Specific biomarkers of myocardial inflammation and remodeling processes as predictors of mortality in high‐risk patients undergoing percutaneous mitral valve repair (MitraClip)

Abstract

Background

Specific matrix metalloproteinases (MMP‐2, MMP‐9) and inflammatory biomarkers (hsCRP, IL‐6) were found to be consistently up‐regulated in severe mitral valve regurgitation (MR) and are associated with mortality in heart failure patients. The aim of the present study was to examine the prognostic value of biomarkers of cardiac inflammation and remodeling processes in predicting mortality in patients with MR undergoing percutaneous mitral valve repair (PMVR).

Hypothesis

We hypothesize that increased cardiac inflammation and extracellular matrix turnover is predictive for mortality in patients with severe mitral regurgitation undergoing MitraClip.

Methods

A total of 210 consecutive patients undergoing PMVR were included. PMVR was performed according to standard clinical practice. Venous blood samples for biomarker analyses were collected prior to and 6 months after PMVR. Physiological parameters, medication use, safety events, and all‐cause mortality were followed over 12 months.

Results

PMVR was performed successfully in all patients. Twelve months after PMVR there was an effective reduction in the severity of MR (P < 0.001), and an improvement in New York Heart Association class (P < 0.01) was documented. Elevated inflammatory biomarkers (AUC_hsCRP: 0.738 [IQR, 0.626–0.849], P = 0.001; AUC_IL‐6: 0.811 [IQR, 0.724–0.899], P = 0.001) and biomarkers reflecting cardiac remodeling processes (AUC_MMP‐2: 0.723 [IQR, 0.641–0.804], P = 0.001; AUC_MMP‐9: 0.618 [IQR, 0.534–0.701], P = 0.01) were predictors of adverse cardiac events and mortality in patients with congestive heart failure undergoing PMVR.

Conclusions

The present study is the first to identify biomarkers reflecting inflammation (hsCRP, IL‐6) and cardiac remodeling processes (MMP‐2, MMP‐9) as predictors of mortality in high‐risk patients undergoing PMVR.

1. INTRODUCTION

Percutaneous mitral valve repair (PMVR) is an interventional treatment option in patients with severe mitral valve regurgitation (MR) who are at high risk for open‐heart surgery.1 In this approach, both mitral‐valve leaflets are attached with ≥1 clips (MitraClip; Abbott Vascular, Santa Clara, CA), resulting in a so‐called “double‐orifice” mitral valve that provides an improvement in MR.1, 2 PMVR is safe and feasible, and an improvement in MR and physical conditions was achieved in several randomized and nonrandomized multicenter trials.1, 2, 3, 4, 5, 6 MR‐associated volume and pressure overload burdens the left ventricle (LV), which initially leads to LV compensatory adaptation mechanisms.7, 8, 9, 10 These early compensatory changes are gradually replaced by a chronic remodeling process and extracellular matrix (ECM) turnover.11, 12, 13

Increased cardiac collagen synthesis and impaired degradation in response to pressure and volume overload in patients with MR lead to an imbalance of collagen metabolism and contribute to collagen accumulation and consequently cardiac structural remodeling processes, myocardial stiffness, and fibrosis.14, 15, 16 Matrix metalloproteinases (MMPs) are important regulators of ECM turnover and myocardial remodeling processes.14, 15, 16 In particular, MMP‐2 and MMP‐9 levels are elevated in patients with LV volume overload, which contributes to myocardial collagen accumulation and subsequent cardiac dysfunction. Further, both MMPs have predictive value regarding cardiovascular events, including mortality.14, 15, 16 Myocardial remodeling processes are also closely associated with inflammation processes, and published data have shown that extensive inflammation magnifies these cardiac structural alterations and thus may accelerate the progression of heart failure (HF).17, 18, 19, 20 High‐sensitivity C‐reactive protein (hsCRP) is a predictor of cardiovascular events and mortality in patients with HF.17, 18, 19 The hepatic synthesis of CRP is partly induced by inflammatory cytokines such as interleukin‐6 (IL‐6), which is secreted by endothelial and smooth‐muscle cells in response to vascular inflammation.18 In particular, elevated IL‐6 serum levels are linked to LV dysfunction and are accompanied by an increased risk of pathological cardiovascular events.18 Accordingly, IL‐6 may play a pivotal role in MR‐related inflammatory processes and HF progression.18, 19, 20

PMVR contributes to a reduction in the severity of MR and significant changes in hemodynamic conditions in high‐risk patients.1, 2, 3, 4, 5, 6 However, information on risk factors indicating the prognosis in patients undergoing PMVR is currently lacking. Therefore, the aim of the present study was to examine the prognostic value of biomarkers indicating inflammatory and remodeling processes. Specifically, hsCRP and IL‐6, as specific biomarkers for cardiac inflammation, and MMP‐2 and MMP‐9, reflecting myocardial remodeling processes, were analyzed in addition to clinical parameters.

