Postprocedural Troponin Elevation and Mortality After Transcatheter Aortic Valve Implantation.

Background This study sought to investigate the role of postprocedural troponin elevations in mortality prediction after transcatheter aortic valve implantation and to define the threshold at which clinically relevant postprocedure myocardial injury determines mortality. Methods and Results A total of 1333 consecutive patients with transcatheter aortic valve implantation with available postprocedural high-sensitivity cardiac troponin T measurements were included in the analysis. The threshold at which postprocedure myocardial injury determines long-term mortality was identified using restricted cubic spline analysis. A >18.3-fold increase of troponin above the upper reference limit was identified as threshold for relevant postprocedure myocardial injury. Associations remained significant in a landmark analysis between 30 days and 2 years (hazard ratio [HR], 1.61, [95% CI, 1.13-2.28]; P=0.01), after adjusting for known confounders (adjusted HR, 1.90 [95% CI, 1.40-2.57]; P<0001), and in subgroups of patients with coronary artery disease (adjusted HR, 2.17 [95% CI, 1.44-3.29]; P<0.001), renal dysfunction (adjusted HR, 1.88 [95% CI, 1.35-2.62]; P<0.001), and intermediate/high surgical risk (adjusted HR, 2.70 [95% CI, 1.40-5.22]; P=0.003). Conclusions This study determined a troponin threshold for the identification of patients at increased mortality risk after transcatheter aortic valve implantation. The proposed definition of postprocedure myocardial injury advances risk stratification in patients with transcatheter aortic valve implantation and may assist in postprocedural patient management.

European System for Cardiac Operative Risk Evaluation II score

high‐sensitivity cardiac troponin T

Society of Thoracic Surgeons Predicted Risk of Mortality

transcatheter aortic valve implantation

upper reference limit

Clinical Perspective

What Is New?

This study determined a troponin threshold for the identification of patients at increased mortality risk after transcatheter aortic valve implantation.

What Are the Clinical Implications?

The proposed definition of clinically relevant postprocedure myocardial injury advances risk stratification in patients undergoing transcatheter aortic valve implantation and may assist in postprocedural patient management.

Aortic stenosis is the most common acquired valvular heart disease in the Western world, and the prevalence is projected to rise further given the aging of the population. , , , Once symptoms of aortic stenosis occur after a long‐lasting asymptomatic phase of disease progression, aortic stenosis is associated with considerable morbidity, mortality, and healthcare costs. , , Aortic valve replacement either by surgery or as a transcatheter procedure represents the standard therapy after a comprehensive evaluation of the individual surgical risk and comorbid burden. The success of transcatheter aortic valve implantation (TAVI) over the past years has prompted the expansion of TAVI from initially inoperable and high‐risk patients to a younger and low‐risk patient population. , , ,

Postprocedural myocardial injury or infarction has been observed after TAVI and is associated with worse outcomes. While certain degrees of cardiac biomarker elevations after TAVI are considered to occur in almost all patients, , , postprocedural myocardial injury has been observed in up to two thirds of patients, , depending on patients’ baseline risk and the endpoint definition applied. Postprocedural myocardial injury has been related with excess mortality and adverse left ventricular remodeling in most studies, , , , , although some failed to confirm this association. , , In particular, the clinical relevance of only slight cardiac biomarker elevations following TAVI remains a matter of ongoing debate, , and threshold values of high‐sensitivity cardiac troponin indicating clinically relevant postprocedural myocardial injury have not yet been investigated.

This study therefore sought to determine the association between postprocedural myocardial injury and long‐term clinical outcomes after TAVI using data from the prospective Zurich SwissTAVI Registry. In particular, we determine a threshold for the association of postprocedural troponin elevation and mortality after TAVI.

Methods

Data are available from corresponding author upon reasonable request.

