ORIGINAL ARTICLES

Topical Application of a Gentamicin-Collagen Sponge

Combined with Systemic Antibiotic Therapy for the
Treatment of Diabetic Foot Infections of Moderate Severity
A Randomized, Controlled, Multicenter Clinical Trial
Benjamin A. Lipsky, MD*
Michael Kuss, BS
Michael Edmonds, MD
Alexander Reyzelman, DPM
Felix Sigal, DPM

Background: The aim of this pilot study was to determine the safety and potential
benefit of adding a topical gentamicin-collagen sponge to standard of care (systemic
antibiotic therapy plus standard diabetic wound management) for treating diabetic foot
infections of moderate severity.
Methods: We randomized 56 patients with moderately infected diabetic foot ulcers in a
2:1 ratio to receive standard of care plus the gentamicin-collagen sponge (treatment
group, n 38) or standard of care only (control group, n 18) for up to 28 days of
treatment. Investigators performed clinical, microbiological, and safety assessments at
regularly scheduled intervals and collected pharmacokinetic samples from patients
treated with the gentamicin-collagen sponge. Test of cure was clinically assessed 14
days after all antibiotic therapy was stopped.
Results: On treatment day 7, we noted clinical cure in no treatment patients and three
control patients (P .017). However, for evaluable patients at the test-of-cure visit, the
treatment group had a significantly higher proportion of patients with clinical cure than
did the control group (22 of 22 [100.0%] versus 7 of 10 [70.0%]; P .024). Patients in the
treatment group also had a higher rate of eradication of baseline pathogens at all visits
(P  .038) and a reduced time to pathogen eradication (P , .001). Safety data were
similar for both groups.
Conclusions: Topical application of the gentamicin-collagen sponge seems safe and may
improve clinical and microbiological outcomes of diabetic foot infections of moderate severity
when combined with standard of care. These pilot data suggest that a larger trial of this
treatment is warranted. (J Am Podiatr Med Assoc 102(3): 223-232, 2012)

Patients with diabetes are at high risk for foot
ulcerations, which often become infected. These
diabetic foot infections can cause substantial
morbidity because of local pain and tissue destruc-
*Veterans Affairs Puget Sound Health Care System,
University of Washington School of Medicine, Seattle, WA.
Premier Research Group Ltd, Austin, TX.
Kings College Hospital, London, UK.
Center for Clinical Research, Castro Valley, CA.
Pacific Clinical Center, Los Angeles, CA.
ClinicalTrials.gov identifier: NCT 00659646
Corresponding author: Benjamin A. Lipsky, MD, Veterans
Affairs Puget Sound Health Care System, S-111-PCC, 1660 S
Columbian Way, Seattle, WA 98116-1597. (E-mail: balipsky@
uw.edu)
tion, as well as impaired wound healing, and
sometimes lead to lower-extremity amputations.
The validated Infectious Diseases Society of Amer-
ica guideline for diabetic foot infections1 classifies
their clinical severity as mild, moderate, or severe
depending on the size and depth of the infection and
whether there are systemic manifestations or
metabolic perturbations. The guidelines recommen-
dations for selecting empirical antibiotic regimens
are based on this classification. For mild infections,
topical antimicrobial agents may be appropriate.
One randomized controlled trial2 in patients with
mild diabetic foot infections showed that treatment
with a topical antimicrobial peptide (pexiganan)