2. METHODS

2.1. Patients and treatment

Between January 2014 and June 2016, a total of 210 consecutive patients with severe MR who underwent PMVR using the interventional MitraClip system at the Department of Cardiology of the University of Giessen in Germany and at the Kerckhoff Heart and Thorax Center, Bad Nauheim, Germany, were included in this study. All patients who were scheduled for the interventional mitral valve repair using the MitraClip procedure and who were included in the present study were on optimal medical treatment for HF. Clinical history, physical examination, and laboratory test results were assessed for all patients. Physical exam and echocardiography were performed during the clinical routine without the investigators being blinded. Patients with systemic infections, rheumatoid diseases, and malignancies were excluded from the study. Physiological parameters, including New York Heart Association (NYHA) class, medication use, safety events, and all‐cause mortality, were followed over a 12‐month period after the procedure.

All patients provided written informed consent for their participation in the study, and approval was obtained from the institutional review board of the University of Giessen (99/13). The investigation conforms to the principles outlined in the Declaration of Helsinki.

2.2. Laboratory assessment

For determination of the specified biomarkers, venous blood samples were collected at baseline and 6 months after PMVR. Samples were processed immediately and frozen at −80 °C until assay. HsCRP was measured in serum with a high‐sensitivity electrochemiluminescence immunoassay (Elecsys Analyzer 2010; Roche Diagnostics, Mannheim, Germany). The lower detection limit for the hsCRP assay is 0.15 mg/L and the highest concentration measurable is 20.0 mg/L. The lowest concentration measurable with a coefficient of variation of <10% for this assay is 0.3 mg/L.

IL‐6 was measured in serum using a chemiluminescent microparticle immunoassay (Elecsys Analyzer 2010). The lower detection limit of this assay is 1.5 ng/L, the 95th percentile is 7.0 ng/L, and the lowest concentration measurable with a coefficient of variation of <10% is 5.0 ng/L.

MMP‐2 and MMP‐9 were determined in serum using immunoassays (Quantikine Human MMP‐2 and MMP‐9 immunoassays; R&D Systems, Minneapolis, MN). The minimum detectable concentration of the MMP‐2 assay is 0.156 ng/mL; the coefficient of variation is 5.6%. The minimum detectable concentration of the MMP‐9 assay is 0.156 ng/mL; the coefficient of variation is 5.5%.

The analysis of the specified protein biomarkers examined in this study is experimental and not available in an ongoing clinical laboratory setting. Therefore, specially trained technicians performed the biomarker analysis in a core laboratory.

2.3. Statistical analysis

All data for continuous variables are expressed as mean ± SD or as median and interquartile range (IQR), as appropriate. Categorical variables are reported as number and percentage. After testing for normal distribution, values were compared by unpaired Student t test or by Mann–Whitney test, as appropriate. A Fisher exact test or a χ² test was used for categorical variables with nominal scales. Receiver operating characteristic (ROC) curves were assessed for the determination of the performance of the specified biomarkers. Intergroup comparisons were made using the Mann–Whitney test, ANOVA, or correlation and multiple linear regression models. All statistical tests were performed 2‐tailed, and a significance level of P < 0.05 indicated statistical significance. The statistical software SPSS version 20.0 for Windows (IBM Corp., Armonk, NY) was used for all statistical analyses.

3. RESULTS

A total of 210 consecutive patients (104 men [49.6%], mean age 77.4 ± 8.1 years) were included in the present study. The MitraClip procedure was performed with a mean number of 1.7 ± 0.6 MitraClip devices implanted per patient in a single‐staged procedure. Clinical and procedural characteristics of all patients enrolled in the study are shown in Table 1. Prior to the MitraClip procedure, patients had a marked limitation of physical activity (NYHA class III), moderate reduced left ventricular ejection fraction (LVEF; 44.3% ± 16.8%), and elevated BNP values (370 ng/L [IQR, 207–647 ng/L]) and were at high risk for open‐heart surgery (EURO II score, 7.6 ± 3.7).

Table 1.

