Utility of Traditional Circulating and Imaging-Based Cardiac Biomarkers in Patients with Predialysis CKD

Abstract

Cardiac biomarkers, such as cardiac troponin T (cTnT), brain natriuretic peptide (BNP), and N-terminal-pro-BNP (NT-pro-BNP), are commonly used to diagnose acute coronary syndrome and congestive heart failure exacerbation in symptomatic patients. Levels of these biomarkers are frequently chronically elevated in asymptomatic patients with ESRD who are receiving maintenance dialysis. Other imaging biomarkers commonly encountered in nephrologists’ clinical practice, such as coronary artery calcium measured by computed tomography, left ventricular hypertrophy, and carotid intima-media thickness, are also frequently abnormal in asymptomatic patients with ESRD. This article critically reviews the limited observational data on associations between cTnT, BNP, NT-pro-BNP, coronary artery calcium, left ventricular hypertrophy, and carotid intima-media thickness with cardiovascular events and death in non–dialysis-dependent patients with CKD. Although sufficient evidence suggests that these biomarkers may be used for prognostication, the diagnostic utility of cTnT, BNP, and NT-pro-BNP remain challenging in patients with CKD. Decreased renal clearance may affect the plasma levels of these biomarkers, and upper reference limits were originally derived in patients without CKD. Until better data are available, higher cutoffs, or a rise in level compared with previous values, have been proposed to help distinguish acute myocardial infarction from chronic elevations of cTnT in symptomatic patients with CKD. Additionally, it is not known whether these biomarkers are modifiable and amenable to interventions that could change hard clinical outcomes in patients with CKD not yet undergoing long-term dialysis.

Introduction

Cardiovascular disease (CVD) is the leading cause of death in patients with CKD and ESRD, accounting for up to 50% of all deaths (1). Cardiac biomarkers, such as cardiac troponin T (cTnT) and I (cTnI), brain natriuretic peptide (BNP), and N-terminal-pro-BNP (NT-pro-BNP), are commonly used for diagnosing acute myocardial infarction (AMI) and congestive heart failure (CHF) exacerbation. However, chronic elevations of cTnT are observed in 80%–90% of asymptomatic patients with advanced CKD and ESRD (2). cTnT has evolved into an important prognostic factor in dialysis-dependent patients with ESRD, as elevated levels are associated independently with adverse cardiovascular (CV) outcomes (3). Fewer data describe an association between elevated troponins and CVD in patients with non–dialysis-dependent CKD. Other commonly used circulating and imaging-based cardiac biomarkers are also associated with poor CV outcomes in asymptomatic patients with ESRD, but such associations are less clearly established in CKD.

The first aim of this review is to summarize studies that reported associations between traditional cardiac biomarkers, such as cTnT, BNP, NT-pro-BNP, left ventricular mass index (LVMI), coronary artery calcium (CAC) scores, carotid intima-media thickness (cIMT), and clinical outcomes in patients with CKD not yet undergoing maintenance dialysis in an attempt to highlight strengths and limitations of existing data for prognostication. The second aim is to review data that support the utility of these biomarkers for diagnostic purposes in the acute setting. These specific biomarkers were chosen because they are noninvasive tests commonly used in clinical practice. For each biomarker, a general description is given, followed by discussion of levels in CKD, association with outcomes, and, finally, the clinical utility in patients with CKD. Knowledge gaps are identified and areas for future research suggested.

Cardiac Troponin Levels in CKD

Both cTnI and cTnT are biomarkers of cardiac injury that can be measured with standard assays and high-sensitivity (hs) assays, which detect levels about 10-fold lower than the standard assay. However, the upper reference limits for cardiac troponins were originally derived in persons without CKD, and these biomarkers are elevated in up to 80% of patients with asymptomatic CKD and ESRD (2). Troponin elevation in this context does not necessarily indicate acute ischemia from coronary atherosclerosis but may be due to decreased renal clearance or chronic myocardial injury. The mechanisms for this are multifactorial and include myocardial strain from altered hemodynamics, inflammation, endothelial dysfunction, and subendocardial ischemia (3,4) (Figure 1). The effect of renal clearance on circulating troponin concentrations is uncertain (3). Previous literature suggested that cTnT levels, compared with cTnI levels, are more commonly elevated in asymptomatic patients with ESRD (5). Plausible mechanisms for differential elevations include adsorption of cTnI on the dialyzer membrane imparting increased clearance, degradation of the labile cTnI molecule, advanced glycosylation of cTnT imparting decreased clearance, or uremic toxins causing conformational changes in the epitope region and altering the interaction with the assay antibodies (3). Previous clinical data were heavily influenced by differing sensitivities of the cTnT and cTnI assays and are not relevant to contemporary clinical practice. Consensus guidelines, therefore, do not specify a preference for use of cTnI over cTnT in patients with CKD (4). cTnT and cTnI provide largely identical information, and selection between them is typically influenced by laboratory equipment and vendor selection. Unlike the cTnT assay produced by a single manufacturer, cTnI assays are produced by multiple manufacturers using different antibody pairs, and assays are not interchangeable across institutions and studies (6). We therefore chose to focus the following discussion on cTnT.

Figure 1.

Cardiac troponins in patients with CKD. In response to injury, including altered hemodynamics, inflammation, and endothelial dysfunction, cardiac myocytes release troponin T (cTnT) and troponin I (cTnI) into the circulation from the troponin I-troponin T-troponin C (troponin I-T-C) complex present on the thin filament of the contractile apparatus. Serum cTnI and cTnT levels are measured by electrochemiluminescence immunoassay (ECLIA). Compared with cTnI, cTnT assays are standardized. In addition, several reasons explain why serum cTnT, compared with cTnI levels, are elevated in asymptomatic patients with ESRD. In asymptomatic patients with CKD, cardiac troponin levels are associated with various surrogate and hard clinical outcomes.

