, Robert Christenson 4, Joseph Coresh 5, Chris deFilippi 6, James de Lemos 7, Lori Daniels 8, Alan S Go 9, Jiang He 10, Susan Heydati 7, Kuni Matsushita 5, Vijay Nambi 2, Michael Shlipak 11, Jonathan Taliercio 12, Stephen Seliger 4, on behalf of the CRIC Study Investigators
PMCID: PMC8766621  NIHMSID: NIHMS1725815  PMID: 34293394

Abstract

Rationale & Objective:

The utility of conventional upper reference limits (URL) for N-terminal pro brain natriuretic peptide (NT-proBNP) and high-sensitivity cardiac troponin T (hsTnT) in chronic kidney disease (CKD) remains debated. We analyzed the distribution of hsTnT and NT-proBNP in people with CKD in ambulatory settings to examine the diagnostic value of conventional URL in this population.

Study Design:

Observational study.

Setting & Participants:

We studied participants of the Chronic Renal Insufficiency Cohort with CKD and no self-reported history of cardiovascular disease (CVD).

Exposure:

Estimated glomerular filtration rate (eGFR).

Outcomes:

NT-proBNP and hsTnT at baseline.

Analytical Approach:.

We described the proportion of participants above the conventional URL for NT-proBNP (125 pg/ml) and hsTnT (14 ng/L) overall, and by eGFR. We then estimated 99th percentile URL for NT-proBNP and hsTnT. Using quantile regression of the 99th percentile, we modeled the association of eGFR with NT-proBNP and hsTnT.

Results:

Among 2,312 CKD participants, 40% and 43% had levels of NT-proBNP and hsTnT above conventional URL, respectively. In those with eGFR<30 ml/min/1.73 m2, 71% and 68% of participants had concentrations of NT-proBNP and hsTnT above conventional URL, respectively. Amongst all CKD participants, the 99th percentile for NT-proBNP was 3,592(95% CI: 2,470, 4,849) pg/mL and for hsTnT was 126(95% CI: 100, 144) ng/L. Each 15 ml/min/1.73 m2 decrement in eGFR was associated with a ~40% higher threshold for the 99th percentile of NT-proBNP [1.43 (1.21, 1.69)] and hsTnT [1.45 (1.31, 1.60)].

Limitations:

Study included ambulatory patients; we could not test the accuracy of upper reference limits of NT-proBNP and hsTnT in the acute care setting.

Conclusions:

In this ambulatory CKD population with no self-reported history of CVD, a range of 40–88% of participants had concentrations of NT-proBNP and hsTnT above conventional URL, depending on eGFR strata. Developing eGFR-specific thresholds for these commonly-used cardiac biomarkers in the setting of CKD may improve their utility for evaluation of suspected heart failure and myocardial infarction.

Keywords: brain natriuretic peptide, troponin, chronic kidney disease, cardiac biomarkers, heart failure, myocardial infarction

PLAIN LANGUAGE SUMMARY

The cardiac biomarkers N-terminal pro brain natriuretic peptide (NT-proBNP) and high-sensitivity cardiac troponin T (hsTnT) are frequently used clinically to evaluate patients for suspected acute heart failure or myocardial infarction. However, the utility of conventional upper reference limits (URL) for NT-proBNP) and hsTnT in chronic kidney disease (CKD) remains debated. We analyzed the distribution of hsTnT and NT-proBNP in ambulatory CKD participants to examine how conventional URL apply to this population. In this ambulatory CKD population with no self-reported history of CVD, we found that the majority of participants had concentrations of NT-proBNP and hsTnT above conventional URL. Developing thresholds for these commonly-used cardiac biomarkers when used in the setting of CKD may improve their utility for evaluation of suspected heart failure and myocardial infarction.

INTRODUCTION

Cardiac-specific biomarkers N-terminal pro brain natriuretic peptide (NT-proBNP) and high sensitivity cardiac troponin T (hsTnT) are widely used clinically for the rapid diagnosis of acute heart failure (AHF) and myocardial infarction (MI), respectively.1–4 The newer, high sensitive assay for TnT permits earlier diagnoses of MI as compared with prior conventional TnT assays.5–7

An extensive body of research has demonstrated that patients with chronic kidney disease (CKD) are at greater risk of AHF and MI compared with the general population.8, 9 Data from our group and others have reported strong associations of elevations in NT-proBNP and hsTnT with the long-term risk of cardiovascular disease and death in patients with CKD, even when adjusting for estimated glomerular filtration rate (eGFR).10–18 For example, in a cohort of over 3,000 ambulatory persons with CKD free of HF, those with the highest levels of NT-proBNP and hsTnT had greater than 7-fold and 2-fold higher risk, respectively, of developing incident AHF over a median time of 8 years.10 However, asymptomatic CKD patients may have levels of hsTnT and NT-proBNP well above the thresholds used clinically to indicate acute cardiac disease, and it has been suggested that kidney function should be considered when evaluating thresholds and normal ranges of these biomarkers.19–22

The standard approach to defining an upper reference limit of cardiac troponins is to estimate the 99th percentile value in a sample of healthy volunteers without known cardiac disease. For hsTnT, the current manufacturer’s-reported 99th percentile upper reference limit for hsTnT assay is 14 ng/L, derived from a sample of 616 healthy volunteers.23 For NT-proBNP, a variety of center-specific reference ranges have been reported using different subject samples to establish these reference ranges. The current FDA-approved threshold to rule out HF is 125 pg/mL (450 pg/mL for age >75 years), providing a negative predictive value of 99% based on data from select samples, most presenting with acute respiratory symptoms.24 Unlike hsTnT, NT-proBNP upper thresholds have not been determined in asymptomatic, ambulatory population; however this approach may be informative for high risk patients, such as those with CKD, where traditional upper thresholds from these select studies may not apply.

To address this gap in knowledge, we analyzed the distribution of hsTnT and NT-proBNP in a large cohort of ambulatory asymptomatic persons with CKD who were without clinical cardiovascular disease to examine how conventional upper thresholds for NT-proBNP and hsTnT apply to a diverse CKD population. We also explored CKD-specific reference ranges of these biomarkers based on participant eGFR categories, age, sex and race.

