J-shape relationship between normal fasting plasma glucose and risk of type 2 diabetes in the general population: results from two cohort studies

The prevalence of T2D has rapidly increased over the past few decades. The harmful effects and increased healthcare costs associated with it are a growing public concern [1, 2]. According to the International Diabetes Federation, there were approximately 537 million cases of diabetes worldwide in 2019 [1]. Such individuals are at risk of developing associated complications, including heart disease, stroke, retinopathy, peripheral vascular disease, and kidney disease [3]. The health consequences and economic burden of the diabetes epidemic are enormous, with annual spending already exceeding USD 63.7 billion [1]. However, the current therapeutic strategies for diabetes can only partially prevent complications [4]. Therefore, the primary goals are prediction, prevention, and personalised treatment [5, 6].

Fasting plasma glucose (FPG) is closely related to diabetes, not only as an indicator of pancreatic function and insulin resistance [7] but also as a diagnostic criterion. The normal range for FPG was set at 3.9–6.1 mmol/L by the American Diabetes Association (ADA) in 1997 [8] and by the World Health Organization (WHO) in 1999 [4]. This was subsequently revised to 5.6 mmol/L by the ADA in 2003 [9]; however, the WHO standard remained unchanged [10]. Despite lowering the values of normal FPG range, they may not be effective in preventing T2D. Previous studies revealed that individuals with higher FPG, even those with levels in the normal range, are more likely to develop T2D [11‐21]. Nevertheless, previous studies have focused on specific populations (young men, young adults, adults aged ≥ 30 years, healthy workers, middle-aged, and older people aged > 40 years, and nonobese people) [11‐20] or a definite fasting glucose range (5–5.5 mmol/L) [21]. To our knowledge, there is a lack of studies in the general adult population (age ≥ 18 years) are lacking. In addition, the characteristics of T2D are inconsistent across ethnic populations [22].

Therefore, we analysed the relationship between normal FPG and T2D in a general adult population. Considering East Asians have a common ethnicity, they have similar diabetes characteristics and diagnostic criteria [23]. Therefore, both Chinese and Japanese cohorts were included in the analysis together.

In this study, we observed a nonlinear relationship between the normal FPG range (3.9–6.1 mmol/L) and risk of future T2D occurrence in an East Asian population. The risk of T2D increased significantly when FPG exceeded 4.5 mmol/L and 5.2 mmol/L in the Chinese and Japanese populations, respectively. Overall, the rise in T2D risk in these the East Asian populations was insignificant or sluggish until FPG level reached 4.5 mmol/L.

Previous studies involving various populations exhibited similar results [11‐21]. Tirosh et al. [19] reported an increased risk of T2D when FPG rose 87 mg/dL (4.83 mmol/L) in a group of 13 163 men aged 26–45 years tested at the Israeli Ministry of Defense Medical Examination Centre. Nichols et al. [17] determined that multivariate-adjusted HRs for FPG 90–94 mg/dL (5–5.2 mmol/L) and 95–99 mg/dL (5.3–5.5 mmol/dL) were 1.49 and 2.33, respectively in a US population comprising persons aged > 40 years. The difference between these HRs and those for FPG below 85 mg/dL (4.72 mmol/L), was statistically significant. A study involving an Italian population produced similar results [11]. To the best of our knowledge, only two such studies involving Chinese populations have been reported. In one, ROC curves were used to determine optimal FPG cut-off values for T2D prediction in 18 287 people aged ≥ 60 years in Taiwan, China (5.17 mmol/L and 5.11 mmol/L for men and women, respectively [12]. The other was the West China Hospital study involving a nonobese population (BMI < 25 kg/m2) adjusted for age, sex, family history of diabetes, waist circumference, BMI, SBP, and TG [20]. The results indicated a significantly increased risk of T2D in those with FPG > 4.3 mmol/L, which was similar to the cut-off point of 4.5 mmol/L in our study; However, the sample size was small (450 cases). Previous studies involving Japanese populations showed that a nonlinear relationship exists between FPG and T2D risk in people aged > 40 years [16, 18], male workers aged 40 years [15], and Japanese Americans aged 34–75 years [13]. However, the relationship between FPG 90–99 mg/dL (5–5.5 mmol/L) and T2D is controversial. Munekawa et al. [21] suggested that this could be attributed to distinct ages or populations. Consequently, in 2021, they confirmed that FPG 90–99 mg/dL (5–5.5 mmol/L) is associated with an increased risk of T2D using adult physical examination data from the Matsushita cohort in Japan. We found that the cut-off point beyond which the risk of T2D rapidly increased in the Japanese population was 5.2 mmol/L.

