Deep Neural Network for Reducing the Screening Workload in Systematic Reviews for Clinical Guidelines: Algorithm Validation Study.

BACKGROUND: Performing systematic reviews is a time-consuming and resource-intensive process. OBJECTIVE: We investigated whether a machine learning system could perform systematic reviews more efficiently. METHODS: All systematic reviews and meta-analyses of interventional randomized controlled trials cited in recent clinical guidelines from the American Diabetes Association, American College of Cardiology, American Heart Association (2 guidelines), and American Stroke Association were assessed. After reproducing the primary screening data set according to the published search strategy of each, we extracted correct articles (those actually reviewed) and incorrect articles (those not reviewed) from the data set. These 2 sets of articles were used to train a neural network-based artificial intelligence engine (Concept Encoder, Fronteo Inc). The primary endpoint was work saved over sampling at 95% recall (WSS@95%). RESULTS: Among 145 candidate reviews of randomized controlled trials, 8 reviews fulfilled the inclusion criteria. For these 8 reviews, the machine learning system significantly reduced the literature screening workload by at least 6-fold versus that of manual screening based on WSS@95%. When machine learning was initiated using 2 correct articles that were randomly selected by a researcher, a 10-fold reduction in workload was achieved versus that of manual screening based on the WSS@95% value, with high sensitivity for eligible studies. The area under the receiver operating characteristic curve increased dramatically every time the algorithm learned a correct article. CONCLUSIONS: Concept Encoder achieved a 10-fold reduction of the screening workload for systematic review after learning from 2 randomly selected studies on the target topic. However, few meta-analyses of randomized controlled trials were included. Concept Encoder could facilitate the acquisition of evidence for clinical guidelines.

Sodium-glucose co-transporter-2 inhibitors as add-on therapy to insulin for type 1 diabetes mellitus: Systematic review and meta-analysis of randomized controlled trials.

New treatments for type 1 diabetes are an unmet need. We investigated the efficacy and safety of adding sodium-glucose co-transporter-2 (SGLT2) inhibitors to insulin for type 1 diabetes by conducting a meta-analysis of prospective randomized, placebo-controlled trials. A search of electronic databases up to October 2017 identified 1361 studies, of which 14 were investigated (N = 4591). Meta-analysis showed that SGLT2 inhibitor therapy significantly reduced glycated haemoglobin (HbA1c) concentration by 0.4% (95% confidence interval [CI] 0.35, 0.46; P < .001, I2 = 0%), fasting plasma glucose by 1.14 mmol/L (95% CI 0.8,1.47), body weight by 2.68 kg (95% CI 2.0, 3.36), and systolic blood pressure by 3.37 mmHg (95% CI 1.46, 5.28). In addition, bolus insulin decreased by 3.6 units/day (95% CI 2.0, 5.3), and basal insulin decreased by 4.2 units/day (95% CI 2.2, 6.3). Continuous glucose monitoring showed a decrease in glucose excursions compared with placebo, with reduced variation of mean blood glucose, glucose standard deviation, and mean amplitude of glucose excursion. There was no significant increase in the rate of hypoglycaemia or severe hypoglycaemia; however, SGLT2 inhibitor therapy increased diabetic ketoacidosis (odds ratio [OR] 3.38) and genital tract infection (OR 3.44). Add-on SGLT2 inhibitor therapy might be advantageous for type 1 diabetes, but its use should be considered carefully.

Biosimilar vs originator insulins: Systematic review and meta-analysis.

