Effects of Transforming Growth Factor ß2 on Wound Healing in Diabetic Foot Ulcers:

A Randomized Controlled Safety
and Dose-Ranging Trial

 

Martin C. Robson, MD, FACS^*

David L. Steed, MD, FACS^†

John M. McPherson, PhD^‡

Bruce M. Pratt, PhD^‡

For the TGF ß2 Study Group

 

^*Institute for Tissue Regeneration, Repair and Rehabilitation, Bay Pines, FL

^†Department of Surgery, University of Pittsburgh School of Medicine, Pittsburgh, PA

^‡Cell & Protein Therapeutics Department, Genzyme Corporation, Framingham, MA

This clinical trial was funded by Genzyme Tissue Repair, Genzyme Corporation

1 Mountain Rd

Framingham MA, 01701-9322

 

This work was presented in part at the 3rd Joint Meeting of the European Tissue Repair Society and the Wound Healing Society, Bordeaux, France, 1999, and the 6th World Biomaterials Congress, Kamuela, Hawaii, 2000.

 

KEY WORDS:  diabetes, TGF-ß2, ulcer, wound healing

Abstract

This randomized, double blinded, placebo- controlled, multi-center trial was undertaken to establish the safety and effective dose level of transforming growth factor ß2 (TGF-ß2) required to improve the healing of chronic foot ulcers in diabetic patients. Patients with chronic diabetic foot ulcers were randomized into five groups: standardized care (sharp debridement, coverage with nonadherent dressing, and weight off-loading), topical placebo collagen sponge, or topical collagen sponges containing TGF-ß2 at either 0.05 µg/cm2, 0.5 µg/cm2, or 5.0 µg/cm2. All patients receiving sponges also received standardized care. Primary efficacy outcome measures were complete closure of the wound and percentage of wound area reduction at or before week 21. Secondary endpoints were time to wound closure and durability of wound closure. Safety was evaluated by clinical course and adverse event monitoring.

TGF-ß2 was safe at all dosages tested. Within the blinded portion of the trial, patient groups treated with placebo, 0.05 µg/cm2, 0.5 µg/cm2, or 5.0 µg/cm2 TGF-ß2 had closure frequencies of 32%, 58%, 57%, and 61%, respectively. The unblinded standardized care group had a closure frequency of 71%. Percentage of wound area reduction (compared with placebo sponge) at the end of treatment was significantly enhanced in the 5.0 µg/cm2 TGF-ß2 and standardized care groups. Median time to complete closure (compared with placebo sponge) was significantly reduced in the 5.0 µg/cm2 TGF-ß2 and standardized care groups. Durability of complete wound closure at 3 months after successful completion of therapy was comparable among all 5 treatment groups. The high and low doses of TGF-ß2 and the standardized care arm showed significantly improved wound healing frequencies compared with the placebo sponge group. The placebo vehicle (collagen sponge) was previously shown in both animal and clinical studies to have no adverse effect on wound healing. Additionally, the frequency of wound healing in the standardized care arm of this trial was much higher than comparable standardized care results in other large clinical trials of diabetic ulcer wound healing. The data suggest that (1) the standardized care healing rate in this trial may have been a statistical anomaly of small sample size and (2) repeated topical delivery of TGF-ß2 may have a beneficial effect on the healing of chronic diabetic foot ulcers.

Introduction

Transforming growth factor ß2 (TGF-ß2) is one of three isoforms of TGF-ß that has been identified in mammals. Numerous in vitro and in vivo studies, including gene knockout experiments, have provided clear evidence that the TGF-ß family members play central roles in embryonic development, immune modulation, and hard and soft tissue wound repair.1 In soft tissue wounds TGF-ß has been shown to stimulate wound contraction and induce the formation of granulation tissue and extracellular matrix.2

Clinical interest in TGF-ß as a potential agent to stimulate the healing of chronic cutaneous wounds was initially stimulated by results from animal experiments showing that subcutaneous delivery of TGF-ß1 stimulated formation of granulation tissue in both adult and neonatal rats.3–5 Subsequent studies demonstrated that topical administration of TGF-ß1 stimulated the healing of cutaneous wounds in both normal and healing-impaired animals.6 Similar results were obtained with TGF-ß2 in animal models of cutaneous wound repair that healed by primary and secondary intent.7–10

Based on these and other experimental studies in animal models, two phase I/II clinical trials were performed with TGF-ß2 in patients with venous stasis ulcers. In these studies, TGF-ß2 was delivered in a collagen/heparin sponge that had previously been shown to be biocompatible, resorbable, and capable of effective delivery of the growth factor.11 The results of these 2 studies provided evidence that, in combination with compression therapy, repeat topical administration of TGF-ß2 at doses ranging from 0.5–2.5 µg/cm2 in a collagen heparin sponge was safe and could accelerate the rate of wound closure as compared with standard compression therapy alone or compression therapy in combination with the placebo sponge. Furthermore, the data from these trials indicated that the healing rate in the standard of care group was similar to that in the placebo sponge group, suggesting that the collagen/heparin sponge delivery vehicle did not inhibit cutaneous wound healing in this patient group.

