Cohort Profile: The Singapore Multi-Ethnic Cohort (MEC) study

Cohort Profile: The Singapore Multi-Ethnic Cohort (MEC) study Why was the cohort set up? Non-communicable diseases such as type 2 diabetes (T2D) mellitus, coronary artery disease, stroke and cancers, are major contributors to ill health across the world including Asia. These conditions are multi-factorial in origin, often involving complex gene-environment interactions. Singapore is a multi-ethnic island state and provides a useful model to evaluate determinants of the development of chronic diseases in Asian ethnic groups. Three major Asian ethnic groups are represented in Singapore: Chinese, Malays and Indians. The Singapore Multi-Ethnic Cohort (MEC) allows us to better understand how genes and lifestyle may influence health and diseases differently in persons of Chinese, Malay and Indian ethnicity. As these ethnic groups reside in the same Singapore setting, confounding of ethnic differences by differences between countries is avoided. Through the MEC, we hope to improve preventive and therapeutic measures, as well as provide information to advance public health and health education policies for Asian populations. Who is in the cohort? The MEC is a closed cohort which included a total of 14 465 male and female adults. The cohort was formed by combining two existing population-based studies with measurements in the period 2004 to 2007, the Singapore Prospective Study Program (SP2) and the Singapore Cardiovascular Cohort Study (SCCS2), with additional recruitment of participants from 2007 to 2010.1–3 The combined cohort has a good representation of Chinese, Malay and Indian ethnic groups. The SP2 and SCCS2 recruited 8340 participants from four previous cross-sectional studies: Thyroid and Heart Study 1982–84,4 National Health Survey 1992,5 National University of Singapore Heart Study 1993–956 and National Health Survey 1998.7 All studies involved a random sample of Singapore residents aged 21 years and above, with disproportionate sampling stratified by ethnicity to increase the numbers for ethnic minorities, i.e. Malays and Indians. In addition to the participants from SP2 and SCCS2, a further 6125 Singapore residents were recruited into the MEC study through public outreach and referrals from existing cohort members. Invitation to participate was open to any Singapore citizens or long-term residents of age 21 to 75 years. People with a history of heart disease, stroke, cancer and renal failure were excluded at baseline, as these are the outcomes of interest in this prospective cohort for non-communicable diseases. Recruitment drives were carried out at community events, mosques and temples in addition to household visitation to enrich the proportion of Malays and Indians. Ethics approval for the SP2 was provided by the SingHealth Centralised Institutional Review Board (IRB). The rest of the MEC and the follow-up were approved by the National University of Singapore IRB. Written consent was obtained for registry and medical records linkage, future analysis of stored biological samples and future follow-up. All MEC participants were visited at home to complete an interview, and were subsequently invited to undergo a health screening for clinical assessments and collection of blood and urine samples to be stored for future analysis. Of the 14 465 participants of the MEC study, 77% (N = 11 085) completed both the interviews and health screening, whereas the remaining 23% (N = 3380) only completed the interview-administered questionnaire. The sociodemographic and socioeconomic profiles of participants are presented in Table 1. The age of the participants ranged from 21 to 94 years and the mean age was 46 ± 13 years; 56% of the participants were females, and the ethnic composition was 47% Chinese, 26% Malay and 27% Indian. Most of the participants were married (75%). In terms of highest qualification attained, 30% of the participants received primary or lower education, 35% received secondary education, 6% received vocational training and 29% received post-secondary or higher education. Around 89% of the participants lived in public housing and 11% lived in private housing; this is reflective of the general Singapore resident population where the majority resides in government housing. Table 1 Demographic profile of participants at baseline (N = 14 465) Characteristics  Number  Percentage  Age at interview (in years), mean ± SD  46.1  13.3  Gender   Male  6353  43.9   Female  8112  56.1  Ethnicity   Chinese  6814  47.1   Malay  3773  26.1   Indian  3837  26.5   Others  41  0.3  Marital status   Never married  2454  17.0   Currently married  10776  74.5   Separated/divorced  516  3.6   Widowed  710  4.9  Educational statusa   No formal qualification  1313  9.1   Primary  3003  20.8   Secondary  5095  35.3   Vocational training  855  5.9   Post-secondary  2412  16.7   University and above  1775  12.3  Monthly household income (SGD)   Less than $2000  3228  30.4   $2000–$4000  3500  33.0   $4000–$6000  2053  19.4   $6000–$10,000  1292  12.2   More than $10,000  529  5.0  Housing type   Public housing 1–3-room flat  3420  23.7   Public housing 4-room flat  5524  38.2   Public housing 5-room or executive flat  3990  27.6   Private condominium  797  5.5   Landed property  717  5.0  Characteristics  Number  Percentage  Age at interview (in years), mean ± SD  46.1  13.3  Gender   Male  6353  43.9   Female  8112  56.1  Ethnicity   Chinese  6814  47.1   Malay  3773  26.1   Indian  3837  26.5   Others  41  0.3  Marital status   Never married  2454  17.0   Currently married  10776  74.5   Separated/divorced  516  3.6   Widowed  710  4.9  Educational statusa   No formal qualification  1313  9.1   Primary  3003  20.8   Secondary  5095  35.3   Vocational training  855  5.9   Post-secondary  2412  16.7   University and above  1775  12.3  Monthly household income (SGD)   Less than $2000  3228  30.4   $2000–$4000  3500  33.0   $4000–$6000  2053  19.4   $6000–$10,000  1292  12.2   More than $10,000  529  5.0  Housing type   Public housing 1–3-room flat  3420  23.7   Public housing 4-room flat  5524  38.2   Public housing 5-room or executive flat  3990  27.6   Private condominium  797  5.5   Landed property  717  5.0  a Educational status: secondary education (‘O’/‘N’ level), vocational training (attended Institute of Technical Education or obtained National Technical Certificate) and post-secondary education (‘A’ level, polytechnic/diploma). Missings marital status (n = 9), educational status (n = 12), monthly household income (n = 3863) and housing type (n = 17). Table 1 Demographic profile of participants at baseline (N = 14 465) Characteristics  Number  Percentage  Age at interview (in years), mean ± SD  46.1  13.3  Gender   Male  6353  43.9   Female  8112  56.1  Ethnicity   Chinese  6814  47.1   Malay  3773  26.1   Indian  3837  26.5   Others  41  0.3  Marital status   Never married  2454  17.0   Currently married  10776  74.5   Separated/divorced  516  3.6   Widowed  710  4.9  Educational statusa   No formal qualification  1313  9.1   Primary  3003  20.8   Secondary  5095  35.3   Vocational training  855  5.9   Post-secondary  2412  16.7   University and above  1775  12.3  Monthly household income (SGD)   Less than $2000  3228  30.4   $2000–$4000  3500  33.0   $4000–$6000  2053  19.4   $6000–$10,000  1292  12.2   More than $10,000  529  5.0  Housing type   Public housing 1–3-room flat  3420  23.7   Public housing 4-room flat  5524  38.2   Public housing 5-room or executive flat  3990  27.6   Private condominium  797  5.5   Landed property  717  5.0  Characteristics  Number  Percentage  Age at interview (in years), mean ± SD  46.1  13.3  Gender   Male  6353  43.9   Female  8112  56.1  Ethnicity   Chinese  6814  47.1   Malay  3773  26.1   Indian  3837  26.5   Others  41  0.3  Marital status   Never married  2454  17.0   Currently married  10776  74.5   Separated/divorced  516  3.6   Widowed  710  4.9  Educational statusa   No formal qualification  1313  9.1   Primary  3003  20.8   Secondary  5095  35.3   Vocational training  855  5.9   Post-secondary  2412  16.7   University and above  1775  12.3  Monthly household income (SGD)   Less than $2000  3228  30.4   $2000–$4000  3500  33.0   $4000–$6000  2053  19.4   $6000–$10,000  1292  12.2   More than $10,000  529  5.0  Housing type   Public housing 1–3-room flat  3420  23.7   Public housing 4-room flat  5524  38.2   Public housing 5-room or executive flat  3990  27.6   Private condominium  797  5.5   Landed property  717  5.0  a Educational status: secondary education (‘O’/‘N’ level), vocational training (attended Institute of Technical Education or obtained National Technical Certificate) and post-secondary education (‘A’ level, polytechnic/diploma). Missings marital status (n = 9), educational status (n = 12), monthly household income (n = 3863) and housing type (n = 17). Lifestyle, anthropometric measures and biomarkers profiles of study participants, stratified by ethnicity, are presented in Table 2. There were more current smokers in the Malays (24%) than Indians (17%) and Chinese (12%). Self-reported prevalence of T2D was the highest amongst the Indians (16%), followed by Malays (10%) and Chinese (7%). Similarly, Indians had higher HbA1c (6.2 ± 1.3%) than Malays (6.0 ± 1.3%) and Chinese (5.7 ± 0.8%). Prevalence of self-reported hypertension was similar across the three ethnic groups (18–19%, P-value = 0.10), though mean systolic and diastolic blood pressures were higher in Malays and Chinese than Indians. The Chinese had the highest prevalence of self-reported hypercholesterolaemia (Chinese: 28%; Indians: 23%; Malays: 19%), but mean low-density lipoprotein cholesterol (LDL) was the highest in the Malays (Malays: 3.5 ± 0.9 mmol/l; Indians: 3.4 ± 0.9 mmol/l; Chinese: 3.2 ± 0.8 mmol/l). Table 2 Lifestyle, anthropometric measures, and biomarkers profiles of study participants at baseline by ethnicity (N = 14 424) Characteristics  Chinese  Malays  Indians  P-value  Cigarette smoking status (%)   Never smoker  5503 (80.8)  2604 (69.0)  2945 (76.8)  < 0.001   Ex-smoker  514 (7.5)  282 (7.5)  234 (6.1)   Current smoker  797 (11.7)  887 (23.5)  657 (17.1)  Self-reported health conditions (%)   T2Da  436 (6.5)  364 (9.8)  588 (15.6)  < 0.001   Hypertensiona  1285 (19.2)  652 (17.6)  685 (18.2)  0.098   Hypercholesterolaemiaa  1755 (27.5)  672 (18.8)  824 (22.6)  < 0.001  Physical activity (MET-h/week)b, median (IQR)   Moderate- and/or vigorous-intensity leisure-time  5.5 (0.0–15.5)  4.5 (0.0–15.8)  3.5 (0.0–14.0)  < 0.001  Physical activity (MET-h/week)b, mean ± SD   Moderate- and/or vigorous-intensity leisure-time  11.9 ± 23.1  12.8 ± 23.9  11.4 ± 21.1  0.479  Macro nutrient intake (% energy)   Carbohydrate  53.8 ± 6.8  54.5 ± 6.7  54.3 ± 6.5  < 0.001   Protein  15.8 ± 2.3  14.1 ± 2.2  13.2 ± 2.1  < 0.001   Fat  30.1 ± 5.7  31.3 ± 5.8  31.8 ± 5.8  < 0.001  BMI category (kg/m2), (%)   < 18.5  444 (8.6)  113 (3.9)  127 (4.2)  < 0.001   18.5–23.0  2317 (45.1)  621 (21.7)  682 (22.5)   23.0–27.5  1778 (34.6)  1021 (35.7)  1172 (38.7)   ≥ 27.5  603 (11.7)  1107 (38.7)  1050 (34.6)  Waist circumference (cm), (%)   Normal  3492 (67.9)  1287 (45.1)  1155 (38.1)  < 0.001   Highc  1650 (32.1)  1569 (54.9)  1878 (61.9)  Blood pressure (mmHg), mean ± SD   Systolic  127.5 ± 20.6  129.0 ± 20.8  124.5 ± 21.3  < 0.001   Diastolic  75.0 ± 11.3  74.4 ± 11.3  73.0 ± 11.1  < 0.001  Biomarkers, mean ± SD   Triglycerides (mmol/l)  1.3 ± 0.8  1.4 ± 1.1  1.4 ± 1.0  < 0.001   High-density lipoprotein (mmol/l) (HDL)  1.4 ± 0.4  1.2 ± 0.3  1.1 ± 0.3  < 0.001   Low-density lipoprotein (mmol/l) (LDL)  3.2 ± 0.8  3.5 ± 0.9  3.4 ± 0.9  < 0.001   HbA1c (%)  5.7 ± 0.8  6.0 ± 1.3  6.2 ± 1.3  < 0.001  Characteristics  Chinese  Malays  Indians  P-value  Cigarette smoking status (%)   Never smoker  5503 (80.8)  2604 (69.0)  2945 (76.8)  < 0.001   Ex-smoker  514 (7.5)  282 (7.5)  234 (6.1)   Current smoker  797 (11.7)  887 (23.5)  657 (17.1)  Self-reported health conditions (%)   T2Da  436 (6.5)  364 (9.8)  588 (15.6)  < 0.001   Hypertensiona  1285 (19.2)  652 (17.6)  685 (18.2)  0.098   Hypercholesterolaemiaa  1755 (27.5)  672 (18.8)  824 (22.6)  < 0.001  Physical activity (MET-h/week)b, median (IQR)   Moderate- and/or vigorous-intensity leisure-time  5.5 (0.0–15.5)  4.5 (0.0–15.8)  3.5 (0.0–14.0)  < 0.001  Physical activity (MET-h/week)b, mean ± SD   Moderate- and/or vigorous-intensity leisure-time  11.9 ± 23.1  12.8 ± 23.9  11.4 ± 21.1  0.479  Macro nutrient intake (% energy)   Carbohydrate  53.8 ± 6.8  54.5 ± 6.7  54.3 ± 6.5  < 0.001   Protein  15.8 ± 2.3  14.1 ± 2.2  13.2 ± 2.1  < 0.001   Fat  30.1 ± 5.7  31.3 ± 5.8  31.8 ± 5.8  < 0.001  BMI category (kg/m2), (%)   < 18.5  444 (8.6)  113 (3.9)  127 (4.2)  < 0.001   18.5–23.0  2317 (45.1)  621 (21.7)  682 (22.5)   23.0–27.5  1778 (34.6)  1021 (35.7)  1172 (38.7)   ≥ 27.5  603 (11.7)  1107 (38.7)  1050 (34.6)  Waist circumference (cm), (%)   Normal  3492 (67.9)  1287 (45.1)  1155 (38.1)  < 0.001   Highc  1650 (32.1)  1569 (54.9)  1878 (61.9)  Blood pressure (mmHg), mean ± SD   Systolic  127.5 ± 20.6  129.0 ± 20.8  124.5 ± 21.3  < 0.001   Diastolic  75.0 ± 11.3  74.4 ± 11.3  73.0 ± 11.1  < 0.001  Biomarkers, mean ± SD   Triglycerides (mmol/l)  1.3 ± 0.8  1.4 ± 1.1  1.4 ± 1.0  < 0.001   High-density lipoprotein (mmol/l) (HDL)  1.4 ± 0.4  1.2 ± 0.3  1.1 ± 0.3  < 0.001   Low-density lipoprotein (mmol/l) (LDL)  3.2 ± 0.8  3.5 ± 0.9  3.4 ± 0.9  < 0.001   HbA1c (%)  5.7 ± 0.8  6.0 ± 1.3  6.2 ± 1.3  < 0.001  a Participants were asked if they had been diagnosed with T2D, hypertension or high cholesterol by Western doctors. b MET, Metabolic Equivalent of Task. c High waist circumference was defined as > 90 cm for men and > 80 cm for women. Missings smoking status (n = 1), T2D (n = 250), hypertension (n = 269), hyperlipidaemia (n = 813), physical activity (n = 1), macronutrients (n = 264). BMI (n = 3389), waist circumference (n = 3393), blood pressure (n = 3379), triglycerides (n = 3423), HDL (n = 3427), LDL (n = 3475) and HbA1c (n = 5643); most missing values for physical and biochemical variables are due to non-participation in the health screening part of the study. Table 2 Lifestyle, anthropometric measures, and biomarkers profiles of study participants at baseline by ethnicity (N = 14 424) Characteristics  Chinese  Malays  Indians  P-value  Cigarette smoking status (%)   Never smoker  5503 (80.8)  2604 (69.0)  2945 (76.8)  < 0.001   Ex-smoker  514 (7.5)  282 (7.5)  234 (6.1)   Current smoker  797 (11.7)  887 (23.5)  657 (17.1)  Self-reported health conditions (%)   T2Da  436 (6.5)  364 (9.8)  588 (15.6)  < 0.001   Hypertensiona  1285 (19.2)  652 (17.6)  685 (18.2)  0.098   Hypercholesterolaemiaa  1755 (27.5)  672 (18.8)  824 (22.6)  < 0.001  Physical activity (MET-h/week)b, median (IQR)   Moderate- and/or vigorous-intensity leisure-time  5.5 (0.0–15.5)  4.5 (0.0–15.8)  3.5 (0.0–14.0)  < 0.001  Physical activity (MET-h/week)b, mean ± SD   Moderate- and/or vigorous-intensity leisure-time  11.9 ± 23.1  12.8 ± 23.9  11.4 ± 21.1  0.479  Macro nutrient intake (% energy)   Carbohydrate  53.8 ± 6.8  54.5 ± 6.7  54.3 ± 6.5  < 0.001   Protein  15.8 ± 2.3  14.1 ± 2.2  13.2 ± 2.1  < 0.001   Fat  30.1 ± 5.7  31.3 ± 5.8  31.8 ± 5.8  < 0.001  BMI category (kg/m2), (%)   < 18.5  444 (8.6)  113 (3.9)  127 (4.2)  < 0.001   18.5–23.0  2317 (45.1)  621 (21.7)  682 (22.5)   23.0–27.5  1778 (34.6)  1021 (35.7)  1172 (38.7)   ≥ 27.5  603 (11.7)  1107 (38.7)  1050 (34.6)  Waist circumference (cm), (%)   Normal  3492 (67.9)  1287 (45.1)  1155 (38.1)  < 0.001   Highc  1650 (32.1)  1569 (54.9)  1878 (61.9)  Blood pressure (mmHg), mean ± SD   Systolic  127.5 ± 20.6  129.0 ± 20.8  124.5 ± 21.3  < 0.001   Diastolic  75.0 ± 11.3  74.4 ± 11.3  73.0 ± 11.1  < 0.001  Biomarkers, mean ± SD   Triglycerides (mmol/l)  1.3 ± 0.8  1.4 ± 1.1  1.4 ± 1.0  < 0.001   High-density lipoprotein (mmol/l) (HDL)  1.4 ± 0.4  1.2 ± 0.3  1.1 ± 0.3  < 0.001   Low-density lipoprotein (mmol/l) (LDL)  3.2 ± 0.8  3.5 ± 0.9  3.4 ± 0.9  < 0.001   HbA1c (%)  5.7 ± 0.8  6.0 ± 1.3  6.2 ± 1.3  < 0.001  Characteristics  Chinese  Malays  Indians  P-value  Cigarette smoking status (%)   Never smoker  5503 (80.8)  2604 (69.0)  2945 (76.8)  < 0.001   Ex-smoker  514 (7.5)  282 (7.5)  234 (6.1)   Current smoker  797 (11.7)  887 (23.5)  657 (17.1)  Self-reported health conditions (%)   T2Da  436 (6.5)  364 (9.8)  588 (15.6)  < 0.001   Hypertensiona  1285 (19.2)  652 (17.6)  685 (18.2)  0.098   Hypercholesterolaemiaa  1755 (27.5)  672 (18.8)  824 (22.6)  < 0.001  Physical activity (MET-h/week)b, median (IQR)   Moderate- and/or vigorous-intensity leisure-time  5.5 (0.0–15.5)  4.5 (0.0–15.8)  3.5 (0.0–14.0)  < 0.001  Physical activity (MET-h/week)b, mean ± SD   Moderate- and/or vigorous-intensity leisure-time  11.9 ± 23.1  12.8 ± 23.9  11.4 ± 21.1  0.479  Macro nutrient intake (% energy)   Carbohydrate  53.8 ± 6.8  54.5 ± 6.7  54.3 ± 6.5  < 0.001   Protein  15.8 ± 2.3  14.1 ± 2.2  13.2 ± 2.1  < 0.001   Fat  30.1 ± 5.7  31.3 ± 5.8  31.8 ± 5.8  < 0.001  BMI category (kg/m2), (%)   < 18.5  444 (8.6)  113 (3.9)  127 (4.2)  < 0.001   18.5–23.0  2317 (45.1)  621 (21.7)  682 (22.5)   23.0–27.5  1778 (34.6)  1021 (35.7)  1172 (38.7)   ≥ 27.5  603 (11.7)  1107 (38.7)  1050 (34.6)  Waist circumference (cm), (%)   Normal  3492 (67.9)  1287 (45.1)  1155 (38.1)  < 0.001   Highc  1650 (32.1)  1569 (54.9)  1878 (61.9)  Blood pressure (mmHg), mean ± SD   Systolic  127.5 ± 20.6  129.0 ± 20.8  124.5 ± 21.3  < 0.001   Diastolic  75.0 ± 11.3  74.4 ± 11.3  73.0 ± 11.1  < 0.001  Biomarkers, mean ± SD   Triglycerides (mmol/l)  1.3 ± 0.8  1.4 ± 1.1  1.4 ± 1.0  < 0.001   High-density lipoprotein (mmol/l) (HDL)  1.4 ± 0.4  1.2 ± 0.3  1.1 ± 0.3  < 0.001   Low-density lipoprotein (mmol/l) (LDL)  3.2 ± 0.8  3.5 ± 0.9  3.4 ± 0.9  < 0.001   HbA1c (%)  5.7 ± 0.8  6.0 ± 1.3  6.2 ± 1.3  < 0.001  a Participants were asked if they had been diagnosed with T2D, hypertension or high cholesterol by Western doctors. b MET, Metabolic Equivalent of Task. c High waist circumference was defined as > 90 cm for men and > 80 cm for women. Missings smoking status (n = 1), T2D (n = 250), hypertension (n = 269), hyperlipidaemia (n = 813), physical activity (n = 1), macronutrients (n = 264). BMI (n = 3389), waist circumference (n = 3393), blood pressure (n = 3379), triglycerides (n = 3423), HDL (n = 3427), LDL (n = 3475) and HbA1c (n = 5643); most missing values for physical and biochemical variables are due to non-participation in the health screening part of the study. The prevalence of obesity, defined as ≥ 27.5 kg/m2,8 was the highest in Malays (39%), slightly lower in the Indians (35%) and much lower in Chinese (12%). Indians had the highest prevalence of high waist circumference (62%), defined as > 90 cm for men and > 80 cm for women,9 whereas 55% of Malays and 32% of Chinese had high waist circumference. Metabolic equivalents (METs) were obtained from the 2011 Compendium of Physical Activity.10 The median (interquartile range, IQR) weekly leisure-time activity level of moderate and/or vigorous intensity was the highest amongst the Chinese [6 (0–16) MET-h/week], compared with the Malays [5 (0–16) MET-h/week] and Indians [4 (0–14) MET-h/week]. On average, Chinese had a slightly higher protein intake as percentage of total energy intake and lower intake of total fat and carbohydrate than Malays and Indians. Differences between the three ethnic groups were assessed using chi square testing for categorical variables. Comparison of means was done using analysis of variance testing, and comparison of medians was done using the Kruskal-Wallis rank sum test for continuous variables. As shown in Table 2, due to the large number of participants, all differences between ethnic groups had P-values less than 0.001, except hypertension status (P-value = 0.098) and average leisure-time physical activity (P-value = 0.479). How often have they been followed up and what is the attrition like? Baseline recruitment was done between 2004 and 2010. The first wave of in-person follow-up with the cohort participants began in 2011 and ended in 2016. Of the baseline participants, 28% were not contactable (i.e. contact details changed and no updated information available, or frequently travelling, or access to household denied and unable to contact despite six attempts at household visitation) and 2.5% were confirmed to have been lost to follow-up (i.e. deceased, migrated, declined follow-up, lost mental competence to give consent to continue the research, institutionalized or physically unfit to participate). Of the contactable participants, 60% (N = 6112) agreed to participate in the follow-up survey. Data will also be linked to follow-up data from the Singapore National Registry of Diseases Office on chronic diseases (i.e. acute myocardial infarction, cancers, renal failure and stroke) and mortality; 86% of the baseline participants gave consent for data linkage with medical records and national registries. What has been measured? Table 3 summarizes information collected from the MEC study at baseline and follow-up assessments. The baseline interviews assessed sociodemographic variables, detailed lifestyle behaviours, personal and family medical history and medication use, and health-related quality of life (HRQoL). The health examination included measurements of anthropometric characteristics, peripheral neuropathy, branchial and ankle blood pressures and biomarkers in blood and urine. During the follow-up assessments, we collected information on mental health and cognitive function and conducted electrocardiogram, central blood pressure and hand grip strength measurements in addition to the baseline measurements. More details on these measurements are described below. Table 3 Summary of variables collected or derived from the cohort Variables  Baseline  Follow-up  Questionnaire   Demographics  ✓  ✓   Cigarette smoking history  ✓  ✓   Environmental tobacco smoke exposure  ✓  ✓   Alcohol consumption  ✓  ✓   Medication use  ✓  ✓   Medical history  ✓  ✓   Family medical history  ✓  ✓   Women’s health  ✓  ✓   Physical activity  ✓  ✓   Dietary information (FFQ)  ✓  ✓   Health related quality of life (SF-36)  ✓  ✓   Health-related quality of life (EQ-5D)    ✓   Cognitive function (Mini-Mental State Examination)    ✓   Family life    ✓   Sources of psychological stress    ✓   Kessler psychological distress scale (K10)    ✓  Anthropometric measures   Height  ✓  ✓   Weight  ✓  ✓   Waist circumference  ✓  ✓   Hip circumference  ✓  ✓   Skinfold thickness    ✓  Clinical assessment   Light touch assessment measured by monofilament  ✓  ✓   Assessment of foot proprioception by neurothesiometer  ✓  ✓   Brachial and ankle blood pressure  ✓  ✓   Central aortic systolic pressure and arterial pulse waveform indices (rAI, rAP, PRT)    ✓   Bone densitometrya    ✓   Cardiac computer tomographya    ✓   Body fat and lean mass composition by DEXAa    ✓   Subcutaneous abdominal and intra-abdominal fat by CTa    ✓   Electrocardiography    ✓   Hand grip strength    ✓  Blood sample   Fasting or random glucose  ✓  ✓   Creatinine  ✓  ✓   Triglycerides  ✓  ✓   Total cholesterol  ✓  ✓   High-density lipoprotein cholesterol  ✓  ✓   Low-density lipoprotein cholesterol  ✓  ✓   HbA1c  ✓  ✓   Cortisol  ✓  ✓   High-sensitivity C-reactive protein  ✓  ✓   Insulin  ✓     Adiponectin  ✓     Interleukin-6    ✓   Interleukin-1 receptor antagonist    ✓  Urine sample   Protein (semi-quantitative)  ✓     Albumin (semi-quantitative)    ✓  Genetic data   Genotyping array  ✓(subset)     Whole exome sequence  ✓(subset)    Metabolomicsb   Amino acids  ✓(subset)     Acylcarnitines  ✓(subset)     Ceramides  ✓(subset)     Sphingolipids  ✓(subset)    Variables  Baseline  Follow-up  Questionnaire   Demographics  ✓  ✓   Cigarette smoking history  ✓  ✓   Environmental tobacco smoke exposure  ✓  ✓   Alcohol consumption  ✓  ✓   Medication use  ✓  ✓   Medical history  ✓  ✓   Family medical history  ✓  ✓   Women’s health  ✓  ✓   Physical activity  ✓  ✓   Dietary information (FFQ)  ✓  ✓   Health related quality of life (SF-36)  ✓  ✓   Health-related quality of life (EQ-5D)    ✓   Cognitive function (Mini-Mental State Examination)    ✓   Family life    ✓   Sources of psychological stress    ✓   Kessler psychological distress scale (K10)    ✓  Anthropometric measures   Height  ✓  ✓   Weight  ✓  ✓   Waist circumference  ✓  ✓   Hip circumference  ✓  ✓   Skinfold thickness    ✓  Clinical assessment   Light touch assessment measured by monofilament  ✓  ✓   Assessment of foot proprioception by neurothesiometer  ✓  ✓   Brachial and ankle blood pressure  ✓  ✓   Central aortic systolic pressure and arterial pulse waveform indices (rAI, rAP, PRT)    ✓   Bone densitometrya    ✓   Cardiac computer tomographya    ✓   Body fat and lean mass composition by DEXAa    ✓   Subcutaneous abdominal and intra-abdominal fat by CTa    ✓   Electrocardiography    ✓   Hand grip strength    ✓  Blood sample   Fasting or random glucose  ✓  ✓   Creatinine  ✓  ✓   Triglycerides  ✓  ✓   Total cholesterol  ✓  ✓   High-density lipoprotein cholesterol  ✓  ✓   Low-density lipoprotein cholesterol  ✓  ✓   HbA1c  ✓  ✓   Cortisol  ✓  ✓   High-sensitivity C-reactive protein  ✓  ✓   Insulin  ✓     Adiponectin  ✓     Interleukin-6    ✓   Interleukin-1 receptor antagonist    ✓  Urine sample   Protein (semi-quantitative)  ✓     Albumin (semi-quantitative)    ✓  Genetic data   Genotyping array  ✓(subset)     Whole exome sequence  ✓(subset)    Metabolomicsb   Amino acids  ✓(subset)     Acylcarnitines  ✓(subset)     Ceramides  ✓(subset)     Sphingolipids  ✓(subset)    a Imaging was performed only on a sub-sample of 805 Chinese older participants of the follow-up cohort. b Metabolites were measured using a targeted, high-throughput metabolic profile approach using an HPLC online electrospray ionization tandem mass spectrometry.56 Table 3 Summary of variables collected or derived from the cohort Variables  Baseline  Follow-up  Questionnaire   Demographics  ✓  ✓   Cigarette smoking history  ✓  ✓   Environmental tobacco smoke exposure  ✓  ✓   Alcohol consumption  ✓  ✓   Medication use  ✓  ✓   Medical history  ✓  ✓   Family medical history  ✓  ✓   Women’s health  ✓  ✓   Physical activity  ✓  ✓   Dietary information (FFQ)  ✓  ✓   Health related quality of life (SF-36)  ✓  ✓   Health-related quality of life (EQ-5D)    ✓   Cognitive function (Mini-Mental State Examination)    ✓   Family life    ✓   Sources of psychological stress    ✓   Kessler psychological distress scale (K10)    ✓  Anthropometric measures   Height  ✓  ✓   Weight  ✓  ✓   Waist circumference  ✓  ✓   Hip circumference  ✓  ✓   Skinfold thickness    ✓  Clinical assessment   Light touch assessment measured by monofilament  ✓  ✓   Assessment of foot proprioception by neurothesiometer  ✓  ✓   Brachial and ankle blood pressure  ✓  ✓   Central aortic systolic pressure and arterial pulse waveform indices (rAI, rAP, PRT)    ✓   Bone densitometrya    ✓   Cardiac computer tomographya    ✓   Body fat and lean mass composition by DEXAa    ✓   Subcutaneous abdominal and intra-abdominal fat by CTa    ✓   Electrocardiography    ✓   Hand grip strength    ✓  Blood sample   Fasting or random glucose  ✓  ✓   Creatinine  ✓  ✓   Triglycerides  ✓  ✓   Total cholesterol  ✓  ✓   High-density lipoprotein cholesterol  ✓  ✓   Low-density lipoprotein cholesterol  ✓  ✓   HbA1c  ✓  ✓   Cortisol  ✓  ✓   High-sensitivity C-reactive protein  ✓  ✓   Insulin  ✓     Adiponectin  ✓     Interleukin-6    ✓   Interleukin-1 receptor antagonist    ✓  Urine sample   Protein (semi-quantitative)  ✓     Albumin (semi-quantitative)    ✓  Genetic data   Genotyping array  ✓(subset)     Whole exome sequence  ✓(subset)    Metabolomicsb   Amino acids  ✓(subset)     Acylcarnitines  ✓(subset)     Ceramides  ✓(subset)     Sphingolipids  ✓(subset)    Variables  Baseline  Follow-up  Questionnaire   Demographics  ✓  ✓   Cigarette smoking history  ✓  ✓   Environmental tobacco smoke exposure  ✓  ✓   Alcohol consumption  ✓  ✓   Medication use  ✓  ✓   Medical history  ✓  ✓   Family medical history  ✓  ✓   Women’s health  ✓  ✓   Physical activity  ✓  ✓   Dietary information (FFQ)  ✓  ✓   Health related quality of life (SF-36)  ✓  ✓   Health-related quality of life (EQ-5D)    ✓   Cognitive function (Mini-Mental State Examination)    ✓   Family life    ✓   Sources of psychological stress    ✓   Kessler psychological distress scale (K10)    ✓  Anthropometric measures   Height  ✓  ✓   Weight  ✓  ✓   Waist circumference  ✓  ✓   Hip circumference  ✓  ✓   Skinfold thickness    ✓  Clinical assessment   Light touch assessment measured by monofilament  ✓  ✓   Assessment of foot proprioception by neurothesiometer  ✓  ✓   Brachial and ankle blood pressure  ✓  ✓   Central aortic systolic pressure and arterial pulse waveform indices (rAI, rAP, PRT)    ✓   Bone densitometrya    ✓   Cardiac computer tomographya    ✓   Body fat and lean mass composition by DEXAa    ✓   Subcutaneous abdominal and intra-abdominal fat by CTa    ✓   Electrocardiography    ✓   Hand grip strength    ✓  Blood sample   Fasting or random glucose  ✓  ✓   Creatinine  ✓  ✓   Triglycerides  ✓  ✓   Total cholesterol  ✓  ✓   High-density lipoprotein cholesterol  ✓  ✓   Low-density lipoprotein cholesterol  ✓  ✓   HbA1c  ✓  ✓   Cortisol  ✓  ✓   High-sensitivity C-reactive protein  ✓  ✓   Insulin  ✓     Adiponectin  ✓     Interleukin-6    ✓   Interleukin-1 receptor antagonist    ✓  Urine sample   Protein (semi-quantitative)  ✓     Albumin (semi-quantitative)    ✓  Genetic data   Genotyping array  ✓(subset)     Whole exome sequence  ✓(subset)    Metabolomicsb   Amino acids  ✓(subset)     Acylcarnitines  ✓(subset)     Ceramides  ✓(subset)     Sphingolipids  ✓(subset)    a Imaging was performed only on a sub-sample of 805 Chinese older participants of the follow-up cohort. b Metabolites were measured using a targeted, high-throughput metabolic profile approach using an HPLC online electrospray ionization tandem mass spectrometry.56 Baseline measurements Participants completed an interviewer-administered questionnaire at their home, which took approximately an hour. Questionnaires were available in English, Chinese and Malay. Interviewers provided additional explanation when necessary in other languages common to both the participant and the interviewer. Participants were interviewed about their sociodemographic characteristics, personal and family medical history and lifestyle behaviours such as alcohol consumption, sedentary behaviours, cigarette smoking and exposure to environmental tobacco smoke. In addition, all medications used by the participants were recorded. Dietary intake was assessed by a semi-quantitative 169-item food frequency questionnaire (FFQ) that is also used in the Singapore National Nutrition Surveys.11 The physical activity questionnaire asked about the type, frequency and duration of various activities in the transportation, occupation, leisure time and household domains. Transportation activities included walking and cycling, and occupational activities included light, moderate and vigorous occupational activities. Leisure time activities included 48 specific activities, household activities included 15 specific activities, and options for additional activities not provided in the list were recorded through open-ended questions. Based on the physical activity questionnaire12 and FFQ,11 derived variables such as weekly physical activity level (MET-h/week) and intake of macronutrients, micronutrients and daily servings by food groups were calculated. Both questionnaires were validated in the local population.11,12 Nutrient intakes estimated from the FFQ were reasonably well correlated with estimates from repeated 24-h recalls. In addition, associations between food intakes and plasma fatty acids indicated that sources of several fatty acids were captured well by the FFQ.13 For physical activity, vigorous activity agreed well with accelerometer