Keywords
thoracic aortic aneurysm - inguinal hernia - risk factors - screening
Introduction
Thoracic aortic dilatation (TAD) is most often asymptomatic and diagnosed either at
acute presentation or incidentally by examinations primarily performed for other reasons.
The risk posed by TAD is that of a sudden aortic event; aortic dissection or rupture.
Case fatality rate and surgical mortality for acute aortic events are high, approaching
70 and 20%, respectively.[1]
[2] Surgical mortality in elective TAD operations, on the other hand, is low, 0 to 2%.[3]
[4] In absence of specific, and relatively uncommon, risk conditions (e.g., Marfan syndrome
or family history), surgical indication is often based on aortic diameter. Increased
aortic diameter increases the risk of acute events.[5]
[6] Nevertheless, aortic diameter alone is unreliable and insufficient to accurately
predict acute events.[7] Therefore, to diagnose and route more surgical candidates to a strongly preferable
elective operation, screening would be valuable. However, TAD prevalence in the normal
population is comparatively low, approximately 15/100,000 men[8] and screening would profit from, or even demand, useful selection criteria.
Inguinal hernia is a relatively common condition, with reported lifetime prevalence
in men around 27%.[9] Inguinal hernia is much more common in men, related to the embryonal testicular
descent through the inguinal canal. Symptomatic inguinal hernia is usually treated
by surgical repair. Untreated hernia is associated with risk of ileus and incarceration
whereas outcomes of elective surgical repair are excellent.[9] Therefore, previous hernia repair is a useful proxy for inguinal hernia. If associated,
inguinal hernia could serve as an easily detected (by clinical examination and history)
marker, indicating increased risk of otherwise unsuspected TAD.
This study aimed to explore the association between inguinal hernia and TAD in men.
The study hypothesis was that the thoracic aorta is larger in men with previous inguinal
hernia repair compared with controls without hernia. This hypothesis was further divided
into two specific questions. First, is the prevalence of TAD increased in men with
a history of inguinal hernia repair compared with controls? Second, are thoracic aortic
diameters increased in men with a history of inguinal hernia repair compared with
controls? To test the study hypothesis, diameters of specified segments of the thoracic
aorta (root, ascending, isthmus, and descending) and presence of TAD in separate aortic
segments and overall were measured by chest computed tomography (CT) and compared
between patients with previous inguinal hernia and controls. History of inguinal hernia
repair was further assessed by multivariable analysis as an independently associated
predictor of TAD.
Materials and Methods
Study Population and Study Design
The study was approved by the regional research ethics committee (No 2015/1566–31)
and by the local radiation protection committee (No K2402–2015). Written informed
consent was obtained from all study subjects.
From January 2016 to September 2018, the four hundred consecutive adult (> 17 years)
men with previous (5 years or more) inguinal hernia repair at a single adult surgical
unit (Department of Surgery, Ersta Hospital, Stockholm, Sweden) were identified, and
those eligible (i.e., alive and not duplicated) subsequently invited to undergo chest
CT scan and provide self-reported health data. Similarly, another 400 consecutive
men with previous laparoscopic cholecystectomy for symptomatic gallstone disease (cholelithiasis,
a condition not known to have any association with connective tissue dysfunction or
aneurysm development) at the same unit during the same period of time, were identified
as controls. Accordingly, the controls were similar regarding sex, catchment area,
timeframe, and presenting type, that is, as elective general surgery patients treated
at the same facility. No specific exclusion criteria were applied. The minimum 5-year
hiatus between operation and study eligibility was chosen to reproduce the clinical
situation of evaluating the risk of TAD in a previously operated (inguinal hernia
and cholecystectomy alike) patient and to reflect adequate lead time, that is, time
from presence of a marker condition to diagnosis of the target condition (TAD).
Invitations, including study description, health questionnaire, informed consent sheet,
and a prepaid response envelope were sent by mail, with follow-up by phone when needed
for nonresponders or to expand on study details. The health questionnaire asked the
subject to provide numeric data on age, height, and weight and categorical (yes/no
if not stated otherwise) data on smoking (present/former/never), hypertension, myocardial
infarction or angina, chronic obstructive pulmonary disease, diabetes, connective
tissue disease (Marfan, Ehlers-Danlos), abdominal aortic aneurysm (AAA), family history
of any aortic aneurysm, and current medications. Study subjects were identified through
two existing Swedish nationwide, validated, all-inclusive quality registries: “National
Quality Registry for Gallstone Surgery and Endoscopic Retrograde Cholangiopancreatography”
and “National Quality Registry for Inguinal Hernia Surgery,” respectively.
