Patients who participated in a Phase 2 trial of HP802-247 for venous leg ulcers were invited to participate in this 24-week follow-up study to assess the durability of healing, document additional ulcer closures, and evaluate posttreatment safety. Consent was given by 90% (206/228), with 80% (183/228) completing all visits. Blinding was retained from the previous trial in which subjects had been randomized to vehicle or one of four cell therapy regimens. Visits were every 8 weeks. Among the 183 subjects, 43% (21/49) previously treated with cells and entering follow-up with an open wound achieved closure, compared with 35% (7/20) previously treated with vehicle, while 10% (11/106) and 17% (3/18), respectively, experienced reopening of a previously closed wound. Subjects previously treated with cells closed more open wounds than those previously treated with vehicle (OR 1.39, 95% CI 0.47–4.10; p = 0.739), and less subjects with a previously closed wound reopened (OR 0.65, CI 0.16–2.60; p = 0.821); however, these findings were not statistically significant. At the final visit, the difference in proportion of subjects with wounds closed continued to favor the best dose from the prior trial (83% closed vs. 58%, delta 25%). Follow-up beyond 12 weeks is necessary to evaluate the full benefit of this therapy, as treatment with cells may provide stimulus toward healing that persists for up to several weeks following the last application. The results show that the greater proportional benefit achieved by HP802-247 relative to standard care after 12 weeks of treatment persists over a meaningful timeframe.
Venous leg ulcers (VLU) are common sequela of chronic venous insufficiency of the legs, with a prevalence of between 1.65 and 1.74% in the elderly population. Compression therapy with control of infection remains the standard treatment for VLU; however, a substantial percentage (between 30 and 75%) of ulcers do not respond and become chronic.
HP802-247 is a living cell bioformulation consisting of human, growth-arrested, allogeneic keratinocytes and fibroblasts in a self-assembling fibrin matrix. Allogeneic skin cells applied to wounds persist for no more than a few weeks. This has been shown previously with bi-layered constructs,[5, 6] and was confirmed for HP802-247 when applied to acute wounds. Because allogeneic cells do not engraft, there is no need for a complex tissue construct. Instead, it is the cells alone, and the molecular products the cells release, that stimulate healing.
Two Phase 2 dose response trials have been conducted in patients with VLU unresponsive to standard care, where the ratios, concentrations, and frequency of cell application were examined.[9, 10] In each of these studies, treatment of chronic VLU with HP802-247 together with compression therapy resulted in the healing of a greater proportion of ulcers and in a shorter period of time than for vehicle with compression therapy. These differences were significant and clinically meaningful. In the most recent trial using neonatal cells, 0.5 × 106 cells/mL (keratinocyte : fibroblast = 1 : 9) applied every 14 days (0.5M Q14D) was found to be the most efficacious dose, with 70% of ulcers achieving complete closure within 12 weeks with median time to closure = 50 days, compared with 46% and median time to closure of 71 days for vehicle. The primary endpoint in that trial, average percent decrease in wound area over 12 weeks of treatment, included an additional 2-week confirmatory evaluation for subjects who achieved complete wound closure. The present study allowed subjects from this second Phase 2 trial to be observed for an additional 24 weeks of follow-up on completion of their participation in the treatment protocol. Long-term follow-up is essential for the assessment of long-term outcomes including maintenance of healing and continued effects of the treatment.[11, 12]
The objective of this study was to examine the 6-month durability of VLU closures occurring during the Phase 2 treatment protocol, record new closures and reopenings, and to continue to monitor the safety of HP802-247 through analysis of adverse events and status of the periwound area.
Study design and participants
This was a 24-week, noninterventional, observational follow-up study for subjects who had participated in the Phase 2, multicenter, randomized, vehicle-controlled treatment protocol, conducted at 28 outpatient facilities in the US between June 1, 2009 and October 20, 2011. The study adhered to the ethical guidelines of the Declaration of Helsinki, was Institutional Review Board approved, and registered at ClinicalTrials.gov. All subjects provided written informed consent prior to participation.
Subjects were evaluated according to the five treatment groups which had been determined by randomization in the preceding treatment protocol. Blinding was maintained for this follow-up study.
