Chronic heart failure (CHF) progression results in decreased ventricular function, reduced functional capacity, and increased hospitalization and mortality rates.1,2 In addition, the risk increases after hospitalization for decompensated CHF, when the readmission rate is close to 60%.3,4 The mortality rate for hospitalized patients is between 3.8% and 13.0%,4,5 with nearly half of deaths occurring within 5 years following the CHF diagnosis.
The underlying mechanism of CHF progression is ventricular remodeling.6 Although there are a number of causes for the remodeling, the loss of myocytes is the common pathway. Specific markers of myocyte damage are the cardiac troponins T (cTnT) and I (cTnI). The clinical testing of these markers has been responsible for a change in the definition of myocardial infarction.7 Increased levels of both cTnT and cTnI have been reported in patients with CHF and are associated with a worse clinical condition and a poor outcome.8-10
We analyzed whether the low-level loss of cTnT in patients with stable CHF might indicate subclinical progression of CHF and whether the detection of abnormal concentrations of this marker in serial blood sampling could be associated with a poorer long-term outcome.
Patients and methods
We prospectively assessed the relationship between elevated blood levels of cTnT in serial blood samples collected from 115 ambulatory patients with CHF and the incidence of the combined end point of death or hospitalization for worsening heart failure. We also analyzed the pattern of changes in functional capacity, left ventricular ejection fraction (LVEF), and distance covered in the 6-minute walk test. All patients from the heart failure clinic had had a diagnosis of CHF for 30 days or longer and an LVEF below 40%. The diagnosis of CHF was based on a previous history of either acute cardiogenic pulmonary edema, signs and symptoms of volume overload, exertion dyspnea, paroxysmal nocturnal dyspnea, or Framingham criteria during a hospitalization or when first assessed. We excluded patients who had kidney or liver disease, terminal malignancy, a baseline creatinine level above 2.5 mg/dL, or acute coronary syndrome within the previous 3 months.
The functional class was determined based on the metabolic equivalents (METS) attained while the patient was engaged in a particular activity. If the exercise capacity was 7 METS or greater, the patient was considered to be class I; if between 5 and 7 METS, class II; if between 2 and 5 METS, class III; and if below 2 METS, class IV.
On admission to the study and at the 12-month follow-up, subjects were evaluated by a 6-minute walk test and a two-dimensional and M-mode echocardiogram. Ninety-five percent of patients were taking angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, 84% were taking diuretics, 64% were taking antiplatelet agents, 65% were taking beta-adrenergic blocking agents, 63% were taking digoxin (Lanoxin), 41% were taking antithrombotics, 41% were taking amiodarone (Cordarone, Pacerone), and 29% were taking spironolactone (Aldactone).
A third-generation electrochemi-luminescent immunoassay of cTnT (Elecsystroponin-T, Roche Diagnostics, Indianapolis, IN) was used to determine cTnT levels from blood samples taken at baseline and at 3, 6, and 12 months. cTnT levels of 0.02 ng/mL or above were not considered normal, and the lower detection limit was 0.010 ng/mL. Patients who had normal cTnT levels were placed in group 1. Patients with one sample that was not normal were placed in group 2, and those with two or more abnormal samples were placed in group 3. We calculated the area under the receiver operating characteristics (ROC) curve correlating to the greatest concentration of cTnT taken during the follow-up period in an individual patient for the combined end point of death or hospitalization and for the correlation between the highest cTnT concentration and the cutoff of 0.02 ng/mL (Tn index).
Half the patients were in CHF functional class III—IV, and 61.3% had ischemic heart disease. Three fourths of patients were men, and the mean age was 61.2 ± 10.8 years. A median of four samples per patient were taken, and 96.5% had two or more cTnT measurements taken.
According to the grades of myocardial injury, 62 patients (54%) were classified as group 1, 28 (24%) were classified as group 2, and 25 (22%) were classified as group 3. The higher the group number, that is, the greater the number of abnormal samples, the higher the functional class and the number of prior hospitalizations for CHF. In addition, the greater the myocardial injury, the worse renal function was.
As shown in Figure 1, between the first blood test and the 12-month follow-up, functional class became worse in group 3, whereas there was no change in groups 1 and 2. Compared with group 1, the baseline LVEF was lower in groups 2 and 3. All groups showed an increase in LVEF from baseline to 1 year, which was statistically significant for groups 1 and 2 only. At baseline and at 12 months, group 3 patients walked significantly less than patients in groups 1 and 2 did.
In terms of long-term outcomes, worsening CHF was more common in patients with more myocardial damage: 27.4% of patients in group 1, 64.3% of patients in group 2, and 88% of patients in group 3 (P < .001). Hospital admission for CHF was also required more frequently with more abnormal samples: for 76% of patients in group 3, 50% of patients in group 2, and 21% of patients in group 1 (P = .001). The death rate for patients in group 1 was 8.1%, 28.6% for patients in group 2, and 12% for patients in group 3 (P = .032).
