Comparison of late mortality after transcatheter aortic valve implantation versus surgical aortic valve replacement: Insights from a meta-analysis

EJINME-03463; No of Pages 7 European Journal of Internal Medicine xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

European Journal of Internal Medicine journal homepage: www.elsevier.com/locate/ejim

Original Article

Comparison of late mortality after transcatheter aortic valve implantation versus surgical aortic valve replacement: Insights from a meta-analysis Tomo Ando a,⁎, Hisato Takagi b, Representing ALICE (All-Literature Investigation of Cardiovascular Evidence) Group: a b

Detroit Medical Center, Department of Cardiology, Detroit, MI, United States Shizuoka Medical Center, Department of Cardiovascular Surgery, Shizuoka, Japan

a r t i c l e

i n f o

Article history: Received 11 July 2016 Received in revised form 19 November 2016 Accepted 28 January 2017 Available online xxxx Keywords: Transcatheter aortic valve implantation Surgical aortic valve replacement Mortality

a b s t r a c t Introduction: Transcatheter aortic valve implantation (TAVI) has shown non-inferior late mortality in severe aortic stenosis (AS) patients in intermediate to inoperable risk for surgery compared to surgical aortic valve replacement (SAVR). Late outcome of TAVI compared to SAVR is crucial as the number of TAVI continues to increase over the last few years. Methods: A comprehensive literature search of PUBMED and EMBASE were conducted. Inclusion criteria were that [1] study design was a randomized controlled trial (RCT) or a propensity-score matched (PSM) study: [2] outcomes included N 2-year all-cause mortality in both TAVI and SAVR. The random-effects model was utilized to calculate an overall effect size of TAVI compared to SAVR in all-cause mortality. Publication bias was assessed quantitatively with Egger's test. Results: A total of 14 studies with 6503 (3292 TAVI and 3211 SAVR, respectively) were included in the metaanalysis. There was no difference in late all-cause mortality between TAVI and SAVR (HR 1.17, 95%CI 0.98– 1.41, p = 0.08, I2 = 61%). The sub-group analysis of all-cause mortality of RCT (HR 0.93 95%CI 0.78–1.10, p = 0.38, I2 = 40%) and PSM studies (HR 1.44 95%CI 1.15–1.80, p = 0.02, I2 = 35%) differed significantly (p for subgroup differences = 0.002). Meta-regression implicated that increased age and co-existing CAD may be associated with more advantageous effects of TAVI relative to SAVR on reducing late mortality. There was no evidence of significant publication bias (p = 0.19 for Egger's test). Conclusions: TAVI conferred similar late all-cause mortality compared to SAVR in a meta-analysis of RCT but had worse outcomes in a meta-analysis of PSM. © 2017 European Federation of Internal Medicine. Published by Elsevier B.V. All rights reserved.

1. Introduction Since the successful first in man case of transcatheter aortic valve implantation (TAVI) in a patient with inoperable, symptomatic severe aortic stenosis (AS) in 2002 by Cribier et al., the number of TAVI performed worldwide has been dramatically increasing [1–3]. TAVI has shown promising result not only for severe AS patients in high operative risk but also in intermediate surgical risk [4–6]. Patients at intermediate risk are expected to have the longer life expectancy after TAVI compared to those at high or inoperable risk. The data on late outcomes after TAVI is starting to accumulate. Several studies have reported 3 to 7 years of outcome data after TAVI [7–13].

Abbreviations: AS, aortic stenosis; CAD, coronary artery disease; HR, hazards ratio; NOTION, Nordic Aortic Valve Intervention; OR, odds ratio; PARTNER, Placement of AorRTic TraNscathetER; PSM, propensity-matched; RCT, randomized control trial; SAVR, surgical aortic valve replacement; TAVI, transcatheter aortic valve implantation. ⁎ Corresponding author at: 3990, John R, Detroit, MI 48201, United States. E-mail address: [email protected] (T. Ando).

However, a number of studies that have reported comparative late outcomes between TAVI and surgical aortic valve replacement (SAVR) are relatively limited. The United States CoreValve Registry showed allcause mortality favoring TAVI (p = 0.068) during 3 years follow-up, while the Placement of AorRTic TraNscathetER (PARTNER) valve trial have reported similar 5 years all-cause mortality [14,15]. Recently, a meta-analysis of long-term outcomes (N1 year) between TAVI and SAVR has been reported using odds ratio (OR) [16]. However, an estimate of late outcome is better assessed with hazards ratio (HR) than OR [17]. Recent other meta-analyses showed improved mortality in TAVI than SAVR during up to 2 years or median of 2 (range 3 months to 3 years) years follow-up with limited number of studies [18,19]. We have previously published meta-analysis of TAVI vs SAVR using propensity-score analysis and concluded that TAVI had worse outcome compared to SAVR [20]. In the same report, meta-analysis of 4 randomized clinical trials (RCT) [15,21–23] was performed. However, our previous report included in that studies, approximately half (10 studies) had follow-up duration b 2 years [20]. In addition, out of the 4 RCTs included only two had follow-up N2 years [15,22]. It is of great

http://dx.doi.org/10.1016/j.ejim.2017.01.023 0953-6205/© 2017 European Federation of Internal Medicine. Published by Elsevier B.V. All rights reserved.