Baseline, procedural, and postprocedural characteristics

Parameters	Patients, N = 210
Mean age, y	77.4 ± 8.1
Male sex	104 (49.6)
EURO II score	7.6 ± 3.7
Office BP
SBP, mm Hg	124.1 ± 19.2
DBP, mm Hg	72.3 ± 12.5
NYHA class	3.1 ± 0.7
BNP, ng/L	370.1 (207–647)
eGFR, mL/min/1.73 m2	53.5 ± 25
Cr, mg/dL	1.5 ± 0.8
Echocardiographic parameters
MR grade	3 ± 0.3
Vena contracta	6.7 ± 2.4
MR type
Functional	142 (67.6)
Degenerative	40 (19.0)
Combined	28 (11.9)
LVEF, %	44.3 ± 16.8
E/E′	21.2 ± 6.1
sPAP, mm Hg	42.4 ± 22.6
LA, mm	48 ± 15.5
LVEDD, mm	58.9 ± 16.5
PMVR procedure (MitraClip)
Implanted clips, n ± SD	1.7 ± 0.6
Postprocedural parameters, 48 h
MR grade	1–2
MV mean gradient, mm Hg	4.2 ± 1.8

Abbreviations: BNP, brain natriuretic peptide; BP, blood pressure; Cr, creatinine; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; IQR, interquartile range; LA, left atrium; LVEDD, left ventricular end‐diastolic diameter; LVEF, left ventricular ejection fraction; MR, mitral regurgitation; MV, mitral valve; PMVR, percutaneous mitral valve replacement; SBP, systolic blood pressure; SD, standard deviation; sPAP, systolic pulmonary artery pressure.

Data are presented as n (%), mean ± SD, or median (IQR).

Forty‐eight hours after PMVR there was a significant reduction in the severity of MR grade (from 3 to 1–2; P < 0.001). The levels of the pro‐inflammatory cytokine hsCRP (baseline, 4.57 mg/L [IQR, 1.69–11.03 mg/L] vs 6 months, 3.07 mg/L [IQR, 1.63–8.54 mg/L], P = 0.216) and IL‐6 (baseline, 6.7 pg/mL [IQR, 3.9–11.3 pg/mL] vs 6 months, 36.45 pg/mL [IQR, 2.7–9.6], P = 0.404) were not different at the 6‐month follow‐up compared with baseline values (Table 2). Analysis of the levels of MMP‐2 (baseline, 11.6 ng/mL [IQR, 10.3–13.2 ng/mL] vs 6 months, 12.4 ng/mL [IQR, 10.6–14.8 ng/mL], P = 0.38) and MMP‐9 (baseline, 4.02 ng/mL [IQR, 2.6–6.1 ng/mL] vs 6 months, 5.1 ng/mL [IQR, 3.3–7.6], P = 0.10) also revealed no significant change 6 months after PMVR (Table 2).

Table 2.

Biomarker analysis at baseline and 6‐month follow‐up

Biomarkers (n = 53 patients)	Baseline	6 Months	P Value
hsCRP, mg/L	4.57 (1.69–11.03)	3.07 (1.63–8.54)	0.2
IL‐6, pg/mL	6.7 (3.9–11.3)	6.45 (2.7–9.6)	0.4
MMP‐2, ng/mL	11.6 (10.3–13.2)	12.4 (10.6–14.8)	0.4
MMP‐9, ng/mL	4.0 (2.6–6.1)	5.1 (3.3–7.7)	0.2

Abbreviations: hsCRP, high‐sensitivity C‐reactive protein; IL, interleukin; IQR, interquartile range; MMP, matrix metalloproteinase.

Data are presented as median (IQR).

Twelve months after PMVR (mean follow‐up time: 314 days [IQR, 126–421 days]), the improvement in the NYHA class (NYHA: 3 vs 1–2; P = 0.001) remained. A total of 43 (22.5%) patients died within the 12‐month follow‐up period. These patients had impaired renal function (P = 0.01) and somewhat higher baseline BNP serum levels than did survivors; however, the differences in BNP values between the groups did not reach statistical significance (P = 0.14). In addition, there were no significant differences in age or LV function between survivors and nonsurvivors (Table 3). Nonsurvivors had significantly higher baseline serum levels of the inflammatory cytokines hsCRP and IL‐6 and higher concentrations of the MMPs when compared with survivors. These differences were even more pronounced in patients with congestive HF (Table 3).

Table 3.

