We analyzed the Mexican Nutrition Survey 1999 (MNS-99) and the Mexican Health and Nutrition Survey 2006 (MHNS-06). Both the MNS-99 and the MHNS-06 was national, cross-sectional, multi-stage, stratified surveys representative of the country. The MNS-99 was conducted between October 1998 and March 1999. The MHNS-06 was conducted between October 2005 and May 2006. Both surveys had sampling power to disaggregate into urban (population ≥2,500 inhabitants) and rural (population <2,500 inhabitants) areas. The objective of these surveys was to characterize the nutritional status as well as the food and nutrient patterns of the Mexican population. Due to budget restrictions, the MNS-99 only collected information of children and women of reproductive age (12-49y), while the MHNS-06 collected information regarding children, as well as both adult men and women of all ages (≥19y). A detailed description of the sampling procedures and survey methodology has been published elsewhere. 2 31 A total of 1204 and 3552 preschool children (1-4y) were evaluated in 1999 and 2006 respectively. In school age children (5-11y) a total of 2496 and 8716 were evaluated in 1999 and 2006 respectively. Children without complete dietary records were dropped from the analysis presented here.

The MNS-99 and the MHNS-06 samples were drawn to be representative of four regional strata, North, Central, Mexico City and South. The four regional strata, with common geographic and socio economic characteristics, were (1) North: Baja California, Southern Baja California, Coahuila, Durango, Nuevo Leon, Sonora, Sinaloa, Tamaulipas and Zacatecas, (2) Central: Aguascalientes, Colima, Guanajuato, Hidalgo, Jalisco, Mexico, Michoacan, Nayarit, Querétaro, San Luis Potosí and Tlaxcala, (3) Mexico City and (4) South: Campeche, Chiapas, Guerrero, Morelos, Oaxaca, Puebla, Quintana Roo, Tabasco, Veracruz and Yucatan. This regionalization scheme has been used in diverse epidemiologic transition analysis for within country comparisons. 32 33

Informed consent was obtained from each subject or subject's parent or guardian for their participation in the study. The survey protocol was approved by the Ethics Committee of the National Institute of Public Health, Mexico.

Elsewhere there are more detailed descriptions of the dietary data from the 1999 and 2006 surveys. 34 35 The methods are different in the two surveys and this has been addressed in depth in a companion paper. 35 In 1999, a single 24-hour dietary recall was utilized 36 from women and children. In 2006, a previously validated semi quantitative Food Frequency Questionnaire (FFQ) 37 including 101 foods and 14 food groups was used, this questionnaire included the 95% most consumed foods reported in the 24-h dietary recalls collected in the MNS-99. In both cases for children under 12, parents and/or caretakers were involved in handling the surveys with assistance from the children when useful. Standardized personnel applied the FFQ to a nationally-representative sub-sample. Respondents were queried on their dietary intake over the previous 7 days, including the portion size. From these food items a total of 17 beverages were recorded, 13 sweetened and unsweetened beverages (such as tea, coffee, juice, and unique local beverages made with grains, fruit juice and sugar), and four types of milk. This instrument collected information of food intake for the previous seven days in a sub-sample of approximately 16,100 households and all individuals in the household. This sub-sample was representative of the country, country regions, urban and rural areas, and states. Dietitians collected both sets of data and converted food consumption data into grams or milliliters.

Intakes and percentages of energy, carbohydrates, proteins, and fat adequacies reported on the individual level greater than 5 standard deviations from their respective means were excluded from the analysis. Likewise, energy adequacy percentages less than 25% were eliminated, as they were implausible values. Dietary data from 3,959 children in preschool children were obtained of which 407 observations were excluded (10.3%). In school children dietary data were collected from 9,383 and 667 observations (7.1%) were excluded from the analysis. Due to the data elimination during the cleaning process, we calculated new sampling weights.

A socio-economic status (SES) index has been developed in Mexico using a combination of household conditions (i.e., flooring material, ceiling, walls, and number of persons residing in the household), basic services infrastructure (i.e., water source and disposal), and possession of domestic appliances (i.e., radio, television and refrigerator). 38 39 40 This approach was used to create a comparable SES index for the 1999 and 2006 surveys.

Our grouping of beverages that follows was utilized by other scholars. 41 This grouping placed beverages into categories based on caloric intake and overall potential health benefits. Thus we created the following categories: a) High energy (soft drinks, sweetened tea and coffee, sweetened juice and fruit drinks, and atole--a typical Mexican beverage made with sugar, corn meal usually, or other substitutes such as rice flour and corn starch) b) High energy some benefit (whole milk, milk flavored, fruit juice), c) Low energy (sweetened diet beverages, unsweetened tea and coffee) and d) water. At the time of the MHNS-06, only trivial amounts of reduced fat milk was consumed (in the MNS-99 basically no reduced fat milk was available on the market).

