Irrigation is an integral step in the management of soft-tissue injuries
and open fractures, typically following débridement of the injured soft
tissues. Irrigation is often a dogmatic step in initial wound management, yet
several factors including fluid type, fluid volume, and delivery method must
be considered prior to wound irrigation. Recommended fluid volumes have ranged
from 6 to 10
L1-3.
With regard to fluid types, normal saline solution is the standard to which
all other irrigants have been
compared1,4-8.
Additives such as
antibiotics7-14,
surfactants6,
soaps15, and
antiseptics1,7
have all been investigated with mixed results.
Device comparisons have been performed, with bulb syringe irrigation being
the usual standard to which gravity
flow16,17,
pulsed lavage18,
and other high-pressure lavage devices are
compared1,19.
While pulsed lavage is perhaps the most common device used in irrigation, it
has been implicated as causing foreign bodies and bacteria to be driven
further into the
tissues3,20,21,
and it may result in injury to chondrocytes and
osteoblasts22-25.
The superiority of high-pressure devices (e.g., pulsed lavage) over
low-pressure devices (e.g., bulb syringe) is not conclusive and may depend on
different factors such as the degree of initial wound
contamination26.
The definition of what exactly constitutes high and low pressure is not well
established in the literature. However, it is generally accepted that low
pressure is <15 psi (103.4 kPa) and high pressure is >35 psi (241.3
kPa)21.
The so-called gold-standard method of measuring efficacy has been
quantitative culture, which is a consumptive, tissue-destructive process. Loss
of tissue can be minimized by taking fewer and smaller biopsy samples, but
this limits the culture yield. In addition, quantitative culture samples only
a small portion of the wound and may omit a large area of localized
contamination or infection.
The ability to determine the distribution and quantity of bacteria in a
wound is desirable, especially for the purpose of improving techniques and
equipment for irrigation. Transgenic bacteria have been developed with the
gene for luminescence spliced into their own genome. These bioluminescent
bacteria produce light that can be detected with a high-sensitivity
photon-counting camera. Therefore, the relative quantity and distribution of
the bacteria in the wound can be determined.
The purpose of this study was to determine whether pulsed lavage irrigation
is more effective than bulb syringe irrigation in removing bacteria from a
contaminated wound in a large-animal model. We hypothesized that pulsed
lavage, because of the higher pressures it generates, would remove more of the
bacteria from the wound than would bulb syringe irrigation.
General
All procedures were performed in a laboratory accredited by the Association
for Assessment and Accreditation of Laboratory Animal Care after approval of
the protocol was obtained from the Institutional Animal Care and Use
Committee.
Bacterial Preparation
Pseudomonas aeruginosa (ATCC 27317) was genetically engineered to
be luminescent by random chromosomal insertion of the luciferase-luciferin
construct luxCDABE obtained from Photorhabdus luminescens, a nematode
symbiont
bacterium27. With
the lux gene incorporated into the organism, it was then referred to as
Pseudomonas aeruginosa (lux). This bioluminescence became a stable,
heritable genetic trait. For each inoculation procedure, the stock culture was
grown up over eighteen to twenty hours and was diluted to 108
colony-forming units (CFU)/mL in 0.85% NaCl. The concentration was confirmed
by performing a plate count with use of the spread plate technique.
Surgical Procedure—Wound Creation
Twelve castrated, adult male Spanish-Boer goats (Capra hircus) (Talley
Ranch, Uvalde, Texas) were fasted for twenty-four hours, and water was
withheld for twelve hours prior to surgery. Anesthesia was induced with a
combination of ketamine hydrochloride (2.2 to 7.0 mg/kg) and midazolam (0.125
to 0.250 mg/kg) administered intravenously through a 21-gauge needle. After
endotracheal intubation was done, anesthesia was maintained with isoflurane
and supplemental oxygen. An epidural injection of morphine (0.1 mg/kg) diluted
in 0.9% sterile saline solution to a volume of 0.13 mL/kg was given both as an
adjunct to general anesthesia and for its durable postoperative analgesic
effect.
