MMed, Paediatrics, University of the Witwatersrand, Faculty of Health Sciences / Background: To reduce the risks and prevent progression of diabetic complications average blood glucose and glucose variability need to be kept as close to the non diabetic range as possible. Post prandial glucose excursions contribute significantly to average blood glucose and to glycemic variability. Dietary carbohydrate is the primary determinant of meal related blood glucose excursions. Carbohydrate counting is a method of insulin dosing that matches carbohydrate load to insulin dose using a fixed ratio. Many patients and current insulin pumps, calculate insulin delivery for meals based upon a linear carbohydrate to insulin relationship. Hypothesis: A non-linear relationship exists between the amount of carbohydrate consumed and the insulin required to cover it. Rather, an exponential increase in insulin is needed to cover an increasing load of carbohydrate. Aim: To document blood glucose exposure, as measured by AUC, in response to increasing carbohydrate loads on fixed carbohydrate to insulin ratios. Sample and Methods: 5 Type-1 diabetic adolescents and young adults on insulin pump therapy with good control were recruited. Morning basal rates and carbohydrate to insulin ratios were optimized prior to the study start. A Medtronic glucose sensor was worn by each participant for 5 days on which standardized meals of increasing carbohydrate content were consumed. After the 5 days the glucose sensors were downloaded and the glucose area under the curve was analyzed for each carbohydrate load for each participant. Results: Only subjects with 5 days of complete recordings covering the test meals were included for analysis, resulting in 5 complete analyses. Sensor failure and hypoglycaemic v episodes prior to test meals accounted for failures. Increasing carbohydrate loads on a fixed carbohydrate to insulin ratio resulted in increasing glucose area under the curve (AUC).The log (Average AUC) was linear confirming that this relationship is exponential. An Analysis of Covariance performed on the log (AUC) data confirmed a highly significant exponential relationship (p<0.0001) although no significant differences were found between the profiles of the 5 individuals. Late post prandial hypoglycaemia followed carbohydrate loads greater than 60 grams and this was often followed by rebound hyperglycaemia that lasted more than 6 hours. Conclusion: A non linear relationship exists between carbohydrates consumed and the insulin required to cover them when using premeal bolus insulin. This has implications for control of postprandial blood sugars, especially when consuming large carbohydrate loads. Because of the late post prandial hypoglycaemia that follows the larger doses of insulin used with larger amounts of carbohydrate it is not possible to simply increase the amount of the insulin bolus using an exponential formula. Further studies need to be done looking at the optimal ratios of insulin needed for increasing carbohydrate loads, the duration and type of boluses needed to cover these high carbohydrate loads and the possibility of changing the linear equation used in current insulin pumps to one that would better cover the increase in post prandial glucose load with large carbohydrate meals.

A report on a study submitted to the Department of Speech Pathology, Faculty of Arts. University of The Witwatersrand, Johannesburg in Partial fulfilment of the requirements for the Degree of Masters of Arts in Speech Pathology, Johannesburg, 1998. / This study was designed to examine the incidence of dysphagia in a group of fifty diabetic patients. The interrelationship between dysphagia and other complications of diabetes was examined. These included neuropathy, orthostatic dysfunction, renal dysfunction and respiratory disorders. Issues such as type of diabetes, patient age and age of onset were addressed. The utility and sensitivity of an interview schedule devised by the researcher was critically evaluated. Data were obtained through the administration of a standardized open-ended interview schedule coupled with an examination of the patients' hospital files. Results were interpreted and tabulated by the researcher. Results revealed that a significant proportion of diabetic patients are experiencing swallowing difficulties within all three phases of the swallow process. It is postulated that the cause of the dysphagia may be due to severe autonomic neuropathy and consequent vagal denervation. The highest incidence of dysphagia appeared to be amongst Type I diabetics who developed diabetes before the age of forty. The interview schedule was thus able to detect those patients experiencing dysphagia. However it was found to be unsuccessful in determining the severity of the problem. Clinical implications for both the Speech Therapist and the Medical team are discussed. Suggestions for future research are put forward.

