Novel analytical approach for diabetes education based on screening outcomes from Japanese physical examination

Background: We examined the time-course changes in four variables, namely, fasting plasma glucose (FPG) level, fasting insulin level (FIL), homeostasis model assessment beta (HOMA-b) score, and homeostasis model assessment-insulin resistance (HOMA-IR) score. The variables were examined before and after the onset of diabetes mellitus (DM). Methods: In total, 1,333 Japanese men underwent four annual screenings for diabetes from April 2006 to March 2010. DM was dened as an FPG level ³ 126 mg/dL or glycated hemoglobin level ³ 6.5%. The variables were examined in healthy men and those with DM in the body mass index [BMI] ³ 25 and < 25 kg/m 2 groups. The projected FPG level and FIL were assessed based on the simulated HOMA-b and HOMA-IR scores, which is considered a novel analytical approach. Results: In total, 99 (7.4%) men developed DM at any point of the screening periods. The HOMA-b values of the BMI ³ 25 kg/m 2 group were 117.6% ± 72.9%, 60.4% ± 29.8%, and 43.1% ± 28.1% at baseline, onset, and 4 years after the onset, respectively. Thus, the decrease in the values was steep. Meanwhile, the HOMA-IR values of the BMI < 25 kg/m 2 group were 1.7 ± 0.8 and 2.2 ± 1.2 at 1 year before onset and at onset. Hence, there was an important but slight increase in the values. This result indicates that patients presented with reduced insulin secretion and elevated IR during the onset of diabetes. The values of the four variables remained within normal ranges in healthy men. Conclusions: The novel analytical approach has clinical relevance as it provides a visual material for diabetes education, which is useful in providing caution against the risk of developing DM.


Background
Diabetes mellitus (DM) is a group of metabolic disorders characterized by chronic hyperglycemia due to impaired insulin secretion and increased insulin resistance (IR) due to the reduced actions of insulin. The United Kingdom Prospective Diabetes Study16 has shown that the homeostasis model assessment beta (HOMA-β) score decreased by up to approximately 50% at the time of type 2 DM diagnosis. Moreover, the pancreatic β-cell function decreased progressively and gradually about 12 years before the onset of DM. 1 In terms of race, a difference was observed in pancreatic β-cell function. That is, the Japanese population has a lower pancreatic β-cell function than the Western population. Therefore, DM in the Japanese population is predominantly characterized by impaired insulin secretion. However, no study has examined the pancreatic β-cell function of the Japanese population with DM. Thus, this retrospective clinical study aimed to investigate the time-course changes in four variables, namely, fasting plasma glucose (FPG), fasting insulin level (FIL), HOMA-β score, and HOMA-IR score, in Japanese men who underwent screening for DM.

Study population
In total, 11

Statistical analysis
All values were expressed as mean ± standard deviation. The two groups were evaluated according to the unpaired student's t-test using the Statistical Package for the Social Sciences software for Windows version 10.07J (SPSS Japan, Tokyo, Japan). A P value < 0.05 was considered statistically signi cant.

Results
In total, 99 patients developed DM at any point of the screening periods, and the following values were obtained at the time of DM onset. The mean age of the participants was 55.4 ± 9.6 years (53.2 ± 10.0 and 57.2 ± 8.9 years in the BMI ≥ 25 mg/m 2 group and the BMI < 25 kg/m 2 group, respectively). The age of the BMI ≥ 25 kg/m 2 group was signi cantly lower than that of the BMI < 25 kg/m 2 group (P = 0.038).