Please cite this article as: Ando T, Takagi H, Comparison of late mortality after transcatheter aortic valve implantation versus surgical aortic valve replacement: Insights from..., Eur J Intern Med (2017), http://dx.doi.org/10.1016/j.ejim.2017.01.023

2

T. Ando, H. TakagiEuropean Journal of Internal Medicine xxx (2017) xxx–xxx

clinical importance to assess the comparative late outcomes between TAVI and SAVR to better provide physicians and patients the best available evidence to aid their decisions. Therefore, we aimed to compare further late outcomes (≥2 years) through systematic review and meta-analysis using HR to better estimate the effect size of TAVI versus SAVR. 2. Methods 2.1. Literature search strategy A systematic literature search of PUBMED and EMBASE was performed by two independent reviewers (T.A. and H.T.). There were no language limitations and conference abstracts were excluded. Search was performed on March 23rd, 2016 from January 1st, 2004. Search terms were “aortic valve” AND (percutaneous OR transcatheter OR transluminal OR transarterial OR transapical OR transaortic OR transcarotid OR transaxillary OR transsubclavian OR transiliac OR transfemoral OR transiliofemoral OR “Transcatheter Aortic Valve Replacement” [Mesh]) AND (mortality OR death OR deaths OR survival) AND (propensity OR randomized control). Titles and/or abstracts were screened based on inclusion and exclusion criteria. Studies were included when 1: All-cause mortality was reported for more than two years follow-up in both TAVI and SAVR cohort represented by survival curve 2: HR for all-cause mortality or OR were able to abstract or calculate for late events from the provided information 3: if the study design were either RCT or propensity-matched (PSM) cohort study. Exclusion criteria were 1: single arm study of either TAVI or SAVR. 2: when TAVI outcomes were compared with sutureless aortic valve replacement outcomes and 3: conference abstracts. When more than one study was reported in the same database, the study with the longest study was selected. When the follow-up duration was the same, the study with the most cohorts was selected. The meta-analysis was performed in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) and MOOSE (Meta-analysis of Observational Studies in Epidemiology) guidelines [24,25]. 2.2. Study quality assessment Quality of the studies was assessed using Cochrane risk of bias tool for RCT's and Newcastle-Ottawa scale for PSM studies [26,27]. Disagreements were resolved by a discussion by the 2 authors (T.A. and H.T.) to reach a consensus. 2.3. Statistical analysis Categorical values were expressed as percentages and continuous values as a mean ± standard deviation. The generic inverse variance was used to calculate the pooled HR/OR and 95% confidence intervals (CI). When HR was not available from the text, first we used methods proposed by Parmar et al. and when crude event rates were not available, methods proposed by Tierney et al. [28,29] was utilized to estimate the HR and its 95%CI from the Kaplan-Meier curve analysis. Publication bias was assessed visually from the funnel plot and if concern existed for asymmetry, Egger's test was utilized to quantitatively assess the publication bias. Publication bias was not assessed when b 10 studies were included in the meta-analysis for each outcome. Meta-regression analysis was performed by unrestricted maximum likelihood method with continuous variables as a moderator. Variables were pre-specified for the meta-regression analysis. A p-value b 0.05 was considered significant. When 95% CI of either OR or HR for TAVI vs SAVR does not cross one, TAVI (upper limit of 95% CI is below one) or SAVR (lower limit of 95% CI is above on) significantly favors reducing clinical outcomes. Metaanalyses were performed with the Review Manager (RevMan) Version 5.3 (Nordic Cochrane Centre, The Cochrane Collaboration, 2012,

Copenhagen, Denmark) and Comprehensive Meta-analysis version 2 and 3 (Biostat, Englewood, NJ)

3. Results Our search result yielded a total of 616 results. Detailed study flow diagram is shown in Table 1. A total of 14 studies including 6503 (3292 TAVI and 3211 SAVR) patients were finally analyzed in this meta-analysis. Four studies [6,14,15,30] were RCT and 10 were PSM studies [31–40]. There were few significant baseline characteristic differences in certain studies. Age [34,38], surgical risk score [38], previous coronary artery bypass graft [34,38] and peripheral artery disease [34, 38] were higher in TAVI for certain studies while male [34], diabetes [14], and peripheral disease [6] were reported to be more prevalent in SAVR. Characteristics of patients included in each study for TAVI and SAVR are summarized in table 1. Overall, the patients' baselines were well matched in all studies. The flow of the study selection is summarized in Fig. 1. All-cause mortality did not differ between TAVI and SAVR (HR 1.17, 95%CI 0.98–1.41, p = 0.08, I2 = 61%). (Fig. 2) Because we observed high heterogeneity, we performed sensitivity analysis. First, we performed a subgroup analysis based on study design, RCT or a PSM study. Meta-analysis of only the RCT (1895 TAVI and 1866 SAVR) did not show the difference in all-cause mortality between TAVI and SAVR (HR 0.93, 95%CI 0.78–1.10, p = 0.38, I2 = 40%). However, pooled HR of only the PSM studies (1397 TAVI and 1345 SAVR) showed significantly worse mortality in TAVI (HR 1.44, 95%CI 1.15–1.80, p = 0.002, I2 = 35%). There was significant heterogeneity between subgroup (I2 = 89.3%, p for subgroup differences = 0.002). This resulted in non-significant heterogeneity in both groups. Second, study with the largest weight (19.4% = 10.8% for transfemoral +8.6% for transthoracic) [6] was removed but did not significantly alter the result (HR 1.23, 95%CI 0.99985–1.51, p = 0.05, I2 = 60%). Third, studies with a cohort of b 100 were removed (PMID: 26, 30, 32–34) but the result was consistent (HR 1.19, 95%CI 0.94–1.52, p = 0.16, I2 = 73%). Fourth, every each study was removed once at a time. When a study by Deeb et al. [14] was removed, the overall effect was in favor of SAVR (HR 1.23, 95%CI 1.02–1.48, p = 0.03, I2 = 56%). However, when each study was removed at a time for the each subgroup (RCT only meta-analysis and PSM only meta-analysis), it did not significantly affect the result. We performed a meta-regression analysis with a pre-specified clinical baseline to further investigate the source of heterogeneity. As the age or the prevalence of coronary artery disease (CAD) increases, the HR of late mortality (for TAVI vs SAVR) significantly decreases. This result implicates that TAVI was more effective in reducing the HR of late-mortality compared to SAVR in higher age and patients with CAD. The results of meta-regression for other clinical baselines are summarized in Table 2 and were not significant. Late cardiovascular mortality (1895 and 1866 patients from TAVI and SAVR, respectively, from 4 RCT) did not differ between TAVI and SAVR (OR 1.01, 95%CI 0.78–1.29, p = 0.95, I2 = 43%). From the same 4 RCT studies, late stroke (OR 0.83 95%CI 0.65–1.05, p = 0.13, I2 = 0%) and myocardial infarction (OR 0.83 95%CI 0.57–1.21, p = 0.33, I2 = 0%) were also similar between TAVI and SAVR. There was no significant heterogeneity observed for late cardiovascular mortality, myocardial infarction and stroke. Sensitivity analysis (by omitting one study at a time and recalculating the pooled effect size) did not significantly alter the pooled effect size for cardiovascular mortality, stroke, and myocardial infarction. The results are summarized in Fig. 3. Publication bias was assessed for all-cause mortality including 14 studies and did not show significant publication bias (p = 0.19 for Egger's test). For other outcomes, publication bias was not assessed because the included studies were b 10.

