Identication of foods to monitor the sodium content of processed foods for developing a sodium reduction program in the Philippines

In response to the global target for reduction in salt intake, several countries have implemented population sodium reduction strategies. These strategies include identication of major sources of sodium in the diet and reformulation of a set number of products available on the market. This study aimed to identify processed foods that can be targeted for reformulation and whose sodium content can be monitored over time in order to reduce sodium intake in the Philippines. The objectives were to: 1) Estimate per capita sodium intake from minimally processed and processed food groups by income quintile and urban/rural location; 2) Identify foods that contribute to the variance in per capita sodium intake that can be used as indicators for monitoring the sodium content of Philippine processed foods. One-day household food weighing data covering 4880 households from the 2008 National Nutrition Survey was used. Per capita sodium consumption from processed and minimally processed food categories and percentiles of sodium intake from these categories by income quintile and urban/rural location were obtained. The percentage contribution of different food categories to mean per capita sodium intake was calculated. Specic foods that contributed to the variance in sodium intake among Filipinos were identied. in 2008 [4] to 28% in 2013 [5], highlighting the need to reduce sodium intake. The present study identied processed foods that can be targeted for reformulation to achieve reduced salt intake. Important sources of sodium were 13 foods in the processed food group and 2 foods in the minimally processed group, which together accounted for 99.4% of the variance in sodium intake of the entire population. In the processed foods group, the greatest contributors were the following: instant noodles and foods in the following categories: Processed Soup, Sauces & Flavor Enhancers (traditional fermented sh & seafood sauces, table salt); Processed Fish, Meat & Poultry Products (dried/smoked sh & seafood, canned & processed meat/sh/seafood, salted eggs); Baked Products (white bread, pan de sal); Other Noodles & Pasta (wheat and egg noodles); Rice, Cereal & Starch Products (crispy cereal chips and extruded snacks); Fats, Oils & Products (butter, margarine); Milk Products (cheese); Non-alcoholic Beverages (chocolate-based drinks).


Abstract Background
In response to the global target for reduction in salt intake, several countries have implemented population sodium reduction strategies. These strategies include identi cation of major sources of sodium in the diet and reformulation of a set number of products available on the market. This study aimed to identify processed foods that can be targeted for reformulation and whose sodium content can be monitored over time in order to reduce sodium intake in the Philippines. The objectives were to: 1) Estimate per capita sodium intake from minimally processed and processed food groups by income quintile and urban/rural location; 2) Identify foods that contribute to the variance in per capita sodium intake that can be used as indicators for monitoring the sodium content of Philippine processed foods. One-day household food weighing data covering 4880 households from the 2008 National Nutrition Survey was used. Per capita sodium consumption from processed and minimally processed food categories and percentiles of sodium intake from these categories by income quintile and urban/rural location were obtained. The percentage contribution of different food categories to mean per capita sodium intake was calculated. Speci c foods that contributed to the variance in sodium intake among Filipinos were identi ed.

Results
Foods which signi cantly accounted for 99.4% of the variance in sodium intake were 13 types of processed foods and 2 types of minimally processed foods. The category Processed Soup, Sauces, and Flavor Enhancers contributed the greatest proportion to per capita sodium intake. Speci c processed foods that contributed to the variance in per capita sodium intake were instant noodles, traditional fermented condiments and sauces, dried and processed meat, sh, and poultry products, salted eggs, white bread and pan de sal (a traditional Filipino bread), wheat and egg noodles, crispy cereal chips and extruded snacks, butter and margarine, cheese, and chocolate-based beverages.

Conclusion
Identifying processed foods that signi cantly contribute to sodium intake, followed by reformulating and monitoring the sodium content of these foods over time should be considered as one strategy to reduce sodium intake in the Philippines.

Background
Hypertension is a risk factor for cardiovascular disease driven by excess dietary salt intake. The WHO Global Action Plan for the Prevention and Control of Noncommunicable Diseases 2013-2020 [1] set voluntary targets for achievement in 2025 by its Member States, including "a 30% relative reduction in mean population intake of salt/sodium (Na)" towards the recommended level of 2000 mg Na/day (5 g salt/day). This can be achieved by developing "guidelines, recommendations or policy measures that engage different relevant sectors, such as food producers and processors, and other relevant commercial operators, as well as consumers, to reduce the level of salt/sodium added to food (prepared or processed) [1]." In response to the global target for reduction in salt intake, several countries have implemented population sodium reduction strategies. These strategies include identi cation of major sources of sodium in the diet and reformulation of a set number of products available on the market [2]. In the United States, a sodium monitoring program led by USDA tracks "sentinel foods" i.e., foods that contribute to sodium intake in the population and are used as indicators to track changes in the sodium content of processed foods [3]. Since most sodium in the diet comes from processed foods, reducing the amount of sodium in sentinel foods will translate into reduced sodium intake at the population level. The present study aimed to identify processed foods that can be targeted for reformulation to reduce sodium intake among Filipinos, using one-day household food weighing data from the 2008 National Nutrition Survey. The objectives were to: 1) Estimate per capita sodium intake from minimally processed and processed food groups by income quintile and urban/rural location; 2) Identify foods that signi cantly contribute to the variance in per capita sodium intake among Filipinos which can potentially serve as indicator foods to monitor the sodium content of processed foods.