MSc Pharmacology, Faculty of Health Sciences, University of the Witwatersrand / Natural remedies from medicinal plants are considered to be effective and safe alternative treatment for diabetes mellitus. The aim of this study was to demonstrate the hypoglycaemic and antidiabetic activity of the aqueous extract of Icacina tracantha (tuber) (fam Icacinaceae)Ananas cosmos (fam. Bromeliaceae)and Uraria picta (leaves) (fam leguminosae) on an animal model of insulin resistance, a condition which predisposes to type 2 diabetes. The plants have a long history of use as anti-diabetic agents in western Nigeria. Method: 120 male Sprague-Dawley rats were assigned into two major groups. One group was fed on normal rat chow with the other group fed on a high calorie diet for four months a period sufficient for the animals to be fed to attain insulin resistance. The animals were then randomly assigned into different groups (each containing 6 male rats). The plant crude extracts were made by weighing specific dried quantities of each plant, boiling in distilled water for about 2 hours, cooling overnight and separating solid from liquid by filtration. The solution was then poured into preweighed 250 ml beakers and allowed to dry in an oven at a temperature of 60oC. The dried, crude extracts were then weighed out and required doses prepared from the extracts. A non-treated group of animals was used as the control. The mixed dose of extract was administered at 300 mg/kg. Over a 3 week period, all the animals were orally dosed with the different doses of plant extracts daily while metformin was administered through the animals’ drinking water, blood was collected from the tail vein of each rat prior to dosing and thereafter weekly, plasma was preserved and 6 analysed for glucose, insulin, free fatty acid concentrations and calculation of HOMA values to determine insulin sensitivity. During this period, the animals were weighed weekly and food intake was measured every three days. An oral glucose tolerance test (OGTT) was performed after the dosing period and fasting, 0, 30, 60 and 120 minute blood samples were taken and assayed for glucose concentration. Animals were terminated and blood analysed. Statistical analysis: The results were tabulated as mean ± standard deviation and percentage median ± quartile range. The statistical analysis for other parameters was carried out via ANOVA (between groups) and Student’s paired T test (within groups). Only data from percentage median and quartile range was used because of the observed variation in glucose concentration between groups even at baseline values. Statistica software (StatSoft, Tulsa, OK, USA) was used for the analysis. Results: All plant extracts in the study showed differing concentration of significant difference in their effect on the plasma glucose, insulin and free fatty acid concentrations in the rat. The most significant effect was observed on the insulin concentration in the normal rat chow and high calorie diet fed animals. The plant extracts were observed to improve insulin sensitivity in most of the groups. This effect was more significant in the normal rat chow fed rats. The effect of the plant extracts on the weight, food consumed glucose and free fatty acid was minimal and in most of the groups was not significant. Conclusion: In conclusion, the results obtained suggest that the plant extracts may be used to improve insulin resistance in the management of diabetes mellitus.

A research report submitted to the Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, in fulfilment of the requirements for the degree of Master of Medicine in the branch of Internal Medicine, Johannesburg, 2014 / Objective: To evaluate and compare any differences in control, i.e. HbA1C, total cholesterol, BP and BMI in a single group of Type-2 diabetes patients during two time periods, i.e. before and after specialist-supervised care. In addition, to describe differences in the use of anti-platelet and statin therapy for primary cardiovascular prophylaxis. Methods: Patients were recruited from the Diabetes Clinic at Chris Hani Baragwanath Academic Hospital (CHBAH) and the audits of two separate time periods were conducted. The first audit recorded standard of care delivered by registrars from January 2005 to December 2007. The second audit recorded care after the introduction of specialist-supervised care from September 2009 to September 2012. The patients were all insulin-requiring and were required to be seen for at least 24 months during both audit periods. The first recorded HbA1C in (i) 2005 and (ii) from September 2009 triggered the inception of a patient’s assessment periods. Data for at least 80% of parameters had to be available for a patient to be included in the audit. Results: This study showed significant differences using ANCOVA comparing final values for each audit after adjustment for their respective baseline values in respect of HbA1C (p<0.000), SBP (p<0.012) and BMI (p<0.001) after the implementation of an endocrinologist-supervised clinic. The percentage of patients reaching guideline targets, and the use of aspirin and statins, improved as well. Conclusion: This study showed a difference in the level of care delivered by the endocrinologist-supervised clinic as opposed to one which was led by registrars. However, other factors may have contributed to the outcomes, most notably that the consultation time with each patient was longer after the introduction of expert supervision in 2009.

