Trajectories to death in patients with burn injury

Abstract

Background: With unprecedented survival rates in modern burn care, there is increasing focus on optimizing long-term functional outcomes. However, 3% to 8% of patients admitted to burn centers still die of injury. Patterns in which these patients progress to death remain poorly characterized. We hypothesized that burn nonsurvivors will follow distinct temporal distributions and patterns of decline, parallel to the trimodality of deaths previously described for trauma.

Methods: We retrospectively identified all adult deaths from 1995 to 2007 in the National Burn Repository database (n = 5,975) and at our regional burn center (n = 237). We stratified patients by age and analyzed injury and death characteristics. We used objective criteria to allocate nonsurvivors to one of four trajectories: early rapid decline, early organ failure, late sudden death, or late-onset decline.

Results: The greatest concentration of deaths in both samples and age groups occurred within 72 hours of injury and decreased subsequently with no later mortality peak. Death was most often caused by burn shock within the first week of injury, cardiogenic shock or lung injury in Weeks 1 to 2, and sepsis/multiorgan failure after Week 2. In decreasing frequency, trajectories to death fit the pattern of early rapid decline (58%), early organ failure (20%), late-onset decline (16%), and late sudden death (6%).

Conclusion: Most burn deaths follow a pattern of early rapid decline or early organ failure manifested by death or critical illness within several days of the burn. These findings indicate that more than three quarters of burn deaths are attributable to failure or significant decompensation beginning in the resuscitation phase. Sporadic deaths later in hospitalization are uncommon. Despite significant advances in burn resuscitation, our data indicate that ongoing efforts to mitigate deaths in modern burn care should still focus on care improvements in the resuscitation phase.

Level of evidence: Epidemiologic study, level III.