Available via license: CC BY 4.0
Content may be subject to copyright.
ORIGINAL ARTICLE
Biomechanical properties of feline ventral abdominal wall and
celiotomy closure techniques
Fernando S. Reina Rodriguez DVM, MSc
1
|
Conor T. Buckley PhD
2
|
Joshua Milgram DVM, BVSc, DECVS
3
|
Barbara M. Kirby DVM, MS, DACVS, DECVS
1
1
University College Dublin School of
Veterinary Medicine, Dublin, Ireland
2
Trinity College Dublin Centre for
Bioengineering, Dublin, Ireland
3
Koret School of Veterinary Medicine,
Hebrew University of Jerusalem,
Jerusalem, Israel
Correspondence
Fernando S. Reina Rodriguez, Section of
Veterinary Clinical Studies, School of
Veterinary Medicine, University College
Dublin, Belfield, Dublin 4, Ireland.
Email: fernando.reina.vet@gmail.com
Funding information
Veterinary Council of Ireland.
Abstract
Objective: To compare biomechanical properties and mechanism of failure of 3
regions of ventral abdominal wall in cats by using 2 suture materials, 2 suture bite-to-
stitch intervals (SBSI), and full-thickness versus fascia-only closure.
Study Design: Randomized, cadaveric, ex vivo mechanical testing.
Sample Population: 16 adult cat cadavers, 3 samples per cat.
Methods: Three regions of ventral abdominal wall were mechanically tested (N 548
samples). Preumbilical, umbilical (U), and postumbilical (POU) regions were har-
vested by using a template. The thickness of the linea alba was recorded. Six samples
without celiotomy served as controls. Twenty-eight samples were randomized to
SBSI (2 32or535 mm) and suture material (3-0 polyglactin 910 or 3-0 polydiox-
anone) for simple continuous celiotomy closure. Fourteen samples were randomized
to full-thickness or fascia-only closure. Samples were tested by linear distraction; ten-
sile strength and mechanism of failure were recorded. Effects of body weight,
thickness of linea alba, anatomic region, SBSI, type of closure, and suture material
were evaluated by mixed model linear analysis. Load to failure was compared
between males and females, full-thickness and fascia-only closure by independent
ttest, with P<.05 considered statistically significant.
Results: The POU region achieved lower loads to failure. Load to failure was greater
in males compared with females. No difference was detected between full-thickness
and fascia-only closure. Failure most commonly occurred by tearing of suture through
tissues. Tissue failure with suture line loosening occurred mainly in the 5 35-mm
SBSI group.
Conclusion: The POU region is biomechanically weak and may therefore be predis-
posed to incisional herniation.
1
|
INTRODUCTION
Ventral midline celiotomy is commonly performed in small
animals. Long incisions are required for exploration and
complex procedures,
1
whereas short incisions in the caudal
abdomen are used for ovariohysterectomy and cystotomy.
Abstract presented at the Resident Forum, European College of Veterinary
Surgeons Annual Scientific Meeting, July 2016, Lisbon, Portugal.
.......................................................................................................................................................................................
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the orig-
inal work is properly cited.
V
C2017 The Authors Veterinary Surgery published by Wiley Periodicals, Inc. on behalf of American College of Veterinary Surgeons
Veterinary Surgery.2017;1–11. wileyonlinelibrary.com/journal/vsu
|
1
Received: 28 December 2016
|
Accepted: 6 July 2017
DOI: 10.1111/vsu.12751
Complications of ventral midline celiotomy in small animals
include surgical site infection, peritonitis, wound dehiscence,
and incisional herniation with or without eventration.
2
Inci-
sional dehiscence following ventral midline celiotomy may
occur acutely or chronically, weeks to years after surgery.
3
Although a 0.18% rate of incisional hernia has been reported
in small animals following ventral midline celiotomy,
4
this
complication is likely under reported. Incisional complica-
tions were observed in 4% of patients in a retrospective study
of 200 dogs and cats undergoing exploratory laparotomy, but
the incidence of incisional herniation was not specifically
reported.
1
Incisional hernia was reported in 11 of 292 dogs
(3.8%) and 8 of 74 cats (10.8%) in a retrospective study of
types of hernias (C. Bellenger, personal communication,
1990). Major abdominal evisceration injuries were described
in a retrospective study of 8 dogs and 4 cats, located at the
incision for ovariohysterectomy in half of the dogs and all of
the cats.
5
Incisional hernias in domestic cats appear to occur
more frequently after caudal ventral celiotomy.
5
Although
this likely reflects the high prevalence of ovariohysterectomy
among abdominal procedure performed in this species, it is
unknown whether anatomical differences between cranial
and caudal ventral abdominal wall,
6
different thickness along
the linea alba, and/or different biomechanical properties of
abdominal wall muscles
7
could contribute to ventral inci-
sional hernias in the caudoventral abdomen in domestic cats.
Guidelines for ventral midline celiotomy closure in small
animals include recommendations for distance between
suture bites and stitches, but these are based on clinical expe-
rience
4,8
and are not supported by biomechanical testing in
cats. To the best of our knowledge, biomechanical studies
have been performed only in dogs.