Study Population

The study is based on data from the prospective Zurich SwissTAVI Registry. All patients who underwent TAVI at the University Hospital Zurich, Switzerland, between April 2012 and December 2019 were entered into a dedicated database (Zurich SwissTAVI Registry). As previously described, , , , the SwissTAVI Registry is a national, multicenter cohort study, initiated by the Swiss Working Group of Interventional Cardiology and the Swiss Society of Cardiac and Thoracic Vascular Surgery in 2011 and registered at ClinicalTrials.gov (NCT01368250). An independent Clinical Trials Unit is responsible for central data monitoring and verification of data completeness and accuracy. All patients are evaluated for TAVI by a multidisciplinary board of interventional cardiologists, cardiac surgeons, cardiac anesthesiologists, and imaging specialists (ie, the Heart Team). In all patients, demographic, clinical, and procedural characteristics are systematically collected using a web‐based database with standardized case report forms. Electrocardiogram, transthoracic echocardiography, coronary angiography, and cardiac computed tomography were routinely performed before the procedure. Transcatheter aortic valve implantation was performed according to current guidelines and recommendations and using standard techniques in the cardiac catheterization laboratory or the hybrid operating room. Routine laboratory analyses were performed according to the laboratory’s standard operating procedures, and values at baseline and follow‐up were collected in the database. Follow‐up was performed in‐hospital, at 30 days, and yearly thereafter by means of standardized clinical visits or phone calls.

Of 1400 consecutive patients, 1333 (95.2%) had postprocedural measurements of high‐sensitivity cardiac troponin T (hs‐cTnT) levels available and were included in the analysis. The study was approved by the Ethics Committee (Cantonal Ethics Committee Zurich), conducted in full conformance with the Declaration of Helsinki, and all patients provided written informed consent for prospective follow‐up.

Measurement of Cardiac Troponin

Peak hs‐cTnT levels during the index hospitalization for the TAVI procedure were used to define postprocedural myocardial injury. Cardiac troponin was measured using the high‐sensitivity Elecsys cTnT assay (Roche Diagnostics, Mannheim, Germany). Based on the 99th percentile in a healthy population and the requirement of a ≤10% coefficient of variation, the upper reference limit (URL) for hs‐cTnT levels was 14 ng/L.

Definitions

Based on prior studies, , , the primary endpoint was mortality at 2 years. Secondary endpoints included mortality at 30 days, as well as cardiovascular death, cerebrovascular events (stroke or transient ischemic attack), and myocardial infarction at both 30 days and 2 years. Renal dysfunction was defined as estimated glomerular filtration rate <60 mL/min per 1.73 m2. Coronary artery disease was defined as the presence of 1 or more coronary lesions with ≥50% diameter stenosis by visual estimation on the coronary angiogram in vessels ≥1.5 mm in diameter. Patients were dichotomized according to the Society of Thoracic Surgeons Predicted Risk of Mortality (STS‐PROM) Score into a low‐risk (STS‐PROM Score <4%) and an intermediate/high‐risk group (STS‐PROM Score ≥4%). In the SwissTAVI Registry, endpoint definitions are based on the updated standardized endpoint definitions for TAVI of the Valve Academic Research Consortium‐2 and clinical events were reviewed and adjudicated by a dedicated clinical event committee.

Statistical Analysis

Continuous variables are presented as mean and SD, and categorical variables as numbers and percentages, respectively. Baseline and procedural characteristics were compared using χ2 tests for proportions and unpaired t tests for means. The association of postprocedural peak hs‐cTnT levels with mortality was first assessed using univariable Cox regression models. Second, nonparametric restricted cubic splines were used to model the association of the fold increase of postprocedural peak hs‐cTnT levels above the URL with mortality at 2 years. We compared models with different numbers of knots (3, 4, and 5 knots). Because the model with 4 knots showed the best performance based on the Akaike information criterion, a model with 4 flexible knots was compared with a model where the knots were placed at quartiles of the variable. The final model with 4 flexible knots was again determined using the Akaike information criterion. The lower 95% CI was used to determine the ideal cutoff value. In addition, sensitivity analyses were performed to assess the robustness of the retrieved cutoff value. The cohort was then divided based on the above determined cutoff value of postprocedural peak hs‐cTnT levels above the URL, and baseline characteristics compared among groups (above versus below the cutoff value). Kaplan–Meier analysis and univariable and multivariable Cox regression analyses were used to assess the discriminative power of the identified cutoff value, with time zero defined as the date of the TAVI procedure. Subjects who died during the procedure were included in the analysis and survival time was set to 1 day. All variables with P<0.1 in univariable analysis were included in the multivariable model (age, sex, chronic obstructive pulmonary disease, atrial fibrillation, renal failure, peripheral artery disease, and STS‐PROM Score as a continuous variable). The independent association of the cutoff value was also tested for secondary endpoints. Furthermore, interactions of postprocedural peak hs‐cTnT levels above the URL with sex as well as all variables included in the multivariable model were tested by including interaction terms in the corresponding Cox regression models. We used the cox.zph command which is part of the survival package (version 3.2‐7) in the statistical software R to test for the proportional hazard assumption. We also used Poisson regression with robust SEs to calculate rate ratios for in‐hospital events. In an explorative analysis, we determined factors associated with postprocedural myocardial injury using univariable and multivariable logistic regression models. Findings were considered statistically significant at the 0.05 level. All analyses were performed with R software for statistical computing (Version 4.0.2).

