Building digital twins of the human immune system: toward a roadmap.

Digital twins, customized simulation models pioneered in industry, are beginning to be deployed in medicine and healthcare, with some major successes, for instance in cardiovascular diagnostics and in insulin pump control. Personalized computational models are also assisting in applications ranging from drug development to treatment optimization. More advanced medical digital twins will be essential to making precision medicine a reality. Because the immune system plays an important role in such a wide range of diseases and health conditions, from fighting pathogens to autoimmune disorders, digital twins of the immune system will have an especially high impact. However, their development presents major challenges, stemming from the inherent complexity of the immune system and the difficulty of measuring many aspects of a patient's immune state in vivo. This perspective outlines a roadmap for meeting these challenges and building a prototype of an immune digital twin. It is structured as a four-stage process that proceeds from a specification of a concrete use case to model constructions, personalization, and continued improvement.

Development and validation of a Diabetes Risk Score for screening undiagnosed diabetes in Sri Lanka (SLDRISK).

BACKGROUND: Screening for undiagnosed diabetes is not widely undertaken due to the high costs and invasiveness of blood sampling. Simple non-invasive tools to identify high risk individuals can facilitate screening. The main objectives of this study are to develop and validate a risk score for screening undiagnosed diabetes among Sri Lankan adults and to compare its performance with the Cambridge Risk Score (CRS), the Indian Diabetes Risk Score (IDRS) and three other Asian risk scores. METHODS: Data were available from a representative sample of 4276 adults without diagnosed diabetes. In a jack-knife approach two thirds of the sample was used for the development of the risk score and the remainder for the validation. Age, waist circumference, BMI, hypertension, balanitis or vulvitis, family history of diabetes, gestational diabetes, physical activity and osmotic symptoms were significantly associated with undiagnosed diabetes (age most to osmotic symptoms least). Individual scores were generated for these factors using the beta coefficient values obtained in multiple logistic regression. A cut-off value of sum = 31 was determined by ROC curve analysis. RESULTS: The area under the ROC curve of the risk score for prevalent diabetes was 0.78 (CI 0.73-0.82). In the sample 36.3 % were above the cut-off of 31. A risk score above 31 gave a sensitivity, specificity, positive predictive value and negative predictive value of 77.9, 65.6, 9.4 and 98.3 % respectively. For Sri Lankans the AUC for the CRS and IDRS were 0.72 and 0.66 repectively. CONCLUSIONS: This simple non-invasive screening tool can identify 80 % of undiagnosed diabetes by selecting 40 % of Sri Lankan adults for confirmatory blood investigations.

The influence of family history of diabetes on disease prevalence and associated metabolic risk factors among Sri Lankan adults.

AIMS: To describe the influence of family history on diabetes prevalence and associated metabolic risk factors in a nationally representative sample from Sri Lanka. METHODS: A cross sectional national survey was conducted among 5000 adults in Sri Lanka. Family history was evaluated at three levels: (1) parents, (2) grandparents (paternal and maternal) and (3) siblings. A binary-logistic regression analysis controlling for confounders (age, gender, BMI and physical activity) was performed in all patients with 'presence of diabetes' as the dichotomous dependent variable and using family history in father, mother, maternal grandmother/grandfather, paternal grandmother/grandfather, siblings and children as binary independent variables. RESULTS: The sample size was 4485, mean age was 46.1 ± 15.1 years and 39.5% were males. In all adults, the prevalence of diabetes was significantly higher in patients with a family history (23.0%) than those without (8.2%) (P < 0.001). When family history was present in both parents, the prevalence of diabetes was 32.9%. Presence of a family history significantly increased the risk of diabetes [odds ratio (OR): 3.35, 95% confidence interval (CI): 2.78-4.03], obesity (OR: 2.45, 95% CI: 1.99-2.99), hypertension (OR: 1.25, 95% CI: 1.08-1.45) and metabolic syndrome (OR: 2.28, 95% CI: 1.97-2.63). In all adults, the presence of a family history of diabetes in a father (OR: 1.29, 95% CI: 1.02-1.63), mother (OR: 1.23, 95% CI: 1.11-1.36), paternal grandfather (OR: 1.27, 95% CI: 1.14-1.41), siblings (OR: 4.18, 95% CI: 3.34-5.22) and children (OR: 5.47, 95% CI: 2.93-10.19) was associated with a significantly increased risk of developing diabetes. CONCLUSIONS: Family history and diabetes had a graded association in the Sri Lankan population, because the prevalence increased with the increasing number of generations affected. Family history of diabetes was also associated with the prevalence of obesity, metabolic syndrome and hypertension. Individuals with a family history of diabetes form an easily identifiable group who may benefit from targeted interventions.

