Comparative Effectiveness of Clostridial Collagenase

Ointment to Medicinal Honey for Treatment

of Pressure Ulcers

Adrienne M. Gilligan,1,* Curtis R. Waycaster,2 Richard Bizier,3

Bong-Chul Chu,4 Marissa J. Carter,5 and Caroline E. Fife6
1
Truven Health Analytics, Houston, Texas.
2
Smith & Nephew, Fort Worth, Texas.
3
Truven Health Analytics, Cambridge, Massachusetts.
4
Truven Health Analytics, Santa Barbara, California.
5
Strategic Solutions Inc., Cody, Wyoming.
6
Intellicure, Inc, The Woodlands, Texas.

Objective: Compare enzymatic debridement using clostridial collagenase

ointment (CCO) with autolytic debridement using medicinal honey in the
hospital outpatient setting for treating pressure ulcers (PUs).
Approach: Retrospective deidentified electronic health records from 2007–
2013 were extracted from the U.S. Wound Registry. Propensity score matching
Adrienne M. Gilligan, PhD
followed by multivariable analyses was used to adjust for selection bias and
assess treatment effects comparing CCO-treated versus honey-treated PUs. Submitted for publication December 7, 2016.
Accepted in revised form December 15, 2016.
Key outcomes included 100% granulation and epithelialization at 1 year.
*Correspondence: Truven Health Analytics,
Results: Five hundred seventeen CCO-treated PUs (446 patients) were mat- an IBM Company, 2 Riverway #200, Houston, TX
ched to corresponding honey-treated PUs (341 patients). The majority of PUs 77056
(e-mail: [email protected]).
were stage III (CCO 56%, honey 55%). CCO users had significantly fewer total
visits (9.1 vs. 12.6; p < 0.001), fewer total selective sharp debridements (2.7 vs.
4.4; p < 0.001), and fewer PUs receiving negative pressure wound therapy (29%
vs. 38%; p = 0.002) compared with honey.
Innovation: CCO-treated PUs were 38% more likely to achieve 100% granu-
lation compared to honey-treated PUs at 1 year, p = 0.018. Mean days to 100%
granulation were significantly lower for CCO-treated PUs (255 vs. 282 days,
p < 0.001). CCO-treated PUs were 47% ( p = 0.024) more likely to epithelialize
at 1 year compared to PUs treated with honey. Mean days to epithelialization
were significantly lower for PUs treated with CCO at 1 year (288 vs. 308 days;
p = 0.011).
Conclusion: All stages of PUs treated with CCO achieved faster rates of
granulation and subsequent epithelialization compared to PUs treated with
medicinal honey as measured by real-world data collected in the hospital
outpatient department care setting.

Keywords: pressure ulcer, debridement, granulation, epithelialization, clos-

tridial collagenase ointment, medicinal honey

ª Adrienne M. Gilligan, et al., 2017; Published by Mary Ann Liebert, Inc. This Open Access article is
distributed under the terms of the Creative Commons License (http://creativecommons.org/licenses/by/
4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the
original work is properly credited.

ADVANCES IN WOUND CARE, VOLUME 6, NUMBER 4

Mary Ann Liebert, Inc. DOI: 10.1089/wound.2016.0720
j 125
126 GILLIGAN ET AL.

INTRODUCTION dermal ulcers and severely burned areas.25 Col-

Pressure ulcers (PUs) pose substantial clinical lagenase, a proteolytic enzyme derived from Clos-
and economic challenges to patients, their families, tridium histolyticum, breaks down collagen in
and healthcare providers, annually affecting an es- necrotic tissue, selectively removing detritus with-
timated 2.5 million patients in the United States (US) out harming healthy tissue, facilitating granula-
at a substantial cost, estimated to range $9.1–$11.6 tion tissue formation and epithelialization.22,26–31
billion annually.1–3 Seen most commonly in seriously Other clinical advantages of enzymatic debride-
ill patients, the elderly, and infants due to under- ment with CCO include ease of application, mini-
lying conditions impairing movement, PUs are an mal blood loss, and enhanced tissue proliferation.32
indicator of poor overall health status that may Autolytic debridement with support from topically
contribute to mortality risk.4,5 The prevalence of applied medicinal honey is thought to work with the
these high cost adverse events varies considerably body’s natural healing process to cleanse debris and
by clinical setting, ranging from 0.4% to 38% in necrotic tissue from the wound. Its activity has been
acute care hospitals, 2.2% to 23.9% in long-term attributed to its acidity increasing oxygen release
care facilities, and 0% to 17% in home healthcare from hemoglobin, decreasing the activity of destruc-
settings.6–8 tive proteases, and osmotically drawing fluid from
Management of PUs is often complicated by the the wound bed to create an outflow of lymph. While
patient’s underlying disease and associated pain, there is some evidence supporting the use of honey
infection, and interference with functional recov- in wound care, there is however a lack of compara-
ery leading to lengthened hospital stays and in- tive effectiveness data.33,34
creased costs of care.8,9 PU severity is categorized
using a four stage classification system, ranging
from stage I (intact skin, nonblanchable erythema) CLINICAL PROBLEM ADDRESSED
to stage IV (full thickness tissue loss exposing bone, There is a lack of current evidence comparing
tendon, or muscle).10,11 Management requires a enzymatic debridement with medicinal honey-
multimodal care strategy, including nutritional based autolytic debridement as clinically effective
assessment, body positioning, mobilization, and therapies for the management of chronic wounds.
environmental factors (bedding, chairs, medical The objective of this study was to compare clinical
devices, etc.), as well as direct treatments to the outcomes, particularly granulation and epithelial-
wound itself.11 Treatment of chronic wounds such ization, between enzymatic debridement with CCO
as PUs entails a systematic approach to preparing and autolytic debridement with support from me-
the wound bed for healing, consisting of debriding dicinal honey for the treatment and management
unhealthy and nonviable tissue, controlling exu- of PUs in the hospital outpatient environment.
date and edema, controlling infection and bacterial Real-world observational data were used to enable
bioburden, maintaining moisture balance, and more easily and quickly recruiting a larger patient
promoting healthy granulation tissue.12–17 population that is representative of actual clinical
Debridement, the removal of nonviable and practices to support a better understanding of PU
poorly healing tissue, foreign material, and bacte- treatment beyond the controlled environment of a
ria (biofilm) from a wound, is critical to preparing clinical trial.
the wound bed for healing and facilitating wound
epithelialization.18–21 Debridement methods have MATERIALS AND METHODS
been described as surgical or sharp, mechanical, Study design
chemical, biological, autolytic, and enzymatic.16,22 This study of PU patients used a longitudinal
These methods may be used singly or combined observational case–control confirmatory design
serially to optimize the debridement process. Some approach to compare matched treatment cohorts
are considered ‘‘selective,’’ acting only on necrotic receiving CCO (cases) with controls treated with
tissue, or ‘‘nonselective’’ that remove normal, as medicinal honey, analyzing the differences and
well as necrotic, tissue.23 The choice of debride- risk factors in achievement of wound granulation
ment method depends on patient factors, wound and epithelialization over time.
type, likelihood of infection, available medical ex-
pertise, and care setting resources.9,24 Data source
Enzymatic wound debridement with clostridial Data were drawn from the U.S. Wound Registry
collagenase ointment (CCO) has been used since (USWR) database, comprising electronic health
the mid-1960s and is the only FDA approved records (EHRs) from over 100 hospital-based out-
enzymatic agent indicated for debriding chronic patient wound centers in the US and Puerto Rico.
 CLOSTRIDIAL COLLAGENASE VERSUS MEDICINAL HONEY 127

