An omics-based machine learning approach to predict diabetes progression: a RHAPSODY study.

AIMS/HYPOTHESIS: People with type 2 diabetes are heterogeneous in their disease trajectory, with some progressing more quickly to insulin initiation than others. Although classical biomarkers such as age, HbA1c and diabetes duration are associated with glycaemic progression, it is unclear how well such variables predict insulin initiation or requirement and whether newly identified markers have added predictive value. METHODS: In two prospective cohort studies as part of IMI-RHAPSODY, we investigated whether clinical variables and three types of molecular markers (metabolites, lipids, proteins) can predict time to insulin requirement using different machine learning approaches (lasso, ridge, GRridge, random forest). Clinical variables included age, sex, HbA1c, HDL-cholesterol and C-peptide. Models were run with unpenalised clinical variables (i.e. always included in the model without weights) or penalised clinical variables, or without clinical variables. Model development was performed in one cohort and the model was applied in a second cohort. Model performance was evaluated using Harrel's C statistic. RESULTS: Of the 585 individuals from the Hoorn Diabetes Care System (DCS) cohort, 69 required insulin during follow-up (1.0-11.4 years); of the 571 individuals in the Genetics of Diabetes Audit and Research in Tayside Scotland (GoDARTS) cohort, 175 required insulin during follow-up (0.3-11.8 years). Overall, the clinical variables and proteins were selected in the different models most often, followed by the metabolites. The most frequently selected clinical variables were HbA1c (18 of the 36 models, 50%), age (15 models, 41.2%) and C-peptide (15 models, 41.2%). Base models (age, sex, BMI, HbA1c) including only clinical variables performed moderately in both the DCS discovery cohort (C statistic 0.71 [95% CI 0.64, 0.79]) and the GoDARTS replication cohort (C 0.71 [95% CI 0.69, 0.75]). A more extensive model including HDL-cholesterol and C-peptide performed better in both cohorts (DCS, C 0.74 [95% CI 0.67, 0.81]; GoDARTS, C 0.73 [95% CI 0.69, 0.77]). Two proteins, lactadherin and proto-oncogene tyrosine-protein kinase receptor, were most consistently selected and slightly improved model performance. CONCLUSIONS/INTERPRETATION: Using machine learning approaches, we show that insulin requirement risk can be modestly well predicted by predominantly clinical variables. Inclusion of molecular markers improves the prognostic performance beyond that of clinical variables by up to 5%. Such prognostic models could be useful for identifying people with diabetes at high risk of progressing quickly to treatment intensification. DATA AVAILABILITY: Summary statistics of lipidomic, proteomic and metabolomic data are available from a Shiny dashboard at https://rhapdata-app.vital-it.ch .

Systematic Evaluation of Pleiotropy Identifies 6 Further Loci Associated With Coronary Artery Disease.

BACKGROUND: Genome-wide association studies have so far identified 56 loci associated with risk of coronary artery disease (CAD). Many CAD loci show pleiotropy; that is, they are also associated with other diseases or traits. OBJECTIVES: This study sought to systematically test if genetic variants identified for non-CAD diseases/traits also associate with CAD and to undertake a comprehensive analysis of the extent of pleiotropy of all CAD loci. METHODS: In discovery analyses involving 42,335 CAD cases and 78,240 control subjects we tested the association of 29,383 common (minor allele frequency >5%) single nucleotide polymorphisms available on the exome array, which included a substantial proportion of known or suspected single nucleotide polymorphisms associated with common diseases or traits as of 2011. Suggestive association signals were replicated in an additional 30,533 cases and 42,530 control subjects. To evaluate pleiotropy, we tested CAD loci for association with cardiovascular risk factors (lipid traits, blood pressure phenotypes, body mass index, diabetes, and smoking behavior), as well as with other diseases/traits through interrogation of currently available genome-wide association study catalogs. RESULTS: We identified 6 new loci associated with CAD at genome-wide significance: on 2q37 (KCNJ13-GIGYF2), 6p21 (C2), 11p15 (MRVI1-CTR9), 12q13 (LRP1), 12q24 (SCARB1), and 16q13 (CETP). Risk allele frequencies ranged from 0.15 to 0.86, and odds ratio per copy of the risk allele ranged from 1.04 to 1.09. Of 62 new and known CAD loci, 24 (38.7%) showed statistical association with a traditional cardiovascular risk factor, with some showing multiple associations, and 29 (47%) showed associations at p < 1 × 10-4 with a range of other diseases/traits. CONCLUSIONS: We identified 6 loci associated with CAD at genome-wide significance. Several CAD loci show substantial pleiotropy, which may help us understand the mechanisms by which these loci affect CAD risk.