Biomarker levels for survivors and nonsurvivors (N = 210)

	Nonsurvivors, n = 43	Survivors, n = 167	P Value
Mean age, y	77.6 ± 6.6	77.37 ± 8.4	0.69
LVEF, %	42.2 ± 18.1	43.9 ± 18.6	0.29
BNP, ng/L	487 (213–1158)	349 (205–499)	0.14
Cr, mg/dL	1.8 ± 0.7	1.4 ± 0.9	0.01
eGFR, mL/min/1.73 m2	46.4 ± 25.1	55.6 ± 25.4	0.01
Biomarkers
hsCRP, mg/L	7.97 (4.65–31.2)	4.05 (1.4–11.3)	0.002
IL‐6, pg/mL	14.0 (8.4–28.3)	6.7 (3.7–12.4)	0.001
MMP‐2, ng/mL	15.4 (14.1–17.9)	10.6 (8.3–12.6)	0.001
MMP‐9, ng/mL	6.1 (3.8–7.8)	4.7 (2.9–7.5)	0.02
HF subgroup, n = 75	Nonsurvivors, n = 21	Survivors, n = 54	P Value
hsCRP, mg/L	13.6 (5.2–31.4)	3.1 (1.2–11.1)	0.001
IL‐6, pg/mL	19.4 (9.6–39.5)	5.8 (3.4–11.3)	0.001
MMP‐2, ng/mL	13.41 (10.19–17.48)	10.49 (8.11–12.20)	0.004
MMP‐9, ng/mL	8.11 (5.30–8.72)	4.61 (2.67–7.33)	0.02

Abbreviations: BNP, brain natriuretic peptide; Cr, creatinine; eGFR, estimated glomerular filtration rate; HF, heart failure; hsCRP, high‐sensitivity C‐reactive protein; IL, interleukin; IQR, interquartile range; LVEF, left ventricular ejection fraction; MMP, matrix metalloproteinase; SD, standard deviation.

Data are presented as mean ± SD or median (IQR).

ROC analysis was performed to assess the predictive value of baseline serum levels of BNP, hsCRP, IL‐6, and the MMPs as predictors of mortality after PMVR. The AUC values for BNP were 0.612 [IQR, 0.489–0.735], P = 0.079 (Figure 1); for hsCRP, 0.654 [IQR, 0.562–0.746], P = 0.002; and for IL‐6, 0.724 [IQR, 0.641–0.807], P = 0.01 (Figure 2A). For MMP‐2, these values were 0.723 [IQR, 0.641–0.804], P = 0.001; and for MMP‐9, 0.618 [IQR, 0.534–0.701], P = 0.01 (Figure 3A). The predictive value was more enhanced in patients with congestive HF for inflammatory biomarkers hsCRP (0.738 [IQR, 0.626–0.849], P = 0.001); IL‐6 (0.811 [IQR, 0.724–0.899], P = 0.001; Figure 2B), MMP‐2 (0.723 [IQR, 0.641–0.804], P = 0.001), and MMP‐9 (0.618 [IQR, 0.534–0.701], P = 0.01; Figure 3B). In patients with congestive HF, the predictive value for BNP was not statistically significant (AUC: 0.59 [IQR, 0.324–0.84], P = 0.46).

Figure 1.

Predictive value of BNP as a predictor of mortality in patients with severe mitral regurgitation undergoing the MitraClip procedure. Abbreviations: AUC, area under the curve; BNP, brain natriuretic peptide; ROC, receiver operating characteristic

Figure 2.

Predictive value of inflammatory biomarkers as predictors of mortality in (A) all patients and (B) HF patients (LVEF ≤30%). Abbreviations: AUC, area under the curve; HF, heart failure; HFrEF, heart failure with reduced ejection fraction; hsCRP, high‐sensitivity C‐reactive protein; IL‐6, interleukin 6; LVEF, left ventricular ejection fraction; ROC, receiver operating characteristic

Figure 3.

Predictive value of MMPs as predictors of mortality in (A) all patients and (B) HF patients (LVEF ≤30%). Abbreviations: AUC, area under the curve; HF, heart failure; LVEF, left ventricular ejection fraction; MMP, matrix metalloproteinase; ROC, receiver operating characteristic

4. DISCUSSION

The detrimental effect of severe MR on morbidity and mortality is well established.7, 8, 9, 10 In patients with severe MR, myocardial inflammation, ECM turnover, and collagen deposition play a key role in the development and progression of myocardial remodeling processes and fibrosis, which contribute to the progression of HF and mortality.1, 2, 7, 8, 9 The present study is the first to identify biomarkers reflecting inflammation and cardiac remodeling processes as predictors of mortality in high‐risk patients undergoing PMVR.