Use of interactive teaching techniques to introduce mental health training to medical schools in a resource poor setting.

OBJECTIVE: There are currently no practising psychiatrists in Somaliland. In 2007 the first medical students graduated from universities in Somaliland without mental health training. We aimed to pilot an intensive but flexible package of mental health training to all senior medical students and interns using interactive training techniques and to evaluate its effectiveness by assessing knowledge, skills and attitudes. METHODS: Teaching techniques included didactic lectures, case based discussion groups and role playing. Informal feedback informed a flexible teaching package. Assessment tools designed specifically for this course included a pre and post course MCQ exam and an OSCE. Changes in students' attitudes were evaluated using a questionnaire administered before and after the course. In addition, a questionnaire administered following the course evaluated the changes students perceived in their knowledge and attitudes to mental health. RESULTS: The MCQ improved from 50.7% pre course to 64.4% post course (p = 9.73 E-08). Students achieved an average overall OSCE mark of 71%. The pre and post attitudes questionnaire was most significantly different for statements relevant to aetiology, stigma and the overlap between mental and physical health. The statement most strongly agreed with after the course was 'I now understand more about the overlap between mental and physical health'. CONCLUSION: Interactive teaching provided a learning experience for both students and trainers. On site and distance learning based on the teaching described here has widened the scope of the training possible in psychiatry and allowed the provision of regular teaching, supervision and peer support in Somaliland. However, the current lack of local expertise means that important issues of sustainability need to be considered in future work.

Apolipoproteins in diabetes dyslipidaemia in South Asians with young adult-onset diabetes: distribution, associations and patterns.

BACKGROUND: Apolipoproteins B (apoB) and AI (apoAI) are strong predictors of cardiovascular disease (CVD). We describe apolipoprotein distributions and their associations with lipids and diabetes subtype in diabetic young adult South Asians. METHODS: In 995 subjects with diabetes, we measured fasting total cholesterol (TC), high-density lipoprotein cholesterol (HDLC), triglycerides (TG), apoB and apoAI, glycosylated haemoglobin (HbA1c) and glutamic acid decarboxylase antibodies (GADA). Low-density lipoprotein cholesterol (LDLC) and non-HDLC (NHDLC) were calculated. We compared values in subjects aged 15-50 y from the United States National Health and Nutrition Examination Survey (NHANES). RESULTS: Median age and duration of diabetes were 38 (range 14-45) and 4 (0-24) y. Men had significantly higher TC, TG, NHDLC, TC/HDLC, apoB/AI and NHDLC/apoB, and lower apoAI than women. Compared with the reference group, patients with type 1 diabetes had lower TG, apoB:apoAI and HDLC:apoAI, and higher HDLC and apoAI. Patients with type 2 diabetes had higher TG, TC, LDLC, NHDLC, TC:HDL, apoB, apoAI and apoB:apoAI, and lower HDLC, LDLC:apoB and HDLC:apoAI. Among patients with type 2 diabetes, 54% had high apoB (>1.2 g/L) and 33% also had high TG (>1.5 mmol/L). Measures of obesity (body mass index and waist circumference) were weakly correlated with lipid and apoprotein parameters, suggesting a modest contribution to dyslipidaemia. CONCLUSIONS: A large proportion of young adult Sri Lankan patients with type 2 diabetes has a low LDLC:apoB and high apoB and/or TG, suggesting that these patients are at increased risk of CVD.

Diabetes mellitus among young adults in Sri Lanka--role of GAD antibodies in classification and treatment: the Sri Lanka Young Diabetes study.

