D1. An action oriented food monitoring programme

The aims of the project were twofold. In a first part the concentration of different constituents of the food and foodstuffs consumed in Belgium were determined: on the one hand inorganic constituents i.e. 14 minerals and trace elements, on the other hand organic constituents namely total fat and the fatty acid profile. In order to ensure the reliability of the results obtained a thorough quality assurance plan was developed. In a second part mathematical and statistical methods were applied in order to draw relevant conclusions from the large amount of data produced. Moreover these mathematical techniques allow to relate existing data concerning cancer incidences in Belgium and Europe to the foods consumed, the constituents determined in these foods as well as their estimated daily intake.

The different parameters of the quality assurance plan were thoroughly investigated. A validation strategy was developed which was based on well defined method performance characteristics. For the validation of the different atomic absorption spectrometric methods a user-friendly-knowledge-based system was developed. The different analytical methods used were thoroughly validated before the analyses were carried out.

To estimate the daily dietary intake of the specific constituents of food, duplicates of total (24 h) diets were analyzed. This approach was preferred over an estimation based on food consumption surveys because for the latter recent data are not available in Belgium. The 24 h diets were collected during seven consecutive days at five different locations in Belgium. For the study of the total fat and the fatty acid profile a separate sampling plan was elaborated which was based on the BIRNH-study. A total of 150 individual foodstuffs, which are responsible for at least 90% of the daily fat intake, were analyzed.

Since minerals and trace elements are present in almost all foodstuffs the latter were grouped into 16 categories. From these categories about 300 foodstuffs were selected for the analysis of minerals and trace elements.

The analysis results for the minerals and trace elements are discussed, for the duplicate diets and for the individual foodstuffs. When the daily intake is compared with international and Belgian norms, no toxicologic problems are detected for the five non-essential elements (cadmium, lead, chromium, arsenic and aluminium). For the essential minerals and trace elements there are no problems for the intake of copper, manganese and not even for potassium and sodium. The low calcium intake and the low iron intake by women are, on the other hand, important problems. The intake of magnesium, zinc and selenium are marginal. A comparison between the intake of inorganic components on the four different locations where the duplicate menus were collected shows important differences, due to different dietary habits and food preparation methods. For the duplicate menus the measured mineral concentrations are compared with those estimated by the food composition tables. This shows that those tables must be handled with care.

The fat amount and the fatty acid profile were determined in the double menus but also in 150 types of foodstuffs, which are responsible for 90% of the fat intake (BIRNH-study). The results indicate that the total fat consumption in Belgium is very high, namely 42% of the total energy intake. The main categories of foodstuffs that are responsible for this are "fats, oils and sauces" (50%), "meat and poultry" (22%) and "cheese and eggs" (9%). Our study shows that the saturated, the mono- and the polyunsaturated fatty acids contribute for 17%, 14% and 8% to the total energy intake, respectively. When the intake of the fatty acids is compared between different provinces, one does not only observe differences between the French speaking and the Dutch speaking part, but also between the individual provinces.

The mean intake of fatty acids at the different locations where the duplicate menus are collected are of the same order as that computed by the BIRNH study. A comparison with international norms shows that the relative fat intake in Belgium is much higher than the recommended 15-30%.

This is also true for saturated fat, namely 17% in stead of 10%. The mean intake of polyunsaturated fatty acids is also situated above the recommended 3-7%. A comparison with recent EU-recommen-dations shows that no deficiencies exist for the n-6 fatty acids: 90% of the population has an intake higher than the PRI-value (Population Reference Intake). The n-3 fatty acid intake, on the other hand, is quite low. About 37% of the population has an intake lower than the PRI.

Two literature studies were performed to investigate the relation between the intake of the different components and the cancer incidence. It can be concluded that the high fat intake is clearly correlated with breast, colon and prostate cancer. High amounts of n-6 fatty acids are positively correlated with different cancer types, while n-3 fatty acids have a more protective effect. The literature study on minerals and trace elements indicated that increased intake of calcium, selenium and zinc would decrease the risk for different types of cancer, while an increased sodium intake could indirectly increase the risk of stomach cancer. When data from the literature are compared with the results that are obtained during this study, it must be concluded that in Belgium the intake of fat and n-6 fatty acids is clearly high while the intake of n-3 fatty acids is low. For the minerals and trace elements, one can state that the intake of calcium, zinc and selenium is low.

Multivariate statistical techniques were used to correlate specific cancer types with nutritional patterns. These techniques were first applied to European cancer and nutritional data. For each type of cancer the correlation with the intake of specific food groups is given. It is not possible to go here into all results, only examples can be given. It is found for instance that Belgium has one of the highest cancer rates in Europe for certain cancer types (colon, breast,...). Moreover, the Belgian diet is characterised by a high intake of animal fats and a lower intake of vegetables and cereals. The analysis of the Belgian data shows that different cancer patterns occur in different parts of the country and that these differences are not restricted to the north/south axis. For instance the highest cancer mortalities for colon and rectum cancer are found in East- and West-Flanders, where meat intake is high and vegetable intake is low, while oesophagus/larynx cancer are most frequent in Hainaut, where the consumption of alcoholic beverages is high.

The correlation of our analysis results with cancer incidences in Belgium are made difficult because the consumption data date from the years '80-'85 (BIRNH-study), while the most recent mortality data date from 1989. The lag-time between the influence of nutrition and the development of cancer is statistically estimated at about 10 years.

The recommendations derived from this research project are directed towards all persons having responsibility for the formulation of policies concerning food and health. In addition to the food components studied here, there are many others that should be studied in relation with cancer. Continuously updated data for morbidity and mortality are needed. At a technical level only validated techniques should be used for completing food composition tables. The use of multivariate mathematical techniques should be stimulated. Food consumption data for the Belgian population should be obtained on a regular basis. Sensibilisation campaigns are required to improve nutritional habits to prevent disease. Food producing instances should take into account increasing knowledge about the relationship food-diseases.