Development and validation of a cardiovascular risk prediction model in type 1 diabetes.

AIMS/HYPOTHESIS: We aimed to report current rates of CVD in type 1 diabetes and to develop a CVD risk prediction tool for type 1 diabetes. METHODS: A cohort of 27,527 people with type 1 diabetes without prior CVD was derived from the national register in Scotland. Incident CVD events during 199,552 person-years of follow-up were ascertained using hospital admissions and death registers. A Poisson regression model of CVD was developed and then validated in the Swedish National Diabetes Register (n = 33,183). We compared the percentage with a high 10 year CVD risk (i.e., ≥10%) using the model with the percentage eligible for statins using current guidelines by age. RESULTS: The age-standardised rate of CVD per 100,000 person-years was 4070 and 3429 in men and women, respectively, with type 1 diabetes in Scotland, and 4014 and 3956 in men and women in Sweden. The final model was well calibrated (Hosmer-Lemeshow test p > 0.05) and included a further 22 terms over a base model of age, sex and diabetes duration (C statistic 0.82; 95% CI 0.81, 0.83). The model increased the base model C statistic from 0.66 to 0.80, from 0.60 to 0.75 and from 0.62 to 0.68 in those aged <40, 40-59 and ≥ 60 years, respectively (all p values <0.005). The model required minimal calibration in Sweden and had a C statistic of 0.85. Under current guidelines, >90% of those aged 20-39 years and 100% of those ≥40 years with type 1 diabetes were eligible for statins, but it was not until age 65 upwards that 100% had a modelled risk of CVD ≥10% in 10 years. CONCLUSIONS/INTERPRETATION: A prediction tool such as that developed here can provide individualised risk predictions. This 10 year CVD risk prediction tool could facilitate patient discussions regarding appropriate statin prescribing. Apart from 10 year risk, such discussions may also consider longer-term CVD risk, the potential for greater benefits from early vs later statin intervention, the potential impact on quality of life of an early CVD event and evidence on safety, all of which could influence treatment decisions, particularly in younger people with type 1 diabetes.

Use of Non-invasive Parameters and Machine-Learning Algorithms for Predicting Future Risk of Type 2 Diabetes: A Retrospective Cohort Study of Health Data From Kuwait.

Objective: In recent decades, the Arab population has experienced an increase in the prevalence of type 2 diabetes (T2DM), particularly within the Gulf Cooperation Council. In this context, early intervention programmes rely on an ability to identify individuals at risk of T2DM. We aimed to build prognostic models for the risk of T2DM in the Arab population using machine-learning algorithms vs. conventional logistic regression (LR) and simple non-invasive clinical markers over three different time scales (3, 5, and 7 years from the baseline). Design: This retrospective cohort study used three models based on LR, k-nearest neighbours (k-NN), and support vector machines (SVM) with five-fold cross-validation. The models included the following baseline non-invasive parameters: age, sex, body mass index (BMI), pre-existing hypertension, family history of hypertension, and T2DM. Setting: This study was based on data from the Kuwait Health Network (KHN), which integrated primary health and hospital laboratory data into a single system. Participants: The study included 1,837 native Kuwaiti Arab individuals (equal proportion of men and women) with mean age as 59.5 ± 11.4 years. Among them, 647 developed T2DM within 7 years of the baseline non-invasive measurements. Analytical methods: The discriminatory power of each model for classifying people at risk of T2DM within 3, 5, or 7 years and the area under the receiver operating characteristic curve (AUC) were determined. Outcome measures: Onset of T2DM at 3, 5, and 7 years. Results: The k-NN machine-learning technique, which yielded AUC values of 0.83, 0.82, and 0.79 for 3-, 5-, and 7-year prediction horizons, respectively, outperformed the most commonly used LR method and other previously reported methods. Comparable results were achieved using the SVM and LR models with corresponding AUC values of (SVM: 0.73, LR: 0.74), (SVM: 0.68, LR: 0.72), and (SVM: 0.71, LR: 0.70) for 3-, 5-, and 7-year prediction horizons, respectively. For all models, the discriminatory power decreased as the prediction horizon increased from 3 to 7 years. Conclusions: Machine-learning techniques represent a useful addition to the commonly reported LR technique. Our prognostic models for the future risk of T2DM could be used to plan and implement early prevention programmes for at risk groups in the Arab population.