The USWR is certified to meet the Health In- lation tissue formation, epithelialization, or the
formation Technology for Economic and Clinical end of the study period ( December 31, 2012).
Health Act (HITECH Act) standards, including
adherence to wound care quality measures devel- Outcome measures
oped by the USWR as a Qualified Clinical Data The primary outcome of interest was complete
Registry (QCDR) and nationally recognized by the granulation tissue formation for 100% of the wound
Centers for Medicare and Medicaid Services as bed, measured as the percentage of the wound bed
part of the Physician Quality Report System.35,36 with granulation tissue formed. Achievement of
Clinical data are collected at the point of care using 100% granulation (yes/no) and the time to the first
standardized codes and vocabularies with limited recorded encounter noting 100% granulation were
use of free text.37 All wound and ulcer types are measured. If insufficient data were available for
included, such as diabetic foot ulcers, venous sta- granulation, the percentage of necrotic tissue in
sis ulcers, PUs, arterial ulcers, surgical wounds, the wound was assessed, representing the re-
traumatic wounds, vasculitic ulcers, arterial ul- maining part of the wound not yet granulated.
cers, sickle cell ulcers, inflammatory ulcers (e.g., The secondary clinical end point of epithelial-
pyoderma gangrenosum), and ulcers related to ization was evidence of achieving complete epi-
skin disorders such as scleroderma. Ulcers are risk thelialization (yes/no) and time to complete
stratified for outcomes reporting using the Wound epithelialization. Wounds were considered epithe-
Healing Index. Interventions include dressings, lialized when all of the following criteria were met:
compression bandaging, off-loading, cellular- and/ no wound exudate, wound area (length · width)
or tissue-based therapies, hyperbaric oxygen ther- <0.2 square centimeters, and tissue depth <0.1 cm
apy, negative pressure wound therapy, debride- or described as ‘‘partial thickness,’’ the wound
ment, and antibiotics. Outcomes measured include overall was described as controlled or improved
healing or wound closure, surgical closure, death, and not worsening, and the wound bed was char-
and amputations. acterized as epithelialized.
The USWR EHR database was certified to sat-
isfy the conditions set forth in Sections 164.514 Explanatory variables
(a)-(b)1ii of the Health Insurance Portability and The primary explanatory variable was PU
Accountability Act (HIPAA) Privacy Rule regarding treatment with either CCO or honey. Patients were
the determination and documentation of statisti- grouped into mutually exclusive comparison co-
cally deidentified data. Woodlands Institutional horts based on receiving PU treatment with either
Review Board, acting as USWR’s independent In- CCO or honey. The following treatment parame-
stitutional Review Board, has determined that ters were calculated using the date of the initial
retrospective analyses of deidentified HIPAA- wound care visit and each subsequent wound care
compliant data described herein are exempt from visit for PU treatment: number of PUs treated,
the requirement of patient consent. time to first application (days from index date to
the first visit when treatment was applied), days of
Subject selection use (last date of application minus the first date of
Patients were selected who had at least one en- application), number of treatment episodes (de-
counter record with a PU diagnosis code (Interna- fined by gaps between applications of at least
tional Classification of Diseases, Ninth Revision, 60 days), visits with treatment (number of visits
Clinical Modification [ICD-9-CM] diagnosis codes when treatment was applied), application rate
707.00–707.07, 707.09, and 707.20–707.25) in any (days of use divided by application count), and
diagnosis position between January 1, 2007, and percentage of treated visits (application count di-
December 31, 2012. The date of the first found PU vided by total visit count).
diagnosis code during this period was the patient’s The following PU characteristics were captured:
index date. Patients were required to have been arrival score for initial visit (ambulatory, cane/
treated with either CCO or honey and have at least crutches, walker, wheelchair/scooter, and stretch-
one more encounter with a PU diagnosis after the er/bed), number of concurrent wounds, wound age
index PU event. Subjects were excluded who were at first visit, wound surface area at first visit,
treated with both CCO and honey during their wound depth at first visit, PU stage (unstageable,
follow-up, were younger than 18 years old at index, II, III, or IV), PU location, wound exudate level
or whose PUs healed within 2 weeks postindex. (none, minimal, moderate, or heavy), periwound
Patients and their wounds were followed from characteristics (normal, erythematous, macerated,
their index date until the earliest of 100% granu- and other), infection surrogate score (‘‘green event’’
128 GILLIGAN ET AL.