Results

Threshold Definition for Postprocedural Myocardial Injury After TAVI

Out of 1333 patients, 187 (14.0%) patients died during the 2‐year follow‐up. A significant association between postprocedural peak hs‐cTnT levels and mortality at 2 years was observed in univariable Cox regression analysis (hazard ratio [HR], 1.01 [95% CI, 1.01–1.01]; P<0.001). Restricted cubic splines with 4 knots placed at flexible locations were then used to model the relation of postprocedural peak hs‐cTnT levels above the URL with mortality at 2 years (Figure 1). At an 18.3‐fold increase of postprocedural peak hs‐cTnT levels above the URL, the lower end of the CI crossed a relative risk for all‐cause mortality of 1. A ≥18.3‐fold increase of postprocedural peak hs‐cTnT levels above the URL was therefore defined as postprocedural myocardial injury.

Figure 1

Restricted cubic spline analysis to determine the threshold at which postprocedural myocardial injury determines mortality.

Green and red areas represent the 95% CI. Univariable Cox proportional hazard regression with restricted cubic splines was used to flexibly model the association of peak hs‐cTnT levels above URL with mortality at 2 years after transcatheter aortic valve implantation. The minimal threshold at which hs‐cTnT is significantly associated with mortality at 2 years was identified at 18.3‐fold increase above URL. HR indicates hazard ratio; hs‐cTnT, high‐sensitivity cardiac troponin T; and URL, upper reference limit.

Kaplan–Meier analyses showed that mortality at 2 years was significantly higher in patients with postprocedural peak hs‐cTnT levels ≥18.3‐fold above the URL than in those with values below (P<0.001, Figure 2). No significant interaction between postprocedural peak hs‐cTnT levels and sex was observed.

Figure 2

Kaplan–Meier estimates of survival according to the presence/absence of postprocedural myocardial injury.

The HR was adjusted for age, sex, chronic obstructive pulmonary disease, atrial fibrillation, peripheral artery disease, and Society of Thoracic Surgeons Predicted Risk of Mortality (STS‐PROM) Score. aHR indicates adjusted hazard ratio; HR, hazard ratio; and PPMI, postprocedural myocardial injury.

Baseline and Procedural Characteristics

Postprocedural myocardial injury as defined by a ≥18.3‐fold increase of postprocedural peak hs‐cTnT levels occurred in 322 (24.2%) patients (Figure 3). Baseline characteristics according to the presence/absence of postprocedural myocardial injury are given in Table 1. Mean postprocedural peak hs‐cTnT level was 253.8 (±411.3) ng/L and mean postprocedural hs‐cTnT increase above the URL was 18.1 (±29.4)‐fold. Patients with postprocedural myocardial injury were older, had a higher STS‐PROM Score, and a higher European System for Cardiac Operative Risk Evaluation II Score (EuroSCORE II). They more often had prior percutaneous coronary intervention, known coronary artery disease and renal dysfunction, and more often presented with severe dyspnea, high‐grade mitral regurgitation, and pulmonary hypertension. Procedural characteristics according to the presence/absence of postprocedural myocardial injury are given in Table 2. Patients with postprocedural myocardial injury had a longer procedure time and more frequently underwent concomitant coronary revascularization during the TAVI procedure. Postprocedural myocardial injury was observed in 26 out of 31 (83.9%) patients with transapical TAVI.

Figure 3

Increases of high‐sensitivity cardiac troponin T levels following transcatheter aortic valve implantation.

