Post-Mortem Interval Estimation by Microbiological Analysis of Conjunctiva and Vitreous Humor

(1)

Research Project Report

Post-Mortem Interval Estimation by Microbiological Analysis of

Conjunctiva and Vitreous Humor

February 2018 - July 2018

36 EC

Agustina Florencia Brunetti Rolón

Student ID: 11403748

MSc Forensic Science, University of Amsterdam

Examiner: Dr. J.A.J. Hans Breeuwer Supervisor: Dr. Ignasi Galtés Vicente

Research institutes: Institut de Medicina Legal i Ciències Forenses de Catalunya (IMLCFC), Universitat Autònoma de Barcelona (UAB)

Date: October 2018

(2)

2

1. Abstract

The estimation of the time since death or post-mortem interval (PMI) is still one of the pressing, unresolved issues in the field of Forensic Science. Forensic Microbiology has been proven useful to determine the PMI in past cases, and previous studies exploring the relationship between human microbiota and PMI have yielded encouraging results. This research had two aims: determining whether there was a correlation between changes in the microbiota of the conjunctiva and vitreous humor and PMI, and developing a statistic model to estimate unknown PMIs depending on conjunctiva and vitreous humor microbiota. 30 human cadavers were sampled, and Staphylococcus and Streptococcus species were detected and quantified using traditional microbiology methods. Simple and multiple linear regression analyses were performed to study the relationship between the number of microorganisms of each species and the corresponding PMI for each sample, as well as the impact of several qualitative variables. It was concluded that vitreous humor is sterile after death, at least during the range of PMIs that were included in this study. Regarding the conjunctiva samples, Staphylococcus aureus showed the strongest linear relationship between CFUs and PMI, and it was also the best-fitted model generated by multiple linear regression analysis. Overall, a correlation between changes in the microbiota of the conjunctiva and PMI was found, even though the relationship between variables appeared to be from weak to moderate. A statistic model to estimate unknown PMIs was generated for each of the studied microorganisms.

2. Introduction

The estimation of the time since death or post-mortem interval (PMI) is still one of the pressing, unresolved issues in the field of Forensic Science, due to the lack of consensus regarding the certainty of the currently available methods, and its often central role in criminal investigation. The latter suffers especially from this generalized unreliability, as a wrong estimation of the PMI of a victim may lead to wrong convictions in court. Some cases of high media impact, such as the Lundy murders in Palmerston North, New Zealand in 2000 (1), have had the PMI playing an important role for the alibi of the accused, and the use of untrustworthy methods to estimate it has added even more complexity to the development of the trial.

To name a classic example, the measurement of potassium concentration in the vitreous humor is one of the resources that is still currently in use (2)(3), even though it is not free of disagreements on its reliability within the scientific community, and recent research is still

(4)

4 unable to provide a perfect correlation between the concentration of this component and the time since death (4). Therefore, exploring new techniques to estimate the PMI is of upmost importance to the forensic field; ideally, this hypothetical method should be able to estimate the PMI with more certainty than it is possible nowadays, and it should also be able to perform in a non-invasive way that would not interfere with other routine analyses, such as DNA testing. Forensic Microbiology is a slowly growing but promising subfield, which first came into the spotlight due to the use of Bacillus anthracis spores as a biological weapon in the US in 2001 (5)(6). However, biocrime is not the only point of focus of this discipline; fungal spores have been used in the past as trace evidence to link a suspect to a crime scene, and certain microbiology tools can be used to help determine the cause of death, as it is the case with the use of the diatom test to diagnose drowning (7)(8).

Forensic Microbiology has been proven useful to determine the PMI too in past cases, especially by the assessment of fungal growth on cadavers (8). The changes in the microbial communities (or microbiota) of the soil surrounding cadavers has also been recently used to estimate the time since death, based on the growth curve of Firmicutes from human remains (9). Skin microbiota has shown promising results in a recent study as well, where decomposing human bodies were sampled and models of statistical regression were developed to predict the PMI of microbial samples (10). In another experiment performed in mice models, the post-mortem changes in the microbial communities of skin and abdominal cavity were proven to be measurable and repeatable. The authors were able to estimate the PMI within three days (11).

Even though there have already been studies, such as the ones summarized above, that explored the relationship between human microbiota and PMI, eye microbiota was yet to be studied in this context until the start of this study. This research took two different approaches to studying the microbiota of the eye: through conjunctiva samples, and through vitreous humor samples. The reason why eye microbiota was chosen as the subject of the study, aside from its novelty, was that it presents the advantage of being sampleable at the crime scene with no countereffects on other methods and tests that the forensic investigator might want to perform as well. Taking a conjunctiva swab sample is a completely non-obtrusive procedure, and only a tiny amount of vitreous humor is needed to perform microbiological analysis on it, which means that it would not interrupt the routine biochemical analysis of this fluid.

Concerning the conjunctiva of the eye, where microorganisms live symbiotically during life, it has been proven that bacteria can be isolated from it within 6 hours after death. It is worth noting that less bacteria are isolated from earlier samplings (12). On the other hand, the

(5)

5 detection of post-mortem bacterial growth in vitreous humor has been studied before, the results being positive in a majority of samples but presenting a very low concentration of microorganisms per milliliter (13). Moreover, it has been proven that vitreous humor extracted after death does not work as a growth medium for bacteria inoculated in vitro (14), which potentially makes tampering of those samples futile.

Overall, the results obtained by previous experiments on the estimation of the PMI using human microbiota, and the encouraging background knowledge on eye microbiota, were the main foundation of this research which had two defined aims: determining whether there is a correlation between changes in the microbiota of the conjunctiva and vitreous humor and the PMI of the individuals, and, if said correlation were to be found, developing a statistic model based on the obtained results, which could be used to estimate the unknown PMI of individuals depending on their conjunctiva and vitreous humor microbiota.

3. Materials and Methods

3.1. Subjects

30 human cadavers from the Institut de Medicina Legal y Ciències Forenses de Catalunya (IMLCFC) were used in this project. The IMLCFC receives bodies from all over the province of Barcelona which are deemed to need medico-legal autopsies. Previous to the start of the research, the ethics committee of the IMLCFC was consulted, and authorization was provided for the sampling of the bodies.

Regarding the selection of the 30 subjects, they were chosen regardless of sex, age, location or cause of death, as not to make the resulting PMI estimation model specific to a certain demography. The selection criteria were instead based on the subjects having the eye area intact (i.e.: no visible trauma), and on the confidence with which the PMI of the individual could be estimated. Therefore, bodies that arrived from the hospital and had their time of death recorded, and bodies that did not present an advanced state of decomposition, were preferred. The reason why these were the selected criteria was that it was conjectured that eye trauma, especially if it was due to a fall to the ground, could influence the host microbiota of the eye and in turn alter the obtained results. Regarding the preference for PMIs which were easier to estimate, it was so the correlation between amount of microorganisms and PMI could be as exact as possible.

