Participant and wound characteristics of Photo Cohort
Seventeen participants were assessed for eligibility, 16 participants enrolled, and 13 had data available for analysis (2 participants dropped out due to technical difficulties, 1 was excluded because of subsequent enrollment in a clinical trial). The mean age was 26 years (range 3-56 years), 84.6% of participants were White, all had genetic confirmation of RDEB (two COL7A1 mutations), and 76% had the Generalized-Severe or Intermediate Subtype (Table 1).
Adult RDEB participants and/or their caretakers photographed up to 6 target wounds each week for up to 6 months (Figure 1). A total of 734 photos were uploaded for 69 wounds.
At baseline, participants were asked about the duration or age of an open wound. If a wound was reported to be constantly open (never healed) for ³12 weeks, then we classified the target wound as the “chronic open type” at baseline and used subsequent mobile app wound photos to monitor for spontaneous wound closure (defined as an open area < 0.5cm2). If a wound was reported to have been open for <12 weeks duration, we used photos to the 12-week cut-off to determine whether the target wound was the “chronic open” or “recurrent” wound type (Figure 2: Classification Period). After the 12-week classification period, target wounds were observed weekly for spontaneous wound closure for an average of 14 weeks (SD=12), with an average of 7 days between photo uploads (Table 1). In this Photo Cohort, 60.9% (42 out of 69) were classified as chronic open wound type and 27.5% (19 out of 69) as recurrent wounds (Figure 2).
Compared with recurrent wounds, chronic open wounds were larger at baseline (50 vs 5cm2) and present for a longer duration (21.4 vs 2 weeks) (Table 2). Importantly, only 17% of chronic open wounds achieved spontaneous wound closure during the observation period. In contrast, 100% of recurrent wounds achieved spontaneous wound closure at one point during the observation period (100% vs 17%, P<0.0001). The median time to wound closure was 5.7 weeks for recurrent wounds. Of the few chronic open wounds that were observed to have spontaneous wound closure (7 out of 42), the median time for wound closure was close to 6 months (25.7 weeks vs. 5.7 weeks for recurrent wounds, P<0.0001) (Table 2).
Correlation between wound types and pain and itch in the Photo Cohort
At baseline, chronic open wounds were reported to be more painful compared with recurrent wounds (4 vs 3, SD= 2, P<0.001). There was no difference in wound itch between chronic open and recurrent wound types at baseline. During the observation period, participants were asked about wound pain and wound itch at each time photo upload. The change in size of chronic open wounds during follow up correlated with both wound pain (P<0.001) and itch (P<0.001). In contrast, the change in size of recurrent wounds was not significantly associated with changes in wound pain or itch in regression models.
Clinical predictors of spontaneous wound closure in the Photo Cohort
We used generalized estimating equation models to estimate the odds of spontaneous wound closure during observation period. These regression models used clustering to account for the multiple wounds within the same patient (69 wounds from 13 participants). The only clinical predictor associated with spontaneous wound closure was baseline wound size. A smaller baseline wound area was significantly associated with a 12.8-fold greater likelihood of spontaneous closure (12.8, 95% CI: 3.3-48, P<0.01), adjusting for participant age, wound location, and recent wound infection in the past 3 months. Wounds without recent infection had a 3.8-fold (95%CI: 0.91-15.9) greater likelihood of spontaneous closure; however, this did not reach statistical significance. Patient-level covariates (age, sex, genotype, BMI, hemoglobin, and baseline vitamin D levels shown in Table 1) were not significantly associated with spontaneous wound closure in univariate and bivariate analysis adjusting for baseline wound size (data not shown).
Validation of differences in chronic open vs recurrent wound types in a separate cohort
We then confirmed the differences in spontaneous wound closure between chronic open vs recurrent wound types using a separate dataset (Validation Cohort).10–12 Compared to the Photo Cohort, the Validation Cohort participants (N=57) had a clinical diagnosis of RDEB, had wounds in similar anatomic locations, but were slightly younger (84.2%, <18 years of age) (Supplemental Table 1). Only 1 target wound from each participant was photographed and measured each month for up to 3 months. We classified wounds as “recurrent type” or “chronic open type” based on participant self-report of wound age at baseline. Forty-two target wounds were classified as chronic open wounds (73.7%), and these were similarly larger at baseline than recurrent wounds (20 vs 14 cm2). Fewer chronic open wounds achieved spontaneous closure compared with than recurrent wounds (26% vs. 86%, P<0.0001) (Table 2). Of the chronic wounds that were observed to have spontaneous closure, the median time for wound closure was 14.6 weeks for chronic open wounds compared to 8 weeks for recurrent wounds (Table 2, Figure 3). In multivariate models, recurrent wound type was associated with a 29-fold higher-odds of spontaneous closure compared with chronic open wound type, adjusting for participant age, wound location, and baseline wound size (P=0.0045).