Incidence and predictors of type 2 diabetes mellitus during 17 years of follow-up in the Golestan Cohort Study

Abstract

In this study, we aimed to determine the incidence and predictors of type 2 Diabetes Mellitus (T2DM) in the Golestan Cohort Study (GCS). This study is a prospective population-based cohort study conducted in the Golestan province of Iran with the participation of 50,044 people aged 30 to 87 years between 2004 and 2008. Participants were followed up for 17 years for T2DM. The cumulative incidence of T2DM was 13.32% in the GCS. We observed hypertension (HTN) and dyslipidemia (DLP) increased the risk of T2DM 1.16 and 1.63 times relative to the healthy participants (RR: 1.16, 1.63, 95% CI : 1.102–1.222, 1.393–1.928, p < 0.001). For every one-unit increase in the body mass index (BMI), the risk of T2DM increased 1.09 times (RR: 1.09, 95% CI :1.086–1.106, p < 0.001). High-risk waist circumference (WC) increased the risk of T2DM by 1.89 times more than normal WC (RR: 1.89, 95% CI : 1.756–2.053, p < 0.001). Smokers had an 89% lower risk of T2DM than non-smokers (RR: 0.897, 95% CI : 0.814–0.989, p = 0.029). We conclude that environmental factors induce T2DM by affecting body fat. Also, other metabolic diseases could develop T2DM.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-95442-8.

Introduction

In 1980, the World Health Organization (WHO) identified approximately 108 million people with diabetes mellitus (DM), which increased approximately four times in 2014¹. Besides, it is predicted that this number of people in the age range of 20–79 years will reach 629 million in 2045¹. In low-income countries (LICs) and middle-income countries (MICs), most diabetic patients are under 65 years old. In contrast, in high-income countries (HICS), 44% of people with diabetes are over 65 years old¹. DM was the seventh leading cause of disability worldwide in 2016. According to the WHO report in 2019, DM directly caused about 1.5 million deaths. Additionally, 48% of all deaths before the seventh decade were due to DM^2,3.

In 2005, a survey of 70.981 populations aged 25–64 in Iran reported a 7.7% prevalence of DM, 16.8% impaired fasting glucose (IFG), and about 50% undiagnosed DM among Iranian adults⁴. In 2011, the annual incidence of Type 2 Diabetes Mellitus (T2DM) in Iranian was 1%⁵. In Iran, DM, from the eleventh cause of death in 2009, soared to fifth in 2019⁶.

DM describes a disease identified by high blood glucose levels¹. DM leads to life-threatening diseases, including coronary heart disease, stroke, and end-stage renal disease by vascular damage in the heart, kidneys, and brain or microvascular complications, including peripheral vascular disease, retinopathy, neuropathy, and lower-extremity amputations which increases the cost of medical care, reduces the quality of life and increases morbidity and mortality^1,7.

Studies mentioned several parameters for DM development, including socioeconomic status, medical condition, lifestyle, dietary, environmental, and psychological status^7,8. Also, alternating towards sedentary lifestyle and urbanization has been considered a leading cause of rising DM prevalence in the past decades¹.

DM is a health problem because of increasing rate, its complications, healthcare costs, and mortality and morbidity. Therefore, prevention is crucial, and its achievement depends on identifying risk factors and people at risk. In this line, we aimed this study to determine the incidence and predictors of T2DM in the Golestan Cohort Study (GCS) participants.

Method

Study design

This study was conducted at the GCS frame. GCS is a prospective population-based cohort study conducted in the Golestan province of Iran with the participation of 50,044 people aged 30 to 87 between 2004 and 2008. In GCS, 39,399 people from rural areas and 10,645 from urban areas were included.

Data collection

Our exclusion criteria included participants with DM diagnosis or history at the enrollment. Then, the rest of the GCS participants were followed up for 17 years. Diagnosing T2DM in this study, either at the time of entry or during follow-up, is considered according to the patient’s self-reported diagnosis of T2DM by their doctor or the initiation or use of glucose-lowering medication. Self-reports of T2DM in the GCS study had 61.5% sensitivity and 97.6% specificity⁹.

At the beginning of the study, people’s information was obtained based on an interview by a trained general practitioner with a questionnaire. All participants were actively followed up every 12 months by phone, and a separate questionnaire was completed for each person after each phone call. Also, people have been trained to inform the Golestan cohort team in case of hospitalization or new medical cases, and the follow-up time differed from 15 to 20 years, with a median of 17 years. During follow-up, more than 7000 deaths were recorded, and deceased individuals were not excluded from the data. Also, less than 1% (464) of all participants of cohort were lost to follow-up.

