|
| HOME | SUBSCRIPTIONS | CURRENT ISSUE | PAST ISSUES | CARDIOSOURCE | SEARCH | HELP | FEEDBACK |
|
|
|
|||||||||
|
|||||||||||
|
J Am Coll Cardiol, 2001; 38:991-998 © 2001 by the American College of Cardiology Foundation |
* Division of Cardiology, Department of Medicine, Creighton University School of Medicine, Omaha, Nebraska, USA
Long Island College Hospital, Brooklyn, New York, USA
Manuscript received January 30, 2001; revised manuscript received May 21, 2001, accepted June 28, 2001.
Reprint requests and correspondence: Dr. Ijaz A. Khan, Cardiac Center, Creighton University, 3006 Webster Street, Omaha, Nebraska 68131
ikhan{at}cardiac.ctrighton.edu
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionReferences |
|---|
We sought to evaluate the prognostic value of cardiac troponin I (cTnI) in asymptomatic, ambulatory patients with chronic renal failure treated with long-term hemodialysis.
BACKGROUND
Smaller, short-term follow-up studies on this subject have given conflicting results.
METHODS
A total of 126 ambulatory patients with chronic renal failure treated with long-term hemodialysis were followed for two years for all-cause mortality, cardiac mortality, all-cause hospital admissions and cardiac hospital admissions. Serum cTnI was measured before dialysis at the time of study entry.
RESULTS
One hundred two patients had normal serum levels of cTnI (
0.03 ng/ml) and 24 patients had elevated levels (0.015 ± 0.007 vs. 0.053 ± 0.029 ng/ml, p < 0.0001). No significant difference in all-cause mortality (20 vs. 4 deaths), cardiac mortality (4 vs. 1 death), all-cause hospital admissions (1.74 ± 1.72 vs. 1.25 ± 1.19 admissions/patient) or cardiac admissions (0.52 ± 0.89 vs. 0.33 ± 0.76 admissions/patient) was present between the patients with normal cTnI levels and those with elevated cTnI levels. Serum cTnI was not significantly different between patients who died versus those who survived (0.022 ± 0.019 vs. 0.022 ± 0.021 ng/ml). Serum cTnI was not an independent predictor of all-cause mortality, cardiac mortality, all-cause admissions or cardiac admissions. Age (older) and serum albumin (lower) were independent predictors of all-cause mortality, whereas a history of myocardial infarction was an independent predictor of cardiac mortality. Serum sodium (lower) was an independent predictor of all-cause hospital admissions, whereas hypertension and previous myocardial infarction were independent predictors of cardiac admissions. The best predictors of the time to death were age (older) and serum sodium level (lower), irrespective of the serum cTnI levels.
CONCLUSIONS
Cardiac troponin I has a limited role in predicting mortality and hospital admissions in asymptomatic patients with chronic renal failure treated with long-term hemodialysis.
| ||||||||||||
end-stage renal disease and accounts for much of their observed morbidity and mortality; 50% of patients with end-stage renal disease die of cardiovascular events (9). Diagnosing ischemic heart disease in this patient population remains challenging, because many of these patients present with abnormal baseline electrocardiograms, frequently compounded by silent or atypical symptoms and a reduced reliability of creatine kinase-MB isoenzyme (CK-MB), the conventional marker of myocardial necrosis (9–14). Several studies have questioned the usefulness of the cardiac troponins, particularly cTnT, in the presence of renal disease (14–22).
Because cTnI is exclusively of cardiac origin and, unlike CK-MB and cTnT, does not express in the skeletal muscle at any developmental stage, it has been shown to be more specific for the detection of myocardial injury in patients with chronic renal failure (14–16,23–28). However, the prognostic value of elevated serum cTnI to predict subsequent cardiac-related events in asymptomatic patients with renal disease treated with hemodialysis has not been extensively studied with long-term follow-up; the subject has only been studied with follow-up periods limited to 6 to 12 months (29–31). The present study was conducted to evaluate the prognostic value of elevated cTnI in asymptomatic, ambulatory patients with chronic renal failure treated with long-term hemodialysis. The follow-up period extended to two years.
