Time-varying association between physical activity and risk of diabetes in the early and late adulthood: A longitudinal study in a West-Asian country

Time-varying association between physical activity and risk of
diabetes in the early and late adulthood: A longitudinal study in a
West-Asian country
Parisa Naseri a,b, Parisa Amiri a,∗, Hasti Masihay-Akbar a, Navideh Sahebi Vaighana,
Sajad Ahmadizadc, Arash Ghanbariand, Fereidoun Azizi e
a Research Center for Social Determinants of Health, Research Institute for Endocrine Sciences, Shahid Beheshti University of Medical Sciences, Tehran, Iran b Department of Biostatistics, School of Paramedical Sciences, Shahid Beheshti University of Medical Sciences, Tehran, Iran c Department of Biological Sciences in Sport, Faculty of Sports Sciences and Health, Shahid Beheshti University, Tehran, Iran d Prevention of Metabolic Disease Research Center, Research Institute for Endocrine Sciences, Shahid Beheshti University of Medical Sciences, Tehran, Iran e Endocrine Research Center, Research Institute for Endocrine Sciences, Shahid Beheshti University of Medical Sciences, Tehran, Iran
a r t i c l e i n f o
Article history:
Received 15 October 2020
Received in revised form 4 July 2021
Accepted 16 July 2021
Available online xxx
Keywords:
Diabetes
Physical activity
Time-varying
Cox proportional hazard model
a b s t r a c t
Background: The time-varying association between physical activity (PA) and incidence of type 2 diabetes
(T2DM) is still unclear. The present study aimed to investigate this association in the early- and lateadulthood during a 9-year follow-up.
Methods: This study was conducted on 3905 participants in early and late adulthood, using the Tehran
Lipid and Glucose Study (TLGS) dataset. PA was assessed via the Iranian version of Modified Activity
Questionnaire (MAQ). The association between trend of PA and incident T2DM was investigated using
time-varying Cox’s proportional hazard model. Variables including job, education, smoking and body
mass index (BMI) were adjusted in the final model.
Results: The distribution of sex- and age-specific levels of PA changed significantly over time. Compared
with physically inactive women, for older women with high level of PA, the risk of T2DM was 0.64 (95%
CI: 0.43−0.95, P = 0.02) in adjusted model. Moreover, hazard for low PA group was significantly higher
than the moderate group, and for these two groups were significantly higher than high PA level (P < 0.05).
Conclusion: High PA level can postpone the incident T2DM in early-aged and elderly women, over time.
Therefore, gender and age are of great importance in designing the PA modifying programs to prevent
T2DM.
© 2021 Published by Elsevier Ltd on behalf of Primary Care Diabetes Europe.
1. Introduction
Diabetes mellitus counts as one of the most significant health
care problems and challenges tohumanhealthover the last decades
[1]. There is a significant rise in type 2 diabetes (T2DM) in both
developed and developing countries [2]. According to the International Federation of Diabetes report, T2DM has been diagnosed
in 151–415 million individuals between 2000–2015, which is estimated to be increased by 642 million in 2040 [3]. The highest global
prevalence of diabetes has been reported in the Middle East and
North Africa (MENA) region in the adult population (10.9%). In this
∗ Corresponding author at: Research Center for Social Determinants of Health,
Research Institute for Endocrine Sciences, Shahid Beheshti University of Medical
Sciences, P.O. Box: 19395-4763, Tehran, Iran.
E-mail address: amiri@endocrine.ac.ir (P. Amiri).
regard,the prevalence of T2DM for Saudi Arabia, Egypt, United Arab
Emirates, and Tunisia were estimated as 20.2%, 15.6%, 10%, and
9.2%, respectively [4]. Based on a recent World Health Organization report, 10.3% of the Iranian population (9.6% men and 11.1%
women) aged 20–74 years suffer from T2DM [5]. Moreover, age
as an important factor plays a crucial role where the prevalence
of type 2 diabetes is increased by age [6]. In recent decades, the
worldwide prevalence of diabetes has increased considerably in
countries with various income levels; 40% of this increase is related
to population growth and aging [7]. In the USA, the prevalence of
diagnosed and undiagnosed diabetes in people aged over 75 years
was reported to be 28.3% [8]. As a developing country, Iran is experiencing rapid growth in the elderly population [9]. It is estimated
that individuals aged over 65 will grow from 5.7% in 2011 to 9.7%
in 2030 [10]. Approximately 82.5% of those with diabetes will be
living in developing countries, including Iran, by the year 2030 [11].
https://doi.org/10.1016/j.pcd.2021.07.012
1751-9918/© 2021 Published by Elsevier Ltd on behalf of Primary Care Diabetes Europe.
