Ultrasound Therapy for Recalcitrant Diabetic Foot Ulcers: Results of A Randomized, Double-Blind, Controlled, Multicenter Study‚ÄîPart 2
The ultrasound therapy described in this study employs a recently FDA-cleared device for the cleansing and debridement of wounds with an expanded indication (promotes healing). Used as directed, this device delivers an energy of 0.1 W/cm2 to 0.5 W/cm2 within the defined therapeutic range. A recent study27 using an induced diabetic rat, acute excisional wound model noted a statistically significant increase in blood vessel growth and collagen deposition in the ultrasound-treated wounds compared to sham control, validating that the bio-acoustical effects generated by a non-contact, kilohertz ultrasound device are similar to prior results obtained with megahertz ultrasound devices.17
The current report is the first randomized human clinical trial describing the clinical effectiveness of this new approach of delivering kilohertz ultrasound therapy directly to the wound bed employing a non-contact delivery model. Prior ultrasound therapy for the treatment of chronic wounds has focused on periwound tissue and primarily has utilized ultrasound in the megahertz range. The results of this double-blinded, randomized, sham-controlled trial demonstrate a positive effect on the healing of diabetic foot ulcers.
An attempt was made to design a trial that addressed many of the previous shortcomings of both ultrasound and diabetic foot ulcer studies. The methodological filters of Sackett et al, described in Robertson and Baker’s 2001 review,28 were examined and each question/concern was addressed in the design of this trial. An attempt was made to standardize the vascular status of the patient population with extensive, well-defined, reproducible non-invasive vascular parameters. Offloading was controlled with a uniform requirement that only fixed ankle-foot walkers be used during the study. Standard diabetic socks were supplied to all centers for use during the trial. Patients were required to ambulate at least 75% of the time to simulate the potential clinical results that can be expected when this technology is implemented in a real-world clinical environment. Ulcer severity was controlled by limiting enrollment to patients with ulcers that were either Wagner grade one or two
Clinicians were asked to exclude patients with wounds they believed to be clinically infected as well as patients who were receiving antibiotic therapy. Although the results of quantitative culture biopsies taken on all patients enrolled were not revealed to investigators, the post study analysis of the results is intriguing. Despite the absence of clinical signs of infection as determined by the investigator, 86% of the ultrasound-treated patients and 93% of the sham-treated patients demonstrated >105 aerobic bacteria per gram of tissue (55 patient evaluable cohort). Recent work by Cutting and White29 recommends that clinical criteria for wound infections be created for each specific wound etiology. These and other authors have previously described the inherent problems with accurately determining wound infections using either clinical findings or culture results.30,31 Furthermore, whether any particular quantitative value is, in and of itself, indicative of infection remains a matter for debate. Much has been written about the concept of a bioburden continuum and it has been suggested that factors such as the total quantity of bacteria, their virulence, and host factors all contribute to an individual patient’s risk of developing infection.32 At a minimum, clinicians accept that reducing the bioburden in the wound bed leads to a healthier physiological state with a reduced competition for oxygen and nutrients, diminished exudate, and lowered inflammatory cytokines, all of which allow for healing to proceed. This raises the question: Should all wound healing clinical trials begin with a quantitative biopsy to stratify patients and subsequent outcomes when investigational therapies are studied? If the bioburden continuum is truly clinically linked to healing, the answer is likely “yes.” Browne et al33 have shown that wound healing rates were related to quantitative biopsy values in a study of patients with diabetic foot ulcers receiving bioengineered dermal substitutes.
The addition of ultrasound in this study is the only true treatment variable between the two study groups, leading investigators to postulate that an antimicrobial function of the ultrasound device could, in part, explain the positive outcomes despite a lower than planned for total evaluable cohort. Further support for this theory includes the unpublished animal studies conducted by the sponsor, demonstrating bacterial disruption after ultrasound therapy confirmed by scanning electron microscopy.34 Further clinical trials specifically designed to answer these questions are needed. Additional support for the potential antimicrobial effects of ultrasound therapy is the fact that exudation levels decreased over time as the number of ultrasound treatments were performed (see Figure 2). Also, attention should be drawn to the small increase in exudation noted with the initiation of ultrasound therapy. This transient increase is thought to coincide with the re-establishment of an inflammatory response in a previously chronic, recalcitrant wound.
