BACKGROUND: Quantification and comparison of high-sensitivity (hs) cardiac troponin I (cTnI) and cTnT concentrations in chronic kidney disease (CKD) have not been reported. We examined the associations between hs cTnI and cTnT, cardiovascular disease, and renal function in outpatients with stable CKD.
METHODS: Outpatients (n = 148; 16.9% with prior myocardial infarction or coronary revascularization) with an estimated glomerular filtration rate (eGFR) of <60 mL · min−1 · (1.73 m2)−1 had serum cTnI (99th percentile of a healthy population = 9.0 ng/L), and cTnT (99th percentile = 14 ng/L) measured with hs assays. Left ventricular ejection fraction (LVEF) and mass were assessed by echocardiography, and coronary artery calcification (CAC) was determined by computed tomography. Renal function was estimated by eGFR and urine albumin/creatinine ratio (UACR).
RESULTS: The median (interquartile range) concentrations of cTnI and cTnT were 6.3 (3.4–14.4) ng/L and 17.0 (11.2–31.4) ng/L, respectively; 38% and 68% of patients had a cTnI and cTnT above the 99th percentile, respectively. The median CAC score was 80.8 (0.7–308.6), LV mass index was 85 (73–99) g/m2, and LVEF was 58% (57%–61%). The prevalences of prior coronary disease events, CAC score, and LV mass index were higher with increasing concentrations from both hs cardiac troponin assays (P < 0.05 for all). After adjustment for demographics and risk factors, neither cardiac troponin assay was associated with CAC, but both remained associated with LV mass index as well as eGFR and UACR.
CONCLUSIONS: Increased hs cTnI and cTnT concentrations are common in outpatients with stable CKD and are influenced by both underlying cardiac and renal disease.
Chronic kidney disease (CKD),5 with its attendant increase in cardiovascular morbidity and mortality, has come to the forefront of public health (1, 2). Although CKD is associated with an increased risk of cardiovascular disease, (3, 4), interpretation of cardiac biomarkers, particularly cardiac troponin in the setting of CKD, has been controversial (5, 6). Guidelines for the universal definition of myocardial infarction state that an increased value for cardiac troponin is defined as a measurement exceeding the 99th percentile of a healthy reference population. Optimal imprecision (CV) at the 99th percentile should be defined as <10% (7). New high-sensitivity (hs) cardiac troponin assays with detection limits 10–100 times lower than currently available commercial assays meet this imprecision guideline (8). Recently, cardiac troponin T (cTnT) measured in dialysis-dependent patients by an hs assay was shown to be the most powerful predictor of long-term mortality compared to other biomarkers and clinical risk predictors (9). However, in non–dialysis-dependent asymptomatic patients with CKD, both quantification of concentrations and correlation with underlying cardiovascular pathology is unknown. Furthermore, whether any difference in CKD patients exists when hs assays are used to measure cTnI or cTnT remains unknown. Therefore, we designed a multicenter observational study in a stable, ethnically diverse population of CKD patients to quantify concentrations of both cTnI and cTnT using hs assays. We sought to determine the prevalence and extent of increased concentrations in this patient population compared to a healthy control population used to determine the 99th percentile cutoff for each assay and to determine if cardiac troponin concentrations are associated with prevalent cardiovascular disease, the severity of renal disease, or both.
Materials and Methods
Patients (n = 162) with stable CKD (not requiring renal replacement therapy) were prospectively recruited from outpatient nephrology clinics at the Massachusetts General Hospital, Boston; the University of Maryland, Baltimore; and the Baltimore Veteran's Administration Medical Center from 2006 to 2007. Eligibility requirements were age ≥30 years and sustained (≥3 months) reduction in estimated glomerular filtration rate (eGFR) of ≤60 mL · min−1 · (1.73 m2)−1 based on the simplified Modification of Diet in Renal Disease formula (10). Fourteen patients were excluded because of inadequate serum to measure cardiac troponin. Exclusion criteria included stage 5 kidney disease or renal replacement therapy (dialysis or kidney transplant). Although prevalent cardiovascular disease was permitted, a history of myocardial infarction within 90 days of enrollment or a history of coronary artery bypass grafting were both exclusions. To recruit stable asymptomatic to minimally symptomatic patients, those patients were also excluded who had symptoms consistent with greater than New York Heart Association class I heart failure or greater than Canadian Cardiovascular Society class I angina. The study was approved by the institutional review boards of the Massachusetts General Hospital, University of Maryland School of Medicine, and Baltimore Veteran's Administration Medical Center, and all patients provided written informed consent.
