We have shown takeaway food consumption is associated with a poorer diet quality and a higher prevalence of moderate abdominal obesity in young Australian adults. Different socio-economic and lifestyle factors are associated with a higher frequency of takeaway food consumption in men and women.
Differences in the methods used to ascertain takeaway and fast food consumption and the definition of takeaway or fast food used make it difficult to compare findings across studies. The frequency of takeaway food consumption in the current study was higher than that reported in a Mediterranean population (aged 24–75 years) where only 1.1% were consuming fast food at least twice per week but only hamburgers, cheese burgers, Big Macs and French fries were included as fast food . A study in the USA reported 30% of men and 24% of women (aged 20 years and older) had consumed fast food on at least one of the two days studied using 24-hour diet recalls .
The socio-economic and lifestyle characteristics we found to be associated with higher frequency of takeaway food consumption were similar to those found in previous studies: younger age [5, 8], being single [7, 9] and watching more television . However, to our knowledge this is the first study to report characteristics of takeaway food consumption separately for men and women. Being single and spending more time watching TV and sitting were associated with takeaway food consumption in both sexes. In men, being younger and a current smoker were also associated with takeaway consumption whereas in women, there was an association with employment status.
We found men consumed takeaway more frequently than women, which is consistent with some studies [6, 37], but not others [8, 9]. In contrast to a previous study  we found no significant association between takeaway food consumption and alcohol consumption in men or women.
Studies of socio-economic position and diet quality report that people of lower socio-economic status consume diets that are higher in energy dense foods such as takeaway foods . However, our measures of socio-economic status (employment status and education) in this sample of young Australian adults do not support this. Participants who were not in the workforce were not high consumers of takeaway food, possibly because they could not afford to purchase it, and education had no association with takeaway food consumption. Previous studies investigating associations between income and takeaway food consumption have reported mixed results with some studies reporting participants with a high income to be the highest consumers [6, 11], some showing participants with a low income to be the highest consumers , and yet others showing no association . Education also shows mixed results with the majority of studies being consistent with our finding of no association [6, 8, 15], but others have reported positive associations with high education  or low education .
The number of participants achieving individual dietary recommendations was very low and lowest in participants who were eating takeaway food more frequently. This suggests takeaway food is not just an additional food item in an otherwise healthy diet but is associated with a number of other unhealthy eating behaviours, possibly by displacing healthier items from the diet. Our findings are similar to previous studies from the USA and Spain that report a higher frequency of takeaway or fast food consumption is associated with a lower intake of fruit, vegetables and dairy [5, 7–9]. A higher frequency of takeaway food consumption was associated with a lower intake of breads and cereals in men and a lower intake of lean meats and alternatives in women. Overall participants eating takeaway food more frequently met fewer of the dietary recommendations. This supports the previous study in Spain that examined overall diet quality, where participants eating fast food at least twice per week had the lowest adherence to the Healthy Eating Index and the Mediterranean Diet Score .
This is the first study to show that young adults eating takeaway more frequently have a somewhat higher prevalence of moderate abdominal obesity as measured by waist circumference. Women eating takeaway food twice a week or more had a higher prevalence of being overweight or obese as defined by a BMI ≥ 25 kg/m2 and this association remained significant after adjusting for covariates. However, in men, an association was only seen at the higher level of BMI (≥ 30 kg/m2) and the association was not significant. While we were unable to adjust for energy intake, because this was not available from the FFQ, we did take into account key determinants of energy intake by stratifying the analysis by sex and adjusting for age and physical activity levels. Due to the cross-sectional analysis, we cannot be certain of the direction of a causal relationship between takeaway food consumption and abdominal obesity. Although our study sample comes from a cohort study, longitudinal analysis is not possible because comparable dietary data were not collected in childhood. Previous studies have found an association between takeaway and fast food consumption and BMI  and changes in weight over time [5, 7, 12, 39].
There are several limitations with the dietary recommendation analysis. First, the response option for the vegetable question combined four and five serves per day, and the proportion meeting the vegetable recommendation (at least five daily serves) is likely to be lower than that reported here. In addition, compliance with the lean meat and alternatives recommendation may be overestimated due to the large number of items included in this variable. However, previous national data show consumption of meat and alternatives is high in Australian adults . Second, although we excluded from the analysis participants who had not adequately completed the questionnaire (those that failed to complete > 90% of the FFQ), we were left with occasional non-responses to items by the remaining respondents. These were assigned a value of zero on the grounds that a non-response indicated the respondent did not eat that food. However, some of these missing items may have been overlooked by the respondent. If so, this would have resulted in under-estimation of the proportions of respondents meeting the dietary recommendations. It is reassuring that this measurement error did not appear to be differential between the two takeaway food groups, with the exception of the lean meat and alternatives food group. Third, the guidelines recommend consuming wholegrain breads and cereals; apart from bread, the FFQ did not distinguish between wholegrain and non-wholegrain items. Fourth, components of mixed dishes were not included as items in the food groups and may be under-estimated. Mixed dishes are generally difficult to assess using FFQs .
A strength of this study was that we asked about usual takeaway food consumption and, in addition to food available from the main fast food chains (McDonalds, Pizza Hut, KFC etc), our takeaway food variable included other popular takeaway food options such as Indian, Thai and Chinese foods. Furthermore, this is the first study to report associations of takeaway food consumption with lifestyle factors separately for men and women. We did this because we were interested in examining potential sex differences in takeaway food consumption to better understand the predictors of this eating behaviour. Other strengths include the use of a FFQ that has been used in previous national surveys, and examining overall diet quality, which has been done in only one previous study.