Risk assessment tools for identifying individuals at risk of developing type 2 diabetes

Machine Learning Models in Type 2 Diabetes Risk Prediction: Results from a Cross-sectional Retrospective Study in Chinese Adults

Type 2 diabetes mellitus (T2DM) has become a prevalent health problem in China, especially in urban areas. Early prevention strategies are needed to reduce the associated mortality and morbidity. We applied the combination of rules and different machine learning techniques to assess the risk of development of T2DM in an urban Chinese adult population. A retrospective analysis was performed on 8000 people with non-diabetes and 3845 people with T2DM in Nanjing. Multilayer Perceptron (MLP), AdaBoost (AD), Trees Random Forest (TRF), Support Vector Machine (SVM), and Gradient Tree Boosting (GTB) machine learning techniques with 10 cross validation methods were used with the proposed model for the prediction of the risk of development of T2DM. The performance of these models was evaluated with accuracy, precision, sensitivity, specificity, and area under receiver operating characteristic (ROC) curve (AUC). After comparison, the prediction accuracy of the different five machine models was 0.87, 0.86, 0.86, 0.86 and 0.86 respectively. The combination model using the same voting weight of each component was built on T2DM, which was performed better than individual models. The findings indicate that, combining machine learning models could provide an accurate assessment model for T2DM risk prediction.

Effectiveness of a type 2 diabetes prevention program combining FINDRISC scoring and telephone-based coaching in the French population of bakery/pastry employees

Background/objectives

Preventive actions targeting the risk of type 2 diabetes mellitus (T2D) and deployed from the workplace are scarce. This study aimed to measure this T2D risk in a large sample of the bakery/pastry employees in France and to assess the effectiveness of a telephone coaching program in participants with the highest risk.

Subjects/methods

A screening survey using the FINDRISC score was conducted by phone among the employees. Those with a moderate risk (score ≥ 12 and <15; body mass index ≥ 25 kg/m²) or high/very high risk (score ≥ 15) were invited to participate in a 6-month coaching program including 6 monthly interviews together with a final evaluation interview three months later. The effects and impact were evaluated using 8 questions on dietary knowledge/behavior as well as the GPAQ (physical activity) and SF-12 (quality of life) questionnaires.

Results

There were 19,951 employees eligible for screening (age: 38.0 ± 13.5 years, men 49.6%, mean FINDRISC score 5.9 ± 4.4). A high/very high score was found in 4% of individuals. Overall, 1,348 (among 2,018) eligible employees agreed to participate in the coaching program, 630 of whom participated in all interviews. Of the latter, dietary knowledge/behavior (+1.60) and quality of life (+1.83) improved (P < 0.0001), with a favorable trend for physical activity (+0.06, P = 0.0756). Dietary knowledge/behavior continued to improve in the 581 completers (+0.17, P = 0.0001).

Conclusions

This two-step prevention program associating T2D risk estimation and a 6-month telephone coaching was deployed in the French craft bakery/pastry sector with significant adhesion. Such program appears beneficial for enhancing knowledge and mobilizing skills associated with T2D prevention.

Favourable understandability, but poor actionability: An evaluation of online type 2 diabetes risk calculators

OBJECTIVE

The study aim was to identify all freely available online diabetes risk calculators and to evaluate their suitability for patients with low health literacy.

METHODS

Online diabetes risk calculators were identified by an environmental scan. The Patient Education Material Assessment Tool for Printable Materials was used to determine understandability and actionability scores. A high-risk profile was used to compare the risk results obtained with each calculator.

RESULTS

Thirty-five risk calculators were identified; 51% had no described model, 23% reported absolute risk and 31% used visual aids. The estimated risk for the same profile ranged from low to very high. The mean understandability score was 79% (SD = 19%) and the mean actionability score was 42% (SD = 30%).

CONCLUSIONS

Online diabetes risk calculators are generally understandable, but not very actionable, and may not be completely suitable for use by patients with low health literacy. The estimated risk is highly variable depending on the underlying model used for the calculation.

PRACTICE IMPLICATIONS

Patients and healthcare providers need to exercise caution when selecting a diabetes risk calculator.

Suboptimal health status as an independent risk factor for type 2 diabetes mellitus in a community-based cohort: the China suboptimal health cohort study

Background

The prevalence of diabetes, constituted chiefly by type 2 diabetes mellitus (T2DM), is a global public health threat. Suboptimal health status (SHS), a physical state between health and disease, might contribute to the progression or development of T2DM.

Methods

We conducted a prospective cohort study, based on the China Suboptimal Health Cohort Study (COACS), to understand the impact of SHS on the progress of T2DM. We examined associations between SHS and T2DM outcomes using multivariable logistic regression models and constructed predictive models for T2DM onset based on SHS.

Results

A total of 61 participants developed T2DM after an average of 3.1 years of follow-up. Participants with higher SHS scores had more T2DM outcomes (p = 0.036). Moreover, compared with the lowest quartile of SHS scores, participants with fourth, third, and second quartile SHS scores were found to be associated with a 1.7-fold, 1.6-fold, and 1.5-fold risk of developing T2DM, respectively. The predictive model constructed with SHS had higher discriminatory power (AUC = 0.848) than the model without SHS (AUC = 0.795).

Conclusions

The present study suggests that a higher SHS score is associated with a higher incidence of T2DM. SHS is a new independent risk factor for T2DM and has the capability to act as a predictive tool for T2DM onset. The evaluation of SHS combined with the analysis of modifiable risk factors for SHS allows the risk stratification of T2DM, which may consequently contribute to the prevention of T2DM development. These findings might require further validation in a longer-term follow-up study.

Driving Type 2 Diabetes Risk Scores into Clinical Practice: Performance Analysis in Hospital Settings

Electronic health records and computational modelling have paved the way for the development of Type 2 Diabetes risk scores to identify subjects at high risk. Unfortunately, few risk scores have been externally validated, and their performance can be compromised when routine clinical data is used. The aim of this study was to assess the performance of well-established risk scores for Type 2 Diabetes using routinely collected clinical data and to quantify their impact on the decision making process of endocrinologists. We tested six risk models that have been validated in external cohorts, as opposed to model development, on electronic health records collected from 2008-2015 from a population of 10,730 subjects. Unavailable or missing data in electronic health records was imputed using an existing validated Bayesian Network. Risk scores were assessed on the basis of statistical performance to differentiate between subjects who developed diabetes and those who did not. Eight endocrinologists provided clinical recommendations based on the risk score output. Due to inaccuracies and discrepancies regarding the exact date of Type 2 Diabetes onset, 76 subjects from the initial population were eligible for the study. Risk scores were useful for identifying subjects who developed diabetes (Framingham risk score yielded a c-statistic of 85%), however, our findings suggest that electronic health records are not prepared to massively use this type of risk scores. Use of a Bayesian Network was key for completion of the risk estimation and did not affect the risk score calculation (p > 0.05). Risk score estimation did not have a significant effect on the clinical recommendation except for starting pharmacological treatment (p = 0.004) and dietary counselling (p = 0.039). Despite their potential use, electronic health records should be carefully analyzed before the massive use of Type 2 Diabetes risk scores for the identification of high-risk subjects, and subsequent targeting of preventive actions.

Incident Type 2 Diabetes Among Individuals With CKD: Findings From the Chronic Renal Insufficiency Cohort (CRIC) Study

RATIONALE & OBJECTIVE

Few studies have examined incident type 2 diabetes mellitus (T2DM) in chronic kidney disease (CKD). Our objective was to examine rates of and risk factors for T2DM in CKD, using several alternative measures of glycemic control.

STUDY DESIGN

Prospective cohort study.

SETTING & PARTICIPANTS

1,713 participants with reduced glomerular filtration rates and without diabetes at baseline, enrolled in the Chronic Renal Insufficiency Cohort (CRIC) Study.

PREDICTORS

Measures of kidney function and damage, fasting blood glucose, hemoglobin A1c (HbA1c), HOMA-IR (homeostatic model assessment of insulin resistance), demographics, family history of diabetes mellitus (DM), smoking status, medication use, systolic blood pressure, triglyceride level, high-density lipoprotein cholesterol level, body mass index, and physical activity.

OUTCOME

Incident T2DM (defined as fasting blood glucose ≥ 126mg/dL or prescription of insulin or oral hypoglycemic agents).

ANALYTICAL APPROACH

Concordance between fasting blood glucose and HbA1c levels was assessed using κ. Cause-specific hazards modeling, treating death and end-stage kidney disease as competing events, was used to predict incident T2DM.

RESULTS

Overall T2DM incidence rate was 17.81 cases/1,000 person-years. Concordance between fasting blood glucose and HbA1c levels was low (κ for categorical versions of fasting blood glucose and HbA1c = 13%). Unadjusted associations of measures of kidney function and damage with incident T2DM were nonsignificant (P ≥ 0.4). In multivariable models, T2DM was significantly associated with fasting blood glucose level (P = 0.002) and family history of DM (P = 0.03). The adjusted association of HOMA-IR with T2DM was comparable to that of fasting blood glucose level; the association of HbA1c level was nonsignificant (P ≥ 0.1). Harrell's C for the models ranged from 0.62 to 0.68.