CLINICAL AND LABORATORY DATA COLLECTION
At the time of enrollment, patient vital signs were recorded and demographic characteristics and medical history were collected by chart review and interview. Coronary artery disease history was defined as either a prior myocardial infarction or percutaneous coronary revascularization. Blood and urine samples from each study participant were collected and immediately centrifuged, separated into aliquots, and stored at −70 °C for future batched testing. Serum cTnI was measured with a prototype hs assay and a commercial assay (Dimension Vista® 1500); cTnT was measured with a commercial (but not yet available in the US) hs assay and the standard fourth-generation conventional cTnT assay (Elecsys® 2010). The limit of blank for the hs cTnT assay is 3.0 ng/L (11).
Concentrations of hs cTnI [n = 288, 122 males (42%), age range 18–59 years] and hs cTnT [n = 616, 309 males (50.2%), age range 20–71 years] from healthy volunteers defined the 99th-percentile cutoff. For hs cTnI, the reference values were provided by the manufacturer with a reported 99th percentile of 9 ng/L (10% CV at 3 ng/L), and for hs cTnT the reported 99th percentile value was 14 ng/L (10% CV at 13 ng/L) (11; Mary Lou Gantzer, personal communication, on or about March 1, 2011). The 10% CV and 99th percentile in a healthy population for non-hs cTnI are 30 ng/L and 40 ng/L, respectively, and for standard cTnT are 30 ng/L and 10 ng/L, respectively (12, 13).
We quantified cystatin C concentrations (measurement range 0.05–8.0 mg/L; BN ProSpec®, Siemens Healthcare Diagnostics) using previously frozen Li heparin plasma. The estimated glomerular filtration rate calculated with the cystatin C value (eGFRcys) was obtained by use of the Stevens CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration) formula (14). Serum creatinine was measured by use of a modified Jaffe method available on the Dimension RxL system (Siemens Healthcare Diagnostics); this method meets the specified total error goals and is functionally consistent with the isotope dilution mass spectrometry standardization program (15). eGFRcreat was calculated with the creatinine value by using the simplified Modification of Diet in Renal Disease formula as part of a secondary analysis (10). Urine albumin/creatinine ratio (UACR) was also measured by using the Dimension RxL system (Siemens Healthcare Diagnostics).
As an exploratory aim, all-cause mortality was assessed in 100% of participants by reviewing both hospital/outpatient records and the US Death Index through August 2011.
All study participants underwent a research-based 2-dimensional transthoracic echocardiogram to quantify the left ventricular mass index (LVMI), LV volumes, and LV ejection fraction (LVEF). All images were quantified by a single reviewer blind to the clinical data at a core laboratory at the University of Maryland. LVMI and LV volumes and LVEF were quantified as previously described (16).
Each study participant also underwent a research-based cardiac scan to assess for coronary artery calcification (CAC). Computed tomography examinations were performed on a 64-slice scanner [Sensations 64 (Siemens Medical Solutions) at the Massachusetts General Hospital; Brilliance 64 (Philips Healthcare) at the University of Maryland] with standardized protocols. Calcifications were quantified with dedicated scoring software (Brilliance Workspace, Philips Healthcare) at the University of Maryland by 1 observer who was blind to the clinical and laboratory data and used the method of Agatston et al. (17).
