You are on page 1of 5

INTRODUCTION:

Atrial fibrillation is the most common cardiac rhythm disorder affecting the general
population and a major cause of stroke, heart failure, and generalised mortality. In
fact, there were a combined 50 million individuals with AF or atrial flutter in 2020,
with adults aged 55 years or older experiencing an overall lifetime risk of AF as high
as 37%. Studies demonstrate substantial AF-related healthcare expenditure due in
large part to long-term pharmaceutical treatment and hospitalizations. However,
projections in the USA and EU slate dramatic 2-fold increases in AF prevalence by
the years 2030 and 2060 respectively. These alarming trends have been largely
attributed to an aging global population and rising rates of modifiable risk factors,
leading to what will amount to a public health crisis in the coming decades, without
adequate intervention.

MATERIALS AND METHODS:

119 consecutive patients that had been diagnosed with either PAF or PeAF and
underwent a catheter ablation procedure at ‘Spitaul Clinic de Recuperare Cluj-
Napoca’ during the period of November 2020 to August 2022 were selected.
Multislice CT-angiography was performed on all patients prior to catheter ablation,
from which cardiac 3D-reconstructions were obtained and imported into the CARTO
3 Mapping System. Software utilities were used to isolate the LA and LAA from
adjacent anatomical structures as well as delineate their borders.

Measurement parameters were chosen based on a variety of factors, namely, that


they could be measured or extrapolated to a reasonable degree of accuracy using
the aforementioned means of visualisation, that they captured clinically relevant
aspects of size and that they coincided with variables measured in other
contemporary studies as to form a basis of comparison. A margin line representing
the common border of the LA and LAA, and thus the boundary of the LAA orifice,
was manually placed on each 3D-reconstruction model based on anatomical
landmarks described in other studies. The LAA orifice major and minor axes were
each measured using two manually selected points lying on the margin line,
representing the longest and shortest diameters of the orifice. The LAA orifice
circumference variable was derived using the software perimeter function,
corresponding directly to the length of the orifical boundary. LAA depth was
measured using two points corresponding to the longest straight-line distance
between the orifical boundary and a section of LAA lying on the z-axis of an orifical
plane projection that used the orifical boundary point as the origin. In taking depth
measurements, care was placed to respect physical boundaries of the varying LAA
morphologies that were observed and to avoid clipping. LAA Area, total area
(combined LA and LAA areas), LAA% of total area, and LA and LAA volume were
calculated by the software engine based on anatomical delimitations that were
manually set.

Raw data, including demographic information and measurements were collected in


Microsoft Excel and imported into IBM SPSS Statistics 29 for analysis. Initially, a Chi-
square test was used to investigate the association between the categorical
variables. Shapiro-Wilk tests were performed on all the variables, in respect to the
two categories of diagnosis, to assess for normal distribution. Missing data analysis
and Little’s MCAR test were employed to review the randomness of the missing
dependent data points and to evaluate the feasibility of leveraging an imputation
model solution to predict their values. Given the non-normal distribution of the
variables in regard to diagnosis, a PMM imputation model approach was utilised,
resulting in multiple iterations of imputed data sets being produced that were then
pooled. The pooled data set formed the basis on which Mann-Whitney U tests were
applied, and test statistic and p values were calculated. The significance level for all
tests was p < 0.05. Results for non-normally distributed continuous variables were
presented as medians (IQR). Categorical variables were presented as numbers and
percentages.

RESULTS:

119 patients (mean age = 60 ± 9.6 years, M:F = 69 male: 50 female) were included
in the study. 81 patients were diagnosed with PAF (68.1%) and 38 were diagnosed
with PeAF (31.9%). No correlation between sex and diagnosis of either PAF or PeAF
was observed (x2 (1, N = 119) = .614, p = .433). Age was also not significantly
different among patient groups (Mdn PAF = 62 years, Mdn PeAF = 61.5 years, z
= .416, p = .677).
The following table details the morphological parameters of the LA and LAA that
were compared among the PAF and PeAF patient groups.