Liothyronine use in a 17 year observational population-based study - the tears study.

OBJECTIVE: To look at adverse outcomes for patients on liothyronine compared to l-thyroxine. Some trials have examined the relative merits of liothyronine but none have looked at adverse outcomes in large numbers. STUDY DESIGN: An observational study of all patients prescribed thyroid hormone replacement in Tayside Scotland (population 400 000) from 1997 to 2014. PATIENTS: A study group of patients having ever used liothyronine (n = 400) was compared to those who had only used l-thyroxine (n = 33 955). All patients were followed up until end-point, death or leaving Tayside. MEASUREMENTS: Mortality rates and admissions with cardiovascular disease, atrial fibrillation, fractures, breast cancer and mental diseases were compared. Incident use of bisphosphonates, statins, antidepressants and antipsychotics was compared. RESULTS: Compared to patients only taking l-thyroxine, those using liothyronine had no increased risk of cardiovascular disease [hazard ratio (HR) 1·04; 95% CI 0·70-1·54], atrial fibrillation (HR 0·91: 0·47-1·75), or fractures (HR 0·79: 0·49-1·27) after adjusting for age. There was no difference in the number of prescriptions for bisphosphonates or statins. There was an increased risk of new prescriptions for antipsychotic medication (HR 2·26: 1·64-3·11 P < 0·0001) which was proportional to the number of liothyronine prescriptions. There was a non-significant trend towards an increase in breast cancer and new use of antidepressant medications. During follow-up, median TSH was higher for patients on l-thyroxine alone (2·08 vs 1·07 mU/L; P < 0·001). CONCLUSION: For patients taking long-term liothyronine we did not identify any additional risk of atrial fibrillation, cardiovascular disease or fractures. There was an increased incident use of antipsychotic medication during follow-up.

Paradoxical lower serum triglyceride levels and higher type 2 diabetes mellitus susceptibility in obese individuals with the PNPLA3 148M variant.

BACKGROUND: Obesity is highly associated with elevated serum triglycerides, hepatic steatosis and type 2 diabetes (T2D). The I148M (rs738409) genetic variant of patatin-like phospholipase domain-containing 3 gene (PNPLA3) is known to modulate hepatic triglyceride accumulation, leading to steatosis. No association between PNPLA3 I148M genotype and T2D in Europeans has been reported. Aim of this study is to examine the relationship between PNPLA3 I148M genotypes and serum triglycerides, insulin resistance and T2D susceptibility by testing a gene-environment interaction model with severe obesity. METHODS AND FINDINGS: PNPLA3 I148M was genotyped in a large obese cohort, the SOS study (n = 3,473) and in the Go-DARTS (n = 15,448), a T2D case-control study. Metabolic parameters were examined across the PNPLA3 I148M genotypes in participants of the SOS study at baseline and at 2- and 10-year follow up after bariatric surgery or conventional therapy. The associations with metabolic parameters were validated in the Go-DARTS study. Serum triglycerides were found to be lower in the PNPLA3 148M carriers from the SOS study at baseline and from the Go-DARTS T2D cohort. An increased risk for T2D conferred by the 148M allele was found in the SOS study (O.R. 1.09, 95% C.I. 1.01-1.39, P = 0.040) and in severely obese individuals in the Go-DARTS study (O.R. 1.37, 95% C.I. 1.13-1.66, P = 0.001). The 148M allele was no longer associated with insulin resistance or T2D after bariatric surgery in the SOS study and no association with the 148M allele was observed in the less obese (BMI<35) individuals in the Go-DARTS study (P for interaction  = 0.002). This provides evidence for the obesity interaction with I48M allele and T2D risk in a large-scale cross-sectional and a prospective interventional study. CONCLUSIONS: Severely obese individuals carrying the PNPLA3 148M allele have lower serum triglyceride levels, are more insulin resistant and more susceptible to T2D. This study supports the hypothesis that obesity-driven hepatic lipid accumulation may contribute to T2D susceptibility.