Standard descriptive statistics were used to assess baseline characteristics. Continuous data are shown as the median [interquartile range (IQR)] or mean (SD), and categorical data as the number (percentage). Frequency histograms were plotted to compare the distribution of hs cTnI and cTnT between healthy control individuals and study participants with CKD. Clinical characteristics were compared across quartiles of hs cTnI and hs cTnT by use of 1-way ANOVA or Cuzik's nonparametric trend test for continuous variables, and the score test for trend for binary variables. The pair-wise correlation between hs cTnI or cTnT and renal function measures (eGFRcys and UACR) were estimated with Pearson's correlation coefficients. Simple and multiple linear regression (adjusting for age, sex, race, diabetes, coronary artery disease history, smoking, eGFRcys, and UACR) were developed to estimate the association between cardiac imaging (CAC and LVMI) and hs cTnI or cTnT concentrations. In addition, univariate and multivariate linear regression (adjusting for age, sex, race, diabetes, coronary disease history, and LVMI) analyses were performed to determine correlations between renal function (eGFRcys and eGFRcreat, dependent variables) and hs cardiac troponin concentrations (independent variables). We also performed simple and multiple linear regression for the association of UACR and both hs cardiac troponin assays, adjusting for age, sex, race, diabetes, coronary disease history, LVMI, and eGFRcys. Cumulative survival was compared across subgroups defined by tertiles of hs cTnI and hs cTnT with the Kaplan–Meier method and the log-rank test for trend. Cox proportional hazard models were developed to estimate the association of hs cardiac troponin concentrations and survival, adjusting for demographics and the severity of renal disease. No formal sample size calculation was performed for this pilot-level study. Post hoc power estimation suggested 90% power to detect a linear correlation coefficient of at least 0.23, with a 5% type I error rate. We performed statistical analysis using MedCalc for Windows, version 18.104.22.168 (MedCalc Software), Stata version 11 (Statacorp), and the graphing functions of a spreadsheet program (Excel 2003, Microsoft).
ASSOCIATION OF CARDIAC TROPONINS WITH CLINICAL CHARACTERISTICS AND CARDIAC IMAGING
Table 1 shows baseline clinical, laboratory, and imaging characteristics for the 148 CKD study participants. The median age was 63 (IQR 56–71) years, with 102 (68.9%) males, nearly 92% of patients with hypertension, and a little over half of patients with a history of diabetes. The mean (SD) eGFRcys for the study participants was 37.3 (16.7) mL · min−1 · (1.73 m2)−1. The LVEF was typically within the reference range (>55%) with a median of 58% (IQR 56%–61%). The CAC score, although frequently increased [median of 80.8 (IQR 0.5–315.2)], was within the reference interval (<10) in 46 (31.1%) of study participants. There were no differences between the sites (Baltimore sites combined vs Massachusetts General Hospital) with respect to age, eGFRcys, coronary disease history, or hs cTnI or cTnT concentrations.
With the use of the hs assays, the median cTnI concentration was 6.3 (IQR 3.4–14.4) ng/L and the median cTnT concentration was 17.0 (IQR 11.2–31.9) ng/L. The distribution of concentrations of cTnI and cTnT in the CKD study cohort compared to the healthy adult controls are shown in Fig. 1. Notably, 38% of the patients with CKD had an hs cTnI concentration above the 99th percentile of the healthy controls, whereas 68% had an hs cTnT concentration above the 99th percentile. In contrast, the use of the contemporary cTnI and cTnT assay for each analyzer showed that only 12 (8.1%) of patients had a cTnI value above the 99th percentile and 24 (16.2%) had a cTnT value above the 99th percentile (see Fig. 1, A and B, in the Data Supplement that accompanies the online version of this article at http://www.clinchem.org/content/vol58/issue9).
Tables 2 and 3 show trends for clinical, laboratory, and cardiac imaging characteristics across progressive quartiles of hs cTnI and cTnT, respectively. Progressively higher hs cTnI and cTnT quartiles were significantly associated with advancing age, greater prevalence of male sex, diabetes, history of coronary disease, greater use of antiplatelet medications, and higher CAC scores. Furthermore, progressively higher hs cTnT quartiles were associated with greater LVMI (P = 0.005) with a trend for increased LVMI across progressive hs cTnI quartiles.
MULTIVARIATE ASSOCIATION WITH CARDIAC IMAGING MEASURES
We developed linear regression models to evaluate the strength of the association of the hs cardiac troponin assays with CAC and LVMI using the hs cardiac troponin concentrations as continuous predictor variables and adjusting for potential confounders. Online Supplemental Table 1 shows the unadjusted and adjusted association of both hs cardiac troponin assays and measures of renal function with the extent of CAC. In an unadjusted analysis, hs cTnI, hs cTnT, and eGFRcys, but not UACR, were associated with CAC. After we adjusted for demographic variables, neither hs cardiac troponin assay remained associated with CAC. To test if CAC was relevant to epicardial coronary disease in these CKD patients, history of coronary disease was entered into a statistical model adjusted for demographics, eGFRcys, and hs cardiac troponin concentrations. A history of coronary disease remained correlated to CAC (β̂) = 0.24, P = 0.001), whereas neither measures of renal function nor concentrations of hs cTnI or cTnT were independently associated with CAC.