4,9
In these studies, the
external leaf of rectus abdominis sheath has been identified
as the primary strength-holding layer for ventral paramedian
incision closure in dogs.
9
To the best of our knowledge,
fascia-only closure has not been previously reported in
domestic cats. The force that the feline abdominal wall must
resist to prevent wound dehiscence is unknown. Establish-
ment of feline anatomical and biomechanical data may pro-
vide evidence to support recommendations for ventral
midline celiotomy closure in cats.
Our study seeks (1) to compare the biomechanical
properties and mechanism of failure of discrete segments
of feline ventral abdominal wall and linea alba (preumbili-
cal region [PU], umbilical region [U], and postumbilical
region [POU]), (2) to compare the load to failure and
mechanism of failure of celiotomy closure in discrete seg-
ments of feline ventral abdominal wall by using 3/0 poly-
dioxanone and 3/0 polyglactin 910 and 2 suture-bite-stitch
intervals (SBSI), and (3) to compare the load to failure and
mechanism of failure of celiotomy closure in discrete seg-
ments of feline ventral abdominal wall using full-thickness
suture bites or fascia-only suture bites in the closure. We
hypothesized that (1) the POU region of feline ventral
abdominal wall is biomechanically weak compared with
the PU and U regions, (2) there is no biomechanical differ-
ence in celiotomy closure between the 2 suture materials
tested, (3) there is no biomechanical difference between 2
SBSIs tested, and (4) there is no biomechanical difference
between full-thickness and fascia-only suture bites in feline
celiotomy closure.
2
|
MATERIALS AND METHODS
Full ethical approval was obtained from the institutional ani-
mal research ethics committee of the University College
Dublin School of Veterinary Medicine, Dublin, Ireland.
2.1
|
Tissue collection and sample preparation
Adult domestic cats euthanized for reasons unrelated to this
study with no evidence of systemic or local disease affecting
abdominal wall (eg, obesity, cachexia, recent abdominal sur-
gery, abdominal trauma, or neoplasia) were donated. Cadav-
ers were stored frozen at 2808C within 30 minutes after
euthanasia and thawed at room temperature 24 hours prior to
testing. Body weight (BW; kg), gender, and breed were
recorded. With the cat in dorsal recumbency, hair was
clipped and the abdominal wall was inspected for bruising,
asymmetry, wounds, defects, or other evidence warranting
exclusion. Skin and subcutaneous tissue were removed. Ven-
tral abdominal wall was harvested by transverse incisions at
caudal margin of xyphoid and cranial margin of pubis and
longitudinal incisions 10 cm lateral to the linea alba on each
side (Figure 1A). Immediately after harvest, the falciform lig-
ament was removed, and the specimen was pinned to card-
board. Craniocaudal length was measured with a ruler and
recorded.
Three regions of interest (ROI) were defined, PU, U, and
POU, for each abdominal wall sample. One sample of each
ROI was obtained from each cat by using a “dog bone”
shaped template (polylactic acid) manufactured by 3D printer
(Ultimaker 21; Ultimaker BV, Geldermalsen, the Nether-
lands). The shape of the template was based on recommenda-
tions for tensile testing of materials with elastic properties.
10
The template was 1.5 cm long with 5-cm shoulders and 2.5-
cm width at its center (Figure 2A). Each ROI sample in the
study was harvested with a scalpel blade after securing the
template transversely to the body wall while centered directly
over the linea alba. The sample at the linea alba was con-
firmed to measure 2.5 cm in width with digital calipers (Fish-
erbrand Traceable digital calipers; Thermo Fisher Scientific,
Waltham, Massachusetts) immediately after harvest. In addi-
tion, the thickness of the linea alba was measured at the cen-
ter of each sample with digital calipers and recorded. Care
2
|
REINA RODRIGUEZ ET AL.
was taken to avoid compression of tissue with pressure of
the calipers. Samples were immersed in a phosphate buffered
saline bath to avoid desiccation and stored at 48C prior to
biomechanical testing performed within 6 hours after sample
preparation.
Samples were randomly assigned to 3 groups (control,
SBSI/suture material, and type of closure) by using computer
software (SPSS Statistics, Version 20; IBM, Armonk, New
York). Control samples were left intact. Celiotomy samples
were bisected by scalpel on the linea alba. Celiotomy sam-
ples were closed by simple continuous suture secured with
starting square knot of 5 throws and ending square knot of 6
throws.
11
After completing knots, suture material was cut,
leaving 3-mm ends. Length of suture material used was cal-
culated by adding lengths of remnants plus remaining length
on needle (all measured with digital calipers) and subtracting
from suture material length recorded on suture packet. Suture
length to wound length ratio (SL:WL) was calculated as
length of suture material used for closure divided by standar-
dized wound length (2.5 cm).