Based on these preclinical and early clinical results, a phase II dose-ranging study was initiated in patients with chronic diabetic foot ulcers using the same collagen/heparin delivery used in the venous stasis ulcer trial. Diabetic foot ulcers were chosen for further clinical testing of TGF-ß2 for a variety of reasons, including (1) a large patient population with these types of chronic wounds (a minimum of 700,000 patients per year in the United States);12 (2) the high degree of morbidity associated with chronic diabetic foot ulcers, eg, gangrene and amputation; and (3) the belief that induction of granulation tissue and stimulation of wound contraction could potentially enhance the healing of these types of cutaneous wounds. This phase II, randomized, blinded, placebo-controlled, multicenter clinical trial was designed to assess both the safety and efficacy of TGF-ß2 applied twice weekly, in single use lyophilized collagen/ heparin sponges, at doses of 0.05, 0.5, or 5.0 µg/cm2 of wound surface area for up to 20 weeks. The healing times for the various TGF-ß2 treatment groups were compared with those for patients receiving the placebo collagen/heparin sponge alone as well as a group of patients treated with only standard care. This same standard care was applied to all patients enrolled in this trial.

Methods

This randomized, double-blinded, placebo-controlled, multicenter trial was conducted at 15 centers in the United States in accordance with the principles of Good Clinical Practice and the Declaration of Helsinki, as well as with the approval of institutional review boards and written consent from all patients. The trial was conducted between December 1995 and October 1998.

Patients were eligible for inclusion if they were at least 18 years of age, had diabetes mellitus and a neuropathic ulcer present for at least 8 weeks on the plantar surface of the forefoot, toes, metatarsals, or dorsum of the foot. Additionally, after debridement, the ulcer must have been between 1cm2 and 20 cm2 in area and full thickness without exposed bone or tendon. The patient must also have had adequate peripheral arterial circulation as determined by an ankle/brachial index between 0.7 and 1.3, or a transcutaneous oxygen pressure measurement on the foot of 30 mm Hg or more. Exclusion criteria included radiographically documented osteomyelitis, clinical infection of the ulcer, use of systemic steroids within the previous 30 days, HgA_1c greater than 13%, serum creatinine greater than 2.5 mg/dL or serum albumin less than 2 mg/dL.

For patient assignment, Genzyme Corporation created computer-generated treatment randomization lists in which sequential patient numbers were randomly assigned to 1 of 5 treatment groups. Two lists of treatment randomization were provided to each clinical center: one list for patients with ulcers between 1 and 5 cm2 and the second for patients with ulcers larger than 5 cm2 but smaller than 20 cm2. Patients were consecutively assigned a patient number from the appropriate randomized list as they entered the trial. The 5 treatment groups were (1) standardized care (sharp debridement, coverage with nonadherent dressing, and weight off-loading from the affected foot), (2) topical placebo collagen sponge, and topical collagen sponges containing recombinant human TGF-ß2 at (3) 0.05 µg/cm2, (4) 0.5 µg/cm2, or (5) 5.0 µg/cm2. All patients who received sponges also received standardized care. Double blinding for all 4 groups receiving collagen sponges was maintained by coded labeling of all collagen-sponge packaging materials. The coded labeling key, generated by the Biometrics department at Genzyme Corporation, was unavailable to other Genzyme staff, clinical investigators, or patients. The standardized care group could not be blinded since these patients did not receive a collagen sponge like the other four treatment groups.

At the first treatment visit of eligible consenting patients, baseline physical examination and clinical laboratory measurements were obtained. Specific ulcer evaluations included photography, size, and depth measurements. For all patients, the ulcer was sharply debrided and the ulcer edges beveled. For patients receiving collagen sponge, the sponge was trimmed to fit the wound size and placed directly onto the moist wound bed. For all patients, the ulcer was then covered with non-adherent dressing and wrapped with gauze bandaging. All patients were educated on the need for pressure relief (off-loading) from the ulcer site. The method of pressure relief was not mandated by the study design. A variety of pressure relief devices and methods were permitted as long as no pressure was placed on the ulcer and the device did not touch the ulcer site.

