Guozhi Jiang

Use of Net Reclassification Improvement (NRI) Method Confirms The Utility of Combined Genetic Risk Score to Predict Type 2 Diabetes

Background Recent genome-wide association studies (GWAS) identified more than 70 novel loci for type 2 diabetes (T2D), some of which have been widely replicated in Asian populations. In this study, we investigated their individual and combined effects on T2D in a Chinese population. Methodology We selected 14 single nucleotide polymorphisms (SNPs) in T2D genes relating to beta-cell function validated in Asian populations and genotyped them in 5882 Chinese T2D patients and 2569 healthy controls. A combined genetic score (CGS) was calculated by summing up the number of risk alleles or weighted by the effect size for each SNP under an additive genetic model. We tested for associations by either logistic or linear regression analysis for T2D and quantitative traits, respectively. The contribution of the CGS for predicting T2D risk was evaluated by receiver operating characteristic (ROC) analysis and net reclassification improvement (NRI). Results We observed consistent and significant associations of IGF2BP2, WFS1, CDKAL1, SLC30A8, CDKN2A/B, HHEX, TCF7L2 and KCNQ1 (8.5×10−18<P<8.5×10−3), as well as nominal associations of NOTCH2, JAZF1, KCNJ11 and HNF1B (0.05<P<0.1) with T2D risk, which yielded odds ratios ranging from 1.07 to 2.09. The 8 significant SNPs exhibited joint effect on increasing T2D risk, fasting plasma glucose and use of insulin therapy as well as reducing HOMA-β, BMI, waist circumference and younger age of diagnosis of T2D. The addition of CGS marginally increased AUC (2%) but significantly improved the predictive ability on T2D risk by 11.2% and 11.3% for unweighted and weighted CGS, respectively using the NRI approach (P<0.001). Conclusion In a Chinese population, the use of a CGS of 8 SNPs modestly but significantly improved its discriminative ability to predict T2D above and beyond that attributed to clinical risk factors (sex, age and BMI).

Genetic and clinical variables identify predictors for chronic kidney disease in type 2 diabetes

Type 2 diabetes and chronic kidney disease (CKD) may share common risk factors. Here we used a 3-stage procedure to discover novel predictors of CKD by repeatedly applying a stepwise selection based on the Akaike information criterion to subsamples of a prospective complete-case cohort of 2755 patients. This cohort encompassed 25 clinical variables and 36 genetic variants associated with type 2 diabetes, obesity, or fasting plasma glucose. We compared the performance of the clinical, genetic, and clinico-genomic models and used net reclassification improvement to evaluate the impact of top selected genetic variants to the clinico-genomic model. Associations of selected genetic variants with CKD were validated in 2 independent cohorts followed by meta-analyses. Among the top 6 single-nucleotide polymorphisms selected from clinico-genomic data, three (rs478333 of G6PC2, rs7754840 and rs7756992 of CDKAL1) contributed toward the improvement of prediction performance. The variant rs478333 was associated with rapid decline (over 4% per year) in estimated glomerular filtration rate. In a meta-analysis of 2 replication cohorts, the variants rs478333 and rs7754840 showed significant associations with CKD after adjustment for conventional risk factors. Thus, this novel 3-stage approach to a clinico-genomic data set identified 3 novel genetic predictors of CKD in type 2 diabetes. This method can be applied to similar data sets containing clinical and genetic variables to select predictors for clinical outcomes.