measurements, whereas modest agreement was observed for moderate-intensity activity. Participants were given the option to fast for the health examination. For those who did, they were told to fast for 8–12 h before their appointment. Fasting status was recorded by phlebotomists who performed venepuncture to collect blood for both biochemical analyses and biobanking. Random spot urine samples, collected mid-stream, were measured semi-quantitatively for protein level and any excess was stored in the biobank. Height was measured without shoes on a portable stadiometer (SECA 200 series, Germany) with the head in the Frankfurt Plane position. Participants were instructed to remove objects from their pockets such as wallets, keys and mobile phones before being measured for weight on SECA digital scales (SECA 700 series, Germany). A stretch-resistant tape was used to measure waist circumference at the mid-point between the last rib and iliac crest, and hip circumference at the greater trochanter of the femur. Systolic and diastolic blood pressures were measured after participants rested for 5 min, using an automated digital monitor to obtain two readings (Dinamap Carescape V100, General Electric). A third reading was performed if the difference between the first two readings was greater than 10 mmHg or 5 mmHg for systolic blood pressure and diastolic blood pressure, respectively. For the minority of the participants whose blood pressure exceeded the range of the digital monitor, a sphygmomanometer (Accoson, UK) was used. The ankle-brachial index was calculated to assess risk of peripheral artery disease, as follows. The systolic blood pressure from the brachial artery in the right arm and from the dorsalis pedis or posterior tibial artery in both ankles were measured using sphygmomanometers (Accoson, UK) in the supine position (head elevated at 15° to 20°). A handheld Doppler (Hadeco, Japan) was used to detect the pulses at the arteries. If the pulse at the dorsalis pedis artery could not be detected, the posterior tibial artery would be probed. Blood pressure in the left arm would be measured if it could not be done in the right arm. All the systolic blood pressure measurements were repeated once. Peripheral neuropathy was measured using neurothesiometer and monofilament tests. Participants lay almost supine for comfort, with their feet bare and eyes closed for the sensory functions assessment on their feet. Foot proprioception was assessed with a neurothesiometer (Horwell, UK) where a vibration-emitting probe was applied to the apex of the big toe and medial malleolus of both feet. The voltage was gradually increased from zero until the participants indicated verbally that they could feel the vibration. The voltage reading was recorded for both sites and both feet. Light touch assessment was performed by using a 10g (5.07) monofilament (Sensory Testing System, USA) on five least calloused plantar sites per foot—the distal great toe, third toe and fifth toe and the first and fifth metatarsal heads. The number of sites that the participants could feel was recorded for each foot. Follow-up measurements In the cohort follow-up, the health examination included four additional procedures, including psychological distress (Kessler K10),14,15 cognitive function (Mini Mental State Examination),16 body composition from skinfold thickness, and hand grip strength. Using a Holtain/Tanner skinfold caliper, skinfold thickness was measured at the left triceps, left biceps, sub-scapular, supra-iliac and calf regions. Three measurements were made on each site while the participant was standing. An electrocardiogram, 10 leads (Nihon Kohden ECG-1350K, Japan), was performed on each participant resting supine with head slightly elevated about 15° to 20°. For the measurement of the central aortic systolic pressure and arterial pulse waveform (A-PULSE CASPro Lite, HealthSTATS, Singapore), measurements were performed predominantly on the left arm with the participant seated and the left arm resting on a table at chest level. To assess their hand grip strength, participants were asked to stand with their arms free at the side and to grip the hand dynamometer (TAKEI A5401, Japan) as hard as they could with one hand. Three readings were recorded for each arm. A sub-sample of the follow-up cohort (805 Chinese, aged 50 and older) underwent more intensive assessments which included bone densitometry and body fat and lean mass composition, and cardiac computed tomography (CT).17 Biochemical analyses and biobanking Blood and urine samples were kept at the collection site at 4°C and transported at 4–8°C by a van to the biochemistry and biobank laboratories within 4–6 h of sample collection. Biochemical analyses of the blood samples were performed on the same day as collection. The biobank laboratory processed the blood and urine samples on the same day except for urine samples that were collected in the afternoon, which would be stored at 4°C and processed the next day. Blood samples were analysed at the biochemistry laboratory of the National University Hospital from April 2007 to October 2007 and November 2010 to August 2016. Analyses were performed at the Singapore General Hospital biochemistry laboratory for the period October 2007–November 2010. Both laboratories are accredited by the College of American Pathologists. Fresh blood samples were analysed on the day of collection for glucose, lipids, creatinine and HbA1c. Supplementary Table 1 (available as Supplementary data at IJE online) provides laboratory details [analyser platform, assay methodology, measurement range and coefficients of variation (within and between day)] of the fresh blood analyses reported by the laboratories. As of present, the biological specimens are stored in the tissue repositories of the National University Hospital and Saw Swee Hock School of Public Health. Supplementary Table 2 (available as Supplementary data at IJE online) shows the types of biospecimen stored. Genome-wide single nucleotide polymorphism (SNP) data are available for a subset of the MEC cohort from a combination of Illumina genome-wide genotyping arrays and customized arrays such as Oncoarray18 and iCOGS.19 Samples genotyped on the genome-wide arrays have been imputed to 1000G Phase 3 reference panels.20 Whole-exome sequence data are also available for some of the participants.21 Quality control and data processing Interviewers and personnel who performed the health examination were trained and assessed for competency before they could begin data collection. This was to ensure standardized interviewing and consistency in performing the health examination and collection of samples. As the cohort is multi-ethnic in nature, the interviewers and personnel are proficient in at least one of the local languages (English, Chinese, Malay, Chinese dialects) verbally. Completed questionnaires were sampled for verification with 20% of the participants on the data collected. All data forms, including consent forms, were inspected for missing data before they were subjected to double data entry by two independent staff. A third person resolved contradictions between entries by verifying with the source document. The data forms were optically scanned and archived following the completion of data entry. Research data were de-identified, key-coded and maintained in research databases accessible only to database administrators. Database administrators performed further data cleaning of the research data by checking for consistency between related variables and data range before releasing the data for research analysis. Personal information of participants and study visit records were maintained in different databases that only the fieldwork management team could access for the purpose of follow-up. What has it found? Key findings and publications This large multi-ethnic cohort has provided us an opportunity to examine differences in health conditions, HRQoL and behaviours between major Asian ethnic groups within a relatively homogeneous environment and access to health care in Singapore. The prevalence of T2D and other metabolic-related outcomes is increasing worldwide, with the greatest increases predicted in East (Chinese and Malays) and South Asian (Indians) populations in Asia.22 Within the MEC study, we compare and evaluate different metabolic biomarkers and their contributions to disease outcomes between the three ethnic groups. For example, Indians had the greatest insulin resistance, followed by Malays and lastly Chinese. The difference between Chinese and Malays could be explained by differences in body mass index (BMI). However, the greater insulin resistance in Indians could only be partly explained by higher BMI, waist circumference and inflammation.23 In addition, BMI had a greater influence on insulin resistance, C-reactive protein and adiponectin in Chinese than the other major Asian ethnic groups.24 These suggest that ethnicity modifies the relationship between adiposity and pathways (e.g. insulin resistance, inflammation and adiponectin) involved in the pathogenesis of T2D and cardiovascular diseases (CVD), and may lead to greater increases in T2D and CVD in Chinese than other ethnic groups as a result of increasing BMI. Longitudinal follow-up in the cohort allows us to examine the utility of conventional diagnostic criteria in their prediction of long-term prognosis and complications. When we examined if the current T2D diagnostic criteria based on fasting glucose levels reflect peripheral neuropathy and chronic kidney disease, we showed that both microvascular complications gradually increased in relation to fasting plasma glucose, beginning at levels below the existing diagnostic threshold for T2D of 7.0 mmol/l. For chronic kidney disease, these associations persisted after adjustment for other risk factors. These findings suggest that current diagnostic thresholds for T2D may have limited sensitivity for identifying individuals with microvascular complications in an Asian population.1 Another prospective study examined changes in weight and waist circumference over the follow-up period and their associations with demographic and socioeconomic factors.25 Those who gained the most weight were younger adults, likely to be of higher socioeconomic status (SES) and ethnic minority group, and had the lowest BMI at baseline. The data provided evidence for obesity prevention strategies to be initiated at a younger age and with specific considerations for the minority ethnic and higher SES groups. Record linkage with national registries enabled the identification of cases with ischaemic heart disease (IHD) in the cohort. We showed that not only did Indians have a higher susceptibility for developing T2D than Chinese and Malays, but also the risk of IHD in Indians with T2D was higher than the other groups.2 In another study, we demonstrated that having metabolic syndrome with or without central obesity confers the same risk for IHD in Asians. We went on to show that when using the International Diabetes Federation definition of metabolic syndrome, which includes central obesity as an essential component, considerably fewer individuals at risk of IHD in this cohort would be identified than if central obesity was regarded as an optional component as defined by the American Heart Association/National Heart, Lung, and Blood Institute.26 With detailed validated questionnaires on dietary intakes and physical activity, we can assess associations of dietary and physical activity patterns with health outcomes including HRQoL, and how these associations differ across ethnic groups. We observed that a combination of brief television watching and regular vigorous physical activity was strongly associated with lower insulin resistance.27 Interestingly, the association between television watching and insulin resistance appeared to be mediated by eating habits and adiposity rather than lower physical activity. With regards to dietary factors, coffee consumption was associated with less insulin resistance, whereas high consumption of rice and noodles was associated with greater insulin resistance and hyperglycaemia.28,29 Using the Short Form 36 version 2 (SF-36v2) and Short Form 6D (SF-6D) instrument surveys validated in MEC,30 Leow et al. found that HRQoL levels differed between ethnic groups: Chinese had higher physical health component summary scores than Malays and Indians, whereas Malays displayed higher mental health component summary scores than Chinese.31 The same study also discovered that the effects of gender, stroke and musculoskeletal conditions on HRQoL may vary by ethnicity. Other studies have further examined the relationship between HRQoL and T2D and its complications and comorbidities. Venkataraman et al. found that the HRQoL of patients with T2D was affected by complications and not T2D per se, with peripheral neuropathy even in mild form being associated with the greatest reduction in HRQoL.32 These findings support the improvement of T2D management to prevent or delay the onset of complications. In patients with T2D, hypertension and dyslipidaemia, disease awareness was found to be associated with lower HRQoL whereas undiagnosed disease was not, which may partly explain why the undiagnosed did not seek medical care.33 Understanding these key modifiable lifestyle patterns can help frame lifestyle recommendations, targeted interventions and health promotion policies. Insights as to why people remain undiagnosed with chronic diseases can help inform strategies for preventive health screening. Another study by Venkataraman et al. identified Malay ethnicity, lower educational status, lower family cohesion and unhealthy diet and physical activity patterns as some key barriers to participation in health screening. It suggested enhancing the cultural competence of preventive health services, as one of the strategies to improve the health screening participation of these groups and reduce the proportions in the population that remain undiagnosed with chronic diseases.34 Finally, the diverse ethnic groups represented in MEC enabled mapping genetic diversity across these populations35 and discovery of genetic variants associated with various disease outcomes and related biomarkers. For example, a common genetic variant (10%) at PAX4 gene was found to be associated with T2D in East Asians, whereas the variant is extremely rare in other global populations.21,36,37 We have led and contributed to multiple discovery efforts in East and South Asians and trans-ethnic large-scale genetic association analyses, and identified many genetic variants associated with blood pressure,38–40 anthropometric measures,41–44 kidney function,45 adiponectin levels,46 glycaemic traits47–49 and blood lipid profiles.50,51 In addition to metabolic disorders, MEC participants have been ethnicity-matched to breast cancer cases of a patient cohort to act as controls in genetic association studies.52,53 About 80 scientific papers based on the MEC data have been published up to May 2017 [https://blog.nus.edu.sg/sphs/publications/]. Data have also been shared for international collaborative projects such as the Non-Communicable Disease Risk Factor Collaboration, formerly known as the Global Burden of Metabolic Risk Factors of Chronic Diseases Collaborating Group, and the Asia Pacific Cohort Studies Collaboration, culminating in reports on aetiological associations and regional and global trends in cardiovascular diseases, obesity and T2D.54,55 What are the main strengths and weaknesses? A major strength of the MEC study is that it represents three major ethnic groups in Asia who are at risk of major health outcome changes. The extensive interviewer-administered questionnaire covered a wide range of health determinants, medical history and quality of life with minimal missing data. The physical activity and FFQ used in the MEC study were validated in independent samples of the