Discussion
The risk of developing DM increases linearly, which is parallel to the severity of obesity. [3][4][5][6][7] However, although the prevalence of obesity is high in Western countries, the prevalence of DM in Asian countries, including Japan, is similar to that in Western countries. 8 In the current study, the mean BMI of the group with DM in Japan was 24.5 ± 3.0 kg/m 2 . Meanwhile, that of patients with DM in the United States is approximately 30 kg/m 2 , indicating a difference in the phenotype of DM in terms of race. 9 In addition to environmental factors (e.g., aging, excessive nutrition, and lack of physical activity and genetic predisposition, impaired insulin secretion in pancreatic β-cells and increased insulin resistance in the liver and peripheral tissues causes type 2 DM. 10 and 40% in obese and lean patients with DM, respectively. 19 Thus, decreased pancreatic β-cell volume contributed to the low HOMA-β scores. The patients with DM in the BMI ≥ 25 kg/m 2 group had elevated HOMA-IR scores at baseline, and the healthy men in the BMI ≥ 25 kg/m 2 group had HOMA-IR scores ranging from 2.0 to 2.2 during the four annual screening periods. We considered that both reduced insulin secretion and elevated insulin resistance contributed to disease onset in patients with DM in the BMI ≥ 25 kg/m 2 group. In contrast, Gautier et al. 20 have reported that an increase in waist circumference was associated with the onset of DM in patients with a BMI < 25 kg/m 2 . An increase in abdominal circumference re ects an elevated insulin resistance. The current study supports this nding because a slight elevation in insulin resistance without a further decrease in insulin secretion contributed to disease onset in patients with DM in the BMI < 25 kg/m 2 group. That is, the pathogenic patterns differ between obese and nonobese patients with DM in the Japanese population. Figure 3 shows a novel analytical approach that can be used to project FPG levels and FIL based on the simulated HOMA-β and HOMA-IR scores. Consequently, patients with DM in the BMI ≥ 25 kg/m 2 group had a signi cant increase in FPG level from baseline to 4 years after onset, and this result is in accordance with the time-course changes in HOMA-β scores. Furthermore, they exhibited no remarkable change in FIL from 1 year before onset to onset. However, a slight increase in HOMA-IR score (probably attributed to an increase in FPG) was observed. In contrast, patients with DM in the BMI < 25 kg/m 2 group had an increase in FIL from 1 year before onset to onset. However, such increase does not result in an elevated FPG level, which is represented by a slight elevation in HOMA-IR score.
In healthy men and patients with DM in the BMI ≥ 25 and < 25 kg/m 2 groups, the values of the four variables remained at extremely narrow ranges during the four annual screening periods. Obesity is a risk factor for insulin resistance, and our study supports this concept because the BMI ≥ 25 kg/m 2 group had higher HOMA-IR scores than the BMI < 25 kg/m 2 group at baseline. In contrast, prior to onset, patients with DM in the BMI ≥ 25 kg/m 2 group and the BMI < 25 kg/m 2 group had HOMA-IR values > 2.5 and 1.5, respectively.
In the DECODE study, approximately 31.3% of patients with DM were missed during the screening for DM because only FPG level was used and OGTT was not performed. 21 However, we did not perform the tolerance test, and the patients were diagnosed with DM based on the FPG and HbA 1c values alone.
Therefore, we did not accurately assess this cohort of patients with DM.
The functions of pancreatic β-cells are regulated by pancreatic β-cell volume and insulin secretion by βcells. The HOMA-β score indicates basal insulin secretion by β-cells. However, we cannot refer to additional insulin secretion by β-cells because the tolerance tests for glucose, glucagon, and arginine were not performed. Moreover, the pancreatic β-cell volume cannot be used as it cannot be identi ed unless necropsy is conducted. Due to the retrospective nature of the study and the fact that only male adults were included in the study, selection bias might have affected the study results.

Conclusions
We developed a novel analytical approach that has clinical relevance as it provides a visual material for diabetes education. Moreover, this approach is useful in providing caution against the risk of developing  Figure 1 Time-course changes in HOMA-β scores according to BMI Solid line: the BMI ≥ 25 kg/m2 group, broken line: the BMI < 25 kg/m2 group a:Diabetic patient group, b:Healthy man group BMI, body mass index; HOMA-, homeostasis model assessment beta model This is a list of supplementary les associated with this preprint. Click to download. FCSTROBEchecklistv4combined25.8.docx