The studied variables included age, blood group, sex, marital status, ethnicity, education, residential area, wealth score, smoking, alcohol and opium consumption, Body Mass Index (BMI), Hip Circumference (HC), Waist Circumference (WC), Waist to Hip Ratio (WHR), Waist to Height Ratio (WHtR), dyslipidemia (DLP), and hypertension (HTN).

In this study, demographic indicators and basic information of patients, including age, sex, education, ethnicity, marital status, residential area, history of alcohol consumption, history of tobacco use, and history of opium use, were collected by a structured lifestyle questionnaire. Also, WC, HC, blood pressure (BP), weight, and height measured by a trained technician.

Blood groups were identified by the blood samples collected from the studied subjects and classified as four groups: A, B, AB, and O. Education was divided into literate and illiterate. The residential area was divided into two parts: urban and rural. DLP was considered positive if anti-lipid drugs were used. In this study, BMI was obtained by dividing weight based on kilograms by height based on square meters. According to the WHO definition, a BMI lower than 18.5 kg/m2 is considered underweight, 18.5 < BMI < 25 kg/m2 is normal, 25 < BMI < 30 kg/ m2 is overweight, and a BMI higher than 30 kg/m² is obese¹⁰. The measured WC was divided into two groups: standard and high risk. According to the classification mentioned in the adult treatment panel (ATP) III, men with WC > 102 and women with WC > 88 were considered high-risk people¹¹. The ethnicity included Turkmen and non-Turkman (Fars, Turk, Sistani, Baloch, Kurd, and Afghani). Wealth score is calculated based on the highest level of education and appliance ownership, including bath in the residence, personal vehicles (car, motorbike), TV (black and white or color), refrigerator, freezer, vacuum cleaner, and washing machine to indicate socioeconomic status¹². People who used cigarettes and other tobacco (nass, hookah, and pipe) once or more a week for at least six months or more were known as a smoker and were divided into two groups: never smokers and ever smokers. Opium users were considered those who have used opium and its derivatives at least once a week and for six months or more. Alcohol consumption was defined as consuming beer, homemade wine, or imported wine at least once a month for six months or more. The subjects’ systolic and diastolic BP was measured twice in each arm, and a sitting position, and the average was included in the study. People were defined as HTN who either had a physician’s diagnosis of high BP or those who took antihypertensive drugs or had the criteria of average systolic BP above 140 mm Hg or average diastolic BP above 90 mm Hg¹³.

Statistical analyses

The nominal and categorical variables were presented as counts and percentages. Also, continuous variables with normal distribution were described by mean ± standard deviation (SD). The chi-square test was used to compare nominal or categorical variables, and the independent t-test was used to compare continuous variables between participants with T2DM and without T2DM. We used simple and multiple univariate Log-Binomial regression with Enter method from Generalized Linear Models to assess the relationship between different variables and the incidence of T2DM. Predictors with a P-value less than 0.2 in simple models were selected to enter the multiple models. To avoid collinearity, WHR and WHtR variables highly correlated with BMI, WC, and HC were not included in multiple regression. Crude and adjusted risk ratios (RR) were calculated with their 95% confidence intervals (95% CI). The cumulative incidence of T2DM in age groups (< 40, 40–49, 50–59, 60–69, and ≥ 70 years) in every sex was calculated by dividing the number of new cases of T2DM by the total number of subjects in that group. To compute incidence proportion, we used the definition given by Jewell. The proportion of a defined population calculated who become new cases of the T2DM before the end of the interval to all who are at risk for the disease at the beginning of a specified time interval. Sometimes, it is referred to as the cumulative incidence proportion since this quantity includes all individuals who become cases over the entire interval¹⁴. P-values of less than 0.05 are considered significant. Data analysis was done using SPSS 26.

Ethics

The ethical review committees of the Digestive Disease Research Center, Shariati Hospital, Tehran University of Medical Sciences, Tehran, Iran (DDRC), International Agency for Research on Cancer, Lyon, France (IARC) and Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, MD, USA (NCI) approved GCS’s protocol and informed consent. Also, The Research Ethics Committee of Golestan University of Medical Sciences approved this study (Approval ID: IR.GOUMS.REC.1402.035). GCS has been carried out in accordance with Declaration of Helsinki. Informed consent obtained from participants at the enrollment of GCS.