Severe MR, which aggravates HF symptoms and increases the risk of mortality, is a common problem in patients with reduced LVEF.7, 8, 9, 10 Although the MitraClip procedure is safe and feasible in patients who are at high risk for open‐heart surgery, valid clinical predictors of overall survival and rehospitalization have not yet been identified.21, 22, 23, 24 Risk stratification is partly based on clinical evaluation and the assessment of specific risk scores. Importantly, risk scores that are currently used in routine clinical practice, such as the European System for Cardiac Operative Risk Evaluation (EuroSCORE),25 are mainly based on clinical and biochemical variables. However, established clinical risk models for short‐ and mid‐term procedural outcome are generally based on cohorts of patients undergoing cardiac surgery and poorly predict outcome in patients undergoing interventional procedures such as aortic valve replacement or mitral valve repair.2, 7, 22, 24, 25, 26, 27, 28, 29, 30, 31 In the present study, patients who died within the 12‐month follow‐up period had significantly poorer renal function and tended to have higher BNP levels at baseline, confirming the results from previous studies.9, 21, 22 However, there were no differences in age, LV function, or LV end‐diastolic diameter between survivors and nonsurvivors, which essentially reflects the dilemma in evaluation of clinical outcome after PMVR in high‐risk patients. As previously described, myocardial inflammation and advanced remodeling processes are associated with poor prognosis in HF13, 16, 17, 18, 19, 32; accordingly, indicators of these processes were significantly elevated in patients who died within 12 months after the MitraClip procedure in this study.

The predictive value of the biomarkers of inflammation and cardiac remodeling investigated here was even more pronounced than that of BNP, the most investigated biomarker in HF. Moreover, the predictive value of hsCRP, IL‐6, and the specified MMPs was even stronger in high‐risk patients with congestive HF and severely reduced LV function. Patients who were scheduled for the MitraClip procedure and who were included in the present study were on optimal medical treatment, and thus the patients' volume status was recompensated before the procedure; this might have impacted the predictive value of baseline BNP values. Nevertheless, the fact that the predictive value of the inflammatory biomarkers and MMP was also more pronounced than BNP under these conditions confirms the diagnostic sensitivity of the specified biomarkers in this study; this was further underlined in patients with severe HF.

Previous studies presented data indicating post‐PMVR improvement in clinical symptoms and LV function as well as cardiac reverse remodeling processes using imaging methods or volumetric analysis by echocardiography and cardiac magnetic resonance imaging.2, 26, 27, 33, 34, 35 The results of the biomarker assessment in this study, however, did not provide any information confirming an effect of interventional edge‐to‐edge mitral‐valve repair on structural alterations that were described previously by imaging methods. The lack of an effect of MitraClip‐based PMVR on inflammatory and remodeling biomarker levels might be explained by the characteristics of the present cohort of patients, who presented with primarily functional MR that was mainly associated with reduced LVEF and an enlarged LV end‐diastolic diameter. This assumption is supported by several published studies that presented cardiac recovery processes after MitraClip‐based PMVR, although they excluded patients with reduced LVEF and were, thus, not entirely comparable with the present data. Accordingly, further analyses are required to examine the association between cardiac magnetic resonance imaging and biomarker alterations after PMVR to confirm the beneficial effect of the MitraClip procedure on cardiac remodeling processes, also in patients with reduced LVEF, that are of prognostic relevance for patients with MR and HF.

4.1. Study limitations

The present study was not randomized and did not have a control group, which can be considered a major limitation. The present study included patients with degenerative MR as well as functional MR and did not discriminate between different underlying pathophysiology; this should also be considered a limitation of the study. The results of the study were based on a pretreatment and post‐treatment biomarker analysis and are therefore exploratory in nature.

5. CONCLUSION

This is the first study to identify biomarkers reflecting cardiac inflammation and ECM remodeling processes as valid predictors of mortality. Therefore, assessment of such biomarkers may provide helpful information for risk stratification and prognosis in high‐risk patients undergoing PMVR. Further, larger‐scale studies will be necessary to confirm the results of the present study.

Dörr O, Walther C, Liebetrau C, et al. Specific biomarkers of myocardial inflammation and remodeling processes as predictors of mortality in high‐risk patients undergoing percutaneous mitral valve repair (MitraClip). Clin Cardiol. 2018;41:481–487. 10.1002/clc.22900