Choline is an essential nutrient for humans, though it can also be synthesised endogenously by the human liver [1
]. Choline is thought to belong to the B-group complex of vitamins. Discovered in 1962, choline was isolated by the boiling of pig bile following a renewed interest in its deficiency status [2
]. These early studies concluded that humans did not need to eat choline from food sources. Later research uncovered important links between choline transport via the placenta [2
]. These study findings also uncovered an increased need for choline in certain population groups, e.g., males and post-menopausal females, resulting in the development of dietary guidance for this nutrient [3
]. It is speculated that oestrogen plays a role in the biosynthesis of choline [4
]. During pregnancy, however, females with low levels of vitamin B12 or folic acid may be at risk of inadequate choline, due to its role in methyl group metabolism [5
]. Choline donates methyl groups to homocysteine to form methionine in methyl group metabolism. DNA methylation increases for cell differentiation and organogenesis during embryogenesis and early postnatal life [5
]. The literature suggests that maternal choline supplementation is positively related to offspring neurocognitive outcomes [6
Given DNA methylation is an epigenetic modification which regulates the patterns of gene expression, DNA methylation patterns are critical in mammalian development and later physiological function in adulthood [8
]. Additionally, choline is a precursor to compounds that maintain human health, including cell membrane constituents (phospholipid and sphingomyelin) and neurotransmitters (acetylcholine) [5
]. Choline also plays a role in lipid and cholesterol transport [5
]. Studies suggests that low intakes of choline have been linked to cardiovascular disease [10
], neurological disorders such as Alzheimer’s disease [11
] and fatty liver disease [12
To determine levels of consumption within a population, suitable food composition data is required. Food composition databases are used in practice across a variety of areas, including food policy formulation, food labelling and dietary assessment [13
]. Food composition databases allow for food information reported by population groups during a dietary assessment to be translated into nutrient information. Databases also create a framework for food grouping systems that can be used for dietary pattern analyses in relation to the nutrients [14
]. Choline and its esters are widely found in foods. Although choline can be formed in the liver, the majority of people need to consume choline from dietary sources to meet their requirements [3
]. While many studies have used data from existing food composition databases, it is imperative that regionally-specific food composition data is used, due to the variation in food harvesting, processing and preparation conditions in different countries [15
]. Following the development of dietary guidance values for choline during the early 2000s, the United States Department of Agriculture sought to create choline values for its standard reference food composition database. This development also created retention factors for losses of choline due to cooking, with variations of between 70 and 100% retention. Where no data were available, choline retention factors were imputed from other B-group vitamins due to their similarities [16
]. Such developments have provided an impetus for other countries to create data for their food supply, though until recently, no data was available for Australian foods. The aim of this study was to develop an Australian choline database, and to apply this database to Australian population data to identify choline consumption patterns by age and gender.
The present findings, the first Australian representative intake estimates of choline at a national level, suggest that the estimated choline intakes in Australia (265.18 ± 1.3 mg; 151–311 mg) are below the values reported in the current literature though there is considerable variation in the methods applied between countries. For Australians aged 19–64 years, the mean choline intakes were 310.54 mg/day and 247.65 mg/day for males and females, respectively. The estimated choline intake based on the National Health and Nutrition Examination Survey 2013–14 in the USA reported mean choline intakes between 359 mg/day to 426 mg/day for males aged 20–69 years, and 275 mg/day to 296 mg/day for females [28
]. The European adult male intake estimates ranged from 357 mg/day to 468 mg/day (18–65 years), and 291 mg/day to 374 mg/day for females [29
]. The Australian intakes of choline for childbearing females and women during pregnancy and lactation were also less than the intakes revealed from the published studies. For example, the estimated choline intakes of females aged 18–40 years was 316 mg/day in New Zealand [30
]. During pregnancy in Canada it was 347 mg/day [31
], and Latvian values ranged from 336mg/d to 356 mg/day [29
]. However, the comparability of these choline intakes should be interpreted with caution. Each of these values are higher than the reported Australian values, though this may be due to the use of borrowed food composition data. Even though the USDA choline food composition database was applied in these studies [28
], different releases of the databases may contribute to the discrepancy between intake estimates. As the USDA choline database is the most comprehensive database for choline in foods globally, many countries will apply this data to the reported food intakes of their own country. Best practice recommendations for food intake assessment are for the use of regionally-appropriate food composition data when analysing food intakes. This was applied within the current study though a large proportion of the data was still borrowed, due to the limited Australian studies. This reliance on borrowed data stresses the need for more studies focused on the analysis of foods for their choline composition. These compositional analyses should align with the AOAC analytical methods, and be suited to the food matrix that is being analysed. By growing the number of Australian studies with choline content analyses, researchers can become less reliant on borrowed data, and lessen the need to create assumptions when matching a reported food type to a food composition value. For example, this was evident in this study for the mammalian game meats food group, which includes kangaroo meat. No published studies were identified, and borrowed data was required for the database development. Kangaroo meat is unique to Australia and, therefore, the closest equivalent meat type needed to be matched to the kangaroo values in the Australian AUSNUT 2011-13 food composition database, as suggested by the FAO INFOODS food matching guidelines [22
]. Subsequently, this may also impact the food group level outcomes as outlined below.