SBSI/suture material group was divided into 4 subgroups
based on SBSI and suture material used for celiotomy clo-
sure. SBSI was designated 2 32or535, indicating a dis-
tance of 2 mm or 5 mm, respectively, from suture bite to
lateral edge of the linea alba (grossly identified as the junc-
tion between the linea alba and the external leaf of rectus
FIGURE 1 A, Cat in dorsal recumbency after skin removal from ventral abdomen. B, “Dog bone”shaped samples indicating suture bite and stitch
interval. C, Diagram of needle penetration for full-thickness and fascial-only closure
REINA RODRIGUEZ ET AL.
|
3
abdominis muscle sheath) and a distance of 2 mm or 5 mm,
respectively, between suture bites (Figure 1B). Distance from
suture bite to the lateral edge of the linea alba and between
suture bites was standardized and marked on each sample by
using hectograph paper, a marker, and a ruler for accurate
needle penetration and suture spacing. A template was drawn
with lateral edges of the linea alba marked as solid lines and
points marking distance (mm) between suture bites and dis-
tance (mm) from lateral edges of the linea alba marked for
needle penetration for SBSI (Figure 2B). Celiotomy was
closed by using either polyglactin 910 USP 3-0 (Vicryl; Ethi-
con, Somerville, New Jersey) or polydioxanone USP 3-0
(PDS II; Ethicon) on a 26-mm SH taper needle. In “type of
closure”group, celiotomy was closed by using simple con-
tinuous 2 32 SBSI and polydioxanone USP 3-0 on a 26-
mm SH taper needle with full-thickness or fascia-only tissue
bites (Figure 1C).
2.2
|
Biomechanical testing
Samples were mounted in a testing machine (MTM; Zwick
Roell Z005; Zwick/Roell GmbH, Ulm, Germany) equipped
with a 100 N load cell. Sample shoulders were secured with
custom-made clamps attached to stationary and moving anvils
of the MTM, equidistant from the linea alba and oriented per-
pendicular to the direction of distraction (Figure 2C). Each
clamp consisted of two 5 35-cm metal plates with cheese
grater blades secured to their inner surfaces and compressed
the sample shoulders by tightening 2 screws 3 cm apart. Rota-
tion of the sample in the grips prior to testing allowed uniform
distribution of load across the sample width.
Samples were preloaded to 1 N and then tested under
vertical displacement uniaxial loading at 0.5 mm/second to
failure. Strain (%), tissue elongation (mm), and load to failure
(N) were recorded for each test. Stress was calculated using
the formula r5F
A; where rrepresents stress, Fis applied
load (N), and A(mm
2
) is initial cross sectional area calcu-
lated by multiplying the width (25 mm) and thickness of
each sample. Mechanism of failure was recorded by digital
camera (VG-160; Olympus, Shinjuku, Tokyo, Japan). All
tests were performed and interpreted by a single investigator.
Load to failure was defined as the maximum force (N)
recorded before a sharp decline on load displacement curve
followed by automatic termination of test. Test data were
transferred automatically by computer software to an Excel
2010 document (Microsoft, Redmond, Washington). Mecha-
nism of failure was evaluated and classified after completion
of all biomechanical tests.
2.3
|
Statistical analysis
Data were analyzed in statistical software (SPSS Statistics,
Version 20; IBM). A test for normality was performed for all
variables. Descriptive statistics were performed to determine
mean BW, thickness of linea alba, and load to failure. Spear-
man’s rank correlation coefficient was calculated to evaluate
correlation between BW, gender, and thickness of linea alba.
Mixed model linear analysis was performed to evaluate
effect of suture material, SBSI, type of closure, and anatomi-
cal location (repeat measures/fixed effects) on load to failure
(dependent variable). BW, gender, and thickness of linea
alba were subjects/random effects in the statistical model.
FIGURE 2 A, “Dog bone”shaped sample over template. B, 5 35 celiotomy sample. Points marked on external sheath of rectus abdominis to iden-
tify site of needle penetration during closure. The 2 parallel lines represent edges of the linea alba. C, Sample mounted in materials testing machine
4
|
REINA RODRIGUEZ ET AL.
Stepwise regression analysis was performed to remove non-
significant variables from the statistical model; P<.05 was
considered statistically significant. One-way ANOVA for the
final statistical model was performed to compare mean load
to failure between groups, followed by post hoc analysis. An
independent ttest was performed to compare mean load to
failure between full-thickness and fascia-only closure and
load to failure for samples obtained from male and female
cats. P<.05 was considered significant.
3
|
RESULTS
Eighteen adult cat cadavers were evaluated for inclusion in
the study. Two cats were excluded, 1 with evidence of
abdominal trauma and 1 with a large subcutaneous mass
firmly adhered to and invading the ventral abdominal wall.
Sixteen cats (N 516) were included in the study. All speci-
mens were abandoned cats donated from rescue centers.
Information provided from the rescue centers was limited to
estimated age (range, 2-8 years) and date of euthanasia.
There were 7 females (3 entire, 4 neutered) and 9 males (5
entire, 4 neutered). Breeds included 9 domestic shorthaired
cats, 6 domestic longhaired cats, and 1 Abyssinian. The
mean 6SD BW was 3.41 61.16 kg (range, 1.30-5.28).
Four female cats used in our study had undergone ovario-
hysterectomy. We did not observe evidence of previous sur-
gical incision, suture material, scar, or adhesions on
inspection of the ventral abdominal wall prior to harvesting
the samples. We concluded that either a previous celiotomy
incision was completely healed without visible evidence of
incision or a flank approach had been used. These specimens
were included in the study.