count over treatment time of PU, where ‘‘green ables presented as the count and percentage of
events’’ occur when there is an antibiotic order, a patients and continuous variables providing the
laboratory test for bacterial culture, provider’s number of observations, the mean, SD, and me-
EHR note of infection or oozing green pus from the dian. Statistical tests of significance for observed
wound), total visit count, visit frequency (visit differences between treatment groups were con-
count divided by days from first to last visit), total ducted using Chi-square tests for categorical vari-
number of selective sharp debridements (visits ables and t-tests for continuous variables.
with Current Procedural Terminology [CPT] codes Multivariate models were estimated postmatch-
97597, 97598, or 11040–11047), debridement rate ing to control for any remaining imbalances be-
(debridement count divided by days from first to tween cases and controls. Cox proportional hazard
last visit), and concurrent therapies, including hy- models were estimated to identify risk factors for
perbaric oxygen, negative pressure wound therapy, the time from index date to granulation or epithe-
becaplermin gel, and cellular tissue-derived prod- lialization. The assumption of proportional hazards
ucts (OASIS, Apligraf, Dermagraft, Integra, and was tested plotting Kaplan–Meier observed sur-
PriMatrix). vival curves and comparing them with Cox pre-
Other explanatory variables included the follow- dicted curves from the same variable. Logistic
ing patient demographics and clinical characteris- regression models were estimated to examine the
tics: age, sex, payer, ethnicity, smoker, receiving impact of patients’ demographic, clinical, and treat-
palliative care, pain medication, antibiotics, antire- ment characteristics on granulation and epithe-
jection drugs (immunocompromised), paralyzed, lialization within a fixed time period (e.g., 1-year
home healthcare, comorbidities (heart failure, coro- follow-up). All models controlled for patients’ de-
nary heart disease, end-stage renal disease [ESRD], mographic and clinical characteristics at baseline to
peripheral vascular disease, hypertension, and estimate hazard and odds ratios (HR and OR) of
diabetes). each factor on granulation and epithelialization.
The threshold of statistical significance for all
Analysis analyses was set a priori at 0.05. Statistical ana-
Wounds receiving PU treatment with CCO were lyses were conducted using STATA SE, version 13
matched 1:1 to PUs treated with honey using the (STATA Corp, College Station, TX). Data man-
following propensity score–based matching ap- agement and analytic file building were conducted
proach. The goal of propensity score matching is to using SAS software, version 9.4 (SAS Institute,
reduce bias by statistically emulating the compar- Inc., Cary, NC).
ator balance inherent in randomized clinical trials
(RCTs) so that the distribution of observed baseline
demographic and clinical characteristics is similar RESULTS
between treatment groups.38 Propensity scores for The final matched PU treatment cohorts meet-
each wound were modeled using logistic regression ing all selection criteria each included 517 wounds
with a binary indicator for membership in the CCO from 446 CCO-treated patients and 341 honey-
or honey PU cohorts as the dependent variable and treated patients found in the USWR from 2007
a vector of the following independent variables at through 2012 (Table 1).
index or first PU visit: patient age, wound age, wound Patients were on average aged 64 to 66 years
surface area, arrival score, number of concurrent with the majority older than 65 years (CCO 59.0%,
wounds, infection surrogate score, ESRD (yes/no), PU honey 52.8%; Table 2). Around two-thirds of pa-
stage, and patient paralyzed (yes/no). CCO-treated tients were ambulatory on arrival (CCO 66.4%,
PUs (cases) were matched to honey-treated control honey 63.6%), with over half receiving home
PUs using the nearest neighbor matching technique, healthcare (CCO 59.2%, honey 56.0%) and over
enforcing a caliper of 0.25 times the standard devia- half receiving antibiotics (CCO 57.4%, honey
tion (SD) of the propensity score to maximize the 55.7%). The most commonly noted comorbidity was
number of available controls (honey-treated PUs) hypertension (both 18%) followed by coronary
for each CCO-treated PU.39 The balance achieved heart disease (CCO 13.7%, honey 7.0%; p = 0.003)
by propensity score matching was assessed using and diabetes (both 11%). A higher percentage of
standardized differences comparing the pre- and post- patients in the honey cohort were prescribed pain
match distributions of the independent variables in- medication (CCO 17.0%, honey 22.6%; p = 0.052)
cluded in the propensity score model.38 compared with the CCO cohort (Table 2).
Dependent and independent variables were Wound assessment at first visit (Table 3) found
summarized descriptively, with categorical vari- that the majority of PUs in the matched cohorts
 CLOSTRIDIAL COLLAGENASE VERSUS MEDICINAL HONEY 129

Table 1. Patient and wound sample attrition applying selection criteria

Patients Wounds

Selection Criteria n % n %
a b
PU diagnosis found January 1, 2007—December 31, 2012 and ‡1 encounter in the USWR after the index date 9,202 100.0 20,330 100.0
At least 1 application of CCO or medicinal honey 2,711 29.5 7,823 38.5
At least 18 years old on the index date 2,685 29.2 7,752 38.1
Not treated with both CCO and medicinal honey 2,639 28.7 4,512 22.2
Healed within 2 weeks or had only one visit record in USWR 2,639 28.7 4,512 22.2
CCO cohort before matching 2,297 25.0 3,993 19.6
Medicinal honey cohort before matching 342 3.7 519 2.6
CCO cohort after matching 446 4.8 517 2.5
Medicinal honey cohort after matching 341 3.7 517 2.5
a
PU diagnosis required one of these ICD-9-CM diagnosis codes: 707.00–707.07, 707.09, or 707.20–707.25.
b
The first found PU diagnosis is the patient’s index date.
CCO, clostridial collagenase ointment; PU, pressure ulcer; USWR, U.S. Wound Registry.

