Type 2 Diabetes Genetic Variants and Risk of Diabetic Retinopathy

Impact of the trans-ancestry polygenic risk score on type 2 diabetes risk, onset age and progression among population in Taiwan

Type 2 diabetes (T2D) prevalence in adults at a younger age has increased but the disease status may go unnoticed. This study aimed to determine whether the onset age and subsequent diabetic complications can be attributed to the polygenic architecture of T2D in the Taiwan Han population. A total of 9,627 cases with T2D and 85,606 controls from the Taiwan Biobank were enrolled. Three diabetic polygenic risk scores (PRSs), PRS_EAS and PRS_EUR, and a trans-ancestry PRS (PRS_META), calculated using summary statistic from East Asian and European populations. The onset age was identified by linking to the National Taiwan Insurance Research Database, and the incidence of different diabetic complications during follow-up was recorded. PRS_META (7.4%) explained a higher variation for T2D status. And the higher percentile of PRS is also correlated with higher percentage of T2D family history and prediabetes status. More, the PRS was negatively associated with onset age (β = -0.91 yr), and this was more evident among males (β = -1.11 vs. -0.76 for males and females, respectively). The hazard ratio of diabetic retinopathy (DR) and diabetic foot were significantly associated with PRS_EAS and PRS_META, respectively. However, the PRS was not associated with other diabetic complications, including diabetic nephropathy, cardiovascular disease, and hypertension. Our findings indicated that diabetic PRS which combined susceptibility variants from cross-population could be used as a tool for early screening of T2D, especially for high-risk populations, such as individuals with high genetic risk, and may be associated with the risk of complications in subjects with T2D. NEW & NOTEWORTHY Our findings indicated that diabetic polygenic risk score (PRS) which combined susceptibility variants from Asian and European population affect the onset age of type 2 diabetes (T2D) and could be used as a tool for early screening of T2D, especially for individuals with high genetic risk, and may be associated with the risk of diabetic complications among people in Taiwan.

Prediction of diabetic kidney disease risk using machine learning models: A population-based cohort study of Asian adults

Background

Machine learning (ML) techniques improve disease prediction by identifying the most relevant features in multidimensional data. We compared the accuracy of ML algorithms for predicting incident diabetic kidney disease (DKD).

Methods

We utilized longitudinal data from 1365 Chinese, Malay, and Indian participants aged 40-80 y with diabetes but free of DKD who participated in the baseline and 6-year follow-up visit of the Singapore Epidemiology of Eye Diseases Study (2004-2017). Incident DKD (11.9%) was defined as an estimated glomerular filtration rate (eGFR) <60 mL/min/1.73 m2 with at least 25% decrease in eGFR at follow-up from baseline. A total of 339 features, including participant characteristics, retinal imaging, and genetic and blood metabolites, were used as predictors. Performances of several ML models were compared to each other and to logistic regression (LR) model based on established features of DKD (age, sex, ethnicity, duration of diabetes, systolic blood pressure, HbA1c, and body mass index) using area under the receiver operating characteristic curve (AUC).

Results

ML model Elastic Net (EN) had the best AUC (95% CI) of 0.851 (0.847-0.856), which was 7.0% relatively higher than by LR 0.795 (0.790-0.801). Sensitivity and specificity of EN were 88.2 and 65.9% vs. 73.0 and 72.8% by LR. The top 15 predictors included age, ethnicity, antidiabetic medication, hypertension, diabetic retinopathy, systolic blood pressure, HbA1c, eGFR, and metabolites related to lipids, lipoproteins, fatty acids, and ketone bodies.

Conclusions

Our results showed that ML, together with feature selection, improves prediction accuracy of DKD risk in an asymptomatic stable population and identifies novel risk factors, including metabolites.

Funding

This study was supported by the National Medical Research Council, NMRC/OFLCG/001/2017 and NMRC/HCSAINV/MOH-001019-00. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Prognostic models of diabetic microvascular complications: a systematic review and meta-analysis

BACKGROUND

Many prognostic models of diabetic microvascular complications have been developed, but their performances still varies. Therefore, we conducted a systematic review and meta-analysis to summarise the performances of the existing models.