A research report submitted to the Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, in partial fulfillment of the requirements for the degree of Master of Medicine in the branch of Internal Medicine Johannesburg, 2014 / Background Diabetes is a global epidemic. The International Diabetes Federation estimates that there are at least 285 million diabetics worldwide and this is estimated to grow to over 440 million by 2030 1 . A study was conducted at the Helen Joseph Hospital Diabetic clinic in an attempt to identify predictors of glycaemic control and to compare the level of care to the 2012 Society for Endocrinology, Metabolism and Diabetes of South Africa (SEMDSA) guidelines. Methods Patients were recruited from the Helen Joseph Hospital Diabetic clinic. To be included the patient had to be part of the coloured (mixed race) community, be willing to give informed consent, be older than 18 years, have an HBA1C taken within 6 months, have a diagnosis of Type 2 diabetes mellitus and be a clinic attendee for at least 1 year. Pregnant patients, Type 1 diabetic patients, patients with a psychotic disorder or aphasia were excluded. Data collection consisted of face-to-face interviews, review of treatment, medication knowledge evaluation, a short examination and collection of recent blood results. Statistical analysis was done by stratifying patients into two groups by using the mean HBA1C. Variables with a p < 0.1 from this analysis were used in a logistic regression model. In addition, the correlation between continuous variables were tested. A comparison was made between the level of care and the 2012 SEMDSA guidelines. v Results A total of 100 patients were recruited into the study. The mean age was 62.8 years with mean duration of diabetes of 15.8 and clinic attendance of 10.9 years. The group had very poor education level and the median income of R1200 per month was also low. The mean HBA1C was found to be 9.74%, well above the target recommended by SEMDSA. Knowledge of diabetes with respect to management and complications was very poor. Age > 50 years (OR 0.372 CI 0.06-2.26), estimated glomerular filtration rate ≥ 60 ml/min/1.73m2 (OR 0.90 CI 0.25-3.27), experiencing a microvascular complication (OR 0.73 CI 0.11-5.07) or any other diabetic complication (OR 0.56 CI 0.07-4.38) and having experienced a hypoglycaemic episode (OR 0.31 CI 0.09-1.10) predicted better glycaemic control. Duration of diabetes < 10 years (OR 1.36 CI 0.37-5.02), diastolic blood pressure ≥ 70 mmHg (OR 2.80 CI 0.80-9.78), aspirin dosage ≥ 150 mg daily (OR 6.47 CI 1.60-26.05), simvastatin dosage = 40 mg daily (OR 2.35 CI 0.31-18.10) and body mass index > 25 kg/m2 (OR 1.09 CI 0.49-2.41) all predicted a poorer glycaemic result. HBA1C was found to positively correlate with diastolic blood pressure (p = 0.0024, r = 0.31). Systolic blood pressure positively correlated with diastolic blood pressure (p < 0.0001, r = 0.56). Apart from correlating with systolic blood pressure and HBA1C, diastolic blood pressure also positively correlated with the triglyceride level (p = 0.0003, r = 0.36). Positive correlations between total cholesterol, triglycerides, HDL-C and LDL-C were found. As expected, body mass index and waist circumference correlated positively (p < 0.0001, r = 0.82). Level of care was not at the level recommended by the 2012 SEMDSA guidelines. Only 6% of patients met the waist circumference goal. Only 15% of patients achieved blood v i pressure goal. Most of the patients (86%) who qualified for aspirin did not receive it. In the group of patients receiving aspirin 33% did not qualify. According to the SEMDSA guidelines, most of the patients not receiving a statin (90%) should have been on statin therapy. Only 23.5% of patients on statins were at lipid goal. The frequency of laboratory testing did not meet SEMDSA guidelines. There were 31 (31%) patients without a urea, creatinine and electrolyte test for the previous year and 37 (37%) patients without a lipogram for the previous year. Only 21 patients had a listed urine albumin/creatinine ratio and only 33% of these had been done in the previous year. Conclusions Various new variables were identified in the search for predictors of glycaemic control. It was surprising to find that education level, monthly income, smoking status and knowledge of diabetes did not have a statistical impact on glycaemic control. Increased age, duration of diabetes, glomerular filtration rate, hypoglycaemic frequency and diabetic complications experienced were associated with improved glycaemic control. Increased diastolic blood pressure, aspirin dosage, statin dosage and body mass index were associated with worse glycaemic control. The standard of care in the clinic was found on the whole to be inferior to the level of care recommended by SEMDSA.