had 0% granulation (CCO 77.2%, honey 77.0%) (SD) total visits were required by CCO-treated PUs
with >4 concurrent wounds (CCO 56.7%, honey (CCO 9.1 [9.9] visits, honey 12.6 [16.6] visits;
51.6%). Wound age was most commonly <28 days p < 0.001), with treatment administered at an av-
(CCO 46.8%, honey 43.7%), with mean (SD) wound erage (SD) of 50.6% (31.5%) and 41.8% (31.1%) of
depth of 0.5 (0.8) cm, and minimal exudate (CCO visits, respectively ( p = 0.085). These observations
42.7%, honey 46.4%). PUs were predominantly at remained relatively consistent for each of the sub-
stage III (CCO 56.1%, honey 55.3%) with an in- sets of patients who achieved 100% granulation at
fection surrogate score of 0 (CCO 40.0%, honey 1 year and those with epithelialization at 1 year.
36.6%) and found most commonly in the sacrum or
buttock areas. Granulation results at 1 year
Honey-treated PUs underwent significantly more A significantly greater percentage of CCO-
selective sharp debridements, with 51.5% receiving treated PUs achieved 100% granulation at 1 year
a total of three or more compared with 29.6% of (CCO 42.0%, honey 35.2%; p = 0.025) (Table 5). PUs
CCO PUs ( p < 0.001) and similarly higher mean treated with CCO were 38% more likely to achieve
(SD) numbers of selective sharp debridements 100% granulation at 1 year compared to honey-
(CCO 2.7 [5.2], honey 4.4 [5.8]; p < 0.001), as well as treated PUs based on logistic regression modeling
wounds receiving negative pressure wound ther- (OR 1.384, 95% confidence limit [CL] 1.057–1.812,
apy (CCO 29.0%, honey 38.3%; p = 0.002) (Table 4). p = 0.018) (Table 6).
Patients were treated (days of use) for a mean The Cox regression found that CCO-treated
(SD) of 34.0 (54.2) days and 33.6 (60.2) days, re- PUs had a 30% significantly higher probability of
spectively ( p = 0.904). Significantly fewer mean achieving 100% granulation compared to honey-
treated PUs, with a HR of 1.302 (95% CL 1.056–
Table 2. Patient demographic and clinical characteristics
1.605, p = 0.013). CCO-treated PUs achieved 100%
Demographic and Clostridial Medicinal Standardized granulation within a significantly shorter mean
Clinical Characteristics Collagenase Honey p Differencea (SD) time frame compared to honey (CCO 255.3
No. of patients 446 341 [129.8] days, honey 282.2 [127.4] days; p < 0.001)
Age, mean (SD) 66.2 (20.3) 63.6 (20.5) 0.072 N/A (Table 5). Figure 1 graphically shows the prob-
Age group 0.212 ability of achieving 100% granulation during
18–40 14.1% 17.0% 7.96 the 1-year follow-up, comparing CCO and honey
41–64 26.9% 30.2% 7.31
65+ 59.0% 52.8% 12.48
cohorts.
Sex 0.469
Female 47.8% 45.2% Epithelialization results at 1 year
Male 52.2% 54.8%
Logistic regression modeling of epithelialization
Ethnicity 0.027 at 1 year (Table 7) found that CCO-treated PUs
Caucasian 66.4% 73.0%
African American 12.1% 9.4% were 47% more likely to epithelialize compared to
Hispanic 12.3% 6.7% honey-treated PUs (OR 1.467, 95% CL 1.051–
Other/Unknown 9.2% 10.9% 2.047, p = 0.024). Significantly higher proportions
a
Used in propensity score matching. of PUs treated with CCO achieved epithelialization
N/A, not applicable; SD, standard deviation. at 1 year (28.2% vs. 21.3%, p = 0.009) (Table 5).
130 GILLIGAN ET AL.

Table 3. Wound demographic and clinical characteristics Table 4. Treatment characteristics

Clostridial Medicinal Standardized Clostridial Medicinal

Wound Characteristics Collagenase Honey p Differencea Treatment Characteristic Collagenase Honey p

Total wounds 517 517 All wounds, n (%) 517 (100) 517 (100)
Granulation % at first visit, 0.941 Days to first application, mean (SD) 18.1 (65.3) 53.6 (143.2) <0.001
mean (SD) Days of use, mean (SD) 34.0 (54.2) 33.6 (60.2) 0.904
0% 77.2% 77.0% Total visits, mean (SD) 9.1 (9.9) 12.6 (16.6) <0.001
0.1–50% 11.6% 9.5% Treated visits, mean (SD) 3.3 (3.6) 3.1 (2.8) 0.201
>50% 11.3% 13.5% % Treated visits, mean (SD) 50.6 (31.5) 41.8 (31.1) 0.085