Biosimilar insulins have expanded the treatment options for diabetes. We compared the clinical efficacy and safety of biosimilar insulins with those of originator insulins by conducting a meta-analysis. A random-effects meta-analysis was performed on randomized controlled trials comparing biosimilar and originator insulins in adults with diabetes. Studies were obtained by searching electronic databases up to December 2017. Ten trials, in a total of 4935 patients, were assessed (2 trials each on LY2963016, MK-1293, Mylan's insulin glargine and SAR342434, and 1 trial each on FFP-112 and Basalog). The meta-analysis found no differences between long-acting biosimilar and originator insulins with regard to reduction in glycated haemoglobin at 24 weeks (0.04%, 95% confidence interval [CI] -0.01, 0.08; P for efficacy = .14, I2 = 0%) or at 52 weeks (0.03%, 95% CI -0.04, 0.1), or reduction in fasting plasma glucose (0.08 mmol/L, 95% CI 0.36, 0.53), hypoglycaemia (odds ratio 0.99, 95% CI 0.96, 1.03), mortality, injection site reactions, insulin antibodies and allergic reactions. Analyses stratified by type of diabetes and prior insulin use yielded similar findings. Similarly, no significant differences were found between short-acting biosimilar and originator insulins. In summary, our meta-analysis showed no significant differences in clinical efficacy and safety, including immune reactions, between biosimilar and originator insulins. Biosimilar insulins can increase access to modern insulin therapy and reduce medical costs.

Genome-wide association study identifies three novel loci for type 2 diabetes.

Although over 60 loci for type 2 diabetes (T2D) have been identified, there still remains a large genetic component to be clarified. To explore unidentified loci for T2D, we performed a genome-wide association study (GWAS) of 6 209 637 single-nucleotide polymorphisms (SNPs), which were directly genotyped or imputed using East Asian references from the 1000 Genomes Project (June 2011 release) in 5976 Japanese patients with T2D and 20 829 nondiabetic individuals. Nineteen unreported loci were selected and taken forward to follow-up analyses. Combined discovery and follow-up analyses (30 392 cases and 34 814 controls) identified three new loci with genome-wide significance, which were MIR129-LEP [rs791595; risk allele = A; risk allele frequency (RAF) = 0.080; P = 2.55 × 10(-13); odds ratio (OR) = 1.17], GPSM1 [rs11787792; risk allele = A; RAF = 0.874; P = 1.74 × 10(-10); OR = 1.15] and SLC16A13 (rs312457; risk allele = G; RAF = 0.078; P = 7.69 × 10(-13); OR = 1.20). This study demonstrates that GWASs based on the imputation of genotypes using modern reference haplotypes such as that from the 1000 Genomes Project data can assist in identification of new loci for common diseases.

A single-nucleotide polymorphism in ANK1 is associated with susceptibility to type 2 diabetes in Japanese populations.

To identify a novel susceptibility locus for type 2 diabetes, we performed an imputation-based, genome-wide association study (GWAS) in a Japanese population using newly obtained imputed-genotype data for 2 229 890 single-nucleotide polymorphisms (SNPs) estimated from previously reported, directly genotyped GWAS data in the same samples (stage 1: 4470 type 2 diabetes versus 3071 controls). We directly genotyped 43 new SNPs with P-values of <10(-4) in a part of stage-1 samples (2692 type 2 diabetes versus 3071 controls), and the associations of validated SNPs were evaluated in another 11 139 Japanese individuals (stage 2: 7605 type 2 diabetes versus 3534 controls). Combined meta-analysis using directly genotyped data for stages 1 and 2 revealed that rs515071 in ANK1 and rs7656416 near MGC21675 were associated with type 2 diabetes in the Japanese population at the genome-wide significant level (P < 5 × 10(-8)). The association of rs515071 was also observed in European GWAS data (combined P for all populations = 6.14 × 10(-10)). Rs7656416 was in linkage disequilibrium to rs6815464, which had recently been identified as a top signal in a meta-analysis of East Asian GWAS for type 2 diabetes (r(2) = 0.76 in stage 2). The association of rs7656416 with type 2 diabetes disappeared after conditioning on rs6815464. These results indicate that the ANK1 locus is a new, common susceptibility locus for type 2 diabetes across different ethnic groups. The signal of association was weaker in the directly genotyped data, so the improvement in signal indicates the importance of imputation in this particular case.

Genetic architecture of type 2 diabetes.

Genome-wide association studies (GWAS) have identified over 70 loci associated with type 2 diabetes (T2D). Most genetic variants associated with T2D are common variants with modest effects on T2D and are shared with major ancestry groups. To what extent the genetic component of T2D can be explained by common variants relies upon the shape of the genetic architecture of T2D. Fine mapping utilizing populations with different patterns of linkage disequilibrium and functional annotation derived from experiments in relevant tissues are mandatory to track down causal variants responsible for the pathogenesis of T2D.

