Novel subgroups of prediabetes and the associations with outcomes in health professionals: a data–driven cluster analysis

Novel subgroups of prediabetes and the associations with outcomes in health professionals: a data–driven cluster analysis | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Novel subgroups of prediabetes and the associations with outcomes in health professionals: a data–driven cluster analysis Apinya Surawit, Phongthana Pasookhush, Sureeporn Pumeiam, Pichanun Mongkolsucharitkul, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6351460/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background: This study examined prediabetic clusters and their associations with type 2 diabetes (T2D) and cardiovascular disease (CVD) using variables from metabolic syndrome, glycemic measures, and blood lipids. Methods: A total of 1,016 prediabetic individuals were classified into four clusters using k-means clustering. Weibull proportional hazards models estimated T2D and CVD risk, and T2D polygenic risk scores (PRS) were analyzed to refine risk within each cluster. Results: Four clusters were identified: the metabolic syndrome prediabetes (MESPD) cluster, characterized by elevated BMI and adverse lipid profiles, had the highest T2D risk (HR 5.86). The mild age-related prediabetes (MARPD) cluster, associated with older age, showed an increased T2D risk. In contrast, the low–risk prediabetes (LORPD) cluster exhibited the lowest risk, suggesting that a reduced metabolic burden may confer greater disease stability. PRS were used to refine risk stratification, with the MESPD cluster showing a significant genetic predisposition to T2D. PRS also enhanced predictive accuracy for the LORPD cluster, providing additional insights into genetic factors. Conclusions: The findings highlight the importance of precision medicine by identifying prediabetic subgroups with varying risks for T2D. Incorporating genetic data, the study improves models and offers insights for future research and interventions to prevent prediabetes progression. Health sciences/Endocrinology/Endocrine system and metabolic diseases Health sciences/Diseases/Endocrine system and metabolic diseases Health sciences/Diseases/Metabolic disorders Health sciences/Diseases Health sciences/Endocrinology Prediabetes Type 2 diabetes Cardiovascular disease Cluster analysis Figures Figure 1 Figure 2 Figure 3 Figure 4 INTRODUCTION Prediabetes, defined as elevated blood glucose levels that do not yet meet the diagnostic criteria for type 2 diabetes mellitus (T2D), affects nearly one–third of the Asian population and a substantial proportion of individuals worldwide (1, 2). Annually, approximately 5–10% of those with prediabetes progress to T2D (3), underscoring the urgency for a deeper understanding of prediabetes. Considerable emphasis has been placed on identifying effective methods to delay or prevent T2D onset among individuals with prediabetes. The Da Qing Diabetes Prevention Study, initiated in 1986, was the first major trial to demonstrate a significant impact, showing a 51% reduction in diabetes incidence following a 6–year intervention involving diet, exercise, or both, along with a reduction in cardiovascular disease (CVD) events among participants with prediabetes (4). The Finnish Diabetes Prevention Study, conducted in 1993, and the US Diabetes Prevention Program, launched in 1999, each demonstrated a 58% reduction in T2D incidence following approximately three years of lifestyle intervention (5). Despite these findings, debate persists regarding the optimal strategies for managing prediabetes, highlighting a need for continued research (6). Classifying prediabetes into distinct subgroups is essential to balance the health benefits of interventions against their economic cost–effectiveness, facilitating targeted and effective preventive measures for individuals at risk of T2D and related complications. Recent work by Wagner et al. (7) identified six distinct sub–phenotype clusters in Caucasian individuals with an elevated risk of T2D and its complications. However, physiological differences exist between Asian and Caucasian populations; for instance, the Thai population exhibits a higher prevalence of impaired β–cell function and increased susceptibility to the metabolic effects of obesity (8). A refined classification system could enable the identification of individuals at highest risk for T2D and its complications, facilitating individualized preventive interventions. Thus, investigating prediabetes sub–phenotypes within the Thai population could guide tailored preventive and therapeutic strategies for T2D. Accordingly, we aimed to classify prediabetes into sub–phenotypes through unsupervised, data–driven cluster analysis. We will examine clinical characteristics, T2D progression risk, associated complications, and genetic factors within each sub–phenotype. We hypothesize that distinct sub–phenotypes of prediabetes correlate variably with T2D and its complications, underscoring the need for tailored interventions. METHODS Study setting and study population We utilized secondary data from the Siriraj Healthy (SIH) study, conducted at Siriraj Hospital, Mahidol University, Bangkok, Thailand (9). The SIH study is a large, longitudinal cohort that includes various health professionals, such as doctors, nurses, pharmacists, and medical technicians, as well as support staff, including drivers, engineers, security officers, clerks, and academic personnel such as lecturers, researchers, and research assistants. Participants were recruited during their annual health checkups, with the registration period spanning from September 2017 to December 2020. Initially, 4,976 participants were examined at baseline. The inclusion criteria for our analysis were individuals aged ≥ 18 years who met prediabetes criteria (based on only HbA1c range of 5.7–6.4% or 39–46 mmol/L) (10), had no prior diagnosis of T2D, and had complete baseline data, including BMI, hemoglobin A1C (HbA1c), fasting blood glucose (FBG), lipid profiles, and waist circumference (WC). Individuals with any type of diabetes and CVD were excluded. Of the initial participants, 3,935 were excluded for various reasons, including 3,446 with normal glycemic status, 103 diagnosed with T2D, 176 with self–reported or drug–treated diabetes or HbA1c ≥ 6.5%, 1 with pre–existing cardiovascular disease, and 209 with uncertain self–reported T2D or missing baseline data. This process resulted in a cohort of 1,041 prediabetic individuals. After excluding 25 participants due to missing follow-up data, 1,016 prediabetic subjects were included in the analysis (Figure S1 ). All individuals were followed up, contributing a total of 4,180 person–years of follow–up. The remaining individuals with normal glycemic status were used as a control group in a genome–wide association study (GWAS) to compare genetic traits or risk factors with those of individuals with prediabetes. Outcome definitions of T2D, complications, and glycemic progression T2D was defined based on the first medical record indicated T2D during follow–up, using the ICD–10 codes E10 to E14, or when a HbA1c level of 6.5% (48 mmol/mol) or higher was recorded during follow–up, according to the American Diabetes Association guidelines (10). The diagnosis of CVD was determined during follow-up using specific ICD–10 codes. Codes I20 to I25 correspond to coronary heart disease and related conditions such as angina pectoris, myocardial infarction (heart attack), and other ischemic heart diseases. These codes encompass a range of cardiovascular conditions affecting the heart's blood vessels. Codes I60 to I64 pertain to cerebrovascular diseases, including stroke and transient cerebral ischemia. Individuals with known previous events were excluded. Glycemic progression was defined as the mean HbA1c during the study period, from the date of registry to a maximum of 7 years. Cluster analysis Variables for the clustering model were selected based on the premise that subjects develop T2D according to information extracted from their medical records. Seven routine clinical parameters were applied for this analysis: age at inclusion, BMI, WC, FBG, HbA1c, TG, and HDL–C. We applied the k–means clustering method with a k value of 4, using the k–means function in R version 4.3.3, allowing a maximum of max 10,000 iterations. To determine the optimal number of clusters, we used UMAP-based dimensionality reduction to visualize the resulting cluster distributions. We then applied the Elbow method and silhouette width for each clustering, varying the number of clusters from three to ten. Cluster stability was assessed using the Jaccard bootstrap method, with 1,000 resampling iterations performed via the R function clusterboot from the fpc package (v.2.2–12). We identified four clusters as the optimal solution, with a mean (SD) Jaccard similarity of 0.89 (0.07) across all clusters (Supplementary method 1, Figure S2, Table S1 ). Pairwise comparisons of the clustering variables between these four clusters are shown in Figure S3, where most differences reaching statistical significance after Bonferroni adjustment (p < 0.001). However, half of these differences were small or negligible, as indicated by Cohen’s d value (Table S2). Genotyping and genetic imputations In the SIH cohort, 3,960 individuals were selected for genotyping using the Infinium Asian Screening Array–24 v1.0 BeadChip (Illumina, San Diego, CA, USA). This array contains probes for 659,184 single nucleotide polymorphisms (SNPs). SNPs with call rate lower than 90% were excluded from further analysis, leaving 659,184 SNPs. Sample with call rate lower than 97% (n = 54) and those with inconsistent sex information between demographic and genotype data (n = 24) were excluded. Genotypes from the SIH cohort underwent quality controlled (QC) using PLINK software v1.9/2 (11). After QC, 3,879 individuals and 609,901 SNPs remained. Genotype imputation was performed on approximately 15 million genotypes using the Genomes Asia reference panels (12). Autosomal SNPs were pre–phased using SHAPEIT5 and imputed using IMPUTE5. The imputed genotypes were filtered based on minor allele frequency (MAF > 0.01), Hardy–Weinberg equilibrium (HWE, p–value > 1e–06), kinship (KING > 0.086) (13), and imputation info scores (INFO > 0.8). This left a final dataset of 3,080,191 variants from 3,768 high–quality samples (see Supplementary method 2 and Figures S4–5). These samples were used in GWAS to examine associations between genetic markers and phenotypic traits classified by clusters. Significant variants were visualized as Manhattan plots and QQ plots using the QQMAN package in R (14). Genome–wide PRS for T2D In this study, we evaluated all currently available T2D PRS listed in the polygenic score (PGS) catalog (data retrieval cut–off date: September 2024). We compared their performance in an independent cohort (146 PGSs) using the base summary statistics and PGS evaluations. After identifying the best–performing PGS, we evaluated its performance in the overall study population using a sum-of-effect weighting approach (Supplementary method 3). Association analysis between the T2D group (n = 374) and the control group (n = 3,505) was performed using logistic regression, adjusting for age, sex, and the first 10 principal components (PCs) in R. We also compared the proportion of the overlapping variants among the top 15 PGSs with the highest area under the receiver operating characteristic (AUROC) values (Table S3). The PGS catalog entry with ID: PGS004106 (15) achieved the highest AUROC value of 0.692 (95% confidence interval (CI) = 0.663–0.722) and exhibited an increased risk of T2D with an odds ratio (OR) of = 1.87 (95% CI = 1.35–2.59, p–value < 0.001) (Table S4). Therefore, we selected the best–performing model, PGS004106, which demonstrated strong performance across major population clusters and is suitable for use in individual risk assessments for T2D. Statistical analysis Baseline characteristics of participants, stratified by cluster, are presented as the mean and SD for continuous variables and as proportions for binary and categorical variables. The risk of developing T2D and CVD across clusters was estimated using multivariable Weibull proportional hazards regression models, adjusted for potential confounding variables. The Weibull model was selected because it had the lowest Akaike Information Criterion (AIC) (16) and provided the best fit for the data (Supplementary method 4 and Table S5). In the model assessing the risk of developing T2D (Model 1), adjustments were made for age at inclusion, sex, current smoking status, alcohol consumption, hypertension, dyslipidemia, systolic blood pressure (SBP), diastolic blood pressure (DBP), LDL–C, TC, and simple method for quantifying metabolic syndrome (siMS). For the model assessing the risk of CVD (Model 2), additional adjustments included MAU/Cr. Genetic risk scores were calculated based on the number of risk alleles weighed by their effect sizes, as reported in previous GWAS. The distribution of PRS between individuals with prediabetes and T2D was compared using the Mann-Whitney U Test due to nonparametric nature of the data. All statistical analyses were performed using Stata (Intercooled, version 17, Stata Corp, College Station, TX) and R 4.3.3, p–value less than 0.05 was considered as statistically significant. Ethics statement The study was conducted in accordance with the principles established by the Declaration of Helsinki and was approved by the Ethics Committee of the Human Research Protection Unit, Faculty of Medicine Siriraj Hospital, Mahidol University board (COA: Si 647/2016). All participants voluntarily provided written informed consent before enrollment. RESULTS Distribution of the clinical features by clusters Participants were categorized into four clusters based on baseline characteristics (Table 1 , Fig. 1 ): Cluster 1 (Mild obesity-related prediabetes (MORPD); 340 patients, 33.5%) featured moderate BMI and WC; Cluster 2 (Low-risk prediabetes (LORPD); 272 patients, 26.8%) was characterized by lower BMI and WC, representing a low–risk prediabetes profile; Cluster 3 (Metabolic syndrome prediabetes (MESPD); 140 patients, 13.8%) exhibited high BMI, elevated WC, a poor lipid profile, and elevated blood pressure, indicative of metabolic syndrome prediabetes; and Cluster 4 (Mild age-related prediabetes (MARPD; 264 patients, 26.0%) was defined by older age at onset and moderate metabolic disturbances, designated as mild age–related prediabetes. Table 1 Baseline characteristics of study participants in four clusters Parameters Cluster 1 MORPD Cluster 2 LORPD Cluster 3 MESPD Cluster 4 MARPD p-value (n = 340, 33.5%) (n = 272, 26.8%) (n = 140, 13.8%) (n = 264, 26.0%) Female 210 (59.8%) 241 (85.8%) 80 (55.2%) 166 (62.9%) < 0.001 Age (year) 36.62 ± 6.88 37.17 ± 9.07 36.91 ± 7.18 47.43 ± 7.24 < 0.001 BMI (kg/m 2 ) 27.56 ± 3.25 21.73 ± 2.53 33.43 ± 4.98 25.61 ± 2.97 < 0.001 Overweight 58 (16.5%) 41 (14.6%) 4 (2.8%) 58 (22.0%) Obesity 266 (75.8%) 28 (10.0%) 135 (93.1%) 152 (57.6%) WC (cm) 89.58 ± 6.97 73.88 ± 7.05 102.87 ± 10.78 85.71 ± 6.78 < 0.001 Male < 90, Female < 80 102 (29.1%) 245 (87.2%) 4 (2.8%) 113 (42.8%) Male ≥ 90, Female ≥ 80 238 (67.8%) 27 (9.6%) 136 (93.8%) 151 (57.2%) SBP (mmHg) 123.05 ± 12.13 115.75 ± 12.68 131.39 ± 14.44 125.48 ± 14.82 < 0.001 DBP (mmHg) 75.86 ± 11.27 70.17 ± 9.70 81.65 ± 10.54 76.35 ± 10.34 < 0.001 FBG (mmol/L) 5.01 ± 0.33 4.86 ± 0.33 5.60 ± 0.64 5.47 ± 0.44 < 0.001 HbA1c (mmol/L) 39.80 ± 1.17 39.60 ± 1.01 42.11 ± 2.17 42.18 ± 2.06 < 0.001 Lipid profiles TC (mmol/L) 5.07 ± 0.91 5.08 ± 