The development of left ventricular remodeling (LVR) after myocardial infarction is associated with a high risk of heart failure and death. LVR is difficult to predict, and limited information is available on the association of cardiac biomarkers and LVR. Growth-differentiation factor-15 (GDF-15) is induced during heart failure development and, in animals models, might influence the different processes involved in cardiac remodeling. The aim of the present investigation was to assess the association between the serum levels of GDF-15 within the first 24 hours of ST-segment elevation myocardial infarction and the development of subsequent LVR at 12 months of follow-up. This prospective study included 97 patients with ST-segment elevation myocardial infarction undergoing primary percutaneous coronary intervention. Echocardiography was performed in all patients within the first 96 hours of admission and at 12 months of follow-up. LVR was defined as a >20% increase in the left ventricular end-diastolic volume at 12 months of follow-up compared to baseline. Blood samples for the determination of GDF-15 and brain natriuretic peptide were obtained within the first 24 hours after symptom onset. According to the pre-established criteria, 21 patients (22%) had LVR. Patients with LVR had greater levels of GDF-15 at study entry (median 3,439 pg/ml, interquartile range 2,391 to 6,168 vs median 1998 pg/ml, interquartile range 1,204 to 3,067, respectively; p <0.001). Multivariate analysis showed that GDF-15 (odds ratio 10.1, 95% confidence interval 2.5 to 40.1, p <0.001) and treatment with angiotensin-converting enzyme inhibitors (odds ratio 3.9, 95% confidence interval 1.2 to 12.3, p <0.01) were independents predictors of LVR. Receiving operating characteristics analysis showed an area under the curve of 0.77 for GDF-15 (95% confidence interval 0.67 to 0.84, p <0.001). In conclusion, the results of the present study have identified GDF-15 as an independent marker of LVR in patients with ST-segment elevation myocardial infarction.
Growth-differentiation factor-15 (GDF-15) is a member of the transforming growth factor-β cytokine superfamily that plays a key role in development and regulation of the cellular responses to stress signals and inflammation and tissue repair after acute injury. In animal models, GDF-15 is induced in the heart in response to ischemia-reperfusion injury, pressure overload, and heart failure, possibly by way of proinflammatory cytokine and oxidative stress-dependent signaling pathways. The GDF-15 level at admission is a strong predictor of mortality in patients with ST-segment elevation myocardial infarction (STEMI) reperfused by primary angioplasty and is associated with decreased myocardial salvage and subsequent adverse clinical outcomes. The relation between increased GDF-15 levels and left ventricular remodeling (LVR) has not been assessed systematically in patients with STEMI. The aim of the present study was to assess the association between the serum levels of GDF-15 within the first 24 hours of STEMI and the development of subsequent LVR at 12 months of follow-up.
Methods
From July 1, 2007 to October 30, 2009, 173 patients admitted with a first STEMI to a tertiary care hospital and reperfused within 6 hours of symptom onset were prospectively included in the present study. The diagnosis of STEMI was established in the presence of typical ischemic chest pain lasting ≥30 minutes, electrocardiographic ST-segment elevation ≥0.1 mV in ≥2 leads in the same vascular territory or new-onset left bundle branch block, and an increase in troponin I to ≥2 times greater than the normal range. Of the 173 patients, 76 were excluded from the analysis because of a history of myocardial infarction (n = 23), hypertensive heart disease (n = 14), valvular heart disease (n = 9), a history of heart failure (n = 13), impaired renal function (glomerular filtration rate <60 ml/min/1.73 m 2 ; n = 14), and prostate cancer (n = 3). These exclusion criteria were chosen because these conditions can be associated with an elevation of left ventricular diastolic filling pressures, changes in the left ventricle end-diastolic volume, or nonspecific serum elevation of GDF-15. Hence, the data from 97 patients were analyzed. The ethics committee of our institution approved the research protocol. All patients gave written informed consent for inclusion. Percutaneous revascularization was considered successful if the residual stenosis was <50% and the flow in the culprit vessel after percutaneous coronary intervention was Thrombolysis In Myocardial Infarction grade ≥2. Every patient received medical treatment with aspirin, clopidogrel. and statins, and most received β blockers and angiotensin-converting enzyme inhibitors.
