This cross-sectional study was conducted in Gondar city, Northwest Ethiopia in July 2005. Gondar is a zonal capital city located 750kms north of Addis Ababa in Amhara Region. The city has a longitude and latitude of 12°36′N 37°28′E. Based on figures from the Ethiopian Central Statistical Agency in 2005, Gondar has an estimated total population of 194,773 of whom 97,625 were males and 97,148 were females. Sample size was calculated based on expected estimates of 50% of BMI < 18.5, 95% confidence limits, and a 5% marginal error, the required sample was 384. Probability sampling in a form of simple random and two-stage probability sampling method was used for selecting the required size. The first stage of the sampling was started by selecting kebeles (smallest administrative unit) using simple random sampling. At the second stage, a random sample of households was selected based on a sampling frame from the 1994 census and adapted for recent population changes.

Out of 384 participants, data of 28(7%) of the study participants were incomplete and excluded of the statistical analysis. Nutritional status and dietary intake indicators was primary variables of interest. In addition, a structured questionnaire was used to collect information on socio-demographic variables including sex, age, religion, marital status, occupation, educational status and monthly family income. Monthly family income was estimated by combining incomes reported for husband, wife, son and/or daughter. The inclusion criteria for participation were age >18 year, not acutely ill at the time of survey and not diagnosed for chronic illnesses. Ethical approval for this study was obtained from the Research Ethics Committee of the University of Gondar. Informed consent was obtained from all subjects.

Body weight (kg) was measured using an electronic scale to the nearest 10 g, and standing height was measured using a wall stadiometer to the nearest 0.1 cm. Subjects were instructed to take off their shoes before performing these measurements. Body Mass Index (BMI) was calculated as body weight (kg)/height (m²). The classifications of BMI applied in this study were recommended by the World Health Organization (WHO) 13 BMI values of <18.5 kg/m² and >25 kg/m² represented thinness and overweight, respectively. An acceptable weight was considered to fall within these two extremes. Waist and hip circumferences were measured with a flexible steel metric tape at the nearest 0.5 cm. Central obesity was also calculated and defined on the basis of WHR. The cut-off value of central obesity was considered high risk WHR= >0.80 or waist measurement >80% of hip measurement for women for females and >0.95 for males that is >95% for men indicates central (upper body) obesity and is considered high risk for diabetes & CVS disorders. A WHR below these cut-off levels is considered low risk 13 .

Data were collected by face-to-face interview using a structured Food Frequency Questionnaire (FFQ) modified from the Helen Keller International FFQ that was used previously in Ethiopia, to estimate meat and vegetable consumption that was in addition to the staple food intake 14 . The FFQ included eight food categories (Meat, Egg, Fish, Fat rich food, Vegetables, Fruits, Diary products, Sweet food) and was designed to obtain qualitative information about the usual food consumption patterns with an aim to assess the frequency with which certain food items or groups are consumed during a specific time period 15 . All frequency variables were coded as never or hardly ever, once a month, 2–3 times a month, once a week, 2–3 times a week, 4–6 times a week, and at least once a day.

The respondents were asked to recall the exact food intake of the previous day. Detailed descriptions of all foods including recipes and beverages consumed were recorded. Quantities of food consumed were estimated in household measures. One single 24-h recall was collected for every participant. Only one adult individual was selected from a house hold. For the transformation of household measurements and centimetres into grams, the portion sizes were weighed with a digital household dietary scale (Omron Electronic kitchen scale, Omron, Tokyo, Japan). Information from the 24-h protocols was entered and analyzed with Microsoft EXCEL software. The various food items mentioned in the recall were transformed into their corresponding weight of raw food ingredients. Ethiopian food composition tables 16 or food composition table for use in Africa 17 , for those not available in the former, was used to calculate energy and nutrients content. Major nutrients in the food composition tables were measured. The data were subsequently converted into the amount of energy and nutrient intake per individual per day. Relative validity of 24-h recall was determined by comparison data obtained from the same participants using a food-frequency questionnaire. Furthermore, three 24-h recalls were repeated in 10% of the sample. The dietary results are under preparation.

Adequacy of the macronutrients and micronutrients intake was evaluated according to the Dietary Reference Intakes (DRI) of The Institute of Medicine of The National Academies 18 . The reported energy intakes were compared with estimated minimal energy requirements to assess adequacy. Basal Metabolic Rate (BMR) was estimated using the sex and age specific equations of FAO/WHO/UNU expert consultations. The BMR was then multiplied by a factor which stands for physical activity level for each individual 19 .

As a measure of overall nutrient adequacy, mean adequacy ratio (MAR) was calculated as the mean of the nutrient adequacy ratios (NARs) for the intake of energy and nine nutrients (protein, calcium, iron, phosphorus, retinol, thiamin, riboflavin, niacin, ascorbic acid), each truncated at 1 so that a nutrient with a high NAR could not compensate for a nutrient with a low NAR 20 .

The mean ± SD daily nutrient intake was computed and tabulated. The mean intakes of energy, macronutrients and micronutrients were compared between men and women by independent sample t-test. Chi square test of proportion was used to determine the percentage of participants with intakes at or below the recommended daily allowance and adequate intakes. Correlation test was tested to examine the relationship between socioeconomic factors on dietary intake and selected nutritional variables. All statistical analyses were undertaken using SPSS version 13. P values less than 0.05 were considered statistically significant.

