Meta-analysis of longitudinal comparison of transcatheter versus surgical aortic valve replacement in patients at low to intermediate surgical risk

Abstract

Background:

Surgical aortic valve replacement (SAVR) is the commonly used approach for aortic valve replacement (AVR) in patients with aortic stenosis at low or intermediate surgical risk. However, transcatheter aortic valve replacement (TAVR) has emerged as an alternative to SAVR for AVR. This meta-analysis aims to assess the comparative efficacy and safety of TAVR versus SAVR in low-to-intermediate surgical risk patients by analyzing temporal trends in the outcomes of TAVR and SAVR at various follow-up intervals, providing a more detailed understanding.

Methods:

A thorough literature search was performed across PubMed/MEDLINE, Embase, and the Cochrane Library from their inception up to May 2024 to identify eligible randomized controlled trials (RCTs). Clinical outcomes were evaluated using a random-effects model to pool risk ratios (RRs) with 95% CIs.

Results:

A total of 17 studies reporting data at different follow-ups for nine trials were included (n=9092). No statistically significant difference was observed between TAVR and SAVR for reducing all-cause death at 30 days, 1 year, and 2 years but significantly increased risk with TAVR at 5 years or longer follow-up (RR=1.13, 95% CI: 1.03–1.23). However, TAVR was associated with a significantly decreased risk for cardiac death at 1-year follow-up (RR=0.79, 95% CI: 0.64–0.96) and comparable risk for cardiac death at 30 days, 2 years, and 5 years or longer follow-up when compared with SAVR. No statistically significant difference was observed between TAVR and SAVR for reducing the risk of myocardial infarction (MI) at 30 days, 1 year, 2 years, and 5 years or longer follow-up.

TAVR was associated with a significantly lower risk of major bleeding events at 30 days (RR=0.38, 95% CI: 0.21–0.67); lower risk of acute kidney injury (AKI) at 30 days (RR=0.38, 95% CI: 0.26–0.54) and 1 year (RR=0.58, 95% CI: 0.41–0.82) and lower risk of new onset or worsening atrial fibrillation (AF) at 30 days (RR=0.25, 95% CI: 0.18–0.34), 1 year (RR=0.26, 95% CI: 0.16–0.41) and 2 years (RR=0.32, 95% CI: 0.20–0.49) when compared with SAVR. However, TAVR was associated with a significantly increased risk of permanent pacemaker implantation (PPI) at 30 days (RR: 2.62, 95% CI: 1.40–4.91), at 1 year (RR: 2.19, 95% CI: 1.24–3.87), at 2 years (RR: 2.74, 95% CI: 1.31–5.71), and beyond 5 years (RR: 1.95, 95% CI: 1.20–3.15). TAVR was also associated with a significantly increased risk of prosthetic valve thrombosis at 2 years (RR=2.70, 95% CI: 1.08–6.71), though no significant association was observed at 30 days, 1 year, or 5 years. Similarly, no significant differences were observed in aortic-valve reintervention rates at 30 days, 2 years, or 5 years, but TAVR showed a significantly increased risk at 1 year (RR=1.98, 95% CI: 1.21–3.24). TAVR was associated with a significantly increased risk of major vascular complications at 30 days (RR=2.37, 95% CI: 1.38–4.04) and a significantly increased risk of TIA at 2 years (RR: 1.43, 95% CI: 1.02–2.00, I ²=0%). The risk of hospitalizations was comparable between the groups.

Conclusion:

TAVR and SAVR demonstrated comparable rates of all-cause death up to 2 years of follow-up. However, at 5 years or longer follow-up, TAVR was associated with a higher risk of all-cause death. While TAVR showed certain procedural advantages, such as a lower risk of major bleeding, AKI, and new-onset or worsening AF, the choice between TAVR and SAVR in patients with low or intermediate surgical risk should consider long-term outcomes, with SAVR potentially being more favorable due to better survival observed on longer follow-up durations.

Introduction

Highlights

What is already known

• Transcatheter aortic valve replacement (TAVR) is associated with improved clinical outcomes in patients at high surgical risk when compared with surgical aortic valve replacement (SAVR).

• New trials have shown that TAVR is noninferior to SAVR in patients at low to intermediate surgical risk at 1 year of follow-up. However, data about long-term safety and efficacy is limited.

What this study adds

• A longitudinal comparison of TAVR versus SAVR demonstrated a comparable risk of all-cause death at 30 days, 1 year, and 2 years of follow-up. However, at 5 years or longer follow-up TAVR was associated with an increased risk of all-cause death.

• While TAVR showed certain procedural advantages, the choice between TAVR and SAVR in patients with low or intermediate surgical risk should consider long-term outcomes, with SAVR potentially being more favorable due to better survival observed on longer follow-up durations.