Higher cutoffs than used in persons without CKD for the diagnosis of AMI were suggested in patients with CKD and ESRD. A cTnT cutoff of 350 ng/L (>10-fold higher than the recommended cutoff for general use) had the best sensitivity (95%) and specificity (97%) for AMI in 284 patients with ESRD presenting with chest pain (7). In 89 patients with asymptomatic CKD stages 3–5, the 95th percentile for hs-cTnT was 139 ng/L, >10-fold higher than that derived in the general population (8), with levels increasing across higher CKD stages. Another study reported that the specificity of a cutoff of >14.0 ng/L, as recommended for diagnosis of AMI in the general population, was much lower in those with an eGFR of ≤60 ml/min per 1.73 m² (54%) versus >60 ml/min per 1.73 m² (87%) (9). A higher cutoff of >43.2 ng/L had a much higher specificity (88%) in those with an eGFR≤60 ml/min per 1.73 m². In addition to higher cutoffs, a rise in troponins compared with previous chronically elevated values, or a rise and/or fall using serial measurements, has been proposed to help distinguish AMI from chronic elevations of cTnT in patients with advanced CKD or ESRD (4,10,11).

Cardiac Troponins and Surrogate Outcomes

cTnT levels were associated with CV events and all-cause mortality in asymptomatic patients with ESRD (12). Although fewer data extend similar associations to patients with nondialysis CKD, several studies reported correlations between cTnT or hs-troponin T (hs-cTnT) with surrogate and hard outcomes (Table 1) (2,13–20). Cross-sectional studies revealed an association between higher cTnT and decreased eGFR, as well as measures of left ventricular hypertrophy (LVH). An analysis of the Chronic Renal Insufficiency Cohort (CRIC) reported detectable hs-cTnT (≥3 ng/L) among 81% (Table 1) (2). hs-cTnT was associated with higher LVMI across all LVMI categories, and lower ejection fraction, mainly in the lowest category (≤35%) (2). In another cross-sectional report from CRIC, the highest quartile of hs-cTnT (>24 ng/L) compared with undetectable levels was associated with the presence of LVH and left ventricular systolic dysfunction (LVSD) (21). Among asymptomatic outpatients with CAD and eGFR<60 ml/min per 1.73 m², elevated hs-cTnT was associated with higher LVMI, lower creatinine- and cystatin-based eGFRs, and higher urine albumin-to-creatinine ratio (UACR) (Table 1) (17). Correlations between cTnT and eGFR were confirmed in British outpatients with atherosclerotic renovascular disease (16). A Japanese study of nondiabetic patients with CKD reported that those with echocardiographic evidence of left ventricular diastolic dysfunction (LVDD) had a significantly higher hs-cTnT level than those without (19).

Table 1.

Studies reporting associations of cardiac troponin T and high-sensitivity cardiac troponin T with outcomes in CKD

Study (Reference)	Patients (n)	Study Design	Sample	Outcomes
Dubin et al. (2)	2464 CRIC	Cross-sectional	Asymptomatic outpatients, eGFR 20–70 ml/min per 1.73 m2	hs-cTnT independently associated with lower eGFR; aOR, 2.83 (95% CI, 2.41 to 3.33) for eGFR<30 ml/min per 1.73 m2 versus >60 ml/min per 1.73 m2, higher LVMI, and lower LVEF
Mishra et al. (21)	3243 CRIC	Cross-sectional	Asymptomatic outpatients with eGFR<60 ml/min per 1.73 m2 and CAD	hs-cTnT>24 pg/ml versus undetectable independently associated with LVH (aOR, 2.43 [95% CI, 1.44 to 4.09]) and LVSD (aOR, 1.4 [95% CI, 1.2 to 1.7]), but not LVDD
deFilippi et al. (17)	148 (50% diabetic)	Cross-sectional	Outpatients from United States with known CAD and eGFR<60 ml/min per 1.73 m2	hs-cTnT independently associated with LVMI, decreased GFR, and increased UACR
Kitagawa et al. (19)	93 (nondiabetic)	Cross-sectional	Japanese inpatients with nondiabetic CKD stages 1–5	hs-cTnT≥9 pg/ml and BNP≥20 pg/ml were best cutoffs for severe LVDD
Abbas et al. (13)	222	Longitudinal	Asymptomatic British outpatients with CKD stages 3–5	Detectable versus undetectable cTnT conferred increased risk for all-cause mortality (uOR, 3.47 [95% CI, 1.27 to 10.39]) (n=23)
Goicoechea et al. (14)	176	Longitudinal	Asymptomatic Spanish outpatients; n=128 with creatinine clearance <60 ml/min	Detectable versus undetectable cTnT increased hazard of CV event (uHR, 12.3 [95% CI, 4.91 to 31.02]) (n=21)
Chrysochou et al. (16)	82	Longitudinal	Asymptomatic British outpatients with ARVD at single center	cTnT independently associated with all-cause mortality (uHR, 3.9 [95% CI, 1.8 to 8.5])
Scheven et al. (15)	8121 PREVEND	Longitudinal	Asymptomatic Dutch outpatients; 18% with CKD (UACR>30 mg/g or eGFR<60 ml/min per 1.73 m2)	hs-cTnT independently associated with CV events (adjusted for eGFR, albuminuria, and CV risk factors) (aHR, 1.18 [95% CI not given; P=0.03])
Hasegawa et al. (18)	442	Longitudinal	Asymptomatic Japanese outpatients with eGFR<60 ml/min per 1.73 m2	hs-cTnT≥33 versus ≤9 pg/ml conferred increased risk for CV events (aHR, 6.18 [95% CI, 1.38 to 27.7]) (n=63)