3.1
|
Thickness of the linea alba
The thickness of the linea alba (mean 6SD) ranged from 0.6
to 1.98 mm (1.13 60.40). The thickness ranged ranged from
0.71 to 1.98 mm (1.39 60.4) in the PU region, 0.63 to
1.63 mm (1.08 60.33) in the U region, and 0.6 to 1.4 mm
(0.91 60.32) in the POU region (Figure 3A). This thickness
decreased gradually from cranial to caudal in all cats
(P5.002), with thicker samples obtained from the PU region
compared with those from the POU region, except in speci-
men 2, in which the thickness was identical in the PU and U
regions. There was a moderate correlation between BW and
thickness of the linea alba (R50.427, P5.02). No correla-
tion was observed between gender and thickness of the linea
alba (P5.179) or between gender and BW (P5.06).
3.2
|
Group distribution
Forty-eight samples (N 548) were tested biomechanically
(3 samples per cat). Six samples from 2 cats (n 56) served
as controls, with 2 samples from each ROI. Forty-two sam-
ples were obtained from the remaining 14 cats (14 samples
from each ROI). Twenty-eight of these 42 samples (n 528)
were allocated to the SBSI/suture material group (n 528),
with 7 samples (n 57) in each subgroup, depending on
FIGURE 3 A, Thickness of the linea alba (mm) per region of interest
(ROI). B, Load to failure (N) for control and celiotomy samples. C, Load
to failure per ROI. Results are mean 6SD. *P<.05.
REINA RODRIGUEZ ET AL.
|
5
suture material and SBSI used for celiotomy closure. Four-
teen samples (n 514) were allocated to type of closure group
(n 514), with 7 samples (n 57) undergoing full-thickness
closure and the remaining samples (n 57) undergoing
fascia-only closure. The number of samples from each ROI
was variable between groups due to randomization.
3.3
|
Load to failure
SL:WL was >4 in all specimens. Load to failure (mean 6
SD) ranged from 11.60 to 121.35 N (45.34 623.81) among
the 48 samples. Load to failure for control samples ranged
from 33.38 to 121.35 N (68.88 63.51) and was consistently
highest in the PU region, lower in the U region, and lowest
in the POU region. Load to failure (N) was 37.93 65.27 for
samples closed with polydioxanone, 34.55 64.29 for sam-
ples closed with polyglactin 910, 42.05 65.45 with a 2 32
SBSI, and 3.43 63.41 with a 5 35 SBSI. Load to failure
(N) was 48.18 61.15 for full-thickness closure samples,
58.73 67.30 for fascia-only closure samples, 52.64 65.19
for samples obtained from males, and 35.95 63.25 for sam-
ples obtained from females. Load to failure was not affected
by suture material (P5.488), SBSI (P5.07), or type of clo-
sure (P5.533). Mean load to failure did not differ between
suture material and SBSI groups (P5.082). Load to failure
for gender, including neutering status, did not show signifi-
cant differences between males and females (P5.102).
However, an independent ttest comparing load to failure
between males and females without considering neutering
status showed lower load to failure in females compared
with males (P5.024). No significant differences were
observed in load to failure of samples obtained from the 4
female neutered cats included in the study. Data are summar-
ized in Table 1.
In the mixed model, no effect of suture material, SBSI,
or type of closure was observed in load to failure (Figure
3B). When these variables were removed from the statistical
model (stepwise regression analysis), anatomical location
showed significance as a dependent variable with decreased
load to failure in POU region samples (P5.049) compared
with PU (P5.864) and U (P5.064) regions. One-way
ANOVA for anatomical location excluding suture material,
SBSI, and type of closure showed decreased load to failure
of the POU region (P5.009) compared with the PU and the
U regions (Figure 3C). No statistical differences were
observed between the PU and U regions (P5.34). BW con-
tributed to 33.4% of variability in load to failure, and thick-
ness of linea alba contributed to 35.4%.
3.4
|
Type of failure
Type of failure was classified as (1) body wall failure distant
from linea alba or celiotomy site, (2) failure of body wall by
linear tears in muscle or fascia perpendicular to linea alba at
site of suture material penetration(s) with suture line in celiot-
omy remaining intact, and (3) tissue failure at suture line with-
out linear tears with loosening of suture line resulting in gap
between incised wound edges and with intact knots. All con-
trol group samples (n 56) failed by muscle tearing distant
from the linea alba (Figure 4A). Two 2 32 SBSI samples (1
polydioxanone and 1 polyglactin 910), one 5 35 polyglactin
910 sample, and 3 fascia-only closure samples failed by mus-
cle tearing distant from the linea alba, similarly to control sam-
ples. Linear tears in muscle or fascia at points of suture
penetrationoccurredin5ofseven232 polyglactin 910 sam-
ples, 6 of seven 2 32 polydioxanone samples, 3 of seven 5
35 polydioxanone samples, 4 of 7 fascia-only closure sam-
ples, and all 7 full-thickness closure samples (Figure 4B). Tis-
sue failure without linear tears with loosening of suture line
with knots intact occurred in 1 of seven 2 32 polyglactin 910
samples, 6 of seven 5 35polyglactin910samples,and4of
seven 5 35 polydioxanone samples (Figure 4C). Failure by
muscle or fascial tearing was most commonly observed in the
232 SBSI samples, whereas failure by loosening of suture
lineoccurredmainlyin535 SBSI samples, independent of
suture material used for closure.
