Dermatologic Therapy, Vol.

19, 2006, 383390

Printed in the United States All rights reserved

Copyright Blackwell Publishing, Inc., 2006

DERMATOLOGIC THERAPY
ISSN 1396-0296

Wound bed score and its

correlation with healing of chronic
wounds
Blackwell Publishing Inc

VINCENT FALANGA*, LILIANA J. SAAP*, & ALEXANDER OZONOFF

*Department of Dermatology and Skin Surgery, Roger Williams Medical
Center, Providence, Rhode Island, Departments of Dermatology and
Biochemistry and Department of Biostatistics, Boston University, Boston,
Massachusetts

ABSTRACT: Adequate wound bed preparation is essential for healing of leg ulcers, and consists of
controlling exudate and edema, decreasing the bacterial burden, promoting healthy granulation
tissue, and removing necrotic tissue. Currently, there is no classification system for wound bed preparation that has predictive value. Based on past work and the authors experience, we have now
developed and tested a new classification system that scores the following parameters: healing edges
(wound edge effect), presence of eschar, greatest wound depth/granulation tissue, amount of exudate
amount, edema, peri-wound dermatitis, peri-wound callus and or fibrosis, and a pink/red wound
bed. Each parameter receives a score from 0 (worst score) to 2 (best score), and all the parameter
scores are added for a total score. Each wound can have a maximum score of 16 (the best score
possible), to a minimum score of 0 (the worst score possible). We used this wound bed score (WBS)
system in a study of 177 patients with venous ulcers who had been prospectively treated with and
randomized to either conventional therapy (compression alone) or a living bilayered skin construct
(BSC). We evaluated serial photographs at baseline to determine whether the results would be predictive of complete wound closure and could validate the WBS. We found that wounds that ultimately
achieved full closure had a statistically significant higher WBS than those that did not heal
(p = 0.0012). This was also true when separating wounds by treatment modality: standard therapy
(p = 0.044) and treatment with a BSC (p = 0.011). When dividing the WBS in the following quartile
groups: scores 410, 1012, 1213, and 1316, the percentage of healed wounds correlated with the
WBS (p = 0.0008). For all wounds, a one unit increase in total WBS resulted on average in a 22.8%
increase in odds of healing (OR = 1.228). This WBS seems to have validity in predicting complete
wound closure in wounds treated with either standard therapy or advanced modalities, such as BSC.
If confirmed and widely adopted in this and other types of wounds, it could be a useful tool in both
the clinical and research setting.
KEYWORDS: bioengineered skin, prognosis, skin equivalents, wound, wound bed preparation, wound
healing

Introduction
Address correspondence and reprint requests to: Vincent
Falanga, MD, FACP, Professor of Dermatology and
Biochemistry, Department of Dermatology, Roger Williams
Medical Center, 50 Maude Street, Providence, RI 02908, or
email: vfalanga@bu.edu.
Dr Saap is presently with Affiliated Dermatology Cosmetic
Surgery Center, Dublin, OH.

It is estimated that in the United States, between

400,000 and 500,000 patients are affected by
venous leg ulcers. However, the true prevalence is
probably higher and is likely to become larger
with the ever-increasing elderly population.
Venous leg ulcers account for 8090% of all
chronic wounds of the lower extremity (1). Adequate

383

Falanga et al.

wound bed preparation is essential for healing

of chronic wounds, and consists of controlling
exudate and edema, decreasing the bacterial
burden, promoting healthy granulation tissue,
and removing necrotic tissue (13). Currently, there
is no classification system for wound bed preparation that can predict ultimate wound closure
(3). In 2000, we pioneered the concept of wound
bed preparation and proposed a classification
system for it (35). This scoring system consisted
of a wound bed appearance score and an exudate
score that were added together to give each
wound a total score (Table 1). For example, a score
of B2 would signify a granulation tissue between
50% and 100% and a moderate amount of exudate. The appearance score took into account the
amount of granulation tissue, fibrinous material,
and eschar. That classification system remains
very useful for describing the wound bed. However, we wanted to know whether it would also
have predictive value for ultimate wound closure.
Therefore, in further work (6) we applied that
classification system to venous ulcers treated either
with standard therapy or a bilayered bioengineered
skin construct (BSC). The data showed that
minimal exudate and lack of an eschar closely
correlated with increased healing. However, we
found that the extent of granulation tissue and
fibrinous material were not as helpful and did not
add substantially to the predictive value of that
scoring system.
We have now devised a wound bed score (WBS)
that reflects our previous work and clinical experience. The WBS includes the following parameters
that are not limited to the absolute wound bed alone

and also includes the surrounding diseased skin:

(a) healing edges (wound edge effect); (b) black
eschar; (c) greatest wound depth/granulation tissue; (d) amount of exudate, edema, peri-wound
dermatitis, peri-wound callus and or fibrosis; and
(e) a pink wound bed (Table 2). Each individual
parameter receives a score from 0 (worst score) to
2 (best score), and all the parameter scores are
added for a total score. A wound can have a
maximum WBS of 16 (the best possible score), to
a minimum WBS of 0 (the worst possible score).
Some representative examples are shown in
FIGS 1 and 2. In this report, we applied this scoring system to a group of patients with venous
ulcers randomized to treatment over a 24-week
period with either conventional therapy (compression bandages alone) or also receiving a living
BSC. The results of this study are very promising,
showing that the WBS is valid in assessing wound
bed preparation and predicting whether ultimate
wound closure occurs.

Patients and methods

Patients
The details of the clinical trial used for the present
analysis of venous ulcers have been previously
published (4). We made use of data generated
from the pivotal controlled, randomized trial
that led to Food and Drug Administration approval
of a bilayered BSC (also known as Apligraf, Organogenesis, Inc., Canton, MA) in the treatment venous
ulcers (4). Photographs from a total of 177 venous

Table 1. Initial Wound Bed Preparation Score Developed in 2000a

Wound bed characteristics

Wound bed
appearance score
A
B
C
D

Granulation tissue (%)

Fibrinous tissue

Eschar

100
50100
< 50
Any amount

+
+
+

Wound exudate
score

Extent of
control

Exudate
amount

Dressing
requirement

Fully

None/minimal

2
3

Partially
Uncontrolled

Moderate amount
Very exudative wound

No absorptive dressings required

If feasible, dressings could stay on for up to a week
Dressing changes required every 23 days
Absorptive dressing changes required at least daily

Adapted from Falanga V. Classifications for wound bed preparation and stimulation of chronic wounds. Wound Rep Reg 2000;
8 : 347352. Copyright V. Falanga, 2000.

384

Wound bed score for healing

Table 2. New Wound Bed Score ( WBS) and Its Individual Featuresa
Wound bed score
characteristics
Healing edges
Black eschar
Greatest wound
Depth/granulation
Tissue
Exudate amount
Edema
Peri-wound
Dermatitis
Peri-wound
Callus/fibrosis
Pink wound bed

None
> 25% of wound
surface area
Severely depressed or
raised when compared to
peri-wound skin
Severe
Severe
Severe

25 75%
0 25%

> 75%
None

Moderate

Flushed or
almost even

Moderate
Moderate
Moderate

Severe

Moderate

None

50 75%

None/mild
None/mild
None or
minimal
None or
minimal
> 75%

The total WBS adds each individual score for each characteristic to give a total score. The maximum possible score (best score) is
16. The minimum possible score (worst score) is 0.

FIG. 1. Steep wound edges. This is an example of a venous

ulcer with pink and adequate granulation tissue but with
steep edges. This feature is strongly associated with impaired
healing. No eschar is present.

ulcer patients at day 1 were evaluated and scored.

Of these patients, 83 were treated with standard
accepted conventional therapy (compression alone)
and 94 were treated with BSC. The inclusion criteria
for these patients were the following: ages 1885;
venous insufficiency, history of nonhealing venous
ulcers of longer than 1 months duration; and
venous ulcers extending through the epidermis
into dermal tissue, but without exposed bone or
tendon. The main exclusion criteria were: venous
ulcer size area of less than 1 cm2 or more than
10 20 cm, arterial insufficiency (ABI < 0.65), and
medical conditions known to impair healing (4).
Informed consent was obtained of all patients
in accordance to each study centers human

FIG. 2. Healing wound edges (edge effect). This is a venous

ulcer that is going to heal. The thick oval shows a portion of
the wound demonstrating the edge effect whereas the thin
oval shows a part of the wound with no edge effect. This
particular wound has no black eschar. The thick black arrow
points to the wound bed that is raised in comparison to the
peri-wound skin, whereas the thin black arrow points to the
wound bed that is flush with the peri-wound skin. This wound
also shows an excellent example of peri-wound fibrosis as
demonstrated by the blue arrow.

