Ce travail avait pour objectif d'étudier l'intérêt et les limites de l'analyse de deux matrices alternatives que sont la bile et l'humeur vitrée (HV), en toxicologie médicolégale. Pour chacune des deux matrices, une revue de la littérature visait à investiguer les connaissances utiles à leur application en toxicologie médicolégale. Une place importante de ces revues est réservée à l'anatomie et la physiologie de l'HV et du système biliaire ainsi qu'aux mécanismes de distribution des xénobiotiques dans ces matrices. La partie expérimentale décrit trois études: deux menées sur populations autopsiques et une associant expérimentations animales et études de populations autopsiques. Les deux premières ont permis de proposer des outils statistiques d'interprétation des concentrations de méprobamate mesurées dans ces matrices. Ils peuvent être utilisés dans diverses situations à la place ou en complément de l'interprétation des concentrations sanguines: cadavre exsangue, putréfaction avancée du corps, redistribution post mortem des xénobiotiques… La troisième étude concernait six molécules (diazépam, citalopram, cyamémazine, morphine, caféine et méprobamate). Les molécules détectées dans le sang l'étaient systématiquement dans l'HV et la bile aussi bien dans les prélèvements des populations autopsiques que ceux issus des expérimentations animales. Les concentrations vitréennes chez l'animal et chez l'homme étaient systématiquement corrélées aux concentrations sanguines, exceptées celles de cyamémazine et de citalopram chez l'homme. Pour la bile, une corrélation significative était observée pour le méprobamate et la caféine chez l'homme et l'animal. Il ressort de ces résultats, que l'analyse de l'HV et de la bile permettent de disposer d'informations relatives à la nature des molécules absorbées et à leur rôle dans la survenue du décès / The present study sought to assess the interest and limitations of analyzing two alternative matrices, bile and vitreous humor (VH), in forensic toxicology. For each matrix, a literature review established the state of knowledge relating to their forensic application. The review placed special focus on the anatomy and physiology of VH and the biliary system and the mechanisms of xenobiotic distribution within the specific matrix. The experimental sections describe three studies: two performed on autopsy populations, and one associating autopsy populations to an animal model. The first two studies resulted in statistical tools for interpreting meprobamate concentrations in these matrices, which can be used as alternatives or complements to blood concentrations in various situations: exsanguination, advanced putrefaction, postmortem xenobiotic redistribution, etc. The third study focused on 6 molecules: diazepam, citalopram, cyamemazine, morphine, caffeine and meprobamate. Molecules detected in blood were also systematically detected in VH and bile samples from both the autopsy and animal populations. Animal VH and blood levels showed systematic correlation. In autopsy samples, cyamemazine and citalopram showed no such correlation. In bile, significant correlations with blood concentrations were found for meprobamate and caffeine in both the autopsy and animal populations. This study confirmed the interest of postmortem analysis of bile and VH. Results show that analyzing bile and VH sheds light on drugs intake and on their implication in cause of death

Identification of human or nonhuman skeletal remains is important in assisting the police and law enforcement officers for the investigation of forensic cases. Identification of bone can be difficult, especially in fragmented remains. It has been reported that 25 to 30% of medicolegal cases, which involved nonhuman skeletal remains have been mistaken for human. In such cases, histomorphometric method was used to identify human and nonhuman skeletal remains. However, literature has shown that histomorphometric data for human and nonhuman bone were insufficient. Additionally, age estimation in bone may help in the identification of human individual, which can be done by using a histomorphometric method. Age estimation is based on bone remodeling process, where microstructural parameters have strong correlations with age. Literature showed that age estimation has been done on the American and European populations. However, little work has been done in the Asian population. The aims of this project were thus, to identify human and nonhuman bone, and to estimate age in human bones by using histomorphometric analysis. In this project, 64 human bones and 65 animal bones were collected from the mortuary of the Universiti Kebangsaan Malaysia Medical Centre and the Zoos in Malaysia, respectively. A standard bone preparation was used to prepare human and nonhuman bone thin sections for histomorphometric assessment. Assessments were made on the microstructural parameters such as cortical thickness, medullary cavity diameter, osteon count, osteon diameter, osteon area, osteon perimeter, Haversian canal diameter, Haversian canal area, Haversian canal perimeter, and Haversian lamella count per osteon by using image analysis, and viewed under a transmitted light microscope. The microstructural measurements showed significant differences between human and nonhuman samples. The discriminant functions showed correct classification rates for 81.4% of cases, and the accuracy of identification was 96.9% for human and 66.2% for animal. Human age estimation showed a standard error of estimate of 10.41 years, comparable with those in the literature. This study project offers distinct advantages over currently available histomorphometric methods for human and nonhuman identification and human age estimation. This will have significant implications in the assessment of fragmentary skeletal and forensic population samples for identification purposes.