METHODS

Prognostic models of diabetic microvascular complications were retrieved from PubMed and Scopus up to 31 December 2020. Studies were selected, if they developed or internally/externally validated models of any microvascular complication in type 2 diabetes (T2D).

RESULTS

In total, 71 studies were eligible, of which 32, 30 and 18 studies initially developed prognostic model for diabetic retinopathy (DR), chronic kidney disease (CKD) and end stage renal disease (ESRD) with the number of derived equations of 84, 96 and 51, respectively. Most models were derived-phases, some were internal and external validations. Common predictors were age, sex, HbA1c, diabetic duration, SBP and BMI. Traditional statistical models (i.e. Cox and logit regression) were mostly applied, otherwise machine learning. In cohorts, the discriminative performance in derived-logit was pooled with C statistics of 0.82 (0.73‑0.92) for DR and 0.78 (0.74‑0.83) for CKD. Pooled Cox regression yielded 0.75 (0.74‑0.77), 0.78 (0.74‑0.82) and 0.87 (0.84‑0.89) for DR, CKD and ESRD, respectively. External validation performances were sufficiently pooled with 0.81 (0.78‑0.83), 0.75 (0.67‑0.84) and 0.87 (0.85‑0.88) for DR, CKD and ESRD, respectively.

CONCLUSIONS

Several prognostic models were developed, but less were externally validated. A few studies derived the models by using appropriate methods and were satisfactory reported. More external validations and impact analyses are required before applying these models in clinical practice.

SYSTEMATIC REVIEW REGISTRATION

PROSPERO CRD42018105287.

Multiple Single Nucleotide Polymorphism Testing Improves the Prediction of Diabetic Retinopathy Risk with Type 2 Diabetes Mellitus

Diabetic retinopathy (DR) is one of the most frequent causes of irreversible blindness, thus prevention and early detection of DR is crucial. The purpose of this study is to identify genetic determinants of DR in individuals with type 2 diabetic mellitus (T2DM). A total of 551 T2DM patients (254 with DR, 297 without DR) were included in this cross-sectional research. Thirteen T2DM-related single nucleotide polymorphisms (SNPs) were utilized for constructing genetic risk prediction model. With logistic regression analysis, genetic variations of the FTO (rs8050136) and PSMD6 (rs831571) polymorphisms were independently associated with a higher risk of DR. The area under the curve (AUC) calculated on known nongenetic risk variables was 0.704. Based on the five SNPs with the highest odds ratio (OR), the combined nongenetic and genetic prediction model improved the AUC to 0.722. The discriminative accuracy of our 5-SNP combined risk prediction model increased in patients who had more severe microalbuminuria (AUC = 0.731) or poor glycemic control (AUC = 0.746). In conclusion, we found a novel association for increased risk of DR at two T2DM-associated genetic loci, FTO (rs8050136) and PSMD6 (rs831571). Our predictive risk model presents new insights in DR development, which may assist in enabling timely intervention in reducing blindness in diabetic patients.

Relationship between Clinical Features of Diabetic Retinopathy and Systemic Factors in Patients with Newly Diagnosed Type II Diabetes Mellitus

BACKGROUND

We investigated the relationship between clinical features of diabetic retinopathy (DR) and systemic factors in patients with newly diagnosed type II diabetes mellitus (T2DM).

METHODS

Retrospective review of newly diagnosed T2DM-patients who underwent complete ophthalmic examinations at the time of T2DM diagnosis were conducted. We reviewed DM related systemic factor data and investigated systemic factors related to the presence of DR at T2DM diagnosis. In DR patients, the relationship between DR severity and systemic factors was analyzed.

RESULTS

Of 380 patients, forty (10.53%) patients had DR at the initial ophthalmologic examination. Glycated hemoglobin (HbA_1C), fasting plasma glucose (FPG), urine albumin to creatinine ratio (UACR), and urine microalbumin level were significantly higher in DR patients than in patients without DR. In the multivariate logistic regression analysis, high HbA_1C was a significant risk factor for the presence of DR at new T2DM diagnosis (odds ratio, 2.372; P < 0.001). HbA_1C, FPG, UACR, and urine microalbumin level showed significantly positive correlations with DR severity .