Chest Computed Tomography and Aortic Measurements
Radiological examinations were performed between October 2016 and April 2019 at a
single unit (Department of Radiology, Ersta Hospital, Stockholm, Sweden). They were
performed as conventional chest CT with 5 mm axial slices with standard multiplanar
(sagittal and coronal) reformations, low dose of radiation (1.6–4.8 mSv), and without
intravenous contrast agent. This approach was chosen for study subject safety, but
also a part of the study rationale, to reflect a clinical “real-world” scenario comprised
of chest CT examinations performed primarily for nonaortic reasons. For the purpose
of this study, achieved CT imaging quality (exemplified in [Fig. 1]) was considered adequate and clinically relevant.
Fig. 1 Thoracic aortic diameter measurements in low-dose chest computed tomography without
intravenous contrast at (A) the aortic root (sagittal plane, sinus of Valsalva); (B) ascending aorta (transversal plane at pulmonary artery bifurcation); (C) aortic isthmus (coronal plane 10 mm from origin of left subclavian artery); (D) mid-descending aorta (transversal plane at pulmonary artery bifurcation).
Thoracic aortic diameters (in millimeters) were measured in prespecified aortic segments
by two independent radiologists blinded to clinical data. The prespecified segments
were: the aortic root at the mid-level of the sinuses of Valsalva; the mid-ascending
aorta at the level of the pulmonary artery bifurcation; the aortic isthmus just distal
to the origin of the left subclavian artery; and the mid-descending aorta at the level
of the pulmonary artery bifurcation ([Fig. 1]). The aortic root generated a single measurement from sagittal images, other segments
generated two orthogonal measurements. The single largest diameter obtained by any
of the two observers in each aortic segment of each patient was selected and compared
between the subjects with previous inguinal hernia repair and the control subjects.
TAD was defined as aortic diameter > 45 mm for the aortic root and ascending aorta
and diameter > 35 mm for the aortic isthmus and descending aorta, respectively. The
rationale for using the largest obtained aortic diameter was to avoid undue underestimation
of aortic diameter applying a “ruling-in” rather than a “ruling-out” strategy, similar
to clinical practice. Cutoff diameters for TAD (> 45 mm for the proximal aorta; > 35 mm
for the distal aorta) were chosen to represent an aortic dilatation clearly above
upper normal limits[10]
[11]
[12]
[13] and to be of clinical relevance since, in most circumstances, this would entail
further radiological study and follow-up, even if criteria for surgical repair are
not immediately met. For the proximal aorta, aortic diameter was further indexed to
body surface area (calculated using the DuBois and DuBois formula), generating the
aortic size index (cm/m2). The reported[5] threshold for moderate aortic event risk (≥ 2.08 cm/m2) was applied to define TAD according to the aortic size index.
Statistical Methods
Numbers were presented as counts with percentages and as means with standard deviations
(SDs; for normally distributed variables) or medians with interquartile range (for
nonnormally distributed variables). Univariate analyses were performed with Student's
t-test for continuous variables and Pearson's chi-square or Fisher's exact test for
categorical variables. Spearman's r was used to estimate correlation. Multivariable analysis, with TAD as the dependent
variable, was performed using logistic regression. Generating a predictive model,
variables with p-value < 0.10 in univariate analysis were included, followed by stepwise backward
exclusion to retain significantly (p < 0.05) associated variables. Logistic regression results were reported as odds ratios
(ORs) with 95% confidence interval. Sample size calculations were performed for prevalence
of TAD and for differences in aortic diameters, respectively. For 80% power with α = 0.05 and assuming 10% TAD in the study group versus 4% in the control group, required
sample size was 276 in each group. For 95% power with α = 0.01 assuming normal aortic size (34 mm) in the control group and increased size
by 1 SD (38 mm) in the study group, required sample size was 38 in each group. Based
on power calculations, sample size of 300 in each group was judged desirable, and
accommodating for 25% declining study invitation, 400 in each group were identified.