Please cite this article as: Ando T, Takagi H, Comparison of late mortality after transcatheter aortic valve implantation versus surgical aortic valve replacement: Insights from..., Eur J Intern Med (2017), http://dx.doi.org/10.1016/j.ejim.2017.01.023

T. Ando, H. TakagiEuropean Journal of Internal Medicine xxx (2017) xxx–xxx

3

Table 1 Characteristics summary of included studies. Author

Year

Design

Cohort

Age

Male

Surgical risk score

TAVI

SAVR

TAVI

SAVR

TAVI

SAVR

TAVI

SAVR

2.9 ± 1.6a 8.4 ± 4.0b 7.3 ± 3.0a 17.7 ± 13.0b 9.9 ± 6.9b 23 ± 15b 5.8 ± 2.1a 11.8 ± 3.3a 29.3 ± 16.5b 8.2 ± 4.2a 19.5 ± 6.7b 11.1 ± 2.8a 24 ± 6b 8.7 ± 2.7b 12.1 ± 10.0a 36.4 ± 17.4b NRc 18.7 ± 11.1b 6.6a (5.4–10) 29.9 ± 14.0b

3.1 ± 1.7a 8.9 ± 5.5b 7.5 ± 3.3a 18.8 ± 13.2b 9.9 ± 6.4b 20 ± 14b 5.8 ± 1.9a 11.7 ± 3.5a 29.2 ± 15.6b 8.3 ± 4.4a 19.2 ± 7.4b 10.4 ± 3a 19 ± 6b 8.8 ± 2.8b 7.1 ± 5.2a 22.2 ± 17.5b NR 18.3 ± 14.0b 6.1a (4.8–10.3) 26.4 ± 12.9b

Sφndergaard [30]

2016

RCT

145

135

79.2 ± 4.9

79.0 ± 4.7

53.8

52.6

Deeb [14]

2016

RCT

391

359

83.2 ± 7.1

83.3 ± 6.4

52.9

52.4

Fraccaro [31] Johansson [34] Leon [6] Mack [15]

2016 2016 2016 2015

PSM PSM RCT RCT

415 166 1011 348

415 125 1021 351

83.7 ± 2.6 80 ± 9 81.5 ± 6.7 83.6 ± 6.8

83.7 ± 2.9 78 ± 6 81.7 ± 6.7 84.5 ± 6.4

40.0 51 54.2 57.8

38.1 63 54.8 56.7

Muneretto[35]

2015

PSM

204

204

80 ± 2

80 ± 3

55.4

52

Papadopoulos [36]

2014

PSM

40

40

81 ± 4

80 ± 3

73

73

Schymik [37] Wendt [38]

2015 2015

PSM PSM

216 62

216 51

78.3 ± 5.2 78.7 ± 5.9

78.2 ± 4.6 71.1 ± 10.8

46.3 30.6

51.4 25.5

Zweng [40] Holzhey [33] Fusari [32] Wilbring [39]

2015 2012 2012 2012

PMC PSM PSM PSM

44 167 30 53

44 167 30 53

82.3 ± 4.5 80.5 ± 4.6 80.5 (75–83) 78.1 ± 5.5

82.2 ± 4.4 79.8 ± 5.4 77.5 (77–81) 77.6 ± 2.7

41 35.3 20 65

32 35.3 46.7 66

Author

Sφndergaard [30] Deeb [14] Fraccaro [31] Johansson [34] Leon [6] Mack [15] Muneretto [35] Papadopoulos [36] Schymik [37] Wendt [38] Zweng [40] Holzhey [33] Fusari [32] Wilbring [39] Author

Sφndergaard [30] Deeb [14] Fraccaro [31] Johansson [34] Leon [6] Mack [15] Muneretto [35] Papadopoulos [36] Schymik [37] Wendt [38] Zweng [40] Holzhey [33] Fusari [32] Wilbring [39]

HTN (%)

DM (%)

PAD (%)

Transfemoral (%)

Pulmonary disease (%)

TAVI

SAVR

TAVI

SAVR

TAVI

SAVR

TAVI

SAVR

TAVI

SAVR

71.0 95.1 NR NR NR NR 63.2 45 NR 91.9 82 84.4 93.3 NR

76.3 96.1 NR NR NR NR 66.1 40 NR 88.2 84 87.4 76.7 NR

17.9 34.8 19.3 24 37.7

20.7 45.1 18.6 16 34.2 NR 26.4 35 NR 43.1 36 44.3 6.7 43.4

4.1 41.0 17.3 52 27.9 43.2 21 33 5.1 52.4 NR 27.5 40 NR

6.7 42.0 17.3 39 32.9 41.6 22.6 27 6.9 29.4 NR 25.7 30 NR

96.5 82.8 NR 45 76.3 70.1 74.5 0 NR NR NR 0 53.3 NR

– – – – – – – – – – – – – –

11.7 NR 17.6 18 31.8 43.7 27.4 23 9.3 25.8 NR 13.8 30 9.4

11.9 NR 15.2 15 30.0 43.0 26.4 20 8.8 33.3 NR 13.8 33.3 7.5

30.3 42 NR 38.7 32 39.5 23.3 52.8

BMI (kg/m2)