Aortic stenosis (AS) is one of the most prevalent forms of valvular heart disease in the senior population¹. The frequency of severe, potentially fatal AS will rise as the population ages². With a 12.4% incidence of AS in North America and Europe, it may be inferred that over 291 000 candidates have received aortic valve replacements². It has been demonstrated that people with AS have reduced cardiac outflow, which raises cardiac workload and leads to heart failure and left ventricular hypertrophy. A prior study found that individuals with symptoms of AS, such as angina, syncope, or heart failure, experienced yearly death rates of close to 25%³. Therefore, individuals with AS require efficient treatment options.

Surgical aortic valve replacement (SAVR) is the present treatment protocol for severe symptomatic AS⁴. However, it has been demonstrated that individuals with calcified aortas, scars from prior cardiac surgeries, or advanced age are at a higher risk for SAVR. This increased risk in the patients is due to inherent challenges in surgery, such as the need to address adhesions, extended procedure times, and complications related to clamping, rather than indicating a decrease in the effectiveness of valve replacement. Hence, there have been significant advancements in therapy alternatives. Many studies^5–7 have shown transcatheter aortic valve replacement (TAVR) as a novel alternative strategy for inoperable or high surgical-risk AS patients due to its less invasive nature^8,9. However, it is unclear if these findings apply to AS patients with low to intermediate surgical risk too. Two recent trials DEDICATE (Randomized, Multicenter, Event-Driven Trial of TAVR vs. SAVR in Patients with Symptomatic Severe Aortic-Valve Stenosis) and NOTION-2 (Nordic Aortic Valve Intervention), have shown conflicting outcomes among patients with low or intermediate surgical risk undergoing TAVR versus SAVR^10,11. Although the DEDICATE trial showed a significant reduction in all-cause death with TAVR at 1 year of follow-up¹⁰, the NOTION-2 trial showed a comparable rate of all-cause death and the composite endpoint (all-cause death, stroke, and rehospitalization) at 1 year of follow-up with TAVR versus SAVR¹¹. These conflicted findings warrant the need for a meta-analysis to comprehensively synthesize the available evidence regarding not only the comparative efficacy but also the long-term outcomes of TAVR versus SAVR in patients with severe AS at low or intermediate surgical risk. While short-term outcomes have been evaluated extensively, there is a lack of consolidated data addressing the differences between the two groups over extended follow-up periods. Moreover, a comparison between the two treatment options at various time intervals has yet to be conducted. To bridge this literature gap and to determine the most optimal course of treatment in this patient subset, we conducted a meta-analysis comparing the two treatment options.

Methods

This systematic review and meta-analysis has been reported in concordance with guidelines provided by preferred reporting items for systematic review and meta-analyses (PRISMA, Supplemental Digital Content 1, http://links.lww.com/JS9/D564)¹². Approval from the institutional review board was not required since the included studies were publicly available with patient information de-identified.

Data sources and search strategy

An electronic search of Cochrane Library, PubMed/MEDLINE, and Embase was conducted for relevant randomized controlled trials (RCTs) and their follow-up reports, that aimed to assess the safety and efficacy of SAVR compared with TAVR in patients with low surgical risk from their inception through 10th May 2024, without any time or language restrictions. Medical subject headings (MESH) terms along with Boolean operators were used to devise an effective search strategy for each database (Table S1, Supplemental Digital Content 1, http://links.lww.com/JS9/D564). We also performed a manual search for additional relevant studies using references of previous systematic reviews and meta-analyses to ensure no important publications were overlooked.

The following predefined inclusion criteria were used: (i) RCTs; (ii) short (<1 year), intermediate (2–4 years), or long-term (≥5 years) follow-up reports of RCTs; (iii) RCTs that included adult male or female patients who were >18 years old; (iv) patients with severe AS, classified as having a low or intermediate surgical risk; (v) compared clinical outcomes of SAVR with TAVR; and (vi) assessed at least one of the predetermined clinical outcomes. Low surgical risk was defined as a mean STS score of <4% and/or a logistic EuroSCORE of <10%. Intermediate surgical risk was defined as a mean STS score of 4–8% and/or a logistic EuroSCORE of 10–20%. In the case of trials reporting outcomes at two different intervals for the same follow-up strata (e.g. for long-term follow-up, if outcomes were separately reported for 5 and 6-year follow-up, the data was extracted and pooled for longer follow-up). Post-hoc analyses of parent RCTs were excluded. The detailed inclusion and exclusion criteria of our meta-analysis are given in Table S2 (Supplemental Digital Content 1, http://links.lww.com/JS9/D564).

Study selection

All articles retrieved from the systematic search were exported to EndNote Reference Manager (Version X7.5; Clarivate Analytics, Philadelphia, Pennsylvania, 2016) where duplicates were sought and removed. The remaining articles were then assessed at the title and abstract level by two independent investigators (A.A. and F.A.), after which full texts were read to confirm relevance. Any disagreements were resolved by mutual discussion with a third investigator (M.A.).