Journal of the American Podiatric Medical Association  Vol 102  No 3  May/June 2012 223
produced clinical outcomes similar to those of an
oral antibiotic (ofloxacin). However, there are no
topical antimicrobials approved for treating diabetic
foot infections, and their role in the treatment of
chronic wounds remains controversial.3 For most
diabetic foot infections, systemic antibiotic drug
therapy, either oral or parenteral, is prescribed.
However, the need for better treatment approaches
is highlighted by reported clinical failure rates
greater than 20%.1
One potential approach is to use standard
systemic therapy combined with an effective topical
delivery system for ensuring high local antibiotic
concentrations at the site of infection. We are aware
of no other randomized controlled trials investigat-
ing any such combination for treating diabetic foot
infections.
The gentamicin-collagen sponge (Innocoll Tech-
nologies Ltd, Athlone, Ireland) is a promising
antibiotic delivery system for the management of
diabetic foot infections4 that is composed of
gentamicin sulfate, a water-soluble broad-spectrum
aminoglycoside antibiotic, uniformly dispersed in a
type I collagen matrix. Gentamicin provides con-
centration-dependent bactericidal activity in vitro
against most strains of aerobic gram-negative and
gram-positive and facultative anaerobic gram-nega-
tive pathogens found in diabetic foot infections,5
including methicillin-resistant Staphylococcus au-
reus.6 Historically, there have been concerns that
the topical application of gentamicin cream to skin
ulcers may cause or propagate resistance to this
agent,7 but one recent study8 of ocular isolates
found no increased resistance to gentamicin com-
pared with other commonly used ocular antibiotics.
Of note is that the creams and ointments indicated
for skin infections contain only 0.1% gentamicin,
whereas the gentamicin-collagen sponge contains
27% gentamicin. This high dose is intended to
reduce the likelihood of resistance propagation
and to increase its efficacy.
The gentamicin-collagen sponge is approved for
use in Europe as a biodegradable surgical implant
for the adjuvant treatment of localized bone or soft-
tissue infections. The product has also been used
topically for the treatment of chronic leg ulcers.9
Topical (as opposed to implantable) administration
allows for repeated, rather than single, applications,
and should theoretically, maintain high concentra-
tions of gentamicin in the wound environment over
the course of treatment while avoiding potentially
serious toxic effects associated with systemic
administration. This trial was designed to determine
whether adding daily application of the gentamicin-
224 May/June 2012  Vol 102  No 3  Journal of the American Podiatric Medical Association
collagen sponge to standard of care would improve
resolution of infection in patients with diabetic foot
infections of moderate severity.
Research Design and Methods
Eligible individuals for this randomized, controlled,
open-label, multicenter pilot study were diabetic
patients aged 18 to 80 years with a single,
moderately infected lower-extremity ulcer. We
defined a moderately infected ulcer by the Infec-
tious Diseases Society of America guideline criteria,
ie, a patient who is systemically well and metabol-
ically stable with a foot wound showing two or
more manifestations of inflammation (purulence or
erythema, pain, tenderness, warmth, or induration)
and one or more of the following characteristics:
cellulitis extending 2 cm, lymphangitic streaking,
spread beneath the superficial fascia, deep tissue
abscess, gangrene, and involvement of muscle,
tendon, joint, or bone.1 We graded the ulcers using
the Lipsky Wound Scoring Scale,10 and we excluded
patients with an ulcer that could not be completely
covered with a 10 3 10-cm gentamicin-collagen
sponge. We also excluded patients who had
received any antimicrobial therapy in the preceding
2 weeks and those with ischemia of the lower limb
(defined as an ankle-brachial index ,0.7 or .1.3
unless they had either transcutaneous oxygen
pressure or toe pressure 40 mm Hg) and, at
institutional review board request, patients with a
hemoglobin A1c level of at least 10.0%.
The study compared daily topical application of
the gentamicin-collagen sponge combined with
systemic antibiotic therapy (a daily oral or intrave-
nous dose of 750 mg of levofloxacin or alternative
antimicrobial therapy, as determined by suscepti-
bility testing) with systemic antibiotic therapy
alone. We selected levofloxacin as the first-choice
systemic antibiotic based on its proven efficacy in
complicated skin and soft-tissue (including diabetic
foot) infections,11 its listing among the recommend-
ed agents in the Infectious Diseases Society of
America guidelines,1 and the convenience of its
once-daily dosing. Two different sizes of gentami-
cin-collagen sponge were provided to investigators:
5 3 5 cm with 50 mg of gentamicin sulfate
(equivalent to 32.5 mg of gentamicin base) and 10
3 10 cm with 200 mg of gentamicin sulfate
(equivalent to 130 mg of gentamicin base). The
larger size was applied only to patients whose
ulcers could not be completely covered with the
smaller size, ensuring that the dose of gentamicin
per unit area of wound surface was more uniform
across patients. We chose not to administer placebo
collagen sponges to patients in the control group
because of the concern that a placebo sponge could
potentially harbor bacteria and bias the results in
favor of the active treatment. Therefore, to reduce
the complexity of this pilot study, we chose an
open-label design.
Patients in both arms also received standard
diabetic wound management, including sharp sur-
gical debridement at each visit where appropriate,
pressure off-loading as applicable, and daily dress-
ing changes using a nonadherent, moisture-perme-
