Wound‐healing trajectories as outcome measures of venous stasis ulcer treatment

Abstract

Outcome measures of venous ulcer healing are not uniformly accepted. Stringent criteria of 100% closure fail to provide information of healing over the entire span of repair. Wound‐healing trajectories (plot of percentage of wound closure versus time of wound treatment) were constructed for 232 patients treated in eight clinical trials at two independent wound care/research centres. Trajectories were constructed for ulcers that totally healed (100% closure) and those that did not (<100% closure) over a 20‐week period. Kaplan–Meier survival plots determined the percentage of patients achieving total healing versus time of treatment. The wound‐healing trajectories were almost identical for patients achieving complete ulcer healing, as were the trajectories for patients with less than 100% closure. The trajectories for the ulcers healing completely were significantly different from those with <100% closure. Only 60% of all patients achieved 100% closure by 20 weeks. Using linear regression, it was predicted that it would take 31 weeks for all patients to achieve total healing. Total healing is an inadequate outcome measure for healing of venous stasis ulcers. Clinical trials using this measure would require excessive time periods. As wound‐healing trajectories for patients treated at two centres mimic one another, shifting the wound‐healing trajectory from one of impaired healing to one of a more ideal healing course may be considered a better outcome measure for determining healing of venous stasis ulcers.

Introduction

Chronic cutaneous ulceration of the lower extremity can result in an indolent non healing wound (1). Although the causes of such ulcerations can be varied and often multifactorial, the most common cause (70–90%) is venous insufficiency 1, 2, 3). Venous ulcers have a prevalence of approximately 1% in the United States (approximately 2 million persons), with approximately 50% of the patients having an ulcer history of 10 years (2, 4). Despite standard therapy with compression dressings, only 50–60% of venous stasis ulcers heal completely in 6 months, and the recurrence rate is as high as 70% (5, 6). The direct and indirect costs associated with the treatment of these ulcers are considerable (7). Patients affected with venous ulcers experience a decrease in their overall health‐related quality of life, including severe pain, impaired mobility, limited work capacity and negative emotions (8, 9).

Treatment of venous stasis ulcers varies from multiple surgical procedures to the use of various topical agents combined with external pressure support to the legs (1). In the United States, the standard compression dressing is either the Unna boot, a zinc‐oxide‐impregnated inelastic dressing, the Duke boot with an added elastic bandage or the multiple‐ply compression dressing (5, 10). Even with good compression therapy, approximately 40–50% of ulcers do not achieve complete closure within 6 months of therapy (5, 9). Because of this low healing rate and a recidivism rate of nearly 70%, multiple wound‐healing agents, cytokine growth factors and cell‐based devices have been used to determine if closure of venous ulcers can be facilitated without operative intervention (1, 9, 11, 12, 13).

The cumulative goal of much wound‐healing research is to discover products and processes that can accelerate wound healing in humans (14). One problem in clinical trials to date attempting to discover such therapeutic agents has been trial methodology including determination of outcome endpoints. Investigators have not routinely used statistically robust analytical methods that directly, unambiguously and unequivocally demonstrate that the agent under evaluation actually promotes wound healing (15). Different studies have suggested various outcome measures for healing of open chronic wounds such as venous stasis ulcers [e.g. incidence of complete wound closure (16, 17), percentage of initial wound remaining (18, 19), wound perimeter remaining (18, 20), various definitions of wound‐healing velocities 21, 22, 23, 24) and wound‐healing trajectories (25)].

The United States Food and Drug Administration (USFDA) and a committee of the Wound Healing Society have suggested the most stringent outcome criteria to determine efficacy for a new wound‐healing agent, that of total (100%) wound closure (25). Stromberg et al. (16) of the USFDA have stated, ‘With respect to outcome measurements, complete rather than partial closure of wounds provides the best objective evidence of clinical benefit and therefore is the preferred outcome endpoint’. They did state that in certain indications, outcome endpoints of a clinical trial other than complete (100%) closure may be considered as long as they are validated measurements of patient benefit. In a follow‐up publication, the FDA Wound Healing Clinical Focus Group stated that measurement of partial healing, even when prospectively defined, cannot be used as primary evidence of clinical efficacy (17).

