Incidence and risk factors of vascular complications in people with impaired fasting glucose: a national cohort study in Korea

This study aimed to evaluate the risk of vascular complications of impaired fasting glucose (IFG). This population-based study included 425,608 participants from the National Health Screening Cohort in Korea in 2003 and 2004 who were followed-up until 2015. The participants were classified into normal, IFG, and diabetes groups based on fasting plasma glucose levels. Incidence rate (per 1000 person-year) was evaluated for the following vascular complications: cardiovascular (ischemic heart disease, cerebrovascular disease, arterial and capillary disease), renal, and retinal diseases. Hazard ratios (HR) of IFG for diabetes were estimated after adjusting for patient characteristics. Among the 88,330 IFG participants, the incidence of cardiovascular, chronic renal and retinal diseases were 11.52, 0.47, and 1.08 per 1000 person-years, respectively. Furthermore, IFG patients with a family history of diabetes, past history of hypertension, and high body mass index had significantly increased risk of vascular complications [adjusted HR, cardiovascular: 1.39 (95% CI 1.33–1.46); renal: 2.17 (95% CI 1.66–2.83); and retinal: 1.14 (95% CI 0.98–1.32)]. IFG patients have a substantial risk of cardiovascular, chronic renal and retinal diseases. Therefore, early preventative interventions are beneficial, especially for those with high-risk factors, in whom should emphasize on maintaining a healthy lifestyle, early screening and continuous follow-up.

Considering the long follow-up period (12 years maximum), the latest data of each participant was evaluated and used for sensitivity analyses. The most recent records of the various vascular complications represented our study outcomes. Additionally, based on the Standards of Medical Care in Diabetes-2019, the IFG group was further classified into the high-and low-risk subgroups. Participants with more than one of the following risk factors were considered high-risk: family history of DM, past history of hypertension, or body mass index (BMI) > 23 kg/m 2 (Asian).
Past history of DM was identified either based on the 10th edition of the International Classification of Diseases (ICD-10 codes E10-E14) between 2002 and 2003, or through self-administered questionnaire. The study flowchart is presented in Fig. 1.

Definitions of vascular complications.
Two or more hospitalization days and death records based on ICD-10 codes were used as the diagnosis of complications. The follow-up period ended when the participant was diagnosed with vascular complications, died, or when the study ended (December 31, 2015). Diabetic vascular complications were classified into macrovascular and microvascular complications (Supplementary Table 1). Vascular complications were also classified into three types: cardiovascular diseases, chronic renal diseases, and retinal diseases. The disease codes were as follows: ischemic heart diseases (ICD-10: I20-I25), cerebrovascular diseases (ICD-10: I60-I68), and arterial and capillary diseases (ICD-10: I70-I79); chronic renal diseases (ICD-10: N18-N19); and retinal diseases (ICD-10: H35-H36). Retinal diseases have many reasons, including congenital diseases. Therefore, exclusion criteria were set for excluding retinal diseases by other reasons (H35.1: immature retinopathy, H35.5: hereditary retinitis).
Statistical analysis. Patient characteristics were presented as mean ± SD (standard deviation) for continuous variables and as frequency for categorical variables. The incidence rates (cases per 1000 person-years) were calculated for the NFG, IFG, and DM groups. In addition to the crude incidence rate, adjusted hazard ratio (HR) was estimated using the Cox proportional-hazards model (reference = NFG group). Adjusted variables were selected stepwise with p-value < 0.05. The final adjusted variables included age, sex, BMI, alanine aminotransferase (ALT), total cholesterol, systolic blood pressure (BP), past history of hypertension, family history of DM, www.nature.com/scientificreports/ smoking status, and drinking habits. For subgroup analyses, adjusted HR was estimated in terms of the low-and high-risk groups. Rate and ratios were presented with 95% confidence intervals. p-value of < 0.05 were considered statistical significance. The proportionality and linearity of the Cox proportional-hazards model were tested with the Schoenfeld residuals and the negative-log hazard plots. All statistical analyses were performed using SAS Enterprise 7.1 (NHIS remote connection).
Ethical approval. This project was approved by the Institutional Review Board of the Korea University (IRB: KUIRB-2019-349-1). Informed consent was waived as de-identified data were used throughout the analysis. All methods were carried out in accordance with relevant guidelines and regulations.

