The study was conducted in Nsukka urban. Nsukka is located in the northern part of Enugu State, Southeast, Nigeria with a total population of 309,633 people as at 2006 national census increasing at an annual rate of 3.0%. Major occupation includes farming, trading and civil service. Major crops and livestock consumed are cassava, yam, maize, cocoyam, rice and sweet potato, poultry, pigs, goats and sheep.
Study design and participants
The study employed retrospective cross-sectional cohort design in the study of energy status and factors associated with energy balance of young adults (20–39 years). The study population comprised of all free living non pregnant non lactating young adults (20–39 years) in Nsukka urban. Those who refused to be included by not signing informed consent or unable to supply data for three consecutive days were also excluded.
Sample size calculation
Sample size for the study was calculated using modified Cochran’s formula: N = 4P (1-P)/W2. Margin of error (5%), non-response rate (5%) and p value of 17.0% which is the prevalence of obesity among urban Nigerian adults was used to obtain a sample size of 240 .
A multi-stage probability sampling technique was used in selecting the respondents. In stage one, two (2) wards (Ihe and Mkpunano) out of 4 wards that make up Nsukka urban were selected using simple random sampling technique by balloting without replacement. In the second stage, one community (Onuiyi from Ihe and Umuakashi from Mkpunano) was selected from each ward by simple random sampling. In stage three, urban settlements (Onuiyi from Onuiyi and Army Barracks from Umuakashi) in the two communities were identified and included (on the basis of population density and ease of access to transport). Stage four involved systematic random selection of every 5th living house in the area. Probability proportional to size was adopted. In the fifth stage, one household was selected from each house by simple random sampling technique. In the sixth and final stage, only two young adults within the ages of 20–39 years were selected from each selected household by simple random sampling using balloting without replacement. Where there was only one eligible adult, a second household was selected from the same house and if there was none, the next house was selected and stages five and six repeated.
Ethical clearance and consent to participate
Ethical approval for the study was obtained from Health Research Ethical Committee, University of Nigeria Teaching Hospital (UNTH) Ituku-Ozalla, Enugu State (NHREC/05/01/2008B-FWA00002458-1RB00002323). After details of the study were explained to them, respondents were requested to sign an informed consent form indicating their willingness to participate in the study.
Data collection methods
A validated questionnaire was used to obtain data on socio-demographic, dietary habits and lifestyle characteristics of respondents. WHO global physical activity questionnaire administered by trained interviewers was used to assess physical activity level of the respondents.
Weight was measured to the nearest 0.1 kg with 120 kg capacity Hanson’s bathroom weighing scale. Participants stood erect in minimal clothing with arms hanging by the sides and no shoes on. Height (in cm) was taken with height meter rule with bare feet parallel to each other and heels, buttocks, shoulders and back of head touching the height meter rule. Body mass index (BMI, kg/m2) derived as weight to height ratio (weight in kg/height in metre squared) was used to classify subjects into underweight (< 18.5), normal weight (18.5 – 24.9), overweight (25.0 – 29.9) and obese (> 30). Waist circumference (WC) in centimetres was measured at the end of expiration using a flexible, non-stretchable tape placed at the midpoint between the top of the iliac crest and lower margin of the last palpable rib while participants stood upright. Hip circumference (in cm) was measured around the widest portion of the buttocks. Ratio of waist to hip circumference (WHR) was calculated. WC > 94 cm in males and > 80 cm in females were taken as abdominal obesity; WHR > 0.85 for females and > 0.90 for males were considered high (health) risks .
Three 24-h dietary recall involving two weekdays and one weekend day and a total of 6 meals per day was conducted by trained interviewers to determine the energy intake of the respondents [13, 14]. Respondents were requested to describe the types, brand names and quantity of ingredients, method of preparation/cooking and portion sizes of foods (meals, snacks and beverages/drinks) consumed during the period under study whether at home or outside the home. Quantification of the reported foods and beverages/drinks was achieved with weights and volumes of household measures (cups, glasses, bowls, jugs, spoons, plates, slices) and food items/models of different sizes. Estimated amounts were weighed using kitchen scales and the results recorded in grams. Macronutrient (protein, carbohydrate and fat) values of the foods were obtained from West African and Nigerian food composition tables and results of food/diet analysis reported in journal articles [15,16,17,18]. These were used to estimate the energy values of each food/snack and beverage/drink consumed based on Atwater factor of 4, 4 and 9 kcal/g for protein, carbohydrate and fat, respectively. The values for the three days were summed up and divided by three to obtain the mean daily energy intake. The mean values were used in statistical analysis.
