Early mobilisation after transfemoral transcatheter aortic valve implantation: results of the MobiTAVI trial

We prospectively included all consecutive patients undergoing TF-TAVI from September 2016 until August 2018 at the Amsterdam University Medical Centre (Amsterdam UMC, location AMC), a high-volume tertiary centre in Amsterdam, the Netherlands. In patients with symptomatic aortic valve stenosis, decisions regarding treatment, access route and valve selection were at the discretion of our multidisciplinary TAVI team. These decisions were part of regular clinical care and based on pre-operative screening, including computed tomography angiography, cardiac echocardiography and diagnostic coronary catheterisation, all performed in accordance with the most recent guidelines [19, 20]. After the decision to perform TAVI using the transfemoral approach, patients were assigned randomly to two pre-defined weekdays at the discretion of our planning bureau, which had no insight into the expected difficulty of the procedure or the patients’ health status. The operators were assigned to the two pre-defined week days weeks before the patients were. The Institutional Review Board approved this study with a waiver, and the trial was registered in the Dutch Trial Register (NTR 6098).

The standard approach for TAVI was a fully percutaneous transfemoral approach using local anaesthesia. We followed regular hospital protocol regarding the pre-procedural administration of heparin and protamine, based on the weight of the patients and the measured activated clotting time. For vascular closure, the double-ProGlide preclose technique (Abbott Vascular, CA, USA) and the Manta closure device (Essential Medical, Exton, PA, USA) were used for valve introduction [21‐24]. The non-valve side was closed with either a single ProGlide or an Angio-Seal (Terumo Medical Corporation, NJ, USA). Afterwards, SafeGuards (Merit Medical, South Jordan, UT, USA) were placed on both groins; the devices were deflated after 2 h and removed after 4 h according to hospital protocol.

We developed a strict protocol to assess patient eligibility for early mobilisation and to guarantee patient safety. Patients could be excluded at three different time points during the hospital stay (Fig. 1; see Electronic Supplementary Material, Table S1, for complete checklist). The first time point, T1, was assessed before the procedure, whereas T2 was assessed during and directly after the procedure. After 4 h, following consultation with the operator and physical examination of the patient, T3 was assessed. After passing the three time points, the patient was deemed eligible for early mobilisation and was either allocated to the early mobilisation group (EARLY), i.e. ambulation within 4–6 h after the procedure, or to the regular hospital protocol (REGULAR), which consisted of supine bed rest until the next morning. Allocation was performed based on the time at which the procedure ended; all patients in whom the procedure was finished before 1,300 hours were allocated to the EARLY group, and all patients after 1,300 hours to the REGULAR group. The reason for choosing this design was twofold: (1) to increase clarity and feasibility for the medical staff and (2) to increase the safety of the patients in the EARLY group, who in this manner would ambulate during the fully staffed day shift.

Baseline characteristics including data from the pre-operative screening were prospectively collected in the AMC TAVI database. The primary endpoint of this trial was the safety of early ambulation, consisting of the presence of vascular (access site) complications and access site bleedings (according to the VARC‑2 criteria [22]). The secondary endpoint was the combined incidence of in-hospital outcomes. In-hospital outcomes included post-operative pain, scored with the Visual Analogue Scale (VAS, whereby post-operative pain was defined as VAS >3 [23, 24]), post-operative delirium (confirmed by a geriatric internist), clinically diagnosed infections (defined as the clinical suspicion with conclusive laboratory [increase in C‑reactive protein or leucocytes] or conclusive microbiology findings), and unplanned urinary catheter use (defined as urinary catheter use in patients who were hospitalised without a urinary catheter before TAVI). As a secondary safety endpoint, fall incidents were registered. Lastly, the duration of the hospital stay was evaluated and was defined as the number of days from the TF-TAVI to the day the patient was discharged to home.

The primary and the secondary endpoint were compared between the EARLY and REGULAR group. The secondary endpoint was analysed as a composite of the in-hospital complications (incidence of pain, infection, delirium and unplanned urinary catheter use). Moreover, all individual in-hospital outcomes were compared between the EARLY and REGULAR group. Categorical variables are presented as numbers with percentages and compared between both groups using Pearson’s chi-squared test. Continuous data were checked for normality, and are presented as mean with standard deviation or median with interquartile range and compared using an unpaired Student’s t-test or Mann-Whitney U‑test as appropriate. A double-sided p-value <0.05 was considered significant. All analyses were performed using SPSS software (version 24.0 for Windows, SPSS, Inc., Chicago, IL, USA). Since this was a first-time study and no comparable studies are available, no reasonable assumptions could be made regarding the expected incidences of the primary and secondary outcomes. Therefore, we did not perform a sample size analysis for the primary or the secondary endpoint.