Within the first 96 hours of admission, echocardiography was performed in all patients. Follow-up was completed at 12 months with the performance of another echocardiographic study. The echocardiographic data were obtained using a commercially available system (Philips Medical Systems iE33, Andover, Massachusetts). Echocardiograms were obtained by experienced ultrasonographers and repeated by the same operators, who were unaware of the GDF-15 and brain natriuretic peptide (BNP) measurements and clinical data. A standard imaging protocol was used with apical 4- and 2-chamber views. Two-dimensional echocardiograms of the left ventricular short axis were recorded from the left parasternal region at the mitral valve, midpapillary muscle level, and apex. The left ventricular volume and ejection fraction were calculated using the modified Simpson’s rule. The ratio of early diastolic mitral flow velocity to early diastolic mitral annulus velocity was used as a surrogate for the left ventricular filling pressure. The mean value of 3 measurements of the technically best cardiac cycles was taken from each examination, and the left ventricular volume was corrected for the body surface area. The intraobserver variability in the evaluation of the end-diastolic volume and end-systolic volume was 3.5% and 3.3%, respectively. The corresponding interobserver variabilities were 3.9% and 5.5%. An analysis of diastolic function was performed to evaluate the mitral flow pattern with pulsed Doppler (E and A waves, E/A ratio, and deceleration time of the E wave). LVR at 12 months of follow-up was defined as a >20% increase in the left ventricular end-diastolic volume compared to baseline.
Blood samples were taken on admission. BNP was measured using a high-sensitivity, quantitative sandwich enzyme immunoassay (ALPCO, Salem, New Hampshire). In this assay, the lowest detection limit of BNP is 11.6 pg/ml. The coefficient of variation was 4.1% and 5.1% for intra- and interassay variability, respectively. The serum GDF-15 concentrations were measured using a commercial enzyme-linked immunosorbent assay (BioVendor GmB, Heidelberg, Germany). In this assay, the lowest detection limit of GFD-15 is 30.2 pg/ml. The coefficient of variation was 4.3% and 7.8% for intra- and interassay variability, respectively. Troponin I was determined immunoenzymatically using a technique based on sandwich enzyme-linked immunosorbent assay (Boehringer Mannheim, Mannheim, Germany). The coefficient of variation was 2.2% and 5.9% for the intra- and interassay variability, respectively.
The results for normally distributed continuous variables are expressed as the mean ± SD, and continuous variables with non-normal distribution are presented as the median and interquartile range. Categorical data are expressed as percentages. An analysis of normality of the continuous variables was performed using the Kolmogorov-Smirnov test. Differences between groups were assessed using an unpaired 2-tailed t test and the Mann–Whitney U test for continuous variables, as appropriate. Categorical data and proportions were analyzed using the chi-square or Fisher’s exact test, when required. The GDF-15 and BNP levels had a non-normal distribution and were therefore logarithmically transformed before regression analysis to fulfill the conditions required for this type of analysis. We assessed the independent predictors of LVR at 12 months of follow-up using a binary logistic regression analysis that included age, gender, infarct size, anterior location, left ventricular ejection fraction, treatment with angiotensin-converting enzyme inhibitors and β blockers at 12 months of follow-up, and the GDF-15 and BNP levels. Backward stepwise selection was used to derive the final model for which a significance level of 0.1 and 0.05 was chosen to exclude and include terms, respectively. The optimal cutoff point of GDF-15 to predict LVR was calculated using receiving operating characteristics analysis. Differences were considered statistically significant if the null hypothesis could be rejected with >95% confidence. All probability values are 2 tailed. The Statistical Package for Social Sciences, version 15 (SPSS, Chicago, Illinois) was used for all calculations.
