Acute coronary syndrome (ACS) remains one of the most important causes of mortality, despite advances in diagnosis and treatment.1 It is therefore important to perform risk assessment in ACS patients to improve long-term survival.

Matrix metalloproteinases (MMPs), a component of the extracellular matrix, play a role in tissue remodeling in both physiological and pathophysiological processes.2 Dysregulation of MMP expression contributes to tissue proteolysis and thus to thinning and rupturing of the atherosclerotic plaque, which is one of the main scenarios for the occurrence of acute coronary syndrome.3 It has been shown that MMP-9 levels increase significantly in various vascular conditions, especially acute myocardial infarction (AMI).4

Many indicators of adverse prognosis after AMI exist, some of which can be used not only for prognostic assessment but also for early diagnosis, such as creatinine kinase and troponins, which indicate cardiomyocyte necrosis.5–7 Previous studies showed that inflammation markers like C-reactive protein and serum amyloid A can be used for risk determination in post-AMI patients.8

It is thought that MMP-9 may play a role not only in the disease process but also in predicting adverse events in patients with coronary artery disease (CAD).9,10 MMP-9 may affect myocardial fibrosis via transforming growth factor beta signaling-induced periostin and connective tissue growth factor expression.11 In addition, MMP-9 was found to be three times higher in the infarct site in post-AMI patients, suggesting that MMP-9 is an important indicator for cardiac remodeling and prognosis in patients with AMI.12 However, existing data on the specific impact of baseline plasma MMP-9 levels and its utility as a prognostic marker after AMI are limited and have been conflicting.13–15 This study was designed to determine the clinical value of MMP-9 in predicting two-year adverse cardiovascular events in ST-elevation myocardial infarction (STEMI) patients who underwent primary percutaneous coronary intervention (PCI).

After removing cases based on the exclusion criteria, a total of 204 patients remained in the study. There were no differences between the high and low MMP-9 groups in terms of baseline characteristics or laboratory values (Tables 1 and 2), and data on coronary intervention were similar between the study groups. Moreover, cardiac medical therapy after discharge was prescribed in around 90% of both groups, without statistical difference (Table 3). Nonetheless, cardiogenic shock (31.4% vs. 1.8%, p<0.001), need for inotropic therapy (34.3% vs. 3.0%, p<0.001) and intra-aortic balloon pump (IABP) use (14.3% vs. 1.8%, p=0.003) were more frequent in patients with no-reflow compared to patients with reflow.

In our study, the cardiovascular mortality rate was 7.3% in-hospital, 9.3% at one-year follow-up and 11.2% at two-year follow-up. Although in-hospital cardiovascular mortality was higher in the high MMP-9 group than the low MMP-9 group, this was not statistically significant. However, the high MMP-9 group had a significantly higher incidence of cardiopulmonary resuscitation, use of inotropes, cardiogenic shock, IABP use and no-reflow phenomenon. Furthermore, cardiogenic shock (31.4% vs. 1.8%, p<0.001), need for inotropic therapy (34.3% vs. 3.0%, p<0.001) and IABP use (14.3% vs. 1.8%, p=0.003) were more frequent in patients with no-reflow compared to patients with reflow. When long-term outcomes were analyzed, one-year cardiovascular mortality was significantly higher in the high MMP-9 group (13.7% vs. 4.8% for the low MMP-9 group, p=0.030), unlike in-hospital outcomes (10.8% vs. 3.9%, p=0.060). Furthermore, when the follow-up period was extended to two years, the difference in cardiovascular mortality between the groups was also statistically significant (17.6% vs. 4.9%, p=0.004; Table 4), with more serious outcomes for the high MMP-9 group. In parallel with the above-mentioned results, we found that the high MMP-9 group had 3.5-fold higher odds for two-year cardiovascular mortality than the low MMP-9 group, independently of age and LVEF (Table 5).

From the secondary endpoints, advanced heart failure was compared between the groups in order to investigate the ability of MMP-9 to predict left ventricular remodeling. While there was no significant difference at the end of the first year (9.8% vs. 3.9%, p=0.098), there was a statistically significant higher prevalence of advanced heart failure in the high MMP-9 group at two-year follow-up (16.7% vs. 5.9%, p=0.015).

In ROC analysis, the area under the curve for MMP-9 ≥12.92 ng/ml was 0.675 for predicting two-year-cardiac mortality with 74.4% sensitivity and 55.6% specificity (Figure 1). The Kaplan-Meier survival plot for two-year CV mortality for the high and low MMP-9 groups is presented in Figure 2.