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Assays. All of the serum samples were centrifuged for 5 min at 2,300 revolutions per minute to remove any particulate matter. Plunger-type filters were used to remove any remaining fibrin. Each tube was labeled with the patient number and accession number. The technicians who performed the assays were had no knowledge of the clinical data. Specimens were re-analyzed at random or to verify an abnormal result to eliminate technical errors. Collected data were summarized on spreadsheets, with the data sorted in ascending order by both the patient number and accession number.
Cardiac troponin I was measured with the ACCESS troponin I assay (Sanofi Diagnostic Pasteur, Chaska, Minnesota). The ACCESS troponin I assay is a two-site sandwich immunochemiluminescent assay using two mouse monoclonal antibodies as captured and labeled antibodies that recognize different epitopes unique to the human cTnI isoform. The cTnI assay has no detectable cross reactivity with human skeletal muscle troponin I. By this method, the serum cTnI level in a normal healthy population is 0.03 ng/ml, as compared with 0.4 ng/ml by the conventionally used Stratus II fluorometric enzyme immunoassay (Dade International, Miami, Florida). Creatine kinase, MB isoenzymes were measured with the ACCESS CK-MB assay (Sanofi Diagnostic Pasteur). The ACCESS CK-MB assay is a two-site sandwich immunochemiluminescent assay using two mouse monoclonal antibodies that specifically recognize CK-MB. By this method, the reference normal limit of CK-MB is 4 ng/ml.
The analytical sensitivity of each assay was defined as the smallest concentration that can be distinguished from zero. Ten replicates of the zero calibrator were run for cTnI and CK-MB. The analytical sensitivity was calculated using the 10-zero calibrator replicate relative light units measured 2 SD from the zero-curve fit. Both assays generated analytical sensitivity values within package-insert specifications; the analytical sensitivity of cTnI was 0.006 ng/ml, against an expected value of 0.03 ng/ml, and that of CK-MB was 0.09 ng/ml, against an expected value of 0.3 ng/ml. The precision within runs was calculated based on 10 replicates of BioRad’s tri-level cardiac control. The procedure adopted included one analyte per tray, three control levels per analyte and 10 replicates per control level. The imprecision within runs for both assays was <10%.
Serum concentrations of creatinine, calcium, potassium, sodium, albumin, bicarbonate, chloride and blood urea nitrogen were measured on a Hitachi 747 Analyzer (Hitachi, Indianapolis, Indiana), using routine methods according to the manufacture’s protocols.
Follow-up.
The study group was classified into two groups according to the pre-dialysis serum cTnI level on initial enrollment: 1) patients with normal serum cTnI levels (0.03 ng/ml); and 2) patients with elevated serum cTnI levels (>0.03 ng/ml). All patients were followed up for two years from the time of their initial enrollment in the study. The end points studied were: 1) all-cause mortality; 2) cardiac mortality; 3) all-cause hospital admissions; and 4) cardiac hospital admissions. Hospital records, outpatient clinic records and interviews with the patient, family or primary physician were used for confirmation of the events. In case of death, the cause of death was verified by reviewing medical records or death certificates. Death due to acute myocardial infarction, pulmonary edema, cardiogenic shock or malignant arrhythmias was deemed cardiac. Sudden unexplained death was also considered as cardiac. Admission secondary to acute myocardial infarction, chest pain suggestive of an acute coronary syndrome, congestive heart failure, arrhythmias and vascular diseases were considered as cardiac. Admissions related to vascular access for dialysis were not included in the admission counts.
Statistical analysis.