Energy intake was estimated using a comprehensive nutrient composition database compiled from diverse references. Major outliers were reviewed case by case and corrected when possible. 42 43 44 45 46 Using the MNS-99 and MHNS-06 we generated measures of percent calories from beverages, proportion of beverages that include some type of caloric sweetener by age, gender, urban or rural locations, geographic areas and socio-economic status index tertile. Consumption patterns were examined using a variety of characterizations, including a) the dichotomous characterization as consumed beverage for the proportion who consume each type of beverage; b) per capita consumption of each beverage for the age-gender grouping [in milliliters (ml) and kilocalories (kcal)]; c) the amount of beverages consumed by each consumer of these beverages (in ml and kcal); and the percent contribution to total energy intake (among consumers and per-capita). To describe consumption (ml and kcal), arithmetic means were calculated. Differences across energy intake by socio-demographic factors were evaluated using ANOVA with Tukey's adjustment for multiple comparisons (termed in STATA linear combination estimators). Significant differences between consumption trends between 1999 and 2006 were evaluated using a T-test for independent samples; a p-value < 0.05 was considered significant. All calculations were weighted by expansion factors and adjusted for the complex sampling survey design using the STATA 9.0 SVY module for complex surveys. 47

Consumption of caloric beverages was analyzed by beverage group for pre-school and school children. At the national level, the per capita calories from beverages in pre-school children increased from 161 kcal in 1999 to 310 kcal in 2006 (p < 0.05). Similarly, in school-age children, consumption increased from 185 kcal to 323 kcal per capita from 1999 to 2006 (p < 0.05) (Figure 1).

Pre-school children consumed approximately 794 ml of beverages per day, including plain water in 2006. These beverages contributed to 27.8% of the total reported energy intake. From all beverage types in this age group, the "high sugar some benefit" group, which included mainly whole milk, provided the largest number and proportion of calories (21% of the total estimated daily energy intake) (Table 1).

School age children consumed 1254 ml per capita of beverages in 2006. These beverages contributed to 20.7% of the total energy intake. In contrast with pre-school children, consumption of high-sugar beverages was more important (10.3% of the total energy intake as opposed to only 6.6% in the younger children) (p = 0.01). The proportional role of sweetened beverages, such as fruit water, sugar-added orange juice, sweetened and flavored water, increased also (Table 2).

School age children's milk consumption was slightly lower (144 kcal) than pre-school children's consumption (220 kcal). This is equivalent to a 34.5% lower intake (p < 0.01). A total of 607 ml of plain water was consumed per capita in school age children. In pre-school children only 243 ml was consumed.

Thus, the main change in beverage consumption for pre-school children occurred in whole milk and juice water. In school children the change in consumption of whole milk is important, but flavored waters and soft drinks changed dramatically.

Tables 3 and 4 describe consumption of energy beverages and their contribution to the diet. In pre-school children across all the socio-demographic variables whole milk was the most important caloric beverage. It contributes at least 200 kcal per capita daily with the exception of rural areas and low socio-economic status children. In contrast, children from urban areas consume 236 kcal. Children from Mexico City consume 262 kcal and those in high socio-economic status consume 259 kcal. Next in importance are the sugar-sweetened juices which contribute 29 kcals in the low socio-economic status children and 51 kcal in those from the high SES (p < 0.05). In the population from the low SES tertile, calories from coffee and tea with sugar, and atole were higher than in the highest SES tertile (p < 0.05), in rural than urban areas (p < 0.05) and south region than in the other country regions (p < 0.05). Males had a significantly higher intake compared to females from juices and sweetened beverages (p < 0.05).

In school children the contribution of whole milk to total beverage caloric intake is lower than in pre-school children across all socio-demographic factors, and important differences were observed; urban children had a 58% higher consumption than rural children, and higher SES tertile children had a 57% higher consumption than those in the lower SES tertile. The contribution of soft drinks was higher in this age group compared to the pre-school children, and overall caloric consumption from beverages was higher with the exception of children from the south country region.

Children residing in the north region consumed almost double the energy from soft drinks.

Using results from this study and a comparable one from adults, the Minister of Health organized a beverage guidance panel to review and organize a set of actions for the government to reduce energy from beverages as one component of its obesity prevention campaign for all age groups. 59 A comparable paper on adult beverage trends has been published. 35 The Minister has worked with the government to institute a number of actions already.

One action already taken has been to replace all whole milk in the feeding programs of approximately twenty million Mexicans with 1.5% milk. The government social support programs--Oportunidades, Liconsa and PAL-- distribute foods and/or provide cash transfers. Initially for a 2-month trial the program gave reduced fat milk only to senior citizens. This worked well and now all individuals receiving milk or using vouchers obtain 1.5% milk.

A study is underway in Mexico City to work out ways to provide potable water in a series of schools and use this pilot to launch this first in several states, and then nationally. This is being led by the National Institute of Public Health. Ultimately, the Secretary of Education will allow only skim milk and water in the schools. School breakfast programs are removing flavored milk and using reduced fat milk only.

First and foremost, the government is considering a range of regulations and taxation plans related to reducing added sugar in beverages. 58 In addition, these authors are involved with others in Mexico in creating a Front of the Package Labeling system that will label only beverages with limited saturated fat and added sugars with a healthy beverage label. This is an ongoing activity sponsored jointly by the Finance Ministry, the Food Standards Agency and the Ministry of Health that will be finalized late in 2009 and implemented sometime in 2010.