Another difference between the studies reporting population level choline intakes may be the dietary assessment method used in the survey, and the method of analysis. Studies have shown differences in data between the use of day 1, and day 2, as well as an average of days 1 and 2, and also the use of a regression model to represent usual intakes. This study applied usual intake analyses to the NNPAS data, which considered the reporting of both days 1 and 2. There are fluctuations around individual intakes reflecting true eating habits in free-living conditions. Thus, estimation of usual dietary intake will create a value that is more representative of habitual intakes, in comparison to using only one day of data.
Australian intakes for both genders and across the majority of age groups were not aligned with the AI levels. Similar trends of suboptimal choline intake were also found in the US population for different age groups, suggesting that only 11% of the US population aged ≥2 years met the AI for choline [32
Although the literature suggested that dietary choline intakes were largely contributed by the variation of energy intake [33
], the developed AI levels were largely from experimental studies, and were based on 7 mg/kg body weight. These levels were based on a 76 kg male and a 61 kg female [1
] to determine the targets. Due to the high rates of obesity within the Australian population, it is anticipated that a 7 mg/kg body weight intake level would not be achieved. The pregnancy AI values for choline add an additional 11 mg/day, based on the assumption that no additional choline is produced by the foetus or the placenta. The demands of choline during pregnancy and lactation increase [5
]. Choline, as a supply of methyl groups plays a critical role in stem cell proliferation and apoptosis, thereby influencing the structure and function of the brain and spinal cord in relation to a risk of neural tube defects and the lifelong memory function of the foetus and infant [5
]. The added requirements for choline during pregnancy are of a greater concern, given the extremely low levels of choline for the pregnant and lactating females in this study. Despite this, two in every three reported intakes for infants met the AI levels. This may be related to parental influence in eating patterns during this life stage, and the relatively small amounts of food required. The reporting of infant intakes by parents may also be influenced by a desirability bias, whereby the parent may unconsciously increase the amount of particular foods reported.
The major food source for choline in this study was found to be egg products and dishes for children and lactating females. This shifted for the adult intakes, though egg-related food groups continued to be within the top five ranked food groups. Foods derived from animal sources, such as eggs, meat, fish and milk, generally contain more choline than plant-based foods, such as grain, vegetable and fruit in per unit weight [25
]. In the USA, the major food sources of choline intake were meat, poultry and fish, grain-based mixed meals, dairy and eggs [28
]; where meat, milk, grain, eggs and their derived products, composite dishes and fish were the main contributors in European countries [29
]. The foods contributing to choline intake in New Zealand were eggs, red meat, milk and bread [30
]. Outcomes of this study were most comparable with those of New Zealand. Interestingly, due to the detailed food group categorisation in the present analyses, the top-ranked food group position for adults 19–64 years of age was found to be from the mammalian game meat food group, contributed by kangaroo meat, which is unique to Australian intakes. As outlined above, kangaroo meat did not have analysed choline values in the food composition database, requiring food data to be matched and borrowed from the USDA. Further, only a small proportion of the population reported consuming kangaroo meat, but the portions consumed were larger (~210g) by comparison to the portions consumed from the eggs food group (~90g) for example. Similarly, for pregnant females the other sea and fresh water foods food group contributions had a substantial impact on choline intakes. Again, this was reported by a small proportion of the population, likely due to pregnancy intake guidelines for this population group [35
While some similarities were seen between the Australian food group contributions and other countries, the discrepancy between the Southern and Northern hemisphere findings may be due to the level of detail of food items that were reported. Instead of grouping food items based on the dominant nutrients or ingredients, for example cereal and meat, the food items in the present analyses were categorised on the basis of similar animal/plant species or family, or sharing similar cooking methods and derived from a nested hierarchical food group classification system [36
]. The present study provides more detail on the actual foods consumed by the respondents in the NNPAS, which may have a more practical focus for dietary recommendations. This level of detail may also be indicative of the differences between the published intake studies for choline to date. Recipe calculations were also performed in the present study. Eggs are often used as an ingredient item in a number of dishes which may be categorised within food groups other than the egg products, and dishes food group. Creating a method to quantify the use of eggs within other food items of the food composition tables will allow for other smaller egg contributions to be included. Such an analysis at a recipe and food level may in future provide a more accurate estimate of total population intakes of eggs, which may also be apparent for milk when used as an ingredient. For the purpose of this study, however, the choline content would have been accounted for within each of the food groups used in the analyses.
There are a number of limitations to this study. Development of the choline database required substantial use of professional judgement. While quality assurance measures were implemented, some matches may require local knowledge of the food supply to ensure the matches are correct. Further, while population intakes were based on nationally-representative intake data, it was based on self-reported data only, which may be susceptible to bias. Adjustments of intake were also not made for the analyses reported in this study.
Food group predictors of Australian population choline intakes have not been reported in the published literature to date. This study found that meat, poultry and game products and dishes were the main predictive food group for choline intake, followed by egg products and dishes and cereal-based products and dishes. While these food groups were the primary predictors for choline intake, they predicted a combined 52.27% variation in choline intakes. The remaining predictors were weaker, and spanned across a wide range of food groups. These food groups were significantly influenced by covariates of the model, namely the gender, level of education and the energy intake of the respondents. These factors have also been associated with diet quality, whereby females compared with males [37
] and those with higher levels of education [38
] are reported to consume a higher quality diet.
In conclusion this study has provided preliminary data for a choline database to be used with Australian foods. It has aligned existing published values with entries of the Australian food composition survey database and applied these to recent consumption data to estimate population intakes. This has provided insight into food-based sources of choline for Australia. The data for the choline content of Australian food items should be updated over time to provide more accurate estimates of intake, particularly as newer consumption survey data becomes available.