Continuous variables were expressed as the mean value ± SD and were analyzed by the Student t test. Categorical variables were expressed as percentages and were analyzed by the chi-square or Fisher exact test, as appropriate. A two-tailed p value 0.05 was considered significant. Kaplan-Meier survival analysis was performed to evaluate significant differences in all-causes death and cardiac death between patients with normal cTnI levels and those with elevated cTnI levels. Receiver-operating characteristic (ROC) curves of serum cTnI, as a predictor of all-cause mortality, cardiac mortality, all-cause hospital admissions and cardiac hospital admissions, were plotted. Logistic regression analysis was performed to identify independent predictors of all-cause mortality, cardiac mortality, all-cause hospital admissions and cardiac hospital admissions. Cox regression analysis was used to determine independent variables as the best predictors of time to death. All of the statistical analyses were performed using SPSS, version 7.0 (SPSS Inc., Chicago, Illinois).
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
0.03 ng/ml) and 24 had elevated levels (>0.03 ng/ml). The serum cTnI values in the group of patients with normal levels ranged from 0 to 0.029 ng/ml (mean [±SD] 0.015 ± 0.007 ng/ml), and the serum cTnI values in the group of patients with elevated levels ranged from 0.031 to 0.14 ng/ml (mean [±SD] 0.053 ± 0.029 ng/ml; p < 0.0001). There was no significant difference in baseline clinical characteristics between the patients with normal cTnI levels and those with elevated cTnI levels (Table 1). Similarly, there was no significant difference in the baseline biochemical profile between the two groups, except for a high CK-MB value in the group with elevated cTnI (Table 2).
|
|
|
Kaplan-Meier survival analysis showed no significant differences in all-cause mortality (p = 0.7292 by the log-rank statistic) or cardiac mortality (p = 0.2319 by the log-rank statistic) between patients with normal cTnI levels and those with elevated cTnI levels. The ROCs of serum cTnI as a predictor of all-cause mortality, cardiac mortality, all-cause hospital admissions and cardiac hospital admissions are shown in Figures 1, 2, 3 and 4. The data points on all of these ROC plots are near 45°, indicating that serum cTnI is not a useful tool for predicting any of these end points. Independent predictors of death were older age (65.9 ± 11.5 vs. 58.1 ± 15.7 years in dead vs. alive patients; p = 0.008 by the independent t test) and lower serum albumin levels (3.64 ± 0.32 vs. 3.85 ± 0.36 g/dl in dead vs. alive patients; p = 0.01 by the independent t test). There was no difference in the serum cTnI levels between patients who died and those who survived (0.022 ± 0.019 vs. 0.022 ± 0.021 ng/ml; p = 0.925). In logistic regression analysis, serum cTnI was not an independent predictor of all-cause mortality, cardiac mortality, all-cause hospital admissions or cardiac hospital admissions. Serum albumin was the best predictor of all-cause mortality (the lower levels were associated with higher mortality), and a history of myocardial infarction was the best predictor of cardiac mortality. The serum sodium level was an independent predictor of all-cause hospital admissions (lower levels were associated with more admissions). Hypertension and previous myocardial infarction were independent predictors of cardiac hospital admissions. The best predictors of the time to death by Cox regression analysis were age (older) (p < 0.005) and serum sodium level (lower) (p < 0.005), irrespective of the serum cTnI levels.
|
|
|
|
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Elevation of cTnI in renal failure. Although cTnI has been considered the most specific among the currently available biochemical markers of myocardial damage, its value in the presence of chronic renal failure has been questioned (28,30,32). The cause of nonspecific elevation of cTnI is not clear. The current hypotheses addressing such elevation in chronic renal failure include microinjury to the myocardium that is undetected by conventional cardiac imaging, cardiotoxicity due to changes in osmolarity and/or ion fluxes, increased cardiac preload and myocardial stretch due to volume overload and nonischemic cardiac injury due to calcium and oxalate deposition in the heart (30,32). There is little evidence to suggest that the false-positive rate of cTnI is related to the magnitude of renal insufficiency, because no relationship between the serum cTnI concentration and the serum creatinine value or creatinine clearance rate has been demonstrated, and no significant differences has been found between pre- and post-dialysis values of serum cTnI (28–30,33).