Results
a. Per capita sodium intake from different food categories Table 1 shows the mean per capita sodium intake and percentile distribution of sodium intake from minimally processed and processed food groups. Mean per capita intake exceeded the WHO recommendation of 2000 mg sodium, with rural households ingesting more sodium than urban households. Median sodium intake was highest for Processed Soup, Sauces & Flavor Enhancers, with half of the population consuming > 1416 mg Na from this food category alone. Median intake was highest among the highest income households. Table 1 Mean per capita intake by urban/rural location and percentile distribution of sodium (mg/day) ingested from minimally processed and processed food groups by income quintile and urban/rural location  -Q1  50  76  101  240  51  53  87  114  244  61  53  86  112  244  60   -Q2  60  76  104  192  44  62  85  115  244  53  62  81  110  207  48   -Q3  60  78  101  194  41  62  82  110  190  47  61  80  106  193  45   -Q4  58  74  95  187  37  61  81  113  319  52  58  76  100  219  42   -Q5  51  68  88  170  38  51  78  100  279  49  51  69  92  205     Per capita Na intake (mg/day) Percentage contribution of processed and minimally processed foods to per capita sodium intake Minimally processed foods. Among the highest income quintiles, minimally processed food categories that contributed the most sodium were Fish, Meat and Poultry followed by Rice, Cereals and Starches. In lower income quintiles, (Q1 in urban and rural areas, and Q2 in rural areas), Rice, Cereals and Starches contributed the most sodium followed by Fish, Meat and Poultry. In both urban and rural locations, Vegetables and Fruits contributed minimal amounts (< 1%) of sodium. Foods that contributed the least amount to sodium intake were Milk followed by Beans, Nuts and Seeds.  Table 3 shows the results of multiple regression analysis. A total of 15 foods (13 foods belonging to the processed food group, and 2 foods belonging to the minimally processed group) explained the variance in per capita sodium intake. Minimally processed foods that contributed signi cantly to sodium were cooked white rice and ready-to-eat prepared foods ( sh, meat, poultry, organ meats). All other foods with signi cant contributions belonged to the processed food group. Among all minimally processed and processed foods, cooked white rice contributed the greatest amount of sodium (i.e., consumption of one gram of rice increased per capita sodium intake by 0.79 mg) followed by instant noodles (i.e., consumption of one gram instant noodles increased per capita sodium intake by 0.02 mg). This was followed by traditional condiments (fermented sh/seafood sauce), table salt, and processed meat, sh, poultry products.

Discussion
The prevalence of hypertension among adult Filipinos aged 20 years and above increased from 16% in 2003 to 21% in 2008 [4] to 28% in 2013 [5], highlighting the need to reduce sodium intake. The present study identi ed processed foods that can be targeted for reformulation to achieve reduced salt intake. Important sources of sodium were 13 foods in the processed food group and 2 foods in the minimally processed group, which together accounted for 99.4% of the variance in sodium intake of the entire population. In the processed foods group, the greatest contributors were the following: instant noodles and foods in the following categories: Processed soup, sauces and avor enhancers. Within this category, table salt and traditional fermented sh/seafood sauces were the signi cant contributors to sodium intake. In 2008, coarse salt was the most commonly consumed condiment in the Philippines, with 64.9% of households consuming an average of 3 grams salt per day [10], equivalent to 1200 mg Na. Philippine shrimp paste contains 13-14 g Na/100 g [11]. The percentage of households consuming these traditional fermented foods in 2008 [12] was: bagoong isda (fermented anchovy), ginamos (fermented shrimp) -10.1%; patis ( sh sauce) -6.1%; bagoong alamang (shrimp paste) -4.7%. In a study among 1789 women, Lee found that salty condiments added during cooking or at the table accounted for 76.3% of sodium intake [13]. The most signi cant source of sodium was table salt, contributing 53.3% for women who consumed < 4600 mg/day of sodium and 66.5% for women who consumed higher amounts of sodium.

