Left ventricular (LV) remodeling after myocardial infarction (MI) indicates a high risk of heart failure and death. However, LV remodeling is difficult to predict, and limited information is available on the association of cardiac biomarkers with LV remodeling. Our aim was to study the association of B-type natriuretic peptide (BNP), cardiac troponin I (cTnI), and C-reactive protein with LV remodeling after MI. We designed a prospective multicenter study including 246 patients with a first anterior Q-wave MI. Serial echocardiographic studies were performed at hospital discharge and 3 months and 1 year after MI; quantitative analysis was performed at a core echocardiographic laboratory. Blood samples for determination of BNP, cTnI, and C-reactive protein levels were obtained at hospital discharge and the 1-month, 3-month, and 1-year follow up visits. One-year echocardiographic follow-up was obtained in 226 patients. End-diastolic volume increased from 52.3 ± 13.8 ml/m 2 at baseline to 62.3 ± 18.4 ml/m 2 at 1 year (p <0.0001); LV remodeling (>20% increase in end-diastolic volume) was observed in 87 patients (38%). At baseline, we found significant univariate relations between LV remodeling and the 3 biomarkers. During follow-up, high BNP levels and persistently detectable levels of cTnI were associated with LV remodeling. In multivariate analysis, none of the 3 biomarkers at baseline was independently predictive of LV remodeling. In contrast, during follow-up, high BNP and positive cTnI were independently associated with LV remodeling. In conclusion, circulating cardiac biomarkers may reflect pathophysiologic processes implicated in LV remodeling after MI. Determination of BNP and cTnI during follow-up can help refine risk stratification.
Degree of left ventricular (LV) dilation after myocardial infarction (MI) is useful for risk stratification because LV remodeling is a well-documented surrogate, indicating a high risk for heart failure and cardiovascular death. The ability to predict LV remodeling is, however, difficult; although several risk factors are identified including infarct size and anterior infarct location, variability persists, with some patients developing important LV dilation despite the absence of evident risk factors and, conversely, other patients, classified as high risk during the initial evaluation, not developing LV remodeling. We hypothesized that levels of 3 cardiac biomarkers, B-type natriuretic peptide (BNP), cardiac troponin I (TnI), and C-reactive protein (CRP), might correlate with structural changes occurring in the left ventricle in response to an MI, and that repeated assessment of these biomarkers in stabilized patients after MI could help clinicians predict LV remodeling more effectively. These 3 biomarkers assess different pathways implicated in the pathogenesis of LV remodeling: BNP is increased in response to LV overload; TnI increase may indicate myocyte injury; and CRP is a marker of inflammation. Although increased levels of each of these markers are associated with major cardiac events after MI, no or at best limited information is available on their association with LV remodeling. We designed the Remodelage Ventriculaire–2 (REVE-2) study to test this hypothesis prospectively in patients after a first anterior MI. Repeat echocardiographic examinations were performed during the first year after MI to assess LV remodeling and serial blood sampling was performed to determine cardiac biomarkers.
Methods
REVE-2 was a multicenter study that enrolled 246 patients with anterior wall Q-wave MI from 8 centers in France ( Appendix ) from February 2006 to September 2008. Inclusion criteria were hospitalization within 24 hours after symptom onset and ≥3 LV segments of the infarct zone that were akinetic at predischarge echocardiography. Exclusion criteria were inadequate echographic image quality, life-limiting noncardiac disease, significant valvular disease, or previous Q-wave MI. The research protocol was approved by the ethics committee of the Centre Hospitalier et Universitaire de Lille Lille, France), and written informed consent was obtained from each patient. The protocol required serial echographic studies at hospital discharge (day 3 to day 7) and 3 months and 1 year after MI; serial blood sampling was performed at hospital discharge (day 3 to day 7) and 1 month, 3 months, and 1 year after MI. For those patients who did not return at the designated times, information on their clinical status was obtained through telephone interviews.