LIMITATIONS

Limited number of outcome events; predictors limited to measures taken at baseline.

CONCLUSIONS

The T2DM incidence rate among individuals with CKD is markedly higher than in the general population, supporting the need for greater vigilance in this population. Measures of glycemic control and family history of DM were independently associated with incident T2DM.

Urinary N-Telopeptide as Predictor of Onset of Menopause-Related Bone Loss in Pre- and Perimenopausal Women

The menopause transition (MT) is a period of rapid bone loss and has been proposed to be a time-limited window for early intervention to prevent permanent microarchitectural damage and reduce the risk of subsequent fracture. To intervene early, however, we first need to be able to determine whether menopause-related bone loss is about to begin, in advance of substantial bone loss. The objective of this study was, therefore, to assess whether urinary N-telopeptide (U-NTX) in pre- or early perimenopause can predict the onset of menopause-related bone loss. Repeated U-NTX measurements were obtained during pre- and early perimenopause in 1243 participants from the Study of Women's Health Across the Nation (SWAN). We examined the ability of U-NTX to predict the onset of significant menopause-related bone loss (categorical outcome, yes versus no) at the lumbar spine (LS) and femoral neck (FN), defined as annualized bone mineral density (BMD) decline at a rate faster than the smallest detectable change in BMD over the 3 to 4 years from the time of U-NTX measurement. Adjusting for age, race/ethnicity, body mass index, urine collection time, starting BMD, and study site in multivariable, modified Poisson regression, every standard deviation increment in U-NTX, measured at baseline in early perimenopausal women, was associated with an 18% and 22% greater risk of significant bone loss at the LS (p = 0.003) and FN (p = 0.003), respectively. The area under the receiver-operator curve for predicting LS and FN bone loss was 0.72 and 0.72, respectively. In mixed-effects analysis of all repeated measures of early perimenopausal U-NTX over follow-up, U-NTX predicted onset of bone loss at the LS (p = 0.002) but not at the FN. We conclude that U-NTX can be used early in the MT to determine if a woman is about to experience significant LS bone loss before there has been substantial skeletal deterioration. © 2018 The Authors. JBMR Plus is published by Wiley Periodicals, Inc. on behalf of the American Society for Bone and Mineral Research.

Preconception risk of gestational diabetes: Development of a prediction model in nulliparous Australian women

AIM

To develop a prediction model for preconception identification of women at risk of gestational diabetes mellitus (GDM).

METHODS

Data from a prospective cohort, the Australian Longitudinal Study on Women's Health, were used. Nulliparous women aged 18-23 who reported a pregnancy up to age 37-42 were included. Preconception predictors of GDM during a first pregnancy were selected using logistic regression. Regression coefficients were multiplied by a shrinkage factor estimated with bootstrapping to improve prediction in external populations.

RESULTS

Among 6504 women, 314 (4.8%) developed GDM during their first pregnancy. The final prediction model included age at menarche, proposed age at future first pregnancy, ethnicity, body mass index, diet, physical activity, polycystic ovary syndrome, and family histories of type 1 or 2 diabetes and GDM. The model showed good discriminative ability with a C-statistic of 0.79 (95% CI 0.76, 0.83) after internal validation. More than half of the women (58%) were classified to be at risk of GDM (>2% predicted risk), with corresponding sensitivity and specificity values of 91% and 43%.

CONCLUSIONS

Nulliparous women at risk of GDM in a future first pregnancy can be accurately identified based on preconception lifestyle and health-related characteristics. Further studies are needed to test our model in other populations.

Development and validation of risk models to predict the 7-year risk of type 2 diabetes: The Japan Epidemiology Collaboration on Occupational Health Study

AIMS/INTRODUCTION

We previously developed a 3-year diabetes risk score in the working population. The objective of the present study was to develop and validate flexible risk models that can predict the risk of diabetes for any arbitrary time-point during 7 years.

MATERIALS AND METHODS

The participants were 46,198 Japanese employees aged 30-59 years, without diabetes at baseline and with a maximum follow-up period of 8 years. Incident diabetes was defined according to the American Diabetes Association criteria. With routine health checkup data (age, sex, abdominal obesity, body mass index, smoking status, hypertension status, dyslipidemia, glycated hemoglobin and fasting plasma glucose), we developed non-invasive and invasive risk models based on the Cox proportional hazards regression model among a random two-thirds of the participants, and used another one-third for validation.

RESULTS

The range of the area under the receiver operating characteristic curve increased from 0.73 (95% confidence interval 0.72-0.74) for the non-invasive prediction model to 0.89 (95% confidence interval 0.89-0.90) for the invasive prediction model containing dyslipidemia, glycated hemoglobin and fasting plasma glucose. The invasive models showed improved integrated discrimination and reclassification performance, as compared with the non-invasive model. Calibration appeared good between the predicted and observed risks. These models performed well in the validation cohort.

CONCLUSIONS

The present non-invasive and invasive models for the prediction of diabetes risk up to 7 years showed fair and excellent performance, respectively. The invasive models can be used to identify high-risk individuals, who would benefit greatly from lifestyle modification for the prevention or delay of diabetes.

Metabolomic profile of insulin resistance in patients with multiple sclerosis is associated to the severity of the disease

BACKGROUND

Dysglycemia and adiposity have been related to disability in patients with multiple sclerosis. The objective of this work was to determine the prevalence and characteristics of insulin resistance in patients with multiple sclerosis using the metabolomics Quantose score.

METHODS

A total of 64 patients were accrued in the study. A blood sample was drawn to estimate the Quantose score, which is derived from fasting measurements of insulin, α-hydroxybutyrate, linoleoyl-glycerophosphocholine, and oleate, three nonglucose metabolites shown to correlate with insulin-stimulated glucose disposal.

RESULTS

Insulin resistance was documented in 33 out of 64 patients and it was found in association with the degree of disability and the time from diagnosis. Patients with the secondary progressive form of the disease showed the highest prevalence.

CONCLUSION

Insulin resistance is frequent in patients with multiple sclerosis and might contribute to metabolic complications and general disability. Early markers of dysglycemia should be sought for in these patients to avoid additional deterioration of their quality of life.

Diabetes risk scores for Hispanics living in the United States: A systematic review

AIM

Undiagnosed diabetes is more prevalent among racial/ethnic minorities in the United States (U.S.). Despite the proliferation of risk scores, few have been validated in Hispanics populations. The aim of this study is to systematically review published studies that developed risk scores to identify undiagnosed Type 2 Diabetes Mellitus based on self-reported information that were validated for Hispanics in the U.S.

METHODS

The search included PubMed, EMBASE, Cochrane and CINAHL from inception to 2016 without language restrictions. Risk scores whose main outcome was undiagnosed Type 2 diabetes reporting performance measures for Hispanics were included.

RESULTS

We identified three studies that developed and validated risk scores for undiagnosed diabetes based on questionnaire data. Two studies were conducted in Latin America and one in the U.S. All three studies reported adequate performance (area under the receiving curve (AUC) range between0.68and 0.78). The study conducted in the U.S. reported a higher sensitivity of their risk score for Hispanics than whites. The limited number of studies, small size and heterogeneity of the combined cohorts provide limited evidence of the validity of risk scores for Hispanics.

CONCLUSIONS

Efforts to develop and validate risk prediction models in Hispanic populations in the U.S are needed, particularly given the diversity of thisfast growing population. Healthcare professionals should be aware of the limitations of applying risk scores developed for the general population on Hispanics.

Predicting risk of substantial weight gain in German adults-a multi-center cohort approach

Background

A risk-targeted prevention strategy may efficiently utilize limited resources available for prevention of overweight and obesity. Likewise, more efficient intervention trials could be designed if selection of subjects was based on risk. The aim of the study was to develop a risk score predicting substantial weight gain among German adults.

Methods

We developed the risk score using information on 15 socio-demographic, dietary and lifestyle factors from 32 204 participants of five population-based German cohort studies. Substantial weight gain was defined as gaining ≥10% of weight between baseline and follow-up (>6 years apart). The cases were censored according to the theoretical point in time when the threshold of 10% baseline-based weight gain was crossed assuming linearity of weight gain. Beta coefficients derived from proportional hazards regression were used as weights to compute the risk score as a linear combination of the predictors. Cross-validation was used to evaluate the score's discriminatory accuracy.

Results

The cross-validated c index (95% CI) was 0.71 (0.67-0.75). A cutoff value of ≥475 score points yielded a sensitivity of 71% and a specificity of 63%. The corresponding positive and negative predictive values were 10.4% and 97.6%, respectively.

Conclusions

The proposed risk score may support healthcare providers in decision making and referral and facilitate an efficient selection of subjects into intervention trials.