Group I: Vehicle applied every 7 days for 12 weeks + 4-layer compression (Vehicle)
Group II: 0.5 × 106 cells/mL every 7 days for 12 weeks + 4-layer compression (0.5M Q7D)
Group III: 5.0 × 106 cells/mL every 7 days for 12 weeks + 4-layer compression (5M Q7D)
Group IV: 0.5 × 106 cells/mL every 14 days, vehicle on alternate weeks, for 12 weeks + 4-layer compression (0.5M Q14D)
Group V: 5.0 × 106 cells/mL every 14 days, vehicle on alternate weeks, for 12 weeks + 4-layer compression (5M Q14D)
This study was open to all 227 subjects who had participated in the Phase 2 treatment study and had received at least one application of study treatment. Key eligibility criteria for the Phase 2 treatment study included a diagnosis of venous reflux confirmed by Doppler ultrasound; a VLU between 2 and 12 cm2, located between the knee and the ankle at or above the malleolus, without exposed tendon, muscle, or bone; and duration of 6–104 weeks. Subjects must also have had adequate perfusion to the target limb (Great toe pressure ≥ 50 mmHg, or ABI ≥ 0.8, or TCPO2 ≥ 40 mmHg). Full details of original study entry criteria are published elsewhere.
Randomization and masking
In the current trial, subjects were grouped according to the treatment assignment of the previous treatment trial. Randomization details for the treatment study are described in detail elsewhere. Masking from the double-blind treatment trial was maintained for the duration of this follow-up study, and trial sponsor, trial monitors, statisticians, investigators, center personnel, and subjects were masked to treatment group assignments until data verification was completed and the database locked.
The primary outcome variable was the persistence of wound closure as assessed by the target wound status (open, closed, or reopened). Secondary outcome was the number of subjects with target wound closed for the first time. Wound status was determined by visual inspection. Periwound area was assessed for clinical changes such as erythema, dermatitis, edema, maceration, or excoriation; these were recorded as safety outcome variables.
Subjects were enrolled at the final visit of the Phase 2 dose response trial which also served as Visit 0 of the present study. Subjects returned for Visits 1–3 at 8-week (56 day ± 14) intervals. At each visit, wound status, adverse events related to the target ulcer, target wound therapies or treatments, and status of the periwound area were recorded.
Subjects who had not healed their wound during the prior dose response trial were expected to receive standard medical care for their ulcer during this follow-up study, as directed by the Investigator. The present study did not specify or restrict any particular course of treatment, nor did it provide any therapeutics.
All enrolled subjects were evaluable for efficacy. The nonparametric Fisher's Exact test was used to analyze both the primary and secondary endpoints. The test was performed separately for each pairwise comparison of an active (cells) regimen vs. vehicle. Statistical significance was set at 5%. SAS version 9.1.3 (SAS Institute Inc., Cary, NC) was used for all analyses. Missing data due to either missed visits or to subject discontinuation were treated as missing. No imputation was performed.
Subject disposition and demographics
There were 206 (91%) of the 227 qualified subjects from the Phase 2 dose response trial who were enrolled in this follow-up study. The groups (from the original study) with the least and greatest efficacy in the treatment trial, Vehicle and 0.5M Q14D, had the greatest number of subjects who withheld consent for participation (6 and 8, respectively), while the other groups had only 2 or 3 subjects who withheld consent. Those who did consent were representative of the overall treatment groups (from the treatment study) with respect to the proportion with healed wounds. The greatest difference in proportion healed between those eligible to enroll and those enrolled (6%) was in the 0.5M Q14D group (Table 1).
Table 1. Subject demographics and wound baseline characteristics
Treatment group assignment from Phase 2 treatment study
*All of the 227 subjects in the ITT population of the previous study.
0.5M = 0.5 × 106 cells/mL; 5M = 5.0 × 106 cells/mL; Q14D = every 14 days; Q7D = every 7 days; SD = standard deviation.
Wound status (all subjects, at end of treatment study)*
Wound status (at enrollment in current study)
Of the 206 subjects enrolled, 183 (89%) completed all visits and 23 discontinued early (Figure 1). The main reasons for discontinuation were withdrawal of consent and lost to follow-up. As with enrollment, dropouts were reasonably uniform in number across the groups. Demographic variables were well balanced at baseline. These are presented along with baseline wound characteristics in Table 1.