local population with similar characteristics.11,12 The MEC study also included a detailed physical examination, biochemical analyses of blood samples and storage of blood and spot urine samples. A potential limitation of the MEC cohort was the substantial non-response during the follow-up visit. This can partly be resolved by passive follow-up using linkage to disease registries that have good coverage of the Singapore population. Can I get hold of the data? Where can I find out more? Researchers can visit the Saw Swee Hock School of Public Health website [https://blog.nus.edu.sg/sphs/] for information on submitting a request for data and/or samples. Supplementary Data Supplementary data are available at IJE online. Profile in a nutshell Singapore is a multi-ethnic island state and provides a useful model to evaluate determinants of the development of chronic diseases in Asian ethnic groups. Baseline recruitment was done between 2004 and 2010. The cohort includes 14 465 male and female participants aged 21 years and above, with 47% Chinese, 26% Malay and 27% Indian ethnicity. The first follow-up began in 2011 and ended in 2016; 28% of the baseline participants were not contactable and 2.5% were confirmed to have been lost to follow-up. Of the participants that were contactable, 60% agreed to participate in the follow-up survey. Sociodemographic, lifestyle behaviours, personal and family medical history, medication use and health-related quality of life information was collected at baseline. Measurements of anthropometric characteristics, peripheral neuropathy, brachial and ankle blood pressure and biomarkers in blood and urine were also collected. During the follow-up, additional information on mental health and cognitive function was collected and electrocardiogram, central blood pressure and hand grip strength measurements were conducted; 86% of all baseline participants gave consent for data linkage, allowing disease follow-up through medical records and national registries. Researchers can visit the Saw Swee Hock School of Public Health website [https://blog.nus.edu.sg/sphs/] for information on submitting a request for data and/or samples. Author Contributions K.H.X.T. contributed to the analysis and interpretation of data and drafted the manuscript. L.W.L.T. contributed to design of the study, management of the study and drafted the manuscript. J.J.M.L., K.S.C. and E.S.T. designed the study. X.S. and R.M.V.D. critically revised the manuscript. Acknowledgements We are grateful to the Singapore residents who volunteered time, data and samples for this study and to the community leaders who had facilitated our recruitment drives. We wish to thank the fieldwork and data management team for their dedicated work on this study. Funding The MEC study was supported by the National Medical Research Council (grant 0838/2004), Biomedical Research Council (grants 03/1/27/18/216, 05/1/21/19/425 and 11/1/21/19/678), Ministry of Health, Singapore, National University of Singapore and National University Health System, Singapore. Conflict of interest: None declared. References 1 Nang EE, Khoo CM, Tai ES et al.   Is there a clear threshold for fasting plasma glucose that differentiates between those with and without neuropathy and chronic kidney disease?: the Singapore Prospective Study Program. Am J Epidemiol  2009; 169: 1454– 62. Google Scholar CrossRef Search ADS PubMed  2 Yeo KK, Tai BC, Heng D et al.   Ethnicity modifies the association between diabetes mellitus and ischaemic heart disease in Chinese, Malays and Asian Indians living in Singapore. Diabetologia  2006; 49: 2866– 73. Google Scholar CrossRef Search ADS PubMed  3 Lee J, Heng D, Ma S, Chew SK, Hughes K, Tai ES. The metabolic syndrome and mortality: the Singapore Cardiovascular Cohort Study. Clin Endocrinol (Oxf)  2008; 69: 225– 30. Google Scholar CrossRef Search ADS PubMed  4 Hughes K, Yeo PP, Lun KC et al.   Cardiovascular diseases in Chinese, Malays, and Indians in Singapore. II. Differences in risk factor levels. J Epidemiol Community Health  1990; 44: 29– 35. Google Scholar CrossRef Search ADS PubMed  5 Tan CE, Emmanuel SC, Tan BY, Jacob E. Prevalence of diabetes and ethnic differences in cardiovascular risk factors. The 1992 Singapore National Health Survey. Diabetes Care  1999; 22: 241– 47. Google Scholar CrossRef Search ADS PubMed  6 Hughes K, Aw TC, Kuperan P, Choo M. Central obesity, insulin resistance, syndrome X, lipoprotein(a), and cardiovascular risk in Indians, Malays, and Chinese in Singapore. J Epidemiol Community Health  1997; 51: 394– 99. Google Scholar CrossRef Search ADS PubMed  7 Cutter J, Tan BY, Chew SK. Levels of cardiovascular disease risk factors in Singapore following a national intervention programme. Bull World Health Organ  2001; 79: 908– 15. Google Scholar PubMed  8 WHO Expert Consultation. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet  2004; 363: 157– 63. CrossRef Search ADS PubMed  9 World Health Organization. Waist Circumference and Waist-Hip Ratio: Report of a WHO Expert Consultation . Geneva: WHO, 2008. 10 Ainsworth BE, Haskell WL, Herrmann SD et al.   2011 Compendium of Physical Activities: a second update of codes and MET values. Med Sci Sports Exerc  2011; 43: 1575– 81. Google Scholar CrossRef Search ADS PubMed  11 Deurenberg-Yap M, Li T, Tan WL, van Staveren WA, Deurenberg P. Validation of a semiquantitative food frequency questionnaire for estimation of intakes of energy, fats and cholesterol among Singaporeans. Asia Pac J Clin Nutr  2000; 9: 282– 88. Google Scholar CrossRef Search ADS PubMed  12 Nang EE, Gitau Ngunjiri SA, Wu Y et al.   Validity of the International Physical Activity Questionnaire and the Singapore Prospective Study Program physical activity questionnaire in a multiethnic urban Asian population. BMC Med Res Methodol  2011; 11: 141. Google Scholar CrossRef Search ADS PubMed  13 Seah JY, Gay GM, Su J et al.   Consumption of red meat, but not cooking oils high in polyunsaturated fat, is associated with higher arachidonic acid status in Singapore Chinese Adults. Nutrients  2017; 9: 101. Google Scholar CrossRef Search ADS   14 Kessler RC, Andrews G, Colpe LJ et al.   Short screening scales to monitor population prevalences and trends in non-specific psychological distress. Psychol Med  2002; 32: 959– 76. Google Scholar CrossRef Search ADS PubMed  15 Kessler RC, Barker PR, Colpe LJ et al.   Screening for serious mental illness in the general population. Arch Gen Psychiatry  2003; 60: 184– 89. Google Scholar CrossRef Search ADS PubMed  16 Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res  1975; 12: 189– 98. Google Scholar CrossRef Search ADS PubMed  17 Nang EE, van Dam RM, Tan CS et al.   Association of television viewing time with body composition and calcified subclinical atherosclerosis in Singapore Chinese. PLoS One  2015; 10: e0132161. Google Scholar CrossRef Search ADS PubMed  18 Amos CI, Dennis J, Wang Z et al.   The OncoArray Consortium: A network for understanding the genetic architecture of common cancers. Cancer Epidemiol Biomarkers Prev  2017; 26: 126– 35. Google Scholar CrossRef Search ADS PubMed  19 Sakoda LC, Jorgenson E, Witte JS. Turning of COGS moves forward findings for hormonally mediated cancers. Nat Genet  2013; 45: 345– 48. Google Scholar CrossRef Search ADS PubMed  20 Auton A, Brooks LD, Durbin RM et al.   A global reference for human genetic variation. Nature  2015; 526: 68– 74. Google Scholar CrossRef Search ADS PubMed  21 Fuchsberger C, Flannick J, Teslovich TM et al.   The genetic architecture of type 2 diabetes. Nature  2016; 536: 41– 47. Google Scholar CrossRef Search ADS PubMed  22 Whiting DR, Guariguata L, Weil C, Shaw J. IDF diabetes atlas: global estimates of the prevalence of diabetes for 2011 and 2030. Diabetes Res Clin Pract  2011; 94: 311– 21. Google Scholar CrossRef Search ADS PubMed  23 Gao H, Salim A, Lee J, Tai ES, van Dam RM. Can body fat distribution, adiponectin levels and inflammation explain differences in insulin resistance between ethnic Chinese, Malays and Asian Indians? Int J Obes (Lond)  2012; 36: 1086– 93. Google Scholar CrossRef Search ADS PubMed  24 Khoo CM, Sairazi S, Taslim S et al.   Ethnicity modifies the relationships of insulin resistance, inflammation, and adiponectin with obesity in a multiethnic Asian population. Diabetes Care  2011; 34: 1120– 26. Google Scholar CrossRef Search ADS PubMed  25 Ong SK, Fong CW, Ma S et al.   Longitudinal study of the socio-demographic determinants of changes in body weight and waist circumference in a multi-ethnic Asian population. Int J Obes (Lond)  2009; 33: 1299– 308. Google Scholar CrossRef Search ADS PubMed  26 Lee J, Ma S, Heng D et al.   Should central obesity be an optional or essential component of the metabolic syndrome? Ischemic heart disease risk in the Singapore Cardiovascular Cohort Study. Diabetes Care  2007; 30: 343– 47. Google Scholar CrossRef Search ADS PubMed  27 Nang EE, Salim A, Wu Y, Tai ES, Lee J, Van Dam RM. Television screen time, but not computer use and reading time, is associated with cardio-metabolic biomarkers in a multiethnic Asian population: a cross-sectional study. Int J Behav Nutr Phys Act  2013; 10: 70. Google Scholar CrossRef Search ADS PubMed  28 Rebello SA, Chen CH, Naidoo N et al.   Coffee and tea consumption in relation to inflammation and basal glucose metabolism in a multi-ethnic Asian population: a cross-sectional study. Nutr J  2011; 10: 61. Google Scholar CrossRef Search ADS PubMed  29 Zuniga YL, Rebello SA, Oi PL et al.   Rice and noodle consumption is associated with insulin resistance and hyperglycaemia in an Asian population. Br J Nutr  2014; 11: 1118– 28. Google Scholar CrossRef Search ADS   30 Tan ML, Wee HL, Lee J et al.   The Short Form 36 English and Chinese versions were equivalent in a multiethnic Asian population. J Clin Epidemiol  2013; 66: 759– 67. Google Scholar CrossRef Search ADS PubMed  31 Leow MK, Griva K, Choo R et al.   Determinants of Health-Related Quality of Life (HRQoL) in the multiethnic Singapore population - a national cohort study. PLoS One  2013; 8: e67138. Google Scholar CrossRef Search ADS PubMed  32 Venkataraman K, Wee HL, Leow MK et al.   Associations between complications and health-related quality of life in individuals with diabetes. Clin Endocrinol (Oxf)  2013; 7: 865– 73. Google Scholar CrossRef Search ADS   33 Venkataraman K, Khoo C, Wee HL et al.   Associations between disease awareness and health-related quality of life in a multi-ethnic Asian population. PLoS One  2014; 9: e113802. Google Scholar CrossRef Search ADS PubMed  34 Venkataraman K, Wee HL, Ng SH et al.   Determinants of individuals' participation in integrated chronic disease screening in Singapore. J Epidemiol Community Health . 2016 Jun 10. doi: 10.1136/jech-2016-207404. 35 Saw WY, Tantoso E, Begum H et al.   Establishing multiple omics baseline for three Southeast Asian ethnic groups in the Singapore Integrative Omics Cohort. Nat Commun  2017; 8: 653. Google Scholar CrossRef Search ADS PubMed  36 Cho YS, Chen CH, Hu C et al.   Meta-analysis of genome-wide association studies identifies eight new loci for type 2 diabetes in east Asians. Nat Genet  2011; 44: 67– 72. Google Scholar CrossRef Search ADS PubMed  37 Mahajan A, Go MJ, Zhang W et al.   Genome-wide trans-ancestry meta-analysis provides insight into the genetic architecture of type 2 diabetes susceptibility. Nat Genet  2014; 46: 234– 44. Google Scholar CrossRef Search ADS PubMed  38 Kato N, Takeuchi F, Tabara Y et al.   Meta-analysis of genome-wide association studies identifies common variants associated with blood pressure variation in east Asians. Nat Genet  2011; 43: 531– 38. Google Scholar CrossRef Search ADS PubMed  39 Kelly TN, Takeuchi F, Tabara Y et al.   Genome-wide association study meta-analysis reveals transethnic replication of mean arterial and pulse pressure loci. Hypertension  2013; 62: 853– 59. Google Scholar CrossRef Search ADS PubMed  40 Kato N, Loh M, Takeuchi F et al.   Trans-ancestry genome-wide association study identifies 12 genetic loci influencing blood pressure and implicates a role for DNA methylation. Nat Genet  2015; 47: 1282– 93. Google Scholar CrossRef Search ADS PubMed  41 Wen W, Cho YS, Zheng W et al.   Meta-analysis identifies common variants associated with body mass index in east Asians. Nat Genet  2012; 44: 307– 11. Google Scholar CrossRef Search ADS PubMed  42 Wen W, Zheng W, Okada Y et al.   Meta-analysis of genome-wide association studies in East Asian-ancestry populations identifies four new loci for body mass index. Hum Mol Genet  2014; 23: 5492– 504. Google Scholar CrossRef Search ADS PubMed  43 He M, Xu M, Zhang B et al.   Meta-analysis of genome-wide association studies of adult height in East Asians identifies 17 novel loci. Hum Mol Genet  2015; 24: 1791– 800. Google Scholar CrossRef Search ADS PubMed  44 Wen W, Kato N, Hwang JY et al.   Genome-wide association studies in East Asians identify new loci for waist-hip ratio and waist circumference. Sci Rep  2016; 6: 17958. Google Scholar CrossRef Search ADS PubMed  45 Okada Y, Sim X, Go MJ et al.   Meta-analysis identifies multiple loci associated with kidney function-related traits in east Asian populations. Nat Genet  2012; 44: 904– 09. Google Scholar CrossRef Search ADS PubMed  46 Wu Y, Gao H, Li H et al.   A meta-analysis of genome-wide association studies for adiponectin levels in East Asians identifies a novel locus near WDR11-FGFR2. Hum Mol Genet  2014; 23: 1108– 19. Google Scholar CrossRef Search ADS PubMed  47 Chen P, Takeuchi F, Lee JY et al.   Multiple nonglycemic genomic loci are newly associated with blood level of glycated hemoglobin in East Asians. Diabetes  2014; 63: 2551– 62. Google Scholar CrossRef Search ADS PubMed  48 Hwang JY, Sim X, Wu Y et al.   Genome-wide association meta-analysis identifies novel variants associated with fasting plasma glucose in East Asians. Diabetes  2015; 64: 291– 98. Google Scholar CrossRef Search ADS PubMed  49 Chen P, Ong RT, Tay WT et al.   A study assessing the association of glycated hemoglobin A1C (HbA1C) associated variants with HbA1C, chronic kidney disease and diabetic retinopathy in populations of Asian ancestry. PLoS One  2013; 8: e79767. Google Scholar CrossRef Search ADS PubMed  50 Teslovich TM, Musunuru K, Smith AV et al.   Biological, clinical and population relevance of 95 loci for blood lipids. Nature  2010; 466: 707– 13. Google Scholar CrossRef Search ADS PubMed  51 Spracklen CN, Chen P, Kim YJ et al.   Association analyses of East Asian individuals and trans-ancestry analyses with European individuals reveal new loci associated with cholesterol and triglyceride levels. Hum Mol Genet  2017; 26: 1770– 84. Google Scholar CrossRef Search ADS PubMed  52 Michailidou K, Hall P, Gonzalez-Neira A et al.   Large-scale genotyping identifies 41 new loci associated with breast cancer risk. Nat Genet  2013; 45: 353– 61, 61e1-2. Google Scholar CrossRef Search ADS PubMed  53 Garcia-Closas M, Couch FJ, Lindstrom S et al.   Genome-wide association studies identify four ER negative-specific breast cancer risk loci. Nat Genet  2013; 45: 392– 98, 8e1-2. Google Scholar CrossRef Search ADS PubMed  54 NCD Risk Factor Collaboration (NCD-RisC). Worldwide trends in blood pressure from 1975 to 2015: a pooled analysis of 1479 population-based measurement studies with 19.1 million participants. Lancet  2017; 389: 37– 55. CrossRef Search ADS PubMed  55 Woodward M, Huxley R, Ueshima H, Fang X, Kim HC, Lam TH. The Asia Pacific cohort studies collaboration: a decade of achievements. Glob Heart  2012; 7: 343– 51. Google Scholar CrossRef Search ADS PubMed  56 Haus JM, Kashyap SR, Kasumov T et al.   Plasma ceramides are elevated in obese subjects with type 2 diabetes and correlate with the severity of insulin resistance. Diabetes  2009; 58: 337– 43. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png International Journal of Epidemiology Oxford University Press

Loading next page...