Although the reason for the elevation in serum cTnI in patients with renal disease is not clear, it may represent the problems in standardization of cTnI assays. Up to 20-fold variation in cTnI mass determination may be observed for a given patient sample when measured by different systems (34). The manufacturers quote different assay cut-off points and different studies using the same manufacturer’s assay have used different cut-off points (4,17). The heterogeneous nature and biochemical complexity of the serum forms of cTnI and the difference in the epitome recognition by different methods have also hindered the standardization of serum cTnI assays (34). Consequently, significant ambiguity often exists in the clinical interpretation of serum cTnI levels. Furthermore, recent studies have shown differential stability of cTnI with cTnI-troponin C complexes, which may also affect the assay results (35). In addition, binding sites of test antibodies could be altered due to changes in protein kinase C properties in patients with renal disease, because protein kinase C phosphorylates troponin I at many different sites (35).
Cardiac troponin I and prognosis. Although troponin I is a part of the troponin-tropomyosin complex of the contractile apparatus in myocytes, the exact nature of its release from the contractile apparatus of the striated muscle is not fully understood; probably it is released from the cytosolic pool into the circulation after necrosis (19). Because cTnI does not normally circulate in the blood, or circulates only in minute amounts, and is 13 times more abundant in the myocardium than is CK-MB, on the basis of weight, the signal-to-noise ratio associated with cTnI is much more favorable for the detection of minor amounts of cardiac necrosis (36). Elevation of cTnI has been shown to be an important factor for the short- and long-term prognoses of cardiac mortality and nonfatal cardiac events in patients with unstable angina (7). As more research is focused on the clinical utility of cTnI, its role as a prognosis indicator is expanding. It has been shown that elevated cTnI in acute myocardial infarction predicts lower reperfusion rates and an increased risk of primary angioplasty failure (37). Furthermore, it can identify a group of patients with acute coronary syndromes who would benefit from early catheter-based coronary interventions (38). Cardiac troponin I has also been shown to identify a subgroup of patients who are at high risk of graft failure after cardiac transplantation; thus, it may be used in the early detection of rejection-related myocardial injury/necrosis (39). Although not proven conclusively, microinjury to the myocardium is a frequently proposed mechanism for elevation of troponins in the absence of clinical, ECG or echocardiographic abnormalities. Histologic findings strongly suggest that minor myocardial cell injury takes place more frequently in high-risk groups of patients who have a worse prognosis (4,6,40).
Cardiac troponin I and prognosis in renal failure. Although there have been a few previous reports on the ability of cTnI to predict future cardiac events in patients with chronic renal failure who are free of cardiac symptoms, the published data are sparse and the follow-up periods have been short (29–31). These studies have demonstrated mixed results. Apple et al. (29) retrospectively studied 16 patients receiving long-term hemodialysis to evaluate the prognostic significance of both cTnI and cTnT over a period of one year. All of these patients had type I diabetes for a mean period of 24 years. The end point evaluated was the occurrence of fatal myocardial infarction. The assay used to measure serum cTnI was enzyme-linked immunosorbent assay, and the upper reference limit used was of 0.8 µg/l (ng/ml). At the time of enrollment, three patients had serum cTnI levels above the upper reference limit (1.2, 2.4 and 1.6 µg/l) and 12 patients had serum cTnT levels above the upper reference limit. All three patients with elevated serum cTnI levels had a history of ischemic heart disease (unstable angina/coronary artery bypass graft [n = 1], myocardial infarction [n = 1] and unstable angina [n = 1]). Patients who had elevated cTnI levels also had elevated cTnT and CK-MB. During one-year follow-up, four patients had a fatal myocardial infarction, including those three with elevated serum cTnI levels; however, it was not analyzed whether cTnI was an independent predictor of the fatal myocardial infarction.
