Objective Performing transfemoral transcatheter aortic valve implantation (TAVI) without general anaesthesia (GA) has been increasingly adopted. We sought to study the impact of GA and non-GA approaches on procedural outcome and 30-day and 1-year mortality in transfemoral TAVI.
Methods The UK TAVI registry holds information for every TAVI procedure in the UK. We analysed the data for patients implanted during 2013–2014 using either an Edwards Sapien or a Medtronic CoreValve prosthesis. Propensity score-matching analysis was performed to adjust for confounding factors.
Results 2243 patients were studied (aged 81.4±7.5 years, 1195 males). 1816 (81%) underwent TAVI with GA and 427 (19%) without GA. Transoesophageal echocardiography (TOE) was used in 92.3% of GA and 12.4% of non-GA cases (p<0.001). There was no significant difference in the rate of successful valve deployment (GA 97.2% vs non-GA 95.7%, p=0.104) and in the incidence of more than mild aortic regurgitation (AR) at the end of the procedure (GA 5.6% vs non-GA 7.0%, p=0.295). However, procedure time was longer (131±60 vs 121±60mins, p=0.002) and length of stay was greater (8.0±13.5 vs 5.7±5.5 days, p<0.001) for GA cases. 30-day and 1-year mortality rates did not differ between the GA and non-GA cases. After propensity matching, these results remained unchanged. A second propensity analysis (adjusted for mode of anaesthesia) did not show an association between use of TOE and rate of successful valve deployment or frequency of significant AR. Neither was TOE associated with a longer procedural time or greater length of stay.
Conclusion Procedure outcome, and 30-day and 1-year mortality are not influenced by mode of anaesthesia. However, GA is associated with longer procedure duration and greater length of stay.
- aortic stenosis
- transcatheter valve interventions
Statistics from Altmetric.com
The advent of transcatheter aortic valve implantation (TAVI) has revolutionised the field of aortic stenosis.1 TAVI has become the standard treatment for higher risk and inoperable patients with severe aortic stenosis.1 The results from the Placement of Aortic Transcatheter Valve 2 trial that included patients at intermediate risk for surgical aortic valve implantation, promises even further growth in the number of TAVI procedures in upcoming years.2 Initially, the size of delivery catheters was such that a surgical approach needing a general anaesthesia (GA) was the preferred option during TAVI. However, technological advances in catheters and devices, and the increased operator experience, have led to a strong trend towards performing TAVI under conscious sedation. The choice of GA versus no GA in TAVI varies considerably among centres, and the impact of mode of anaesthesia on the outcomes is uncertain. While a randomised trial in this area remains extremely unlikely, data from large registries can address this area. In this cohort study, we sought to assess the impact of GA and non-GA approaches on procedural outcome and complication rates predischarge, and mortality at 30 days and 1 year.
TAVI has been performed in the UK since 2007. The UK TAVI registry is hosted by the National Institute of Cardiovascular Outcomes Research (NICOR) and records information for every TAVI case in the UK through a web-based interface as previously described.3 From the beginning of 2013, a new dataset that includes additional variables, such as frailty score and mobility status, has been in place. For this reason and given the rapid advances in the transcatheter therapy field, we elected to limit the study to the cohort of patients who underwent transfemoral TAVI during 2013–2014. Survival data for Scotland and Northern Ireland were not available, therefore only patients from England and Wales were included in this study, however procedures from these two countries comprised the majority of the TAVI cases. In order to minimise potential bias, only cases that were treated with Sapien XT and Sapien 3 (Edwards Lifesciences, Irvine, California, USA) and the CoreValve (Medtronic, Minneapolis, Minnesota, USA) were included.
The choice of GA versus non-GA was at the discretion of the local hospitals, after discussion in a multidisciplinary heart team meeting. Information regarding the reasons behind the choice of a specific method of anaesthesia was not available in the dataset. Canadian Study of Health and Aging frailty score >3 was used to define vulnerable and frail patients in this study. Grade III or IV calcification of the ascending aorta is defined as significant in the UK TAVI database. Procedure time was defined as the time from the entry to exit of the patient from the catheterisation laboratory. Procedural complications were adjudicated according to Valve Academic Research Consortium-2 (VARC-2) criteria. Aortic regurgitation (AR) at the end of procedure was assessed by transoesophageal echocardiography (TOE), transthoracic echocardiography (TTE) or aortic root angiography and graded as mild, moderate and severe. Cases that entered the study had less than 5% missing data that were not imputed and are shown in the online supplementary tables 1. Mortality tracking was available for 97.3% of study population. All patients provided written consent for the TAVI procedure.