No. of concurrent wounds 0.289 No. of debridements–total <0.001

1 18.0% 18.2% 0.50 0.0 39.7% 21.7%
2 11.6% 14.9% 9.71 1.0 18.8% 13.7%
3 13.7% 15.3% 4.39 2.0 12.0% 13.3%
4+ 56.7% 51.6% 10.11 3.0–4.0 12.2% 19.9%
‡5.0 17.4% 31.3%
Wound age (days) at first visit 0.787
0–28 46.8% 43.7% 6.22 No. of debridements per 4 weeks <0.001
28.1–75 22.4% 23.8% 3.21 <2.0 80.7% 67.3%
75.1–182.5 12.8% 13.9% 3.41 2.0–3.9 14.5% 23.6%
>182.5 18.0% 18.6% 1.50 ‡4.0 4.8% 9.1%
Wound surface area (sqcm) 0.777 Concurrent therapies 0.042
at first visit Negative pressure wound therapy 29.0% 38.3%
£0.5 21.9% 20.3% 3.79 Hyperbaric oxygen therapy 2.1% 1.2%
0.6–2.0 21.7% 22.1% 0.94 Becaplermin gel or cellular tissue products 1.4% 4.4%
2.1–10 32.1% 31.9% 0.41
>10 18.2% 20.7% 6.36
Missing 6.2% 5.0% 5.05 DISCUSSION
Wound depth (cm) at first visitb, 0.5 (0.8) 0.5 (0.8) 0.767 This comparative analysis of enzymatic de-
mean (SD)
PU stage 0.989
bridement with CCO and autolytic debridement
Unstageable 3.3% 3.1% 1.10 with support from medicinal honey found for the
II 13.5% 13.9% 1.12 primary outcomes of granulation and epithelial-
III 56.1% 55.3% 1.56 ization that CCO-treated wounds were 38% more
IV 27.1% 27.7% 1.30
likely to achieve 100% granulation at 1 year and
PU location 0.086
Sacrum/buttock 32.3% 36.6%
in a significantly shorter time frame (255 vs.
Other site 25.7% 28.6% 282 days) and 47% more likely at 1 year to achieve
Heel 19.7% 15.7% epithelialization also in shorter time frames (289 vs.
Hip/thigh 8.5% 6.8% 308 days) compared to matched PUs treated with
Ankle 7.5% 5.0%
Back 6.2% 7.4%
honey. Healthcare resource use was also lower for
Wound exudate level 0.048
CCO-treated PUs compared to honey with signifi-
at first visit cantly fewer total visits (9.1 vs. 12.6), fewer total
None 27.1% 20.5% selective sharp debridements (2.7 vs. 4.4), and fewer
Minimal 42.7% 46.4% PUs receiving negative pressure wound therapy
Moderate 26.3% 27.1%
Heavy 3.9% 6.0%
(29% vs. 38%). These findings of more complete
and faster granulation and epithelialization with
a
Used in propensity score matching.
b
Missing results for 12% of clostridial collagenase and 16.8% of
medicinal honey cohorts
sqcm, square centimeter. Table 5. Observed granulation and epithelialization at 1 year

Clostridial Medicinal
The Cox regressions found that CCO-treated Collagenase Honey
PUs had a 42% higher probability of epithelializa- Outcomes n = 517 n = 517 p
tion at 1 year (HR 1.421, 95% CL 1.0909–1.8516, 100% Granulation at 1 year, n (%) 217 (42.0) 182 (35.2) 0.025
p = 0.009). Figure 2 contrasts the probability of Weeks to 100% granulation at 1 year, 36.5 (20.0) 40.3 (18.2) 0.001
CCO and honey PUs achieving epithelialization mean (SD)
Days to 100% granulation at 1 year, 255.3 (129.8) 282.2 (127.4) <0.001
during the 1-year follow-up period. The mean (SD) mean (SD)
number of days for achieving epithelialization was Epithelialization at 1 year, n (%) 146 (28.2) 110 (21.3) 0.009
significantly lower for CCO-treated PUs treated Weeks to epithelialization at 1 year, 41.2 (18.4) 44.0 (16.7) 0.011
with CCO than those treated with honey at 1 year mean (SD)
Days to epithelialization at 1 year, 288.6 (128.9) 308.1 (116.6) 0.011
(288.6 [128.9] days vs. 308.1 [116.6] days, p = 0.011; mean (SD)
Table 5).
 CLOSTRIDIAL COLLAGENASE VERSUS MEDICINAL HONEY 131

Table 6. Logistic regression for 100% granulation at 1 year methods requiring the shortest durations and
Lower Upper fewest clinical visits and, thereby, the two most
PU Treatment Characteristic OR 95% CI 95% CI p cost-effective debridement methods. Autolytic de-
Clostridial collagenase (vs. honey) 1.384 1.057 1.812 0.018
bridement was said to take the longest to work,
Patient has home healthcare 1.934 1.463 2.557 <0.001 relying on the patient’s own cellular mechanisms to
Wound age >182.5 daysa 0.656 0.438 0.983 0.041 remove necrotic tissue.17
Infection surrogate score 3–6a 0.623 0.416 0.932 0.021 There has been a paucity of comparative studies
Infection surrogate score 2a 0.517 0.325 0.824 0.005
2.0–3.9 debridements/4 weeksa 0.609 0.414 0.896 0.012
with medicinal honey, and this study is the first we
‡4.0 debridements/4 weeksa 0.265 0.125 0.563 0.001 are aware of comparing enzymatic debridement to
PU location—heela 0.567 0.374 0.860 0.008 autolytic debridement with medicinal honey. A
Patient age 80+ yearsa 0.505 0.323 0.790 0.003 number of studies have compared enzymatic de-
Covariates shown are those significantly ( p £ 0.05) affecting the odds of bridement with CCO to other debridement meth-
100% granulation.
a
ods, primarily other autolytic modalities, in terms
Reference values: wound age 0–28 days, infection surrogate score 0,
debridements/4 weeks <2.0, PU location buttock, patient age 18–40 years. of both treatment efficacy and cost. A US-based,
CI, confidence interval; OR, odds ratio. prospective, randomized, parallel-group, open-
label exploratory study of diabetic foot ulcers with
CCO (vs. honey) bear important implications from 55 participants receiving CCO plus serial sharp
a clinical perspective, as well as impacts to the debridement compared with supportive care plus
economic burden of this costly adverse health serial sharp debridement found that CCO-treated
condition. ulcers healed more quickly and completely while
Medical complications that can develop from costing *$300 less.40 An economic analysis based
chronic wound PUs can wreak havoc on patients’ on a RCT of PU therapy in a long-term care setting
health and quality of life. Mechanisms to support found CCO more cost-effective relative to autolysis
faster and more complete PU healing may prevent with hydrogel dressings.41 Muller et al. in an RCT
these sequelae from occurring sooner, providing more of patients, with grade IV heel pressure sores,
days of relief for the patient and their care team. found that 91.7% of the collagenase patients were
PUs can impose on the patient, their caregivers, treated successfully compared with 63.6% of the
and healthcare professionals a costly cycle of hos- patients in the hydrocolloid group ( p < 0.005) with
pitalizations, clinic visits, and home healthcare more rapid closure, further showing collagenase
requiring extensive management resources from a treatment to be more cost-effective than the
trained multidisciplinary team. In their 2015 hydrocolloid treatment owing to the shorter closure
study, Woo et al. found the cost of debridement time.42 An RCT of long-term care facility PU pa-
dependent on the resources required over the tients comparing CCO to autolysis demonstrated
length of time needed to achieve a clean wound bed, that at 42 days of therapy 85% of the CCO-treated
with surgical sharp and enzymatic debridement patients achieved a clean wound bed without initial