4
|
DISCUSSION
The main findings of our study are that (1) the POU region of
the feline ventral abdominal wall was biomechanically weaker
than the PU and U regions, (2) the thickness of the linea alba
progressively decreased from cranial to caudal, (3) no differ-
ence was detected between incisions closed with polydiaxa-
none or polyglactin 910 suture materials or between full-
thickness and fascia-only closures, and (4) the only difference
identified between incisions consisting of small versus large
bites consisted of tissue failure in which suture loosening with-
out fascial tears was observed more frequently in the large bite
group.
In a mixed model linear analysis adjusted for BW and linea
alba thickness, the POU region sustained lower strain to fail-
ure, supporting our hypothesis that the POU region is biome-
chanically weaker than the PU and U regions. In dogs, the
internal leaf of rectus abdominis sheath disappears, and the
aponeurosis of the internal abdominal oblique muscle joins the
external leaf of the rectus abdominis sheath. This sheath is cov-
ered dorsally only by a thin continuation of transversalis fascia
and by the peritoneum in the caudal third of the abdominal
wall.
6
Similar anatomical findings in the cat may explain the
lower strain to failure for the POU region observed in our
study. Different characteristics and intrinsic properties of
abdominal wall muscles have been described experimentally
in dogs.
7
Similarities in domestic cats could contribute to dif-
ferences observed during tensile tests. To the best of our
6
|
REINA RODRIGUEZ ET AL.
TABLE 1 Data from 48 samples
a
Thickness linea
alba (mm) Load to failure (N) Strain to failure (%)
Stress at
failure (N/mm
2
) SL:WL
Group
Specimen
number
Body
weight
(kg) PU U POU PU U POU PU U POU PU U POU PU U POU
Control (n 56) 1 2.50 1.09 0.98 0.64 74.08 55.29 33.38 116.25 125.44 27.87 2.72 2.26 2.09
7 5.28 1.41 0.90 0.68 121.35 82.85 67.65 55.31 76.11 64.32 3.44 3.68 3.98
Suture
material
SBSI
232 Polyglactin
910 (n 57)
2 2.24 0.89 34.57 93.15 1.55 8.2
3 2.80 1.21 38.14 24.84 1.26 11.6
5 2.58 1.02 23.32 99.91 0.91 9.6
9 1.30 0.85 24.76 56.65 1.17 8.0
10 2.46 1.26 67.73 93.69 2.15 9.7
11 5.20 1.37 43.54 121 1.27 11.2
12 4.36 1.70 56.19 35.86 1.32 7.4
232 Polydioxanone
(n 57)
2 2.24 0.74 47.50 78.64 2.57 7.6
4 3.94 0.63 0.62 32.78 20.24 152.43 56.24 2.08 1.31 6.8 6.8
5 2.58 0.60 23.88 45.02 1.59 7.5
6 4.50 1.45 76.06 86.65 2.10 8.4
8 2.52 1.55 21.55 26.03 0.56 7.7
11 5.50 1.50 78.44 144.74 2.09 11.3
535 Polyglactin
910 (n 57)
2 2.24 0.89 43.33 42.64 1.95 4.3
3 2.8 0.82 0.70 22.20 14.74 35.26 20.28 1.08 0.84 5.2 5.7
4 3.94 0.71 24.84 24.81 1.40 6.4
5 2.58 0.91 30.37 142.00 1.34 6.1
9 1.30 0.60 11.60 34.56 0.77 7.4
(Continues)
REINA RODRIGUEZ ET AL.
|
7
TABLE 1 (Continued)
Thickness linea
alba (mm) Load to failure (N) Strain to failure (%)
Stress at
failure (N/mm
2
) SL:WL
Group
Specimen
number
Body
weight
(kg) PU U POU PU U POU PU U POU PU U POU PU U POU
10 2.46 1.12 48.34 47.77 1.73 6.7
535 Polydioxanone
(n 57)
6 4.50 1.04 0.71 43.93 27.31 48.61 109.9 1.69 1.54 6.2 4.2
8 2.52 0.75 0.62 22.96 17.57 54.90 94.83 1.22 1.13 4.3 4.8
9 1.30 0.66 28.24 63.38 1.71 6.3
10 2.46 1 39.31 62.14 1.57 5.7
11 5.20 1.98 51.28 50.65 1.04 6.6
Type of
closure
Full thickness
(n 57)
12 4.36 1.40 46.10 30.10 1.32
13 4.44 1.39 28.60 18.44 0.82
14 3.70 1.78 1.52 19.58 54.29 1.96 12.07 0.45 1.43
15 2.80 1.45 1.25 97.60 27.89 11.28 6.12 2.69 0.89
16 3.94 0.99 63.24 3.51 2.55
Fascia only
(n 57)
12 4.36 1.63 59.39 51.29 1.46
13 4.44 1.94 1.51 65.75 83.94 23.84 59.10 1.36 2.22
14 3.70 1.32 43.95 4.14 1.33
15 2.80 1.35 81.03 11.97 2.40
16 3.94 1.93 1.19 41.26 35.79 5.12 2.96 0.86 1.20
POU, postumbilical; PU, preumbilical; SBSI, suture bite to stitch interval; SL:WL, suture length to wound length ratio; U, umbilical.
a
Empty cell indicates data not recorded.