(6)

6 A summary of the subjects regarding sex, age, location, manner of death and estimated PMI at the time of autopsy can be found in Table 1. The subjects are sorted by date of arrival to the Institute, and therefore date of sampling. There were 20 males and 10 females in the sample, all in the age groups from 35 to ≥75 years old; they were collected from either their homes, the public road or the hospital, and the manner of death was natural in most cases. The 30 subjects were received in the Institute from February to May of 2018 (Table 1).

Table 1. Summary of sample, sorted by date of sampling of the subjects.

Subject Date of _sampling Sex Age _{(years) Location} Manner of _death Estimated PMI _(hours)

1 08/02/18 M 45-54 Home Natural 10 - 11 2 08/02/18 M 45-54 Home Suicide 22.5 - 34.5 3 08/02/18 M 55-64 Home Natural 12.25 - 13.25 4 08/02/18 M 65-74 Home Natural 6 - 7 5 08/02/18 F 45-54 Hospital Accident 29 6 08/02/18 M ≥75 Hospital Accident 21 7 20/02/18 M ≥75 Home Natural 44.75 - 50.75 8 12/03/18 M 55-64 Road Suicide 31.5 - 32.5 9 19/03/18 M 55-64 Home Natural 45.25 - 51.25 10 20/03/18 M ≥75 Hospital Natural 21.25 11 09/04/18 M 55-64 Home Accident 23.75 - 24.75 12 09/04/18 M 55-64 Hospital Natural 24 13 09/04/18 F 65-74 Home Natural 22.5 - 35.5 14 09/04/18 M 65-74 Home Natural 34 - 37 15 11/04/18 M ≥75 Hospital Natural 24.25 16 12/04/18 M 65-74 Road Natural 22 - 24 17 17/04/18 M 45-54 Home Natural 22 - 24 18 17/04/18 M 35-44 Home Natural 20.25 - 28.25 19 17/04/18 M 55-64 Home Natural 29.25 - 33.25 20 30/04/18 M 55-64 Home Natural 91.5 - 115.5 21 02/05/18 F 65-74 Home Natural 30.5 - 34.5 22 03/05/18 F 65-74 Home Natural 20.75 - 24.75 23 03/05/18 M 45-54 Home Suicide 21.5 - 22.5 24 08/05/18 F 35-44 Home Natural 25.5 - 26.5 25 08/05/18 F 55-64 Home Natural 23.5 - 27.5 26 09/05/18 F ≥75 Home Natural 26.5 - 34.5 27 09/05/18 F 35-44 Home Suicide 46 - 70 28 09/05/18 F 55-64 Home Natural 28.5 - 29.5 29 15/05/18 M 45-54 Home Natural 20.75 - 24.75 30 15/05/18 F 35-44 Hospital Indeterminate 19.5 3.2. Sampling

Each body was sampled twice: first when it arrived to the IMLCFC, and second right before the autopsy. The first sampling time will hereby be referred to as t0, and the second sampling time

(7)

7 Both the sampling of conjunctiva and the sampling of vitreous humor were performed using DELTALAB's STUART swabs with transport medium (15). In the case of the conjunctiva samples, they were taken by swabbing the tarsal conjunctiva from both eyes, using a different swab for each. In the case of the vitreous humor samples, they were taken by swabbing the leftovers of fluid in the barrel of the syringe after extraction for routine analysis. Conjunctiva was sampled both in t0 and in t1, while vitreous humor was only sampled in t1.

Overall, five samples (four conjunctiva samples and one vitreous humor sample) were taken from each subject. The samples taken during the week were preserved at 2°C until they were taken to the Animal Sanitation and Anatomy Department of the Veterinary Faculty of the Universitat Autònoma de Barcelona (UAB) for plating and incubation. The samples were transported in refrigerated conditions.

It is worth noting that vitreous humor was only sampled from cases 1 to 8 and in case 20, due to the lack of positive results. This issue will be detailed further in the Results section of the report.

Control samples were taken from the fridges where the bodies were stored, as to discard the possibility of the conjunctiva samples being hypothetically contaminated by microorganisms present in the environment. The control samples were taken from a fridge chosen at random by direct sedimentation of air particles onto suitable agar plates, as suggested by the Centers for Disease Control and Prevention (16). The plates were placed in pairs, both outside and inside the bags that contained the bodies, and they were incubated immediately after the test was performed.

3.3. Microbiological analysis

The microorganisms that were chosen for detection and quantification were Staphylococcus aureus, Staphylococcus epidermidis, Streptococcus pneumoniae and Streptococcus pyogenes, due to their habitual presence in eye microbiota (17)(18) and the possibility to identify them with traditional microbiology methods.

As for the plates that were used, Baird-Parker Agar was chosen for being a selective medium for isolating and quantifying Staphylococcus species (19), and Columbia Agar with 5% Sheep Blood (namely Blood Agar) was chosen for being a nutritive medium where hemolysis can be easily observed, which is a useful feature to differentiate between Streptococcus pyogenes and Streptococcus pneumoniae (20). Regarding the plates that were used for the control samples, Trypticase Soy Agar (21) and Sabouraud Glucose Agar (22) were chosen, the former for being a

(8)

8 very general medium that supports the growth of a wide range of microorganisms, and the latter for being suitable for the isolation of fungi (which occur very frequently in fridges).

The plating of the samples was performed in a laminar flow cabinet to ensure sterility of the environment and reduce the risk of contamination. Each swab was inoculated on a Blood Agar plate and a Baird-Parker Agar plate, in that order, since Baird-Parker Agar is a more restrictive medium and inoculating it first could lead to the loss of bacteria. The plates were inoculated by gently dragging the swab on the agar surface in a zigzag pattern, and they were incubated right after.

The incubating conditions were aerobic with a 37°C temperature for the conjunctiva samples, and microaerophilic (5% CO2 atmosphere) with a 37°C temperature for the vitreous humor

samples, as to replicate the environment of the inside of the eyeball. As for the incubation time, it oscillated between 48h and 72h, depending on the observed bacterial growth on the plates. Regarding the control samples, the incubating conditions were aerobic with a 37°C temperature for the Trypticase Soy Agar plates, and aerobic with a 28°C temperature for the Sabouraud Glucose Agar plates. The incubation time was of 72 hours.

All the plates were read immediately after incubation, by counting the colony-forming units (CFU) and evaluating a series of macroscopic and microscopic features which are summarized in Table 2. Note that the coagulase and catalase test were performed following the protocols by the American Society for Microbiology (23)(24). Regarding microscopic observation, it was performed using a light microscope after staining with the Gram Stain, as suggested by the protocol also by the American Society for Microbiology (25).

Table 2. Summary of the macroscopic and microscopic features that were evaluated to identify each

microorganism on its suitable plate.