Result

From 50,044 participants included in the GCS, people diagnosed with T2DM at the beginning of the study were excluded, then among 40,083 non-T2DM participants during 17 years of follow-up, 5342 people developed T2DM. The cumulative incidence of T2DM was 13.32% in the GCS. Incidence of T2DM was 12.49% and 16.1% in rural and urban residents, respectively. Also, incidence of T2DM was 12% in Turkman ethnicity and 17.56% in non-Turkman ethnics. The cumulative incidence of T2DM in women was higher than in men of the same age (in age group < 40 years 9.77% vs. 7.14%, in 40–49 years group 15.64% vs. 9.79%, in 50–59 years group 17.55% vs. 10.9%, in 60–69 years group 14.3% vs. 10.01%, and in age > 70 years 11.82% vs. 8.33%). The cumulative incidence of T2DM in people in the sixth decade of their life at the time of entering the study in both sexes was higher than in other age groups. The trend of observed incidence in people under 50 years old has an upward trend, the highest in people aged 50–59 years, and a downward trend after 60 years old, which can be seen in Fig. 1.

Fig. 1.

Incidence proportion of T2DM by age and gender (in percent).

The baseline characteristics of participants with T2DM incidence and without T2DM incidence are shown in Table 1.

Table 1.

Baseline characteristics of people with T2DM incidence and without T2DM incidence, N (%) in GCS participants.

Variables	People with T2DM incidence N = 5342	People without T2DM incidence N = 34,741	p-value
Sex			< 0.001
Male	1770 (33.13)	15,819 (45.53)
Female	3572 (66.86)	18,922 (54.46)
Age (mean ± SD)	68.3 ± 8.3	68.6 ± 9	0.01
Age groups			< 0.001
< 40 years	28 (0.52)	266 (0.76)
40–49 years	2576 (48.22)	16,853 (48.51)
50–59 years	1849 (34.61)	10,742 (30.92)
60–69 years	691 (12.93)	5037 (14.49)
≥ 70 years	198 (3.7)	1843 (5.3)
Marital status			0.03
Single	642 (12.01)	3827 (11.01)
Married	4700 (87.98)	30,914 (88.98)
Ethnicity			< 0.001
Turkman	3633 (68)	26,723 (76.92)
Non-Turkman	1709 (31.99)	8018 (23.07)
Education			0.03
Illiterate	3733 (69.88)	23,762 (68.39)
Literate	1609 (30.11)	10,979 (31.6)
Residence			< 0.001
Rural	3854 (72.14)	26,992 (77.69)
Urban	1488 (27.85)	7749 (22.3)
Smoking			< 0.001
Yes	588 (11)	6373 (18.34)
No	4754 (88.99)	28,368 (81.65)
Alcohol			0.008
Yes	146 (2.73)	1194 (3.43)
No	5196 (97.26)	33,547 (96.56)
Opium			< 0.001
Yes	539 (10.08)	5814 (16.73)
No	4803 (89.91)	28,927 (83.26)
BMI (mean ± SD)	29.4 ± 5.4	26 ± 5.2	< 0.001
BMI			< 0.001
Underweight (< 18.5)	84 (1.57)	1810 (5.2)
Normal (18.5–24.9)	956 (17.89)	14,075 (40.51)
Overweight (25-29.9)	1917 (35.88)	11,525 (33.17)
Obese (> 30)	2330 (43.61)	7328 (21.09)
WC			< 0.001
Normal	1339 (25.06)	18,750 (53.97)
High-risk	4003 (74.93)	15,990 (46.02)
HC (mean ± SD)	103.29 ± 9.68	98.74 ± 8.94	< 0.001
WHR (mean ± SD)	0.99 ± 0.07	0.94 ± 0.07	< 0.001
WHtR (mean ± SD)	0.64 ± 0.08	0.58 ± 0.08	< 0.001
HTN			< 0.001
Yes	2175 (40.71)	10,821 (31.14)
No	3162 (68.55)	23,886 (68.75)
DLP			< 0.001
Yes	100 (1.87)	195 (0.56)
No	5242 (98.12)	34,545 (99.43)
Blood group			0.171
A	1748 (32.72)	11,603 (33.39)
B	1515 (2.84)	9379 (26.99)
AB	496 (9.28)	3397 (9.77)
O	1583 (29.63)	10,358 (29.81)
Wealth score (mean ± SD)	0.02 ± 0.2	-0.003 ± 0.2	< 0.001

T2DM: Type 2 Diabets Mellitus; GCS: Golestan Cohort Study; BMI: body mass index; WC: waist circumstance; HTN: hypertension; DLP: dyslipidemia, HC; hip circumstance, WHR; waist hip ratio, WHtR; waist height ratio.