As shown in Figure 2, 63% of patients in group 1 had hospitalization-free survival at 18 months, as did 46% of patients in group 2 and 17% of patients in group 3 (P = .001). This pattern remained for those patients with clinically diagnosed ischemic disease (event-free survival in group 1 compared with the other two groups was 57% and 33%, respectively; P = .011) and for patients with no ischemic disease (75% versus 0%, respectively;
P = .005). In individuals with a creatinine level below 1.2 mg/dL (50%), 68% of patients in group 1 had event-free survival compared with 47% in groups 2 and 3 (P = .001).
To explore the value of cTnT as a continuous variable, we built a ROC curve for the highest concentration of cTnT for an individual patient. The corresponding area under the curve was 0.68 to prognosticate the combined end point of death or hospitalization, and 0.020 ng/mL was chosen as the cutoff (sensitivity, 62%; specificity, 71%). The incidence of events based on the level of Tn index (< 1, 1—1.9, 2–2.9, 3–3.9, ≥ 4) was 29%, 54%, 60%, 80%, and 78%, respectively (P = .006). Each level of Tn index was associated with decreased survival free of hospitalization for CHF in the proportional hazards model (Table).
The prognostic value of cTnT monitoring was explored by a mean of two Cox proportional hazards models. The variables that independently correlated with the primary end point in the first model were the number of abnormal cTnT results (hazard ratio [HR], 1.6; 95% confidence interval [CI], 1.1—2.4), functional class III–IV (HR, 2.3; 95% CI, 1.1–4.6), and hospitalization for CHF in the previous year (HR, 2.1; 95% CI, 1.1–4.1). In the second model, the Tn index was incorporated, and the variables associated with events were functional class III–IV (HR, 2.87; 95% CI, 1.4–5.8), Tn index (1.09; 95% CI, 1.01–1.18), and hospitalization in the previous year (HR, 2.39; 95% CI, 1.3–4.5). An increase of 0.020 ng/mL of cTnT, therefore, was associated with a 9% increased incidence in the combined end point of death or hospitalization for worsening heart failure.
The main finding of this study
was that serial measurements of cTnT levels help in the detection of myocardial damage in outpatients with CHF and can identify a long-term high-risk population.
Troponins and CHF have been related in different clinical settings. In other studies, myocardial injury was present in 25% to 50% of stable patients with CHF.10,11 In decompensated CHF, high levels of troponin correlated with a poorer clinical and hemodynamic condition, a more severely
remodeled left ventricle, and poorer short-term and long-term prognoses.9,12-14 Little is known, however, about serial cTnT measurements. In patients with nonischemic dilated cardiomyopathy, three evolutionary patterns of serial cTnT monitoring were found in a retrospective study: normal levels, high initial levels that fell during follow-up, and consistently high levels.8 The lowest long-term survival correlated with consistently high levels of cTnT. For patients with ischemic heart disease greater than 60%, we contribute two new methods of monitoring cTnT:
according to the quantitative level
and according to the number of abnormal tests.
Patients with a higher incidence of worsening heart failure and death were identified by measurements of cTnT levels. Patients who had an increased number of abnormal cTnT levels had decreased hospitalization-free survival at 18 months. Both the number of
samples with increased cTnT levels and cTnT levels in any sample were independently associated with the combined end point. This fact has been
previously shown in other studies, particularly in decompensated CHF.11-14 In one study, detectable levels of cTnI correlated with worsening left ventricular dysfunction, increased brain natriuretic peptide (BNP) levels, a worse prognosis in advanced CHF, and impaired hemodynamics.10 The combined abnormal levels of cTnI and BNP also correlated with a substantial increase in long-term mortality.
In our study, the renal impairment may be related to the high cTnT levels. The increased creatinine level, how-ever, was mild to moderate, and the third-generation assay used in the study was highly specific. cTnT level has been shown to be an independent predictor of mortality in patients with renal impairment.15
We speculate that two different but additive mechanisms of myocardial damage can be found during the evolution of CHF. In a clinically stable condition, up to half of patients may have ongoing cellular injury, with a low increase in troponins, promoted
by a loss of the cytoplasmatic pool due to higher cell membrane permeability or as a result of other factors, causing apoptosis or a small amount of necrosis. These patients are at high risk because they are more susceptible to myocardial injury. During the decompensated state, the release of cTnT can be triggered by a new process associated with myocardial injury (subendocardial ischemia or subclinical microinfarct) or by minor events in patients with more susceptible myocardium (diet transgression or fluid overload). Independent of the underlying mechanism, the short-term and long-term prognoses of this subgroup are worse.
Our study showed that increased cTnT levels in one or more samples from patients with stable CHF were associated with an increase in the incidence of the combined end point of hospitalization or death. In addition, the number of abnormal cTnT samples and a quantitative Tn index were also independent predictors of events. These results indicate that monitoring for myocardial injury may help to identify high-risk patients.
The challenge for future investigations should be to test the efficacy of several drugs on the production or mitigation of myocardial injury and their impact on risk and clinical benefit.