AIMS/HYPOTHESIS: Diabetes mellitus is increasing among young adult South Asians. The aim of this study was to determine the prevalence and phenotypic characteristics of diabetes subtypes based on GAD65 autoantibody (GADA) status in those with young adult-onset diabetes in Sri Lanka. METHODS: Clinical, metabolic and GADA data were available for 992 consecutively recruited individuals with diabetes aged < or =45 years (age at diagnosis 16-40 years). Participants were classified according to the following definitions: type 1 diabetes, insulin-dependent <6 months from diagnosis; latent autoimmune diabetes in adults (LADA), GADA-positive, age > or =30 years and insulin-independent > or =6 months from diagnosis; type 2 diabetes, GADA-negative and insulin-independent > or =6 months from diagnosis. RESULTS: The median (interquartile range) age at diagnosis and diabetes duration were 33.0 (29.0-36.1) and 4.0 (1.1-7.1) years, respectively; 42.1% were male. GADA positivity was seen in 5.4% of participants (n = 54) and GADA levels negatively correlated with age at diagnosis (p < 0.0001), BMI (p < 0.0001) and time to insulin requirement (p = 0.006). Type 1 diabetes, type 2 diabetes and LADA were present in 7.0%, 89.7% and 2.6%, respectively. The remaining 0.7% of the participants were GADA-positive, insulin independent > or =6 months from diagnosis and were diagnosed at age <30 years. The metabolic syndrome and homeostasis model assessment of beta cell function (HOMA %B) were lowest in GADA-positive type 1 diabetes and increased progressively in latent autoimmune diabetes, GADA-negative type 1 diabetes and type 2 diabetes. Among those requiring insulin, 69.2% had fasting C-peptide levels in the lowest quartile, whereas only 19.5% were GADA-positive (p < 0.0001). CONCLUSIONS/INTERPRETATION: The prevalence of GADA-positive autoimmune diabetes is low among individuals with young adult-onset diabetes in Sri Lanka. Young-onset diabetic phenotypes appear as a continuum from autoimmune type 1 diabetes to type 2 diabetes.

Prevalence and clinical characteristics of maternally inherited diabetes and deafness caused by the mt3243A > G mutation in young adult diabetic subjects in Sri Lanka.

AIMS: The maternally inherited mt3243A > G mutation is associated with a variable clinical phenotype including diabetes and deafness (MIDD). We aimed to determine the prevalence and clinical characteristics of MIDD in a large South Asian cohort of young adult-onset diabetic patients from Sri Lanka. METHODS: DNA was available from 994 subjects (age of diagnosis 16-40 years, age at recruitment < or = 45 years). Mutation screening was performed using a QRT-PCR method on an ABI 7900HT system using sequence-specific probes. Samples with heteroplasm > or = 5.0% were considered positive. RESULTS: Nine (four males) mutation-positive subjects were identified (prevalence 0.9%). They were diagnosed at a younger age (25.9 +/- 4.8 years vs. 31.9 +/- 5.6 years, P = 0.002) and were lean (body mass index [BMI] 18.7 +/- 2.7 kg/m(2) vs. 24.7 +/- 4.0 kg/m(2), P < 0.001) compared to NMCs. One mutation-positive subject (11.1%) had metabolic syndrome, compared to 633 (64.3%) of NMCs. Insulin therapy within 6 months of diagnosis was used in four (44.0%) carriers compared to 6.9% of NMCs (P = 0.002). Combined screening criteria of any two of maternal history of diabetes, personal history of hearing impairment and family history of hearing impairment only identified five (55%) of the carriers, with a positive predictive value of 7.4%. CONCLUSIONS: The prevalence of mt3243A > G mutation among young adult-onset diabetic subjects from Sri Lanka was 0.9%. Our study demonstrates that a maternal family history of diabetes and either a personal and/or family history of deafness only distinguish half of patients with MIDD from Sri Lankan subjects with young-onset diabetes.

Prevalence and projections of diabetes and pre-diabetes in adults in Sri Lanka--Sri Lanka Diabetes, Cardiovascular Study (SLDCS).