Please cite this article as: P. Naseri, et al., Time-varying association between physical activity and risk of diabetes in the early and late
adulthood: A longitudinal study in a West-Asian country, Prim. Care Diab., https://doi.org/10.1016/j.pcd.2021.07.012
ARTICLE IN PRESS G Model
PCD-1081; No. of Pages7
P. Naseri et al. Primary Care Diabetes xxx (xxxx) xxx–xxx
Since no definite treatment for T2DM is yet available, primary
prevention is crucial[12]. It has been well-documented that regular
physical activity, as one ofthemostimportant priorities for improving well-being, plays a crucial role in preventing T2DM through
increasing energy expenditure, resting metabolic rate, and body
fat loss [13]. Furthermore, regular exercise reduces insulin resistance and hyperinsulinemia [14], which their etiological role for
T2DM is substantial [15]. Existing data show that a third of adults
globally do not participate in physical activity during their lifespan
[16]. A national study in Iran revealed that 40% of the population
(31.6% of men and 48.6% of women) have a low level of physical
activity [17] and only 14–16 percent of Iranian women aged 45–75
years meet the recommended levels of physical activity [18]. A
population-wide study indicated that physical inactivity is more
prevalent in the middle- and older-aged people [19], where 65% of
adults aged 75 and older reportno leisurephysical activity [20]. Previous studies indicated the importance of physical inactivity as an
important predictor for survivaltime in older adults than inmiddleaged groups [21]. Although the evidence indicated that a healthy
lifestyle is an effective program for aging, older persons do not meet
required health-promoting behavior such as physical activity [22].
According to previous investigations, the association between
physical activity and T2DM is still controversial. Some observational cohort studies [23,24] and systematic reviews indicated
that increasing daily physical activity could decrease the progression and incidence of T2DM [25]. In contrast, few studies have
reported no significant association between physical activity and
T2DM [26,27]. Most previous studies assessed physical activity at
baseline, and further studies are required to investigate the timevarying effect of physical activity on the incidence of T2DM.
Physical activity is likely to change over time [28], though it
seems that not much attention has been paid to its variability in the
previous studies. To our knowledge, only two studies considered
physical activity changes on the incidence of T2DM in the specific
populations, including older British men and American employees, during less than ten years follow-ups. One study targeted
high-risk adults in China to explore the effectiveness of intensive
lifestyle modifications, including physical activity, on the incidence
ofthe disease during a long-term follow-up period [29]. The current
study is the first effort to investigate the time-varying association
between physical activity and risk of T2DM in a general population
of West-Asian men and women in their middle and late adulthood
during a 9-year follow-up.
2. Methods
2.1. Study design and participants
This study has been conducted in the framework of Tehran lipid
and glucose study (TLGS), a population-based cohort to determine
the prevalence of non-communicable diseases (NCDs) risk factors,
conducted on a representative sample of residents of district 13 of
Tehran. The TLGS was designed with two main junctures: Phase I, a
cross-sectional study of the prevalence of NCDs and their risk factors, implemented from 1999 to 2001; and prospective follow-ups
starting from Phase II, along with lifestyle interventions and triennial data recollection. Four follow-up re-exams have been carried
out from 2001 until 2015. Using multistage cluster random sampling methods, 15,005 individuals aged three years and over were
recruited in the study at baseline. All socio-demographic, behavioral, anthropometric, and clinical data were collected through
face-to-face interviews by trained examiners.
The current longitudinal analysis has been conducted on data
from the second to fifth follow-up examinations of the TLGS. From
7349 participants without diabetes (phase 2), 3905 individuals
Participants without diabetes
who were included for
longitudinal study
(n=7349)
Included for longitudinal
analysis
(n=3905, 59.5%)
Loss to follow up
(n=782, 12%)
Included:
(n=6567, 88%)
Missing covariates across
phases(n=2662, 40.5%)
Fig. 1. Flowchart showing study population selection method; subjects were
selected from the population of the TLGS.
