Significance of intracranial gas on post-mortem computed tomography in traumatic cases in the context of medico-legal opinions

Traditional autopsy is the primary method used by forensic pathologists for examination of a decedent, but it is prone to certain limitations and occasionally fails to reveal important details [1]. With the rapid advancement of postmortem computed tomography (PMCT), forensic pathologists have gained a powerful new method to complement traditional autopsy [2]. Regardless of the autopsy technique used and the experience of the person conducting the autopsy, it is practically impossible to assess the accumulation of gas in certain parts of the body – including, but not limited to, the skull. Accumulated gas can be reliably detected only via imaging studies [3, 4]. For the above reasons, medico-legal opinions based solely on traditional autopsies disregard the possible presence of intracranial gas (ICG). However, with the popularization of PMCT, ICG have been increasingly recognized. The caveat of adding this finding to a forensic opinion is that the significance of this pathology in the medico-legal context must be known. According to our knowledge, there have been no publications so far that have dealt with this specific issue. Numerous clinical studies on detecting ICG in living patients indicate that this finding may be significant in determining the mechanism and circumstances of death [5‐15]. Due to the fact that ICG is mainly observed in cases of traumatic deaths [16], the goal of this study was to determine the frequency and importance of ICG in this group of cadavers in the context of medico-legal opinions.

The study was conducted retrospectively on PMCT scans acquired in the period between Nov 2014 and Nov 2016. PMCT scans were obtained with a 16-row Astelion CT scanner (Toshiba). In each case unenhanced CT scans were performed, with 1 mm thick slices acquired at 120 V with automatic exposure control (AEC). The pitch factor was 1.438 for the trunk and 0.688 for the head. Cadavers were scanned in a supine position, with the use of the standard protocol, including acquiring scans of the head and neck, the torso, and the lower limbs (if needed). The examination was performed without opening the protective plastic bag or altering the position of the body inside. The PMCT scans were analyzed using OsiriX application (OsiriX MD v.0.8.1, Pixmeo SARL- 266 Rue Bernex, Switzerland). The assessment was conducted independently from the traditional autopsy. Each scan was assessed by a board-certified forensic pathologist with experience in forensic radiology. Ambiguous images were consulted with a board-certified radiologist. The results were reported in a standardized questionnaire.

Initially, a total of 492 scans were evaluated. Ultimately, 159 cases of traumatic deaths were included in the study, after exclusion of cases with evident opening of the skull (excerebrations, open fractures, neurosurgical operations) and cases with extensive putrefaction. The degree of putrefactive development was assessed on the basis of changes described in the autopsy reports, with each pertinent autopsy conducted at least one day after the PMCT examination. Only cases with no, or slight, signs of decomposition, such as green discoloration of the lower abdomen, were included in the study. The mean age of the deceased was 50.4 years (0.01–90 years) and 69.8% (n = 111) of all cases were male. The largest number of cases were victims of traffic accidents 52.8% (n = 84), followed by fall from height 26.41% (n = 42), blunt injury (low force) 11.95% (n = 19), stab or cut wounds 7.54% (n = 12), and gunshot wounds 1.25% (n = 2). The primary data was collected with regard to the presence of ICG, which was categorized into two groups based on its location: PNC, defined as gas present between the cerebral surface and the inner table of the cranium (regardless of the gas volume), and IVG, defined as the gas present inside cerebral arteries or venous sinuses (Fig. 1). There were 22 cases in which both forms of gas collection were present simultaneously. Additionally, PNC was assessed semi-quantitatively based on cross-sectional images of the head in the PMCT and described as small (PNC I – gas visible in the form of a narrow band in the frontal part of the cranial cavity), moderate (PNC II – gas visible in the form of a distinct band covering the anterolateral parts of the cranial cavity) or large (PNC III – gas-filled space occupies much of the anterior part of the cranial cavity and extends to the back of the skull) (Fig. 2). Selective analyses were also performed, with the exclusion of cases in which PMCT results revealed the presence of intrahepatic gas, as in several publications this was an indication of an ongoing putrefactive processes. A total of 93.8% of evaluated cadavers underwent a traditional autopsy performed by an experienced forensic pathologist; in these cases, pertinent data was also obtained from the autopsy reports. The analyzed data included the subject’s age, sex, mechanism of death, cause of death, time from death to PMCT, time from trauma to death (whether or not death took place in a hospital), the time from trauma to PMCT, body position at the time of injury (vertical or not), the presence of gas in other body spaces and organs, and certain organ and bone injuries (selected with regard to the possible impact on the occurrence of ICG.

