We found significant associations between METS-IR and incidence of CKD, as well as between HOMA-IR and incidence of CKD using the data from a large community-based prospective cohort with a 14-year follow-up. Our results suggest that there was no significant difference in the predictive abilities of METS-IR and HOMA-IR for CKD prevalence at baseline. Considering the comparable predictive abilities of METS-IR and HOMA-IR for CKD prevalence, our findings suggest that it may be more practical to use lower cut-off values of 1.3 for HOMA-IR and 35.5 for METS-IR in clinical settings, rather than the widely accepted cut-off value of 2.5 for defining IR.

Our results showed that, contrastingly, METS-IR had a higher predictive ability for CKD incidence compared to HOMA-IR (Heagerty's iAUC: 0.775 vs. 0.767). Even after the subgroup analysis of potential risk factors for CKD incidence, METS-IR remained a better predictor. Furthermore, the combination of both indices did not significantly enhance the predictive power of CKD incidence compared with using METS-IR alone. 

There could be several reasons for the superior predictive ability of METS-IR over HOMA-IR for CKD incidence. One possible explanation is that METS-IR is better at describing peripheral IR compared to HOMA-IR [11]. CKD could be more closely associated with peripheral IR than hepatic IR. This is because post-receptor signal defects, such as reduced PI3K/Akt activity in skeletal muscle and adipose tissue, are recognized as the primary cause of insulin resistance in patients with CKD [7, 25]. Previous studies have demonstrated that dysfunction in the PI3K/Akt signaling pathway may lead to the inhibition of anti-lipolytic function, a rise in free fatty acid release, elevation in serum TG levels, and the onset of ectopic fat deposition, ultimately leading to lipotoxicity and IR [26].

Second, the superiority of METS-IR is its effectiveness as a screening and predictive tool for diseases related to metabolic syndrome [27, 28]. Metabolic syndrome is closely linked to CKD, as the kidney is a highly sensitive target organ for metabolic syndrome [29]. Numerous studies have provided significant evidence linking IR and chronic inflammation to metabolic syndrome, which can lead to anomalies in lipid and glucose metabolism [30,31,32].

Third, METS-IR includes BMI, which can provide valuable insights into the nutritional status of individuals with CKD. Given that many patients with CKD are affected by protein-energy malnutrition, they may present with a micro-inflammatory state, low BMI, progressive skeletal muscle wasting, and insufficient nutritional and caloric intake [33]. Studies have indicated that individuals with CKD frequently suffer from anorexia, leading to a decrease in daily food intake, which can result in malnutrition and reduced plasma albumin levels, negatively impacting muscle protein synthesis and metabolism [34]. Therefore, METS-IR is deemed to be a more reliable predictor of CKD than HOMA-IR, since it consists of three direct components of metabolic syndrome (glucose, TG, and HDL-C levels) and one indirect component (BMI). A study conducted among the Chinese population demonstrated that METS-IR is linked to CKD and albuminuria. For each 1-unit increase in METS-IR, the risk of both CKD and albuminuria rose by 2%. It has also been reported that the higher the METS-IR, the higher the risk of CKD (odds ratio (OR): 1.02, 95% CI 1.01–1.03) [35].

Fourth, METS-IR reflects the impact of chronic inflammation on IR in individuals with CKD. In the early stages of CKD, patients typically have elevated levels of circulating inflammatory cytokines, such as tissue necrosis factor alpha (TNF-α), interleukin-6, interferon gamma, and lipopolysaccharide. These circulating inflammatory cytokines are produced by various organs in the body, including the kidneys, adipocytes, liver, or muscles [36, 37]. The activation of p44/42 kinase by TNF-α suppresses early insulin receptor signaling, which disrupts insulin's antilipolytic function and triggers the production of free fatty acids through lipolysis in adipose tissues, leading to increased serum TG levels [38]. Additionally, HDL-C suppresses the production of several pro-inflammatory cytokines and chemokines, and reduces the expression of adhesion molecules, demonstrating its anti-inflammatory properties [39]. Therefore, METS-IR is considered to be a better indicator of chronic inflammation than HOMA-IR.

The final possible explanation for the superior predictive power of METS-IR over HOMA-IR is that Koreans typically have a smaller pancreatic volume and higher pancreatic fat content than their Western counterparts with similar BMIs and body fat indices; this can result in reduced pancreatic secretions. Therefore, using HOMA-IR may underestimate IR in Koreans [40].

Our study revealed an interesting finding that the predictive power of both METS-IR and HOMA-IR for CKD incidence increased until year 8 of follow-up, but decreased after year 10, leading to a similar predictive power. This suggests that factors other than IR may affect CKD development over time. One possible factor is the function of pancreatic beta cells, which have been shown to be impaired without a change in pancreatic beta cell mass in patients with CKD. Furthermore, beta cell dysfunction alone may be enough to cause glucose intolerance [41]. Moreover, Koreans are known to have a genetic trait that lowers their insulin secretion compared to their Western counterparts [42]. Therefore, in the presence of IR, beta cells typically increase insulin secretion several-fold as a compensatory response [43]. However, Koreans are often unable to increase pancreatic insulin secretion sufficiently, regardless of obesity [44]. Second, the causes of IR in CKD are complex and can be influenced by multiple factors. Various risk factors, such as physical inactivity, oxidative stress, vitamin D deficiency, metabolic acidosis, anemia, and microbial toxins, may contribute to IR in CKD. Additionally, the impact of these factors on patients with CKD may vary over time. [4] For example, a study has shown that in patients with CKD, vitamin D deficiency can inhibit the secretion of insulin in response to glucose stimulation. Additionally, the same study found that vitamin D supplementation can raise insulin levels in vivo [45]. Furthermore, several risk factors, such as autoimmune diseases, genetic disorders, environmental pollution, and increased prevalence of DM and HTN, have been identified as important contributors to CKD [46,47,48,49]. Further studies are required to identify the precise molecular mechanisms responsible for the pathogenesis of IR in CKD and investigate other high-risk factors to predict CKD.

HOMA-IR has been validated and is a useful surrogate index to measure IR in clinical application; however, the serum insulin is not a routine laboratory measurement in usual clinical settings. Therefore, potential alternative indicators of IR, such as triglyceride-glucose (TyG) index, have also been studied and validated. In the present study, the AUC for HOMA-IR and TyG were not significantly different at baseline (p for AUC comparison = 0.693), and the AUC for HOMA-IR and TyG were also not significantly different at baseline (p for AUC comparison = 0.382) (Additional file 1: Figure S1). Interestingly, Harrell’s C index for TyG was significantly greater than that of HOMA-IR, but there was no significant difference in Harrell's C index between METS-IR and TyG. Additionally, Heagerty's iAUC for TyG was significantly higher than that of HOMA-IR (p = 0.007); however, Heagerty's iAUC for METS-IR was significantly higher than that of TyG (p = 0.011) (Additional file 1: Tables S1, S2).

This study had some limitations. Firstly, the study population was limited to Koreans; therefore, the findings may not be applicable to other ethnic groups. Second, since we used metabolic parameters and anthropometric measurements which were measured in a baseline survey, changes in METS-IR and HOMA-IR during the follow-up period could not be considered. Despite these weaknesses, to the best of our knowledge, this study is the first to investigate the effect of METS-IR and HOMA-IR on CKD incidence using large-scale population-based data.

In the present study, we found that both METS-IR and HOMA-IR have a high predictive power for CKD development, but METS-IR was superior to HOMA-IR. Given its convenience and economic feasibility compared to HOMA-IR, METS-IR could be an effective tool for early detection and prevention of CKD in the Korean population.