8
|
REINA RODRIGUEZ ET AL.
knowledge, this is the first study reporting the tensile strength
of the feline ventral abdominal wall. In control samples, the
POU region reached lower loads to failure compared with the
PU and U regions in the same animal. However, only 2 sam-
ples of each region with intact linea alba were tested, preclud-
ing statistical analysis and representing a limitation of our
study. All control samples failed distant to the linea alba,
highlighting the high intrinsic resistance to failure of intact
linea alba. In our study, the thickness of the linea alba
decreased gradually from cranial to caudal in all cats, being
thinner and biomechanically weaker in the POU region than
in the PU and U regions. In a human cadaveric study, fascial
thickness contributed 32% of variability in suture pull-out
force in the linea alba.
12
In our study, thickness of linea alba
accounted for 35.4% of variability in load to failure, suggest-
ing that linea thickness plays a role in strength of celiotomy
closure in cats.
No difference in breaking strength of celiotomy samples
was observed between the 2 suture materials (polydioxanone
and polyglactin 910) used in our study. There is no experimen-
tal or clinical evidence of a continuous suture pattern increas-
ing the risk of wound disruption.
2
Continuous suture patterns
distribute tension along the entire incisional length, decreasing
strain at individual suture bites.
13
Simple continuous closure of
the ventral abdominal wall is well accepted in dogs and
cats,
4,14
likely because it is faster than simple interrupted clo-
sure
4
and does not increase the risk of wound disruption.
3
We
did not detect differences in breaking strengths for celiotomies
closed with simple continuous fascia-only or full-thickness
bites in cats. To the best of our knowledge, fascia-only closure
has not been previously reported in the domestic cat. The
external leaf of the rectus abdominis sheath has been shown to
serve as the primary strength-holding layer for ventral parame-
dian incision closure in dogs. This concept is based on a lack
of difference in biomechanical properties of fascia-only and
full-thickness closure, which was attributed to incorporating
the external leaf in both groups.
9
In our study, fascia-only clo-
sure is biomechanically equivalent to full-thickness closure in
the cat, likely for the same reason. Fascia-only closure could
potentially reduce postoperative incisional swelling and pain in
feline patients after abdominal surgery. Inclusion of muscle in
sutured wounds is considered to increase pain during the post-
operative period.
2
SL:WL 4:1 has been recommended for median celiot-
omy closure in humans,
15-17
given that incisional herniation
has been found more common in wounds closed with SL:WL
<4:1.
16,17
A recent trend in human median abdominal closure
favors small bites over large bites while observing this mini-
mal SL:WL of 4:1.
18,19
Our study was designed to provide
SL:WL >4:1. We arbitrarily chose 2-mm and 5-mm bites,
representing small and large bite closure, respectively. Advan-
tages of small bite closure include reduced incidence of surgi-
cal site infection, better distribution of tension across the
celiotomy closure, and reduced incidence of incisional hernia-
tion.
17-19
Large SBSI may cause the suture to cut through or
compress soft tissues when tension is applied, making the
FIGURE 4 A, Failure distant from celiotomy closure in a 2 32 polydioxanone sample. B, Failure by muscle/fascial tearing in a 2 32polyglactin
910 sample. C,Failure by loosening of suture line with intactknots
REINA RODRIGUEZ ET AL.
|
9
stitch loosen and allowing wound edges to separate.
17,18
We
observed this pattern of failure most commonly in our 5 35
SBSI celiotomy groups. No differences in breaking strength
were observed in our celiotomy samples comparing the 2
SBSI used; however, evaluation of influence of SBSI in multi-
variable analysis and mean load to failure between suture
material/SBSI samples approached significance. A larger sam-
ple could have revealed significant differences between the 2
SBSI used.
In humans, sutures engaging the transition zone between
linea alba and rectus sheath markedly increase strength of
celiotomy closure.
20
In small animals, this transition zone
has not been described, and the distance between the linea
alba and the site of needle penetration during celiotomy clo-
sure is rarely reported
16
or is based on recommendations
rather than biomechanical testing.
4,8,21
We designed our
study to engage the transition zone in all celiotomies, meas-
uring suture bites from the lateral edge of the linea alba. In
this setting, no difference was detected between closures
incorporating small or large bites.
Muscle/fascial tearing was the most common type of fail-
ure observed for all celiotomy samples in our study, without
any knot failure. Instead, cracks initiated by stitches at sites of
needle penetration extended longitudinally toward the celiot-
omy. The biomechanical significance of this type of failure is
unknown but has been previously described in an equine
bursting model in which fascial failure was the main failure
mode observed for the 2 suture intervals used.