Persistence of wound closure
At the completion of follow-up, at least three-fourths of all wounds that were closed at the beginning of this 24-week follow-up study maintained closure throughout the study, regardless of treatment group. With the exception of the 5M Q7D group, persistence of closure was ≥81% (Table 2). When all wounds closed at the beginning of the study that were also closed at the end of the study are considered, there is a difference of 10% between the Vehicle and the 0.5M Q14D group, 83 and 93% remaining closed, respectively.
Table 2. Distribution of the three possible outcomes for subjects* who entered the follow-up study with a closed wound (reopen, reopen and reclose, remain closed)
Wound status n (%)
*Includes all subjects completing final follow-up visit; does not include subjects discontinuing study prior to final follow-up visit.
At the completion of follow-up, 41% of ulcers open at the start of the study had achieved closure. The number closed by treatment group is listed in Table 3. Over the duration of the study, for each treatment group, a greater number of open wounds closed, compared with the number of closed wounds that reopened, with the greatest difference in the 0.5M Q14D group. There was net gain in wound closures for all treatment groups by the completion of the follow-up study ranging from an additional 4 to 13%. Subjects previously treated with cells closed more open wounds than those previously treated with vehicle (OR 1.39, 95% CI 0.47–4.10; p = 0.739), and less subjects with a previously closed wound reopened (OR 0.65, CI 0.16–2.60; p = 0.821); however, these findings were not statistically significant (Figure 2). The difference between the optimal dose identified in the previous trial (0.5M Q14D) and the Vehicle group remained essentially unchanged at 25% better (83% vs. 58% closed at the end of the current study compared with 70 and 46% closed at the end of the original treatment study) in favor of the cell treated group (Figure 3).
Table 3. Ratio of new wound closures to wound re-openings*
Percent of open wounds that closed (n)
Percent of closed wounds that reopened (n)
*Does not include subjects discontinuing study prior to final follow-up visit.
Table 3 summarizes the numbers and proportions of subjects with open wounds that closed and with closed wounds that reopened at the end of follow-up.
Use of compression
Depending on the treatment group, compression stockings or bandages were used by 30 to 72% of the subjects with wounds or to maintain closure throughout follow-up. Those groups with >80% persistence of closure had compression (stockings or bandages) used for 33, 72, 58, and 52% of subjects (Vehicle, 0.5M Q14D, 0.5M Q7D, 5M Q14D, respectively). The lowest utilization of compression was 30%, recorded for the 5M Q7D, the treatment group with the lowest percentage (74%) of ulcers with persistent closure. Similarly, for those subjects with closed wounds that reopened, the lowest use of compression was for the 5M Q7D group (33%).
Periwound signs and symptoms
The status of the periwound area was assessed at each of the follow-up visits as a safety variable. Periwound skin that was erythematous, eczematous, edematous, macerated, excoriated, or any combination of these signs and symptoms was recorded by the Investigator. The number of subjects with at least one of these findings was maximum (74) at Visit 1 (8 weeks), decreasing at subsequent visits to 62 and 56 at Visits 2 and 3 (16 and 24 weeks). The number of subjects with at least one finding was similar across all treatment groups, although generally highest for the Vehicle group and lowest for the 0.5M Q14D group.
During this noninterventional follow-up study, Investigators were asked to record only those adverse events judged as related to test article and to continue to follow ongoing adverse events originating in the Phase 2 treatment study. Only one new adverse event (infection of target ulcer) was considered by the Investigator to be related to test article. This was for a subject in the 5M Q14D treatment group. Despite the assessment of relatedness for this single case, target ulcer infection was reported for all treatment groups, with 4 for the Vehicle group and only 1 for each of the active groups, and none of these other cases were considered to be related to test article.
In this Phase 2 follow-up trial, we report durable and persistent healing of refractory venous ulcers treated with a novel allogeneic cell therapy, HP802-247. Only 12% (overall) of closed ulcers reopened at any point in the 6 months of follow-up and 43% of open wounds treated with cell therapy closed during the 6-month follow-up. We also found the difference in the proportion of wounds closed between Vehicle and the best cell dose (0.5M Q14D) at the end of follow-up remained the same as was seen at the end of the 12-week treatment period. This suggests a persistent benefit of cell therapy.