able dressing followed by a secondary saline-
moistened gauze dressing. To keep the dressings
in place, roll gauze was wrapped circumferentially
around the wound and secured with nonirritating
tape.
Investigators provided patients with levofloxacin
tablets, gentamicin-collagen sponges as necessary,
their choice of off-loading device, and dressings for
standard wound management for daily outpatient
treatment between study visits. Investigators also
instructed patients or their caregivers in use of the
off-loading device and method of dressing change,
including dressing removal and wound irrigation
with sterile saline.
Patients were treated for at least 7 days and
continued treatment until the investigator deter-
mined that all signs and symptoms of infection had
resolved to a maximum of 28 days. The test-of-cure
and safety assessment was scheduled 2 weeks after
discontinuation of treatment, for a total study
duration of up to 42 days.
This study was approved by either Quorum
Review IRB (Seattle, WA) or a review board at the
trial center, and the studies were conducted in
accordance with the Declaration of Helsinki and
Good Clinical Practice guidelines. Patients at eight
sites in the United States and one site in the United
Kingdom were enrolled in the study. At their first
visit, they underwent screening procedures, includ-
ing routine laboratory tests and culture of a tissue
specimen of their ulcer. Patients who were deemed
eligible provided written informed consent and
were randomized in a 2:1 ratio to the treatment or
control group using an interactive voice response
system. Randomization was not stratified by any
baseline characteristic. Enrolled patients all re-
ceived their allocated treatment while awaiting
baseline laboratory results. Once these results were
available, patients who were no longer deemed
eligible due to an out-of-range laboratory value or a
levofloxacin-resistant isolate were withdrawn from
the study.
Journal of the American Podiatric Medical Association  Vol 102  No 3  May/June 2012
The investigator performed clinical and microbi-
ological assessments at regular study visits while
the patient was undergoing antibiotic therapy (ie,
days 3, 7, 10, 14, 21, and 28) and again at the follow-
up visit. Clinical outcomes were defined as follows:
1) clinical cure: resolution of all baseline signs and
symptoms of infection; 2) clinical improvement:
improvement of at least 1, but not all, of the baseline
signs and symptoms of infection; and 3) failure: no
improvement of any of the baseline signs and
symptoms of infection.
A third party blinded to the treatment regimen
determined microbiological outcomes by isolated
pathogen. These outcomes were classified as
documented eradication, presumed eradica-
tion, documented persistence, presumed per-
sistence, or unknown. The evaluator also as-
signed a by-patient outcome of microbiological
success or microbiological failure based on
culture results obtained at the final visit or early
termination.
We performed minimal inhibitory concentration
testing on all isolated pathogens to determine
susceptibility to systemic concentrations of levo-
floxacin, linezolid, vancomycin, and oxacillin for
gram-positive species and levofloxacin, aztreonam,
and piperacillin/tazobactam for gram-negative spe-
cies. We also performed definitive gentamicin
minimal inhibitory concentration testing against all
isolated pathogens up to 1,024 lg/mL to mimic high
local concentrations.
After wound debridement, the investigators grad-
ed the ulcer using the Lipsky wound score at
baseline and at all of the subsequent visits. In
addition, study personnel who were blinded to the
treatment regimen obtained wound tracings. A
centralized imaging facility (Canfield Scientific,
Fairfield, NJ) calculated the wound surface area
from the tracings using a validated custom applica-
tion based on Image-Pro Plus (Media Cybernetics,
Bethesda, MD).
For all of the patients who received the gentami-
cin-collagen sponge, on day 7 we collected predose
trough pharmacokinetic samples within the 30
minutes before application to determine whether
there was any systemic gentamicin accumulation.
We also measured serum gentamicin concentrations
for up to 8 hours after sponge application on day 7
for all five patients who received the 10 3 10-cm
sponge and for the first three patients who received
the 5 3 5-cm sponge.
We collected safety data, including physical
findings, vital signs, and laboratory assessments, at
scheduled intervals during the study, and we
225
recorded all adverse events. Using standard defini-
tions, an adverse event was any clinically unfavor-
able and unintended sign (including abnormal
laboratory findings), symptom, or disease temporal-
ly associated with the treatment, regardless of
whether it was causally related to the treatment. A
serious adverse event was any adverse event that
resulted in death, was life threatening, required
inpatient hospitalization or prolongation of the
existing hospitalization, resulted in permanent
disability or incapacity, caused a congenital abnor-
mality, or was an important medical event. We
designated each adverse event based on its clinical
severity as mild (caused no limitation of usual
activities), moderate (caused some limitation of
usual activities), or severe (prevented or severely
limited usual activities). We also assessed its
relationship to treatment as either definitely
related, probably related, unlikely related, or Results
not related.
To improve patient enrollment, we made several
amendments to the original protocol, including
widening the patient age range, decreasing the
number of patients requiring postdose pharmacoki-
netic sampling, and reducing the duration of
postdose sampling. In addition, we added the
exclusion criteria history of epilepsy and tendon
disorders related to fluoroquinolone administra-