Baird-Parker Agar

Microorganism Macroscopic features Microscopic features

Staphylococcus aureus Black colony with halo,

coagulase +, catalase + Gram+, grape-like clusters, no spores

Staphylococcus epidermidis Black colony without halo,

coagulase -, catalase + Gram+, grape-like clusters, no spores

Columbia Agar with 5% Sheep blood

Microorganism Macroscopic features Microscopic features

Streptococcus pneumoniae White/gray colonies, α

hemolysis Gram+, pairs, no spores

Streptococcus pyogenes White/gray colonies, β

(9)

9 3.4. Data preparation

Prior to the analyses, the numbers of CFUs obtained from right and left eye at each sampling time (t0 and t1)were averaged. The obtained numbers were then converted from CFUs to

CFUs/mm2_{, considering 5 mm}2_{as the average area for the tarsal conjunctiva (26).}

The numbers of hours from death to t0 and hours from death to t1 (i.e.: the PMI of the subjects

at each sampling time) also had to be averaged. Except in the cases where the bodies arrived from the hospital, and therefore had a registered time of death, the numbers of hours from death was averaged from the PMI range estimated by the forensic pathologist at the time of autopsy, which can be found in Table 1.

3.5. Statistical analysis

Simple linear regression analysis was used to study the relationship between the number of microorganisms of each species and the corresponding hours after death for each sample. Additionally, multiple linear regression analysis was used to study the impact of several qualitative variables from the different subjects. The statistical analysis of the obtained results was performed using NCSS 12 (27).

4. Results

In Table 3, the number of hours between samplings, between each sampling and plating, and between plating and reading, are summarized. The raw data sorted by subject can be found in the Annex (Annex 1).

The amount of hours between samplings depended on when the body arrived at the IMLCFC, since the time of the autopsies was fixed to around 9 AM and 16 PM; the values go from 5.25 to 21.25 hours (Table 3). Regarding the hours between samplings and plating, and the hours from plating to reading (i.e.: the incubation time of the samples), they depended on the availability of the microbiology laboratory; the values go from 15 to 135, 2 to 195, and 48 to 120 hours for the times between t0 and plating, t1 and plating, and plating to reading respectively (Table 3).

These differences to plating time between subjects pose no influence over the obtained results, since all samples were refrigerated from sampling to plating. In the case of the hours between samplings, the variability between cases was beneficial, since it allowed for a wider range of PMIs to be studied.

(10)

10 Table 3. Summary of the number of hours between samplings, samplings to plating, and plating to

reading.

Hours t0 to t1

Hours t0

to plating Hours tto plating 1 Hours plating to reading (Incubation time)

Minimum value 5.25 15 2 48

Maximum value 21.25 135 195 120

In Table 4, the number of hours between death and both sampling times (i.e.: the PMIs at t0 and

at t1), after data preparation, are summarized. The raw data sorted by subject can be found in

the Annex (Annex 2). The values go from 2 to 43.5 hours for the PMI at t0, and from 6.5 to 103.5

hours for the PMI at t1 (Table 4).

Table 4. Summary of the PMIs of the subjects at t0 and at t1.

PMI at t0 PMI at t1

Minimum value 2 6.5

Maximum value 43.5 103.5

4.1. Microbiological analysis results

The results of the microbiological analysis can be consulted in Table 5, which summarizes the number of CFUs and negative results (i.e.: number of CFUs was 0) obtained for each microorganism species at each sampling time, from each eye. The raw data sorted by subject can be found in the Annex (Annex 3). The obtained values range from 0 to as many as 197 CFUs; it is also worth noting that 256 out of 416 results were negative. The number of CFUs was significantly higher at t1 than at t0 only in the case of S. aureus results (p = 0.0439, α = 0.05),

while there was no significant difference between right and left eye for any of the studied microorganisms (Table 5). The results of these two significance tests can be consulted in the Annex (Annex 4).

Table 5. Summary of the results of CFU quantification from conjunctiva samples. (S. aur.: Staphylococcus

aureus, S. epi.: Staphylococcus epidermidis, S. pne.: Streptococcus pneumoniae, S. pyo.: Streptococcus pyogenes.)

Right eye _{S. aur. S. epi. S. pne. S. pyo. S. aur. S. epi. S. pne. S. pyo.}Left eye

At t0 Minimum CFU value 0 0 0 0 0 0 0 0

Maximum CFU value 13 54 178 165 8 87 197 110

At t1 Minimum CFU value 0 0 0 0 0 0 0 0

Maximum CFU value 61 174 165 142 47 38 162 27

Negative results 34/52 30/52 37/52 29/52 36/52 31/52 39/52 20/52 Total negative results 256/416

(11)

11 Table 5 only includes the quantification of CFUs from conjunctiva samples, since vitreous humor samples only yielded one positive result, from subject 20, as depicted in Fig. 1. The colony was presumptively identified as Streptococcus pyogenes; complete lysis of the red cells in the medium (β hemolysis) can be observed. Since only one positive result was obtained, it was not used for any further calculations.

Fig. 1. Presumptively identified Streptococcus pyogenes CFU, on a Blood Agar plate, from the vitreous

humor sample from subject 20.

Regarding the control samples, all eight agar plates turned out negative, as depicted in Fig. 2.

Fig. 2. Control samples, all negative. The plates on subfigures (a) and (b) were placed inside the body bag

for sampling, and the plates on subfigures (c) and (d) were placed outside. (T: Trypticase Soy Agar, S: Sabouraud Glucose Agar.)

(12)

12 Lastly, examples of CFUs of the chosen microorganisms can be seen in Fig. 3 and Fig. 4, on their suitable plates and displaying the macroscopic features that were evaluated for their identification. In Fig. 3, S. aureus CFUs are those with a halo around the black colony, such as the two circled on the right side of the picture, while S. epidermidis CFUs are those without a halo around the black colony, such as the one circled on the left side of the picture. In Fig. 4, S. pyogenes CFUs are those with a β hemolysis halo around the colony, such as the one circled on the upper part of the picture, while S. pneumoniae CFUs are those with an α hemolysis halo around the colony, such as the one circled below the S. pyogenes CFU. The β hemolysis halo cannot be appreciated clearly, since the picture was not backlighted. However, the α hemolysis halo can be perceived as a darker shade of red spread in a half moon shape. Most of the CFUs had α hemolysis, which resulted in the halos fusing and spreading throughout the plate.

Fig. 3. Baird-Parker Agar plate from conjunctiva sample taken from subject 3. (S. aureus: Staphylococcus

aureus, S. epidermidis: Staphylococcus epidermidis.)

Fig. 4. Blood Agar plate from conjunctiva sample taken from subject 11. (S. pneumoniae: Streptococcus

(13)

13 4.2. Statistical analysis results

Four sets of data of 52 observations each, from the conjunctiva results only, were analyzed: one set for each microorganism. The complete sets of data can be found in the Annex (Annex 5). Fig. 5 shows the results of the simple linear regression analyses. Four graphs were generated, one for each microorganism. The noticeable amount of observations on the Y axes are due to more than half of the plates being negative, as mentioned above. It is also remarkable that the only two plots that show an evident positive correlation between hours since death and CFUs are the graphs for S. aureus and S. epidermidis, while S. pneumoniae appears to show a slightly negative correlation and S. pyogenes does not seem to display a correlation that is either positive or negative.