Prior CABG (%)

CAD (%)

TAVI

SAVR

TAVI

SAVR

TAVI

SAVR

NR 29.4 4.3 49 23.6 42.5 NR NR NR 87.1 23 NR NR 73.6

NR 31.5 3.6 32 25.6 43.6 NR NR NR 64.7 18 NR NR 69.8

NR NR 25.9 ± 4.3 27 ± 5 28.6 ± 6.2 NR 26.9 ± 5.3 NR NR 27.1 ± 4.1 NR 26.3 ± 4.6 24.1 (23.2–26.2) 27.9 ± 4.0

NR NR 26.2 ± 4.1 27 ± 4 28.3 ± 6.2 NR 26.8 ± 3.2 NR NR 26.6 ± 3.7 NR 26.3 ± 3.9 26.3 (22.9–29.4) 27.3 ± 4.2

NR 75.4 NR NR 69.2 74.9 25.9 83.0 48.1 NR NR NR 36.7 100

NR 76.0 NA NR 66.5 76.9 24.0 75.0 48.1 NR NR NR 33.3 100

TAVI: transcatheter aortic valve implantation, SAVR: surgical aortic valve replacement, RCT: randomized clinical trial, PSM: propensity matched, HTN, hypertension, DM: diabetes mellitus, PAD: peripheral artery disease, BMI: body mass index, CABG: coronary artery bypass graft, CAD: coronary artery disease NR: not reported, NA: not applicable. a Society of thoracic surgeons score. b Logistic EuroSCORE. c Age, body mass index and surgical risk score is expressed in either mean ± standard deviation or interquartile range.

Study quality for RCTs showed risk of bias was low for all the studies. Newcastle-Ottawa score showed N7 for all included studies for the PSM studies. 4. Discussion Our major findings of this meta-analysis were that 1: Late (≥2 years) all-cause mortality of TAVI was comparable to SAVR. However, the all-

cause mortality of RCT and PSM studies differed significantly. 2: Metaregression implicated that as the age or proportion of CAD increased, TAVI was more advantageous in reducing mortality. Late all-cause mortality in TAVI was comparable to SAVR. However, the all-cause mortality of RCT and PSM studies differed significantly. Siontis et al. reported 13% relative risk reduction in mortality [18] in TAVI than SAVR with 4 RCTs that reported follow-up duration only up to 2 years [4,6,22,30]. Another recent meta-analysis by Siemieniuk

Please cite this article as: Ando T, Takagi H, Comparison of late mortality after transcatheter aortic valve implantation versus surgical aortic valve replacement: Insights from..., Eur J Intern Med (2017), http://dx.doi.org/10.1016/j.ejim.2017.01.023

4

T. Ando, H. TakagiEuropean Journal of Internal Medicine xxx (2017) xxx–xxx

Fig. 1. Flow chart for study selection.

et al. focused mainly on the outcomes of low-intermediate surgical risk cohorts for aortic valve replacement [19]. They concluded TAVI conferred 21% reduction in mortality risk up to 3 years [6,14,30]. We have included the data of longest follow-up duration in order to assess the late-mortality. This is the reason for discrepancy from previous metaanalysis and what makes our results unique and important. TAVI patients included in the RCT generally have more strict inclusion/exclusion criteria than the real world practice where patient candidacy for TAVI is decided by the “heart team” and would likely be one of the main reasons for the discrepancy. There were some clinical features set for exclusion criteria in the RCTs but not in the PSM studies. Severe renal insufficiency and/or chronic dialysis patients, patients with a pre-existing prosthetic heart valve, prior cardiac surgery were excluded from the included RCTs [6,14,15,30] These clinical features were often included in the PSM studies included in this meta-analysis and also may account for

the difference observed between RCT and PSM subgroup metaanalysis. Outcome of TAVI is known to improve with operator/center experience. Several data suggest that roughly 30–80 cases of experiences were required to achieve better outcomes [41,42]. Most of the centers included in the analysis were large volume, highly experienced centers in both TAVI and SAVR. TAVI has increased not only the volume of SAVR but also led to improved outcomes [43]. Sensitivity analysis showed that when the study removed by Deeb et al. [14] with the largest weight in favor of TAVI was removed, the pooled HR became significant. The authors have considered the favorable outcome in TAVI to low peri-procedural complications, lower patient-prosthesis mismatch and better improvement in health status. Also, patients with a pre-existing prosthetic heart valve, which could negatively affect TAVI outcomes [44], were also excluded in the study by Adams et al. [45]. Our meta-regression results suggest that the follow-up duration did not affect mortality of TAVI compared to SAVR. This implies that the follow-up duration did not affect the mortality and TAVI could confer similar long-term mortality compared to SAVR. However, the data for long-term valve durability is still limited for the transcatheter prosthetic valves and further longer follow-up duration could affect the outcomes, although the available data up to 5 years shows similar or even better valve performance than bioprostheses of SAVR [14,15]. Recently, a meta-analysis by Gargiulo et al. reported long-term outcomes in TAVI versus SAVR. They defined long-term mortality as N1 year and used OR to summarize and compare the effect size of TAVI and SAVR for long-term mortality. They reported OR 1.03 and 95%CI 0.65–1.62 for the meta-analysis of RCTs, OR 1.70 and 95%CI 1.23–2.35 for PSM cohort studies and OR 1.28 with 95%CI 0.97–1.69 for the total study included [16]. Our study used HR, which better assess, especially outcomes for long-term mortality, and resulted in more narrow 95%CI. In addition, heterogeneity for the meta-analysis for RCT in our study was non-significant as opposed to result by Gargiulo et al. However, the result was the same for both meta-analysis of RCT, PSM, and both study design combined. Assessment of long-term outcomes in TAVI is crucial to further expand the indication of TAVI to low surgical risk patients as these patients would have longer life expectancy. The Nordic Aortic Valve Intervention (NOTION) trial included patients at low risk and mean surgical risk was lower than the previous RCTs but the 2-year all-cause mortality was similar between TAVI and SAVR [30]. In the exploratory sub-group analysis from the NOTION trial, patients with STS-PROM score b 4%, TAVI conferred similar mortality to SAVR [30]. However, Rosato et al. reported worse mortality and the higher rate of major cardiac and cerebrovascular events in TAVI compared to SAVR in patients with EuroSCORE II b 4% [46]. Their study did not exclude subjects on dialysis therapy, history of cardiac surgery and prior stroke and that may have led to the difference from the results of NOTION trial. Valve durability is another issue that requires long-term follow-up data. Gulino et al. reported that there was 4.6% of progression of paravalvular regurgitation from mild to moderate and 3.2% developed prosthetic valve failure (moderate/severe restenosis, endocarditis with severe intraprosthetic aortic regurgitation and moderate transvalvular regurgitation) during 4-year follow-up with the CoreValve [13]. Others reported durability on the Sapien valve. Mack et al. demonstrated similar mean valve area and mean gradient between TAVI and SAVR over 5years follow-up. In addition, there was no case that required surgical intervention for valve deterioration in both groups [15]. Although these results appear promising, bioprosthesis used for SAVR have longer durability experience. Bourguignon and colleagues reported 19.7 years of expected valve durability with the Carpentier-Edwards Perimount pericardial bioprosthesis [47]. Therefore, TAVI still requires data for longer follow-up, especially given recent evidence of similar outcome in intermediate patients [6]. Meta-regression implicated that increased age and co-existing CAD may be associated with more advantageous effects of TAVI relative to