Pros and cons of indigenous fermented sauces
Traditional fermented salted products, while contributing signi cantly to sodium intake of Filipinos, are an important part of the food culture in the Philippines. Commonly used indigenous sauces are fermented sh and seafood sauces (patis or sh sauce, bagoong or sh/shrimp paste), soy sauce. These products are generally produced with high levels of salt, up to 25% for sh sauces and 11 to 25% for soy sauce [14] [15]. High levels of salt and low pH are important to suppress the growth of pathogenic microorganisms and enable bacterial degradation of proteins, carbohydrates, and nucleic acids. In spite of their high sodium content, these fermented sauces were shown to have functional effects. Japanese style fermented soy sauce (shoyu) showed antiallergic, antimicrobial, antihypertensive, and anticarcinogenic effects [16] [17]. Fermented foods contain live microorganisms and therefore comprise a good source of probiotics. Lactic acid bacteria were found in fermented sh (ranging from 3.48 to 5.43 log cfu/g) while aerobic bacteria were found in sh sauce (ranging from 4.92 to 5.53 log cfu/g [18]).
Fermentation-derived microorganisms have the potential to in uence gut microbiota diversity, structure, and function and increase the amount of nutrients such as vitamins and other bioactive molecules produced from microbial metabolism that are not present in the original food [18]. These bacteria may also secrete anti-microbial agents, degrade anti-nutritive compounds, produce short-chain fatty acids from indigestible carbohydrates, and contribute to immune homeostasis [18] [19] [20]. A study on the composition of shrimp pastes produced in some parts of the Philippines showed these foods were good sources of omega-3 fatty acids, iron, zinc, and calcium [11]. Due to their extensive use, forti cation of condiments and seasonings is also seen as a cost-effective intervention to address micronutrient de ciencies in Southeast Asia [21] [22]. Studies in young children and adult women suggested that forti cation of sauces ( sh sauce, soy sauce) can effectively address iron and iodine de ciencies [23] [24].
Processed sh, meat, poultry products. Foods that contributed signi cantly to sodium intake were dried/smoked sh and seafood, canned/processed meat, sh & seafood, and salted eggs. In 2008, consumption of sh and sh products was 110 grams per capita. Canned sardines (containing approximately 521 mg Na/100g) [8] was consumed by 15.3% of households with mean consumption of 8 grams per capita per day [10]. Dried and smoked sh was consumed by 20.5% of households [12]. Dried sh contains ≈ 7000 mg Na per 100 g [25]. Filipinos aged 60 + years ate the most sh and sh products (15.6% of total food consumption), followed by those aged 20 to 59 years (14.7% of total consumption) [12].
Consumption of meat and meat products in 2008 was 83 grams per capita. The Family Income and Expenditure Survey (FIES) showed that household food expenditures on meats increased by 4 to 5 percentage points from 1965 to 2000. The biggest growth in expenditure was for processed meats, increasing by 2.7% during the same period [26]. In 2003, processed meat products (hotdogs, meatloaf, sausages) represented nearly 30% of per capita meat intake [27].

Limitations of the study
The study examined only 2008 national food consumption data. Data from multiple successive surveys should be examined since the market for processed foods is dynamic, with products constantly being introduced, reformulated, or taken out. In spite of this, the present study is the currently the only one that identi es sodium-contributing foods for development of population sodium reduction initiatives. The consumption of processed foods among Filipinos has increased over time. For instance, the demand for instant noodles in the Philippines increased from 3400 million servings in 2016 to 4470 million servings in 2020 [28]. For processed meat, the average volume per person is expected to amount to 3.9 kg in 2021 and the market is expected to grow annually by 1.89% from 2021 to 2025 [29]. During this pandemic, sodium intake is expected to increase further. Food relief packs distributed nationwide by the Department of Social Welfare and Development contain rice, corned beef, sardines, and chocolate energy drink or coffee [30]. Corned beef, sardines, and chocolate beverage are among the foods identi ed in this study which signi cantly contribute to the variance in sodium intake of Filipinos.

Conclusions
Indicator foods that can be targeted for reformulation to reduce population sodium intake among Filipinos are instant noodles, traditional fermented condiments and sauces, and processed meat, sh, and poultry products. Other processed foods with signi cant contributions to the variance in sodium intake and whose consumption can be reduced via consumer education and reformulation (e.g., "stealth" reductions) are table salt, alcoholic beverages, white bread and pan de sal (a traditional bread), crispy cereal chips and extruded snacks, butter and margarine, cheese, and chocolate based beverages.