Another question posed by this trial design was whether the depth or level of debridement had an impact on overall healing rates. A large multicenter, randomized controlled trial35 found a positive correlation between healing in centers frequently performing debridement. However, whether the extent of debridement has an impact on healing rates is not known. In this study, frequency of debridement and debridement CPT codes were recorded to determine the extent of debridement. No statistically significant differences were noted between the two groups. One potential limitation of this method of analyzing debridement depth is the potential for over- or undercoding. Future study designs should explore this phenomenon because the small patient population in this study may have precluded finding any significant effects. The importance of differentiating maintenance debridement from surgical/sharp debridement also has been noted. The use of ultrasound therapy did not decrease the number of sharp/surgical debridements performed.
Prior ultrasound studies either did not include a control device or did not describe the control device or its physical parameters. Although the sham device in this study delivered equivalent saline volumes and pressure when used at the appropriate distance from the wound bed, the need to treat the two study groups at different distances from the wound bed with the two therapies caused confusion and resultant protocol violations. Despite hands-on training sessions, an instructional video, and frequent clinical monitor site visits, multiple protocol violations occurred. Notwithstanding the smaller-than-planned-for remaining efficacy cohort, the ultrasound therapy-managed group had a significantly higher proportion of healed ulcers than the sham treatment group. The impact of treatment on bioburden could explain these findings. Another postulate is that sham therapy may have had a negative effect on healing. The overall sham intent-to-treat group (133 patients) experienced a 22% healing rate at 12 weeks, compared to a 14.3% healing rate at 12 weeks of the efficacy group (55 patients). These values are in reasonable agreement with Margolis’36 published meta-analysis that examines control group outcomes of recently published diabetic foot ulcer trials. Current results also suggest that the sham device did not significantly impede healing. The discrepancy between the intent-to-treat and the evaluable sham healing rates may be related to the increased debridement effects secondary to holding the sham device too close to the wound bed as observed at clinical sites where protocol violations occurred.
Ultrasound therapy has been shown to be a useful adjunct to standard of care for the treatment of diabetic foot ulcers. When used appropriately, 40.7% of treated wounds were healed after 12 weeks of care compared to 14.3% of wounds healed in the sham treatment group. The device was found to be safe, well-tolerated by patients, and easy to administer. When used in clinical practice, confusion about the distance of the device from the wound, which occurred during the study, is not a factor because clinicians only have access to an active unit. Despite the decrease in the total evaluable population, statistically significant differences were observed. The proportion of ulcers healed in this study compares favorably with prior randomized controlled diabetic foot ulcer studies.37,38 Specifically, after 12 weeks, 30% of diabetic foot ulcers treated with bio-engineered tissue (Dermagraft*; Smith & Nephew, Largo, Fla.) were healed compared to 18.3% control-treated ulcers and 56% of ulcers treated with Apligraf® (Organogenesis, Canton, Mass.) were healed compared to 38% of control ulcers. A healing rate meta-analysis of patients receiving control treatments in randomized controlled trials for diabetic foot ulcers found that at 12 weeks the mean healing rate was 24.2%.35
The results of this study suggest the need for further research, including assessing the impact of quantitative biopsy results at enrollment, debridement depth and impact on healing, as well as the potential antimicrobial action of this ultrasound device. Better methods of quantifying the debridement process must be evaluated in order to accurately compare study results. In general, designing clinical studies for wound healing is difficult and surrogate endpoints might enable trials to be constructed with valid outcomes measures other than total closure. Many randomized controlled clinical wound care studies have failed to deliver the outcomes anticipated by pre-clinical work, an observation that warrants further study.39
The authors acknowledge the MIST Diabetic Foot study group: Jeffrey Jensen, DPM; Michael Lerner, DPM; Joseph McCullouch, PhD; Jerry Fabricant, DPM; Jason Hanft, DPM; Lawrence Harkless, DPM; Jeffrey Karr, DPM; Pamela Unger, PT; William Ennis, DO; Adam Landsman, DPM; Pamela Houghton, PhD, PT; Joi Massey, PT; David Armstrong, DPM; Richard Pollack, DPM; Phil Formann, DPM; Mark Melin, MD; Oscar Alvarez, PhD; Alexander Reyzelman, DPM; Michael Dellacorte, DPM; Peter Salza, MD; Gerad Furst, DPM; Neal Mozen, DPM; Morris Prigoff, DPM; Teresa Conner-Kerr. PhD, PT.
The authors also are grateful to Royce Medical for supplying the fixed ankle orthotic walkers used in this study.