p
Variable PAF PeAF z value
value
LAA orifice major axis 23.600 (19.300, 25.200 (22.450,
2.534 .011*
(mm) 25.800) 28.775)
LAA orifice minor axis 16.100 20.050
2.637 .008*
(mm) (13.450,19.200) (14.175,22.050)
LAA orifice 7.800 (6.200, 9.100 (7.100,
1.893 .058
circumference (cm) 9.100) 10.300)
26.100 (21.150, 25.200 (20.900,
LAA depth (mm) .516 .606
29.300) 30.425)
17.500 (11.500, 20.450 (15.150,
LAA area (cm) 2.269 .023*
23.150) 27.100)
261.950
226.600 (198.100,
Total area (cm) (226.475, 2.502 .012*
268.550)
284.600)
6.900 (6.050, 7.300 (5.700,
LAA area % .761 .447
9.150) 10.525)
162.250
129.200 (102.500,
LA volume (mL) (144.075, 4.743 < .001*
150.950)
206.200)
7.200 (5.750, 10.000 (7.325,
LAA volume (mL) 2.682 .007*
10.700) 11.625)

The LAA orifice major axis (p = .011), LAA orifice minor axis (p = .008), LAA area (p =
.023), total area (p = .012), LA volume (p = < .001) and LAA volume (p = .007) were
all significantly greater in patients with PeAF than in those diagnosed with PAF.
Adversely, no statistical significance was noted in the LAA orifice circumference, LAA
depth, or LAA area % (p = >.005).
DISCUSSION:

Our study demonstrated that patients with PeAF had larger LAA orifice major and
minor axes compared to their PAF counterparts. Additionally, LAA and combined LA/
LAA areas as well as LA and LAA volumes were also significantly larger in the latter
group. These results corroborate the findings of other authors and illustrate the
precision and repeatability of using 3D-CT reconstruction techniques to measure the
different parameters of the AF-stricken LAA.

LAA occlusion is a contemporary therapeutic technique that has proven useful in


abating thromboembolic complications in patients with AF. Clinicians sizing occlusion
devices may derive significant value from the observations contained within this
study and others like it to more accurately assess and predict device dimensions
based on their patient’s diagnosis.

This study was limited in its ability to accurately assess LAA orifice circumference in
particular, due to 3D-CT reconstruction artifacts that were encountered on a few
occasions. In most instances, the left pulmonary veins formed a continuous structure
with the LAA, frequently deforming the posterior aspect of the LAA orifice. In the
absence of a method by which to remove these artifacts consistently, manual LAA
orifice boundary placement was adjusted in these circumstances only, to bypass
these erroneous sections, albeit as tightly as possible as to minimise the degree of
deviation from the expected circumference. Due to the fact that the same margin line
was also utilised in software-based calculations of LAA area, LAA area%, LA and
LAA volume for these patients, an element of error was also introduced into the
values of these measurements. However, the relative impact on these variables can
be assumed to be much lower as the area and volume loss caused by deviating the
orifice border to avoid artifacts is negligible in comparison to the relative effect on
circumference lengthening that occurs when performing the same process. In
situations in which artifacts were either more extensive or conferred a level of
deviation that was considered unacceptable, certain variables could not be
determined by the software. The devised approach to deal with this fact was to
impute the missing values based on a model that factored in the particular variable’s
distribution and relationship with all the other variables both dependent and
independent.

No correlation was observed between sex and diagnosis. Furthermore, it was not in
the scope of this study to compare LA and LAA dimensions between sexes, as the
physical characteristics of patients were not collected and thus standardisation by
body surface area could not be performed. Sex-focused comparisons have been
explored in other studies yet are more frequently encountered in the context of AF
and non-AF patient groups, as are the majority of studies pertaining to LAA
dimensions. As such, this study provides novel insight into the changes in these
dimensions as seen in different types of AF, contributing to a better understanding as
it pertains to their effect on the pathophysiology of AF progression.

Beyond the dimensions of the LA and LAA, the morphology of the LAA has been
demonstrated to be an important indicator of embolic stroke. Despite the observation
of these morphologies during data collection, these aspects did not comprise the
scope of this study.

CONCLUSIONS:

In this study, patients with PeAF had a greater LAA orifice major axis and minor axis,
LAA area and total combined LA/ LAA area, LA and LAA volume when compared to
patients with PAF. However, LAA orifice circumference, LAA depth and LAA% area
did not change significantly among these patient groups, nor was there an
association between gender and diagnosed PAF or PeAF.

You might also like