CRIC, Chronic Renal Insufficiency Cohort; PREVEND, Prevention of Renal and Vascular End Stage Diseases; CAD, coronary artery disease; ARVD, atheromatous renovascular disease; UACR, urinary albumin-to-creatinine ratio; hs-cTnT, highly sensitive cardiac troponin T; aOR, adjusted odds ratio; 95% CI, 95% confidence interval; LVMI, left ventricular mass index; LVEF, left ventricular ejection fraction; LVH, left ventricular hypertrophy; LVSD, left ventricular systolic dysfunction; LVDD, left ventricular diastolic dysfunction; BNP, brain natriuretic peptide; uOR, unadjusted odds ratio; CV, cardiovascular; uHR, unadjusted hazard ratio; aHR, adjusted hazard ratio.

Cardiac Troponins and Hard Outcomes

Limited prospective data are available regarding the association of cTnT with CV outcomes in nondialysis patients with CKD. In a British study, cTnT was detectable (≥10 ng/L) in 43% of asymptomatic patients with CKD stages 3–5 (13). Detectable cTnT was associated with increased all-cause mortality at 19 months (Table 1) (13). Similar results for the association of cTnT with increased CV events were reported by Goicoechea et al. in Spanish patients with creatinine clearance <60 ml/min (Table 1) (14). Given low event rates, however, these studies were limited by lack of multivariable analysis and adjustment for confounders (13,14,16). More recently, reports from larger cohorts showed an independent association between hs-cTnT and CV events among patients with CKD in adjusted analyses (Table 1) (15,18).

Clinical Utility of Cardiac Troponins in CKD

In summary, because troponin upper reference limits were originally derived in non-CKD samples, knowledge gaps exist in establishing consensus regarding appropriate diagnostic cutoff values in patients with CKD, as well as the required magnitude of the threshold of change in serial values. The updated consensus definition of AMI requires a rise and/or fall in serial levels, with at least one value above the 99th percentile of the upper reference limit, in addition to appropriate electrocardiographic changes, imaging consistent with myocardial damage, or new regional wall abnormalities (4). However, it does not specify different thresholds for defining AMI in patients with CKD. Nonetheless, it seems reasonable to consider higher threshold values in patients with CKD or rely more heavily on assessment of serial changes to confirm AMI diagnosis. There are no recommendations to support a specific threshold of change in patients with CKD, although recent data in 19 patients with ESRD support the use of a ≥20% change for hs-cTnT (10,11) a value that exceeds analytical variation alone (6). For prognostic purposes, it appears that detectable compared with undetectable troponins portend higher risk for future death and CV events. Future research needs to ascertain whether further work-up or intervention is warranted when clinicians find a detectable troponin in asymptomatic patients with CKD.

BNP and NT-pro-BNP in CKD

NT-pro-BNP and BNP are commonly tested in symptomatic patients suspected of acute CHF exacerbation. In one study, they were elevated in 56% of asymptomatic patients with CKD (22). Pre-pro-BNP is synthesized within the cardiac myocytes in response to ventricular wall stress and stretch (23). After removal of a signaling peptide within the cytosol, pro-BNP is further cleaved into the inactive form (NT-pro-BNP) and the active hormone (BNP) at the time of release from the myocyte or in the circulation (Figure 2). NT-pro-BNP is more stable, with a longer half-life, and may be a better biomarker for chronic volume expansion or stress than is BNP (23). Reduced renal function decreases the fractional plasma clearance of both BNP and NT-pro-BNP, and studies reported correlations between graded elevations in these peptides and declining eGFR or advancing CKD stages (Table 2) (22,24–32). The clearance of NT-pro-BNP is predominantly renal, while BNP is also degraded systemically (Figure 2) (23). This may explain the observed correlation of reduced eGFR to a greater extent with NT-pro-BNP than with BNP (23,24), and the increased ratio of NT-pro-BNP/BNP with advancing CKD stages (30), a finding not borne out by all studies. One study reported an equal dependence on renal clearances for both peptides, although most participants had a GFR≥30 ml/min per 1.73 m² (33), suggesting that clearance may be similar for both until renal function deteriorates to advanced stages.

Figure 2.

Brain natriuretic peptides in patients with CKD. In response to increased stretch or tension, left ventricular myocytes release brain natriuretic peptide (BNP) and N-terminal-pro-BNP (NT-pro-BNP) from precursors. BNP is an active molecule with a short plasma half-life and is degraded in the circulation by enzymatic action. NT-pro-BNP is the inactive form of BNP, with a longer half-life. It is primarily cleared by the kidneys. Reduced eGFR correlates to a greater extent with elevated plasma NT-pro-BNP than to BNP levels. Increased NT-pro-BNP/BNP ratio correlates with advancing CKD stages, especially if the eGFR is <30 ml/min per 1.73 m². However, both BNP and NT-pro-BNP are associated with surrogate and hard clinical outcomes in asymptomatic patients with CKD.

Table 2.