(14)

14 Table 6 shows the correlation coefficient R for each microorganism, as well as the R2_{, slope,}

y-intercept, and estimated equation. Table 7 shows the results for the significance tests, which were run with a null hypothesis (H0) that the slopes were zero, to determine whether or not

there was a relationship between the variables.

Table 6. Results of the simple linear regression tests and obtained equations for the prediction of hours

since death from CFUs. (s.e.: standard error, Hrs: hours since death.)

Microorganism R R2 Slope

(s. e.) Y-intercept (s. e.) Equation Staphylococcus aureus 0.3159 0.0998 (1.3929) 3.2791 (2.3116) Hrs = (22.1608) + (3.2791) * CFUs 22.1608 Staphylococcus epidermidis 0.0942 0.0089 (0.6753) 0.4518 (2.4370) Hrs = (23.4533) + (0.4518) * CFUs 23.4533 Streptococcus pneumoniae -0.0319 0.0010 (0.2390) -0.0539 (2.5926) Hrs = (24.3092) + (-0.0539) * CFUs 24.3092 Streptococcus pyogenes 0.0019 0.0000 (0.6941) 0.0096 (2.4527) Hrs = (24.0215) + (0.0096) * CFUs 24.0215

Table 7. Results of the significance tests, run with an Alpha (α) of 0,05. H0 = The slope of the simple linear

regression is 0.

Microorganism t-value p-value Reject H0? Linear relationship?

Staphylococcus aureus 2.3541 0.0225 Yes Yes

Staphylococcus epidermidis 0.6690 0.5065 No No

Streptococcus pneumoniae -0.2254 0.8226 No No

Streptococcus pyogenes 0.0138 0.9891 No No

The results obtained from the S. aureus model show the strongest correlation between CFUs and hours since death (R = 0.3159) out of the four studied microorganisms (Table 6); it is also the only one where the significance test was successful (p-value = 0.0225 < 0.05; H0 was

rejected), and the existence of a linear relationship between variables, albeit not too strong, could be confirmed. The significance test was run considering an Alpha (α) of 0.05 (Table 7). All other three studied microorganisms yielded very low correlation values (R = 0.0942 for S. epidermidis, -0.0319 for S. pneumoniae, and 0.0019 for S. pyogenes)(Table 6). Their significance tests also determined that the relationship between CFUs and hours since death for their sets of data was not linear, since their p-values were higher than 0.05 and therefore H0 could not be

rejected (Table 7). This lack of a linear relationship does not necessarily indicate that the number of CFUs of those microorganisms and the PMI of the individual are independent, though. A more complex, non-linear model might be required to determine how those variables are related in those three cases.

Multiple linear regression analyses were performed in order to study the relationship between the aforementioned quantitative variables, and a series of qualitative variables from the

(15)

15 subjects. The qualitative variables that were chosen to be evaluated were: sex, age, location of the body, manner of death, and state of the eyes.

Five age groups were formed, as not to have too many categories for one variable: 35 to 44, 45 to 54, 55 to 64, 65 to 74 and ≥75. Regarding location, there were three possible locations where the bodies were found: at their homes, on the public road, and in the hospital. Locations such as car, truck, sports center and school were all grouped under “home” to simplify the calculation. The possible manners of death were natural, accident, suicide or indeterminate. Lastly, the state of the eyes could be open, closed or mixed. The number of subjects for each of the aforementioned qualitative variables can be consulted in the Annex (Annex 6).

Table 8 shows the R2_{, coefficient of variation and F-ratio for each microorganism. S. aureus}

yielded the highest R2_{out of the four microorganisms (0.4007), as well as the lowest coefficient}

of variation (0.6443); however, the difference between values yielded by each microorganism are not too broad. Regarding the F-ratio, S. aureus yielded the highest value as well (1.337). These three values are indicative that the model generated by the S. aureus results has a better fit than the ones generated by the other three studied microorganisms (Table 8).

Table 8. Results of the multiple linear regression tests for each microorganism. (CV: coefficient of

variation.) Microorganism R2 _CV _F-ratio Staphylococcus aureus 0.4007 0.6443 1.337 Staphylococcus epidermidis 0.3209 0.6859 0.945 Streptococcus pneumoniae 0.3510 0.6705 1.082 Streptococcus pyogenes 0.3220 0.6854 0.950

Table 9 summarizes the impact of each of the qualitative variables on the results yielded by each microorganism. In all cases, the variable that signified the highest R2_{lost if it were to be}

removed, and therefore the most influential variable, was the location where the bodies were found. This lost for the R2 _{value goes as far as 0.2137 for the}_{S. pneumoniae results, which is}

quite high considering that the overall R2 _{was 0.3510. On the other hand, the least influential}

qualitative variable in all four cases appears to be the sex of the subject, even reaching a value of 0 for S. aureus (Table 9). These results were considered to be quite interesting, since it was conjectured that the state of the eyes (i.e. whether they were open, closed, or one each at the time of sampling) would be the qualitative variable to have the greatest impact on the R2_;

however, the test places it behind the location of the body and the age of the subject, which suggests that whether the eyes of the subject were open or closed is not as influential to the microbial communities of the eye as the environment where the body is found, or the age of the host.

(16)

16 Table 9. Values of R2_{lost for each microorganism if each qualitative variable is removed from the}

calculations.

Microorganism Location of _{the body} Age State of the _eyes Manner of _death Sex

Staphylococcus aureus 0.1477 0.0559 0.0420 0.0303 0.0000

Staphylococcus epidermidis 0.1910 0.1034 0.0462 0.0166 0.0083

Streptococcus pneumoniae 0.2137 0.1398 0.0604 0.0180 0.0032

Streptococcus pyogenes 0.1901 0.1162 0.0467 0.0168 0.0081

5. Discussion

Even though several studies regarding the thanatomicrobiome had been published (9)(10)(28)(29)(30), up to the start of this project the relationship between eye microbiota and PMI was yet to be researched. In addition, most of the studies on microbial communities and determination of the time since death were performed in animal models (11)(31)(32)(33)(34) instead of human subjects, which constitutes a liability, since microbiota studies are not always translatable between species.

The eye was the chosen location to sample not only because of the novelty, but also because of its practicality. The sampling of vitreous humor is already routine, since it is used to estimate the PMI from its concentration of potassium ions, and taking a conjunctiva sample with a swab is a simple, quick, and non-invasive procedure. It is also worth noting that, out of the whole ocular surface, the conjunctiva yields the higher number of microorganisms when using traditional culture methods (35), which made it the perfect sampling location for this study.

Vitreous humor was quickly discarded as a good sample to predict PMI from microbiological analysis. Only one of the plated samples was positive, and therefore it has been theorized that the result was obtained due to contamination during the extraction of the vitreous humor. No calculations were made with said result. It was concluded that vitreous humor is sterile after death, at least during the range of PMIs that were included in this study (the earliest sampling was two hours after death (t0 subject 16), and the latest was almost five days after death (t1

subject 20, coincidentally the only positive vitreous humor sample.))

Regarding the control samples, all the plates turned out negative, which suggests that the possibility of the results being affected by microorganisms present in the fridges where the bodies were stored can be safely discarded.