0.82 5.20 ± 1.14 5.28 ± 0.81 0.016 TG (mmol/L) 1.28 ± 0.58 0.86 ± 0.44 2.44 ± 1.71 1.38 ± 0.63 < 0.001 HDL-C (mmol/L) 1.32 ± 0.25 1.93 ± 0.41 1.09 ± 0.26 1.41 ± 0.34 < 0.001 LDL-C (mmol/L) 3.15 ± 0.80 2.76 ± 0.74 3.06 ± 0.94 3.21 ± 0.79 < 0.001 siMS score 4.83 ± 0.40 4.75 ± 0.44 5.54 ± 0.96 5.02 ± 0.42 < 0.001 Urine eGFR (mL/min/1.73m2) 103.77 (91.72, 111.63) 101.18 (89.17, 108.67) 103.01 (94.58, 115.00) 95.71 (84.30, 104.33) < 0.001 MAU/Urine Cr ratio (mg/g) 3.9 (2.5, 6.6) 3.4 (2.5, 5.7) 5.6 (3.1, 10.3) 4.5 (2.6, 8.3) 0.032 Underlying diseases Hypertension, yes 15 (4.3%) 9 (3.2%) 22 (15.2%) 33 (12.5%) < 0.001 Dyslipidemia, yes 40 (11.4%) 31 (11.0%) 28 (19.3%) 61 (23.1%) < 0.001 Family history Heart failure, yes 24 (6.8%) 16 (5.7%) 6 (4.1%) 27 (10.2%) 0.887 Cancer, yes 64 (18.2%) 52 (18.5%) 25 (17.2%) 53 (20.1%) 0.007 Stroke, yes 21 (6.0%) 20 (7.1%) 6 (4.1%) 36 (13.6%) 0.003 Lifestyle Current smoker, yes 29 (8.3%) 7 (2.5%) 17 (11.7%) 17 (6.4%) 0.001 Alcohol consumption, yes 225 (64.1%) 161 (57.3%) 92 (63.4%) 140 (53.0%) 0.003 Data are expressed as mean ± SD, number (percentage), or median (interquartile range). BMI, body mass index; WC, waist circumference; SBP, systolic blood pressure; DBP, diastolic blood pressure; FBG, fasting blood glucose; HbA1c, hemoglobin A1c; TC, total cholesterol; TG, triglyceride; HDL-C, high density lipoprotein cholesterol; LDL-C, low density lipoprotein cholesterol; eGFR, estimated glomerular filtration rate; MAU/Urine Cr ratio, microalbumin/creatinine ratio; siMS score, simple method for quantifying metabolic syndrome; MORPD, mild obesity-related prediabetes; LORPD, low risk prediabetes; MESPD, metabolic syndrome prediabetes; MARPD, mild age-related prediabetes. Cross-tabulation of prediabetes definitions—classified based on HbA1c categories (5.7–6.4% or 39–46 mmol/L)—revealed that all participants in the MORPD, LORPD, and MARPD clusters were identified according to these HbA1c criteria. The prediabetes in the MESPD cluster had abnormal FBG and HbA1c simultaneously. The LORPD cluster had the highest proportion of female participants (85.8%), while the MESPD cluster had the lowest proportion of females (55.2%). Age differed significantly across clusters, with the MARPD cluster having the highest average age (47.43 ± 7.24 years). BMI was highest in the MESPD cluster (33.43 ± 4.98 kg/m²) and lowest in the LORPD cluster (21.73 ± 2.53 kg/m²), reflecting different obesity profiles across clusters. Similarly, WC was highest in the MESPD cluster (102.87 ± 10.78 cm) and lowest in the LORPD cluster (73.88 ± 7.05 cm). BP varied significantly among clusters, with the MESPD cluster exhibiting the highest systolic (131.4 ± 14.4 mmHg) and diastolic (81.7 ± 10.5 mmHg) blood pressure, suggesting an elevated risk for hypertension. FBG and HbA1c were elevated in the MESPD and MARPD clusters compared to the other clusters, indicating a higher risk of diabetes progression. Lipid profile analysis revealed significant differences across clusters: the MESPD cluster had the highest triglyceride (2.44 ± 1.71 mmol/L) and the lowest HDL–C (1.09 ± 0.26 mmol/L), indicating a more adverse lipid profile. Additionally, the MARPD cluster had the lowest estimated glomerular filtration rate (eGFR), median 95.71 ± mL/min/1.73m², which may suggest declining kidney function. The MESPD cluster also had the highest microalbuminuria/urine creatinine ratio, further indicating renal stress. Overall, these baseline characteristics underscore the distinct metabolic and demographic profiles of each prediabetes cluster. Associations of prediabetes clusters with major diseases A total of 1,016 participants were followed over a median period of 4.5 years to monitor the incidence of T2D and CVD. The MESPD cluster exhibited the highest risk for T2D, with an incidence rate of 86.9 cases per 1,000 person–years and 32.9% of participants diagnosed with T2D during follow–up. Similarly, for CVD, the MESPD cluster had an incidence rate of 5.7 cases per 1,000 person–years, compared to lower rates of 2.1 and 1.8 cases per 1,000 person–years in the MORPD and LORPD clusters, respectively (Table S6). Weibull proportional hazards models were used to compare the incidence of T2D and CVD across clusters, with the MORPD cluster as the reference group. The MESPD cluster had a significantly higher risk of developing T2D, with an adjusted hazard ratio (HR) of 5.86 (95% CI: 2.54–13.51, p < 0.001) compared to the LORPD cluster. The MARPD cluster also demonstrated an increased risk of T2D, with an adjusted HR of 4.36 (95% CI: 1.96–9.73, p < 0.001). In contrast, the MORPD cluster showed no significant difference in diabetes risk compared to the LORPD cluster (adjusted HR: 1.74, 95% CI: 0.77–3.94, p = 0.183). For CVD, there is no evidence of a significantly different risk among the four groups compared to the LORPD cluster (Table 2 , Fig. 2 , and Table S7). Table 2 Weibull proportional hazards models comparing risk of type 2 Diabetes and CVD by cluster Events Clusters Participants Person-years at risk Rate per 1000 Unadjusted HR (95% CI) Adjusted HR (95% CI) Type 2 Diabetes† Cluster 1 MORPD 340 1424 15.4 1.95 (0.89–4.24) 1.74 (0.77–3.94) Cluster 2 LORPD 272 1137 7.9 1.00 (ref.) 1.00 (ref.) Cluster 3 MESPD 140 529 86.85 11.09*** (5.43–22.67) 5.86*** (2.54–13.51) Cluster 4 MARPD 264 1089 41.33 5.21*** (2.54–10.65) 4.36*** (1.96–9.73) CVD‡ Cluster 1 MORPD 340 1423 2.8 1.59 (0.29–8.73) 0.42 (0.06–3.10) Cluster 2 LORPD 272 1137 1.8 1.00 (ref.) 1.00 (ref.) Cluster 3 MESPD 140 525 5.7 3.22 (0.54–19.27) 0.77 (0.09–6.75) Cluster 4 MARPD 264 1080 7.4 1.95 (0.89–19.84) 0.69 (0.11–4.32) *p < 0.05, **p < 0.01, ***p < 0.001. HR, hazard ratio; 95% CI, 95% confidence interval. †Adjusted for baseline age at inclusion, sex, current smoker, alcohol consumption, hypertension, dyslipidemia, systolic blood pressure (SBP), diastolic blood pressure (DBP), low density lipoprotein cholesterol (LDL-C), total cholesterol (TC) and siMS score. ‡ Adjusted for baseline self-reported family history of heart disease, age at inclusion, sex, current smoker, alcohol consumption, hypertension, dyslipidemia, SBP, DBP, LDL-C, TC, and MAU/Urine Cr ratio. MORPD, mild obesity-related prediabetes; LORPD, low risk prediabetes; MESPD, metabolic syndrome prediabetes; MARPD, mild age-related prediabetes. The study investigated transition patterns across distinct prediabetes clusters, focusing on outcomes such as normal glucose regulation (NGR), persistence of prediabetes, and progression to T2D. In the MORPD cluster, 245 (75.9%) individuals reverted to NGR, while 22 (6.8%) progressed to T2D. The LORPD cluster showed the highest reversion rate to NGR, with 219 (88.3%) individuals, and the lowest progression rate to T2D, with 9 (3.6%), indicating a lower risk of diabetes progression. In contrast, the MESPD cluster demonstrated the lowest reversion rate to NGR (52, 40.6%), with a significant proportion progressing to T2D (46, 35.9%), indicating the highest risk of T2D onset. Lastly, in the MARPD cluster, 111 (45.7%) individuals reverted to NGR, while 45 (18.5%) developed T2D, reflecting a moderate risk of progression to T2D (Fig. 3 and Table S8). During the follow-up period, changes in the biochemical parameters used to monitor prediabetes varied across the clusters. In particular, MARPD and MESPD showed a pronounced increase in both HbA1c and FBG, suggesting a higher likelihood of progressing to T2D compared with other clusters. Notably, MESPD had elevated TG and lower HDL-C, both of which are associated with an increased risk of CVD (Figure S6). Participants in the MARPD cluster had the highest cumulative incidence of T2D during the median follow-up period of 3.92 years, reflecting the cluster’s high-risk metabolic profile. PGS among different clusters The PRS analysis in individuals with prediabetes revealed notable patterns. The MESPD cluster exhibited the highest PRS, with a median value of 1.24, indicating a potentially elevated genetic risk, whereas the healthy control group displayed the lowest score, with a median value of 1.13 (Table S9 and Fig. 4 a). When comparing clusters in terms of T2D follow–up, the LORPD cluster had a significantly higher PGS score in the T2D subgroup compared to the other clusters, indicating a strong predictive ability of PRS for T2D for this cluster (Table S9 and Fig. 4 b). Additionally, the majority of MESPD individuals belonged to the highest PRS quintile with 9% in quintile 1 to 33% in quintile 4 (Fig. 4 c). Manhattan plots indicate no genome–wide significance SNPs (p < 5 × 10^–8) across all clusters (MORPD, LORPD, MESPD, and MARPD) (Figure S7). DISCUSSION This study presents a data–driven approach to classify prediabetes into distinct sub–phenotypes, offering a refined understanding of the heterogeneity with prediabetic states. Our findings identified four prediabetes subgroups—MORPD, LORPD, MESPD, and MARPD—each with unique clinical and metabolic characteristics and varied risks of progression to T2D and CVD. These clusters underscore the complexity of prediabetes and suggest potential benefits of tailored prevention strategies to prevent the development of T2D in the Thai population. The MESPD cluster, which was markedly characterized by metabolic syndrome–like features, showed the highest incidence of T2D. This high–risk profile is documented by elevated BMI, WC, and adverse atherogenic lipid profiles, consistent with previous studies linking metabolic syndrome components with increased cardiometabolic risk (17, 18). The substantially increased hazard ratios for T2D, even after adjusting for covariates, indicate severe metabolic disruption in MESPD, suggesting a need for targeted and intensive preventive measures. In contrast, the LORPD cluster, with lower BMI and WC, had the lowest risk of T2D and CVD, suggesting that metabolic burden plays a significant role in disease progression among prediabetic individuals (19, 20). The MARPD cluster, comprising older participants with relatively modest metabolic derangements, exhibited an elevated cumulative incidence of T2D over time. This finding suggests that age, combined with mild metabolic dysfunction, may significantly contribute to T2D risk. Prediabetes was common in community setting cohort study (21), particularly among aging population, and current evidence indicates that cardiovascular disease and mortality may be more critical prevention targets in this group than short–term diabetes progression (21). A systematic review of primarily middle–aged adults reported a 17% 6–year cumulative incidence of diabetes for those with HbA1c–defined prediabetes (Hba1c 5.7–6.4%) and a 22% incidence for individuals with impaired fasting glucose (IFG 100–125 mg/dL) (22). However, study in older populations, such as a Swedish cohort with 12 years of follow–up, found that HbA1c–defined prediabetes often remained stable or even regressed more frequently to normoglycemia than progressed to diabetes (23). Regression to normoglycemia appears more common in IFG–defined prediabetes than HbA1c–defined cases, likely due to greater variability in fasting glucose. These findings underscore the importance of addressing age–related metabolic risks alongside diabetes prevention in older adults. Similar to previous studies (24, 25), we observed a substantial number of individuals reverting from prediabetes to NGR. Nearly 80% of those in the MORPD and LORPD clusters, and approximately 40% in the MESPD and MARPD clusters, reverted to NGR. Additionally, more than one five of each prediabetes cluster sustained their prediabetes status over 3.5–year follow–up. This confirmed the higher risk of T2D progression in the MESPD and MARPD individuals. The Whitehall II study similarly reported that a majority of people with HbA1c–defined prediabetes persisted in that state over a 5–year period (25). These distinct cluster profiles reveal underlying risk factors and highlight the need for tailored management strategies, which could enhance prevention efforts and potentially halt the progression from prediabetes to diabetes. Prediabetes was associated with significantly worse liver function. It is proposed that T2D results from an excess of liver fat, which leads to an increased supply of fat to the pancreas, causing dysfunction in both organs and ultimately resulting in the progression of diabetes (26). MESPD cluster was associated with a higher risk of T2D. Although we did not measure liver function in our study, the higher BMI, WC, and hypertriglyceridemia may indicate greater levels of visceral fat accumulation among participants in the MESPD cluster. In contrast, the MESPD cluster did not demonstrate an increased risk of CVD. This finding may be explained by the relatively short follow-up period and the lower average age of participants (mean 39.5 years), which led to fewer observed events. Nonetheless, previous studies have indicated that metabolic dysfunction can be detrimental especially when accompanied by lipid abnormalities that promote CVD (27). Consequently, extended follow-up is warranted to fully elucidate the long-term risk of CVD. International medical organizations have defined prediabetes for over 10 years. However, there is still controversy over whether it should be classified as a distinct pathological condition. Prediabetes is characterized by various physiological abnormalities (28, 29), and the metabolic environment can lead to a broad range of glycemic fluctuations, spanning from normoglycemia on one end to diabetes mellitus on the other, depending on the stage of the process. In the present study, we analyzed only those with prediabetes due to the priority of diabetes prevention in this group. Several key trial have demonstrated that treating prediabetes with effective interventions could significantly alter the progression of T2D. Nonetheless, efforts to implement diabetes prevention in clinical practice have encountered some difficulties. Interventions often require substantial resources, and prediabetic individuals may be unaware of their hyperglycemic condition, or the effects may be subtle (30). Our classification of prediabetes may aid in identifying metabolic heterogeneity and guiding targeted interventions. Prediabetes in the high–risk MESPD and MARPD clusters should be prioritized due to their elevated risk of T2D and related complications. However, while clinical variables provide useful insights, they may lack precision, as prediabetes can shift between clusters over time. Advancing precision medicine for T2D prevention will require integrating multidimensional data, such as clinic, multiomics, and sensor–based behavioral information. Future studies should explore targeted interventions, like aerobic exercise and caloric restriction, to determine the most effective health benefits for each subgroup.