CONCLUSION

In patients with newly diagnosed T2DM, 10.53% have DR at initial ophthalmologic examination and high HbA_1C, FPG, UACR and urine microalbumin levels. These factors are significantly positively correlated with DR severity. Therefore, more careful fundus examination is needed for newly diagnosed T2DM patients with high HbA_1C, FPG, UACR, and urine microalbumin levels.

Serum uric acid is independently associated with diabetic nephropathy but not diabetic retinopathy in patients with type 2 diabetes mellitus

BACKGROUND

This study aims to investigate the relationship between serum uric acid (SUA) and the severity of diabetic nephropathy (DN) and diabetic retinopathy (DR) in patients with type 2 diabetes mellitus (T2DM).

METHODS

A total of 2961 patients were enrolled in the present cross-sectional study. The severity of DN was determined by 24-hour urinary albumin excretion (UAE), which was classified as normal (UAE <30 mg/24 h), microalbuminuria (UAE: 30-299 mg/24 h), and macroalbuminuria (≥300 mg/24 h). The severity of DR was determined by non-mydriatic retinal photography and was classified as non-diabetic retinopathy (NDR), non-proliferative diabetic retinopathy (NPDR), and proliferative DR (PDR).

RESULTS

Patients with high SUA levels (≥420 μmol/L for males and ≥360 μmol/L for females) had a significantly higher prevalence of DN (UAE ≥30 mg/24 h, 39.3% vs 26.3%; p < 0.001), higher UAE levels (140 ± 297 vs 63 ± 175 mg/24 h; p < 0.001), and lower estimated glomerular filtration rate (eGFR; 79.3 ± 26.8 vs 96.8 ± 19.6 mL/min/1.73 m; p < 0.001), when compared with patients with normal SUA levels. However, the prevalence of DR, NPDR, or PDR did not differ. Furthermore, the concentration of SUA was higher in patients with higher severity of DN (all, p < 0.001) and patients with PDR (compared with NDR or NPDR, p < 0.05). SUA levels were positively associated with male gender, body mass index, the use of diuretics, triglyceride, low-density lipoprotein, and UAE levels, whereas they were negatively correlated with high-density lipoprotein, fasting blood glucose, glycosylated hemoglobin, and eGFR. After adjustment, SUA remained significantly associated with UAE (r = 0.069, p < 0.001).

CONCLUSION

For patients with T2DM, higher SUA levels are associated with higher UAE, lower eGFR, and higher prevalence of DN, but not DR.

Technical and imaging factors influencing performance of deep learning systems for diabetic retinopathy

Deep learning (DL) has been shown to be effective in developing diabetic retinopathy (DR) algorithms, possibly tackling financial and manpower challenges hindering implementation of DR screening. However, our systematic review of the literature reveals few studies studied the impact of different factors on these DL algorithms, that are important for clinical deployment in real-world settings. Using 455,491 retinal images, we evaluated two technical and three image-related factors in detection of referable DR. For technical factors, the performances of four DL models (VGGNet, ResNet, DenseNet, Ensemble) and two computational frameworks (Caffe, TensorFlow) were evaluated while for image-related factors, we evaluated image compression levels (reducing image size, 350, 300, 250, 200, 150 KB), number of fields (7-field, 2-field, 1-field) and media clarity (pseudophakic vs phakic). In detection of referable DR, four DL models showed comparable diagnostic performance (AUC 0.936-0.944). To develop the VGGNet model, two computational frameworks had similar AUC (0.936). The DL performance dropped when image size decreased below 250 KB (AUC 0.936, 0.900, p < 0.001). The DL performance performed better when there were increased number of fields (dataset 1: 2-field vs 1-field-AUC 0.936 vs 0.908, p < 0.001; dataset 2: 7-field vs 2-field vs 1-field, AUC 0.949 vs 0.911 vs 0.895). DL performed better in the pseudophakic than phakic eyes (AUC 0.918 vs 0.833, p < 0.001). Various image-related factors play more significant roles than technical factors in determining the diagnostic performance, suggesting the importance of having robust training and testing datasets for DL training and deployment in the real-world settings.

Analysis of association between common variants of uncoupling proteins genes and diabetic retinopathy in a Chinese population

BACKGROUND

The aim of this study was to explore the association between diabetic retinopathy (DR) and the variants of uncoupling proteins (UCPs) genes in a Chinese population of type 2 diabetes, in total and in patients of different glycemic status separately.

