Type 2 diabetes mellitus (T2DM) has become a major public health problem worldwide. The pathophysiology of T2DM is characterized by insulin resistance (IR) and β-cell dysfunction (1, 2). After disease onset, β-cell function progressively declines over time. Thus, exploring potential risk factors of β-cell dysfunction is important for preventing or delaying the development of diabetes (3, 4).

Obesity is a major risk factor for IR and T2DM (5). Although body mass index (BMI) is an internationally recognized index for diagnosing obesity, some studies have shown that obesity defined by BMI is remarkably heterogenous, and people with similar BMI do not have the same level of T2DM risk (6, 7). Abdominal obesity, specifically visceral adipose tissue (VAT), is associated with a greater risk of developing T2DM than peripheral obesity because expanded visceral fat stores affect insulin metabolism by releasing free fatty acids into the portal circulation, which may reduce the hepatic clearance of insulin, thus leading to IR and hyperinsulinemia (8). Further, Chinese have more visceral fat (VF) than Caucasians with the same BMI (9, 10). Therefore, it is significant to discover the differences of insulin secretion and sensitivity in Chinese with different types of obesity.

Measurement of VF accumulation is essential for the diagnosis of obesity. The visceral fat area (VFA) measured by dual bioelectrical impedance analysis (dual-BIA) is a simple and reliable method to estimate VF accumulation. Dual-BIA measures the bioelectrical impedance of the entire abdomen and its surface with a dual current path, which is considered better than the conventional BIA using only one current path and has high correlation with computed tomography (CT), a gold standard for VF accumulation (11, 12). To our knowledge, no population-based studies have examined the associations of β-cell function with VAT evaluated by dual-BIA.

The aim of this study was to examine the association of abdominal obesity assessed by dual-BIA with basal and post-load β-cell function, and clarify whether VAT and subcutaneous adipose tissue (SAT) have the same predictive effect on insulin secretion and sensitivity in Chinese patients with T2DM.

The basic clinical characteristics of participants are summarized in Table 1 . A total of 300 patients, 221 (73.67%) men and 79 (26.33%) women, with a mean age of 51.25 ± 11.96 years were included for data analysis. The median (25th, 75th percentile) VFA and SFA of the subjects were 100.00 (77.00, 127.00) cm² and 171.00 (138.25, 211.00) cm², respectively.

Table 1 also shows the characteristics of subjects according to BMI and VFA levels. Among the 130 participants with normal weight, 99 had normal VF and 31 had high VF. Among the 170 participants who were overweight/obese, 49 had normal VF and 121 had high VF. The participants in the overweight/obese-high VF group were younger, had higher SBP, DBP, Hs-CRP, TG, FINS, INS_2h, FCP, CP_2h, FPIS, SPIS, HOMA2-IR, shorter diabetes duration, lower HDL and Gutt-ISI compared to the normal weight-normal VF group (all p < 0.05). In addition, the participants in the overweight/obese-high VF group had higher DBP, Hs-CRP, FPG, FCP, HOMA2-IR and lower Gutt-ISI than those with normal VF (all p < 0.05).

Spearman’s correlation analysis ( Table 2 ) showed that BMI, VFA and SFA were positively correlated with FPIS (r = 0.302, 0.318, 0.294, all p < 0.001), SPIS (r = 0.152, 0.170 and 0.144, p = 0.008, 0.003 and 0.013), HOMA2-β (r = 0.119, 0.135 and 0.121, p = 0.039, 0.019 and 0.037), HOMA2-IR (r = 0.388, 0.418 and 0.432, all p < 0.001), and negatively correlated with Gutt-ISI (r = -0.271, -0.292 and -0.308, all p < 0.001), while there was no significant correlation between V/S ratio and the indexes of insulin secretion and sensitivity.

Multivariate linear regression was used to analyze the relationship between dependent (FPIS, SPIS, HOMA2-β, HOMA2-IR and Gutt-ISI) and predictor variables (VFA and SFA) ( Table 3 ). After adjusted for gender and diabetes duration, either VFA or SFA was significantly and positively associated with FPIS (standard β = 0.229 and 0.149, p < 0.001 and p = 0.012), HOMA2-IR (standard β = 0.400 and 0.393, both p < 0.001) but negatively associated with Gutt-ISI (standard β = -0.199 and -0.158, p = 0.001 and p = 0.007). After further adjustment for BMI and VFA, the positive association of SFA with FPIS, HOMA2-IR and the inverse association of SFA with Gutt-ISI disappeared. After further adjustment of BMI and SFA, the positive association of VFA with FPIS (standard β = 0.225, p = 0.010), SPIS (standard β = 0.198, p = 0.024), HOMA2-IR (standard β = 0.211, p = 0.008) still remained significant but the negative association of VFA with Gutt-ISI disappeared.

To understand the indexes of insulin secretion and sensitivity associated with different obesity patterns, we developed multiple linear regression models ( Table 4 ). Independent variables were normal weight-normal VF, normal weight-high VF, overweight/obese-normal VF, and overweight/obese-high VF. Dependent variables were FPIS, SPIS, HOMA2-β, HOMA2-IR and Gutt-ISI. After adjustment of gender and diabetes duration, FPIS, HOMA2-IR were higher and Gutt-ISI was lower in the overweight/obese-high VF group when compared to the normal weight-normal VF group and overweight/obese-normal VF group, respectively.

