In the absence of readily available physiological models of human growth, the effects of oestradiol on the human C28/I2 chondrocyte cell line were studied. The classical oestrogen receptors, ERα and ERβ, were shown to be expressed in both murine and human chondrocyte cell lines. Oestradiol and related chemicals, which alter the function of the oestrogen receptors (ER), were exploited to tease out the different functions of each ER in the growth plate. In the absence of foetal bovine serum, oestradiol had no effect on proliferation, differentiation or apoptosis of chondrocyte cells in monolayer culture or on the growth of the foetal metatarsal culture system. In addition, oestradiol did not convey a protective effect on chondrocytes exposed to the pro-inflammatory cytokines, tumour necrosis factor-α (TNF-α) and interleukin-1β (IL-1β) in monolayer culture. However, endogenous oestrogen appears to play an important role in maintaining chondrocyte proliferation in monolayer culture and mineralisation in metatarsal culture as reflected by the inhibitory effects of Faslodex, the non-specific ER antagonist, on chondrocytes and metatarsals in culture. In the presence of methyl-piperidino-pyrazole (MPP), a selective ERα antagonist, and raloxifene, a selective oestrogen receptor modulator with higher ERβ binding affinity, a reduction in chondrocyte proliferation and increase in apoptosis was observed in murine and human chondrocytes. Similarly, a marked reduction in linear growth occurred when foetal murine metatarsals were exposed to MPP and raloxifene in combination. A less marked reduction in growth was observed in MPP-treated metatarsals. These findings suggest that the oestrogen receptors may have opposing actions in the growth plate with ERβ acting like a brake on chondrocyte growth and ERα promoting growth. ERβ may regulate cell proliferation through control of cell cycle modulators affecting G1/S phase transition as MPP and raloxifene in combination reduced cyclin E and p53 levels on Western blot analysis. The aim of the second part of my thesis was to investigate the effect of oral oestrogen on linear growth in girls with primary ovarian insufficiency (POI). A retrospective review of girls with POI treated at a tertiary endocrinology clinic over an 11 year period was performed. As expected the majority of girls with POI had Turner syndrome (TS; 83.7%). Non-TS associated POI was rare and the leading cause was iatrogenic secondary to the effects of total body irradiation for bone marrow transplantation (12.8%). A significant proportion of these girls developed POI after full pubertal development so few cases were available to investigate the effect of oestrogen on growth. The oral oestrogen regime followed in individual patients with TS was highly variable so it was not possible to assess the effects of dose on height velocity or bone maturation in this retrospective audit. However, the second clinical study examined in detail the effect of oestrogen on growth in TS girls who received a standardised course of oral ethinylestradiol for pubertal induction and a standard dose of growth hormone (10 mg/m2/week). These girls participated in a prospective randomised double-blind placebo-controlled multi-centre study of growth promoting treatment in TS. The girls were initially randomised to oxandrolone or placebo at 9 years of age and further randomised to oral ethinylestradiol at 12 or 14 years of age. The results of this study are embargoed until published. The laboratory effects of oestradiol found in this thesis suggest that ERα may stimulate or maintain growth, and ERβ may inhibit growth. The obvious question is how these observations might be involved in the complex relationship between puberty, oestrogen and height velocity in humans. As affinity studies show that the half maximal effective concentration (EC50) of ERα is achieved at slightly lower concentrations of oestradiol than ERβ it is conceivable that the ERα effect could predominate at lower systemic oestradiol concentrations and that ERβ could become more important at higher concentrations for example in later puberty. Alternatively, it is possible that the expression of ERα reduces or ERβ increases in the growth plate after reaching peak height velocity.