METHODS

This case-control study included a total of 3107 participants from two datasets, among which 662 were DR patients (21.31%). Eighteen tag single nucleotide polymorphisms (SNPs) of UCP1, UCP2, and UCP3 were selected as genetic markers. TaqMan probes, Sequenom MassARRAY MALDI-TOF mass spectrometry platform and Affymetrix Genome-Wide Human SNP Array were used for genotyping. Online SHEsis software was used for association analysis. Bonferroni correction was used for multiple comparisons correction.

RESULTS

Three SNPs of UCP1: rs7688743 (A allele, OR = 1.192, p = 0.013), rs3811787 (T allele, OR = 0.863, p = 0.023), and rs10011540 (G allele, OR = 1.368, p = 0.004) showed association with DR after the adjustment of glucose, but only rs10011540 was marginally significantly associated with DR when Bonferroni correction was strictly applied (p_adj = 0.048). In patients with uncontrolled glucose, rs7688743 (A allele, p = 0.012, OR = 1.309), rs10011540 (G allele, p = 0.033, OR = 1.432), and rs3811787 (T allele, p = 0.022, OR = 0.811) were associated with DR, while in participants with well controlled glucose, the rs2734827 of UCP3 was associated with DR (A allele, p = 0.017, OR = 0.532). Rs3811787 of UCP1 showed a protective effect to sight threatening DR (T allele, p = 0.007, OR = 0.490), and the association existed after the adjustment for environmental factors and the correction. In patients with uncontrolled glucose, the rs3811787 of UCP1 (T allele, p = 0.017, OR = 0.467) and the rs591758 of UCP3 (C allele, p = 0.026, OR = 0.103) were associated with STDR. While in those with well controlled glucose, only the rs7688743 of UCP1 showed a protective effect (A allele, p = 0.024, OR = 0.049). None of the associations remain significant when Bonferroni correction was strictly applied (all p < 0.05).

CONCLUSIONS

The rs10011540 and rs3811787 of the UCP1 gene was marginally significantly associated with DR in Chinese type 2 diabetes patients. There might be different mechanisms of DR development in patients with different glycemic status.

MicroRNA-29b-3p Promotes Human Retinal Microvascular Endothelial Cell Apoptosis via Blocking SIRT1 in Diabetic Retinopathy

Background

Diabetic retinopathy (DR) is a main complication of diabetes mellitus (DM). Recent studies have implicated microRNAs in human retinal microvascular endothelial cell (HRMEC) dysfunction. In this study, we aim to investigate the apoptotic promotion of miR-29b-3p by blocking SIRT1 in HRMEC for DR situation.

Method

Blood samples were obtained from DR patients and controls. Dual-luciferase reporter assay using HEK-293T cells was performed to show the direct interaction of miR-29b-3p and the 3′UTR of SIRT1. HRMECs were exposed to 5.5 mmol/L of glucose (normal control), 5.5 mmol/L of glucose and 24.5 mmol/L of mannitol (osmotic pressure control), 30 mmol/L of glucose [hyperglycemia (HG)], 150 μmol/L of CoCl₂ (hypoxia), and 30 mmol/L of glucose plus 150 μmol/L of CoCl₂ (HG-CoCl₂). To identify the regulating relationship between miR-29b-3p and SIRT1, HRMECs were transfected with miR-29b-3p mimics/inhibitors or their negative controls. SRT1720 was used as a SIRT1 agonist. Cell viability was assessed with the cell counting kit-8 (CCK-8) assay, and apoptotic cells were stained by one-step terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay kit. Gene and protein expression were assayed by quantitative real-time reverse transcriptase-PCR (RT-qPCR) and western blotting separately.