Data extraction, outcomes, and quality assessment

Two investigators (A.A. and Z.A.N.) autonomously extracted data from the selected studies on prespecified collection forms. Data were extracted, including trial name, publication year, number of patients in TAVR and SAVR groups, mean age of patients, percentage of males among the study participants, details of the patient population, STS score, intervention subtype, duration of follow-up, and outcome measures.

The primary outcomes included all-cause death and cardiac death. The secondary outcomes included; myocardial infarction (MI), transient ischemic attack (TIA), stroke, valve endocarditis, prosthetic-valve thrombosis, permanent pacemaker implantation (PPI), acute kidney injury, new-onset or worsening atrial fibrillation (AF), which was defined as an increase in the severity, frequency, or duration of AF episodes, new-onset left bundle-branch block, major vascular complications, major bleeding, hospitalization, coronary obstruction, cardiogenic shock and aortic-valve reintervention. All outcomes except periprocedural were assessed on short (<1 year), intermediate (2–5 years), and long-term (≥5 years) follow-ups. A similar methodology has been used in prior studies^13–15.

To assess the quality of included RCTs, we used the Cochrane risk of bias tool (RoB 2.0)¹⁶. The risk of bias was assessed across the following domains: randomization, deviations from intended variation, missing outcome data, measurement of outcome, and selection of reported results. The trials were scored based on a high, some concerns, or low risk of bias in each domain. Traffic light plots were created using the Robvis tool¹⁷.

Statistical analysis

Statistical analysis was performed by pooling RR with corresponding 95% CIs for each clinical outcome. Data was pooled using the random-effects model in the R version 4.4.1. Forest plots were created to visualize the results of pooling. Heterogeneity across studies was evaluated using Higgins I ², and a value of I ²=25–50% was considered mild, 50–75% as moderate, and I ²>75% as severe¹⁸. The total number of included studies was less than 10; hence, the publication bias could not be assessed using funnel plots. A P-value <0.05 was considered statistically significant in all cases.

Results

A systematic literature search of databases/registers identified 3350 records. The duplicate studies were removed and 2285 articles were screened based on their titles and abstracts by two independent investigators. This process identified 2175 articles that were excluded and full texts of 110 articles were retrieved for screening. The review of full texts determined 17 reports that met our prespecified inclusion criteria. The PRISMA flowchart Figure S1 (Supplemental Digital Content 1, http://links.lww.com/JS9/D564) summarizes the screening and study selection process.

The meta-analysis included nine parent RCTs^{10,11,19–25} and eight follow-up reports^26–33 reporting data at 30 days, 1 year, 2 years, and 5 years or longer follow-up. The included trials reported data for 9092 patients with AS at low to intermediate surgical risk. Four thousand six hundred twenty-one patients underwent TAVR while SAVR was performed in 4471 patients. The mean age of patients was 77.43 years (SD: ±3.48) and 77.51 (SD: ±3.57) in the TAVR and SAVR groups, respectively. Male patients constituted about 57.25% of the study sample. Four RCTs randomized low-risk patients^11,20,22,25, two trials included low-to-intermediate risk patients^10,19, and three reported data for intermediate-risk patients^21,23,24. Although the NOTION trial¹⁹ included both low and intermediate-risk patients, over 80% of the randomized patients had low surgical risk based on STS score. The details of STS score and baseline characteristics are provided in Table 1 and Table S3 (Supplemental Digital Content 1, http://links.lww.com/JS9/D564). Of the nine included trials, seven were at high-risk of bias, and two were judged to have some concerns. The source of this bias was mainly due to deviations from intended interventions in all studies as it was impossible to blind the treatment approach, its influence on the medical staff was unavoidable. The details of bias assessment are provided in Figures S2 and S3 (Supplemental Digital Content 1, http://links.lww.com/JS9/D564).

Table 1.

Baseline characteristics of the included studies and reports.