In another study, Mockel et al. (30) evaluated the prognostic significance of cTnI and cTnT in 40 patients with chronic renal failure. Twenty of these patients had pre–end-stage renal disease and another 20 had end-stage renal disease and were treated with hemodialysis. Patients with any symptoms of angina pectoris during the two weeks before enrollment or an acute coronary syndrome in the four weeks before enrollment were not included in the study. Serum cTnI was measured by immunofluorescence assays made by two different manufacturers—Dade Stratus II (Dade International) and Behring OPUS Plus (Behring Diagnostics, Frankfurt, Germany). The cTnI cut-off values used in this study were 0.4 µg/l (ng/ml) for the Dade Stratus II assay and 1.6 µg/l (ng/ml) for the Behring OPUS Plus assay. In patients with pre–end-stage renal disease, 20% of patients had elevated cTnI levels by the Dade Stratus II assay and 10% by the Behring OPUS Plus assay, and in patients with end-stage renal disease treated with hemodialysis, these rates were 55% and 15%, respectively. Most of these elevated levels were modest. The follow-up period was nine months. The primary end points studied were acute myocardial infarction, all-cause mortality and hospital admission. One patient had an acute myocardial infarction, 5 patients died and 22 patients were admitted to the hospital during the nine-month follow-up period. The logistic regression model showed that serum cTnI measured by any assay was not a predictor of any of the primary end points, considered individually or in combination of two or all three end points.
In a pilot study, Ropollo et al. (31) examined the usefulness of cTnI and cTnT as predictors of subsequent cardiac events in 49 patients with chronic renal failure treated with dialysis (48 on hemodialysis and 1 on peritoneal dialysis). An additional 83 patients with renal failure who were not on dialysis were also examined. Serum cTnI was measured by the Behring OPUS immunofluorescence assay, and the cut-off value used was 0.5 ng/ml. The end points evaluated were unstable angina, acute myocardial infarction and cardiac death. The follow-up period was six months. Among the 49 dialysis patients, three had elevated cTnI at entry (0.94, 1.34 and 14.5 ng/ml). Six patients had significant cardiac events (myocardial infarction [n = 4], arrhythmia [n = 1] and unstable angina [n = 1]) during the six-month follow-up period. Of these six patients, three had elevated cTnI serum levels, but the other three had normal cTnI serum levels. Three patients died during the six-month follow-up period. Of these, two had elevated serum levels of cTnI. All three patients with elevated troponin I also had elevation of cTnT, and, similarly, all six patients who developed cardiac events had elevation of cTnT. All 49 patients on dialysis, although asymptomatic at the time of enrollment, were had chronic stable angina. Other co-morbid conditions reported in this group of patients were hypertension (n = 46), previous myocardial infarction (n = 10), congestive heart failure (n = 10), diabetes (n = 22), hyperlipidemia (n = 13) and arrhythmias (n = 6). Of the 83 patients with renal failure who were not on dialysis, two had elevated cTnI, but neither of them had any adverse cardiac events or died during the six-month follow-up period, and two patients had myocardial infarction, but neither of them had elevated troponin I levels. Logistic regression analysis was not performed to determine whether cTnI was an independent predictor of acute myocardial infarction, unstable angina or cardiac death.
In the present study, in which patients with chronic stable angina were not included, we examined 126 patients with end-stage renal disease treated with long-term hemodialysis and prospectively followed for two years for mortality and hospital admissions. Mortality in patients with chronic renal failure and elevated cTnI was compared with that in patients with chronic renal failure and normal cTnI by Kaplan-Meier survival analysis, and no difference was found between the two groups. The ROC plots failed to demonstrate serum cTnI as a useful tool for predicting any of the end points. In addition, logistic regression analysis was used to examine the independent predictive value of cTnI for all four end points, and it was found that cTnI was not an independent predictive of any end point. Older age and lower serum albumin levels were independent predictors of all-cause mortality, whereas a history of myocardial infarction was an independent predictor of cardiac mortality. A lower serum sodium level was an independent predictor of all-cause hospital admissions, whereas hypertension and previous myocardial infarction were independent predictors of cardiac hospital admissions. Moreover, Cox regression analysis showed that the cTnI level was not predictive of the time to death; rather, older age and lower serum sodium levels were predictive of the time to death.
