Enzymatic Debridement of Chronic Nonischemic Diabetic Foot Ulcers: Results of a Randomized, Controlled Trial

Objective. The aim of this randomized, controlled multicenter trial was to evaluate the clinical outcomes associated with the use of clostridial collagenase ointment (CCO) for up to 12 weeks in 215 patients with type 1 or type 2 diabetes mellitus and a neuropathic, nonischemic diabetic foot ulcer (DFU). Materials and Methods. Patients were randomized into either a group receiving CCO applied once daily at a thickness of ~2 mm or a group receiving standard care (SC) consisting of a daily application of a hydrogel as needed to maintain a moist ulcer environment. All ulcers were covered with a nonadherent foam dressing that was changed once daily, and sharp debridement was allowed when the investigators deemed it medically warranted. Outcome measures included the percent change in ulcer area and the effect of baseline wound microbiota on subsequent healing. Patients with ulcers that showed no decrease in size after 4 weeks were crossed over to the other treatment group. Results. The wound area decreased relative to baseline for both the CCO group (-60%, P < .0001; -65%, P < .0001) and the control group (-50%, P = .0001; -51%, P = .0001) after 6 and 12 weeks, respectively. While the intergroup differences at 6 and 12 weeks did not reach statistical significance (P = .3801; P = .3606), mean percent reductions for the CCO group were greater than the control at all 12 time points (averages: -55%; -41%, respectively). Overall closure rate was 21% and 41% (weeks 6 and 12, P = .3705; P = .2358) with no significant differences between groups. However, the DFUs that showed no improvement at 4 weeks (N = 24, 12/group) were crossed over to the other treatment group. A numerically greater proportion of subjects who switched to CCO achieved closure (33%) than for those who switched to SC (8%). Baseline biopsy showed that despite the absence of clinical signs of infection, all ulcers were heavily colonized by 1 to 5 species of bacteria. No adverse events were assessed by the investigators as related to either treatment. Conclusion. These results confirm observations from previous studies demonstrating positive outcomes associated with enzymatic debridement with CCO following 6 weeks of treatment.

Successful management and healing of diabetic foot ulcers (DFUs) can be uniquely challenging, especially within the context of chronic and hard-to-heal wounds. Potential impediments to healing that DFUs face include poorly perfused tissue,¹ lack of or inadequate blood glucose control, patient with poor nutritional status,² high risk of bacterial contamination leading to biofilm and infection,³ concomitant comorbidities such as renal disease, and poor offloading.⁴ Facilitation of healing should address and correct these potential barriers to the extent possible in concert with wound bed preparation in order to maximize the chance for healing to occur.⁵ Debridement, a keystone of wound bed preparation, may be accomplished by various means including surgical, mechanical, biological, or enzymatic. The overall goal of debridement is to change the environment of the wound bed through removal of nonviable tissue, reduction of bacterial load, and (depending on methodology) providing a wound edge that can reepithelialize the ulcer.^6-8

Previous work has shown wound bed appearance of DFUs enzymatically debrided with clostridial collagenase ointment (CCO), as assessed and scored visually, is indistinguishable from mechanical plus serial sharp debridement, but the CCO-debrided DFU had greater reductions in area over the assessment period.⁹ Similarly, when CCO is used in conjunction with serial sharp debridement, better outcomes were achieved than for serial sharp debridement used with passive moist wound care methods.¹⁰

The present study was designed to corroborate earlier reports of the beneficial effects of enzymatic debridement with CCO in the treatment of DFUs.^9,10 Specifically, additional data were collected to demonstrate a significant and meaningful reduction in wound area achieved over the 12-week treatment period. Consistent with previous work,¹⁰ enzymatic debridement was employed in the current study alongside periodic sharp debridement when deemed medically warranted.

