Results
In the absence of randomised controlled trial data, prospectively
collected observational data offer reliable alternative for such a
comparison. In this regard, we reported a large series of SC and TA
cases over a long period and each limitation of no-randomized
observational study was robustly addressed by a propensity score
analysis. From the UK TAVI registry, 1,506 patients were suitable for
our analysis. The TA arm grouped 1,216 patients, while the SC collected
290 patients. Demographic characteristics of the 1,506 patients are
summarized in Table I. To reduce biases which are related to confounding
variables and to the observational study per se, we performed a
propensity score matching weighted analysis for LogEuroSCORE II, valve
type (balloon-expandable/self-expandable), presence and severity of
coronary artery disease (one, vs two, vs three coronary arteries),
access route (SC vs. TA), heart rhythm (atrial fibrillation vs. sinus
rhythm), and year of implantation (2015 vs. 2007). Stroke, major/minor
vascular complications, major/minor bleeding, tamponade, permanent
pacemaker implantation, acute kidney impairment within 7 days, renal
replacement therapy, emergency valve in valve needs, paravalvular leak,
balloon re-dilatation, and in operating room, 30-days, and 12-month
mortality were all treated as dichotomous post-operative outcomes and
their analysis is summarized in Table II. In the Cox proportional hazard
model, 1,263 patients were fit to be entered and the results are
presented in Table III. The median age of patients was 80 (IQR 75–85)
years and 30% were female. SC access patients were marginally older
than TA ones. Significantly more men were approached via SC route (p =
0.04), while the Logistic EuroSCORE was almost similar among groups (p =
0.06). The body mass index (BMI) distribution was comparable. SC access
was almost exclusively used for the self-expandable device (93.9%),
whereas most TA cases were for a balloon-expandable device (89.8%).
Acute Kidney Injury stage III within 7 days after procedure and the
renal replacement therapy rate were almost similar between groups.
Conversely, the rate of permanent pacemaker implantation (PPI) in the SC
(28%) was significantly greater than that in the TA (11.0%) group (p =
0.02), but this data did not affect the overall survival at our
analysis. Vascular complications and major peripheral vascular injuries
that required vascular surgery repair were observed with comparable
frequency within groups, 1.0% in SC and 0.6% in TA. While in SC they
were purely related to the access, in TA they occurred as complication
of femoral percutaneous access for concomitant coronary intervention (4
cases) or accidental injury of the femoral artery, which occurred during
the placement of the venous temporary intracavitary pacemaker catheter
(8 cases). The rates of in-hospital acute myocardial infarction (within
72 hours after the procedure), in-hospital TIA, and stroke were not
significantly different among groups (p = 1.0, p = 0.25, and p = 0.09,
respectively). The rates of moderate and severe paravalvular leakage,
balloon re-dilatation after valve deployment, emergency valve in valve
procedure, bleeding, cardiac tamponade, emergency cardiac surgery, and
in-operating room death were not significantly different among groups.
Conversely, we found that in the SC the procedural time (193.24 ±77.3529
minutes) was significantly longer (p= 0.04) than that in TA (123.10
±55.12 minutes), meanwhile the average length of hospital stay was
significantly reduced (p = 0.01) in SC group (9.8 ±7.5 days) compared to
TA (13.3 ±7.5 days). Moreover, the 30-days mortality was not
significantly different in SC compared to TA, while the 12-months
mortality was significantly higher in SC than in TA (p = 0.01).
Estimates of long-term survival are represented by Kaplan-Meier curves.
Long-term survival in the SC and TA groups at 8-year follow up was not
significantly different (p = 0.77), as it is represented in Figure 1.
Finally, based on our analysis, there was no significant difference in
long-term survival between patients who received balloon-expandable
devices compared to those who received self-expandable devices (p =
0.26), as it is shown in Figure 2.