Figure 1. Probability of 100% granulation over 365 days (Kaplan–Meier curve).

132 GILLIGAN ET AL.

Table 7. Logistic regression for epithelialization at 1 year after treatment initiation with greater wound clo-
Lower Upper sure at 1 and 2 years (22% vs. 11% after 1 year and
PU Treatment Characteristic OR 95% CL 95% CL p 27% vs. 14% after 2 years, respectively).32 Our re-
Clostridial collagenase (vs. honey) 1.467 1.051 2.047 0.024
sults in terms of achieving 100% granulation and
2.0–3.9 visits/4 weeksa 1.547 1.025 2.335 0.038 epithelialization at 1 year were similar in both the
Patient has home healthcare 1.422 1.005 2.014 0.047 proportion of patients and the time to closure for
Male 0.689 0.488 0.972 0.034 CCO in these studies and for honey as an autolytic
Wound depth 0.1–0.3 cma 0.620 0.407 0.946 0.027
Wound depth ‡0.3 cma 0.559 0.339 0.921 0.022
support method.
Wound age >182.5 days 0.530 0.308 0.911 0.022 This analysis examined outcomes of using en-
Infection surrogate score 2a 0.524 0.303 0.937 0.029 zymatic debridement and autolytic debridement
Infection surrogate score >6a 0.454 0.237 0.869 0.017 support with medicinal honey. The choice of de-
Age 80+ yearsa 0.493 0.270 0.899 0.021
Wound surface area >10.0 sqcma 0.471 0.258 0.861 0.014
bridement using any of the available methods de-
Ethnicity—othera 0.446 0.230 0.866 0.017 pends on the patient’s medical condition and the
‡4 debridements/4 weeks 0.417 0.176 0.988 0.047 condition of each wound. For each patient, these
Patient has paralysis 0.375 0.176 0.800 0.011 are unique. The clinician’s decision requires con-
Arrival score 10a 0.288 0.092 0.896 0.032
sideration of the patient’s goals and condition, the
Covariates shown are those significantly ( p £ 0.05) affecting the odds of wound characteristics, the care setting and the
100% granulation.
a
Reference values: visits/4 weeks <2, wound depth £0.1 cm, wound age
talents of the other care team members, and im-
0–28 days, infection surrogate score 0, patient age 18–40 years, wound portantly, evidence of the efficacy and limitations
surface area £0.5 sqcm, ethnicity white, debridements/4 weeks <2.0, and of the various modalities available.
arrival score 0.
Limitations
or concomitant sharp debridement compared with The USWR EHR data used in this study were
29% in the autolysis arm.43 In a 1999 study by collected from over 100 contributing hospital out-
Mosher of elderly long-term care residents, CCO patient wound care clinics’ billing records to support
compared to autolysis, wet-to-dry dressings (me- reimbursement and were not collected for research
chanical debridement), or fibrinolysin debridement purposes. There may be variability in the data re-
had the highest probability of achieving a clean ported among the various contributing clinics. Di-
wound bed, low infection rate, and lowest total agnoses or procedures may be subject to coding
cost.44 A recent observational, retrospective study error, for which the extent of miscoding or under-
compared treatment of stage IV PUs using enzy- coding that could result in bias is unknown, and may
matic debridement plus selective debridement also result in measurement error in ICD-9-CM or
with selective debridement alone, finding the CCO CPT based variables. Analyses using text fields in-
treatment arm significantly more likely to achieve stead of discrete numeric data may be subject to
greater and faster epithelialization 1 and 2 years error resulting from incorrect interpretation of the

Figure 2. Probability of complete epithelialization over 365 days (Kaplan–Meier curve).