Supplementary file 4
Categorical data were presented as percentages and numerical data as mean±SD. For comparison of two groups, χ2 was used for the categorical data, and independent sample T test for normally distributed data, and a Mann-Whitney U test for skewed data. Propensity score-matching analysis was used to adjust for the confounding factors. Significantly different baseline characteristics and important predictors of outcome identified from previous studies were used in the model. Using the neighbourhood method and a calliper of 0.01 (for each treatment case, the selected control was within 0.01 SD of the propensity score) each GA TAVI case was matched to a non-GA TAVI patient in a 1:1 fashion. The variables included in the model were age, estimated glomerular filtration rate (eGFR), body mass index (BMI), significant ascending aorta calcification, previous cardiac surgery, Canadian Study of Health and Aging frailty score, left ventricular function, New York Heart Failure Association (NYHA) class, procedure urgency, procedural imaging, delivery approach and valve type. Kaplan-Meier survival analysis was performed to assess the impact of mode of anaesthesia on mortality at 30 days and 1 year. A separate propensity score matching was performed according to the intraprocedural imaging, and all patients who underwent TAVI with intraprocedural TOE were matched in a 1:1 fashion to a control case without intraprocedural TOE using the neighbourhood method and a calliper of 0.01. The model included age, BMI, Canadian Study of Health and Aging frailty score, NYHA class, left ventricular function, previous percutaneous coronary intervention, ascending aorta calcification, procedure urgency, mode of anaesthesia, delivery approach and valve type. Statistical analysis was performed using SPSS software V.23.0 (Released 2015; IBM SPSS Statistics for Mac).
The rates of GA and non-GA transfemoral TAVIs from the beginning of the use of TAVI in all four countries of the UK are summarised in figure 1. The data span 2007 to the end of 2014 and show a statistically significant increase in the number of non-GA procedures in this period (p=0.001). The study population comprised 2243 patients treated by transfemoral TAVI using the Edwards Sapien and Medtronic CoreValve in England and Wales during the two calendar years 2013 and 2014 (figure 2).
Baseline characteristics of the study population
The procedure was performed under the GA in 1816 (81%) and non-GA in 427 (19%) patients (table 1). Non-GA TAVI patients were older (82.0±6.8 vs 81.2±6.7 years, p=0.043), frailer (65.2.0% vs 58.6%, p=0.013) and had higher rate of chronic kidney disease stage ≥3 (56.0 % vs 49.1%, p=0.011). A higher BMI (27.9±5.9 vs 27.1±5.0 Kg/m², p=0.014) and a higher number of patients with ejection fraction <50% (38.1% vs 31.6%, p=0.012) and previous cardiac surgery (29.7% vs 24.5%, p=0.037) were seen in those treated using GA.
Procedural characteristics and outcome
Edwards Sapien valves were more frequently used in the GA than non-GA patients (61.9% vs 56.2%, p=0.032) (table 2). Successful valve deployment was achieved in 97.2% and 95.7% of GA and non-GA cases, respectively (p=0.104). Incidence of more than mild AR at the end of the procedure was numerically higher in the non-GA TAVI group (7% vs 5.6% in the non-GA TAVI) but did not reach statistical significance (p=0.295). There was no difference in the rate of reported procedural complications (table 2), however procedure time (131±60 vs 121±60 min, p=0.002) and the average length of stay (LOS) in hospital (8±13.5 vs 5.7±5.5 days, p<0.001) were longer in the GA group.
Propensity score matching according to the mode of anaesthesia
Owing to differences in baseline characteristics between the GA and non-GA patients, a propensity score-matching analysis according to the mode of anaesthesia was performed resulting in two comparable groups of 153 patients. The baseline and procedural characteristics were similar between the two matched groups of TAVI with and without GA as shown in tables 3 and 4. While the numerical number of patients with more than mild AR at the end of the procedure remained higher in the non-GA group, a significant statistical difference was not reached. A longer procedure time and a greater average LOS in hospital in the GA TAVI were the only significant differences between the matched GA and non-GA groups.
Impact of the mode of anaesthesia on mortality
Propensity score matching according to intraprocedural TOE
While 92.3% of the GA TAVI patients had intraprocedural TOE, 7.7% did not. Intraprocedural TOE was performed in 12.4% of the non-GA cases. This allowed propensity score-matching analysis according to the use of intraprocedural TOE resulting in two comparable groups of 163 patients. Adjusted for the mode of anaesthesia, there was no significant difference in the rate of successful valve deployment (97% vs 97.5%, p≥0.999), and more than mild AR at the end of the procedure (6.1% vs 4.0%, p=0.201) or other procedural complications between the TOE and no TOE cases, respectively. Of note, the procedure time (128±66 vs 137±61 mins, p=0.174) and the average LOS in hospital (8.7±7.7 vs 7.1±6.2 days, p=0.137) were not different either (figure 5, online supplementary tables 2-5 and online supplementary figures 1 and 2).