METHODS

Study population

We studied adults with mild to moderate CKD (eGFR 20–70 ml/min/1.73 m2) enrolled in the Chronic Renal Insufficiency Cohort (CRIC) Study. A total of 3,939 participants were enrolled into the CRIC study between June 2003 and August 2008 at seven clinical centers across the United States.25, 26 Inclusion and exclusion criteria have been previously described.25 Participants on maintenance dialysis or with a kidney transplant were not included at cohort entry. CRIC also excluded participants with advanced heart failure (HF), defined as New York Heart Association Class III or IV, on cohort entry. All study participants provided written informed consent, and the study protocol was approved by institutional review boards at each site.

For the present analysis, we excluded participants who had insufficient stored sera at baseline to measure NT-proBNP and hsTnT. We also excluded participants with a self-reported history of HF, coronary artery disease, stroke, prior revascularization, peripheral vascular disease and atrial fibrillation at cohort entry. After applying these exclusions, 2,312 participants were analyzed (Figure 1)

Figure 1.

Figure 1.

Open in a new tab

Derivation of the study cohort

Biomarker Measurements

NT-proBNP and hsTnT were originally measured at from EDTA plasma stored at −70°C and collected from the baseline study visit. Both assays were performed using chemiluminescent microparticle on the Elecsys 2010 (Roche Diagnostics, Indianapolis, IN) in 2008. However, to calibrate these measures to a newer instrument, we remeasured a subset of both NT-proBNP and hsTnT in 2017 on the Roche E601. We remeasured NT-proBNP and hsTnT in 100 random samples. Additionally, for hsTnT, we remeasured any sample with a value <5ng/L since the limit of detection for the newer instrument was low (<3 ng/L for the limit of detection and 2.5 ng/L for the limit of blank).27 For both NT-proBNP and hsTnT, we developed and applied a Deming regression28 to calibrate the 2008 measures with the 2017 NT- measures. Deming regression is an extension of linear regression in which both the “X” and “Y” variables are presumed to have random measurement error; but does not assume that there is the same amount of measurement error in both “X” and “Y”. The following equations were applied: calibrated NT-proBNP = −17.25 + 1.02 * uncalibrated NT-proBNP and calibrated hsTnT = 2.77 + 0.99 * uncalibrated hsTnT.

For the purposes of this analysis, participants with undetectable levels of each biomarker were set to half of the lower limit of detection (2.5 pg/mL for NT-proBNP, 1.5 ng/mL for hsTnT) and included in all analyses

Covariates

At the baseline visit and each annual visit, participants provided information on their sociodemographic characteristics, medical history, medication usage, and lifestyle behaviors. Race/ethnicity was categorized as non-Hispanic white, non-Hispanic black, Hispanic, and other. Current smoking (yes/no) was determined by self-report. Diabetes mellitus was defined as a fasting glucose >126 mg/dL, a non-fasting glucose >200 mg/dL, or use of insulin or other antidiabetic medication. Anthropometric measurements and blood pressure (BP) were assessed using standard protocols. Body mass index (BMI) was derived as weight in kg divided by height in meters squared. Serum creatinine was measured using an enzymatic method on an Ortho Vitros 950 at the CRIC Central Laboratory and standardized to isotope dilution mass spectrometry-traceable values29, 30 and the CKD-EPI equation was used to estimate GFR.31

Research transthoracic echocardiograms were performed at Year 1. Assessments of cardiac structure and function were performed using echocardiography according to American Society of Echocardiography guidelines and were quantified at the CRIC Central Echocardiography Laboratory at the University of Pennsylvania. Left ventricular hypertrophy (LVH) was defined as ≥ 50 grams for males and ≥47 grams for females using the Cornell criteria. Reduced left ventricular ejection fraction (LVEF) was defined as <50%.

Statistical analyses

We first described characteristics of the study sample across eGFR levels with means and standard deviations for continuous variables or frequencies and percentages for categorical variables. We then generated scatterplots of NT-proBNP and hsTnT against eGFR. Using quantile regression of the 99th percentile, we modeled the association of linear eGFR with log-transformed NT-proBNP and hsTnT, first unadjusted, then adjusting for age, age2 (since the association with age and the biomarker level was non-linear), sex, race, body mass index (BMI), diabetes mellitus, and hypertension. These covariates were chosen based on biological plausibility. The exponentiated coefficients were then scaled to present the fold-change associated with a 15 mL/min/1.73m2 decrement in eGFR. We then generated kernel density plots illustrating the distribution of NT-proBNP and hsTnT by eGFR stratum (<30, 30–44, 45–59 and ≥60 ml/min/1.73 m2), age, sex and race based on previous studies that suggest that demographics are also important determinants of levels of these cardiac biomarkers.

We described the proportion of participants above the traditional upper reference limits for NT-proBNP and hsTnT overall, and across strata of eGFR and demographics. We then calculated the 99th percentile for NT-proBNP and hsTnT overall, by eGFR strata, and by combinations of sex, race and age by eGFR strata; 95% confidence intervals for each of these estimates were calculated from the 2.5th and 97.5th percentiles of the distribution obtained from the nonparametric bootstrap with 5000 replicates. In secondary analyses, we calculated the 95th percentile for NT-proBNP and hsTnT.

We performed two sensitivity analyses. In the first, we included participants with prevalent cardiovascular disease in the primary analyses, bringing the total number of participants to 3868. We chose to exclude these participants in the primary analyses since comparable studies to examine reference limits of cardiac biomarkers have typically included populations free of prevalent cardiac diseases. In a secondary sensitivity analysis in which participants who had either LVH or reduced LVEF by a research echocardiogram at Year 1 were excluded, leaving 942 participants for analysis. The rationale for this analysis was that previous studies have demonstrated associations of increased LV mass and decreased LVEF with levels of NT-proBNP and hsTnT in patients with CKD19.

All analyses were performed using R 3.6.2 (R Foundation for Computing, Vienna, Austria).

RESULTS

Characteristics of study population

Among the 2,312 participants free of cardiovascular disease, 378 (16.3%) had eGFR<30 ml/min/1.73 m2, 766 (33.1%) had eGFR 30–44 ml/min/1.73 m2, 732 (31.7%) had eGFR 45–59 ml/min/1.73 m2, and 436 (18.9%) had eGFR ≥60 to <70 ml/min/1.73 m2. Participants with lower baseline eGFR were older, more likely to be black or Hispanic, were more likely to have hypertension and diabetes, and have a higher blood pressure (Table 1). Further, the median values of NT-proBNP and hsTnT were higher among participants with lower eGFR.

Table 1.