tion, which were necessitated by a new levoflox-
acin black box warning.
The primary efficacy end point for this pilot study
was the percentage of patients with a clinical
outcome of clinical cure on day 7. Assuming that
70% of treatment patients and 40% of control
patients would achieve this outcome, we calculated
that we needed a sample size of 50 treatment
patients and 25 control patients to achieve 70%
power to detect a difference between the two
groups at the P .05 significance level using a 2-
sided v2 test and a 2:1 randomization ratio.
Statistical calculations suggested that we would
need a substantially larger sample size to test for
treatment superiority at the test-of-cure visit given
the higher rates of clinical cure expected. There-
fore, we selected the day 7 visit as the primary end
point to satisfy an appropriate sample size for a
pilot study.
Secondary efficacy end points included the
following: percentage of patients with clinical cure
at all time points other than day 7, percentage with a
positive clinical response (defined as clinical cure
or clinical improvement), percentage with pathogen
eradication at each time point (each of which we
summarized descriptively and compared across
226 May/June 2012  Vol 102  No 3  Journal of the American Podiatric Medical Association
treatments using the v2 test), change in Lipsky
wound score, total wound surface area (which we
compared across treatments using 2-sample t tests),
time to clinical cure, time to positive clinical
response, and time to eradication of baseline
pathogens (which we calculated for each patient
and summarized using the Kaplan-Meier method,
with log-rank tests used to compare treatment
groups). We summarized antibiotic susceptibility
testing of isolated pathogens descriptively by the
specific antibiotic agent. For patients with postdose
pharmacokinetic assessments, we plotted the indi-
vidual gentamicin serum concentration time profiles
and corresponding mean profiles for each sponge
size. Safety evaluations included summary reports
for the incidence and severity of all adverse events,
as described previously herein.
Between April 1, 2008, and the last patients final
visit in May 2009, we enrolled 56 patients; 38
patients were randomized to the treatment group
and 18 to the control group. We decided to stop the
trial before reaching the targeted sample size
because enrollment in this pilot study was more
difficult than we anticipated. This difficulty was
largely related to the inclusion and exclusion
criteria being too restrictive, particularly the re-
quirements that patients have only a single ulcer
and not have received systemic or topical antimi-
crobial therapy in the preceding 2 weeks. We also
found that many randomized patients were subse-
quently deemed ineligible once baseline laboratory
results became available.
The disposition of all enrolled patients is sum-
marized in Figure 1. Of 56 patients randomized, 33
(59%) completed the study and 23 (41%) were
discontinued prematurely. The most common rea-
son for discontinuation (20 of 23 patients) was
ineligibility because the patient did not meet all of
the inclusion and exclusion criteria. We deemed one
patient ineligible on the day of randomization for
having an underlying medical condition that would
interfere with result interpretation, and we deemed
19 patients (11 in the treatment group and 8 in the
control group) ineligible once baseline laboratory
results became available: 14 due to excessively high
hemoglobin A1c levels, one of whom also had
elevated aspartate aminotransferase and alanine
aminotransferase levels; one due to high serum
creatinine levels; one due to high alanine amino-
transferase levels; and three due to levofloxacin-
resistant isolates. Because such a high proportion of
Figure 1. CONSORT 2010 flow diagram of patient disposition. The following six patient populations were
defined for this study: 1) modified intention-to-treat (mITT) population: all patients who had been randomized
to receive and had taken any dose of gentamicin-collagen sponge or antibiotic therapy and had not early
terminated from the study for failing to comply with the study criteria; 2) micro-modified intention-to-treat
(mmITT) population: all patients with a positive baseline pathogen culture who were randomized to receive the
gentamicin-collagen sponge or any antibiotic therapy and had taken any dose of gentamicin-collagen sponge
or antibiotic therapy and had not early terminated from the study for failing to comply with study criteria; 3) per-
protocol (PP) population: all randomized patients who were not early terminated from the study for failing to
comply with the study entry criteria and who did not deviate* significantly from the protocol; 4) per-protocol
population for mmITT (mPP): all randomized patients who were not early terminated from the study for failing
to comply with the study entry criteria who had a positive baseline pathogen culture and who did not deviate*
significantly from the protocol; 5) pharmacokinetic (PK) population: all randomized patients who received the
gentamicin-collagen sponge and who had serum samples collected for determination of gentamicin
concentrations on day 7 (this included the predose [trough] measurements); and 6) safety population: all
randomized patients who received any dose of gentamicin-collagen sponge or antibiotic therapy. *Major
protocol violations or deviations were identified during the review of the study data before the database lock.
All of the efficacy analyses were to be based on both the mITT and PP populations. All of the microbiology
analyses were to be based on both the mmITT and mPP populations. All of the PK analyses were based on
the PK population. Note: Results for the PP analysis were similar to those for the mITT analysis. For the
purpose of brevity, only the mITT, PK, and safety populations are discussed.