Of the 384 study participants, 356 were studied (93% response rate). Of the 28 excluded, 8 were absent on the day of the interview, 5 adults refused on behalf of their household and data for the rest was not complete. The details of socio-demographic characteristics of the study participants are presented in Table1. Of the 356 participants, 255 (71.3%) were females and 101 (28.7%) were males. Their mean age was 37.3 years (SD = 13.1; range 18 – 80 years). A substantial majority of them were Christians (90.2%) and married (57%). About 14% of the sample had no formal education. An additional 26.1% had received some years of primary school education, whereas only 16.8% had tertiary level training. The majority of males (54.5%) were government employees while that of females were house wives (37.6%). Twenty four percent of the sample lived with a monthly income of less than 30 US dollar and additional 28% lived for less than 60 US dollar per month.

Mean anthropometrical measures are presented in Table2. Men were taller, heavier and had higher waist to hip ratio compared to women (P < 0.05). Overweight subjects composed 21.3% of the total sample population, whereas obese subjects composed 5.9% of the above sample. Among females, 19.6% were overweight and 7.1% were obese, whereas among males, 25.7% were overweight and only 3% were obese. The cutoff points used for classification of participants as overweight and obese were similar to those introduced by the WHO 14 . Waist circumference was higher in males than females (P < 0.05) but hip circumference was higher in females (Table2).

Table3 compares mean daily intakes of food in men and women. Mean overall consumption of Fish, fruits and vegetables, separately and totally, was lower in large proportion of the subjects. Oil and butter was eaten daily by most subjects (n = 310). Meat, fish, sweets, milk and yoghurt were consumed in significantly higher amounts by women (P < 0.05). However, although not statistically significant, men's mean consumption of oil and butter was higher than women's.

^aResults are expressed as the mean ± SD of the participants consuming different frequencies of each food items in the study period. NS: not significant.

Table4 shows the average daily per capita consumption of the various food items by the sample population in comparison to previous data from a national survey. More than half of respondents reported intake of energy-dense food and alcohol. One fourth of men (25/101) and 6.6% (17/255) women reported having consumed alcoholic beverages (beer, tela or katikala) during the previous day. Only two men reported to have consumed katikala (home brewed liquor). Beer consumption was reported by 10 women and 22 men (mean intake 1047 ml and 762 ml, respectively) while tela consumption was reported by four men (mean 488 ml) and seven women (mean 779 ml).

Mean energy, macronutrient and fiber intakes of the study subjects and comparison of percentage contribution to total energy from macronutrients is presented in Table5. Mean energy intakes was significantly higher in men participants (3001 vs 2510 kcal/day, P = 0.007). However, the mean energy intake for both men and women was not significantly different from the estimated mean energy requirement (2234 vs 2167, P = 0.3). About 45.5% (162/356) of the participants had reported energy intake within 80-120% of the estimated requirement while 18.0% (64/356) reported its intake above 120% of the estimated requirement. Reported energy intake was higher than estimated energy requirement in 122 study participants. The mean fat, protein and carbohydrate intake (g/day) was 80, 79 and 320 and their percentage contribution for total energy was 33.0%, 14.1% and 52.9%, respectively. Men had significantly higher energy and macronutrient intake than women (P < 0.001). Protein intake was inadequate (<0.8 g/kg/day) in 11.2% (40/356) of the participants. Only 2.8% (10/356) reported carbohydrate intake below the Recommended Dietary Allowances (RDA) (130 g/day). About a third (31.7%, 13/356) of the study subjects had fat intake which contributed for less than 30% of total energy per day. Mean dietary fiber intake (19 g/day) did not meet the prudent dietary recommendation (38 g/day for men and 25 g/day for women).

a: mean ± SD, b: median (range), c: Ranges of population energy intake goal as % of total energy, *: estimated energy requirement.

**: recommended dietary allowances.

Average intake of minerals and vitamins of the subjects and prevalence of inadequate micronutrient intakes, that were computed based on RDA reference values, are presented in Table6. Inadequate intakes of calcium, retinol, thiamin, riboflavin, niacin and ascorbic acid were seen in 90.4%, 100%, 73%, 92.4%, 86.2% and 95.5% of the participants whereas intakes of iron and phosphorus were found to be adequate except in a few subjects 0.3% and 1.4%, respectively (Table7). Mean MAR was 0.74 for the total sample. A diet that covers the recommended intake for all nutrients has a MAR of 1 and a MAR below one indicates lower than the recommended intake for one or more nutrients 21 . A significantly higher proportion of women were deficient in calcium, thiamin and niacin compared to men while the proportion of inadequate retinol, riboflavin and ascorbic acid intakes were similar between the two sexes.

a: mean ± SD, b: median (range), c: number (proportion) of subject with inadequate intake.

**: recommended dietary allowances.

*: nutrient adequacy ratio, **: mean adequacy ratio, : recommended dietary allowances.

The correlation between socioeconomic variables and frequency of food consumption was tested. Income at the time of data collection had a positive correlation with BMI and level of education (p < 0.01). In addition, the correlation between income and frequency of consumption for most of the foods were significantly positive (p < 0.01). The correlation coefficients between BMI and the food consumption frequency were significantly positive for meat (r = 0.36; p < 0.01), egg (r = 0.177; p < 0.01), vegetables (r = 0.252; p < 0.01), fruits (r = 0.263; p < 0.01), sweets (r = 0.124; p < 0.05) and Milk (r = 0.217; p < 0.01). Similarly, the association between level of education and food consumption frequency was positive, except for oil and butter (r = −0.082) (Table8).

** Correlation is significant at the 0.01 level (2-tailed).

* Correlation is significant at the 0.05 level (2-tailed).