Result

MiR-29b-3p was upregulated to 3.2-fold, and SIRT1 protein was downregulated to 65% in DR patients. Dual-luciferase reporter assay showed the direct interaction of miR-29b-3p and SIRT1. HRMECs were identified as >95% positive for CD31 and von Willebrand factor (vWF). MiR-29b-3p and Bax/Bcl-2 ratio was upregulated, whereas SIRT1 was downregulated in HRMECs in the HG-CoCl₂ condition. Decreased cell viability and upregulated apoptosis were also found in HRMECs of the HG-CoCl₂ condition. Upregulated miR-29b-3p decreased the expression of SIRT1 and increased the ratio of Bax/Bcl-2, whereas downregulated miR-29b-3p increased the expression of SIRT1 protein and downregulated the ratio of Bax/Bcl-2. SRT1720 rescued miR-29b-3p-induced HRMEC apoptosis via upregulating the expression of SIRT1 protein.

Conclusion

The dysregulation of miR-29b-3p/SIRT1 is a potential mechanism of HRMEC apoptosis in DR. MiR-29b-3p/SIRT1 may be a potential therapeutic target for DR.

Ultra-performance liquid chromatography-mass spectrometry-based metabolomics reveals Huangqiliuyi decoction attenuates abnormal metabolism as a novel therapeutic opportunity for type 2 diabetes

Background: As a typical chronic metabolic disease, type 2 diabetes mellitus causes a heavy health-care burden to society. In this study, we applied the metabolomics strategy to explore the potential molecular mechanism of the Huangqiliuyi decoction (HQLYD) for type-2 diabetes (T2D). Ultra-performance liquid chromatography-mass spectrometry (UPLC-MS) combined with pattern recognition methods was utilized to select specific metabolites closely associated with HQLYD. Biomarker pathway analysis and biological network were utilized to uncover the therapeutic effect and action mechanism related to HQLYD. A total of twenty-five biomarkers were identified in the animal model, in which sixteen biomarkers are associated with HQLYD treatment for T2D. They attenuated the abnormalities of metabolic pathways such as phenylalanine, tyrosine and tryptophan biosynthesis, phenylalanine metabolism, and the citrate cycle. HQLYD also significantly elevated the serum FINS and SOD, GSP-x level in the liver and kidney, and reduced the serum TC, TG, HDL, LDL, urea, Scr, AST, ALT, FBG, IRS, MDA, and CAT level. We found that the therapeutic mechanism of HQLYD against T2D affected amino acid metabolism, glucose metabolism and lipid metabolism. Metabolomics revealed that the Huangqiliuyi decoction attenuates abnormal metabolism as a novel therapeutic opportunity for type 2 diabetes.

Multiethnic Genome-Wide Association Study of Diabetic Retinopathy Using Liability Threshold Modeling of Duration of Diabetes and Glycemic Control

To identify genetic variants associated with diabetic retinopathy (DR), we performed a large multiethnic genome-wide association study. Discovery included eight European cohorts (n = 3,246) and seven African American cohorts (n = 2,611). We meta-analyzed across cohorts using inverse-variance weighting, with and without liability threshold modeling of glycemic control and duration of diabetes. Variants with a P value <1 × 10-5 were investigated in replication cohorts that included 18,545 European, 16,453 Asian, and 2,710 Hispanic subjects. After correction for multiple testing, the C allele of rs142293996 in an intron of nuclear VCP-like (NVL) was associated with DR in European discovery cohorts (P = 2.1 × 10-9), but did not reach genome-wide significance after meta-analysis with replication cohorts. We applied the Disease Association Protein-Protein Link Evaluator (DAPPLE) to our discovery results to test for evidence of risk being spread across underlying molecular pathways. One protein-protein interaction network built from genes in regions associated with proliferative DR was found to have significant connectivity (P = 0.0009) and corroborated with gene set enrichment analyses. These findings suggest that genetic variation in NVL, as well as variation within a protein-protein interaction network that includes genes implicated in inflammation, may influence risk for DR.

Development and Validation of a Deep Learning System for Diabetic Retinopathy and Related Eye Diseases Using Retinal Images From Multiethnic Populations With Diabetes

IMPORTANCE

A deep learning system (DLS) is a machine learning technology with potential for screening diabetic retinopathy and related eye diseases.

OBJECTIVE

To evaluate the performance of a DLS in detecting referable diabetic retinopathy, vision-threatening diabetic retinopathy, possible glaucoma, and age-related macular degeneration (AMD) in community and clinic-based multiethnic populations with diabetes.

