Experimental Gerontology 64 (2015) 70–77

Contents lists available at ScienceDirect

Experimental Gerontology journal homepage: www.elsevier.com/locate/expgero

Diabetes mellitus and its association with central obesity and disability among older adults: A global perspective Stefanos Tyrovolas a,b, Ai Koyanagi a,b, Noe Garin c,d, Beatriz Olaya a,b, Jose Luis Ayuso-Mateos b,e, Marta Miret b,e, Somnath Chatterji f, Beata Tobiasz-Adamczyk g, Seppo Koskinen h, Matilde Leonardi i, Josep Maria Haro a,b,⁎ a

Parc Sanitari Sant Joan de Déu, Universitat de Barcelona, Fundació Sant Joan de Déu, Dr. Antoni Pujades, 42, 08830 Sant Boi de Llobregat, Barcelona, Spain Instituto de Salud Carlos III, Centro de Investigación Biomédica en Red de Salud Mental, CIBERSAM, Monforte de Lemos 3-5, Pabellón 11, 28029 Madrid, Spain Pharmacy Department, Hospital de la Santa Creu i Sant Pau, Institut d'Investigacions Biomèdiques Sant Pau (IIB Sant Pau), Sant Antoni Maria Claret 167, 08025 Barcelona, Spain d Universitat Autònoma de Barcelona, Bellaterra 08193 Cerdanyola del Vallès, Spain e Department of Psychiatry, Universidad Autónoma de Madrid, Instituto de Investigación Sanitaria Princesa (IP), Hospital Universitario la Princesa, Madrid, Spain f Department of Health Statistics and Information Systems, World Health Organization, Geneva, Switzerland g Department of Medical Sociology, Jagiellonian University Medical College, Krakow, Poland h National Institute for Health and Welfare, Helsinki, Finland i Neurology, Public Health and Disability Unit, Neurological Institute “Carlo Besta” Foundation IRCCS (Istituto di ricovero e cura a carattere scientifico), Milan, Italy b c

a r t i c l e

i n f o

Article history: Received 1 December 2014 Received in revised form 11 February 2015 Accepted 13 February 2015 Available online 16 February 2015 Section Editor: Holly M. Brown-Borg Keywords: Diabetes mellitus Disability Waist-to-height ratio Waist circumference Obesity Older adults

a b s t r a c t The aim of the study was to evaluate the association between various factors and diabetes type II (DM) with a particular emphasis on indicators of central obesity, and to compare the effect of DM on disability among elder populations (≥50 years old) in nine countries. Data were available for 52,946 people aged ≥18 years who participated in the WHO Study on global AGEing and adult health and the Collaborative Research on Ageing in Europe studies conducted between 2007 and 2012. DM was defined as self-report of physician diagnosis. Height, weight, and waist circumference were measured. Disability status was assessed with the WHODAS II questionnaire. The overall prevalence of DM was 7.9% and ranged from 3.8% (Ghana) to 17.6% (Mexico). A 10 cm increase in waist circumference and waist-to-height ratio of N 0.5 were associated with a significant 1.26 (India) to 1.77 (Finland), and 1.68 (China, Spain) to 5.40 (Finland) times higher odds for DM respectively. No significant associations were observed in Mexico and South Africa. DM was associated with significantly higher disability status in all countries except Mexico in the model adjusted for demographics and smoking. The inclusion of chronic conditions associated with diabetes in the model attenuated the coefficients in varying degrees depending on the country. A considerable proportion of the studied older population had DM. Central obesity may be a key factor for the prevention of DM among older populations globally. Prevention of DM especially among the older population globally may contribute to reducing the burden of disability. © 2015 Elsevier Inc. All rights reserved.

1. Introduction According to the World Health Organization, approximately 347 million people have diabetes mellitus (DM) worldwide and 80% of diabetics live in low- and middle-income countries (WHO, 2013). Results from the World Health Survey (WHS) indicated that the global prevalence of DM is almost 4% in populations 18 years or older (Liu et al., 2012). Prevalence figures of 20% in males and almost 17% in females among adults aged ≥65 years have been reported in the US (SHIELD study), and studies from Europe have reported even higher figures among those in this age group (nearly 30%) (Bays et al., 2007; Tyrovolas et al., 2009). These figures are expected to increase at ⁎ Corresponding author at: Parc Sanitari Sant Joan de Déu, Fundació Sant Joan de Déu, CIBERSAM, Dr. Antoni Pujadas, 42, 08830 Sant Boi de Llobregat, Barcelona, Spain. E-mail address: [email protected] (J.M. Haro).

http://dx.doi.org/10.1016/j.exger.2015.02.010 0531-5565/© 2015 Elsevier Inc. All rights reserved.

alarming rates by the year 2025, because of population growth, aging, increase in unhealthy lifestyle patterns (i.e., sedentary life, unhealthy nutrition, etc.), and obesity. DM has been associated with cardiovascular disease (CVD), as well as with blindness, micro- and macro-vascular disease, kidney failure and stroke events (Toutouzas et al., 2005). In the past years, multiple studies have shown that sociodemographic, bio-clinical, and lifestyle factors (i.e., obesity, physical activity, education, etc.) are related to the development of DM in varying degrees (Wood, 2001; Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus, 1997; Pitsavos et al., 2007). Furthermore, global studies have reported that increased body mass index (BMI) is one of the factors most strongly associated with DM (Liu et al., 2012). However, studies have also shown that BMI may be a poorer predictor of mortality and metabolic diseases such as DM, compared to waist circumference or waist-to-height ratio which are considered to be a closer reflection of central obesity (Qiao and