randomized patients (20 of 56 [36%]) was deemed
ineligible, we defined a modified intention-to-treat
population used for efficacy analyses to include
only the 36 eligible patients.
The three eligible modified intention-to-treat
patients who did not complete the study were all
randomized to the treatment group. One clinically
improved patient was withdrawn on day 9 due to a
serious adverse event (hypoglycemia) unrelated to
the gentamicin-collagen sponge, another was with-
drawn for failing to apply the sponges (protocol
noncompliance), and the third achieved clinical
cure at day 14 but did not attend the test-of-cure
visit and was deemed lost to follow-up. One

Journal of the American Podiatric Medical Association  Vol 102  No 3  May/June 2012 227
additional patient was deemed clinically cured at
day 14 but was subsequently hospitalized because
of tendon rupture attributed to oral levofloxacin
(both legs casted). The patient completed the
follow-up visit, although the cast prevented final
clinical assessments. Therefore, of the 36 patients in
the modified intention-to-treat population, 32 were
evaluable for test of cure at the follow-up visit.
Demographic characteristics of enrolled patients
were not significantly different between the groups
(Table 1). However, assessment of baseline charac-
teristics in the modified intention-to-treat popula-
tion revealed an unfortunate imbalance in baseline
wound severity. Baseline Lipsky wound scores were
significantly higher in the treatment group than in
the control group (median, 17 versus 12; P .011),
and the mean baseline wound surface area was
almost fourfold larger for patients randomized to
the treatment group (5.11 cm2) than to the control
group (1.24 cm2). Of the 26 patients in the modified
intention-to-treat population randomized to the
treatment group, 21 received the 5 3 5-cm sponge
and five received the 10 3 10-cm sponge.
Pathogens isolated from wounds at baseline are
summarized in Table 2. The most commonly
identified baseline pathogen for both groups was S
aureus, about equally divided between methicillin-
resistant and methicillin-sensitive species. The
investigators did not switch any patient from oral
levofloxacin to an alternative antibiotic regimen on
the basis of their culture and antibiotic susceptibil-
ity results.
Efcacy
On day 7 (the primary end point), no patients in the
treatment group and three in the control group
achieved clinical cure (P .017). On days 10, 14, and
21, the control group had a nonsignificantly higher
cumulative percentage of patients with clinical cure
compared with the treatment group. At the test-of-
cure visit, however, the evaluable patients in the
treatment group had a significantly higher propor-
tion of patients with clinical cure than did the
control group (22 of 22 [100%] versus 7 of 10 [70.0%],
P .024). The treatment group also had a
nonsignificantly higher cumulative percentage of
patients with clinical cure at the end-of-treatment
visit than did the control group (24 of 26 [92.3%]
versus 7 of 10 [70.0%], P .119). Using Kaplan-Meier
estimates of time to clinical cure, 75% of patients
achieved an outcome of clinical cure after 28 days in
the treatment group compared with 40 days in the
control group.
228 May/June 2012  Vol 102  No 3  Journal of the American Podiatric Medical Association
The proportion of patients with baseline patho-
gen eradication on day 3 was significantly higher in
the treatment group than in the control group (20
of 26 [76.9%] versus 1 of 8 [12.5%], P , .001). The
pathogen eradication rate continued to be signifi-
cantly higher in the treatment group throughout
the remainder of the study (P  .038 for each
visit). Using Kaplan-Meier estimates, 75% of pa-
tients in the treatment group achieved baseline
pathogen eradication 6 days after initiation of
treatment; a comparative control group value
could not be calculated because baseline pathogen
eradication was observed for only one control
patient during treatment. As shown in Figure 2, by
day 21, the probability of achieving baseline
pathogen eradication in the treatment group was
0.923 compared with 0.125 in the control group; the
overall log-rank test comparing the likelihood of
baseline pathogen eradication by treatment groups
was significantly in favor of the treatment group (P
, .001).
Although the study was not powered to detect
significant changes in Lipsky wound score or wound
surface area from baseline, the difference in
absolute reduction of the Lipsky wound score at
the final visit was significantly in favor of the
treatment group (median, 13 versus 8, P .042),
although the difference in percentage reduction was
not significant (median, 77% versus 64%, P
.376). However, there was no apparent trend
suggesting either accelerated or delayed wound
healing in the treatment group compared with in the
control group based on changes in wound surface
area from baseline. Furthermore, there was no
statistically significant difference in the percentage
of patients who achieved complete wound closure
at the end of treatment.
Pharmacokinetics
For all of the patients treated with the gentamicin-
collagen sponge on day 7, the trough serum
gentamicin concentration was undetectable (ie,
0.19 lg/mL, the lowest concentration that the
analytical method can quantify). Postdose serum
gentamicin concentrations were quantifiable for
three of the five patients who received the 10 3
10-cm sponge; the highest peak concentration
recorded was 1.22 lg/mL 2 hours postdose for the
patient who had the largest baseline wound area
(39.3 cm2). Postdose concentrations were undetect-
able for the three patients who received the 5 3 5-
cm sponge.
Table 1. Patient Demographic Characteristics
Parameter Treatment group (n 38) Control group (n 18)