					Sample size		Age (mean±SD)		Males (%)		Hypertension (%)		Diabetes (%)		STS score		Previous MI (%)
Trial	Publication year	Follow-up	Valve for TAVR	AVA (cm2)	TAVR	SAVR	TAVR	SAVR	TAVR	SAVR	TAVR	SAVR	TAVR	SAVR	TAVR	SAVR	TAVR	SAVR
NOTION	2015	1 year	Medtronic CoreValve System(TM)	–	145	135	79.2±4.9	79.0±4.7	54	53	71	76	18	21	2.9	3.1	6	4
NOTION	2016	2 years	Medtronic CoreValve System(TM)	–	123	113	79.1±4.8	79.0±4.7	54	53	71	76	18	21	2.9	3.1	6	4
NOTION	2024	10 years	Medtronic CoreValve System(TM)	–	139	135	79.2±4.9	79.0±4.7	53	53	72	78	17	21	2.9	2.9	6	4
NOTION 2	2024	1 year	–	0.7/0.7	187	183	71.1±3.1	71.0±3.2	64	62	68	72	20	22	1.1	1.1a	7	3
Evolut Low Risk Trial	2019	1 year	3.6% CoreValve/74.1% Evolut R/22.3% Evolut PRO	0.8/0.8	725	678	74.1±5.8	73.6±5.9	64	66	85	83	31	31	1.9	1.9a	7	5
Evolut Low Risk Trial	2022	2 years	3.6% CoreValve/74.1% Evolut R/22.3% Evolut PRO	0.8/0.8	730	684	74.1±5.8	73.7±5.9	64	66	85	82	31	31	2	1.9a	7	5
SURTAVI trial	2017	2 years	84% CoreValve bioprosthesis/16% Evolut R bioprosthesis	–	864	796	79.9±6.2	79.7±6.1	58	55	93	90	34	35	4.4	4.5	15	14
SURTAVI trial	2020	2 years	84% CoreValve bioprosthesis/16% Evolut R bioprosthesis	–	864	796	79.9±6.2	79.7±6.1	58	55	–	–	34	35	4.4	4.5	–	–
SURTAVI trial	2022	5 years	84% CoreValve bioprosthesis/16% Evolut R bioprosthesis	–	503	426	79.9±6.2	79.7±6.1	54	53	–	–	36	33	4.2	4.4	13	15
STACCATO trial	2012	30 days	Edwards Lifesciences SAPIEN	0.7/0.7	34	36	80±3.6	82±4.4	27	33	–	–	3	8	3.1	3.4	–	–
PARTNER 2	2016	2 years	Edwards Lifesciences SAPIEN XT heart-valve system	0.7/0.7	1011	1021	81.5±6.7	81.7±6.7	54	55	–	–	38	34	5.8	5.8	18	18
PARTNER 2	2020	5 years	Edwards Lifesciences SAPIEN XT heart-valve system	0.7/0.7	1011	1021	81.5±6.7	81.7±6.7	54	55	–	–	38	34	5.8	5.8	18	18
UK TAVI trial	2022	1 year	Edwards Lifesciences, Medtronic, St Jude Medical/ Abbott	0.7/0.7	458	455	81 (78–84)	81( 78–84)b	54	53	72	72	23	25	2.6	2.7	8	7
PARTNER 3	2019	1 year	SAPIEN 3 system	0.8/0.8	496	454	73.3±5.8	73.6±6.1	68	71	–	–	31	30	1.9	1.9	6	6
PARTNER 3	2021	2 years	SAPIEN 3 system	0.8/0.8	491	426	73.3±5.8	73.6±6.1	68	71	–	–	31	30	1.9	1.9	6	6
PARTNER 3	2023	5 years	SAPIEN 3 system	0.8/0.8	469	401	73.3±5.83	73.6±6.1	68	71	–	–	31	30	1.9	1.9	6	6
DEDICATE-DZHK6	2024	1 year	–	–	701	713	74.3±4.6	74.6±4.2	56	57	85	87	34	33	1.8	1.9a	5	8

^aSTS-PROM score was used.

^bData are reported as median and interquartile range.

AVA, aortic valve area; MI, myocardial infarction; TAVR, transcatheter aortic valve replacement; SAVR, surgical aortic valve replacement; STS, Society of Thoracic Surgeons; STS-PROM, Society of Thoracic Surgeons–Procedural Risk of Mortality.

Results of the meta-analysis

Primary outcomes

All-cause death: No significant association was observed of all-cause death with TAVR or SAVR at 30 days (RR: 0.94, 95% CI: 0.61–1.45, I ²=15%, seven studies), at 1 year (RR: 0.78, 95% CI: 0.61–1.01, I ²=42%, eight studies), and at 2 years (RR: 0.97, 95% CI: 0.84–1.12, I ²=0%, five studies). However, beyond 5 years of follow-up, TAVR was associated with a significantly increased risk of all-cause death (RR: 1.13, 95% CI: 1.03–1.23, I ²=15%, four studies, Fig. 1).

Figure 1.

Forest plot showing pooled effect size for all-cause death comparing transcatheter aortic valve replacement versus surgical aortic valve implantation.

Cardiac death: No significant association was observed of cardiac death with TAVR or SAVR at 30 days (RR: 0.95, 95% CI: 0.66–1.37, I ²=4%, six studies), at 2 years (RR: 0.90, 95% CI: 0.75–1.08, I ²=0%, five studies), or beyond 5 years (RR: 1.07, 95% CI: 0.96–1.20, I ²=0%, four studies). However, at 1 year of follow-up, TAVR was associated with a significantly decreased risk of cardiac death (RR: 0.79, 95% CI: 0.64–0.96, I ²=0%, 8 studies, Fig. 2).

Figure 2.

Forest plot showing pooled effect size for cardiac death comparing transcatheter aortic valve replacement versus surgical aortic valve implantation.