A threshold-free summary index of prediction accuracy for censored time to event data

Prediction performance of a risk scoring system needs to be carefully assessed before its adoption in clinical practice. Clinical preventive care often uses risk scores to screen asymptomatic population. The primary clinical interest is to predict the risk of having an event by a prespecified future time t0 . Accuracy measures such as positive predictive values have been recommended for evaluating the predictive performance. However, for commonly used continuous or ordinal risk score systems, these measures require a subjective cutoff threshold value that dichotomizes the risk scores. The need for a cutoff value created barriers for practitioners and researchers. In this paper, we propose a threshold-free summary index of positive predictive values that accommodates time-dependent event status and competing risks. We develop a nonparametric estimator and provide an inference procedure for comparing this summary measure between 2 risk scores for censored time to event data. We conduct a simulation study to examine the finite-sample performance of the proposed estimation and inference procedures. Lastly, we illustrate the use of this measure on a real data example, comparing 2 risk score systems for predicting heart failure in childhood cancer survivors.

The Extending Spectrum of NPC1-Related Human Disorders: From Niemann-Pick C1 Disease to Obesity

The Niemann-Pick type C1 (NPC1) protein regulates the transport of cholesterol and fatty acids from late endosomes/lysosomes and has a central role in maintaining lipid homeostasis. NPC1 loss-of-function mutations in humans cause NPC1 disease, a rare autosomal-recessive lipid-storage disorder characterized by progressive and lethal neurodegeneration, as well as liver and lung failure, due to cholesterol infiltration. In humans, genome-wide association studies and post-genome-wide association studies highlight the implication of common variants in NPC1 in adult-onset obesity, body fat mass, and type 2 diabetes. Heterozygous human carriers of rare loss-of-function coding variants in NPC1 display an increased risk of morbid adult obesity. These associations have been confirmed in mice models, showing an important interaction with high-fat diet. In this review, we describe the current state of knowledge for NPC1 variants in relationship to pleiotropic effects on metabolism. We provide evidence that NPC1 gene variations may predispose to common metabolic diseases by modulating steroid hormone synthesis and/or lipid homeostasis. We also propose several important directions of research to further define the complex roles of NPC1 in metabolism. This review emphasizes the contribution of NPC1 to obesity and its metabolic complications.

HAPT2D: high accuracy of prediction of T2D with a model combining basic and advanced data depending on availability

OBJECTIVE

Type 2 diabetes arises from the interaction of physiological and lifestyle risk factors. Our objective was to develop a model for predicting the risk of T2D, which could use various amounts of background information.

RESEARCH DESIGN AND METHODS

We trained a survival analysis model on 8483 people from three large Finnish and Spanish data sets, to predict the time until incident T2D. All studies included anthropometric data, fasting laboratory values, an oral glucose tolerance test (OGTT) and information on co-morbidities and lifestyle habits. The variables were grouped into three sets reflecting different degrees of information availability. Scenario 1 included background and anthropometric information; Scenario 2 added routine laboratory tests; Scenario 3 also added results from an OGTT. Predictive performance of these models was compared with FINDRISC and Framingham risk scores.

RESULTS

The three models predicted T2D risk with an average integrated area under the ROC curve equal to 0.83, 0.87 and 0.90, respectively, compared with 0.80 and 0.75 obtained using the FINDRISC and Framingham risk scores. The results were validated on two independent cohorts. Glucose values and particularly 2-h glucose during OGTT (2h-PG) had highest predictive value. Smoking, marital and professional status, waist circumference, blood pressure, age and gender were also predictive.

CONCLUSIONS

Our models provide an estimation of patient's risk over time and outweigh FINDRISC and Framingham traditional scores for prediction of T2D risk. Of note, the models developed in Scenarios 1 and 2, only exploited variables easily available at general patient visits.

Predictors of undiagnosed prevalent type 2 diabetes - The Danish General Suburban Population Study

AIMS

To investigate how self-reported risk factors (including socioeconomic status) predict undiagnosed, prevalent type 2 diabetes mellitus (T2DM). To externally validate Leicester Risk Assessment Score (LRAS), Finnish Diabetes Risk Score (FINDRISC) and Danish Diabetes Risk Score (DDRS), and to investigate how these predict a European Heart SCORE≥5% in a Danish population study.

METHODS

We included 21,205 adults from the Danish General Suburban Population Study. We used relative importance calculations of self-reported variables in prediction of undiagnosed T2DM. We externally validated established prediction models reporting ROC-curves for undiagnosed T2DM, pre-diabetes and SCORE.

RESULTS

More than 20% of people with T2DM were undiagnosed. The 7 most important self-rated predictors in sequential order were high BMI, antihypertensive-therapy, age, cardiovascular disease, waist-circumference, fitness compared to peers and family disposition for T2DM. The Area Under the Curve for prediction of undiagnosed T2DM was 77.1 for LRAS; 75.4 for DDRS and 67.9 for FINDRISC. AUCs for SCORE was 75.1 for LRAS; 62.3 for DDRS and 54.3 for FINDRISC.

CONCLUSIONS

BMI and self-reported cardiovascular disease are important risk factors for undiagnosed T2DM. LRAS performed better than DDRS and FINDRISC in prediction of undiagnosed T2DM and SCORE≥5%. SCORE performed best in predicting pre-diabetes.

Diabetes Risk Assessment in Latinas: Effectiveness of a Brief Diabetes Risk Questionnaire for Detecting Prediabetes in a Community-Based Sample

OBJECTIVE

Numerous validated questionnaires use self-reported data to quantify individuals' risk of having diabetes or developing it in the future. Evaluations of these tools have primarily used nationally representative data, limiting their application in clinical and community settings. This analysis tested the effectiveness of the American Diabetes Association (ADA) risk questionnaire for identifying prediabetes in a community-based sample of Latinas.

METHODS

Data were collected using the ADA risk questionnaire and assessing A1C. Among 204 participants without diabetes, we examined the association between individual characteristics and glycemic status. We then calculated the performance characteristics (sensitivity, specificity, positive predictive value [PPV], and negative predictive value [NPV]) of the ADA risk questionnaire for detecting prediabetes, using A1C results as the gold standard to define the outcome.

RESULTS

All participants were women of self-reported Hispanic/Latino ethnicity. Their mean ADA risk score was 5.6 ± 1.6. Latinas who had prediabetes were older, with significantly higher rates of hypertension and a higher ADA risk score than those without prediabetes. At a risk score ≥5-the threshold for high risk set by the ADA-the questionnaire had the following test performance characteristics: sensitivity 77.8%, specificity 41.7%, PPV 76.2%, and NPV 43.9%.

CONCLUSION

The ADA risk questionnaire demonstrates reasonable performance for identifying prediabetes in a community-based sample of Latinas. Our data may guide other groups' use of this tool in the same target population. Future research should examine the effectiveness of this questionnaire for recruiting diverse populations into diabetes prevention programs. In addition, unique diabetes risk assessment tools for specific target populations are needed and may outperform questionnaires developed using nationally representative data.

Undetected dysglycaemia common in primary care patients treated for hypertension and/or dyslipidaemia: on the need for a screening strategy in clinical practice. A report from EUROASPIRE IV a registry from the EuroObservational Research Programme of the European Society of Cardiology

Background

Dysglycaemia defined as type 2 diabetes (T2DM) and impaired glucose tolerance (IGT), increases the risk of cardiovascular disease (CVD). The negative impact is more apparent in the presence of hypertension and/or dyslipidaemia. Thus, it seems reasonable to screen for dysglycaemia in patients treated for hypertension and/or dyslipidaemia. A simple screening algorithm would enhance the adoption of such strategy in clinical practice.

Objectives

To test the hypotheses (1) that dysglycaemia is common in patients with hypertension and/or dyslipidaemia and (2) that initial screening with the Finnish Diabetes Risk Score (FINDRISC) will decrease the need for laboratory based tests.

Methods

2395 patients (age 18–80 years) without (i) a history of CVD or TDM2, (ii) prescribed blood pressure and/or lipid lowering drugs answered the FINDRISC questionnaire and had an oral glucose tolerance test (OGTT) and HbA1c measured.

Results

According to the OGTT 934 (39%) had previously undetected dysglycaemia (T2DM 19%, IGT 20%). Of patients, who according to FINDRISC had a low, moderate or slightly elevated risk 20, 34 and 41% and of those in the high and very high-risk category 49 and 71% had IGT or T2DM respectively. The OGTT identified 92% of patients with T2DM, FPG + HbA1c 90%, FPG 80%, 2hPG 29% and HbA1c 22%.

Conclusions

(1) The prevalence of dysglycaemia was high in patients treated for hypertension and/or dyslipidaemia. (2) Due to the high proportion of dysglycaemia in patients with low to moderate FINDRISC risk scores its initial use did not decrease the need for subsequent glucose tests. (3) FPG was the best test for detecting T2DM. Its isolated use is limited by the inability to disclose IGT. A pragmatic strategy, decreasing the demand for an OGTT, would be to screen all patients with FPG followed by OGTT in patients with IFG.