Total energy expenditure (TEE) was determined as the sum of resting energy expenditure (REE), energy expenditure of activity (EEA) and diet-induced energy expenditure (DEE) based on three days’ assessment. Mean of the total energy expenditure (kcal/day) for the three days was used in statistical analysis.
Resting energy expenditure (kcal/day) was obtained through Harris-Benedict’s predictive equation [14, 19] for males (66.5 + (13.75 × weight in kg) + (5.003 × height in cm) – (6.75 × age in years)) and females (655.1 + (9.563 × weight in kg) + 1.850 × height in cm) – (4.676 × age in years)).
Energy expenditure of activities (EEA) was obtained by multiplying REE with physical activity factor of the respondents’ physical activity level (PAL). PAL was determined with WHO global physical activity questionnaire that provided detailed report of types, intensity, frequency and duration (in minutes) of all physical activities (exercise and non-exercise) performed daily for three (3) consecutive days by the respondents [14, 20]. Total physical activity was calculated by summing all the minutes spent on each physical activity category and categorised accordingly into sedentary/light (less than 30 min a day), moderately active (regularly active or accumulated ≥ 30 min per day) and vigorously active (greater intensity activity in ≥ 8–10 min’ bouts in a day) . Physical activity level factor of 1.4 for sedentary/light activity, 1.70 for moderately active, 2.0 for vigorously active were used to account for individual energy expenditure of activity [22, 23]. Energy expenditure of activity (EEA) = Activity factor × REE (kcal/day). Mean of the three days’ values was used in statistical analysis.
Diet-induced energy expenditure was calculated as 10% of total calories consumed in a day [4, 22]. Mean of the three days’ values was used in statistical analysis.
Mean of three days’ energy expenditure was subtracted from the mean of three days’ energy intake to obtain energy balance and interpreted thus: energy intake > energy expenditure = positive energy balance; energy intake < energy expenditure = negative energy balance; energy intake = energy expenditure = equilibrium (energy balance).
Outcome and predictor variables
The binary outcome variable is energy balance (positive or negative) whereas the exposure variables (covariates) were socioeconomic (age, sex, education, occupation, marital status and income), dietary (skipping meals, number of meals consumed in a day, weekly snack consumption and eating outside the home), lifestyle variables (alcohol consumption, smoking of cigarette/substances), body mass index and waist circumference. Relationships between the outcome and exposure variables were assessed at both the binary and multivariate logistic regression. After examining the individual effects of the above 14 exposure variables at the binary level, they were entered simultaneously into the multivariate logistic model to evaluate the effect of each of the covariates on the outcome variable when other covariates are held constant. Crude and adjusted odds ratios were reported for each of the covariate evaluated.
Data collected were entered into Microsoft excel, validated, cleaned and sorted before being transported into IBM Statistical Product and Service Solutions (version 21) computer software for descriptive and inferential statistical analysis. Descriptive statistics (frequencies and percentages) was used for general characteristics, anthropometric and physical activity levels of the adults. Chi square test was used to evaluate the relationship between categorical variables (anthropometric parameters and physical activity level of the respondents by age and sex as well as the relationship of these parameters with energy intake, expenditure and balance). Means and standard deviations were used for energy intake, expenditure and balance. T-test was used to assess relationships between energy intake, expenditure and balance, and sex, waist circumference and waist hip ratio. Whereas analysis of variance was used to compare the energy parameters among four age groups of the adults and assess the relationship of mean energy intake, expenditure and balance with anthropometric parameters and physical activity level. Binary logistic regression analysis was employed to evaluate associations between the outcome variable and the predictor variables. Since binary logistic regression analysis does not control confounding effects, multivariate logistic regression analysis was conducted to correct for simultaneous effects of multiple factors and control the effects of confounding variables on the response variable. The adjusted odds ratios were used to define the independent strength of the associations. Significance was accepted at 95% precision (P < 0.05).