Dietary Guidelines for Americans 2015–2020 have recommended replacing sugar-sweetened beverages (SSBs) with plain drinking water [1
]. The 2006 proposed guidance system for beverage consumption in the US also recommended choosing water over other beverages [2
]. Analyses of 24 h dietary intakes from the most recent National Health and Nutrition Examination Survey (NHANES 2011–2016) suggest that these recommendations may have been effective. In recent years, plain drinking water, bottled and tap, has been replacing SSBs in the US diet [3
]. The main sources of drinking water in the US have been tap water at home (288 mL/d), tap water away from home (301 mL/d), and bottled water from supermarkets and grocery stores (339 mL/d) [3
]. Most SSBs have also come from stores. Stores contributed far more SSBs to the US diet than fast food restaurants, full service restaurants, and schools combined [4
The 2015 Dietary Guidelines for Americans recommended a shift to reduce added sugar consumption to less than 10 percent of calories per day [1
]. Among the suggested strategies were drinking SSBs less often, reducing SSB volume, or replacing SSBs with plain water on specific eating or drinking occasions [1
]. Successful implementation of those strategies may require a better understanding of water and SSB consumption patterns during the day. The timing of and the frequency of drinking bouts and the amounts of fluids consumed can vary across population subgroups. Replacing SSBs with drinking water can also be challenging if the established SSB and water consumption patterns differ by age, race/ethnicity, or socioeconomic status (SES).
For example, past analyses of NHANES data have shown a significant effect of age. Teenagers and young adults consumed the most fruit juices, SSBs, and water. Adults and older adults consumed much less SSBs but drank more coffee, tea, and alcohol [3
]. Education and incomes also played a role. In past studies, lower-income groups consumed more regular soda; whereas higher-income groups tended to drink more diet soda [4
]. Similarly, the consumption of whole milk was associated with lower SES; higher-income groups consumed more skim and reduced-fat milk [4
A socio-economic gradient was recently observed for the consumption of tap water [3
]. Analyses of NHANES 2011–2016 data showed, for the first time, that most tap water was consumed by groups of higher education and incomes [3
]. This may be the result of powerful new marketing campaigns that hope to change the way that Americans think about water, bottled and tap [6
]. The newly observed social gradient may also be a direct result of the “Flint effect” and the growing distrust of municipal water systems in low-income areas and among communities of color [7
Aligning daily beverage choices with healthy eating patterns is a key component of many dietary intervention programs [1
]. However, such dietary strategies may need to build on existing beverage consumption patterns and the timing of water and beverage consumption during the day. Here, the available data are limited. Only a few studies on children in the UK and in France have examined water and beverage consumption patterns by meal and time of day [11
]. Earlier US based studies have examined sourcing locations but not by meal type or time of day [4
The timing of beverage consumption in the course of the day may have additional implications for adequate hydration. There is an emerging mythology about the correct time to drink water in the course of the day. One strategy is to drink water 30 minutes before a meal, during a meal. and after a meal, but no more [13
]. Another is to drink water early in the morning, soon after waking up [14
]. Additional recommendations are to drink water before, during, and after a workout, before a bath, and just before going to bed at night. Drinking water at the correct time is alleged to help prevent stomach pain, irritable bowel syndrome, fatigue, overeating, high blood pressure, and even heart attack and stroke [15
]. However, evidence in support of those strategies is limited.
One recent suggestion was that mild dehydration may occur in a transient manner when water and fluids are not consumed, either because of poor access to water or beverages or because of poor drinking and eating habits [16
]. The present study explored daily fluctuations in water intakes from water, beverages, and foods in a large and nationally representative sample of children and adults in the US.
2. Materials and Methods
2.1. Dietary Intake Databases
Consumption data for drinking water, beverages, and foods came from 3 cycles of the nationally representative National Health and Nutrition Examination Surveys (NHANES), corresponding to years 2011–2012, 2013–2014, and 2015–2016 [17
]. The three NHANES cycles provided a nationally representative sample of 7453 children (aged 4–18 y) and 15,263 adults (aged ≥19 y).
