Pourquoi devons-nous évaluer l’alimentation des sujets à risque vasculaire ?

Validation and reproducibility of a short food frequency questionnaire for cardiovascular prevention

BACKGROUND

Diet strongly influences cardiovascular risk. Dietary evaluation is a major issue in cardiovascular prevention, but few simple tools are available. Our team previously validated a short food frequency questionnaire; a new version of this questionnaire (Cardiovascular Dietary Questionnaire 2 [CDQ-2]) is easier to complete and more reliable.

AIMS

To validate CDQ-2 in comparison with the original version, and to test its reproducibility.

METHODS

CDQ-2 has 17 closed-ended questions; it provides a global dietary score that is a combination of specific scores for saturated, monounsaturated and omega-3 fatty acids, and fruit and vegetables. CDQ-2 was validated against the original version in two groups, who completed both questionnaires: 99 patients with cardiovascular risk factors and 50 healthy subjects. Reproducibility was assessed with 27 health professionals who completed the questionnaire twice, with a 1-month interval.

RESULTS

The correlation coefficients of the scores between the two questionnaires ranged from 0.65 (monounsaturated fatty acids) to 0.93 (fruit and vegetables) (all P<0.001). The percentage of subjects classified in the same quartile by both questionnaires ranged from 56% (omega-3 fatty acids) to 78% (fruit and vegetables). The percentage of subjects classified in the same or adjacent quartile ranged from 91% to 99%. The intraclass correlation coefficients, which assessed reproducibility, ranged from 0.61 (fruit and vegetables) to 0.88 (saturated fatty acids) (P<0.001).

CONCLUSIONS

This new version of the short dietary questionnaire shows good reproducibility and correlations with the original version; use and reliability are improved, which makes CDQ-2 a valuable tool for cardiovascular prevention.

Evaluation of short food-frequency questionnaires to assess the dietary pattern associated with atherosclerotic cardiovascular diseases

OBJECTIVES

Cardiovascular diseases are strongly related to dietary habits. Diet can be assessed using dedicated questionnaires that can be self-completed by subjects but with the risk of errors.

AIM

To compare the completion error rate of two questionnaires designed to assess dietary pattern linked to cardiovascular diseases and to study the correlation between the two questionnaires.

MATERIALS AND METHODS

Two questionnaires were used to assess dietary patterns of students: the 14-item Food-Frequency-Questionnaire (FFQ) that was validated against biomarkers, and the Cardiovascular-Dietary-Questionnaire 2 (CDQ2), which is a 19-item-FFQ derived from the previous 14-item FFQ. Both questionnaires assessed the intake of various food groups associated with either favourable or unfavourable effects on cardiovascular risk. A global dietary score was calculated for each questionnaire.

RESULTS

FFQ and CDQ2 were completed by 150 sport degree students. In the case of FFQ, 111 questionnaires out of 150 (74.0%) were incomplete compared to only 1 CDQ2 out of 150 (0.7%) (P<0.001). The correlation coefficient between the overall CDQ2 score and the FFQ dietary score was 0.53 (P<0.01).

CONCLUSION

The self-completion of CDQ2 compared to FFQ was associated with far less errors. There was a significant correlation between CDQ2 and FFQ. Preference should be given to CDQ2 in clinical practice and in studies where dietary pattern are evaluated without any interviewer.

[Good reproducibility of a 14-item food frequency questionnaire for cardiovascular prevention in students]

Diet is a modifiable risk factor of atherosclerosis. A 14-item food frequency questionnaire (FFQ) has been developed. The reproducibility of this FFQ is unknown in a student population whereas its use could be of interest.

METHODS

This FFQ allows calculating different scores for different food groups involved in cardiovascular disease. The vascular dietary score (VDS) can be calculated. The VSD ranges from -17 to +19. The higher the VSD, the better diet. Reproducibility was assessed in sports faculty students using mean tests comparing measurement 1 and 2 (minimum time interval ≥ 7 days) and intra-class correlation (ICC) tests.

RESULTS

Thirty students (50% men) were included in a French Sports Faculty. Time between two FFQ assessments was 19 ± 9 days. Mean VSD was 0.50 ± 3.70 for the first assessment and 0.30 ± 3.14 for the second one (non significant). Any score for each food group was statistically significant between the first and the second measurement. ICC of VSD was 0.68 [95% confidence interval: 0.43-0.83].

CONCLUSION

This FFQ that assesses a risky vascular diet has good reproducibility. This tool could be useful for large studies involving students.

Dietary pattern, a modifiable risk factor that can be easily assessed for atherosclerosis vascular disease prevention in clinical practice

OBJECTIVES

Nutrition is one of the modifiable risk factors of atherosclerosis vascular diseases (VD). We aimed to (i) evaluate the dietary patterns associated with VD in clinical practice using a validated FFQ; and (ii) determine potential independent sociodemographic and behavioural factors that are involved in such dietary patterns. The FFQ determined the vascular dietary score (VDS; ranges from -17 to 19) for each subject.

DESIGN

Cross-sectional study.

SETTING

A general practitioner's (GP) office in 2009.

SUBJECTS

A total of 250 French subjects (18-84 years old).

RESULTS

A total of 21 % had a favourable vascular diet (VDS ≥ 8), 79 % needed to improve their diet (VDS < 8) and 21 % had a risky vascular diet (VDS ≤ -1). A step-by-step multivariate linear regression analysis with stepwise selection was performed using the VDS as a dependent variable. Significant variables were: age (β = 0·495, P < 0·0001), men (β = -0·282, P < 0·0001), 'sport ≥ 1 h/week' (β = 0·253, P = 0·001), 'walking 20 min/d' (β = 0·161, P = 0·012), 'former smoker' (β = 0·118, P = 0·029), previous nutritional advice (β = 0·105, P = 0·049), 'alcohol ≥ 20 g/d' (β = -0·216, P < 0·0001) and 'primary school' (β = -0·156, P = 0·010). The R2 coefficient of this model was 0·347 (P < 0·0001). In all, 88·7 % of the subjects found the evaluation very interesting and 89·6 % believed that the GP should perform it.

CONCLUSIONS

Simple dietary assessment for VD prevention can be easily performed in clinical practice to allow physicians to give objective and rapid advice for each patient. Age, educational status, alcohol consumption, gender and physical activity are associated with the VDS. Compliance with such evaluation was found to be very high, which should encourage larger dietary screening in the population in order to reduce the impact of VD.