Figure 1.

Receiver operating characteristic (ROC) curve showing the ability of matrix metalloproteinase-9 level to predict two-year cardiovascular mortality. AUC: area under the ROC curve; CI: confidence interval; MMP: matrix metalloproteinase.

(0.15MB).

Figure 2.

Kaplan-Meier analysis showing two-year cardiovascular mortality according to matrix metalloproteinase-9 level. MMP-9 cut-off: 12.92 pg/ml. MMP: matrix metalloproteinase.

(0.09MB).

In our study, higher MMP-9 levels were associated with a marked increase in two-year cardiovascular mortality. Furthermore, as a secondary outcome, cases of advanced heart failure were significantly more frequent in the high MMP-9 group during two-year follow-up. Another notable aspect of the study was that, while there was no significant difference in in-hospital cardiovascular mortality between the groups, it was significantly different in long-term follow-up and was more marked as the follow-up period increased. Furthermore, it was noted that high MMP-9 was one of the independent predictors of two-year cardiovascular mortality (odds ratio [OR]: 3.5; 95% confidence interval [CI]: 1.12-11.35). Age, no-reflow phenomenon and LVEF were the other independent predictors of two-year cardiovascular mortality.

Levels of MMP-9, which is associated with plaque instability and rupture and with left ventricular remodeling due to its role in collagen-cleaving and degrading extracellular matrix proteins in plaque, have proved useful in predicting adverse cardiovascular events. Kaptoge et al.20 reported that high MMP-9 levels were associated with adverse cardiovascular events in a population-based study. In another prospective analysis, MMP-9 levels were high in patients with AMI, but were not identified as independent predictors due to their close relationship with other cardiovascular risk factors, such as C-reactive protein and smoking.15 Previous studies had established that serum MMP-9 levels could also be associated with both severity and progression of CAD.21,22 Furthermore, Blackenberg et al.23 demonstrated that plasma MMP-9 was a predictor of future CV mortality in patients with a history of CAD, independently of potential confounders. Although MMP-9| contributes to determining prognosis in population-based studies and individuals with coronary artery disease, there are conflicting data on predicting adverse CV events after AMI. For this reason, we felt the need to perform a study to clarify this situation.

As mentioned above, matrix metalloproteinases promote plaque instability and left ventricular remodeling, and in this study we set out to analyze the clinical responses to these two effects. Firstly, the role of MMP-9 in plaque instability was investigated through different parameters affecting CV mortality and morbidity. The effect of MMP-9 levels on plaque rupture was noted and the no-reflow phenomenon was seen more often in patients with ruptured plaques in previous studies.24,25 For this reason, in our study, the relation of MMP-9 levels with no-reflow was investigated, and no-reflow was statistically more frequent in the high MMP-9 group (p=0.048), as expected. Moreover, inadequate coronary flow may cause a domino effect, and no-reflow during the procedure may lead to cardiogenic shock or the need for IABP or inotropes, development of advanced heart failure and ultimately cardiovascular mortality. On the basis of this hypothesis, these consequences were analyzed separately in our study, which showed that the use of IABP (p=0.019), inotropic therapy (p=0.023) and cardiogenic shock (p=0.012) were more frequent in the high MMP-9 group. Furthermore, no-reflow was an independent predictor of two-year cardiovascular mortality (OR: 3.7; 95% CI: 1.19-11.15).