Study participants. This study was registered at ClinicalTrials.gov as NCT02111291. Eligible patients were 18 years or older with a diagnosis of type 1 or 2 diabetes mellitus requiring medications to control blood glucose levels. Criteria for study eligibility included neuropathic foot ulcers of at least 6 weeks in duration located on the plantar surface with an area between 0.5 cm² and 10 cm², ankle-brachial index (ABI) > 0.7, no clinical signs or symptoms of ulcer infection, serum albumin ≥ 2.0 g/dL, and glycated hemoglobin ≤ 12%. Patients were excluded for any of the following reasons: ulcer did not require debridement, uncontrolled bleeding disorder, infection with systemic toxicity, an ulcer that could not be offloaded, active cellulitis of the target ulcer, deep tissue abscess, gangrene, or infection of muscle, tendon, joint, or bone. This study was performed in compliance with the ethical principles of the Declaration of Helsinki and Good Clinical Practice. The study protocol, investigators, and consent documents were reviewed and approved by accredited institutional review boards (IRBs). Sterling Institutional Review Board, Atlanta, GA, was the central IRB for the study. All patients provided written informed consent before taking part in the study.

Study design. This was a randomized, parallel group, open-label, multicenter clinical study (eFigure 1), carried out at 26 sites in the United States and in Canada. The study took place between May 2014 and December 2015. Randomization to treatment was centralized using a computer-generated sequence. 

Patients were evaluated for eligibility at a screening visit, which included a blood draw for assessment of hematology and blood chemistry. Eligible patients reported for the baseline/randomization visit 1 to 8 days later. The baseline wound area measurement was carried out following sharp surgical debridement for all patients. Patients were then randomly assigned to treatment in an equal allocation ratio to either the CCO (Collagenase SANTYL Ointment; Smith & Nephew, Fort Worth, TX) group or to the control group (standard care [SC]). Patients received a CCO application once daily at a thickness of ~2 mm on the DFU, which was then covered with a nonadhesive foam dressing (ALLEVYN Non-Adhesive; Smith & Nephew, Fort Worth, TX) in the CCO group. The ulcers of patients randomized to the control group received a daily application of a hydrogel (3M Tegaderm Hydrogel; 3M Health Care, St Paul, MN) as needed to maintain wound moisture, and the ulcers were covered with the nonadhesive foam dressing like the CCO group. Ulcer treatment was applied at weekly study visits by the Investigator or designated staff and patient-applied each day between visits. While the application of hydrogel was at the discretion of the investigators, hyperbaric and negative pressure therapies were not allowed. In either group, treatment was given for up to 12 weeks. Sharp debridement was permitted for both treatment groups at the investigators’ discretion throughout the treatment period. All patients agreed to comply with offloading; an offloading boot or other appropriate offloading device was provided for all patients. At each weekly visit, patients were instructed to cleanse the wound daily with sterile saline, dressing changes, and application of either the CCO or the SC treatment. It is generally accepted in clinical practice that ulcer treatment should be changed or modified if significant progress toward healing has not been achieved after 4 weeks, therefore following 4 weeks of treatment (study visit 5), the patients in either group with ulcers that had not decreased in size changed therapy to the opposite treatment. 

Study assessments. Granulation was assessed visually, and wound area was measured at each study visit. Area was measured using the ARANZ Silhouette digital image capture and wound measurement device (ARANZ Medical, Christchurch, New Zealand).  

If neuropathy had not been previously documented in their medical record, it was confirmed by the patient’s inability to sense a 10 g nylon monofilament on 3 of 5 tested areas on the plantar surface. Adequate perfusion to the affected foot was confirmed by an ABI > 0.7 or, for an ABI > 1.2, a foot that was warm to the touch with a palpable pulse.

A biopsy was taken from the target wound bed using a 4-mm punch at visit 1 for quantitative bacteriology assessment at a certified clinical lab (LabCorp Clinical Trials Services, Cranford, NJ).

Safety was assessed through analysis of adverse events. Both volunteered and elicited adverse events were collected at each of the 13 study visits. 