S. Tyrovolas et al. / Experimental Gerontology 64 (2015) 70–77

Nyamdorj, 2010; Hadaegh et al., 2006; Cai et al, 2013; Schneider et al., 2010). Central obesity (known as accumulated visceral adipose tissue) is strongly related to increased mortality and various clinical conditions (such as insulin resistance, dyslipidemia, and hypertension) (Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus, 1997). Recently, one study indicated that the prevalence of central obesity in the older European populations was quite high, close to 40% in males and 60% in females (Saaristo et al., 2008). To date, global epidemiological data on the role of waist-to-height ratio on DM among older populations are scarce (Liu et al., 2012; Yusuf et al, 2004; Espelt et al., 2013). Furthermore, a few studies on diabetes epidemiology in low- and middle-income countries do exist (Tao et al., 2013; Amoah et al, 2002; Assari, 2014; Assari et al., 2014) but the majority of these studies are small, and do not always include older individuals, or do not follow a common research protocol. Also, despite the fact that DM is related to major causes of disability (Anton et al., 2013), until now, there has been little research on the cross-country differences of the effects of DM on disability especially among older populations (Assari et al, 2014). This is an important research gap as disability is known to be associated with various health deficiencies, quality of life as well as institutionalization, and use of health services (Guralnik et al., 1994; Branch and Jette, 1982). The Collaborative Research on Ageing in Europe (COURAGE) and WHO Study on global AGEing and adult health (SAGE) studies, from which our data was derived, are among the few large populationbased nationally-representative health studies (e.g., WHS study) that apply standard design and survey procedures across all survey populations. Given the rapid increase of DM globally, the metabolic complexity of the disease and its related co-morbidities, as well as the lack of global evidence on the association between DM and disability among older populations, the aim of the present study was to evaluate the association between central obesity including waist circumference, waist-toheight ratio and DM, as well as to assess the effect of DM on disability among older adults (≥ 50 years old) in nine high-, middle- and lowincome countries from Asia, Africa, Europe, and Latin America.

2. Research design and methods 2.1. The SAGE and COURAGE surveys The SAGE survey was conducted between 2007 and 2010 in China, Ghana, India, Mexico, Russia, and South Africa, and the COURAGE survey was conducted between 2011 and 2012 in Finland, Poland, and Spain. The two surveys followed similar methodologies and used the same standardized questionnaire to collect information on health and wellbeing among adult non-institutionalized populations. Both studies were population-based household surveys including adults ≥18 years of age, with oversampling of those 50 years or older. Nationallyrepresentative samples, with no replacement, were selected by multistage clustered sampling. The response rate ranged from 51% (Mexico) to 93% (China), and 53% (Finland) to 70% (Spain) in the SAGE and COURAGE studies respectively. Trained interviewers collected data through face-to-face interviews. For those who were unable to participate in the survey due to limited cognitive function, information was obtained through a proxy respondent using a shorter questionnaire. These participants were excluded from the current analysis as most information pertaining to the current analysis were not collected. Sampling weights were generated to adjust for the population structure reported by the United Nations Statistical Division and the census of the National Institute of Statistics for the SAGE and COURAGE respectively. The research review board of each location and the WHO Ethical Review Committee provided ethical approval to conduct the study. Informed consent was obtained from all participants. Further details of the two surveys are provided elsewhere (Basu and Millet, 2013; Perales et al., 2014).

71

2.1.1. Clinical and anthropometric measures Height and weight were measured with the use of a stadiometer and a routinely calibrated electronic weighting scale respectively. Waist circumference was measured using an inelastic tape at the navel level, and recorded to the nearest 0.1 cm. Waist-to-height ratio was calculating by dividing the waist circumference (cm) by height (cm), and was dichotomized as ≤0.5 and N0.5 (Browning et al., 2010). BMI was calculated as weight in kilograms divided by height in meters squared. BMI was categorized as the following: b 18.5 kg/m2 (underweight), 18.5–24.9 kg/m2 (normal weight), 25.0–29.9 kg/m2 (overweight), 30.0–34.9 kg/m2 (obesity class I), and ≥35.0 kg/m2 (obesity class II+). Blood pressure was measured 3 times in the SAGE survey and 2 times in COURAGE survey, with a less than one-minute interval using standard protocols. Mean systolic and diastolic pressure were obtained by calculating the mean of all the available measurements. Hypertension was a dichotomous variable and was defined as at least either one of the following: mean systolic blood pressure ≥ 140 mm Hg, mean diastolic blood pressure ≥ 90 mm Hg, and self-reported medical diagnosis of hypertension. The diagnosis of angina was based on the algorithms of the Rose questionnaire (Rose, 1962) and/or self-reported diagnosis. Depression was based on the algorithms for DSM-IV major depressive disorder and/or self-reported diagnosis. Cataract was defined as having cloudy or blurry vision and vision problems with light, such as glare from bright lights, or halos around lights in the past 12 months and self-reported diagnosis in the past five years. The diagnosis of arthritis, DM, and stroke were based on self-reported diagnosis. 2.1.2. Socio-demographic, dietary and other lifestyle characteristics Education was based on the highest level of education completed and was categorized as primary or less, secondary, and tertiary or higher. Wealth quintiles were created based on country-specific income. Respondents were also categorized as living in either urban or rural areas. Level of physical activity was assessed with the Global Physical Activity Questionnaire using conventional cut-offs and categorized as low, moderate, and high (http://www.who.int/chp/steps/GPAQ/en/). Information on smoking habits was obtained with two questions: “Have you ever smoked tobacco or used smokeless tobacco?” and “Do you currently use (smoke, sniff or chew) any tobacco products such as cigarettes, cigars, pipes, chewing tobacco or snuff?” Those who answered ‘no’ to the first question were considered ‘never’ smokers, and those who answered ‘yes’ to both questions were considered ‘current’ smokers. Those who answered ‘yes’ to the first question but ‘no’ to the second were categorized as having ‘quit’. 2.1.3. Functioning and disability The 12-item validated version of the World Health Organization Disability Assessment Schedule 2.0 (WHODAS 2.0) was used to assess functioning and disability (Ustün et al., 2010). This scale ranged from 0 (no disability) to 100 (maximum disability). 2.2. Statistical analysis Data were available for 52,946 individuals. After the exclusion of those aged b50 years, the final analytical sample size was 42,116 (China 13,175, Finland 1452, Ghana 4305, India 6560, Mexico 2313, Poland 2910, Russia 3938, South Africa 3838, Spain 3625). Countrywise analyses were conducted to account for the heterogeneity between countries. Age and sex adjusted, and crude prevalence of DM was calculated. In addition, the age and sex adjusted prevalence of the country-wise highest decile of the WHODAS II score by DM status was calculated. The highest decile of the WHODAS score was used as this cut-off has been suggested to represent significant clinical disability (Andrews et al., 2009). All age and sex adjusted prevalence were calculated using the United Nation population pyramids for the year 2010 (http://esa.un.org/wpp/Excel-Data/population.htm) as the standard population.