Studies reporting associations of brain natriuretic protein and N-terminal B-type natriuretic peptide with outcomes in CKD

Study (Reference)	Patients (n)	Study Design	Sample	Outcomes
Yasuda et al. (29)	485	Longitudinal	Japanese patients from single center with CKD stages 3–5	Elevated BNP>86 pg/ml associated with doubling of creatinine or ESRD
Spanaus et al. (30)	177	Longitudinal	Central European nondiabetic outpatients with CKD	NT-pro-BNP, but not BNP, associated with doubling of creatinine or ESRD
Vickery et al. (24)	213	Cross-sectional	Predialysis CKD stages 3–5	LVMI positively correlated with BNP and NT-pro-BNP (r=0.34 and 0.39)
Lee et al. (31)	256	Cross-sectional	Korean predialysis CKD, outpatients or from emergency department	NT-pro-BNP associated with LVSD independent of eGFR; higher NT-pro-BNP cutoffs for worse CKD stages: 2165 pg/ml in CKD 3, 4740 pg/ml in CKD 4, and 15,892 pg/ml in CKD 5
Yang et al. (32)	207	Cross-sectional	Chinese outpatients with hypertension	LVDD and NT-pro-BNP associated with worse eGFR and UACR; LVDD positively correlated with Log-NT-pro-BNP (r=0.54)
Mishra et al. (20)	3232 CRIC	Cross-sectional	Patients with CKD and without heart failure	Highest versus lowest quartile of NT-pro-BNP associated with LVH and LVSD (aOR, 2.7 [95% CI, 1.8 to 4.0] and 2.7 [95% CI, 1.7 to 4.5])
Khan et al. (34)	54	Cross-sectional	Outpatients with predialysis CKD	Both NT-pro-BNP and BNP correlated with LVH and CAD
Bruch et al. (27)	142	Longitudinal	Stable German outpatients with CHF	Elevated NT-pro-BNP levels independently associated with NYHA class and inversely associated with eGFR and LVEF
Astor et al. (35)	994 AASK	Longitudinal	African Americans with eGFR 20–65 ml/min per 1.73m2 and hypertension	Elevated versus undetectable NT-pro-BNP associated with CV events (aHR, 4.0 [95% CI, 2.1 to 7.6])
Horii et al. (26)	1083	Longitudinal	Japanese patients with CKD stages 1–5 and CV disease on cardiac catheterization; ACS and acute CHF were excluded	BNP and NT-pro-BNP associated with death and CV composite; composite event AUC 0.720 for NT-pro-BNP and 0.666 for BNP; cutoffs for composite: CKD 1–3: BNP 91 pg/ml, NT-pro-BNP 260 pg/ml; CKD 4–5: BNP 157 pg/ml, NT-pro-BNP 5112 pg/ml
deFilippi et al. (22)	207	Cross-sectional	VA outpatients with predialysis CKD stages 1–5	NT-pro-BNP >490 pg/ml independently associated with prior CAD events (n=67; AUC of 0.69)
Fu et al. (36)	999	Longitudinal	Chinese patients with CAD age >60 yr; 358 with CKD (eGFR<60 ml/min per 1.73 m2)	NT-pro-BNP associated with all-cause death if eGFR<60 ml/min per 1.73 m2 (aHR, 1.54 [95% CI, 1.32 to 1.80])
Bruch et al. (37)	341	Longitudinal	German outpatients with stable CHF, 183 with CKD	Elevated NT-pro-BNP independently associated with CV events (including death) in patients with stable CHF with and without CKD; cutoff for both, 1474 pg/ml
Tarnow et al. (38)	386	Longitudinal	Danish outpatients with type 1 diabetes, 198 with diabetic nephropathy	Elevated NT-pro-BNP independently associated with death in patients with diabetic nephropathy (aRR, 2.49 [95% CI, 1.22 to 5.08])
Anwaruddin et al. (40)	599	Cross-sectional	Dyspneic patients suspected of having CHF presenting to urban emergency department, 207 with CKD (eGFR<60 ml/min per 1.73 m2)	Elevated NT-pro-BNP and eGFR inversely correlated; NT-pro-BNP independently associated with 60-day mortality in patients with and without CKD
Oterdoom et al. (39)	3840	Longitudinal	Dutch outpatients, 606 renal transplant recipients	NT-pro-BNP, decreased eGFR, and hypertension medication use associated with death in both renal transplant and nontransplant patients

AASK, African American Study of Kidney Disease and Hypertension; CHF. congestive heart failure; ACS, acute coronary syndrom; VA, Veterans Affairs; NT-pro-BNP, N-terminal B-type natriuretic peptide; 95% CI, 95% confidence interval; NYHA, New York Heart Association; aHR, adjusted hazard ratio; AUC, area under the curve; aRR, adjusted relative risk.

BNP and Surrogate Outcomes in CKD

Elevated levels of both BNP and NT-pro-BNP are associated with abnormal echocardiographic findings in patients with CKD (Table 2) (20,24,31,32,34). Among those with eGFR<60 ml/min per 1.73 m², LVMI positively correlated with BNP and NT-pro-BNP levels (24). NT-pro-BNP was independently associated with presence of LVSD in patients with CKD (31) (Table 2). Higher gradations in NT-pro-BNP cutoffs to detect LVSD were reported for increasing CKD stages (31). A Chinese study showed that LVDD positively correlated with Log-NT-pro-BNP (32) (Table 2). In the CRIC study, the highest compared with lowest quartile of NT-pro-BNP was associated with a 3-fold higher odds of LVH and LVSD (20) (Table 2).

