Date of Completion


Embargo Period



debridement, diabetes, wound, ulcer, foot, systematic review, meta-analysis, meta-regression

Major Advisor

Tania Huedo-Medina, PhD

Associate Advisor


Associate Advisor

Martin Cherniak, MD, MPH

Associate Advisor

Richard Stevens, PhD

Associate Advisor

Nicholas Warren, ScD

Field of Study

Public Health


Doctor of Philosophy

Open Access

Open Access



Diabetic Foot ulceration has devastating complications. These include amputations resulting in poor quality of life, and serious infections including osteomyelitis and life threatening sepsis. Diabetic wounds can be protracted, take significant time to heal, and they can recur after they have healed. They can be very costly and consume healthcare resources. These consequences have serious public health and clinical implications. Debridement is often used as a standard of care in efforts to help avert these consequences. It is used to remove nonviable or necrotic tissue such as nonviable tissue in order to facilitate the wound healing process and help prevent these disabling outcomes. What is/are the most effective method(s) of debridement remains unclear? This systematic review of the literature on debridement of diabetic foot ulcers synthesizes all experimental evidence in an effort to help answer this important question.

Foot ulceration affects 15% of diabetics at some point in their lives. The prevalence of diabetic foot ulcers is 4.6% in the UK, 8.3% in the US, and includes 7% of the world’s population. The non-healing wound increases the risk of amputations, complicating infections, healthcare costs, and reduces quality of life. Debridement is regarded as an effective intervention to accelerate ulcer healing and to decrease the risk of serious complications.

Current published literature is unclear on which specific method of debridement interventions have the optimal effect on these important public health and clinical implications including: amputation rates, complicating infection rates, quality of life, cost of care) and clinical implications (wound healing rates, wound recurrence rates, and time to complete healing.

Analyzing moderators or prognostic risk factors can facilitate the development of population-specific guidelines or recommendations on the effects of debridement. This can promote better understanding of which groups may benefit from debridement based on prognostic factors. This understanding could increase adherence to the common practices used in diabetic wound care including debridement, provided the evidence supports its efficacy.


The current systematic review and meta-analysis was conducted in order to obtain overall effect sizes of all debridement interventions on the following outcomes (Amputation frequency, Complicating wound infection rates such as Methicillin Resistant Staphylococcus Aureus (MRSA), Quality of life, Proportion of ulcers healing, Proportion of ulcer recurrence, Cost, and Time to complete healing). The goal was to evaluate the variability and consistency of these effects across the current literature on this topic. Any significant variability across the current literature was investigated using moderator analysis based on study-specific and sample-specific characteristics.

Does the use of any form of debridement in diabetic foot ulcers demonstrate benefit over any other form of debridement including standard gauze dressings with respect to amputation frequency, complicating infection rates, quality of life, cost, proportion of ulcers healed, recurrence rates, and time to healing? Are any prognostic or other moderating factors predictive of benefit in some populations or groups? This study summarizes and synthesizes the evidence in a comprehensive qualitative systematic review and quantitative systematic review/meta-analysis of all randomized control trials (RCT’s) on this research question.


A comprehensive literature search was conducted to retrieve articles that met the following inclusion criteria.

Inclusion/Exclusion Criteria

The following inclusion/exclusion criteria were utilized:

a) Individual studies, Systematic reviews (SR’s) and/or meta-analyses (MA’s) that included randomized controlled trials (RCT’s) on debridement of diabetic foot ulcers. Comparison of any method of debridement (i.e. the removal of nonviable tissue from the wound, by either mechanical or non-mechanical debridement) with control or an alternative method of debridement were included The search included any form of debridement but did not include studies on Negative Pressure Wound Therapy (NPWT). NPWT includes mechanical debridement but has other functions.

b) Adult Type 1, or Type 2 diabetics with ischemic, neuropathic, or neuro-ischemic diabetic foot ulcers. The wounds were not limited in severity or in grading system utilized including Wagner Wound Grade, and the Texas Classification systems.

c) There were no other limitations based on age, gender, country, healthcare setting, or language.

d) RCT’s, and Systematic reviews/Meta-analyses that included other wound types i.e. venous stasis ulcers, arterial insufficiency ulcers in non-diabetics, pressure ulcers, and atypical ulcers were excluded.

e) Studies that were nonrandomized (observational studies) were excluded. Systematic Reviews/Meta-analyses that were limited to nonrandomized studies were excluded. All systematic reviews were retrieved along with RCT’s for purposes of comparison and contrasting them with the results of this review.