research committee, and the study protocols

conformed to the ethical guidelines of the 1975
Declaration of Helsinki.
Treatment and evaluation
Evaluation of the photographs was performed by
a blinded observer who had had no involvement
in the trial. The photographs to be evaluated were

385

Falanga et al.

those taken at day 1 of the study, prior to assignment to either of the two treatment arms. The
control venous ulcer patients were treated with
standard therapy consisting of a nonadherent
primary dressing (Tegapore, registered trademark
3M Health Care, St Paul, MN), a secondary gauze
pressure bolster, a zinc paste bandage (Unna
boot), and a self-adherent elastic wrap (Coban,
3M Health Care); or with the BSC applied directly
to the wound, followed by the nonadherent
primary dressing, cotton gauze dressing folded
as a bolster, and the same elastic wrap. Patients
received up to five applications of BSC within the
first 21 days of the study. Re-application was
performed if less than 50% of the wound had reepithelialized (4). The original trial had a total of
146 patients treated with the BSC and 129 in the
control group. However, evaluable photographs
for the purpose of the present study were only
available from 83 control patients and 94 BSCtreated patients. This was either the result of
pictures not being taken at day 1 or patients leaving
the study prematurely so that healing or lack of
healing could not be assessed. Patients were
followed for a total of 12 months, and healing was
determined as 100% epithelialization with no
wound drainage by week 24.
Statistical analysis
The WBS was applied to each patient photograph,
as explained earlier. Healing percentages for each
score were determined by using the case report
forms provided for each patient. The independent
samples t-test was used to compare mean WBS
between healed and not healed wounds in all
patients, including control and BSC-treated
patients. This test was also used to compare mean
WBS between wounds that healed and those that
did not heal based on initial wound duration of
either less than 1 year to greater than 1 year prior
to enrollment into the protocol. The CochranArmitage trend test was used to evaluate the
association and trend between the WBS and complete wound closure. Logistic regression was used
to model total WBS as a predictor for healing status
(as a binary outcome) and to model total WBS
prior to treatment as a predictor for the duration
of wound. Finally, probability of healing was
modeled with a stepwise logistic regression model,
which includes each individual characteristic
one at a time, with an entry/removal threshold
of p = 0.10. This procedure allowed the present
authors to assess each individual characteristic
for its predictive power in the model.

386

Results
As stated earlier, photographs had not been submitted at day 1 for all study participants. Therefore, we were not able to analyze all the patients
initially enrolled in the study trial. These dropouts were not included in the data set for analysis
as there was no complete data for these subjects
to analyze. This also means that there was no way
to assess potential differential drop-out rates for
any of the subgroups used in the analyses. For
these reasons, we decided to omit the survival
analysis, although we examined association between
time-to-heal and bed score with other methods.
Therefore, photographs were evaluated for 177
patients at day 1 before any treatment was instituted. Of these 177 patients, 94 were treated with
BSC and 83 were treated with standard therapy
(control) as described in the methods section.
Using the independent samples t-test, we
found that the WBS at baseline was higher in the
subjects who healed (WBS = 11.83) compared to
those who did not heal by the 24-week mark
(WBS = 10.62; p = 0.0012). We also saw similar
results after separating the patients according to
treatment. Mean WBS score for patients treated
with BSC was significantly higher in patients who
healed (11.87) versus those patients who did not
heal (10.59) (p = 0.0113). In patients treated with
standard therapy, mean WBS score was also significantly higher in patients who healed (11.79)
versus those who did not heal (10.66) (p = 0.0439).
We then took the total WBS for all wounds (control
and treatment groups) and divided them into the
following four groups using roughly equiprobable
cut-off values or quartiles: 410, 1012, 1213, and
1316. We found a statistically significant trend
showing higher healing percentages for the WBS in
the higher quartiles (p = 0.0008 Cochran-Armitage
trend test) (FIG. 3).
Using logistic regression to model total WBS
(for both control and BSC-treated wounds) as a
predictor for healing status, we obtained the odds
ratio of 1.228 (p = 0.002). This can be interpreted as
meaning that one unit increase in total WBS
resulted on average in a 22.8% increase in the
odds of healing. For example, increasing the total
WBS from 10 to 13 leads to a 1.228 or 1.85-fold
increase in the odds of healing. The odds ratio for
wounds treated with BSC alone was 1.253
(p = 0.015) and for standard therapy alone was 1.202
(p = 0.010). Once again, this means that a one unit
increase in WBS leads to a 25.3% and a 20.2%
increase in odds of healing for wounds treated
with BSC and standard therapy, respectively.