(58.6% women and 41.4% men) who had completed sets of data
were followed until phase 5 for an average period of 9 years (Fig. 1).
Participants were categorized into two age groups including, 1)
early adulthood (20−40 years) and 2) middle-aged and elderly (≥40
years) who were 837 (39.4%) men and 1144 (49.97%) women. The
mean survival time was 144.76 months. Informed written consent
was obtained from all participants. This study was approved by
the Ethical Committee of Research Institute for Endocrine Sciences,
Shahid Beheshti University of Medical Sciences.
2.2. Measurements and definitions
Physical activity levels, including leisure time and occupational
activities, were assessed using a reliable and validated Iranian versionoftheModifiedActivity Questionnaire (MAQ)[30]. Participants
reported their physical activities and frequency and duration for
each activity over the past 12 months. Total numbers of minutes/year for all leisure time physical activities were summed up
and then divided by 52 to estimate total leisure time physical activities per week. The metabolic equivalent (MET) of total leisure time
physical activity for each person was then calculated by multiplying each activity’s number of minutes/week by its MET. One MET
is set at 3.5 ml of oxygen consumed per kg of body weight per
minute and represents the resting metabolic rate [30]. According
to the questionnaire, participants were asked to report the number of months and hours they participated in physical activity at
work (standing, housework, work activities) over the past year. The
assessment of occupational activity was based on summing up the
number of hours per week of light, moderate, and high intensity
activities and multiplying by 60 to express minutes per week of
occupational activity over the past year. Final occupational (METminutes/week) activity was calculated by multiplying the number
of minutes per week for each of the three categories of occupational activity by METs (10). Total physical activity was recorded by
adding leisure time physical activity to occupational activity. Levels of physical activity were defined as low (MET < 600 min/wk),
moderate (MET 600–2999 min/wk), and high (MET ≥ 3000 min/wk)
[31].
Weight was measured to the nearest 0.1 kg using a digital scale,
without shoes and minimum clothes. Height was measured using a
stadiometer. Body mass index (BMI) was calculated as weight (kg)
divided by height in square meters. The same trained examiner
made all measurements.
Type 2 diabetes was defined based onfasting plasma glucose test
(FPG), post-challenge plasma glucose (PCPG), and the presence of
2
ARTICLE IN PRESS G Model
PCD-1081; No. of Pages7
P. Naseri et al. Primary Care Diabetes xxx (xxxx) xxx–xxx
Table 1
General characteristics of participants according to sex and age groups at baseline.
Men (n = 1616) Women (n = 2289)
≤ 40 year n(%) >40 year n(%) P-value ≤ 40 year n(%) >40 year n(%) P-value
Education
Primary 163 (19.5) 353 (45.3) < 0.001 297 (26.0) 805 (70.3) < 0.001
Secondary 456 (54.5) 275 (35.3) 608 (53.1) 272 (23.8)
Higher 218 (26.0) 151 (19.4) 239 (20.9) 68 (5.9)
Job status
Employed 744 (88.9) 553 (71.0) < 0.001 201 (17.6) 87 (7.6) < 0.001
Unemployed 92 (11) 16 (2.1) 936 (81.8) 928 (81.0)
Unemployed with Income 1 (0.1) 210 (27.0) 7 (0.6) 130 (11.4)
Smoking
Non-smoker 504 (60.2) 444 (57.0) 0.001 1105 (96.6) 1081 (94.4) 0.041
Ex-smoker 125 (14.9) 172 (22.1) 15 (1.3) 26 (2.3)
Current smoker 208 (24.9) 163 (20.9) 24 (2.1) 38 (3.3)
Physical activity
Low 369 (44.1) 343 (44.0) 0.364 365 (31.9) 325 (28.4) 0.184
Moderate 267 (31.9) 269 (34.5) 643 (56.2) 675 (59.0)
High 201 (24.0) 167 (21.4) 136 (11.9) 145 (12.7)
BMI (kg/m2) 25.7 (4.2) 26.7 (4.7) < 0.001 26.5 (3.8) 29.5 (4.3) < 0.001
at least one of these criteria: FPG ≥ 7 mmol/L, or 2-h PCPG ≥ 11.1
mmol/L or taking anti-diabetic medication based on the American
DiabetesAssociation (ADA)[32]. Education level was defined as primary, secondary and higher. Smoking habits were categorized into
never smokers, former smokers, and current smokers. Job status
was categorized in three sub categories: employed, unemployed,
and unemployed with income.