Proportion of patients across categories of postprocedural increases of hs‐cTnT above the URL. hs‐cTnT indicates high‐sensitivity cardiac troponin T; and URL, upper reference limit.

Table 1

Baseline Characteristics According to the Presence of Postprocedural Myocardial Injury Defined as an 18.3‐Fold Increase of Postprocedural Troponin Levels Above the Upper Reference Limit

	No PPMI (N=1011)	PPMI (N=322)	Total (N=1333)	P value
Age				0.006
No.	1011	322	1333
Mean (SD)	80.2 (7.6)	81.5 (7.6)	80.5 (7.7)
Sex				0.51
Male	549 (54.3%)	168 (52.2%)	717 (53.8%)
Female	462 (45.7%)	154 (47.8%)	616 (46.2%)
BMI				0.17
No.	1011	321	1332
Mean (SD)	27.1 (5.1)	26.7 (4.8)	27.0 (5.1)
Diabetes				0.32
Yes	270 (26.7%)	77 (23.9%)	347 (26.0%)
No	741 (73.3%)	245 (76.1%)	986 (74.0%)
Hypertension				0.96
Yes	780 (77.2%)	248 (77.0%)	1028 (77.1%)
No	231 (22.8%)	74 (23.0%)	305 (22.9%)
Dyslipidemia				0.97
Yes	498 (49.3%)	159 (49.4%)	657 (49.3%)
No	513 (50.7%)	163 (50.6%)	676 (50.7%)
Coronary artery disease				0.005
Yes	474 (46.9%)	180 (55.9%)	654 (49.1%)
No	537 (53.1%)	142 (44.1%)	679 (50.9%)
Myocardial infarction				0.77
Yes	95 (9.4%)	32 (9.9%)	127 (9.5%)
No	916 (90.6%)	290 (90.1%)	1206 (90.5%)
PCI				0.03
Yes	225 (22.3%)	91 (28.3%)	316 (23.7%)
No	786 (77.7%)	231 (71.7%)	1017 (76.3%)
Pacemaker implantation				0.92
Yes	77 (7.6%)	24 (7.5%)	101 (7.6%)
No	934 (92.4%)	298 (92.5%)	1232 (92.4%)
Atrial fibrillation				0.27
Yes	196 (19.5%)	58 (18.2%)	254 (19.0%)
No	813 (80.5%)	263 (81.8)	1076 (81.0)
Pulmonary hypertension				0.01
Yes	34 (3.4%)	21 (6.8%)	55 (4.2%)
No	956 (96.6%)	286 (93.2%)	1242 (95.8%)
COPD				0.22
Yes	121 (12.0%)	47 (14.6%)	168 (12.6%)
No	890 (88.0%)	275 (85.4%)	1165 (87.4%)
Cerebrovascular event				0.48
Yes	114 (11.3%)	41 (12.7%)	155 (11.6%)
No	897 (88.7%)	281 (87.3%)	1178 (88.4%)
Renal failure				<0.001
Yes	643 (63.7%)	252 (78.5%)	895 (67.2%)
No	367 (36.3%)	69 (21.5%)	436 (32.8%)
Dyspnea				0.03
NYHA I	175 (17.3%)	45 (14.0%)	220 (16.5%)
NYHA II	346 (34.2%)	92 (28.6%)	438 (32.9%)
NYHA III	421 (41.6%)	153 (47.5%)	574 (43.1%)
NYHA IV	69 (6.8%)	32 (9.9%)	101 (7.6%)
EuroSCORE II				<0.001
No.	955	294	1249
Mean (SD)	4.3 (4.0)	5.8 (6.3)	4.7 (4.7)
STS‐PROM Score				<0.001
No.	1009	321	1330
Mean (SD)	4.3 (3.2)	5.7 (4.7)	4.7 (3.7)
Troponin at baseline (ng/L)				<0.001
No.	773	257	1030
Median (IQR)	23.0 (15.0–40.0)	42.0 (21.0–88.0)	26.0 (16.0–49.0)
Postprocedural peak troponin (ng/L)				<0.001
No.	1011	322	1333
Median (IQR)	112.0 (74.0–159.0)	433.5 (311.8–746.5)	139.0 (86.0–252.0)
Postprocedural peak CK (U/L)				<0.001
No.	992	319	1311
Mean (SD)	151.3 (300.0)	776.1 (3569.4)	303.3 (1798.0)
LVEF (%)				0.921
No.	1002	317	1319
Mean (SD)	53.9 (13.7)	53.8 (14.0)	53.9 (13.8)
Transaortic mean pressure gradient (mm Hg)				0.29
No.	986	306	1292
Mean (SD)	40.6 (16.1)	41.7 (17.7)	40.8 (16.5)
Aortic valve area (cm2)				0.89
No.	911	284	1195
Mean (SD)	0.7 (0.2)	0.7 (0.3)	0.7 (0.2)
Aortic regurgitation grade				0.54
None/mild	859 (87.0%)	271 (86.3%)	1130 (86.9%)
Moderate	80 (8.1%)	23 (7.3%)	103 (7.9%)
Severe	48 (4.9%)	20 (6.4%)	68 (5.2%)
Mitral regurgitation grade				0.02
None/mild	760 (77.2%)	214 (69.3%)	974 (75.3%)
Moderate	178 (18.1%)	75 (24.3%)	253 (19.6%)
Severe	46 (4.7%)	20 (6.5%)	66 (5.1%)
Tricuspid regurgitation grade				0.58
None/mild	866 (88.7%)	271 (87.1%)	1137 (88.3%)
Moderate	79 (8.1%)	31 (10.0%)	110 (8.5%)
Severe	31 (3.2%)	9 (2.9%)	40 (3.1%)
Aspirin				0.27
Yes	561 (55.5%)	190 (59.0%)	751 (56.3%)
No	450 (44.5%)	132 (41.0%)	582 (43.7%)
Statin				0.49
Yes	581 (57.5%)	192 (59.6%)	773 (58.0%)
No	430 (42.5%)	130 (40.4%)	560 (42.0%)
β‐Blocker				0.90
Yes	492 (48.7%)	158 (49.1%)	650 (48.8%)
No	519 (51.3%)	164 (50.9%)	683 (51.2%)
ACE inhibitor				0.65
Yes	312 (30.9%)	95 (29.5%)	407 (30.5%)
No	699 (69.1%)	227 (70.5%)	926 (69.5%)
Diuretics				0.004
Yes	557 (55.1%)	207 (64.3%)	764 (57.3%)
No	454 (44.9%)	115 (35.7%)	569 (42.7%)