Dressing changes and additional sponge placements were required twice weekly. At least one of the dressing changes/treatments per week was performed at the clinic, while the other was usually performed at home by the patient or caregiver. At each weekly clinic visit, the patient was monitored for adverse events and the ulcer was evaluated for complete closure. If the ulcer was not closed, the ulcer was evaluated for area, depth, and overall condition. The areas of unclosed ulcers were obtained using digitized image analysis (Bio-Imaging Technologies, Inc., West Trenton, NJ) of ulcer outline tracings made on double thickness plastic sheets. The depth of unclosed ulcers was determined by orthogonal placing of a sterile cotton-tipped applicator into the deepest portion of the ulcer and measuring how much of the applicator was below the surface of the surrounding skin. For nonhealed ulcers, sharp debridement was performed as necessary, fresh sponges were placed in the wound, and the wound was re-dressed. If, however, clinical infection of the ulcer or osteomyelitis was observed, treatment was suspended and the infection was treated according to best judgment of the physician. If the infection resolved within the 20 week intervention period, treatment could be resumed. However, missed treatments, ie, application of sponges, could not be made up. Complete closure was defined as 100% re-epithelialization of the wound with no visible broken areas in the epidermis and no drainage from the ulcer site. At week 21 or at the time of complete healing, the patients entered the follow-up phase of the trial. During this period, the patients continued pressure relief measures for at least 1 month and returned to the clinic for monthly visits, for a total of 3 months follow-up. At each visit, the patient’s ulcer was evaluated for durability of healing and adverse events. If the ulcer was not healed, it was measured for area and depth.

 

Outcome Assessments

The primary endpoint measures of this trial were the (1) proportion of patients with complete ulcer closure at or before week 21, and (2) percentage of ulcer area reduction by week 21. The secondary endpoint measures were (1) the proportion of patients with complete closure of the ulcer at each weekly visit, (2) the percentage of ulcer area reduction at each weekly visit, (3) time to wound closure, and (4) 3-month follow-up assessment of durability of wound closure.

 

Safety

Patients were evaluated monthly for vital signs, urinalysis, clinical chemistry, and CBC, and weekly for concomitant medications and ulcer status. They were evaluated at baseline and end of treatment for HgA_1c and 24-hour creatinine clearance. Additionally, the first 120 patients enrolled in this trial also received baseline and end-of-treatment retinal examinations and fundus photography. Because no treatment-related adverse ocular effects were observed in patients treated with sponges (placebo or TGF-ß2 containing), eye examinations were not required for the remaining patients.

 

Statistical Analyses

Baseline demographic parameters were summarized by treatment group. For safety data, continuous variables were compared with baseline using Wilcoxon’s signed-rank. For the primary efficacy endpoints, comparisons were made between the placebo sponge group and the other 4 groups, ie, standardized care and the 3 TGF-ß2-containing sponge groups. The proportion of patients with complete ulcer closure was computed for each treatment group, and the 95% confidence interval constructed. Comparisons to the placebo sponge group were made using linear logistic regression analysis. The percentage of wound area reduction was compared using non-parametric analysis. For the secondary endpoint of median time to complete ulcer closure, the comparison was made with the log-rank test. All endpoint analyses were conducted on intent to treat patients. Additionally, the Cox regression model was used to analyze the effects of baseline covariates on time to complete ulcer closure. Durability of wound closure was tabulated by treatment groups and compared by chi square contingency test.

Results

Overall, 177 patients were enrolled in 14 centers. Of these, 150 (85%) patients completed the treatment phase of the study and 139 (79%) completed the 3 month follow-up phase. The study profile and patient disposition are summarized in Figure 1. No differences in patient dropout rate during the trial or follow-up period were observed among the treatment groups. Patient demographic data are summarized in Table 1. Patient randomization appeared to be successful, in so far as no statistically significant differences between treatment groups were observed for any demographic or ulcer baseline characteristics. The average ulcer size within groups ranged from 2.1 to 2.7cm2, with 87% of the ulcers less than 5 cm2.

