Study design. A prospective, randomized, blinded, multicenter, parallel study was conducted to compare TCOT with moist wound therapy (MWT) to MWT alone (hereafter referred to as the control) in the treatment of chronic DFUs.
Ethics. This was an Investigational Device Exemption (IDE) study under US CFR 812.2(c)(2). (FDA regulation states medical devices with 510[k[ clearance, used or investigated in accordance with the approved indications for use, are exempt from IDE requirement.) However, all aspects of the study were conducted according to the principles of Good Clinical Practice, the Declaration of Helsinki (1989), the provisions specified in Title 21 Parts 50, 54, 56, and 812 of the US Code of Federal Regulations, the protocol, and all federal, state, and local laws of pertinent regulatory authorities. The study was approved by Western IRB (July 2009). The study was registered with clinicaltrials.gov (NCT01291160). Enrollment of the first patient started October 2009, and the study was completed in November 2012.
Study sites. The study research team was led by a lead principal investigator (PI). Neogenix LLC (Norwood, MA) sponsored the trial; data were managed by Amarex Clinical Research (Germantown, MD). Strategic Solutions Inc (Cody, WY) managed statistical analysis. A Clinical Events Committee (CEC) monitored the safety aspects of the study and included 4 US and Canadian medical doctors, including the lead PI. Blood tests were performed primarily at clinical labs used by the investigators. Wound tracings and photographs were processed by Jill S. Kawalec, PhD, Research Division Head at Kent State University, College of Podiatric Medicine, which acted as a central wound core lab for all sites.Twenty-two (22) investigators from 22 US, Puerto Rican, and Canadian wound clinics participated in the project.
Study inclusion and exclusion criteria. Inclusion criteria stipulated participants must be either gender; 20 to 90 years of age; have a diagnosis of type 1 or type 2 diabetes mellitus; and have a nonhealing, full-thickness, University of Texas Classification of Diabetic Foot Ulcers Class IA of at least 4 but not >52 weeks’ duration measuring 1 cm2 to 10 cm2 in area and located at or below the malleoli. Patients with partial amputation up to and including a transmetatarsal amputation and an ankle-brachial index (ABI) ≥ 0.7 on the study limb; transcutaneous partial pressure oxygen >40 mm Hg; a toe pressure 40 mm Hg; or a Doppler waveform consistent with adequate flow in the foot (biphasic or triphasic waveforms) at screening also were included.
Patients were excluded from participating if their wounds had a duration >52 weeks; there was evidence of gangrene or evidence of active Charcot’s foot on the study limb; they were scheduled to undergo vascular surgery, angioplasty, or thrombolysis; they had infected target ulcers accompanied by cellulitis, known or suspected osteomyelitis, or other clinical evidence of infection; the index ulcer had exposed tendons, ligaments, muscle, or bone; the ulcers were present between toes; the target limb was infected at screening or baseline; if they had a history of malignancy on the study limb; they had used oral or IV antibiotic/antimicrobial agents or medications used within 2 days (48 hours) of baseline; were taking steroids; had received growth factor therapy (eg, autologous platelet-rich plasma gel, becaplermin, bilayered cell therapy, dermal substitute, extracellular matrix) within 2 weeks of the screening date; were pregnant; the total surface area of the ulcer was >10 cm2 at the screening visit as measured by a member of the study staff; they were undergoing renal dialysis; they had known immune insufficiency other than diabetes mellitus; the ulcer decreased in area by >30% during the run-in period; they had a history of peripheral vascular repair within the 30 days of baseline; they were currently receiving or had received radiation or chemotherapy within 3 months of randomization; they had known or “patient-reported” alcohol or substance abuse within 3 months before baseline; they were currently enrolled or participated within 30 days of baseline in another investigational device, drug, or biological trial; they were allergic to a broad-spectrum of primary and secondary dressing materials, including occlusive dressings and the adhesives on such dressings; they had a Chopart amputation; and/or they had an active malignancy except nonmelanoma skin cancer.
The participant and/or caregiver had to be willing and able to learn and perform the duties of dressing changes and demonstrate the ability to do so. If the patient had a history of alcohol or substance abuse within 6 months before the baseline period, proof of treatment needed to be provided.