The NHANES 24-hour recall uses a multi-pass method, conducted by a trained interviewer using a computerized interface. Respondents report the types and amounts of all food and beverages consumed in the preceding 24 hours, from midnight to midnight [18
]. Respondents first identify a quick list of foods and beverages, reporting both meal occasion and time of day. A more detailed cycle then records the amounts consumed, followed by a final probe for any often-forgotten foods. Day 1 interviews are conducted by trained dietary interviewers in a mobile examination center. Day 2 interviews are conducted by telephone some days later [19
For children 4–5 y, dietary recall is completed entirely by a proxy respondent (i.e., a parent or guardian with knowledge of the child’s diet) [19
]. Children 6–11 y are primary respondents, but a proxy respondent is present and able to assist. Children 12–19 y are primary respondents but can be assisted by an adult who has knowledge of their diet [19
]. We used a combination of the 1-day value and the 2-day mean to make use of all available dietary data. This method included all NHANES participants, even those without a second recall.
2.2. Participant Characteristics
NHANES participants were stratified by sex and age. The age group cut-points were: 4–8, 9–13, 14–18, 19–30, 31–50, 51–70, and >70 y. These age groups generally correspond to the age groups used by the IOM. Race/ethnicity was defined as non-Hispanic white, non-Hispanic black, Mexican American, Other Hispanic, and other/mixed race. Family income-to-poverty ratio (IPR) is an index of socioeconomic status; the cut-points for the IPR were <1, 1–1.99, 2–3.49, and ≥3.5.
2.3. Water and Beverage Categories
Plain drinking water was split into tap and bottled. Beverages were classified into 15 categories: milk and milk beverages, milk substitutes (soy milk), citrus juices, non-citrus juices, diet soda, regular soda, ready-to-drink tea, ready-to-drink (RTD) coffee, fruit drinks, sports drinks, energy drinks, hot tea/coffee, alcoholic beverages, flavored, carbonated or enhanced water, and supplemental beverages. The present analyses of water intakes from beverages were for beverages only; for example, milk consumed with cereal (i.e., not as a beverage) was counted in the food category. The USDA Food and Nutrient Database for Dietary Studies (FNDDS), used to establish energy and nutrient content of individual diets, has been revised in parallel to each NHANES cycle [20
The NHANES 24-hour recall for each participant provides information on the amount in grams of each food and beverage consumed. The present results were for mL of water content from selected beverages, and not for the volume of the beverages themselves (which may not be 100% water). Moisture from foods was calculated as well.
2.4. Water and Beverages by Meal Type and Time of Day
Eating occasions were classified as follows: breakfast, morning snack, lunch, afternoon snack, dinner, and evening snacks. That information was obtained by self-reporting. The distribution of water and beverage intakes was also captured by time of day. The 24 h temporal profile of water and beverage consumption was framed in 3 h intervals, starting at 06:00 (6 a.m.) until midnight. The time intervals were 06:00–09:00; 09:00–12:00; 12:00–15:00, 15:00–18:00, 18:00–21:00, 21:00–00:00, and 00:00–06:00.
2.5. Data Availability and Ethical Approval
The necessary IRB approval for NHANES was obtained by the National Center for Health Statistics (NCHS) [21
]. Adult participants provided written informed consent. Parental/guardian written informed consent was obtained for children. Children/adolescents ≥12 y provided additional written consent. All NHANES data are publicly available on the NCHS and USDA websites [17
]. Per University of Washington (UW) policies, public data that do not involve “human subjects” and their use requires neither IRB review nor an exempt determination. Such data may be used without any involvement of the Human Subjects Division or the UW Institutional Review Board.
2.6. Statistical Analyses
The survey-weighted mean intakes of total water were evaluated overall and by age group, sex, race/ethnicity, and family income-to-poverty ratio. All analyses accounted for the complex survey design of NHANES and reflected the dietary behaviors of the US adult population from 2011 to 2016.