The proportion of patients in each treatment group with complete wound closure is shown in Figure 2 and Table 2. Within the blinded portion of the trial, TGF-ß2 at doses of 0.05 µg/cm2 (25 of 43, 58%, P=0.046) and 5.0 µg/cm2 (27 of 44, 61%, P=0.025), low and high dose respectively, produced statistically significant increases in the proportion of patients with total wound closure as compared with the placebo sponge-treated patients. The effect of the middle dose of TGF-ß2, 0.5 µg/cm2, approached but did not achieve statistical significance (25 of 44, 57%, P=0.056). The unblinded, standardized care treatment arm also achieved a highly significant increase (17 of 24, 71%, P=0.009) in total wound closure. A number of exploratory statistical analyses were performed to understand the basis of the unexpectedly high healing rate of patients in the standardized care arm. However, no covariates were identified that distinguished the standardized care arm from the other treatment groups.

The effects of TGF-ß2 on wound surface area reduction are summarized in Table 3. Within the blinded portion of the trial, TGF-ß2 at 5.0 µg/cm2 (high dose) achieved a statistically significant effect in wound surface area reduction (85%; P=0.041) when compared with the placebo sponge group, where complete closure of a wound is scored as 100%. The unblinded, standardized care treatment arm also achieved a significant effect (79%; P=0.047) in wound surface area reduction. Additionally, among wounds not closed at 21 weeks, the 5.0 µg/cm2 TGF-ß2 group (n=5) showed a 72% ± 30% reduction in wound area while the standardized care group (n=5) showed only a 25% ± 35% reduction in wound area.

The effects of TGF-ß2 on median time to complete wound closure are summarized in Table 4. Within the blinded portion of the trial, TGF-ß2 at 5.0 µg/cm2 (high dose) achieved a statistically significant reduction in median time to complete wound closure (P = 0.030) when compared with the placebo sponge group. The unblinded, standardized care treatment arm also achieved a significant effect (P = 0.009) in reduction of median time to complete wound closure. To further evaluate this endpoint, Cox regression analysis was performed to determine the effects of selected demographic and baseline covariates, ie, treatment arm, ulcer age, ulcer size, and serum albumin, age, creatinine clearance, HgA_1c, and smoking status, on the time to complete wound closure. These results are reported in Table 5. Of these covariates, TGF-ß2 (5.0 µg/cm2), standardized care, and serum albumin were significant positive covariates of time to complete wound closure, while baseline ulcer age and size were significant negative covariates of time to complete wound closure.

The durability of wound healing was assessed among patients whose wounds had completely closed at or before 20 weeks of treatment. Among these patients, the persistence of wound closure was evaluated at 3 months after completion of treatment. These results are summarized in Table 6. The mean proportion of wounds that remained healed was equivalent among all 5 treatment groups and was 68% (69 of 101).

All doses of TGF-ß2, as well as placebo sponge, were well tolerated in this study. A total of 572 adverse events were reported among 130 patients. The frequency of patients with one or more adverse events was similar in all treatment groups. The adverse events with highest frequency are reported in Table 7. Three patients died during the course of the study. The causes of death (myocardial infarction and respiratory failure, acute myocardial infarction, and heart attack) were judged by the clinical investigator not to be related to study treatment. Of the other 117 serious adverse events reported by 64 patients, only four were judged to have a possible relationship to the study drug, and none were judged to have definite relationships to the study drug. Of these four serious adverse events (occurring in three patients) three were cellulitis or infection of study or non-study ulcers, and one was possible osteomyelitis at the study ulcer site. All three of these patients recovered without sequelae.

The first 120 patients in the trial also received pretreatment and post-treatment ophthalmological examinations. Ten of these patients experienced adverse or serious adverse ocular events; however, none were judged to be treatment related. Consequently, with permission of the regulatory agency (FDA), ophthalmological examinations were not performed on the remainder of the patients. Four patients experienced serious adverse renal events. Again, however, all were judged to be unrelated to treatment.

Discussion

This trial was designed to test the safety and efficacy of the topical administration of TGF-ß2 for the treatment of diabetic neurotrophic foot ulcers. Within the blinded portion of this trial, this objective was achieved.

TGF-ß1, the most common human isoform of TGF-ß, is produced by many cell types in the body but is present in the highest concentrations in platelets and bone.13 Endogenous TGF-ß1 has been shown to be involved in several phases of the wound healing response: suppressing certain aspects of the acute inflammatory response, facilitating angiogenesis, stimulating extracellular matrix production, and stimulating matrix contraction. TGF-ß2, which is not generally present in platelets but is present in bone, exhibits essentially the same wound-healing properties as does TGF-ß1.2,7,9 These biologic activities of TGF-ß and its biodistribution in platelets and bone, along with healing deficits that have been reported in TGF-ß1 knockout mice, have led researchers to the conclusion that TGF-ß plays a central role in orchestrating the normal wound healing response following injury.14 As noted above, numerous preclinical experiments have demonstrated that topical use of either TGF-ß1 or TGF-ß2 in models of normal or impaired cutaneous wound healing significantly enhances the wound healing response.

