Postmortem bacterial culture is useful in forensic medicine for determining the cause of death as well as in clinical medicine for studying infectious diseases. However, there are many concerns about the use of postmortem bacterial culture. In addition, there are no specific guidelines, recommendations, or checklist items that address performance standards for such tests [1]. Theoretically, a positive test could be a result of true positive sample, agonal spread, postmortem transmigration, or contamination [2]. Agonal spread is not a cause of death: during the process of death or when circulation is artificially maintained, the integrity of the mucosal surface may be compromised by ischemia and/or hypoxia, leading to bacterial invasion. Spread of bacteria in this manner is called agonal spread. Bacterial growth in agonal spread is likely mixed and includes both pathogens and commensals. In postmortem transmigration, bacteria migrate from the mucosal surface into blood after death during putrefaction. In the case of contamination, bacteria are introduced into the sample during its collection [2]. The main postmortem artifacts are contamination and transmigration, as agonal spread seems to be a theoretical concept [2]. Pure growth of a pathogen in blood or cerebrospinal fluid may contribute to death [2]. Although postmortem bacterial culture has gained value recently, there are no standard tests for culture and identification. In Korea, the genetic method had been used for postmortem bacterial culture and identification in forensic medicine. In this study, we compared the genetic and biochemical methods using the same samples and aimed to identify the more useful method for postmortem bacterial culture and identification in autopsy practice.

A total of 225 autopsies were performed at National Forensic Service Gwangju Institute in the study period. For 34 autopsies, both genetic and biochemical methods for postmortem bacterial culture and identification were performed. Data on the cause of death, detected bacteria, kinds of samples, autopsy findings including histological examinations, and antemortem information were collected from these cases. A total number of 78 samples including 32 blood samples and 46 other samples were obtained from the 34 cases.

A microbiological autopsy is performed as follows: as soon as the pericardium is opened using sterile instruments, 70% alcohol is applied to the inferior vena cava and blood is drawn from the heart with a sterile syringe. Cerebrospinal fluid is sampled by puncture of the posterior cerebellomedullary cistern after disinfection of the skin using 70% alcohol. Sputum is sampled from main bronchus before the evisceration of the lung. Other fluids such as pleural effusion, pericardial fluid, and ascites are collected using transport medium swabs with caution to prevent contamination. All instruments were autoclaved before use.

For detecting bacteria, both genetic and biochemical methods were used. In the genetic method, bacterial cultures were performed using blood agar and MacConkey's agar in both aerobic and anaerobic conditions. After culture, bacterial colonies were obtained were subjected to bacterial identification using FAST MicroSEQ 500 16S rRNA Bacterial Identification Kit (Applied Biosystems, Foster City, CA, USA) in the genetic methods. In the biochemical method, bacterial cultures were performed in the same way as the genetic method, and bacterial identifications were conducted using Vitek2 system (BioMérieux, Marcy l'Etoile, France).

The age at death ranged from 2 days to 79 years (average, 41 years; 19 males and 15 females). Of the 34 cases, bacteria were identified in 28 cases (82.4%). Of the 32 samples for which total blood bacterial culture was performed, 23 samples (71.9%) showed positive results. Of the 46 other samples, bacteria were identified in 34 samples (73.9%). The main bacteria identified in the blood were Enterococcus faecium, Escherichia coli, Acinetobacter baumannii, and Enterococcus faecalis. The main bacteria identified in the other samples were Staphylococcus aureus, Escherichia coli, Pseudomonas aeruginosa, and Streptococcus pneumoniae. From the 78 samples obtained from 34 cases, 74 samples (30 blood samples and 44 other samples) were analyzed using both genetic and biochemical methods for postmortem bacterial culture and identification. In the remaining four samples, either one method for postmortem bacterial culture and identification was used. These 74 samples were used to compare the result of the two methods (Table 1). The average time until reporting of result was 15.3 days (range, 7 to 24 days) in the genetic method, and the average time required for the biochemical method was 5.9 days (range, 2 to 14 days). In some samples, bacteria were only detected by one method: in four samples (21.1%) bacteria, by the genetic method and in 15 samples (78.9%), by the biochemical method. Of the 34 cases, only eight cases (23.5%) showed same results for the two methods, and of the 74 comparable samples, only 28 samples (37.8%) showed same results for the two methods. Only nine of the 30 blood samples showed same results: eight of the nine samples showed a negative result and only one showed presence of Enterococcus faecium by both methods. Among the 44 other samples, 19 samples showed same results by both methods, of which 12 samples showed negative results and seven showed positive results (Table 2). Twenty-six of the 34 cases showed inconsistent bacterial culture results (Table 3).