72

S. Tyrovolas et al. / Experimental Gerontology 64 (2015) 70–77

The baseline characteristics were compared between non-diabetics and diabetics. Continuous variables were presented as mean ± SD and categorical variables as proportions. Comparisons of continuous and categorical variables by the presence of DM within countries were performed using Student's t-tests and chi2 tests respectively. Two multiple logistic regression models were constructed to assess the association between two measures of central obesity (waist circumference and waist-to-height ratio) and DM as the outcome. Separate models were constructed to account for collinearity between the two measures. The first model included waist circumference and the second model included waist-to-height ratio while both models adjusted for sex, age, education, wealth, location, physical activity, smoking, angina, arthritis, cataract, depression, hypertension, and stroke. In addition, since it is possible that the inclusion of different control variables in the model might have affected the association between central obesity and diabetes in different ways, we also conducted a hierarchical analysis that examined the effect of including different covariates in the model sequentially by comparing the waist-to-height odds ratios in the univariate and subsequent models. The models that were constructed were the following: Model 1 — univariate; Model 2 — adjusted for demographic factors (sex, age, education, wealth, settlement type); Model 3 — adjusted for demographic factors and lifestyle factors (physical activity and smoking); Model 4 — adjusted for demographic, lifestyle factors, and chronic conditions (angina, arthritis, cataract, depression, hypertension, and stroke). Next, a linear regression model was constructed to assess the association between DM and disability (WHODAS II score as a continuous variable ranging from 0 to 100) as the outcome. A base model which adjusted for sex, age, education, wealth, location, and smoking was constructed. To the base model, angina, arthritis, cataract, depression, stroke, and BMI were included individually in separate models to assess their individual effect as mediators in the association between DM and disability. A final model included all six conditions simultaneously in the model. BMI was used in this analysis rather than the other two measures of central obesity as exploratory analysis showed that only very high BMI values (i.e., ≥35 kg/m2) were associated with disability. Age was included in the models as a continuous variable. The selection of the control variables was based on past literature (Tyrovolas et al., 2009; Gregg et al., 2000). The sample weighting and the complex

study design were taken into account in all analyses to generate nationally-representative estimates. Results from logistic and linear regression models are presented as odds ratios and b-coefficients respectively, and their 95% confidence intervals (CIs). All reported p-values were based on two-sided tests. The level of statistical significance was set at p b 0.05. Stata software, version 12.1 was used for all analyses (Stata Corp. LP, College Station, Texas). 3. Results The crude overall prevalence of diabetes was 7.9%, with figures ranging from 3.8% (Ghana) to 17.6% (Mexico) (Fig. 1). Little difference was observed between the sex and age adjusted, and the crude prevalence. Overall, individuals ≥65 years old had higher prevalence of DM compared to their younger counterparts (10.5% vs. 6.3%, p ≤ 0.001). Table 1 presents the unadjusted association between sociodemographic, lifestyle, clinical, anthropometric, and psychological characteristics among diabetics and non-diabetics. The urban inhabitants had significantly higher prevalence of DM in China, Ghana, India, and South Africa. Across most countries, compared with non-diabetics, the diabetics were significantly more likely to have angina, arthritis, cataract, hypertension, and stroke. They also had significantly higher BMI, waist-to-height ratio, and waist circumference. Table 2 illustrates the association between waist-to-height ratio and DM estimated by multivariate logistic regression. Αfter adjusting for various confounders, weight-to-height ratio of N 0.5 was associated with significantly higher odds for DM in all countries except Mexico and South Africa. The significant odds ratios ranged from 1.68 in China (95% CI 1.35–2.09) and Spain (95% CI 1.03–2.74) to 5.40 (95% CI 2.06– 14.12) in Finland. Similarly, significant associations between waist circumference and DM were observed in all countries except Mexico and South Africa. The odds ratio (95% CI) for DM associated with a 10 cm increase in waist circumference was: China 1.33 (1.23–1.43); Finland 1.77 (1.54–2.02); Ghana 1.28 (1.12–1.46); India 1.26 (1.16–1.37); Mexico 0.82 (0.65–1.03); Poland 1.68 (1.50–1.88); Russia 1.40 (1.25–1.57); South Africa 1.08 (0.99–1.18); Spain 1.30 (1.16–1.44) (data available only in text). The results of the hierarchical models are presented in Table A1 (Appendix A). There was some between-country variation in the magnitude of the change in ORs after inclusion of the various blocks

Age and sex adjusted

Crude

25

20

15 % 10

5

0

China

Finland

Ghana

India

Age and sex adjusted

6.5

11.5

3.9

6.9

Crude

6.6

12.9

3.8

6.9

Mexico

Poland

Russia

S. Africa

Spain

18.2

13.1

6.6

9.3

14.4

17.6

13.6

7

9.2

16.3

Fig. 1. Age and sex adjusted, and crude prevalence of diabetes mellitus by country. Abbreviation: S. Africa, South Africa. All values are calculated based on weighted sample. Age and sex adjusted prevalence was calculated using the United Nation population pyramids for the year 2010 as the standard population. Bars denote upper value of 95% confidence interval.

Table 1 Baseline characteristics of individuals 50 years or older with and without diabetes mellitus by country.

Sex Age (years)

China

Diabetes

No

Male p-Value⁎ Mean (SD) p-Value⁎

50.2 0.009 62.3 (16.4) b0.001 4.3 32.2 63.5 b0.001 16.8 18.4 20.3 23.3 21.3 b0.001 45.6 b0.001 45.3 27.0 27.7 b0.001 63.3 30.2 6.4 b0.001 8.8 b0.001 21.4 b0.001 2.5 b0.001 1.3 0.992 59.5 b0.001 2.8 b0.001 0.9 0.105 67.1 b0.001 0.53 (0.13) b0.001 84.5 (19.3) b0.001

Education

≥Tertiary Secondary ≤Primary p-Value⁎

Wealth

Poorest Poorer Middle Rich Richest p-Value⁎ Urban p-Value⁎

Location Physical activity

Smoking

Angina

High Moderate Low p-Value⁎ Never Current Quit p-Value⁎ Yes p-Value⁎ Yes p-Value⁎ Yes p-Value⁎

Arthritis Cataract Depression Hypertension Stoke 2

BMI (kg/m ) Waist-to-height ratio

Waist circumference (cm)

Yes p-Value⁎ Yes p-Value⁎ Yes p-Value⁎ ≥35 p-Value⁎ N0.5 p-Value⁎ Mean (SD) p-Value⁎ Mean (SD) p-Value⁎

Finland Yes 43.2 65.3 (15.6) 8.1 40.5 51.4 7.0 14.2 21.8 26.1 30.8 74.5 31.3 32.4 36.2 75.9 15.8 8.3 17.5 30.0 7.4 1.3 76.2 6.0 1.6 82.1 0.56 (0.13) 89.6 (20.5)