Previously, two Phase 2 clinical trials of HP802-247 showed that growth-arrested human allogeneic keratinocytes and fibroblasts applied in a fibrin spray provide an effective means of healing chronic VLU.[9, 10] The second of these two trials, using neonatal cells, was designed to record healing, including wound closure, over a maximum of 12 weeks, and established that the most efficacious dose/frequency was 0.5M Q14D. While this time interval was sufficient to compare the different dose levels and dose frequencies, durability of closure requires longer follow-up than the 2-week postclosure included in the treatment trial. The current follow-up study provided an additional 24 weeks of observation, immediately following the completion of treatment, sufficient to distinguish durable wound healing from transient wound coverage, determine if the cells affect the durability of wound closure, and monitor for adverse effects on surrounding tissue.
Depending on baseline characteristics of wounds treated, at least 25% of VLU are expected to remain open following application of good standard care. Treatment of these ulcers with HP802-247 resulted in complete healing of 64% (all active groups combined) over the 12-week treatment period compared with 46% in the Vehicle group. The best dose/frequency, 0.5M Q14D, achieved complete healing in 70%. Continued healing and additional new closures were recorded for all treatment groups during this follow-up study, with 52% of open ulcers at study entry closed at the end of follow-up for the three treatment groups showing best efficacy in the previous trial (vs. 35% for Vehicle, p = 0.270, two-sided Fisher's exact test). Although this difference does not reach statistical significance, it is reasonable to conclude that continued healing beyond 12 weeks was due in part to the effectiveness of compression bandaging, while the difference between active and Vehicle groups was due to beneficial effects of treatment with HP802-247. We have previously shown that HP802-247 cells persist for approximately 14 days following application to acute wounds. Improved healing during follow-up was therefore not due to continued presence of the growth-arrested cells; this healing was instead a downstream result of prior effects set in motion by the treatment on the host tissue.
The majority of subjects entering this follow-up study with closed wounds showed persistent closure at each visit. Persistence of closure was essentially the same for subjects who had previously been treated with cells, or treated with vehicle. This suggests that healing mechanisms were quite similar for those ulcers that healed, regardless of treatment assignment and likely dependent on endogenous cellular activity.
Subjects who had been treated with cells in the treatment trial showed a higher net gain (wound closure vs. wound reopening) compared with those exposed to vehicle. The best dose, 0.5M Q14D, maintained a 19 to 25% advantage in closure over the vehicle control group at each follow-up visit. The achievement of complete wound closure in over 80% of chronic VLU as seen with HP802-247 treatment (0.5M Q14D) at the end of this follow-up study is a singularly impressive outcome.
The idea of delivering skin cells to a wound for the purpose of either engraftment or stimulating reepithelialization by the host is not new. Methods have included autologous or allogeneic source material and simple suspensions or complex constructs for delivery. Autologous therapies include, for example, constructs consisting of layered, differentiated keratinocytes produced from keratinocyte progenitors, and cell suspensions in fibrin. Autologous methods typically require a period of 2–3 weeks for cell expansion before the cells can be applied to the wound. Allogeneic cells have the advantage that they can be prepared for use, and stored until needed. Currently, marketed products include a bilayered construct that includes keratinocytes and dermal fibroblasts in distinct layers, and an artificial dermis containing fibroblasts. In contrast, HP802-247 bears no histological similarity to differentiated skin. The spray-delivered neonatal cells as applied are therefore in a less differentiated state, by design, than the cellular components of engineered skin substitutes. Growth-arrest of the HP802-247 cells by gamma-irradiation is also thought to contribute to the overall efficacy by focusing cellular activities away from growth and replication.
The findings of the present study support a conclusion that closure achieved with cells is persistent and closure with either cells, or vehicle in the treatment trial are equally durable. For wounds not yet healed, prior treatment may have provided a continued stimulus toward healing that is persistent for up to several weeks following the last application.
Source of Funding: Healthpoint Biotherapeutics provided financial support and was the sponsor of the study.
Conflicts of Interest: R.S.K., W.A.M., and R.J.S. received consulting fees for their assistance with protocol design and data interpretation. Y.Z. received consulting fees for biostatistical analysis and data interpretation. H.B.S., J.E.D., D.I.C., and T.D.L. are employed by Healthpoint Biotherapeutics. H.B.S., J.E.D., and D.I.C. hold adjunct appointments at the University of North Texas Health Science Center in Fort Worth, TX.