Dissertation submitted to the Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, in fulfilment of the requirements for the degree of Master of Science In Medicine / The interaction between miglitol, an a-glucosidase inhibitor used as an adjunct therapy in diabetes, and phenytoin, an anticonvulsant primarily used in the treatment of epilepsy, was studied over 26 days. Twenty-four healthy male volunteers took part in a placebo controlled, double blind, cross-over study in two phases, to determine the effects of multiple 100mg doses of miglitol on a single 400mg dose of phenytoin sodium. Miglitol or placebo was administered three times daily from Day 1 to Day 5. Phenytoin was administered as a single dose on Day 3 of each phase, after which blood samples were taken at regular intervals. A washout period of 14 days separated the two phases. / IT2018

Introdução: Não foram encontrados na literatura dados sobre a prevalência de diabetes mellitus em mães de crianças com fissura labiopalatina. Dada a relevância do tema esse foi o principal objetivo da presente investigação. Método: Após aprovação do comitê de ética e pesquisa e obtenção do consentimento livre e esclarecido, foram entrevistadas 325 mulheres, mães biológicas de crianças com fissuras labiopalatinas com idades de 0 a 3 anos matriculadas no Hospital de Reabilitação de Anomalias Craniofaciais da Universidade de São Paulo, HRAC-USP. Foi aplicado questionário com 24 questões onde obteve-se a identificação das mães, o tipo de diabetes, os principais sinais clínicos, as comorbidades associadas, hipertensão, obesidade, medicamentos e drogas lícitas e ilícitas usadas durante a gravidez. Também foram aferidos a pressão arterial e o perímetro abdominal das mães e anotados os valores da glicemia de jejum na primeira consulta pré-natal. Resultados: Os resultados obtidos mostraram que 88 mulheres apresentavam diabetes mellitus, sendo 78 com diabetes mellitus gestacional, 05 com diabetes mellitus tipo 1 e 05 com diabetes mellitus tipo 2. A prevalência de diabetes mellitus em mães de crianças com fissuras labiopalatinas foi de 27,08%, comparativamente maior (p<0,01) aos valores encontrados na população adulta brasileira (7,6%). Dessas 88 mulheres, foram excluídos os fatores que de alguma forma pudessem influenciar o aparecimento de fissuras labiopalatinas e obteve-se o percentual de 16%, onde a hiperglicemia materna nessas mães foi o único fator provável no aparecimento dessas fissuras. Conclusão: A prevalência de diabetes mellitus em mães de crianças com fissuras labiopalatinas foi de 27,08%. (p<0,01). Extraídos os fatores tidos como confundidores, tais como medicamentos usados na gestação, álcool, tabaco, drogas ilícitas, obesidade e hipertensão arterial, restaram 52 mães que tinham somente a hiperglicemia materna como fator isolado, num total de 16%. / Introduction: there was no data found on the literature regarding the prevalence of diabetes mellitus in mothers of children born with cleft lip and palate. Given the relevance of the subject, that was the main objective of this investigation. Method: after approval by the research and ethics committee, and the free, informed consent firmed by the mothers, interviews were conducted with 325 women biological mothers of children aged between 0 to 3 years, born with cleft lip and palate, registered at Hospital for Rehabilitation of Craniofacial Anomalies, University of São Paulo, HRAC-USP. A 24-question survey was applied to collect data regarding the mothers identification, type of diabetes, clinical data, associated comorbidities, hypertension, obesity, medication and legal or illegal drugs used during pregnancy. Blood pressure and abdominal circumference were measured, and fasting blood glucose levels measured at the first prenatal visit. Results: It was found that 88 women had diabetes mellitus, 78 of which had gestational diabetes, 05 had type 1 diabetes and 05 had type 2 diabetes the prevalence of diabetes in mothers of children with cleft lip and palate was 27.08%, a number significantly higher (p<0.01) than the rates found among the Brazilian adult population (7.6%). From this group of 88 women we have suppressed the factors that could contribute to the occurrence of cleft lip and palate in some way, obtaining a prevalence of 16% in which maternal hyperglycemia was the only likely cause of cleft lip and palate. Conclusion: The prevalence of diabetes mellitus in mothers of children born with cleft lip and palate was 27.08% (p<0.01). Suppressing other potentially confusing factors like medication used during pregnancy, alcohol, tobacco, illegal drugs, obesity and hypertension, there were 52 mothers who had only maternal hyperglycemia as an isolated factor, representing 16% of the total studied population.

Type 2 diabetes, as a complex disease, has a range of genetic and environmental factors that underpin its aetiology. It is hoped that the emerging study of epigenetic processes will provide the necessary mechanistic insight into the genetic and environmental interactions that, to date, are poorly understood. This thesis considers the role of DNA methylation, an epigenetic modification, in the aetiology of type 2 diabetes. A range of different genome-wide and whole genome techniques are applied to a study of established type 2 diabetes and experimental models (human and animal) of fetal programming. Samples from a recent genome-wide association study of type 2 diabetes were used to identify DNA methylation patterns at areas of genetic variation associated with disease risk. Analysis of data from methylated DNA immunoprecipitation and microarray identified a genetic-epigenetic interaction in the FTO gene. At this locus, the presence or absence of a SNP created or abrogated a CpG site capable of methylation and further analyses highlighted possible functional relevance via enhancer activity. Models of fetal programming were then used to identify whether variation in DNA methylation may underlie the ‘programmed’ phenotype of diabetes and related cardiometabolic disease. Pre-existing human models of programming via maternal vitamin B12 deficiency and maternal famine exposure have been used to generate exploratory evidence of such mechanisms. Whole genome-based techniques (Medip-seq and Illumina 450k methylation array) were used to profile DNA methylation in whole blood samples from the offspring born to each of these studies. Custom bioinformatic analysis was performed to identify differences in methylation between offspring exposed versus unexposed to the in utero environmental insult. Technical replication and validation studies are ongoing to confirm or refute the presence of regions of differential methylation. Finally, this thesis considers whether a state of ‘over nutrition’ gestational diabetes, may play a role in fetal programming. This condition is of increasing prevalence across the world and is characterised by maternal hyperglycaemia and insulin resistance, often resulting in fetal overgrowth. A mouse model using an inbred strain (Lepr) of mice induced a programmed phenotype of glucose intolerance and obesity in aged offspring born to mothers with gestational diabetes. Medip-seq performed on the livers of late gestation mouse embryos identified differential methylation in cases vs. controls, located at genomic regions with potential functional relevance. A human cohort of women with gestational diabetes was collected to develop further hypothesis around the multiple environmental factors that could interact in pregnancy. Prevalent nutritional deficiencies of vitamin D, iron and one-carbon metabolites were found in women with and without gestational diabetes recruited from a local antenatal clinic. This thesis presents preliminary findings that variation in DNA methylation may be involved in the genetic and environmental risk of type 2 diabetes. The work presented highlights how future studies must incorporate integrated genetic, epigenetic and functional analysis with sufficient sample size if their results are to be translatable to diverse populations at risk of diabetes.