Now delving into the conjunctiva results, the simple linear regression analyses revealed that only S. aureus displayed a linear relationship between hours since death and number of CFUs (p-value

(17)

17 = 0.0225 < 0.05). While S. epidermidis, S. pneumoniae and S. pyogenes did not display a linear relationship between said variables (p-values = 0.5065, 0.8226 and 0.9891 respectively), that does not mean the variables are definitely independent from each other, as was mentioned above. A different model may be necessary to determine the relationship between variables for those microorganisms; models which are not linear, however, are beyond the scope of this study. Even so, the S. aureus model does not show a strong linear relationship (R = 0.3159). The S. pneumoniae model even shows a slightly negative correlation instead of a positive one (R = -0.0319).

Regarding the multiple linear regression analyses, again the S. aureus model appears to be the best-fitted one, with a R2_{of 0.4007. It also shows the lowest coefficient of variation out of the}

four models (0.6443) and the highest F-ratio (1.337), which are indicators of a well-fitted model and statistical significance, respectively. However, the conclusions that can be drawn from these values are not strong either. It is also worth noting that there was not such a huge difference between models generated by multiple linear regression analysis, as there was in the ones generated by simple linear regression analysis; for example, the R2 _{of the other three models}

stood at 0.3209, 0.3510 and 0.3220, which are not too far away from the S. aureus value. A similar tendency can be seen for the coefficient of variation and F-ratio values. Overall, S. aureus showed the most encouraging results again, but the relationship between variables turned out weak once more.

Another interesting result was found during the multiple linear regression analyses, when evaluating the values of R2_{lost if each qualitative variable was removed. Said analysis revealed}

that the location of the body was the most influential variable in that sense, for all four models. It might indicate that the microbiota of the eye after death is heavily influenced by the environment. The following variable with the highest value, again in all four models, was the age of the subject, which suggests that the microbial community varies in life as the individual ages. Next was the state of the eyes, which was expected to have a heavier influence on the R2_{. It still}

showed a loss from 0.0420 to 0.0604, which is relatively significant considering that the R2_values

ranged from 0.3209 to 0.4007. Following the state of the eyes was the manner of death, and the least significant variable for the R2_{calculation was the sex of the individual, going as far as having}

no influence at all if removed from the model (in the case of the S. aureus results). In respect of the manner of death, it was conjectured whether the cases that were classified as indeterminate would pose a problem in the interpretation of the results, but seeing that the variable as a whole had such little influence over the model, those suppositions have been discarded.

(18)

18 One of the issues that might have been detrimental to the obtained results was the sample size and diversity. While the IMLCFC receives several bodies each day, not all of them were suitable for this experiment. Subjects with the eye area compromised were discarded, as well as those without the eye area at all. Subjects for which it was difficult to estimate the time of death confidently, such as those in an advanced state of decomposition, were avoided as well. Sampling the bodies as soon as they arrived at the facility also proved to be a disadvantage when it came to the overall sample size, since samples were only taken from 9:00 AM to 5:00 PM, but the Institute received bodies all day and night long. The diversity of the sample was a factor which was impossible to control as well, and it was also conditioned by the fact that the Institute only receives bodies in need of a medico-legal autopsy (unlike a hospital, where it would have been easier to sample subjects of all ages and the time of death would have been registered in every case).

Regarding the reading of the plates, the four kinds of microorganisms that were studied in this project were identified presumptively, which means that the identification was based on colony morphology, growth on selective media, gram stains, and up to three tests (which were coagulase and catalase tests in this case) (36). Presumptive identification can reach species level, but it is more prone to errors than identification performed through more sophisticated methods such as 16S rRNA gene sequencing (37). For example, the existence of coagulase-negative S. aureus has been reported before (38), which in turn means that there is a possibility of mistakenly identifying a coagulase-negative S. aureus as S. epidermidis or as any other species of coagulase-negative Staphylococci. This would not happen when using a gene sequencing technique, which would identify the species of the isolate beyond any doubts. Working with plates also presents another series of problems; for example, while different agar plates have their own recommended incubation temperature, atmosphere, and growth time (19)(20)(21)(22), certain strains of bacteria can take a longer time to grow, which means that it is possible that some of the plates that were read as negative needed to be incubated longer in order to yield a positive result. The availability of the microbiology laboratory at the UAB was irregular at times, which in turn affected the incubation hours of the different samples.

Lastly, it has been reported that a series of factors can alter the eye microbiota, at least during life (17). For example, any infections and other conditions that can affect the tear film have an influence over the microbiota of the ocular surface, such as dry eye syndrome or the use of contact lenses (which account for a higher bacterial concentration than in healthy, non-contact lenses wearing subjects). It has also been reported that the use of certain ophthalmic antibiotics results in an increase in S. epidermidis (39) and other Staphylococcus species (40). These factors

(19)

19 were not considered, as they were not registered in the medical history of the subjects (when available). In any case, no studies were found that reviewed whether these factors that influence eye microbiota in life also have an effect after death.

As said above, up to the beginning of this project (February 2018) there were no other studies which investigated the relationship between eye microbiota and PMI. However, in April 2018, Pechal et al. published an article (41) which, even though it aimed at determining whether the thanatomicrobiome reflected the microbiome of the individual while alive and could be used to infer the individual’s health status, can be used to compare results to the ones obtained by this research. Contrary to this study, Pechal et al. used 16S rRNA sequencing for microorganism identification and quantification.

Out of the locations that were tested in their investigation, the rectum and the eyes showed the highest phylogenetic abundance. Coupled with the fact that unique taxa to each location were found, this indicates that there are other microorganisms than the ones tested in this study that could have the potential to be good PMI indicators, even though they decrease after 48 h from death. Another good quality of the location was pointed out to be that neither the subject's diet or weight seems to have an influence over the bacterial community, as it does in locations such as the gut.

Another interesting result from Pechal et al.'s article was how the predominant taxa of each location changed within 48 h from death. For the eyes, Staphylococcus species were abundant both before and after 48 h from death, but they showed an increase in abundance after the 48 h mark. On the other hand, Streptococcus species were not as abundant as Staphylococcus species, and they showed a decrease after 48 h. This correlates well with the results obtained in this research; the studied Staphylococcus species (S. aureus and S. epidermidis) showed the highest positive correlations with the PMI of the subjects, while the Streptococcus species (S. pneumoniae and S. pyogenes) showed the lowest values, with even a negative correlation in the case of S. pneumoniae.

6. Conclusion

This research was conceived with two objectives: determining whether there is a correlation between changes in the microbiota of the conjunctiva and vitreous humor and the PMI of the individuals, and, if said correlation were to be found, developing a statistic model based on the

(20)

20 obtained results, which could be used to estimate the unknown PMI of individuals depending on their conjunctiva and vitreous humor microbiota.