Considerations in the design of randomized trials to screen for type 2 diabetes

Background Randomized controlled trials (RCTs) are the most robust and valid approach to evaluate screening for diseases. Many in the diabetes research community have advocated sole reliance on RCTs for designing diabetes screening policies. However, the challenges of conducting RCTs of screening for type 2 diabetes may have been underappreciated. Purpose Discuss the key theoretical concepts and practical challenges of designing and conducting RCTs of diabetes screening. Methods Narrative and critical review of the literature pertaining to the theory and practice of designing and conducting RCTs of diabetes screening. Results We present the theoretical basis of a diabetes screening trial, using concepts developed mainly in studies of cancer screening and illustrations from the Cambridge component of the Anglo Danish Dutch Study of Intensive Treatment In peOple with screeN-detected diabetes in primary care (ADDITION-Cambridge), the only extant trial of diabetes screening. We examine design issues, including the appropriate trial question, choice of design, and duration of follow-up, and address aspects of trial implementation, including recruitment, randomization, endpoint determination, sample size requirements, and screening interval. Limitations The limited number of trials of diabetes screening did not permit us to illustrate many of the practical difficulties one encounters when implementing theoretical concepts. Conclusion When diabetes screening trials are planned, we suggest careful consideration to potential areas of practical difficulty, especially the need for particularly large sample sizes and extended follow-up, and the choice of appropriate outcomes and screening intervals.

Diabetes in China: Epidemiology and Genetic Risk Factors and Their Clinical Utility in Personalized Medication

The incidence of type 2 diabetes (T2D) has rapidly increased over recent decades, and T2D has become a leading public health challenge in China. Compared with European descents, Chinese patients with T2D are diagnosed at a relatively young age and low BMI. A better understanding of the factors contributing to the diabetes epidemic is crucial for determining future prevention and intervention programs. In addition to environmental factors, genetic factors contribute substantially to the development of T2D. To date, more than 100 susceptibility loci for T2D have been identified. Individually, most T2D genetic variants have a small effect size (10-20% increased risk for T2D per risk allele); however, a genetic risk score that combines multiple T2D loci could be used to predict the risk of T2D and to identify individuals who are at a high risk. Furthermore, individualized antidiabetes treatment should be a top priority to prevent complications and mortality. In this article, we review the epidemiological trends and recent progress in the understanding of T2D genetic etiology and further discuss personalized medicine involved in the treatment of T2D.

Derivation and external validation of a clinical version of the German Diabetes Risk Score (GDRS) including measures of HbA1c

OBJECTIVE

The German Diabetes Risk Score (GDRS) is a diabetes prediction model which only includes non-invasively measured risk factors. The aim of this study was to extend the original GDRS by hemoglobin A1c (HbA1c) and validate this clinical GDRS in the nationwide German National Health Interview and Examination Survey 1998 (GNHIES98) cohort.

RESEARCH DESIGN AND METHODS

Extension of the GDRS was based on the European Prospective Investigation into Cancer and Nutrition (EPIC)-Potsdam study with baseline assessment conducted between 1994 and 1998 (N=27 548, main age range 35-65 years). Cox regression was applied with the original GDRS and HbA1c as independent variables. The extended model was evaluated by discrimination (C-index (95% CI)), calibration (calibration plots and expected to observed (E:O) ratios (95% CI)), and reclassification (net reclassification improvement, NRI (95% CI)). For validation, data from the GNHIES98 cohort with baseline assessment conducted between 1997 and 1999 were used (N=3717, age range 18-79 years). Missing data were handled with multiple imputation.

RESULTS

After 5 years of follow-up 593 incident cases of type 2 diabetes occurred in EPIC-Potsdam and 86 in the GNHIES98 cohort. In EPIC-Potsdam, the C-index for the clinical GDRS was 0.87 (0.81 to 0.92) and the overall NRI was 0.26 (0.21 to 0.30), with a stronger improvement among cases compared with non-cases (NRI_cases: 0.24 (0.19 to 0.28); NRI_non-cases: 0.02 (0.01 to 0.02)). Almost perfect calibration was observed with a slight tendency toward overestimation, which was also reflected by an E:O ratio of 1.07 (0.99 to 1.16). In the GNHIES98 cohort, discrimination was excellent with a C-index of 0.91 (0.88 to 0.94). After recalibration, the calibration plot showed underestimation of diabetes risk in the highest risk group, while the E:O ratio indicated overall perfect calibration (1.02 (0.83 to 1.26)).

CONCLUSIONS

The clinical GDRS provides the opportunity to apply the original GDRS as a first step in risk assessment, which can then be extended in clinical practice with HbA1c whenever it was measured.

Comparison of traditional diabetes risk scores and HbA1c to predict type 2 diabetes mellitus in a population based cohort study

AIMS

Compare performances of diabetes risk scores and glycated hemoglobin (HbA1c) to estimate the risk of incident type 2 diabetes mellitus (T2DM) in Northeast Germany.

METHODS

We studied 2916 subjects (20 to 81years) from the Study of Health in Pomerania (SHIP) in a 5-year follow-up period. Diabetes risk scores included the Cooperative Health Research in the Region of Augsburg (KORA) base model, the Danish diabetes risk score and the Data from the Epidemiological Study on the Insulin Resistance syndrome (D.E.S.I.R) clinical risk score. We assessed the performance of each of the diabetes risk scores and the HbA1c for 5-year risk of T2DM by the area under the receiver-operating characteristic curve (AUC) and calibration plots.

RESULTS

In SHIP, the incidence of T2DM was 5.4% (n=157) in the 5-year follow-up period. Diabetes risk scores and HbA1c achieved AUCs ranging from 0.76 for the D.E.S.I.R. clinical risk score to 0.82 for the KORA base model. For diabetes risk scores, the discriminative ability was lower for the age group 55 to 74years. For HbA1c, the discriminative ability also decreased for the group 55 to 74years while it was stable in the age group 30 to 64years old.

CONCLUSIONS

All diabetes risk scores and the HbA1c showed a good prediction for the risk of T2DM in SHIP. Which model or biomarker should be used is driven by its context of use, e.g. the practicability, implementation of interventions and availability of measurement.

Derivation and validation of the prediabetes self-assessment screening score after acute pancreatitis (PERSEUS)

BACKGROUND AND AIM

Approximately 40% of patients develop abnormal glucose metabolism after a single episode of acute pancreatitis. This study aimed to develop and validate a prediabetes self-assessment screening score for patients after acute pancreatitis.

METHODS

Data from non-overlapping training (n=82) and validation (n=80) cohorts were analysed. Univariate logistic and linear regression identified variables associated with prediabetes after acute pancreatitis. Multivariate logistic regression developed the score, ranging from 0 to 215. The area under the receiver-operating characteristic curve (AUROC), Hosmer-Lemeshow χ² statistic, and calibration plots were used to assess model discrimination and calibration. The developed score was validated using data from the validation cohort.

RESULTS

The score had an AUROC of 0.88 (95% CI, 0.80-0.97) and Hosmer-Lemeshow χ² statistic of 5.75 (p=0.676). Patients with a score of ≥75 had a 94.1% probability of having prediabetes, and were 29 times more likely to have prediabetes than those with a score of <75. The AUROC in the validation cohort was 0.81 (95% CI, 0.70-0.92) and the Hosmer-Lemeshow χ² statistic was 5.50 (p=0.599). Model calibration of the score showed good calibration in both cohorts.

CONCLUSION

The developed and validated score, called PERSEUS, is the first instrument to identify individuals who are at high risk of developing abnormal glucose metabolism following an episode of acute pancreatitis.

Validation of the Choking Risk Assessment and Pneumonia Risk Assessment for adults with Intellectual and Developmental Disability (IDD)

BACKGROUND

Risk assessments are needed to identify adults with intellectual and developmental disability (IDD) at high risk of choking and pneumonia.

AIM

To describe the development and validation of the Choking Risk Assessment (CRA) and the Pneumonia Risk Assessment (PRA) for adults with IDD.

METHODS

Test items were identified through literature review and focus groups. Five-year retrospective chart reviews identified a positive choking group (PCG), a negative choking group (NCG), a positive pneumonia group (PPG), and a negative pneumonia group (NPG). Participants were tested with the CRA and PRA by clinicians blind to these testing conditions.

RESULTS

The CRA and PRA differentiated the PCG (n=93) from the NCG (n=526) and the PPG (n=63) from the NPG (n=209) with high specificity (0.91 and 0.92 respectively) and moderate to average sensitivity (0.53 and 0.62 respectively). Further analyses revealed associations between clinical diagnoses of dysphagia and choking (p=0.043), and pneumonia (p<0.001).

CONCLUSIONS

The CRA and PRA are reliable, valid risk indicators for choking and pneumonia in adults with IDD. Precautions for mitigating choking and pneumonia risks can be applied selectively thus avoiding undue impacts on quality of life and unnecessary interventions for low risk individuals.