Smoking and opium usage had a negative correlation with BMI (-0.21 and − 0.244), WC (-0.120 and − 0.181), and HC (-0.117 and − 0.186) significantly (p < 0.001). Also, alcohol consumption negatively correlated with BMI (-0.05) significantly (p < 0.001).

All variables except the blood group were significant predictors for T2DM in the simple logistic regression model. After adjustment, the multiple logistic regression model revealed that Turkman ethnicity decreased 0.64 times the risk of T2DM developing compared to the non-Turkman ethnicity (RR: 0.645, 95% CI: 0.611–0.682, p < 0.001). In addition, HTN and DLP increased the risk of T2DM 1.16 and 1.63 times relative to the healthy participants, respectively (RR: 1.16, 1.63, 95% CI: 1.102–1.222, 1.393–1.928, p < 0.001). Every unit raising in BMI increased risk of T2DM 1.09 times (RR: 1.096, 95% CI: 1.086–1.106, p < 0.001) whereas HC decreased T2DM risk 0.97 times (RR: 0.978, 95% CI: 0.973–0.983, p < 0.001). Also, high-risk WC increased risk 1.89 times more than the normal group (RR: 1.898, 95% CI: 1.756–2.053, p < 0.001). Smoking decreased 0.89 times T2DM developing risk compared to non-smoking (RR: 0.897, 95% CI: 0.814–0.989, p = 0.029). —Table 2.

Table 2.

Crude and adjusted risk ratio (RR) and 95% confidence interval (CI) of T2DM predictors in GCS participants.

Variables	Simple regression			Multiple regression
	RR	95% CI	P- value	RR	95% CI	P-value
Constant	-	-		0.088	0.057–0.135	< 0.001
Ethnicity			< 0.001			< 0.001
Non-Turkman	1			1
Turkman	0.681	0.646–0.718		0.645	0.611–0.682
Sex			< 0.001			0.054
Male	1			1
Female	1.578	1.496–1.665		0.934	0.871–1.001
HTN			< 0.001			< 0.001
No	1			1
Yes	1.432	1.361–1.506		1.160	1.102–1.222
DLP			< 0.001			< 0.001
No	1			1
Yes	2.573	2.190–3.023		1.639	1.393–1.928
WC			< 0.001			< 0.001
Normal	1			1
High-risk	3.004	2.833–3.185		1.898	1.756–2.053
Marital status			0.029			0.172
Married	1			1
Single	1.089	1.009–1.175		0.945	0.872–1.025
Smoking			< 0.001			0.029
No	1			1
Yes	0.589	0.542–0.639		0.897	0.814–0.989
Age (increased per year)	0.997	0.994–1.000	0.024	1.001	0.997–1.004	0.721
Education			0.030			0.723
Literate	1			1
Illiterate	1.062	1.006–1.122		1.012	0.947–1.082
Residence			< 0.001			0.710
Urban	1			1
Rural	0.776	0.734–0.820		0.988	0.928–1.052
Wealth score (increased per unit)	1.735	1.537–1.958	< 0.001	0.957	0.833–1.099	0.532
Alcohol			0.009			0.065
No	1			1
Yes	0.812	0.696–0.949		1.165	0.991–1.370
Opium			< 0.001			0.173
No	1			1
Yes	0.596	0.547–0.649		0.938	0.856–1.028
BMI (increased per kg/m2)	1.093	1.089–1.096	< 0.001	1.096	1.086–1.106	< 0.001
HC (increased per cm)	1.041	1.039–1.043	< 0.001	0.978	0.973–0.983	< 0.001
Blood group						-
O	1
A	0.988	0.927–1.052	0.700
B	1.049	0.983–1.120	0.152
AB	0.961	0.875–1.056	0.408

T2DM: Type 2 Diabetes Mellitus; GCS: Golestan Cohort Study; HTN: hypertension; DLP: dyslipidemia; WC: waist circumstance; BMI: body mass index; HC: hip circumstance.

Significant values are in [bold].