Common variants near ATM are associated with glycemic response to metformin in type 2 diabetes.

Metformin is the most commonly used pharmacological therapy for type 2 diabetes. We report a genome-wide association study for glycemic response to metformin in 1,024 Scottish individuals with type 2 diabetes with replication in two cohorts including 1,783 Scottish individuals and 1,113 individuals from the UK Prospective Diabetes Study. In a combined meta-analysis, we identified a SNP, rs11212617, associated with treatment success (n = 3,920, P = 2.9 × 10(-9), odds ratio = 1.35, 95% CI 1.22-1.49) at a locus containing ATM, the ataxia telangiectasia mutated gene. In a rat hepatoma cell line, inhibition of ATM with KU-55933 attenuated the phosphorylation and activation of AMP-activated protein kinase in response to metformin. We conclude that ATM, a gene known to be involved in DNA repair and cell cycle control, plays a role in the effect of metformin upstream of AMP-activated protein kinase, and variation in this gene alters glycemic response to metformin.

Peroxisome proliferator-activated receptor-delta genotype influences metabolic phenotype and may influence lipid response to statin therapy in humans: a genetics of diabetes audit and research Tayside study.

CONTEXT: Previous studies have identified a single-nucleotide polymorphism in the gene encoding peroxisome proliferator-activated receptor-delta (PPARD), rs2016520, that is associated with changes in metabolic disease in some but not all studies, which suggests that PPARD agonists may have therapeutic benefits for the treatment of metabolic disorders, including dyslipidemia, type 2 diabetes, and obesity. OBJECTIVE: The objective of the study was to determine whether rs2016520 or other single-nucleotide polymorphism in the PPARD locus influenced the risk of developing various characteristics of metabolic disease. DESIGN: Haplotype tagging analysis across PPARD was performed in 11,074 individuals from the Welcome Trust U.K. Type 2 Diabetes Case Control Collection. RESULTS: In subjects with and without type 2 diabetes, rs2016520 was associated with body mass index, high-density lipoprotein cholesterol, leptin, and TNFalpha and was dependent on gender. CONCLUSION: The current results suggest differential effects of PPARdelta in males and females.

The FTO gene is associated with an atherogenic lipid profile and myocardial infarction in patients with type 2 diabetes: a Genetics of Diabetes Audit and Research Study in Tayside Scotland (Go-DARTS) study.

BACKGROUND: Common variation in the fat mass and obesity (FTO)-related gene is associated with increased body fat and susceptibility to type 2 diabetes. We hypothesized that this would also associate with metabolic phenotypes of insulin resistance and increased risk of cardiovascular morbidity and mortality. METHODS AND RESULTS: FTO rs9939609 genotype was determined in 4897 patients with type 2 diabetes in the prospective Genetics of Diabetes Audit and Research Study in Tayside Scotland study. The A allele was associated with lower plasma high-density lipoprotein cholesterol (mean difference, 0.03 mmol/L; P=0.008), higher triglycerides (0.1 mmol/L, P=0.007), higher atherogenic index of plasma (0.03, P=0.003), and, as expected, increased body mass index (0.77 kg/m(2), P=8.8 x 10(-6)). During a mean follow-up of 3.6 years, the A allele was also associated with increased risk (hazard ratio, 2.36; CI, 1.49 to 3.74; P=0.0002) of fatal and nonfatal myocardial infarction (total of 324 events) in a model, including baseline age, gender, prevalent myocardial infarction, smoking status, statin, and insulin use. This association diminished but remained significant when obesity-related traits, such as body mass index, glycohemoglobin, and lipid parameters, were also included (hazard ratio, 2.01; CI, 1.18 to 3.45, P=0.011). There was a strong interaction of FTO genotype and statin use and cardiovascular outcome (P=0.001), such that cardiovascular morbidity and mortality was completely abrogated in individuals who were prescribed statins. CONCLUSIONS: The increased fat mass in carriers of the A allele of rs9939609 of FTO is associated not only with increased risk of type 2 diabetes, but also with an increase in atherogenic lipid profile and myocardial infarction in these patients. This variant may, therefore, in the future contribute to more effective targeting of specific preventative therapy.

