Diabetic foot ulcer healing is predictable with the Wound, Ischemia, and foot Infection (WIfI) classification system when used alone or with multivariable risk-adjustment analysis, according to.
The research was conducted by, of Johns Hopkins University, Baltimore, and her colleagues as a retrospective study using prospective database information from enrolled type 1 and type 2 medication-dependent diabetic patients presenting to the multidisciplinary diabetic limb preservation service at Johns Hopkins Hospital from June 2012 to July 2017. The cohort of 310 patients with diabetic foot ulcer (DFU) in the study had a median age of 59 years and was composed of 60.3% men, with 60.0% of patients being black.
Infectious disease, plastic surgery, and orthopedic foot and ankle consultations were provided as needed. Individuals with evidence of peripheral artery disease (PAD) were provided lower extremity revascularization as determined to be appropriate by the primary vascular surgeon.
The 709 presented DFUs were assessed by x-ray imaging and follow-up MRI as needed. Wounds were debrided to clean margins and antibiotic treatments were administered as appropriate. At each visit the primary team assessed and assigned each wound a WIfI classified stage of 1-4 according to the calculation based on previously accepted Society of Vascular Surgery definitions, with PAD considered separately in final multivariable model analysis.
The association between WIfI stage and wound characteristics and healing was tested by univariable analysis. Multivariable Cox proportional hazards models that included sociodemographic, comorbidity, and wound characteristics were subsequently created to test WIfI stage as an independent predictor for wound healing after adjusting for those variables. Differences between models were related to wound location.
Most of the treated wounds occurred on toes, with the least common wound location being the leg/ankle. Of the 709 treated wounds, 32.4% (n = 230) were WIfI stage 1, 19.9% (n = 141) were stage 2, 25.2% (n = 179) were stage 3, and 22.4% (n = 329) were stage 4.
Differences between the stages included larger increases in mean wound area size, wound depth, and mean time from wound onset to initial assessment as WIfI stages increased from 1 to 4.
Healed wounds were defined as “maintained complete epithelialization with the restoration of sustained functional and anatomic continuity for 6 weeks after complete healing.”
The researchers found that wound healing time significantly increased with increasing WIfI stage, with a mean wound healing time of 96.9 days for WIfI stage 1 wounds, increasing to 195.1 days for WIfI stage 4 wounds (P less than .001). The authors found a likelihood of 94.1% for stage 1 wounds to be healed at 1 year, decreasing to a low of 67.4% for stage 4 wounds (P less than .001).
In univariable and risk-adjusting multivariable analysis, WIfI stage had an independent negative association with wound healing. With inclusion of risk adjustment, the probability of wound healing at 1 year was significantly lowered for stage 4 wounds, compared with stage 1 wounds (hazard ratio, 0.44). The three most prominent independently associated factors associated with poorer wound healing results include concomitant PAD (HR, 0.73), increasing wound area (HR, 0.99 per 1 cm2 area increase), and longer time from wound onset to initial assessment (HR, 0.97 per month). The strongest predictors for poor wound healing were increasing wound area (z score, –3.14), WIfI stage 3 (z score, –3.11), and WIfI stage 4 (z score, –5.40).
In this expanded study of previous work, the authors stated that they were the first to provide validating evidence for use of the WIfI classification system in giving “wound healing prognoses regardless of patient risk factors, comorbidities, and wound location.” Their findings also demonstrated that this classification system has broader applications than its original purpose to provide prognostic information and risk expectations for major amputation for patients presenting with foot wounds, Dr. Hicks and her colleagues concluded.
The authors reported no conflicts of interest.
SOURCE: Hicks CW et al. .