 CLOSTRIDIAL COLLAGENASE VERSUS MEDICINAL HONEY 133

data. Patient medical history was limited to

KEY FINDINGS
the patients’ visits to a specific provider and
This observational study examined the clinical impact of enzymatic de-
clinic, such that patients’ outpatient treat-
bridement with collagenase ointment relative to autolytic debridement with
ment for other comorbidities and by other
medicinal honey for management of PUs.
clinics (e.g., non-USWR clinics) was un-
known. Patients with a qualifying PU di-  Enzymatic debridement with collagenase ointment provided more com-
agnosis before the study period may have plete granulation of PUs at 1 year compared with autolytic debridement
been incorrectly categorized as incident. using medicinal honey.
Multivariate analyses of outcomes provided  Epithelialization at 1 year was more complete and occurred faster for
further adjustment for potentially con- patients using collagenase ointment than those using medicinal honey.
founding covariates used in the matching  Enzymatic debridement of PUs with collagenase ointment offered clini-
process and an additional degree of ro- cally significant advantages over medicinal honey in all stages of PUs.
bustness in cohort comparisons; however,
there is always the potential for unmea-
sured confounders, especially since infor-
mation such as comorbidity severity and other Health Analytics). This study was sponsored by
sociodemographic variables (e.g., socioeconomic Smith & Nephew, Fort Worth, TX.
status, basal metabolic index, smoking status, and
so on) were unavailable. The USWR EHR is a con-
venience sample of contributing clinics treating PUs AUTHOR DISCLOSURE AND GHOSTWRITING
in the US and so findings in this study may not be A.M.G., R.B., and B.-C.C. are employees of
generalizable to other types of chronic wounds or to Truven Health Analytics, which was paid by Smith
patient populations outside of the wound care clinics & Nephew in connection with the development of
contributing data to the USWR EHR. this article. M.J.C. and C.E.F. were paid consul-
tants to Smith & Nephew in connection with the
INNOVATION (CONCLUSIONS) development of this article. C.R.W. is an employee
of Smith & Nephew.
This is the first retrospective observational
study comparing the clinical impact of enzymatic
debridement using CCO with autolytic debride- ABOUT THE AUTHORS
ment support using medicinal honey for PU man- Adrienne M. Gilligan, PhD, is a Researcher at
agement. It found CCO-treated PUs significantly Truven Health Analytics. She is also an adjunct
more likely to achieve 100% granulation and epi- assistant professor at the University of North
thelialization at 1 year and at significantly faster Texas Health Sciences Center College of Pharmacy.
rates compared to honey-treated PUs, with fewer Richard Bizier, BS, is a Senior Programmer
healthcare visits, less debridement, and less use of Analyst at Truven Health Analytics. Bong-Chul
negative pressure wound therapy. Enzymatic de- Chu, PhD, is a Senior Biostatistician at Truven
bridement with CCO provided statistically signifi- Health Analytics. Marissa J. Carter, PhD, MA, is
cant greater clinical benefit than honey for treating the Chief Executive Officer for Strategic Solutions,
PUs at all PU stage levels. Future research evalu- Inc. Caroline E. Fife, MD, is the Chief Medical
ating cost-effectiveness of CCO compared to honey Officer at Intellicure, Inc. and the Medical Director,
in this population is warranted. Wound Clinic at St. Luke’s Hospital. Curtis R.
Waycaster, RPh, PhD, is the Director of Health
ACKNOWLEDGMENTS Economics and Outcomes Research at Smith &
AND FUNDING SOURCES Nephew. He is also an adjunct assistant professor
Editorial/medical writing support for this article at the University of North Texas Health Sciences
was provided by Jay Margolis, PharmD (Truven Center College of Pharmacy.

REFERENCES
1. Preventing Pressure Ulcers in Hospitals: Are We hospital/pressureulcertoolkit/putool1.html (last vention of pressure sores in burn patients. J Burn
Ready for This Change? A toolkit for improving accessed October 28, 2016). Care Rehabil 2004;25:388–410.
quality of care. Content last reviewed October 2014.
Agency for Healthcare Research and Quality, 2. Gordon MD, Gottschlich MM, Helvig EI, et al. 3. Kuhn BA. Balancing the pressure ulcer cost and
Rockville, MD. www.ahrq.gov/professionals/systems/ Review of evidence-based practice for the pre- quality equation. Nurs Econ 1992;10:353–359.
134 GILLIGAN ET AL.

