Carbon monoxide and cyanide poisoning in fire related deaths in Victoria, Australia

Abstract

Objective: This study was undertaken to examine the association of hydrogen cyanide and carboxyhaemoglobin in victims of fire related deaths in Australia. The secondary aim was to document demographic data about Australian fire related deaths.

Methods: An observational retrospective study was undertaken of autopsy reports from the Victorian Institute of Forensic Medicine. Reports of fire related deaths were electronically searched using the terms burns, "smoke" or "fire" as a cause of death in the calender years 1992 to 1998. Data on the circumstances of the fire and results of toxicological screening were obtained on 178 persons. Additional whole blood cyanide levels were determined if blood samples were available in storage. Demographics of the victims were analysed, as well as the relationship between carboxyhaemoglobin and whole blood cyanide levels.

Results: Most (82%) of the victims died at the scene, whilst 32 victims died after a period of hospitalisation (hours to weeks). Suicide as a result of self-immolation was the reported cause of death in 32 cases. Most of the fires were in houses (114) and cars (29). The blood ethanol level was zero in 112 cases; the remaining cases (53) had a mean level of 0.17%. Other central nervous system (CNS) depressants were recorded in 49 of the 134 cases that received a complete toxicological screen. Carboxyhaemoglobin levels were measured in only 154 of 178 cases. The carboxyhaemoglobin level was zero in 43 cases. The remaining cases (111) had a mean level of 40%; with 44 cases having a level greater than 50%, a level considered to be potentially lethal. Whole blood hydrogen cyanide levels were measured in only 138 of 178 cases. The hydrogen cyanide level was zero in 52 cases. The remaining cases (86) had a mean level of 1.65 mg/L; with 11 cases having a level greater than 3.0 mg/L (potentially fatal). Blood ethanol levels were significantly correlated with both carboxyhaemoglobin (R = 0.22, P < 0.01) and cyanide (R = 0.36, P < 0.001). In addition, a significant correlation (r = 0.34) between carboxyhaemoglobin and hydrogen cyanide levels was noted.

Conclusions: This study showed a correlation between elevated blood ethanol and whole blood cyanide levels (r = 0.36, p < 0.001) and between elevated carboxyhaemoglobin and hydrogen cyanide levels (r = 0.34). Although the mean cyanide level was 1.3 mg/L (above the level some consider potentially toxic) in those cases with a carboxyhaemoglobin level of greater than 10%, there is insufficient data to permit recommendations for clinical care. Further studies are required on those victims that reach hospital alive.