The goat was placed supine on the operating table, and the left lower
extremity was shaved and aseptically prepared. The limb was then draped free
with a sterile stockinette over the hoof and lower leg. The tibial tubercle
was marked, and a 5-cm skin incision approximately 1 cm lateral to, and at the
level of, the tubercle was made and extended distally to the level of the
medial periosteum and fascia overlying the anterior and lateral leg
compartments. The lateral compartment was elevated from its attachment to the
lateral aspect of the tibia after the lateral compartment fascia was incised
with use of electrocautery. The fascia was also elevated from the superficial
surface of the anterior and lateral compartments. The medial tibial periosteum
was exposed and incised longitudinally throughout the length of the skin
incision and parallel to the incision in the anterior compartment fascia. This
incision was measured so that a 6-mm strip of periosteum was left intact on
the anteromedial aspect of the tibia. The more posterior portion of periosteum
was elevated with a blunt periosteal elevator and retracted medially. A
partial medial cortical injury measuring 1.2 cm in diameter was created in the
tibia with use of a 3-mm drill-bit on a twist drill and a small osteotome.
Care was taken to avoid breaching the cortical wall and entering the medullary
canal. Three Kelly clamps were spaced evenly over a 5-cm segment of the
anterior compartment muscles and were left in place for three minutes to
induce a standardized crush injury to the anterior compartment muscles.
Concurrently, electrocautery was used to create thermal damage to the
intervening muscle between the clamps. Thus, the wound rendered was complex,
involving injury to muscle, fascia, periosteum, and bone
(Fig. 1).
The wound was inoculated with 1 mL of >108 CFU/mL of
Pseudomonas aeruginosa (lux), which was spread evenly over the wound
surfaces with a cotton-tipped applicator soaked in the same inoculum. The
wound was left open for a five-minute period after which it was dressed open
with a cover sponge, a rolled gauze dressing, and Vetrap bandaging tape (3M
Animal Care Products, St. Paul, Minnesota).
Postoperative Care
After surgery, the goats recovered in their pens and were allowed activity
ad libitum. If an animal demonstrated any discomfort, a fentanyl citrate patch
(Duragesic 50) was placed on the neck area to control postoperative pain. The
goats were killed six hours after inoculation with a concentrated solution of
pentobarbital sodium (90 mg/kg administered intravenously). Six hours was
chosen as the incubation time because luminescence at six hours was visible
and measurable by the camera and the removal of bacteria by means of
irrigation was still possible. In addition, the current standard of care
dictates that the initial débridement and irrigation should be
performed within a few hours after the
injury3. Moreover,
in a recent clinical trial, the average interval between the time that an open
fracture was sustained to the time of irrigation was approximately six
hours8.
Imaging Procedure
Bioluminescent bacteria emit light in proportion to their
number28,29.
This allowed the use of a specialized imaging system contained in a light-free
enclosure to quantify the amount of bacteria in the wound. A photon-counting
camera (Charge Couple Device [CCD] Imaging System Model C2400; Hamamatsu
Photonics, Hamamatsu-City, Japan), which recognizes the event created by a
single photon hitting the photocathode of an image intensifier, was used. By
accumulating many images containing binary photon information, a luminescent
image is generated. Superimposition of this image onto a gray-scale background
image yields information on the location and intensity in terms of photon
number. The camera was connected to a computer system with Windows 98
(Microsoft, Redmond, Washington) through an image processor (model C5510,
Argus-20; Hamamatsu Photonics). Argus-20 Interface software (version 1.10;
Hamamatsu Photonics) and AquaCosmos basic software (version 1.30; Hamamatsu
Photonics) were used to acquire images and to process the image data
collected. Each day, prior to imaging, a background noise image was made to
ensure that background light was kept below a known minimum level.
The goat was placed supine on the operating table within the light-free
enclosure. The camera was placed directly over the wound, and the leg was
secured to the camera apparatus through an adapted external fixation system
(Synthes, Paoli, Pennsylvania) with use of two 5-mm external fixator pins
placed as far proximally and distally as possible from the wound location. The
hip and knee joints were flexed to 90° in the imaging apparatus so that
the wound surface was 9 cm directly beneath the camera lens. Gelpi retractors
were placed at the proximal and distal wound margins.