BNP and Hard Outcomes in CKD

BNP and NT-pro-BNP are also associated with hard outcomes in CKD (Table 2). In a Japanese study, both BNP and NT-pro-BNP were associated with death and the composite of death and CV events. On the basis of the areas under the curve, the authors concluded that NT-pro-BNP may be a superior marker to BNP for composite events in patients with CKD stages 4 and 5 (versus stages 1–3), although a formal statistical test was not used to determine whether the curves significantly differed (26). Among the African American Study of Kidney Disease and Hypertension cohort, those with elevated NT-pro-BNP had a four times higher hazard for CV events than those with undetectable levels (Table 2) (35). The association was significantly stronger in those with than without proteinuria (interaction P=0.05) (35). In Chinese patients with known CAD, NT-pro-BNP was associated with all-cause death if eGFR was <60 ml/min per 1.73 m² (36). In addition, the NT-pro-BNP cutoff associated with mortality was higher in patients with CKD (2584 pg/ml) than in persons without CKD (370 pg/ml) (36). Several other studies reported similar associations between NT-pro-BNP, CV events, and all-cause death (Table 2) (37–39).

Clinical Utility of NT-pro-BNP and BNP in CKD

To summarize, NT-pro-BNP and BNP can be used for prognostication in patients with CKD because elevated levels are associated with both adverse surrogate and hard outcomes in this population. However, most studies included asymptomatic samples, and clinicians are still left with the important question of how to best interpret elevated BNP and NT-pro-BNP levels for acute CHF diagnosis in symptomatic patients. A study of patients presenting with dyspnea revealed that NT-pro-BNP may be a useful diagnostic test for CHF in patients with and without CKD, although the diagnostic cutoff was higher in those with eGFR<60 ml/min per 1.73 m² (>1200 pg/ml) than in those with eGFR≥60 ml/min per 1.73 m² (>450 pg/ml if age <50 years; >900 pg/ml if age ≥50 years) (40). More prospective, well controlled studies are needed to confirm these findings.

CAC in CKD

CAC as measured by computed tomography is a noninvasive measurement of the burden of coronary atherosclerosis. Patients with CKD have higher CAC scores compared with age-matched controls without CKD, and patients with CKD without baseline calcification exhibit higher incidence rates of developing future de novo CAC (41,42). Cross-sectional analyses have reported a graded relationship between lower eGFR and increasing CAC (41). These associations were attenuated after adjustment for traditional CV risk factors, such as diabetes, but remained statistically significant for patients with an eGFR<30 ml/min per 1.73 m² (42). It is not entirely clear whether a decline in eGFR plays a mechanistic role for developing de novo CAC and CAC progression. Interestingly, several analyses reported higher baseline CAC and CAC progression to be associated with eGFR decline and worsening proteinuria (43–45) (Table 3). A plausible explanation may be that the progression of CAC and CKD are collinear because of the presence of similar risk factors for both disease processes.

Table 3.

Studies reporting association of coronary artery calcium with outcomes in CKD

Study (Reference)	Patients (n)	Study Design	Sample	Outcomes
Chang et al. (43)	279	Longitudinal	Korean middle-aged outpatients with CKD, excluding those with CAD	CAC was associated with annual decrease in eGFR, even after adjustment for age, sex, baseline eGFR, albumin, and UPCR
Garland et al. (44)	125	Longitudinal	Consecutive Canadian outpatients from single center with CKD stages 3–5	Log-transformed CAC correlated with decline in eGFR of ≥5% (r=0.22); higher baseline CAC associated with higher odds of decline in eGFR ≥5% at 1 yr
Maahs et al. (45)	1066	Longitudinal	CACTI study cohort with asymptomatic CVD	Worse UACR and eGFR associated with CAC progression
Russo et al. (46)	341, 60 diabetic	Longitudinal	Single-center Italian inpatients and outpatients with CKD stages 2–5 and well controlled hypertension	CAC prevalence higher in patients with diabetes versus nondiabetic patients; patients with diabetes with CKD had higher annualized percentage increase in CAC and CV events
Russo et al. (50)	181	Longitudinal	Italian CKD stages 2–5 outpatients without symptomatic CVD	Compared with patients with baseline CAC score ≤100 AU, those with a score >100 AU had a higher hazard of CV death or MI (aHR, 4.11 [95% CI, 1.77 to 9.57]) (29 events)
Chiu et al. (47)	225, all diabetic	Longitudinal	Proteinuric patients with mean UPCR of 2.7 and eGFR of 52 ml/min per 1.73 m2; 70% Latino	Those with highest (compared with lowest) quartile baseline CAC had a higher hazard of all-cause death (aHR, 2.61 [95% CI, 1.23 to 5.54]) (54 events)
Nguyen et al. (51)	281, 42 diabetic	Longitudinal	Belgian kidney transplant recipients (98% white) from single center	Baseline CAC score was associated with CV composite (aHR, 1.40 [95% CI, 1.12 to 1.75]) (31 events, 8 CV deaths)

CAC, coronary artery calcium; CACTI, Coronary Artery Calcification in Type 1 Diabetes Mellitus; UPCR, urinary protein-to-creatinine ratio; AU, Agatston units; MI, myocardial infarction; aHR, adjusted hazard ratio; 95% CI, 95% confidence interval.