4. Berlowitz DR, Brandeis GH, Anderson J, Du W, 20. Steed DL, Donohoe D, Webster MW, Lindsley L. facility level-of-service coding in outpatient wound
Brand H. Effect of pressure ulcers on the survival Diabetic ulcer study group. Effect of extensive centers: results of a multicenter study. Ostomy
of long-term care residents. J Gerontol A Biol Sci debridement and treatment on the healing of dia- Wound Manage 2012;58:20–22, 24, 26–28.
Med Sci 1997;52A:M106–M110. betic foot ulcers. J Am Coll Surg 1996;183:61–64.
36. U.S. Wound Registry. The largest complete patient
5. Thomas DR, Goode PS, Tarquine PH, Allman RM. 21. Levin ME. Pathogenesis and general management registry of chronic wound care data in the world.
Hospital-acquired pressure ulcers and risk of of foot lesions in the diabetic patient. In: Bowker https://www.uswoundregistry.com/index.aspx (last
death. J Am Geriatr Soc 1996;44:1435–1440. JH, Pfeifer MA, eds. Levin and o’Neal’s the Dia- accessed October 12, 2016).
betic Foot. St. Louis, MO: Mosby, 2001:219–260.
6. National Pressure Ulcer Advisory Panel Board of 37. Horn SD, Torres A, Jr., Willson D, Dean JM,
Directors. Pressure ulcers in America: prevalence, 22. Ramundo J, Gray M. Collagenase for enzymatic Gassaway J, Smout R. Development of a pediatric
incidence, and implications for the future: an debridement: a systematic review. J Wound Ost- age- and disease-specific severity measure. J
executive summary of the National Pressure Ulcer omy Continence Nurs 2009;36(Suppl 6):S4–S11. Pediatr 2002;141:496–503.
Advisory Panel Monograph. Adv Skin Wound Care
23. Beitz JM. Wound debridement: therapeutic op- 38. D’Agostino RB, Jr. Propensity score methods for
2001;14:208–215.
tions and care considerations. Crit Care Nurs Clin bias reduction in the comparison of a treatment to
7. Lyder CH. Pressure ulcer prevention and man- North Am 2012;24:239–253. a non-randomized control group. Stat Med 1998;
agement. JAMA 2003;289:223–226. 17:2265–2281.
24. Moore J, Jensen P. Assessing the role and impact
8. Graves N, Birrell F, Whitby M. Effect of pressure of enzymatic debridement. Podiatry Today 2004;17: 39. Baser O. Too much ado about propensity score
ulcers on length of hospital stay. Infect Control 54–61. models? Comparing methods of propensity score
Hosp Epidemiol 2005;26:293–297. matching. Value Health 2006;9:377–385.
25. Collagenase Santyl Ointment. Prescribing In-
9. Carter MJ. Cost-effectiveness research in wound formation. Fort Worth, TX: Smith & Nephew, Inc., 40. Motley TA, Gilligan AM, Lange DL, Waycaster CR,
care: definitions, approaches, and limitations. 2014. Dickerson JE. Cost-effectiveness of clostridial
Ostomy Wound Manage 2010;56:48–59. collagenase ointment on wound closure in pa-
26. Shi L, Carson D. Collagenase santyl ointment: a
tients with diabetic foot ulcers: economic analysis
10. National Pressure Ulcer Advisory Panel. selective agent for wound debridement. J Wound
of results from a multicenter, randomized, open-
NPUAP pressure injury stages. NPUAP 2016 Sta- Ostomy Continence Nurs 2009;36 (6 Suppl):S12–
label trial. J Foot Ankle Res 2015;8:7.
ging Consensus Conference. www.npuap.org/ S16.
resources/educational-and-clinical-resources/npuap- 41. Waycaster C, Milne CT. Economic and clinical
27. Karagol BS, Okumus N, Dursun A, Karadag N,
pressure-injury-stages/ (last accessed October 28, benefit of collagenase ointment compared to a hy-
Zenciroglu A. Early and successful enzymatic de-
2016). drogel dressing for pressure ulcer debridement in a
bridement via collagenase application to pinna in
long-term care setting. Wounds 2013;25:141–147.
11. National Pressure Ulcer Advisory Panel, European a preterm neonate. Pediatr Dermatol 2010;28:
Pressure Ulcer Advisory Panel and Pan Pacific 600–601. 42. Muller E, van Leen MW, Bergemann R. Economic
Pressure Injury Alliance. Prevention and Treat- evaluation of collagenase-containing and hydro-
28. Shi L, Ramsay S, Ermis R, Carson D. pH in the
ment of Pressure Ulcers: Quick Reference Guide. colloid dressing in the treatment of pressure ul-
bacteria-contaminated wound and its impact on
Haesler E, ed. Osborne Park, Australia: Cambridge cers. Pharmacoeconomics 2001;19:1209–1216.
clostridium histolyticum collagenase activity: im-
Media, 2014. www.npuap.org/wp-content/
plications for the use of collagenase wound de- 43. Milne CT, Ciccarelli AO, Lassy M. A comparison
uploads/2014/08/Quick-Reference-Guide-DIGITAL-
bridement agents. J Wound Ostomy Continence of collagenase to hydrogel dressings in wound
NPUAP-EPUAP-PPPIA-Jan2016.pdf (last accessed
Nurs 2011;38:514–521. debridement. Wounds 2010; 22:270–274.
October 28, 2016).
29. Marazzi M, Stefani A, Chiaratti A, Ordanini MN, 44. Mosher BA, Cuddigan J, Thomas DR, Boudreau
12. Phillips TJ, Ouahes N. Leg ulcers. Curr Probl
Falcone L, Rapisarda V. Effect of enzymatic de- DM. Outcomes of 4 methods of debridement using
Dermatol 1995;7:109–142.
bridement with collagenase on acute and chronic a decision analysis methodology. Adv Wound Care
13. Falanga V. Classifications for wound bed prepa- hard-to-heal wounds. J Wound Care 2006;15: 1999;12:81–88.
ration and stimulation of chronic wounds. Wound 222–227.
Repair Regen 2000;8:347–352.
30. Ramundo J, Gray M. Enzymatic wound debride- Abbreviations and Acronyms
14. Brem H, Balledux J, Sukkarieh T, Carson P, Fa- ment. J Wound Ostomy Continence Nurs 2008;35:
AHRQ ¼ Agency for Healthcare Research
langa V. Healing of venous ulcers of long duration 273–280.
and Quality
with a bilayered living skin substitute: results
31. Ostlie DJ, Juang D, Aguayo P, et al. Topical silver CCO ¼ clostridial collagenase ointment
from a general surgery and dermatology depart-
sulfadiazine vs collagenase ointment for the treat- CL ¼ confidence limit
ment. Dermatol Surg 2001;27:915–919.
ment of partial thickness burns in children: a pro- CPT ¼ current procedural terminology
15. Schultz GS, Sibbald RG, Falanga V, et al. Wound spective randomized trial. J Pediatr Surg 2012;47: EHR ¼ electronic health record
bed preparation: a systematic approach to wound 1204–1207. ESRD ¼ end-stage renal disease
management. Wound Repair Regen 2003;11:1–28. HIPAA ¼ Health Insurance Portability
32. Carter MJ, Gilligan AM, Waycaster CR, Fife CE.
and Accountability Act
16. Sibbald RG, Woo K, Ayello E. Wound bed prepa- Treating pressure ulcers with clostridial collage-
HR ¼ hazard ratio
ration: DIM before DIME. Wound Heal S Afr nase ointment: results from the US Wound Reg-
ICD-9-CM ¼ International Classification
2008;1:29–34. istry. Wound Repair Regen 2016;24:904–912.
of Diseases, Ninth Revision,
17. Woo KY, Keast D, Parsons N, Sibbald RG, Mitt- 33. Molan P, Rhodes T. Honey: a biologic wound Clinical Modification
mann N. The cost of wound debridement: a Ca- dressing. Wounds 2015;27:141–151. OR ¼ odds ratio
nadian perspective. Int Wound J 2015;12:402–407 PU ¼ pressure ulcer
34. Molan PC. The evidence and the rationale for the
RCT ¼ randomized clinical trial
18. Anderson I. Debridement methods in wound care. use of honey as a wound dressing. Wound Pract
SD ¼ standard deviation
Nurs Stand 2006;22:65–72. Res 2011;19:204–221.
sqcm ¼ square centimeter
19. Steed DL. Debridement. Am J Surg 2004;187: 35. Fife CE, Wall V, Carter MJ, Walker D, Thomson B. USWR ¼ U.S. Wound Registry
71S–74S. Examining the relationship between physician and