The consumption of water and beverages was evaluated for the entire population and for population sub-groups. Survey-weighted means and corresponding standard errors were obtained. All analyses were conducted using SAS software, version 9.4 (SAS Institute Inc., Cary NC, USA) by using SURVEYREG, SURVEYMEANS, and SURVEYFREQ procedures.
The present results are among the first to document the timing of water and beverage intakes around the clock in a large and representative NHANES 2011–2016 sample of US children and adults. The present results have important implications for the promotion of healthy beverage choices, notably the ongoing attempts to replace SSBs with plain drinking water.
There is very little science on population water consumption patterns during the day. One recent study, conducted in Greece, explored the fluctuation in water intakes and hydration indices during the day, looking for signs of transient dehydration in a sample of healthy adults [16
]. While water intakes did go up and down during the day, as they did here, the term fluctuation generally refers to an unpredictable and irregular rising and falling. As the present results show, the timing of water and beverage consumption followed predictable patterns. SSBs were consumed with lunch and dinner and in the afternoon but rarely at breakfast or the mid-morning snack. Water was consumed largely in the morning and rarely at night. Adults drank coffee in the morning and alcohol in the evening [22
]. Beverage choices and consumption patterns varied with age. Whereas children aged 4–18 y consumed water between 18:00 and 21:00, adults were more likely to consume less water and more alcohol in the same time slot. Furthermore, while children consumed milk in the morning, adults tended to drink tea or coffee during this time.
The present data add to past work on the impact of caloric beverages consumed separately or with a meal on total energy intakes. In experimental studies, when caloric beverages were presented shortly before or with a pizza meal, no energy compensation was observed. Caloric beverages consumed freely at meal times added calories to the meal [23
]. By contrast, other studies showed that the presentation of a stand-alone liquid preload reduced energy intakes at the test meal; however, the effects were sometimes inconsistent and full energy compensation was rarely observed [25
Excessive SSB consumption is thought to contribute to childhood obesity [26
]. Dietary Guidelines for Americans have stressed the importance of healthier beverage choices to be made throughout the day [1
]. Ensuring access to safe, free drinking water in schools is an important CDC initiative that is intended to increase water consumption, help maintain hydration and reduce energy intake when substituted for SSBs [27
]. School-based strategies to replace SSBs with plain drinking water have ranged from limiting sales in cafeterias, vending machines, and competitive food outlets to featuring teachers as competitive role models [9
The present analyses support the CDC initiative but for a different reason. The CDC report notes that more than 95% of children are enrolled in schools and typically spend 6 h at school each day. We note that those times are in the morning (typically), which are the peak times for water consumption. There seems to be less competition during the morning snack from other beverages. Another productive strategy would be to promote water consumption with the school lunch meal.
We were surprised to see that water did not figure prominently in the afternoon snack. Though the total amount of water consumed was comparable to that during the morning snack, those beverages were SSBs, tea, and alcohol (for adults). Promoting water consumption by children in the afternoon may be a potential intervention strategy.
By contrast, promoting more water consumption in the morning might be a viable strategy for adults. There are opportunities to increase water consumption at lunch. By contrast, water is unlikely to displace morning coffee, especially with older adults or the afternoon tea [22
The social gradient in water consumption has been addressed before [28
]. Plain drinking water, bottled and tap, accounts for 38% of daily water intake from all sources including food moisture. The intake is slightly lower for lower income and minority populations. Hispanic Americans drink more bottled water and tap water but other minorities do not [3
]. One issue is public trust in the municipal water system—the provision of safe water in schools and community settings is critical to the adoption of healthier beverage choices [27
This study had limitations. First, the NHANES 2011–2016 data are based on dietary self-reports, still the default practice in large population-based studies. Second, the within-day variations by meal and time interval were close but not quite the same. For example, the consumption of water was high between 06:00 and 12:00, but it was associated not with breakfast but with a morning snack. People who reported consuming a morning snack did not necessarily have breakfast. Finally, the NHANES 2011–2016 data are cross sectional, not allowing for causal inferences to be made. The potential impact on water consumption patterns on health outcomes of interest cannot be determined.