2.3. Statistical analysis
Participant characteristics across sex and age categories were
summarized as mean ± standard deviation and frequencies (percentages) for continuous and categorical variables, respectively.
Baseline characteristics for men and women were compared
between age categories using independent samples t-test for continuous and chi-square tests for categorical variables. The physical
activity trend in different sex- age groups was evaluated using
Generalized Estimating Equation (GEE). Data were analyzed using
time-varying Cox’s proportional hazard model. When a covariate
changes over time during the follow-up period, we encountered
with time-varying covariate. After restructuring the dataset, Cox’s
regression model using stcox command in stata was fitted; also, we
considered a standard error for cluster estimator using vce (cluster
clustvar) option.
The association between the trend of physical activity and incidence of type 2 diabetes was assessed using four models, including
1) unadjusted (univariate model), 2) adjusted for job status and
education, 3) adjusted for smoking status, and 4) adjusted for all
mentioned confounders, as well as BMI. The gender- and agespecific Hazard Ratios (HRs) and their 95%confidence intervals (CIs)
were calculated for predictors. Statistical analyses were performed
using STATA version 14. The level of significance in all statistical
analyses was set at P < 0.05. and education, 3) adjusted for smoking status, and 4) adjusted for all mentioned confounders, as well
as BMI. The gender- and age-specific Hazard Ratios (HRs) and their
95% confidence intervals (CIs) were calculated for predictors. Statistical analyses were performed using STATA version 14. The level
of significance in all statistical analyses was set at P < 0.05.
3. Results
Baseline characteristics of 3905participants according to gender
and age groups are presented in Table 1. Physical activity levels
were not statistically different in study groups for both men and
women. However, other covariates, including education, job status,
smoking, and BMI, differed between age categories in both men and
women (P < 0.05).
Levels of physical activity based on sex and age groups across
follow-up examinations (from second to fifth) are shown in Table 2.
Except for women aged over 40 years (P = 0.37), the distribution
of physical activity level was significantly changed over time in
all study groups (P < 0.001). Based on the physical activity levels, means of diabetes survival time were 143.3, 145.1, and 144.8
months for low, moderate, and high levels of physical activity,
respectively.
The results of time-varying Cox’s models are presented in
Table 3. Unadjusted HR for diabetes showed that compared to the
low physical activity group, in women aged ≥ 40 years, the incidence of diabetes in high and moderate levels of physical activity
were 0.59 (95% CI: 0.40−0.88, P = 0.01) and 0.78 (95% CI: 0.60−1.01,
P = 0.06), respectively (model 1). After adjusting for job status and
education, HR for diabetes incidence was 0.63 (95% CI: 0.42−0.93,
P = 0.02) in women over 40 years with high physical activity compared to those with low physical activity (model 2). Considering
smoking status (model 3) and BMI (model 4) as adjusting factors,
HRs for women over 40 years with high physical activity were 0.63
(95% CI: 0.43−0.94, P = 0.02) and 0.64 (95% CI: 0.43−0.95, P = 0.02)
during the follow-up time.
The cumulative hazard functions of diabetes for low/moderate
and high physical activity levels are demonstrated in Fig. 2. Hazard for the low physical activity group was significantly (P < 0.05)
higher than the moderate group, and for these two groups were
significantly higher than the high physical activity level (P < 0.05).
Regarding Fig. 2, HR (95% CI) for diabetes incidence were 0.80
(0.62−1.04) and 0.64 (0.43−0.95) in moderate and high physical
activity levels compared to low physical activity group, respectively.
4. Discussion
This study aimed to examine the effect of physical activity
changes on the incidence of T2DM in a large group of Iranian urban
populationinbothearly and late adulthood over a 9-year follow-up.
Our baselinefindings indicatedinsufficientphysical activity inmost
participants, which, despite its positive slope over time, was far
from the criteria defined for optimal physical activity. In addition,
after adjusting for potential socio-behavioral confounders,the incidence of T2DM was significantly lower in middle-aged and elderly
women with moderate and high physical activity compared to their
counterparts with lower levels of physical activity. No association
3
ARTICLE IN PRESS G Model
PCD-1081; No. of Pages7
P. Naseri et al. Primary Care Diabetes xxx (xxxx) xxx–xxx
Table 2
Distribution of levels of physical activity based on sex and age groups.