Progress in genetics of type 2 diabetes and diabetic complications.

Type 2 diabetes results from a complex interaction between genetic and environmental factors. Precision medicine for type 2 diabetes using genetic data is expected to predict the risk of developing diabetes and complications and to predict the effects of medications and life-style intervention more accurately for individuals. Genome-wide association studies (GWAS) have been conducted in European and Asian populations and new genetic loci have been identified that modulate the risk of developing type 2 diabetes. Novel loci were discovered by GWAS in diabetic complications with increasing sample sizes. Large-scale genome-wide association analysis and polygenic risk scores using biobank information is making it possible to predict the development of type 2 diabetes. In the ADVANCE clinical trial of type 2 diabetes, a multi-polygenic risk score was useful to predict diabetic complications and their response to treatment. Proteomics and metabolomics studies have been conducted and have revealed the associations between type 2 diabetes and inflammatory signals and amino acid synthesis. Using multi-omics analysis, comprehensive molecular mechanisms have been elucidated to guide the development of targeted therapy for type 2 diabetes and diabetic complications.

Identification of serum metabolome signatures associated with retinal and renal complications of type 2 diabetes.

BACKGROUND: Type 2 diabetes is a common disease around the world and its major complications are diabetic retinopathy (DR) and diabetic kidney disease (DKD). Persons with type 2 diabetes with complications, especially who have both DR and DKD, have poorer prognoses than those without complications. Therefore, prevention and early identification of the complications of type 2 diabetes are necessary to improve the prognosis of persons with type 2 diabetes. The aim of this study is to identify factors associated with the development of multiple complications of type 2 diabetes. METHODS: We profiled serum metabolites of persons with type 2 diabetes with both DR and DKD (N = 141) and without complications (N = 159) using a comprehensive non-targeted metabolomics approach with mass spectrometry. Based on the serum metabolite profiles, case-control comparisons and metabolite set enrichment analysis (MSEA) were performed. RESULTS: Here we show that five metabolites (cyclohexylamine, P = 4.5 × 10-6; 1,2-distearoyl-glycero-3-phosphocholine, P = 7.3 × 10-6; piperidine, P = 4.8 × 10-4; N-acetylneuraminic acid, P = 5.1 × 10-4; stearoyl ethanolamide, P = 6.8 × 10-4) are significantly increased in those with the complications. MSEA identifies fatty acid biosynthesis as the type 2 diabetes complications-associated biological pathway (P = 0.0020). CONCLUSIONS: Our metabolome analysis identifies the serum metabolite features of the persons with type 2 diabetes with multiple complications, which could potentially be used as biomarkers.

In the management of type 2 diabetes, prevention and early identification of diabetes complications are important. In particular, people with type 2 diabetes with diabetic retinopathy (DR), affecting the eye, and diabetic kidney disease (DKD), have poorer outcomes than those without complications and need early intervention. Here, we comprehensively profiled blood metabolites, or breakdown products of the biological processes occurring in the body, of people with type 2 diabetes with both DR and DKD and those without complications. We found that five metabolites were significantly increased in those with complications, and we identified a specific metabolic pathway associated with having complications. Our analysis identified the blood metabolite features of people with type 2 diabetes with multiple complications, which could potentially be used as markers in the future.

Variations with modest effects have an important role in the genetic background of type 2 diabetes and diabetes-related traits.