Serum kidney injury molecule 1 and ß2-microglobulin perform as well as larger biomarker panels for prediction of rapid decline in renal function in type 2 diabetes.

AIMS/HYPOTHESIS: As part of the Surrogate Markers for Micro- and Macrovascular Hard Endpoints for Innovative Diabetes Tools (SUMMIT) programme we previously reported that large panels of biomarkers derived from three analytical platforms maximised prediction of progression of renal decline in type 2 diabetes. Here, we hypothesised that smaller (n ≤ 5), platform-specific combinations of biomarkers selected from these larger panels might achieve similar prediction performance when tested in three additional type 2 diabetes cohorts. METHODS: We used 657 serum samples, held under differing storage conditions, from the Scania Diabetes Registry (SDR) and Genetics of Diabetes Audit and Research Tayside (GoDARTS), and a further 183 nested case-control sample set from the Collaborative Atorvastatin in Diabetes Study (CARDS). We analysed 42 biomarkers measured on the SDR and GoDARTS samples by a variety of methods including standard ELISA, multiplexed ELISA (Luminex) and mass spectrometry. The subset of 21 Luminex biomarkers was also measured on the CARDS samples. We used the event definition of loss of >20% of baseline eGFR during follow-up from a baseline eGFR of 30-75 ml min-1 [1.73 m]-2. A total of 403 individuals experienced an event during a median follow-up of 7 years. We used discrete-time logistic regression models with tenfold cross-validation to assess association of biomarker panels with loss of kidney function. RESULTS: Twelve biomarkers showed significant association with eGFR decline adjusted for covariates in one or more of the sample sets when evaluated singly. Kidney injury molecule 1 (KIM-1) and ß2-microglobulin (B2M) showed the most consistent effects, with standardised odds ratios for progression of at least 1.4 (p < 0.0003) in all cohorts. A combination of B2M and KIM-1 added to clinical covariates, including baseline eGFR and albuminuria, modestly improved prediction, increasing the area under the curve in the SDR, Go-DARTS and CARDS by 0.079, 0.073 and 0.239, respectively. Neither the inclusion of additional Luminex biomarkers on top of B2M and KIM-1 nor a sparse mass spectrometry panel, nor the larger multiplatform panels previously identified, consistently improved prediction further across all validation sets. CONCLUSIONS/INTERPRETATION: Serum KIM-1 and B2M independently improve prediction of renal decline from an eGFR of 30-75 ml min-1 [1.73 m]-2 in type 2 diabetes beyond clinical factors and prior eGFR and are robust to varying sample storage conditions. Larger panels of biomarkers did not improve prediction beyond these two biomarkers.

Apolipoprotein CIII and N-terminal prohormone b-type natriuretic peptide as independent predictors for cardiovascular disease in type 2 diabetes.

BACKGROUND AND AIMS: Developing sparse panels of biomarkers for cardiovascular disease in type 2 diabetes would enable risk stratification for clinical decision making and selection into clinical trials. We examined the individual and joint performance of five candidate biomarkers for incident cardiovascular disease (CVD) in type 2 diabetes that an earlier discovery study had yielded. METHODS: Apolipoprotein CIII (apoCIII), N-terminal prohormone B-type natriuretic peptide (NT-proBNP), high sensitivity Troponin T (hsTnT), Interleukin-6, and Interleukin-15 were measured in baseline serum samples from the Collaborative Atorvastatin Diabetes trial (CARDS) of atorvastatin versus placebo. Among 2105 persons with type 2 diabetes and median age of 62.9 years (range 39.2-77.3), there were 144 incident CVD (acute coronary heart disease or stroke) cases during the maximum 5-year follow up. We used Cox Proportional Hazards models to identify biomarkers associated with incident CVD and the area under the receiver operating characteristic curves (AUROC) to assess overall model prediction. RESULTS: Three of the biomarkers were singly associated with incident CVD independently of other risk factors; NT-proBNP (Hazard Ratio per standardised unit 2.02, 95% Confidence Interval [CI] 1.63, 2.50), apoCIII (1.34, 95% CI 1.12, 1.60) and hsTnT (1.40, 95% CI 1.16, 1.69). When combined in a single model, only NT-proBNP and apoCIII were independent predictors of CVD, together increasing the AUROC using Framingham risk variables from 0.661 to 0.745. CONCLUSIONS: The biomarkers NT-proBNP and apoCIII substantially increment the prediction of CVD in type 2 diabetes beyond that obtained with the variables used in the Framingham risk score.

