Intravascular gas distribution in the upper abdomen of non-traumatic in-hospital death cases on postmortem computed tomography

Abstract

Objectives To investigate the occurrence of intravascular gas in the liver, kidneys, spleen, and pancreas by postmortem computed tomography (PMCT) in cases of non-traumatic in-hospital deaths and elucidate the relationship between the PMCT data and clinical information or autopsy results. Methods The study included 45 cadavers of patients who died while receiving treatment in our academic tertiary-care hospital between April and December 2009. All subjects underwent PMCT and conventional autopsy. The appearance of postmortem gas in the liver, kidney, spleen, and pancreas was assessed using PMCT and scored using a subjective scale (liver, L0–L3; kidney, K0–K2; spleen, S0–S1; and pancreas, P0–P1), and the distribution of gases in the vessels of the liver (arteries, veins, and portal veins) was analyzed. The relationship between the gas score and time elapsed since death, cardiopulmonary resuscitation (CPR), administration of antibiotics, a history of bacteremia, or cause of death was assessed statistically. Results Positive correlations were found between administration of CPR and liver and kidney gas scores ( P = 0.008 and 0.002, respectively), but not with spleen and pancreas gas ( P = 0.291 and 0.535, respectively). No significant relationship between distribution of gas in the vessels of the liver and CPR was found. No other significant correlations between gas and any of the other parameters described above were found. While significant correlations were detected in no-CPR cases between liver gas, kidney gas, spleen gas, and pancreas gas ( P < 0.001 for all six combinations), no correlation between these parameters was detected in the CPR cases. Conclusions The present study was the first statistical analysis of intravascular gas in the liver, kidneys, spleen, and pancreas by using PMCT in non-traumatic in-hospital death cases. The results showed that PMCT in the presence and absence of CPR reveals differences in intraorgan gas distribution. In addition, the detection of intraorgan gas on PMCT cannot be used to predict time elapsed since death, and it is not affected by the administration of antibiotics, a history of bacteremia, and cause of death. Awareness of these postmortem changes is important for the accurate interpretation of PMCT results.