AIMS: To determine the prevalence of diabetes mellitus and pre-diabetes (impaired fasting glucose and impaired glucose tolerance) in adults in Sri Lanka. Projections for the year 2030 and factors associated with diabetes and pre-diabetes are also presented. METHODS: This cross-sectional study was conducted between 2005 and 2006. A nationally representative sample of 5000 adults aged >or= 18 years was selected by a multi-stage random cluster sampling technique. Fasting plasma glucose was tested in all participants and a 75-g oral glucose tolerance test was performed in non-diabetic subjects. Prevalence was estimated for those > 20 years of age. RESULTS: Response rate was 91% (n = 4532), males 40%, age 46.1 +/- 15.1 years (mean +/- standard deviation). The age-sex standardized prevalence (95% confidence interval) of diabetes for Sri Lankans aged >or= 20 years was 10.3% (9.4-11.2%) [males 9.8% (8.4-11.2%), females 10.9% (9.7-12.1%), P = 0.129). Thirty-six per cent (31.9-40.1%) of all diabetic subjects were previously undiagnosed. Diabetes prevalence was higher in the urban population compared with rural [16.4% (13.8-19.0%) vs. 8.7% (7.8-9.6%); P < 0.001]. The prevalence of overall, urban and rural pre-diabetes was 11.5% (10.5-12.5%), 13.6% (11.2-16.0%) and 11.0% (10.0-12.0%), respectively. Overall, 21.8% (20.5-23.1%) had some form of dysglycaemia. The projected diabetes prevalence for the year 2030 is 13.9%. Those with diabetes and pre-diabetes compared with normal glucose tolerance were older, physically inactive, frequently lived in urban areas and had a family history of diabetes. They had higher body mass index, waist circumference, waist-hip ratio, systolic/diastolic blood pressure, low-density lipoprotein cholesterol and triglycerides. Insulin was prescribed to 4.4% (2.7-6.1%) of all diabetic subjects. CONCLUSIONS: One in five adults in Sri Lanka has either diabetes or pre-diabetes and one-third of those with diabetes are undiagnosed.

Integrating genetics and genomics to identify new leads for the control of Eimeria spp.

Eimerian parasites display a biologically interesting range of phenotypic variation. In addition to a wide spectrum of drug-resistance phenotypes that are expressed similarly by many other parasites, the Eimeria spp. present some unique phenotypes. For example, unique lines of Eimeria spp. include those selected for growth in the chorioallantoic membrane of the embryonating hens egg or for faster growth (precocious development) in the mature host. The many laboratory-derived egg-adapted or precocious lines also share a phenotype of a marked attenuation of virulence, the basis of which is different as a consequence of the in ovo or in vivo selection procedures used. Of current interest is the fact that some wild-type populations of Eimeria maxima are characterized by an ability to induce protective immunity that is strain-specific. The molecular basis of phenotypes that define Eimeria spp. is now increasingly amenable to investigation, both through technical improvements in genetic linkage studies and the availability of a comprehensive genome sequence for the caecal parasite E. tenella. The most exciting phenotype in the context of vaccination and the development of new vaccines is the trait of strain-specific immunity associated with E. maxima. Recent work in this laboratory has shown that infection of two inbred lines of White Leghorn chickens with the W strain of E. maxima leads to complete protection to challenge with the homologous parasite, but to complete escape of the heterologous H strain, i.e. the W strain induces an exquisitely strain-specific protective immune response with respect to the H strain. This dichotomy of survival in the face of immune-mediated killing has been examined further and, notably, mating between a drug-resistant W strain and a drug-sensitive H strain leads to recombination between the genetic loci responsible for the specificity of protective immunity and resistance to the anticoccidial drug robenidine. Such a finding opens the way forward for genetic mapping of the loci responsible for the induction of protective immunity and integration with the genome sequencing efforts.

Integrating centralized and decentralized organization structures: an education and development model.

Organization change efforts have led to critically examining the structure of education and development departments within hospitals. This qualitative study evaluated an education and development model in an academic health sciences center. The model combines centralization and decentralization. The study results can be used by staff development educators and administrators when organization structure is questioned. This particular model maximizes the benefits and minimizes the limitations of centralized and decentralized structures.

Clinical features of systemic lupus erythematosus in Sri Lankan patients: results from a lupus clinic.