A similar analysis was done for LVMI, as detailed in online Supplemental Table 2. In an unadjusted analysis, both hs cardiac troponin assays and UACR, but not eGFRcys, were significantly associated with LVMI. However, in fully adjusted models, only the hs cardiac troponin assays, but neither measure of renal function, remained significantly correlated with LVMI.
ASSOCIATION OF HS CARDIAC TROPONINS WITH RENAL FUNCTION
We assessed the association of both hs cardiac troponin assays with measures of renal function. Fig. 2 shows the correlation of hs cTnI and hs cTnT with eGFRcys. For both markers there was a moderate but significant correlation with renal function (hs cTnI, r = −0.29, P < 0.003; hs cTnT, r = −0.48, P < 0.001). Online Supplemental Fig. 2, A and B, shows the association of hs cTnI and hs cTnT with renal function by quartiles of eGFRcys. The association of hs cTnT with eGFR was numerically larger than for hs cTnI. When both hs cTnI and hs cTnT were entered simultaneously as independent variables, the association of hs cTnI with eGFR was no longer significant (P = 0.7), whereas hs cTnT remained significantly correlated with eGFR (P < 0.001). Each cardiac troponin was significantly associated with UACR (hs cTnI, β̂ = 0.33, P < 0.001; hs cTnT, β̂ = 0.45, P < 0.001). In unadjusted and multivariate analysis (adjusted for variables that were significantly associated with each measure of renal function) hs cTnI and hs cTnT individually remained significantly associated with eGFRcys (and to a lesser extent with eGFRcreat) and UACR (see online Supplemental Table 3). We also assessed the association of the hs cardiac troponin concentrations with dichotomous combinations of eGFRcys and UACR divided at their median value for the study population. Median values for both cardiac troponin assays rose progressively from study participants with eGFRcys above the median and UACR below the median to study participants with eGFRcys below the median and UACR above the median (Fig. 3). This upward shift in hs cardiac troponin in the presence of higher amounts of albuminuria, irrespective of eGFRcys, appeared more marked for hs cTnT than hs cTnI.
CONCORDANCE OF HS CARDIAC TROPONIN ASSAYS
Study participants were divided into 4 groups on the basis of concentrations greater than or less than or equal to the 99th percentile of the healthy population without CKD for each hs cardiac troponin assay. The concordance between the 2 assays was significant, but only moderate (κ = 0.40, P < 0.001), with 31.1% of study participants having discordant results; the majority (25.7%) of the discordance was a high hs cTnT and low hs cTnI concentration (see online Supplemental Table 4). To further explore the characteristics of this discordance, hs cTnT vs hs cTnI values were plotted on a scatter plot with the lines representing the 99th percentile for each assay placed for reference (see online Supplemental Fig. 3). There were outliers for both assays for which a value twice the 99th percentile of 1 assay was present, whereas for the other assay, the value of the cardiac troponin concentration was measured below the 99th percentile. This was more common for hs cTnT than hs cTnI. Lastly, in an exploratory analysis we looked for differences between the concordant and discordant groups (see online Supplemental Table 5). The numbers of study participants in the discordant groups were small, but in the presence of an increase of only 1 of 2 cardiac troponin concentrations there was a trend for more advanced age, greater prevalence of diabetes, and higher LV mass and CAC scores compared to study participants with low concentrations of both cTnI and cTnT.
HS CARDIAC TROPONIN ASSAYS AND ALL-CAUSE MORTALITY
A total of 34 deaths occurred over a median of 4.8 years of follow-up. Kaplan–Meier plots for tertiles of each assay are shown in online Supplemental Fig. 4, A and B. There is a significant increase in mortality with higher concentrations, but this isn't seen until the third quartile, which starts for each assay above the 99th percentile of the healthy population without CKD. The Cox regression analysis for each hs cardiac troponin assay is shown in online Supplemental Table 6. Adjustment for demographic variables only modestly attenuated the hazard ratios for all-cause mortality irrespective of whether the assay concentrations were analyzed categorically as tertiles or as continuous variables. In contrast, the addition of either measure of renal function moderately to markedly diminished the hazard ratios for mortality, suggesting significant colinearity between markers of renal function and the hs cardiac troponin assays for predicting the risk of death.