Given these observations and the early positive clinical experience with TGF-ß2 in the treatment of venous stasis ulcers,11 a phase II trial in patients with diabetic foot ulcers was performed. The primary efficacy endpoints in this trial were frequency of complete wound closure and percentage of wound area reduction within 20 weeks of initiation of therapy. Secondary endpoints included time to complete wound closure and durability of wound closure at 3 months after completion of therapy. The primary efficacy endpoint was based on guidance from the FDA that complete wound closure was the most important and definitive endpoint for demonstration of efficacy and securing product approval of a biologically active agent that would claim to stimulate wound closure of chronic cutaneous wounds.15

The doses of TGF-ß2 chosen for this study were based on the previous clinical experience in venous stasis ulcers and several different animal studies that had shown stimulation of the wound healing response across a very broad dose range.8,16–18 Patients who received TGF-ß2-containing sponges achieved a frequency of healing of approximately 59% by the end of the study, independent of the dose of TGF-ß2 received. This healing rate of 59% compares very favorably with the aggregate healing rate of 43% reported for the use of platelet derived growth factor (PDGF; Regranex®) in the treatment of diabetic foot ulcers.19 Based on wound area reduction and median time to closure and covariate analysis, the highest dose of TGF-ß2, 5.0 µg/cm2, was shown to be most effective. The lowest dose of TGF-ß2 used in this study (0.05 µg/cm2) was expected to provide no effect or minimum effect. However, it also showed a significant healing rate of 57% at 21 weeks, albeit with a discernibly lower healing frequency at an intermediate time point. The efficacy of this low dose may be due to TGF-ß2-mediated stimulation of endogenous production of TGF-ß1 by cells in the wound healing environment.

The healing frequency observed in the unblinded standardized care arm of this trial (71%) was unexpected. Two interpretations are possible for this observation in this context. First, that for some unknown reason, the patients in the standardized care arm did better than patients in previously reported diabetic foot trials, or second, that these unexpectedly good healing rates are an anomaly of the treatment allotments. The authors believe it is the latter for several reasons. This result is inconsistent with published results from other wound healing trials that employed similar or nearly identical protocols of standardized care. In a meta-analysis of 586 individuals enrolled in the placebo/standard care arm of five randomized clinical trials, Kanter and Margolis19 found the healing rate to be 33%. Other chronic ulcer trials have reported standardized care healing rates between 30% and 40%.20–21 In a report covering 11 clinical trials of patients with diabetic neurotrophic foot ulcers, Robson et al21 reported that the healing rate of patients not receiving an experimental drug (ie, control patients) was 39%.21

The result from the present clinical trial is also inconsistent with the general clinical experience of the principal investigators of this trial. To probe the basis for this high standardized care healing rate, a number of exploratory statistical analyses were performed, eg, debridement frequency, clinical site variation, smoking status. However, none of these analyses revealed any demographic or treatment differences between the standard care and placebo treatment arms that could explain the observed differences in healing rates between these two groups. We therefore suggest that the 71% healing rate observed in the standardized care arm of this trial is a statistical anomaly of small sample size (n = 24 in the standardized care arm, distributed across 10 centers) and is probably representative of the healing rate of a standardized care arm in a larger clinical trial.

The healing frequency observed in the blinded, placebo sponge arm of this trial was 32%. This observation is also open to two interpretations. When compared with the healing frequencies of the standardized care and TGF-ß2 treatment arms, one possible interpretation is that the placebo sponge drastically inhibited wound healing and that the TGF-ß2 treatment partially reversed this effect. However, based on a variety of published clinical and preclinical data, we believe that this interpretation would be incorrect. First, previously published clinical data, most notably the PDGF trials (33% placebo healing rate, n = 254), have shown that placebo healing rates are in the 30% to 40% range.22 Furthermore, this range of healing rates is representative of the clinical experience of the principal investigators in the present trial.21 Second, no wound healing inhibition was seen with the clinical use of this sponge formulation in venous stasis ulcers.11 Furthermore, in that same clinical study, comparable rates of wound closure were observed in both placebo sponge and standardized care treatment arms, corroborating that the sponge caused no deleterious effect on healing. Third, in preclinical studies, use of this collagen sponge formulation also did not inhibit wound healing.23 The alternative interpretation is to conclude that the 32% healing rate among patients in the placebo sponge treatment arm is a reasonable estimate of the placebo healing rate for chronic skin ulcers and that the collagen sponge does not have any adverse effects on healing rates in this patient population. The authors favor this interpretation.