Table 1 General considerations of samples Click for larger image

Table 2 Postmortem bacterial findings in consistent cases Click for larger image

Table 3 Postmortem bacterial findings in inconsistent cases Click for larger image

Since the advent of scientific postmortem examination in the 19th century, numerous studies have been performed on postmortem bacterial culture. However, there is no clear conclusion of its effectiveness. Although postmortem bacterial culture is valuable in understanding infectious disease, epidemiological investigation, and postmortem investigation and autopsy, no specific guidelines and recommendations for postmortem bacterial culture have been established [1]. In Korea, the genetic method for postmortem bacterial culture and identification is used in forensic medicine; however, it is not used as much for determining the cause of death due to several reasons. In this study, we compared the results of the genetic method and the biochemical method to identify the appropriate method and protocol for postmortem bacterial culture in forensic medicine.

Of the 34 cases, 28 cases showed positive bacterial culture and six cases showed negative results, which was possibly due to viral infections, unsuitable samples, or absence of infectious disease. The average reporting time for the biochemical method (5.9 days) was shorter than that required for the genetic method (15.3 days) because the genetic method involves the use of complex testing procedures such as polymerase chain reaction after bacterial culture. For some cases, only one of the methods detected bacteria: four cases (21.1%) for the genetic method and 15 cases (78.9%) for the biochemical method. Therefore, we believe the biochemical method is more sensitive than the genetic method, and the use of liquid transfer medium in the biochemical method could account for the differences between the two methods. Seventy-four samples from 34 cases were analyzed using both genetic and biochemical methods, and only 28 samples showed consistent results. Of the 28 samples with consistent results for bacterial culture tests, 20 samples showed negative results, indicating that majority of results identified by the two methods was inconsistent.

In addition to its usefulness for determining the cause of death in forensic medicine, postmortem bacterial culture testing is valuable in clinical medicine for several reasons including identification of the pathogens that were unidentified during the antemortem period and confirmation of the antemortem infectious disease such as infective endocarditis [3]. Moreover, an autopsy or postmortem bacterial culture could help in evaluation of the range and severity of the infectious disease identified during the antemortem period [4] by providing valuable data to identify the pathogen if the deceased could not provide appropriate data before demise, and help in evaluation of antibacterial treatment before death [5]. Finally, postmortem bacterial culture is valuable for alleviating nosocomial infections by identifying the pathogen.

In order to increase the value of postmortem bacterial culture, forensic pathologists should select appropriate cases for autopsy and interpret the results of postmortem bacterial culture testing while considering other information obtained from tests such as autopsy and histologic findings. Further, postmortem bacterial culture testing should be managed by a clinical bacteriologist to maintain standard protocols for sample handling and storage, use of appropriate method for bacterial culture and identification, and proper analysis of the culture result. The sampling must be conducted as soon as possible: although the recommended period is within 24-48 hours after death [1], it is not mandatory. The method used for sampling must be as sterile as possible [1]. Bodily fluid samples such as blood must be sampled before tissue sampling, and the tissue samples must be taken before organ extraction. To accomplish this, the instruments for autopsy should be autoclaved or disposable, and it ≥2 samples may need to be sampled from a single case. In addition, an appropriate sample for the postmortem bacteria culture test must be taken. For blood culture, Staphylococcus aureus, Streptococcus pneumoniae, or Escherichia coli is considered a pathogen, and if one of them is the only detected bacteria, it should be considered significant [1]. The cerebrospinal fluid is not sampled in the all autopsies of adults, but recognized as a valuable sample for sudden death of infants [6]. For this study, Streptococcus pneumoniae was detected in meningitis cases by both methods, and no bacteria were detected in the cerebrospinal fluid for most adults. The collected sample must be placed in appropriate media and sent to the laboratory as soon as possible. For samples in which the presence of bacteria is unexpected, such as blood and cerebrospinal fluid, appropriate bacterial culture may be performed before process of identification. Clinical and pathological data after the autopsy must be considered when analyzing the postmortem bacterial culture results. As mentioned above, the postmortem bacterial culture including selection of appropriate sample, collection of sample, culture and identification, and analysis of the result must be managed by clinical bacteriologist.

In conclusion, this study confirmed that in Korea, the results obtained by the genetic method for postmortem bacterial identification using FAST MicroSEQ 500 16S rRNA Bacterial Identification Kit (Applied Biosystems) and those obtained by the biochemical method using the conventional method with the Vitek2 system (BioMérieux) were inconsistent. We also analyzed the causes of the inconsistency and provided acceptable conditions and protocols for postmortem bacterial culture and identification. To improve the use of postmortem bacterial culture and identification in forensic medicine in Korea, further research should be performed.