The problem with using complete healing (100% closure) as an outcome for therapeutic efficacy of chronic wounds is that Falanga (5) has reported that only 50–60% of venous ulcers heal in a 6‐month period. Margolis et al. (26) have been able to improve that figure to 81% in 6 months. Robson et al. using a linear regression curve of Kaplan–Meier healing curves on 160 patients with diabetic foot ulcers demonstrated that it would take 9 months for all patients to reach 100% wound closure (25). Therefore, if 100% wound closure is the only outcome endpoint to determine efficacy, it would require a clinical trial of many months duration to determine efficacy of a new wound treatment. The practical difficulty and expense of conducting a closely monitored wound‐healing trial of such a duration is immense (25).

Hokanson et al. (15) proposed an alternative to complete closure when evaluating a wound‐healing agent. They proposed an analytical strategy that combined a mathematical model to approximate the actual wound closure measurements and an endpoint analysis similar to that used in failure‐time studies. They then showed that the analysis worked when applied to large groups of experimental wound‐healing studies (15, 25, 27, 28). The analysis uses the wound‐healing trajectory (plot of percentage of wound closure versus time of wound treatment) to evaluate the effect of treatment. Once the wound‐healing trajectory is plotted, the time required for the percentage healing to reach an arbitrary fraction of 100% closure can be determined. The distributions of these fractional closure times and statistical differences in these distributions can then be determined using statistical methods similar to those used for failure‐time or survival analyses (15, 29). Our group previously reported the usefulness of wound‐healing trajectories as predictors of effectiveness of therapeutic agents for diabetic foot ulcers (25). The purpose of this report is to demonstrate the usefulness of the wound‐healing trajectory as an outcome measure for venous stasis ulcer treatment.

Methods

Over a 10‐year period, two multidisciplinary wound research teams, one based at the University of Pittsburgh, Pittsburgh, PA and one based at the University of South Florida/Bay Pines Department of Veterans Affairs Medical Center, Bay Pines, FL, participated in eight venous ulcer clinical trials. A total of 232 patients were enrolled in the trials. All patients, regardless of the particular clinical trial shared several basic requirements. To be enrolled in the trials, each patient must be diagnosed as having a venous stasis ulcer with the following criteria: (i) the leg ulcer must be located between the malleoli distally and the mid‐calf proximally; (ii) there must be a degree of lipodermatosclerosis; (iii) distal palpable pulses must be present with an ankle‐brachial index of ≥0·8 and (iv) there must be evidence of a venous aetiology for the ulcer by venous duplex Doppler ultrasound scanning/imaging or a similar validated methodology. All patients shared three standard treatment regimens: (i) all had debridement of all necrotic tissue in and/or surrounding the ulcer prior to study entry and as needed during the trial; (ii) all were treated with extremity compression dressings and (iii) all had frequent periodic evaluation by an experienced wound care/research team.

Because it was not our intent to focus on the analysis of data from any particular clinical trial, the minimal description of experimental data was sequentially measured venous stasis ulcers in patients treated with various therapeutic wound‐healing agents, their respective vehicle controls or standardised care (25). At weekly intervals, the wound perimeters were traced on clear plastic sheets and the wound areas calculated planimetrically. The ulcers were measured for a defined period determined by the parameters of the particular clinical trial, most commonly 20 weeks (140 days).

As total ulcer healing (100% wound closure) was the efficacy standard used in the majority of the clinical trials, wound‐healing trajectories (plots of percentage of wound closure versus time of wound treatment) were constructed for patients who achieved total healing (100% closure) and those who did not (<100% closure) over a 20‐week period at the two centres. Percentage healing refers to the diminishment of wound area relative to day zero as follows: [area (day 0) − area (day x)/area (day 0)] × 100 (25). In addition, the percentage of patients who achieved total healing was plotted versus time of wound treatment (days) for both centres to determine how patients with venous stasis ulcers receiving at least debridement, compression dressings and frequent monitoring heal their ulcers.