Abnormal skin in toe webs is a marker for abnormal glucose metabolism. A cross-sectional survey among 1,849 adults in Finland

Diabetes is undiagnosed disease and easy screening tools for it are warranted. Because foot complications are usual in diabetes, we aimed to test hypothesis that skin abnormalities are found already from patients who are not aware of having diabetes, by studying the possible association between unhealthy toe web skin and abnormal glucose metabolism. 1,849 cases without previously diagnosed diabetes participated to the 46-year follow-up study of the Northern Finland Birth Cohort. A skin investigation was performed for all, and abnormal skin findings in toe web spaces were taken as explanatory variables. Abnormal glucose tolerance was the main outcome and it was tested with an oral glucose tolerance test (OGTT), glycosylated haemoglobin fraction (HbA_1c) Values are numbers (percentages) of sub and fasting blood glucose. The participants who had any abnormal skin findings in toe webs were associated with 2.5-fold (OR 2.5, 95% CI 1.3–4.9) and 6-fold (OR 6.2, 1.4–27.6) increased risk of having previously undiagnosed diabetes detected by a 2-hour OGTT and HbA_1c, respectively. The predictive power of toe web findings was comparable with FINDRISC score. Abnormal skin findings in the toe webs show increased risk of occult diabetes, and may, thus serve as an additional sign of undiagnosed diabetes.

Individual- and Area-Level SES in Diabetes Risk Prediction: The Multi-Ethnic Study of Atherosclerosis

INTRODUCTION

The purpose of this study was to evaluate if adding SES to risk prediction models based upon traditional risk factors improves the prediction of diabetes.

METHODS

Risk prediction models without and with individual- and area-level SES predictors were compared using the prospective Multi-Ethnic Study of Atherosclerosis. Cox proportional hazards models were utilized to estimate hazard ratios for SES predictors and to generate 10-year predicted risks for 5,021 individuals without diabetes at baseline followed from 2000 to 2012. C-statistics were used to compare model discrimination, and the proportion of individuals reclassified into higher or lower risk categories with the addition of SES predictors was calculated. The accuracy of risk prediction by SES was assessed by comparing observed and predicted risks across tertiles of the SES variables. Statistical analyses were performed in 2015-2016.

RESULTS

Over a median of 9.2 years of follow-up, 615 individuals developed diabetes. Individual- and area-level SES variables did not significantly improve model discrimination or reclassify substantial numbers of individuals across risk categories. Models without SES predictors generally underestimated risk for low-SES individuals or individuals residing in low-SES areas (underestimates ranging from 0.31% to 1.07%) and overestimated risk for high-SES individuals or individuals residing in high-SES areas (overestimates ranging from 0.70% to 1.30%), and the addition of SES variables largely mitigated these differences.

CONCLUSIONS

Standard diabetes risk models may underestimate risk for low-SES individuals and overestimate risk for those of high SES. Adding SES predictors helps correct this systematic misestimation, but may not improve model discrimination.

IDF Diabetes Atlas: Global estimates for the prevalence of diabetes for 2015 and 2040

AIM

To produce current estimates of the national, regional and global impact of diabetes for 2015 and 2040.

METHODS

A systematic literature review was conducted to identify data sources on the prevalence of diabetes from studies conducted in the period from 1990 to 2015. An analytic hierarchy process was used to select the most appropriate studies for each country, and estimates for countries without data were modelled using extrapolation from similar countries that had available data. A logistic regression model was used to generate smoothed age-specific estimates, which were applied to UN population estimates.

RESULTS

540 data sources were reviewed, of which 196 sources from 111 countries were selected. In 2015 it was estimated that there were 415 million (uncertainty interval: 340-536 million) people with diabetes aged 20-79years, 5.0 million deaths attributable to diabetes, and the total global health expenditure due to diabetes was estimated at 673 billion US dollars. Three quarters (75%) of those with diabetes were living in low- and middle-income countries. The number of people with diabetes aged 20-79years was predicted to rise to 642 million (uncertainty interval: 521-829 million) by 2040.

CONCLUSION

Diabetes prevalence, deaths attributable to diabetes, and health expenditure due to diabetes continue to rise across the globe with important social, financial and health system implications.

Comparison of machine-learning algorithms to build a predictive model for detecting undiagnosed diabetes - ELSA-Brasil: accuracy study

CONTEXT AND OBJECTIVE:

Type 2 diabetes is a chronic disease associated with a wide range of serious health complications that have a major impact on overall health. The aims here were to develop and validate predictive models for detecting undiagnosed diabetes using data from the Longitudinal Study of Adult Health (ELSA-Brasil) and to compare the performance of different machine-learning algorithms in this task.

DESIGN AND SETTING:

Comparison of machine-learning algorithms to develop predictive models using data from ELSA-Brasil.

METHODS:

After selecting a subset of 27 candidate variables from the literature, models were built and validated in four sequential steps: (i) parameter tuning with tenfold cross-validation, repeated three times; (ii) automatic variable selection using forward selection, a wrapper strategy with four different machine-learning algorithms and tenfold cross-validation (repeated three times), to evaluate each subset of variables; (iii) error estimation of model parameters with tenfold cross-validation, repeated ten times; and (iv) generalization testing on an independent dataset. The models were created with the following machine-learning algorithms: logistic regression, artificial neural network, naïve Bayes, K-nearest neighbor and random forest.

RESULTS:

The best models were created using artificial neural networks and logistic regression. -These achieved mean areas under the curve of, respectively, 75.24% and 74.98% in the error estimation step and 74.17% and 74.41% in the generalization testing step.

CONCLUSION:

Most of the predictive models produced similar results, and demonstrated the feasibility of identifying individuals with highest probability of having undiagnosed diabetes, through easily-obtained clinical data.

How to improve opportunistic screening by using EMRs and other data. The prevalence of undetected diabetes mellitus in target population in Croatia

OBJECTIVES

Opportunistic screening for type 2 diabetes (T2D) has not been adopted as part of routine practice. The aim of the study was to investigate the yield of opportunistic target screening for T2D in Croatia and to evaluate the process of screening by using data from electronic medical record.

STUDY DESIGN

We conducted opportunistic screening in 23 general practitioners (GPs) in a population of 13,344 patients aged 45-70 years.

METHODS

First, after excluding patients with T2D, patients with risk factors for T2D were derived from the electronic medical record and GP's assessment during the preconsultation phase. Second, those with data about normoglycemia in past three years were excluded. Remaining patients started the consultation phase during their usual visit, when they were offered capillary fasting plasma glucose testing in the next consultation.

RESULTS

Prevalence of T2D was 10.9% (new 1.4%). A total of 5568 (46.1%) patients had risks and 2849 (51.2%) had data about normoglycemia in the last three years. Using those data, number needed to invite to screening (NNI) was reduced to half: from 46.1% to 22.5%. One hundred eighty-four patients were screened positive for T2D in two capillary fasting plasma glucose tests (yield 9.8%). Number needed to screen (NNS) in order to detect one T2D was 10.3 patients. Among risks for T2D, overweight was the best predictive factor for undiagnosed T2D (odds ratio [OR]: 2.11, confidence interval [CI]:1.41-3.15, P < .001). Logistic regression showed that in targeted population, overweight patients with a family history in fold were 2.5 times more likely to have T2D (OR: 2.54, CI 1.78-.61, P < .001).

CONCLUSIONS

Total yield in targeted population was 1,4%. By using data about normoglycemia from EMRs, NNI was reduced by half and NNS was 10.3 patients. Our findings suggest the model for improvement in opportunistic screening.

Predicting Risk of Type 2 Diabetes by Using Data on Easy-to-Measure Risk Factors

INTRODUCTION

Statistical models for assessing risk of type 2 diabetes are usually additive with linear terms that use non-nationally representative data. The objective of this study was to use nationally representative data on diabetes risk factors and spline regression models to determine the ability of models with nonlinear and interaction terms to assess the risk of type 2 diabetes.

METHODS

We used 4 waves of data (2005-2006 to 2011-2012) on adults aged 20 or older from the National Health and Nutrition Examination Survey (n = 5,471) and multivariate adaptive regression splines (MARS) to build risk models in 2015. MARS allowed for interactions among 17 noninvasively measured risk factors for type 2 diabetes.

RESULTS

A key risk factor for type 2 diabetes was increasing age, especially for those older than 69, followed by a family history of diabetes, with diminished risk among individuals younger than 45. Above age 69, other risk factors superseded age, including systolic and diastolic blood pressure. The additive MARS model with nonlinear terms had an area under curve (AUC) receiver operating characteristic of 0.847, whereas the 2-way interaction MARS model had an AUC of 0.851, a slight improvement. Both models had an 87% accuracy in classifying diabetes status.

CONCLUSION

Statistical models of type 2 diabetes risk should allow for nonlinear associations; incorporation of interaction terms into the MARS model improved its performance slightly. Robust statistical manipulation of risk factors commonly measured noninvasively in clinical settings might provide useful estimates of type 2 diabetes risk.

A community-based primary prevention programme for type 2 diabetes mellitus integrating identification and lifestyle intervention for prevention: a cluster randomised controlled trial

Background

Prevention of type 2 diabetes mellitus (T2DM) is a global priority; however, there is a lack of evidence investigating how to effectively translate prevention research into a primary care setting.