Wound bed score for healing

FIG. 3. Total WBS and its correlation with healing for venous
ulcers. These data were for all venous ulcers (treated either
with standard therapy or a bioengineered skin substitute).

For all patients, regardless of treatment, we

evaluated the WBS and correlated it to time to
heal. It was found that a higher WBS, although not
statistically significant, did show a trend toward
shorter healing time (r = 0.146, p = 0.15). We then
took all the patients that did heal and tested the
association between time to heal and total WBS
using the Wilcoxon rank-sum test. Time to heal
was categorized into low (= 8 weeks) and high
(> 8 weeks) time to heal groups. Once again, we
found a strong correlation trend between higher
WBS and lower time to heal (p = 0.064).
We further evaluated WBS for all wounds and
correlated it to wound duration before treatment.
Wounds with a duration of less than 1 year had
a higher mean WBS (11.81) than wounds that
had been present for more than a year (10.84)
(p = 0.0086). We estimated an odds ratio of 1.1179
(p = 0.010) when logistic regression was used to
model total WBS as a predictor for wound healing
for those wounds with a duration of less than 1
year. Once again, this can be interpreted as follows: a one-unit increase in total WBS prior to
treatment will result in an increase of 17.9% in odds
of healing.
When separating for treatment, wounds treated
with BSC also showed a statistically significant
higher mean WBS for wounds with less than 1
year in duration (12.05) than wounds present for
more than a year (10.78) (p = 0.0128). The odds ratio
was 1.260 (p = 0.016) for wounds with a healing
duration of less than 1 year. For wounds treated
with standard therapy, there was a trend, although
not statistically significant, showing higher WBS

(11.62) in wounds with less than 1 year in duration

than wounds with more than 1 year duration
(WBS = 10.92; p = 0.2156).
Noting that wound duration before treatment
is an important factor in predicting ultimate
wound closure, we decided to modify the WBS
slightly by adding wound duration (a historical
parameter) as an additional parameter (Table 3).
This was done to determine whether this would
also have an effect on total WBS and ultimate
healing. Using this modified WBS, we found that
regardless of treatment those wounds that healed
still had a higher mean WBS (13.14) than those
that did not heal (WBS = 11.15; p = 0.0001). Logistic
regression to model probability of healing estimated
an odds ratio of 1.312 (p < 0.0001), indicating that
one unit increase in total WBS results, on average,
in a 31.2% increase in odds of healing. Using this
WBS, wounds present for less than 1 year also had
a higher mean total WBS (13.79) than wounds present
for longer than 1 year (WBS = 10.86; p = 0.0001).
Logistic regression modeling estimated an odds
ratio of 1.628 (p < 0.0001) for wounds with an initial
duration that was less than 1 year.
When separating by treatment modality, wounds
treated with BSC that went on to heal had a
higher mean WBS (12.85) than those that did not
heal (WBS = 11.17; p = 0.0041). We estimated an
odds ratio of 1.25 (p = 0.006) when modeling
probability of healing for this group. Wounds with
a duration of less than 1 year also had a higher
mean WBS (14.0) than those present for longer
than a year (WBS = 10.78; p = 0.0001). We estimated
an odds ratio of 1.898 when the probability of
healing for wounds with duration of less than 1 year
was modeled.
Patients treated and healed with standard
control therapy had a higher mean total WBS
(13.46) than those who did not heal (WBS = 11.11;
p = 0.0001), with an estimated odds ratio of 1.391
(p = 0.0005) when we modeled the probability of
healing for this group. Similarly, control wounds
with an initial duration of less than 1 year also
had a higher mean total WBS (13.62) than those
wounds with a duration longer than 1 year
(WBS = 10.97; p = 0.0001), with an estimated odds
ratio of 1.447 (p = 0.0001) when the healing probability for wounds with duration of less than 1 year
was modeled. However, as also discussed later, we
want to make it clear that we do not want to add
duration to the ultimate WBS because this is a
historical and nondynamic parameter that cannot
be modified to improve ultimate healing.
Finally, we wanted to see which characteristics
in the total WBS had the most predictive value for

387

Falanga et al.