Statistical methodology

Survival analyses were done using the Kaplan–Meier method, with both log‐rank and Wilcoxon statistics to test between groups (15, 30, 31). Calculations and graphs were made using jmp software (SAS Institute Inc. Cary, NC, USA) and Sigma Plot Software (SPSS Inc., Chicago, IL, USA).

Results

A total of 232 patients with venous stasis ulcers met inclusion/exclusion criteria and were entered into institutional review board‐approved prospectively randomised clinical trials. Of all the patients entered into these trials at the two centres, 60% achieved total ulcer healing (100% closure) within 20 weeks.

The wound‐healing trajectories were almost identical for patients achieving complete healing at the two centres (Figure 1). This was true whether one evaluated the mean or the median percentage healing versus time of treatment. The trajectories for patients with less than 100% closure over the period of the trials were also remarkably similar. However, the trajectories of the patients achieving total ulcer healing were markedly different from those not achieving 100% closure (Figure 1).

Figure 1.

Curves depicting (A) mean and (B) median values for percentage of initial wound, by study centre. Circles indicate patients who achieved total ulcer healing (100% closure); squares indicate patients who did not achieve complete healing. Because of differences in clinical trial designs, data for all 232 patients were not available at all time points.

Survival analysis of patients entering the eight clinical trials at the two centres who achieved total ulcer healing each week is shown in Figure 2. This was broken down for each individual centre, and those analyses revealed that 60% of all patients at the two centres entered into a prospectively randomised clinical trial and receiving the standard care of debridement, compression dressings, and regular evaluations healed their wounds 100% by 20 weeks. There was no statistical difference between the two centres when the survival analyses were compared (log‐rank P = 0·54; Wilcoxon P = 0·52). Almost no total healing occurred until 21 days (3 weeks). After 21 days, a linear relationship occurred, with 5% more patients achieving 100% closure per week (20% per month).

Figure 2.

Kaplan–Meier healing curves depicting the percentage of patients achieving 100% wound closure (A) by study centre and (B) at both sites combined. Test between centres in (A) log‐rank P = 0·54; Wilcoxon P = 0·52. Derived coefficients for linear regression curve in (B) y = 0·49 × −7·37, r² = 0·98.

Although none of the eight trials of venous stasis ulcers conducted at the two centres were of greater than 20‐week duration, a linear regression analysis was performed to predict when all patients would achieve 100% closure. Based on the linear relationship demonstrated in Figure 2, this would occur at 219 days (31 weeks).

Discussion

Outcome measures or clinical endpoints are necessary both to evaluate efficacy of potential new treatments undergoing investigation, as well as to determine effectiveness of currently used therapies. Presently, the only outcome acceptable for new treatments for venous stasis ulcers is total ulcer healing (100% closure) (17). A recent review of total healing rates for venous ulcers reported that only 29–81% of ulcers achieved 100% closure in 24 weeks (32). Our data showed that only 60% of 232 patients achieved total ulcer healing in 20 weeks (Figure 2). The data suggest that all patients receiving a level of care of at least debridement of necrotic tissue, adequate extremity compression and regular evaluations will achieve 100% wound closure by 8 months. Therefore, if 100% wound closure is the best efficacy outcome measure, a clinical trial of approximately 8 months would allow the best determination of this efficacy endpoint. This would be impractical and prohibitively expensive (25). Using this outcome requirement, only one cytokine growth factor and two cell‐based devices have been approved by the USFDA for treatment of chronic dermal ulcers. The tremendous costs involved to obtain approval coupled with slow cost recovery due to reimbursors using the same total healing standard make it difficult for small wound care companies to survive development of novel therapies (33). Two sponsors of recently approved cell‐based devices have asked the US Bankruptcy Court for approval to reorganise or restructure (33).