Conclusions

In summary, the data from the primary and secondary efficacy endpoints, as well as the safety data, indicate that the highest dose of TGF-ß2, 5 µg/cm2, is safe and, in this clinical trial, provided the greatest clinical benefit to patients. We also conclude that 5 µg/cm2 should be selected as the treatment dose in a follow-up phase III pivotal clinical trial designed to confirm the efficacy of this therapy in the treatment of diabetic foot ulcers. The design of a phase III trial with TGF-ß2 presents interesting challenges; it can be proposed that commercialization of a growth factor-based product for treatment of diabetic foot ulcers should provide a benefit to risk profile that is at least as good as Regranex®, the only FDA approved growth factor product for treatment of such wounds. We would suggest, therefore, that a phase III trial directly comparing the efficacy of TGF-ß2 with that of Regranex®, which would be consistent with the recommendations of the recently modified Declaration of Helsinki,24 would provide the necessary information to judge the utility of TGF-ß2 in treating diabetic foot ulcers.

Clinical Investigators

(Listed alphabetically) Charles Baxter, MD, Charles Baxter Wound Center, Dallas TX; Todd R. Bushman, DPM, Diabetic Foot Centers of America, Hermitage, TN; Elliot Chaikof, MD, Emory University Hospital, Atlanta GA; Kelman Cohen, MD, Medical College of Virginia, Richmond VA; Vincent Falanga, MD, University of Miami, School of Medicine, Miama FL; Sushil Gupta, MD, Raul Landa, MD, Metrowest Surgical Associates, PC, Framingham MA; Linda Harris, MD, John Ricotta, M.D., Millard Fillmore Hospital, Buffalo NY; Jeffrey Jensen, DPM, Diabetic Foot and Wound Center, Denver CO; Roy Kroeker, DPM, Diabetic Foot Center, Fresno, CA; Benjamin Lipsky, MD, Seattle VA Medical Center, Seattle WA; Richard Pollak, DPM, San Antonio Podiatry Associates, P.C, San Antonio TX; Martin Robson, MD, Institute for Tissue Regeneration, Repair and Rehabilitation, Bay Pines, FL; David Steed, MD, University of Pittsburgh Medical Center, Pittsburgh, PA; Daniel Walsh, MD, Dartmouth Hitchcock Medical Center, Lebanon NH; Samuel Wilson, MD, UCI Medical Center, Orange CA.

Acknowledgements

We acknowledge the important contributions and assistance of Beverly Accomando for biostatistical analyses, Doris Peterkin for program management, and Susan McAllister, Dana Hollingsworth, and Susan Szambelan for management of the clinical trial. We also acknowledge the editorial comment and review of Beverly Accomando, Doris Peterkin, and Susan McAllister.

 

References

1. Lawrence DA: Transforming growth factor-beta: a general review. Eur Cytokine Net 7:363-374, 1996.

2. Grande JP: Role of transforming growth factor-b in tissue injury and repair. Proc Soc Exp Biol Med 214:27–40, 1997.

3. Sporn MB, Roberts AB, Shull JH, et al: Polypeptide transforming growth factors isolated from bovine sources and used for wound healing in vivo. Science 219:1329–1331, 1983.

4. Sporn MB, Roberts AB, Wakefield LM, Assoian RK: Transforming growth factor-beta: biological function and chemical structure. Science 233:532–534, 1986.

5. Roberts AB, Sporn MB, Assoian RK, et al: Transforming growth factor type beta: Rapid induction of fibrosis and angiogenesis in vivo and stimulation of collagen formation in vitro. Proc Nat Acad Sci USA 83:4167–4171, 1986.

6. Roberts AB: Transforming growth factor b: Activity and efficacy in animal models of wound healing. Wound Repair Regen 3:408–418, 1995.

7. Ksander GA, Pratt BM, Desilets-Avis P, et al: Inhibition of connective tissue formation in dermal wounds covered with synthetic, moisture vapor-permeable dressings and its reversal by transforming growth factor-beta. J Invest Dermatol 95:195–201, 1990.