Cardiovascular Disease, Cancer, and Mortality Among People With Type 2 Diabetes and Alcoholic or Nonalcoholic Fatty Liver Disease Hospital Admission.

OBJECTIVE: To describe associations between alcoholic liver disease (ALD) or nonalcoholic fatty liver disease (NAFLD) hospital admission and cardiovascular disease (CVD), cancer, and mortality in people with type 2 diabetes mellitus (T2DM). RESEARCH DESIGN AND METHODS: We performed a retrospective cohort study by using linked population-based routine data from diabetes registry, hospital, cancer, and death records for people aged 40-89 years diagnosed with T2DM in Scotland between 2004 and 2013 who had one or more hospital admission records. Liver disease and outcomes were identified by using ICD-9 and ICD-10 codes. We estimated hazard ratios (HRs) from Cox proportional hazards regression models, adjusting for key risk factors. RESULTS: A total of 134,368 people with T2DM (1,707 with ALD and 1,452 with NAFLD) were studied, with a mean follow-up of 4.3 years for CVD and 4.7 years for mortality. Among those with ALD, NAFLD, or without liver disease hospital records 378, 320, and 21,873 CVD events; 268, 176, and 15,101 cancers; and 724, 221, and 16,203 deaths were reported, respectively. For ALD and NAFLD, respectively, adjusted HRs (95% CIs) compared with the group with no record of liver disease were 1.59 (1.43, 1.76) and 1.70 (1.52, 1.90) for CVD, 40.3 (28.8, 56.5) and 19.12 (11.71, 31.2) for hepatocellular carcinoma (HCC), 1.28 (1.12, 1.47) and 1.10 (0.94, 1.29) for non-HCC cancer, and 4.86 (4.50, 5.24) and 1.60 (1.40, 1.83) for all-cause mortality. CONCLUSIONS: Hospital records of ALD or NAFLD are associated to varying degrees with an increased risk of CVD, cancer, and mortality among people with T2DM.

Modelling cumulative exposure for inference about drug effects in observational studies.

PURPOSE: To demonstrate a modelling approach that controls for time-invariant allocation bias in estimation of associations of outcome with drug exposure. METHODS: We show that in a model that includes terms for both ever-exposure versus never-exposure and cumulative exposure, the parameter for ever-exposure represents the effect of time-invariant allocation bias, and the parameter for cumulative exposure represents the effect of the drug after adjustment for this unmeasured confounding. This assumes no stepwise effect of the drug on the event rate, no reverse causation, and no unmeasured time-varying confounders. We demonstrated this by modelling the effect of statins on cardiovascular disease, for which the true effect has been well characterised in randomised trials, using time-updated Cox regression models in a national cohort of Type 2 diabetes patients. RESULTS: The crude hazard ratio associated with ever-use of statins was 1.13 in a standard cohort analysis comparing exposed with unexposed person-time intervals. When ever-never use and cumulative exposure are modelled jointly, the effect of statins can be estimated from the cumulative exposure parameter (hazard ratio 0.97 per year of exposure, 95% CI 0.97 to 0.98). The ever-exposed term (hazard ratio 1.20, 1.16 to 1.23) in this model can be interpreted as estimating the allocation bias. CONCLUSIONS: Where stepwise effects on the risk of adverse events are unlikely, as for instance for effects on risk of cancer, joint modelling of ever-never and cumulative exposure can be used to study the effects of multiple drugs and to distinguish causal effects from confounding by allocation.

Risk of acute kidney injury and survival in patients treated with Metformin: an observational cohort study.