Methods
Aim, design and setting. The study identi ed processed foods that can be targeted for reformulation and whose sodium content can be monitored over time to achieve reduced sodium intake in the Philippines. This cross-sectional study examined per capita food consumption obtained from one-day household food weighing data of 4880 households participating in the 2008 National Nutrition Survey.
Sampling method. The 2008 National Nutrition Survey used a strati ed multi-stage sampling design. In the rst stage, primary sampling units were selected from 17 regions and 79 provinces throughout the country. In the second stage, enumeration areas were identi ed from primary sampling units. Finally speci c households from each enumeration area were selected, comprising a total of 4880 households (median 5 members per household) nationwide.
Characteristics of the sample. The sample comprised 43% urban and 57% rural households with more households belonging to lower income groups, re ecting the country's socioeconomic classi cation as a low middle income country. Table 4 shows the distribution of households by location of residence and wealth quintile. Data collection. One-day household weighing of food items from breakfast through supper, including snacks was conducted. Digital dietetic scales were calibrated using a one-kg standard weight. On the day of weighing, all items were weighed before cooking or serving including: raw as purchased foods to be cooked for each meal and snacks, food served and eaten raw, cooked and processed foodstuff served directly on the dining table. Leftover foods were weighed and, together with the weights of plate wastes and foods given out, were deducted from the sum of weighed food to obtain the actual amount of food consumed by the household [31]. A food inventory was also conducted. Nonperishable food items that might be used anytime of the day such as coffee, sugar, salt, cooking oil, and other condiments were weighed at the beginning and end of the food weighing day. Foods eaten by household members who ate outside their homes were recalled and recorded to complete the household's food record. Sample weighing of similar food items eaten out was performed for validation purposes [31].
Data analysis. Prior to statistical analysis, the following steps were taken: 1) Creation of a food composition database for sodium; 2) Grouping of all foods consumed into 2 categories: Minimally Processed Foods and Processed Foods/Food Products.
Development of a food composition database for sodium. The Philippine food composition table does not provide nutrient values for sodium. Hence, the sodium content of all foods consumed was estimated from values derived from different food composition tables, using the process described by INFOODS. The INFOODS guidelines for food matching [32] guided the selection of appropriate foods from which to borrow sodium values, in the most appropriate source of compositional data. Values for sodium consumption were then computed by multiplying each food's sodium content by the amount consumed by the entire household.
Grouping of foods into minimally processed and processed food categories. Almost all foods consumed in the Filipino diet are processed or cooked to a certain extent prior to ingestion. FAO [33] recommended that the level of food processing should be taken into account when examining food consumption data, so as to inform the development and implementation of food-based guidelines and approaches to the prevention of chronic diseases. The NOVA food classi cation system [33] [34] developed by researchers in Brazil, classi es foods according to the nature, degree, and purpose of processing. The present study used a modi ed version of the NOVA classi cation wherein foods were classi ed into two main groups and each group was further classi ed into subgroups: 1) Minimally Processed Foods (subgroups comprised cooked/prepared whole foods, e.g., boiled rice and tubers, whole sh/meat/chicken dishes, milk (fresh liquid and whole milk powder), raw or cooked whole vegetables and fruits), and 2) Processed Foods (subgroups comprised processed and preserved/salted food products, foods made from processed ingredients).
All foods consumed by survey households were listed. Similar foods were grouped into speci c subgroups (a total of 18 subgroups or categories were created for 1306 individual food items). Each food category was classi ed as belonging to either the minimally processed or processed groups (shown in Table 5). This classi cation was done to allow the development of recommendations for sodium reduction that correspond to dietary patterns of the entire population. Per capita consumption of sodium from different food subcategories was obtained by summing the total amount of sodium (in milligrams) ingested by the entire household divided by the number of consumption units. Percentiles of sodium intake (P25, P50, P75, P99) from different food subcategories and interquartile range (IQR) were obtained using STATA. The percentage contribution of different food categories to mean per capita intake was calculated using the ratio of means wherein mean sodium intake from a speci c category was divided by mean per capita sodium intake.
Multiple regression analysis was used to identify speci c foods that contributed to the variance in sodium intake for the entire population.
Sodium intake values from speci c foods in the different categories shown in Table 5 were transformed logarithmically. Thus the form of the regression model tted is different foods across food groups, and ε is the error term that represents the variation not due to food consumption, including measurement errors. The signi cant variables were obtained by backward elimination. Variables in the equation were retained at 5% level of signi cance. To account for heteroskedasticity, the linearized robust standard errors were produced. Outliers and in uential observations were excluded from the analysis.

Declarations
Ethics approval and consent to participate