Women
Age Groups ≤ 40 P-value > 40 P-value
Phase 2 Phase 3 Phase 4 Phase 5 Phase 2 Phase 3 Phase 4 Phase 5
Physical activity
<0.001 0.370
Low 365 (31.9) 377 (33.0) 274 (24.0) 268 (23.4) 325 (28.4) 336 (29.3) 291 (25.4) 337 (29.4)
Moderate 643 (56.2) 522 (45.6) 638 (55.8) 674 (58.9) 675 (59.0) 565 (49.3) 642 (56.1) 685 (59.8)
High 136 (11.9) 245 (21.4) 232 (20.3) 202 (17.7) 145 (12.7) 244 (21.3) 212 (18.5) 123 (10.7)
Men
Phase 2 Phase 3 Phase 4 Phase 5 Phase 2 Phase 3 Phase 4 Phase 5
Physical activity
0.001 <0.001
Low 369 (44.1) 374 (44.7) 378 (45.2) 323 (38.6) 343 (44.0) 275 (35.3) 303 (38.9) 237 (30.4)
Moderate 267 (31.9) 262 (31.3) 237 (28.3) 258 (30.8) 269 (34.5) 286 (36.7) 276 (35.4) 328 (42.1)
High 201 (24.0) 201 (24.0) 222 (26.5) 256 (30.6) 167 (21.4) 218 (28.0) 200 (25.7) 214 (7.5)
*P value was obtained using Generalized Estimating Equation (GEE) approach.
Table 3
Hazard ratios for development (incidence) of diabetes based on sex and age groups during study period.
Age groups Men Women
≤ 40 >40 ≤ 40 >40
HR (95% CI) P-value HR (95% CI) P-value HR (95% CI) P-value HR (95% CI) P-value
Model 1 Physical activity
Low (Ref) (Ref) (Ref) (Ref) (Ref) (Ref) (Ref) (Ref)
Moderate 0.90 (0.53−1.52) 0.70 0.91 (0.66−1.25) 0.58 0.96 (0.64−1.43) 0.85 0.78 (0.60−1.01) 0.06
High 1.13 (0.68−1.08) 0.63 0.92 (0.64−1.33) 0.68 0.94 (0.55−1.60) 0.82 0.59 (0.40−0.88) 0.01
Model 2 Physical activity
Low (Ref) (Ref) (Ref) (Ref) (Ref) (Ref) (Ref) (Ref)
Moderate 0.87 (0.52−1.46) 0.60 0.88 (0.64−1.02) 0.43 0.93 (0.62−1.40) 0.76 0.80 (0.62−1.03) 0.09
High 1.02 (0.61−1.69) 0.92 0.90 (0.63−1.30) 0.60 0.92 (0.54−1.57) 0.78 0.63 (0.42−0.93) 0.02
Model 3 Physical activity
Low (Ref) (Ref) (Ref) (Ref) (Ref) (Ref) (Ref) (Ref)
Moderate 0.88 (0.52−1.48) 0.64 0.86 (0.62−1.18) 0.36 0.93 (0.62−1.39) 0.73 0.80 (0.61−1.03) 0.09
High 1.02 (0.61−1.70) 0.91 0.89 (0.62−1.27) 0.53 0.92 (0.54−1.57) 0.77 0.63 (0.43−0.94) 0.02
Model 4 Physical activity
Low (Ref) (Ref) (Ref) (Ref) (Ref) (Ref) (Ref) (Ref)
Moderate 0.99 (0.57−1.72) 0.99 0.88 (0.64−1.21) 0.46 0.82 (0.54−1.23) 0.34 0.80 (0.62−1.04) 0.10
High 1.21 (0.72−2.03) 0.47 0.92 (0.63−1.34) 0.68 0.83 (0.48−1.42) 0.50 0.64 (0.43−0.95) 0.02
Hazard ratios (HR, 95% CI) for incidence of diabetes calculated using extended Cox survival analysis.
Model 1: Physical activity effect, unadjusted HR (95% CI) for diabetes incidence. Model 2: Physical activity effect, adjusted for job status and education. Model 3: Physical
activity effect, adjusted for job status, education and smoking.