New users of metformin are at low risk of incident cancer: a cohort study among people with type 2 diabetes.

OBJECTIVE: The antidiabetic properties of metformin are mediated through its ability to activate the AMP-activated protein kinase (AMPK). Activation of AMPK can suppress tumor formation and inhibit cell growth in addition to lowering blood glucose levels. We tested the hypothesis that metformin reduces the risk of cancer in people with type 2 diabetes. RESEARCH DESIGN AND METHODS: In an observational cohort study using record-linkage databases and based in Tayside, Scotland, U.K., we identified people with type 2 diabetes who were new users of metformin in 1994-2003. We also identified a set of diabetic comparators, individually matched to the metformin users by year of diabetes diagnosis, who had never used metformin. In a survival analysis we calculated hazard ratios for diagnosis of cancer, adjusted for baseline characteristics of the two groups using Cox regression. RESULTS: Cancer was diagnosed among 7.3% of 4,085 metformin users compared with 11.6% of 4,085 comparators, with median times to cancer of 3.5 and 2.6 years, respectively (P < 0.001). The unadjusted hazard ratio (95% CI) for cancer was 0.46 (0.40-0.53). After adjusting for sex, age, BMI, A1C, deprivation, smoking, and other drug use, there was still a significantly reduced risk of cancer associated with metformin: 0.63 (0.53-0.75). CONCLUSIONS: These results suggest that metformin use may be associated with a reduced risk of cancer. A randomized trial is needed to assess whether metformin is protective in a population at high risk for cancer.

SLC2A9 is a newly identified urate transporter influencing serum urate concentration, urate excretion and gout.

Uric acid is the end product of purine metabolism in humans and great apes, which have lost hepatic uricase activity, leading to uniquely high serum uric acid concentrations (200-500 microM) compared with other mammals (3-120 microM). About 70% of daily urate disposal occurs via the kidneys, and in 5-25% of the human population, impaired renal excretion leads to hyperuricemia. About 10% of people with hyperuricemia develop gout, an inflammatory arthritis that results from deposition of monosodium urate crystals in the joint. We have identified genetic variants within a transporter gene, SLC2A9, that explain 1.7-5.3% of the variance in serum uric acid concentrations, following a genome-wide association scan in a Croatian population sample. SLC2A9 variants were also associated with low fractional excretion of uric acid and/or gout in UK, Croatian and German population samples. SLC2A9 is a known fructose transporter, and we now show that it has strong uric acid transport activity in Xenopus laevis oocytes.

Apolipoprotein E genotypes are associated with lipid-lowering responses to statin treatment in diabetes: a Go-DARTS study.

BACKGROUND: Apolipoprotein E (APOE) genotypes have been associated with variations in plasma-lipid levels and with response to statins, although the influence of APOE on the response to statins remains controversial, especially in patients with diabetes. We sought to evaluate the association of the APOE genotype with the low-density lipoprotein cholesterol (LDLc)-lowering response to statins, in a large population-based cohort of patients with diabetes. METHODS AND RESULTS: A total of 1383 patients, commencing statins between 1990 and 2006, were identified from the Genetics of Diabetes Audit and Research in Tayside Scotland database. Statin response was determined both by the minimum LDLc achieved, and by the failure of the patients to reach a clinical target LDLc (< or =2 mmol/l). APOE genotype and potential confounding covariates were entered into the linear and logistic regression models. RESULTS: We found an association of APOE genotypes with both baseline and treatment responses. E2 homozygotes achieved lower LDLc levels (mean 0.6; confidence interval: 0.1-1.1 mmol/l) than E4 homozygotes (mean 1.7; confidence interval: 1.4-1.9 mmol/l; P=2.96 x 10). Minimum LDLc was associated by a linear trend with genotype. This relationship remained statistically significant after adjustment for baseline LDLc, adherence, duration, dose, smoking, and age. None of the E2 homozygotes failed to reach the target LDLc, compared with 32% of the E4 homozygotes (P=5.3 x 10). CONCLUSION: This study demonstrates the potential clinical value of the APOE genotype as a robust marker for LDLc responses to statin drugs, which might contribute to the identification of a particularly drug-resistant subgroup of patients. This marker provides information over and above baseline lipid levels.