Clinical Assessment of a Biofilm-Disrupting Agent for the Management of Chronic Wounds Compared With Standard of Care

A Therapeutic Approach

Daniel Kim, MD; William Namen II, DPM; January Moore, BA; Mauricia Buchanan, BSN; Valerie Hayes, PhD; Matthew F. Myntti, PhD; Albert Hakaim, MD

Disclosures

Wounds. 2018;30(5):120-130. 

In This Article

Results

Demographics

Forty-three patients were enrolled in the study with 32 completing all study visits. The study visits ended after each patient finished 12 weeks of treatment or the patient's wound was healed. The first patient visit occurred in September 2014 and the last patient study visit was in March 2016.

At the end of the enrollment period, 22 patients were randomized to the experimental group and 21 were randomized to the control group; however, 12 of the control patients crossed-over to the experimental group due to the wound worsening or failure to heal/lack of improvement, resulting in an experimental product group of 34 patients and a control group of 21 patients (Table 2). Patients ranged in age from 32 to 91 years (average, 62 years). The age range of the wounds in the study ranged from 1 month to 20 years (average, 21.2 months). Wound size ranged from 1 cm2 to 114 cm2 (average wound area, 10 cm2).

The patient population had a high number of comorbidities (Table 3), with all patients except 1 presenting with 1 or more. These comorbidities were not statistically significant factors affecting wound closure or healing rates when analyzed by ANOVA.

Wound Size Reduction

The primary endpoint was the decrease in wound area at 12 weeks (Figure 2). The statistical power of the wound closure percentage was greater than 99% (type I error of 0.05 comparing experimental to control), indicating that sufficient patients were evaluated to yield meaningful results.

Figure 2.

Comparison of normalized wound area reduction over time. After 12 weeks of treatment with the experimental product, the average wound area reduction was 72% versus 15% in the control group (P < .01). P value for treatment time was <.01. Histogram shows that wound areas were continuing to decrease at 12 weeks compared with the plateau in the control group. Bars depict one-standard error from the mean; grouping bars depict groups that are equivalent. Treatment bars that do not fall under the same grouping bar are statistically distinct (P < .05).

Figure 3.

Comparison of normalized wound area reduction over time, including only those control patients who were not crossed over. After 12 weeks of treatment with the experimental product, the average wound area reduction was 72% versus 53% in the control group (P < .02). P value for treatment time was < .01. Histogram shows that wound areas were continuing to decrease at 12 weeks compared with the plateau in the control group. Bars depict one-standard error from the mean; grouping bars depict groups that are equivalent. Treatment bars that do not fall under the same grouping bar are statistically distinct (P < .05).

The mean and median values for the percent wound area at each time point, with all patients included in the control group, is demonstrated in Table 4. The average wound area reduction in the experimental group was 72% ± 8% at 12 weeks; this was statistically significant in comparison with the control (P < .01) and for treatment time (P < .01). The wound healing reduction of the experimental group was 2.44 times greater than the control. In the case of median wounds, which removes the bias from the outliers, the difference is even greater, with median wound size for the experimental group being 90% compared with 5% for the control. The wound reduction rate increased as the treatment time progressed, indicating that those wounds that were not yet closed were progressing towards closure.

The number of patients at each time point as well as the mean and median values for the percent wound area at each time point, with the crossover patients removed from the control group, is shown in Table 5.

The average wound area reduction of the experimental product was 72% ± 8% at 12 weeks; this was statistically significant versus control (P < .01) and for treatment time (P < .01). Even including only those patients that were showing improvement with the control (and not including those that failed treatment with the control [ie, crossed over]), the wound healing reduction was 36% greater than the control product. In the case of median wound size reduction, which removes the bias from the outliers, the difference is even greater, with the median wound size reduction for the experimental group being 90% compared with 50% for the control. As such, median improvement with the experimental product is 1.8 times greater than the control. Moreover, the wound reduction rate in the experimental group increased as the treatment time progressed, indicating that those wounds that were not yet closed were still progressing towards closure.

Wound Closure

The secondary endpoint was the increase in the percentage of wounds that were closed after 12 weeks of treatment (Figure 4, Figure 5). The use of the experimental agent improved the success rate for chronic wound healing. The statistical power of the wound closure percentage was 90% (type I error of 0.05 comparing experimental product to vehicle), indicating that sufficient patients were evaluated to yield meaningful results.

Figure 4.

Comparison of percentage of healed wounds over time. After 12 weeks of treatment with the experimental product, 52% of patients achieved wound closure (P<.01) versus 16% in the control group. P value for treatment time was <.01. In the experimental group, 43% of the wounds were closed at 8 weeks versus 17% in the control group. Bars depict one-standard error from the mean; grouping bars depict groups that are equivalent. Treatment bars that do not fall under the same grouping bar are statistically distinct (P < .05).

Figure 5.

Comparison of percentage of healed wounds over time (excluding control patients that were not crossed over to experimental). After 12 weeks of treatment with the experimental product, 52% of patients achieved wound closure versus 33% in the control group (not statistically significant). P value for treatment time was < .01. In the experimental group, 43% of the wounds were closed at 8 weeks versus 33% in control group. Bars depict one-standard error from the mean; grouping bars depict groups that are equivalent. Treatment bars that do not fall under the same grouping bar are statistically distinct (P < .05).

After 12 weeks of treatment with the experimental product, 52% of patients achieved wound closure. This was statistically significant versus the control (P < .01) and for treatment time (P < .001). The percentage of patients with healed wounds was 3.12 times greater in the experimental group than the control (17% closure). Also, it is interesting to note that 40% of the wounds treated with the experimental agent were closed within 8 weeks of once per every other day application; in fact, wound closure started to occur as soon as 2 weeks in the case of a wound that had failed to heal with the use of numerous treatments for more than 20 years. The wound size range for the group with healed wounds was 0.7 cm2 to 7.3 cm2 (average, 2.3 cm2), and the age range of the wounds was 1.5 to 240 months (average, 29 months).

Biofilm Analysis

There were 90 bacterial and 4 fungal species found in the wounds. Only 5% of patients had fungi in their wounds. Of the 90 bacteria, only 17 were found in at least 10% of patients. Figure 6 shows a histogram of these bacteria. The mean number of bacteria present in the wound was 2.9 species per wound (range, 0–13 bacteria). There was no statistically significant relationship between the bacteria present in the wound or number of bacterial species present and wound healing or healing rates. An ANOVA of the bacterial load also was not statistically significant for wound closure or wound healing.

Figure 6.

Histogram of the most frequently found bacteria in the study wounds. Bacterial species were detected in at least 10% of the patients' wounds.

A PCR analysis for vancomycin-resistant and methicillin-resistant genes resulted in an incidence rate too low to provide enough power for statistical evaluation.

Complications

All patients enrolled in this study presented with 1 or more comorbidities and 60% had diabetes (N = 26). There were no unanticipated problems.

The root cause of all adverse events was evaluated (Table 6). There were 2 adverse events that were directly related to product application in the control group; one was an allergic reaction to the control product with skin desquamation, and the other was burning sensations. In both cases, the control antibiotic ointment was discontinued, and then the issues resolved and did not reoccur with the experimental product. There were no adverse events that were attributed to the experimental product, although 1 patient was discontinued in the group due to wound site pain, which was present prior to the study treatment application and persisted with the new treatment of PolyMem (Ferris Mfg. Corp, Fort Worth, TX) dressings.

There were 11 study discontinuations: 4 patients in the experimental group and 7 in the control (Table 6).

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....