Burn index, burn characteristics and carboxyhemoglobin levels in indoor fire-related deaths: Significance and interpretation of the autopsy findings

Abstract

IntroductionThe Burn Index (BI) is a significant clinical prognostic parameter for patients with burns. It simultaneously considers major mortality risk factors: age and burns extensivity. Despite the inability to distinguish between ante- and post-mortem burns, their characteristics on autopsy might indicate if a significant thermal injury occurred before the onset of death. We investigated whether autopsy BI, burn extensivity, and severity could tell whether burns were the concurrent cause of fire-related death (FRD), even if the body remained in a fire. Material and methodsTen-year retrospective study analyzed FRD that occurred at the scene in a confined space. Soot aspiration was the main inclusion criterion. Autopsy reports were reviewed for demographic data, burn characteristics (degree, Total Body Surface Area burned- TBSA), coronary artery disease, and blood ethanol. We calculated the BI as a sum of the victim's age and percentage of TBSA affected by 2nd, 3rd and 4th-degree burns. Cases were divided into two groups: those with COHb≤ 30% and with COHb> 30%. Subjects with burned TBSA≤ 40% were analyzed separately afterward. ResultsThe study included 53 males (71.6%) and 21 females (28.4%). No significant difference in age was observed between groups (p > 0.05). COHb≤ 30% had 33, and COHb> 30% had 41 victims. BI and burns extensivity (TBSA) had significant negative correlation with COHb values (ρ = −0.581, p < 0.01 and ρ = −0.439, p < 0.01, respectively). Both were significantly higher in subjects with COHb≤ 30% compared to those with COHb> 30% (140.7 ± 29.57 vs. 95.49 ± 38.49, p < 0.01 and 98 (13−100) vs. 30 (0−100), p < 0.01, BI and TBSA respectively). BI had excellent and TBSA fair performance for detection of subjects with COHb≤ 30% on ROC curve analysis (AUCs 0.821, p < 0.001 and 0.765, p < 0.001), with optimal cut-off values: BI≥ 107 (sensitivity 81.3%, specificity 70.7%) and TBSA≥ 45 (sensitivity 84.8%, specificity 70.7%). On logistic regression analysis BI≥ 107 was independently associated with COHb≤ 30% values (aOR 6; 95%CI 1.55–23.37). The same holds for the presence of 3rd-degree burns (aOR 5.9; 95%CI 1.45–23.99). In the subgroup of subjects with TBSA≤ 40% burned, those with COHb≤ 50% were significantly older than victims with COHb> 50% (p < 0.05). Here BI≥ 85 was a particularly good predictor for detection of subjects with COHb≤ 50% (AUC=0.913, p < 0.001, 95% CI 0.813–1.00; sensitivity 90.9%, specificity 81%). ConclusionThe BI≥ 107, TBSA≥ 45% burned, and 3rd-degree burns observed on autopsy point to a significantly higher odds that limited CO intoxication occurred, and burns should be considered a concurrent cause of indoor FRD. When less than 40% of TBSA was affected, BI≥ 85 indicated sub-lethal CO poisoning.