8. Ksander GA, Ogawa Y, Chu GH, et al: Exogenous transforming growth factor-beta 2 enhances connective tissue formation and wound strength in guinea pig dermal wounds healing by secondary intent. Ann Surg 211:288–294, 1990.

9. Cox DA, Kunz S, Cerletti N, et al: Wound healing in aged animals: Effects of locally applied transforming growth factor beta 2 in different model systems. EXS 61:287–295, 1992.

10. Wright TE, Hill DP, Ko F, et al: The effect of TGF-ß2 in various vehicles on incisional wound healing. Int J Surg Invest 2:133–43, 2000.

11. Robson MC, Phillip LG, Cooper DM, et al: Safety and effect of transforming growth factor-ß2 treatment of venous stasis ulcers. Wound Repair Regen 3:157–167, 1995.

12. Arrow, AK: Biotechnology in Wound Care, 4th ed. Los Angeles, Calif: Wedbush Morgan Securities; 1999.

13. Roberts AB, Sporn MB: The transforming growth factor-bs. In, MB Sporn, AB Roberts, eds: Peptide Growth Factors and Their Receptors, Handbook of Experimental Pharmacology. Berlin, Germany: Springer-Verlag; 1990:419–472.

14. Brown R, Ormsby I, Doetschman T, Greenhalgh D: Wound healing in the transforming growth factor-ß1-deficient mouse. Wound Repair Regen 3:25–36, 1995.

15. Stromberg K, Chapekar MS, Goldman BA, et al: Regulatory concerns in the development of topical recombinant opthalmic and cutaneous wound healing biologies. Wound Repair Regen 2:155–164, 1994.

16. Mustoe TA, Pierce GF, Morishima C, Deuel TF: Growth factor-induced acceleration of tissue repair through direct and inductive activities in a rabbit dermal ulcer model. J Clin Invest 87:694–703, 1991.

17. Mustoe TA, Pierce GF, Thomason A, et al: Accelerated healing of incisional wounds in rats induced by transforming growth factor-ß. Science 237:1333–1336, 1987.

18. Jones SC, Curtsinger LJ, Whalen JD, et al: Effect of topical recombinant TGF-ß on healing of partial thickness injuries. J Surg Res 51:344–352, 1991.

19. Kantor J, Margolis DJ: Expected healing rates for chronic wounds. Wounds 12:155–158, 2000.

20. Gentzkow GD, Iwasaki SD, Jensen JL, et al: Improved healing of diabetic foot ulcers after grafting with a living human dermal replacement. Wounds 11:77–84, 1999.

21. Robson MC, Hill DP, Woodske ME, Steed DL: Wound healing trajectories as predictors of effectiveness of therapeutic agents. Arch Surg 135:773–777, 2000.

22. Smiell JM, Wieman TJ, Steed DL, et al: Efficacy and safety of becaplermin (recombinant human platelet-derived growth factor-BB) in patients with nonhealing, lower extremity diabetic ulcers: A combined analysis of four randomized studies. Wound Repair Regen 7:335–346, 1999.

23. McPherson JM: The utility of collagen-based vehicles in delivery of growth factors for hard and soft tissue wound repair. Clin Mat 9:225–234, 1992.

24. Enserink M: Helsinki’s new clinical rules: Fewer placebos, more disclosure. Science 290:418–419, 2000.

 

 

     

 

Figure 1. Study Profile and Patient Disposition

 

   Completed Trt. Phase

n=36

 

Lost to Follow-up

n=1

 

Completed Follow-up

n=35

 

   Completed Trt. Phase

n=38

 

Lost to Follow-up

n=4

 

Completed Follow-up

n=34

 

Randomized

n = 177

 

   Completed Trt. Phase

n=40

 

Lost to Follow-up

n=1

 

Completed Follow-up

n=39

 

   Completed Trt. Phase

n=15

 

Lost to Follow-up

n=1

 

Completed Follow-up

n=14

 

   Completed Trt. Phase

n=21

 

Lost to Follow-up

n=4

 

Completed Follow-up

n=17

 

 

 

Table 1. Patient Demographic Data and Ulcer Baseline Characteristics

 

 

           Standardized                                            0.05 µg/cm2              0.5 µg/cm2            5.0 µg/cm2

                        Care                   Placebo                    TGF-ß2                    TGF-ß2                    TGF-ß2                             

 

 

  Gender (%)

    Male                 92                            82                            77                            77                          77

 

 

 

  Race (%)

    Caucasian       88                            82                            67                            77                          73

    Black                 4                              0                            12                              9                            5

 

  Smoking Status (%)

    Yes                 17                              9                            23                              7                          23

 

 

  Ulcer Age (weeks)*   59 (103)                    41 (47)                    51 (64)                    59 (74)                  54 (72)

 

*Measures given are mean (standard deviation).