All statistical analyses were performed using Stata v16 (Stata Corp., College Station,
TX).
Results
The study participants' characteristics and comorbid conditions, from the self-reporting
questionnaire, are presented in [Table 1].
Table 1
Clinical characteristics of study subjects with previous inguinal hernia repair versus
previous cholecystectomy (controls)
|
Hernia repair
n = 216
|
Cholecystectomy
n = 254
|
p-Value
|
|
Age, years
|
70 ± 10
|
66 ± 9
|
< 0.001
|
|
Height, cm
|
179 ± 7
|
179 ± 7
|
0.72
|
|
Weight, kg
|
83 ± 11
|
89 ± 13
|
< 0.001
|
|
Hypertension
|
106 (49)
|
114 (45)
|
0.36
|
|
Ischemic heart disease
|
29 (14)
|
17 (7)
|
0.012
|
|
Smoking
Never
Previous
Current
|
99 (46)
107 (50)
9 (4)
|
130 (52)
109 (43)
13 (5)
|
0.36
|
|
Chronic obstructive pulmonary disease
|
22 (10)
|
27 (11)
|
0.85
|
|
Diabetes
|
25 (12)
|
32 (13)
|
0.75
|
|
Abdominal aortic aneurysm
|
5 (2)
|
1 (0.4)
|
0.065
|
|
Connective tissue disorder
|
1 (0.5)
|
1 (0.4)
|
0.90
|
|
Family history of aortic disease
|
12 (6)
|
8 (3)
|
0.19
|
Note: Mean ± standard deviation (SD) or n (%). p-Value using Student's t-test or Pearson's chi-square test.
Overall, groups were comparable, but subjects in the inguinal hernia repair group
were slightly older, weighed less, and more often had self-reported ischemic heart
disease compared with the control (cholecystectomy) group. Mean elapsed time from
operation to study participation was 8.5 years in the hernia repair group versus 10.5
years in controls. A total of 470/800 (59%) participants finalized the study with
complete data: Two hundred sixteen (216/400, 54%) with previous hernia repair and
254/400 (63%) with previous cholecystectomy. However, from study design it followed
that 82 men were ineligible: 62 in the hernia group (24 operated twice, 38 dead at
the time of study) and 20 in the control group (dead), corresponding to 65% overall
participation, 64% (216/338) in the hernia group and 66% (254/380) in the control
group ([Fig. 2]).
Fig. 2 Flowchart with exclusions forming the complete data study groups of subjects with
previous hernia repair (n = 216) and controls (cholecystectomy, n = 254).
At the time of identification, 58 men were deceased; 38/58 (65%) operated for inguinal
hernia (38/400 = 9.5% of the hernia group) and 20/58 (35%) with cholecystectomy (20/400 = 5%
of the control group), p = 0.025. Cardiovascular (nonaortic) death was most common in both groups: 15/38 (39%)
in the hernia group and 15/20 (75%) in the cholecystectomy group, followed by malignancy
(12/38 vs. 4/20). The only aortic death, an infrarenal aneurysm rupture, occurred
in the hernia group.
Characteristics of invitees providing questionnaire self-reported health information
only, that is, not undergoing CT and thus excluded from analysis of aortic measurements,
did not differ significantly from the study participants with complete data, indicating
possible random distribution of nonparticipants. Apart from weight (80 ± 9 vs. 92 ± 13 kg,
p = 0.001) there were no statistically significant intergroup differences (data not
shown).
Thoracic Aortic Diameters
As illustrated in [Fig. 3] and detailed in [Table 2], the hernia repair group had a mean aortic diameter slightly but statistically significantly
larger at the isthmus and descending aorta: 31 ± 3 versus 30 ± 3 mm (p = 0.008) and 30 ± 4 versus 29 ± 3 mm (p = 0.002), respectively. In the proximal aortic segments (root, ascending), there
were no statistically significant differences in mean diameters. Overall, interobserver
differences were small (1–2 mm differences) and correlation satisfactory (Spearman's
r 0.56–0.92), with most divergences and lowest correlation noted for measurements of
the aortic root ([Table 3]).