22
Thetypeof
needle used could have influenced this type of failure. How-
ever, we used a fine, atraumatic needle of consistent size. It is
unknown if different needle type or size could produce differ-
ent results.
Tensile testing of connective tissue is logistically difficult
because low friction at the grip–tissue interface often leads to
slippage prior to failure.
23
Cryoclamps have been used for ten-
sile testing of the linea alba of horses
24
because they allow bet-
ter gripping of tissue and increase friction between tissue and
grips. The custom-made clamps used in our study were previ-
ously used in a biomechanical study of muscles in other spe-
cies.
25
No slippage was reported in the Chism et al
25
study
with the use of dumbbell-shaped specimens, as recommended
for tensile testing of elastomers
10
and cheese grater clamps. No
slippage at tissue–grip interface was observed in our study.
Fresh feline cadavers were not available for our study.
However, we do not believe that freezing affected our results
because of its lack of influence on the breaking strength of the
canine linea alba.
26
In the Rath et al
26
study, samples of
abdominal wall were frozen after harvest, with muscle contrac-
tion caused by freezing artifact and more difficulties to secure
samples in grips reported. In our study, whole cadavers were
frozen and thawed intact prior to sampling and testing. We did
not observe changes in the abdominal wall or experience diffi-
culty in securing samples in grips. After thawing, samples
were collected and preserved in a saline bath under refrigera-
tion. This protocol, reported in another study with samples
kept up to 24 hours,
27
yields a state of mechanical stability
making the samples useful for analysis. Nonetheless, limita-
tions of our study remain inherent to its ex vivo design. Suture
degradation, surgical site infection, prolonged inflammation,
suture failure during cyclic and repetitive tension, and other
patient-related factors that contribute to incisional hernia in
vivo following ventral midline celiotomy closure
2,3,13,16,17,19
could not be evaluated in our ex vivo study. We also tested
discrete segments of ventral abdominal wall, providing an idea
of the dynamic properties of the 3 regions tested and their bio-
mechanical properties but making extrapolations of results to
the entire ventral abdominal wall impossible. Another limita-
tion relates to the method of tensile testing. We used uniaxial
loading in a single test to failure with continuous distraction.
This design differs from the clinical scenario, in which strains
are cyclic and dependent on the animal’smovementduringthe
postoperative period. Cyclic strain on an incision may result in
different results compared with continuous distraction. We
chose a low speed of distraction (0.5 mm/second, 30 mm/
minute) for our study because of the viscoelastic nature of soft
tissue specimens. In previous studies, a different speed of dis-
traction was used,
12,24,26
making direct comparisons of results
impossible. In addition, 2 types of forces act on the ventral
abdominal wall, intra-abdominal pressure acting on the deep
aspect and linear traction exerted by flank muscles.
27
We eval-
uated elasticity and elongation in discrete segments to mimic
linear traction of flank muscles. However, the interaction of
surrounding muscles of the entire ventral abdominal wall was
not evaluated. Different results may be obtained by testing
bursting strength of the abdomen and celiotomy closure techni-
ques and by evaluating tensile forces in the entire ventral
abdominal wall. A combination of tensile and bursting tests
may be more representative of forces that celiotomy closure
encounters after surgery in vivo. Another limitation is the small
number of specimens used and samples tested for variables
used in the statistical model. Due to randomization, the number
of samples from each ROI and the specimen where the sam-
ples were obtained varied among groups. A minimum of 20
samples per ROI (equivalent to 20 cats) would have been
required to obtain 80% statistical power. The coefficient of var-
iation between the different specimens tested represents
another limitation of the study. Biological tissue has a high
degree of natural variation, both interindividual and intraindi-
vidual. A large number of tests and a uniform sample popula-
tion are required to extrapolate results to the clinical scenario.
Although ex vivo studies can aid in determining adequate
suture techniques for celiotomy closure in domestic cats, pro-
spective studies evaluating the 2 SBSI that we used or other
SBSI combinations and complications associated with their
use, including incidence of incisional herniation, would be
10
|
REINA RODRIGUEZ ET AL.
required to determine optimal technique for celiotomy closure
in this species.
In conclusion, the POU region of ventral abdominal wall
in domestic cats is biomechanically weaker than the PU and
U regions under uniaxial loading in this model. Considering
the biomechanical results and types of failure observed in
our study, we favor simple continuous closure using 2 32-
mm SBSI and including the external leaf of rectus abdominis
muscle sheath in cats. However, additional biomechanical
and clinical studies are warranted before specific celiotomy
closure recommendations can be made.
CONFLICT OF INTEREST
The authors declare no conflicts of interest related to this
report.
ORCID
Joshua Milgram DVM, BVSc, DECVS http://orcid.org/
0000-0002-5949-5939
REFERENCES
[1] Boothe HW, Skater MR, Hobson HP, Fossum TW, Jung C.
Exploratory celiotomy in 200 nontraumatized dogs and cats. Vet
Surg. 1992;21:452-457.
[2] De Rooster H. Celiotomy. In: Griffon D, Hamaide A, eds. Com-
plications in Small Animal Surgery. 1st ed. Hoboken, NJ:
Wiley-Blackwell; 2016:355-361.