Statistical analysis. The intent-to-treat (ITT) population, defined as all patients who were randomized and had a valid baseline ulcer measurement, was used for primary inference. Statistical Analysis System Version 9.0 (SAS Institute, Cary, NC) was used to perform the analyses. Hypothesis testing was performed at the 5% significance level. All P values were rounded to 4 decimal places; P values < .0001 are presented as P < .0001. All group comparisons from analysis of variance or analysis of covariance (ANCOVA) models were based on type III sums of squares. Between-group comparisons of the average percent change in wound area from baseline to each treatment week of visits 2 to 13 were analyzed using two-way ANCOVA. Paired t-tests were employed to evaluate the significance of changes from baseline in wound area within each treatment group. A Cochran-Mantel-Haenszel non-parametric test with adjustment for investigative site was used to examine the differences in categorical variables (eg, proportion of closed ulcers). Chi-squared tests were used to compare demographic and baseline ulcer characteristics. Missing values for wound area measures or ulcer closures due to early discontinuation, missed visits, or complete wound closure were imputed using the technique of last observation carried forward. Area was imputed as 0 for closed ulcers.  

Sample size. Previous studies^9,10 of patients with DFUs disclosed differences in ulcer area reduction from baseline from 0.30 cm² to 0.80 cm² for CCO treatment compared with various modes of supportive care for up to 6 weeks of treatment. Assuming an effect size of 0.40, a sample size of 100 patients per group was calculated to provide statistical power of 80% to detect a between treatment difference at a significance level of 0.05 (2-sided). The study was therefore planned to enroll 200 patients with an allocation ratio of 1:1 for CCO versus SC. 

At the 26 centers between the United States and Canada, 215 subjects were enrolled in this trial. Patient disposition is diagrammed in eFigure 2. 

Baseline wound and demographic characteristics for the 209 ITT evaluable patients comprising the study population are presented in eTable 1. As is typical in previous US-based DFU studies,^9,10 enrolled patients were predominantly elderly, male, white, and non-Hispanic/Latino. Similar to previous studies,^9,10 baseline ulcer size was approximately 1.5 cm² and were plantar. While there were no statistically significant differences in these characteristics observed between groups, the percentage of females in the CCO group was about 11% more than the control, and this difference approached statistical significance. 

Although the choice to use a daily application of hydrogel for the control group was at the investigators’ discretion, most control subjects did receive the hydrogel daily at all (87%) or some (95%) dressing changes.

Offloading was required as a condition for enrollment. The great majority of ulcers in both the CCO (90%) and the control (92%) groups were offloaded using the sponsor-supplied device with PegAssist insoles (PegAssist, OrthoWedge; DARCO, Huntington, WV). Other appropriate devices were allowed at the investigators’ discretion.

Wound area. Mean percent reduction in wound area was -60% versus -50% at week 6 and -65% versus -51% at week 12 for the CCO and control groups, respectively. While the difference between groups was not statistically significant, within-group reduction in ulcer area from baseline was statistically significant in the CCO group after 1 week of treatment and in the control group after 2 weeks of treatment (CCO, P < .0001; control, P = .0275). Mean percent reductions for the CCO group were greater than those in the control at each of the 12 time points (12-week averages: -55%; -41%), with the greatest difference at week 8 (-64%; -41%). eFigure 3 shows the decrease in ulcer area at week 6 and week 12 for the CCO group for the present study as well as for 2 previous trials.^9,10  

Overall closure rate was 21% and 41% (weeks 6 and 12), with no significant differences between groups (P = .3705; P = .2358).

Wound appearance. No significant differences were observed in the proportion of ulcers achieving well-granulated status at any of the visits, with both groups achieving about 79% of ulcers well granulated by week 12.

Sharp debridement. No statistically significant difference between groups was noted in the frequency of sharp debridement; both groups had approximately 0.8 debridements per subject per week.