Both traditional and nontraditional CV risk factors are associated with the presence and severity of CAC in patients with CKD who are not undergoing dialysis. Traditional factors explored included advanced age, white race, male sex, higher body mass index, and diabetes mellitus, in particular (46–48). A retrospective study of patients with stages 2–5 CKD with well controlled BP reported higher prevalence of CAC in patients with diabetes than in nondiabetic patients (77% versus 33%) (Table 3) (46), and another study found faster progression in CAC among diabetic patients with CKD than in those without diabetes (48). We previously reported in a multiethnic, population-based asymptomatic cohort that three nontraditional risk factors—calcium-phosphorus product, homocysteine, and osteoprotegerin—were independently associated with high CAC scores, and diminished the magnitude of the association between the presence of CKD and elevated CAC, suggesting that they may play mechanistic roles in the development of CAC (49). Others reported similar associations between elevated serum phosphorus and CAC in patients with CKD (48).

CAC and Clinical Outcomes

Fewer data are available on unfavorable clinical implications of CAC in predialysis CKD versus ESRD samples. The few observational studies reporting associations of CAC with adverse outcomes are limited by low event rates, limited follow-up, or ethnic homogeneity (Table 3) (46,47,50,51). A study of a predominantly Latino diabetic cohort reported that those in the highest compared with the lowest quartile of baseline CAC had a higher hazard of all-cause mortality at 39 months (47). During a 25-month follow-up, there was four times higher risk of CV death or AMI among outpatients with stages 2–5 CKD and baseline CAC scores >100 Agatston units (AU) compared with ≤100 AU (50). Finally, in renal transplant recipients, CAC score assessed at the inception of the cohort was associated with the composite of CV death, AMI, stroke, transient ischemic attack, and revascularization at 2.3 years (Table 3) (51). However, models were overadjusted for the few events in the last two studies (50,51).

Clinical Utility of CAC in CKD

CAC is being used as a screening test to assess risk of future CV events in patients without CKD who are at intermediate CV risk because it may add to the prognostic utility of the Framingham Risk Score (52,53). Asymptomatic persons without CKD and without CAC have a very low risk of CV events, whereas those with scores >400 AU have elevated risk similar to that in patients with diabetes or peripheral vascular disease (54). Studies in patients without CKD reported a strong correlation between CAC and total atherosclerotic plaque burden at the individual level (r=0.90) (55). Although current guidelines do not recommend the routine use of CAC for risk stratification, they do recommend its use to inform treatment decision-making in patients without CKD if a risk-based treatment decision is uncertain after quantitative risk assessment using traditional CV risk factors (56). However, it is too early to recommend the standard use of CAC for risk stratification in patients with CKD because it remains unclear whether such calcific lesions in a coronary artery segment increase or decrease biomechanical stability of atherosclerotic plaques in CKD (57). Similarly, it is not known whether increased CAC or its progression truly plays a mechanistic role in the development of future CV events or is merely a surrogate for other CV risk factors in patients with CKD. Finally, there are not enough data to show that CAC is a modifiable risk factor in CKD. For example, it is not known whether the reduction of calcium or phosphate using various binders persistently influences regression of CAC in CKD or whether CAC regression translates to improved outcomes (58).

LV Mass or LV Dysfunction in CKD

LVH and abnormal LV function, based on echocardiographic parameters, are highly prevalent among patients with CKD who initiate dialysis. According to a Canadian cohort, 74% have LVH, 36% have LV dilation, and 15% have LVSD (59). Higher baseline LVMI is associated with severity of CKD as well as progression, but it is not clear whether this is independent of high BP.

In a cross-sectional study of diabetic patients, severity of CKD stage paralleled increases in LVMI and decreases in LV ejection fraction (LVEF) (Table 4) (60). Patients with CKD stages 3–5 and LVH had lower eGFR and greater proteinuria than patients without LVH, as well as a weak inverse correlation between LVMI and eGFR (61). However, in multivariable models that included systolic BP and body mass index, eGFR was not independently associated with LVH (61). Another cross-sectional study did report a correlation between urinary protein-to-creatinine ratio and LVMI, independent of systolic BP, although a similar correlation was not observed with eGFR (62). These studies were limited by lack of controls without CKD. Interestingly, there were higher LV mass and greater degree of LVDD, but no difference in LVEF, among patients with CKD compared with age- and sex-matched controls according to univariate analyses (63). However, pulse pressure was significantly higher in patients with CKD than in controls, which could account for the observed differences (63). Finally, three prospective longitudinal studies reported changes in LV geometry to independently correlate with eGFR decline and progression to ESRD (Table 4) (64–66).

Table 4.

Studies reporting associations of left ventricular mass and function with outcomes in CKD

Study (Reference)	Patients (n)	Study Design	Sample	Outcomes
Chen et al. (60)	285	Cross-sectional, all diabetic	Taiwanese outpatients	Stepwise increases in LVMI and decreases in LVEF corresponded to higher CKD stages
Nitta et al. (61)	1185	Cross-sectional, 41% diabetic	Japanese outpatients with CKD stages 3–5	Echocardiography-based LVMI correlated with eGFR (r=0.18); patients with LVH had lower GFR and more proteinuria compared with those without
McQuarrie et al. (62)	49	Cross-sectional	British outpatients with CKD stages 2–4	Log-PCR correlated with LVMI by cMRI (r=0.52); proteinuria explained 23% of LVMI variance
Chen et al. (64)	415	Longitudinal, 53% diabetic	Taiwanese outpatients with CKD stages 3–5	cLVH measured by echocardiography associated with progression to ESRD (aHR, 2.03 [95% CI, 1.00 to 4.10])
Chen et al. (65)	540	Longitudinal, 50% diabetic	Taiwanese outpatients with CKD stages 3–5	Those with higher uric acid and LVMI had higher hazard of progression to dialysis and higher odds of rapid decline in eGFR (aHR, 1.83 [95% CI, 1.01 to 3.33]; aOR, 2.23 [95% CI, 1.06 to 4.70])
Park et al. (66)	3866 MESA	Longitudinal, 11% diabetic	eGFR>60 ml/min per 1.73 m2	During a median follow-up of 4.8 yr, each SD higher LV concentricity was associated with a 9% and 8% decline in eGFRcr and eGFRcys
Silberberg et al. (59)	91	Longitudinal	Canadian patients from single center with incident ESRD	Those with highest versus lowest quintile of LVMI at baseline experienced higher hazards of all-cause mortality and CV mortality (aHR, 2.9 [95% CI, 1.3 to 6.9] and 2.7 [95% CI, 0.9 to 8.2])
Chen et al. (67)	505	Longitudinal, 56% diabetic	Taiwanese outpatients with CKD stages 3–5	Every g/m2 increase in LVMI and LVEF<55% versus ≥55% were associated with increased CV events (aHR, 1.006 [95% CI, 1.002 to 1.010] and 2.01 [95% CI, 1.01 to 3.74])