[3] Claeys S. Dehiscence. In: Griffon D, Hamaide A, eds. Compli-
cations in Small Animal Surgery. 1st ed. Hoboken, NJ:
Wiley-Blackwell; 2016:57-63.
[4] Crowe DT. Closure of abdominal incisions using a continuous
polypropylene suture: clinical experience in 550 dogs and cats.
Vet Surg. 1978;7:74-77.
[5] Gower SB, Weisse CW, Brown DC. Major abdominal eviscera-
tion injuries in dogs and cats. J Am Vet Med Assoc. 2009;234:
1566-1572.
[6] Evans HE, de Lahunta A. Miller’s Anatomy of the Dog. 4th ed.
St. Louis, MO, Elsevier Saunders; 2013:224-227.
[7] Farkas GA, Rochester DF. Characteristics and functional signifi-
cance of canine abdominal muscles. J Appl Physiol. 1998;65:
2427-2433.
[8] Bellenger CR. Abdominal wall. In: Slatter D 3rd, ed. Textbook
of Small Animal Surgery. Philadelphia, PA: Elsevier Saunders;
2003:405-413.
[9] Rosin E, Richardson S. Effect of fascial closure technique on
strength of healing abdominal incision in the dog: a biomechani-
cal study. Vet Surg. 1987;16:269-272.
[10] Davis JR. Tensile Testing. 2nd ed. Novelty, OH: ASM Interna-
tional; 2004:161-164.
[11] Rosin E, Robinson GM. Knot security of suture materials. Vet
Surg. 1989;18:269-273.
[12] Campbell JA, Temple WJ, Frank CB, Huchcroft SA. A biomechani-
cal study of suture pullout in linea alba. Surgery. 1989;106:888-892.
[13] Poole GV. Mechanical factors in abdominal wound closure: the
prevention of fascial dehiscence. Surgery. 1985;97:631-639.
[14] Runk A, Allen SW, Mahaffey A. Tissue reactivity to poliglecap-
rone 25 in the feline linea alba. Vet Surg. 1999;28:466-471.
[15] Jenkins TP. The burst abdominal wound: a mechanical approach.
Br J Surg. 1976;63:873-876.
[16] Israelsson LA, Johnson T. Suture length to wound length ratio
and healing of midline laparotomy incisions. Br J Surg. 1993;
80:1284-1286.
[17] Israelsson LA, Millbourn D. Prevention of incisional hernias:
how to close a midline incision. Surg Clin North Am. 2013;93:
1027-104.
[18] Deerenberg EB, Harlaar JJ, Steyerberg EW, Lont HE, van Doorn
HC, Heisterkamp J. Small bites versus large bites for closure of
abdominal midline incisions (STITCH): a double-blind, multi-
center, randomized controlled trial. Lancet. 2015;386:1254-126.
[19] Millbourn D, Cengiz Y, Israelsson LA. Effect of stitch length on
wound complications after closure of midline incisions: a
randomized controlled trial. Arch Surg. 2009;144:1056-1059.
[20] Tera H, Aberg C. Tissue strength of structures involved in
musculo-aponeurotic layer sutures in laparotomy incisions. Acta
Chir Scand. 1976;142:349-355.
[21] Freeman LJ, Pettit GD, Robinette JD, Lincoln JD, Person MW.
Tissue reaction to suture material in the feline linea alba: a retro-
spective, prospective, and histologic study. Vet Surg. 1987;16:
440-445.
[22] Hassan KA, Galuppo LD, van Hoogmoed LM. An in vitro com-
parison of 2 suture intervals using braided absorbable loop
suture in the equine linea alba. Vet Surg. 2006;35:310-314.
[23] Bowser JE, Elder SH, Rashmir-Raven AM, Swiderski CE. A
cryogenic clamping technique that facilitates ultimate tensile
strength determinations in tendons and ligaments. Vet Comp
Orthop Traumatol. 2011;24:370-373.
[24] Chism PN, Latimer FG, Patton CS, Rohrbach BW, Blackford
JT. Tissue strength and wound morphology of the equine linea
alba after ventral median celiotomy. Vet Surg. 2000;29:145-151.
[25] Mohammadkhah M, Murphy P, Simms CK. The in vitro passive
elastic response of chicken pectoralis muscle to applied tensile
and compressive deformation. J Mech Behav Biomed Mater.
2016;62:468-48.
[26] Kirpensteijn J, Fingland RB, Boyer JE, Kennedy GA, Klemm
RD, Debowes RM. Comparison of stainless steel fascial staples
and polypropylene suture material for closure of the linea alba in
dogs. Vet Surg. 1993;22:464-472.
[27] Rath AM, Attali P, Dumas JL, Goldlust D, Zhang J, Chevrel JP.
The abdominal linea alba: an anatomo-radiologic and biome-
chanical study. Surg Radiol Anat. 1996;18:281-288.
How to cite this article: Reina Rodriguez FS, Buckley
CT, Milgram J, Kirby BM. Biomechanical properties
of feline ventral abdominal wall and celiotomy closure
techniques. Veterinary Surgery. 2017;00:1-11. https://
doi.org/10.1111/vsu.12751
REINA RODRIGUEZ ET AL.
|
11