Model 4: Physical activity effect, adjusted for job status, education, smoking and BMI.
was observed between physical activity changes and the incidence
of T2DM in men and younger women.
The present study’s findings showed that many participants did
not meet the recommended level of physical activity (at least 150
minutes of moderate-intensity aerobic physical activity or ≥ 600
MET/min throughout the week) in their daily lives. This is consistent with a national study in Iran that revealed low physical
activity levels among at least 40% of Iranian adults [17]. Our results
showed that insufficient physical activity is more frequent in men
than women in the TLGS population; though, physical inactivity is
considerably higher in women than men in most countries, including Iran [33]. Although a large percentage of the TLGS population
is still inactive, the prevalence of low physical activity showed a
decreasing trend through follow-ups [34]; this contrasts the trend
at the national level, where insufficient physical activity has been
increased [35].
In the present study, the more active adults are generally at
lower risk for developing diabetes, which supports the findings of
previous studies [5,25]. A meta-analysis provided strong evidence
of an inverse dose-response relationship between physical activity and diabetes [36]. Several mechanisms could explain the link
between physical activity and diabetes. The indirect effect of physical activity includes improving the energy balance and preventing
obesity, hypertension, and dyslipidemia. However, physical activity directly enhances glucose homeostasis and insulin sensitivity.
In addition, long-term physical activity leads to positive changes in
the structure and function of skeletal muscles, which may confer
diabetes prevention. Therefore, regular physical activity is needed
to benefit from both short- and long-term protective effects of
physical activity [37]. As a lifestyle component and as a result of
aging, physical activity could be changed [28]. However, few studies have investigated the longitudinal effects of different physical
activity levels on the risk of T2DM. These studies were conducted on
British men in their elderly period, American employees, and Chinese adults [38,29]; they showed an adverse association between
time-varying physical activity and incidence of T2DM, which confirms the findings of the present study.
Although the current results indicated a significant adverse
association between physical activity and the incidence of T2DM,
further sex-specific analysis revealed an age-dependent causal
effect between variables only in women. In our study, the nonsignificant association between physical activity levels and T2DM
in men is consistent with two previous studies [39,40] and in
contrast with others that showed protective effects of physical
4
ARTICLE IN PRESS G Model
PCD-1081; No. of Pages7
P. Naseri et al. Primary Care Diabetes xxx (xxxx) xxx–xxx
Fig. 2. The cumulative hazard function of diabetes for all three physical activity
levels.
activity on T2DM in male participants [41,42]. There are different
explanations to justify the observed gender differences in the relationship between physical activity and the incidence of diabetes.
One explanation might be the variety in domains and intensity
of physical activity. The physical activity domain is an important
dimension thatis given less attention. Previous studies showed that
work-related physical activity constitutes the largest proportion
of the total weekly physical activity in the Iranian population. On
the other hand, leisure time physical activity — a domain largely
responsible for the glucoregulatory benefits of physical activity
[43] — is decreasing in Iran [44]. Moreover, considering the negative association between work-hours and the amount of LTPA [45],
the more work-related physical activity than LTPA in Iranian men
might be the reason for less obvious benefits from physical activity
in this gender. Intensity of physical activity is another important
factor. The intensity threshold, which is required for diabetes prevention might also be sex- specific [39]. Studies show that higher
intensity of physical activity is needed to reduce risk of diabetes in
middle aged men and women, while, low intensity physical activity
might be effective in elderly individuals [46].
There is also gender differences in biological hormones through
the lifespan. Sexhormones significantly impact energymetabolism,
fat distribution, and vascular function; as women hit menopause,
the protective effects of feminine hormones on adiposity and
insulin resistance are eliminated, causing weight gain [47]. The
weight gain in postmenopausal women could itself be a motivation to increase physical activity. On the other hand, available data
suggests a stronger protective effect of increased physical activity against T2DM in individuals at higher risk of diabetes, such as
obese persons [48]. This is exactly the case for the stronger effects
of physical activity on DM in older women in the current study.
Furthermore, a part of the risk reduction among older women
in the more active group may result from other behavioral factors. Healthy behaviors such as less smoking and a healthier diet
are more prevalent in older women in the TLGS population [49].