The aim of the present study was to explore the role of variations with modest effects (previously identified by a large-scale meta-analysis in European populations) in the genetic background of type 2 diabetes (T2D) and diabetes-related traits in a Japanese population. We enrolled 2632 Japanese subjects with T2D and 2050 non-diabetic subjects. We analyzed nine single-nucleotide polymorphisms (SNPs), including rs340874 (PROX1), rs4607517 (GCK), rs2191349 (DGKB-TMEM195), rs7034200 (GLIS3), rs10885122 (ADRA2A), rs174550 (FADS1), rs11605924 (CRY2), rs10830963 (MTNR1B) and rs35767 (IGF1). rs340874 (PROX1) and rs174550 (FADS1) were significantly associated with T2D (P=0.0078, OR: 1.12; and P=0.0071, OR: 1.12, respectively). Subjects with more risk alleles related to nine SNPs had an increased risk of T2D (P=0.0017), as well as a higher fasting plasma glucose level (P=0.018), higher HbA(1c) level (P=0.013) and lower HOMA-ß (P=0.033) compared with subjects who had fewer risk alleles. We identified a significant association of a SNP of FADS1 and a SNP near PROX1 with T2D in a Japanese population. The present findings suggest that inclusion of SNPs with a tendency to increase the disease risk captured more of the genetic background of T2D than that revealed by only assessing significant SNPs.

Metabolic surgery in treatment of obese Japanese patients with type 2 diabetes: a joint consensus statement from the Japanese Society for Treatment of Obesity, the Japan Diabetes Society, and the Japan Society for the Study of Obesity.

Bariatric surgery has been shown to have a variety of metabolically beneficial effects for patients with type 2 diabetes (T2D), and is now also called metabolic surgery. At the 2nd Diabetes Surgery Summit held in 2015 in London, the indication for bariatric and metabolic surgery was included in the "algorithm for patients with type T2D". With this background, the Japanese Society for Treatment of Obesity (JSTO), the Japan Diabetes Society (JDS) and the Japan Society for the Study of Obesity (JASSO) have formed a joint committee to develop a consensus statement regarding bariatric and metabolic surgery for the treatment of Japanese patients with T2D. Eventually, the consensus statement was announced at the joint meeting of the 38th Annual Meeting of JSTO and the 41st Annual Meeting of JASSO convened in Toyama on March 21, 2021. In preparing the consensus statement, we used Japanese data as much as possible as scientific evidence to consider the indication criteria, and set two types of recommendation grades, "recommendation" and "consideration", for items for which recommendations are possible. We hope that this statement will be helpful in providing evidence-based high-quality care through bariatric and metabolic surgery for the treatment of obese Japanese patients with T2D. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s13340-021-00551-0.

Correction to: Metabolic surgery in treatment of obese Japanese patients with type 2 diabetes: a joint consensus statement from the Japanese Society for Treatment of Obesity, the Japan Diabetes Society, and the Japan Society for the Study of Obesity.

[This corrects the article DOI: 10.1007/s13340-021-00551-0.].

Correction to: Metabolic surgery in treatment of obese Japanese patients with type 2 diabetes: a joint consensus statement from the Japanese Society for Treatment of Obesity, the Japan Diabetes Society, and the Japan Society for the Study of Obesity.

[This corrects the article DOI: 10.1007/s13340-021-00551-0.].

Structural basis of ethnic-specific variants of PAX4 associated with type 2 diabetes.

Recently, we conducted genome-wide association studies of type 2 diabetes (T2D) in a Japanese population, which identified 20 novel T2D loci that were not associated with T2D in Europeans. Moreover, nine novel missense risk variants, such as those of PAX4, were not rare in the Japanese population, but rare in Europeans. We report in silico structural analysis of ethnic-specific variants of PAX4, which suggests the pathogenic effect of these variants.

Genotype-Structure-Phenotype Correlations of Disease-Associated IGF1R Variants and Similarities to Those of INSR Variants.

We previously reported genotype-phenotype correlations in 12 missense variants causing severe insulin resistance, located in the second and third fibronectin type III (FnIII) domains of the insulin receptor (INSR), containing the α-ß cleavage and part of insulin-binding sites. This study aimed to identify genotype-phenotype correlations in FnIII domain variants of IGF1R, a structurally related homolog of INSR, which may be associated with growth retardation, using the recently reported crystal structures of IGF1R. A structural bioinformatics analysis of five previously reported disease-associated heterozygous missense variants and a likely benign variant in the FnIII domains of IGF1R predicted that the disease-associated variants would severely impair the hydrophobic core formation and stability of the FnIII domains or affect the α-ß cleavage site, while the likely benign variant would not affect the folding of the domains. A functional analysis of these variants in CHO cells showed impaired receptor processing and autophosphorylation in cells expressing the disease-associated variants but not in those expressing the wild-type form or the likely benign variant. These results demonstrated genotype-phenotype correlations in the FnIII domain variants of IGF1R, which are presumably consistent with those of INSR and would help in the early diagnosis of patients with disease-associated IGF1R variants.