Objectives

(1) To develop and validate a risk score to identify individuals at high risk of T2DM in the UK; and (2) to establish whether or not a structured education programme targeting lifestyle and behaviour change was clinically effective and cost-effective at preventing progression to T2DM in people with prediabetes mellitus (PDM), identified through a risk score screening programme in primary care.

Design

A targeted screening study followed by a cluster randomised controlled trial (RCT), with randomisation at practice level. Participants were followed up for 3 years.

Setting

A total of 44 general practices across Leicestershire, UK. The intervention took place in the community.

Participants

A total of 17,972 individuals from 44 practices identified through the risk score as being at high risk of T2DM were invited for screening; of these, 3449 (19.2%) individuals attended. All received an oral glucose tolerance test. PDM was detected in 880 (25.5%) of those screened. Those with PDM were included in the trial; of these, 36% were female, the average age was 64 years and 16% were from an ethnic minority group.

Intervention

Practices were randomised to receive either standard care or the intervention. The intervention consisted of a 6-hour group structured education programme, with an annual refresher and regular telephone contact.

Main outcome measures

The primary outcome was progression to T2DM. The main secondary outcomes were changes in glycated haemoglobin concentrations, blood glucose levels, cardiovascular risk, the presence of metabolic syndrome, step count and the cost-effectiveness of the intervention.

Results

A total of 22.6% of the intervention group did not attend the education and 29.1% attended all sessions. A total of 131 participants developed T2DM (standard care,n = 67; intervention,n = 64). There was a 26% reduced risk of T2DM in the intervention arm compared with standard care, but this did not reach statistical significance (hazard ratio 0.74, 95% confidence interval 0.48 to 1.14;p = 0.18). There were statistically significant improvements in glycated haemoglobin concentrations, low-density lipoprotein cholesterol levels, psychosocial well-being, sedentary time and step count in the intervention group. The intervention was found to result in a net gain of 0.046 quality-adjusted life-years over 3 years at a cost of £168 per patient, with an incremental cost-effectiveness ratio of £3643 and a probability of 0.86 of being cost-effective at a willingness-to-pay threshold of £20,000.

Conclusions

We developed and validated a risk score for detecting those at high risk of undiagnosed PDM/T2DM. We screened > 3400 people using a two-stage screening programme. The RCT showed that a relatively low-resource pragmatic programme may lead to a reduction in T2DM and improved biomedical and psychosocial outcomes, and is cost-effective.

Limitations

Only 19% of those invited to screening attended, which may limit generalisability. The variation in cluster size in the RCT may have limited the power of the study.

Future work

Future work should focus on increasing attendance to both screening and prevention programmes and offering the programme in different modalities, such as web-based modalities. A longer-term follow-up of the RCT participants would be valuable.

Trial registration

Current Controlled Trials ISRCTN80605705.

Funding

The National Institute for Health Research Programme Grants for Applied Research programme.

The utility of diabetes risk score items as predictors of incident type 2 diabetes in Asian populations: An evidence-based review

BACKGROUND

The prevalence of type 2 diabetes mellitus is rising, with many Asian countries featured in the top 10 countries with the highest numbers of persons with diabetes. Reliable diabetes risk scores enable the identification of individuals at risk of developing diabetes for early intervention.

OBJECTIVES

This article aims to identify common risk factors in the risk scores with the highest discrimination; factors with the most influence on the risk score in Asian populations, and to propose a set of factors translatable to the multi-ethnic Singapore population.

METHODS

A systematic search of PubMed and EMBASE databases was conducted to identify studies published before August 2016 that developed risk prediction models for incident diabetes.

RESULTS

12 studies were identified. Risk scores that included laboratory measurements had better discrimination. Coefficient analysis showed fasting glucose and HbA1c having the greatest impact on the risk score.

CONCLUSION

A proposed Asian risk score would include: family history of diabetes, age, gender, smoking status, body mass index, waist circumference, hypertension, fasting plasma glucose, HbA1c, HDL-cholesterol and triglycerides. Future research is required on the influence of ethnicity in Singapore. The risk score may potentially be used to stratify individuals for enrolment into diabetes prevention programmes.

Prediabetes: A Worldwide Epidemic

Prediabetes, defined by blood glucose levels between normal and diabetic levels, is increasing rapidly worldwide. This abnormal physiologic state reflects the rapidly changing access to high-calorie food and decreasing levels of physical activity occurring worldwide, with resultant obesity and metabolic consequences. This is particularly marked in developing countries. Prediabetes poses several threats; there is increased risk of developing type 2 diabetes mellitus (T2DM), and there are risks inherent to the prediabetes state, including microvascular and macrovascular disease. Studies have helped to elucidate the underlying pathophysiology of prediabetes and to establish the potential for treating prediabetes and preventing T2DM.

Validation of the German Diabetes Risk Score among the general adult population: findings from the German Health Interview and Examination Surveys

OBJECTIVE

To evaluate the German Diabetes Risk Score (GDRS) among the general adult German population for prediction of incident type 2 diabetes and detection of prevalent undiagnosed diabetes.

METHODS

The longitudinal sample for prediction of incident diagnosed type 2 diabetes included 3625 persons who participated both in the examination survey in 1997-1999 and the examination survey in 2008-2011. Incident diagnosed type 2 diabetes was defined as first-time physician diagnosis or antidiabetic medication during 5 years of follow-up excluding potential incident type 1 and gestational diabetes. The cross-sectional sample for detection of prevalent undiagnosed diabetes included 6048 participants without diagnosed diabetes of the examination survey in 2008-2011. Prevalent undiagnosed diabetes was defined as glycated haemoglobin ≥6.5% (48 mmol/mol). We assessed discrimination as area under the receiver operating characteristic curve (ROC-AUC (95% CI)) and calibration through calibration plots.

RESULTS

In longitudinal analyses, 82 subjects with incident diagnosed type 2 diabetes were identified after 5 years of follow-up. For prediction of incident diagnosed diabetes, the GDRS yielded an ROC-AUC of 0.87 (0.83 to 0.90). Calibration plots indicated excellent prediction for low diabetes risk and overestimation for intermediate and high diabetes risk. When considering the entire follow-up period of 11.9 years (ROC-AUC: 0.84 (0.82 to 0.86)) and including incident undiagnosed diabetes (ROC-AUC: 0.81 (0.78 to 0.84)), discrimination decreased somewhat. A previously simplified paper version of the GDRS yielded a similar predictive ability (ROC-AUC: 0.86 (0.82 to 0.89)). In cross-sectional analyses, 128 subjects with undiagnosed diabetes were identified. For detection of prevalent undiagnosed diabetes, the ROC-AUC was 0.84 (0.81 to 0.86). Again, the simplified version yielded a similar result (ROC-AUC: 0.83 (0.80 to 0.86)).

CONCLUSIONS

The GDRS might be applied for public health monitoring of diabetes risk in the German adult population. Future research needs to evaluate whether the GDRS is useful to improve diabetes risk awareness and prevention among the general population.

Urinary N-telopeptide and Rate of Bone Loss Over the Menopause Transition and Early Postmenopause

The purpose of this study was to assess the ability of urinary N-telopeptide (U-NTX) to gauge rate of bone loss across and after the menopause transition (MT). U-NTX measurement was measured in early postmenopause in 604 participants from the Study of Women's Health Across the Nation (SWAN). We examined the association between U-NTX and annualized rates of decline in lumbar spine and femoral neck bone mineral density (BMD) across the MT (1 year before the final menstrual period [FMP] to time of U-NTX measurement), after the MT (from time of U-NTX measurement to 2 to 4 years later), and over the combined period (from 1 year before FMP to 2 to 4 years after U-NTX measurement). Adjusted for covariates in multivariable linear regression, every standard deviation (SD) increase in U-NTX was associated with 0.6% and 0.4% per year faster declines in lumbar spine and femoral neck BMD across the MT; and 0.3% (lumbar spine) and 0.2% (femoral neck) per year faster declines over the combined period (across and after the MT) (all p < 0.01). Each SD increase in U-NTX was also associated with 44% and 50% greater risk of fast bone loss in the lumbar spine (defined as BMD decline in the fastest 16% of the distribution) across the MT (p < 0.001, c-statistic = 0.80) and over the combined period (across and after the MT) (p = 0.001, c-statistic = 0.80), respectively. U-NTX measured in early postmenopause is most strongly associated with rates of bone loss across the MT, and may aid early identification of women who have experienced fast bone loss during this critical period. © 2016 American Society for Bone and Mineral Research.

A Systematic Review of Biomarkers and Risk of Incident Type 2 Diabetes: An Overview of Epidemiological, Prediction and Aetiological Research Literature

Background

Blood-based or urinary biomarkers may play a role in quantifying the future risk of type 2 diabetes (T2D) and in understanding possible aetiological pathways to disease. However, no systematic review has been conducted that has identified and provided an overview of available biomarkers for incident T2D. We aimed to systematically review the associations of biomarkers with risk of developing T2D and to highlight evidence gaps in the existing literature regarding the predictive and aetiological value of these biomarkers and to direct future research in this field.