No 45.5 0.160 64.5 (12.2) b0.001 27.9 56.3 15.8 b0.001 21.7 26.7 19.4 17.9 14.4 0.001 75.9 0.690 43.2 31.8 25.0 0.008 35.9 17.6 46.6 0.389 9.8 b0.001 43.6 0.001 3.0 0.087 14.6 0.073 64.0 b0.001 4.1 0.007 6.6 b0.001 78.9 b0.001 0.56 (0.09) b0.001 93.8 (16.1) b0.001

Ghana Yes 50.7 67.8 (11.5) 12.5 60.6 26.9 32.6 30.8 19.9 10.6 6.2 74.5 33.0 30.2 36.8 33.2 14.8 51.9 25.6 55.9 5.2 20.3 82.8 8.4 17.7 97.0 0.63 (0.10) 105.3 (17.0)

No 52.6 0.061 64.3 (19.9) 0.683 3.2 20.8 76.0 b0.001 18.6 19.5 20.8 20.3 20.7 b0.001 40.1 b0.001 62.8 12.3 25.0 b0.001 75.0 10.9 14.1 0.242 12.7 0.835 13.8 0.634 3.2 0.244 7.8 0.127 58.9 b0.001 2.3 b0.001 3.5 0.004 54.8 b0.001 0.52 (0.15) b0.001 84.5 (23.8) b0.001

India Yes 44.5 64.7 (17.9) 13.4 30.0 56.6 10.7 9.4 13.6 27.4 39.0 65.5 36.4 19.2 44.4 78.6 6.1 15.3 13.4 15.3 5.1 11.6 78.8 13.7 8.4 82.4 0.58 (0.17) 93.1 (25.7)

No 50.3 0.001 61.5 (13.6) 0.701 4.5 17.7 77.8 b0.001 19.1 20.0 19.2 19.3 22.4 b0.001 27.5 0.001 52.6 22.6 24.8 0.468 44.8 50.9 4.3 b0.001 16.3 b0.001 17.3 b0.001 11.2 b0.001 15.8 0.242 35.5 b0.001 1.9 0.126 0.8 0.126 54.0 b0.001 0.51 (0.12) b0.001 80.5 (19.0) b0.001

Mexico Yes 60.5 61.8 (14.0) 14.1 33.0 53.0 5.2 13.1 13.2 24.3 44.2 48.4 47.9 24.9 27.2 51.4 38.4 10.1 25.6 30.5 19.9 18.7 65.2 3.1 1.8 77.9 0.55 (0.12) 88.5 (18.5)

No 47.2 0.684 62.3 (18.2) 0.283 7.6 11.0 81.4 0.355 15.6 27.1 16.3 16.8 24.2 0.131 76.8 0.145 38.8 23.3 37.9 0.644 61.8 19.4 18.8 0.704 5.2 0.015 9.0 0.849 2.6 0.010 15.8 0.336 59.8 0.076 3.4 0.024 8.3 0.167 95.9 0.446 0.62 (0.14) 0.475 96.9 (21.0) 0.383

Poland Yes 44.1 63.9 (16.6) 10.4 18.3 71.3 13.4 15.1 17.5 16.2 37.8 86.1 45.5 18.4 36.1 55.5 24.1 20.4 13.8 8.5 7.9 22.0 71.2 8.4 4.6 97.5 0.61 (0.13) 95.4 (20.7)

No 43.3 0.642 63.5 (12.7) b0.001 16.7 59.3 24.0 0.001 24.0 16.8 18.5 22.7 18.0 0.133 68.8 0.736 49.6 19.1 31.2 0.147 44.9 26.7 28.4 0.093 16.8 b0.001 29.8 b0.001 4.4 b0.001 13.3 0.530 63.9 b0.001 4.5 0.072 7.9 b0.001 77.4 b0.001 0.56 (0.11) b0.001 92.9 (18.0) b0.001

Russia Yes 44.8 68.3 (13.9) 9.3 59.2 31.5 30.1 18.9 17.8 20.4 12.8 69.9 42.1 23.0 34.9 42.5 22.3 35.2 34.0 42.3 10.7 14.7 91.3 7.5 21.6 95.3 0.63 (0.11) 103.2 (18.5)

No 39.7 0.081 63.8 (15.4) 0.038 18.3 74.4 7.4 0.706 16.2 19.8 18.6 20.4 24.9 0.178 72.4 0.440 57.8 16.1 26.1 0.265 68.7 22.5 8.8 b0.001 35.8 b0.001 29.3 0.031 6.5 b0.001 5.3 0.017 70.9 b0.001 4.8 0.606 10.5 b0.001 74.6 b0.001 0.56 (0.16) b0.001 91.6 (25.9) b0.001

Yes 29.2 65.9 (15.1) 18.5 72.5 9.1 14.9 16.2 25.9 24.7 18.4 77.4 56.6 12.1 31.2 81.1 7.3 11.5 57.6 42.2 15.1 9.8 87.0 5.7 27.4 94.1 0.63 (0.15) 101.7 (22.3)

South Africa

Spain

No

No

45.0 0.005 61.4 (18.3) 0.020 5.9 22.3 71.8 0.444 21.0 20.5 18.7 19.4 20.3 0.010 63.9 0.002 29.1 12.2 58.7 0.194 65.6 24.7 9.7 0.006 7.6 b0.001 22.7 b0.001 2.4 b0.001 4.5 0.150 77.4 b0.001 3.7 0.116 22.9 b0.001 77.3 0.427 0.58 (0.28) 0.002 91.3 (42.0) 0.002

Yes 33.0 63.4 (17.7) 5.2 27.5 67.4 11.6 14.8 17.6 25.5 30.4 78.9 20.4 13.4 66.2 79.0 14.7 6.3 21.9 44.9 6.3 7.7 89.7 7.0 39.2 80.9 0.62 (0.30) 97.9 (47.5)

45.5 0.040 65.6 (14.8) b0.001 12.0 26.6 61.4 b0.001 22.9 22.3 19.9 19.3 15.7 b0.001 84.2 0.248 31.8 37.4 30.9 0.003 52.5 20.7 26.8 0.052 6.3 b0.001 27.0 b0.001 4.9 b0.001 25.5 0.009 60.2 b0.001 3.8 0.001 7.2 b0.001 86.7 b0.001 0.59 (0.12) b0.001 95.8 (19.2) b0.001

Yes 50.6 69.4 (13.8) 4.7 20.4 75.0 19.5 32.0 22.2 16.3 9.9 82.1 24.5 39.9 35.6 57.8 15.0 27.2 15.8 37.3 10.6 31.5

S. Tyrovolas et al. / Experimental Gerontology 64 (2015) 70–77

Country

77.3 9.5 13.8 94.4 0.63 (0.12) 102.1 (19.3)

Abbreviation: BMI, body mass index. Data are in % unless otherwise stated. Percentages are proportion of those in that diabetes category who have that characteristic. All values are calculated based on weighted sample. ⁎ p-Values for the difference between those with and without diabetes mellitus within each country. p-Values are obtained by Student's t-tests and Chi2 tests for continuous and categorical variables respectively. 73

74

S. Tyrovolas et al. / Experimental Gerontology 64 (2015) 70–77

Table 2 Country-wise correlates of diabetes mellitus estimated by multivariate logistic regression.