TIBCO Software Inc. (2017) Statistica (data analysis software system, version 13, Statsoft Polska, ul. Kraszewskiego 36, Kraków) was used for statistical analysis. The chi-square test and Fischer’s exact test were used to compare groups of categorical variables. The influence of continuous data on study groups was assessed using a two-sample t-test for parametric variables and the Mann-Whitney U test for non-parametric variables. The chi-square or Kolmogorov-Smirnov tests were used in order to determine the distribution of data. The phi coefficient (Φ) was used to determine the correlation between data. The results were considered statistically significant when the adjusted p values were less than 0.05 (p < 0.05).

There were no statistically significant differences between the subgroups characterized by a given extent of PNC with respect to age, time from injury to death, or time from injury to PMCT, even though the time varied in the latter case. The median time between injury and PMCT stratified by PNC size was 4 days for PNC I, 2 days for PNC II, and 1.5 days for PNC III. Only 7 people with PNC died in hospital, all of whom had PNC I or PNC II. There was one case of in-hospital survival of over 24 h (PNC II), in the remaining cases, regardless of the extent of PNC, survival time was limited to a few hours after trauma. The incidence of gas in specific body spaces and organs is presented in Table 4; the frequency of injuries to internal organs and bones in relation to PNC size is presented in Table 5. There was no statistically significant correlation between PNC size and the incidence of gas in extracranial areas of the body. Nevertheless, a direct positive relationship was observed between the size of PNC and the incidence of gas in the liver – with 50% of PNC I cases, 65.71% of PNC II cases, and 80% of PNC III cases showing gas in the liver. No such relationship was observed in regard to other places in the body which were examined for gas accumulation. Only one statistically significant relationship was found in regard to organ and bone injuries – the relationship between PNC grades and brain injuries (p = 0.027), with a correlation coefficient Φ = 0.35. Cerebral injuries were observed in all cases of PNC III. Increasing grades of PNC were also found to coexist with increasing numbers of the following injuries: cranial vault fractures, skull bases fractures, and depressed fractures of the skull. All cases of PNC III were accompanied by a fracture of the cranial vault (similarly to the case of brain injuries). Despite this, this relationship was not statistically significant.