The history of newborn resuscitation in the twentieth century presented thus far in the writings of practitioner-historians describes a ‘hands-off’ attitude to newborn care prior to the 1950s. These practioner-historians tend to recount a positivist narrative with the rapid expansion of newborn care after WWII and the eventual logical uptake of endotracheal intubation and positive pressure resuscitation as the most effective method for treating asphyxia neonatorum. This thesis challenges this positivist narrative my examining the resuscitation of the newborn in Britain and America during the interwar period through to the late 1960s. It uncovers a much more complex and non-linear narrative for the development of newborn resuscitation during the twentieth century, uncovering some interesting themes which the practitioner-histories have not addressed. These themes include the interactions between neonatal and fetal physiologists and their research with clinicians and clinical practice, and the role of new groups of clinicians, the paediatricians and anaesthetists, in newborn resuscitation during this period. Many of the practitioner-histories ridicule what they deem to be ‘failed’ resuscitation techniques, seeing them as ‘deveiations’ from the eventual widespread adoption of positive pressure methods. My analysis of both the clinical and scientific debates surrounding both the use of positive pressure methods and some of these ‘failed’ techniques provides a more complex and detailed story. Two techniques in particular, intragastric oxygen and hyperbaric oxygen, provide useful case-studies to reflect on the factors which influenced the development of newborn resuscitation during the twentieth-century. One important factor which is analysed in detail is the formation of a network of scientists and clinicians with a shared interest in the neonate, which emerged during the 1950s. This ‘neonatal network’ has been identified and mapped, and its actions are discussed in detail. The thesis argues that the neonatal network played a fundamental role in directing neonatal research and care during the 1950s and 1960s. The case of newborn resuscitation is used to highlight the interactions of the network members. The history of newborn resuscitation is used to reflect on some wider themes of late-twentieth century medicine. It highlights the divided role of the post-war academic clinician, who was responsible for both clinical care and research. It also illustrates common trends such as the move towards super-specialization in medicine, the increasingly technological nature of medical care and the growing authority of science in the clinic. The research has analysed a variety of sources including the archives of the Ministry of Health, Medical Research Council, Scottish Home and Health Department, the Neonatal Society and National Birthday Trust Fund. Oral histories have been used to map the relationships forged between key actors. A variety of published resources, including journal articles, textbooks and conference proceedings, have also been studied to track both the changes in neonatal care and changes in the physiological understanding of the newborn.

The use of antiretroviral therapy (ART) during pregnancy and lactation has significantly reduced the rate of mother-to-child transmission (MTCT) of HIV. However, pregnancy is known to affect the pharmacokinetics of many drugs, including key antiretroviral (ARV) drugs. In addition, ARV use during lactation raises questions about unintended exposure of breastfed infants to maternal drugs through breast milk. For drugs with significant genetic contribution to observed pharmacokinetic variability, we hypothesised that polymorphisms in drug disposition genes may accentuate or attenuate pregnancy-induced changes and/or breastfed infants’ exposure. HIV positive pregnant women and nursing mothers taking efavirenz (EFV)- or nevirapine (NVP)-based ART were recruited from three hospitals in Benue State, Nigeria. A novel strategy involving a preliminary pharmacogenetic association study was used to investigate the magnitude of pregnancy-induced changes in EFV and NVP pharmacokinetics in women stratified by single nucleotide polymorphisms (SNPs) in disposition genes. EFV apparent clearance (CL/F) was higher and AUC0-24, Cmax and Cmin were significantly lower in pregnant compared with postpartum women. When stratified based on the SNP with the highest predictive power, pregnant women with CYP2B6 516GG genotype were especially at risk. In the NVP cohort, exposure was also significantly lower in pregnant compared with postpartum women. When stratified based on composite CYP2B6 516G > T and 983T > C genotypes, Cmin was