(31). Future studies may investigate the types of interventions, such as aerobic exercise and dietary caloric restriction, that provide the greatest health benefits for individuals with prediabetes across various clusters. Differences in polygenetic risks support classification in this study. For instance, the higher T2D risk in the MESPD cluster correlated with an elevated PRS quintile, indicating a significant genetic contribution. The LORPD cluster also had elevated PRS scores among individuals who developed T2D, underscoring the predictive value of genetic risk assessment, even in low clinical risk groups. Higher PRS quintile have been linked to increased T2D risk across diverse populations, including East Asians (32). These findings suggest that genetic information support stratification of prediabetes patients into distinct clinical subgroups for targeted prevention, especially if future studies confirm differential responses to interventions or specific risks of progression. Although genetic factors significantly influence T2D risk (33), lifestyle interventions remain crucial, reinforcing the need for a holistic approach to diabetes prevention and management. Notably, the lack of genome–wide significant SNPs in the Manhattan plots suggests that although genetic predisposition contributes to T2D risk, other factors—such as environmental influences, lifestyle choices, and epigenetic changes—also play a significant role in the progression from prediabetes to T2D. Interventions in adults with prediabetes have shown effectiveness in reducing the risk of progression to diabetes, with lifestyle improvements being particularly impactful (34). The Diabetes Prevention Program (DPP) trial demonstrated that an intensive lifestyle intervention—focused on weight loss and increased physical activity—and, to a lesser extent, metformin use significantly reduced diabetes risk in adults aged 25 years and older (mean age 51) who were at high risk of diabetes progression (35). Consequently, current ADA guidelines recommend lifestyle interventions for adults with prediabetes (defined by HbA1c levels of 5.7–6.4%, fasting glucose levels of 100–125 mg/dL, or 2–hour glucose levels of 140–199 mg/dL) to achieve at least a 7% reduction in initial body weight and at least 150 minutes per week of moderate–intensity physical activity. Metformin is recommended for patients under 60 years with a BMI of 35 or greater or for women with a history of gestational diabetes (36, 37). These findings support the potential for lifestyle interventions to promote not only diabetes prevention but also prediabetes and diabetes remission (38). Evidence suggests that intensive lifestyle modifications can lead to significant improvements in glycemic control, allowing some individuals with prediabetes to revert to normoglycemia and achieve prediabetes remission, while others with early–stage diabetes may achieve diabetes remission (39). Given the lower risk of diabetes progression observed in older adults in this study, relative to mortality risk, aggressive pharmacologic interventions may offer limited benefit and could lead to unintended consequences, such as overdiagnosis, anxiety, or insurance implications. This underscores the importance of prioritizing safe, feasible lifestyle interventions that confer broad health benefits beyond diabetes prevention, especially in older adults (40). In this study, we employed a robust, data-driven clustering approach that combined k–means with UMAP, validated by the Elbow method, silhouette width, and Jaccard similarity, to reliably classify prediabetic individuals into distinct sub-phenotypes. Our findings are further supported by an extensive follow-up period and advanced statistical models like Weibull proportional hazards regression. Furthermore, incorporating polygenic risk scores (PRS) provided additional insights into the genetic contributions to type 2 diabetes risk. Despite the strengths of this study, several limitations should be acknowledged. The current definition of prediabetes neither reflects the sub-phenotypes of T2D pathophysiology nor accurately predicts future metabolic trajectories. Additionally, the relatively short follow-up period may limit the ability to capture long-term trends in disease progression, highlighting the need for further longitudinal studies to validate our findings. Conclusion In conclusion, our findings underscore the importance of stratifying individuals with prediabetes into distinct clusters based on specific metabolic and genetic profiles. This stratification enables more precise preventive strategies, particularly for high–risk groups such as the MESPD and MARPD clusters. Integrating PRS further strengthens early detection and personalizes intervention for those most vulnerable to T2D. Future studies should validate these findings across diverse populations and investigate the complex interactions among genetic, metabolic, and lifestyle factors that contribute to disease risk. Integrating genetic and metabolic profiling paves the way for precision–based management of prediabetes, establishing a new standard for proactive and tailored healthcare interventions. Declarations Acknowledgments: This research project is a collaboration between the Siriraj Medical Research Center and the Department of Preventive Medicine which is responsible for the staff health screening program. The authors gratefully acknowledge Ruengpung Sutthent and Sith Sathornsumetee, for her guidance for initiation of this project. We thank the members of the SIH Study Group, SPHERE staff members, and research assistants. The voluntary participation of all participants is highly appreciated. We are also grateful to the management of Siriraj Medical Research Center for providing office space, a recruitment center, and a biobank. Members of SIH study group: Winai Ratanasuwan, Keerati Charoencholvanich, Bhoom Suktitipat, Manop Pithukpakorn, Prapat Suriyaphol, Rungroj Krittayaphong, Prasert Auewarakul, Chalermchai Mitrpant, Boonrat Tassaneetritap, Mayuree Homsanit, and Naravat Poungvarin Members of SPHERE group: Sureeporn Pumeiam, Bonggochpass Pinsawas, Pichanun Mongkolsucharitkul, Apinya Surawit, Tanyaporn Pongkunakorn, Sophida Suta, Thamonwan Manosan, Suphawan Ophakas, and Korapat Mayurasakorn Members of Biobank: Somruedee Chatsiricharoenkul, Parichart Permpikul, Duangthip Apiratmontree, Sutee Udomchotphruet, and Pattranit Onsing We thank Sissades Tongsima, Tassathorn poonsin, and Pongsakorn Wangkumhang from National Biobank of Thailand for providing genome reference panel and advising the imputation steps. Funding: This work was supported by the Faculty of Medicine Siriraj Hospital, Mahidol University (R016034006). Additional funding support including infrastructure, staff and utilities, was provided by Faculty of Medicine Siriraj Hospital, Mahidol University. The funder had no role in the study design, data collection and analysis, decision to publish, and preparation of the manuscript Competing interests: None declared. Consent for publication: Not applicable. Data availability: Individual participant data will be shared with researchers in a deidentified or anonymized format upon submitting a research proposal and requesting data access to Associate Professor Korapat Mayurasakorn (contact person: [email protected] ). Data will be made available for analyses as approved by the data access committee. Patient consent for publication: Not required Author Contributors: A.S. and K.M. conceived of the analysis, developed the analysis plan, and wrote the initial drafts of the manuscript. A.S. and P.P. conducted the statistical analysis. A.S., S.P., S.S., P.M., B.P., S.O., and K.M. contributed with the data collection. K.M., A.S., P.P., S.P., S.S., P.M., B.P., N.V., P.V. and S.O. reviewed and revised the manuscript for important intellectual content. A.S. and K.M. oversees study implementation, assisted in writing and editing the paper and prepared the manuscript for publication; All authors have read the manuscript and are in agreement with the decision to submit the manuscript for journal publication. References Association AD. Diagnosis and classification of diabetes mellitus. Diabetes Care. 2010;33:S62–S9. Yip WCY, Sequeira IR, Plank LD, Poppitt SD. Prevalence of pre–diabetes across ethnicities: a review of impaired fasting glucose (IFG) and impaired glucose tolerance (IGT) for classification of dysglycaemia. Nutrients. 2017;9(11). Tabák AG, Herder C, Rathmann W, Brunner EJ, Kivimäki M. Prediabetes: a high–risk state for diabetes development. The Lancet. 2012;379(9833):2279–90. Gong Q, Zhang P, Wang J, Ma J, An Y, Chen Y, et al. Morbidity and mortality after lifestyle intervention for people with impaired glucose tolerance: 30–year results of the da qing diabetes prevention outcome study. Lancet Diabetes Endocrinol. 2019;7(6):452–61. Lindström J, Louheranta A, Mannelin M, Rastas M, Salminen V, Eriksson J, et al. The Finnish Diabetes Prevention Study (DPS): lifestyle intervention and 3–year results on diet and physical activity. Diabetes Care. 2003;26(12):3230–6. Piller C. Dubious diagnosis. Science. 2019;363(6431):1026–31. Wagner R, Heni M, Tabak AG, Machann J, Schick F, Randrianarisoa E, et al. Pathophysiology–based subphenotyping of individuals at elevated risk for type 2 diabetes. Nat Med. 2021;27(1):49–57. Do HD, Lohsoonthorn V, Jiamjarasrangsi W, Lertmaharit S, Williams MA. Prevalence of insulin resistance and its relationship with cardiovascular disease risk factors among Thai adults over 35 years old. Diabetes Res Clin Pract. 2010;89(3):303–8. Mongkolsucharitkul P, Surawit A, Manosan T, Ophakas S, Suta S, Pinsawas B, et al. Metabolic and genetic risk factors associated with pre–diabetes and type 2 diabetes in Thai healthcare employees: A long–term study from the Siriraj Health (SIH) cohort study. PLoS One. 2024;19(6):e0303085. Association AD. American diabetes association standards of care in diabetesd–2024. Diabetes Care. 2024;47:S1–S321. Chang CC, Chow CC, Tellier LC, Vattikuti S, Purcell SM, Lee JJ. Second–generation PLINK: rising to the challenge of larger and richer datasets. Gigascience. 2015;4:7. Wall JD, Stawiski EW, Ratan A, Kim HL, Kim C, Gupta R, et al. The Genome Asia 100K project enables genetic discoveries across Asia. Nature. 2019;576(7785):106–11. Manichaikul A, Mychaleckyj JC, Rich SS, Daly K, Sale M, Chen W–M. Robust relationship inference in genome–wide association studies. Bioinformatics. 2010;26(22):2867–73. Turner S. qqman: an R package for visualizing GWAS results using Q–Q and manhattan plots2014. Monti R, Eick L, Hudjashov G, Läll K, Kanoni S, Wolford BN, et al. Evaluation of polygenic scoring methods in five biobanks shows larger variation between biobanks than methods and finds benefits of ensemble learning. Am J Med Genet A. 2024;111(7):1431–47. Bedrick EJ, Tsai C–L. Model selection for multivariate regression in small samples. Biometrics. 1994;50(1):226–31. Biswas T, Townsend N, Gupta RD, Ghosh A, Rawal LB, Mørkrid K, et al. Clustering of metabolic and behavioural risk factors for cardiovascular diseases among the adult population in South and Southeast Asia: findings from WHO STEPS data. Lancet Reg Health Southeast Asia. 2023;12. Laucyte–Cibulskiene A, Chen C–H, Cockroft J, Cunha PG, Kavousi M, Laucevicius A, et al. Clusters of risk factors in metabolic syndrome and their influence on central blood pressure in a global study. Sci Rep. 2022;12(1):14409. Andréasson K, Edqvist J, Adiels M, Björck L, Lindgren M, Sattar N, et al. Body mass index in adolescence, risk of type 2 diabetes and associated complications: a nationwide cohort study of men. eClinicalMedicine. 2022;46. Bae JC, Cho NH, Kim JH, Hur KY, Jin SM, Lee MK. Association of body mass index with the risk of incident type 2 diabetes, cardiovascular disease, and all–cause mortality: a community–based prospective study. Endocrinol Metab (Seoul). 2020;35(2):416–24. Rooney MR, Rawlings AM, Pankow JS, Echouffo Tcheugui JB, Coresh J, Sharrett AR, et al. Risk of progression to diabetes among older adults with prediabetes. JAMA Intern Med. 2021;181(4):511–9. Richter B, Hemmingsen B, Metzendorf MI, Takwoingi Y. Development of type 2 diabetes mellitus in people with intermediate hyperglycaemia. Cochrane Database Syst Rev. 2018;10(10):Cd012661. Shang Y, Marseglia A, Fratiglioni L, Welmer AK, Wang R, Wang HX, et al. Natural history of prediabetes in older adults from a population–based longitudinal study. J Intern Med. 2019;286(3):326–40. Sallar A, Dagogo–Jack S. Regression from prediabetes to normal glucose regulation: state of the science. Exp Biol Med. 2020;245(10):889–96. Vistisen D, Kivimäki M, Perreault L, Hulman A, Witte DR, Brunner EJ, et al. Reversion from prediabetes to normoglycaemia and risk of cardiovascular disease and mortality: the Whitehall II cohort study. Diabetologia. 2019;62(8):1385–90. Taylor R. Type 2 diabetes and remission: practical management guided by pathophysiology. J Intern Med. 2021;289(6):754–70. Hill MA, Yang Y, Zhang L, Sun Z, Jia G, Parrish AR, et al. Insulin resistance, cardiovascular stiffening and cardiovascular disease. Metabolism. 2021;119:154766. Blond MB, Færch K, Herder C, Ziegler D, Stehouwer CDA. The prediabetes conundrum: striking the balance between risk and resources. Diabetologia. 2023;66(6):1016–23. Barbu E, Popescu MR, Popescu AC, Balanescu SM. Phenotyping the prediabetic population–a closer look at intermediate glucose status and cardiovascular disease. Int J Mol Sci. 2021;22(13). Echouffo–Tcheugui JB, Selvin E. Prediabetes and what it means: The epidemiological evidence. Annu Rev Public Health. 2021;42(Volume 42, 2021):59–77. Del Prato S. Heterogeneity of diabetes: heralding the era of precision medicine. Lancet Diabetes Endocrinol. 2019;7(9):659–61. Lee H, Choi J, Kim JI, Watanabe RM, Cho NH, Park KS, et al. Higher genetic risk for type 2 diabetes is associated with a faster decline of β–cell function in an east asian population. Diabetes Care. 2024;47(8):1386–94. Udler MS, Kim J, von Grotthuss M, Bonàs–Guarch S, Cole JB, Chiou J, et al. Type 2 diabetes genetic loci informed by multi–trait associations point to disease mechanisms and subtypes: A soft clustering analysis. PLOS Medicine. 2018;15(9):e1002654. Peña A, Olson ML, Hooker E, Ayers SL, Castro FG, Patrick DL, et al. Effects of a diabetes prevention program on type 2 diabetes risk factors and quality of life among latino youths with prediabetes: a randomized clinical trial. JAMA Network Open. 2022;5(9):e2231196–e. Knowler WC, Fowler SE, Hamman RF, Christophi CA, Hoffman HJ, Brenneman AT, et al. 10–year follow–up of diabetes incidence and weight loss in the diabetes prevention program outcomes study. Lancet. 2009;374(9702):1677–86. Boltri JM, Tracer H, Strogatz D, Idzik S, Schumacher P, Fukagawa N, et al. The national clinical care commission report to congress: leveraging federal policies and programs to prevent diabetes in people with prediabetes. Diabetes Care. 2023;46(2):e39–e50. Association AD. Prevention or delay of type 2 diabetes: standards of medical care in diabetes—2020. Diabetes Care. 2019;43(Supplement_1):S32–S6. Lean MEJ, Leslie WS, Barnes AC, Brosnahan N, Thom G, McCombie L, et al. 5–year follow–up of the randomised Diabetes Remission Clinical Trial (DiRECT) of continued support for weight loss maintenance in the UK: an extension study. Lancet Diabetes Endocrinol. 2024;12(4):233–46. Rosenfeld RM, Kelly JH, Agarwal M, Aspry K, Barnett T, Davis BC, et al. Dietary interventions to treat type 2 diabetes in adults with a goal of remission: An expert consensus statement from the american college of lifestyle medicine. Am J Lifestyle Med. 2022;16(3):342–62. Bradley MD, Arnold ME, Biskup BG, Campbell TM, 2nd, Fuhrman J, Guthrie GE, et al. Medication deprescribing among patients with type 2 diabetes: a qualitative case series of lifestyle medicine practitioner protocols. Clin Diabetes. 2023;41(2):163–76. Additional Declarations No competing interests reported. Supplementary Files Supplementfile.