Table 2
Aortic diameters (mm), mean (standard deviation), and range
|
Hernia repair
n = 216
|
Cholecystectomy
n = 254
|
p-Value
|
|
Aortic root
|
36.8 (3.7)
28–51
|
36.5 (3.3)
29–48
|
0.32
|
|
Ascending aorta
|
38.3 (4.5)
28–59
|
37.8 (3.9)
27–49
|
0.24
|
|
Aortic isthmus
|
31.0 (3.1)
23–41
|
30.2 (2.9)
23–40
|
0.0075
|
|
Descending aorta
|
29.6 (3.8)
23–56
|
28.8 (2.8)
22–38
|
0.015
|
Table 3
Interobserver differences in aortic diameters per aortic segment
|
Median (IQR)
|
Range
|
≥ 5 mm difference, n (%)
|
Correlation (r)
|
|
Aortic root
|
2 (1–4)
|
0–14
|
91 (19)
|
0.56
|
|
Ascending aorta
|
1 (1–2)
|
0–9
|
11 (2.3)
|
0.92
|
|
Aortic isthmus
|
2 (1–3)
|
0–8
|
27 (5.7)
|
0.79
|
|
Descending aorta
|
1 (0–2)
|
0–9
|
9 (1.9)
|
0.88
|
Abbreviation: IQR, interquartile range.
Fig. 3 (A, B) Boxplot of aortic diameters in millimeters (mm) for aortic segments. Inguinal hernia
repair, green; control (cholecystectomy), blue. Thresholds for thoracic aortic dilatation
at 46 mm (root and ascending aorta; upper panel) and 36 mm (isthmus and descending
aorta; lower panel) indicated by red dashed lines.
Thoracic Aortic Dilatations
TAD, on the other hand, was more common in the aortic root with borderline statistical
significance (p = 0.052) and ascending aorta (p < 0.001) of hernia repair subjects ([Table 4]).
Table 4
Prevalence of thoracic aortic dilatation for aortic segments, individual and combined,
n (%)
|
Hernia repair
n = 216
|
Cholecystectomy
n = 254
|
p-Value
|
|
Aortic root > 45 mm
|
6 (2.8)
|
1 (0.4)
|
0.052
|
|
Ascending aorta > 45 mm
|
14 (6.5)
|
5 (2.0)
|
0.013
|
|
Proximal aorta
|
16 (7.4)
|
6 (2.4)
|
0.001
|
|
Aortic Isthmus > 35 mm
|
19 (8.8)
|
15 (5.9)
|
0.28
|
|
Descending aorta > 35 mm
|
10 (4.6)
|
6 (2.4)
|
0.21
|
|
Distal aorta
|
21 (9.7)
|
19 (7.5)
|
0.39
|
Note: Proximal aorta = Root + ascending; Distal aorta = Isthmus + descending.
Dilatation of the proximal aorta, the distal aorta, and all aortic segments combined,
respectively, were all significantly more prevalent in the hernia repair group, presenting
overall TAD in 50/216 (23%) versus 27/254 (11%), p < 0.001. Furthermore, TADs with indication for intervention based on aortic diameter
according to current guidelines were encountered in three subjects (3/216, 1.4%),
all in the hernia repair group.
Aortic diameter indexed to body surface area (aortic size index) indicated an even
more pronounced prevalence of (medium-risk and above) TAD in the hernia group: 31/216
(14%) versus 10/254 (4%), p = 0.0001 in the aortic root and 46/216 (21%) versus 32/254 (13%), p = 0.012 in the ascending aorta.
Independent Association between Hernia Repair and TAD
In multivariable analysis, previous hernia repair was found independently associated
with ascending aortic dilatation (OR 4.6) and proximal aortic dilatation (OR 5.3),
respectively, but not to dilatations of other aortic segments ([Table 5]). Specifically, no variables independently associated with aortic root TAD could
be identified: all cases appeared in the hernia group.