This thesis describes studies on the analysis of arsenic (As) in human biological samples, mainly urine but also hair and fingernails using inductively coupled plasma mass spectrometry (ICP-MS) and graphite furnace atomic absorption spectrometry (GF-AAS). The relationship between ethnicity and arsenic metabolism was investigated for the first time for a population in the United Kingdom. This investigation has been carried out through comparative analysis of arsenic in human urine, hair and fingernails in volunteers from three different ethnic groups (Whites, Asians and Somali Black-Africans) who are only exposed to background levels of arsenic. Results obtained with 63 volunteers showed ethnic differences in urinary arsenic excretion as well as differences in arsenic levels in fingernail samples. The averages of total arsenic levels for the Somali Black-Africans (urine 7.2 µg/g creatinine; fingernails 723 µg/kg) are significantly (P< 0.05) different from both the Asians (urine 20.6 µg/g creatinine; fingernails 153.9 µg/kg) and Whites (urine 24.5 µg/g creatinine; fingernails 177.0 µg/kg). The Somali group also shows a higher percentage (50%) of dimethylarsinate (DMA) and a lower percentage (48%) of arsenobetaine (AB), compared to Asians (16% DMA and 83% AB) and Whites (22% DMA and 77% AB). The effect of fasting on urinary arsenic species distribution was also investigated by monitoring urine samples from 29 Ramadan fasting volunteers, with each volunteer providing a sample at the beginning (RF1) and at the end (RF2) of an approximately 12 hours fast. The results obtained showed the frequency of MA detection for RF2 was 12 and 2-fold higher than for the non-fasting and RF1 groups, respectively. This suggests fasting may alter the pattern of arsenic metabolism and excretion. However, there was no significant difference (P> 0.05) in the average of total level of arsenic for RF1 (18.3 µg/g creatinine) and RF2 (17.7 µg/g creatinine). A relationship between excretion of arsenic and selenium in individuals exposed to background levels of arsenic and selenium was investigated through analysis of urine samples from 93 volunteers from Leicester, UK. A positive correlation between arsenic and selenium was found and the As:Se ratio was 0.7 ± 0.4. The intra-individual variation of As:Se ratio does not alter significantly over time, as determined by monitoring urine samples from a volunteer over a period of one year. Furthermore, within a single day, with urine samples collected at the beginning and after a 12-hour fast, the As:Se ratio was found to be similar (0.7 ± 0.5). These findings suggest a close relationship between these two metalloids, the biological significance of which needs to be explored in the future.

Drug facilitated assault (DFA) is an increasing problem in the UK. The crime often occurs through the surreptitious administration of a drug into a victims drink, rendering the victim unable to resist the assault. The detection of these drugs in a biological specimen from the victim is one of the most challenging facets of forensic chemistry. Drug concentrations can be very low, as often only a single dose is administered, and the pharmacodynamics of commonly employed drugs further hinders the testing process. The research presented in this work shows the development of several new assays for the detection of flunitrazepam, gamma-hydroxybutyrate (GHB) and ethyl glucuronide (EtG) in a variety of biological matrices. New methods of drug testing in blood and urine are demonstrated, as well as interesting developments in the field of hair testing. Using hair to detect drug exposure allows a much wider window of detection than the more traditional matrices of blood and urine. New methods are presented in this work using gas chromatography-tandem mass spectrometry (GCMS/MS) to detect drugs in hair. Validation data is presented along with the results of authentic DFA testing. All aspects of the drug testing procedure have been evaluated, from new extraction techniques utilising water instead of solvents, to novel clean up stages involving the unique combination of SFE and SPME. Several confirmation techniques are explored including single quadrupole, triple quadrupole and ion trap mass spectrometry. In addition to developing assays for DFA cases, the versatility of this type of analytical chemistry is explored in two population studies. The first study evaluates alcohol consumption between two groups; drugs users and non drug users in medico-legal cases. There is an anecdotal belief amongst drug clinic staff that alcohol use is lower in drugs users than it is in non drug users. This study presents the first scientific confirmation of this belief through EtG (an alcohol metabolite) testing in hair of the two groups. The second study investigates whether there is a correlation between EtG and cocaethylene (a metabolite of cocaine only produced in the presence of alcohol) in cocaine users. Results f this study suggest that there is no positive correlation between the two compounds. The research presented in this thesis aims to further the analytical science surrounding FA investigation and provide accurate, sensitive and reliable methodology for drug esting in blood, urine and hair.