The flowchart of patient distribution at the different time points and allocation to subgroup are shown in Fig. 1. The total study population consisted of 309 patients who underwent successful TF-TAVI, ambulating before the procedure and consenting to study participation. The main reason for ineligibility and thus exclusion 4 h after the procedure (T3) was that possible early ambulation was considered to be too hazardous, because of difficult vascular closure (n = 55, 35%), as decided by the operator. Nine of 159 (5.7%) of these patients had a closure device failure according to the VARC-II criteria. Thereafter, residual bleeding/‘oozing’ (n = 53, 33%), the presence of any systolic femoral murmur (n = 21, 13%) and the presence of a transvenous temporary pacemaker (n = 19, 12%) were the most prominent reasons for exclusion after 4 h. In the 21 patients deemed ineligible because of a systolic femoral murmur, a false aneurysm was found in 7 patients and was treated accordingly.

The eligible population had a mean age of 80 years, 48% were male and had a mean Society of Thoracic Surgeons Predicted Risk of Mortality (STS-PROM) score of 3.781 ± 1.842, reflecting contemporary practice in a lower-risk TF-TAVI population. These 150 eligible patients were allocated to either the EARLY (n = 73) or REGULAR group (n = 77), as previously described. Two eligible patients were not willing to ambulate early; no further ‘cross-over’ happened between the EARLY and REGULAR group. Baseline characteristics of the subgroups are shown in Tab. 1, and were equally distributed in the subgroups, except for a lower EuroSCORE II and slightly better estimated renal function (expressed as estimated glomerular filtration rate) in the EARLY group. There were no significant differences in pre-procedural medical regimen (i.e. anti-aggregation or anti-coagulation) between the two subgroups.

Procedural characteristics and outcome are shown in Tab. 2 and were similarly distributed in the EARLY and REGULAR group. The vast majority of the patients were treated using the third-generation balloon-expandable SAPIEN 3 (Edwards Lifesciences, Irvine, CA, USA) prosthesis (95%), with similar distribution in the two groups regarding valve type and valve size. All patients were treated using a fully percutaneous approach and local analgesia only. For arterial closure on the valve introduction side, double ProGlides were most frequently used. For the contralateral side, a single ProGlide or an Angio-Seal was used most frequently.

The outcomes regarding the primary and secondary endpoint are shown in Fig. 2 and Tab. 2. Time to mobilisation was four-fold longer in the patients following regular hospital protocol (4 h 49 min ± 31 min vs 20 h 7 min ± 3 h 6 min for EARLY vs REGULAR, p < 0.0001). There was no difference regarding the primary (safety) endpoint between the EARLY and REGULAR group. No major vascular or bleeding complications occurred in either group. The incidence of minor vascular complications, all minor bleedings, was similar in both groups (5.5% vs 7.8% for EARLY vs REGULAR, respectively, p = 0.570).

Lastly, the duration of the hospital stay in the total study cohort was relatively short (median 3 days) and statistically similar in the EARLY and REGULAR group.

Of the total cohort, 49% were deemed eligible for early mobilisation. Since this was a first-time trial, we predominantly focused on feasibility and safety, and thus were very strict in excluding patients considered to be at increased risk for complications after possible early ambulation. This cautious approach was also taken for the actual ambulation, which was performed under direct supervision of the nursing staff, taking into consideration the increased risk for falling incidents in this elderly, frail population. Considering the low number of minor vascular complications and the total absence of fall incidents, we succeeded in selecting patients for safe early ambulation. We believe that these results could be extrapolated to the patients who were treated in the afternoon, reasoned from the total absence of major complications (which require intervention by an interventional radiologist/vascular surgeon, preferably performed during daytime). Lastly, we report on a relatively low-risk population, when compared to large randomised trials like the PARTNER 2A and PARTNER 3, SURTAVI and CoreValve Low Risk Trial [3‐7]. Accordingly, our results and protocol could be used in other hospitals to introduce the possibility of early mobilisation after contemporary lower-risk TF-TAVI.

Most of the excluded patients at T3 (n = 159, 4 h post-procedure) were deemed ineligible for early mobilisation because of a difficult arterial closure so that early mobilisation was considered hazardous. Of these patients, 9 of 159 (5.7%) had a closure device failure according to the VARC‑2 criteria [19]. These criteria state that a failed closure device placement only accounts for ‘closure device failure’ when another (second) closure device is used. The actual number of failed closure devices was higher (n = 32/159, 20.1% of the excluded patients; and 32/309, 10.3% of the total study population). These failing closure devices were treated with additional (manual) compression of the femoral artery, and patients were excluded accordingly, being at increased risk for bleeding complications in the case of early ambulation. Newer closure devices may increase the number of successful closures, especially in these old and calcified femoral arteries, and thus enlarge the proportion of patients eligible for early mobilisation [25‐27]. Further studies could elaborate on the correlation between the quality of the peripheral vasculature (i.e. calcification burden) and the rate of successful closures, to ensure the maximum chance of successful closure and thus the possibility for early mobilisation.

The second reason for exclusion at T3 was residual ‘bleeding’, which was defined as any blood loss or active bleeding at the access site after 4 h (T3). Some of these cases probably were actually venous ‘oozing’, caused by the absence of a venous closure device. One could consider the possibility of adding a venous closure device to the procedural protocol, especially when used in combination with an additional cutaneous suture, which will increase eligibility for early mobilisation. Lastly, 21 patients were deemed ineligible because of a systolic femoral murmur; all underwent ultrasonography of the suspected femoral artery. In only 7 patients was a false aneurysm found and treated accordingly. In hindsight, the remaining 14 patients could have been eligible for early mobilisation, after the negative vascular ultrasound.