Waist-to-height ratio N0.5 Angina Arthritis Cataract Depression Hypertension Stroke

China

Finland

Ghana

India

Mexico

Poland

Russia

S. Africa

Spain

1.68⁎⁎⁎ (1.35–2.09) 1.53⁎⁎ (1.15–2.03) 1.22 (0.99–1.52) 1.71⁎⁎

5.40⁎⁎⁎ (2.06–14.12) 2.08⁎⁎ (1.25–3.44) 1.27 (0.86–1.88) 1.43 (0.65–3.14) 1.73⁎

2.37⁎⁎⁎ (1.45–3.88) 1.22 (0.68–2.18) 0.90 (0.51–1.61) 1.00 (0.34–2.97) 1.21 (0.68–2.17) 2.14⁎⁎⁎ (1.41–3.25) 4.19⁎⁎⁎ (2.02–8.67)

2.14⁎⁎⁎ (1.47–3.12) 1.64⁎⁎ (1.14–2.37) 1.63⁎⁎ (1.13–2.35) 1.76⁎⁎

1.43 (0.30–6.74) 2.42 (0.97–6.03) 0.72 (0.36–1.45) 4.18⁎

4.45⁎⁎⁎ (2.08–9.50) 2.03⁎⁎⁎ (1.43–2.88) 1.27 (0.75–2.15) 2.15⁎⁎

1.15 (0.69–1.91) 3.30⁎⁎⁎ (1.78–6.13) 2.10⁎⁎ (1.33–3.31) 4.30⁎⁎⁎

(1.15–2.67) 1.35 (0.96–1.90) 2.54⁎⁎⁎ (1.63–3.95) 0.89 (0.38–2.09)

(1.14–15.26) 1.43 (0.69–2.95) 1.86⁎ (1.06–3.25) 1.90 (0.71–5.08)

3.47⁎⁎⁎ (2.03–5.94) 1.62⁎ (1.10–2.38) 1.15 (0.79–1.67) 1.61 (0.94–2.76) 1.03 (0.68–1.56) 3.92⁎⁎⁎ (2.48–6.21) 0.90 (0.45–1.78)

(1.29–3.60) 0.73 (0.30–1.79) 1.85⁎ (1.07–3.20) 0.78 (0.36–1.73)

(2.02–9.14) 1.13 (0.49–2.60) 1.74 (0.98–3.10) 0.91 (0.39–2.15)

1.68⁎ (1.03–2.74) 1.64⁎⁎ (1.17–2.29) 1.26 (0.94–1.70) 1.31 (0.87–1.97) 1.41⁎

(1.18–2.48) 0.80 (0.34–1.85) 1.94⁎⁎⁎ (1.52–2.48) 1.43⁎ (1.01–2.01)

(1.03–2.90) 1.73⁎ (1.05–2.87) 1.44 (0.71–2.91)

(1.04–1.91) 1.68⁎⁎⁎ (1.26–2.24) 1.62 (0.84–3.10)

Abbreviation: S. Africa, South Africa. Data are odds ratio (95% confidence interval). Model is mutually adjusted for all covariates in the table in addition to sex, age, education, wealth, location, physical activity, and smoking. ⁎ p b 0.05. ⁎⁎ p b 0.01. ⁎⁎⁎ p b 0.001.

of covariates. More specifically, in Finland and Ghana, the influence of the inclusion of demographic factors was larger than in other countries. In addition, in some countries, education (Finland, Ghana, India, Poland and Spain) and income (China, India and Russia) were more strongly associated with DM than others. The age and sex adjusted prevalence of the country-wise highest decile of the WHODAS II score by the presence of DM is illustrated in Fig. 2. The difference between diabetic and non-diabetic participants was particularly pronounced in Finland, Ghana, Poland, Russia, and Spain where diabetic individuals had much higher levels of disability. The association between DM and disability assessed by the WHODAS II estimated by multivariate linear regression is shown in Table 3. In the model adjusting for socio-demographics and smoking, diabetes was associated with significant 2.46 (95% CI 1.49–3.43) (China) to 6.56 (95% CI 2.67–10.45) (Ghana) higher scores on the WHODAS II scale with the exception of Mexico. The inclusion of individual conditions resulted in the attenuation of the coefficients in varying degrees in most countries. For example, when angina or arthritis was

Diabetes mellitus

No

included in the model, the association between DM and disability lost significance in India and South Africa. A loss of significance was also observed when cataract or depression was included in the model in India. The association between DM and disability remained significant even after the inclusion of all six conditions in China, Finland, Ghana, Poland, Russia, and Spain. 4. Discussion This study evaluated the prevalence of DM, and its association with various factors such as waist circumference, waist-to-height ratio, and disability among elderly individuals in nine low-, middle-, and highincome countries. A large difference in the prevalence of DM was found in older populations among different countries ranging from 3.8% (Ghana) to 17.6% (Mexico). Apart from Mexico, a high prevalence was also observed in Spain (16.3%), Poland (13.6%), and Finland (12.9%). Multivariate analysis revealed that central obesity (measured either by waist circumference or waist-to-height ratio) was significantly

Yes

30

25

20

% 15

10

5

0

China

Finland

Ghana

India

Mexico

Poland

Russia

South Africa

Spain

Fig. 2. Age and sex adjusted prevalence of the country-wise highest decile of WHODAS II scores by presence of diabetes mellitus. Bars denote upper value of 95% confidence interval. All values are calculated based on weighted sample. Age and sex adjusted prevalence was calculated using the United Nation population pyramids for the year 2010 as the standard population.