Methods and Findings

We systematically searched PubMed MEDLINE (January 2000 until March 2015) and Embase (until January 2016) databases for observational studies of biomarkers and incident T2D according to the 2009 PRISMA guidelines. We also searched availability of meta-analyses, Mendelian randomisation and prediction research for the identified biomarkers. We reviewed 3910 titles (705 abstracts) and 164 full papers and included 139 papers from 69 cohort studies that described the prospective relationships between 167 blood-based or urinary biomarkers and incident T2D. Only 35 biomarkers were reported in large scale studies with more than 1000 T2D cases, and thus the evidence for association was inconclusive for the majority of biomarkers. Fourteen biomarkers have been investigated using Mendelian randomisation approaches. Only for one biomarker was there strong observational evidence of association and evidence from genetic association studies that was compatible with an underlying causal association. In additional search for T2D prediction, we found only half of biomarkers were examined with formal evidence of predictive value for a minority of these biomarkers. Most biomarkers did not enhance the strength of prediction, but the strongest evidence for prediction was for biomarkers that quantify measures of glycaemia.

Conclusions

This study presents an extensive review of the current state of the literature to inform the strategy for future interrogation of existing and newly described biomarkers for T2D. Many biomarkers have been reported to be associated with the risk of developing T2D. The evidence of their value in adding to understanding of causal pathways to disease is very limited so far. The utility of most biomarkers remains largely unknown in clinical prediction. Future research should focus on providing good genetic instruments across consortia for possible biomarkers in Mendelian randomisation, prioritising biomarkers for measurement in large-scale cohort studies and examining predictive utility of biomarkers for a given context.

Nurse Practitioner Perceptions of a Diabetes Risk Assessment Tool in the Retail Clinic Setting

IN BRIEF This article describes a study to gain insight into the utility and perceived feasibility of the American Diabetes Association's Diabetes Risk Test (DRT) implemented by nurse practitioners (NPs) in the retail clinic setting. The DRT is intended for those without a known risk for diabetes. Researchers invited 1,097 NPs working in the retail clinics of a nationwide company to participate voluntarily in an online questionnaire. Of the 248 NPs who sent in complete responses, 114 (46%) indicated that they used the DRT in the clinic. Overall mean responses from these NPs indicated that they perceive the DRT as a feasible tool in the retail clinic setting. Use of the DRT or similar risk assessment tools in the retail clinic setting can aid in the identification of people at risk for type 2 diabetes.

The ACE2 G8790A Polymorphism: Involvement in Type 2 Diabetes Mellitus Combined with Cerebral Stroke

BACKGROUND

We aimed to investigate the correlations between ACE2 polymorphisms and type 2 diabetes mellitus (T2DM) combined with cerebral stroke (CS).

METHODS

A total of 346 patients treated or hospitalized in our hospital were enrolled, including 181 cases without cerebrovascular complications (T2DM group) and 165 cases combined with CS (T2DM + CS group); 284 healthy individuals were selected as the control group. PCR-RFLP and ELISA were used to analyze ACE2 G8790A polymorphisms and serum ACE2 levels, respectively.

RESULTS

Significant differences were observed in the genotype/allele frequency of ACE2 G8790A between the T2DM + CS and control groups, and the T2DM and T2DM + CS groups, and in the genotype frequency of ACE2 G8790A between the T2DM and the control groups. The A allele may increase the risk of T2DM combined with CS. The AA genotype may also increase the risk of T2DM combined with CS (OR = 3.733, 95%CI = 2.069-6.738; OR = 3.597, 95%CI = 1.884-6.867). Serum ACE2 levels showed statistically significant differences among the groups. Systolic pressure and diastolic pressure were protective factors of T2DM combined with CS.

CONCLUSION

The ACE2 G8790A polymorphism in T2DM patients was correlated with CS, and the A allele might be a risk factor of T2DM combined with CS.

Validation of a nomogram for predicting regression from impaired fasting glucose to normoglycaemia to facilitate clinical decision making

BACKGROUND

In Hong Kong, fasting plasma glucose (FPG) is the most popular screening test for diabetes mellitus (DM) in primary care. Individuals with impaired fasting glucose (IFG) are commonly encountered.

OBJECTIVES

To explore the determinants of regression to normoglycaemia among primary care patients with IFG based on non-invasive variables and to establish a nomogram for the prediction of regression from IFG.

METHODS

This cohort study consisted of 1197 primary care patients with IFG. These subjects were invited to repeat a FPG test and 75-g 2-hour oral glucose tolerance test (2h-OGTT) to determine the glycaemia change. Normoglycaemia was defined as FPG <5.6 mmol/L and 2h-OGTT <7.8 mmol/L. Stepwise logistic regression model was developed to predict the regression to normoglycaemia with non-invasive variables, using a randomly selected training dataset (810 subjects). The model was validated on the remaining testing dataset (387 subjects). Area under the receiver operating characteristic curve (AUC) and Hosmer-Lemeshow test were used to evaluate discrimination and calibration of the model. A nomogram was constructed based on the model.

RESULTS

After a mean follow-up period of 6.1 months, 180 subjects (15.0%) had normoglycaemia based on the repeated FPG and 2h-OGTT results at follow-up. Subjects without central obesity or hypertension, with moderate-to-high-level physical activity and a lower baseline FPG level, were more likely to regress to normoglycaemia. The prediction model had acceptable discrimination (AUC = 0.705) and calibration (P = 0.840).

CONCLUSION

The simple-to-use nomogram could facilitate identification of subjects with low risk of progression to DM and thus aid in clinical decision making and resource prioritization in the primary care setting.

Comparison of screening scores for diabetes and prediabetes

AIMS

There are numerous risk or screening scores for the prediction of type-2 diabetes mellitus (DM). In contrast, few scores are available for preDM. In this paper, we compare the two screening scores from the American Diabetes Association (ADA) and Centers for Disease Control and Prevention (CDC) that can be used for DM as well as preDM.

METHODS

Adult participants (N=9391) without known DM from the National Health and Nutrition Examination Surveys 2009-12 were included. We fitted the factors/items in the ADA and CDC scores in logistic regression with the outcomes of undiagnosed DM, preDM, and combination, and assessed the association and discrimination accuracy. We also evaluated the suggested cutpoints that define high risk individuals. We mimicked the original models/settings but also tested various deviations/modifications often encountered in practice.

RESULTS

Both scores performed well and robustly, while the ADA score performed somewhat better (e.g., AUC=0.77 for ADA and 0.73-0.74 for CDC for DM; 0.72-0.74 and 0.70-0.71 for preDM). The same predictors and scoring rules seem to be reasonably justified with different cutpoints for DM and preDM, which can make usage easier and consistent. Some factors such as race and HDL/LDL cholesterols may be useful additions to health education.

CONCLUSIONS

Current DM education and screening focus on the prevention and management of DM. The ADA and CDC scores could further help when we identify individuals at high risk for preDM, and teach the importance of preDM during which lifestyle intervention can be effective and urgently needed.

Quantifying the Balance Between Total Bone Formation and Total Bone Resorption: An Index of Net Bone Formation

CONTEXT

Bone gain vs loss across the skeleton loss depends on the balance between total bone formation and total bone resorption.

OBJECTIVE

The objective of the study was to determine whether resorption and formation markers can be combined to gauge net bone formation across the skeleton.

DESIGN

The study included a cohort followed up across menopause transition (Study of Women's Health Across the Nation).

SETTING AND PARTICIPANTS

Community-dwelling women, 42-52 years old, premenopausal or early perimenopausal at baseline, participated in the study.

OUTCOME

The study included the following measures: 1) bone balance index (BBI) created by estimating the relationship between resorption (urinary N-telopeptide) and formation (osteocalcin) markers when the total formation equals the total resorption in 685 women with stable bone mineral density (BMD) (>5 y before the final menstrual period [FMP]) and applying this relationship to measured bone turnover markers in 216 women beginning to lose bone (≤2 y from FMP); and 2) annualized percentage declines over the following 3-4 years in the lumbar spine (LS) and femoral neck (FN) BMD.

RESULTS

Adjusted for covariates, the BBI was greater (more favorable) in women with a greater body mass index (P = .03) and lower (less favorable) in women closer to the FMP (P = .007). Each SD decrement in BBI was associated with 0.27%/y faster LS BMD decline (P 0.04) and a 38% higher odds of faster-than-average loss of LS bone mass (P = .008, c-statistic 0.76). BBI was not associated with decline in FN BMD. Urinary N-telopeptide alone was not associated with either LS or FN BMD decline.

CONCLUSIONS

An index that quantifies net bone formation vs resorption can be created from bone turnover markers and may help identify individuals at high risk for LS bone loss.

Simple non-laboratory- and laboratory-based risk assessment algorithms and nomogram for detecting undiagnosed diabetes mellitus

BACKGROUND

The aim of the present study was to develop a simple nomogram that can be used to predict the risk of diabetes mellitus (DM) in the asymptomatic non-diabetic subjects based on non-laboratory- and laboratory-based risk algorithms.