BACKGROUND: Whether metformin precipitates lactic acidosis in patients with chronic kidney disease (CKD) remains under debate. We examined whether metformin use was associated with an increased risk of acute kidney injury (AKI) as a proxy for lactic acidosis and whether survival among those with AKI varied by metformin exposure. METHODS: All individuals with type 2 diabetes and available prescribing data between 2004 and 2013 in Tayside, Scotland were included. The electronic health record for diabetes which includes issued prescriptions was linked to laboratory biochemistry, hospital admission, death register and Scottish Renal Registry data. AKI events were defined using the Kidney Disease Improving Global Outcomes criteria with a rise in serum creatinine of at least  26.5 µmol/l or a rise of greater than 150% from baseline for all hospital admissions. Cox Regression Analyses were used to examine whether person-time periods in which current metformin exposure occurred were associated with an increased rate of first AKI compared to unexposed periods. Cox regression was also used to compare 28 day survival rates following first AKI events in those exposed to metformin versus those not exposed. RESULTS: Twenty-five thousand one-hundred fourty-eight patients were included with a total person-time of 126,904 person years. 4944 (19.7%) people had at least one episode of AKI during the study period. There were 32.4 cases of first AKI/1000pyrs in current metformin exposed person-time periods compared to 44.9 cases/1000pyrs in unexposed periods. After adjustment for age, sex, diabetes duration, calendar time, number of diabetes drugs and baseline renal function, current metformin use was not associated with AKI incidence, HR 0.94 (95% CI 0.87, 1.02, p = 0.15). Among those with incident AKI, being on metformin at admission was associated with a higher rate of survival at 28 days (HR 0.81, 95% CI 0.69, 0.94, p = 0.006) even after adjustment for age, sex, pre-admission eGFR, HbA1c and diabetes duration. CONCLUSIONS: Contrary to common perceptions, we found no evidence that metformin increases incidence of AKI and was associated with higher 28 day survival following incident AKI.

Biomarkers of rapid chronic kidney disease progression in type 2 diabetes.

Here we evaluated the performance of a large set of serum biomarkers for the prediction of rapid progression of chronic kidney disease (CKD) in patients with type 2 diabetes. We used a case-control design nested within a prospective cohort of patients with baseline eGFR 30-60 ml/min per 1.73 m(2). Within a 3.5-year period of Go-DARTS study patients, 154 had over a 40% eGFR decline and 153 controls maintained over 95% of baseline eGFR. A total of 207 serum biomarkers were measured and logistic regression was used with forward selection to choose a subset that were maximized on top of clinical variables including age, gender, hemoglobin A1c, eGFR, and albuminuria. Nested cross-validation determined the best number of biomarkers to retain and evaluate for predictive performance. Ultimately, 30 biomarkers showed significant associations with rapid progression and adjusted for clinical characteristics. A panel of 14 biomarkers increased the area under the ROC curve from 0.706 (clinical data alone) to 0.868. Biomarkers selected included fibroblast growth factor-21, the symmetric to asymmetric dimethylarginine ratio, ß2-microglobulin, C16-acylcarnitine, and kidney injury molecule-1. Use of more extensive clinical data including prebaseline eGFR slope improved prediction but to a lesser extent than biomarkers (area under the ROC curve of 0.793). Thus we identified several novel associations of biomarkers with CKD progression and the utility of a small panel of biomarkers to improve prediction.

Association between body mass index and onset of hypertension in men and women with and without diabetes: a cross-sectional study using national health data from the State of Kuwait in the Arabian Peninsula.

OBJECTIVE: Obesity contributes directly to the risk of diabetes and hypertension. Effective management of diabetes is essential to prevent or delay the onset of comorbid hypertension. In this study, we delineate the association body mass index (BMI) has with risk and age at onset of hypertension and explore how this association is modulated by sex and the pre-existing condition of diabetes. DESIGN: Cross-sectional study using retrospective data. SETTING: Kuwait Health Network that integrates primary health and hospital laboratory data into a single system. PARTICIPANTS: We considered 3904 native Kuwaiti comorbid individuals who had the onset of type 2 diabetes prior to that of hypertension, and 1403 native Kuwaiti hypertensive individuals with no incidence of diabetes. These participants have been regularly monitored for BMI, glycated haemoglobin and blood pressure measurements. Mean variance in BMI per individual over the period from registration is seen to be low. MAIN OUTCOME MEASURES: Association between age at onset of hypertension and BMI (as measured at hypertension diagnosis); HRs for developing hypertension. RESULTS: Risk of hypertension increases with obesity levels, and is higher in patients with diabetes than in patients without diabetes but of similar obesity levels. Age at onset of hypertension is inversely related to BMI; this relationship is seen to be stronger in men compared to women (slope estimate in men, -0.62 years per unit increase in BMI; in women -0.18) and in individuals (particularly women) with diabetes compared to those without (slope estimate in women, -0.39 vs -0.18, p<0.001; in men -0.66 vs -0.62; p=0.66). CONCLUSIONS: The observation that the presence of diabetes doubles the slope of inverse relationship between hypertension onset age and BMI in women (while the slope is high in men irrespective of diabetes status) leads to a possible proposition that pre-existing diabetes narrows down sex-specific differences in the impact of obesity on blood pressure.