When we used hs assays, cTnT and cTnI concentrations were detectable in almost all of our stable outpatients with CKD. Furthermore, irrespective of the assay, hs cardiac troponin concentrations were above the 99th percentile of a healthy general population in a large proportion of patients with CKD, frequently several-fold higher than the 99th percentile. The etiologies of these increases in these patients with CKD appear to be multifactorial, related to cardiac and renal pathology. This finding was further emphasized by our exploratory analysis of hs cardiac troponin concentrations and all-cause mortality, for which the strong association between higher hs cardiac troponin concentrations by either assay was markedly attenuated when we adjusted for measures of renal function.
Unadjusted associations were present for both hs cardiac troponin assays in association with coronary disease and 2 different markers of renal function. Increased CAC scores are common in CKD independent of markers of renal function or mineral metabolism, such as fibroblast growth factor 23 (16, 18). However, after we adjusted for demographic factors, the association between hs cTnI, hs cTnT and CAC was lost. This implies either that epicardial coronary disease plays a minor role in cardiac troponin increase, or CAC may be a poor surrogate for coronary disease in patients with CKD. We cannot comment on the presence or severity of lipid-rich “soft” atherosclerotic plaques in our study participants, which have been independently related to concentrations of hs cTnT in other populations (19) and may account for increased hs cardiac troponin concentrations in this population. In contrast, the robust independent association of both cardiac troponin assays with LVMI suggests that prevalent ongoing subclinical myocyte cell death (in the form of oncosis, apoptosis, or autophagy) is present among patients with stage 3–4 CKD.
The independent association of both hs cardiac troponin assays with eGFR and UACR in this study identifies potential renal-specific mechanisms of cardiac injury that are not identified by cardiac imaging. The finding that eGFRcys and UACR are synergistically associated with higher hs cardiac troponin concentrations, particularly cTnT, is supported by epidemiologic evidence that the presence of albuminuria is an additional risk factor for cardiovascular events in the presence of an eGFR <60 mL · min−1 · (1.73 m2)−1 (20–22).
In patients with CKD, potential mechanisms to explain the colinearity of measures of renal function and myocardial injury with cardiac troponin concentrations include both increased cardiac injury and decreased renal clearance. A third explanation for increased cardiac troponin concentrations, increased cardiac troponin production from noncardiac sources, has been previously refuted (23). In support of increased myocardial injury are the cardiac histologic findings resulting from uremia in partially nephrectomized rats, including increased LV mass with decreased capillary density and increased interstitial fibrosis (24). Recently, fibroblast growth factor 23, a bone hormone–regulating renal phosphate excretion that is up regulated in CKD and associated with increased LVMI in this cohort and others (17, 25), has been shown to have a causative role in increasing LV mass in the setting of experimentally induced CKD (25). This finding may indicate an important link between measures of impaired renal function and increased concentrations of cardiac troponin for mortality. In contrast, a study in which heart failure patients with increased cardiac troponin concentrations, with and without CKD, were investigated demonstrated similar increases in the transcardiac gradient for cardiac troponin concentrations from the aorta to the coronary sinus, but higher aortic cardiac troponin concentrations in patients with CKD compared to patients without CKD (26). This finding is indirect evidence in support of decreased renal clearance of cardiac troponin.
In this study we confirmed that cTnI and cTnT, when measured by using conventional assays (with the same immunochemistry platforms used to measure the hs assay), are detectable only in a small minority of clinically stable CKD patients not on dialysis. There remains controversy with respect to whether there are inherent differences between cTnI and cTnT in the setting of renal disease (27). Lamb et al. introduced the concept that differences between the 2 cardiac troponins were related to differences in assay sensitivity and showed that the proportion of nondialysis CKD patients with detectable cTnI and cTnT were similar (approximately one-third) when a sensitive (not hs) version of cTnI was used (28). In our analysis of concordance using 99th percentile values derived from healthy non-CKD populations, the majority of discordant values were associated with high cTnT/low cTnI concentrations. The small number of study participants in these discordant groups precludes definitive conclusions regarding differences in these study participants compared to study participants with concordant hs cardiac troponin results.