A standard echographic imaging protocol was used based on apical 4- and 2-chamber views; 2-dimensional echocardiograms of the LV short axis were recorded from the left parasternal region at the mitral valve, midpapillary muscle, and apex. All echocardiograms were analyzed at the Lille Core Echocardiographic Laboratory (Lille, France), as previously described. LV volumes and ejection fraction were calculated using a modified Simpson rule. To evaluate regional systolic function, the left ventricle was divided according to a 16-segment model. For each segment, wall motion was scored from 1 (normal) to 4 (dyskinetic) and the wall motion score index was derived.
For each patient, plasma (ethylenediaminetetra-acetic acid was used as anticoagulant) and serum were collected in glass tubes at the 4 time points indicated earlier. Plasma and serum were processed within 2 hours, and samples were divided into aliquots and stored at −80°C. Samples underwent no more than 2 freeze/thaw cycles before analysis in a core laboratory (Lille, France).
BNP was measured in plasma samples using a fully automated 2-site sandwich immunoassay on an Advia Centaur (Siemens Diagnostic, Zurich, Switzerland). The lowest concentration measurable with this assay with a ≤20% coefficient of variation is 2.5 pg/ml. The precision of this technique is 2.3% to 4.7%. Cardiac TnI was measured in plasma samples using a 3-site sandwich immunoassay on an Advia Centaur. The lowest concentration measurable with this assay with a ≤20% coefficient of variation is 0.015 ng/ml. Cardiac TnI in the 99th percentile in healthy subjects is 0.05 ng/ml, with 10% coefficient of variation for this threshold. The precision of this technique is 2.7% to 5.3%. CRP was measured in serum samples using a sensitive latex-enhanced immunonephelometric method with the automated BNII nephelometer (Siemens Diagnostic). The minimum sensitivity for this assay is 0.15 mg/L.
STATA 9.0 (STATA Corporation, College Station, Texas) was used for statistical analysis. Results are presented as mean ± SD, median with 25 th and 75 th percentiles, or frequency expressed as a percentage. Levels of BNP, cTnI, and CRP were used as continuous variables except for cTnI levels at 1 month, 3 months, and 1 year, which were categorized as positive (≥0.05 ng/ml) or negative (<0.05 ng/ml). In case of recurrent MI during follow-up, the biological measurements obtained <1 month after the repeat acute event were excluded from statistical analysis (n = 1 patient). Variables with skewed distribution were log-transformed before being used as continuous variables in statistical analyses. Continuous variables were compared with paired or unpaired Student’s t test or with simple linear regression, as appropriate. Discrete variables were compared using chi-square analysis. A p value <0.05 was considered statistically significant. Independent correlates of change in end-diastolic volume were identified by multiple linear regression. Variables with a p value <0.10 on univariate analysis were entered in the multivariate model. Colinearity was excluded by a correlation matrix between candidate predictors. Linearity and continuity assumptions were assessed by plotting residuals against independent variables. One multivariate model was constructed for each study period (baseline, 1 month, 3 months, and 1 year). To illustrate our findings, LV remodeling was defined as a >20% increase in end-diastolic volume from baseline to 1-year follow-up examination, a definition used previously to indicate severe remodeling. The optimal BNP cut-off value for prediction of LV remodeling was determined by receiver operating characteristic curve analysis. Sample size calculations were based on a 2-sided alpha error of 0.05 and 80% power. Based on unpublished preliminary data from our institution showing BNP levels of 68 ± 48 pg/ml 1 month after MI, we calculated that 200 patients would provide sufficient power to detect a 30% difference in BNP between patients with LV remodeling and patients without LV remodeling.
Results
Characteristics of the 246 patients who formed the study population are listed in Table 1 .