below target in most patients with combined CYP2B6 516GG and 983TT during pregnancy and postpartum. Cmin was below target in at least 50% of pregnant women with one or two variant alleles, compared with 0% in postpartum women. The intensive pharmacokinetics of EFV and NVP in breast milk and pharmacogenetic predictors were described for the first time. Breast milk pharmacokinetic parameters of EFV in breast milk differed significantly between patient groups stratified by CYP2B6 516G > T. The median time averaged milk-to-plasma concentration (M/P) ratio was 1.10 (range: 0.57-1.71) and the paediatric dose weight-adjusted exposure index was 4.05% (1.08-13.8). The resulting infant plasma concentration was influenced by CYP2B6 516G > T, highest up to 8 days of age at 1590 ng/mL (190-4631) and decreased by about 90% in the age stratum 9 days to 3 months. NVP AUC0-12, Cmax and Cmin in breast milk were significantly lower in patients with composite CYP2B6 516GG/983TT than those with at least one variant allele. The M/P ratio was 0.88 (0.74-1.2) and paediatric dose weight-adjusted exposure index was 3.64% (1.99-9.88). Infant plasma concentration differed significantly based on CYP2B6 516G > T/983T > C and CYP3A4 20230G > A (*1G), highest in those exposed through both breast milk and post-exposure prophylaxis compared with either alone. A breastfeeding physiologically-based pharmacokinetic (PBPK) model to predict infant exposure to maternal drugs through breast milk was developed and validated, with over 90% of all individual observed data points within the predictive interval. This thesis presents details about five different studies where these findings were observed. Their clinical implications in the context of current knowledge and practice were also explored.

Metabolite concentrations are fundamental biomarkers of disease. With increasing interest in personalised medicine, this work assessed the accuracy of non-invasive metabolite quantification with magnetic resonance spectroscopy (MRS) using a combination of simulations, phantom and in vivo data. No optimal echo time (TE) was found for measuring a range of key metabolites with quantification accuracy generally influenced more by data quality than TE choice. The T2 relaxation times of water and metabolites with MRS dominated by a singlet could be estimated using 2 TEs and were found to be significantly different in paediatric brain tumours compared with normal brain, varying between tumour types. The T2 relaxation times of paediatric brain tumours were significantly shorter at 3T compared with 1.5T. Metabolite concentrations for individual patients were most affected by changes in the T2 relaxation time of water which is quick to measure. A clinical JPRESS protocol was developed which aids assignment of overlapping metabolites using changes of MRS with TE. Overall, measurement of MRS with a short TE reduces inaccuracies associated with variability in metabolite T2 and does not tend to lead to worse quantification of overlapping resonances. Further improvements in concentration accuracy can be obtained by measuring case-specific water T2.

Profiles of sleep disturbance and sleep quality of children with the specific neurodevelopmental disorders of Smith-Magenis syndrome (SMS), Angelman syndrome (AS), autism spectrum disorder (ASD) and tuberous sclerosis complex (TSC) and the relationships between behavioural and health characteristics, age and sleep were described in these groups. Interview data demonstrated that children with AS's sleep disturbance had a negative impact on both parents and children. A homogeneous sleep disturbance profile of severe night waking and early morning waking affected over 70% of children with SMS but more heterogeneous profiles were found for children with AS, TSC and ASD using cross-group questionnaire data comparisons and when compared to typically developing (TD) children. A heightened risk of sleep-related breathing disorders was identified for children with AS and SMS. Compared to TD children, children with SMS had significantly earlier morning wake times and children with AS and SMS had significantly earlier bedtimes according to actigraphy and sleep diary data. Increased daytime sleepiness in children with SMS was associated with increased overactivity and impulsivity. This thesis includes the largest samples of actigraphy data for children with SMS and AS to date. The importance of aetiology of intellectual disability in the profiling of sleep disturbance was evidenced. Areas for further assessment and intervention include sleep-related breathing disorders for children with AS and SMS and individualized assessment of circadian rhythm disorders for both groups.