Results
During Compared to the NFG group, IFG and DM groups were significantly associated with an increased risk of microvascular and macrovascular complications, with the DM group demonstrating the highest incidence and risk of complications in all variables, followed by the IFG group (p < 0.001). The IFG and DM groups also showed increasing trends in relative significance between the risk of vascular complications with fasting plasma glucose levels.
General characteristics of study population. Table 1 shows the general characteristics of the study population. The proportion of males was significantly higher than females in all groups. DM participants tended to be older, exercised more, had higher systolic BP, higher BMI, as well as higher fasting plasma glucose, AST, and ALT levels (p < 0.001). Family history of DM and past history of hypertension were also significantly higher in the DM group than NFG or IFG group (p < 0.001). The IFG group contained more current smokers, consumed more alcohol, and higher total cholesterol levels as compared to the other groups. www.nature.com/scientificreports/ Incidence and risk of vascular complications. Incidence and risk of vascular complications of the 3 groups are presented in Table 2. The DM group was significantly associated with the highest risk of vascular complications, followed by the IFG group, and then the NFG group (p < 0.001).
Similar to those observed in cardiovascular diseases, the risk of chronic kidney diseases increased as fasting plasma glucose levels increased. Although the absolute incidence rate of chronic renal diseases was generally lower (NFG: 0.35; IFG: 0.47; and DM: 2.11 per 1000 person-year), the IFG group had higher risk of chronic renal diseases (adjusted HR 1.12, 95% CI 1.00-1.26) than CVD. Furthermore, the DM group showed the strongest association with chronic renal diseases (adjusted HR 2.45, 95% CI 3.16-3.76).
The risk of retinal diseases was not significantly different between the NFG and IFG groups (adjusted HR 1.02, 95% CI 0.95-1.12), but was higher in the DM group (adjusted HR 2.71, 95% CI 2.53-2.90).