Quantitative bacteriology. Ulcer biopsies were obtained from 159 (76%) of the 209 ITT evaluable patients at the baseline visit. Patients eligible for enrollment were required to have an ulcer that was not infected based on clinical assessment. Despite this, baseline biopsies of the referenced ulcer bed revealed that all ulcers carried significant bacterial bioburdens ranging between 1 x 10⁴ CFU/g to 3 x 10⁷ CFU/g per species isolated. The 2 groups were essentially identical in this regard. The mean minimum and maximum CFU/g per patient was 1.2 x 10⁶ and 7.1 x 10⁶ for the CCO, and 1.5 x 10⁶ and 7.4 x 10⁶ for the control. A total of 67 different bacterial species were identified from the 159 patients, with each patient positive for 1 to 5 species. Of the 159 biopsies, 39 positive cultures yielded a single species, 57 yielded 2 species, 41 yielded 3 species, 16 yielded 4 species, and 6 yielded 5 species. Two-thirds of the organisms identified (258/391) were accounted for by 8 species in 7 genera (eTable 2). Bacterial load had a slight negative association with ulcer closure (r = -0.053 and r = -0.15, for the CCO and control groups, respectively) for patients who stayed on the original treatment; neither was statistically significant. For all patients who crossed over to the alternate treatment at week 4, the negative association between level of bioburden and ulcer closure rate was stronger (r = -0.694 and r =  -0.456, for the CCO and control groups, respectively), suggesting wound bioburden was a key influence on healing outcomes for these stalled ulcers. 

The proportion of closed ulcers for patients positive for specific organisms is shown in eFigure 4. Bacteria associated with a lower-than-overall-average proportion closed (< 34%, considering only the subgroup with biopsy data) were termed “inhibitory.” Except for Pseudomonas aeruginosa and Acinetobacter baumannii, these were facultative anaerobes. It should be noted that similar inhibition of closure when present was generally seen for a given organism across both treatment groups. 

Stalled ulcers. Ulcers that showed no improvement following 4 weeks of active study treatment (N = 24, 12/group) were crossed over to the other treatment group. A numerically (albeit not statistically significant) greater proportion of subjects who switched to the CCO group achieved closure (33%) than for those who switched to the control (8%).   

Adverse events. A total of 315 adverse events were recorded from 129 patients over the course of the study. These were evenly distributed between the 2 treatment groups, albeit slightly higher in the control group (144 for CCO; 171 for control); the investigators determined none were related to the treatment. All adverse events recorded were those normally expected from this population under study.

Given the bioburden present in these wounds, the treatment applied to the ulcers could either positively or negatively influence the incidence of target ulcer infection. To determine if treatment of the target ulcer with CCO or SC made the ulcer more or less susceptible to infection, all adverse events for infections or infections stated to be associated with the target ulcer (including adverse events recorded as cellulitis, gangrene, or osteomyelitis) were reviewed. There were 9 of these infections (8.5%) for the CCO group and 15 (13.8%) for the control.

The difficulty in healing DFUs usually cannot be attributed to a single factor. Instead, the chronicity and recalcitrance of these ulcers are often due to multiple pernicious factors operating concomitantly to impede progress toward healing. These include the well-known sequelae of diabetes (eg, peripheral neuropathy and impaired renal function), commonly co-occurring conditions such as peripheral arterial disease and ischemic extremities, and environmental and compliance issues that include poor nutrition, inadequate or no offloading, and smoking. The wound bed itself is often characterized by the presence of devitalized and necrotic tissues, senescent cells, foreign material, and biofilm-forming bacteria. While many of these factors require intervention by medical or surgical specialists or significant lifestyle changes, dysfunction in the wound bed can be addressed through debridement. 