Please cite this article as: Ando T, Takagi H, Comparison of late mortality after transcatheter aortic valve implantation versus surgical aortic valve replacement: Insights from..., Eur J Intern Med (2017), http://dx.doi.org/10.1016/j.ejim.2017.01.023

T. Ando, H. TakagiEuropean Journal of Internal Medicine xxx (2017) xxx–xxx

5

Fig. 2. Random-effects model forest plot for all-cause mortality RCT: randomized control trial, PSM: propensity-matched.

SAVR on reducing late mortality. Considering the less invasiveness of TAVI over SAVR, especially the transfemoral approach, it is understandable that TAVI could be more advantageous than SAVR in higher-age patients. Age has been reported to be associated with worse outcomes in both TAVI and SAVR [9,48,49]. Subgroup analysis from the RCT of TAVI vs SAVR did not show the difference in higher age group [15,30]. This discordance may be partially explained by the difference statistical methods utilized. Also, there could be a difference in other comorbidities between TAVI and SAVR in these older patients because it is an ad-hoc analysis. Few studies have suggested comparable outcomes of TAVI in nonagenarian patients [50,51]. These studies suggested similar outcomes, however, Arsalan et al. reported worse one-year mortality in nonagenarian (adjusted HR 1.20, p b 0.001) [52]. CAD has been associated with worse outcomes in SAVR [53,54] but with conflicting result in TAVI [55–57]. One meta-analysis reported similar mid-term mortality in patients with the history of coronary artery bypass graft surgery in TAVI and SAVR [58]. Although our metaregression analysis suggests that TAVI may be more effective than SAVR in CAD patients with severe AS, heterogeneity of the definition of CAD and lack of information of the peri-operative revascularization

Table 2 Meta-regression analysis. Variable

Number of studies

Slope

Lower limit

Upper limit

p-Value

Age Men DM STS Logistic EuroSCORE Follow up length BMI CAD PAD Publication year

14 14 12 8 11 14 8 9 12 14

−0.091 −0.0088 −0.0082 0.035 0.015 0.11 −0.16 −0.014 0.0029 −0.054

−0.15 −0.027 −0.031 −0.050 −0.020 −0.017 −0.40 −0.0026 −0.012 −0.29

−0.0033 0.0088 0.015 0.12 0.038 0.24 0.074 −0.0023 0.0018 0.089

0.0021 0.33 0.48 0.18 0.53 0.088 0.18 0.019 0.70 0.33

DM: diabetes, STS: society of thoracic surgeon, BMI: body mass index, CAD: coronary artery disease, PAD: peripheral artery disease.

make this result hard to interpret. Therefore, this result should be considered hypothesis generating and confirmed in future studies. Our results should be viewed in with several limitations. First, not all the studies included were RCT and the result of the meta-analysis of RCT and PSM differed significantly. This is likely due to the difference in patient selection between RCT and PSM, which more reflects the “real world” practice. This finding may be utilized to improve the outcomes in TAVI compared to SAVR. Although we only included PSM studies as non-randomized studies, PSM studies are more susceptible to confounders and biases. Indeed, there were several differences in baseline comorbidities between TAVI and SAVR cohorts, which may have potentially affected the result. However, the numbers of studies with difference in baseline characteristics were limited. Therefore, our results should be viewed with caution. Meta-regression results are exploratory and the results should be viewed with caution. Second, only 4 RCT were included for other clinical outcomes besides all-cause mortality, hence the results may not be robust to conclude the findings. Third, although relevant articles were searched rigorously and strict inclusion/exclusion criteria were defined, publication bias could not be completely excluded. However, there was no evidence of publication bias on Egger's test. Forth, data on late cardiovascular mortality, stroke and myocardial infarction were available only from the RCTs and therefore 4 studies were included. However, there was no significant heterogeneity and sensitivity analysis showed consistent result, suggesting robust result. Lastly, the only maximum of 5 years has been assessed for the follow-up duration and longer TAVI valve durability needs to be observed.