Neurodevelopmental conditions like Attention Deficit Hyperactivity Disorder (ADHD) and Autism Spectrum Disorder (ASD) impact a significant number of children and adults worldwide. Currently, the means of diagnosing of such conditions is carried out by experts, who employ standard questionnaires and look for certain behavioural markers through manual observation. Such methods are not only subjective, difficult to repeat, and costly but also extremely time consuming. However, with the recent surge of research into automatic facial behaviour analysis and it's varied applications, it could prove to be a potential way of tackling these diagnostic difficulties. Automatic facial expression recognition is one of the core components of this field but it has always been challenging to do it accurately in an unconstrained environment. This thesis presents a dynamic deep learning framework for robust automatic facial expression recognition. It also proposes an approach to apply this method for facial behaviour analysis which can help in the diagnosis of conditions like ADHD and ASD. The proposed facial expression algorithm uses a deep Convolutional Neural Networks (CNN) to learn models of facial Action Units (AU). It attempts to model three main distinguishing features of AUs: shape, appearance and short term dynamics, jointly in a CNN. The appearance is modelled through local image regions relevant to each AU, shape is encoded using binary masks computed from automatically detected facial landmarks and dynamics is encoded by using a short sequence of image as input to CNN. In addition, the method also employs Bi-directional Long Short Memory (BLSTM) recurrent neural networks for modelling long term dynamics. The proposed approach is evaluated on a number of databases showing state-of-the-art performance for both AU detection and intensity estimation tasks. The AU intensities estimated using this approach along with other 3D face tracking data, are used for encoding facial behaviour. The encoded facial behaviour is applied for learning models which can help in detection of ADHD and ASD. This approach was evaluated on the KOMAA database which was specially collected for this purpose. Experimental results show that facial behaviour encoded in this way provide a high discriminative power for classification of people with these conditions. It is shown that the proposed system is a potentially useful, objective and time saving contribution to the clinical diagnosis of ADHD and ASD.

Background and aims: Leukaemia is a cancer of the blood and is the most common type of childhood cancer with almost 500 new cases every year in the UK. There is a vast amount of research exploring the experiences of parents of children with cancer, however, there is less research focussed on parents whose children are in the maintenance phase of Acute Lymphoblastic Leukaemia (ALL) treatment. This treatment phase usually occurs in the first or second year following diagnosis but can start much earlier. The maintenance phase still involves active treatment but with less visits to and stays in hospital. At this stage, parents have been through the most intense segment of the treatment regimen and may have encountered various Khalighyprotocol may therefore have valuable advice to provide to other parents regarding coping and useful supports. This study aimed to explore the coping style and patterns of accessing support in mothers and fathers who have a child diagnosed with ALL and who are in the maintenance phase of treatment. The aim was also to explore whether there were differences between mothers and fathers with regards to coping and support. Methods: Three mothers and two fathers were interviewed separately using a semi structured interview. These interviews were then transcribed and analysed using interpretative phenomenological analysis (IPA). Results: Four main themes were identified through interviews with parents: the parental role; internal coping strategies; external coping strategies; and looking to the future. Conclusions: Parents were found to use a variety of coping strategies and accessed various sources of support to help them to deal with their child’s ALL diagnosis. These coping strategies could be shared with parents who are new to the haematology service. There were however, significant difficulties with recruitment which means it was not possible to compare mothers and fathers in terms of coping style and types of support accessed. This highlights that the recruitment strategy needs revising if further research is to be conducted in this area.