Myocardial infarction in type 2 diabetes using sodium-glucose co-transporter-2 inhibitors, dipeptidyl peptidase-4 inhibitors or glucagon-like peptide-1 receptor agonists: proportional hazards analysis by deep neural network based machine learning.

Aims: Some hypoglycemic therapies are associated with lower risk of cardiovascular outcomes. We investigated the incidence of cardiovascular disease among patients with type 2 diabetes using antidiabetic drugs from three classes, which were sodium-glucose co-transporter-2 inhibitors (SGLT-2is), glucagon-like peptide-1 receptor agonists (GLP-1RAs) and dipeptidyl peptidase-4 inhibitors (DPP-4is).Materials and methods: We compared the risk of myocardial infarction (MI) among these drugs and developed a machine learning model for predicting MI in patients without prior heart disease. We analyzed US health plan data for patients without prior MI or insulin therapy who were aged ≥40 years at initial prescription and had not received oral antidiabetic drugs for ≥6 months previously. After developing a machine learning model to predict MI, proportional hazards analysis of MI incidence was conducted using the risk obtained with this model and the drug classes as explanatory variables.Results: We analyzed 199,116 patients (mean age: years), comprising 110,278 (58.6) prescribed DPP-4is, 43,538 (55.1) prescribed GLP-1RAs and 45,300 (55.3) prescribed SGLT-2is. Receiver operating characteristics analysis showed higher precision of machine learning over logistic regression analysis. Proportional hazards analysis by machine learning revealed a significantly lower risk of MI with SGLT-2is or GLP-1RAs than DPP-4is (hazard ratio: 0.81, 95% confidence interval: 0.72-0.91, p = .0004 vs. 0.63, 0.56-0.72, p < .0001). MI risk was also significantly lower with GLP-1RAs than SGLT-2is (0.77, 0.66-0.90, p = .001).Limitations: All patients analyzed were covered by US commercial health plans, so information on patients aged ≥65 years was limited and the socioeconomic background may have been biased. Also, the observation period differed among the three classes of drugs due to differing release dates.Conclusions: Machine learning analysis suggested the risk of MI was 37% lower for type 2 diabetes patients without prior MI using GLP-1RAs versus DPP-4is, while the risk was 19% lower for SGLT-2is versus DPP-4is.

Clinical usefulness of multigene screening with phenotype-driven bioinformatics analysis for the diagnosis of patients with monogenic diabetes or severe insulin resistance.

AIMS: Monogenic diabetes is clinically heterogeneous and differs from common forms of diabetes (type 1 and 2). We aimed to investigate the clinical usefulness of a comprehensive genetic testing system, comprised of targeted next-generation sequencing (NGS) with phenotype-driven bioinformatics analysis in patients with monogenic diabetes, which uses patient genotypic and phenotypic data to prioritize potentially causal variants. METHODS: We performed targeted NGS of 383 genes associated with monogenic diabetes or common forms of diabetes in 13 Japanese patients with suspected (n = 10) or previously diagnosed (n = 3) monogenic diabetes or severe insulin resistance. We performed in silico structural analysis and phenotype-driven bioinformatics analysis of candidate variants from NGS data. RESULTS: Among the patients suspected having monogenic diabetes or insulin resistance, we diagnosed 3 patients as subtypes of monogenic diabetes due to disease-associated variants of INSR, LMNA, and HNF1B. Additionally, in 3 other patients, we detected rare variants with potential phenotypic effects. Notably, we identified a novel missense variant in TBC1D4 and an MC4R variant, which together may cause a mixed phenotype of severe insulin resistance. CONCLUSIONS: This comprehensive approach could assist in the early diagnosis of patients with monogenic diabetes and facilitate the provision of tailored therapy.