Values are given as mean and SD or numbers and percentages. Renal failure was defined as estimated glomerular filtration rate <60 mL/min per 1.73 m2. ACE indicates angiotensin‐converting enzyme; BMI, body mass index; CK, creatine kinase; COPD, chronic obstructive pulmonary disease; IQR, interquartile range; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association; PCI, percutaneous coronary intervention; PPMI, postprocedural myocardial injury; and STS‐PROM, Society of Thoracic Surgeons Predicted Risk of Mortality.

Table 2

Procedural Characteristics According to the Presence of Postprocedural Myocardial Injury Defined as an 18.3‐Fold Increase of Postprocedural Troponin Levels Above the Upper Reference Limit

	No PPMI (N=1011)	PPMI (N=322)	Total (N=1333)	P value
Access site				<0.001
No.	1010	322	1332
Femoral	983 (97.3%)	288 (89.4%)	1271 (95.4%)
Transapical	5 (0.5%)	26 (8.1%)	31 (2.3%)
Subclavian	20 (2.0%)	8 (2.5%)	28 (2.1%)
Aortic	2 (0.2%)	0 (0.0%)	2 (0.2%)
Prosthesis type				0.005
Edwards Sapien 3	289 (28.6%)	86 (26.9%)	375 (28.2%)
SJM Portico	232 (23.0%)	58 (18.1%)	290 (21.8%)
Medtronic Evolut R	161 (15.9%)	42 (13.1%)	203 (15.3%)
Medtronic CoreValve	106 (10.5%)	47 (14.7%)	153 (11.5%)
Edwards Sapien XT	66 (6.5%)	38 (11.9%)	104 (7.8%)
Concomitant PCI				0.006
Yes	47 (4.7%)	28 (8.7%)	75 (5.6%)
No	963 (95.3%)	294 (91.3%)	1257 (94.4%)
Procedure time (min)				0.003
No.	366	166	532
Mean (SD)	55.7 (42.6)	67.4 (41.7)	59.4 (42.6)
Postprocedure mean trans‐prosthetic pressure gradient (mm Hg)				0.10
No.	975	286	1261
Mean (SD)	8.1 (4.6)	8.6 (5.3)	8.2 (4.8)
Postprocedure mean trans‐prosthetic pressure gradient (mm Hg) binary				0.75
<20 mm Hg	951 (97.5%)	278 (97.2%)	1229 (97.5%)
≥20 mm Hg	24 (2.5%)	8 (2.8%)	32 (2.5%)
Device success				0.28
Device failure (VARC‐2)	140 (15.3%)	49 (18.0%)	189 (15.9%)
Device success (VARC‐2)	775 (84.7%)	223 (82.0%)	998 (84.1%)