Age (years)
Mean (SD) 57.9 (11.47) 54.7 (12.80)
Median (range) 58.0 (2480) 54.5 (2981)
Sex (no. [%])
Male 23 (60.5) 15 (83.3)
Female 15 (39.5) 3 (16.7)
Race (no. [%])
American Indian or Alaskan Native 1 (2.6) 0
Black 4 (10.5) 3 (16.7)
Native Hawaiian or other Pacific Islander 1 (2.6) 0
White 32 (84.2) 15 (83.3)
Ethnicity (no. [%])
Hispanic or Latino 12 (31.6) 5 (27.8)
Not Hispanic or Latino 26 (68.4) 13 (72.2)
BMI
Mean (SD) 32.38 (6.500)a 32.67 (5.796)
Median (range) 32.30 (21.144.8)a 31.70 (23.745.1)

Abbreviation: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared).
a
Mean and median BMI values in the treatment group are based on measurements for 37 patients.

Safety discontinued taking the study drug because of an

Of the 56 patients enrolled, 16 (28.6%) experienced
one or more adverse events during the study. The
proportion of patients experiencing any adverse
event was similar for the treatment (11 of 38
[28.9%]) and control (5 of 18 [27.8%]) groups. The
most common adverse events occurring in at least
two patients per group were infected skin ulcer,
tinea pedis, and increased blood creatinine concen-
tration. Other than increased blood creatinine level,
there were no clinically or statistically significant
changes in laboratory test values or vital signs. Most
adverse events were mild or moderate in severity,
but there were six serious adverse events: five in the
treatment group (hypoglycemia, renal failure, cellu-
litis, tendon rupture, and wound hemorrhage), all of
which resolved, and one in the control group
(infected skin ulcer), which resolved with sequelae.
Only one patient in each group experienced an
adverse event that was considered definitely or
probably related to the study drug: moderate renal
failure in the treatment group (also reported as a
serious adverse events) and moderate drug intoler-
ance in the control group. Both of these related
adverse events resolved by the final visit. The
serious adverse event of tendon rupture reported
in the treatment group was attributed to the
concomitant oral levofloxacin therapy. No deaths
occurred during the study, and only one patient
adverse event.
Discussion
Using topical antimicrobial therapy for chronically
infected wounds has great appeal.3 These agents
have typically been applied alone to treat relatively
minor infections, although calcium sulfate pellets
impregnated with tobramycin have been used as an
adjunct to systemic antibiotics for treating diabetic
foot infections, with or without osteomyelitis.12,13
We have not been able to find information on any
other randomized controlled studies that have
examined the benefit of topical combined with
systemic antibiotic therapy for diabetic foot infec-
tions. This study was designed to determine
whether adding the gentamicin-collagen sponge to
standard of care (systemic antibiotic therapy plus
standard diabetic wound management) improves
the rate of clinical cure of infection in diabetic
patients with an infected foot ulcer of moderate
severity.
We designed this pilot study to use clinical cure
on day 7 as the primary end point and noted that the
cure rate was significantly higher for the control
group. The mean wound surface area at baseline,
however, was nearly fourfold greater, and the
Lipsky wound scores were significantly higher, in

Journal of the American Podiatric Medical Association  Vol 102  No 3  May/June 2012 229
Table 2. Pathogens Isolated From Diabetic Foot Wounds at Baseline
mITT populationa Safety populationb
Pathogen Treatment (n 26) Control (n 10) Treatment (n 38) Control (n 18)

Gram positive
b-Hemolytic Streptococcus 5 (19.2) 2 (20.0) 7 (18.4) 3 (16.7)
Coagulase-negative Staphylococcus spp 7 (26.9) 0 10 (26.3) 3 (16.7)
Vancomycin-susceptible Enterococcus faecalis 2 (7.7) 4 (40.0) 4 (10.5) 5 (27.8)
Gram-positive cocci (NOS) 0 0 1 (2.6) 0
Peptostreptococcus spp 1 (3.8) 1 (10.0) 2 (5.3) 1 (5.6)
Methicillin-resistant Staphylococcus aureus 7 (26.9) 5 (50.0) 9 (23.7) 5 (27.8)
Methicillin-sensitive Staphylococcus aureus 6 (23.1) 3 (30.0) 10 (26.3) 5 (27.8)
Streptococcus pyogenes 0 1 (10.0) 0 1 (5.6)
Gram negative
Acinetobacter sp 1 (3.8) 0 1 (2.6) 2 (11.1)
Alcaligenes faecalis 1 (3.8) 0 1 (2.6) 0
Bacteroides fragilis 0 1 (10.0) 1 (2.6) 1 (5.6)
Burkholderia sp 1 (3.8) 0 1 (2.6) 0
Enterobacter cloacae 1 (3.8) 0 1 (2.6) 1 (5.6)
Klebsiella oxytoca 3 (11.5) 0 3 (7.9) 2 (11.1)
Morganella morganii 0 0 0 1 (5.6)
Ochrobactrum anthropi 1 (3.8) 0 1 (2.6) 0
Proteus sp 4 (15.4) 0 5 (13.2) 0
Providencia rettgeri 0 0 1 (2.6) 0
Pseudomonas aeruginosa 4 (15.4) 0 4 (10.5) 1 (5.6)
Serratia marcescens 2 (7.7) 0 2 (5.3) 0
Other gram negatives 1 (3.8) 1 (10.0) 1 (2.6) 1 (5.6)
Other
Candida albicans 1 (3.8) 0 1 (2.6) 0

Abbreviations: mITT, modified intention-to-treat; NOS, not otherwise specified.

a
The mITT population consisted of all patients randomized to receive the gentamicin-collagen sponge or any antimicrobial therapy
who took any dose of gentamicin-collagen sponge or systemic antibiotic therapy and did not early terminate from the study for failing
to comply with the inclusion and exclusion criteria.
b
The safety population consisted of all randomized patients who received any dose of gentamicin-collagen sponge or systemic
antibiotic therapy.

patients randomized to receive the gentamicin-

P collagen sponge compared with the control group.
Figure 2. Probability of achieving baseline pathogen
eradication in the treatment and control groups.
Although occurring by chance, this imbalance in
baseline wound severity may be attributable to the
smaller-than-expected number of eligible patients
and a lack of stratification by any baseline wound
characteristic. These factors likely biased the
results in favor of the control group, especially at
the early visits. Note that despite this baseline
imbalance and the relatively small sample size, the
treatment group showed significantly higher rates of
eradication of baseline pathogens from day 3
onward and, most important, clinical cure at the
test-of-cure visit. In fact, all treatment group
patients who were evaluable at the test-of-cure visit
achieved clinical cure, and we also observed a