Baseline characteristics of included participants by level of estimated glomerular filtration rate (eGFR, ml/min/1.73 m2) in the CRIC study

eGFR < 30 eGFR 30–44 eGFR 45–59 eGFR ≥ 60
N 378 766 732 436
Age (years) 56.3 (12.1) 57.3 (11.8) 56.3 (10.8) 49.8 (10.9)
Male 189 (50) 387 (51) 412 (56) 231 (53)
Race/ethnicity
 White 115 (30) 301 (39) 339 (46) 226 (52)
 Black 157 (42) 299 (39) 282 (39) 155 (36)
 Hispanic 91 (24) 135 (18) 76 (10) 35 (8)
 Other 15 (4) 31 (4) 35 (5) 20 (5)
Highest level of education
 Less than HS 117 (31) 191 (25) 101 (14) 36 (8)
 HS 86 (23) 133 (17) 126 (17) 65 (15)
 Some college 95 (25) 223 (29) 209 (29) 136 (31)
 College or above 80 (21) 219 (29) 296 (40) 198 (45)
Current smoking 52 (14) 101 (13) 70 (10) 54 (12)
Height (cm) 166.4 (9.7) 167.7 (9.8) 169.3 (9.6) 170.5 (9.6)
Weight (kg) 88.5 (23.6) 90.1 (22.2) 90.7 (24.1) 90.2 (22.2)
BMI (kg/m2) 32.0 (8.2) 32.0 (7.7) 31.6 (8.1) 31.0 (7.1)
Diabetes at baseline 187 (49) 371 (48) 286 (39) 124 (28)
Hypertension 349 (92) 688 (90) 614 (84) 272 (62)
Antihypertensive medications 357 (94) 715 (93) 632 (86) 294 (67)
Statins 202 (53) 367 (48) 338 (46) 137 (31)
Systolic blood pressure (mmHg) 132.6 (23.5) 129.2 (21.4) 125.9 (19.9) 120.4 (18.1)
LDL (mg/dL) 104.7 (37.6) 104.0 (35.9) 108.1 (34.4) 110.7 (32.5)
HDL (mg/dL) 46.8 (14.7) 48.0 (15.8) 49.8 (16.9) 50.0 (16.9)
eGFR (mL/min/1.73m2) 24.7 (3.9) 38.1 (4.3) 52.0 (4.4) 69.5 (8.8)
24-hour urine protein (g), median (IQR) 0.7 (0.2–2.2) 0.3 (0.1–1.2) 0.1 (0.1–0.4) 0.1 (0.1–0.2)
Left ventricular hypertrophy by Year 1 echocardiogram** 147 (39) 221 (29) 194 (27) 82 (19)
LVEF < 50% by Year 1 echocardiogram 34 (9) 83 (11) 84 (11) 58 (13)
NT-proBNP (pg/mL), median (IQR) 255.6 (109.6–513.1) 111.9 (54.7–250.8) 64.9 (25.8–160.2) 35.4 (9.2–91.0)
Undetectable (<5 pg/mL) NT-proBNP* 7 (2) 24 (3) 74 (10) 78 (18)
hsTNT (ng/L), median (IQR) 21.4 (11.8–41.0) 14.3 (8.8–23.6) 11.0 (7.3–17.0) 7.6 (5.3–11.5)
Undetectable (< 3 ng/L) hsTNT* 1 (0) 5 (1) 12 (2) 20 (5)
Open in a new tab

Entries are mean (SD) for continuous variables or N (%) for categorical variables, except as noted.

*

Undetectable limits are 5 pg/mL for NT-proBNP and 3 ng/L for hsTNT;

**

Defined as LVM (Cornell) ≥ 50 for males, LVM (Cornell) ≥ 47 for females

Distribution of NT-proBNP and hsTnT by eGFR category

Density plots demonstrated that the distribution of NT-proBNP and hsTnT varied by eGFR (Figure 2). The distribution of each biomarker was shifted to the right (e.g. corresponding to higher levels) for those with lower eGFR, older age, male sex and black race; differences in distributions were qualitatively largest for eGFR strata (Figure 2 and Figure S1)

Figure 2:

Figure 2:

Open in a new tab

Density plots for NT-proBNP and hsTNT by eGFR stratum

Application of traditional upper thresholds of NT-proBNP and hsTnT among participants with CKD

If traditional upper threshold limits of NT-proBNP (125 pg/ml) and hsTnT (14 ng/L) were applied to this CKD cohort overall, then 40% and 43%, respectively, of these ambulatory asymptomatic persons would exceed these reference limits. In those with eGFR<30 ml/min/1.73 m2 specifically, 71% and 68% of participants would have values of NT-proBNP and hsTnT, respectively, above these upper reference limits (Table 2). The proportion of participants exceeding these traditional upper reference limits of NT-proBNP was greater in lower eGFR categories across all subgroups of age, race and sex (Table 2).

Table 2:

Proportion of CKD participants above conventional upper reference limits for NT-proBNP and hsTnT

Category N (%) with NT-proBNP > 125 pg/mL N (%) with hsTnT > 14 ng/L
Overall 922 (40) 986 (43)
eGFR
 < 30 269 (71) 256 (68)
 30 – 44 349 (46) 392 (51)
 45 – 59 222 (30) 261 (36)
 ≥ 60 82 (19) 77 (18)
Male
 < 30 125 (66) 156 (83)
 30 – 44 155 (40) 259 (67)
 45 – 59 93 (23) 187 (45)
 ≥ 60 39 (17) 62 (27)
Female
 < 30 144 (76) 100 (53)
 30 – 44 194 (51) 133 (35)
 45 – 59 129 (40) 74 (23)
 ≥ 60 43 (21) 15 (7)
Black
 < 30 105 (67) 106 (68)
 30 – 44 120 (40) 162 (54)
 45 – 59 70 (25) 123 (44)
 ≥ 60 28 (18) 43 (28)
Non-black
 < 30 164 (74) 150 (68)
 30 – 44 229 (49) 230 (49)
 45 – 59 152 (34) 138 (31)
 ≥ 60 54 (19) 34 (12)
Age < 60
 < 30 142 (68) 128 (61)
 30 – 44 163 (43) 177 (47)
 45 – 59 117 (29) 137 (33)
 ≥ 60 54 (16) 58 (17)
Age ≥ 60
 < 30 127 (76) 128 (76)
 30 – 44 186 (48) 215 (56)
 45 – 59 105 (33) 124 (39)
 ≥ 60 28 (30) 19 (20)
Open in a new tab

Association between eGFR and NT-proBNP and hsTnT among participants with CKD

Scatterplots demonstrated a monotonic, inverse association between eGFR and the 99th percentile of N-proBNP and hsTnT (Figure S2). In unadjusted models, each 15-ml/min/1.73 m2 decrement in eGFR was associated with a 2.02 (1.57, 2.60) fold and 1.42 (1.18, 1.70) fold higher 99th percentile of NT-proBNP and hsTnT, respectively. In multivariable models that were adjusted for age, age2, sex, race, BMI, diabetes, and hypertension, the 99th percentile values of NT-proBNP and hsTnT were 1.43 (1.21, 1.69) fold and 1.45 (1.31, 1.60) fold higher, respectively, per 15 ml/min/1.73 m2 lower value of baseline eGFR.

