Microalbuminuria as a simple predictor of incident diabetes over 8 years in the Korean Genome and Epidemiology Study (KoGES)

Microalbuminuria (MAU) is a common subclinical disease and related with cardiovascular outcome both in diabetic and non-diabetic patients. However, there is rare data about the effect of MAU on the development of diabetes. Thus, we aimed to investigate whether MAU is associated with the development of incident diabetes. A total of 3385 subjects without diabetes (1503 men and 1882 women; mean age, 53 years) who participated in the Ansung–Ansan cohort study from 2001–2002 (baseline) to 2011–2012 (fifth follow-up visit) were followed for a mean of 8 years. The prevalence of MAU at baseline was 10.8% (365 patients), and the incidence of newly developed diabetes during the follow-up period was 15.3% (56 patients) in subjects with MAU. The hazard ratio (HR) for development of diabetes was 1.43 (95% confidence interval (CI) 1.07–1.91, p-value 0.016), independent of traditional risk factors for diabetes including pre-diabetes, age, obesity, and family history. The impact of MAU on diabetes was also significant in the non-pre-diabetic population (HR 2.08, 95% CI 1.07–4.03, p-value 0.031). In conclusion, our results show that incident MAU is associated with future development of diabetes and could be an early marker for diabetes, even in the non-prediabetic population.

and performed according to the principles of the Declaration of Helsinki. All data generated during this study are available.
Clinical and biochemical parameters. Study data included medical history, physical examination, information provided by a questionnaire, anthropometric measurements, and laboratory measurement. Information on medical history, family history, current use of medications, weekly alcohol consumption, and smoking status were obtained from all participants using a standard questionnaire.
Waist circumference was measured three times at the midpoint between the bottom of the ribcage and the top of the iliac crest using a fiberglass tape measure. Blood pressure was measured in the sitting position after 5 min of rest using a standard mercury sphygmomanometer.
Body mass index (BMI) was calculated as weight divided by height squared (kg/m 2 ). Physical activity was classified into the following three categories: none, irregular (≤2 episodes/week), and regular (≥3 episodes/week) exercise. One episode of exercise was defined as exercising for at least 30 min. The International Obesity Task Force (IOTF) and World Health Organization (WHO) regional office for the Western Pacific region recommend defining obesity in Asians as a BMI ≥ 25 kg/m 2 . Subsequently, the Korean Society for the Study of Obesity (KSSO) adopted this definition. Thus, subjects were classified as "generally obese" if their BMI was ≥25 kg/m 2 . Abdominal obesity was defined as a waist circumference ≥90 cm for male subjects and ≥85 cm for female subjects according to the criteria of the modified National Cholesterol Education Program's Adult Treatment Panel III (NCEP ATP III) Asian criteria for metabolic syndrome. Collected blood samples were delivered to and analysed at a central laboratory (Seoul Clinical Laboratories, Seoul, Korea). Plasma glucose, total cholesterol, triglycerides, and HDL cholesterol levels were determined using a Hitachi 747 chemistry analyser (Hitachi, Tokyo, Japan). The LDL cholesterol level was calculated using Friedewald's equation. The HbA1c level was measured by high-performance liquid chromatography on a Variant II instrument (BioRad Laboratories, Hercules, CA).
Definitions of MAU, diabetes, and prediabetes. MAU was defined as an UACR of 30-300 mg/g. We excluded overt albuminuria, defined as UACR > 300 mg/g. diabetes was defined as one of the following: (1) self-reported history of diabetes; (2) HbA1c level ≥6.5%; or (3) use of an antidiabetic agent or insulin. Prediabetes was defined as a fasting plasma glucose level of 110 to 125 mg/dl or an HbA1c concentration of 5.7-6.4%.
Statistical analyses. Data are expressed as mean and SD or as the number and percentage. Comparisons of baseline variables with respect to the presence or absence of MAU were analysed using Student's t test for continuous variables; categorical variables were analysed using the x 2 test. We calculated the hazard ratios (HRs) for incident diabetes using Cox proportional hazards models with potential confounding variables. We included traditional risk factors for diabetes such as age, obesity, family history of diabetes, smoking status, physical activity and prediabetes in univariate analyses. Variables with p < 0.15 in the univariate analysis were entered into multivariate analysis. We developed two different models to estimate the risk of new onset diabetes. The basic model was based only on the parameters that were easy to assess without biochemical testing. The risk factors for the incidence of diabetes in the basic model were: age, family history of diabetes and obesity. Clinical model was created by adding MAU to the basic model. For each model, the area under a receiver operating characteristic curve (AROC) was calculated. Comparison of AUC curves was calculated using Delong method. This value represents an estimate of the probability that a model assigns a higher risk to those who develop diabetes within an 8-year follow up than to those who do not. All analyses were performed using SPSS Statistics for Windows version 18.0 (IBM, Armonk, NY). For all tests, a P value < 0.05 was considered to indicate a statistically significant difference.