 
/lp/ou_press/cohort-profile-the-singapore-multi-ethnic-cohort-mec-study-GTWGb0RtuI
Publisher
Oxford University Press
Copyright
© The Author(s) 2018; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association
ISSN
0300-5771
eISSN
1464-3685
D.O.I.
10.1093/ije/dyy014
Publisher site
See Article on Publisher Site

Abstract

Why was the cohort set up? Non-communicable diseases such as type 2 diabetes (T2D) mellitus, coronary artery disease, stroke and cancers, are major contributors to ill health across the world including Asia. These conditions are multi-factorial in origin, often involving complex gene-environment interactions. Singapore is a multi-ethnic island state and provides a useful model to evaluate determinants of the development of chronic diseases in Asian ethnic groups. Three major Asian ethnic groups are represented in Singapore: Chinese, Malays and Indians. The Singapore Multi-Ethnic Cohort (MEC) allows us to better understand how genes and lifestyle may influence health and diseases differently in persons of Chinese, Malay and Indian ethnicity. As these ethnic groups reside in the same Singapore setting, confounding of ethnic differences by differences between countries is avoided. Through the MEC, we hope to improve preventive and therapeutic measures, as well as provide information to advance public health and health education policies for Asian populations. Who is in the cohort? The MEC is a closed cohort which included a total of 14 465 male and female adults. The cohort was formed by combining two existing population-based studies with measurements in the period 2004 to 2007, the Singapore Prospective Study Program (SP2) and the Singapore Cardiovascular Cohort Study (SCCS2), with additional recruitment of participants from 2007 to 2010.1–3 The combined cohort has a good representation of Chinese, Malay and Indian ethnic groups. The SP2 and SCCS2 recruited 8340 participants from four previous cross-sectional studies: Thyroid and Heart Study 1982–84,4 National Health Survey 1992,5 National University of Singapore Heart Study 1993–956 and National Health Survey 1998.7 All studies involved a random sample of Singapore residents aged 21 years and above, with disproportionate sampling stratified by ethnicity to increase the numbers for ethnic minorities, i.e. Malays and Indians. In addition to the participants from SP2 and SCCS2, a further 6125 Singapore residents were recruited into the MEC study through public outreach and referrals from existing cohort members. Invitation to participate was open to any Singapore citizens or long-term residents of age 21 to 75 years. People with a history of heart disease, stroke, cancer and renal failure were excluded at baseline, as these are the outcomes of interest in this prospective cohort for non-communicable diseases. Recruitment drives were carried out at community events, mosques and temples in addition to household visitation to enrich the proportion of Malays and Indians. Ethics approval for the SP2 was provided by the SingHealth Centralised Institutional Review Board (IRB). The rest of the MEC and the follow-up were approved by the National University of Singapore IRB. Written consent was obtained for registry and medical records linkage, future analysis of stored biological samples and future follow-up. All MEC participants were visited at home to complete an interview, and were subsequently invited to undergo a health screening for clinical assessments and collection of blood and urine samples to be stored for future analysis. Of the 14 465 participants of the MEC study, 77% (N = 11 085) completed both the interviews and health screening, whereas the remaining 23% (N = 3380) only completed the interview-administered questionnaire. The sociodemographic and socioeconomic profiles of participants are presented in Table 1. The age of the participants ranged from 21 to 94 years and the mean age was 46 ± 13 years; 56% of the participants were females, and the ethnic composition was 47% Chinese, 26% Malay and 27% Indian. Most of the participants were married (75%). In terms of highest qualification attained, 30% of the participants received primary or lower education, 35% received secondary education, 6% received vocational training and 29% received post-secondary or higher education. Around 89% of the participants lived in public housing and 11% lived in private housing; this is reflective of the general Singapore resident population where the majority resides in government housing. Table 1 Demographic profile of participants at baseline (N = 14 465) Characteristics  Number  Percentage  Age at interview (in years), mean ± SD  46.1  13.3  Gender   Male  6353  43.9   Female  8112  56.1  Ethnicity   Chinese  6814  47.1   Malay  3773  26.1   Indian  3837  26.5   Others  41  0.3  Marital status   Never married  2454  17.0   Currently married  10776  74.5   Separated/divorced  516  3.6   Widowed  710  4.9  Educational statusa   No formal qualification  1313  9.1   Primary  3003  20.8   Secondary  5095  35.3   Vocational training  855  5.9   Post-secondary  2412  16.7   University and above  1775  12.3  Monthly household income (SGD)   Less than $2000  3228  30.4   $2000–$4000  3500  33.0   $4000–$6000  2053  19.4   $6000–$10,000  1292  12.2   More than $10,000  529  5.0  Housing type   Public housing 1–3-room flat  3420  23.7   Public housing 4-room flat  5524  38.2   Public housing 5-room or executive flat  3990  27.6   Private condominium  797  5.5   Landed property  717  5.0  Characteristics  Number  Percentage  Age at interview (in years), mean ± SD  46.1  13.3  Gender   Male  6353  43.9   Female  8112  56.1  Ethnicity   Chinese  6814  47.1   Malay  3773  26.1   Indian  3837  26.5   Others  41  0.3  Marital status   Never married  2454  17.0   Currently married  10776  74.5   Separated/divorced  516  3.6   Widowed  710  4.9  Educational statusa   No formal qualification  1313  9.1   Primary  3003  20.8   Secondary  5095  35.3   Vocational training  855  5.9   Post-secondary  2412  16.7   University and above  1775  12.3  Monthly household income (SGD)   Less than $2000  3228  30.4   $2000–$4000  3500  33.0   $4000–$6000  2053  19.4   $6000–$10,000  1292  12.2   More than $10,000  529  5.0  Housing type   Public housing 1–3-room flat  3420  23.7   Public housing 4-room flat  5524  38.2   Public housing 5-room or executive flat  3990  27.6   Private condominium  797  5.5   Landed property  717  5.0  a Educational status: secondary education (‘O’/‘N’ level), vocational training (attended Institute of Technical Education or obtained National Technical Certificate) and post-secondary education (‘A’ level, polytechnic/diploma). Missings marital status (n = 9), educational status (n = 12), monthly household income (n = 3863) and housing type (n = 17). Table 1 Demographic profile of participants at baseline (N = 14 465) Characteristics  Number  Percentage  Age at interview (in years), mean ± SD  46.1  13.3  Gender   Male  6353  43.9   Female  8112  56.1  Ethnicity   Chinese  6814  47.1   Malay  3773  26.1   Indian  3837  26.5   Others  41  0.3  Marital status   Never married  2454  17.0   Currently married  10776  74.5   Separated/divorced  516  3.6   Widowed  710  4.9  Educational statusa   No formal qualification  1313  9.1   Primary  3003  20.8   Secondary  5095  35.3   Vocational training  855  5.9   Post-secondary  2412  16.7   University and above  1775  12.3  Monthly household income (SGD)   Less than $2000  3228  30.4   $2000–$4000  3500  33.0   $4000–$6000  2053  19.4   $6000–$10,000  1292  12.2   More than $10,000  529  5.0  Housing type   Public housing 1–3-room flat  3420  23.7   Public housing 4-room flat  5524  38.2   Public housing 5-room or executive flat  3990  27.6   Private condominium  797  5.5   Landed property  717  5.0  Characteristics  Number  Percentage  Age at interview (in years), mean ± SD  46.1  13.3  Gender   Male  6353  43.9   Female  8112  56.1  Ethnicity   Chinese  6814  47.1   Malay  3773  26.1   Indian  3837  26.5   Others  41  0.3  Marital status   Never married  2454  17.0   Currently married  10776  74.5   Separated/divorced  516  3.6   Widowed  710  4.9  Educational statusa   No formal qualification  1313  9.1   Primary  3003  20.8   Secondary  5095  35.3   Vocational training  855  5.9   Post-secondary  2412  16.7   University and above  1775  12.3  Monthly household income (SGD)   Less than $2000  3228  30.4   $2000–$4000  3500  33.0   $4000–$6000  2053  19.4   $6000–$10,000  1292  12.2   More than $10,000  529  5.0  Housing type   Public housing 1–3-room flat  3420  23.7   Public housing 4-room flat  5524  38.2   Public housing 5-room or executive flat  3990  27.6   Private condominium  797  5.5   Landed property  717  5.0  a Educational status: secondary education (‘O’/‘N’ level), vocational training (attended Institute of Technical Education or obtained National Technical Certificate) and post-secondary education (‘A’ level, polytechnic/diploma). Missings marital status (n = 9), educational status (n = 12), monthly household income (n = 3863) and housing type (n = 17). Lifestyle, anthropometric measures and biomarkers profiles of study participants, stratified by ethnicity, are presented in Table 2. There were more current smokers in the Malays (24%) than Indians (17%) and Chinese (12%). Self-reported prevalence of T2D was the highest amongst the Indians (16%), followed by Malays (10%) and Chinese (7%). Similarly, Indians had higher HbA1c (6.2 ± 1.3%) than Malays (6.0 ± 1.3%) and Chinese (5.7 ± 0.8%). Prevalence of self-reported hypertension was similar across the three ethnic groups (18–19%, P-value = 0.10), though mean systolic and diastolic blood pressures were higher in Malays and Chinese than Indians. The Chinese had the highest prevalence of self-reported hypercholesterolaemia (Chinese: 28%; Indians: 23%; Malays: 19%), but mean low-density lipoprotein cholesterol (LDL) was the highest in the Malays (Malays: 3.5 ± 0.9 mmol/l; Indians: 3.4 ± 0.9 mmol/l; Chinese: 3.2 ± 0.8 mmol/l). Table 2 Lifestyle, anthropometric measures, and biomarkers profiles of study participants at baseline by ethnicity (N = 14 424) Characteristics  Chinese  Malays  Indians  P-value  Cigarette smoking status (%)   Never smoker  5503 (80.8)  2604 (69.0)  2945 (76.8)  < 0.001   Ex-smoker  514 (7.5)  282 (7.5)  234 (6.1)   Current smoker  797 (11.7)  887 (23.5)  657 (17.1)  Self-reported health conditions (%)   T2Da  436 (6.5)  364 (9.8)  588 (15.6)  < 0.001   Hypertensiona  1285 (19.2)  652 (17.6)  685 (18.2)  0.098   Hypercholesterolaemiaa  1755 (27.5)  672 (18.8)  824 (22.6)  < 0.001  Physical activity (MET-h/week)b, median (IQR)   Moderate- and/or vigorous-intensity leisure-time  5.5 (0.0–15.5)  4.5 (0.0–15.8)  3.5 (0.0–14.0)  < 0.001  Physical activity (MET-h/week)b, mean ± SD   Moderate- and/or vigorous-intensity leisure-time  11.9 ± 23.1  12.8 ± 23.9  11.4 ± 21.1  0.479  Macro nutrient intake (% energy)   Carbohydrate  53.8 ± 6.8  54.5 ± 6.7  54.3 ± 6.5  < 0.001   Protein  15.8 ± 2.3  14.1 ± 2.2  13.2 ± 2.1  < 0.001   Fat  30.1 ± 5.7  31.3 ± 5.8  31.8 ± 5.8  < 0.001  BMI category (kg/m2), (%)   < 18.5  444 (8.6)  113 (3.9)  127 (4.2)  < 0.001   18.5–23.0  2317 (45.1)  621 (21.7)  682 (22.5)   23.0–27.5  1778 (34.6)  1021 (35.7)  1172 (38.7)   ≥ 27.5  603 (11.7)  1107 (38.7)  1050 (34.6)  Waist circumference (cm), (%)   Normal  3492 (67.9)  1287 (45.1)  1155 (38.1)  < 0.001   Highc  1650 (32.1)  1569 (54.9)  1878 (61.9)  Blood pressure (mmHg), mean ± SD   Systolic  127.5 ± 20.6  129.0 ± 20.8  124.5 ± 21.3  < 0.001   Diastolic  75.0 ± 11.3  74.4 ± 11.3  73.0 ± 11.1  < 0.001  Biomarkers, mean ± SD   Triglycerides (mmol/l)  1.3 ± 0.8  1.4 ± 1.1  1.4 ± 1.0  < 0.001   High-density lipoprotein (mmol/l) (HDL)  1.4 ± 0.4  1.2 ± 0.3  1.1 ± 0.3  < 0.001   Low-density lipoprotein (mmol/l) (LDL)  3.2 ± 0.8  3.5 ± 0.9  3.4 ± 0.9  < 0.001   HbA1c (%)  5.7 ± 0.8  6.0 ± 1.3  6.2 ± 1.3  < 0.001  Characteristics  Chinese  Malays  Indians  P-value  Cigarette smoking status (%)   Never smoker  5503 (80.8)  2604 (69.0)  2945 (76.8)  < 0.001   Ex-smoker  514 (7.5)  282 (7.5)  234 (6.1)   Current smoker  797 (11.7)  887 (23.5)  657 (17.1)  Self-reported health conditions (%)   T2Da  436 (6.5)  364 (9.8)  588 (15.6)  < 0.001   Hypertensiona  1285 (19.2)  652 (17.6)  685 (18.2)  0.098   Hypercholesterolaemiaa  1755 (27.5)  672 (18.8)  824 (22.6)  < 0.001  Physical activity (MET-h/week)b, median (IQR)   Moderate- and/or vigorous-intensity leisure-time  5.5 (0.0–15.5)  4.5 (0.0–15.8)  3.5 (0.0–14.0)  < 0.001  Physical activity (MET-h/week)b, mean ± SD   Moderate- and/or vigorous-intensity leisure-time  11.9 ± 23.1  12.8 ± 23.9  11.4 ± 21.1  0.479  Macro nutrient intake (% energy)   Carbohydrate  53.8 ± 6.8  54.5 ± 6.7  54.3 ± 6.5  < 0.001   Protein  15.8 ± 2.3  14.1 ± 2.2  13.2 ± 2.1  < 0.001   Fat  30.1 ± 5.7  31.3 ± 5.8  31.8 ± 5.8  < 0.001  BMI category (kg/m2), (%)   < 18.5  444 (8.6)  113 (3.9)  127 (4.2)  < 0.001   18.5–23.0  2317 (45.1)  621 (21.7)  682 (22.5)   23.0–27.5  1778 (34.6)  1021 (35.7)  1172 (38.7)   ≥ 27.5  603 (11.7)  1107 (38.7)  1050 (34.6)  Waist circumference (cm), (%)   Normal  3492 (67.9)  1287 (45.1)  1155 (38.1)  < 0.001   Highc  1650 (32.1)  1569 (54.9)  1878 (61.9)  Blood pressure (mmHg), mean ± SD   Systolic  127.5 ± 20.6  129.0 ± 20.8  124.5 ± 21.3  < 0.001   Diastolic  75.0 ± 11.3  74.4 ± 11.3  73.0 ± 11.1  < 0.001  Biomarkers, mean ± SD   Triglycerides (mmol/l)  1.3 ± 0.8  1.4 ± 1.1  1.4 ± 1.0  < 0.001   High-density lipoprotein (mmol/l) (HDL)  1.4 ± 0.4  1.2 ± 0.3  1.1 ± 0.3  < 0.001   Low-density lipoprotein (mmol/l) (LDL)  3.2 ± 0.8  3.5 ± 0.9  3.4 ± 0.9  < 0.001   HbA1c (%)  5.7 ± 0.8  6.0 ± 1.3  6.2 ± 1.3  < 0.001  a Participants were asked if they had been diagnosed with T2D, hypertension or high cholesterol by Western doctors. b MET, Metabolic Equivalent of Task. c High waist circumference was defined as > 90 cm for men and > 80 cm for women. Missings smoking status (n = 1), T2D (n = 250), hypertension (n = 269), hyperlipidaemia (n = 813), physical activity (n = 1), macronutrients (n = 264). BMI (n = 3389), waist circumference (n = 3393), blood pressure (n = 3379), triglycerides (n = 3423), HDL (n = 3427), LDL (n = 3475) and HbA1c (n = 5643); most missing values for physical and biochemical variables are due to non-participation in the health screening part of the study. Table 2 Lifestyle, anthropometric measures, and biomarkers profiles of study participants at baseline by ethnicity (N = 14 424) Characteristics  Chinese  Malays  Indians  P-value  Cigarette smoking status (%)   Never smoker  5503 (80.8)  2604 (69.0)  2945 (76.8)  < 0.001   Ex-smoker  514 (7.5)  282 (7.5)  234 (6.1)   Current smoker  797 (11.7)  887 (23.5)  657 (17.1)  Self-reported health conditions (%)   T2Da  436 (6.5)  364 (9.8)  588 (15.6)  < 0.001   Hypertensiona  1285 (19.2)  652 (17.6)  685 (18.2)  0.098   Hypercholesterolaemiaa  1755 (27.5)  672 (18.8)  824 (22.6)  < 0.001  Physical activity (MET-h/week)b, median (IQR)   Moderate- and/or vigorous-intensity leisure-time  5.5 (0.0–15.5)  4.5 (0.0–15.8)  3.5 (0.0–14.0)  < 0.001  Physical activity (MET-h/week)b, mean ± SD   Moderate- and/or vigorous-intensity leisure-time  11.9 ± 23.1  12.8 ± 23.9  11.4 ± 21.1  0.479  Macro nutrient intake (% energy)   Carbohydrate  53.8 ± 6.8  54.5 ± 6.7  54.3 ± 6.5  < 0.001   Protein  15.8 ± 2.3  14.1 ± 2.2  13.2 ± 2.1  < 0.001   Fat  30.1 ± 5.7  31.3 ± 5.8  31.8 ± 5.8  < 0.001  BMI category (kg/m2), (%)   < 18.5  444 (8.6)  113 (3.9)  127 (4.2)  < 0.001   18.5–23.0  2317 (45.1)  621 (21.7)  682 (22.5)   23.0–27.5  1778 (34.6)  1021 (35.7)  1172 (38.7)   ≥ 27.5  603 (11.7)  1107 (38.7)  1050 (34.6)  Waist circumference (cm), (%)   Normal  3492 (67.9)  1287 (45.1)  1155 (38.1)  < 0.001   Highc  1650 (32.1)  1569 (54.9)  1878 (61.9)  Blood pressure (mmHg), mean ± SD   Systolic  127.5 ± 20.6  129.0 ± 20.8  124.5 ± 21.3  < 0.001   Diastolic  75.0 ± 11.3  74.4 ± 11.3  73.0 ± 11.1  < 0.001  Biomarkers, mean ± SD   Triglycerides (mmol/l)  1.3 ± 0.8  1.4 ± 1.1  