A black-and-white image of the wound was made, and this was followed by a
photon count of the same region. This entire wound photon count was quantified
as relative luminescent units and was displayed in a pseudospectrum ranging
from red (most intense) to blue (least intense), with black representing no
photon detection. At this time, locations in the periosteum, fascia, and
anterior muscle compartment with the most photon detection were identified.
Each of the three regions was marked by a surgical staple (Visistat; Weck
Closure Systems, Research Triangle Park, North Carolina), and an additional
sequence of black-and-white and photon images was taken to confirm staple
placement. A 6-mm-square soft-tissue biopsy sample was taken from above each
staple with use of a number-11 scalpel blade (Bard-Parker; Becton Dickinson,
Franklin Lakes, New Jersey). In the case of the fascia and periosteum,
full-thickness samples of each tissue were harvested. Approximately 4-mm-thick
sections of the muscle were harvested for quantitative cultures.
Irrigation Sequence
The goats were randomly assigned to either the group that had irrigation
with the bulb syringe (Kendall, Mansfield, Massachusetts) or the group that
had pulsed lavage (InterPulse irrigation system; Stryker Instruments,
Kalamazoo, Michigan) such that both groups consisted of six goats. Each wound
was then sequentially irrigated with 3 L of normal saline solution with the
selected device for a total of 9 L. Device tips were held 1 cm from the wound
surface, and the pulsed lavage device was operated at its highest setting. The
pulsed lavage system used a high-flow tip attachment (model 210-14) with a
maximum pressure of 19 psi (131 kPa) (minimum, 6 psi [41.4 kPa]) and maximum
flow rate of 1025 mL/min. After each 3-L increment, both a black-and-white and
a photon-count image were made. After images were made of the condition
following irrigation with 9 L of saline solution, tissue samples were
harvested from each of the three locations inferior to the previously placed
surgical staple for quantitative cultures.
Data Analysis
Raw data were collected in the form of relative luminescent units generated
by the charge-coupled device camera and image processor as well as from
quantitative cultures. All data were saved in Excel XP software (Microsoft).
For the relative luminescent units, the AquaCosmos imaging software provided a
count of relative luminescent units for the entire field within view of the
camera. The quantitative culture results were expressed in terms of CFU/mL and
were converted to CFU/g of tissue.
Scatter plots were created of relative luminescent units compared with the
average CFU/g of the periosteum, fascia, and muscle. Linear regression was
performed, and correlation coefficients were calculated. Ratios of relative
luminescent units after each volume of irrigation compared with the baseline
relative luminescent units were calculated to include standard errors of the
mean. A pre hoc power analysis demonstrated that six goats per group were
required to achieve beta of 0.80 and an ability to identify a 25% difference
between the bulb syringe and pulsed lavage groups with respect to the ratio of
the relative luminescent units after irrigation to the baseline relative
luminescent units. Alpha was set at 0.05, and data are presented as the mean
and the standard error of the mean. All ratios between the bulb syringe and
pulsed lavage groups were analyzed with use of a hierarchical mixed-model
analysis of variance. Post hoc testing was performed with t tests for further
within-group comparisons of ratios with use of the Bonferroni correction. SAS
statistical software (SAS Institute, Cary, North Carolina) was used for all
statistical calculations.
Pulsed lavage decreased bacterial luminescence more than bulb syringe
irrigation with use of 6 and 9 L of saline solution (p = 0.04), but no
difference was found after 3 L of irrigation (p = 0.06)
(Table I). With 9 L of
irrigation, bulb syringe irrigation reduced the bioluminescence by 51%,
whereas a 70% reduction was seen with use of pulsed lavage. With quantitative
cultures, similar decrements were observed after 9 L of irrigation (Tables
II and
III). A within-group comparison
of the percentage of relative luminescent units reduced following the various
volumes of irrigation showed significant decreases (p = 0.002) in the ratio
of the remaining luminescence between the two groups at all irrigation volumes
tested (Fig. 2).
