Are your Nutrition Facts Reliable?

Retour aux articles

Last April 7, 2015, the online Montreal news media La Presse+ published a report revealing major discrepancies between the nutritional data resulting from laboratory analyses of five food products and the nutrition information declared in the nutrition facts tables of the same five foods. The article led us to believe that Nutrition Facts tables are unreliable as the energy contribution (Calories), as well as the fat, sugar and salt contents of foods sold on the market is sometimes much higher than the values declared in the Nutrition Facts table.

That same morning, before I even had a chance to read the report, I received a call from radio station 104.7 FM, asking me to comment live on the information published in La Presse+. Considering the general interest for the topic and that the short interview did not allow me to address the issue with great depth, I decided to make it the subject of the next issue of Reporter ACC.

How can one explain these discrepancies?

The specialists interviewed in the La Presse + report all indicated that the regulations allow a 20% gap between the real nutritional value and the declared value. However, consumers, and food manufacturers who pay for the nutritional analyses of their products, deserve clearer explanations. Why do the regulations allow a 20 % gap? How can one justify such a margin of error?

Several factors contribute to the variability of nutritional values. For example:

1.  Raw materials

Foods of animal origin

We can presume that a manufacturer of tuna sandwiches always uses canned tuna, of the same brand and the same variety. This helps deliver a product with consistent quality. Nevertheless, variations in the nutritional value of the tuna can be explained by the fact that the tuna can:

  • have been subject to different environmental conditions: it may come from different regions, have been caught at different times of year, have been exposed to different food supplies;
  • be male or female; young or old;
  • have different genetics (just like brothers and sisters of the same family are not all of the same size, some are leaner than others)

All these factors may have an impact on the nutritional composition of the fish and eventually on food products prepared with the fish.

In the case of meat and poultry, important nutritional variations occur within cuts from a single animal. For example, beef cuts from the round (hip) are fatter than rib meat. This is a matter of animal morphology and is absolutely normal. Similarly, chicken wings are fatter than chicken legs, which are fatter than breast meat.

Foods of plant origin

Fruits and vegetables, legumes, grains, nuts and seeds all come from the earth. Like animal products, plant food products may be subjected to different environmental conditions. The nutritional value of foods of plant origin can vary according to:

  • their genetic composition (there are for example hundreds of varieties of lettuce, tomatoes, etc.)
  • the presence of other plants and animals roaming in the vicinity, which can have an impact on the quality of the soil that nourishes the plants;
  • weather conditions (very sunny and dry or very rainy);
  • the growing methods (field, greenhouse, hydropony, monoculture vs polyculture, etc.);
  • the time of harvest (fresh local strawberries in July may not be the same as strawberries imported in December).


2.  Processing of raw materials

When raw foods are refined or otherwise processed, they generally lose some of their nutritional value. White flour, for example, loses a significant part of its nutritional value during the refining process. Even though it is enriched, only a portion of the lost nutrients are added back. Similarly, peeling and trimming fruits and vegetables removes a good part of the fibre and other nutrients.

Canned and frozen fruits and vegetables may also lose some of their nutritional value. Fortunately, commercially canned and frozen produce lose very little of their nutritional value because the processing is done at the time of harvest, at the peak of freshness. Consequently, they are often as nutritious, if not more than their fresh counterparts that have been harvested a few days back.

Dehydrated fruits and vegetables are both:

  • more dense in some nutrients (protein, fat, carbohydrate, fibre) and Calories;
  • poorer in some vitamins: dehydrated fruits can lose as much as 50% of their vitamin A content, and 80% of their vitamin C content.


3.  Processing method

  1. Home-made style or industrial manufacturing

    One can easily imagine that the proportion of ingredients in a lot of hand-made tuna sandwiches prepared on a production line will not always be uniform, even if the same measuring tools are used. For example, the amount of tuna salad measured with an ice cream scoop will not always be identical from one sandwich to another. The scoop can be more or less full, and the amount of tuna salad adhering to the scoop can vary from one sandwich to the other. Also the amount of butter or margarine spread on the bread can vary from one slice to another. This is especially true if artisan-type bread is used, and that slices in the centre of the loaf are bigger than the slice at the ends. Not only will the size of the bread vary, but the larger the slice the more butter or margarine (or whatever spread is used) is likely to be spread on each slice.


    Also, if lettuce and tomato slices are added to the sandwich, these will not always weigh exactly the same. With varying amounts, the nutritional data per serving will necessarily vary as well. Ideally, all ingredients should be weighed for consistency. However, weighing each ingredient separately for each sandwich (and trimming tomatoes to the required weight for each sandwich) would be counterproductive.


    It should also be noted that a margin of error of 20% may appear large. However, the percentage must be put in perspective. For example, 20% of 350 mg of sodium (15% DV) represents 70 mg of sodium or 0.18 g of salt, i.e. about 1/6 teaspoon (a pinch). In terms of sodium, it seems a lot, but in terms of salt, the 20% margin is easily exceeded.


