Dietary differences between elderly Iranians living in Sweden and Iran a cross-sectional comparative study

All Iranian-born persons aged 60-84 settled in the township of Kista, Stockholm (N = 270), were invited to participate in the study via a letter written in both Swedish and Farsi. One hundred and seventy-six persons (65 percent) agreed to participate in the study. Interviews were conducted face to face in Farsi, using a questionnaire based on material produced for the Swedish Annual Level of Living Survey by Statistics Sweden [13].

The non-response analysis, which was followed up by phone call to persons who refused to participate, showed no significant differences between participants and non-participants with regard to sex, education and self-reported weight and height. However, the non-participants were slightly older (p < 0.05) (mean age = 72.8 ± 7.7 years) than the participants (mean age = 70.5 ± 7.0 years).

In order to have the same age group as the Iranian group living in Iran, we excluded four men and 12 women older than 80 years. Of the 160 who participated, 8 had no dietary data. Thus, 152 individuals remained in the final sample.

The reference group consisted of 52 Iranians aged 60-80, within the framework of Tehran Lipid and Glucose Study (TLGS), a prospective study performed on residents of district 13 of Tehran with the aim of determining the prevalence of non-communicable disease risk factors and developing a healthy lifestyle to improve these risk factors [14]. However the Iranian population in Iran is rather young; in 2005 in Iran, the elderly population aged 60 and older was 6.5 percent of the whole population [15]. In addition in the TLGS population which included representative sample of 15005 participants aged 3 years and over of the urban population of Tehran, 12 percent were 60 years and older. Furthermore 1476 persons were selected randomly for dietary assessment in the TLGS population. 52 subjects (3.5 percent) aged 60 years and over were willing and able to participate in the dietary survey and all are included in this study.

A semi-quantitative food frequency questionnaire (FFQ) for the elderly Iranians living in Stockholm was adapted from the dietary survey of the Swedish population which was carried out by the Swedish National Food Administration in collaboration with Statistics Sweden [16]. The FFQ was originally designed to assess dietary intake in terms of energy and macronutrient intake (total fat, carbohydrate and protein) as well as intake of different food groups, and was modified to exclude prepared dishes and typically Swedish food that are not commonly eaten by Iranians. The modified FFQ consisted of a list of 114 food items and the standard serving size for each item. The FFQ was conducted in a day centre for Iranian seniors in Kista or in the participant's home face-to-face during 30 minutes by the first author (AK) of the present study. AK is familiar with the Persian cuisine and language. If a male participant were unable to describe the content of the food, we ascertained the content by asking the wife or a daughter. Photos from a portion guide [17] were shown to the participants in order to estimate portion sizes of the following food items: bread, cooked rice, mixed salad, fat spread on bread, feta cheese, and cooked pasta. Frequency of consumption was obtained by asking open-ended questions on how often a particular food item was consumed, i.e. per day/week/month or during the last year. Participants were also asked about the fat percentage of dairy products and margarine. Brand name of items like bread, jam, breakfast cereal, cookies/cakes, fruit juices and ice cream were asked to obtain a close match with items listed in the Swedish food tables. An open question at the end of the interview allowed individuals to report any other eaten foods not listed in the FFQ. The 114 food items were divided into 18 food categories, which were combined into six major food groups that matched the TLGSs classification of food groups (Table 1) [18]. Items that did not fit into any of the six groups were individually categorized (ice-cream, jam, nuts, sugar, sweets, potato chips, popcorn, cheese doodles, chocolate, honey, soft beverages and alcoholic beverages) and assessed for their content of macronutrients. Based on the reported amounts we calculated daily intakes of macronutrients and each food item in grams. Daily grams were computed for each food item as the daily frequency, e.g. monthly and weekly frequencies were divided by 30 and 7, respectively.

Data obtained from the FFQ were entered into DIETIST XP software [19], which is based on food-tables from the Swedish National Food Administration [20] that includes almost 1600 items and 50 nutrients.

Doubly-labeled water method can accurately evaluate total energy expenditure and can be use for assessing the validity of reported total energy intake [21, 22]. However, because of the high cost of this method, validation of reported dietary intakes are most frequently measured by Goldberg cut-off method which can be used to evaluate the mean population bias in reported energy intake [23]. Therefore, for each participant in Sweden, the FFQ was validated by the Goldberg cut-off [24], which was revised in the year 2000 [23]. The model is based on reported energy intake (EIrep), estimated basal metabolic rates (BMRest), and physical activity level (PAL).