Background: Reducing neonatal mortality rates [NMR] (deaths/1,000 live births within 28 days of delivery) is a key global health goal. Using comparable data from Ghana (West Africa) and Scotland, I investigated NMR, specific causes of death and risk factors in the two countries. By identifying the main causes of excess mortality in Ghana and where they occur, it is hoped more effective strategies can be developed. Methods: This thesis used a multilevel mixed methods study design. Data on live births were obtained from three Health and Demographic Surveillance Systems (HDSS) in the north, middle and south of Ghana respectively: Navrongo (2004-12; 17,016 live births, 320 deaths); Kintampo (2005-10; 11,207 live births, 140 deaths); Dodowa (2006-14; 21,647 live births, 135 deaths). Comparable Scottish data were obtained from the Information Services Division (1992 to 2015; 1,278,846 live births, 2,783 deaths). Each dataset was analysed by neonatal death (dead/alive), using univariate and multivariable logistic regression. The multivariable analyses adjusted for maternal demographic and obstetric characteristics. Missing data were analysed using multiple imputation techniques. Data analyses were complemented by a researcher-developed questionnaire survey of 71 maternity care providers in the three regions of Ghana followed by face-to-face in-depth interviews with 48 maternity care providers who had experience of prematurity, birth asphyxia, neonatal infection and neonatal death. Results: The NMRs in the three HDSS were: Navrongo: 18.8; Kintampo: 12.5; and Dodowa 6.2 and in Scotland it was 2.2; the NMR in both countries is reducing. More than 99% of the neonatal deaths in Scotland occurred in the first week compared to 74% in Ghana. The leading causes of neonatal deaths (NMR) in Ghana were infection (4.3), asphyxia (3.7) and prematurity (2.2). In Scotland, they were congenital malformations (0.6), asphyxia (0.4) and prematurity (0.3). Only 88 deaths (0.07) of neonatal deaths in Scotland were due to infection. Ninety-eight percent of babies born in Scotland were born in a health facility compared to 60% of babies born in Ghana (hospital: 38.1%; clinic: 21.1%). In Ghana, babies born in hospitals had a higher risk of neonatal mortality compared to those born at home (NMR-hospital: 15.6; clinic: 7.1; home: 11.8). Most of the neonatal deaths in Ghana occurred at home (54%); there were more deaths among babies who were born in a hospital but died at home (hosp/home) compared to those born at home but died in a hospital (home/hosp). Asphyxia was the leading cause of death among hosp/hosp, and infection was the leading cause of death among hosp/home, home/home and home/hosp. Neonatal mortality in Ghana was largely influenced by where mothers sought maternity service, or the type of personnel who provided maternity care service. Mothers and babies who were cared for in hospitals by doctors and midwives received relatively better care and proper management of birth complications. Those who were cared for in clinics received basic delivery services and management of uncomplicated asphyxia. Mothers and babies who were cared for at home by traditional birth attendants (TBA) received poor care and poor management of neonatal illnesses based on traditional approaches which increased the risk of death. Women’s maternity choices were influenced by wider societal factors including prominent cultural values, family hierarchical structures and the cost of maternity services, and individual/ family factors including place of residence and availability of transport and beliefs about the cause of disease. Conclusion: There is considerable opportunity for reducing NMR in Ghana, especially deaths due to asphyxia and infections. Most uncomplicated deliveries should be performed by midwives in community clinics. The number of community maternity clinics should gradually be increased to enable home deliveries by TBAs to be phased out. Facilities should be improved for delivery and postnatal care in hospitals and the proportion of sick babies managed by health care workers trained in their care should be increased. Regular postnatal checks in the community by trained staff should be standard.