Table 3. Modified WBS with the addition of wound duration prior to treatment
Wound bed score
characteristics
Healing edges
Black eschar
Greatest wound
Depth/granulation
Tissue
Exudate amount
Edema
Peri-wound
Dermatitis
Peri-wound
Callus/fibrosis
Pink wound bed
Wound duration
Prior to treatment

None
> 25% of wound
surface area
Severely depressed or
raised when compared to
peri-wound skin
Severe
Severe
Severe

25 75%
0 25%

> 75%
None

Moderate

Flushed or
almost even

Moderate
Moderate
Moderate

Severe

Moderate

None
Greater or equal to
1 year

50 75%

None/mild
None/mild
None or
minimal
None or
minimal
> 75%
Less than
1 year

The total WBS adds each individual score for each characteristic to give a total score.
The maximum possible score (best score) is 18.
The minimum possible score (worst score) is 0.

Table 4. Characteristics of components in the total WBS with the most predictive value for ultimate
wound healing, as assessed by a stepwise logistic regression model
Effect
Edges (2 versus 0)
Edges (1 versus 0)
P.callus/fibrosis (2 versus 0)
P.callus/fibrosis (1 versus 0)
Duration in years (= 1 versus > 1)

OR

High odds group

p value

95% CI

1.783
0.744
9.335
1.998
4.579

2
1
2
1
=1

0.055
0.133
0.006
0.334
< 0.0001

(0.63, 5.08)
(0.26, 2.17)
(2.03, 42.83)
(0.79, 5.05)
(2.24, 9.35)

= 0.10 used as stopping threshold for stepwise selection.

eventual wound closure. To achieve this goal, we

modeled the probability of healing with a stepwise
logistic regression model that includes each individual characteristic one at a time, with an entry/
removal threshold of p = 0.10. This procedure allowed
us to assess each individual characteristic for its
predictive power in the model. The stepwise procedure selected wound edge effect, peri-wound callus/
fibrosis, and wound duration prior to treatment as
independent significant predictors in the model.
Although these characteristics have the most predictive capability and were selected out by this
model, all characteristics in the WBS contribute to
some degree in predicting wound closure (Table 4).

Discussion
Adequate wound bed preparation is essential for
healing of chronic wounds. Some of the principles

388

involved in this concept had been known for a

long time but had never been truly integrated in
one unified approach. These principles include
exudate and edema control, diminishing the
bacterial burden, stimulating healthy granulation
tissue, and debridement of fibrinous material and
necrotic tissue. Correction of the pathophysiological
abnormalities, whenever possible, is also essential.
In fact, wound bed preparation, as originally
outlined (2), emphasized correction of biological
abnormalities as well as long-established surgical
principles. Many clinical attributes of the appearance
of wound bed are related to these principles and
other interventions. For example, one often speaks
of fibrinous material (or slough) covering the
wound bed, or whether there is proper and adequate granulation tissue. However, it would be of
even greater importance to actually be able to
classify the wound bed based on some important
parameters that might be predictive of ultimate

Wound bed score for healing

outcome. We have learned about the importance

of these parameters mostly from clinical experience, but there are few clinical trials that have
evaluated the importance of each characteristic
(79). Importantly, it remains unclear what specific endpoints in the appearance of the wound
would improve healing or suggest alternative
treatments with advanced therapeutic modalities.
About 5 years ago, in formally outlining the
concept of wound bed preparation, we proposed
a classification system to evaluate wound bed
preparation (2,10). The published classification is
indeed helpful in describing the wound, as it
captures the state of the epidermis, granulation
tissue, and the degree of exudate. In formulating
that classification, we relied on the traditional view
from purely clinical experience that the amount
of granulation tissue may be critical. However,
either this parameter is not as important as initially
thought, or it might be that other attributes are
equally if not more critical to ultimate wound closure. Thus, initial analysis using that classification
system showed that some of the parameters,
whereas clinically useful in documenting the
course of the wound, did not have the prognostic
significance we had hoped for (6) (FIG. 4). The
analysis showed that exudate control and minimal
eschar led to improved healing. However, somewhat surprisingly, minimal fibrinous material and
maximal granulation tissue did not necessarily
lead to improved healing.
Therefore, we felt that we needed to go back to
the drawing board and use our clinical experience
to identify parameters that seemed to correlate

FIG. 4. Percentage of healed venous ulcers and correlation

with initial the scoring system published in 2000 (2).

