The care of individuals with spinal cord injury (SCI) has been designated as a special area of interest for the Department of Veterans Affairs (DVA). Pressure ulcers (PrUs) are a frequent complication in Veterans with SCI due to denervation and immobility. Most Veterans with SCI will develop at least one serious PrU during their lifetime. The mean annual cost of a new PrU in SCI is very high (>$100,000). Wound size is the most consistent predictor of healing/non-healing of other chronic wounds (e.g., diabetic foot and venous leg ulcers) in the non-SCI population. However, there is limited evidence for PrU healing in persons with SCI. We found no published studies on the clinical correlates of PrU healing in SCI.
Our objective was to develop evidence-based parameters for PrU healing in Veterans with SCI using data collected in a double-blind placebo-controlled trial (VA CSP #535 "Anabolic steroid therapy on PrU healing in persons with SCI"). Previous literature showed that burn patients treated with Oxandrolone improved body composition, preserved muscle mass and reduced hospital stay time. However,CSP#535 patients with wounds treated with Oxandrolone did no better than controls. CSP#535 remains the largest known prospective, detailed database of Veterans with SCI and PrUs. Primary data collected for the study included PrU characteristics including serial wound measurements, nutritional status, metabolic values, clinical laboratory assessments, immune function, endogenous hormone levels and inflammatory markers.
We used Delphi survey techniques to query experts in SCI and PrUs using the VHA national woundcare listserv and other national experts. Participants were asked to rank CSP#535 variables associated with wound healing. In the second round survey, participants received aggregate feedback on the first round results. The panel's responses guided variable selection for subsequent regression analyses. The CSP began with a 4-week screening phase (n=775); non-healers (n=209) became eligible for the intervention (treatment). In our study, we only used data for subjects with PrU size >0 at baseline (start of screening).
The primary outcome of our analyses was wound healing, defined as "change in area measured in cm2." Outcome measures included: 1) complete [100%] healing (area = 0 cm2) and 2) "50% healing" (e.g., >50% decrease in area).
Wound healing was modeled at 4, 8, 12, 16, 20, and 24 weeks as a function of patient factors (age, race, education, smoking, alcohol use), SCI characteristics (etiology, injury level, ASIA score), wound characteristics (location, duration, tunneling, undermining, odor), and other clinical factors (prealbumin, glycated hemoglobin) based on the Delphi Panel results.
For 4 week outcomes, multivariable analyses included all screening participants; subsequent outcomes included only treatment phase participants. Due to small number of events, we used Poisson regression to analyze complete wound healing with center-level clustering corrected using Huber-White sandwich estimators. For 50% healing, we used generalized estimating equations (GEE) to estimate random-intercept logistic regression models for cluster correction.
The Delphi survey was sent to SCI and/or national PrU experts (n=451), with response rates of 41.0% and 27.3% respectively for the first and second rounds. Most (70%+) respondents were nurses, a third of which dealt primarily with SCI patients. On average, respondents had 13.5 years of wound care experience. Delphi participants identified patient, PrU and treatment characteristics associated with wound healing/non-healing. Highest rated PrU characteristics included PrU number/size, history, stage, location, tunneling, infection, pain and color. Top treatment characteristics associated with healing included: SCI center and use of adjunctive therapies (e.g., vacuum-assisted therapy, electrical stimulation). Highest rated patient characteristics associated with PrU healing included: smoking status, compliance with treatment recommendations, weight, level of injury/ASIA score, high or low BMI, race, and education. Across both rounds, the highest rated predictors included patient non-compliance, current smoking status, SCI level, and PrU duration/severity.
Of those who were screened, wound size >0 for 748 subjects and, of those entered treatment, 177 had wound size > 0. Most of those screened were white, with more than a high school education and a mean age of 57 (sd=11) years. Over 50% had complete paraplegia. PrUs were most frequently located on the ischium or perineum. The average PrU size at baseline was 15.4 (sd=19.3 cm2) for the screened sample and 16.8 (sd=17.9 cm2) for those treated. The mean PrU duration at baseline was 36.0 (sd=67.9 weeks) (median = 18 weeks). About half of the screening patients had PrUs with baseline tunneling (54%) and undermining (52%) and 4% achieved complete healing at the end of the 4-week screening phase.
A small number of screening patients achieved complete wound closure at 4 weeks (n=26/662, 4%) and 8 weeks (n=9/163, 5.5%). In contrast, 41% and 66% of the screening sample achieved 50% and 25% healing in 4 weeks. Of those who were treated, higher rates of complete healing were achieved by 8 (11.5%), 12 (19.5%), 16 (24.7%), and 20 (31%) weeks after randomization.
Multivariable analyses showed that undermining was significantly associated with worse healing at 8 weeks (incidence rate ratio [IRR] = 0.03; 95% CI, 0.01 - 0.08) and at 24 weeks (IRR = 0.52; 95% CI, 0.27 - 0.99). Baseline smoking status was associated with worse healing at 4 weeks (IRR = 0.28; 95% CI, 0.13 - 0.62) but not at 24 weeks (IRR = 0.98; 95% CI, 0.55 - 1.73). No other demographic, SCI, or PrU wound characteristics predicted 100% healing. For both 50% and 25% healing, ischial or perineal location were significantly associated with better healing, undermining with worse healing, and better nutrition (prealbumin 10 mg/dL) with better healing. By 8th week after the start of screening, 57 (sd=25%) (range 30 - 90%) of those who started the treatment phase achieved at least 50% healing across SCI centers. Although strongly endorsed by the Delphi participants, neither patient non-compliance nor SCI level predicted wound healing.
Directions for Future Research: The small sample size and low rates of complete healing limited our ability to make policy or clinical inferences for many of the variables we examined. Full implementation of the SCI-PUMT (Pressure Ulcer Monitoring Tool) is ongoing. These questions can be more fully examined using the larger sample that will be available once the SCI-PUMT is more fully implemented across the 24 VA SCI centers.
None at this time.
Brain and Spinal Cord Injuries and Disorders
Treatment - Efficacy/Effectiveness Clinical Trial