Development and validation of a food-frequency questionnaire to assess short-term antioxidant intake in athletes

Journal Title (Medline/Pubmed accepted abbreviation): Int. J. Sport Nutr. Exerc. Metab.
Year: 2011
Volume: 21
Issue: 2
Page numbers: 105-112
doi (if applicable):

Summary of Background and Research Design

Background: Assessment of an athlete’s antioxidant intake is important for researchers as well as sports nutritionists. There are a few common methods used to determine antioxidant intake:

Food frequency questionnaire (FFQ): asks how often one consumes a particular food (ex. once a week, once a month) and how much one consumes as an average serving. The food choices in the questionnaire are often specific to the study. For example, in this case, specific antioxidant sources like berries, tea, and supplements were individually inquired about. This method is cheap and easy, but it might not capture all the significant antioxidant sources in the person’s diet.

7-day diary: a participant records their total food intake for 7 days. It is more time-consuming for both the subject (time needed to record food intake) and the analyst (time required to determine antioxidant values for food consumed). In addition, it is often difficult to find antioxidant values for specific foods in nutrient composition databases. However, this method does provide more specific information and much more information on overall food intake than does a food frequency questionnaire.

Total antioxidant capacity (TAC): the assessment of how well a sample (either a food or a participant’s blood sample) can reduce (the opposite of oxidize) a chemical sample such as iron. Foods with reported TACs can help analysts assess antioxidant intake in either the FFQ or the 7-day diary. The TAC of a subject’s blood is more objective and accurate than the diaries and questionnaires when assessing antioxidant status. However, measurements of TAC are in vitro measurements and may not reflect actual in vivo conditions within the body.

Research goal: To develop an FFQ assessing antioxidant intake in young, active individuals; correlate its precision with a 7-day food diary; and assess its accuracy by comparing it to the antioxidant capacity of the participants’ blood.

Subjects: Participants were healthy rowers, age 17-36 y. There were 113 participants total- 56 male and 57 female. Not all volunteers participated in all parts of the study (96 completed the questionnaire and blood test, 81 completed the 7-day food record and questionnaire, and 63 completed the 7-day food record and blood test).

Protocol: Subjects first completed the FFQ and provided a blood sample for assessment of antioxidant content of the blood. Then, they completed a 7-day food diary. Of those that completed the questionnaire, 20 completed the FFQ again 1 week later to assess reliability.
                  The FFQ included foods that contained greater than 0.1 mmol ferric-reducing ability of plasma (FRAP) per serving (measurement of antioxidant capacity) and commonly consumed foods with antioxidants. Some foods were grouped together if they were found to have similar antioxidant contents. There were 45 questions about 75 foods and dietary supplements. The FFQ included questions regarding physical activity and competition.

Summary of research findings
  • The majority of the antioxidant intakes came from cereal (31% according to the food diary and 19% from the FFQ), fruits and berries (9.8% from the diary and 21% from the FFQ), and vegetables (9.6% from the diary and 21% from the FFQ).
  • The mean bias between the questionnaire and the food record (criterion method), as determined by linear regression modeling, was 4.8%. However, the bias associated with the questionnaire increased greatly for those with very high (120 mmol/day; 73% underestimation by questionnaire) or very low (20 mmol/day; 42% overestimation by questionnaire) antioxidant intakes.
  • There was a high correlation (r = 0.5-0.7) between the diary and the FFQ for cereals, alcoholic drinks (wine and beer), and vitamin C supplements. There was a moderate correlation (r = 0.3-0.5) for coffee and tea and a low correlation (r = 0.1-0.3) for fruit, fruit juices, and chocolate.
  • There was a correlation of r = 0.25 between the plasma sample antioxidant scores and the FFQ. This correlation was described by the authors as “small but clear”. However, there was an “insubstantial” correlation between the plasma score and the food diary (r = 0.15). The results were similar when the numbers were normalized to energy intake.
  • When the FFQ was retested after 1 week, the correlations were at least moderate for all food groups except chocolate, indicating reliability of the questionnaire.

Interpretation of findings/Key practice applications

The FFQ that was developed to assess antioxidant status in young, active individuals has been validated and may be useful for researchers as well as coaches and sports nutritionists. This research highlighted that, although cereals have lower antioxidant contents than fruits and vegetables, they can be consumed in high enough quantities to serve as the primary source of antioxidants. This observation applies to athletes in particular because of their higher utilization of carbohydrates compared with non-exercising individuals.


Antioxidants are not all equal in regard to uptake from the gut, bioavailability to the body cells, potency, etc. Moreover, there are physiological differences among people that further complicate these measures. Therefore, it is difficult to assess how much of antioxidant intake actually can be utilized by the body. Plasma antioxidant scores have been met with some criticism; however, at this point, it is unlikely that any noninvasive technique will surpass the accuracy of plasma methods.
Google Tracking Google Plus Tracking Twitter Tracking