Patients were screened and sent for laboratory assessments after signing the informed consent document. Screening assessment consisted of 1) medical history; 2) inclusion/exclusion criteria matching (see Table 1); 3) vital statistics (height, weight, blood pressure, systolic blood pressure in arm and ankle, respiratory and pulse rates, and body temperature); 4) wound assessment; and 5) SC wound treatment (removal of necrotic or infected tissue, wound cleansing, establishment of adequate blood circulation, maintenance of a moist wound environment, offloading for plantar ulcers, management of wound infection, nutritional support where needed, and blood glucose control).
If a potential participant passed the screening, he/she was scheduled for a baseline visit 1 week after the screening date and sent for laboratory assessments, including glycated hemoglobin (HbA1c), creatinine, and complete blood count panel. The study participant was randomized providing the HbA1c was <12%, creatinine level was <3 mg/dL, and wound area had not decreased >30% since the screening visit. Baseline assessments included diabetes history, smoking habits, allergies, vital signs, wound location, monofilament testing, wound tracing using Visitrak or Dermal Map mylar grid and felt tip pen, wound photograph, measurement of largest length and width and depth of the wound using a disposable ruler, and wound characteristics (wound edge, base color, periwound conditions, periwound color, edema, drainage amount and type, percentage of granulation tissue, and pain level). When plantar ulcers were present, the study participant also was instructed to offload and was provided the appropriate product.
Randomization and blinding. Study participants who continued to meet the inclusion criteria were enrolled into the study by the site investigator or coordinator and randomized to 1 of the 2 treatment groups. The randomization was centralized with a 1:1 ratio (active treatment group: control treatment group), and 2 stratification factors were used: 1) wound size at baseline (≥1 cm2 to ≤5 cm2 and >5 cm2 to 10 cm2) and 2) patient age at baseline (<65 years, ≥65 years). These stratification factors also were used in analyzing the results of the trial. The randomization schedule was prepared by Precision Sciences Inc (Phoenix, AZ). The actual randomization assignment was made via a web-based centralized system (WebView, Zifo Technologies, Lindenhurst, IL).
All study participants, the investigators, and site staff were blinded to the treatment. In addition, the evaluators who processed the tracings and photographs also were blinded. The active treatment group received the TCOT device, and the control group received a sham “device.” TCOT treatment consisted of continuous administration of 98+% oxygen to the wound site using a 15-day device changed every 15 days. Sham units were prepared by assembling the “device” without the oxygen-generating fuel cell assembly. This was done by Sparton Corporation (Plaistow, NH) (blinded for review), which received a copy of the 1:1 randomization schedule to enable the company to prepare kits with either working device units or sham units. The sponsor also was blinded to this randomization process. Allocation concealment was successfully achieved because the devices looked the same regardless of assignment (for example, A or B).
At each study visit, a member of the study staff recorded any adverse experiences, obtained laboratory samples, measured the study wound, collected level of pain information using the Wong-Baker FACES Pain Rating Scale, administered the Standard Form-12 version 2 (SF-12v2) quality-of-life questionnaire to assess patient-reported physical and mental health, performed wound care treatment, and assessed the study wound.
Clinical evaluation and procedures. The device was applied after wound assessment (including recording of exudate and dressing type and wound characteristics as described earlier), cleansing, debridement, and any other wound management of the study participant’s wound site as follows: using sterile techniques, the cannula was removed from the cannula packet and attached to the Luer Lock located on the side of the device and tightened clockwise to the maximum finger strength. After determining where the device would be most comfortably worn by the study participant, the clinician measured along the path of the cannula to ensure the cannula was long enough to reach from the wound to the device. The tip of the cannula was placed at the center of the cleaned and debrided wound bed and secured to the chosen dressing with Transpore plastic tape (3M, Minneapolis, MN).
A dressing selection guide (hydrocolloid or alginate and foam dressing) was included in the protocol to be selected based on the exudate level of the wound. Clinical judgment was used to determine which dressing was best suited for each study participant. To prevent the cannula from embedding into the periwound or wound bed, a padding (gauze or a foam) was placed underneath the cannula on the participant’s extremity in the periwound area. The tip of the cannula and the entire wound bed was covered with the chosen dressing and secured with Tegaderm (3M, Minneapolis, MN) transparent film dressing; the film ensured the oxygen would be trapped in the wound bed. A 2-inch margin of film dressing along the edges of the wound was recommended, and the film was pinched around the cannula to reduce any chances of oxygen leakage. The transparent film dressing covered the entire length of the primary dressing. The study participant then was scheduled for a weekly visit for dressing change, change of the spent TCOT device (every other week), and wound assessment. All visits and dressing and device changes were done in outpatient clinics.