230 May/June 2012  Vol 102  No 3  Journal of the American Podiatric Medical Association
statistically significant greater reduction in Lipsky
wound score from baseline at this visit. Taken
together, these findings suggest that when used in
combination with systemic antibiotic therapy, the
gentamicin-collagen sponge may be more effective
for the treatment of diabetic foot infections than is
systemic therapy alone.
Based on previous studies of gentamicin concen-
trations found in wound drainage fluids after
surgical implantation of the gentamicin-collagen
sponge,14 we thought that local concentrations up
to 1,000 lg/mL in the wound environment were
possible after topical application. According to the
results of susceptibility testing, such a concentra-
tion would exceed the gentamicin minimal inhibi-
tory concentrations for 91.6% of the organisms (87
of 95) isolated from the infected wounds at
baseline. This likely accounts for the faster and
more effective pathogen eradication observed with
the gentamicin-collagen sponge and, perhaps, the
superior clinical cure rate at the test-of-cure visit.
However, combining gentamicin with levofloxacin
could conceivably have a synergistic effect against
wound pathogens if both achieve therapeutic
concentrations. Few studies have addressed the
potential for synergy between fluoroquinolones and
aminoglycosides. Most older studies showed indif-
ference against gram-negative organisms,15 but in
one study,16 gatifloxacin was synergistic with
gentamicin against some drug-resistant pathogens,
including Pseudomonas aeruginosa, and in another
study17 the bactericidal effect of moxifloxacin was
increased against methicillin-resistant S aureus
when combined with gentamicin. Of note, one
investigation18 demonstrated synergy between ami-
noglycosides and fluoroquinolones against some
biofilm-forming P aeruginosa strains.
The pharmacokinetic results confirmed that
patients had low serum gentamicin concentrations.
Thus, patients should be able to derive the benefit of
high local concentrations of gentamicin without
being at risk for the potentially serious adverse
effects of therapeutic serum concentrations. These
results support findings from other studies where
surgical implantation of one or more gentamicin-
collagen sponges has been used to treat various
localized bone and soft-tissue infections while
maintaining low systemic exposure and avoiding
adverse effects.19-21
By measuring the change in wound surface area
from baseline, we found no apparent trend to
suggest either accelerated or delayed wound heal-
ing in the treatment group compared with the
control group. The study was not, however,
Journal of the American Podiatric Medical Association  Vol 102  No 3  May/June 2012
adequately powered to detect such differences.
The low incidence of drug-related adverse events
and analysis of all of the safety variables seem to
suggest that daily application of the gentamicin-
collagen sponge for up to 28 days is safe and well
tolerated.
Conclusions
There are several limitations of this study. First,
because of difficulties with enrollment and patient
eligibility, the number of patients analyzed was
relatively small. Second, because of concerns about
applying placebo collagen sponges to an infected
wound, the study was not blinded. Third, differenc-
es in the size and severity of the wounds at baseline
compromised the interpretation of the primary
outcome. Nevertheless, the results suggest that the
gentamicin-collagen sponge is safe and potentially
beneficial when added to systemic antibiotic ther-
apy for the treatment of diabetic foot infections of
moderate severity. We believe that these results
provide a strong basis for conducting a larger
blinded clinical trial, and the valuable experience
gained will allow us to overcome the limitations of
this pilot study.
Acknowledgment: Premier Research Group Ltd
contributed to the writing of this manuscript. We
thank the following investigators and their study
staff who enrolled patients in the study: Dr. Barbara
Aung (Aung Foot Health Clinics & Advanced Wound
Care Center, Tucson, AZ), Dr. Kevin Brattain
(Methodist Medical Center, Peoria, IL), Dr. Walter
DCosta (Santa Rosa, CA), Dr. Michael Edmonds
(Kings College Hospital, London, UK), Dr. Jeffrey
Karr (Karr Foot Kare, Lakeland, FL), Dr. Susan
Kemp (Clinical Trials of America, Inc., Shreveport,
LA), Dr. Max McLaughlin (Wilmax Clinical Re-
search, Inc., Mobile, AL), Dr. Robert Mendicino
(Western Pennsylvania Hospital, Pittsburgh, PA),
Dr. Alexander Reyzelman (Center for Clinical
Research, Castro Valley, CA), Dr. Bhavesh Shah
(South Texas Foot Institute, San Antonio, TX), Dr.
Robert Shouey (Harrisonburg Foot Clinic, Harrison-