  2. The cooking method (steamed, boiled, roasted, fried, etc.); the temperature, and the cooking and cooling times; as well as the moisture in the air may affect the:
    • Fat content of the product;
    • The density of the product, which can be more or less dense, more or less rich in nutrients;
    • Vitamin content of the product.
  3. Fermentation of food products (such as yogurt, sauerkraut, tofu, etc.) with bacterial cultures or other microorganisms may also alter the nutritional value of a food product.


4.  Nutritional analysis method

The nutritional composition of a product, i.e. its energy value (Calories) and nutrient content (protein, fat, carbohydrates, vitamins and minerals) may be determined either by laboratory analysis or by database analysis.

  • Laboratory analysis (chemical breakdown of the product)

Laboratory analysis of food products is generally recommended for nutrition labelling purposes, as it most accurately reflects the actual product composition. However, it does not reflect seasonal varieties or variations in the processing methods.

In laboratory analyses, each nutrient is analyzed according to specific procedures, following standardized methods. The procedure involves several dozens of steps, each of which carries a margin of error.

  • Database analysis (computerized analysis of the product formula)

A database analysis consists of a series of computations based on the product’s written recipe and the nutrient contribution of each ingredient. The margin of error is greater than with laboratory analysis, because it cumulates the error margins of each nutrients of each ingredient.

To reduce the margins of error as much as possible, it is important to use the nutritional data of the specific ingredients composing the product and not those of a generic food. For example, if a tuna sandwich contains brand X cheddar cheese, with a specific fat, moisture and salt content, it is recommended to use the nutritional data of brand X cheddar cheese rather than the nutritional data of an average cheddar cheese found in the database.

It is important however to make sure you have accurate data from the cheese manufacturer. You cannot use the rounded values from the Nutrition Facts table. The amounts of each nutrient must be provided with the degree of precision (i.e. same number of significant figures) corresponding to the accuracy of the analytical methodology used to produce the nutrition information. Rounded values, as found in Nutrition Facts tables on food labels are not sufficiently precise to be used for nutritional analysis, for labelling purposes.

Manufacturers of ingredients to be used in the manufacturing of other foods are required, by regulation, to provide the nutritional data of the foods they sell with the degree of precision corresponding to the analytical methodology used to produce the data. However, this appears to be a weak link. Food manufacturers, especially small companies with low production volumes who buy their ingredients in wholesale outlets (rather than having the foods delivered at their production plants), sometimes have difficulty obtaining this information, and end up using the rounded nutritional values on the labels to perform their analyses. If the nutritional analysis of the finished product is based on rounded values for several ingredients, the error margin will likely be higher than the tolerated 20%.


5.  The person conducting the nutritional analysis

The nutritional analysis of food products for nutritional labelling purposes requires high methodological standards.

For laboratory analyses, one must master and apply the AOAC (Association of Official Analytical Chemists) standardized techniques. For this purpose, ACC Label generally recommends that laboratory nutritional analyses of food products be carried out by a reputable laboratory fully accredited by the Standards Council of Canada. Be aware however that a laboratory may be accredited for the analysis of some, but not necessarily all the nutrients required in the Nutrition Facts table. It is therefore important to make sure of the competence of the laboratory for each of the required nutrients.

For database analyses, several factors that may affect the analysis results must be considered. It is not only a matter of inputting a product formula in a computer program and printing the results. One must have the knowledge and sufficient expertise to validate the nutritional data submitted by the ingredient manufacturers and the competence to take account of the various factors that can alter the results. Often, it requires the wisdom to acknowledge that a database analysis is simply not suitable to the type of product, as the margins of error would be too large.


6.  Rounding rules

Health Canada established rounding rules for the declaration of each nutrient in the Nutrition Facts table. These rounding rules take into account the degree of precision of the analytical methods, the amount of nutrient in the product, and nutrient-content claims made or that could be made. For example, the rounding rules require that we round to the nearest:

  • 5 calories increment the energy value of a product containing 5 to 50 calories, but to 10 calories increment the energy value of a product containing more than 50 calories.
  • 1 gram increment the fat, carbohydrate and protein contents. However, if the product contains less than 0.5 g of fat or protein, the values must generally be rounded to the nearest 0.1 g increment. As for carbohydrates, the value would have to be declared as zero if the product contains less than 0.5 grams.
  • 1 mg, 5 mg or 10 mg increment the sodium content, depending on the product’s sodium content. However, if the product contains less than 5 mg sodium, and meets the requirement for a sodium-free claim, the sodium content may be declared as zero in the Nutrition Facts table.



In short, the margin of error of 20% granted by Health Canada for the nutrition information declared on the labels of food products accounts for several of the natural variability of food products, the impact of food processing methods on the nutrient content, margins of errors inherent to the analytical methods and rounding rules.

The margin of error may appear large, but considering the many variables, it is actually quite reasonable.

In order to provide reliable nutritional labelling, and to avoid a margin of error greater than 20%, food manufacturers must standardize their recipes and manufacturing processes. Additionally, if they do not have nutrition analysis experts on their team, they should outsource the nutritional analysis to experts, i.e.:

  • nutritionists specialized in nutritional analyses; or
  • laboratories fully accredited by the Standards Council of Canada.

On our blog

Mandatory Front-of-Pack Nutrition Labelling

En savoir plus [+]