Debridement is a therapeutic intervention with the goal of clearing away impediments to healing. Although it is accepted that debridement involves the removal of foreign bodies, nonviable tissues, and calluses, broader definitions have included infection control, biofilm removal, inflammation reduction, visualization and probing for tunneling and undermining, and removal of dysfunctional or senescent cells.^11-17  

All various debridement modalities can, with varying efficiency, accomplish the removal of nonviable tissue and rapidly so when used in conjunction with sharp methods. Previous studies^9,10,16 comparing CCO with mechanical or autolytic methods have shown, at least in terms of gross appearance of the wound bed, similar results can be obtained in similar time frames. Consistent with these previous reports,^9,10,16 the present study also found no difference in the visual assessment of ulcers as “well granulated” between the CCO or the control group. However, progress toward closure is not solely dependent on the visual appearance of the wound.¹⁰ Galperin et al¹⁶ showed wound area reduction was associated with a permissive biochemical environment (ie, resolution of inflammation). Brem et al¹³ have suggested the difference between a stalled wound and a healing wound may only be apparent using histological techniques. The accomplishment of other debridement goals such as inflammation resolution or changing the balance between senescent and nonsenescent cells cannot be assessed visually, but progress toward closure can be measured through reduction in ulcer size (a downstream result).

While surgical sharp debridement is generally considered to be the “gold standard,” some situations can restrict its use. These include patients with bleeding disorders, intolerance for associated pain, or the unavailability of appropriately qualified personnel to perform the sharp debridement. The daily use of CCO as the sole means of wound debridement has previously been shown to be an efficacious treatment for pressure injuries and for DFUs.^9,18 It is interesting to contrast the results obtained in Tallis et al⁹ (where the ulcers received a single sharp debridement at baseline, but not thereafter) with Motley et al¹⁰ and the present study, both of which allowed adjunctive serial sharp debridement. While good outcomes were achieved for CCO in the absence of repeated sharp debridement (54% reduction in area at 12 weeks), consistently better outcomes of at least 65% reduction were achieved when CCO was used in conjunction with sharp.

In the present study, treatment with CCO resulted in an average reduction in wound area of 60% following 6 weeks of treatment. This is consistent with a previous study¹⁰ with similar inclusion/exclusion criteria (68% at 6 weeks). Continued treatment with the CCO beyond 6 weeks provided only modest additional benefit. After 12 weeks of treatment, the mean reduction in area was 65%, while for control was 51% (an additional 5% and 1% over 6 weeks). The fact that the greatest differences occur within the first half of the treatment period underscore the important role that the CCO plays in initial wound bed preparation.

Enzymatic debridement with CCO has also been shown to result in fewer stalled ulcers.  Tallis et al⁹ showed there was only one-third as many stalled ulcers (8% vs 25%, defined as ≤ 10% reduction in area at 12 weeks) with 4 weeks of CCO.⁹ In the present study, patients were crossed over to the alternate treatment if their ulcer had not changed in size following 4 weeks of treatment. While there were 12 patients in each group that switched treatments, those who switched to CCO were 4 times more likely to achieve closure over the ensuing 12 weeks (33% vs 8%).

The presence of biofilm-forming bacteria is considered to be one of the most significant impediments to healing.^3,7 Lantis et al¹⁹ reported that 87% of chronic venous ulcers from a large clinical study were found to be culture positive for 1 or more species of bacteria. James et al³ found 60% of chronic wounds (including pressure, diabetic foot, and venous leg ulcers) contained biofilm. In the present study, nearly all of the patients that were biopsied at baseline were found to be critically colonized (≥ 1 x 10⁵ CFU/g) with various bacteria. In agreement with both previous studies,^3,19 staphylococci and in particular Staphylococcus aureus was the most prevalent organism encountered followed by Enterococcus. While it is not possible to rule out an inhibitory effect on healing for any of the organisms detected, E. faecalis and P. aeruginosa were associated with a particularly low rate of healing. Interestingly, Enterococcus spp and Pseudomonas spp were also identified as significantly inhibiting venous leg ulcer healing.¹⁹ The presence of nonviable, necrotic tissue provides a niche for bacteria to become established and grow. It may also limit the body’s cellular defenses to fight infection. Removal of this tissue through enzymatic debridement may help to reduce bioburden and thereby remove a significant barrier to healing.¹¹ In line with this possibility, the adverse event data from this study showed that subjects treated with the CCO had a lower incidence of infection as assessed by clinical signs and symptoms.