BMRest was estimated from the Schofield equations [25] for persons older than 60 years as follows:

Data on self-reported leisure-time physical activity was recorded during the interviews, and subjects were categorized into 5 levels of physical activity: very light (almost no activity at all), light (walking, non-strenuous cycling or gardening approximately once a week), moderate (regular activity at least once a week, e.g., walking, bicycling, or gardening; or walking to work 10-30 min/d), active (regular activities more than once a week, e.g., intensive walking, bicycling, or sports), and very active (strenuous activities several times a week). PAL was systematically estimated for each subject according to a method developed by Johansson et al. [26].

Subjects were identified as under-reporters, acceptable-reporters, or over-reporters according to their EIrep/BMRest and the lower and upper cut-offs which were calculated as follows [23]:

Where s.d. min is - 2 for the 95 percent lower confidence limit, s.d. max is + 2 for the 95 percent upper confidence limit. S is the factor that accounts for the variation in EI, BMR, and PAL and is given by:

CVwEI is the within-subject coefficient of variation (CV) in EI, d is the number of days of diet assessment, CVwB is the CV of repeated BMR measurements, and CVtP is the total CV in PAL. CVtP is the CV derived from the mean and standard deviation of a study and includes true between-subject variation, an element of within-subject variation and methodological errors. The CVs used were 23 percent for CVwEI, 8.5 percent for CVwB and 15 percent for CVtP [16]. Subjects were defined as under-reporters if the ratio EIrep/BMRest was below the lower 95 percent confidence limits (< 0.56), acceptable-reporters if the ratio was within the lower 95 percent confidence limits and the upper 95 percent confidence limits (0.56 - 1.78), and over-reporters if the ratio was above the upper 95 percent confidence limits (> 1.78). Approximately 80 percent of the participants were acceptable-reporters, 14 percent were over-reporters, and 6 percent were under-reporters. After exclusion of subjects who were not acceptable reporters, 121 subjects (41 men and 80 women) remained in the final data analysis.

Role of migration process on health behaviour is likely to be associated with adoption processes to positive and negative lifestyles of the host country [10, 33, 35]. In Sweden, the elderly population, like others in the general Swedish population has nutritionally adequate diets in relation to recommended levels [36]. Hence higher intake of fruit, vegetables, fiber and dairy products among Iranians living in Sweden, can be a part of the adoption process to the Swedish dietary habits. This level of fruit and vegetable consumption equates with current Swedish food administration who recommends to eat 500 games of fruit and vegetables every day [37]. Moreover, the observed lower intake of carbohydrates among Iranians living in Sweden was reflected in lower amount of bread and grain consumed by this group. However, the fact that Iranians living in Sweden had higher fiber intake than Iranians living in Iran, suggest that the major part of the consumed fiber in Sweden should be a result of high consumption of fruit and vegetable which are major sources of soluble fiber. Moreover consumption of soluble fiber is associated with positive effect on plasma LDL-cholesterol, glucose level and decreased risk for cardiovascular disease [38].

Higher total fat and lower carbohydrate consumption suggests that Iranians living in Sweden have adopted dietary habits that are not in line with current nutritional recommendation in Sweden. A similar pattern of edible fat consumption was also observed in a study from Norway indicating that Pakistani women in Oslo substitute butter/ghee with margarine or vegetable oil because they believed that vegetable oil has less fat than butter [39]. For this reason we argue that the observed edible fat consumption can be due to misunderstanding about the beneficial effect of vegetable oil (e.g. olive and canola oil) among Iranians living in Sweden who may believe that vegetable oil is low fat and does not increase weight. Higher total fat intake including fat rich foods such as meat products and dairy is a point of concern in light of the high prevalence of obesity, among Iranians living in Sweden and can have an unfavourable impact on cardiovascular health of this immigrant group.

Food choice might also be influenced by limited economic resources among low-income groups [40, 41], which in turn leads to fat rich and energy dense dietary options, which are inexpensive and good tasting [42]. This may explain the high consumption of fat rich food items such as meat and dairy products among Iranians living in Sweden with very low incomes. Because most of Iranians living in Sweden in this study arrived in Sweden late in life and can therefore not be eligible for full retirement pension. However, the explanation that costs is more relevant than nutritional concerns on personal dietary options, contrast with the Norwegian study which indicate that food prices were not the most important predictor of dietary changes among immigrant women [39]. Though it is difficult to determine factors that may influence food choice among this group of immigrant in Sweden with different dietary habits compared to the population of their origin country.