Search methods

The search included: The Cochrane Wounds Group Specialized Register; The Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library); Ovid MEDLINE, PubMed, EMBASE, EBSCO, CINAHL, and Web of Science.

Selection criteria

Randomized controlled trials (RCTs) evaluating any method of debridement used in diabetic foot ulcers. There was no restriction on articles/trials based on language or publication status.

Data collection and analysis

Data extraction and assessment of study quality were undertaken by two independent reviewers and referred to a methods expert and content expert when there was disagreement. When necessary, if disagreements were not resolved they were referred to the Wounds Group to resolve any remaining discordance between reviewers.

The primary outcomes of interest included: 1) Amputation rates, 2) Complicating wound infection rates, and 3) Quality of life.

The secondary outcomes of interest included: 4) Proportion of participants with ulcers completely healed, 5) Time to complete healing, 6) Proportion of ulcers recurring after healing, and 7) Cost of treatment.

These outcomes have direct bearing on clinical and public health implications including morbidity and mortality. These consequences cause significant hardships for individuals with wounds. A major amputation (above or below knee) is considered by experts to be a predictor of increased 5-year mortality.


Descriptive Statistics

Sample sizes

This review included an analysis of thirty studies with a total of 2654 participants. All 30 studies reported total sample size for each of the included studies. The mean sample size for the included studies was 152 (SD = 119) participants. The included studies ranged from sample sizes of 18 to 619 participants.

Range of follow-up and study period duration

The range of follow up was 10 days to 24 weeks for the included studies. The study period or duration ranged from 1992 – 2012 for the included studies.

Participant characteristics (age, gender, ethnicity)

The mean ages for the samples in the included studies ranged from 52.1 years through 69.3 years. A total of 23 studies reported mean age, 1 study reported a median age of 59.5 years (Roberts 2001), while 6 other studies did not report age. The mean age for the sample of studies was 59.01 (SD = 4.31) years.

Gender was reported in 21/30 (70%) of studies. The number of male participants ranged from 12 to 240, while the number of female participants ranged from 1 to 88 for the reported studies.

Ethnicity was reported in 5/30 (16.7%) studies.

Socioeconomic status was reported in 1/30 (3.3%) study.

Geographic location and healthcare setting

A majority of the studies were conducted in the US or Europe 21/30 (70%) and published in English 28/30 (93%). The study settings included outpatient or specialized clinics 17/30 (56.7%), hospital settings 8/30 (26.7%), and both inpatient-hospital and outpatient settings 2/30 (6.7%), and was unclear in 5/30 (16.7%) studies.

Wound severity

Thirteen studies reported on wound severity, which included wounds up to Wagner grade 4, and wounds up to Texas classification Grade 3. The size of the wound was reported in 20/30 (67%) studies. Depth of wound was specified in 5/30 (16.7%) studies. A total of 14/30 (70%) studies reported on wound duration which ranged from 0 – 60 weeks.

Clinical prognostic factors

8/30 (26.7%) studies reported on hemoglobin a1c which ranged from 7.25% – 9.25%. 14/30 (70%) studies reported on duration of diabetes which ranged from 13 to 21 years. The proportion of baseline peripheral arterial insufficiency was reported in 9/30 (30%) studies. BMI was reported in 5/30 (16.7%) studies.

Table 1 Descriptive summaries of the 30 included studies used in this systematic review and meta-analysis.

Table 1 of Descriptive Statistics

Total number of studies


Total number of participants


Sample size range

18 to 619

Average sample size per study


Total Range of follow up

10 days to 24 weeks

Total Study period or duration

1992 - 2012

Studies reporting age

24/30 (70%)

Mean age (range)

52.1 – 69.3 years

Total number of studies reporting gender

21/30 (70%)

Range of number of males

12 to 240

Range of number of females

1 to 88

Number of studies reporting ethnicity

5/30 (16.7%)

Number of studies reporting socioeconomic status

1/30 (3.3%)

Geographic setting

Europe and US (70%)

Publication Language

English 93%

Healthcare setting

Hospital 8/30 (26.7%)

Outpatient 17/30 (56.7%)

Both 2/30 (6.7%)

Studies reporting wound size (area)

20/30 (67%)

Studies reporting wound duration

14/30 (70%)

Studies reporting Hemoglobin a1c (Hgba1c)

8/30 (26.7%)