Only one of the vitreous humor samples yielded a positive result, and contamination of said sample was suspected. It was concluded that vitreous humor is sterile after death, at least during the range of PMIs that were included in this study. Regarding the conjunctiva samples, S. aureus showed the strongest linear relationship between CFUs and hours since death, out of the four studied microorganisms. It was also the best-fitted model generated by multiple linear regression analysis. Therefore, a correlation between changes in the microbiota of the conjunctiva and the PMI of the individual was indeed found, even though the relationship between variables appeared to be from weak to moderate (R = 0.3159 in simple linear regression analysis, and R2_{= 0.4007 in multiple linear regression analysis). Regarding the second goal of the}

research, the model that could be used to estimate the unknown PMI of individuals would be the obtained equation from the simple linear regression analysis of the S. aureus results (Hrs = (22.1608) + (3.2791) * CFUs), which were the most reliable ones.

In conclusion, while this pilot study did not find strong correlations between the chosen microorganisms and the PMI of the individuals, the results were encouraging enough to warrant a larger research on the topic, especially with the support of the results by Pechal et al. Using more refined methods for the detection and identification of microorganisms in the conjunctiva samples, such as 16S rRNA sequencing, has been proven to yield a higher microbial diversity (42) and therefore a wider field of study to find other good PMI indicators. It would also provide a more reliable and exact quantification for the microorganisms that this study found to have the potential to be useful as PMI indicators, such as S. aureus. Future research should fix issues that were detrimental in this pilot study, such as sample size and sample diversity. It should also expand the scope to other taxa of microorganisms, especially the ones which seem to be unique to the eye microbiota (41).

Overall, Forensic Microbiology as a discipline definitely has the potential to help with the still pressing issue of finding a reliable method of time of death estimation. Further research will help solidify the foundation of this still growing field, which will most likely become a powerful tool for the forensic scientist in the future.

(21)

21

7. Bibliography

1. Radio New Zealand. Inside the case that gripped the nation for 17 years.

https://www.radionz.co.nz/news/national/342053/inside-the-case-that-gripped-the-nation-for-17-years; 2017 Accessed 12 July 2018.

2. Rognum TO, Holmen S, Musse MA, Dahlberg PS, Stray-Pedersen A, Saugstad OD, Opdal SH. Estimation of time since death by vitreous humor hypoxanthine, potassium, and ambient temperature. Forensic Sci Int. 2016;262:160–165.

3. Ding Y, Li X, Guo Y, Duan W, Ling J, Zha L, Yan J, Zou Y, Cai J. Estimation of postmortem interval by vitreous potassium evaluation with a novel fluorescence aptasensor. Sci Rep. 2017;7:1868.

4. Ansari N, Lodha A, Menon SK. Vitreous humor in relation to time since death: A review. Eur J Forensic Sci. 2017;4:240440.

5. Brachman PS. Bioterrorism: an update with a focus on anthrax. Am J Epidemiol. 2002;155:981–987.

6. Jernigan DB, Raghunathan PL, Bell PB, Brechner R, Bresnitz EA, Butler JC, Cetron M, Cohen M, Doyle T, Fischer M, et al. Investigation of Bioterrorism-Related Anthrax, United States, 2001: Epidemiologic Findings. Emerg Infect Dis. 2002;8:1019–1028. 7. Lehman DC. Forensic Microbiology. Clin Lab Sci. 2012;25:114–119.

8. Lehman DC. Forensic Microbiology. Clin Microbiol Newsl. 2014;36:49–54.

9. Adserias-Garriga J, Hernandez M, Quijada N, Rodriguez Lazaro D, Steadman D, Garcia-Gil J. Daily thanatomicrobiome changes in soil as an approach of postmortem interval estimation: An ecological perspective. Forensic Sci Int. 2017;278:388–395.

10. Johnson H, Donovan T, Guzman S, Khan Z, Parziale J, DeBruyn J, Lents N. A Machine Learning Approach for Using the Postmortem Skin Microbiome to Estimate the Postmortem Interval. PLoS One. 2016;11:e0167370.

11. Metcalf JL, Parfrey LW, Gonzalez A, Lauber CL, Knights D, Ackermann G, Humphrey GC, Gebert MJ, Van Treuren W, Berg-Lyons D, et al. A microbial clock provides an accurate estimate of the postmortem interval in a mouse model system. eLife. 2013;2:e01104.

(22)

22 12. Reddy S, Paul G. Bacterial flora of conjunctiva after death. Int J Ophthalmol.

2013;6:632–636.

13. Harper D. A comparative study of the microbiological contamination of postmortem blood and vitreous humour samples taken for ethanol determination. Forensic Sci Int. 1989;43:37–44.

14. Egger S, Buxbaum A, Georgopoulos M, Scholda C, Vecsei V, Huber-Spitzy V, Georgopoulos A. Bacterial growth in human vitreous humor. Exp Eye Res. 1997;65:791–795.

15. DELTALAB. STUART. Estéril. http://www.deltalab.es/producto/stuart-esteril/; 2016 Accessed 12 July 2018.

16. Guidelines for Environmental Infection Control in Health-Care Facilities. Background F. Environmental Sampling.

https://www.cdc.gov/infectioncontrol/guidelines/environmental/background/samplin g.html; 2015 Accessed 16 July 2018.

17. Lu LJ, Liu J. Human Microbiota and Ophthalmic Disease. Yale J Biol Med. 2016;89:325– 330.

18. Miller D, Iovieno A. The role of microbial flora on the ocular surface. Curr Opin Allergy Clin Immunol. 2009;9:466–470.

19. BD Instructions For Use – Ready-To-Use Plated Media. BD™ Baird-Parker Agar. http://www.bd.com/resource.aspx?IDX=25497; 2011 Accessed 12 July 2018. 20. BD Instructions For Use – Ready-To-Use Plated Media. BD™ Columbia Agar with 5%

Sheep Blood. http://www.bd.com/resource.aspx?IDX=8968; 2013 Accessed 12 July 2018.

21. BD Instructions For Use – Ready-To-Use Plated Media. BD™ Trypticase Soy Agar. http://www.bd.com/resource.aspx?IDX=8989; 2014 Accessed 16 July 2018.

22. BD Instructions For Use – Ready-To-Use Plated Media. BD™ Sabouraud Glucose Agar. http://www.bd.com/resource.aspx?IDX=8984; 2013 Accessed 16 July 2018.

23. American Society for Microbiology. Coagulase test protocol.

http://www.asmscience.org/content/education/protocol/protocol.3220; 2010 Accessed 12 July 2018.

(23)

23 24. American Society for Microbiology. Catalase test protocol.

http://www.asmscience.org/content/education/protocol/protocol.3226; 2010 Accessed 12 July 2018.

25. American Society for Microbiology. Gram Stain Protocols.

http://www.asmscience.org/content/education/protocol/protocol.2886; 2005 Accessed 15 October 2018.

26. Anatomy of Conjunctiva.

http://www.eophtha.com/eophtha/Anatomy/anatomyofconjunctiva.html; 2017 Accessed 27 August 2018.

27. Statistical Software | Sample Size Software | NCSS. https://www.ncss.com/; 2018 Accessed 27 August 2018.

28. Belk A, Xu ZZ, Carter DO, Lynne A, Bucheli S, Knight R, Metcalf JL. Microbiome Data Accurately Predicts the Postmortem Interval Using Random Forest Regression Models. Genes (Basel). 2018;9:104.