99th and 95th percentile of NT-proBNP and hsTnT by eGFR category

In this CKD sample overall, the 99th percentile for NT-proBNP was 3,592 (95% CI: 2,470, 4,849) pg/mL and for hsTnT was 126 (95% CI: 100, 144) ng/L (Table 3). Notably, there were graded increases in the 99th percentile values across lower eGFR categories for both NT-proBNP and hsTnT . The 99th percentile for NT-proBNP was 1677 (95% CI 600, 6282) pg/mL, 1887 (95% CI 1146, 2922) pg/mL, 2921 (95% CI 2034, 4715) pg/mL, and 8402 (95% CI: 4526, 15460) pg/mL for eGFR ≥60, 45–59, 30–44 and <30 ml/min/1.73 m2, respectively. The 99th percentile for hsTnT was 51 (95% CI 35, 62) ng/L, 97 (95% CI 69, 120) ng/L, 127 (95% CI 97, 149) ng/L, and 219 (95% CI: 136, 300) ng/L for eGFR ≥60, 45–59, 30–44 and <30 ml/min/1.73 m2, respectively.

Table 3:

99th and 95th percentiles of NT-proBNP and hsTnT by eGFR category

Category 95th % NT-proBNP 99th % NT-proBNP 95th % hsTNT 99th % hsTNT
Overall 1039 (891, 1169) 3592 (2470, 4849) 58 (53, 64) 126 (100, 144)
eGFR
 < 30 2523 (1735, 3424) 8402 (4526, 15460) 93 (77, 122) 219 (136, 300)
 30 – 44 1130 (824, 1482) 2921 (2034, 4715) 59 (50, 68) 127 (97, 149)
 45 – 59 682 (503, 907) 1887 (1146, 2922) 43 (35, 53) 97 (69, 120)
 ≥ 60 317 (238, 516) 1677 (600, 6282) 26 (22, 31) 51 (35, 62)
Male
 < 30 3675 (2243, 5085) 9671 (4706, 25000) 118 (84, 158) 241 (145, 300)
 30 – 44 1374 (847, 1880) 3729 (2134, 5351) 82 (63, 98) 144 (103, 187)
 45 – 59 641 (415, 878) 1400 (984, 2880) 54 (41, 67) 106 (75, 133)
 ≥ 60 354 (212, 605) 2527 (572, 8902) 31 (24, 38) 56 (36, 62)
Female
 < 30 1742 (1107, 2407) 5749 (2165, 15460) 66 (49, 79) 189 (77, 734)
 30 – 44 826 (588, 1258) 1922 (1518, 2950) 36 (28, 44) 65 (47, 136)
 45 – 59 737 (477, 1099) 2291 (1149, 3626) 27 (22, 35) 63 (41, 100)
 ≥ 60 292 (192, 516) 1159 (441, 5850) 19 (13, 25) 33 (23, 76)
Black
 < 30 2046 (1217, 3281) 7073 (2436, 15460) 103 (77, 147) 293 (128, 734)
 30 – 44 968 (601, 1470) 2151 (1486, 2953) 60 (49, 67) 101 (68, 142)
 45 – 59 658 (408, 1099) 2135 (1109, 4632) 48 (36, 61) 91 (61, 111)
 ≥ 60 373 (198, 632) 1737 (483, 6358) 32 (23, 39) 51 (36, 62)
Non-black
 < 30 2851 (1754, 4415) 9037 (3888, 22147) 86 (68, 118) 178 (115, 293)
 30 – 44 1212 (788, 1571) 3727 (1928, 5147) 58 (48, 75) 134 (99, 162)
 45 – 59 687 (497, 897) 1755 (1051, 2622) 39 (32, 52) 98 (64, 128)
 ≥ 60 295 (202, 483) 1811 (485, 6801) 23 (18, 27) 45 (27, 65)
Age < 60
 < 30 2673 (1657, 4077) 10144 (3109, 24332) 94 (71, 130) 252 (118, 695)
 30 – 44 1114 (724, 1539) 3057 (1630, 4829) 77 (56, 96) 143 (103, 171)
 45 – 59 687 (408, 926) 1688 (1091, 2922) 55 (40, 70) 109 (90, 146)
 ≥ 60 366 (204, 591) 2578 (606, 6358) 25 (21, 34) 54 (35, 62)
Age ≥ 60
 < 30 2266 (1388, 4415) 6508 (2917, 10736) 93 (68, 136) 187 (111, 274)
 30 – 44 1110 (737, 1564) 2698 (1668, 5351) 50 (44, 58) 82 (59, 124)
 45 – 59 659 (491, 978) 2093 (1060, 3647) 34 (28, 42) 56 (43, 63)
 ≥ 60 269 (214, 338) 431 (265, 689) 27 (20, 35) 35 (27, 39)
Open in a new tab

Entries are estimates (95% bootstrapped CIs) of the 95th and 99th percentile

Within demographic categories of sex, race (black vs. non-black) and age (<60 or ≥60 years), there was an even more pronounced difference in the 99th percentile of hsTnT in each eGFR strata, with higher 99th thresholds observed in subgroups of participants with lower eGFR who were men, non-Black or younger in age (Table 3).

The 95th percentile for NT-proBNP was 1039 (891, 1169) pg/mL and for hsTnT was 58 (53, 64) ng/L overall, with higher values in participants with lower eGFR, who were men, non-Black, and younger in age (Table 3) The 95th percentile for NT-proBNP was 317 (95% CI 238, 516) pg/mL, 682 (95% CI 503, 907) pg/mL, 1130 (95% CI 824, 1482) pg/mL, and 2523 (95% CI: 1735, 3424) pg/mL for eGFR ≥60, 45–59, 30–44 and <30 ml/min/1.73 m2, respectively. The 95th percentile for hsTnT was 26 (95% CI 22, 31) ng/L, 43 (95% CI 35, 53) ng/L, 59 (95% CI 50, 68) ng/L, and 93 (95% CI: 77, 122) ng/L for eGFR ≥60, 45–59, 30–44 and <30 ml/min/1.73 m2, respectively.