Supplementary file 3
Our study shows no significant difference in procedural success, complication rate and mortality at 30 days and 1 year between TAVI performed with or without GA. However, TAVI with GA was associated with a longer procedure time and LOS in hospital.
The variation in the adoption of a non-GA approach in TAVI is notable for being large, ranging from almost 0% to 100% across Europe.4 5 The UK has been relatively slow to move to non-GA TAVI with only 35% of the transfemoral TAVI cases performed under non-GA in 2014. Likewise, GA remains the predominant approach to TAVI in the USA.6
Apart from the resolved issue of the size of the catheters, other rationale for performing TAVI under GA include the use of intraprocedural TOE to compensate for limitations of fluoroscopy, accurate assessment of AR at the end of the procedure and the detection of potential life-threatening complications. Balloon aortic valvuloplasty also carries a small but real risk of serious complications and may be tolerated better with GA.7 There is evidence from the early years that intraprocedural TOE was associated with a reduction in fluoroscopy time.8 Moreover, in a recent single-centre study comparing TTE and TOE-guided TAVI, a higher incidence of second valve implantation during the TTE TAVI procedures was observed although this was not independently associated with the rate of paravalvular AR at the time of discharge.9 In contrast, performing TAVI without GA allows neurological monitoring, avoids invasive ventilation and might be associated with lower labour costs. Small single-centre observational studies have also shown that performing TAVI under conscious sedation may be associated with better haemodynamic stability and less fluid use during the procedure.10–13
The focus of studies derived from multicentre and national registries, however, has been on the impact of GA and non-GA approaches on the overall outcomes. To date, there have been four reports concerning the mode of anaesthesia in TAVI, of which two included only Medtronic CoreValve14 15 and two a combination of Medtronic CoreValve and Edwards Sapien valves.4 16 Notably, these studies do not go beyond the end of 2012. While higher numbers of older and frailer patients and those with chronic kidney disease stage ≥3 in the non-GA TAVI may seem intuitive, a reason for a higher prevalence of previous cardiac surgery, left ventricular impairment and higher BMI in the GA TAVI group is not as obvious. Noteworthy is the fact that the distribution of these variables in the GA and non-GA patients, among the studies, appears discordant. Similar to our report, in Transcatheter Valve Treatment Sentinel (TCVT) registry, the patients in the non-GA group were older, and the rate of previous coronary artery bypass grafting (CABG) surgery was higher in the GA group, whereas in contrast to our study they observed no difference in the BMI, left ventricular ejection fraction and eGFR between the GA and non-GA groups.4 16 In keeping with our report, there were higher rate of patients with previous CABG and impaired left ventricular function in the GA group in ADVANCE registry.14 A higher logistic EuroSCORE was observed in the GA cases in the ADVANCE, Italian CoreValve and TCVT registries but not in the French Aortic National CoreValve and Edwards 2 (FRANCE 2) registry. These observed differences among the GA and non-GA TAVI patients across different reports may reflect the variability in real world practice, the presence of other confounders and may imply the lack of a consistent approach in choice of anaesthesia mode in TAVI.
In our study, Edwards Sapien valves were more frequently used with GA TAVI as opposed to CoreValve, an observation that was also present in the FRANCE 2 and TCVT registries.4 16 Parallel to this, we also observed a significantly higher rate of predeployment balloon valvuloplasty in the GA cases although this difference was not seen in the FRANCE 2 and TCVT registries and was no longer observed in our cohort after the propensity-matching analysis adjusted for the type of the valve. The need for rapid pacing for deployment of the balloon expanding Edward Sapien valves may be an explanation for the higher number of these valves in the GA group.
In contrast to previous reports from the registries, we did not observe any differences in the VARC-2 defined procedural complications rate between the GA and non-GA cases. Acute kidney injury stage III and bleeding, even after propensity score matching, was higher in the GA patients in the Italian CoreValve registry.15 The report from the TCVT registry also showed a higher rate of renal replacement therapy in the GA TAVI group,4 and there was a higher rate of vascular complications in the non-GA TAVI cohort in the ADVANCE registry.14
The longer procedural time and greater LOS in hospital associated with the GA and the similar cumulative all-cause mortality in the GA and non-GA cases are consistent with published single-centre studies and the reports from large registries. However, the average LOS in hospital in the ADVANCE and SOURCE 3 registries was not significantly different between the GA and non-GA groups.17
The effect of mode of anaesthesia on AR at the end of the procedure
The FRANCE 2 registry reported significantly higher incidence of more than mild AR in the non-GA TAVI patients16 as did the TCVT registry.4 A meta-analysis comprising 5919 patients also showed an association between the non-GA TAVI and paravalvular AR.18 One suggested explanation for the findings in the FRANCE 2 registry was the lack of intraprocedural TOE in the non-GA TAVI. Yet, the difference in AR was no longer apparent after the TOE-adjusted propensity score-matching analysis. We observed a numerically higher rate of more than mild AR in the non-GA TAVI group before and after propensity matching although this did not reach a statistical significance. The UK TAVI database had incomplete postdischarge data and as such, we could not assess the impact of the mode of anaesthesia on AR at follow-up.