29. Javan GT, Finley SJ, Can I, Wilkinson JE, Hanson JD, Tarone AM. Human Thanatomicrobiome Succession and Time Since Death. Sci Rep. 2016;6:29598. 30. DeBruyn JM, Hauther KA. Postmortem succession of gut microbial communities in

deceased human subjects. PeerJ. 2017;5:e3437.

31. Burcham ZM, Hood JA, Pechal JL, Krausz KL, Bose JL, Schmidt CJ, Benbow ME, Jordan HR. Fluorescently labeled bacteria provide insight on post-mortem microbial

transmigration. Forensic Sci Int. 2016;264:63–69.

32. Dickson GC, Poulter RT, Maas EW, Probert PK, Kieser JA. Marine bacterial succession as a potential indicator of postmortem submersion interval. Forensic Sci Int. 2011;209:1– 10.

33. Pechal JL, Crippen TL, Benbow ME, Tarone AM, Dowd S, Tomberlin JK. The potential use of bacterial community succession in forensics as described by high throughput metagenomic sequencing. Int J Legal Med. 2014;128:193–205.

34. Heimesaat MM, Boelke S, Fischer A, Haag LM, Loddenkemper C, Kühl AA, Göbel UB, Bereswill S. Comprehensive postmortem analyses of intestinal microbiota changes and bacterial translocation in human flora associated mice. PLoS One. 2012;7:e40758.

(24)

24 35. Willcox M. Characterization of the normal microbiota of the ocular surface. Exp Eye

Res. 2013;117:99–105.

36. Pathology: Microbiology - Medi-Cal. https://files.medi-cal.ca.gov/; 2008 Accessed 5 September 2018.

37. Boughner LA, Singh P. Microbial Ecology: Where are we now? Postdoc J. 2016;4:3–17. 38. Matthews KR, Roberson J, Gillespie BE, Luther DA, Oliver SP. Identification and

Differentiation of Coagulase-Negative Staphylococcus aureus by Polymerase Chain Reaction. J Food Prot. 1997;60:686–688.

39. Dave SB, Toma HS, Kim SJ. Changes in ocular flora in eyes exposed to ophthalmic antibiotics. Ophthalmology. 2013;120:937–941.

40. Fontes BM, Muccioli C, Príncipe AH, Finamor LP, Höfling-Lima AL, Belfort R. Effect of chronic systemic use of trimethoprim-sulfamethoxazole in the conjunctival bacterial flora of patients with HIV infection. Am J Ophthalmol. 2004;138:678–679.

41. Pechal JL, Schmidt CJ, Jordan HR, Benbow ME. A large-scale survey of the postmortem human microbiome, and its potential to provide insight into the living health condition. Sci Rep. 2018;8:5724.

42. Ozkan J, Nielsen S, Diez-Vives C, Coroneo M, Thomas T, Willcox M. Temporal Stability and Composition of the Ocular Surface Microbiome. Sci Rep. 2017;7:9880.

(25)

25

8. Annex

Annex 1: Raw data sorted by subject regarding number of hours between samplings, between samplings and plating, and between plating and reading.

Raw data of the number of hours between samplings, samplings to plating and plating to reading, per subject. (—: sample not taken.)

Subject Hours t_{to t}₁ 0 Hours t_{to plating}0 Hours t_{to plating}1 Hours plating to reading _{(Incubation time)}

1 — — 195 72 2 — — 194.75 3 — — 194.5 4 — — 194 5 — — 193.75 6 — — 191.75 7 19.5 69 50 72.25 8 21.25 94.75 73.5 72.75 9 18.5 92.5 74.25 ₁₂₀ 10 5.25 73.5 68.25 11 14.25 40.5 26.25 48 12 14.25 26.25 13 14 26.5 14 14.5 26 120 15 13.5 40 26.25 ₇₂ 16 21 23.5 2.5 17 13.75 135 121.25 72 18 19 20 — — 92 71.75 21 13.5 40 26.5 22 ₁₄ _16.5 _2.5 23 24 _13.5 _64.5 ₅₁ 71.5 25 26 13.5 40.5 27 27 28 29 13 15 2 48.5

(26)

26 Annex 2: Raw data sorted by subject regarding number of hours between death and both sampling times, after data preparation. Note that subjects 1 to 6 and subject 20, which were only sampled at t1, were included as well; even though they have only one observation each,

they were useful to add more values for the generation of the graphs.

Raw data of the number of hours between death and both sampling times, per subject. (—: sample not taken.)

Subject Hours death _{to t} 0 Hours death to t1 1 — 10.5 2 — 28.5 3 — 12.75 4 — 6.5 5 — 29 6 — 21 7 28.25 47.75 8 10.75 32 9 29.75 48.25 10 16 21.25 11 10 24.25 12 9.75 24 13 15 29 14 21 35.5 15 10.75 24.25 16 2 23 17 9.5 23 18 10.75 24.25 19 17.75 31.25 20 — 103.5 21 19 32.5 22 8.75 22.75 23 8 22 24 12.5 26 25 12 25.5 26 41 30.5 27 43.5 58 28 15.5 29 29 9.75 22.75 30 6.5 19.5

(27)

27 Annex 3: Raw data sorted by subject regarding number of CFUs obtained for each microorganism species at each sampling time, from each eye, in conjunctiva samples.

Raw data of CFU quantification in conjunctiva samples, per subject. (S. aur.: Staphylococcus aureus, S. epi.: Staphylococcus epidermidis, S. pne.: Streptococcus pneumoniae, S. pyo.: Streptococcus pyogenes, —: sample not taken.)

Subject

Right eye Left eye

t0 t1 t0 t1

S.

aur. S. epi. S. pne. S. pyo. S. aur. S. epi. S. pne. S. pyo. S. aur. S. epi. S. pne. S. pyo. S. aur. S. epi. S. pne. S. pyo.

1 — — — — 0 0 0 0 — — — — 0 0 0 0 2 — — — — 1 0 0 0 — — — — 0 0 0 0 3 — — — — 0 0 0 0 — — — — 11 1 0 2 4 — — — — 0 0 0 0 — — — — 0 0 0 0 5 — — — — 0 0 0 0 — — — — 0 0 0 0 6 — — — — — — — — — — — — — — — — 7 0 0 0 0 7 2 0 10 0 1 0 0 10 2 0 27 8 0 54 0 1 0 0 0 1 0 12 0 0 1 14 0 0 9 0 0 136 0 0 3 0 0 0 0 0 0 0 0 0 12 10 2 0 0 165 1 0 0 142 4 0 0 0 0 0 0 0 11 8 7 8 1 61 174 109 10 8 3 6 0 33 38 140 0 12 0 0 0 0 0 0 0 0 1 4 0 0 0 4 0 0 13 0 0 46 0 0 0 4 0 1 3 10 0 1 0 37 0 14 0 1 61 0 1 2 153 0 0 1 0 0 0 2 29 0 15 5 2 14 0 28 5 0 0 3 3 0 0 3 2 0 0 16 0 6 0 0 2 2 3 3 0 4 0 0 0 10 13 0 17 0 0 0 7 0 0 0 1 0 0 0 0 0 0 0 0 18 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 19 1 1 0 0 0 1 0 1 0 0 0 2 0 0 0 6 20 — — — — 7 33 0 32 — — — — 47 28 0 0 21 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 22 3 27 125 0 3 2 165 0 0 0 173 34 1 0 0 0 23 0 0 139 3 0 1 125 2 0 0 197 0 0 1 114 0 24 13 20 0 22 0 4 0 2 0 0 0 0 0 2 0 0 25 0 0 0 1 0 0 0 0 0 0 0 0 0 27 0 7 26 0 0 178 1 0 1 0 0 0 5 145 0 0 6 162 0 27 0 0 0 0 0 0 0 0 1 2 0 0 0 0 0 1 28 1 17 112 0 1 0 0 4 2 87 121 110 24 34 134 14 29 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 30 0 0 0 2 0 1 0 4 0 0 0 4 0 6 0 20