Sensitivity analysis: inclusion of participants with prevalent cardiovascular disease.

When we included participants with prevalent CVD in the analyses, we found that the proportion of participants who exceeded convention URL for NT-proBNP and hsTnT was even higher than that reported in the primary analyses (52% for NT-proBNP and 53% for hsTnT overall) (Table S1). Moreover, the 95% and 99% thresholds were also higher in this population (Table S2).

Sensitivity analysis: exclusion of participants with subclinical cardiovascular disease

Among the 942 participants without LVH or reduced LVEF, the 99th percentile for NT-proBNP was 1300 (95% CI: 878, 1509) pg/mL and for hsTnT was 75 (95% CI: 59, 102) ng/L, which was lower than that observed in the primary analysis without excluding subclinical cardiac disease (Table S3). Similar to the primary analysis, there was a graded association of higher 99th percentile level of each biomarker with lower eGFR. For NT-proBNP, the 99th percentile ranged from 444 to 5,394 pg/mL; and for hsTnT, it ranged from 34 to 122 ng/L across eGFR strata.

DISCUSSION

In this analysis of ambulatory CKD participants free of clinical cardiovascular disease, we found that ~40% of participants were above the upper threshold values for NT-proBNP and hsTnT that were derived from general patient or volunteer populations; this proportion was as high as 88% among subgroups of participants in the lowest eGFR category. Our data suggest application of traditional thresholds for NT-proBNP and hsTnT may not be appropriate for patients with CKD. Although persons with CKD have higher risk for AHF and MI, the current biomarker thresholds may exaggerate their likelihood of having these diagnoses to the exclusion of alternative possibilities. Alternatively, eGFR specific or individualized thresholds could be developed and ultimately be incorporated into clinical practice to inform the interpretation of NTproBNP and/or hsTnT measurements in CKD patients presenting with acute breathlessness or chest pain to more accurately revise the probability of AHF and MI, events for which CKD patients are at markedly greater risk of developing.

Despite data demonstrating strong associations of elevated NT-proBNP and hsTnT with long-term risk of adverse cardiovascular and renal events in patients with CKD,11, 27, 32 there remains uncertainty on how to best to apply the current upper reference limits in CKD patients for the diagnosis of acute cardiac disease. Often, elevations in these biomarkers are dismissed as simply due entirely to the effect of reduced kidney excretion, or alternatively low-level elevations in symptomatic CKD patients may be chronic rather than a reflection of acute cardiac pathology. While there has been substantial debate about the degree to which kidney excretion impacts the circulating concentrations of these biomarkers,14, 33–38 decreased GFR cannot account entirely for the elevations.22 For example, in a study of 165 subjects undergoing renal arteriography with invasive renal plasma flow measurements, the median fractional renal extraction of NT-proBNP was 0.16 (IQR 0.09 to 20) to 0.18 (IQR 0.12 to 0.22).39 Similarly, studies of TnT have also shown partial, but not complete, kidney clearance. In a study of heart failure patients undergoing renal vein catherization the extraction index of TnT was 8–19%.40 Therefore, the notion that elevations in NT-proBNP and hsTnT are entirely related to reduced renal clearance are not likely true. In this analysis, we found elevation in hsTnT and NT-proBNP even in individuals with a preserved eGFR>60. In addition, greater TnT levels in asymptomatic ambulatory CKD patients were strongly related to greater LVH, even after accounting for kidney function, suggesting that greater myocardial damage explains greater TnT levels.19, 41 Thus, there is a need to define the optimal thresholds of NT-proBNP and hsTnT in CKD patients to guide clinical management, considering the potential effects of reduced kidney excretion on the interpretation of these biomarkers to identify patients with true acute cardiac pathology. This approach has been used in other populations, including older adults.42

The 99th percentile upper threshold derived from healthy populations has been used for diagnostic thresholds for some biomarkers including hsTnT. In our study, we found that there were marked differences in the 99th percentile for NT-proBNP and hsTnT across eGFR strata. For example, in those with advanced CKD (<30 ml/min/1.73m2), the 99th percentile for NT-proBNP and hsTnT were substantially higher (8400 pg/mL and 219 ng/L, respectively) than rule in thresholds used in the non-CKD population (450 pg/mL and 14 ng/L, respectively). Even in participants with mild and moderate CKD, the 99th percentiles were markedly higher compared with those used in the general population. This was also demonstrated when we reported high proportions of participants with levels of NT-proBNP and hsTnT above the traditional upper reference limits, even in these ambulatory, asymptomatic subjects. After excluding those with LVH or reduced LVEF, we still found that the 99th percentiles for NT-proBNP and hsTnT were still much higher across eGFR strata as compared with the thresholds reported in the general population (albeit lower than the primary population in our study).

While previous studies have not specifically defined upper reference limits for troponin T in CKD patients across a wide range of eGFR, they have shown that elevations in troponin are associated with worse short-term outcomes in patients with CKD. In a post-hoc analysis of the ACUITY trial (which randomized patients with non-ST elevation acute coronary syndrome to different anti-thrombotic regimens), patients with CKD with abnormal troponin (defined as any value above the reference limit used in all populations) had higher rates of MI and 30-day mortality.34 In a study of patients presenting with suspected acute coronary syndrome, 19% of patients had CKD (with mean eGFR of 43 ml/min/1.73 m2). The positive predictive value and specificity of traditional thresholds was lower in the CKD patients vs. those without CKD.43 In a systematic review of 23 studies of patients with CKD, applying the reference values used in the general population, the sensitivity of TnT for diagnosis of acute coronary syndrome ranged from 71–100% and specificity ranged from 31–86%.36 The varying sensitivity and, generally, lower specificity of troponin in these patients (compared to studies in patients without CKD) may have been due to the application of traditional reference ranges in CKD patients. Our study augments this previous literature by studying a wide range of CKD stages, including advanced stages of CKD, and proposing alternative URL for hsTnT derived from ambulatory CKD patients free of self-reported cardiovascular disease.