| Age (years) | 57.0 ± 13.8 |
| Women | 46 (19%) |
| Body mass index (kg/m 2 ) | 27.1 ± 4.6 |
| Hypercholesterolemia (treated or total cholesterol >250 mg/dl) | 82 (33%) |
| Hypertension (treated or blood pressure >160/90 mm Hg) | 89 (36%) |
| Current smokers | 116 (47%) |
| Diabetes mellitus | 51 (21%) |
| Previous angina pectoris | 14 (6%) |
| Previous percutaneous coronary intervention | 6 (2%) |
| Initial reperfusion therapy | |
| Primary percutaneous coronary intervention | 128 (52%) |
| Thrombolysis alone | 28 (11%) |
| Thrombolysis and rescue percutaneous coronary intervention | 59 (24%) |
| No reperfusion | 31 (13%) |
| Killip class ≥II | 79 (32%) |
| Coronary angiography during hospitalization | 243 (99%) |
| Multivessel disease | 98 (40%) |
| Percutaneous coronary intervention during hospitalization | 212 (86%) |
| Stent implantation | 208 (85%) |
Baseline echocardiographic study was performed at a mean of 4.0 ± 1.5 days after the MI. During the 1-year follow-up period, 3 patients died (all from cardiovascular causes) and 1 patient underwent heart transplantation. In addition, 3 patients had recurrent MIs during the follow-up period. Echocardiographic follow-up was achieved in 226 of the 242 eligible patients (93%, 1 patient was lost to follow-up and 15 patients declined the repeat echographic studies or had inadequate image quality for measurement of LV volumes). Table 2 lists changes in major echocardiographic variables throughout the study period. LV remodeling was documented by a significant increase in end-diastolic volume from baseline to 3 months (p <0.0001) and from 3 months to 1 year (p <0.0001). Similar findings were observed for end-systolic volume from baseline to 3 months (p = 0.0012). Median percent LV remodeling, defined as percent change in end-diastolic volume from baseline to 1 year, was 14%. LV remodeling was observed in 87 patients (38% of population with echocardiographic follow-up) when it was defined as a >20% increase in end-diastolic volume from baseline to 1 year. Major medications used during the study are presented in Table 2 . The proportion of patients who received secondary preventive treatments was very large and remained stable across the 3 time points, except for a slight decrease in the proportion of patients receiving angiotensin-converting enzyme inhibitors with a parallel increase in the proportion of patients receiving angiotensin II receptor blockers. The most frequently prescribed β blocker was bisoprolol (n = 162 patients) with doses of 5.5 ± 3.3, 5.8 ± 3.3, 6.7 ± 3.3, and 7.1 ± 3.1 mg at baseline, 1 month, 3 months, and 1 year, respectively. The most frequently prescribed angiotensin-converting enzyme inhibitors were perindopril (n = 101 patients) with doses of 4.8 ± 2.5, 5.2 ± 2.3, 5.6 ± 2.4, and 6.2 ± 2.4 mg at baseline, 1 month, 3 months, and 1 year, respectively, and ramipril (n = 96 patients) with doses of 6.1 ± 3.1, 6.3 ± 3.2, 7.1 ± 3.2, and 7.7 ± 2.9 mg at baseline, 1 month, 3 months, and 1 year, respectively.
| Variable | Baseline | 3 Months | 1 Year |
|---|---|---|---|
| Heart rate (beats/min) | 71.3 ± 13.8 | 61.3 ± 10.9 ⁎ | 62.4 ± 11.7 ⁎ |
| Systolic blood pressure (mm Hg) | 110.3 ± 15.3 | 120.9 ± 18.8 ⁎ | 124.7 ± 16.7 ⁎ ∥ |
| Diastolic blood pressure (mm Hg) | 62.3 ± 11.4 | 72.3 ± 10.2 ⁎ | 73.2 ± 10.0 ⁎ |
| End-diastolic volume (ml/m 2 ) | 52.3 ± 13.8 | 59.5 ± 16.5 ⁎ | 62.3 ± 18.4 ⁎ ¶ |
| End-systolic volume (ml/m 2 ) | 26.6 ± 10.4 | 28.2 ± 11.8 † | 29.0 ± 14.5 ‡ |
| Ejection fraction | 49.7 ± 8.4 | 53.7 ± 9.1 ⁎ | 55.2 ± 10.2 ⁎ ¶ |