The correlation coefficients for the relative luminescent units compared
with the average CFU/g of the periosteum, muscle, and fascia were 0.96, 0.83,
and 0.92 for the condition before irrigation, the condition after irrigation,
and the pooled data, respectively. Qualitative analysis indicated that the
bacterial luminescence was not homogeneous throughout the wound
(Fig. 3). However, there was
not a distinguishable pattern regarding which area or tissue within the wound
had the highest amount of bacterial luminescence.
Anovel model of a contaminated orthopaedic wound was created to test the
hypothesis that irrigation with pulsed lavage would be more effective than
bulb syringe irrigation in removing bacteria from a contaminated wound. The
use of bioluminescent bacteria allowed the changes in quantity and
distribution of bacteria in the entire wound to be visualized and quantified
noninvasively over multiple volumes of irrigant.
The use of bioluminescent bacteria as a research tool has a relatively
brief but successful history. Bacterial bioluminescence and bacterial quantity
had correlation coefficients of 0.98 and 0.99 in a mouse foreign-body model
and a mouse hind-limb infection
model28,29.
The previous studies with use of bioluminescent bacteria to create
contaminated wounds either consumed the entire sample of tissue after
imaging28 or
utilized an implanted foreign
body29. The overall
relative luminescent units compared with bacterial count correlation
coefficient from the current study (0.92) compares extremely well with the
correlation coefficients reported for the other biologic models, especially
considering that the current musculoskeletal wound imaged was more complex
than that evaluated in the mouse models.
Many studies have compared bulb irrigation and pulsed lavage irrigation;
however, none have provided conclusive proof of the superiority of pulsed
lavage over bulb syringe irrigation in a contaminated musculoskeletal wound
with use of a pulsed lavage device that produces pressures similar to those
routinely used in the treatment of extremity injuries. In an in vitro model of
stainless-steel screws coated with slime-producing bacteria, Anglen et
al.4 subjected the
screws to either bulb irrigation or pulsed lavage irrigation. After irrigation
with 1 L of normal saline solution, the decrease in bacterial number was 100
times greater with pulsed lavage than with bulb syringe irrigation. This is a
compelling difference; however, it did not assess bacterial clearance in an
actual wound. In another study, a small-animal model of a paraspinal
contaminated wound demonstrated that pulsed lavage removed more bacteria than
did bulb syringe
irrigation16. The
pulsed lavage used in that study, a 50 psi (345 kPa) Water Pik device (Water
Pik Technologies, Newport Beach, California), produced more pressure than that
available in the commercial device that we tested and that is commonly used to
irrigate wounds. In a study with a similar rat model, Hamer et
al.30 compared
gravity flow, bulb syringe, and pulsed lavage in wounds contaminated with both
Staphylococcus aureus and Escherichia coli. The authors
concluded that pulsed lavage reduced bacterial culture counts and clinical
infection rates more than gravity flow and bulb syringe irrigation. It should
be noted that the pulsed lavage device used in their study also produced 50
psi (345 kPa). Unfortunately, the results from these studies cannot be
directly extrapolated to clinical wound care either because the models used
were not relevant to trauma-related musculoskeletal injuries that are
routinely débrided and irrigated or because the pulsed lavage devices
used in the studies generated pressures well above what is routinely used in
the operating room during irrigation (approximately =20 psi [=137.9 kPa]
with pulsed lavage irrigation and approximately 1 psi [6.9 kPa] for bulb
syringe irrigation).
Both a volume effect and a device effect are evident when bacterial
luminescence counts are compared before irrigation and after irrigation. The
volume effect is demonstrated by the observation that there is a significant
difference in relative luminescent units between each volume within each
device group (p = 0.02). However, it appears that the amount of reduction
in bacterial luminescence diminishes with each 3-L iteration
(Fig. 2). This volume effect
has been seen clinically as well. Gustilo and
Anderson31 observed
that rates of infection in open fractures decreased with increasing volumes of
irrigation. Conversely, in an in vitro model, Gainor et
al.32 noted no
significant decrease in bacterial counts after irrigation with 100 mL to 1 L
of normal saline solution or after irrigation with 1 to 10 L. In that study,
bacteria were placed on strips of beef and were allowed fifteen minutes of
incubation before irrigation with a Water Pik (Water Pik Technologies). The
lack of volume effect seen in that study may be attributed to two factors: (1)
the short incubation time, which did not allow for bacterial adhesion, and (2)
the high-pressure pulsatile device. The six-hour incubation period that was
chosen for our study more closely mimics the current standard of care for this
type of injury.