Similarly, there have been limited studies evaluating the performance of traditional NT-proBNP URLs in patients with CKD. In a study of patients presenting with dyspnea, nearly all the patients with CKD had NT-proBNP levels above the clinical threshold, suggesting low specificity of these conventional thresholds for diagnosing AHF.44 In another study of 599 patients presenting with dyspnea, a NT-proBNP cut-off of >450 pg/mL had 85% sensitivity and 88% specificity for a diagnosis of acute heart failure in the non-CKD population.45 In comparison, for those patients with eGFR<60 ml/min/1.73 m2, a higher NT-proBNP threshold of 1,200 pg/mL was required to achieve a specificity of 72% while maintaining a sensitivity of 89%.45 However, there were too few patients (n=19) with eGFR<30 to identify optimal cut-points for this subgroup. These data urge the need for eGFR-specific thresholds in interpretation of levels of NT-proBNP. Our study proposes possible thresholds based on data from relatively healthy ambulatory CKD patients free of known heart failure and other cardiac diseases. Further work is needed to test the performance of these thresholds to diagnose acute decompensated heart failure in patients presenting with acute dyspnea..

We also noted differences across sex, race and age within our eGFR strata with higher 99th percentile upper reference limits in participants who were male, not of Black race and younger. Other studies have also suggested demographic-based thresholds.42, 46–48 In particular, guideline committees endorse age-specific thresholds for interpretation of NT-proBNP and hsTnT.24, 42

Elevations in NT-proBNP and hsTnT in CKD patients are likely to reflect at least in part chronic cardiac pathology, and may be helpful in determining cardiovascular prognosis in CKD patients if eGFR specific clinical thresholds are developed. Elevations in these cardiac biomarkers may also be due to CKD-specific risk factors for cardiovascular disease including proteinuria, anemia, inflammation and alterations in mineral metabolism. CKD specific reference ranges would identify patients who are appropriate for more aggressive observation, diagnosis, and treatment. Patients with CKD have high rates of cardiovascular disease; and despite this risk, few diagnostic tools are available in the acute or chronic care setting to guide therapeutic decisions as well as aid in determining prognosis. Future studies are needed to evaluate operating characteristics (sensitivity, specificity, positive predictive value and negative predictive value) of applying these alternative cut-offs for NT-proBNP and hsTnT in CKD patients for identification of heart failure and myocardial infarction in the acute setting. There may be a “trade-off” between sensitivity and specificity of applying higher URL for these cardiac biomarkers. Further, more nuanced thresholds may be a step towards more personalized medicine, where a “one size fits all” strategy for diagnostic testing may not be appropriate. Future studies should investigate whether more personalized approaches (including demographics, kidney function and baseline values in determining clinical thresholds) would improve the accuracy of cardiac biomarker interpretation in the acute care setting.

There were some notable strengths of our study. Our study included a large, multi-center, well characterized population of ambulatory CKD patients. NT-proBNP and hsTnT were measured in batch using rigorous methods and contemporary instruments. The study population was large, diverse and had a wide range of eGFR. There were some limitations as well. The population included asymptomatic ambulatory patients; therefore we could not test the accuracy of various upper reference limits of NT-proBNP and hsTnT in the acute care setting to identify ACS and AHF. These finding apply to CKD patients similar to that in the CRIC study, which had few patients with glomerular disease and excluded patients with polycystic kidney disease. For certain strata, the sample size was limited which may affect the stability of the estimates and limit generalizability. Although we attempted to select a sample of CKD patients free of cardiac diseases by excluding those with known cardiovascular disease and subclinical cardiovascular disease by echocardiography, participants may still have had underlying unrecognized cardiovascular disease. The population was composed of research volunteers recruited largely from CKD clinics, such that the results may not be generalizable to all CKD patients in the U.S.

In conclusion, 40–88% of ambulatory CKD patients without clinical cardiovascular disease have concentrations of NT-proBNP and hsTnT above current upper reference limits. Developing individualized upper reference limits for these commonly used cardiac biomarkers may guide identification and treatment of acute cardiovascular disease in patients with CKD.

Supplementary Material

1

Support:

This study was supported by R01 DK103612 (to Dr Bansal). Roche Diagnostics provided partial funding for the NT-proBNP and hsTnT assays. Funding for the CRIC Study was obtained under a cooperative agreement from National Institute of Diabetes and Digestive and Kidney Diseases (U01DK060990, U01DK060984, U01DK061022, U01DK061021, U01DK061028, U01DK060980, U01DK060963, U01DK060902 and U24DK060990). In addition, this work was supported in part by: the Perelman School of Medicine at the University of Pennsylvania Clinical and Translational Science Award NIH/NCATS UL1TR000003, Johns Hopkins University UL1 TR-000424, University of Maryland GCRC M01 RR-16500, Clinical and Translational Science Collaborative of Cleveland, UL1TR000439 from the National Center for Advancing Translational Sciences (NCATS) component of the National Institutes of Health and NIH roadmap for Medical Research, Michigan Institute for Clinical and Health Research (MICHR) UL1TR000433, University of Illinois at Chicago CTSA UL1RR029879, Tulane COBRE for Clinical and Translational Research in Cardiometabolic Diseases P20 GM109036, Kaiser Permanente NIH/NCRR UCSF-CTSI UL1 RR-024131, Department of Internal Medicine, University of New Mexico School of Medicine Albuquerque, NM R01DK119199 . The funders had no role in study design, data collection, analysis, reporting, or the decision to submit for publication.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Financial Disclosure: The authors declare that they have no relevant financial interests.