5. Conclusions In this meta-analysis, TAVR was associated with similar late allcause mortality compared to SAVR in the RCT but had worse all-cause mortality in a meta-analysis from PMS. Exploratory meta-regression suggested that with an increase in age and proportion of CAD, TAVI may be a better mode of aortic valve replacement for severe AS. These findings need to be confirmed with further studies.

Please cite this article as: Ando T, Takagi H, Comparison of late mortality after transcatheter aortic valve implantation versus surgical aortic valve replacement: Insights from..., Eur J Intern Med (2017), http://dx.doi.org/10.1016/j.ejim.2017.01.023

6

T. Ando, H. TakagiEuropean Journal of Internal Medicine xxx (2017) xxx–xxx

Fig. 3. Random-effects model forest plot for cardiovascular mortality, stroke and myocardial infarction.

Disclosure Authors have no disclosures. Conflict of interest statement All authors have no conflict of interest to disclose. Acknowledgements To my wife and daughter for giving me pleasure every day. References [1] Cribier A, Eltchaninoff H, Bash A, Borenstein N, Tron C, Bauer F, et al. Percutaneous transcatheter implantation of an aortic valve prosthesis for calcific aortic stenosis: first human case description. Circulation 2002;106:3006–8. [2] Reinöhl J, Kaier K, Reinecke H, Schmoor C, Frankenstein L, Vach W, et al. Effect of availability of transcatheter aortic-valve replacement on clinical practice. J Med]– >N Engl J Med 2015;373:2438–47. [3] Holmes Jr DR, Nishimura RA, Grover FL, Brindis RG, Carroll JD, Edwards FH, et al. Annual outcomes with transcatheter valve therapy: from the STS/ACC TVT registry. J Am Coll Cardiol 2015;66:2813–23. [4] Kodali SK, Williams MR, Smith CR, Svensson LG, Webb JG, Makkar RR, et al. Twoyear outcomes after transcatheter or surgical aortic-valve replacement. J Med]–>N Engl J Med 2012;366:1686–95. [5] Thourani VH, Kodali S, Makkar RR, Herrmann HC, Williams M, Babaliaros V, et al. Transcatheter aortic valve replacement versus surgical valve replacement in intermediate-risk patients: a propensity score analysis. Lancet 2016;387:2218–25. [6] Leon MB, Smith CR, Mack MJ, Makkar RR, Svensson LG, Kodali SK, et al. Transcatheter or surgical aortic-valve replacement in intermediate-risk patients. N Engl J Med 2016;374:1609–20. [7] Ruparelia N, Latib A, Buzzatti N, Giannini F, Figini F, Mangieri A, et al. Long-term outcomes after transcatheter aortic valve implantation from a single high-volume center (the Milan experience). J Cardiol]–>Am J Cardiol 2016;117:813–9.

[8] Salinas P, Moreno R, Calvo L, Sánchez-Recalde Á, Jiménez-Valero S, Galeote G, et al. Long-term follow-up after transcatheter aortic valve implantation for severe aortic stenosis. Rev Esp Cardiol (Engl Ed) 2016;69:37–44. [9] Duncan A, Ludman P, Banya W, Cunningham D, Marlee D, Davies S, et al. Long-term outcomes after transcatheter aortic valve replacement in high-risk patients with severe aortic stenosis: the U.K. Transcatheter Aortic Valve Implantation Registry. JACC Cardiovasc Interv 2015;8:645–53. [10] Barbanti M, Schiltgen M, Verdoliva S, Bosmans J, Bleiziffer S, Gerckens U, et al. Threeyear outcomes of transcatheter aortic valve implantation in patients with varying levels of surgical risk (from the CoreValve ADVANCE Study). J Cardiol]–>Am J Cardiol 2016;117:820–7. [11] Escárcega RO, Lipinski MJ, Baker NC, Magalhaes MA, Minha S, Torguson R, et al. Analysis of long-term survival following transcatheter aortic valve implantation from a single high-volume center. J Cardiol]–>Am J Cardiol 2015;116:256–63. [12] Toggweiler S, Humphries KH, Lee M, Binder RK, Moss RR, Freeman M, et al. 5-year outcome after transcatheter aortic valve implantation. J Am Coll Cardiol 2013;61: 413–9. [13] Gulino S, Barbanti M, Deste W, Immè S, Aruta P, Bottari V, et al. Four-year durability of clinical and haemodynamic outcomes of transcatheter aortic valve implantation with the self-expanding CoreValve. EuroIntervention 2015;11 (pii: 20140923-06). [14] Deeb GM, Reardon MJ, Chetcuti S, Patel HJ, Grossman PM, Yakubov SJ, et al. 3-year outcomes in high-risk patients who underwent surgical or transcatheter aortic valve replacement. J Am Coll Cardiol 2016;67:2565–74. [15] Mack MJ, Leon MB, Smith CR, Miller DC, Moses JW, Tuzcu EM, et al. 5-year outcomes of transcatheter aortic valve replacement or surgical aortic valve replacement for high surgical risk patients with aortic stenosis (PARTNER 1): a randomised controlled trial. Lancet 2015;385:2477–84. [16] Gargiulo G, Sannino A, Capodanno D, Barbanti M, Buccheri S, Perrino C, et al. Transcatheter aortic valve implantation versus surgical aortic valve replacement: a systematic review and meta-analysis. Ann Intern Med 2016 Jun 7. [17] Deeks JJ, Higgins JP, Altman DG. Analysing data and undertaking meta-analyses. In: Higgins JP, Green S, editors. Cochrane handbook for systematic reviews of interventions version 5.1.0The Cochrane Collaboration; 2011 [updated March 2011, Chapter 9, Available from: www.cochrane-handbook.org]. [18] Siontis GC, Praz F, Pilgrim T, Mavridis D, Verma S, Salanti G, et al. Transcatheter aortic valve implantation vs. surgical aortic valve replacement for treatment of severe aortic stenosis: a meta-analysis of randomized trials. Eur Heart J 2016 Jul 7. [19] Siemieniuk RA, Agoritsas T, Manja V, Devji T, Chang Y, Bala MM, et al. Transcatheter versus surgical aortic valve replacement in patients with severe aortic stenosis at