 

 

Figure 2. Patients with complete wound closure.

 

 

 

 

Table 2. Effect of TGF-ß2 on Complete Wound Closure at Week 21

 

 

           Standardized                                            0.05 µg/cm2              0.5 µg/cm2            5.0 µg/cm2

                        Care                   Placebo                    TGF-ß2                    TGF-ß2                    TGF-ß2                             

 

 

  Not Healed*        7 (29%)       15 (68%)                 18 (42%)                 19 (43%)               17 (39%)

 

  P-value^†        0.009                                                       0.046                       0.056                     0.025

 

*Ten patients listed as either “not done” (ND) or “not applicable” (NA) at the week 21 visit were recorded as “not healed”. ND/NA patients are noted for each treatment group.

^†P value from Mantel-Haenszel Chi-square vs placebo sponge.

 

Table 3. Effect of TGF-ß2 on Wound Area Reduction at Week 21

 

 

           Standardized                                            0.05 µg/cm2              0.5 µg/cm2            5.0 µg/cm2

                        Care                   Placebo                    TGF-ß2                    TGF-ß2                    TGF-ß2                             

 

 

 

  P value^†     0.047                                                       0.065                       0.116                     0.041

 

*Mean wound surface area reduction in percentages; wounds completely closed are scored as 100%.

^†P value of treatment group mean vs placebo sponge mean, from a non-parametric comparison.

 

Table 4. Effect of TGF-ß2 on Time to Wound Closure

 

 

           Standardized                                            0.05 µg/cm2              0.5 µg/cm2            5.0 µg/cm2

                        Care                   Placebo                    TGF-ß2                    TGF-ß2                    TGF-ß2                             

 

 

  P-value^†        0.009                                                       0.133                       0.085                     0.030

 

*Median time to wound closure (weeks)

^†P value from log-rank test of treatment group vs placebo sponge group.

NA=Value could not be computed.

 

 

 

Table 5. Effect of Baseline Covariates on Time to Wound Closure

 

 

  *†TGF-ß2 (0.05 µg/cm2)         0.696 (0.4293‡)     0.105

  †TGF-ß2 (0.5 µg/cm2)         0.814 (0.430‡)       0.059

  †TGF-ß2 (5.0 µg/cm2)         0.991 (0.427‡)       0.020

  †Standardized care   1.313 (0.452)                     0.004

  Baseline Ulcer Age (weeks) -0.006 (0.002‡)       0.003

  Baseline Ulcer Size (cm2)   -0.146 (0.055‡)       0.008

  Serum Albumin (g/dL)            0.735 (0.328‡)       0.025

 

*Cox regression model with stepwise selection was used to identify demographic and baseline covariates which, in addition to treatment, were predictors of time to complete wound closure. Age, sex, creatinine clearance, HgA_1c level, and smoking status were not significant covariates.

†Comparison with placebo sponge group.

‡Standard error.

 

 

 

Table 6. Durability of Wound Healing for Ulcers With Complete Closure at or Before 21 Weeks

 

 

           Standardized                                            0.05 µg/cm2              0.5 µg/cm2            5.0 µg/cm2

                        Care                   Placebo                    TGF-ß2                    TGF-ß2                    TGF-ß2                             

 

  Healed at 3 Months     11 (65%)                   5 (71%)                 16 (64%)                 17 (68%)               20 (74%)

 

  Not Healed at 3 Months       2 (12%)                   2 (29%)                   6 (24%)                   5 (20%)                 6 (22%)

 

  Not Evaluated     4 (24%)           0 (0%)                   3 (12%)                   3 (12%)                   1( 4%)

 

P=0.66; ß2 contingency test, df=8.

 

Table 7. Incidence of Adverse Events

 

 

           Standardized                                            0.05 µg/cm2              0.5 µg/cm2            5.0 µg/cm2

                        Care                   Placebo                    TGF-ß2                    TGF-ß2                    TGF-ß2                             

 

  Patients with one                          631                            77                            86                            68                          70

 

  Most common adverse events

 

 

 

 

 

 

    Pharyngitis        0                            14                            12                              7                         11