In some scenarios, the appropriate irrigation volume to be used may depend
not only on the severity and contamination of the wound (which would dictate
that high volumes be used) but also on the resources available. The possible
need to ration available irrigant solutions while caring for combat
casualties18 or
domestic mass casualties only underscores the practical need to understand the
effectiveness of different irrigation techniques over various volumes. In this
light, the device effect becomes more significant as demonstrated by the
observation that the pulsed lavage reduced bacterial luminescence
significantly more than the bulb syringe at volumes of 6 L and 9 L (p =
0.04). After 3 L of irrigation, the pulsed lavage decreased luminescence by
52% and bulb syringe irrigation decreased it by 33% of baseline. With only six
animals in each group, these amounts were not significantly different (p =
0.06). The first 3 L of pulsed lavage irrigation and 9 L of bulb syringe
irrigation reduced luminescence by 52% and 51%, respectively
(Table I). The similar
reduction in bacteria implies that pulsed lavage irrigation at lower volumes
may be as effective as higher volumes with use of the bulb syringe.
Considering that resources are limited in austere environments, such as
forward surgical teams on the battlefield, these data suggest that pulsed
lavage devices may be more desirable despite the fact that they weigh more and
are more expensive than bulb syringes.
Qualitatively, this model provides data regarding the spatial distribution
of bacteria not previously possible with quantitative cultures
(Fig. 3). The distribution of
luminescence was not homogeneous, although the bacteria were evenly applied
initially over the entire surface area of the wound. A pattern could not be
discerned to describe which type of tissue allowed the most bacterial
adherence or best growth. This asymmetric distribution supports the notion
that typical quantitative sampling methods have an inherent degree of error,
rendering their culture results less reliable. This finding should lend
caution to the conclusions provided by studies that use quantitative cultures
solely to assess bacterial load.
The current study differs from the standard of care for contaminated wounds
because it did not include débridement, and this may be considered a
weakness. We support débridement as an essential step in open wound
management; however, its inclusion in this study would have added additional
procedural variability. Although the predominant bacterial species present
during initial wound management are generally gram-positive, gram-negative
bacteria, such as Pseudomonas aeruginosa, are also present up to 33%
of the time at the initial débridement of open
fractures31. During
model development, we attempted to use gram-positive strains, but the bacteria
did not emit enough photons to capture using our current camera system. Care
should be taken in extrapolating the current results to conditions involving
other organisms until other bacterial species have been used in the model.
Another limitation of our study is that it was acute in nature and did not
follow contaminated wounds for any time after the initial irrigation to
evaluate for signs of clinical infection. It is possible that the pulsed
lavage, due to the higher pressure (19 psi [131 kPa]), may have propelled the
bacteria into the tissue, perhaps resulting in an increased rate of clinical
infection. The same pulsed lavage device has been shown to drive bacteria into
soft tissue in an ex vivo
model21. Pulsed
lavage could have also increased local inflammation or destroyed viable
tissue; neither was assessed in the present study. Ongoing studies at our
institution with use of bioluminescent gram-positive organisms, as well as
survival studies that allow the wounds to be followed for clinical signs of
infection, are planned. Further study is also ongoing with regard to the
evaluation of irrigant types and additives as well as device prototypes.
Transgenic bioluminescent bacteria provide a powerful tool for the
noninvasive study of extremity wound injuries. Irrigation with pulsed lavage
is more effective than irrigation with a bulb syringe in this large-animal
model of a complex, contaminated musculoskeletal wound and may allow smaller
volumes of irrigant to be used. ?