Values are given as mean and SD or numbers and percentages. PCI indicates percutaneous coronary intervention; PPMI, postprocedural myocardial injury; and VARC‐2, Valve Academic Research Consortium‐2.

Postprocedural Myocardial Injury After TAVI and Outcomes

In‐hospital outcomes of patients with and without postprocedural myocardial injury are given in Table S1. Postprocedural myocardial injury was significantly associated with an excess risk of mortality at 2 years (HR, 2.36 [95% CI, 1.76–3.15]; P<0.001, Table 3 and Table S2). The association of postprocedural myocardial injury with 2‐year mortality remained significant irrespective of the presence (adjusted HR, 2.17 [95% CI, 1.44–3.29]; P<0.001) or absence (adjusted HR, 2.67 [95% CI, 1.20–5.93]; P=0.03) of coronary artery disease (Figure 4A), as well as the presence (adjusted HR, 1.88 [95% CI, 1.35–2.62]; P<0.001) or absence (adjusted HR, 2.21 [95% CI, 1.09–4.50]; P=0.03) of renal dysfunction (Figure 4B). Furthermore, the association of postprocedural myocardial injury with 2‐year mortality was significant both in patients with low (adjusted HR, 1.76 [95% CI, 1.13–2.75]; P=0.015) and intermediate/high surgical risk (adjusted HR, 2.70 [95% CI, 1.40–5.22]; P<0.001, Figure 4C). This association remained significant in a univariable landmark analysis between 30 days and 2 years (HR, 1.61 [95% CI, 1.13–2.28]; P=0.01, Figure 5), while in multivariable landmark analysis, the association was not significant (adjusted HR, 1.22 [95% CI, 0.84–1.76]; P=0.32).

Table 3

Outcomes According to Presence of Postprocedural Myocardial Injury as Defined by an 18.3‐Fold Increase Above the Upper Reference Limit

Outcome	No PPMI	PPMI	HR (95% CI)	P value	Adjusted HR (95% CI)	P value
30 d
All‐cause mortality	16 (1.6%)	33 (10.2%)	6.74 (3.71–12.20)	<0.001	6.09 (3.30–11.20)	<0.001
Cardiovascular death	9 (0.9%)	14 (4.3%)	5.16 (2.23–11.90)	<0.001	4.48 (1.90–10.60)	<0.001
Cerebrovascular events	20 (2.0%)	12 (3.7%)	2.03 (0.99–4.15)	0.06	1.76 (0.85–3.68)	0.14
Myocardial infarction	2 (0.2%)	5 (1.6%)	8.39 (1.63–43.30)	0.007	9.92 (1.85–53.10)	0.007
Acute kidney failure	25 (2.5%)	35 (10.9%)	4.80 (2.87–8.03)	<0.001	3.95 (2.33–6.71)	<0.001
2 y
All‐cause mortality*	107 (10.6%)	80 (24.8%)	2.36 (1.76–3.15)	<0.001	1.90 (1.40–2.57)	<0.001
Cardiovascular death	34 (3.4%)	34 (10.6%)	3.31 (2.06–5.33)	<0.001	2.58 (1.58–4.23)	0.008
Cerebrovascular events	28 (2.8%)	13 (4.0%)	1.58 (0.82–3.05)	0.32	1.34 (0.68–2.62)	0.40
Myocardial infarction	8 (0.8%)	8 (2.5%)	3.35 (1.26–8.93)	0.02	3.41 (1.24–9.38)	<0.001
Acute kidney failure	42 (4.2%)	43 (13.4%)	3.50 (2.29–5.36)	<0.001	2.91 (1.88–4.52)	<0.001

Reported are numbers of first events (%), HRs with corresponding 95% CI from Cox regression models. Multivariable Cox regression models were adjusted for age, sex, chronic obstructive pulmonary disease, peripheral artery disease, renal dysfunction, and Society of Thoracic Surgeons Predicted Risk of Mortality (STS‐PROM) Score. HR indicates hazard ratio; and PPMI, periprocedural myocardial injury.