REFERENCES

  • 1.Morrison LK, Harrison A, Krishnaswamy P, Kazanegra R, Clopton P, Maisel A. Utility of a rapid B-natriuretic peptide assay in differentiating congestive heart failure from lung disease in patients presenting with dyspnea. Journal of the American College of Cardiology. 2002;39(2): 202–209. [DOI] [PubMed] [Google Scholar]
  • 2.Holzmann MJ. Acute myocardial infarction can be ruled out with a single high-sensitivity cardiac troponin T level. Evidence Based Medicine. 2017;22(6): 226–226. [DOI] [PubMed] [Google Scholar]
  • 3.Katus HA, Remppis A, Neumann FJ, et al. Diagnostic efficiency of troponin T measurements in acute myocardial infarction. Circulation. 1991;83(3): 902–912. [DOI] [PubMed] [Google Scholar]
  • 4.McCullough PA, Nowak RM, McCord J, et al. B-type natriuretic peptide and clinical judgment in emergency diagnosis of heart failure: analysis from Breathing Not Properly (BNP) Multinational Study. Circulation. 2002;106(4): 416–422. [DOI] [PubMed] [Google Scholar]
  • 5.Boeddinghaus J, Twerenbold R, Nestelberger T, et al. Clinical Validation of a Novel High-Sensitivity Cardiac Troponin I Assay for Early Diagnosis of Acute Myocardial Infarction. Clinical chemistry. 2018;64(9): 1347–1360. [DOI] [PubMed] [Google Scholar]
  • 6.Reichlin T, Hochholzer W, Bassetti S, et al. Early diagnosis of myocardial infarction with sensitive cardiac troponin assays. The New England journal of medicine. 2009;361(9): 858–867. [DOI] [PubMed] [Google Scholar]
  • 7.Reiter M, Twerenbold R, Reichlin T, et al. Early diagnosis of acute myocardial infarction in patients with pre-existing coronary artery disease using more sensitive cardiac troponin assays. European heart journal. 2012;33(8): 988–997. [DOI] [PubMed] [Google Scholar]
  • 8.Go AS, Chertow GM, Fan D, McCulloch CE, Hsu CY. Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. The New England journal of medicine. 2004;351(13): 1296–1305. [DOI] [PubMed] [Google Scholar]
  • 9.Bansal N, Katz R, Robinson-Cohen C, et al. Absolute Rates of Heart Failure, Coronary Heart Disease, and Stroke in Chronic Kidney Disease: An Analysis of 3 Community-Based Cohort Studies. JAMA cardiology. 2017;2(3): 314–318. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Bansal N, Zelnick L, Go A, et al. Cardiac Biomarkers and Risk of Incident Heart Failure in Chronic Kidney Disease: The CRIC (Chronic Renal Insufficiency Cohort) Study. Journal of the American Heart Association. 2019;8(21): e012336. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Lamprea-Montealegre JA, Zelnick LR, Shlipak MG, et al. Cardiac Biomarkers and Risk of Atrial Fibrillation in Chronic Kidney Disease: The CRIC Study. Journal of the American Heart Association. 2019;8(15): e012200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Tuegel C, Katz R, Alam M, et al. GDF-15, Galectin 3, Soluble ST2, and Risk of Mortality and Cardiovascular Events in CKD. American journal of kidney diseases : the official journal of the National Kidney Foundation. 2018;72(4): 519–528. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Landray MJ, Emberson JR, Blackwell L, et al. Prediction of ESRD and death among people with CKD: the Chronic Renal Impairment in Birmingham (CRIB) prospective cohort study. American journal of kidney diseases : the official journal of the National Kidney Foundation. 2010;56(6): 1082–1094. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Michos ED, Wilson LM, Yeh HC, et al. Prognostic value of cardiac troponin in patients with chronic kidney disease without suspected acute coronary syndrome: a systematic review and meta-analysis. Annals of internal medicine. 2014;161(7): 491–501. [DOI] [PubMed] [Google Scholar]
  • 15.Bruch C, Fischer C, Sindermann J, Stypmann J, Breithardt G, Gradaus R. Comparison of the prognostic usefulness of N-terminal pro-brain natriuretic Peptide in patients with heart failure with versus without chronic kidney disease. The American journal of cardiology. 2008;102(4): 469–474. [DOI] [PubMed] [Google Scholar]
  • 16.Colbert G, Jain N, de Lemos JA, Hedayati SS. Utility of traditional circulating and imaging-based cardiac biomarkers in patients with predialysis CKD. Clinical journal of the American Society of Nephrology : CJASN. 2015;10(3): 515–529. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Desai AS, Toto R, Jarolim P, et al. Association between cardiac biomarkers and the development of ESRD in patients with type 2 diabetes mellitus, anemia, and CKD. American journal of kidney diseases : the official journal of the National Kidney Foundation. 2011;58(5): 717–728. [DOI] [PubMed] [Google Scholar]
  • 18.Vickery S, Webb MC, Price CP, John RI, Abbas NA, Lamb EJ. Prognostic value of cardiac biomarkers for death in a non-dialysis chronic kidney disease population. Nephrology, dialysis, transplantation : official publication of the European Dialysis and Transplant Association - European Renal Association. 2008;23(11): 3546–3553. [DOI] [PubMed] [Google Scholar]
  • 19.deFilippi C, Seliger SL, Kelley W, et al. Interpreting cardiac troponin results from high-sensitivity assays in chronic kidney disease without acute coronary syndrome. Clinical chemistry. 2012;58(9): 1342–1351. [DOI] [PubMed] [Google Scholar]
  • 20.deFilippi CR, Seliger SL, Maynard S, Christenson RH. Impact of renal disease on natriuretic peptide testing for diagnosing decompensated heart failure and predicting mortality. Clinical chemistry. 2007;53(8): 1511–1519. [DOI] [PubMed] [Google Scholar]
  • 21.End C, Seliger SL, deFilippi CR. Interpreting cardiac troponin results from highly sensitive assays in patients with chronic kidney disease: acute coronary syndromes and beyond. Coronary artery disease. 2013;24(8): 720–723. [DOI] [PubMed] [Google Scholar]
  • 22.Parikh RH, Seliger SL, deFilippi CR. Use and interpretation of high sensitivity cardiac troponins in patients with chronic kidney disease with and without acute myocardial infarction. Clinical biochemistry. 2015;48(4–5): 247–253. [DOI] [PubMed] [Google Scholar]
  • 23. https://diagnostics.roche.com/us/en/products/params/elecsys-troponin-t-high-sensitive-tnt-hs.html.
  • 24.Dickstein K, Cohen-Solal A, Filippatos G, et al. ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure 2008: the Task Force for the Diagnosis and Treatment of Acute and Chronic Heart Failure 2008 of the European Society of Cardiology. Developed in collaboration with the Heart Failure Association of the ESC (HFA) and endorsed by the European Society of Intensive Care Medicine (ESICM). European heart journal. 2008;29(19): 2388–2442. [DOI] [PubMed] [Google Scholar]