burg, VA), Dr. Felix Sigal (Pacific Clinical Center,
Los Angeles, CA), Dr. Scott Soulier (Jean Brown
Research, Salt Lake City, UT), Dr. Dean Vayser
(Therapeutics Clinical Research, San Diego, CA),
and Dr. Stephanie Wu (National Center for Limb
Preservation at Advocate Lutheran General Hospi-
tal, Niles, IL).
231
Financial Disclosure: This study was funded in
whole by Innocoll Technologies Ltd.
Conflict of Interest: Dr. Lipsky has served as a
consultant to Innocoll Technologies Ltd. Mr. Kuss
was involved in protocol development, research,
and data analysis for this clinical trial. Drs.
Edmonds, Reyzelman, and Sigal were investigators
in the study and received research funding from
Innocoll Technologies Ltd.
References
1. LIPSKY BA, BERENDT AR, DEERY HG, ET AL: Diagnosis and
treatment of diabetic foot infections. Clin Infect Dis 39:
885, 2004.
2. LIPSKY BA, HOLROYD KJ, ZASLOFF M: Topical versus
systemic antimicrobial therapy for treating mildly
infected diabetic foot ulcers: a randomized, controlled,
double-blinded, multicenter trial of pexiganan. Clin
Infect Dis 47: 1537, 2008.
3. LIPSKY BA, HOEY C: Topical antimicrobial therapy for
treating chronic wounds. Clin Infect Dis 49: 1541, 2009.
4. GRIFFIS CD, METCALFE S, BOWLING FL, ET AL: The use of
gentamycin-impregnated foam in the management of
diabetic foot infections: a promising delivery system?
Expert Opin Drug Deliv 6: 639, 2009.
5. CITRON DM, GOLDSTEIN EJC, MERRIAM CV, ET AL: Bacteri-
ology of moderate-to-severe diabetic foot infections and
in vitro activity of antimicrobial agents. J Clin Microbiol
45: 2819, 2007.
6. KOTLUS BS, WYMBS RA, VELLOZZI EM, ET AL: In vitro activity
of fluoroquinolones, vancomycin, and gentamicin
against methicillin-resistant Staphylococcus aureus
ocular isolates. Am J Ophthalmol 142: 726, 2006.
7. CHATTOPADHYAY B, TELI JC: Gentamicin-resistant Staphy-
lococcus aureus after topical therapy in general
practice. J Hosp Infect 2: 278, 1981.
8. AFSHARI NA, MA JJ, DUNCAN SM, ET AL: Trends in
resistance to ciprofloxacin, cefazolin, and gentamicin
in the treatment of bacterial keratitus. J Ocul Pharmacol
Ther 24: 217, 2008.
9. FORMIGA A, NEVES J, PAULINO A, ET AL: Treatment of leg
ulcers with collagen gentamicin implant (preliminary
study), in Proceedings Eurosurgery 2000, edited by E
Duren, A Tuzuner, I Sayek, Monduzzi, Istanbul, 2000.
10. LIPSKY BA, POLIS AB, LANTZ K, ET AL: The value of a wound
232 May/June 2012  Vol 102  No 3  Journal of the American Podiatric Medical Association
score for diabetic foot infections in predicting treatment
outcome: a prospective analysis from the SIDESTEP
trial. Wound Repair Regen 17: 671, 2009.
11. GRAHAM DR, TALAN DA, NICHOLS RL, ET AL: Once-daily,
high-dose levofloxacin versus ticarcillin-clavulanate
alone or followed by amoxicillin-clavulanate for com-
plicated skin and skin-structure infections: a random-
ized, open-label trial. Clin Infect Dis 35: 381, 2002.
12. ARMSTRONG DG, FINDLOW AH, OYIBO SO, ET AL: The use of
absorbable antibiotic-impregnated calcium sulphate
pellets in the management of diabetic foot infections.
Diabet Med 18: 942, 2001.
13. SALGAMI EV, BOWLING EV, WHITEHOUSE RW, ET AL: Use of
tobramycin-impregnated calcium sulphate pellets in
addition to oral antibiotics: an alternative treatment to
minor amputation in a case of diabetic foot osteomye-
litis. Diabetes Care 30: 181, 2007.
14. STEMBERGER A, GRIMM H, BADER F, ET AL: Local treatment
of bone and soft tissue infections with the collagen-
gentamicin sponge. Eur J Surg Suppl 578: 17, 1997.
15. NEU H: Synergy of fluoroquinolones with other antimi-
crobial agents. Rev Infect Dis 11 (suppl 5): S1025, 1989.
16. DAWIS MA, ISENBERG HD, FRANCE KA, ET AL: In vitro activity
of gatifloxacin alone and in combination with cefepime,
meropenem, piperacillin and gentamicin against multi-
drug-resistant organisms. J Antimicrob Chemother 51:
1203, 2003.
17. MAGGIOLO F, CAPRA R, BOTTURA P, ET AL: In-vitro activity of
moxifloxacin combined with other antibacterials
against methicillin-resistant Staphylococcus aureus. J
Chemother 12: 195, 2000.
18. CERNOHORSKA L, VOTAVA M: Antibiotic synergy against
biofilm-forming Pseudomonas aeruginosa. Folia Micro-
biol (Praha) 53: 57, 2008.
19. IPSEN T, JRGENSEN PS, DAMHOLT V, ET AL: Gentamicin-
collagen sponge for local applications: 10 cases of
chronic osteomyelitis followed for 1 year. Acta Orthop
Scand 62: 592, 1991.
20. JRGENSEN LG, SRENSEN TS, LORENTZEN JE: Clinical and
pharmacokinetic evaluation of gentamycin containing
collagen in groin wound infections after vascular
reconstruction. Eur J Vasc Surg 5: 87, 1991.
21. LEYH RG, BARTELS C, Sievers HH: Adjuvant treatment of
deep sternal wound infection with collagenous genta-
mycin. Ann Thorac Surg 68: 1648, 1999.