Discussion
Large and small vessel diseases are major complications of diabetes. Diabetic patients are prone to coronary heart disease, stroke, ischemic heart disease, cerebrovascular disease, peripheral vascular disease, retinopathy, and chronic kidney disease [11][12][13][14] . In a variety of ways, these conditions can be attributed to prolonged hyperglycemia.
This study included participants aged 40-79 years old with different blood glucose categories with a 12-year follow-up period, and examined the incidence and risk of vascular complications in terms of cardiovascular, www.nature.com/scientificreports/ chronic renal, and retinal diseases. The DM group demonstrated the highest incidence and risk of all 3 vascular complications, followed by the IFG group. To our knowledge, this is the first study that compared the risk of vascular complications using incidence rates between different fasting blood glucose groups. It was confirmed that not only DM patients, but even those with IFG, had a substantial risk of vascular complications. This study also showed that past history of hypertension, family history of DM and high BMI were high-risk factors for vascular complications in IFG patients. Based on a previous systematic review 15 on the relationship between pre-diabetes and CVD risk, 8 publications reported that IFG (100-125 mg/dL) showed a relative rate (RR) of 1.18 (95% CI 1.09-1.28). Another study 16 argued that interventions for vascular diseases should be indicated if intermediate hyperglycemia actually increased the risk of CVD by 20%. Diabetic complications involving atherosclerotic diseases have been more concerning than glucose-centric small vessel diseases 17,18 , and the advancement of arteriosclerosis among prediabetic and diabetic patients have been much faster than that in normal individuals. Another study supported that this accelerated progression of arteriosclerosis often lead to coronary heart disease and stroke 19 , reporting an increased CVD risk in IFG and DM patients with adjusted HR of 1.03 (95% CI 1.09-1.28) and 1.39 (95% CI 1.33-1.46), respectively. In line with this, our cohort study confirmed that efforts to lower blood sugar levels to normal ranges can reduce the risk of CVD. Aneurysm is an important vascular disease. However, in this study, the risk factors of aneurysm remain unclear because there were no statistically significant differences.
The issue of whether intermediate hyperglycemia (pre-diabetes) can induce diabetic microvascular disease has been discussed in detail in the WHO report of 2006 20 , which concluded that increased plasma glucose level is a predicting factor for microvascular disease and mortality. Many studies have reported that small vessel disease is very much related to hyperglycemia, and the only way to prevent or delay microvascular diseases in pre-diabetic patients is to prevent the development of diabetes 21,22 . Chronic renal failure usually occurs about 15-30 years after the onset of type 1 and 2 diabetes 23 , so the incidence rate is low but the potential risk is quite high. Although the absolute incidence rate of chronic renal disease was generally lower, our data corroborated the association between IFG and the high risk of renal complications. In line with this, a relatively high prevalence of decreased kidney function was shown in prediabetic patients of the United State 24 . A cross-sectional study demonstrated that glomerular filtration rate (GFR) is elevated among prediabetic and newly diagnosed diabetic individuals both at baseline and during at least 4 years of follow-up. These findings were consistent with early diabetic glomerular hyperfiltration 25 . Similar to our study, a population-based study in China categorized glycemia into diagnosed diabetes, undiagnosed diabetes, prediabetes, normal glycemia, and estimated the prevalence of kidney diseases. Table 3. Comparison of risk of vascular complications between the high-risk and low-risk subgroups of the IFG group. IFG Impaired fasting glucose. High-risk participants with one or more of the following risk factors: family history of DM, past history of hypertension, BMI > 23 kg/m 2 . Incidence rate in cases per 1000 personyears. Adjusted for age, sex, ALT, total cholesterol, systolic BP, past history of smoking and drinking. www.nature.com/scientificreports/ The prevalence of albuminuria, decreased kidney function and chronic kidney diseases increased with higher glycemic levels 26 . Our findings showed that the risk of retinal disease increased in the high-risk IFG group. Funagata et al. demonstrated that retinopathy was associated with high BMI, and both impaired glucose tolerance and IFG 27 . Retinopathy may occur even before the development of type 2 diabetes, and was associated with hypertension and high BMI, which are the main features of metabolic syndrome 28,29 .
IFG with high-risk factors such as family history of diabetes, history of hypertension and high BMI had higher incidence of vascular complications, and were associated with elevated risk of cardiovascular, retinal and chronic renal diseases. Diabetes and its complications are complex and multifactorial, involving both environmental and genetic components 30 . Genetic determinants of diabetes are likely to be more common in patients with early onset diabetes 31 . Numerous genetic studies have demonstrated a clear genetic component to both diabetes and its complications [32][33][34] . The risk factors for individual and concurrent vascular complications include family history, hypertension, age, family history, duration of diabetes, and BMI. It is well known that obesity, in particular, predisposes to type 2 diabetes, a condition in which even moderate weight loss can improve insulin resistance and chronic hyperglycemia, both of which are related to microvascular complications 35,36 . Taken together, these data suggest that several variables significantly associate with one or more microvascular and macrovascular complications.
Study limitations and strengths. This study had several limitations. First, fasting blood glucose may have been misclassified as only one measurement was taken at baseline. However, fasting plasma glucose was evaluated after no caloric intake for at least 8 h and false-positive cases were therefore likely to be rare. Second, a combination of 75 g oral glucose tolerance test (OGTT) and glycosylate hemoglobin (HbA1c) is required for a more accurate diagnosis of impaired fasting glucose. However, OGTT and HbA1c were not included due to a lack of NHIS-HEALS data. Third, the use of blood pressure and lipid drugs were not distinguished among these participants and the clustering of multiple complications was not considered. Despite the use of a large, nationally representative cohort with a long follow-up period which covered a substantial range of validated vascular complication cases, further studies using more accurate classification of diabetic complications and the effect of drugs are needed to validate our findings (Supplementary Table 1).

Conclusions
This longitudinal study revealed that impaired fasting glucose increased the risk of vascular complications (cardiovascular disease, chronic renal disease, retinal disease). Early preventative interventions are beneficial, especially among IFG patients with high-risk factors such as family history of DM, past history of hypertension, and high BMI. We recommend that a healthy lifestyle, early screening, and continuous follow-up should be emphasized in individuals with impaired fasting glucose.