Secondary outcomes

Major bleeding: TAVR was associated with significantly reduced risk of major bleeding at 30 days (RR: 0.38, 95% CI: 0.21–0.67, I ²=95%, seven studies, Figure S4, Supplemental Digital Content 1, http://links.lww.com/JS9/D564). Reporting beyond 30 days was omitted as major bleeding is a periprocedural event. Any increased risk observed after SAVR at 1, 2, or 5 years primarily reflects higher bleeding risk within the first 30 days.

Myocardial Infarction: No significant association was observed of myocardial infarction with TAVR or SAVR at 30 days (RR: 0.72, 95% CI: 0.47–1.10, I ²=0%, six studies), at 1 year (RR: 0.85, 95% CI: 0.63–1.14, I ²=0%, eight studies), at 2 years (RR: 0.99, 95% CI: 0.74–1.33, I ²=0%, five studies), or beyond 5 years (RR: 1.14, 95% CI: 0.74–1.76, I ²=49%, four studies, Figure S5, Supplemental Digital Content 1, http://links.lww.com/JS9/D564).

Cardiogenic shock: TAVR was associated with a significantly reduced risk of cardiogenic shock at 30 days (RR: 0.34, 95% CI: 0.19–0.59, I ²=0%, two studies, Figure S6, Supplemental Digital Content 1, http://links.lww.com/JS9/D564). This outcome was not reported at any other time-points.

Major vascular complications: TAVR was associated with a significantly increased risk of major vascular complications at 30 days (RR: 2.37, 95% CI: 1.38–4.04, I ²=75%, six studies, Figure S7, Supplemental Digital Content 1, http://links.lww.com/JS9/D564). This outcome was not reported beyond 30 days of follow-up as it is a periprocedural complication and any increased risk after SAVR mainly reflects the complications within the initial 30 days.

Acute kidney injury: TAVR was associated with a significantly reduced risk of acute kidney injury stage II or III at 30 days (RR: 0.38, 95% CI: 0.26–0.54, I ²=0%, six studies) and at 1 year (RR: 0.58, 95% CI: 0.41–0.82, I ²=0%, four studies, Figure S8, Supplemental Digital Content 1, http://links.lww.com/JS9/D564). As only one trial reported this outcome for 2 years of follow-up and no trials for 5 years, a meta-analysis could not be performed at these time points.

Stroke: No significant association was observed of stroke with TAVR or SAVR at 30 days (RR: 0.79, 95% CI: 0.58–1.07, I ²=22%, seven studies), at 1 year (RR: 0.93, 95% CI: 0.63–1.38, I ²=53%, eight studies), at 2 years (RR: 0.90, 95% CI: 0.75–1.09, I ²=0%, five studies), or beyond 5 years (RR: 0.93, 95% CI: 0.72–1.22, I ²=53%, four studies, Figure S9, Supplemental Digital Content 1, http://links.lww.com/JS9/D564).

Transient ischemic attack: No significant association was observed of TIA with TAVR or SAVR at 30 days (RR: 1.34, 95% CI: 0.75–2.37, I ²=0%, six studies), at 1 year (RR: 1.40, 95% CI: 1.00–1.96, I ²=0%, seven studies), or beyond 5 years (RR: 1.33, 95% CI: 1.00–1.77, I ²=0%, three studies, Figure S10, Supplemental Digital Content 1, http://links.lww.com/JS9/D564). However, at 2 years of follow-up, TAVR was associated with a significantly increased risk of TIA (RR: 1.43, 95% CI: 1.02–2.00, I ²=0%, four studies) compared to SAVR.

Valve endocarditis: No significant association was observed of valve endocarditis with TAVR or SAVR at 30 days (RR: 0.93, 95% CI: 0.16–5.35, I ²=0%, three studies), at 1 year (RR: 0.95, 95% CI: 0.53–1.70, I ²=0%, seven studies), at 2 years (RR: 0.45, 95% CI: 0.12–1.66, I ²=67%, four studies), or beyond 5 years (RR: 0.93, 95% CI: 0.47–1.84, I ²=57%, three studies, Figure S11, Supplemental Digital Content 1, http://links.lww.com/JS9/D564).

Permanent pacemaker implantation: TAVR was associated with increased risk of PPI at 30 days (RR: 2.62, 95% CI: 1.40–4.91, I ²=86%, seven studies), at 1 year (RR: 2.19, 95% CI: 1.24–3.87, I ²=82%, seven studies), at 2 years (RR: 2.74, 95% CI: 1.31–5.71, I ²=93%, five studies), and beyond 5 years (RR: 1.95, 95% CI: 1.20–3.15, I ²=91%, four studies, Figure S12, Supplemental Digital Content 1, http://links.lww.com/JS9/D564).