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6351460","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":453085312,"identity":"46ae0d15-bd44-4a53-b6e7-1de825606cac","order_by":0,"name":"Apinya Surawit","email":"","orcid":"","institution":"Siriraj Population Health and Nutrition Research Group (SPHERE), Research Group and Research Network Division, Research Department, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok","correspondingAuthor":false,"prefix":"","firstName":"Apinya","middleName":"","lastName":"Surawit","suffix":""},{"id":453085313,"identity":"2905857a-8d74-4ae3-8651-8bd7067bff87","order_by":1,"name":"Phongthana Pasookhush","email":"","orcid":"","institution":"Siriraj Population Health and Nutrition Research Group (SPHERE), Research Group and Research Network Division, Research Department, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok","correspondingAuthor":false,"prefix":"","firstName":"Phongthana","middleName":"","lastName":"Pasookhush","suffix":""},{"id":453085314,"identity":"09d634ed-126b-4b70-b1dd-e5a37bd65623","order_by":2,"name":"Sureeporn Pumeiam","email":"","orcid":"","institution":"Siriraj Population Health and Nutrition Research Group (SPHERE), Research Group and Research Network Division, Research Department, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok","correspondingAuthor":false,"prefix":"","firstName":"Sureeporn","middleName":"","lastName":"Pumeiam","suffix":""},{"id":453085315,"identity":"d95b1018-dfd1-4342-9ae1-e04794073084","order_by":3,"name":"Pichanun Mongkolsucharitkul","email":"","orcid":"","institution":"Siriraj Population Health and Nutrition Research Group (SPHERE), Research Group and Research Network Division, Research Department, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok","correspondingAuthor":false,"prefix":"","firstName":"Pichanun","middleName":"","lastName":"Mongkolsucharitkul","suffix":""},{"id":453085316,"identity":"d9d6eebc-fce5-486d-b257-036c8059f03f","order_by":4,"name":"Sophida Suta","email":"","orcid":"","institution":"Siriraj Population Health and Nutrition Research Group (SPHERE), Research Group and Research Network Division, Research Department, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok","correspondingAuthor":false,"prefix":"","firstName":"Sophida","middleName":"","lastName":"Suta","suffix":""},{"id":453085317,"identity":"4b6bb0f8-4226-4254-8c9a-2ed31812b9df","order_by":5,"name":"Bonggochpass Pinsawas","email":"","orcid":"","institution":"Siriraj Population Health and Nutrition Research Group (SPHERE), Research Group and Research Network Division, Research Department, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok","correspondingAuthor":false,"prefix":"","firstName":"Bonggochpass","middleName":"","lastName":"Pinsawas","suffix":""},{"id":453085319,"identity":"19d16d2f-c902-4e95-b473-bbe615e9e8fc","order_by":6,"name":"Suphawan Ophakas","email":"","orcid":"","institution":"Siriraj Population Health and Nutrition Research Group (SPHERE), Research Group and Research Network Division, Research Department, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok","correspondingAuthor":false,"prefix":"","firstName":"Suphawan","middleName":"","lastName":"Ophakas","suffix":""},{"id":453085320,"identity":"46de676b-e029-4d78-b38d-23f99434cc7e","order_by":7,"name":"Nutchavadee Vorasan","email":"","orcid":"","institution":"Genetic Epidemiology Group, Research Department, Faculty of Medicine Siriraj Hospital, Mahidol University","correspondingAuthor":false,"prefix":"","firstName":"Nutchavadee","middleName":"","lastName":"Vorasan","suffix":""},{"id":453085322,"identity":"c16bd14a-1f1b-4e49-b8a5-c9bc8625cfad","order_by":8,"name":"Parit Voharnsuchon","email":"","orcid":"","institution":"Postgraduate Education Devision, Faculty of Medicine Siriraj Hospital, Mahidol University","correspondingAuthor":false,"prefix":"","firstName":"Parit","middleName":"","lastName":"Voharnsuchon","suffix":""},{"id":453085323,"identity":"e8c2bf6b-bf96-4522-b04b-cd21f626f3be","order_by":9,"name":"Korapat Mayurasakorn","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABEUlEQVRIiWNgGAWjYBACCSQ24wMGhgM8DAeYG8BcA/xawNLMBhAtjGAtEsRoYQNyDjAQ1CLZ3vvwc0HNHwbd9jNmVTdq7sjwHW9sYPhRw1BnjkOLNM9xY+kZxwwYzM7kmN3OOfaMR/LMwQbGnmMMEpYN2LXISaQBtbEBtRwAaWE7zGNwI7GBgbcB6LADOLUw/+b5B9Ry/o1Zcc4/iBbGv3i0SEuksUnztgG13MgxY85tg2hhxmeLZM8xNuuZfcY8ZjeeFUvn9h0G++WwzDEJyQ04tEgcb2O+XfBNTs7sfPLGzznfDtvzHW8++PBNjQ0/LltAgBmIeRgYOBARcQA1WWDXAgTsD/CqGgWjYBSMgpELALGzXEzM9VQVAAAAAElFTkSuQmCC","orcid":"","institution":"Siriraj Population Health and Nutrition Research Group (SPHERE), Research Group and Research Network Division, Research Department, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok","correspondingAuthor":true,"prefix":"","firstName":"Korapat","middleName":"","lastName":"Mayurasakorn","suffix":""}],"badges":[],"createdAt":"2025-04-01 09:08:20","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6351460/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6351460/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":82349355,"identity":"7ae7f5fc-72c9-4c77-a910-c9ad9e8a6491","added_by":"auto","created_at":"2025-05-09 10:46:48","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1249981,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDistribution of the cluster feature variables by clusters\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDistributions according to baseline (a) fasting blood glucose (FBG) (b) hemoglobin A1c (HbA1c), (c) high density lipoprotein cholesterol (HDL-C), (d) triglyceride (TG), (e) age at inclusion, (f) body mass index (BMI), and (g) waist circumference (WC) for each cluster and (h) distribution of participants. To classify patients into diabetes subgroups, a k-means analysis was performed using seven clustering variables. In the population, mild obesity-related prediabetes (MORPD) was the most frequent cluster (33.5%), followed by low-risk prediabetes (LORPD) (26.8%), metabolic syndrome prediabetes (MESPD) (13.8%), and mild age-related prediabetes (MARPD) (26.0%). MORPD had the highest BMI and intermediate metabolic control. MARPD had an older age and relatively lower WC at diagnosis. Individuals in the MESPD cluster had an age at diagnosis intermediate between MORPD and MARPD, the highest triglycerides, the lowest HDL-C, high HbA1c, high BMI, and high WC. LORPD had the lowest BMI, the highest HDL-C, the lowest WC, and the best metabolic control.\u003c/p\u003e","description":"","filename":"Figure1.Distributionoftheclusterfeaturevariablesbyclusters.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6351460/v1/b6e9e22bce0c9ec4f79b76dd.jpg"},{"id":82351329,"identity":"617134ac-a3ce-402a-bb4f-ecbf740db519","added_by":"auto","created_at":"2025-05-09 10:54:49","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":5133583,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eProgression of disease over time by cluster.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe cumulative incidences of (a) type 2 diabetes (T2D) and (b) cardiovascular disease (CVD) were calculated. The incidence of T2D and CVD was determined by dividing the number of incident cases by the total observation duration (person-years) in each cluster. Incident T2D and CVD between clusters were analyzed using Cox proportional hazards models. For each outcome, the cluster with the lowest incidence was used as the reference group. Model (A) was adjusted for baseline sex, current smoker status, alcohol consumption, hypertension, dyslipidemia, systolic blood pressure (SBP), diastolic blood pressure (DBP), low density lipoprotein cholesterol (LDL-C), total cholesterol (TC) and siMS score. Model (B) was adjusted for baseline self-reported family history of heart disease, sex, current smoker status, alcohol consumption, hypertension, dyslipidemia, SBP, DBP, LDL-C, TC, and MAU/Urine Cr ratio. Mild obesity-related prediabetes (MORPD), low-risk prediabetes (LORPD), metabolic syndrome prediabetes (MESPD), and mild age-related prediabetes (MARPD).\u003c/p\u003e","description":"","filename":"Figure2.Progressionofdiseaseovertimebycluster.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6351460/v1/eeb9ae5295890d747c72efec.jpg"},{"id":82349353,"identity":"3b58a90d-24cd-4bca-ba7e-4fa23d54441e","added_by":"auto","created_at":"2025-05-09 10:46:48","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":4483301,"visible":true,"origin":"","legend":"\u003cp\u003eTransitions from prediabetes clusters at baseline to normal glucose regulation (NGR), prediabetes clusters and type 2 diabetes clusters at follow-up.\u003c/p\u003e","description":"","filename":"Figure3.Transitionsfromprediabetesclusters.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6351460/v1/3c68310e1172b2d76a75fccb.jpg"},{"id":82352760,"identity":"74c6606f-8ff2-480d-957f-3c03c4127cb7","added_by":"auto","created_at":"2025-05-09 11:02:49","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":6150738,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDistribution and association of type 2 diabetes (T2D) polygenic risk scores (PRS) with prediabetes subtypes.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe density plot and box plot show the distribution of T2D PRS across five different subtypes: No diabetes, Mild obesity-related prediabetes (MORPD), low-risk prediabetes (LORPD), metabolic syndrome prediabetes (MESPD), and mild age-related prediabetes (MARPD) each represented by a different color (a). The compares T2D PRS between individuals with prediabetes and type 2 diabetes (T2D) within the four subtypes (b). The proportion of individuals falling into four T2D PRS quartiles (1-4) across the five subtypes (c).\u003c/p\u003e","description":"","filename":"Figure4.Distributionandassociationoftype2diabetesT2DpolygenicriskscoresPRSwithprediabetessub.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6351460/v1/fb65cab4b565d7900fb60014.jpg"},{"id":82349359,"identity":"dfdb4e2a-3f97-4846-88f4-b51251c1fdae","added_by":"auto","created_at":"2025-05-09 10:46:48","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1246093,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementfile.docx","url":"https://assets-eu.researchsquare.com/files/rs-6351460/v1/1664623d4b06303aa2f23032.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Novel subgroups of prediabetes and the associations with outcomes in health professionals: a data–driven cluster analysis","fulltext":[{"header":"INTRODUCTION","content":"\u003cp\u003ePrediabetes, defined as elevated blood glucose levels that do not yet meet the diagnostic criteria for type 2 diabetes mellitus (T2D), affects nearly one\u0026ndash;third of the Asian population and a substantial proportion of individuals worldwide (1, 2). Annually, approximately 5\u0026ndash;10% of those with prediabetes progress to T2D (3), underscoring the urgency for a deeper understanding of prediabetes. Considerable emphasis has been placed on identifying effective methods to delay or prevent T2D onset among individuals with prediabetes. The Da Qing Diabetes Prevention Study, initiated in 1986, was the first major trial to demonstrate a significant impact, showing a 51% reduction in diabetes incidence following a 6\u0026ndash;year intervention involving diet, exercise, or both, along with a reduction in cardiovascular disease (CVD) events among participants with prediabetes (4). The Finnish Diabetes Prevention Study, conducted in 1993, and the US Diabetes Prevention Program, launched in 1999, each demonstrated a 58% reduction in T2D incidence following approximately three years of lifestyle intervention (5). Despite these findings, debate persists regarding the optimal strategies for managing prediabetes, highlighting a need for continued research (6).\u003c/p\u003e \u003cp\u003eClassifying prediabetes into distinct subgroups is essential to balance the health benefits of interventions against their economic cost\u0026ndash;effectiveness, facilitating targeted and effective preventive measures for individuals at risk of T2D and related complications. Recent work by Wagner et al. (7) identified six distinct sub\u0026ndash;phenotype clusters in Caucasian individuals with an elevated risk of T2D and its complications. However, physiological differences exist between Asian and Caucasian populations; for instance, the Thai population exhibits a higher prevalence of impaired β\u0026ndash;cell function and increased susceptibility to the metabolic effects of obesity (8). A refined classification system could enable the identification of individuals at highest risk for T2D and its complications, facilitating individualized preventive interventions. Thus, investigating prediabetes sub\u0026ndash;phenotypes within the Thai population could guide tailored preventive and therapeutic strategies for T2D. Accordingly, we aimed to classify prediabetes into sub\u0026ndash;phenotypes through unsupervised, data\u0026ndash;driven cluster analysis. We will examine clinical characteristics, T2D progression risk, associated complications, and genetic factors within each sub\u0026ndash;phenotype. We hypothesize that distinct sub\u0026ndash;phenotypes of prediabetes correlate variably with T2D and its complications, underscoring the need for tailored interventions.\u003c/p\u003e"},{"header":"METHODS","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy setting and study population\u003c/h2\u003e \u003cp\u003eWe utilized secondary data from the Siriraj Healthy (SIH) study, conducted at Siriraj Hospital, Mahidol University, Bangkok, Thailand (9). The SIH study is a large, longitudinal cohort that includes various health professionals, such as doctors, nurses, pharmacists, and medical technicians, as well as support staff, including drivers, engineers, security officers, clerks, and academic personnel such as lecturers, researchers, and research assistants. Participants were recruited during their annual health checkups, with the registration period spanning from September 2017 to December 2020.\u003c/p\u003e \u003cp\u003eInitially, 4,976 participants were examined at baseline. The inclusion criteria for our analysis were individuals aged\u0026thinsp;\u0026ge;\u0026thinsp;18 years who met prediabetes criteria (based on only HbA1c range of 5.7\u0026ndash;6.4% or 39\u0026ndash;46 mmol/L) (10), had no prior diagnosis of T2D, and had complete baseline data, including BMI, hemoglobin A1C (HbA1c), fasting blood glucose (FBG), lipid profiles, and waist circumference (WC). Individuals with any type of diabetes and CVD were excluded. Of the initial participants, 3,935 were excluded for various reasons, including 3,446 with normal glycemic status, 103 diagnosed with T2D, 176 with self\u0026ndash;reported or drug\u0026ndash;treated diabetes or HbA1c\u0026thinsp;\u0026ge;\u0026thinsp;6.5%, 1 with pre\u0026ndash;existing cardiovascular disease, and 209 with uncertain self\u0026ndash;reported T2D or missing baseline data. This process resulted in a cohort of 1,041 prediabetic individuals. After excluding 25 participants due to missing follow-up data, 1,016 prediabetic subjects were included in the analysis (Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). All individuals were followed up, contributing a total of 4,180 person\u0026ndash;years of follow\u0026ndash;up. The remaining individuals with normal glycemic status were used as a control group in a genome\u0026ndash;wide association study (GWAS) to compare genetic traits or risk factors with those of individuals with prediabetes.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eOutcome definitions of T2D, complications, and glycemic progression\u003c/h3\u003e\n\u003cp\u003e T2D was defined based on the first medical record indicated T2D during follow\u0026ndash;up, using the ICD\u0026ndash;10 codes E10 to E14, or when a HbA1c level of 6.5% (48 mmol/mol) or higher was recorded during follow\u0026ndash;up, according to the American Diabetes Association guidelines (10). The diagnosis of CVD was determined during follow-up using specific ICD\u0026ndash;10 codes. Codes I20 to I25 correspond to coronary heart disease and related conditions such as angina pectoris, myocardial infarction (heart attack), and other ischemic heart diseases. These codes encompass a range of cardiovascular conditions affecting the heart's blood vessels. Codes I60 to I64 pertain to cerebrovascular diseases, including stroke and transient cerebral ischemia. Individuals with known previous events were excluded. Glycemic progression was defined as the mean HbA1c during the study period, from the date of registry to a maximum of 7 years.