While a priori power calculations provided a sample size estimate of N = 200 as adequate to detect a statistically significant difference between groups, the calculation was based on an incorrect assumption of the variability in wound area reductions; thus, the study was underpowered. Nevertheless, the results shown are in overall agreement with several other similar studies,^9,10,16 supporting their validity. All patients were required to wear an offloading device; however, it is likely that compliance was less than universal. Offloading is known to be critical for healing plantar ulcers. This limitation is mitigated by the blinded randomization that should allocate noncompliant patients to the 2 groups in similar proportions.

The credibility of a scientific observation is closely connected with its reproducibility. Repeatable experiments are necessary to test hypotheses; confidence is established through repetition.²⁰ Despite the fact that statistical significance (between group comparison) was not achieved, the present study builds on previous reports^9,10,16 demonstrating positive outcomes associated with enzymatic debridement with CCO when used to treat chronic, hard-to-heal DFUs. Consistent with this previous work,^9,10,16 wound bed appearance was improved and this improvement equaled that achieved by other debridement modalities. In accord with the current report, these previous studies^9,10,16 have also shown good progress toward closure as measured by mean percent decrease in wound area is achieved with CCO in 4 or 6 weeks. Indeed, at weeks 6 and 12 postrandomization, ulcer size reductions were virtually the same (≥ 60%) as was seen in Motley et al.¹⁰  

The authors thank the members of the study group for their contributions: Joseph Caporusso, DPM; Patrick S. Agnew, DPM; Travis A. Motley, DPM; Mher Vartivarian, DPM; Shawn M. Cazzell, DPM; Gregory Tovmassian, DPM; Christina Morin, DPM; Cyaandi Dove, DPM; Carl C. VanGils, DPM; Naohiro Shibuya, DPM; Dean Vayser, DPM, ACFAS; Evaristus Oshiokpekhai, DPM; Maria Ann Kasper, DPM; Susan E. Kemp, MD; Bradley M. Lamm, DPM; John C. Lantis, II, MD; Juan Carlos Jimenez, MD; Laurie M. Parsons, MD; Robert Galiano, MD; John R, Clements, DPM; Joseph Cavorsi, MD; Christopher J. Gauland, DPM; Ali R. Davis, DPM; Perry Mayer, MD; Lewis H. Freed, DPM; and Manuel  J. Sone, DPM. In addition,  Art Leffe, DSc, and Luc Duvel, PhD, provided valuable encouragement. 

From the UCLA Medical Center, Sylmar, CA; Coastal Podiatry Inc, Virginia Beach, VA; The Mayer Institute, Hamilton, Ontario, Canada; Carilion Clinic, Roanoke, VA; Complete Family Foot Care, McAllen, TX; Smith & Nephew Inc, Fort Worth, TX; Graduate School of Biomedical Sciences, University of North Texas Health Science Center, Fort Worth, TX; and Department of Pediatrics, University of North Texas Health Science Center

Address correspondence to:
Jaime E. Dickerson, Jr, PhD
Smith and Nephew
5600 Clear Fork Main St
Fort Worth, TX 76109   
jedickerson@live.com

Disclosure: All financial support for this study was provided by Smith & Nephew (Fort Worth, TX), the study sponsor. Drs. Lange, Dickerson, and Slade were all employees of Smith & Nephew at the time of the study. Drs. Agnew, Mayer, and Clements have once been or are currently speakers for Santyl (Smith & Nephew).