with healing. The parameters we determined to

be promising were the following: healing edges
(wound edge effect), black eschar, greatest wound
depth/granulation tissue, exudate amount, edema,
peri-wound dermatitis, peri-wound callus and or
fibrosis, and a pale or pink wound bed. It should
be noted that a very red granulation tissue may
often indicate bacterial infection. The exudate
amount and black eschar had been shown in our
previous analysis to be significant. We have noted
that wounds with an established wound edge effect,
that is, a wound edge that has faint epithelial cells
gently sloping towards the wound (no steep edge),
have a higher chance of healing. Interestingly,
wounds that have fibrinous material may still heal
as long as the wound is flush with the edges and
not depressed. On the other spectrum, impaired
healing is seen with wounds that have abundant
but raised granulation tissue.
Wound bed exudate is important, but so is
peri-wound dermatitis. Peri-wound callus has been
shown to be important in diabetic foot ulcers, and
not in venous ulcers and without the presence of
pressure forces. This is the reason we combined
peri-wound callus with peri-wound fibrosis. In
our clinical experience, wounds with peri-wound
fibrosis also have difficulty healing. Finally, a
wound bed that has a pink, but not red color may
be more desirable. An intensely red granulation
tissue may actually reflect increased bacterial
colonization and inflammation. Taken together,
these considerations based on previous analyses
and clinical experience led us to propose a new
WBS that could be tested for wound bed preparation and prognostic value.
In this report, we found that the new proposed
WBS can predict healing. Regardless of whether
they were treated with compression alone or with
a living BSC, wounds with a higher total WBS had
a higher probability of complete wound closure
healing. In support of other work and clinical
experience, wounds with duration of less than 1
year also had a higher WBS. Although all the combined parameters of the WBS were important in
improving the probability of healing, the following three parameters had the strongest predictive
power: wound edge effect, peri-wound callus/
fibrosis and wound duration prior to initiation of
treatment.
In conclusion, we have devised a WBS that,
when used at baseline, appears promising in
predicting ultimate wound bed closure. This WBS
may be useful both in clinical studies and in clinical practice to optimize treatment modalities and
design and test new treatments. Further studies

389

Falanga et al.

are going to be required to use the WBS prospectively in cohorts of patients with different types of
chronic wounds. Ideally, one would like the WBS
to be applicable to wound bed preparation in
general, regardless of the etiology of the chronic
wound.

4.

5.

6.

Acknowledgments
This study was funded by NIH grants AR42936,
AR46557, DK067836 (VF), and the Wound Biotechnology Foundation.

References
1. Phillips TJ, Ouahes N. Leg ulcers. Curr Probl Dermatol
1995: 7: 4 109 142.
2. Falanga V. Classifications for wound bed preparation and
stimulation of chronic wounds. Wound Repair Regen 2000:
8: 347352.
3. Brem H, Balledux J, Sukkarieh T, Carson P, Falanga V. Healing of venous ulcers of long duration with a bilayered living

390

7.

8.

9.

10.

skin substitute: results from a general surgery and dermatology department. Dermatol Surg 2001: 27: 915919.
Falanga V, Margolis D, Alvarez O, et al. Healing of venous
ulcers and lack of clinical rejection with an allogeneic cultured
human skin equivalent. Arch Dermatol 1998: 134: 293300.
Falanga V. A classification system for evaluating graft
appearance and early clinical outcome of bioengineered skin
in human wounds. J Invest Dermatol 2000: 114839.
Donohue K, Saap L, Falanga V. Abstract: Clinical classification of wound bed preparation and its correlation for
healing of venous ulcers. Presented at the 63rd Annual
Meeting for the Society of Investigative Dermatology. May
15 18, 2002. Los Angeles, California.
Steed DL, Donohoe D, Webster MW, Lindsley L. Effect of
extensive debridement and treatment on the healing of
diabetic foot ulcers. Diabetic Ulcer Study Group. J Am Coll
Surg 1996: 183: 6164.
Saap LJ, Donohue K, Falanga V. Clinical classification of
bioengineered skin use and its correlation with healing of
diabetic and venous ulcers. Dermatol Surg 2004: 30: 1095
1100.
Saap LJ, Falanga V. Debridement performance index and its
correlation with complete closure of diabetic foot ulcers.
Wound Repair Regen 2002: 10: 354359.
Falanga V. A classification system for evaluating graft
appearance and early clinical outcome of bioengineered
skin in human wounds. J Invest Dermatol 2000: 114: 839.