Table 5
Multivariable analysis (logistic regression) of independent predictors for thoracic
aortic dilatation in each aortic segment (root, ascending, isthmus, descending) and
segments combined (proximal, distal)
|
Odds ratio
|
95% Confidence interval
|
p-Value
|
|
Aortic root:
|
|
No independent predictors found
|
|
|
|
|
Ascending aorta:
|
|
|
|
|
Hernia repair
|
4.6
|
1.5–14
|
0.007
|
|
Age, y
|
1.09
|
1.03–1.15
|
0.003
|
|
Weight, kg
|
1.07
|
1.03–1.12
|
0.001
|
|
Proximal aorta:
|
|
|
|
|
Hernia repair
|
5.3
|
1.8–15
|
0.003
|
|
Age, y
|
1.09
|
1.03–1.15
|
0.002
|
|
Weight, kg
|
1.07
|
1.03–1.11
|
0.001
|
|
Aortic isthmus:
|
|
|
|
|
Age, y
|
1.15
|
1.09–1.21
|
p < 0.0001
|
|
Weight, kg
|
1.05
|
1.02–1.09
|
0.005
|
|
Abdominal aortic aneurysm
|
8.7
|
1.4–57
|
0.022
|
|
Descending aorta:
|
|
Age, y
|
1.11
|
1.04–1.18
|
0.001
|
|
Weight, kg
|
1.05
|
1.002–1.10
|
0.042
|
|
Distal aorta
|
|
|
|
|
Age, y
|
1.15
|
1.10–1.21
|
< 0.0001
|
|
Weight, kg
|
1.05
|
1.02–1.09
|
0.001
|
|
Abdominal aortic aneurysm
|
6.9
|
1.1–44
|
0.042
|
The 460 chest CT examinations yielded 28 (6.1%) pulmonary findings with indication
for further study and/or follow-up. Of these, two (0.4%) were found malignant.
Discussion
This study is the first of its kind, evaluating the association between inguinal hernia
and TAD by comparing aortic diameters and prevalence of TAD in subjects with a history
of inguinal hernia repair versus controls undergoing chest CT for this specific purpose
in a cross-sectional study design suitable to compare prevalence.
Differences in mean aortic diameters were generally very small, but aortic diameters
were statistically significantly larger in the distal parts of the aorta (the isthmus
and descending aorta) in the hernia repair group. These differences might be of limited
clinical relevance by themselves, and should be interpreted with caution, even if
sample sizes sufficed to achieve adequate power. Overall, the findings did not support
the study hypothesis of a generally increased thoracic aortic diameter in men with
previous inguinal hernia repair.
Much more importantly, TAD was significantly more prevalent in the hernia group, in
the root as well as the ascending parts and at a 16/216 (7.4%) prevalence, almost
three times as common in the proximal aortic segments compared with controls. In the
distal parts of the aorta, differences were smaller and not statistically significant.
For the ascending and proximal aorta, inguinal hernia repair was the most prominent
(OR 4–5) independent predictor in multivariable analysis, further strengthening the
association. Furthermore, TAD with diameters fulfilling criteria for intervention
were found only in the hernia repair group. These findings do support the study hypothesis
and strongly suggest that prevalence of proximal TAD can be expected to be considerable
in a population of 70+ years aged men with a history of inguinal hernia repair, with
implications for risk profiling, TAD detection, and screening. For perspective, a
TAD (> 4.9 cm) was found on only 12/4129 (0.34%) in a population study,[11] indicating the low yield of unselected population-based screening.
The only other variables independently associated with TAD were age and weight ([Table 5]). These findings were consistent with previous observations of larger aortic diameter
with increasing age[10]
[11]
[12]
[13]
[14] and body surface area.[5]
[11] Accordingly, indexing the aortic diameter to body surface area produced even more
pronounced differences, since weight (but not height) was higher in the control group.
Self-reported presence of AAA was associated only with isthmic or distal aortic dilatation.
This association has been described previously[15] and has not been suggested to encompass the proximal aorta.
Several previous studies have reported an association between inguinal and other abdominal
wall hernias including incisional hernias, with AAA,[16]
[17] but in a large population-based study such an association could not be confirmed.[18] While still interesting from a pathophysiological mechanistic perspective, a relationship
between hernia and AAA is of relatively less clinical importance, since full-scale
AAA screening programs are already operational in risk groups defined by sex and age.