Impact of hypertension on the association of BMI with risk and age at onset of type 2 diabetes mellitus: age- and gender-mediated modifications.

AIMS: Given that BMI correlates with risk of Type 2 diabetes mellitus (T2DM), and that hypertension is a common comorbid condition, we hypothesize that hypertension augments significantly the impact of obesity on T2DM onset. METHODS: We obtained data on T2DM in Kuwaiti natives from Kuwait Health Network Registry. We considered 1339 comorbid individuals with onset of hypertension preceding that of T2DM, and 3496 non-hypertensive individuals but with T2DM. Multiple linear regressions, ANOVA tests, and Cox proportional hazards models were used to quantify the impact of hypertension on correlation of BMI with age at onset and risk of T2DM. RESULTS: Impact of increasing levels of BMI on age at onset ot T2DM is seen augmented in patients diagnosed with hypertension. We find that the slope of the inverse linear relationship between BMI and onset age of T2DM is much steep in hypertensive patients (-0.69, males and -0.39, females) than in non-hypertensive patients (-0.36, males and -0.17, females). The decline in onset age for an unit increase of BMI is two-fold in males than in females. Upon considering BMI as a categorical variable, we find that while the mean onset age of T2DM in hypertensive patients decreases by as much as 5-12 years in every higher BMI categories, significant decrease in non-hypertensive patients exists only when severely obese. Hazard due to hypertension (against the baseline of non-hypertension and normal weight) increases at least two-fold in every obese category. While males have higher hazard due to hypertension in early adulthood, females have higher hazard in late adulthood. CONCLUSION: Pre-existing condition of hypertension augments the association of BMI with Type 2 diabetes onset in both males and females. The presented results provide health professionals directives on the extent of weight-loss required to delay onset of Type 2 diabetes in hypertensive versus non-hypertensive patients.

Predictive models to assess risk of type 2 diabetes, hypertension and comorbidity: machine-learning algorithms and validation using national health data from Kuwait--a cohort study.

OBJECTIVE: We build classification models and risk assessment tools for diabetes, hypertension and comorbidity using machine-learning algorithms on data from Kuwait. We model the increased proneness in diabetic patients to develop hypertension and vice versa. We ascertain the importance of ethnicity (and natives vs expatriate migrants) and of using regional data in risk assessment. DESIGN: Retrospective cohort study. Four machine-learning techniques were used: logistic regression, k-nearest neighbours (k-NN), multifactor dimensionality reduction and support vector machines. The study uses fivefold cross validation to obtain generalisation accuracies and errors. SETTING: Kuwait Health Network (KHN) that integrates data from primary health centres and hospitals in Kuwait. PARTICIPANTS: 270 172 hospital visitors (of which, 89 858 are diabetic, 58 745 hypertensive and 30 522 comorbid) comprising Kuwaiti natives, Asian and Arab expatriates. OUTCOME MEASURES: Incident type 2 diabetes, hypertension and comorbidity. RESULTS: Classification accuracies of >85% (for diabetes) and >90% (for hypertension) are achieved using only simple non-laboratory-based parameters. Risk assessment tools based on k-NN classification models are able to assign 'high' risk to 75% of diabetic patients and to 94% of hypertensive patients. Only 5% of diabetic patients are seen assigned 'low' risk. Asian-specific models and assessments perform even better. Pathological conditions of diabetes in the general population or in hypertensive population and those of hypertension are modelled. Two-stage aggregate classification models and risk assessment tools, built combining both the component models on diabetes (or on hypertension), perform better than individual models. CONCLUSIONS: Data on diabetes, hypertension and comorbidity from the cosmopolitan State of Kuwait are available for the first time. This enabled us to apply four different case-control models to assess risks. These tools aid in the preliminary non-intrusive assessment of the population. Ethnicity is seen significant to the predictive models. Risk assessments need to be developed using regional data as we demonstrate the applicability of the American Diabetes Association online calculator on data from Kuwait.