We also found that hs cTnT appears to be more strongly associated with renal measures of glomerular filtration and vascular permeability/injury (UACR) than hs cTnI. As Lamb et al. previously observed, and our findings with hs assays extend, an hs cTnI assay narrows, but does not close, this gap for prevalence of increased concentrations (above the 99th percentile of a healthy general population cohort) between cTnI and cTnT in this setting (28). The moderate size of this study precludes definitive conclusions that one hs cardiac troponin assay is more correlated with renal function than the other. The fact that both cardiac troponin assays remain independently correlated with eGFR calculated by both cystatin C– and creatinine-based formulas as well as another measure of renal function, UACR, supports the finding that the extent of measured cardiac injury by hs cardiac troponin assays is associated with the extent of renal impairment. Lastly, our finding that the increased risk of all-cause mortality for outpatients with CKD does not begin until values for both hs cardiac troponin assays are above the 99th percentile of a healthy population should also be interpreted with caution given the small sample size.
Our study had several limitations to consider. First, despite the multicenter, multiethnic study population, patients were recruited from hospital-based CKD specialty clinics and therefore may not have represented the CKD population seen in community practices or presenting to the emergency department with chest pain or the equivalent. Second, as discussed above, the correlation of coronary calcium with coronary artery disease is uncertain. Third, although the Jaffe method is commonly used to measure serum creatinine in clinical laboratories, the issue of interfering substances in serum can lead to the overestimation of serum creatinine by as much as 15%–25% by various Jaffe methodologic applications (15). However, we relied predominantly on the eGFRcys for correlation to the cardiac troponin concentrations, which, as measured by an immunoassay, is not subject to the same interference. Lastly, we could not determine to what extent impaired clearance vs increased production accounted for increased hs cardiac troponin concentrations.
When hs cardiac troponin assays are used to diagnosis acute coronary syndromes in patients with CKD, baseline values above the 99th percentile of a healthy population should be anticipated irrespective of whether cTnI or T is measured. As a result, an increased reliance on change in concentrations between sequential measures may be necessary, particularly for low concentration increases that are of uncertain prognostic importance (29). When evaluating those patients with CKD, clinicians should be aware of the potential for the presence of hs cardiac troponin results due to nonacute coronary syndrome mechanisms, while keeping in mind the excess risk related to acute ischemic heart disease in such patients; integrating the results of hs cardiac troponin assays together with clinical assessment, as has been recommended, remains the optimal approach for use of these sensitive assays (7).
The authors would like to thank Mary Lou Gantzer, PhD, formerly of Siemens Healthcare Diagnostics, for measuring the blood samples for hs cTnI.
↵5 Nonstandard abbreviations:
- chronic kidney disease;
- high sensitivity;
- cardiac troponin T;
- estimated glomerular filtration rate;
- eGFR calculated by using cystatin C;
- eGFR calculated by using creatinine;
- urine albumin/creatinine ratio;
- LV mass index;
- LV ejection fraction;
- coronary artery calcification;
- interquartile range.
Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article.
Authors' Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest:
Employment or Leadership: None declared.
Consultant or Advisory Role: C. deFilippi, Roche Diagnostics Corporation; S.L. Seliger, Roche Diagnostics Corporation; R.H. Christenson, Siemens Healthcare Diagnostics; J. Januzzi, Roche Diagnostics Corporation and Critical Diagnostics.
Stock Ownership: None declared.
Honoraria: C. deFilippi, Roche Diagnostics Corporation and Siemens Healthcare Diagnostics; R.H. Christenson, Siemens Healthcare Diagnostics; J. Januzzi, Roche Diagnostics Corporation and Siemens Healthcare Diagnostics.
Research Funding: Roche Diagnostics Corporation, Siemens Healthcare Diagnostics, Critical Diagnostics, and Thermo Fisher; C. deFilippi, Dade Behring (now Siemens Healthcare Diagnostics).
Expert Testimony: None declared.
Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, or preparation or approval of manuscript.
- Received for publication March 15, 2012.
- Accepted for publication June 20, 2012.
- © 2012 The American Association for Clinical Chemistry