Previous epidemiological studies [26, 27] demonstrated that the incidence of diabetes progressively increases after the age of 40 years, suggesting the necessity of further research in younger populations. Therefore, we included people aged ≥ 18 years in our study and performed a stratified analysis to further elucidate the relationship between normal FPG range and risk of T2D. We found a stronger correlation between FPG and T2D in the following groups of people: age < 45 years, BMI (18.5–24 kg/m²), no hypertension, no smoking, and no drinking. Previous studies showed that age [28], obesity [29], hypertension [30], and smoking [31] are risk factors for T2D development. Consequently, the association between FPG and T2D may be weakened by the presence of these factors; however, drinking enhanced it, which might have been because the study population comprised light to moderate drinkers. Previous studies have shown that drinking is a double-edged sword for T2D; light to moderate drinking improves insulin sensitivity, while heavy drinking inhibits gluconeogenesis [32]. Thus, the effect of FPG on T2D is amplified in the drinking population. Nevertheless, the specific mechanisms regarding regulation of FPG and T2D by the factors mentioned above are unclear and need to be further explored.

Currently, the pathogenesis of T2D is not fully understood. Studies in European and American populations have shown that insulin resistance triggers T2D and ultimately leads to islet β-cell failure [33]. However, Asians, including the Chinese and Japanese, tend to have lower insulin levels and mild insulin resistance at T2D onset [23, 34, 35]. Mitsui et al. [36] reported that people with 100–109 mg/dL (5.6–6.1 mmol/L) FPG exhibit defective insulin secretion and insulin resistance. Therefore, the significantly increased risk of T2D associated with FPG > 4.5 mmol/L in the Chinese population and FPG > 5.2 mmol/L in the Japanese population observed in this study might have been due to defective insulin secretion and insulin resistance. At the cellular level, these homeostatic dysregulations cause cellular stress and mitochondrial malfunction [37, 38]. Increased reactive oxygen species (ROS) generation, mitochondrial membrane depolarisation, and decreased adenosine triphosphate synthesis are symptoms of mitochondrial malfunction [39]. To some extent, mitochondrial dysfunction may precede the emergence of insulin resistance [40]. Mitochondrial dysfunction affects insulin endocytosis in a dose- and time-dependent manner, reducing the expression of insulin receptor signalling [41]. Antioxidants, such as resveratrol and quercetin, reduce oxidative stress by scavenging ROS produced because of mitochondrial dysfunction [42, 43], thus aiding T2D treatment. Antioxidants act at an earlier stage of T2D pathogenesis than traditional insulinotropic agents, insulin sensitisers, and sodium-glucose cotransport protein 2 inhibitors and can serve as a new strategy for T2D treatment, with associated mitochondria-targeted antioxidant drugs showing good tolerability. Nevertheless, evidence for glycaemic control remains limited [44]. This study also provides a theoretical basis for the early application of mitochondria-targeted antioxidant therapy in the normoglycaemic phase.

This study comprising over 200 000 general East Asians showed that a J-shaped relationship exists between normal FPG range and T2D risk, with inflexion points of 4.5 mmol/L and 5.2 mmol/L for the Chinese and Japanese populations, respectively. These results are expected to draw clinicians' attention to the risk of diabetes in a population exhibiting healthy glycaemic index currently. FPG levels below 4.5 mmol/L significantly reduce the risk of progression to T2D. Clinical trial interventions are still required to further validate our results. Future research, regarding early primary prevention and management of populations with healthy glucose levels currently is necessary.