OBJECTIVES: To find the common clinical features, pattern of visceral involvement, treatment received and outcome in patients diagnosed as having systemic lupus erythematosus (SLE) on American Rheumatological Association (ARA) criteria. SETTING: Clinic for patients referred or admitted to the University Medical Unit, National Hospital of Sri Lanka, Colombo, with diagnosed or suspected SLE. DESIGN AND METHODS: A prospective descriptive study. Clinical features of patients collected at time of registration in the clinic were maintained in a database. Patients were followed up prospectively and changes recorded. Data were analysed after 3 years of follow up. RESULTS: Of the 111 patients registered during this period, 96 (86%) were clinically diagnosed as having SLE. Of these, 77 patients (80%) satisfied ARA criteria for diagnosis of SLE. 72 were females (93%). The mean age of patients who satisfied the ARA criteria was 32 years (range 11 to 58), and the mean duration of disease 7 years (range 1 to 15). The commonest presentation was with mucocutaneous features (98%) and alopecia in 87%. Systemic features were found in 92% of patients. 67 (87%) of patients had visceral involvement with 60 (78%) having it at time of diagnosis. 53 (69%) had renal, 42 (54%) haematological, 33 (42%) neurological, 12 (16%) cardiac and 8 patients pulmonary involvement. Five patients died during the 3-year follow up and 2 developed chronic renal failure. Three patients underwent successful pregnancy after diagnosis of SLE. CONCLUSIONS: Our study confirmed the wide variability of clinical features seen in SLE. Alopecia and visceral involvement were common in Sri Lankan patients.

Pregnancy following renal transplantation in Sri Lanka.

OBJECTIVE: To assess the outcome and complications of pregnancy following renal transplantation in Sri Lanka. METHODS: Ten pregnancies following transplantation managed between January 1993 and July 1999 by the University Obstetrics and Gynaecology Unit, De Soysa Hospital for Women, Colombo were reviewed. RESULTS: Five women had planned pregnancy with an average duration from transplantation to conception of 2.3 (+/- 0.2) years; five had an unplanned pregnancy within 12 months of transplantation. All were treated with immunosuppressives, with none developing rejection. In the planned pregnancy group, 3 developed pregnancy induced hypertension and 3 impaired glucose tolerance. All delivered mature healthy babies with an average birth weight of 2.6 (+/- 0.3) kg. In the unplanned group, 1 developed cholestatic jaundice and delivered a growth retarded baby at 36 weeks. Another developed severe pulmonary oedema at 34 weeks (due to a past myocardial infarction) resulting in a fresh stillbirth. Two others has mid-trimester foetal deaths complicating severe diabetes mellitus. The conception at 3 months after transplantation developed diabetes mellitus and pregnancy induced hypertension, and delivered a live growth retarded baby. None had deterioration of renal function. CONCLUSION: Although a successful outcome is possible with stringent pre-pregnancy selection, maternal morbidity and foetal wastage can be high in those without.

A new monospecific ovine Fab fragment antivenom for treatment of envenoming by the Sri Lankan Russell's viper (Daboia Russelii Russelii): a preliminary dose-finding and pharmacokinetic study.

Russell's viper is the most important cause of life-threatening snake bite and acute renal failure in Sri Lanka. Only equine polyspecific antivenoms imported from India are available. They have not proved effective clinically or in clearing venom antigenemia and they frequently cause reactions. In an attempt to reduce mortality and morbidity, a new monospecific ovine Fab fragment antivenom (PolongaTab; Therapeutic Antibodies, Inc., London, United Kingdom) was raised against Sri Lankan Russell's viper venom. In a preliminary dose-finding study in 35 patients, an initial dose of 3-4 g restored blood coagulability permanently and stopped systemic bleeding, even in severely envenomed patients. Venom antigenemia disappeared within 1 hr of antivenom treatment but recurred, probably as a result of continued absorption of venom from the site of the bite, after the rapid clearance of therapeutic antibody. Twelve patients (34%) experienced early reactions that were usually mild and always responded to epinephrine.

The quality of function of renal allografts is associated with donor age.

The quality of renal allograft function was assessed by prospective measurement of creatinine clearance at 1 year (n = 197) and at 3 years (n = 115) after cadaveric renal transplantation in a cohort of 268 patients treated with triple therapy immunosuppression. Donor age (P < 0.0012) and recipient age (P < 0.01) were independently associated with creatinine clearance both at 1 and at 3 years. In patients with donor age above 50 years and recipient age above 45 years, the mean creatinine clearance was 32.7 (SD 10.4) ml/min (n = 27). When the donor age was below 30 years and recipient age below 45 years, the mean creatinine clearance was 55.6 (SD 14.4) ml/min (n = 47, P < 0.001). However, in these patients there was no significant association between graft function and many of the factors known to influence graft survival, such as HLA matching, sensitisation of the recipient, and the occurrence of rejection. In conclusion, the quality of renal allograft function declined with increasing donor and recipient age in our patients, whilst immunological factors were not significantly associated with function in surviving grafts.