S. Tyrovolas et al. / Experimental Gerontology 64 (2015) 70–77

75

Table 3 Country-wise associations between diabetes mellitus and disability assessed by WHODAS II estimated by multivariate linear regression. China

Finland

Ghana

India

Mexico

Poland

Russia

S. Africa

Spain

Diabetes, cataract

2.46⁎⁎⁎ (1.49–3.43) 2.14⁎⁎⁎ (1.23–3.05) 2.28⁎⁎⁎ (1.34–3.21) 2.26⁎⁎⁎

5.54⁎⁎⁎ (2.96–8.11) 4.76⁎⁎⁎ (2.19–7.34) 4.96⁎⁎⁎ (2.45–7.46) 5.48⁎⁎⁎

6.56⁎⁎ (2.67–10.45) 6.40⁎⁎⁎ (2.65–10.15) 6.55⁎⁎⁎ (2.71–10.38) 6.46⁎⁎

6.36⁎⁎ (2.52–10.20) 5.19⁎⁎ (1.55–8.83) 5.66⁎⁎ (1.95–9.36) 5.86⁎⁎

4.24⁎ (0.93–7.55) 2.62 (−0.79–6.03) 2.19 (−0.99–5.38) 3.86⁎

5.97⁎⁎⁎ (3.88–8.07) 5.03⁎⁎⁎ (2.97–7.09) 5.21⁎⁎⁎ (3.17–7.24) 5.67⁎⁎⁎

(1.29–3.23) 2.47⁎⁎⁎

(2.91–8.04) 4.82⁎⁎⁎

(2.58–10.35) 6.39⁎⁎⁎

(3.10–9.20) 6.13⁎⁎⁎

(2.12–9.59) 5.97⁎⁎

(0.54–7.18) 4.00⁎

(3.58–7.77) 4.97⁎⁎⁎

(1.51–3.43) 2.17⁎⁎⁎ (1.17–3.17) 2.48⁎⁎⁎ (1.37–3.59) 1.73⁎⁎

(2.48–7.15) 5.41⁎⁎⁎ (2.81–8.01) 4.11⁎⁎ (1.47–6.75) 2.52⁎

(2.66–10.11) 4.31⁎ (0.97–7.65) 5.33⁎⁎⁎ (2.18–8.48) 3.81⁎

(3.03–9.23) 6.22⁎⁎⁎ (3.18–9.26) 5.30⁎⁎⁎ (2.23–8.37) 4.05⁎⁎

(2.26–9.67) 6.36⁎⁎ (2.59–10.12) 6.04⁎⁎ (1.81–10.27) 4.22⁎

(0.11–4.93)

(0.86–6.75)

(1.11–6.98)

(0.50–7.93)

(0.66–7.35) 4.02⁎ (0.76–7.28) 4.17⁎ (0.88–7.47) 1.32 (−1.76–4.40)

(3.06–6.88) 5.63⁎⁎⁎ (3.46–7.79) 5.12⁎⁎⁎ (2.98–7.25) 2.89⁎⁎

(0.69–2.77)

1.55 (−1.31–4.41) 0.74 (−2.21–3.69) 1.62 (−1.18–4.43) 1.11 (−1.73–3.95) 1.23 (−1.64–4.10) 1.17 (−1.69–4.02) 1.17 (−2.00–4.34) −0.25 (−3.30–2.79)

6.29⁎⁎⁎ (3.23–9.34) 4.78⁎⁎ (1.91–7.65) 5.68⁎⁎⁎ (2.67–8.70) 6.15⁎⁎⁎

Diabetes, depression

3.01⁎ (0.20–5.81) 1.95 (−0.75–4.65) 2.05 (−0.59–4.69) 2.57 (−0.19–5.32) 2.36 (−0.20–4.92) 2.90⁎ (0.08–5.71) 2.71⁎ (0.06–5.36) 0.50 (−1.71–2.70)

Diabetesa Diabetes, angina Diabetes, arthritis

Diabetes, stroke Diabetes, BMI Diabetes, all conditions

(0.97–4.81)

Abbreviations: S. Africa, South Africa; BMI, body mass index. Data are beta-coefficients (95% confidence intervals). Higher scores correspond to higher levels of disability. a Base model is adjusted for age, sex, education, wealth, location, and smoking. Other models included the individual condition to the base model. The final model is adjusted for all six conditions simultaneously. ⁎ p b 0.05. ⁎⁎ p b 0.01. ⁎⁎⁎ p b 0.001.

associated with DM (except Mexico and South Africa). In addition, DM was associated with higher odds for disability as measured by the WHODAS II score in all countries (except Mexico) in the base model. A few cross-country studies on the prevalence of DM exist (Liu et al., 2012; Yusuf et al., 2004; Espelt et al., 2013; Tao et al., 2013; Assari et al., 2014), however, only a limited number (Liu et al., 2012; Yusuf et al., 2004; Espelt et al., 2013; Tao et al., 2013; Assari et al., 2014) of these studies used the same study design, limiting comparability between countries. Additionally, none of these studies focused on older adults among which the prevalence of DM is increasing in parallel with the speed of aging. Recent data has shown that the prevalence of selfreported DM is approximately 17% and 20% among females and males, respectively in the US (Bays et al., 2007). Furthermore, in Europe, the prevalence of DM among males in the 6th decade of their life varied from 11% to 19% and from 13% to 25% in females of the same age (Rathmann et al., 2005). In addition, recent data from China and Ghana have reported DM prevalence to be 9% and 6% respectively (Amoah et al., 2002; Tao et al., 2013) for the population over 65 years old. The aforementioned are in accordance to the figures observed in our data. CVD risk factors (i.e., hypertension, obesity, DM, hypercholesterolemia, etc.) accumulate throughout the aging process (Tyrovolas et al., 2011). The burden of CVD risk factors varies between populations, as well as within population, with elderly people having the highest burden (Ford et al., 2002). Our study revealed that hypertension and central obesity were the two most consistent factors associated with greater likelihood of DM. The independent association between hypertension and DM is in line with results of other studies (Bays et al., 2007; Tyrovolas et al, 2009). The pathway of endothelial dysfunction could explain the association between hypertension and DM, since various markers of endothelial dysfunction has been reported in association with new-onset of DM (Meigs et al., 2004). Moreover, aging is generally associated with increase in central obesity and abdominal adiposity as well as fat accumulation in internal organs (Rabkin, 2007). It has been reported that body weight increases between 0.30 and 0.55 kg each year between the 4th and 6th decades of life (Guo et al., 1999). Previous studies indicated that body fat and muscle mass ratio in older populations are reversed in young adults (Rivlin, 2007). Greater waist circumference and weight-to-height ratio have been reported to be strongly related with increased visceral fat (Dobbelsteyn et al., 2001). In turn, increased visceral adipose tissue and central obesity promote insulin