DESIGN, SETTING, AND PARTICIPANTS

Diagnostic performance of a DLS for diabetic retinopathy and related eye diseases was evaluated using 494 661 retinal images. A DLS was trained for detecting diabetic retinopathy (using 76 370 images), possible glaucoma (125 189 images), and AMD (72 610 images), and performance of DLS was evaluated for detecting diabetic retinopathy (using 112 648 images), possible glaucoma (71 896 images), and AMD (35 948 images). Training of the DLS was completed in May 2016, and validation of the DLS was completed in May 2017 for detection of referable diabetic retinopathy (moderate nonproliferative diabetic retinopathy or worse) and vision-threatening diabetic retinopathy (severe nonproliferative diabetic retinopathy or worse) using a primary validation data set in the Singapore National Diabetic Retinopathy Screening Program and 10 multiethnic cohorts with diabetes.

EXPOSURES

Use of a deep learning system.

MAIN OUTCOMES AND MEASURES

Area under the receiver operating characteristic curve (AUC) and sensitivity and specificity of the DLS with professional graders (retinal specialists, general ophthalmologists, trained graders, or optometrists) as the reference standard.

RESULTS

In the primary validation dataset (n = 14 880 patients; 71 896 images; mean [SD] age, 60.2 [2.2] years; 54.6% men), the prevalence of referable diabetic retinopathy was 3.0%; vision-threatening diabetic retinopathy, 0.6%; possible glaucoma, 0.1%; and AMD, 2.5%. The AUC of the DLS for referable diabetic retinopathy was 0.936 (95% CI, 0.925-0.943), sensitivity was 90.5% (95% CI, 87.3%-93.0%), and specificity was 91.6% (95% CI, 91.0%-92.2%). For vision-threatening diabetic retinopathy, AUC was 0.958 (95% CI, 0.956-0.961), sensitivity was 100% (95% CI, 94.1%-100.0%), and specificity was 91.1% (95% CI, 90.7%-91.4%). For possible glaucoma, AUC was 0.942 (95% CI, 0.929-0.954), sensitivity was 96.4% (95% CI, 81.7%-99.9%), and specificity was 87.2% (95% CI, 86.8%-87.5%). For AMD, AUC was 0.931 (95% CI, 0.928-0.935), sensitivity was 93.2% (95% CI, 91.1%-99.8%), and specificity was 88.7% (95% CI, 88.3%-89.0%). For referable diabetic retinopathy in the 10 additional datasets, AUC range was 0.889 to 0.983 (n = 40 752 images).

CONCLUSIONS AND RELEVANCE

In this evaluation of retinal images from multiethnic cohorts of patients with diabetes, the DLS had high sensitivity and specificity for identifying diabetic retinopathy and related eye diseases. Further research is necessary to evaluate the applicability of the DLS in health care settings and the utility of the DLS to improve vision outcomes.

CDKAL1 rs7756992 is associated with diabetic retinopathy in a Chinese population with type 2 diabetes

Diabetic retinopathy (DR) is a major microvascular complication of diabetes. Susceptibility genes for type 2 diabetes may also impact the susceptibility of DR. This case-control study investigated the effects of 88 type 2 diabetes susceptibility loci on DR in a Chinese population with type 2 diabetes performed in two stages. In stage 1, 88 SNPs were genotyped in 1,251 patients with type 2 diabetes, and we found that ADAMTS9-AS2 rs4607103, WFS1 rs10010131, CDKAL1 rs7756992, VPS26A rs1802295 and IDE-KIF11-HHEX rs1111875 were significantly associated with DR. The association between CDKAL1 rs7756992 and DR remained significant after Bonferroni correction for multiple comparisons (corrected P = 0.0492). Then, the effect of rs7756992 on DR were analysed in two independent cohorts for replication in stage 2. Cohort (1) consisted of 380 patients with DR and 613 patients with diabetes for ≥5 years but without DR. Cohort (2) consisted of 545 patients with DR and 929 patients with diabetes for ≥5 years but without DR. A meta-analysis combining the results of stage 1 and 2 revealed a significant association between rs7756992 and DR, with the minor allele A conferring a lower risk of DR (OR 0.824, 95% CI 0.743–0.914, P = 2.46 × 10⁻⁴).