\u003c/p\u003e\n\u003ch3\u003eCluster analysis\u003c/h3\u003e\n\u003cp\u003eVariables for the clustering model were selected based on the premise that subjects develop T2D according to information extracted from their medical records. Seven routine clinical parameters were applied for this analysis: age at inclusion, BMI, WC, FBG, HbA1c, TG, and HDL\u0026ndash;C. We applied the k\u0026ndash;means clustering method with a k value of 4, using the k\u0026ndash;means function in R version 4.3.3, allowing a maximum of max 10,000 iterations. To determine the optimal number of clusters, we used UMAP-based dimensionality reduction to visualize the resulting cluster distributions. We then applied the Elbow method and silhouette width for each clustering, varying the number of clusters from three to ten. Cluster stability was assessed using the Jaccard bootstrap method, with 1,000 resampling iterations performed via the R function clusterboot from the fpc package (v.2.2\u0026ndash;12). We identified four clusters as the optimal solution, with a mean (SD) Jaccard similarity of 0.89 (0.07) across all clusters (Supplementary method 1, Figure S2, Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). Pairwise comparisons of the clustering variables between these four clusters are shown in Figure S3, where most differences reaching statistical significance after Bonferroni adjustment (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). However, half of these differences were small or negligible, as indicated by Cohen\u0026rsquo;s d value (Table S2).\u003c/p\u003e\n\u003ch3\u003eGenotyping and genetic imputations\u003c/h3\u003e\n\u003cp\u003eIn the SIH cohort, 3,960 individuals were selected for genotyping using the Infinium Asian Screening Array\u0026ndash;24 v1.0 BeadChip (Illumina, San Diego, CA, USA). This array contains probes for 659,184 single nucleotide polymorphisms (SNPs). SNPs with call rate lower than 90% were excluded from further analysis, leaving 659,184 SNPs. Sample with call rate lower than 97% (n\u0026thinsp;=\u0026thinsp;54) and those with inconsistent sex information between demographic and genotype data (n\u0026thinsp;=\u0026thinsp;24) were excluded. Genotypes from the SIH cohort underwent quality controlled (QC) using PLINK software v1.9/2 (11). After QC, 3,879 individuals and 609,901 SNPs remained. Genotype imputation was performed on approximately 15\u0026nbsp;million genotypes using the Genomes Asia reference panels (12). Autosomal SNPs were pre\u0026ndash;phased using SHAPEIT5 and imputed using IMPUTE5. The imputed genotypes were filtered based on minor allele frequency (MAF\u0026thinsp;\u0026gt;\u0026thinsp;0.01), Hardy\u0026ndash;Weinberg equilibrium (HWE, p\u0026ndash;value\u0026thinsp;\u0026gt;\u0026thinsp;1e\u0026ndash;06), kinship (KING\u0026thinsp;\u0026gt;\u0026thinsp;0.086) (13), and imputation info scores (INFO\u0026thinsp;\u0026gt;\u0026thinsp;0.8). This left a final dataset of 3,080,191 variants from 3,768 high\u0026ndash;quality samples (see Supplementary method 2 and Figures S4\u0026ndash;5). These samples were used in GWAS to examine associations between genetic markers and phenotypic traits classified by clusters. Significant variants were visualized as Manhattan plots and QQ plots using the QQMAN package in R (14).\u003c/p\u003e\n\u003ch3\u003eGenome–wide PRS for T2D\u003c/h3\u003e\n\u003cp\u003eIn this study, we evaluated all currently available T2D PRS listed in the polygenic score (PGS) catalog (data retrieval cut\u0026ndash;off date: September 2024). We compared their performance in an independent cohort (146 PGSs) using the base summary statistics and PGS evaluations. After identifying the best\u0026ndash;performing PGS, we evaluated its performance in the overall study population using a sum-of-effect weighting approach (Supplementary method 3). Association analysis between the T2D group (n\u0026thinsp;=\u0026thinsp;374) and the control group (n\u0026thinsp;=\u0026thinsp;3,505) was performed using logistic regression, adjusting for age, sex, and the first 10 principal components (PCs) in R. We also compared the proportion of the overlapping variants among the top 15 PGSs with the highest area under the receiver operating characteristic (AUROC) values (Table S3). The PGS catalog entry with ID: PGS004106 (15) achieved the highest AUROC value of 0.692 (95% confidence interval (CI)\u0026thinsp;=\u0026thinsp;0.663\u0026ndash;0.722) and exhibited an increased risk of T2D with an odds ratio (OR) of =\u0026thinsp;1.87 (95% CI\u0026thinsp;=\u0026thinsp;1.35\u0026ndash;2.59, p\u0026ndash;value\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Table S4). Therefore, we selected the best\u0026ndash;performing model, PGS004106, which demonstrated strong performance across major population clusters and is suitable for use in individual risk assessments for T2D.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eBaseline characteristics of participants, stratified by cluster, are presented as the mean and SD for continuous variables and as proportions for binary and categorical variables. The risk of developing T2D and CVD across clusters was estimated using multivariable Weibull proportional hazards regression models, adjusted for potential confounding variables. The Weibull model was selected because it had the lowest Akaike Information Criterion (AIC) (16) and provided the best fit for the data (Supplementary method 4 and Table S5). In the model assessing the risk of developing T2D (Model 1), adjustments were made for age at inclusion, sex, current smoking status, alcohol consumption, hypertension, dyslipidemia, systolic blood pressure (SBP), diastolic blood pressure (DBP), LDL\u0026ndash;C, TC, and simple method for quantifying metabolic syndrome (siMS). For the model assessing the risk of CVD (Model 2), additional adjustments included MAU/Cr.\u003c/p\u003e \u003cp\u003eGenetic risk scores were calculated based on the number of risk alleles weighed by their effect sizes, as reported in previous GWAS. The distribution of PRS between individuals with prediabetes and T2D was compared using the Mann-Whitney U Test due to nonparametric nature of the data. All statistical analyses were performed using Stata (Intercooled, version 17, Stata Corp, College Station, TX) and R 4.3.3, p\u0026ndash;value less than 0.05 was considered as statistically significant.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eEthics statement\u003c/h3\u003e\n\u003cp\u003e The study was conducted in accordance with the principles established by the Declaration of Helsinki and was approved by the Ethics Committee of the Human Research Protection Unit, Faculty of Medicine Siriraj Hospital, Mahidol University board (COA: Si 647/2016). All participants voluntarily provided written informed consent before enrollment.\u003c/p\u003e"},{"header":"RESULTS","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eDistribution of the clinical features by clusters\u003c/h2\u003e \u003cp\u003eParticipants were categorized into four clusters based on baseline characteristics (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e): Cluster 1 (Mild obesity-related prediabetes (MORPD); 340 patients, 33.5%) featured moderate BMI and WC; Cluster 2 (Low-risk prediabetes (LORPD); 272 patients, 26.8%) was characterized by lower BMI and WC, representing a low\u0026ndash;risk prediabetes profile; Cluster 3 (Metabolic syndrome prediabetes (MESPD); 140 patients, 13.8%) exhibited high BMI, elevated WC, a poor lipid profile, and elevated blood pressure, indicative of metabolic syndrome prediabetes; and Cluster 4 (Mild age-related prediabetes (MARPD; 264 patients, 26.0%) was defined by older age at onset and moderate metabolic disturbances, designated as mild age\u0026ndash;related prediabetes.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eBaseline characteristics of study participants in four clusters\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"11\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c11\" colnum=\"11\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colspan=\"2\" morerows=\"1\" nameend=\"c2\" namest=\"c1\" rowspan=\"2\"\u003e \u003cp\u003eParameters\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003eCluster 1\u003c/p\u003e \u003cp\u003eMORPD\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003eCluster 2\u003c/p\u003e \u003cp\u003eLORPD\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003eCluster 3\u003c/p\u003e \u003cp\u003eMESPD\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003eCluster 4\u003c/p\u003e \u003cp\u003eMARPD\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c11\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003ep-value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;340, 33.5%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;272, 26.8%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;140, 13.8%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;264, 26.0%)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eFemale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e210\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(59.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e241\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e(85.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e(55.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e166\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e(62.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eAge (year)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e36.62\u0026thinsp;\u0026plusmn;\u0026thinsp;6.88\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e37.17\u0026thinsp;\u0026plusmn;\u0026thinsp;9.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e36.91\u0026thinsp;\u0026plusmn;\u0026thinsp;7.18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003e47.43\u0026thinsp;\u0026plusmn;\u0026thinsp;7.24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eBMI (kg/m\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e27.56\u0026thinsp;\u0026plusmn;\u0026thinsp;3.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e21.73\u0026thinsp;\u0026plusmn;\u0026thinsp;2.53\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e33.43\u0026thinsp;\u0026plusmn;\u0026thinsp;4.98\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003e25.61\u0026thinsp;\u0026plusmn;\u0026thinsp;2.97\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOverweight\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e58\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(16.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e(14.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e(2.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e58\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e(22.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eObesity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e266\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(75.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e(10.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e135\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e(93.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e152\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e(57.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eWC (cm)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e89.58\u0026thinsp;\u0026plusmn;\u0026thinsp;6.97\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e73.88\u0026thinsp;\u0026plusmn;\u0026thinsp;7.05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e102.87\u0026thinsp;\u0026plusmn;\u0026thinsp;10.78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003e85.71\u0026thinsp;\u0026plusmn;\u0026thinsp;6.78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMale\u0026thinsp;\u0026lt;\u0026thinsp;90, Female\u0026thinsp;\u0026lt;\u0026thinsp;80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e102\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(29.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e245\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e(87.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e(2.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e113\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e(42.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMale\u0026thinsp;\u0026ge;\u0026thinsp;90, Female\u0026thinsp;\u0026ge;\u0026thinsp;80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e238\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(67.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e(9.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e136\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e(93.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e151\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e(57.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eSBP (mmHg)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e123.05\u0026thinsp;\u0026plusmn;\u0026thinsp;12.13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e115.75\u0026thinsp;\u0026plusmn;\u0026thinsp;12.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e131.39\u0026thinsp;\u0026plusmn;\u0026thinsp;14.44\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003e125.48\u0026thinsp;\u0026plusmn;\u0026thinsp;14.82\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eDBP (mmHg)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e75.86\u0026thinsp;\u0026plusmn;\u0026thinsp;11.27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e70.17\u0026thinsp;\u0026plusmn;\u0026thinsp;9.70\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e81.65\u0026thinsp;\u0026plusmn;\u0026thinsp;10.54\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003e76.35\u0026thinsp;\u0026plusmn;\u0026thinsp;10.34\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eFBG (mmol/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e5.01\u0026thinsp;\u0026plusmn;\u0026thinsp;0.33\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e4.86\u0026thinsp;\u0026plusmn;\u0026thinsp;0.33\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e5.60\u0026thinsp;\u0026plusmn;\u0026thinsp;0.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003e5.47\u0026thinsp;\u0026plusmn;\u0026thinsp;0.44\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eHbA1c (mmol/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e39.80\u0026thinsp;\u0026plusmn;\u0026thinsp;1.17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e39.60\u0026thinsp;\u0026plusmn;\u0026thinsp;1.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e42.11\u0026thinsp;\u0026plusmn;\u0026thinsp;2.17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003e42.18\u0026thinsp;\u0026plusmn;\u0026thinsp;2.06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eLipid profiles\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eTC (mmol/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e5.07\u0026thinsp;\u0026plusmn;\u0026thinsp;0.91\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e5.08\u0026thinsp;\u0026plusmn;\u0026thinsp;0.82\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e5.20\u0026thinsp;\u0026plusmn;\u0026thinsp;1.14\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003e5.28\u0026thinsp;\u0026plusmn;\u0026thinsp;0.81\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e0.016\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eTG\u0026nbsp;(mmol/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e1.28\u0026thinsp;\u0026plusmn;\u0026thinsp;0.58\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e0.86\u0026thinsp;\u0026plusmn;\u0026thinsp;0.44\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e2.44\u0026thinsp;\u0026plusmn;\u0026thinsp;1.71\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003e1.38\u0026thinsp;\u0026plusmn;\u0026thinsp;0.