MESA, Multi-Ethnic Study of Atherosclerosis; cMRI, cardiac magnetic resonance imaging; cLVH, concentric LVH; aHR, adjusted hazard ratio; 95% CI, confidence interval; aOR, adjusted odds ratio; eGFR_cr, eGFR calculated using serum creatinine; eGFR_cys, eGFR calculated using cystatin C.

LV Mass, LV Dysfunction, and Clinical Outcomes

LVMI was independently associated with increased all-cause and CV mortality in patients initiating dialysis in a prospective study, even after adjustment for age, CAD, diabetes mellitus, and systolic BP (59). These findings were extended to outpatients with CKD stages 3–5, in whom higher LVMI and LVEF <55% versus ≥55% at baseline were associated with CV events, including death, AMI, sustained ventricular arrhythmia, hospitalization for unstable angina, congestive heart failure, transient ischemic attack, or stroke at 26 months (Table 4) (67).

Clinical Utility of LV Mass or LV Dysfunction in CKD

Although these data suggest that LVH is associated with CKD progression and CV events, elevated SBP and pulse pressure, which are highly prevalent in this patient population, may be major confounders in these analyses. In addition, lack of well controlled prospective studies limit the utility of echocardiographic parameters in predicting outcomes in clinical practice. Future studies need to analyze how changes in LV mass and function may be used to prognosticate hard clinical outcomes.

cIMT in CKD

The cIMT has become a frequently studied sonographic marker of early atherosclerotic changes in vessels. The thickening of the intima-media complex not only reflects a local vessel change in the carotid but could indicate a systemic change in all vessel walls. It may also predict future risk for CV events. The ease and safety of this imaging study allow its use as a potential new biomarker for systemic atherosclerosis in high-risk patient populations, such as predialysis patients with CKD. Several studies, mostly cross-sectional, suggested that cIMT measurements were elevated in CKD individuals, as reviewed later in this article (Table 5).

Table 5.

Studies reporting associations of carotid intima-media thickness with outcomes in CKD

Study (Reference)	Patients (n)	Study Design	Sample	Outcomesa
Zoungas et al. (68)	159 CKD, 159 control	Case-control	Australian outpatients with SCr≥0.40 mmol/L versus controls	Mean cIMT in CKD, 0.89±0.17 mm versus 0.73±0.13 mm in controls (P<0.05)
Lemos et al. (69)	122	Cross-sectional	Brazilian nondiabetic outpatients with CKD stages 2–5	cIMT, 0.62±0.19 (eGFR<60 ml/min per 1.73 m2) versus 0.53±0.10 mm (eGFR>60 ml/min per 1.73 m2) (P=0.03)
Tanaka et al. (70)	1003	Cross-sectional	Japanese outpatients with and without CKD, nondialysis	In multivariable regression analysis, lower eGFR correlated with mean maximum cIMT after adjustment for age and sex (r=−0.104; P<0.001)
Aggarwal et al. (71)	60	Longitudinal	Indian outpatients with CKD stages 1–5, including ESRD	cIMT in left CCA at baseline: 0.5±0.08 mm in CKD stages 1–2, 0.7±0.10 mm in CKD stages 3–4, and 0.8±0.16 mm in CKD stage 5 (P<0.001)
Zhou et al. (72)	227	Cross-sectional	Chinese outpatients with CKD, nondialysis	CKD stage associated with increased cIMT: CKD stages 1–2, 0.64±0.18 mm; CKD stage 3, 0.74±0.25 mm; CKD stage 4, 0.81±0.25 mm; CKD stage 5, 0.86±0.20 (P<0.01)
Marcos et al. (76)	117	Longitudinal	Brazilian outpatients with CKD stages 2–4	Patients with cIMT>0.6 mm had lower eGFR (P=0.01) ; no association with CV events or death
Szeto et al. (73)	203	Longitudinal	Chinese outpatients with CKD stages 3–4	Event-free survival from CV death, nonfatal MI or stroke, unstable angina admission, PCI, CHF, or TIA was 94.4%, 89.8%, 77.7%, and 65.9% for cIMT quartiles 1, 2, 3, and 4, respectively (log-rank test P<0.01)
Desbien et al. (74)	3364	Longitudinal	German outpatients with and without CKD	Each 1-ml/min per 1.73 m2 decrease in creatinine clearance (aHR, 1.04 [95% CI, 1.02 to 1.23]) or 0.1-mm increase in cIMT (aHR, 1.15 [95% CI, 1.11 to 1.93]) was associated with fatal and nonfatal vascular events
Kim et al. (75)	182	Longitudinal	Korean nondiabetic, asymptomatic outpatients with eGFR<60 ml/min per 1.73 m2	Carotid plaque associated with fatal or nonfatal ACS or stroke (OR, 7.80 [95% CI, 1.45 to 45.97]), but cIMT not significant in multivariable analysis
Adeseun et al. (77)	220 CRIC	Cross-sectional	CKD outpatients with GFR 20–70 ml/min per 1.73 m2	CAC, carotid plaque, or cIMT ability to discriminate prevalent CVD was not significant (c-statistics = 0.67, 0.64, and 0.61, respectively) (P>0.05)
Matsushita et al. (78)	6553 MESA	Longitudinal	Outpatients without CVD, 1284 with CKD	CAC (aHR, 1.69 [95% CI, 1.45 to 1.97]) performed better than cIMT (aHR, 1.12 [95% CI, 1.00 to 1.25]) for prediction of CHD and heart failure in CKD

