Wounds and survival in noncancer patients
BACKGROUND: Currently available prognostic models for noncancer patients lack high levels of discrimination. Therefore, the quest for additional prognostic factors must continue. To date, none have utilized the occurrence of wounds as aprognostic factor. METHODS: As a prospective observational study, based on a sequential case series of 189 advanced noncancer patients, all wounds were documented. One hundred seventy patients were followed until their deaths. Univariate and multivariate survival analyses were performed using hazard ratios(HRs) derived from Cox proportional hazard models. RESULTS: Seventy-eight percent of patients presented with at least one wound at referral. Patients with wounds displayed worse overall survival than those without wounds (p = 0.009). Survivalanalysis for the full postreferral period revealed a violation of the proportional hazards assumption for pressure ulcers and the Palliative Performance Scale version 2 (PPSv2). In order to address this finding, early deaths (within 14 days of referral) were assessed separately from later deaths (more than 14 days after referral). After controlling for the co-occurrence of wounds, age, gender, Charlson comorbidity index, and PPSv2, pressure ulcers wereassociated with statistically significant increased risk of death for patients of sufficient health to survive at least 2 weeks after referral (HR 2.42, p = 0.003), while other wounds were associated with greater levels of mortality overthe full postreferral period (HR 1.96, p = 0.0001). DISCUSSION: The occurrence of pressure ulcers and other wounds are correlated with reduced survival in patients with advanced noncancer illness. These data merit incorporation into existing prognostic models or used in conjunction with them to enhance prognostic accuracy.