Please cite this article as: Ando T, Takagi H, Comparison of late mortality after transcatheter aortic valve implantation versus surgical aortic valve replacement: Insights from..., Eur J Intern Med (2017), http://dx.doi.org/10.1016/j.ejim.2017.01.023

T. Ando, H. TakagiEuropean Journal of Internal Medicine xxx (2017) xxx–xxx

[20]

[21]

[22]

[23]

[24]

[25]

[26] [27]

[28] [29] [30]

[31]

[32]

[33]

[34]

[35]

[36]

[37]

[38]

low and intermediate risk: systematic review and meta-analysis. BMJ 2016;28(354): i5130. Takagi H, Umemoto T, ALICE (All-Literature Investigation of Cardiovascular Evidence) Group. Worse survival after transcatheter aortic valve implantation than surgical aortic valve replacement: a meta-analysis of observational studies with a propensity-score analysis. Int J Cardiol 2016;220:320–7. Nielsen HH, Klaaborg KE, Nissen H, Terp K, Mortensen PE, Kjeldsen BJ, et al. A prospective, randomised trial of transapical transcatheter aortic valve implantation vs. surgical aortic valve replacement in operable elderly patients with aortic stenosis: the STACCATO trial. EuroIntervention 2012;8:383–9. Reardon MJ, Adams DH, Kleiman NS, Yakubov SJ, Coselli JS, Deeb GM, et al. 2-year outcomes in patients undergoing surgical or self-expanding transcatheter aortic valve replacement. J Am Coll Cardiol 2015;66:113–21. Thyregod HG, Steinbrüchel DA, Ihlemann N, Nissen H, Kjeldsen BJ, Petursson P, et al. Transcatheter versus surgical aortic valve replacement in patients with severe aortic valve stenosis: 1-year results from the all-comers NOTION randomized clinical trial. J Am Coll Cardiol 2015;65:2184–94. Moher D, Liberati A, Tetzlaff J, Altman DG, PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann Intern Med 2009;151:264–9. Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, et al. Metaanalysis of observational studies in epidemiology: a proposal for reporting. Metaanalysis of Observational Studies in Epidemiology (MOOSE) group. JAMA 2000; 283:2008–12. Jüni P, Altman DG, Egger M. Systematic reviews in health care: Assessing the quality of controlled clinical trials. BMJ 2001;323:42–6. Kim SY, Park JE, Lee YJ, Seo HJ, Sheen SS, Hahn S, et al. Testing a tool for assessing the risk of bias for nonrandomized studies showed moderate reliability and promising validity. J Clin Epidemiol 2013;66:408–14. Parmar MK, Torri V, Stewart L. Extracting summary statistics to perform meta-analyses of the published literature for survival endpoints. Stat Med 1998;17:2815–34. Tierney JF, Stewart LA, Ghersi D, Burdett S, Sydes MR. Practical methods for incorporating summary time-to-event data into meta-analysis. Trials 2007 Jun 7;8:16. Søndergaard L, Steinbrüchel DA, Ihlemann N, Nissen H, Kjeldsen BJ, Petursson P, et al. Two-year outcomes in patients with severe aortic valve stenosis randomized to transcatheter versus surgical aortic valve replacement: the all-comers Nordic aortic valve intervention randomized clinical trial. Circ Cardiovasc Interv 2016 Jun;9(6) [pii: e003665]. Fraccaro C, Tarantini G, Rosato S, Tellaroli P, D'Errigo P, Tamburino C, et al. Early and midterm outcome of propensity-matched intermediate-risk patients aged ≥80 years with aortic stenosis undergoing surgical or transcatheter aortic valve replacement (from the Italian Multicenter OBSERVANT Study). J Cardiol]–>Am J Cardiol 2016; 117:1494–501. Fusari M, Bona V, Muratori M, Salvi L, Salis S, Tamborini G, et al. Transcatheter vs. surgical aortic valve replacement: a retrospective analysis assessing clinical effectiveness and safety. J Cardiovasc Med (Hagerstown) 2012;13:229–41. Holzhey DM, Shi W, Rastan A, Borger MA, Hänsig M, Mohr FW. Transapical versus conventional aortic valve replacement–a propensity-matched comparison. Heart Surg Forum 2012;15:E4–8. Johansson M, Nozohoor S, Bjursten H, Ragnarsson S, Götberg M, Kimblad PO, et al. Late survival and heart failure after transcatheter aortic valve implantation. Asian Cardiovasc Thorac Ann]–>Asian Cardiovasc Thorac Ann 2016;24:318-235. Muneretto C, Alfieri O, Cesana BM, Bisleri G, De Bonis M, Di Bartolomeo R, et al. A comparison of conventional surgery, transcatheter aortic valve replacement, and sutureless valves in “real-world” patients with aortic stenosis and intermediate- to high-risk profile. J Thorac Cardiovasc Surg 2015;150:1570–7. Papadopoulos N, Schiller N, Fichtlscherer S, Lehmann R, Weber CF, Moritz A, et al. Propensity matched analysis of longterm outcomes following transcatheter based aortic valve implantation versus classic aortic valve replacement in patients with previous cardiac surgery. J Cardiothorac Surg 2014;9:99. Schymik G, Heimeshoff M, Bramlage P, Herbinger T, Würth A, Pilz L, et al. A comparison of transcatheter aortic valve implantation and surgical aortic valve replacement in 1,141 patients with severe symptomatic aortic stenosis and less than high risk. Catheter Cardiovasc Interv 2015;86:738–44. Wendt D, Al-Rashid F, Kahlert P, El-Chilali K, Demircioglu E, Neuhäuser M, et al. Conventional aortic valve replacement or transcatheter aortic valve implantation in patients with previous cardiac surgery. J Cardiol 2015;66:292–7.