1.4 ± 1.0  < 0.001   High-density lipoprotein (mmol/l) (HDL)  1.4 ± 0.4  1.2 ± 0.3  1.1 ± 0.3  < 0.001   Low-density lipoprotein (mmol/l) (LDL)  3.2 ± 0.8  3.5 ± 0.9  3.4 ± 0.9  < 0.001   HbA1c (%)  5.7 ± 0.8  6.0 ± 1.3  6.2 ± 1.3  < 0.001  Characteristics  Chinese  Malays  Indians  P-value  Cigarette smoking status (%)   Never smoker  5503 (80.8)  2604 (69.0)  2945 (76.8)  < 0.001   Ex-smoker  514 (7.5)  282 (7.5)  234 (6.1)   Current smoker  797 (11.7)  887 (23.5)  657 (17.1)  Self-reported health conditions (%)   T2Da  436 (6.5)  364 (9.8)  588 (15.6)  < 0.001   Hypertensiona  1285 (19.2)  652 (17.6)  685 (18.2)  0.098   Hypercholesterolaemiaa  1755 (27.5)  672 (18.8)  824 (22.6)  < 0.001  Physical activity (MET-h/week)b, median (IQR)   Moderate- and/or vigorous-intensity leisure-time  5.5 (0.0–15.5)  4.5 (0.0–15.8)  3.5 (0.0–14.0)  < 0.001  Physical activity (MET-h/week)b, mean ± SD   Moderate- and/or vigorous-intensity leisure-time  11.9 ± 23.1  12.8 ± 23.9  11.4 ± 21.1  0.479  Macro nutrient intake (% energy)   Carbohydrate  53.8 ± 6.8  54.5 ± 6.7  54.3 ± 6.5  < 0.001   Protein  15.8 ± 2.3  14.1 ± 2.2  13.2 ± 2.1  < 0.001   Fat  30.1 ± 5.7  31.3 ± 5.8  31.8 ± 5.8  < 0.001  BMI category (kg/m2), (%)   < 18.5  444 (8.6)  113 (3.9)  127 (4.2)  < 0.001   18.5–23.0  2317 (45.1)  621 (21.7)  682 (22.5)   23.0–27.5  1778 (34.6)  1021 (35.7)  1172 (38.7)   ≥ 27.5  603 (11.7)  1107 (38.7)  1050 (34.6)  Waist circumference (cm), (%)   Normal  3492 (67.9)  1287 (45.1)  1155 (38.1)  < 0.001   Highc  1650 (32.1)  1569 (54.9)  1878 (61.9)  Blood pressure (mmHg), mean ± SD   Systolic  127.5 ± 20.6  129.0 ± 20.8  124.5 ± 21.3  < 0.001   Diastolic  75.0 ± 11.3  74.4 ± 11.3  73.0 ± 11.1  < 0.001  Biomarkers, mean ± SD   Triglycerides (mmol/l)  1.3 ± 0.8  1.4 ± 1.1  1.4 ± 1.0  < 0.001   High-density lipoprotein (mmol/l) (HDL)  1.4 ± 0.4  1.2 ± 0.3  1.1 ± 0.3  < 0.001   Low-density lipoprotein (mmol/l) (LDL)  3.2 ± 0.8  3.5 ± 0.9  3.4 ± 0.9  < 0.001   HbA1c (%)  5.7 ± 0.8  6.0 ± 1.3  6.2 ± 1.3  < 0.001  a Participants were asked if they had been diagnosed with T2D, hypertension or high cholesterol by Western doctors. b MET, Metabolic Equivalent of Task. c High waist circumference was defined as > 90 cm for men and > 80 cm for women. Missings smoking status (n = 1), T2D (n = 250), hypertension (n = 269), hyperlipidaemia (n = 813), physical activity (n = 1), macronutrients (n = 264). BMI (n = 3389), waist circumference (n = 3393), blood pressure (n = 3379), triglycerides (n = 3423), HDL (n = 3427), LDL (n = 3475) and HbA1c (n = 5643); most missing values for physical and biochemical variables are due to non-participation in the health screening part of the study. The prevalence of obesity, defined as ≥ 27.5 kg/m2,8 was the highest in Malays (39%), slightly lower in the Indians (35%) and much lower in Chinese (12%). Indians had the highest prevalence of high waist circumference (62%), defined as > 90 cm for men and > 80 cm for women,9 whereas 55% of Malays and 32% of Chinese had high waist circumference. Metabolic equivalents (METs) were obtained from the 2011 Compendium of Physical Activity.10 The median (interquartile range, IQR) weekly leisure-time activity level of moderate and/or vigorous intensity was the highest amongst the Chinese [6 (0–16) MET-h/week], compared with the Malays [5 (0–16) MET-h/week] and Indians [4 (0–14) MET-h/week]. On average, Chinese had a slightly higher protein intake as percentage of total energy intake and lower intake of total fat and carbohydrate than Malays and Indians. Differences between the three ethnic groups were assessed using chi square testing for categorical variables. Comparison of means was done using analysis of variance testing, and comparison of medians was done using the Kruskal-Wallis rank sum test for continuous variables. As shown in Table 2, due to the large number of participants, all differences between ethnic groups had P-values less than 0.001, except hypertension status (P-value = 0.098) and average leisure-time physical activity (P-value = 0.479). How often have they been followed up and what is the attrition like? Baseline recruitment was done between 2004 and 2010. The first wave of in-person follow-up with the cohort participants began in 2011 and ended in 2016. Of the baseline participants, 28% were not contactable (i.e. contact details changed and no updated information available, or frequently travelling, or access to household denied and unable to contact despite six attempts at household visitation) and 2.5% were confirmed to have been lost to follow-up (i.e. deceased, migrated, declined follow-up, lost mental competence to give consent to continue the research, institutionalized or physically unfit to participate). Of the contactable participants, 60% (N = 6112) agreed to participate in the follow-up survey. Data will also be linked to follow-up data from the Singapore National Registry of Diseases Office on chronic diseases (i.e. acute myocardial infarction, cancers, renal failure and stroke) and mortality; 86% of the baseline participants gave consent for data linkage with medical records and national registries. What has been measured? Table 3 summarizes information collected from the MEC study at baseline and follow-up assessments. The baseline interviews assessed sociodemographic variables, detailed lifestyle behaviours, personal and family medical history and medication use, and health-related quality of life (HRQoL). The health examination included measurements of anthropometric characteristics, peripheral neuropathy, branchial and ankle blood pressures and biomarkers in blood and urine. During the follow-up assessments, we collected information on mental health and cognitive function and conducted electrocardiogram, central blood pressure and hand grip strength measurements in addition to the baseline measurements. More details on these measurements are described below. Table 3 Summary of variables collected or derived from the cohort Variables  Baseline  Follow-up  Questionnaire   Demographics  ✓  ✓   Cigarette smoking history  ✓  ✓   Environmental tobacco smoke exposure  ✓  ✓   Alcohol consumption  ✓  ✓   Medication use  ✓  ✓   Medical history  ✓  ✓   Family medical history  ✓  ✓   Women’s health  ✓  ✓   Physical activity  ✓  ✓   Dietary information (FFQ)  ✓  ✓   Health related quality of life (SF-36)  ✓  ✓   Health-related quality of life (EQ-5D)    ✓   Cognitive function (Mini-Mental State Examination)    ✓   Family life    ✓   Sources of psychological stress    ✓   Kessler psychological distress scale (K10)    ✓  Anthropometric measures   Height  ✓  ✓   Weight  ✓  ✓   Waist circumference  ✓  ✓   Hip circumference  ✓  ✓   Skinfold thickness    ✓  Clinical assessment   Light touch assessment measured by monofilament  ✓  ✓   Assessment of foot proprioception by neurothesiometer  ✓  ✓   Brachial and ankle blood pressure  ✓  ✓   Central aortic systolic pressure and arterial pulse waveform indices (rAI, rAP, PRT)    ✓   Bone densitometrya    ✓   Cardiac computer tomographya    ✓   Body fat and lean mass composition by DEXAa    ✓   Subcutaneous abdominal and intra-abdominal fat by CTa    ✓   Electrocardiography    ✓   Hand grip strength    ✓  Blood sample   Fasting or random glucose  ✓  ✓   Creatinine  ✓  ✓   Triglycerides  ✓  ✓   Total cholesterol  ✓  ✓   High-density lipoprotein cholesterol  ✓  ✓   Low-density lipoprotein cholesterol  ✓  ✓   HbA1c  ✓  ✓   Cortisol  ✓  ✓   High-sensitivity C-reactive protein  ✓  ✓   Insulin  ✓     Adiponectin  ✓     Interleukin-6    ✓   Interleukin-1 receptor antagonist    ✓  Urine sample   Protein (semi-quantitative)  ✓     Albumin (semi-quantitative)    ✓  Genetic data   Genotyping array  ✓(subset)     Whole exome sequence  ✓(subset)    Metabolomicsb   Amino acids  ✓(subset)     Acylcarnitines  ✓(subset)     Ceramides  ✓(subset)     Sphingolipids  ✓(subset)    Variables  Baseline  Follow-up  Questionnaire   Demographics  ✓  ✓   Cigarette smoking history  ✓  ✓   Environmental tobacco smoke exposure  ✓  ✓   Alcohol consumption  ✓  ✓   Medication use  ✓  ✓   Medical history  ✓  ✓   Family medical history  ✓  ✓   Women’s health  ✓  ✓   Physical activity  ✓  ✓   Dietary information (FFQ)  ✓  ✓   Health related quality of life (SF-36)  ✓  ✓   Health-related quality of life (EQ-5D)    ✓   Cognitive function (Mini-Mental State Examination)    ✓   Family life    ✓   Sources of psychological stress    ✓   Kessler psychological distress scale (K10)    ✓  Anthropometric measures   Height  ✓  ✓   Weight  ✓  ✓   Waist circumference  ✓  ✓   Hip circumference  ✓  ✓   Skinfold thickness    ✓  Clinical assessment   Light touch assessment measured by monofilament  ✓  ✓   Assessment of foot proprioception by neurothesiometer  ✓  ✓   Brachial and ankle blood pressure  ✓  ✓   Central aortic systolic pressure and arterial pulse waveform indices (rAI, rAP, PRT)    ✓   Bone densitometrya    ✓   Cardiac computer tomographya    ✓   Body fat and lean mass composition by DEXAa    ✓   Subcutaneous abdominal and intra-abdominal fat by CTa    ✓   Electrocardiography    ✓   Hand grip strength    ✓  Blood sample   Fasting or random glucose  ✓  ✓   Creatinine  ✓  ✓   Triglycerides  ✓  ✓   Total cholesterol  ✓  ✓   High-density lipoprotein cholesterol  ✓  ✓   Low-density lipoprotein cholesterol  ✓  ✓   HbA1c  ✓  ✓   Cortisol  ✓  ✓   High-sensitivity C-reactive protein  ✓  ✓   Insulin  ✓     Adiponectin  ✓     Interleukin-6    ✓   Interleukin-1 receptor antagonist    ✓  Urine sample   Protein (semi-quantitative)  ✓     Albumin (semi-quantitative)    ✓  Genetic data   Genotyping array  ✓(subset)     Whole exome sequence  ✓(subset)    Metabolomicsb   Amino acids  ✓(subset)     Acylcarnitines  ✓(subset)     Ceramides  ✓(subset)     Sphingolipids  ✓(subset)    a Imaging was performed only on a sub-sample of 805 Chinese older participants of the follow-up cohort. b Metabolites were measured using a targeted, high-throughput metabolic profile approach using an HPLC online electrospray ionization tandem mass spectrometry.56 Table 3 Summary of variables collected or derived from the cohort Variables  Baseline  Follow-up  Questionnaire   Demographics  ✓  ✓   Cigarette smoking history  ✓  ✓   Environmental tobacco smoke exposure  ✓  ✓   Alcohol consumption  ✓  ✓   Medication use  ✓  ✓   Medical history  ✓  ✓   Family medical history  ✓  ✓   Women’s health  ✓  ✓   Physical activity  ✓  ✓   Dietary information (FFQ)  ✓  ✓   Health related quality of life (SF-36)  ✓  ✓   Health-related quality of life (EQ-5D)    ✓   Cognitive function (Mini-Mental State Examination)    ✓   Family life    ✓   Sources of psychological stress    ✓   Kessler psychological distress scale (K10)    ✓  Anthropometric measures   Height  ✓  ✓   Weight  ✓  ✓   Waist circumference  ✓  ✓   Hip circumference  ✓  ✓   Skinfold thickness    ✓  Clinical assessment   Light touch assessment measured by monofilament  ✓  ✓   Assessment of foot proprioception by neurothesiometer  ✓  ✓   Brachial and ankle blood pressure  ✓  ✓   Central aortic systolic pressure and arterial pulse waveform indices (rAI, rAP, PRT)    ✓   Bone densitometrya    ✓   Cardiac computer tomographya    ✓   Body fat and lean mass composition by DEXAa    ✓   Subcutaneous abdominal and intra-abdominal fat by CTa    ✓   Electrocardiography    ✓   Hand grip strength    ✓  Blood sample   Fasting or random glucose  ✓  ✓   Creatinine  ✓  ✓   Triglycerides  ✓  ✓   Total cholesterol  ✓  ✓   High-density lipoprotein cholesterol  ✓  ✓   Low-density lipoprotein cholesterol  ✓  ✓   HbA1c  ✓  ✓   Cortisol  ✓  ✓   High-sensitivity C-reactive protein  ✓  ✓   Insulin  ✓     Adiponectin  ✓     Interleukin-6    ✓   Interleukin-1 receptor antagonist    ✓  Urine sample   Protein (semi-quantitative)  ✓     Albumin (semi-quantitative)    ✓  Genetic data   Genotyping array  ✓(subset)     Whole exome sequence  ✓(subset)    Metabolomicsb   Amino acids  ✓(subset)     Acylcarnitines  ✓(subset)     Ceramides  ✓(subset)     Sphingolipids  ✓(subset)    Variables  Baseline  Follow-up  Questionnaire   Demographics  ✓  ✓   Cigarette smoking history  ✓  ✓   Environmental tobacco smoke exposure  ✓  ✓   Alcohol consumption  ✓  ✓   Medication use  ✓  ✓   Medical history  ✓  ✓   Family medical history  ✓  ✓   Women’s health  ✓  ✓   Physical activity  ✓  ✓   Dietary information (FFQ)  ✓  ✓   Health related quality of life (SF-36)  ✓  ✓   Health-related quality of life (EQ-5D)    ✓   Cognitive function (Mini-Mental State Examination)    ✓   Family life    ✓   Sources of psychological stress    ✓   Kessler psychological distress scale (K10)    ✓  Anthropometric measures   Height  ✓  ✓   Weight  ✓  ✓   Waist circumference  ✓  ✓   Hip circumference  ✓  ✓   Skinfold thickness    ✓  Clinical assessment   Light touch assessment measured by monofilament  ✓  ✓   Assessment of foot proprioception by neurothesiometer  ✓  ✓   Brachial and ankle blood pressure  ✓  ✓   Central aortic systolic pressure and arterial pulse waveform indices (rAI, rAP, PRT)    ✓   Bone densitometrya    ✓   Cardiac computer tomographya    ✓   Body fat and lean mass composition by DEXAa    ✓   Subcutaneous abdominal and intra-abdominal fat by CTa    ✓   Electrocardiography    ✓   Hand grip strength    ✓  Blood sample   Fasting or random glucose  ✓  ✓   Creatinine  ✓  ✓   Triglycerides  ✓  ✓   Total cholesterol  ✓  ✓   High-density lipoprotein cholesterol  ✓  ✓   Low-density lipoprotein cholesterol  ✓  ✓   HbA1c  ✓  ✓   Cortisol  ✓  ✓   High-sensitivity C-reactive protein  ✓  ✓   Insulin  ✓     Adiponectin  ✓     Interleukin-6    ✓   Interleukin-1 receptor antagonist    ✓  Urine sample   Protein (semi-quantitative)  ✓     Albumin (semi-quantitative)    ✓  Genetic data   Genotyping array  ✓(subset)     Whole exome sequence  ✓(subset)    Metabolomicsb   Amino acids  ✓(subset)     Acylcarnitines  ✓(subset)     Ceramides  ✓(subset)     Sphingolipids  ✓(subset)    a Imaging was performed only on a sub-sample of 805 Chinese older participants of the follow-up cohort. b Metabolites were measured using a targeted, high-throughput metabolic profile approach using an HPLC online electrospray ionization tandem mass spectrometry.56 Baseline measurements Participants completed an interviewer-administered questionnaire at their home, which took approximately an hour. Questionnaires were available in English, Chinese and Malay. Interviewers provided additional explanation when necessary in other languages common to both the participant and the interviewer. Participants were interviewed about their sociodemographic characteristics, personal and family medical history and lifestyle behaviours such as alcohol consumption, sedentary behaviours, cigarette smoking and exposure to environmental tobacco smoke. In addition, all medications used by the participants were recorded. Dietary intake was assessed by a semi-quantitative 169-item food frequency questionnaire (FFQ) that is also used in the Singapore National Nutrition Surveys.11 The physical activity questionnaire asked about the type, frequency and duration of various activities in the transportation, occupation, leisure time and household domains. Transportation activities included walking and cycling, and occupational activities included light, moderate and vigorous occupational activities. Leisure time activities included 48 specific activities, household activities included 15 specific activities, and options for additional activities not provided in the list were recorded through open-ended questions. Based on the physical activity questionnaire12 and FFQ,11 derived variables such as weekly physical activity level (MET-h/week) and intake of macronutrients, micronutrients and daily servings by food groups were calculated. Both questionnaires were validated in the local population.11,12 Nutrient intakes estimated from the FFQ were reasonably well correlated with estimates from repeated 24-h recalls. In addition, associations between food intakes and plasma fatty acids indicated that sources of several fatty acids were captured well by the FFQ.13 For physical