resistance in the liver, adipose tissue, and muscles and this may explain the almost consistent association between central obesity and DM observed in the countries studied. An almost consistent association between DM and disability status was observed throughout the multivariate analysis. The nonsignificant result observed in Mexico for example, is in line with a previous study which also found non-significant associations in some countries although the age group was not the same as in our study (Assari et al., 2014). Several studies have demonstrated the role of diabetes on physical and functional disability (Assari et al., 2014; Gregg et al., 2000; Chiu et al., 2011). Recent longitudinal data have indicated that DM could consistently, as well as independently, change the physical functioning status during the process of aging (Chiu et al., 2011). Specifically, it has been proposed that DM may increase the risk of disability through interrelated complications, such as CVDs, micro- and macro-vascular disease, vision loss, stoke events, kidney failure, and neuropathy (Toutouzas et al., 2005). Large-scale population studies have reported that having DM is equally disabling as having coronary artery disease (Haffner et al., 1998). However, to the best of our knowledge, almost no global studies have reported an association between DM and disability status using the WHODAS II score — a validated measurement that assesses not only physical impairments, but also activity limitations that affect actions or behaviors of daily living and social participation that could affect various aspects of life. In the model adjusting for demographics and smoking, DM was associated with disability in all countries except Mexico. The inclusion of the chronic conditions lead to an attenuation of the coefficients in most countries but a large variation in its magnitude existed. A recent cross-country study on DM and activities of daily living also reported inconsistent associations among the investigated countries (Assari et al., 2014). This could be attributed to the differences in the prevalence and severity of the co-morbidities associated with DM between countries (Assari et al., 2014; Gregg et al., 2000). The fact that DM was still associated with higher risk for disability even after adjustment for angina, arthritis, cataract, depression, stroke, and BMI point to the possibility of residual confounding by conditions such as retinopathy, nephropathy, and neuropathy that our analysis could not adjust for. Moreover, the results of the multivariate analysis suggest the importance of DM on disability status among different populations and raise some concerns regarding prevention measures and early treatment among older adults with DM.

76

S. Tyrovolas et al. / Experimental Gerontology 64 (2015) 70–77

4.1. Strengths and limitations The present study has several strengths. It is one of the few studies to evaluate the effect of central obesity on DM using large nationallyrepresentative samples of older people around the world. Moreover, it is one of the first studies reporting a consistent association between DM and disability status by using the validated WHODAS II score. In terms of the limitations, although self-report of diseases has been shown to demonstrate good agreement with medical records in developed countries (Kriegsman et al., 1996), in settings with limited access to health care systems, under-reporting of DM is likely to have occurred. This may explain the particularly low prevalence of DM observed in Ghana. However, the estimates obtained in our study were similar to previously reported figures. In addition, in resource-limited areas, people may only have diseases detected when they are symptomatic or more severe (Levesque et al., 2013). Thus, heterogeneity in terms of the severity of DM may have existed in the countries studied. Next, despite the fact that healthy dietary patterns are strongly associated with DM, the survey did not include a detailed dietary assessment. Thus, their independent and confounding effects remain unknown. Finally, since this was a cross-sectional study, temporal relationships or causality cannot be established. For example, an inverse association between smoking and DM was observed, and this may have been due to the modification of smoking habits following physician's advice. 5. Conclusions The prevalence of DM among adults aged 50 years or over was high overall with large regional variations. DM was associated with hypertension, central obesity, and disability status in almost all of the countries studied. This work highlighted the importance of central obesity control on DM prevention among the elderly population globally. Given that DM has increased its impact worldwide in terms of disability in the last 20 years (Vos et al., 2012), DM prevention and management through specific medication and healthy lifestyle habits may constitute an effective mean for reducing disability among the older population. However, at the same time, our results suggest that the magnitude of the association between central obesity and diabetes or the association between diabetes and disability among older adults, may not be the same across all countries and the influence of different covariates on this association may differ by country. The importance of the identification of context-specific social and behavioral determinants of health outcomes has been emphasized (Assari, 2014; Assari et al., 2014; Assari and Lankarani, 2014). Thus, future studies should also focus on the factors that lead to the between-country differences observed. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.exger.2015.02.010. References Amoah, A.G., Owusu, S.K., Adjei, S., 2002. Diabetes in Ghana: a community based prevalence study in Greater Accra. Diabetes Res. Clin. Pract. 56, 197–205. Andrews, G., Kemp, A., Sunderland, M., et al., 2009. Normative data for the 12 item WHO Disability Assessment Schedule 2.0. PLoS One 4, e8343. Anton, S.D., Karabetian, C., Naugle, K., Buford, T.W., 2013. Obesity and diabetes as accelerators of functional decline: can lifestyle interventions maintain functional status in high risk older adults? Exp. Gerontol. 48, 888–897. Assari, S., 2014. Cross-country variation in additive effects of socio-economics, health behaviors, and comorbidities on subjective health of patients with diabetes. J. Diabetes Metab. Disord. 21 (13), 36. Assari, S., Lankarani, M.M., 2014. Association between heart disease and subjective health in ten North, Middle, and South American countries. Int. J. Travel Med. Glob. Health 2, 141–147. Assari, S., Lankarani, R.M., Lankarani, M.M., 2014. Cross-country differences in the association between diabetes and disability. J. Diabetes Metab. Disord. 13, 3.