METHODS

Anthropometric data, plasma fasting glucose, full lipid profile, exercise habits, and family history of DM were collected from Chinese non-diabetic subjects aged 18-70 years. Logistic regression analysis was performed on a random sample of 2518 subjects to construct non-laboratory- and laboratory-based risk assessment algorithms for detection of undiagnosed DM; both algorithms were validated on data of the remaining sample (n = 839). The Hosmer-Lemeshow test and area under the receiver operating characteristic (ROC) curve (AUC) were used to assess the calibration and discrimination of the DM risk algorithms.

RESULTS

Of 3357 subjects recruited, 271 (8.1%) had undiagnosed DM defined by fasting glucose ≥7.0 mmol/L or 2-h post-load plasma glucose ≥11.1 mmol/L after an oral glucose tolerance test. The non-laboratory-based risk algorithm, with scores ranging from 0 to 33, included age, body mass index, family history of DM, regular exercise, and uncontrolled blood pressure; the laboratory-based risk algorithm, with scores ranging from 0 to 37, added triglyceride level to the risk factors. Both algorithms demonstrated acceptable calibration (Hosmer-Lemeshow test: P = 0.229 and P = 0.483) and discrimination (AUC 0.709 and 0.711) for detection of undiagnosed DM.

CONCLUSION

A simple-to-use nomogram for detecting undiagnosed DM has been developed using validated non-laboratory-based and laboratory-based risk algorithms.

RISK OF NEW ONSET TYPE II DM IN MDD PATIENTS RECEIVING SECOND-GENERATION ANTIPSYCHOTICS TREATMENT: A NATIONWIDE COHORT STUDY

BACKGROUND

Second-generation antipsychotics (SGA) augmentation treatment has showed better efficacy in patients with major depressive disorder (MDD). However, the association between SGA and diabetes mellitus (DM) in MDD patients deserves further investigation. The study aimed to examine the risk of new onset type II DM in MDD patients receiving SGA treatment.

METHODS

From the Psychiatric Inpatient Medical Claim Dataset, MDD patients treated with SGA continuously for more than 8 weeks were analyzed in a 1:1 propensity score matched pair sample to 1,049 patients that had never been treated with SGA. Patients were followed up to 5 years based on ICD-9 CM codes indicating incident type II DM. Cumulative incidences of type II DM were calculated and the Cox proportional hazards model with competing risk was applied to determine the risk factors for type II DM onset.

RESULTS

Cumulative incidences of new-onset type II DM between the two groups were similar. Use of SGA showed no significant increase in risk for new-onset type II DM (hazard ratio [HR] = 0.898; 95% confidence interval [CI], 0.605-1.334; P-value = 0.596). Increased risk for type II DM was shown to be associated with aging (per year) (HR = 1.039; 95% CI, 1.026-1.053; P-value < 0.001) and history of hyperlipidemia (HR = 2.323; 95% CI, 1.469-3.675; P-value < 0.001).

CONCLUSIONS

This study indicated that there is no significant difference in the risk of developing type II DM between MDD patients with and without SGA exposure. More studies focused on the benefit-risk assessment of SGA treatment in patients with MDD are warranted.

Triglyceride-glucose index (TyG index) in comparison with fasting plasma glucose improved diabetes prediction in patients with normal fasting glucose: The Vascular-Metabolic CUN cohort

AIMS

We evaluated the potential role of the triglyceride-glucose index (TyG index) as a predictor of diabetes in a White European cohort, and compared it to fasting plasma glucose (FPG) and triglycerides.

METHODS

4820 patients of the Vascular-Metabolic CUN cohort (VMCUN cohort) were examined and followed up for 8.84years (±4.39). We performed a Cox proportional hazard ratio with repeated-measures analyses to assess the risk of developing type 2 diabetes across quartiles of FPG, triglycerides and the TyG index (ln[fasting triglycerides (mg/dl)×fasting plasma glucose (mg/dl)/2]), and plotted a receiver operating characteristics (ROC) curve for discrimination.

RESULTS

There were 332 incident cases of type 2 diabetes involving 43,197.32person-years of follow-up. We observed a progressively increased risk of diabetes in subjects with TyG index levels of 8.31 or more. Among those with normal fasting glucose at baseline, <100mg/dl, subjects with the TyG index in the fourth quartile were 6.87 times more likely to develop diabetes (95% CI, 2.76-16.85; P for trend<0.001), as compared with the bottom quartile. The areas under the ROC curves (95% CI) were 0.75 (0.70-0.81) for TyG index, 0.66 (0.60-0.72) for FPG and 0.71 (0.65-0.77) for TG, in subjects with normal fasting glucose (p=0.017).

CONCLUSIONS

Our data suggest that the TyG index is useful for the early identification of individuals at risk of type 2 diabetes. The TyG index seems to be a better predictor than FPG or triglycerides of the potential development of type 2 diabetes in normoglycemic patients.

Development and Validation of a Risk-Score Model for Type 2 Diabetes: A Cohort Study of a Rural Adult Chinese Population

Some global models to predict the risk of diabetes may not be applicable to local populations. We aimed to develop and validate a score to predict type 2 diabetes mellitus (T2DM) in a rural adult Chinese population. Data for a cohort of 12,849 participants were randomly divided into derivation (n = 11,564) and validation (n = 1285) datasets. A questionnaire interview and physical and blood biochemical examinations were performed at baseline (July to August 2007 and July to August 2008) and follow-up (July to August 2013 and July to October 2014). A Cox regression model was used to weigh each variable in the derivation dataset. For each significant variable, a score was calculated by multiplying β by 100 and rounding to the nearest integer. Age, body mass index, triglycerides and fasting plasma glucose (scores 3, 12, 24 and 76, respectively) were predictors of incident T2DM. The model accuracy was assessed by the area under the receiver operating characteristic curve (AUC), with optimal cut-off value 936. With the derivation dataset, sensitivity, specificity and AUC of the model were 66.7%, 74.0% and 0.768 (95% CI 0.760–0.776), respectively. With the validation dataset, the performance of the model was superior to the Chinese (simple), FINDRISC, Oman and IDRS models of T2DM risk but equivalent to the Framingham model, which is widely applicable in a variety of populations. Our model for predicting 6-year risk of T2DM could be used in a rural adult Chinese population.

Lifestyle interventions in preventing new type 2 diabetes in Asian populations

The aim of this study was to review current evidence on interventional studies aimed at the prevention of type 2 diabetes in Asian population with lifestyle interventions. Prevalence of type 2 diabetes sharply increased in most Asian countries during the last decades. This issue has now also relevant implication for Europe where different surveys are also consistently revealing an higher prevalence of type 2 diabetes and other and major CVD risk factors among subjects originating from Asian Countries than in the native population. Nutrition and lifestyle transition seem to play a role in disclosing the predisposition for the development of type 2 diabetes and great interest is now shown toward the possibility to intervene with lifestyle intervention on at risk populations. A meta-analysis of Randomized Controlled Trials showed that lifestyle interventions are highly effective also in the Asian population. All studies were, however, conducted with an individual approach based on the identification of high-risk individuals. When ethnic minority groups have to be addressed, an approach directed to the community rather than to the individual might, however, be more effective. This review reinforces the importance for policy-makers to consider the involvement of the whole community of minority immigrant groups with lifestyle intervention programs.

Ten-year Diabetes Risk Forecast in the Capital of Jordan: Arab Diabetes Risk Assessment Questionnaire Perspective-A Strobe-Complaint Article

The prevalence of diabetes in Jordan has been increasing. The early diagnosis of diabetes is vital to slow its progression. The Arab Risk (ARABRISK) screening tool is a self-administered questionnaire used to determine people who are at high risk for developing diabetes. This study aimed to identify people at high risk for developing type 2 diabetes by using the ARABRISK in the capital of Jordan.A cross-sectional study was conducted with a convenience sample of people in the capital of Jordan. The ARABRISK screening tool was administered to identify the participants' risk for developing diabetes. In addition to descriptive statistics, percentages of the ARABRISK categories were represented, and an independent samples t test was used to explore the differences between men and women. A total of 513 participants with a mean age of 51.94 (SD = 10.33) were recruited; 64.9% of the participants were men (n = 333).The total ARABRISK score ranged from 0 to 25 with a mean score of 12.30 (SD = 4.76). Using the independent samples t test, women (mean = 13.25, SE = 0.10) had significantly higher ARABRISK total scores than men did (mean = 12.95, SE = 0.09), t(141) = -2.23, P = 0.03 in the "moderate risk" category. All of the items in the ARABRISK questionnaire were found to be good predictors of the ARABRISK total scores. Among them, age, body mass index (BMI), and high blood glucose (HBG) were the best predictors as indicated by the standardized regression coefficient (β). Older age, obesity, elevated weight circumference, absence of daily physical activity, daily consumption of fruits/vegetables, presence of high blood pressure (HBP), and HBG were significantly associated with increased odds of high ARABRISK total scores. Neither a history of gestational diabetes nor a positive family history was associated with an increased odds of high ARABRISK total scores.By identifying risk factors in these participants, interventions and lifestyle changes can be suggested and implemented to reduce the risk and incidence of diabetes.