The molecular basis for the use of synthetic growth factors (GFs) in tissue reparation has been poorly investigated. More recently, autologous derived platelet rich growth factor has gained popularity in the field of regenerative/ reparative medicine, mostly because it fits the description of an ideal naturally existing constellation of GFs. However, its efficacy remains controversial. Hence, this study is designed to further elucidate the physiological role of PRP in treating chronic diabetic foot ulcers. Platelet -rich plasma (PRP) and Platelet -poor plasma (PPP) were prepared from blood samples taken from healthy donors and diabetic patients through the use of platelet collecting and concentrating system. The GFs released were measured through immunoassay technique. The effects of the varying concentrations of PRP/PPP in culture media was assessed through tissue culture assay (proliferation, cell migration and angiogenesis assay) on human epithelia keratinocyte, dermal fibroblast and umbilical vein endothelia cell. Furthermore, immuno-histochemistry technique was used to evaluate the differentiation, proliferation, migration and extracellular matrix alterations occurring along wound margins of patients with chronic diabetic ulcers following PRP treatment. A significant difference was observed when the expression of platelet derived growth factor-AA, epidermal growth factor, vascular endothelia growth factor, transforming growth factor and thrombospodin-I released from PRP/PPP were compared between the two groups. There was a significant proliferative, migratory and angiogenic effect of PRP over PPP in the tissue culture assay; however this effect was most prominent with 5% PRP. Overall, hyperproliferative keratin, CD44 and β1-integrin were upregulated in diabetic ulcer keratinocytes as compared with normal foot skin. The clinical study showed that 3 of the 7 diabetic foot ulcer patients treated with PRP achieved complete wound re-epithelisation. We have been able to demonstrate through in vitro studies that PRP has a positive biological effect which mimics normal physiological tissue reparation process.

Type 2 Diabetes, a metabolic disorder associated with chronic hyperglycaemia and hyperlipidaemia, is characterised by an impairment of insulin secretion and production and β-cell death. This β-cell dysfunction is determined by different factors, among which inflammatory processes, characterised by increased expression of pro-inflammatory cytokines and chemokines. Although some molecular mechanisms have been proposed to be involved in this β-cell dysfunction, they fail to explain the whole process. In this thesis, a combined approach of microarray, RNAseq, RT-qPCR and western blot will be used to elucidate the pathways affected under glucolipotoxicity, in order to discover novel molecules involved in the pathogenesis of T2D. We found that INS-1 cells exposed to 27 mM glucose, 200 μM oleic acid, 200 μM palmitic acid, show an overexpression of CD40, a TNF receptor involved in inflammation (more than 300% p<0.01), both at RNA and protein level. These data were validated in cultured human islets (p<0.05) and in islets of mice fed a high fat diet (p<0.05). We showed also that siRNA downregulation of CD40 is associated with increase in insulin secretion (p<0.05), revealing a potential new role of this receptor in β-cells. In addition, RNAseq analysis revealed a wide list of molecules differentially expressed in glucolipotoxicity, in particular molecules involved in inflammation, insulin/IGF pathway, fatty acids-cholesterol metabolism and biosynthesis. We focused our attention on potential novel targets, including the thyroid pathway, unknown microRNAs and novel genes, in order to discover new pathways involved in the impairment of insulin secretion in T2D. This work will open the way to future studies aiming to characterise these molecules and to understand their role in the insulin secretion process. Interesting candidates can then be used in the future as potential targets for the development of new and specific therapeutic strategies.