METS is a prospective cohort study examining physical activity and weight change in 5 African-origin cohorts from Ghana, Jamaica, South Africa, Seychelles, and the United States of America (USA) (N = 500 at each site). METS was initiated in 2009. In total, 2506 participants of African origin, aged 25–45 years were recruited. Approximately 50% of participants were women (N = 1339), and all participants self-identified as African-origin. Sites were categorized using the 2010 United Nations’ Human Development Index (HDI) [15]. In 2010, Ghana was ranked as low HDI (0.554), South Africa was ranked as moderate (0.645), Seychelles and Jamaica were ranked as high (0.747 and 0.712, respectively), and the USA was ranked as very high (0.914). Data for the current study were collected during the baseline clinic visit, between the months of January 2009 and December 2011. The protocol for METS was approved by the Institutional Review Board of Loyola University Chicago, IL, USA (LU#200038); the Committee on Human Research Publication and Ethics of Kwame Nkrumah University of Science and Technology, Kumasi, Ghana; the Research Ethics Committee of the University of Cape Town, South Africa; the Board for Ethics and Clinical Research of the University of Lausanne, Switzerland; the Research and Ethics Committee, Ministry of Health and Social Development, Seychelles Public Health Department, Republic of Seychelles, and the Ethics Committee of the University of the West Indies, Kingston, Jamaica. A written informed consent was obtained from all participants before enrollment.
Study inclusion criteria included participants who self-identified as African-origin and considered themselves healthy. Participants with obvious infectious disease, HIV positive individuals, women who were either lactating or pregnant, and participants unable to engage in normal physical activity were not eligible for the study due to the impact these conditions had on CM risk.
METS participants presented annually for a health check, which included CM risk measurements, and self-reported alcohol consumption. Two alcohol consumption metrics were used to estimate drinking behavior: the number of drinks per day and the frequency of drinking. The clinical measurements were used to estimate the prevalence of CM risk at each of the five sites.
Clinical measurements
All measurements were collected at outpatient research clinics in the respective communities. Weight (kg) and height (m) were collected while participants wore light clothing and no shoes, using identical procedures and equipment models across all five sites. This information was used to calculate body mass index (BMI) (kg/m2). BMI was used both continuously and categorically as “not overweight” (BMI < 25 kg/m2) and “overweight/obese (BMI ≥ 25 kg/m2). Waist circumference was also collected using tape measures and recorded in both centimeters and inches.
Blood pressure
Blood pressure (BP) measurements were performed by trained personnel using an automatic digital monitor (Omron model HEM-747Ic). Measurements were taken with the antecubital fossa at heart level. BP was measured twice during the research visit with at least an hour in between measurements. BP was measured in triplicate each time, thus resulting in a total of six BP measurements. BP was coded to be a dichotomous variable based upon whether the participant was hypertensive. In concordance with the American Heart Association’s definition, elevated blood pressure was identified by systolic blood pressure ≥ 130 mmHg, diastolic blood pressure ≥ 90 mmHg or if the participant was being treated for high blood pressure [26].
Biochemical measures
Blood glucose levels (mg/dL), cholesterol (mg/dL) and triglyceride levels (mg/L) were obtained during the clinic examination. Participants were asked to fast 10–12 h prior to examination, during which fasting blood glucose was measured and blood samples were drawn by venipuncture using standard procedure. Fasting capillary glucose concentrations were determined using finger stick (Accu-check Aviva, Roche). Blood samples were briefly processed, and plasma or serum separated within 2 h of collection. Afterwards, all samples were stored in the laboratory at − 80 °C. Elevated blood glucose was defined as ≥100 mg/dL or receiving treatment for type 2 diabetes, while elevated high triglyceride levels were defined as ≥ 150 mg/dL or receiving treatment [3, 4]. Low HDL cholesterol was defined as < 40 mg/dL in men and < 50 mg/dL in women or receiving treatment [3, 4].
Physical activity
All participants wore an accelerometer (Phillips Respironics, Bend, OR, USA), and instructed to wear it continuously over their right hip. They were advised to only remove it before bathing, showering, or swimming as the device was not waterproof. The monitor measures and records vertical accelerations to represent the intensity of each movement. The intensity and duration of physical activity was summarized in one minute epochs over the course of seven days. Light, moderate and vigorous intensity activity were defined as < 1535 counts per minute (cpm), 1535–3959 cpm and ≥ 3960 cpm, respectively [27]. Physical activity was measured as minutes performed per day.
Questionnaires
All questionnaires were administered by trained personnel, as previously described [14]. The questionnaires inquired on each participant’s health information with a focus on cardiovascular disease and type 2 diabetes. Data collected included dietary habits, medication use, physical activity behavior and age of diagnosis for relevant illnesses. The study also collected data on household characteristics, participant occupation, education, parental education, household assets and amenities. Questions were based on the Core Welfare Indicators Questionnaire from the World Bank [28].
Drinking status
The main exposure for the current study was alcoholic intake, and participants were asked about their alcohol consumption during the staff-administered interviews, and the responses were used to determine patterns of alcohol consumption. The Alcohol Use Disorder Identification Test was used to determine daily alcohol consumption, which includes questions such as: “In the past 12 months, how often have you drink (in days)? Per Week, Per Month, Per Year?”, “On those days that you drank, how many drinks did you have?”, and “On special occasions, how many drinks do you have?” Participants were categorized into 3 alcohol consumption groups, using the guidelines provided by the National Institute of Alcohol Abuse and Alcoholism (NIAAA) [29]. These categories were non-drinkers (who indicated no alcohol consumption), light drinker (who consumed 1–3 drinks daily for men and 1–2 drinks daily for women) and heavy drinker (who had 4 or more drinks every day for men and 3 or more drinks per day for women).
Cardiometabolic risk
The main goal of the study was to measure elevated CM risk. Participants were classified with high CM risk, if they met any three of the following five CM criteria established by the Adult Treatment Panel III [3, 4]. These included abdominal obesity (waist circumference > 102 cm in males and > 88 cm in females); elevated BP (≥130/80 mmHg), or receiving treatment for hypertension; high TG (≥150 mg/dL); low HDL (HDL cholesterol < 40 mg/dL in males and < 50 mg/dL in females); and elevated blood glucose (fasting plasma glucose ≥100 mg/dL).
Statistical analyses
Participant demographic and clinical data were summarized using mean ± standard deviation, both individually and overall. Categorical variables were displayed using N and frequency. The relationship between alcohol consumption (no alcohol, light drinking or heavy drinking) and CM risk was explored using logistic regression, and final results were determined using odds ratio (OR) and 95% confidence intervals (CI). Both daily consumption and drinking frequency were used for these comparisons. Abstaining from alcohol consumption was used as the reference category to which light and heavy drinkers were compared. All models were adjusted for age, sex, smoking status (smoker or non-smoker), self-reported physical activity and site (dummy variables). Statistical analyses were performed using Stata (version 12, Manufacturer, College Station, TX, USA) (year 2011). An alpha p-value of 0.05 was used to denote statistical significance.