In addition to the aforementioned ‘venous’ access site bleedings in the excluded patients, 2 of 4 and 2 of 6 vascular complications in the eligible patients allocated, respectively, to the EARLY and REGULAR group originated from the non-valve introduction side. These could have been related to either the secondary arterial access or to the venous access for the temporary pacemaker. Elimination of the contralateral access site by using radial arterial access and applying left-ventricular pacing via the stiff wire may increase eligibility for early mobilisation.

Our study indicates that early ambulation is safe, and shows a benefit of early mobilisation regarding the in-hospital secondary endpoint, showing a significant two-fold reduction in the incidence of the combined secondary outcomes. In particular, patients who ambulated early experienced less pain and less need for unplanned urinary catheter use, while being on supine bed rest for 15 h less than the patients following the regular protocol. We believe that this combination significantly improves patient comfort. Moreover, our study shows a trend in which early ambulation may potentially decrease the incidence of post-operative delirium and infections, hereby taking into consideration of the fact that we already show a very low incidence of these debilitating complications. These low incidences underline the effect of the practice of ‘minimalist TAVI’ using local analgesia only in contemporary TF-TAVI and, possibly, now subsequent ‘minimalist’ immobilisation.

The FAST-TAVI (NCT02404467) and 3M-TAVI (NCT02287662) provide us with the insights on how to reduce the length of hospital stay, and showing it can be done without any additional risks, supported by a recent systematic review by Kotronias et al. [14, 17, 18]. Our study adds to these results, since a patient needs to be able to ambulate properly in order to go home safely. In this way, our study forms the next step in improving and minimalising the TAVI procedure and subsequent hospitalisation. Our study does not show a reduction in the duration of the hospital stay when early ambulation is performed. This may be partially explained by the fact that the hospitalisation is a median of only 3 days after the procedure we describe, which is relatively short when compared to data in the current literature.

Lastly, while conducting this study we received some quite positive feedback from both patients and the medical staff. Although in-hospital comfort for staff and patients may not be easy to quantify, it is considered a valuable goal, especially when considering the growing number of procedures and patients’ expectations as well as requests for less invasive treatments. Therefore, early ambulation for eligible TF-TAVI patients was included in the regular hospital protocol at our centre directly after completion of the study.

We believe that ‘minimalist’ TAVI and subsequent ‘minimalist’ immobilisation and hospitalisation will be the standard form of care in the very near future, considering the broadening indication, accumulating evidence and exponential gain in experience worldwide [6, 7, 13, 18, 28]. Several procedural changes have already been introduced recently (i.e. local analgesia, fully percutaneous access) and even more could be introduced in the near future, further minimalising the contemporary TAVI procedure. Using left ventricular pacing (instead of transvenous right ventricular pacing) and the radial approach for the secondary arterial access (instead of the contralateral femoral artery) could further diminish vascular complications and increase eligibility for early mobilisation. Additionally, using the jugular vein for the temporary pacemaker lead could enable early mobilisation in patients who are pacemaker-dependent directly after the TF-TAVI. Lastly, technological advances in prostheses (and incrementally decreasing required sheath sizes) and closure devices may further enable early mobilisation in the majority of patients after TF-TAVI. Of these patients, the most elderly, fragile population will probably benefit the most from an early mobilisation protocol. However, the future lower-risk population would probably enlarge the proportion of eligible patients and accordingly increase the overall gain from an early mobilisation protocol. This gain, in combination with further simplifying the procedure and shortening the subsequent hospitalisation, will lower costs and will improve the cost-efficiency of contemporary TAVI.

First, this study was designed as a prospective trial with allocation of treatment based on the time of the procedure, and not as a truly randomised trial. We drafted this design predominantly for safety reasons, since this is the first time early ambulation has been studied in this elderly, frail TAVI population. In this manner we could ensure that the actual ambulation would be performed during the fully staffed day shift. The absence of randomisation could have introduced bias into the patient selection. However, patients were randomly allocated to two pre-defined weekdays by our planning bureau, who did not have any information about the expected complexity of the case or the health status of the patient. This led to a comparable patient population in the EARLY and REGULAR group.

Secondly, this is a single-centre study. This gave us the unique opportunity to perform this study safely. However, due to the relatively small sample size, this may have deprived us of the chance to find any significant differences proving a benefit of early mobilisation for the individual secondary outcomes. Larger, preferably multicentre studies are needed to demonstrate this patient benefit, showing favourable outcomes regarding debilitating post-operative complications like delirium and infections. Nevertheless, we do show a two-fold lower incidence of the combined secondary endpoint when early ambulation is used, which warrants the adoption of such a protocol into contemporary TAVI practice. Furthermore, we predominantly used ProGlides for vascular closure. Extrapolation of our study results should be performed with caution, especially when using different arterial closure methods or when there are different circumstances regarding nursing and medical staff during the day. The adjudication of events in this study was not blinded or performed by a Clinical Event Committee, which raises inherent limitations to our study.