Wound‐healing trajectories as used in this study provide more information about the entire continuum of the wound‐healing process (25). Statistical analyses such as a t‐test performed on data from a single point (e.g. 100% closure) may not provide accurate guidance about the actual effectiveness of new wound care agents over the entire span of the healing process (15). All wounds normalised to initial size begin at 0% healing, and all wounds that heal during the defined period of the trial have 100% healing. Statistical comparisons at these extremes may be non informative and stress the need for statistical techniques that allow dynamic comparisons at biologically meaningful interim values (15, 25). This point is in agreement with Polansky and van Rijswijk who have stated that healing time curves (wound‐healing trajectories) are a ‘moving picture’ of healing that provide more detail than the ‘snapshot approach’ in which only the proportion of patients healed (100% closed) at the end of the study is assessed (25, 31).

Other groups have proposed using wound‐healing curves as endpoints for venous stasis ulcer healing. Attempts to determine healing rates were difficult, because healing of wounds follows an exponential course, with the rate of change of wound area progressively decreasing as the residual wound area approaches total closure (25). Because inclinations of this exponential curve diminish with time, attempts have been made to develop a linear rate of healing. Gilman introduced the concept of using the wound perimeter combined with the wound area to neutralise the effect of varying size wounds in calculating the rate of healing (20). However, researchers using this manipulation to determine a rate of healing found a different initial healing rate and overall healing rate (22). Margolis et al. (22) calculated the initial healing rates over the first 4 weeks of therapy and reported that these rates correlated highly with overall ulcer healing rates. They proposed using the initial healing rates as a surrogate endpoint for clinical trials of different therapeutic agents. Tallman et al. (23) further manipulated the raw data to attempt to get a single healing rate by calculating a rate for each time interval measured and averaging the different intervals. This method was called a mean‐adjusted healing rate. Recently, Hill et al. (32) compared these three methods in a single group of venous stasis ulcer patients and concluded that although initial healing rates have ‘some general prognostic usefulness, their poor predictive performance seems to preclude their use as surrogate endpoints for clinical trials’.

Because the wound‐healing trajectories for venous stasis ulcers treated in eight clinical trials at two totally independent wound care/research centres so closely mimicked one another, it would appear that wound‐healing trajectories can be validated and may be useful to serve as efficacy outcome measures. This hypothesis is strengthened by the fact that the trajectories of patients achieving 100% closure in 20 weeks were so different from patients not achieving total ulcer healing in the same time period (Figure 1). Using the wound‐healing trajectories, the time required for each group of patients to reach an arbitrary percentage of healing, such as 50, 75 and 90%, can be determined (25). These distributions of times to an event (e.g. 75% wound closure) can then undergo standard statistical failure‐time analysis and be just as rigorous as total (100%) ulcer healing statistics (25, 29). Another advantage of the wound‐healing trajectory is that it allows use of raw data and does not require manipulations by introducing the perimeter (20, 21, 22) or a second step of using the perimeter and multiple small time intervals (23, 25).

The FDA Wound Healing Clinical Focus Group has stated that partial healing, per se, is not considered an acceptable wound‐healing claim, because the clinical benefit of statistically significant differences in wound size has not been established (17). One might consider use of a wound‐healing trajectory to determine time to 50% wound healing a measure of partial healing. However, there was a 14‐week difference between the two groups of patients' trajectories to reach 50% healing in as shown in Figure 1. Possibly, this outcome can be validated as an endpoint. In following patients from a four‐arm exogenous cytokine growth factor‐treated pressure ulcer clinical trial, Payne et al. (34) reported that patients who healed ≥85% during a 35‐day treatment period showed an 84·6% incidence of 100% healing at a 1‐year follow‐up examination (34). This compared with only 61% total healing at 1 year of those patients who healed less than 85% during the 35‐day active treatment period (P < 0·05). These data suggest that those on the steeper healing trajectory, although not 100% healed during the trial, did reach a clinically significant result that was sustained for up to a year post‐treatment.

Once wound‐healing trajectories are validated for venous stasis ulcers, the concept of shifting the trajectory of patients with less than 100% closure in a defined period as shown in Figure 1 to the trajectory of patients attaining total ulcer closure could be used to determine new therapeutic agent efficacy. Because the predictive value of such wound‐healing trajectories have been shown for diabetic foot ulcers and pressure ulcers, possibly shorter clinical trials can be constructed relying on specific shifts of the wound‐healing trajectories from impaired healing towards an ideal endpoint (25, 35, 36).