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eHDL-C (mmol/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e1.32\u0026thinsp;\u0026plusmn;\u0026thinsp;0.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e1.93\u0026thinsp;\u0026plusmn;\u0026thinsp;0.41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e1.09\u0026thinsp;\u0026plusmn;\u0026thinsp;0.26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003e1.41\u0026thinsp;\u0026plusmn;\u0026thinsp;0.34\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eLDL-C (mmol/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e3.15\u0026thinsp;\u0026plusmn;\u0026thinsp;0.80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e2.76\u0026thinsp;\u0026plusmn;\u0026thinsp;0.74\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e3.06\u0026thinsp;\u0026plusmn;\u0026thinsp;0.94\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003e3.21\u0026thinsp;\u0026plusmn;\u0026thinsp;0.79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003esiMS score\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e4.83\u0026thinsp;\u0026plusmn;\u0026thinsp;0.40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e4.75\u0026thinsp;\u0026plusmn;\u0026thinsp;0.44\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e5.54\u0026thinsp;\u0026plusmn;\u0026thinsp;0.96\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003e5.02\u0026thinsp;\u0026plusmn;\u0026thinsp;0.42\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eUrine\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eeGFR (mL/min/1.73m2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e103.77\u003c/p\u003e \u003cp\u003e(91.72, 111.63)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e101.18\u003c/p\u003e \u003cp\u003e(89.17, 108.67)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e103.01\u003c/p\u003e \u003cp\u003e(94.58, 115.00)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003e95.71\u003c/p\u003e \u003cp\u003e(84.30, 104.33)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eMAU/Urine Cr ratio (mg/g)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e3.9 (2.5, 6.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e3.4 (2.5, 5.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e5.6 (3.1, 10.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003e4.5 (2.6, 8.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e0.032\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eUnderlying diseases\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eHypertension, yes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(4.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e(3.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e(15.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e33\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e(12.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eDyslipidemia, yes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(11.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e31\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e(11.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e(19.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e61\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e(23.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eFamily history\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eHeart failure, yes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(6.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e(5.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e(4.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e(10.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e0.887\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eCancer, yes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(18.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e52\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e(18.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e(17.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e53\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e(20.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e0.007\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eStroke, yes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(6.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e(7.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e(4.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e36\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e(13.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e0.003\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eLifestyle\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eCurrent smoker, yes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(8.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e(2.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e(11.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e(6.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eAlcohol consumption, yes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e225\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(64.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e161\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e(57.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e92\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e(63.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e140\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e(53.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e0.003\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"11\" nameend=\"c11\" namest=\"c1\"\u003e \u003cp\u003eData are expressed as mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SD, number (percentage), or median (interquartile range).\u003c/p\u003e \u003cp\u003eBMI, body mass index; WC, waist circumference; SBP, systolic blood pressure; DBP, diastolic blood pressure; FBG, fasting blood glucose; HbA1c, hemoglobin A1c; TC, total cholesterol; TG, triglyceride; HDL-C, high density lipoprotein cholesterol; LDL-C, low density lipoprotein cholesterol; eGFR, estimated glomerular filtration rate; MAU/Urine Cr ratio, microalbumin/creatinine ratio; siMS score, simple method for quantifying metabolic syndrome; MORPD, mild obesity-related prediabetes; LORPD, low risk prediabetes; MESPD, metabolic syndrome prediabetes; MARPD, mild age-related prediabetes.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eCross-tabulation of prediabetes definitions\u0026mdash;classified based on HbA1c categories (5.7\u0026ndash;6.4% or 39\u0026ndash;46 mmol/L)\u0026mdash;revealed that all participants in the MORPD, LORPD, and MARPD clusters were identified according to these HbA1c criteria. The prediabetes in the MESPD cluster had abnormal FBG and HbA1c simultaneously. The LORPD cluster had the highest proportion of female participants (85.8%), while the MESPD cluster had the lowest proportion of females (55.2%). Age differed significantly across clusters, with the MARPD cluster having the highest average age (47.43\u0026thinsp;\u0026plusmn;\u0026thinsp;7.24 years). BMI was highest in the MESPD cluster (33.43\u0026thinsp;\u0026plusmn;\u0026thinsp;4.98 kg/m\u0026sup2;) and lowest in the LORPD cluster (21.73\u0026thinsp;\u0026plusmn;\u0026thinsp;2.53 kg/m\u0026sup2;), reflecting different obesity profiles across clusters. Similarly, WC was highest in the MESPD cluster (102.87\u0026thinsp;\u0026plusmn;\u0026thinsp;10.78 cm) and lowest in the LORPD cluster (73.88\u0026thinsp;\u0026plusmn;\u0026thinsp;7.05 cm).\u003c/p\u003e \u003cp\u003eBP varied significantly among clusters, with the MESPD cluster exhibiting the highest systolic (131.4\u0026thinsp;\u0026plusmn;\u0026thinsp;14.4 mmHg) and diastolic (81.7\u0026thinsp;\u0026plusmn;\u0026thinsp;10.5 mmHg) blood pressure, suggesting an elevated risk for hypertension. FBG and HbA1c were elevated in the MESPD and MARPD clusters compared to the other clusters, indicating a higher risk of diabetes progression. Lipid profile analysis revealed significant differences across clusters: the MESPD cluster had the highest triglyceride (2.44\u0026thinsp;\u0026plusmn;\u0026thinsp;1.71 mmol/L) and the lowest HDL\u0026ndash;C (1.09\u0026thinsp;\u0026plusmn;\u0026thinsp;0.26 mmol/L), indicating a more adverse lipid profile. Additionally, the MARPD cluster had the lowest estimated glomerular filtration rate (eGFR), median 95.71\u0026thinsp;\u0026plusmn;\u0026thinsp;mL/min/1.73m\u0026sup2;, which may suggest declining kidney function. The MESPD cluster also had the highest microalbuminuria/urine creatinine ratio, further indicating renal stress. Overall, these baseline characteristics underscore the distinct metabolic and demographic profiles of each prediabetes cluster.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eAssociations of prediabetes clusters with major diseases\u003c/h2\u003e \u003cp\u003eA total of 1,016 participants were followed over a median period of 4.5 years to monitor the incidence of T2D and CVD. The MESPD cluster exhibited the highest risk for T2D, with an incidence rate of 86.9 cases per 1,000 person\u0026ndash;years and 32.9% of participants diagnosed with T2D during follow\u0026ndash;up. Similarly, for CVD, the MESPD cluster had an incidence rate of 5.7 cases per 1,000 person\u0026ndash;years, compared to lower rates of 2.1 and 1.8 cases per 1,000 person\u0026ndash;years in the MORPD and LORPD clusters, respectively (Table S6).\u003c/p\u003e \u003cp\u003eWeibull proportional hazards models were used to compare the incidence of T2D and CVD across clusters, with the MORPD cluster as the reference group. The MESPD cluster had a significantly higher risk of developing T2D, with an adjusted hazard ratio (HR) of 5.86 (95% CI: 2.54\u0026ndash;13.51, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) compared to the LORPD cluster. The MARPD cluster also demonstrated an increased risk of T2D, with an adjusted HR of 4.36 (95% CI: 1.96\u0026ndash;9.73, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). In contrast, the MORPD cluster showed no significant difference in diabetes risk compared to the LORPD cluster (adjusted HR: 1.74, 95% CI: 0.77\u0026ndash;3.94, p\u0026thinsp;=\u0026thinsp;0.183). For CVD, there is no evidence of a significantly different risk among the four groups compared to the LORPD cluster (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, and Table S7).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eWeibull proportional hazards models comparing risk of type 2 Diabetes and CVD by cluster\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEvents\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eClusters\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eParticipants\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePerson-years\u003c/p\u003e \u003cp\u003eat risk\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eRate per 1000\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eUnadjusted HR\u003c/p\u003e \u003cp\u003e(95% CI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eAdjusted HR\u003c/p\u003e \u003cp\u003e(95% CI)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003eType 2 Diabetes\u0026dagger;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCluster 1 MORPD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e340\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1424\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e15.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.95 (0.89\u0026ndash;4.24)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.74 (0.77\u0026ndash;3.94)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCluster 2 LORPD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e272\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1137\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e7.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.00 (ref.)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.00 (ref.)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCluster 3 MESPD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e140\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e529\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e86.85\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e11.09*** (5.43\u0026ndash;22.67)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e5.86*** (2.54\u0026ndash;13.51)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCluster 4 MARPD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e264\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1089\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e41.33\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e5.21*** (2.54\u0026ndash;10.65)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e4.36*** (1.96\u0026ndash;9.73)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003eCVD\u0026Dagger;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCluster 1 MORPD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e340\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1423\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.59 (0.29\u0026ndash;8.73)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.42 (0.06\u0026ndash;3.10)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCluster 2 LORPD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e272\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1137\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.00 (ref.)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.00 (ref.)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCluster 3 MESPD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e140\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e525\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e3.22 (0.54\u0026ndash;19.27)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.77 (0.09\u0026ndash;6.75)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCluster 4 MARPD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e264\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1080\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e7.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.95 (0.89\u0026ndash;19.84)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.69 (0.11\u0026ndash;4.32)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"7\" nameend=\"c7\" namest=\"c1\"\u003e \u003cp\u003e*p\u0026thinsp;\u0026lt;\u0026thinsp;0.05, **p\u0026thinsp;\u0026lt;\u0026thinsp;0.01, ***p\u0026thinsp;\u0026lt;\u0026thinsp;0.001. HR, hazard ratio; 95% CI, 95% confidence interval.\u003c/p\u003e \u003cp\u003e\u0026dagger;Adjusted for baseline age at inclusion, sex, current smoker, alcohol consumption, hypertension, dyslipidemia, systolic blood pressure (SBP), diastolic blood pressure (DBP), low density lipoprotein cholesterol (LDL-C), total cholesterol (TC) and siMS score.\u003c/p\u003e \u003cp\u003e\u0026Dagger; Adjusted for baseline self-reported family history of heart disease, age at inclusion, sex, current smoker, alcohol consumption, hypertension, dyslipidemia, SBP, DBP, LDL-C, TC, and MAU/Urine Cr ratio.\u003c/p\u003e \u003cp\u003eMORPD, mild obesity-related prediabetes; LORPD, low risk prediabetes; MESPD, metabolic syndrome prediabetes; MARPD, mild age-related prediabetes.