Modifying Effect of Body Mass Index on Survival in Elderly Type 2 Diabetic Patients: Hong Kong Diabetes Registry

OBJECTIVE There are nonlinear risk associations of body mass index (BMI) with mortality in type 2 diabetes (T2D) and elderly populations although similar information in elderly individuals with T2D are lacking. RESEARCH DESIGN AND METHODS We analyzed prospective data for 3186 Chinese patients with T2D with age 65 years or older. Baseline demographic data, risk factors, complications, and all-cause mortality were captured from the Hong Kong Diabetes Registry and the Hong Kong Hospital Authority Clinical Management System. RESULTS Over a median follow-up period of 6.0 years (medium-term), 816 (25.6%) deaths occurred and at 9.4 years (long-term), 1557 (48.9%) patients had died. Men were more likely to die than women with increased mortality rate with increasing age (morality rates of men with normal BMI at 9-year follow-up in the 65 to 69, 70 to 74, and 75 years or older age groups were 41.8, 70.3, and 101.4 per 1000 person-years, whereas that for women were 35.5, 50.4, and 78.8 respectively). Within each age group, high BMI was associated with increased survival, especially in the 75 years and older age group and with prolonged follow-up period. Using Cox regression analysis, after adjustment for confounders, high BMI (≥ 25.0 kg/m(2)) was associated with reduced risk of death in all subgroups, reaching significance in men in the older age groups at 9-year follow-up (for men 70 to 74 years old, hazard ratio [HR] of mortality was 0.67, 95% confidence interval [CI] 0.48-0.95, for those ≥ 75, HR was 0.62, 95% CI 0.44-0.89) compared with 18.5 to 22.9 kg/m(2) as referent. CONCLUSIONS In Chinese elderly patients with T2D, high BMI protected against mortality, calling for more attention to people with low BMI who might have unmet clinical needs.

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).

Progression to treatment failure among Chinese patients with type 2 diabetes initiated on metformin versus sulphonylurea monotherapy—The Hong Kong Diabetes Registry

AIMS To assess the development of treatment failure in Chinese patients with type 2 diabetes mellitus (T2DM) initiated on metformin or sulphonylurea (SU) monotherapy, with consideration of various potential sources of biases. METHODS A 1:1-matched new metformin and SU user cohort on immortal time and mean propensity score after multiple imputation was selected from a cohort of 5889 Chinese patients with T2DM. Treatment failure was defined as progression to (i) combination oral anti-hyperglycemia drug therapy, (ii) insulin use, or (iii) a treatment haemoglobin A1c (HbA1c) >7.5% (58 mmol/mol). Stratified Cox regression analysis on the matched pairs was employed to examine the associations between initial monotherapy and onset of treatment failure. RESULTS Of 554 new metformin and 840 new SU users, 380 were matched. During a median follow-up duration of 3 years, 173 (45.6%) metformin users and 220 (57.9%) SU users experienced treatment failure (annual failure rates of 15% and 19%, respectively). The median time from monotherapy starting to treatment failure was 3.0 [inter-quartile range (IQR): 1.8-5.4] years for metformin users, versus 1.8 (IQR: 0.9-4.1) years for SU users (p<0.001). Stratified Cox regression analysis showed significantly lower risk of treatment failure for metformin users (HR [95% CI], 0.62[0.47-0.81]; p<0.001). Consistent results were found in analyses based on traditional adjustment schemes with or without imputation. CONCLUSIONS By systematically incorporating new-user design, multiple imputation and matching methods, we found that Chinese patients with T2DM initiated on metformin monotherapy were associated with a significant delay in the onset of treatment failure compared to SU monotherapy.

Variable selection and prediction of clinical outcome with multiply-imputed data via Bayesian model averaging

Multiple imputation (MI) is increasingly used to deal with missing data in medical studies, whilst variable selection and prediction on multiply-imputed data is an area under intense research in statistics. A commonly used strategy is to select a single top model based on the Rubins rules (RR). However, such approaches do not take the model uncertainty into consideration, which might lead to over-confident inferences. In this paper, we extended the Bayesian model averaging method to perform variable selection and prediction under multiple imputation (MI-BMA), which takes into account the uncertainties originated from both the missing data and the model selection. We applied the MI-BMA method to simulated datasets as well as a real data set from a prospective cohort, and demonstrated the advantage of our method as compared with the classical RR stepwise method.