Results
The demographic and clinical data of patients with and without LVR are presented in Table 1 . All patients were treated with primary percutaneous coronary intervention, and successful reperfusion was achieved in 95% of the patients. According to the pre-established criteria, 21 patients (22%) had LVR. No significant differences were found in the cardiovascular risk factors and standard biochemical results between the 2 groups. The use of angiotensin-converting enzyme inhibitors at 12 months of follow-up was more frequent in patients with LVR. The biomarker, GDF-15 (median 2,338 pg/ml, interquartile range 1,385 to 3,602, in the whole cohort) was greater in patients who had developed LVR at 12 months of follow-up than in those who had not. We found no differences in the BNP levels between the 2 groups ( Table 1 and Figure 1 ) . At hospital discharge, the patients received aspirin and/or clopidogrel and statins, together with β blockers and angiotensin-converting enzyme inhibitors. During the follow-up period, 5 patients required hospital readmission because of heart failure within 3 months after STEMI, and none of the patients developed a new myocardial infarction. Multivariate analysis showed that GDF-15 (odds ratio 10.1, 95% confidence interval 2.5 to 40.1, p <0.001) and treatment with angiotensin-converting enzyme inhibitors (odds ratio 3.9, 95% confidence interval 1.2 to 12.3, p <0.01) were independents predictors of LVR development at 12 months of follow-up. Receiver operating characteristic analysis for GDF-15 showed an area under the curve of 0.77 (95% confidence interval 0.67 to 0.84). An optimized cutoff point of 1,881 pg/ml showed 100% sensitivity and 50% specificity. The optimized point was obtained as the value that yielded the best sensitivity and specificity ( Figure 2 ) .
| Variable | LVR | p Value | |
|---|---|---|---|
| Yes (n = 21) | No (n = 76) | ||
| Age (years) | 60.3 ± 12.1 | 65.3 ± 11.1 | 0.07 |
| Men | 18 (86%) | 60 (79%) | 0.75 |
| Hypertension (>140/90 mm Hg) | 12 (57%) | 40 (53%) | 0.8 |
| Hypercholesterolemia (>5.17 mmol/L) | 7 (33%) | 41 (54%) | 0.14 |
| Smoker | 9 (43%) | 30 (39%) | 0.8 |
| Diabetes mellitus | 10 (48%) | 22 (29%) | 0.12 |
| Multivessel coronary artery disease | 10 (48%) | 34 (45%) | 0.81 |
| Left ventricular ejection fraction (%) | 56 ± 10 | 57 ± 12 | 0.9 |
| Anterior wall acute myocardial infarction location | 10 (48%) | 32 (42%) | 0.65 |
| Killip class II–IV | 3 (14%) | 5 (7%) | 0.36 |
| Pain to reperfusion time (min) | 244.3 ± 67.5 | 242.9 ± 70 | 0.93 |
| Body mass index (kg/m 2 ) | 32 ± 4 | 35 ± 8 | 0.10 |
| Treatment at 12 months | |||
| Aspirin | 21 (100%) | 75 (99%) | 0.9 |
| Clopidogrel | 16 (76%) | 57 (75%) | 0.9 |
| Statins | 21 (100%) | 76 (100%) | 1 |
| Angiotensin-converting enzyme inhibitors | 14 (67%) | 24 (32%) | 0.004 |
| Diuretics | 6 (29%) | 14 (18%) | 0.36 |
| β Blockers | 21 (100%) | 68 (89%) | 0.12 |
| Biochemistry findings | |||
| Creatinine (mg/dl) | 0.95 ± 0.2 | 0.96 ± 0.2 | 0.83 |
| Total cholesterol (mg/dl) | 182 ± 41 | 188 ± 50 | 0.67 |
| Peak troponin I (ng/ml) | 71 ± 16 | 61 ± 23 | 0.07 |
| Growth-differentiation factor-15 (pg/ml) | 3,439 (2,391–6,168) | 1,998 (1,204–3,067) | <0.001 |
| Brain natriuretic peptide (pg/ml) | 473 (406–615) | 450 (421–485) | 0.18 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
