Study overview and conduct. IN BALANCE VLU was a prospective, randomized, controlled, parallel-group, multicenter clinical trial conducted between April 2012 and March 2014. The trial was performed in 22 facilities across the US with multispecialty and multidisciplinary teams, including 8 community hospital outpatient wound centers, 5 private wound clinics, 6 university outpatient wound centers, and 3 Veterans Administration outpatient wound centers.
The trial is registered on clinical trials.gov as NCT01549860 and was sponsored by Celleration, Inc (Eden Prairie, MN). Institutional Board Review approval was obtained in accordance with the principles of the Declaration of Helsinki and recognized Good Clinical Practices were followed, including participant informed consent and third-party monitoring. A third-party statistician prepared the randomization program and performed the data analysis for this study. Randomization was computer-generated using a 6-block permuted randomization scheme stratified by study site and automated through the clinical database.
Unique participant IDs were assigned by the clinical database. Study sites labeled all data with the participant ID and redacted any personal health information before submitting any diagnostic tests, ulcer images, or source documents. Data were entered by the study sites using an electronic data capture system (Clindex® EDC Fortress Medical, Minneapolis, MN). Study site personnel received training and were tested on digital wound image procurement, drawing of epithelial boundaries for area measurements, NLFU therapy application, and 4-layer compression wrap application. Study conduct, data management, adjudication practices, and participant safety were overseen by a third-party Data Safety Management Board (DSMB).
Ulcer area measurements. Study endpoints for healing were based on weekly measurements of the nonepithelialized area of the index ulcer using a digital camera (SilhouetteStar® Aranz Medical, Christchurch, NZ). Images were taken by the study site, and epithelial boundaries were electronically drawn using the associated software (SilhoutteConnect® Aranz Medical, Christchurch, NZ). The software performed a calculation of the nonepithelialized area based on the boundaries drawn. Study sites uploaded each deidentified image (labeled with the participant ID) and entered the calculated ulcer area measurement into the clinical database. A third-party, blinded wound expert was sent an automated message when an image was uploaded to the database. The adjudicator was blinded to the study treatment, participant history, and visit sequence and reviewed each individual image to validate the wound boundaries were drawn correctly and to ensure intrareliability. Images collected at enrollment, randomization, after 4 weeks of study treatment, and final study visit were validated for all participants. The adjudicator confirmed that ulcer images were indicated as healed or recidivated.
Inclusion/Exclusion. Eligible participants ranged from 18 to 90 years of age, had a VLU >30 days’ duration, with a size between 4 cm2 to 50 cm2, and a documented etiology of venous stasis with reflux. Participants were required to have demonstrated adequate arterial flow in their index extremity as documented by an ankle brachial index (ABI) 0.8 to 1.2, a biphasic or triphasic Doppler reading, or a TcPO2 or toe pressure >40 mm Hg.
Participants who had received other treatment modalities, including cellular-based tissue products, growth factor therapies, negative pressure therapy, hyperbaric oxygen, or ultrasound therapy, were required to have a minimum of 14 days since the last treatment. Participants were required to be a minimum of 6 weeks post vascular or skin graft procedure. If an ulcer was present for >6 months, a biopsy was required to rule out a malignancy. For participants who presented with multiple ulcers, the largest ulcer that met all other inclusion criteria was selected to be the index ulcer in this study.
Participants were excluded if the index ulcer’s primary etiology was other than venous disease; if the index ulcer was located within 1 cm of another ulcer; if the patient had cellulitis, osteomyelitis, or gangrene; or if there was exposed tendon, muscle, or bone. Participants with more than 5 ulcers on the index limb, an amputation of the study limb above the transmetatarsal region, or with confounding treatments or comorbidities were excluded. Confirmation of participant inclusion/exclusion eligibility was performed by a blinded Physician Advisory Board.
Study phases. The clinical trial was comprised of 4 phases: 1) screening and enrollment; 2) 2-week run-in; 3) 4-week treatment; and 4) 7-week follow-up. Participants whose VLU healed during the trial period were eligible to participate in a 12-month ulcer recidivism registry substudy (see Figure 2).
Screening and enrollment. Participants were evaluated for eligibility based on medical records, including documentation of venous stasis and reflux within the past 12 months and/or performance of a venous duplex scan at enrollment. Participants were screened for arterial blood flow by conducting a bedside or vascular lab ABI, Doppler study, TcPO2, or toe pressure on day of enrollment. If a wound was of >6 months’ duration, a biopsy was performed to rule out other conditions. Upon confirmation of all eligibility criteria, a participant entered the run-in phase of the study.
Run-in. Eligible participants completed a 2-week run-in of standardized treatment. The protocol for SC treatment was based on societal guidelines for VLU9,10 and consisted of compression (30–40 mm Hg), moist wound environment dressings, and debridement at the investigator’s discretion. A 4-layer compression wrap (Profore™ Smith & Nephew, Hull, UK) was provided for the study. If a participant could not tolerate this compression, other forms that met a minimum level of 30 mm Hg were allowed. All participants were seen a minimum of once a week during the run-in phase. Per protocol, participants could be seen up to 3 times per week for ulcer cleansing and debridement, dressing changes, or compression reapplication as deemed medically necessary by the study site investigator.