Discussion

The cumulative incidence of T2DM in this study was 13.32%. Previous reports estimated 8.7% DM prevalence in Iran and other low to middle-income countries¹⁵. Esteghamati et al. (2005) reported a 7.7% prevalence of DM in the 25-64-year-old adult population in Iran⁴. Other countries in the Middle East and North Africa (MENA) region reported higher percentages, including Saudi Arabia, with a 24% prevalence of DM¹⁶.

In our study, women more than men developed T2DM, but female sex was not a predictor. In the Tehran lipid and glucose study (TLGS), the incidence rate of DM after standardizing the age factor was calculated as 9.36% in men and 10.1% in women¹⁷. Also, in a report on the burden of DM in Iran in 2008, women had a higher prevalence of DM compared to men⁴. Also, a cohort study from Korea with 2009 to 2015 follow-up observed that women’s conversion from without DM to DM was 1.01 times higher than men’s¹⁸. Based on the studies, glucose and lipid homeostasis are influenced by sex hormones, and the balance between estrogen and androgen affects metabolism and body composition¹⁹. Also, higher levels of androgen in women led to weight gain and DM incidence¹⁹. Whereas, according to a review study, in 2017, men with 9.1% diagnosed DM had a higher prevalence relative to women with 8.4%¹⁵. Also, a Chinese cross-sectional study in 2013 showed no association between sex and the prevalence of DM²⁰. Therefore, we compared women and men with T2DM in terms of environmental factor (supplementary Table 1). We observed that women who developed T2DM had higher anthropometric indices (BMI, WC, HC, WHtR) and metabolic disease (HTN, DLP). In contrast, men had higher weight and a history of smoking, opium, and alcohol consumption. Our results demonstrated that different incidence rates of T2DM between the two sexes could be due to higher abdominal fat in women that led to metabolic disease. Based on our results, higher physical activity out of the house and smoking, opium, and alcohol consumption in men led to decreased body fat and lower metabolic disease. In this line, European men diagnosed with DM have a lower BMI compared to women¹⁹. Also, BMI in men and WC in women were associated with an incidence of DM that demonstrated a better role of central adiposity compared with general adiposity in women^9,19.

We observed that individuals with an HTN history had 1.16 RR of T2DM developing risk than people with normal BP. Other studies confirmed that HTN was an independent risk factor for DM^7,18,21. One of the pathophysiologic mechanisms in HTN disease is inappropriate activation of the rennin–angiotensin– aldosterone system (RAAS). Angiotensin II is one of the important products of the RAAS that reduces glucose uptake and insulin action in skeletal muscle. Also, RAAS leads to dysfunction of the pancreatic cells by affecting their blood flow²¹; another mechanism could be systemic inflammation that has a role in the development of DM⁷. Another factor in HTN and DM coexistence was obesity and increased visceral adiposity²¹.

One of our findings was 1.63 RR for T2DM in participants with DLP who were diagnosed according to the history of taking antihyperlipidemic drugs. Studies showed that statins increased the risk of new-onset DM. Statins cause DM through insulin resistance and pancreatic beta-cell dysfunction²². Besides, studies showed DLP as a risk factor for DM^18,23. DLP is linked to DM through a superoxide-dependent pathway that peroxidates low-density lipoprotein (LDL) and is associated with Hyperglycemia¹⁶ ), the free fatty acids lead to the gluconeogenesis in the liver, dysfunction of pancreatic beta cells, and insulin resistance²³.

One of our results was higher RR in people with higher anthropometric indices. All of the anthropometric indices (WC, HC, WHtR, WHR, BMI) were significantly higher in people with T2DM rather than without T2DM. T2DM associated factors were different between BMI groups. In underweight participants, ethnicity, urban residence, DLP, and high-risk WC were important while in normal people, smoking decreased this probability. On the other hand, in overweight and obese individuals, metabolic disease in addition to the anthropometric indices had association with T2DM developing (supplementary Table 2). Anthropometric indices had co-linearity together, while after adjusting, every unit increase in the BMI increased the risk of T2DM incidence by 109%. Also, high-risk WC had 1.89 RR higher than the normal group. In contrast, every unit of HC decreased 97% the risk of T2DM incidence. In this line, some studies showed that larger HC was a protective factor in the development of DM^14,19. In different studies, various anthropometric indices were related to the incidence of DM, including higher BMI^15,18, or higher BMI and WC in Chinese population²⁰. An umbrella review of meta-analyses mentioned all the anthropometric indices related to DM, but BMI had a larger effect size and higher significance⁷. Increased intra-abdominal visceral fat in higher BMI and anthropometric indices led to disruption of insulin metabolism by releasing serum-free fatty acids⁷.