| Wall motion score index | 1.90 ± 0.15 | 1.71 ± 0.25 ⁎ | 1.64 ± 0.28 ⁎ ¶ |
| Medications | |||
| Aspirin | 224 (99%) | 219 (97%) | 216 (96%) |
| Clopidogrel | 219 (97%) | 211 (93%) | 210 (93%) |
| β Blockers | 220 (97%) | 213 (94%) | 213 (94%) |
| Angiotensin-converting enzyme inhibitors | 220 (97%) | 209 (92%) § | 195 (86%) ⁎ # |
| Angiotensin II receptor blockers | 1 (0.5%) | 11 (5%) † | 22 (10%) ⁎ # |
| Aldosterone antagonists | 76 (34%) | 75 (33%) | 76 (34%) |
| Diuretics | 54 (24%) | 52 (23%) | 64 (28%) |
| Statins | 213 (94%) | 217 (96%) | 215 (95%) |
As presented in Table 3 , a progressive decrease in BNP levels was observed from a median value of 110 pg/ml at baseline to 35 pg/ml at 1 year; differences in BNP levels were statistically significant between each time point. Cardiac TnI levels were increased in all patients at baseline. One month after the MI, 19% of patients still had detectable levels of cTnI (≥0.05 ng/ml). The median cTnI value in patients with detectable levels at 1-month follow up was 0.07 ng/ml. The proportion of patients who were positive for cTnI decreased to 13% after 3 months and to 7% at 1 year. In total, 23% of patients had ≥1 sample during the long-term follow-up phase (1 month to 1 year) that was positive for cTnI. CRP levels were high at baseline and returned to normal levels during follow-up.
| Timing of Blood Sampling | ||||
|---|---|---|---|---|
| Baseline | 1 Month | 3 Months | 1 Year | |
| B-type natriuretic peptide (pg/ml) | 110 (61–214) | 85 (50–170) ⁎ | 48 (27–104) † ‡ | 35 (17–77) † ‡ § |
| (n = 221) | (n = 218) | (n = 222) | (n = 221) | |
| Cardiac troponin I (ng/ml) | 8.98 (3.19–21.39) | |||
| ≥0.05 ng/ml | 204 (100%) | 39 (19%) † | 27 (13%) † ‡ | 15 (7%) † ‡ § |
| (n = 204) | (n = 203) | (n = 210) | (n = 211) | |
| C-reactive protein (mg/L) | 25.00 (8.76–46.8) | 1.41 (0.69–3.38) † | 1.13 (0.57–2.78) † ‡ | 1.05 (0.52–2.83) † ‡ |
| (n = 219) | (n = 219) | (n = 222) | (n = 220) | |
We first investigated the relations between cardiac biomarkers and LV remodeling as a quantitative variable defined as percent change in end-diastolic volume from baseline to 1-year follow-up ( Table 4 ). There was a significant relation between BNP levels and LV remodeling at each of the 4 time points. The association was mild at baseline and stronger during follow-up, particularly after 3 months. Baseline levels of cTnI were mildly associated with LV remodeling. During follow-up, patients who had persistently detectable levels of cTnI were at higher risk of LV remodeling, with stronger relations observed after 1 month and 3 months. Regarding CRP, there was a significant but modest positive association between baseline levels of CRP and LV remodeling. CRP levels during follow-up were not associated with LV remodeling. Similar results were obtained when LV remodeling was defined as a qualitative variable. Figure 1 displays levels of cardiac biomarkers at each time point in patients without or with LV remodeling defined as a >20% increase in end-diastolic volume from baseline to 1 year.
| Timing of Blood Sampling | ||||
|---|---|---|---|---|
| Baseline | 1 Month | 3 Months | 1 Year | |
| B-type natriuretic peptide | ||||
| β Coefficient | 0.207 | 0.291 | 0.373 | 0.276 |
| p Value | 0.002 | <0.0001 | <0.0001 | <0.0001 |
| Cardiac troponin I | ||||
| β Coefficient | 0.194 | 0.247 | 0.239 | 0.171 |
| p Value | 0.005 | <0.0001 | 0.001 | 0.013 |
| C-reactive protein | ||||
| β Coefficient | 0.185 | 0.009 | 0.060 | −0.030 |
| p Value | 0.006 | 0.899 | 0.377 | 0.659 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
