OBJECTIVE To predict recurrence of pressure ulcers (PrUs) in a high-risk population of veterans with spinal cord injury (SCI). METHODS Cross-sectional observational design. METHODS A convenience sample of 64 subjects from 6 Department of Veterans Affairs (VA) SCI Centers who had been admitted to the hospital for the treatment of stage III-IV pelvic PrUs and were healed at the time of discharge back to the community. METHODS Primary outcome measures were pelvic PrU recurrence, defined as self-reported new skin breakdown (stage II or greater) in the pelvic area (not necessarily in the same location as previous ulcer) and time to recurrence. RESULTS There were no differences between those with/without recurrences with regard to age, age at/level of injury, number of previous ulcers or surgery, rate of or time, to recurrence. Mean age was 56 years; most were white and men, lived at home, and had some college education. Mean time since SCI was 22 years; 28% had tetraplegia; mean number of prior pressure ulcers was 3; and almost one half had a previous ulcer in the same location. The strongest predictor of recurrence in a multivariate logistic regression was African American race (odds ratio = 9.3). Additional predictors included higher scores on the Charlson Co-Morbidity Index (indicating a higher burden of illness), the Salzburg PrU Risk Assessment Scales, and longer sitting time at discharge. CONCLUSIONS Identifying individuals at highest risk for recurrence and developing effective prevention programs are essential rehabilitation goals. We recommend that the unique findings of this exploratory study be considered preliminary until replication of these results is published.