  • 25.Feldman HI, Appel LJ, Chertow GM, et al. The Chronic Renal Insufficiency Cohort (CRIC) Study: Design and Methods. Journal of the American Society of Nephrology : JASN. 2003;14(7 Suppl 2): S148–153. [DOI] [PubMed] [Google Scholar]
  • 26.Lash JP, Go AS, Appel LJ, et al. Chronic Renal Insufficiency Cohort (CRIC) Study: baseline characteristics and associations with kidney function. Clinical journal of the American Society of Nephrology : CJASN. 2009;4(8): 1302–1311. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Bansal N, Zelnick L, Shlipak MG, et al. Cardiac and Stress Biomarkers and Chronic Kidney Disease Progression: The CRIC Study. Clinical chemistry. 2019;65(11): 1448–1457. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Linnet K Necessary sample size for method comparison studies based on regression analysis. Clinical chemistry. 1999;45(6 Pt 1): 882–894. [PubMed] [Google Scholar]
  • 29.Joffe M, Hsu CY, Feldman HI, Weir M, Landis JR, Hamm LL. Variability of creatinine measurements in clinical laboratories: results from the CRIC study. American journal of nephrology. 2010;31(5): 426–434. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Levey AS, Coresh J, Greene T, et al. Expressing the Modification of Diet in Renal Disease Study equation for estimating glomerular filtration rate with standardized serum creatinine values. Clinical chemistry. 2007;53(4): 766–772. [DOI] [PubMed] [Google Scholar]
  • 31.Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate. Annals of internal medicine. 2009;150(9): 604–612. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Bansal N, Hyre Anderson A, Yang W, et al. High-sensitivity troponin T and N-terminal pro-B-type natriuretic peptide (NT-proBNP) and risk of incident heart failure in patients with CKD: the Chronic Renal Insufficiency Cohort (CRIC) Study. Journal of the American Society of Nephrology : JASN. 2015;26(4): 946–956. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Abbas NA, John RI, Webb MC, et al. Cardiac troponins and renal function in nondialysis patients with chronic kidney disease. Clinical chemistry. 2005;51(11): 2059–2066. [DOI] [PubMed] [Google Scholar]
  • 34.Acharji S, Baber U, Mehran R, et al. Prognostic significance of elevated baseline troponin in patients with acute coronary syndromes and chronic kidney disease treated with different antithrombotic regimens: a substudy from the ACUITY trial. Circulation. Cardiovascular interventions 2012;5(2): 157–165. [DOI] [PubMed] [Google Scholar]
  • 35.McCullough PA, Duc P, Omland T, et al. B-type natriuretic peptide and renal function in the diagnosis of heart failure: an analysis from the Breathing Not Properly Multinational Study. American journal of kidney diseases : the official journal of the National Kidney Foundation. 2003;41(3): 571–579. [DOI] [PubMed] [Google Scholar]
  • 36.Stacy SR, Suarez-Cuervo C, Berger Z, et al. Role of troponin in patients with chronic kidney disease and suspected acute coronary syndrome: a systematic review. Annals of internal medicine. 2014; 161(7): 502–512. [DOI] [PubMed] [Google Scholar]
  • 37.Nishimura M, Brann A, Chang KW, Maisel AS. The Confounding Effects of Non-cardiac Pathologies on the Interpretation of Cardiac Biomarkers. Current heart failure reports. 2018;15(4): 239–249. [DOI] [PubMed] [Google Scholar]
  • 38.Palazzuoli A, Masson S, Ronco C, Maisel A. Clinical relevance of biomarkers in heart failure and cardiorenal syndrome: the role of natriuretic peptides and troponin. Heart failure reviews. 2014;19(2): 267–284. [DOI] [PubMed] [Google Scholar]
  • 39.van Kimmenade RR, Januzzi JL Jr., Bakker JA, et al. Renal clearance of B-type natriuretic peptide and amino terminal pro-B-type natriuretic peptide a mechanistic study in hypertensive subjects. Journal of the American College of Cardiology. 2009;53(10): 884–890. [DOI] [PubMed] [Google Scholar]
  • 40.Friden V, Starnberg K, Muslimovic A, et al. Clearance of cardiac troponin T with and without kidney function. Clinical biochemistry. 2017;50(9): 468–474. [DOI] [PubMed] [Google Scholar]
  • 41.Neeland IJ, Drazner MH, Berry JD, et al. Biomarkers of chronic cardiac injury and hemodynamic stress identify a malignant phenotype of left ventricular hypertrophy in the general population. Journal of the American College of Cardiology. 2013;61(2): 187–195. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Gore MO, Seliger SL, Defilippi CR, et al. Age- and sex-dependent upper reference limits for the high-sensitivity cardiac troponin T assay. Journal of the American College of Cardiology. 2014;63(14): 1441–1448. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Miller-Hodges E, Anand A, Shah ASV, et al. High-Sensitivity Cardiac Troponin and the Risk Stratification of Patients With Renal Impairment Presenting With Suspected Acute Coronary Syndrome. Circulation. 2018;137(5): 425–435. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Fabbian F, De Giorgi A, Pala M, Tiseo R, Portaluppi F. Elevated NT-proBNP levels should be interpreted in elderly patients presenting with dyspnea. European journal of internal medicine. 2011;22(1): 108–111. [DOI] [PubMed] [Google Scholar]
  • 45.Anwaruddin S, Lloyd-Jones DM, Baggish A, et al. Renal function, congestive heart failure, and amino-terminal pro-brain natriuretic peptide measurement: results from the ProBNP Investigation of Dyspnea in the Emergency Department (PRIDE) Study. Journal of the American College of Cardiology. 2006;47(1): 91–97. [DOI] [PubMed] [Google Scholar]
  • 46.Christ M, Laule-Kilian K, Hochholzer W, et al. Gender-specific risk stratification with B-type natriuretic peptide levels in patients with acute dyspnea: insights from the B-type natriuretic peptide for acute shortness of breath evaluation study. Journal of the American College of Cardiology. 2006;48(9): 1808–1812. [DOI] [PubMed] [Google Scholar]
  • 47.Knudsen CW, Clopton P, Westheim A, et al. Predictors of elevated B-type natriuretic peptide concentrations in dyspneic patients without heart failure: an analysis from the breathing not properly multinational study. Annals of emergency medicine. 2005;45(6): 573–580. [DOI] [PubMed] [Google Scholar]
  • 48.Maisel AS, Clopton P, Krishnaswamy P, et al. Impact of age, race, and sex on the ability of B-type natriuretic peptide to aid in the emergency diagnosis of heart failure: results from the Breathing Not Properly (BNP) multinational study. American heart journal. 2004;147(6): 1078–1084. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1725815-supplement-1.pdf (664KB, pdf)

RESOURCES