Analysis of a non-diabetic middle-aged population from a large prospective, community-based cohort with a long term follow up revealed a significantly higher prevalence of new onset diabetes and CKD in the high HOMA-IR group than in the other groups. High baseline HOMA-IR was an independent risk factor for both new onset diabetes and CKD regardless of the HbA1c level. However, there was no association between high HOMA-IR and macrovascular events.
Many previous studies have shown the relationship between IR and diabetes [3, 4, 6]. IR refers to reduced responsiveness to insulin in tissues that take up glucose, such as liver, skeletal muscle, and adipose tissue [14, 15]. In compensation for IR, the synthesis of insulin in β cells increases and hyperinsulinemia occurs, leading to impaired glucose disposal . Type 2 diabetes mellitus (T2DM) is induced when there is a combination of insufficiencies in β cell mass and function to meet the demands of IR. A high serum glucose level inhibits the proliferation and de-differentiation of β cells through a process called “glucotoxicity,” which gradually leads to reduced insulin secretion . In the high HOMA-IR group, the HOMA-β-cell value for β cell function decreased more than that in the other groups in our study. Given that individuals with T2DM have a β cell mass and function occasionally preserved within normal range in the early period of T2DM progression, β cell mass and function insufficiencies were relative rather than absolute. Although an insufficient β cell mass is essential for the development of T2DM , it is difficult to accurately measure β cell mass in living people, and insulin secretion capacities widely vary; therefore, β cell mass has limited use as a biomarker for new onset diabetes.
In contrast, IR is commonly observed in most T2DM patients and in individuals with impaired glucose tolerance. IR is reportedly the strongest predictor of T2DM, and diabetes can be prevented by improving IR . IR begins from the very early stage of diabetes and can thus be used as an early biomarker to estimate the risk of new onset diabetes. A recent Saku study  assessed IR and diabetes in 2209 non-diabetic patients. Changes in HOMA-IR were measured in the non-diabetic patients and showed that the incidence of T2DM was high when the changes in HOMA-IR were moderate or high. The Saku study also showed that IR had a strong impact on the development of diabetes. This finding is line with our study. Our study showed a significant relationship between high HOMA-IR and new onset diabetes, even in non-diabetic patients from a larger prospective community-based cohort with ten years of follow up.
BMI is a metabolic disease parameter and is associated with IR and diabetes . In our study, BMI was found to be a significant risk factor for new onset DM but not vascular events. Because BMI is calculated based only on height and weight, it does not seem to represent metabolic status more sensitively than other parameters, such as visceral fat and waist-to-hip ratio [19,20,21]. Vascular disease is also directly affected by factors other than metabolic disease, such as high blood pressure.
Another important finding of the present study was that high HOMA-IR value was found to be an independent risk factor for CKD among non-diabetic individuals, whereas HbA1c was not. Various studies have been conducted on the relationship between CKD and IR [22,23,24,25]. CKD, itself characterised by a low-grade inflammatory state, can cause IR and vice versa. CKD and IR adversely affect each other, accelerating the deterioration of renal function . There are several mechanisms that have been suggested to underly the relationship between CKD and IR, one of which is hyperinsulinemia, which increases oxidative stress, protein glycosylation oxidation, and lipid peroxidation . Hyperinsulinemia causes glomerular hyperfiltration, endothelial dysfunction, and increased vascular permeability. IR with oxidative stress and inflammation is thought to play roles in microalbuminuria development and kidney function impairment . In addition to hyperinsulinemia, inappropriate activation of the renin-angiotensin-aldosterone system may cause renal insufficiency . Eventually, IR can lead to glomerulosclerosis and tubulointerstitial injury. A 3-year prospective cohort study with 7200 patients showed that the incidence of CKD and rate of decrease of eGFR were higher in the high HOMA-IR group with metabolic syndrome . Our study also showed that high HOMA-IR was an independent risk factor for CKD after adjusting for multiple risk factors, including HbA1c and baseline GFR. HbA1c was not a risk factor for CKD in our study. This might be due to the fact that the HbA1c value in our study was within the normal range, unlike the 7% HbA1c standard value for predicting microvascular complications in the UKPDS study .
The incidence of macrovascular events did not differ among the HOMA-IR groups. Some studies have reported a relationship between IR and cardiovascular events. However, it is difficult to directly compare these with our study, since most of these previous studies have a cross-sectional design and involved a few participants or participants who already had atherosclerosis identified as a high-risk factor [22, 27,28,29]. The clinical significance of IR for cardiovascular disease may more likely be as a factor accelerating disease progression in patients with certain risk factors, such as CKD, rather than as an independent risk factor [22, 29]. Another issue with our study was that it focused on a population that was relatively young and healthy, and thus, the risk of cardiovascular events was very low. Major vascular complications begin to develop about 10 years after diabetes diagnosis. Follow-up of the cohort in the present study is still ongoing; thus, we hope to observe very long-term cardiovascular events and influence of high baseline HOMA-IR value.
There are several limitations to this study. First, IR was evaluated using only HOMA-IR. The gold standard for evaluating IR is the hyperinsulinemia-euglycemic glucose clamp technique , but it is clinically difficult to implement and even more difficult to apply in large-scale cohort studies. In contrast, HOMA-IR is widely used to measure IR and has yielded reliable results in many studies [28, 29]. Second, clinical data were obtained through standardised questionnaires by a trained interviewer. However, the incidence of macrovascular events in this relatively healthy cohort was lower than that among people with diabetes. Large cohort studies routinely use standardised questionnaires, and the incidence of macrovascular events in our Korean cohort was similar to that in other ethnic groups without diabetes . Third, the absence of data regarding other microvascular events, such as retinopathy, could be a limitation, although the expected incidences of end-stage DM-related microvascular events in our cohort are very low, as the participants did not have diabetes at baseline. It is difficult to infer the effect of specific medication on clinical events because of the lack of medication data. As mentioned above, participants included in this analysis had a very low incidence of risk factors; thus, the effect of drugs in our cohort is expected to be insignificant.