When the study participant’s wound closed or at the end of 12 weeks of treatment, the treatment was completed and end-of-visit procedures were performed, which included clinical procedures and blood work. If the wound closed on or before the 12-week period, the study participant was scheduled for 2 follow-up visits — the first visit 2 weeks after wound closure and a second follow-up visit 10 weeks after the first follow-up visit, assuming the wound remained closed at the first follow-up visit.
Data collection. Data were collected using Case Report Forms and entered into an electronic database after the patient’s final visit. Original acetate tracings were collected throughout the study and shipped to the wound core lab for analysis. Wound measurement analysis was conducted at the end of the study. Statistical analysis was performed using Statistical Analysis System (SAS™, Cary, NC) statistical software on datasets exported from the electronic database.
Statistical analysis. The sample size was generated using nQuery 6.01 (Statistical Solutions, Boston, MA), based on a clinically meaningful difference of 30% in percentage of healed study participants (28.2% of study participants in the control group and 58.3% of TCOT–treated study participants). The MWT healing rate was calculated as the mean of 6 published DFU studies.21-26 The TCOT healing rate was based on the mean of 2 unpublished case series performed in US and Canadian clinics. Under the above assumptions, 48 study participants per treatment group, totaling 96 study participants in the study, were required to meet the 2-sided Type I error rate of 0.05 and 80% power; anticipating a 40% drop-out rate brought the recommended sample size to 160.
The primary analysis was conducted using the intent-to-treat (ITT) population, defined as the set of randomized study participants who had at least 1 post-randomization efficacy assessment for wound healing. A secondary analysis was based on the per-protocol (PP) population, which consisted of study participants in the ITT population who were not associated with a major protocol violation. Safety assessments were made with the safety population, defined as any study participant receiving treatment after randomization. Continuous variables (age, weight, height, ABI, white blood cell count, creatinine, and HbA1c; and wound area, volume, and duration) were summarized as mean ± standard deviation (SD). Categorical variables (gender, race, diabetes type, neuropathy, smoking, previous amputation, and use of offloading) were summarized as count and percentage with categories as relevant.
Two (2) methods were used to assess treatment efficacy based on the scale and component summary scores derived from the SF-12v2. Analysis of covariance (ANCOVA) was used to compare final visit scores between the treated and the control groups. Baseline score was included as a covariate in the model and treatment arm was included as a fixed effect. Statistical tests were conducted to test whether final±mean scores adjusted for baseline scores differed significantly (P value set at .05) between the active and control groups. In addition to significance testing, Cohen d effect sizes for standardized mean differences were calculated as a way to interpret the magnitude and direction of the difference in the adjusted final SF-12v2 scores between groups.
Efficacy was also assessed using categories of change derived for SF-12v2 physical and mental component scores. Each patient’s final visit SF-12v2 physical and mental component score was categorized as “better,” the “same,” or “worse” than their baseline SF-12v2 score according to the magnitude and direction of the difference in scores between baseline and final visit assessments. In addition, a chi-square test was conducted to test whether the distribution of participants in the 3 categories differed between the study and control groups.
The primary endpoint was defined as complete wound closure by week 12 and was analyzed using chi-squared tests. Time to complete closure was analyzed using Kaplan-Meier analysis in conjunction with the log-rank test. Subgroup analysis was conducted by age using 65 years as the break point.
Adverse events (AEs) were classified by system organ class and preferred term according to MedDRA dictionary (Version 12.0). All AEs that occurred on or after the date of first application of clinical trial treatment were listed and summarized, using frequency counts and percentages, by treatment group. AEs were organized as:
- Overall (regardless of severity or relationship to treatment);
- By severity grade (mild, moderate, or severe); or
- By relationship to clinical trial treatment according to the mapping scheme below:
- Potentially related: included all AEs with a relationship rating of “definitely,” “probably,” or “possibly”;
- Unlikely/not related: included all AEs with a relationship rating of “unlikely” or “unrelated”; or
Serious adverse events (SAEs) were indicated by the investigator as being serious from the question, “Is this considered serious?”