activity, vigorous activity agreed well with accelerometer measurements, whereas modest agreement was observed for moderate-intensity activity. Participants were given the option to fast for the health examination. For those who did, they were told to fast for 8–12 h before their appointment. Fasting status was recorded by phlebotomists who performed venepuncture to collect blood for both biochemical analyses and biobanking. Random spot urine samples, collected mid-stream, were measured semi-quantitatively for protein level and any excess was stored in the biobank. Height was measured without shoes on a portable stadiometer (SECA 200 series, Germany) with the head in the Frankfurt Plane position. Participants were instructed to remove objects from their pockets such as wallets, keys and mobile phones before being measured for weight on SECA digital scales (SECA 700 series, Germany). A stretch-resistant tape was used to measure waist circumference at the mid-point between the last rib and iliac crest, and hip circumference at the greater trochanter of the femur. Systolic and diastolic blood pressures were measured after participants rested for 5 min, using an automated digital monitor to obtain two readings (Dinamap Carescape V100, General Electric). A third reading was performed if the difference between the first two readings was greater than 10 mmHg or 5 mmHg for systolic blood pressure and diastolic blood pressure, respectively. For the minority of the participants whose blood pressure exceeded the range of the digital monitor, a sphygmomanometer (Accoson, UK) was used. The ankle-brachial index was calculated to assess risk of peripheral artery disease, as follows. The systolic blood pressure from the brachial artery in the right arm and from the dorsalis pedis or posterior tibial artery in both ankles were measured using sphygmomanometers (Accoson, UK) in the supine position (head elevated at 15° to 20°). A handheld Doppler (Hadeco, Japan) was used to detect the pulses at the arteries. If the pulse at the dorsalis pedis artery could not be detected, the posterior tibial artery would be probed. Blood pressure in the left arm would be measured if it could not be done in the right arm. All the systolic blood pressure measurements were repeated once. Peripheral neuropathy was measured using neurothesiometer and monofilament tests. Participants lay almost supine for comfort, with their feet bare and eyes closed for the sensory functions assessment on their feet. Foot proprioception was assessed with a neurothesiometer (Horwell, UK) where a vibration-emitting probe was applied to the apex of the big toe and medial malleolus of both feet. The voltage was gradually increased from zero until the participants indicated verbally that they could feel the vibration. The voltage reading was recorded for both sites and both feet. Light touch assessment was performed by using a 10g (5.07) monofilament (Sensory Testing System, USA) on five least calloused plantar sites per foot—the distal great toe, third toe and fifth toe and the first and fifth metatarsal heads. The number of sites that the participants could feel was recorded for each foot. Follow-up measurements In the cohort follow-up, the health examination included four additional procedures, including psychological distress (Kessler K10),14,15 cognitive function (Mini Mental State Examination),16 body composition from skinfold thickness, and hand grip strength. Using a Holtain/Tanner skinfold caliper, skinfold thickness was measured at the left triceps, left biceps, sub-scapular, supra-iliac and calf regions. Three measurements were made on each site while the participant was standing. An electrocardiogram, 10 leads (Nihon Kohden ECG-1350K, Japan), was performed on each participant resting supine with head slightly elevated about 15° to 20°. For the measurement of the central aortic systolic pressure and arterial pulse waveform (A-PULSE CASPro Lite, HealthSTATS, Singapore), measurements were performed predominantly on the left arm with the participant seated and the left arm resting on a table at chest level. To assess their hand grip strength, participants were asked to stand with their arms free at the side and to grip the hand dynamometer (TAKEI A5401, Japan) as hard as they could with one hand. Three readings were recorded for each arm. A sub-sample of the follow-up cohort (805 Chinese, aged 50 and older) underwent more intensive assessments which included bone densitometry and body fat and lean mass composition, and cardiac computed tomography (CT).17 Biochemical analyses and biobanking Blood and urine samples were kept at the collection site at 4°C and transported at 4–8°C by a van to the biochemistry and biobank laboratories within 4–6 h of sample collection. Biochemical analyses of the blood samples were performed on the same day as collection. The biobank laboratory processed the blood and urine samples on the same day except for urine samples that were collected in the afternoon, which would be stored at 4°C and processed the next day. Blood samples were analysed at the biochemistry laboratory of the National University Hospital from April 2007 to October 2007 and November 2010 to August 2016. Analyses were performed at the Singapore General Hospital biochemistry laboratory for the period October 2007–November 2010. Both laboratories are accredited by the College of American Pathologists. Fresh blood samples were analysed on the day of collection for glucose, lipids, creatinine and HbA1c. Supplementary Table 1 (available as Supplementary data at IJE online) provides laboratory details [analyser platform, assay methodology, measurement range and coefficients of variation (within and between day)] of the fresh blood analyses reported by the laboratories. As of present, the biological specimens are stored in the tissue repositories of the National University Hospital and Saw Swee Hock School of Public Health. Supplementary Table 2 (available as Supplementary data at IJE online) shows the types of biospecimen stored. Genome-wide single nucleotide polymorphism (SNP) data are available for a subset of the MEC cohort from a combination of Illumina genome-wide genotyping arrays and customized arrays such as Oncoarray18 and iCOGS.19 Samples genotyped on the genome-wide arrays have been imputed to 1000G Phase 3 reference panels.20 Whole-exome sequence data are also available for some of the participants.21 Quality control and data processing Interviewers and personnel who performed the health examination were trained and assessed for competency before they could begin data collection. This was to ensure standardized interviewing and consistency in performing the health examination and collection of samples. As the cohort is multi-ethnic in nature, the interviewers and personnel are proficient in at least one of the local languages (English, Chinese, Malay, Chinese dialects) verbally. Completed questionnaires were sampled for verification with 20% of the participants on the data collected. All data forms, including consent forms, were inspected for missing data before they were subjected to double data entry by two independent staff. A third person resolved contradictions between entries by verifying with the source document. The data forms were optically scanned and archived following the completion of data entry. Research data were de-identified, key-coded and maintained in research databases accessible only to database administrators. Database administrators performed further data cleaning of the research data by checking for consistency between related variables and data range before releasing the data for research analysis. Personal information of participants and study visit records were maintained in different databases that only the fieldwork management team could access for the purpose of follow-up. What has it found? Key findings and publications This large multi-ethnic cohort has provided us an opportunity to examine differences in health conditions, HRQoL and behaviours between major Asian ethnic groups within a relatively homogeneous environment and access to health care in Singapore. The prevalence of T2D and other metabolic-related outcomes is increasing worldwide, with the greatest increases predicted in East (Chinese and Malays) and South Asian (Indians) populations in Asia.22 Within the MEC study, we compare and evaluate different metabolic biomarkers and their contributions to disease outcomes between the three ethnic groups. For example, Indians had the greatest insulin resistance, followed by Malays and lastly Chinese. The difference between Chinese and Malays could be explained by differences in body mass index (BMI). However, the greater insulin resistance in Indians could only be partly explained by higher BMI, waist circumference and inflammation.23 In addition, BMI had a greater influence on insulin resistance, C-reactive protein and adiponectin in Chinese than the other major Asian ethnic groups.24 These suggest that ethnicity modifies the relationship between adiposity and pathways (e.g. insulin resistance, inflammation and adiponectin) involved in the pathogenesis of T2D and cardiovascular diseases (CVD), and may lead to greater increases in T2D and CVD in Chinese than other ethnic groups as a result of increasing BMI. Longitudinal follow-up in the cohort allows us to examine the utility of conventional diagnostic criteria in their prediction of long-term prognosis and complications. When we examined if the current T2D diagnostic criteria based on fasting glucose levels reflect peripheral neuropathy and chronic kidney disease, we showed that both microvascular complications gradually increased in relation to fasting plasma glucose, beginning at levels below the existing diagnostic threshold for T2D of 7.0 mmol/l. For chronic kidney disease, these associations persisted after adjustment for other risk factors. These findings suggest that current diagnostic thresholds for T2D may have limited sensitivity for identifying individuals with microvascular complications in an Asian population.1 Another prospective study examined changes in weight and waist circumference over the follow-up period and their associations with demographic and socioeconomic factors.25 Those who gained the most weight were younger adults, likely to be of higher socioeconomic status (SES) and ethnic minority group, and had the lowest BMI at baseline. The data provided evidence for obesity prevention strategies to be initiated at a younger age and with specific considerations for the minority ethnic and higher SES groups. Record linkage with national registries enabled the identification of cases with ischaemic heart disease (IHD) in the cohort. We showed that not only did Indians have a higher susceptibility for developing T2D than Chinese and Malays, but also the risk of IHD in Indians with T2D was higher than the other groups.2 In another study, we demonstrated that having metabolic syndrome with or without central obesity confers the same risk for IHD in Asians. We went on to show that when using the International Diabetes Federation definition of metabolic syndrome, which includes central obesity as an essential component, considerably fewer individuals at risk of IHD in this cohort would be identified than if central obesity was regarded as an optional component as defined by the American Heart Association/National Heart, Lung, and Blood Institute.26 With detailed validated questionnaires on dietary intakes and physical activity, we can assess associations of dietary and physical activity patterns with health outcomes including HRQoL, and how these associations differ across ethnic groups. We observed that a combination of brief television watching and regular vigorous physical activity was strongly associated with lower insulin resistance.27 Interestingly, the association between television watching and insulin resistance appeared to be mediated by eating habits and adiposity rather than lower physical activity. With regards to dietary factors, coffee consumption was associated with less insulin resistance, whereas high consumption of rice and noodles was associated with greater insulin resistance and hyperglycaemia.28,29 Using the Short Form 36 version 2 (SF-36v2) and Short Form 6D (SF-6D) instrument surveys validated in MEC,30 Leow et al. found that HRQoL levels differed between ethnic groups: Chinese had higher physical health component summary scores than Malays and Indians, whereas Malays displayed higher mental health component summary scores than Chinese.31 The same study also discovered that the effects of gender, stroke and musculoskeletal conditions on HRQoL may vary by ethnicity. Other studies have further examined the relationship between HRQoL and T2D and its complications and comorbidities. Venkataraman et al. found that the HRQoL of patients with T2D was affected by complications and not T2D per se, with peripheral neuropathy even in mild form being associated with the greatest reduction in HRQoL.32 These findings support the improvement of T2D management to prevent or delay the onset of complications. In patients with T2D, hypertension and dyslipidaemia, disease awareness was found to be associated with lower HRQoL whereas undiagnosed disease was not, which may partly explain why the undiagnosed did not seek medical care.33 Understanding these key modifiable lifestyle patterns can help frame lifestyle recommendations, targeted interventions and health promotion policies. Insights as to why people remain undiagnosed with chronic diseases can help inform strategies for preventive health screening. Another study by Venkataraman et al. identified Malay ethnicity, lower educational status, lower family cohesion and unhealthy diet and physical activity patterns as some key barriers to participation in health screening. It suggested enhancing the cultural competence of preventive health services, as one of the strategies to improve the health screening participation of these groups and reduce the proportions in the population that remain undiagnosed with chronic diseases.34 Finally, the diverse ethnic groups represented in MEC enabled mapping genetic diversity across these populations35 and discovery of genetic variants associated with various disease outcomes and related biomarkers. For example, a common genetic variant (10%) at PAX4 gene was found to be associated with T2D in East Asians, whereas the variant is extremely rare in other global populations.21,36,37 We have led and contributed to multiple discovery efforts in East and South Asians and trans-ethnic large-scale genetic association analyses, and identified many genetic variants associated with blood pressure,38–40 anthropometric measures,41–44 kidney function,45 adiponectin levels,46 glycaemic traits47–49 and blood lipid profiles.50,51 In addition to metabolic disorders, MEC participants have been ethnicity-matched to breast cancer cases of a patient cohort to act as controls in genetic association studies.52,53 About 80 scientific papers based on the MEC data have been published up to May 2017 [https://blog.nus.edu.sg/sphs/publications/]. Data have also been shared for international collaborative projects such as the Non-Communicable Disease Risk Factor Collaboration, formerly known as the Global Burden of Metabolic Risk Factors of Chronic Diseases Collaborating Group, and the Asia Pacific Cohort Studies Collaboration, culminating in reports on aetiological associations and regional and global trends in cardiovascular diseases, obesity and T2D.54,55 What are the main strengths and weaknesses? A major strength of the MEC study is that it represents three major ethnic groups in Asia who are at risk of major health outcome changes. The extensive interviewer-administered questionnaire covered a wide range of health determinants, medical history and quality of life with minimal missing data. The physical activity and FFQ used in the MEC study were validated in independent samples of the