Randomization. Participants that exhibited an index ulcer area reduction <30% during the 2-week run-in were randomized. Participants whose ulcers reduced >30% during the run-in were deemed nonstalled VLUs and were exited before randomization. Randomization eligibility was based on a calculation of percent area reduction made by the clinical database using the enrollment and randomization visit ulcer sizes entered by the study site. Participants meeting the randomization criteria were issued a 1:1 randomization assignment by the clinical database to either SC or NLFU+SC. Study personnel and participants were not blinded to the study treatment.
Study treatment. Participants received their assigned study treatment for 4 weeks. Participants randomized to the control group (SC) continued with the protocol-required SC treatment (wound cleansing, dressing change, compression, debridement as needed) for a minimum of 1 and up to 3 times per week as determined clinically appropriate by the investigator. Participants randomized to the intervention group (NLFU+SC) received the protocol-required SC plus NLFU 3 times per week for a total of 12 treatments. NLFU treatment therapy duration was delivered based on the ulcer size determined at each treatment visit, per manufacturer’s recommendations, starting at 3 minutes for wound areas >10 cm2, with increases of 1 minute for every 10 cm2 increment.20
Follow-up. Upon completion of the treatment phase and evaluation for the primary endpoint, participants were followed for an additional 7 weeks to evaluate secondary endpoints of ulcer closure. The protocol-required SC treatment continued for all participants throughout the follow-up phase. Additional modalities, including crossover to NLFU therapy for SC participants, could be introduced at the discretion of the investigator. A final ulcer measurement was performed at the last study visit, 11 weeks post randomization.
Study variables, data collection, and management. Baseline patient demographics, comorbidities, index limb and compression history, and index ulcer and treatment history were collected from the medical record and entered by the study sites into the clinical database using the EDC system. Clinical characteristics and duplex results documenting venous insufficiency, arterial blood flow measurements, and ulcer biopsies were recorded. At each weekly visit, a single trained research coordinator at each study site performed a digital ulcer measurement and drew the wound epithelial border that was then confirmed by the investigator and adjudicated by a single third-party blinded wound expert. The investigator assessed the ulcer bed tissue and exudate, periwound characteristics, and edema level of the index limb each week. Participant compression adherence and the ulcer treatments performed were entered into the record at each study visit.
Participants completed pain and QOL questionnaires at time of randomization and at the completion of 4 weeks of study treatment. Pain assessments were completed using the Visual Analog Scale (VAS) and recorded in the database via the EDC system.21 VAS measurements were verified for accuracy by the study monitor. QOL was measured using the Short Form-36 Health Survey (SF-36).22 Participant-completed SF-36 forms were collected by the study monitor, and responses were entered into the QualityMetric Health Outcomes™ Scoring Software 4.5. The software-calculated scores were transferred to the SAS version 9.4 software for analysis by the third-party statistician.
Endpoints. The primary trial endpoint was mean percent ulcer area reduction, calculated as the ulcer size differences from randomization to the 4-week, post treatment measurement, for each treatment group.
Secondary endpoints included actual area reduction from randomization to 4 weeks of study treatment, weekly percentage and actual ulcer area reduction, mean ulcer area reduction from randomization to 4 weeks of study treatment stratified by ulcer size and age, ulcer area reduction in ulcers treated with bioengineered skin equivalents in the follow-up, ulcer healing rates and times to healing defined as full epithelialization of the ulcer, ulcer recidivism rates and times, pain score comparison between randomization and 4 weeks of study treatment, and QOL score comparison between randomization and 4 weeks of study treatment.
Statistical analysis. Statistical analysis was performed with SAS version 9.4 software (SAS Institute Inc, Cary, NC), and Fisher’s Exact test or Student’s t-test and the Wilcoxon Rank Sum tests were used to compare the treatment arms with both nonadjusted and adjusted results. A multivariate model was used to test the primary endpoint. The model included published covariates of wound size and wound age23 as well as variables of diabetes mellitus, coronary artery disease, hypertension, anemia, wound location, and compression adequacy. The study had a group-sequential design with one planned interim analysis for superiority at 87 randomized participants. The full study was 80% powered to detect a 15% difference (SD 33%) with a 2-sided overall alpha level of 0.05, requiring 156 randomized participants (78 participants per arm). Statistical significance was defined as P <0.05. Statistical differences between the treatment arms were evaluated for impact to the primary endpoint.
Study revisions. Two revisions to the study protocol occurred during the trial. The first revision occurred at the beginning of enrollment and included removal of the maximum ulcer duration and expanded the ulcer size lower range from 10 cm2 to 4 cm2. All enrolled participants met these criteria. The DSMB guided the second protocol revision in response to eligibility violations and randomization failure rates. Compression requirements per SVS and AVF guidelines were added for the 30-day chronic ulcer period. Documentation requirements of venous stasis and reflux were clarified. New criteria of a biopsy to rule out malignancy for wounds >6 months’ duration and a maximum of 5 wounds on the index limb were added. An evaluation of the impact of protocol revisions to the primary endpoint was conducted with no significant differences detected, and no adjustment for differences was required.
Due to slower-than-anticipated enrollments and randomization rates, the study sponsor decided to stop enrolling, forego the prespecified interim analysis, and report current results as final. No alpha penalty was taken because the data were only given a single look. The data presented in this report meet intent-to-treat principles per International Conference on Harmonization and Food and Drug Administration Guidelines for Statistical Principles.24,25