New onset or worsening AF: TAVR was associated with decreased risk of new onset or worsening AF at 30 days (RR: 0.25, 95% CI: 0.18–0.34, I ²=78%, five studies), at 1 year (RR: 0.26, 95% CI: 0.16–0.41, I ²=84%, six studies), at 2 years (RR: 0.32, 95% CI: 0.20–0.49, I ²=84%, three studies), and beyond 5 years (RR: 0.49, 95% CI: 0.32–0.75, I ²=90%, three studies, Figure S13, Supplemental Digital Content 1, http://links.lww.com/JS9/D564).

Hospitalization: No significant association was observed of hospitalization with TAVR or SAVR at 30 days (RR: 0.73, 95% CI: 0.52–1.03, I ²=48%, four studies), at 1 year (RR: 0.86, 95% CI: 0.67–1.10, I ²=57%, six studies), at 2 years (RR: 0.96, 95% CI: 0.68–1.36, I ²=79%, four studies), or beyond 5 years (RR: 1.13, 95% CI: 0.83–1.52, I ²=77%, three studies, Figure S14, Supplemental Digital Content 1, http://links.lww.com/JS9/D564).

New-onset left bundle-branch block: TAVR was associated with increased risk of new-onset left bundle-branch block 1 year (RR: 2.31, 95% CI: 1.48–3.62, I ²=80%, two studies, Figure S15, Supplemental Digital Content 1, http://links.lww.com/JS9/D564). Only one trial each reported this outcome at 30 days and at 2 years, and no trials reported it at 5 years.

Prosthetic valve thrombosis: No significant association was observed of prosthetic valve thrombosis with TAVR or SAVR at 30 days (RR: 1.48, 95% CI: 0.18–12.04, I ²=0%, two studies), at 1 year (RR: 2.28, 95% CI: 0.78–6.65, I ²=0%, four studies), or at 5 years (RR: 4.10, 95% CI: 0.72–23.36, I ²=43%, two studies). However, at 2 years of follow-up, TAVR was associated with a significantly increased risk of prosthetic valve thrombosis (RR: 2.70, 95% CI: 1.08–6.71, I ²=0%, three studies, Figure S16, Supplemental Digital Content 1, http://links.lww.com/JS9/D564).

Aortic-valve reintervention: No significant association was observed of aortic-valve reintervention with TAVR or SAVR at 30 days (RR: 2.56, 95% CI: 1.00–6.53, I ²=0%, four studies), at 2 years (RR: 1.62, 95% CI: 0.68–3.82, I ²=62%, four studies), or at 5 years (RR: 1.91, 95% CI: 0.89–4.11, I ²=63%, three studies). However, at 1 year of follow-up, TAVR was associated with a significantly increased risk of aortic-valve reintervention (RR: 1.98, 95% CI: 1.21–3.24, I ²=0%, seven studies, Figure S17, Supplemental Digital Content 1, http://links.lww.com/JS9/D564).

Coronary obstruction: No significant association was observed of coronary obstruction with TAVR or SAVR at 30 days (RR: 1.08, 95% CI: 0.42–2.80, I ²=18%, four studies, Figure S18, Supplemental Digital Content 1, http://links.lww.com/JS9/D564). Reporting beyond 30 days was omitted as coronary obstruction is a periprocedural complication and any increased risk after SAVR mainly reflects the complications within the initial 30 days.

Discussion

In this systematic review and meta-analysis of 17 reports and 9092 individuals, we sought to compare TAVR with SAVR in patients with AS who are at low to intermediate surgical risk. This comparison is significant because, while TAVR has become increasingly popular in high-risk patients due to its minimally invasive nature, its role in patients with low to intermediate surgical risk remains debated, with concerns about long-term efficacy, safety, and durability compared to SAVR. Our pooled analysis demonstrated no statistically significant difference between TAVR and SAVR for reducing all-cause death at 30 days, 1 year, and 2 years but significantly increased risk with TAVR at 5 years or longer follow-up. TAVR was associated with a significantly decreased risk for cardiac death at 1-year follow-up and comparable risk at 30 days, 2 years, and 5 years or longer follow-up. TAVR showed a significantly lower risk of major bleeding events at 30 days; however, it was associated with a significantly increased risk of major vascular complications at 30 days and PPI at 30 days, 1 year, and at 2 years; lower risk of AKI at 30 days and 1 year and lower risk of new onset or worsening AF at 30 days, 1 year, and 2 years when compared with SAVR. TAVR was also associated with a significantly reduced risk of cardiogenic shock at 30 days. No statistically significant difference was observed between TAVR and SAVR in reducing the risk of MI, stroke, valve endocarditis, risk of hospitalizations, and TIAs. However, TAVR was associated with a significantly increased risk of TIA at 2 years, and an increased risk of new-onset left bundle-branch block at 1 year. TAVR was also associated with a significantly increased risk of prosthetic valve thrombosis at 2 years and aortic-valve reintervention at 1 year follow-up.