Results
The anthropometric and biochemical characteristics of the subjects are summarised according to the presence of MAU in Table 1. MAU was observed in 365 patients (10.8%) at baseline, and the mean follow up duration was approximately 8 years. The development of diabetes was increased in subjects with MAU compared with subjects without MAU after 2 years of follow up; this gap widened over time (Fig. 1). The mean age, prevalence of hypertension, and level of HbA1C increased as baseline MAU category increased. Although there was no significant difference in the presence of metabolic syndrome, the proportions of subjects with high triglycerides and impaired fasting glucose were higher in the MAU group. There were no differences in activity or family history of diabetes. Of the 365 subjects with MAU, diabetes developed in 56 (15.3%) during the 10-year follow-up period. This proportion was statistically significantly different than that in subjects without MAU. The incidences of composite cardiovascular disease and each component are shown in Table 2; no significant difference was observed according to the presence of MAU. Using a Cox proportional hazards model, we also investigated the clinical impact of MAU on the development of diabetes during the follow-up period. Participants with MAU had a higher risk for development of diabetes (hazard ratio [HR] 1.78, 95% confidence interval [CI] 1.34-2.37, p value < 0.001), as shown in Table 3. Moreover, this significant association remained after adjusting for other risk factors for diabetes including age, prediabetes, general and abdominal obesity, and family history. We next performed subgroup analysis. In the non-prediabetic group, the association was even greater (HR 2.08, 95% CI 1.07-4.03, p value 0.031), even after adjusting for confounding factors. The association between MAU and diabetes was more prominent in younger patients and in nonobese patients (Fig. 2). The AROC for the basic model and the clinical model was evaluated (Fig. 3)

Discussion
In this large prospective, community-based cohort study of Korean adults, we found that MAU was associated with a 1.43-fold higher risk of development of diabetes after adjusting for several well-known risk factors for diabetes including old age, hypertension, IFG, metabolic component, and family history. The impact of MAU on diabetes was more prominent in subjects without prediabetes.
Several studies have investigated the association of MAU and insulin insensitivity or prediabetes 6-8, 11 Dutta et al. reported that MAU is associated with decreased reversal to normoglycemia and increased progression to diabetes with 147 pre-diabetic individuals with 3 months follow up 12 Table 3. Multivariate analysis of the overall prevalence of diabetes and the prevalence of diabetes in nonprediabetic subjects. BMI, body mass index. associated with insulin resistance measuring fasting insulin concentration in nondiabetic subjects 13 . However, most of these were cross-sectional studies or small population with short-term follow up data; few data are available regarding whether MAU is directly associated with overt diabetes in large nondiabetic or population-based cohorts.
There is a growing interest in predictors whose abnormal levels indicate an elevated risk of development of diabetes prior to development of clinical symptoms, since regular heath checkups are becoming increasingly common and often include the acquisition of data regarding left ventricular hypertrophy, MAU, high coronary artery calcium score, and increased carotid intima-media thickness. One previous study showed that these subclinical markers are associated with an elevated risk of CVD, especially in people with metabolic syndrome and diabetes 11 . Recently, MAU was recognized as an emerging marker for cardiovascular disease in individuals with hypertension, obese individuals, and the general population, as well as acting as an emerging marker for diabetes [14][15][16] . The relationship between MAU and insulin insensitivity is based on several pieces of epidemiological evidence, addressing the association of kidney dysfunction with both genetic and non-genetic markers of insulin resistance 13,[17][18][19] . Although many reports have evaluated the relationship between MAU and prediabetes, no clear consensus has yet been reached. In American group, MAU did not predict incident diabetes in pre-diabetic obese subjects 20 . Recently, two cross-sectional studies were performed in Korea. Won et al. 21 reported that MAU was associated with prediabetes, while Kim et al. concluded that the association is probably mediated by hypertension, and MAU is not independently associated with prediabetes 8 . However, both studies were cross-sectional in nature; thus, they could not show a causal relationship between MAU and prediabetes. In addition, Won et al. did not include HbA1c level in the definition of diabetes, which might have led to an underestimation of the real prevalence of diabetes.
A unique aspect of our study is that we showed a direct association between MAU and incidence of diabetes using a large, nationwide cohort with a mean follow up duration of 8 years. After adjusting for traditional risk factors for diabetes including pre-diabetes and components of metabolic syndrome, the association was still significant. Since the impact of MAU was also statistically significant in subjects without prediabetes, we conclude that MAU is potentially an early marker for diabetes, similar to pre-diabetes. Although the precise mechanism of progression to the diabetes in subjects with MAU is not understood, several epidemiologic and experimental studies have reported potential mechanisms by which insulin resistance is linked to kidney dysfunction in both diabetic and non-diabetic subjects. For instance, there were epidemiologic studies showed that MAU may precede the onset of diabetes 12,22 in non-diabetic patients. In addition, insulin insensitivity observed in non-diabetic subjects with MAU 13 causes podocyte apoptosis and detachment from the glomerular basement membrane 23 and insulin insensitivity also plays an important role in the development of albuminuria due to endothelial dysfunction and increased vascular permeability, even in non-diabetic patients 24 .
Although we included a large number of participants extracted by stratified random sampling with proportional allocation among the 1476 categories and tried to adjust for confounding factors, this study had some limitations. First, it was difficult to replicate our findings due to lack of availability of another cohort. Thus, lack of replication of our result is one of the major limitations. Second, urinary albumin was assessed in only a single urine specimen per participant. Since urinary albumin level can exhibit considerable intra-individual variability, the lack of repeated sampling could have missed this variation and is thus a limitation. However, there were several studies to evaluate the diagnostic performance of spot urine for the detection of MAU, since 24-hour urine collection is time and cost consuming. They showed that spot urine was a reliable measure of proteinuria in various subjects with diabetes, prediabetes, chronic kidney disease and rheumatic disease as well as in general population [25][26][27][28][29][30][31] . Third, MAU could transiently develop in the presence of inflammation or stress. However, we could not fully exclude such conditions, since these data were derived from a nationwide heath cohort. Therefore, another limitation is that we might have overestimated the prevalence of MAU. Fourth, this study was restricted to only Koreans. Therefore, these results might not be applicable to other ethnic groups.
In conclusion, incident MAU could be an early marker for diabetes, even in the non-prediabetic population. Thus, physicians should closely monitor all subjects and regularly measure HbA1C level. Patients should also be educated about lifestyle modifications including weight control, alcohol cessation, and regular activity.