Hgba1c (range)

7.25% - 9.25%

Studies reporting on duration of diabetes

14/30 (70%)

Duration of diabetes (range)

13 to 21 years

Studies reporting baseline peripheral arterial insufficiency

9/30 (30%)

Studies reporting BMI

5/30 (16.7%)

Intervention comparisons

Nineteen combinations of debridements or debridement with the control condition were made. This included 12 forms of debridement: 1) sharp, 2) larva, 3) low frequency ultrasound, 4) jet/irrigation lavage, 5) wet to dry gauze, 6) hydrogel, 7) foam, 8) silver based foam dressing, 9) fibrous-hydrocolloid, 10) alginates, 11) honey/jam, and 12) collagenase. The debridements were either compared to each other, or to a gauze/control condition. The control condition included moistened gauze that usually was moistened with saline but could have included an antiseptic (e.g. iodine). The intervention arms were paired with a “standard therapy” (adjunctive wound care measures).

These comparisons included debridement interventions against standard gauze therapy (moistened/saline gauze which may be categorized as a form of autolytic debridement) which was frequently used as a control condition in the included studies OR one form of debridement compared to another form of debridement.

There were significant effects of debridement for some of the outcomes of interest reported in single studies that utilized distinct debridement combination. These combinations could not be pooled in the meta-analysis portion of this systematic review since each of the distinct debridement combinations was only available in one study. These findings are summarized below with respect to the outcomes of interest.

Comparison 1 (Piaggesi 1998) – Sharp surgical debridement demonstrated a statistically significant reduction in quality of life score by 2.2 as compared with nonsurgical management -2.20 (95% CI -3.16 to -1.24), (Heterogeneity: Tau2 = 0.00, χ2 = 0.11, I2=0%). Sharp surgical debridement demonstrated a statistically significant increase in time to complete healing by 82 days as compared with nonsurgical management 81.68 (95% CI 41.07 to 122.29).

Comparison 2 (Goretti 2008) – Superoxide solution demonstrated a significant beneficial effect as compared with standard local treatment using povidone iodine dressing. There was a decrease in time to complete healing by 6 days compared with standard local treatment with povidone iodine -6.00 (95% CI -6.94 to -5.06).

Comparison 5 (Whalley 2001) – Hydrogel purilon as compared with hydrogel intrasite reported a difference of 35% versus 19% in proportion of ulcers healing. It was unclear whether this was significant beneficial effect as there was not enough information reported to make that determination (e.g. no reported counts of events/nonevents).

Comparison 8 (Lalau 2002) – Calcium alginate demonstrated a significant increase in time to complete healing by 2.8 days as compared with gauze 2.80 (95% CI 1.46 to 4.14).

Comparison 12 (Jeffcoate 2009) – Iodine impregnated fiber dressing demonstrated a 45% increase in the number of infections as compared with gauze dressing 1.45 (95% CI 1.13 to 1.86).

Comparison 15 (EhsanUrRehman 2013) – Honey soaked gauze as compared with iodine saline dressing demonstrated a cost difference of 334 Jordanian Dinars in mean difference (MD -334.00, 95% CI -373.99 to -294.01).

Comparison 16 (Hammouri 2004) – Honey/saline dressing demonstrated a statistically significant decrease in treatment cost as compared with iodine peroxide saline – 334.00 (95% CI – 373.99 to – 294.01).

Eight other distinct comparisons that were described in single studies could not be pooled, as there were no other similar intervention comparisons in at least one other study. These studies included: Comparison 3 (Amini 2013) Low frequency ultrasound compared with sharp debridement, Comparison 4 (Markevich 2000) Larvae compared with hydrogel, Comparison 7 (Clever 1995) Polyurethane gel as compared with polyurethane foam, Comparison 9 (Apelqvist 1990) Hydrocolloid as compared with adhesive Zinc, Comparison 11 (Roberts 2001) Foam dressing as compared with saline nonadherent gauze dressing, Comparison 14 (Foster 1994) Hydrocellular polyurethane as compared with Calcium alginate, Comparison 17 (Rhaiem 1998) Sugar Jam with Hydrogen peroxide and topical antibiotic as compared with hydrogen peroxide and topical antibiotic, and Comparison 18 (Belcaro 2010) Silver (standard cleaning and compression management methods) dressing group as compared with (standard cleaning and compression management methods without silver ointment). These comparisons either did not report or did not demonstrate any significant difference in treatment effects for this systematic review’s prespecified outcomes of interest.