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe study investigated transition patterns across distinct prediabetes clusters, focusing on outcomes such as normal glucose regulation (NGR), persistence of prediabetes, and progression to T2D. In the MORPD cluster, 245 (75.9%) individuals reverted to NGR, while 22 (6.8%) progressed to T2D. The LORPD cluster showed the highest reversion rate to NGR, with 219 (88.3%) individuals, and the lowest progression rate to T2D, with 9 (3.6%), indicating a lower risk of diabetes progression. In contrast, the MESPD cluster demonstrated the lowest reversion rate to NGR (52, 40.6%), with a significant proportion progressing to T2D (46, 35.9%), indicating the highest risk of T2D onset. Lastly, in the MARPD cluster, 111 (45.7%) individuals reverted to NGR, while 45 (18.5%) developed T2D, reflecting a moderate risk of progression to T2D (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e and Table S8).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eDuring the follow-up period, changes in the biochemical parameters used to monitor prediabetes varied across the clusters. In particular, MARPD and MESPD showed a pronounced increase in both HbA1c and FBG, suggesting a higher likelihood of progressing to T2D compared with other clusters. Notably, MESPD had elevated TG and lower HDL-C, both of which are associated with an increased risk of CVD (Figure S6). Participants in the MARPD cluster had the highest cumulative incidence of T2D during the median follow-up period of 3.92 years, reflecting the cluster\u0026rsquo;s high-risk metabolic profile.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003ePGS among different clusters\u003c/h2\u003e \u003cp\u003eThe PRS analysis in individuals with prediabetes revealed notable patterns. The MESPD cluster exhibited the highest PRS, with a median value of 1.24, indicating a potentially elevated genetic risk, whereas the healthy control group displayed the lowest score, with a median value of 1.13 (Table S9 and Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea). When comparing clusters in terms of T2D follow\u0026ndash;up, the LORPD cluster had a significantly higher PGS score in the T2D subgroup compared to the other clusters, indicating a strong predictive ability of PRS for T2D for this cluster (Table S9 and Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb). Additionally, the majority of MESPD individuals belonged to the highest PRS quintile with 9% in quintile 1 to 33% in quintile 4 (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ec). Manhattan plots indicate no genome\u0026ndash;wide significance SNPs (p\u0026thinsp;\u0026lt;\u0026thinsp;5 \u0026times; 10^\u0026ndash;8) across all clusters (MORPD, LORPD, MESPD, and MARPD) (Figure S7).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eThis study presents a data\u0026ndash;driven approach to classify prediabetes into distinct sub\u0026ndash;phenotypes, offering a refined understanding of the heterogeneity with prediabetic states. Our findings identified four prediabetes subgroups\u0026mdash;MORPD, LORPD, MESPD, and MARPD\u0026mdash;each with unique clinical and metabolic characteristics and varied risks of progression to T2D and CVD. These clusters underscore the complexity of prediabetes and suggest potential benefits of tailored prevention strategies to prevent the development of T2D in the Thai population.\u003c/p\u003e \u003cp\u003eThe MESPD cluster, which was markedly characterized by metabolic syndrome\u0026ndash;like features, showed the highest incidence of T2D. This high\u0026ndash;risk profile is documented by elevated BMI, WC, and adverse atherogenic lipid profiles, consistent with previous studies linking metabolic syndrome components with increased cardiometabolic risk (17, 18). The substantially increased hazard ratios for T2D, even after adjusting for covariates, indicate severe metabolic disruption in MESPD, suggesting a need for targeted and intensive preventive measures. In contrast, the LORPD cluster, with lower BMI and WC, had the lowest risk of T2D and CVD, suggesting that metabolic burden plays a significant role in disease progression among prediabetic individuals (19, 20).\u003c/p\u003e \u003cp\u003eThe MARPD cluster, comprising older participants with relatively modest metabolic derangements, exhibited an elevated cumulative incidence of T2D over time. This finding suggests that age, combined with mild metabolic dysfunction, may significantly contribute to T2D risk. Prediabetes was common in community setting cohort study (21), particularly among aging population, and current evidence indicates that cardiovascular disease and mortality may be more critical prevention targets in this group than short\u0026ndash;term diabetes progression (21). A systematic review of primarily middle\u0026ndash;aged adults reported a 17% 6\u0026ndash;year cumulative incidence of diabetes for those with HbA1c\u0026ndash;defined prediabetes (Hba1c 5.7\u0026ndash;6.4%) and a 22% incidence for individuals with impaired fasting glucose (IFG 100\u0026ndash;125 mg/dL) (22). However, study in older populations, such as a Swedish cohort with 12 years of follow\u0026ndash;up, found that HbA1c\u0026ndash;defined prediabetes often remained stable or even regressed more frequently to normoglycemia than progressed to diabetes (23). Regression to normoglycemia appears more common in IFG\u0026ndash;defined prediabetes than HbA1c\u0026ndash;defined cases, likely due to greater variability in fasting glucose. These findings underscore the importance of addressing age\u0026ndash;related metabolic risks alongside diabetes prevention in older adults.\u003c/p\u003e \u003cp\u003eSimilar to previous studies (24, 25), we observed a substantial number of individuals reverting from prediabetes to NGR. Nearly 80% of those in the MORPD and LORPD clusters, and approximately 40% in the MESPD and MARPD clusters, reverted to NGR. Additionally, more than one five of each prediabetes cluster sustained their prediabetes status over 3.5\u0026ndash;year follow\u0026ndash;up. This confirmed the higher risk of T2D progression in the MESPD and MARPD individuals. The Whitehall II study similarly reported that a majority of people with HbA1c\u0026ndash;defined prediabetes persisted in that state over a 5\u0026ndash;year period (25). These distinct cluster profiles reveal underlying risk factors and highlight the need for tailored management strategies, which could enhance prevention efforts and potentially halt the progression from prediabetes to diabetes.\u003c/p\u003e \u003cp\u003ePrediabetes was associated with significantly worse liver function. It is proposed that T2D results from an excess of liver fat, which leads to an increased supply of fat to the pancreas, causing dysfunction in both organs and ultimately resulting in the progression of diabetes (26). MESPD cluster was associated with a higher risk of T2D. Although we did not measure liver function in our study, the higher BMI, WC, and hypertriglyceridemia may indicate greater levels of visceral fat accumulation among participants in the MESPD cluster. In contrast, the MESPD cluster did not demonstrate an increased risk of CVD. This finding may be explained by the relatively short follow-up period and the lower average age of participants (mean 39.5 years), which led to fewer observed events. Nonetheless, previous studies have indicated that metabolic dysfunction can be detrimental especially when accompanied by lipid abnormalities that promote CVD (27). Consequently, extended follow-up is warranted to fully elucidate the long-term risk of CVD.\u003c/p\u003e \u003cp\u003eInternational medical organizations have defined prediabetes for over 10 years. However, there is still controversy over whether it should be classified as a distinct pathological condition. Prediabetes is characterized by various physiological abnormalities (28, 29), and the metabolic environment can lead to a broad range of glycemic fluctuations, spanning from normoglycemia on one end to diabetes mellitus on the other, depending on the stage of the process. In the present study, we analyzed only those with prediabetes due to the priority of diabetes prevention in this group. Several key trial have demonstrated that treating prediabetes with effective interventions could significantly alter the progression of T2D. Nonetheless, efforts to implement diabetes prevention in clinical practice have encountered some difficulties. Interventions often require substantial resources, and prediabetic individuals may be unaware of their hyperglycemic condition, or the effects may be subtle (30).\u003c/p\u003e \u003cp\u003eOur classification of prediabetes may aid in identifying metabolic heterogeneity and guiding targeted interventions. Prediabetes in the high\u0026ndash;risk MESPD and MARPD clusters should be prioritized due to their elevated risk of T2D and related complications. However, while clinical variables provide useful insights, they may lack precision, as prediabetes can shift between clusters over time. Advancing precision medicine for T2D prevention will require integrating multidimensional data, such as clinic, multiomics, and sensor\u0026ndash;based behavioral information. Future studies should explore targeted interventions, like aerobic exercise and caloric restriction, to determine the most effective health benefits for each subgroup.(31). Future studies may investigate the types of interventions, such as aerobic exercise and dietary caloric restriction, that provide the greatest health benefits for individuals with prediabetes across various clusters.\u003c/p\u003e \u003cp\u003eDifferences in polygenetic risks support classification in this study. For instance, the higher T2D risk in the MESPD cluster correlated with an elevated PRS quintile, indicating a significant genetic contribution. The LORPD cluster also had elevated PRS scores among individuals who developed T2D, underscoring the predictive value of genetic risk assessment, even in low clinical risk groups. Higher PRS quintile have been linked to increased T2D risk across diverse populations, including East Asians (32). These findings suggest that genetic information support stratification of prediabetes patients into distinct clinical subgroups for targeted prevention, especially if future studies confirm differential responses to interventions or specific risks of progression. Although genetic factors significantly influence T2D risk (33), lifestyle interventions remain crucial, reinforcing the need for a holistic approach to diabetes prevention and management. Notably, the lack of genome\u0026ndash;wide significant SNPs in the Manhattan plots suggests that although genetic predisposition contributes to T2D risk, other factors\u0026mdash;such as environmental influences, lifestyle choices, and epigenetic changes\u0026mdash;also play a significant role in the progression from prediabetes to T2D.\u003c/p\u003e \u003cp\u003eInterventions in adults with prediabetes have shown effectiveness in reducing the risk of progression to diabetes, with lifestyle improvements being particularly impactful (34). The Diabetes Prevention Program (DPP) trial demonstrated that an intensive lifestyle intervention\u0026mdash;focused on weight loss and increased physical activity\u0026mdash;and, to a lesser extent, metformin use significantly reduced diabetes risk in adults aged 25 years and older (mean age 51) who were at high risk of diabetes progression (35). Consequently, current ADA guidelines recommend lifestyle interventions for adults with prediabetes (defined by HbA1c levels of 5.7\u0026ndash;6.4%, fasting glucose levels of 100\u0026ndash;125 mg/dL, or 2\u0026ndash;hour glucose levels of 140\u0026ndash;199 mg/dL) to achieve at least a 7% reduction in initial body weight and at least 150 minutes per week of moderate\u0026ndash;intensity physical activity. Metformin is recommended for patients under 60 years with a BMI of 35 or greater or for women with a history of gestational diabetes (36, 37).\u003c/p\u003e \u003cp\u003eThese findings support the potential for lifestyle interventions to promote not only diabetes prevention but also prediabetes and diabetes remission (38). Evidence suggests that intensive lifestyle modifications can lead to significant improvements in glycemic control, allowing some individuals with prediabetes to revert to normoglycemia and achieve prediabetes remission, while others with early\u0026ndash;stage diabetes may achieve diabetes remission (39). Given the lower risk of diabetes progression observed in older adults in this study, relative to mortality risk, aggressive pharmacologic interventions may offer limited benefit and could lead to unintended consequences, such as overdiagnosis, anxiety, or insurance implications. This underscores the importance of prioritizing safe, feasible lifestyle interventions that confer broad health benefits beyond diabetes prevention, especially in older adults (40).\u003c/p\u003e \u003cp\u003eIn this study, we employed a robust, data-driven clustering approach that combined k\u0026ndash;means with UMAP, validated by the Elbow method, silhouette width, and Jaccard similarity, to reliably classify prediabetic individuals into distinct sub-phenotypes. Our findings are further supported by an extensive follow-up period and advanced statistical models like Weibull proportional hazards regression. Furthermore, incorporating polygenic risk scores (PRS) provided additional insights into the genetic contributions to type 2 diabetes risk. Despite the strengths of this study, several limitations should be acknowledged. The current definition of prediabetes neither reflects the sub-phenotypes of T2D pathophysiology nor accurately predicts future metabolic trajectories. Additionally, the relatively short follow-up period may limit the ability to capture long-term trends in disease progression, highlighting the need for further longitudinal studies to validate our findings.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIn conclusion, our findings underscore the importance of stratifying individuals with prediabetes into distinct clusters based on specific metabolic and genetic profiles. This stratification enables more precise preventive strategies, particularly for high\u0026ndash;risk groups such as the MESPD and MARPD clusters. Integrating PRS further strengthens early detection and personalizes intervention for those most vulnerable to T2D. Future studies should validate these findings across diverse populations and investigate the complex interactions among genetic, metabolic, and lifestyle factors that contribute to disease risk. Integrating genetic and metabolic profiling paves the way for precision\u0026ndash;based management of prediabetes, establishing a new standard for proactive and tailored healthcare interventions.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments:\u003c/strong\u003e This research project is a collaboration between the Siriraj Medical Research Center and the Department of Preventive Medicine which is responsible for the staff health screening program. The authors gratefully acknowledge Ruengpung Sutthent and Sith Sathornsumetee, for her guidance for initiation of this project. We thank the members of the SIH Study Group, SPHERE staff members, and research assistants. The voluntary participation of all participants is highly appreciated. We are also grateful to the management of Siriraj Medical Research Center for providing office space, a recruitment center, and a biobank.