This study examined the cross-sectional associations of abdominal fat distribution with basal and post-load insulin secretion and sensitivity in Chinese patients with T2DM based on different BMI and VF levels. The results demonstrated that (1) after adjustment for gender, diabetes duration, BMI and VFA, some associations of SFA with insulin secretion and sensitivity indices disappeared. Of note, after adjustment of gender, diabetes duration, BMI and SFA, the association of VFA with FPIS, SPIS and HOMA2-IR still remained significant; (2) overweight/obese-high VF patients were more likely to have higher FPIS, HOMA2-IR and lower Gutt-ISI.

Our study found that VFA, rather than SFA, was associated with HOMA2-β, an index reflected basal insulin secretion; VFA was also correlated with post-load insulin secretion, including FPIS and SPIS, independent of SFA and BMI. The influence of obesity on insulin secretion is controversial. Kautzky-Willer et al. (25) found that there was no difference in the dynamic sensitivities to glucose of FPIS and SPIS as studied by the C-peptide minimal model. Bonadonna et al. (26) found an increase in both FPIS and SPIS as assessed by hyperglycemic clamp. Walton (27) et al. and Macor et al. (28) reported that an increased centrality of fat distribution was associated with an elevated SPIS rather than FPIS. Walton (27) et al. and Macor et al. (28) assessed FPIS by intravenous glucose tolerance test (IVGTT) while we assessed FPIS by AST in this study. Progressive impairment in the FPIS to glucose was evident with increasing severity of glucose intolerance; however, patients with T2DM may still have residual β-cell function in response to non-glucose stimulation (29), which may partly explain the association of VFA with FPIS in our study.

In addition to insulin secretion assessed by AST and BMTT, we also used HOMA2-IR and Gutt-ISI to assess insulin sensitivity. In general, HOMA2-IR are derived in the basal state and can therefore be considered to reflects basal or hepatic insulin sensitivity (30), whereas Gutt-ISI is a measure of post–glucose loading insulin resistance and represents both peripheral and hepatic insulin sensitivity, which have a higher correlation with the gold standard method for measuring insulin sensitivity: the euglycemic hyperinsulinemic clamp (31). The results that VFA correlated with HOMA2-IR independent to SFA and BMI suggested that VFA plays important roles in hepatic insulin sensitivity. The association of Gutt-ISI with VFA was disappeared when the confounders were added to SFA and BMI, which indicated that peripheral insulin sensitivity may also affected by SAT. Once SAT reaches its maximal expanding capacity, fatty acids redistribute ectopically in VAT and non-adipose tissues (32–34). Increased VAT leads to an increase in systemic release in resistin and possibly interleukins, and elevated circulating cytokines may play a role in the impairment of muscle insulin response (35). HOMA2-IR was higher and Gutt-ISI was lower in subjects of the overweight/obese-high VF group which confirmed that in subjects with higher VF, even though insulin secretion is higher, their β-cells could not compensate fully for decreased insulin sensitivity, thus leading to diabetes.

This study had several strengths. First, we assessed FPIS and SPIS by AST and BMTT, both of which generate a supraphysiologic insulin secretory response and are less technically demanding than methodologies such as hyperglycemic clamp. AST provides a measure of near-maximal insulin secretion (insulin secretory reserve) (36), while BMTT is easy to administer and is more suitable in β-cell function assessment than OGTT for subjects who have confirmed diabetes. Second, we assessed basal and postprandial insulin sensitivity by employing HOMA2 model and Gutt’s equation, respectively. Although hyperglycemic clamp is the gold standard to assess insulin sensitivity, it is technically challenging, while the indexes in our study are easy to calculate and are suitable for large sample size studies. Third, we divided participants into four groups based on BMI and VF levels, and found that even if subjects had same BMI levels, only those with high VF were associated with the indexes of insulin secretion and sensitivity.

This study had some limitations. First, the number of normal weight-high VF and overweight/obese-normal VF participants in this study was relatively limited; further, T2DM patients with the same BMI may have gender differences in the association of abdominal obesity with β-cell function though we did not find in the present study, a larger population study subgrouped by male and female is needed to confirm our results. Second, the levels of VAT and SAT were measured by dual-BIA rather than CT, which is a gold standard (37, 38). However, CT has problems of complexity, cost and X-ray exposure while dual-BIA, mainly designed to assess VFA and SFA, is simple and may have comparable effectiveness as CT (11, 12). Third, this was a single-center study, and the results might be applicable only to adults with T2DM in southern China.

In summary, the current study suggested that VAT affected basal and post-load insulin secretion and sensitivity. Hence, practitioners should not undermine the risk of IR and β-cell dysfunction in their patients entirely based on BMI, but consider fat distribution as well. For overweight/obese T2DM patients, especially those with VF accumulation, early intervention is needed to reduce VF, in order to delay β-cell dysfunction.

The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by Medical Ethics Committee, Shenzhen Hospital, Southern Medical University (NYSZYYEC202200017). The patients/participants provided their written informed consent to participate in this study.

HH conducted the statistical analyses and wrote the first draft of the manuscript. XZ and XW were involved in the interpretation of data. JZ and YZ contributed to the acquisition of data. LX designed the study and is the guarantor of this work. All authors contributed to the article and approved the submitted version.