(28)

28 Annex 4: Significance tests for CFU data presented in Table 5.

H0 = The CFU counts are not higher in t1 (α = 0.05)

Microorganism t-value p-value Reject H0? Higher counts in t1?

Staphylococcus aureus -1.7231 0.0439 Yes Yes

Staphylococcus epidermidis -0.3377 0.3681 No No

Streptococcus pneumoniae 1.084 0.1405 No No

H0 = There is no difference between right and left eye regarding CFU counts (α = 0.05)

Microorganism t-value p-value Reject H0? Difference between right _{and left eye?}

Staphylococcus aureus -0.0545 0.9566 No No

Staphylococcus epidermidis 0.3027 0.7627 No No

Streptococcus pneumoniae 0.1761 0.8606 No No

Annex 5: Complete sets of data (52 observations each, from conjunctiva results only) used for simple and multiple linear regression analyses.

Set of data for Staphylococcus aureus model.

Hours since

death CFUs Sex State of the eyes Manner of death Location of the body Age

28.25 0 Male Open Natural Home ≥75

10.75 0 Male Closed Suicide Road 55-64

29.75 0 Male Mixed Natural Home 55-64

16 0.6 Male Mixed Natural Hospital ≥75

10 1.6 Male Open Accident Home 55-64

9.75 0.1 Male Closed Natural Hospital 55-64

30 0.1 Female Closed Natural Home 65-74

21 0 Male Closed Natural Home 65-74

10.75 0.8 Male Closed Natural Hospital ≥75

2 0 Male Open Natural Road 65-74

9.5 0 Male Closed Natural Home 45-54

17.75 0.1 Male Open Natural Home 55-64

19 0.1 Female Mixed Natural Home 65-74

8.75 0.3 Female Closed Natural Home 65-74

8 0 Male Closed Suicide Home 45-54

12 0 Female Open Natural Home 55-64

41 0 Female Closed Natural Home ≥75

43.5 0.1 Female Closed Suicide Home 35-44

(29)

29

9.75 0 Male Open Natural Home 45-54

6.5 0 Female Closed Indeterminate Hospital 35-44

28.5 0.1 Male Closed Suicide Home 45-54

29 0 Female Closed Accident Hospital 45-54

47.75 1.7 Male Open Natural Home ≥75

32 0.1 Male Closed Suicide Road 55-64

21.25 0.1 Male Mixed Natural Hospital ≥75

24.25 9.4 Male Open Accident Home 55-64

24 0 Male Closed Natural Hospital 55-64

35.5 0.1 Male Closed Natural Home 65-74

23 0.2 Male Open Natural Road 65-74

32.5 0 Female Mixed Natural Home 65-74

22 0 Male Mixed Suicide Home 45-54

26 0 Female Closed Natural Home 35-44

25.5 0 Female Open Natural Home 55-64

30.5 0 Female Closed Natural Home ≥75

58 0 Female Closed Suicide Home 35-44

Set of data for Staphylococcus epidermidis model.

Hours since

10.75 6.6 Male Closed Suicide Road 55-64

16 0 Male Mixed Natural Hospital ≥75

10 1 Male Open Accident Home 55-64

9.75 0.4 Male Closed Natural Hospital 55-64

21 0.2 Male Closed Natural Home 65-74

(30)

30

19 0 Female Mixed Natural Home 65-74

8 0 Male Closed Suicide Home 45-54

12.5 2 Female Closed Natural Home 35-44

41 0.5 Female Closed Natural Home ≥75

43.5 0.2 Female Closed Suicide Home 35-44

28.5 0 Male Closed Suicide Home 45-54

21.25 0 Male Mixed Natural Hospital ≥75

24 0.4 Male Closed Natural Hospital 55-64

32.5 0 Female Mixed Natural Home 65-74

22 0.2 Male Mixed Suicide Home 45-54

25.5 2.7 Female Open Natural Home 55-64

30.5 0.7 Female Closed Natural Home ≥75

19.5 0.7 Female Closed Indeterminate Hospital 35-44

(31)

31

Hours since

10.75 0 Male Closed Suicide Road 55-64

29.75 13.6 Male Mixed Natural Home 55-64

16 0 Male Mixed Natural Hospital ≥75

9.75 0 Male Closed Natural Hospital 55-64

8 33.6 Male Closed Suicide Home 45-54

43.5 0 Female Closed Suicide Home 35-44

32 0 Male Closed Suicide Road 55-64

21.25 0 Male Mixed Natural Hospital ≥75

24.25 0 Male Closed Natural Hospital ≥75

(32)

32

25.5 0 Female Open Natural Home 55-64

30.5 16.2 Female Closed Natural Home ≥75

Set of data for Streptococcus pyogenes model.

Hours since

10.75 0.1 Male Closed Suicide Road 55-64

16 16.5 Male Mixed Natural Hospital ≥75

9.75 0 Male Closed Natural Hospital 55-64

8 0.3 Male Closed Suicide Home 45-54

12 0.1 Female Open Natural Home 55-64

43.5 0 Female Closed Suicide Home 35-44

(33)

33

24.25 1 Male Open Accident Home 55-64

32.5 0.1 Female Mixed Natural Home 65-74

25.5 0.7 Female Open Natural Home 55-64

30.5 0 Female Closed Natural Home ≥75

58 0.1 Female Closed Suicide Home 35-44

Annex 6: Number of subjects for each qualitative variable used for multiple linear regression analyses.

Number of subjects for each sex.

Male Female Total

33 19 52

Number of subjects for each age group.

35-44 45-54 55-64 65-74 ≥75 Total

8 9 16 11 8 52

Number of subjects for each location of the body.

Home Road Hospital Total

39 4 9 52

Number of subjects for each manner of death.

Natural Accident Suicide Indeterminate Total

40 3 7 2 52

Number of subjects for each state of the eyes.

Open Closed Mixed Total

Post-Mortem Interval Estimation by Microbiological Analysis of Conjunctiva and Vitreous Humor

Research Project Report