\u003c/p\u003e\n\u003cp\u003eMembers of SIH study group: Winai Ratanasuwan, Keerati Charoencholvanich, Bhoom Suktitipat, Manop Pithukpakorn, Prapat Suriyaphol, Rungroj Krittayaphong, Prasert Auewarakul, Chalermchai Mitrpant, Boonrat Tassaneetritap, Mayuree Homsanit, and Naravat Poungvarin\u003c/p\u003e\n\u003cp\u003eMembers of SPHERE group: Sureeporn Pumeiam, Bonggochpass Pinsawas, Pichanun Mongkolsucharitkul, Apinya Surawit, Tanyaporn Pongkunakorn, Sophida Suta, Thamonwan Manosan, Suphawan Ophakas, and Korapat Mayurasakorn\u003c/p\u003e\n\u003cp\u003eMembers of Biobank: Somruedee Chatsiricharoenkul, Parichart Permpikul, Duangthip Apiratmontree, Sutee Udomchotphruet, and Pattranit Onsing\u003c/p\u003e\n\u003cp\u003eWe thank Sissades Tongsima, Tassathorn poonsin, and Pongsakorn Wangkumhang from National Biobank of Thailand for providing genome reference panel and advising the imputation steps.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u003c/strong\u003e This work was supported by the Faculty of Medicine Siriraj Hospital, Mahidol University (R016034006). Additional funding support including infrastructure, staff and utilities, was provided by Faculty of Medicine Siriraj Hospital, Mahidol University. The funder had no role in the study design, data collection and analysis, decision to publish, and preparation of the manuscript\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests:\u0026nbsp;\u003c/strong\u003eNone declared.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication:\u003c/strong\u003e Not applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability:\u0026nbsp;\u003c/strong\u003eIndividual participant data will be shared with researchers in a deidentified or anonymized format upon submitting a research proposal and requesting data access to Associate Professor Korapat Mayurasakorn (contact person: [email protected]). Data will be made available for analyses as approved by the data access committee.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePatient consent for publication:\u0026nbsp;\u003c/strong\u003eNot required\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributors:\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA.S. and K.M. conceived of the analysis, developed the analysis plan, and wrote the initial drafts of the manuscript. A.S. and P.P. conducted the statistical analysis. A.S., S.P., S.S., P.M., B.P., S.O., and K.M. contributed with the data collection. K.M., A.S., P.P., S.P., S.S., P.M., B.P., N.V., P.V. and S.O. reviewed and revised the manuscript for important intellectual content. A.S. and K.M. oversees study implementation, assisted in writing and editing the paper and prepared the manuscript for publication; All authors have read the manuscript and are in agreement with the decision to submit the manuscript for journal publication.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAssociation AD. Diagnosis and classification of diabetes mellitus. Diabetes Care. 2010;33:S62\u0026ndash;S9.\u003c/li\u003e\n\u003cli\u003eYip WCY, Sequeira IR, Plank LD, Poppitt SD. Prevalence of pre\u0026ndash;diabetes across ethnicities: a review of impaired fasting glucose (IFG) and impaired glucose tolerance (IGT) for classification of dysglycaemia. Nutrients. 2017;9(11).\u003c/li\u003e\n\u003cli\u003eTab\u0026aacute;k AG, Herder C, Rathmann W, Brunner EJ, Kivim\u0026auml;ki M. Prediabetes: a high\u0026ndash;risk state for diabetes development. The Lancet. 2012;379(9833):2279\u0026ndash;90.\u003c/li\u003e\n\u003cli\u003eGong Q, Zhang P, Wang J, Ma J, An Y, Chen Y, et al. Morbidity and mortality after lifestyle intervention for people with impaired glucose tolerance: 30\u0026ndash;year results of the da qing diabetes prevention outcome study. Lancet Diabetes Endocrinol. 2019;7(6):452\u0026ndash;61.\u003c/li\u003e\n\u003cli\u003eLindström J, Louheranta A, Mannelin M, Rastas M, Salminen V, Eriksson J, et al. The Finnish Diabetes Prevention Study (DPS): lifestyle intervention and 3\u0026ndash;year results on diet and physical activity. Diabetes Care. 2003;26(12):3230\u0026ndash;6.\u003c/li\u003e\n\u003cli\u003ePiller C. Dubious diagnosis. Science. 2019;363(6431):1026\u0026ndash;31.\u003c/li\u003e\n\u003cli\u003eWagner R, Heni M, Tabak AG, Machann J, Schick F, Randrianarisoa E, et al. Pathophysiology\u0026ndash;based subphenotyping of individuals at elevated risk for type 2 diabetes. Nat Med. 2021;27(1):49\u0026ndash;57.\u003c/li\u003e\n\u003cli\u003eDo HD, Lohsoonthorn V, Jiamjarasrangsi W, Lertmaharit S, Williams MA. Prevalence of insulin resistance and its relationship with cardiovascular disease risk factors among Thai adults over 35 years old. Diabetes Res Clin Pract. 2010;89(3):303\u0026ndash;8.\u003c/li\u003e\n\u003cli\u003eMongkolsucharitkul P, Surawit A, Manosan T, Ophakas S, Suta S, Pinsawas B, et al. Metabolic and genetic risk factors associated with pre\u0026ndash;diabetes and type 2 diabetes in Thai healthcare employees: A long\u0026ndash;term study from the Siriraj Health (SIH) cohort study. PLoS One. 2024;19(6):e0303085.\u003c/li\u003e\n\u003cli\u003eAssociation AD. American diabetes association standards of care in diabetesd\u0026ndash;2024. Diabetes Care. 2024;47:S1\u0026ndash;S321.\u003c/li\u003e\n\u003cli\u003eChang CC, Chow CC, Tellier LC, Vattikuti S, Purcell SM, Lee JJ. Second\u0026ndash;generation PLINK: rising to the challenge of larger and richer datasets. Gigascience. 2015;4:7.\u003c/li\u003e\n\u003cli\u003eWall JD, Stawiski EW, Ratan A, Kim HL, Kim C, Gupta R, et al. The Genome Asia 100K project enables genetic discoveries across Asia. Nature. 2019;576(7785):106\u0026ndash;11.\u003c/li\u003e\n\u003cli\u003eManichaikul A, Mychaleckyj JC, Rich SS, Daly K, Sale M, Chen W\u0026ndash;M. Robust relationship inference in genome\u0026ndash;wide association studies. Bioinformatics. 2010;26(22):2867\u0026ndash;73.\u003c/li\u003e\n\u003cli\u003eTurner S. qqman: an R package for visualizing GWAS results using Q\u0026ndash;Q and manhattan plots2014.\u003c/li\u003e\n\u003cli\u003eMonti R, Eick L, Hudjashov G, L\u0026auml;ll K, Kanoni S, Wolford BN, et al. Evaluation of polygenic scoring methods in five biobanks shows larger variation between biobanks than methods and finds benefits of ensemble learning. Am J Med Genet A. 2024;111(7):1431\u0026ndash;47.\u003c/li\u003e\n\u003cli\u003eBedrick EJ, Tsai C\u0026ndash;L. Model selection for multivariate regression in small samples. Biometrics. 1994;50(1):226\u0026ndash;31.\u003c/li\u003e\n\u003cli\u003eBiswas T, Townsend N, Gupta RD, Ghosh A, Rawal LB, M\u0026oslash;rkrid K, et al. Clustering of metabolic and behavioural risk factors for cardiovascular diseases among the adult population in South and Southeast Asia: findings from WHO STEPS data. Lancet Reg Health Southeast Asia. 2023;12.\u003c/li\u003e\n\u003cli\u003eLaucyte\u0026ndash;Cibulskiene A, Chen C\u0026ndash;H, Cockroft J, Cunha PG, Kavousi M, Laucevicius A, et al. Clusters of risk factors in metabolic syndrome and their influence on central blood pressure in a global study. Sci Rep. 2022;12(1):14409.\u003c/li\u003e\n\u003cli\u003eAndr\u0026eacute;asson K, Edqvist J, Adiels M, Bj\u0026ouml;rck L, Lindgren M, Sattar N, et al. Body mass index in adolescence, risk of type 2 diabetes and associated complications: a nationwide cohort study of men. eClinicalMedicine. 2022;46.\u003c/li\u003e\n\u003cli\u003eBae JC, Cho NH, Kim JH, Hur KY, Jin SM, Lee MK. Association of body mass index with the risk of incident type 2 diabetes, cardiovascular disease, and all\u0026ndash;cause mortality: a community\u0026ndash;based prospective study. Endocrinol Metab (Seoul). 2020;35(2):416\u0026ndash;24.\u003c/li\u003e\n\u003cli\u003eRooney MR, Rawlings AM, Pankow JS, Echouffo Tcheugui JB, Coresh J, Sharrett AR, et al. Risk of progression to diabetes among older adults with prediabetes. JAMA Intern Med. 2021;181(4):511\u0026ndash;9.\u003c/li\u003e\n\u003cli\u003eRichter B, Hemmingsen B, Metzendorf MI, Takwoingi Y. Development of type 2 diabetes mellitus in people with intermediate hyperglycaemia. Cochrane Database Syst Rev. 2018;10(10):Cd012661.\u003c/li\u003e\n\u003cli\u003eShang Y, Marseglia A, Fratiglioni L, Welmer AK, Wang R, Wang HX, et al. Natural history of prediabetes in older adults from a population\u0026ndash;based longitudinal study. J Intern Med. 2019;286(3):326\u0026ndash;40.\u003c/li\u003e\n\u003cli\u003eSallar A, Dagogo\u0026ndash;Jack S. Regression from prediabetes to normal glucose regulation: state of the science. Exp Biol Med. 2020;245(10):889\u0026ndash;96.\u003c/li\u003e\n\u003cli\u003eVistisen D, Kivim\u0026auml;ki M, Perreault L, Hulman A, Witte DR, Brunner EJ, et al. Reversion from prediabetes to normoglycaemia and risk of cardiovascular disease and mortality: the Whitehall II cohort study. Diabetologia. 2019;62(8):1385\u0026ndash;90.\u003c/li\u003e\n\u003cli\u003eTaylor R. Type 2 diabetes and remission: practical management guided by pathophysiology. J Intern Med. 2021;289(6):754\u0026ndash;70.\u003c/li\u003e\n\u003cli\u003eHill MA, Yang Y, Zhang L, Sun Z, Jia G, Parrish AR, et al. Insulin resistance, cardiovascular stiffening and cardiovascular disease. Metabolism. 2021;119:154766.\u003c/li\u003e\n\u003cli\u003eBlond MB, F\u0026aelig;rch K, Herder C, Ziegler D, Stehouwer CDA. The prediabetes conundrum: striking the balance between risk and resources. Diabetologia. 2023;66(6):1016\u0026ndash;23.\u003c/li\u003e\n\u003cli\u003eBarbu E, Popescu MR, Popescu AC, Balanescu SM. Phenotyping the prediabetic population\u0026ndash;a closer look at intermediate glucose status and cardiovascular disease. Int J Mol Sci. 2021;22(13).\u003c/li\u003e\n\u003cli\u003eEchouffo\u0026ndash;Tcheugui JB, Selvin E. Prediabetes and what it means: The epidemiological evidence. Annu Rev Public Health. 2021;42(Volume 42, 2021):59\u0026ndash;77.\u003c/li\u003e\n\u003cli\u003eDel Prato S. Heterogeneity of diabetes: heralding the era of precision medicine. Lancet Diabetes Endocrinol. 2019;7(9):659\u0026ndash;61.\u003c/li\u003e\n\u003cli\u003eLee H, Choi J, Kim JI, Watanabe RM, Cho NH, Park KS, et al. Higher genetic risk for type 2 diabetes is associated with a faster decline of \u0026beta;\u0026ndash;cell function in an east asian population. Diabetes Care. 2024;47(8):1386\u0026ndash;94.\u003c/li\u003e\n\u003cli\u003eUdler MS, Kim J, von Grotthuss M, Bon\u0026agrave;s\u0026ndash;Guarch S, Cole JB, Chiou J, et al. Type 2 diabetes genetic loci informed by multi\u0026ndash;trait associations point to disease mechanisms and subtypes: A soft clustering analysis. PLOS Medicine. 2018;15(9):e1002654.\u003c/li\u003e\n\u003cli\u003ePe\u0026ntilde;a A, Olson ML, Hooker E, Ayers SL, Castro FG, Patrick DL, et al. Effects of a diabetes prevention program on type 2 diabetes risk factors and quality of life among latino youths with prediabetes: a randomized clinical trial. JAMA Network Open. 2022;5(9):e2231196\u0026ndash;e.\u003c/li\u003e\n\u003cli\u003eKnowler WC, Fowler SE, Hamman RF, Christophi CA, Hoffman HJ, Brenneman AT, et al. 10\u0026ndash;year follow\u0026ndash;up of diabetes incidence and weight loss in the diabetes prevention program outcomes study. Lancet. 2009;374(9702):1677\u0026ndash;86.\u003c/li\u003e\n\u003cli\u003eBoltri JM, Tracer H, Strogatz D, Idzik S, Schumacher P, Fukagawa N, et al. The national clinical care commission report to congress: leveraging federal policies and programs to prevent diabetes in people with prediabetes. Diabetes Care. 2023;46(2):e39\u0026ndash;e50.\u003c/li\u003e\n\u003cli\u003eAssociation AD. Prevention or delay of type 2 diabetes: standards of medical care in diabetes\u0026mdash;2020. Diabetes Care. 2019;43(Supplement_1):S32\u0026ndash;S6.\u003c/li\u003e\n\u003cli\u003eLean MEJ, Leslie WS, Barnes AC, Brosnahan N, Thom G, McCombie L, et al. 5\u0026ndash;year follow\u0026ndash;up of the randomised Diabetes Remission Clinical Trial (DiRECT) of continued support for weight loss maintenance in the UK: an extension study. Lancet Diabetes Endocrinol. 2024;12(4):233\u0026ndash;46.\u003c/li\u003e\n\u003cli\u003eRosenfeld RM, Kelly JH, Agarwal M, Aspry K, Barnett T, Davis BC, et al. Dietary interventions to treat type 2 diabetes in adults with a goal of remission: An expert consensus statement from the american college of lifestyle medicine. Am J Lifestyle Med. 2022;16(3):342\u0026ndash;62.\u003c/li\u003e\n\u003cli\u003eBradley MD, Arnold ME, Biskup BG, Campbell TM, 2nd, Fuhrman J, Guthrie GE, et al. Medication deprescribing among patients with type 2 diabetes: a qualitative case series of lifestyle medicine practitioner protocols. Clin Diabetes. 2023;41(2):163\u0026ndash;76.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":false,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Prediabetes, Type 2 diabetes, Cardiovascular disease, Cluster analysis","lastPublishedDoi":"10.21203/rs.3.rs-6351460/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6351460/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground:\u003c/strong\u003e This study examined prediabetic clusters and their associations with type 2 diabetes (T2D) and cardiovascular disease (CVD) using variables from metabolic syndrome, glycemic measures, and blood lipids.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods:\u003c/strong\u003e A total of 1,016 prediabetic individuals were classified into four clusters using k-means clustering. Weibull proportional hazards models estimated T2D and CVD risk, and T2D polygenic risk scores (PRS) were analyzed to refine risk within each cluster.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults:\u003c/strong\u003e Four clusters were identified: the metabolic syndrome prediabetes (MESPD) cluster, characterized by elevated BMI and adverse lipid profiles, had the highest T2D risk (HR 5.86). The mild age-related prediabetes (MARPD) cluster, associated with older age, showed an increased T2D risk. In contrast, the low–risk prediabetes (LORPD) cluster exhibited the lowest risk, suggesting that a reduced metabolic burden may confer greater disease stability. PRS were used to refine risk stratification, with the MESPD cluster showing a significant genetic predisposition to T2D. PRS also enhanced predictive accuracy for the LORPD cluster, providing additional insights into genetic factors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions: \u003c/strong\u003eThe findings highlight the importance of precision medicine by identifying prediabetic subgroups with varying risks for T2D. Incorporating genetic data, the study improves models and offers insights for future research and interventions to prevent prediabetes progression.\u003c/p\u003e","manuscriptTitle":"Novel subgroups of prediabetes and the associations with outcomes in health professionals: a data–driven cluster analysis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-05-09 10:46:44","doi":"10.21203/rs.3.rs-6351460/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"1b609154-fe52-4fa3-b098-96644a81d6a7","owner":[],"postedDate":"May 9th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":48181945,"name":"Health sciences/Endocrinology/Endocrine system and metabolic diseases"},{"id":48181946,"name":"Health sciences/Diseases/Endocrine system and metabolic diseases"},{"id":48181947,"name":"Health sciences/Diseases/Metabolic disorders"},{"id":48181948,"name":"Health sciences/Diseases"},{"id":48181949,"name":"Health sciences/Endocrinology"}],"tags":[],"updatedAt":"2025-05-30T11:23:51+00:00","versionOfRecord":[],"versionCreatedAt":"2025-05-09 10:46:44","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6351460","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6351460","identity":"rs-6351460","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}