Four of the 19 distinct comparisons did include 2 or more studies. These 4/19 comparisons were pooled in meta-analyses for the pre-specified outcomes of interest if reported. This included comparisons 6, 10, 13, and 19.

Pooled data in four separate comparisons including: Comparison 6 Hydrogel compared with gauze (3 studies pooled including: (D'Hemecourt 1998; Jensen 1998; Vandeputte 1997), Comparison 10 Foam dressing compared with Wet to Dry (2 studies pooled including: (Blackman 1994; Mazzone 1993), Comparison 13 Hydrofiber compared with gauze (2 studies pooled including: (Jeffcoate 2009; Piaggesi 2001), and Comparison 19 - Any debridement compared with gauze (10 studies pooled including: (Jeffcoate 2009; Jensen 1998; Piaggesi 2001; Piaggesi 1998; Vandeputte 1997; Lalau 2002; D'Hemecourt 1998; Donaghue 1998; Goretti 2008; Roberts 2001) found no significant beneficial difference, except for the proportion of ulcers completely healed in Comparison 19 - Any debridement as compared with gauze, and Comparison 6 – Hydrogel as compared with gauze.

Comparison 6 - Hydrogel demonstrated a significant beneficial effect as compared with saline gauze. There was a 71% increase in the number of ulcers healed as compared with good wound care 1.71 (95% CI 1.16 to 2.52), (Heterogeneity: Tau2 = 0.00, χ2 = 0.95, I2=0%).

Comparison 19 - Any debridement demonstrated a 17% increase in the number of ulcers healed as compared with saline gauze 1.17 (95% CI 1.00 to 1.36), (Heterogeneity: χ2 = 13.89, I2 = 28%). However, when the two studies available only as abstracts were removed in a subgroup analysis a weaker and nonsignificant beneficial difference was found. 1.12 (95% CI 0.95 to 1.32), (Heterogeneity: χ2 = 2.23, I2 = 55.1%).

There was no significant study heterogeneity that was explained with any of univariate models that were analyzed, except gender. The effects demonstrated low to moderate heterogeneity for the specified outcomes used in this systematic review. Moderators that were analyzed included age, peripheral arterial disease, duration of diabetes, gender, data collection year, and study follow-up duration. All moderators were scrutinized for the recommended number of studies per covariate for each of the prespecified outcomes of interest. There were 6 moderators that satisfied these requirements for the intervention comparison Any debridement as compared with gauze. This analysis was only possible for two of the outcomes of interest this included proportion of infections, and proportion of ulcers healed. There was no significant association or effect using the moderators for either of these two outcomes of interest. A Meta-regression was performed and none of the candidate moderators yielded results any different from the null hypothesis with the exception of the moderator “gender”. This coincides with data that support a gender differential favoring males in the development of wounds in diabetics and amputations having a higher sex predilection among male diabetics. However, the effect was nonsignificant prior to the use of gender as a moderating variable.

Publication bias was assessed and based on the combination of funnel plot and statistical tests (Beggs, Eggers). No significant publication bias was observed despite the fact that 13/30 studies were supported financially by industry.

The GRADE approach was utilized to construct summary of findings tables in order to summarize our conclusions using a structured standardized evidence grading format. This yielded very low to low evidence of efficacy.


Currently there exists weak research evidence to suggest that debridement in one form is more effective in diabetic foot ulcers than other competing forms of debridement or standard gauze for the outcomes of interest in this review. Many of the randomized studies included in this review used small sample sizes that may have been underpowered with too few events/nonevents to make meaningful conclusions. This is evidenced by studies of varying sizes yielding too few events in the intervention arms making it challenging to detect true effects. The included studies often demonstrated significant risk of bias that contributed to the low quality evidence. The studies were variable in the inclusion/exclusion criteria reported.

The findings of researchers could be better supported by following standardized reporting guidelines such as the CONSORT statement. The existing body of literature complicates efforts to synthesize the evidence in systematic reviews. Stakeholders, including patients, physicians, public health professionals, and policy makers, may consider individualized decision making such as indications/contraindications, allergies, tolerability, response, and cost as alternatives pending more definitive standardized RCT’s on this research question. The range of insufficient information in reporting and variation in methods used are summarized in this systematic review. Investigators interested in this research question may benefit from the findings reported in this systematic review as an aide in guiding the design of future randomized studies.

Available for download on Wednesday, April 26, 2017