Although previous meta-analyses have been conducted on this topic^34,35, our study offers new insights. We pooled clinical outcomes at short to long-term follow-ups, providing a better insight regarding the clinical effectiveness of TAVR versus SAVR. Our analysis revealed no statistically significant difference between TAVR and SAVR in reducing all-cause mortality at 30 days, 1 year, and 2 years. This result is consistent with several other studies^34,35 that showed TAVR to be noninferior to SAVR. However, after 5 years or longer follow-up, our analysis found a significantly elevated risk linked to TAVR. Several factors should be considered while interpreting the increased risk of all-cause death with TAVR at 5 years or longer follow-up. NOTION trial investigators reported the longest follow-up data in our pooled analysis for TAVR versus SAVR²⁷. In this study, all consecutive patients were eligible for enrollment in the trial without determining a specific risk profile. As a result, the investigators included a large number of older patients with a mean age of 79 years, although they were predominantly low-risk individuals. Over 80% of the included patients had an STS-PROM score of less than 4% and limited comorbidities. An observational study reported all-cause death rates of 7.1% at 5 years of follow-up and 12.4% at 8 years following isolated SAVR³⁶. In contrast, the NOTION trial had all-cause death rates of 28.9 and 52.6% at 5 and 8 years of follow-up^37,38. After 10 years of follow-up, around 60% of the patients had died, which indicated a higher risk profile of included patients than that determined solely based on STS-PROM score²⁷. In the PARTNER 3 trial, which included younger patients (mean age 73 years) at low surgical risk, the 5-year all-cause mortality rates were 10.0% for TAVR and 8.2% for SAVR³⁹. In the PARTNER 2 trial, which included patients at intermediate surgical risk with a mean age of 81.6 years, all-cause death rates at 5 years were 46.0% in the TAVR group and 42.1% in the SAVR group³¹. It is important to mention that SAPIEN XT was employed in the PARTNER 2 trial which is no longer in clinical use in the USA and has been largely replaced by more advanced generations like the SAPIEN 3 valve. While the second-generation valve, SAPIEN XT, demonstrated procedural success, it was associated with higher rates of paravalvular leakage and aortic regurgitation, which has been linked to poorer long-term outcomes, including increased mortality⁴⁰. The newer valves have significantly improved upon these outcomes^41,42.

TAVR was linked to a notably decreased risk of cardiac mortality at 1 year, indicating a possible early benefit concerning outcomes specific to the heart. Nevertheless, this advantage was not sustained over a year, as TAVR and SAVR had similar risks of cardiac mortality after 30 days, 2 years, and 5 years or more. The reduction in cardiac death at 1 year may reflect a statistical anomaly rather than a true clinical effect, given the difference was not consistent across other time points of 30 days, 2 years, or 5 years. Our findings highlight the need for long-term follow-up studies for TAVR to compare its durability with SAVR, especially as long-term data for SAVR is widely available. A significant challenge in evaluating TAVR outcomes is the regular introduction of new TAVR valve generations. Although long-term data on TAVR exists, they are based on previous generations of valves and are therefore considered outdated due to the advancement in valve designs.

Compared to SAVR, TAVR showed a significantly decreased risk of major bleeding episodes at 30 days. This is likely because TAVR is a less invasive procedure that induces less surgical stress and requires minimal anticoagulant medication⁴³. Additionally, at 30 days, 1 year, and 2 years, TAVR was linked to a decreased risk of AKI as well as a lower risk of new-onset or worsening AF. The potential correlation between the lower risk of AKI and TAVR might be attributed to two factors: the TAVR procedure’s less invasive nature and its ability to prevent cardiopulmonary bypass⁴⁴. The association of TAVR with a lower risk of new-onset or worsening AF is likely due to its less invasive nature, which reduces surgical trauma, oxidative stress, inflammation, and the overall hemodynamic burden of surgery⁴⁵. These results demonstrate the procedural benefits of TAVR, especially concerning lowering perioperative and early postoperative complications. The reduced incidence of cardiogenic shock with TAVR could be due to the procedure’s ability to minimize cardiac stress and preserve hemodynamic stability in the immediate postoperative period attributed to the less intrusive technique⁴⁶. Conversely, TAVR was associated with an increased risk of major vascular complications. This can be linked to the transcatheter procedure, which involves navigating the vascular system, thereby increasing the likelihood of vascular injuries⁴⁷. The transcatheter valve’s location, which may interfere with the heart’s electrical conduction system⁴⁸, might explain the higher risk of permanent pacemaker placement with TAVR observed in our study^49,50. Moreover, the preexisting right bundle branch block and new-onset left bundle branch block, coupled with the older age of patients, are key factors contributing to the increased risk of pacemaker implantation after TAVR⁵¹. The risk of MI, stroke, hospitalizations, and TIAs was comparable between TAVR and SAVR across all time points.