Computed tomography (CT) is an excellent method for diagnosing PNC, as it can help detect as little as 0.5 mL of gas, due to its characteristic very low radiodensity (approx. -1000 HU) [17]. In clinical practice PNC is a commonly diagnosed phenomenon associated with surgical interventions to the skull, spine, or paranasal sinuses. It is especially common as a side effect of operations on the base of the skull [10]. In a study in 131 patients undergoing supratentorial craniotomy, immediately postoperative CT showed PNC in every case [18]. The time of postoperative PNC resolution is not well defined; PNC can still persist up to 3 weeks after surgery [18]. There are also cases of gas emboli in vessels, including cerebral vessels, caused by intravenous puncture [7, 13, 19]. In addition, ICG in the living is observed in barotrauma [19], certain infections, and oncological diseases [20]. However, the main cause of pre-mortem ICG in non-hospital settings is trauma. However, in clinical practice post-traumatic ICG is considered a rare complication, with its reported rates ranging from 0.5% [20] to 9.7% [12] of injury cases (mostly head injuries). There are also published cases of pre-mortem, post-traumatic ICG not associated with heady injury, such as ICG following stab wounds [14]. In medico-legal literature, ICG is only discussed in the context of the mechanism of death by gas embolism which is further accompanied by gas embolism in the cerebral vessels, such as cases of decompression in divers [21], administration of intravenous oxygen [22], or helium intoxication. [23]. Most of the relevant publications mention ICG as one of the pathologies detectable only since the introduction of PMCT [1‐4]. The presence of gas in the skull is also discussed together with other gas locations in PMCT results [16]. The phenomenon of gas visualized with PMCT, mainly in non-traumatic cases [24, 25], has been presented in the literature in the context of attempting to differentiate pre-mortem from post-mortem (i.e. related to decomposition) gas accumulations. The results of these studies are not clear-cut. Some authors emphasize the fact that the majority of post-mortem gas is the result of decomposition and should be treated as an artifact [26], even if there is no evidence of ongoing putrefaction [27]. Other authors highlight the fact that their studies did not confirm the decompositional origin of the gas visible in PMCT [28], especially when the CT scan was performed in the first 24 h after death. Due to the difficulties of definitively determining the origin of gas in decedents [29], the conclusions of these studies do not provide a base for using this PMCT-gathered data for creating medico-legal opinions. Our research, covering a large group of people who died in a traumatic mechanism, showed a much more frequent occurrence of ICG in this group of decedents in relation to the frequency of ICG observed in living people following injury. This phenomenon can be explained by the fact that most people with this complication were shown to have died at the scene of the event [30]. There was also a strong correlation between PNC and chest injuries, especially lung injuries. The mechanism in which damaged lungs can cause the entry of gas into the venous and arterial vessels of the brain has been explained in the literature as due to the rupture of both the alveoli and the surrounding small vessels in cases of blunt chest injuries [24, 31] and damage to larger blood vessels of the chest in cases of penetrating injuries [14]. Our study’s most important finding was that most people presenting with ICG died at the site where they suffered their injuries, and even if their vital functions were maintained long enough for hospitalization, death occurred within 24 h. Our study did not draw a clear-cut conclusion in regard to the mechanism in which the presence of ICG affects the patient’s deterioration. However, several hypotheses can be considered. One of them is the possibility that ICG causes an increase in intracranial pressure in the same way that intracranial hematomas do [6, 12, 19, 32]. The patient’s deterioration may also result from gas embolism, which is present in a significant number of PNC cases and affects both intracranial vessels [12, 33] and the vessels of the neck and heart. The results of our research do not support the statement that the chance of the observed ICG being due to decomposition is high enough to reduce the significance of the finding itself. In fact, evidence of early putrefaction was found more often in the non-ICG group than the ICG one. The median time from trauma to PMCT decreased with the observation of larger PNCs. Additional calculations were performed after excluding cases with the presence of gas in the liver, which, according to the authors of some publications, indicates an ongoing process of decomposition, despite the lack of visible signs on the outside of the body [34‐37]. Nonetheless, this did not significantly change the relationships found in the entire study group, which indicates that the presence of gas in the liver in the investigated cases may have been a consequence of the injury itself and not of putrefactive processes. This can be confirmed in part by certain publications [16, 38, 39]. The relative increase in the rates of aortic damage in the ICG group after the exclusion of cases with intrahepatic gas may be explained by gas not being able to move to the liver due to sudden interruption of blood circulation. Our study indicates that PMCT-based evidence of ICG in traumatic cases with no open skull fractures and no evident signs of decomposition (either on autopsy or PMCT) should be mentioned in medico-legal opinions as one of the significant consequences of trauma that can severely worsen prognosis. This element of prognosis may be particularly important in opinions regarding the withholding of assistance, such as those involving a driver causing a traffic accident and its victim. In such situations, the medical forensic expert usually has to answer the question whether the nature of injuries indicates that the risk of death would have been significantly lower if assistance had been given within a short time after the injury. Our study failed to prove a relationship between ICG and the body position due to difficulties in obtaining this data. Further research in this area is necessary.

It was impossible to examine the chemical composition of gas observed in PMCT in the center where the research was conducted. Furthermore, due to an insufficient amount of relevant data, we were unable to reliably analyze the relationship between ICG and the possible resuscitation efforts or ICG and the victim’s body position at the time of injury.