MESA, Multi-Ethnic Study of Atherosclerosis; CRIC, Chronic Renal Insufficiency Cohort; SCr, serum creatinine; cIMT, carotid intima-media thickness; CCA, common carotid artery; PCI, percutaneous coronary intervention; TIA, transient ischemic attack; aHR, adjusted hazard ratio; 95% CI, 95% confidence interval; OR, odds ratio; aHR, adjusted hazard ratio; CHD, coronary heart disease.

^aValues expressed with a plus/minus sign are the mean±SD.

In a case-control study, case-patients with a serum creatinine ≥0.40 mmol/L had significantly higher cIMT than controls (Table 5) (68). Among patients with CKD stages 3–5, cIMT measurements were significantly higher if eGFR was <60 ml/min per 1.73 m² than >60 ml/min per 1.73 m² (69). Another study reported a weak but statistically significant correlation between lower eGFR and higher mean maximum wall thickness measured along 12 carotid segments, after adjustment for age and sex (70). Two studies revealed small but statistically significant stepwise increases in cIMT measurements with higher CKD stages (Table 5) (71,72).

cIMT and Clinical Outcomes

There are conflicting data on whether cIMT is associated with death or CV events in predialysis patients with CKD. A Chinese study of 203 patients with stages 3 or 4 CKD reported a statistically significant trend for higher adverse CV events for increasing cIMT quartiles (Table 5) (73). In a longitudinal study of 3364 outpatients with and without CKD, lower creatinine clearance and higher cIMT were associated with fatal and nonfatal vascular events (74). However, in a study of nondiabetic outpatients with eGFR<60 ml/min per 1.73 m², carotid plaque burden but not cIMT was associated with fatal or nonfatal acute coronary syndrome (ACS) or stroke (75). Similarly, Marcos et al. did not show a significant association between the severity of cIMT and CV events or death (76). cIMT could not be used to reliably discriminate prevalent CVD in a group of outpatients with CKD (77). Finally, a recent analysis of the Multi-Ethnic Study of Atherosclerosis cohort revealed that CAC was superior to cIMT for CVD prediction in patients with and those without CKD (Table 5) (78).

Clinical Utility of cIMT in CKD

Although studies suggested that cIMT measurements are higher in patients with CKD than in those without CKD, the differences were small and of unclear clinical relevance. In addition, observed increases in cIMT with decreasing eGFR or advancing CKD stages could be confounded by other traditional risk factors that cause CKD progression, such as uncontrolled hypertension or diabetes. At present, cIMT has not proven to be a reliable predictor of hard outcomes in predialysis patients with CKD. Currently, the standardization of cIMT measurement is a major challenge, and it is not routinely recommended in clinical practice for risk assessment in the general population, let alone patients with CKD (56). Further research needs to delineate whether cIMT can be reliably measured in patients with CKD and used as a screening test for CV risk stratification in this patient population.

Conclusion

In summary, Figure 3 outlines the potential uses of an ideal circulating and imaging cardiac biomarker, which should be similar in patients with predialysis CKD. However, given current knowledge gaps, more data need to become available before all of these markers can be reliably used in this patient population. Observational studies reporting associations between cTnT and NT-pro-BNP and decline in eGFR in nondialysis patients with CKD may be confounded by decreased renal clearance of these biomarkers in the setting of advanced CKD. The same traditional and nontraditional factors associated with CAC are likely also correlated with CKD progression. Although the evidence presented suggests that these biomarkers may be used to predict future CV events in asymptomatic patients with CKD, future studies need to confirm reliable cutoffs for the utility of these biomarkers as diagnostic tests in patients presenting with symptoms concerning for ACS or acute CHF. In addition, it remains unclear whether cardiac biomarkers such as cTnT, NT-pro-BNP, BNP, CAC, and cIMT in asymptomatic patients with CKD are modifiable and amenable to interventions to reduce future CV risk. Further studies are needed to inform whether better risk stratification scores that include novel in addition to traditional biomarkers should be developed to quantify CV risk in patients with CKD.

Figure 3.

Utility of cardiac biomarkers in patients with CKD. Ideally, cardiac biomarkers should be useful in diagnosing atherosclerotic cardiovascular disease (ASCVD), risk-stratifying patients according to disease severity, monitoring ASCVD progression, identifying patients who may benefit from early intervention, and/or monitoring response to such interventions. The potential utility of circulating and imaging cardiac biomarkers in patients with CKD is similar. It includes one or many of the features of an ideal biomarker in relation with ASCVD.

Disclosures

J.A.de L. has received grant support and consulting income from Roche Diagnostics and Abbott Diagnostics.