Study Population
The present study was conducted in 2017 with a cross-sectional design, targeting elderly patients under care in geriatric healthcare centers in Isfahan city, Iran. The sample selection was performed as cluster sampling; based on a sample size calculation formula (prevalence: 0.5, standard error: 0.07, design effect: 1.5, significance level: 0.05), the target number of participants in this study was set at 300. The inclusion criteria were as follows: ability to perform daily activities, living in geriatric healthcare centers, having T2DM, and aged above 60 years. In other words, those with the following characteristics were not included: inability to do daily activities, not living in geriatric healthcare centers, not having T2DM, and age less than 60 years. Note that T2DM diagnosis was made by a skilled medical team and based on the medical history, drugs taken, and laboratory data registered in the patients’ file. Next, the participants responded to the questions related to anthropometric information, diet, lifestyle, plus sociodemographic information. Blood samples were taken by a skilled physician from each subject to perform biochemical analyses.
Dietary assessment
The typical dietary intake of each subject was assessed by an experienced dietitian using a 168-item food frequency questionnaire (FFQ). This questionnaire includes a list of typical food items consumed by Iranians. Based on the data obtained, the frequency of food consumption over the last year was also determined on a daily, weekly, and monthly basis for each subject, with the daily consumption of each food item in the last year calculated for every senior participant. Additionally, the level of coffee consumption was quantified for each subject in the last year (cup per day/week/month). Portion sizes of the food items consumed were converted to gram/day through household measurements. To calculate to total calorie, micronutrients, and macronutrients, nutrition IV software (Tinuviel Software, Warrington, UK) was employed.
Note that in this study, no question was asked about method of coffee preparation (espresso, filtered, instant, etc.) from the subjects. However, they responded to questions about the method of coffee consumption such as adding extra materials including sugar, artificial sweeteners, milk, or others. The daily coffee consumption was categorized into three groups based on standard cup size (50 ml): less than 1 cup per day (low), 1-3 cups per day (moderate), and more than 3 cups per day (high). The reliability and validity of FFQ have been demonstrated in previous studies (17). The validation of this questionnaire was also measured following comparison of dietary intake obtained from 24-h dietary recall in 12 consecutive months (one year) with the dietary intake obtained from FFQ. The reliability of the questionnaire was also examined by comparing the food groups obtained from two separate FFQs which had been completed in one year. The results obtained from Pearson correlation statistics revealed acceptable agreement between 24-h recall in 12 consecutive months (one year) and FFQ, as well as between two FFQs (completed in one year), respectively. This questionnaire also showed notable reliability with regards to coffee consumption during a one-year period (18, 19). Generally, the above points suggest that FFQ can also present valid assessment for coffee consumption in the long term.
Basic Characteristics
The lifestyle as well as sociodemographic information of patients included age, gender (male/female), smoking tobacco (yes/no), level of education (<primary school; ≥ primary school), alcohol consumption (yes/no), use of drugs (yes/no), patient-reported physical activity (low, moderate, high), and history of underlying disease (yes/no; except for T2DM). Anthropometric measurements captured weight, height, waist circumference. The weight of each patient was measured by a standard digital balance (accuracy: 100 g) with minimal clothing and barefoot). The patients’ height was measured by a measuring tape (barefoot) with accuracy of 0.5 cm, after which body mass index (BMI) was determined via wt/ht2 formula. Finally, waist circumference was measured through an inelastic measuring tape and in the thinnest part below the ribs.
Biochemical analyses
The blood samples of each participant were collected after 12 h of fasting. This was performed based on standard protocols and by an experienced medical team. Using a proper kit and via enzymatic-colorimetric glucose oxidase method, the fasting plasma glucose (FPG) was measured. The enzymatic-colorimetric method was utilized to measure the lipid profile including total cholesterol (TC), low-density lipoprotein (LDL-C), high-density lipoprotein (HDL-C), and triglyceride (TG). Plasma homocysteine was also measured by ELISA method and via Axis/homocysteine EIA kit (IBL Co., Germany). Finally, systolic and diastolic blood pressure (SBP and DBP) of patients was measured by a proper sphygmomanometer (Microlife BP AG1-20, Espenstrasse 139/ 9443/widnau/Switzerland) by a trained nursing technician.
Statistical analysis
The variables examined in this study were described according to coffee consumption classification, and reported as number, percentage (%), and Mean±SD in relevant tables. The comparison of categorical variables was also done among coffee consumption subgroups via Chi-square. This comparison for continuous variables was performed through One-way ANOVA test and Kruskal-Wallis test. Note that data normality was tested by Kolmogorov-Smirnov test. Also, homogeneity of variances for the data of study was assessed by Leven’s statistics. The relationship between each CVD factor and coffee consumption level (cups/day) was captured using partial correlation test. In this relationship test, the levels of energy intake, caffeine consumption, smoking cigarette, physical activity, BMI, gender, added sugar, history of underlying disease (except for T2DM), history of taking drugs, and level of education were also controlled. All of the analyses performed in this study were done by SPSS software (version 22; Chicago, IL, USA), with the significance level considered p-value<0.05.