local population with similar characteristics.11,12 The MEC study also included a detailed physical examination, biochemical analyses of blood samples and storage of blood and spot urine samples. A potential limitation of the MEC cohort was the substantial non-response during the follow-up visit. This can partly be resolved by passive follow-up using linkage to disease registries that have good coverage of the Singapore population. Can I get hold of the data? Where can I find out more? Researchers can visit the Saw Swee Hock School of Public Health website [https://blog.nus.edu.sg/sphs/] for information on submitting a request for data and/or samples. Supplementary Data Supplementary data are available at IJE online. Profile in a nutshell Singapore is a multi-ethnic island state and provides a useful model to evaluate determinants of the development of chronic diseases in Asian ethnic groups. Baseline recruitment was done between 2004 and 2010. The cohort includes 14 465 male and female participants aged 21 years and above, with 47% Chinese, 26% Malay and 27% Indian ethnicity. The first follow-up began in 2011 and ended in 2016; 28% of the baseline participants were not contactable and 2.5% were confirmed to have been lost to follow-up. Of the participants that were contactable, 60% agreed to participate in the follow-up survey. Sociodemographic, lifestyle behaviours, personal and family medical history, medication use and health-related quality of life information was collected at baseline. Measurements of anthropometric characteristics, peripheral neuropathy, brachial and ankle blood pressure and biomarkers in blood and urine were also collected. During the follow-up, additional information on mental health and cognitive function was collected and electrocardiogram, central blood pressure and hand grip strength measurements were conducted; 86% of all baseline participants gave consent for data linkage, allowing disease follow-up through medical records and national registries. Researchers can visit the Saw Swee Hock School of Public Health website [https://blog.nus.edu.sg/sphs/] for information on submitting a request for data and/or samples. Author Contributions K.H.X.T. contributed to the analysis and interpretation of data and drafted the manuscript. L.W.L.T. contributed to design of the study, management of the study and drafted the manuscript. J.J.M.L., K.S.C. and E.S.T. designed the study. X.S. and R.M.V.D. critically revised the manuscript. Acknowledgements We are grateful to the Singapore residents who volunteered time, data and samples for this study and to the community leaders who had facilitated our recruitment drives. We wish to thank the fieldwork and data management team for their dedicated work on this study. Funding The MEC study was supported by the National Medical Research Council (grant 0838/2004), Biomedical Research Council (grants 03/1/27/18/216, 05/1/21/19/425 and 11/1/21/19/678), Ministry of Health, Singapore, National University of Singapore and National University Health System, Singapore. Conflict of interest: None declared. References 1 Nang EE, Khoo CM, Tai ES et al.   Is there a clear threshold for fasting plasma glucose that differentiates between those with and without neuropathy and chronic kidney disease?: the Singapore Prospective Study Program. Am J Epidemiol  2009; 169: 1454– 62. Google Scholar CrossRef Search ADS PubMed  2 Yeo KK, Tai BC, Heng D et al.   Ethnicity modifies the association between diabetes mellitus and ischaemic heart disease in Chinese, Malays and Asian Indians living in Singapore. Diabetologia  2006; 49: 2866– 73. Google Scholar CrossRef Search ADS PubMed  3 Lee J, Heng D, Ma S, Chew SK, Hughes K, Tai ES. The metabolic syndrome and mortality: the Singapore Cardiovascular Cohort Study. Clin Endocrinol (Oxf)  2008; 69: 225– 30. Google Scholar CrossRef Search ADS PubMed  4 Hughes K, Yeo PP, Lun KC et al.   Cardiovascular diseases in Chinese, Malays, and Indians in Singapore. II. Differences in risk factor levels. J Epidemiol Community Health  1990; 44: 29– 35. Google Scholar CrossRef Search ADS PubMed  5 Tan CE, Emmanuel SC, Tan BY, Jacob E. Prevalence of diabetes and ethnic differences in cardiovascular risk factors. The 1992 Singapore National Health Survey. Diabetes Care  1999; 22: 241– 47. Google Scholar CrossRef Search ADS PubMed  6 Hughes K, Aw TC, Kuperan P, Choo M. Central obesity, insulin resistance, syndrome X, lipoprotein(a), and cardiovascular risk in Indians, Malays, and Chinese in Singapore. J Epidemiol Community Health  1997; 51: 394– 99. Google Scholar CrossRef Search ADS PubMed  7 Cutter J, Tan BY, Chew SK. Levels of cardiovascular disease risk factors in Singapore following a national intervention programme. Bull World Health Organ  2001; 79: 908– 15. Google Scholar PubMed  8 WHO Expert Consultation. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet  2004; 363: 157– 63. CrossRef Search ADS PubMed  9 World Health Organization. Waist Circumference and Waist-Hip Ratio: Report of a WHO Expert Consultation . Geneva: WHO, 2008. 10 Ainsworth BE, Haskell WL, Herrmann SD et al.   2011 Compendium of Physical Activities: a second update of codes and MET values. Med Sci Sports Exerc  2011; 43: 1575– 81. Google Scholar CrossRef Search ADS PubMed  11 Deurenberg-Yap M, Li T, Tan WL, van Staveren WA, Deurenberg P. Validation of a semiquantitative food frequency questionnaire for estimation of intakes of energy, fats and cholesterol among Singaporeans. Asia Pac J Clin Nutr  2000; 9: 282– 88. Google Scholar CrossRef Search ADS PubMed  12 Nang EE, Gitau Ngunjiri SA, Wu Y et al.   Validity of the International Physical Activity Questionnaire and the Singapore Prospective Study Program physical activity questionnaire in a multiethnic urban Asian population. BMC Med Res Methodol  2011; 11: 141. Google Scholar CrossRef Search ADS PubMed  13 Seah JY, Gay GM, Su J et al.   Consumption of red meat, but not cooking oils high in polyunsaturated fat, is associated with higher arachidonic acid status in Singapore Chinese Adults. Nutrients  2017; 9: 101. Google Scholar CrossRef Search ADS   14 Kessler RC, Andrews G, Colpe LJ et al.   Short screening scales to monitor population prevalences and trends in non-specific psychological distress. Psychol Med  2002; 32: 959– 76. Google Scholar CrossRef Search ADS PubMed  15 Kessler RC, Barker PR, Colpe LJ et al.   Screening for serious mental illness in the general population. Arch Gen Psychiatry  2003; 60: 184– 89. Google Scholar CrossRef Search ADS PubMed  16 Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res  1975; 12: 189– 98. Google Scholar CrossRef Search ADS PubMed  17 Nang EE, van Dam RM, Tan CS et al.   Association of television viewing time with body composition and calcified subclinical atherosclerosis in Singapore Chinese. PLoS One  2015; 10: e0132161. Google Scholar CrossRef Search ADS PubMed  18 Amos CI, Dennis J, Wang Z et al.   The OncoArray Consortium: A network for understanding the genetic architecture of common cancers. Cancer Epidemiol Biomarkers Prev  2017; 26: 126– 35. Google Scholar CrossRef Search ADS PubMed  19 Sakoda LC, Jorgenson E, Witte JS. Turning of COGS moves forward findings for hormonally mediated cancers. Nat Genet  2013; 45: 345– 48. Google Scholar CrossRef Search ADS PubMed  20 Auton A, Brooks LD, Durbin RM et al.   A global reference for human genetic variation. Nature  2015; 526: 68– 74. Google Scholar CrossRef Search ADS PubMed  21 Fuchsberger C, Flannick J, Teslovich TM et al.   The genetic architecture of type 2 diabetes. Nature  2016; 536: 41– 47. Google Scholar CrossRef Search ADS PubMed  22 Whiting DR, Guariguata L, Weil C, Shaw J. IDF diabetes atlas: global estimates of the prevalence of diabetes for 2011 and 2030. Diabetes Res Clin Pract  2011; 94: 311– 21. Google Scholar CrossRef Search ADS PubMed  23 Gao H, Salim A, Lee J, Tai ES, van Dam RM. Can body fat distribution, adiponectin levels and inflammation explain differences in insulin resistance between ethnic Chinese, Malays and Asian Indians? Int J Obes (Lond)  2012; 36: 1086– 93. Google Scholar CrossRef Search ADS PubMed  24 Khoo CM, Sairazi S, Taslim S et al.   Ethnicity modifies the relationships of insulin resistance, inflammation, and adiponectin with obesity in a multiethnic Asian population. Diabetes Care  2011; 34: 1120– 26. Google Scholar CrossRef Search ADS PubMed  25 Ong SK, Fong CW, Ma S et al.   Longitudinal study of the socio-demographic determinants of changes in body weight and waist circumference in a multi-ethnic Asian population. Int J Obes (Lond)  2009; 33: 1299– 308. Google Scholar CrossRef Search ADS PubMed  26 Lee J, Ma S, Heng D et al.   Should central obesity be an optional or essential component of the metabolic syndrome? Ischemic heart disease risk in the Singapore Cardiovascular Cohort Study. Diabetes Care  2007; 30: 343– 47. Google Scholar CrossRef Search ADS PubMed  27 Nang EE, Salim A, Wu Y, Tai ES, Lee J, Van Dam RM. Television screen time, but not computer use and reading time, is associated with cardio-metabolic biomarkers in a multiethnic Asian population: a cross-sectional study. Int J Behav Nutr Phys Act  2013; 10: 70. Google Scholar CrossRef Search ADS PubMed  28 Rebello SA, Chen CH, Naidoo N et al.   Coffee and tea consumption in relation to inflammation and basal glucose metabolism in a multi-ethnic Asian population: a cross-sectional study. Nutr J  2011; 10: 61. Google Scholar CrossRef Search ADS PubMed  29 Zuniga YL, Rebello SA, Oi PL et al.   Rice and noodle consumption is associated with insulin resistance and hyperglycaemia in an Asian population. Br J Nutr  2014; 11: 1118– 28. Google Scholar CrossRef Search ADS   30 Tan ML, Wee HL, Lee J et al.   The Short Form 36 English and Chinese versions were equivalent in a multiethnic Asian population. J Clin Epidemiol  2013; 66: 759– 67. Google Scholar CrossRef Search ADS PubMed  31 Leow MK, Griva K, Choo R et al.   Determinants of Health-Related Quality of Life (HRQoL) in the multiethnic Singapore population - a national cohort study. PLoS One  2013; 8: e67138. Google Scholar CrossRef Search ADS PubMed  32 Venkataraman K, Wee HL, Leow MK et al.   Associations between complications and health-related quality of life in individuals with diabetes. Clin Endocrinol (Oxf)  2013; 7: 865– 73. Google Scholar CrossRef Search ADS   33 Venkataraman K, Khoo C, Wee HL et al.   Associations between disease awareness and health-related quality of life in a multi-ethnic Asian population. PLoS One  2014; 9: e113802. Google Scholar CrossRef Search ADS PubMed  34 Venkataraman K, Wee HL, Ng SH et al.   Determinants of individuals' participation in integrated chronic disease screening in Singapore. J Epidemiol Community Health . 2016 Jun 10. doi: 10.1136/jech-2016-207404. 35 Saw WY, Tantoso E, Begum H et al.   Establishing multiple omics baseline for three Southeast Asian ethnic groups in the Singapore Integrative Omics Cohort. Nat Commun  2017; 8: 653. Google Scholar CrossRef Search ADS PubMed  36 Cho YS, Chen CH, Hu C et al.   Meta-analysis of genome-wide association studies identifies eight new loci for type 2 diabetes in east Asians. Nat Genet  2011; 44: 67– 72. Google Scholar CrossRef Search ADS PubMed  37 Mahajan A, Go MJ, Zhang W et al.   Genome-wide trans-ancestry meta-analysis provides insight into the genetic architecture of type 2 diabetes susceptibility. Nat Genet  2014; 46: 234– 44. Google Scholar CrossRef Search ADS PubMed  38 Kato N, Takeuchi F, Tabara Y et al.   Meta-analysis of genome-wide association studies identifies common variants associated with blood pressure variation in east Asians. Nat Genet  2011; 43: 531– 38. Google Scholar CrossRef Search ADS PubMed  39 Kelly TN, Takeuchi F, Tabara Y et al.   Genome-wide association study meta-analysis reveals transethnic replication of mean arterial and pulse pressure loci. Hypertension  2013; 62: 853– 59. Google Scholar CrossRef Search ADS PubMed  40 Kato N, Loh M, Takeuchi F et al.   Trans-ancestry genome-wide association study identifies 12 genetic loci influencing blood pressure and implicates a role for DNA methylation. Nat Genet  2015; 47: 1282– 93. Google Scholar CrossRef Search ADS PubMed  41 Wen W, Cho YS, Zheng W et al.   Meta-analysis identifies common variants associated with body mass index in east Asians. Nat Genet  2012; 44: 307– 11. Google Scholar CrossRef Search ADS PubMed  42 Wen W, Zheng W, Okada Y et al.   Meta-analysis of genome-wide association studies in East Asian-ancestry populations identifies four new loci for body mass index. Hum Mol Genet  2014; 23: 5492– 504. Google Scholar CrossRef Search ADS PubMed  43 He M, Xu M, Zhang B et al.   Meta-analysis of genome-wide association studies of adult height in East Asians identifies 17 novel loci. Hum Mol Genet  2015; 24: 1791– 800. Google Scholar CrossRef Search ADS PubMed  44 Wen W, Kato N, Hwang JY et al.   Genome-wide association studies in East Asians identify new loci for waist-hip ratio and waist circumference. Sci Rep  2016; 6: 17958. Google Scholar CrossRef Search ADS PubMed  45 Okada Y, Sim X, Go MJ et al.   Meta-analysis identifies multiple loci associated with kidney function-related traits in east Asian populations. Nat Genet  2012; 44: 904– 09. Google Scholar CrossRef Search ADS PubMed  46 Wu Y, Gao H, Li H et al.   A meta-analysis of genome-wide association studies for adiponectin levels in East Asians identifies a novel locus near WDR11-FGFR2. Hum Mol Genet  2014; 23: 1108– 19. Google Scholar CrossRef Search ADS PubMed  47 Chen P, Takeuchi F, Lee JY et al.   Multiple nonglycemic genomic loci are newly associated with blood level of glycated hemoglobin in East Asians. Diabetes  2014; 63: 2551– 62. Google Scholar CrossRef Search ADS PubMed  48 Hwang JY, Sim X, Wu Y et al.   Genome-wide association meta-analysis identifies novel variants associated with fasting plasma glucose in East Asians. Diabetes  2015; 64: 291– 98. Google Scholar CrossRef Search ADS PubMed  49 Chen P, Ong RT, Tay WT et al.   A study assessing the association of glycated hemoglobin A1C (HbA1C) associated variants with HbA1C, chronic kidney disease and diabetic retinopathy in populations of Asian ancestry. PLoS One  2013; 8: e79767. Google Scholar CrossRef Search ADS PubMed  50 Teslovich TM, Musunuru K, Smith AV et al.   Biological, clinical and population relevance of 95 loci for blood lipids. Nature  2010; 466: 707– 13. Google Scholar CrossRef Search ADS PubMed  51 Spracklen CN, Chen P, Kim YJ et al.   Association analyses of East Asian individuals and trans-ancestry analyses with European individuals reveal new loci associated with cholesterol and triglyceride levels. Hum Mol Genet  2017; 26: 1770– 84. Google Scholar CrossRef Search ADS PubMed  52 Michailidou K, Hall P, Gonzalez-Neira A et al.   Large-scale genotyping identifies 41 new loci associated with breast cancer risk. Nat Genet  2013; 45: 353– 61, 61e1-2. Google Scholar CrossRef Search ADS PubMed  53 Garcia-Closas M, Couch FJ, Lindstrom S et al.   Genome-wide association studies identify four ER negative-specific breast cancer risk loci. Nat Genet  2013; 45: 392– 98, 8e1-2. Google Scholar CrossRef Search ADS PubMed  54 NCD Risk Factor Collaboration (NCD-RisC). Worldwide trends in blood pressure from 1975 to 2015: a pooled analysis of 1479 population-based measurement studies with 19.1 million participants. Lancet  2017; 389: 37– 55. CrossRef Search ADS PubMed  55 Woodward M, Huxley R, Ueshima H, Fang X, Kim HC, Lam TH. The Asia Pacific cohort studies collaboration: a decade of achievements. Glob Heart  2012; 7: 343– 51. Google Scholar CrossRef Search ADS PubMed  56 Haus JM, Kashyap SR, Kasumov T et al.   Plasma ceramides are elevated in obese subjects with type 2 diabetes and correlate with the severity of insulin resistance. Diabetes  2009; 58: 337– 43. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association

Journal

International Journal of EpidemiologyOxford University Press

Published: Feb 13, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off