Amongst the paediatric haematology/oncology patients attending the Royal Hospital for Sick Children, Glasgow, episodes of viridans streptococcal bacteraemia increased from 12% of al microbiologically documented bacteraemias (i.e. 10/81) in 1993 to 22% (18/83) in 1994. During the first year of this project (which started in December 1994), ITU support was required following the development of viridans streptococcal bacteraemia on 6 occasions, and of these, there were two fatalities. The overall aim of this study was to improve the management of this infection and to explore preventative strategies. Three different approaches were adopted: (1) an extensive epidemiological analysis was undertaken - to include all episodes of viridans streptococcal bacteraemia from December 1994 to December 2000. (2) Phenotypic, followed by genotypic analyses of isolates of viridans streptococci from mouth swabs and blood cultures were carried out to determine whether the mouth was in fact the source of organisms responsible for this infection. (3) Extensive antibiotic susceptibility studies were performed on all isolates of viridans streptococci from blood culture. In total, 69 episodes of viridans streptococcal bacteraemia occurred in 54 children. The infection was more often associated with patients with haematological malignancy (particularly AML), than those with solid tumours, and in the majority (84%) of episodes, the patients suffered from chemotherapy-induced mucositis or other forms of oral compromise. Forty-eight episodes of infection (70% of total) responded well to antimicrobial therapy with no evidence of additional clinical complications. However, in 21 cases (30% of total), pulmonary complications developed, with 8 of these requiring mechanical ventilation and supplemental oxygen. Five of these 8 cases also developed septic shock. S. oralis was the species most commonly isolated from blood culture (63% of total isolates of viridans streptococci) and S. mitis represented 25% of total isolates. Polymicrobial bloodstream infection occurred in 23% of episodes.

Background: The present research is the first multi-modal study of language deficits in long-term PFT survivors to date. In addition to a detailed neuropsychological assessment of language, this thesis investigated neural correlates of language processing by employing functional (fMRI) and diffusion (DTI) magnetic resonance imaging techniques. Method: Twenty-one PFT survivors, aged 16-21, and twenty-two matched healthy volunteers completed a series of neuropsychological assessments, task-based language fMRI study (targeting semantic retrieval and speech articulation), and a DTI study. The patients’ clinical profiles and oro-motor functioning were also analysed. Results: Patients demonstrated significantly poorer performance in semantic content, expressive and receptive skills, verbal memory, reading and writing, visuo-motor coordination and non-verbal intelligence, with elevated internal variability of the linguistic profiles, when compared to controls. Semantically-related language skills contributed more to the between-group differences than non-verbal cognitive skills. A significant association between language and non-verbal cognitive abilities in both patients and controls was unconfounded by the age and disease. On fMRI, patients lacked metabolic response in the pre-central and post-central gyri during semantic retrieval. Broca’s, Wernike’s and Geschwind’s areas responded similarly in both hemispheres, with no significant differences between the groups. In all participants Language Content Index predicted the BOLD response on the border of the left lateral occipital cortex and angular gyrus. Expressive Language Index predicted BOLD response in the right frontal pole, paracingulate gyrus, superior frontal gyrus, and middle frontal gyrus. In healthy controls, articulation of speech was associated with activation in the Crus I and Crus II of the right cerebellar hemisphere. Semantic load triggered activation in the Crus VI and VIIb of the vermis, as well as right lobules V and VI of the cerebellum. DTI revealed a global decrease in the fractional anisotropy and increase in the diffusivity scalars in patients, compared to controls, but not different between those patients that received and did not receive radiotherapy. Patients also demonstrated significant reduction in FA index in the bilateral arcuate fasciculus and increased diffusivity in the bilateral SCP. The FA index in the segments of the left-hemispheric cortico-spinal tract, anterior thalamic radiation, superior longitudinal fasciculus and inferior fronto-occipital fasciculus, positively predicted Language Content Index score in patients. Clinical profiles analysis indicated that younger age at diagnosis, radiotherapy treatment and longer duration of mutism (if present) were associated with the poorest language outcomes. Patients with the longest recovery time demonstrated the best manual dexterity abilities. 38% of the patients that met the criteria for a diagnosable language disorder also had reduced oro-motor functioning and reduced FA within left arcuate fasciculus, compared to the remaining patients. Conclusions: Despite the fact that PFTs do not directly impact cortical language-associated areas, patients, particularly those treated with radiotherapy at a younger age, demonstrate deficits in all aspects of language processing. Semantic processing difficulties in PFT survivors are underpinned by the diminished cortical metabolic response during associated task performance, and microstructural changes in the left-hemispheric white matter. Tumours affecting the right cerebellar hemisphere may further predispose patients to developing difficulties in accessing language semantics.