7

[39] Wilbring M, Tugtekin SM, Alexiou K, Simonis G, Matschke K, Kappert U. Transapical transcatheter aortic valve implantation vs conventional aortic valve replacement in high-risk patients with previous cardiac surgery: a propensity-score analysis. J Cardiothorac Surg]–>Eur J Cardiothorac Surg 2013;44:42–7. [40] Zweng I, Shi WY, Palmer S, MacIsaac A, Whitbourn R, Davis P, et al. Transcatheter versus surgical aortic valve replacement in high-risk patients: a propensity-score matched analysis. Heart Lung Circ 2016;25:661–7. [41] Arai T, Lefèvre T, Hovasse T, Hayashida K, Watanabe Y, O'Connor SA, et al. Evaluation of the learning curve for transcatheter aortic valve implantation via the transfemoral approach. J Cardiol]–>Int J Cardiol 2016;203:491–7. [42] Minha S, Waksman R, Satler LP, Torguson R, Alli O, Rihal CS, et al. Learning curves for transfemoral transcatheter aortic valve replacement in the PARTNER-I trial: success and safety. Catheter Cardiovasc Interv 2016;87:165–75. [43] Patel HJ, Herbert MA, Paone G, Heiser JC, Shannon FL, Theurer PF, et al. The midterm impact of transcatheter aortic valve replacement on surgical aortic valve replacement in Michigan. Ann Thorac Surg 2016;102:728–34. [44] Dvir D, Webb JG, Bleiziffer S, Pasic M, Waksman R, Kodali S, et al. Transcatheter aortic valve implantation in failed bioprosthetic surgical valves. JAMA 2014;312:162–70. [45] Adams DH, Popma JJ, Reardon MJ, Yakubov SJ, Coselli JS, Deeb GM, et al. Transcatheter aortic-valve replacement with a self-expanding prosthesis. N Engl J Med 2014; 370:1790–8. [46] Rosato S, Santini F, Barbanti M, Biancari F, D'Errigo P, Onorati F, et al. Transcatheter aortic valve implantation compared with surgical aortic valve replacement in lowrisk patients. Circ Cardiovasc Interv 2016;9:e003326. [47] Bourguignon T, Bouquiaux-Stablo AL, Candolfi P, Mirza A, Loardi C, May MA, et al. Very long-term outcomes of the Carpentier-Edwards Perimount valve in aortic position. Ann Thorac Surg 2015;99:831–7. [48] Unbehaun A, Pasic M, Drews T, Penkalla A, Dreysse S, Klein C, et al. Transapical aortic valve implantation: predictors of survival up to 5 years in 730 patients. An update†. J Cardiothorac Surg]–>Eur J Cardiothorac Surg 2015;47:281–90. [49] Vasques F, Lucenteforte E, Paone R, Mugelli A, Biancari F. Outcome of patients aged ≥80 years undergoing combined aortic valve replacement and coronary artery bypass grafting: a systematic review and meta-analysis of 40 studies. Am Heart J 2012;164:410–8. [50] Abramowitz Y, Chakravarty T, Jilaihawi H, Kashif M, Zadikany R, Lee C, et al. Comparison of outcomes of transcatheter aortic valve implantation in patients ≥90 years versus b90 years. J Cardiol]–>Am J Cardiol 2015;116:1110–5. [51] Yamamoto M, Meguro K, Mouillet G, Bergoend E, Monin JL, Lim P, et al. Comparison of effectiveness and safety of transcatheter aortic valve implantation in patients aged ≥90 years versus b90 years. J Cardiol]–>Am J Cardiol 2012;110:1156–63. [52] Arsalan M, Szerlip M, Vemulapalli S, Holper EM, Arnold SV, Li Z, et al. Should transcatheter aortic valve replacement be performed in nonagenarians?: insights from the STS/ACC TVT registry. J Am Coll Cardiol 2016;67:1387–95. [53] Beach JM, Mihaljevic T, Svensson LG, Rajeswaran J, Marwick T, Griffin B, et al. Coronary artery disease and outcomes of aortic valve replacement for severe aortic stenosis. J Am Coll Cardiol 2013;61:837–48. [54] Aranki SF, Rizzo RJ, Couper GS, Adams DH, Collins Jr JJ, Gildea JS, et al. Aortic valve replacement in the elderly. Effect of gender and coronary artery disease on operative mortality. Circulation 1993;88:II17–23. [55] Ludman PF, Moat N, de Belder MA, Blackman DJ, Duncan A, Banya W, et al. Transcatheter aortic valve implantation in the United Kingdom: temporal trends, predictors of outcome, and 6-year follow-up: a report from the UK Transcatheter Aortic Valve Implantation (TAVI) Registry, 2007 to 2012. Circulation 2015;131:1181–90. [56] Gilard M, Eltchaninoff H, Iung B, Donzeau-Gouge P, Chevreul K, Fajadet J, et al. Registry of transcatheter aortic-valve implantation in high-risk patients. N Engl J Med 2012;366:1705–15. [57] Linke A, Wenaweser P, Gerckens U, Tamburino C, Bosmans J, Bleiziffer S, et al. Treatment of aortic stenosis with a self-expanding transcatheter valve: the International Multi-centre ADVANCE Study. Eur Heart J 2014;35:2672–84. [58] Ando T, Briasoulis A, Holmes AA, Afonso L, Schreiber T, Kondur A. Transcatheter aortic valve replacement versus surgical aortic valve replacement in patients with previous coronary artery bypass surgery: a systematic review and meta-analysis. J Cardiol]–>Int J Cardiol 2016;215:14–9.

Please cite this article as: Ando T, Takagi H, Comparison of late mortality after transcatheter aortic valve implantation versus surgical aortic valve replacement: Insights from..., Eur J Intern Med (2017), http://dx.doi.org/10.1016/j.ejim.2017.01.023