Study limitations. Unstable angina and acute myocardial infarction were not used as study end points. The diagnosis of unstable angina is purely subjective, and the diagnosis of acute myocardial infarction in patients with chronic renal failure may remain uncertain because of the more ambiguous symptoms and the presence of abnormal baseline ECGs in a number of these patients. The biochemical diagnosis of acute myocardial infarction is also often difficult in these patients because of the frequent false-positive elevations of the cardiac markers. Therefore, the study protocol was limited to the definitive end points of death and hospital admissions.
Conclusions. Cardiac troponin I has a limited role in predicting future mortality and hospital admissions in asymptomatic, ambulatory patients with chronic renal failure treated with long-term hemodialysis. The findings of this study should not be extrapolated to patients with chronic renal failure who have cardiac symptoms.
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. Y.-M. Wang and K.-N. Lai Use of Cardiac Biomarkers in End-Stage Renal Disease J. Am. Soc. Nephrol., September 1, 2008; 19(9): 1643 - 1652. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. A. Khan, B. R. Hemmelgarn, M. Tonelli, C. R. Thompson, and A. Levin Prognostic Value of Troponin T and I Among Asymptomatic Patients With End-Stage Renal Disease: A Meta-Analysis Circulation, November 15, 2005; 112(20): 3088 - 3096. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Babuin and A. S. Jaffe Troponin: the biomarker of choice for the detection of cardiac injury Can. Med. Assoc. J., November 8, 2005; 173(10): 1191 - 1202. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S Troyanov, Q H Ly, E Schampaert, H Ammann, G Lalumiere, F Madore, and S Querin Diagnostic specificity and prognostic value of cardiac troponins in asymptomatic chronic haemodialysis patients: a three year prospective study Heart, September 1, 2005; 91(9): 1227 - 1228. [Full Text] [PDF]
|
|
|
|
 |
|
 D. Duman, S. Tokay, A. Toprak, D. Duman, A. Oktay, and I. C. Ozener Elevated cardiac troponin T is associated with increased left ventricular mass index and predicts mortality in continuous ambulatory peritoneal dialysis patients Nephrol. Dial. Transplant., May 1, 2005; 20(5): 962 - 967. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Boulier, I. Jaussent, N. Terrier, F. Maurice, J.-P. Rivory, L. Chalabi, A.-M. Boularan, C. Delcourt, A.-M. Dupuy, B. Canaud, et al. Measurement of circulating troponin Ic enhances the prognostic value of C-reactive protein in haemodialysis patients Nephrol. Dial. Transplant., September 1, 2004; 19(9): 2313 - 2318. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Hocher, R. Ziebig, C. Altermann, R. Krause, G. Asmus, C.-M. Richter, T. Slowinski, P. Sinha, and H.-H. Neumayer Different Impact of Biomarkers as Mortality Predictors among Diabetic and Nondiabetic Patients Undergoing Hemodialysis J. Am. Soc. Nephrol., September 1, 2003; 14(9): 2329 - 2337. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Beciani, A. Tedesco, A. Violante, S. Cipriani, M. Azzarito, A. Sturniolo, and G. Splendiani Cardiac troponin I (2nd generation assay) in chronic haemodialysis patients: prevalence and prognostic value Nephrol. Dial. Transplant., May 1, 2003; 18(5): 942 - 946. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. S. Jaffe Testing the wrong hypothesis: the failure to recognize the limitations of troponin assays J. Am. Coll. Cardiol., October 1, 2001; 38(4): 999 - 1001. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | SUBSCRIPTIONS | CURRENT ISSUE | PAST ISSUES | CARDIOSOURCE | SEARCH | HELP | FEEDBACK |