Intraprocedural TOE in TAVI
In order to address the trend towards the higher rate of more than mild AR at the end of the procedure in the non-GA TAVI, a propensity score-matching analysis according to the use of intraprocedural TOE was performed. The Brazilian registry also reported a favourable impact on mortality associated with the use of intraprocedural TOE during the TAVI.19 Adjusting for the mode of anaesthesia, the rate of AR at the end of the procedure was not significantly different between those who underwent TAVI with and without intraprocedural TOE; neither was the success of the valve deployment, mortality at 30 days and 1 year. The comparable procedure time and the average LOS in hospital that was observed between the two matched groups of TAVI with and without intraprocedural TOE seem intuitive as performing TOE itself is not expected to lengthen the procedure or hospitalisation after TAVI. However, given that the absolute difference in the average LOS remained the same before and after propensity score-matching analysis, the reduced sample size may have resulted in statistically non-significant difference in the LOS between the two groups. Caution should be undertaken in generalising these results with regard to the use of intraprocedural TOE in TAVI given the small percentage of the GA TAVI cases without TOE and non-GA TAVI cases with TOE and difficulty in separating the effect of the TOE from GA. Nonetheless, in the era that technological advances allow the use of methods such as transnasal or intracardiac echocardiography without GA and considering other existing evidence in support of intraprocedural TOE in TAVI,8 16 19 further clinical studies are required to assess the impact of intraprocedural imaging on the TAVI outcomes.
This is a retrospective study with known limitations including potential confounding factors that might not have been accounted for. The choice of propensity score-matching analysis resulted in two comparable groups of patients, however it significantly reduced the study population, and caution should be taken in generalising the results. Although the data completeness was very high, the data were self-reported by each centre without external validation. However, the mortality tracking record in England and Wales is extremely reliable and provides an object measure of outcome. The rational beyond the choice of the mode of anaesthesia was not recorded in the database which might influence the results of the study. The method of assessment of AR was not specified in the database and the interpretation of the presence and severity of the AR was at the discretion of local centres.
Procedural success, complication rate and mortality in TAVI was not associated with the mode of anaesthesia in the UK TAVI cohort. A GA approach was, however, associated with a longer procedure time and a greater LOS in hospital. There appears to be a wide variation in real-world practice in patient selection for GA versus non-GA TAVI, therefore the choice should be tailored towards each individual patient after discussion in a heart team meeting.
What is already known on this subject?
Other studies have shown that transfemoral transcatheter aortic valve implantation (TAVI) can be successfully performed without general anaesthesia (GA). However, the impact of the mode of anaesthesia on procedural complication rate is inconsistent among published registries.
What might this study add?
This large study represents a more contemporary period reflecting the technological advances in the field and a broader case selection. It shows no difference in procedural success, complication rate, and 30-day and 1-year mortality between the GA and non-GA approaches, and a longer procedure time and greater length of stay in hospital associated with the use of GA in TAVI.
How might this impact on clinical practice?
This study adds to the body of literature showing no harm in real-world use of a non-GA approach in TAVI.
Supplementary file 1
Supplementary file 2
Contributors ME has participated in conception and design of the study, analysis and interpretation of the data, drafted the manuscript and approved the manuscript submitted. OA has participated in analysis and interpretation of the data, revising the draft critically for important intellectual content and has approved the manuscript submitted. RD, JB, PM, EA, PL, OW, DHS have participated in conception and design of the study, revising it critically for important intellectual content and approved the manuscript submitted. MM has participated in the conception and design of the study, analysis and interpretation of the data, revising the draft critically for important intellectual content; and has approved the manuscript submitted.
Funding The research posts of M. Eskandari and O. Aldalati have been funded by the King’s College Hospital Charity, London, United Kingdom.
Competing interests None declared.
Patient consent Detail has been removed from this case description/these case descriptions to ensure anonymity. The editors and reviewers have seen the detailed information available and are satisfied that the information backs up the case the authors are making.
Ethics approval The study was approved internally at King’s College Hospital,London, UK as well as by the Steering Committee of the National Institute for Cardiovascular Outcomes Research who granted the access to the UK TAVI database (un-identifiable data).
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.