The proportional hazards assumption for this variable was violated at 2 years.

Figure 4

Kaplan–Meier estimates of survival according to the presence/absence of postprocedural myocardial injury stratified for coronary artery disease, renal dysfunction, and surgical risk.

A, Kaplan–Meier estimates of survival in patients with and without PPMI stratified according to the presence/absence of coronary artery disease. B, Kaplan–Meier estimates of survival in patients with and without PPMI stratified according to the presence/absence of renal dysfunction. C, Kaplan–Meier estimates of survival in patients with and without PPMI stratified according to low or intermediate/high surgical risk. The hazard ratio in (A through C) was adjusted for age, sex, chronic obstructive pulmonary disease, atrial fibrillation, peripheral artery disease, and Society of Thoracic Surgeons Predicted Risk of Mortality score. aHR indicates adjusted hazard ratio; CAD, coronary artery disease; PPMI, postprocedural myocardial injury; and RF, renal failure.

Figure 5

Landmark analysis of survival between 30 days and 2 years according to the presence/absence of postprocedural myocardial injury.

CI = confidence interval, HR, hazard ratio; PPMI, post‐procedural myocardial infarction.

Postprocedural myocardial injury was also significantly associated with secondary endpoints including cardiovascular death (adjusted HR, 2.58 [95% CI, 1.58–4.23]; P=0.008), myocardial infarction (adjusted HR, 3.41 [95% CI, 1.24–9.38]; P<0.001), and acute kidney failure (adjusted HR, 2.91 [95% CI, 1.88–4.52]; P<0.001) at both 30 days and 2 years (Table 3).

Sensitivity Analyses

Sensitivity analyses testing the robustness of the results are given in Table S3. The estimated threshold remained unchanged when restricting the univariable analysis follow‐up time to 30 days and 5 years. Similarly, the threshold remained unchanged when adjusting the restricted cubic spline analysis for age, sex, chronic obstructive pulmonary disease, renal dysfunction, peripheral artery disease, and the STS‐PROM Score. Changes in inclusion criteria only modestly affected the estimated threshold. When excluding patients with baseline hs‐cTnT levels >70‐fold increase above the URL, the lower bound of the 95% CI crossed the HR of 1 at a 22.7‐fold increase of postprocedural peak hs‐cTnT levels above the URL.

Discussion

This observational study for the first time determined a threshold for the association of postprocedural troponin elevation and long‐term mortality after TAVI. Using spline curve analysis, an 18‐fold hs‐cTnT increase after TAVI was identified as minimum value being significantly associated with mortality at 2 years. The proposed definition of clinically relevant postprocedural myocardial injury advances patient risk stratification after TAVI and assists in postprocedural clinical care.

Postprocedural Myocardial Injury in Patients With TAVI

Although postprocedural myocardial injury is among the most frequent complications occurring after TAVI, the clinical relevance of biomarker elevations remains controversial. , , , While biomarker increases after coronary revascularization have been extensively studied, , , there is a paucity of data on the impact of biomarker increases on prognosis in patients undergoing TAVI, and most studies are limited by their rather small sample size and short‐term follow‐up. Pathophysiological mechanisms underlying the occurrence of postprocedural myocardial injury in patients with TAVI are multifaceted, but only poorly understood. Coronary artery occlusion by the transcatheter heart valve in most severe cases, mechanical compression of the left ventricular outflow tract, distal microembolization of calcium particles during valve manipulation, and myocardial ischemia related to transient hypotension during rapid ventricular pacing are considered to be principally involved, along with direct left ventricular trauma in transapical procedures (Figure 6). , , , , An embolic cause is further supported by cardiac magnetic resonance findings of multifocal small‐sized new myocardial late enhancements after TAVI of subendocardial or intramural localization as well as by in vitro models of aortic valvuloplasty. , Concomitant coronary artery disease further aggravates myocardial oxygen supply–demand mismatch that may occur during procedural phases of hypotension. In this study, patients with myocardial injury after TAVI were older, had higher surgical risk scores, more often presented with coronary artery disease and renal dysfunction, and more frequently had high‐grade mitral regurgitation. The association between known coronary artery disease as well as renal dysfunction and an increased occurrence of myocardial injury after TAVI has previously been reported. , , In these susceptible patients, preventive measures to reduce the amount of myocardial injury may improve prognosis after TAVI.