Based on the findings of the previous study, up to the age of 45,
women have more sedentary behaviors than men; after this age,
the amount of sedentary behaviors becomes equal between the
two genders [44]. The non-significant effect of physical activity in
men and young women in the current study could be justified by
the effects of psychosocial factors. Socio-economic stressors and
how individuals respond to them are gender- and culture-related
and dynamic through the lifespan. In general, young women are
more vulnerable to psychosocial pressure [50]. As they age, Iranian women encounter fewer work-related stressors, while men
-as the main Breadwinner of most Iranian families- are still under
economic pressure even in late adulthood.
The two strengths of our prospective study as the first report
in the Middle East population are that we found physical activity
levels as a time-varying variable that changes through follow up
period and thatthe sample size of men and women in our study was
large. However, there were some limitations in our study. Firstly,
we are unable to extrapolate our findings to the entire population,
especially rural areas. Secondly, in our study, subjective methods were used to measure physical activity, which could impact
the accuracy of physical activity measurements. Thirdly, sedentary
behaviors which could mediate the effect of physical activity were
not measured objectively in the current study. Finally, diet data as
a potential predictor of diabetes was not available.
5. Conclusion
Based on the present study’s findings, it could be concluded
that there is an inverse association between physical activity and
incidence of T2DM in middle-aged women, with no such an association in younger women and men. Therefore, it highlights the
important role of community-based programs to improve physical activity in decreasing the risk of diabetes in middle-aged and
elderly individuals.
Ethics approval and consent to participate
This study was approved by the research ethics committee ofthe
Research Institute for Endocrine Sciences (RIES), Shahid Beheshti
University of Medical Sciences, Tehran, Iran (the ethics approval
code: IR.SBMU.ENDOCRINE.REC.1396.519). Prior to data collection,
both children and parents were informed about the study procedure and its aims and if the child and parent agreed to participate
in the study, parents were asked to sign a written consent form.
Consent for publication
Not applicable.
Availability of data and materials
The datasets used and/or analyzed during the current study are
available from the corresponding authors on reasonable request.
Code availability
Not applicable.
Competing interests
The authors declare that they have no competing interests.
Funding
There are no funding resources for this research.
Authors’ contributions
PN, PA and HM-A designed the study. AG participated in acquisition of data. PN carried out the statistical analysis. PA, PN, HM-A,
and SA contributed to interpretation of data. PA, PN, HM-A, and NSV
drafted the manuscript. FA, PA, AG and SA supervised and revised
the manuscript.All authors read and approved the final manuscript.
5
ARTICLE IN PRESS G Model
PCD-1081; No. of Pages7
P. Naseri et al. Primary Care Diabetes xxx (xxxx) xxx–xxx
Acknowledgement
Authors would like to express their appreciation to all participants who made this study possible and wish to thank Niloofar
Shiva for her thoughtful editing of the manuscript.
References
[1] N. Al Busaidi, P. Shanmugam, D. Manoharan, Diabetes in the Middle East: government health care policies and strategies that address the growing diabetes
prevalence in the Middle East, Curr. Diab. Rep. 19 (2) (2019) 8.
[2] L. Shi, X.-O. Shu, H. Li, H. Cai, Q. Liu, W. Zheng, Y.-B. Xiang, R. Villegas, Physical
activity, smoking, and alcohol consumption in association with incidence of
type 2 diabetes among middle-aged and elderly Chinese men, PLoS One 8 (11)
(2013), e77919.
[3] J. Tuomilehto, J. Lindström, J.G. Eriksson, T.T. Valle, H. Hämäläinen, P. IlanneParikka, S. Keinänen-Kiukaanniemi, M. Laakso, A. Louheranta, M. Rastas,
Prevention of type 2 diabetes mellitus by changes in lifestyle among subjects
with impaired glucose tolerance, N. Engl. J. Med. 344 (18) (2001) 1343–1350.
[4] M. Javanbakht, A. Mashayekhi, H.R. Baradaran, A. Haghdoost, A. Afshin, Projection of diabetes population size and associated economic burden through
2030 in Iran: evidence from micro-simulation Markov model and Bayesian
meta-analysis, PLoS One 10 (7) (2015), e0132505.