Basu, S., Millett, C., 2013. Social epidemiology of hypertension in middle-income countries: determinants of prevalence, diagnosis, treatment, and control in the WHO SAGE study. Hypertension 62, 18–26. Bays, H.E., Bazata, D.D., Clark, N.G., et al., 2007. Prevalence of self-reported diagnosis of diabetes mellitus and associated risk factors in a national survey in the US population: SHIELD (Study to Help Improve Early evaluation and management of risk factors Leading to Diabetes). BMC Public Health 7, 277. Branch, L.G., Jette, A.M., 1982. A prospective study of long-term care institutionalization among the aged. Am. J. Public Health 72, 1373–1379. Browning, L.M., Hsieh, S.D., Ashwell, M., 2010. A systematic review of waist-to-height ratio as a screening tool for the prediction of cardiovascular disease and diabetes: 0.5 could be a suitable global boundary value. Nutr. Res. Rev. 23, 247–269. Cai, L., Liu, A., Zhang, Y., et al., 2013. Waist-to-height ratio and cardiovascular risk factors among Chinese adults in Beijing. PLoS One 8, e69298. Chiu, C.J., Wray, L.A., Ofstedal, M.B., 2011. Diabetes-related change in physical disability from midlife to older adulthood: evidence from 1996–2003 survey of health and living status of the elderly in Taiwan. Diabetes Res. Clin. Pract. 91, 413–423. Dobbelsteyn, C.J., Joffres, M.R., MacLean, D.R., et al., 2001. The Canadian Heart Health Surveys Research Group: a comparative evaluation of waist circumference, waist-to-hip ratio and body mass index as indicators of cardiovascular risk factors. The Canadian Heart Health Surveys. Int. J. Obes. 25, 652–661. Espelt, A., Borrell, C., Palència, L., et al., 2013. Socioeconomic inequalities in the incidence and prevalence of type 2 diabetes mellitus in Europe. Gac. Sanit. 27, 494–501. Ford, E.S., Giles, W.H., Dietz, W.H., 2002. Prevalence of the metabolic syndrome among US adults: findings from the third National Health and Nutrition Examination Survey. JAMA 287, 356–359. Gregg, E.W., Beckles, G.L., Williamson, D.F., et al., 2000. Diabetes and physical disability among older U.S. adults. Diabetes Care 23, 1272–1277. Guo, S.S., Zeller, C., Chumlea, W.C., et al., 1999. Aging, body composition, and lifestyle: the Fels Longitudinal Study. Am. J. Clin. Nutr. 70, 405–411. Guralnik, J.M., Simonsick, E.M., Ferrucci, L., et al., 1994. A short physical performance battery assessing lower extremity function: association with self-reported disability and prediction of mortality and nursing home admission. J. Gerontol. Med. Sci. 49, M85–M94. Hadaegh, F., Zabetian, A., Harati, H., et al., 2006. Waist/height ratio as a better predictor of type 2 diabetes compared to body mass index in Tehranian adult men — a 3.6-year prospective study. Exp. Clin. Endocrinol. Diabetes 114, 310–315. Haffner, S.M., Lehto, S., Ronnemaa, T., et al., 1998. Mortality from coronary heart disease in subjects with type 2 diabetes and in nondiabetic subjects with and without prior myocardial infarction. N. Engl. J. Med. 13, 229–234. Kriegsman, D.M., Penninx, B.W., van Eijk, J.T., et al., 1996. Self-reports and general practitioner information on the presence of chronic diseases in community dwelling elderly. A study on the accuracy of patients' self-reports and on determinants of inaccuracy. J. Clin. Epidemiol. 49, 1407–1417. Levesque, J.F., Mukherjee, S., Grimard, D., et al., 2013. Measuring the prevalence of chronic diseases using population surveys by pooling self-reported symptoms, diagnosis and treatments: results from the World Health Survey of 2003 for South Asia. Int. J. Public Health 58, 435–447. Liu, L., Yin, X., Morrissey, S., 2012. Global variability in diabetes mellitus and its association with body weight and primary healthcare support in 49 low- and middle-income developing countries. Diabet. Med. 29, 995–1002. Meigs, J.B., Hu, F.B., Rifai, N., et al., 2004. Biomarkers of endothelial dysfunction and risk of type 2 diabetes mellitus. JAMA 291, 1978–1986. Perales, J., Martin, S., Ayuso-Mateos, J.L., et al., 2014. Factors associated with active aging in Finland, Poland, and Spain. Int. Psychogeriatr. 15, 1–13. Pitsavos, C., Tampourlou, M., Panagiotakos, D.B., et al., 2007. Association between lowgrade systemic inflammation and type 2 diabetes mellitus among men and women from the ATTICA Study. Rev. Diabet. Stud. 4, 98–104. Qiao, Q., Nyamdorj, R., 2010. Is the association of type II diabetes with waist circumference or waist-to-hip ratio stronger than that with body mass index? Eur. J. Clin. Nutr. 64, 30–34. Rabkin, S.W., 2007. Epicardial fat: properties, function and relationship to obesity. Obes. Rev. 8, 253–261. Rathmann, W., Haastert, B., Icks, A., et al., 2005. The diabetes epidemic in the elderly population in Western Europe: data from population-based studies. Gesundheitswesen 67, 110–114. Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus, 1997. Diabetes Care 20, 1183–1197. Rivlin, R.S., 2007. Keeping the young-elderly healthy: is it too late to improve our health through nutrition? Am. J. Clin. Nutr. 86, 1572S–1576S. Rose, G.A., 1962. The diagnosis of ischaemic heart pain and intermittent claudication in field surveys. Bull. World Health Organ. 27, 645–658. Saaristo, T.E., Barengo, N.C., Korpi-Hyövälti, E., et al., 2008. High prevalence of obesity, central obesity and abnormal glucose tolerance in the middle-aged Finnish population. BMC Public Health 8, 423. Schneider, H.J., Friedrich, N., Klotsche, J., et al., 2010. The predictive value of different measures of obesity for incident cardiovascular events and mortality. J. Clin. Endocrinol. Metab. 95, 1777–1785. Tao, J., Ma, Y.T., Xiang, Y., et al., 2013. Prevalence of major cardiovascular risk factors and adverse risk profiles among three ethnic groups in the Xinjiang Uygur Autonomous Region, China. Lipids Health Dis. 12, 185. Toutouzas, K., Markou, V., Drakopoulou, M., et al., 2005. Patients with type two diabetes mellitus: increased local inflammatory activation in culprit atheromatous plaques. Hell. J. Cardiol. 46, 283–288. Tyrovolas, S., Zeimbekis, A., Bountziouka, V., et al., 2009. Factors associated with the prevalence of diabetes mellitus among elderly men and women living in Mediterranean islands: the MEDIS study. Rev. Diabet. Stud. 6, 54–63.

S. Tyrovolas et al. / Experimental Gerontology 64 (2015) 70–77 Tyrovolas, S., Chalkias, C., Morena, M., et al., 2011. Health care access and prevalence of the metabolic syndrome among elders living in high-altitude areas of the Mediterranean islands: the MEDIS study. Rev. Diabet. Stud. 8, 468–476. Ustün, T.B., Chatterji, S., Kostanjsek, N., et al., 2010. Developing the World Health Organization Disability Assessment Schedule 2.0. Bull. World Health Organ. 88, 815–823. Vos, T., Flaxman, A.D., Naghavi, M., et al., 2012. Years lived with disability (YLDs) for 1160 sequelae of 289 diseases and injuries 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet 380, 2163–2196.

77

Wood, D., 2001. Established and emerging cardiovascular risk factors. Am. Heart J. 141, 49–57. World Health Organization, 2013. Diabetes. Fact sheet no. 312. Available from:, http:// www.who.int/diabetes/en/ (Accessed 4 July, 2014). Yusuf, S., Hawken, S., Ounpuu, S., et al., 2004. Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): case–control study. Lancet 364, 937–952.