Many indicators of adverse prognosis after AMI exist, one of the most important of which is left ventricular remodeling, the second major effect of MMP-9. Infarcted and non-infarcted areas of the ventricle are affected by remodeling, which is caused by maladaptive changes to cellular and extracellular elements of the myocardium. MMP-9 has a role in regulating vascular tone and platelet aggregation, which are involved in adverse left ventricular remodeling.26 Especially following AMI, MMP levels are associated with LV function, as shown by echocardiography, establishing that they potentially affect LV remodeling. Kelly et al.27 showed that higher MMP-9 levels were associated not only with more severe impairment of LV function, but also with more extensive LV remodeling in the follow-up of post-AMI patients. In addition, Wagner et al.28 followed AMI patients treated with primary mechanical reperfusion for two years and found that MMP-9 levels were predictive of late-onset congestive heart failure after AMI. This is supported by other studies, which show that ACS patients with ruptured plaques that contain higher MMP levels compared to those without have larger infarct sizes, which causes more remodeling.29–31 Similarly to the above-mentioned studies, while there was no statistically significant difference in occurrence of advanced heart failure between the high and low MMP-9 groups at one-year follow-up in our study, there was a substantial increase in advanced heart failure correlating with follow-up duration in both groups, due to long-term LV remodeling progression (one-year follow-up, MMP-9 high vs. low group: 9.8% vs. 3.9%, p=0.097; two-year follow-up, MMP-9 high vs. low group: 16.7% vs. 5.9%, p=0.015). In other studies examining the relationship between MMP levels and remodeling, lower MMP-9 levels were seen in patients suffering from dilated cardiomyopathy, ischemic cardiomyopathy with LV assist device, or after cardiac resynchronization therapy.32 As a result, reverse ventricular remodeling in spite of medical treatment accompanied by decreased serum MMP-9 levels indirectly supports our finding that those with high MMP-9 serum levels may have a poor prognosis.

The other essential factors determining prognosis after ACS are reinfarction and revascularization. In this regard, Wang et al.33 studied MMP levels in post-AMI patients and found that increased plasma MMP-9 levels predicted future coronary revascularization after an 18-month follow-up, suggesting a critical role for MMP-9 in atherothrombosis in patients with AMI (hazard ratio: 2.72; 95% CI: 1.24-5.98; p=0.026). In our research, although non-fatal reinfarction and revascularization were more common in those with high MMP-9 levels, this was not statistically significant (p=0.421 and 0.208, respectively). Although we do not have precise data, it can be assumed that there were more fatal MIs in the high MMP-9 group. Besides, statistical significance might have been reached with a larger number of patients and longer follow-up periods. Consequently, we can speculate that circulating MMP-9 levels might have been related to severity of atherosclerosis and plaque rupture, which can result in adverse CV events in post-AMI patients.

There are some limitations to our study. First, this was a single-center analysis, which could lead to selection bias. Second, the number of patients was kept relatively low in order to establish a direct relationship between MMP-9 and adverse clinical outcome. Furthermore, our analysis included only MMP-9 levels. Thus, we cannot comment on other MMPs or tissue inhibitors, which is a large family of extracellular matrix turnover markers, assessment of which might have provided more information about prognosis. However, this study may be a stepping stone in achieving more accurate results by prompting studies on other members of the matrix metalloproteinase group and with more patients. Third, it is difficult to compare our results with other studies on this subject, due to differences in pre-analysis medical therapy, storage techniques and ELISA kits. Also, repeated measurements of MMP-9 were not performed, so changes during follow-up could not be determined. MMP-9 levels were analyzed in samples stored at -70°C, and so we cannot rule out the possibility of protein degradation; it has been demonstrated that storage at this temperature does not prevent the degradation of MMP-9 over time.34 However, it would not affect the differences in results found between the two study groups, since there would be similar degradation in all samples. Fourth, we investigated the predictive value of MMP-9 using a cut-off of 12.92 ng/ml, which was acquired from ROC analysis. Therefore, our results can be assessed only for our cohort, and this cut-off may differ from previous studies. This issue needs validation in other large populations. Nevertheless, this limitation is present in all studies with specific populations on the use of biomarkers in clinical settings. Finally, as the new antiplatelet agents ticagrelor and prasugrel were not covered by social security in our country, we gave clopidogrel to all patients in association with aspirin, so we are unable to determine the exact effect of MMP-9 in AMI patients who use the new antiplatelets. In addition, femoral access was preferred because at the time of recruitment our operators were not experienced in the radial procedure. This approach does not conform to current guidelines. Moreover, the rate of bare-metal stent use in our study was quite high. At the time of patient data collection, drug-eluting stents were not strongly recommended in the guidelines. Also, financial constraints in our country led to a high rate of use of bare-metal stents in that period. However, as shown in Table 3, there was no significant difference between stent types according to MMP-9 groups.

In summary, we determined that high MMP-9 levels were an important predictor of major adverse CV events, including CV mortality and advanced heart failure, in a two-year follow-up period in STEMI patients undergoing primary PCI. Although routine testing for MMP-9 is not recommended in the general population, MMP-9 measurement should be considered for STEMI patients to predict prognosis, and more frequent follow-up may be considered in patients with high MMP-9 levels. More comprehensive investigation is required to clarify this outcome.

No funding was received for this research.

The authors have no conflicts of interest to declare.