[5] A.D. Smith, A. Crippa, J. Woodcock, S. Brage, Physical Activity and Incident Type
2 Diabetes Mellitus: A Systematic Review and Dose–Response Meta-Analysis
of Prospective Cohort Studies, Springer, 2016.
[6] M.B. Davidson, The effect of aging on carbohydrate metabolism: a review of the
English literature and a practical approach to the diagnosis of diabetes mellitus
in the elderly, Metabolism 28 (6) (1979) 688–705.
[7] G. Roglic, WHO Global report on diabetes: a summary, Int. J. Noncommun. Dis.
1 (1) (2016) 3.
[8] A.H. Abdelhafiz, A.J. Sinclair, Diabetes in the elderly, Medicine 47 (2) (2019)
119–122.
[9] M. Noroozian, The elderly population in Iran: an ever growing concern in the
health system, Iran. J. Psychiatry Behav. Sci. 6 (2) (2012) 1.
[10] F. Sharifi, H. Fakhrzadeh, M. Vannaghani, S.M. Arzaghi, M.A. Khoei, F. Farzadfar,
P.T. Tanjani, Prevalence of dementia and associated factors among older adults
in Iran: national elderly health survey (NEHS), Arch Iran Med (AIM) 19 (12)
(2016).
[11] Z.H. Nejhad, H.M. Vardanjani, F. Abolhasani, M. Hadipour, K. Sheikhzadeh, Relative effect of socio-economic status on the health-related quality of life in type
2 diabetic patients in Iran, Diabetes Metab. Syndr. Clin. Res. Rev. 7 (4) (2013)
187–190.
[12] F.B. Hu, J.E. Manson, M.J. Stampfer, G. Colditz, S. Liu, C.G. Solomon, W.C. Willett,
Diet, lifestyle, and the risk of type 2 diabetes mellitus in women, N. Engl. J. Med.
345 (11) (2001) 790–797.
[13] F.B. Hu, Physical activity, sedentary behaviors, and obesity, in: Obesity Epidemiology, Oxford University Press, Inc, New York (NY), 2008, pp. 301–319.
[14] J. Henriksson, Influence of exercise on insulin sensitivity, J. Cardiovasc. Risk 2
(4) (1995) 303–309.
[15] M.F. Saad, W.C. Knowler, D.J. Pettitt, R.G. Nelson, M.A. Charles, P.H. Bennett,
A two-step model for development of non-insulin-dependent diabetes, Am. J.
Med. 90 (1) (1991) 229–235.
[16] P.C. Hallal, L.B. Andersen, F.C. Bull, R. Guthold, W. Haskell, U. Ekelund, Group
LPASW, Global physical activity levels: surveillance progress, pitfalls, and
prospects, Lancet 380 (9838) (2012) 247–257.
[17] A. Esteghamati, O. Khalilzadeh, A. Rashidi, M. Kamgar, A. Meysamie, M. Abbasi,
Physical activity in Iran: results of the third national surveillance of risk factors of non-communicable diseases (SuRFNCD-2007), J. Phys. Act. Health 8 (1)
(2011) 27–35.
[18] A. Didarloo, D. Shojaeizadeh,H.E.Ardebili, S.Niknami, E.Hajizadeh,M.Alizadeh,
Factors influencing physical activity behavior among Iranian women with type
2 diabetes using the extended theory of reasoned action, Diabetes Metab. J. 35
(5) (2011) 513–522.
[19] A.C. King, C. Castro, S. Wilcox, A.A. Eyler, J.F. Sallis, R.C. Brownson, Personal
and environmental factors associated with physical inactivity among different racial–ethnic groups of US middle-aged and older-aged women, Health
Psychol. 19 (4) (2000) 354.
[20] Health UDo, Services H, Physical Activity Fundamental to Preventing Disease,
US Department of Health and Human Services, Office of the Assistant Secretary
for Planning and Evaluation 2002, 2002.
[21] M.A. Davis, J.M. Neuhaus, D.J. Moritz, D. Lein, J.D. Barclay, S.P. Murphy, Health
behaviors and survival among middle aged and older men and women in the
NHANES I epidemiologic follow-up study, Prev. Med. 23 (3) (1994) 369–376.
[22] T. Rakhshani, D. Shojaiezadeh, K.B. Lankarani, F. Rakhshani, M.H. Kaveh, N. Zare,
The association of health-promoting lifestyle wi