Identification of 28 new susceptibility loci for type 2 diabetes in the Japanese population.

To understand the genetics of type 2 diabetes in people of Japanese ancestry, we conducted A meta-analysis of four genome-wide association studies (GWAS; 36,614 cases and 155,150 controls of Japanese ancestry). We identified 88 type 2 diabetes-associated loci (P < 5.0 × 10-8) with 115 independent signals (P < 5.0 × 10-6), of which 28 loci with 30 signals were novel. Twenty-eight missense variants were in linkage disequilibrium (r2 > 0.6) with the lead variants. Among the 28 missense variants, three previously unreported variants had distinct minor allele frequency (MAF) spectra between people of Japanese and European ancestry (MAFJPN > 0.05 versus MAFEUR < 0.01), including missense variants in genes related to pancreatic acinar cells (GP2) and insulin secretion (GLP1R). Transethnic comparisons of the molecular pathways identified from the GWAS results highlight both ethnically shared and heterogeneous effects of a series of pathways on type 2 diabetes (for example, monogenic diabetes and beta cells).

The Gα12/13-coupled receptor LPA4 limits proper adipose tissue expansion and remodeling in diet-induced obesity.

White adipose tissue (WAT) can dynamically expand and remodel through adipocyte hypertrophy and hyperplasia. The relative contribution of these 2 mechanisms to WAT expansion is a critical determinant of WAT function and dysfunction in obesity. However, little is known about the signaling systems that determine the mechanisms of WAT expansion. Here, we show that the GPCR LPA4 selectively activates Gα12/13 proteins in adipocytes and limits continuous remodeling and healthy expansion of WAT. LPA4-KO mice showed enhanced expression of mitochondrial and adipogenesis genes and reduced levels of inhibitory phosphorylation of PPARγ in WAT, along with increased production of adiponectin. Furthermore, LPA4-KO mice showed metabolically healthy obese phenotypes in a diet-induced obesity model, with continuous WAT expansion, as well as protection from WAT inflammation, hepatosteatosis, and insulin resistance. These findings unravel a potentially new signaling system that underlies WAT plasticity and expandability, providing a promising therapeutic approach for obesity-related metabolic disorders.

Glycemic control, mortality, and hypoglycemia in critically ill patients: a systematic review and network meta-analysis of randomized controlled trials.

PURPOSE: It is unclear whether tight glycemic control is warranted in all critically ill adults. We employed network meta-analysis to examine the risk of mortality and hypoglycemia associated with different glycemic control targets in critically ill adults. METHODS: Electronic databases were searched up to 2016 for randomized controlled trials comparing various insulin regimens in critically ill adults with hyperglycemia. Two reviewers independently extracted information and evaluated quality with the Cochrane risk-of-bias tool. Four glycemic control groups were compared: tight (blood glucose: 4.4 < 6.1 mmol/l), moderate (6.1 < 7.8 mmol/l), mild (7.8 < 10.0 mmol/l), and very mild (10.0 to < 12.2 mmol/l). Network meta-analysis was performed by a frequentist approach with multivariate random effects meta-analysis. RESULTS: Thirty-six randomized trials (17,996 patients) were identified. Compared with very mild control, tight control did not reduce the risk of short-term mortality [relative risk (RR) 0.94 (95 % CI 0.83-1.07, p = 0.36)], and neither did mild control [RR 0.88 (0.73-1.06), p = 0.18] or moderate control [RR 1.1 (0.66-1.84), p = 0.72]. However, severe hypoglycemia (<2.2 mmol/l) was more frequent with tight control than very mild control [RR 5.49 (3.22-9.38), p < 0.001] or mild control [RR 4.47 (2.5-8.03), p < 0.001]. Stratified analyses (cause of death, ICU type, time period, or diabetes) did not find significant between-group differences. Ranking analysis revealed the following hierarchy for avoiding death (highest to lowest rank): mild control, tight control, and very mild control. CONCLUSIONS: Network meta-analysis showed no mortality benefit of tight glycemic control in critically ill patients, but fivefold more hypoglycemia versus mild or very mild control.