Figure 6

Schematic illustration of pathophysiological mechanisms contributing to increases of cardiac troponin levels in patients undergoing transcatheter aortic valve implantation.

 

Postprocedural Myocardial Injury and Outcomes in Patients With TAVI

Myocardial injury after TAVI has been related to worse prognosis in most studies. , , To date, thresholds for the definition of postprocedural myocardial injury after TAVI are ambiguous, not well established, and may vary among studies. Given the lack of scientific evidence in the field, the Valve Academic Research Consortium‐2 consensus definition of myocardial infarction after TAVI arbitrarily included a 15‐fold increase in cardiac troponin levels. In this study, spline curve analysis identified a threshold of 18 as optimal cutoff value of postprocedural hs‐cTnT elevations for long‐term mortality prediction, with proven robustness in multiple subgroup, multivariable, and landmark analyses. Relations of postprocedural hs‐cTnT elevations with mortality proved to be significant also in high‐risk patient subsets including those with coronary artery disease, renal dysfunction, and increased surgical risk. Hence, the proposed definition of clinically relevant myocardial injury after TAVI is based on a large contemporary cohort of patients and sound statistical methodology.

Incorporating elevated postprocedural troponin levels complements postprocedural risk assessment and therefore guides further management of patients undergoing TAVI. A comprehensive risk stratification in patients with recent TAVI allows for the identification of patients in need of intense postprocedural care, closer monitoring, and prompt follow‐up. Furthermore, the use of a standardized and appropriate definition of clinically relevant myocardial injury after TAVI would provide uniformity for the comparability of clinical and healthcare studies as well as the assessment of patient outcomes and quality initiatives. Although the need for sex‐related definitions of reference values is widely recognized, no sex‐related differences in the relation of postprocedural peak hs‐cTnT levels and mortality were observed.

Whether myocardial ischemia as reflected by increased hs‐cTnT levels directly drives mortality after TAVI, or whether elevated hs‐cTnT levels represent a surrogate of a highly comorbid patient population at increased risk, remains to be determined. Furthermore, whether a causal relationship between increased rates of in‐hospital complications including vascular complications, renal failure, and repeated unplanned interventions and the occurrence of postprocedural hs‐cTnT elevations exists needs to be further elucidated.

Limitations

Some limitations merit consideration. The prospective observational study enrolled a large contemporary cohort of patients undergoing TAVR and included comprehensive clinical, procedural, and outcome data, along with systematically measured postprocedural hs‐cTnT levels. The study is, however, inherently subject to the limitation of a single‐center design. Furthermore, although established risk factors were incorporated into the multivariable models, we cannot exclude that unmeasured or unknown confounding factors may have affected the observed associations of postprocedural peak hs‐cTnT levels with outcomes after TAVI.

Conclusions

This contemporary study for the first time established a hs‐cTnT threshold for the definition of postprocedural myocardial injury after TAVI. An elevation of hs‐cTnT ≥18‐fold the URL was recognized as most appropriate for the identification of patients with TAVI at increased risk of long‐term mortality. Hence, clinically relevant postprocedural myocardial injury should be incorporated into patient stratification after TAVI to further improve postprocedural patient care, management, and outcomes.

Sources of Funding

The SwissTAVI Registry is supported by a study grant from the Swiss Heart Foundation and the Swiss Working Group of Interventional Cardiology and Acute Coronary Syndromes and is sponsored by research grants from Medtronic, Edwards Lifesciences, Boston Scientific, and Abbott.

Disclosures

The authors have nothing to disclose regarding the content of the paper.

Tables S1–S3

Click here for additional data file.