Our analysis demonstrated no significant differences between the two groups in the risk of prosthetic valve thrombosis at 30 days, 1 year, or 5 years. However, at 2 years, TAVR was associated with a significantly increased risk of prosthetic valve thrombosis. Thrombosis was typically diagnosed through imaging and echocardiographic findings, and although most trials reported the use of antithrombotic or anticoagulant therapy, details regarding whether it led to reintervention were not consistently reported. Similarly, no significant differences were observed in aortic-valve reintervention rates at 30 days, 2 years, or 5 years, but TAVR showed a significantly increased risk at 1 year. Further details regarding reintervention, including whether TAVR explanation was required, were not provided. Several factors, such as valve degeneration or patient baseline characteristics could explain these findings, but further investigation is needed.

This study has some limitations. Firstly, this is a study-level meta-analysis, and the absence of patient-level information made it challenging to evaluate the impact of important moderators, such as the percentage of bicuspid valves. This limits our ability to account for variations in patient characteristics that may influence the outcomes. The effect of AKI on mortality could not be evaluated as well due to limited available data. Secondly, significant heterogeneity was observed in some of the outcomes, which could be attributed to deviations from intended interventions in pooled studies along with a considerable variability in the patients’ characteristics. The variations in the use of valve types, baseline STS scores, and age of patients could have influenced our findings. These variations may affect the generalizability of our findings to broader patient populations. The prognosis of patients with AS who are at low to intermediate surgical risk may be impacted by the lack of information on treatment regimens following TAVR or SAVR in the included trials, further limiting the applicability of our results to clinical practice. Moreover, it is important to note key criticisms of several trials comparing TAVR to SAVR. Trials comparing these patients have been criticized for including concomitant procedures in the surgical arm (such as CABG or other valve interventions), reoperations, and various valve replacement options for patients in the SAVR group, all of which increase the surgical risk and heterogeneity of the patients, compared to TAVR patients which are not impacted by these factors and receive one or two of the most advanced valve options. These confounding variables have made it very difficult to determine the differences in outcomes between patients and likely have contributed to noninferiority findings in some trials as well as conflicting findings in others.

To further validate the effectiveness of TAVR compared to SAVR, future research should focus on long-term, large-scale RCTs with extended follow-up periods. Such studies should include patient-level data to account for individual variations. Moreover, ensuring more homogeneity in study populations and using standardized outcome measures will enhance the reliability of future findings. Detailed reporting of postprocedural management and treatment regimens will also be essential to gaining a better understanding of their impact on outcomes, allowing for more accurate comparisons between TAVR and SAVR. Particularly, for younger patients with a longer life expectancy, studies should address concerns about the long-term durability of TAVR valves and the elevated risks observed at extended follow-ups. The durability of TAVR valves remains an important factor in determining the optimal treatment approach for these patients, as the higher mortality and complications associated with TAVR at longer follow-up suggest SAVR may provide superior long-term advantages in this population. Understanding the economic viability of TAVR compared to SAVR will also be important as TAVR evolves and becomes more widely adopted.

Conclusion

Our meta-analysis concludes that, for patients with aortic stenosis at low to intermediate surgical risk, while TAVR offers significant short-term advantages over SAVR, such as reduced bleeding events, lower risks of AKI, and new-onset or worsening AF, these advantages are outweighed by its poorer long-term outcomes. Both TAVR and SAVR demonstrated comparable rates of all-cause death up to 2 years of follow-up. However, long-term outcomes reveal important drawbacks, including higher mortality with TAVR at extended follow-up periods. TAVR was also associated with significantly increased PPI and major vascular complications. Unlike previous studies, our analysis provides a comprehensive comparison of TAVR and SAVR across a wide range of time-points, highlighting these long-term differences and complications. Given these findings, SAVR can be considered a reliable treatment option for younger people with low to intermediate surgical risk, due to its more favorable medium to long-term outcomes. Future research should focus on improving the durability of TAVR valves and minimizing procedure-specific complications to enhance its long-term efficacy.

Ethical approval

Ethical approval of the study was not required as the data is publicly available and no intervention was performed on patients.

Consent

No consent was needed.

Source of funding

No financial support was received for the study.

Author contribution

M.A.: conceptualization, data curation, and project administration; F.S., R.Y., and R.A.: supervision; A.A. and Z.A.N.: formal analysis of data; A.A., F.A., and H.J.: formal analysis, methodology, and software; A.S., M.A., F.A., and Z.A.N.: writing the original draft; A.S., R.Y., F.S., H.J., E.Z., and R.A.: writing, reviewing, and editing; M.A. and R.A.: visualization and validation.

Conflicts of interest disclosure

The authors declare no conflicts of interest.

Research registration unique identifying number (UIN)

The protocol of review is registered with PROSPERO: CRD42024566228.

Guarantor

Mushood Ahmed.

Data availability statement

Data sets generated during and/or analyzed during the current study are available upon reasonable request to corresponding author.

Provenance and peer review

Not commissioned, externally peer-reviewed.