In a cohort (n = 1,051) of patients undergoing surgery for TAD[19] a higher-than-expected prevalence of inguinal hernia as well as a larger proportion
of inguinal hernia in patients undergoing proximal aortic surgery compared with isolated
aortic valve surgery was reported, indicating a possible association. Recently, a
large study used propensity scoring to match 16,933 hernia patients to 50,799 controls
and found a higher incidence of TAD (6.4 vs. 4.8/10,000 person-years, p = 0.03) in the hernia group.[20] Such population-based epidemiological studies are based on administrative data sources
and coded diagnostic data not intentionally collected to analyze specific queries,
but certainly corroborate findings of smaller, clinical, hypothesis-testing studies.
Molecular mechanistic studies suggest several similarities between connective tissue
deficiencies in hernia and aneurysmatic aortic wall, but there is not yet sufficient
evidence to support a common molecular mechanistic pathway for hernia and TAD.[21]
Presently, hernia or history of hernia repair is not recognized as a risk factor in
current TAD guidelines[22] but could be considered in forthcoming revisions. In contrast, hypertension and
smoking are also associated with TAD, but prevalent in 67 to 86% and 51 to 62%, respectively,[1]
[23] and, although important associated conditions, therefore less useful as discriminating
features to guide screening. Known risk factors for TAD such as Marfan syndrome and
family history, on the other hand, are insufficiently prevalent to help guide selective
screening in normal populations, and, when identified, already subject to dedicated
follow-up. Simple renal cysts have been identified in 38 to 57% of patients with TAD
(compared with 15% in non-TAD) and could also serve as an interesting marker,[24] but unlike hernia, renal cysts cannot be detected by clinical examination and medical
history but require abdominal CT or ultrasound for detection.
Study Limitations
In the present study, only men were included, because of their much higher prevalence
of inguinal hernia. Women can suffer from inguinal and other abdominal wall hernias,
and future studies evaluating a correlation between hernia and TAD in women would
be important to definitively establish inguinal hernia as being associated with TAD
in both sexes.
Importantly, aortic valve cuspidity was not known in this study. Bicuspid aortic valves
(BAVs) are associated with dilated aortic root and ascending aorta. On the other hand,
with BAV, diameters of the distal aorta are normal or even smaller than with a tricuspid
aortic valve.[25] In this study, diameters of the distal aorta were significantly larger ([Fig. 2]) and distal TAD more common in the hernia repair group ([Table 3]), suggesting an effect not driven by presence of BAV.
Conclusion
In conclusion, inguinal hernia, or its proxy previous hernia repair, could serve as
a marker to help make informed decisions about screening or diagnostic TAD workup
to promote TAD detection and timely repair. Study findings supported the hypothesis
of increased prevalence of TAD in men with inguinal hernia. Notwithstanding its clinical
relevance, further mechanistic studies are needed to understand the relationship better.
Editor's Commentary
In their beautifully planned and executed study, Carlestål and colleagues have shown
that inguinal hernia may well be a marker for ascending aortic disease. This is an
important observation that can help to identify asymptomatic ascending aneurysm patients
in the general population. This fits in with other “Guilty Associates” that we have
identified, which should prompt suspicion of and investigation for ascending aneurysm
disease.[1]
The overall findings in the paper from Sweden are so strong that I am not concerned
by the lack of a difference in average diameters. The prevalence of aneurysm was simply
not high enough to be reflected in mean diameters of the two groups.
The aortic root is difficult to measure and, as we have written, definitions and techniques
of assessing diameter in this zone are nonuniform and problematic.[2] I believe the method of estimation used by the authors, and illustrated in [Fig. 1], is appropriate for this type of study.
We thank the Swedish group for this important report. We are also grateful to the
patients for submitting to a limited CT scan just for the purposes of this investigation.
References
1 Elefteriades JA, Sang A, Kuzmik G, Hornick M. Guilt by association: paradigm for
detecting a silent killer (thoracic aortic aneurysm). Open Heart 2015;2(01):e000169
2 Elefteriades JA, Mukherjee SK, Mojibian H. Discrepancies in measurement of the thoracic
aorta: JACC review topic of the week. J Am Coll Cardiol 2020;76(02):201–217