We did not find higher RR for T2DM in opium users and alcohol consumers, while smoking decreased 89% the risk of T2DM. In addition, these individuals, less than non-smokers and non-consumers, had developed T2DM during follow-up significantly. Studies mentioned that moderate alcohol consumption has a protective effect against developing T2DM^7,19. At the same time, other studies revealed alcohol consumption as a risk factor for DM¹⁸. We explained two reasons for our observation: the amount of alcohol consumption could have differently affected DM development, and alcohol had a negative correlation with BMI. Also, in our study, smoking had a negative correlation with BMI and WC and had an RR of 0.89 for T2DM. The role of smoking was important in people with normal BMI (supplementary Table 2). Also, people with underweight and normal BMI smoked cigarettes more than others and had lower incidence of T2DM (Fig. 2.). There is controversy in smoking and DM; based on some studies, smoking is a risk factor for DM through oxidative stress, inflammation, central obesity, resistance to the insulin hormone, and inappropriate functioning of pancreatic beta cells^7,18. On the other hand, a cross-sectional study from Chines with 10,837 participants, observed current smokers had lower BMI relative to the non-smokers²⁴. The mechanism of smoking and weight loss is complex; one of the explanations is appetite-suppressing by cigarettes and controlling weight²⁴. Our colleagues studied opium association with T2DM and BMI in GCS; they showed opium use decreased 17% the incidence of T2DM. This reduction is associated with lower BMI and WC in long-term opium users compared to non-users²⁵. While we observed that opium consumption after adjustment had not significant association with T2DM developing.

Fig. 2.

Incidence of T2DM by BMI and Number of cigarettes smoked per day in men. BMI: body mass index; DM: diabetes mellitus; CI: confidence interval.

After adjusting for anthropometric factors, we did not consider greater RR for T2DM with low socioeconomic factors (low education, urban residence, marriage status, and low wealth score). In contrast, TLGS in 2014 observed low education, especially in men associated with the incidence of DM¹⁷. Also, studies from other countries showed that lower educational levels increased the risk of DM^7,20. In this line, Hu et al. did not find an association between urban residents and DM development²⁰. On the other hand, some studies demonstrated that urban residents had a higher rate of DM than rural residents^4,23. Urban residents were at higher risk of developing DM in the past decades, whereas in the last years, the lifestyle of rural people changed and was similar to that of urban people in terms of diet, smoking, alcohol, opium consumption, and the use of machines instead of physical activity for agriculture. Therefore, these changes could cause less impact on urban residents for T2DM development.

In this cohort study, a sufficient population has been evaluated, with regular long-term follow-up, to assess the incidence of T2DM. The impact of several factors, some of which were previously less evaluated, was investigated. On the other hand, the study population in GCS includes different ethnicities and can be attributed to different ethnicities in Iran. Also, our study had some limitations. We not had some important data from participants, including family history of T2DM, physical activity, and diet. In this study, the diagnosis of T2DM was considered a self-report or using drugs to reduce glucose levels and not a laboratory test. Of course, as mentioned, the sensitivity and specificity of diagnosing T2DM in this way has been evaluated in the sub-cohort in 2011 and found to be acceptable⁹. Also, the time of T2DM diagnosis was not recorded in the data; therefore, we could not calculate person-years.

Conclusion

The cumulative incidence of T2DM in the GCS was 13.32%. The incidence and risk factors of T2DM changed with lapse. Therefore, we must update our knowledge of T2DM factors. We conclude that environmental factors, by affecting anthropometric indices, induce T2DM. Also, other metabolic diseases could develop T2DM.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

“SM contributed to writing the manuscript and data analysis, MNHA contributed to writing the manuscript, GR and MZ contributed to methodology, study design, and revising the manuscript, SGH contributed to data analysis, SZB contributed to revising the manuscript, MM contributed to revising the manuscript. All authors agreed on the final manuscript prior to submission. All authors agreed to be accountable for all aspects of this work.”

Data availability

“In the present study, we obtained the data from the Golestan Cohort Study dataset, which is owned by a third-party organization (The Digestive Disease Research Institute of the Tehran University of Medical Sciences). According to its data-sharing policy, the data of the Golestan Cohort Study may be available by sending requests to the GEMINI Shared Repository (GEM Share) (https://dceg2.cancer.gov/gemshare/). The datasets used and/or analyzed during the current study available from the corresponding author on reasonable request.”

Declarations

Competing interests

The authors declare no competing interests.